$
>
Unemployment and Health International and Interdisciplinary Perspectives Edited by Thomas Kieselbach, Anthony H. Winefield, Carolyn Boyd and Sarah Anderson
+ %
First published in 2006 from a completed manuscript presented to Australian Academic Press 32 Jeays Street Bowen Hills Qld 4006 Australia www.australianacademicpress.com.au © 2006. Copyright for each contribution in the book rests with the listed authors. All responsibility for editorial matter rests with the authors. Any views or opinions expressed are therefore not necessarily those of Australian Academic Press. Reproduction and communication for educational purposes The Australian Copyright Act 1968 (the Act) allows a maximum of one chapter or 10% of the pages of this work, whichever is the greater, to be reproduced and/or communicated by any educational institution for its educational purposes provided that the educational institution (or the body that administers it) has given a remuneration notice to Copyright Agency Limited (CAL) under the Act. For details of the CAL licence for educational institutions contact: Copyright Agency Limited Level 19, 157 Liverpool Street Sydney NSW 2000 Australia Telephone: (02) 9394 7600 Facsimile: (02) 9394 7601 E-mail:
[email protected] Reproduction and communication for other purposes Except as permitted under the Act (for example, a fair dealing for the purposes of study, research, criticism or review) no part of this book may be reproduced, stored in a retrieval system, communicated or transmitted in any form or by any means without prior written permission. All inquiries should be made to the publisher at the address above. National Library of Australia Cataloguing-in-Publication data: Unemployment and health: international and interdisciplinary perspectives. eBook ISBN 9781875378616 1. Unemployment — Health aspects. 2. Unemployment — Psychological aspects. 3. Unemployment — Social aspects. I. Kieselbach, Thomas. II. Winefield, Anthony H. (Anthony Harold), 1937–. 331.137 Cover design by Andrea Rinarelli of Australian Academic Press, Brisbane. Editing and typesetting by Australian Academic Press, Brisbane.
CONTENTS
Foreword
vii
Introduction
ix
SECTION
1
Health Effects of Unemployment Chapter 1 Explanations for Deteriorating Wellbeing in Unemployed People: Specific Unemployment Theories and Beyond
1
Peter A. Creed and Dee Bartrum
Chapter 2 Insecurity, the Restructuring of Unemployment and Mental Health
21
David Fryer
Chapter 3 Justice Concerns and Mental Health During Unemployment
35
Claudia Dalbert
Chapter 4 Quantitative Reviews in Psychological Unemployment Research: An Overview
51
Karsten Ingmar Paul and Klaus Moser
Chapter 5 Unemployment, Secure Employment and Insecure Employment: Differences in Self-Reported Ill Health
61
Bengt Starrin and Staffan Janson
Chapter 6 The Finances–Shame Model and the Relation Between Unemployment and Health
75
Bengt Starrin and Leif R. Jönsson
Chapter 7 Retrenchment and Health Parameters: A Short Report Dimity Pond, Elizabeth Harris, Parker Magin, Amber Sutton, Vanessa Traynor, Kate D’Este and Susan Goode
99
CONTENTS (CONTINUED)
SECTION
1
(CONTINUED) Chapter 8 Suicidal Ideation in the Long-Term Unemployed: A Five-Year Longitudinal Study
109
Bjørgulf Claussen
Chapter 9 Quality of Life of the Employed and Unemployed in Turkey: A Comparative Field Study
119
Yucel Demiral, Alp Ergör, Belgin Unal and Semih Semin
Chapter 10 Health and Lifestyle of Reemployed and Unemployed People Following the Japanese Corporate Reorganisation Law
127
Tatsuya Ishitake and Tsunetaka Matoba
Chapter 11 Do the Health Consequences of Unemployment Differ For Young Men and Women?
135
Anne Hammarström and Urban Janlert
Chapter 12 Mass Lay-Offs and Tolerance for Mental Illness: Racial Differences in the Economy’s Effect on Coerced Treatment
143
Eric Kessell and Ralph Catalano
Chapter 13 Demographic, Occupational and Employer-Related Determinants of Long-Term Unemployment Among Danish Employees
157
Thomas Lund and Merete Labriola
Chapter 14 Organisational Citizenship Behaviours in Relation to Job Status, Job Insecurity, Organisation Commitment and Identification, Job Satisfaction and Work Values N.T. Feather and Katrin A. Rauter
167
CONTENTS (CONTINUED)
SECTION
2
Interventions to Limit the Health Effects of Unemployment: Activation Policies and Empowerment Chapter 15 Active Labour Market Programs for Young Long-Term Unemployed: Psychological Impact of Participation in a Recent Program
181
Anthony H. Winefield and Edgar Carson
Chapter 16 Explaining the Negative Relationship Between Length of Unemployment and the Willingness to Undertake a Job Training: A Self-Determination Perspective
199
Maarten Vansteenkiste, Hans De Witte and Willy Lens
Chapter 17 The Role of Limited Duration Contracts in Labour Market Transition
219
Ola Bergström
Chapter 18 Youth Unemployment and the Risk of Social Exclusion in Six European Countries
233
Thomas Kieselbach
Chapter 19 Can Volunteering Be a Moderator of the Detrimental Effects of Engagement in (Un)Employment?
259
Jacques C. Metzer
Chapter 20 Restructuring and Outplacement in the Netherlands
267
Ellen Heuven, Arnold Bakker and Wilmar Schaufeli
Chapter 21 Unemployment and Activation Policy: The Finnish Experience
285
Simo Mannila
About the Contributors
303
Foreword There is ample circumstantial evidence indicating that unemployment, and underemployment, can cause or contribute to physical, mental and social morbidity. According to the United Nations Convention on Human Rights, the member states ‘recognise the right to work, which includes the right of everyone to the opportunity to gain his living by work which he freely chooses or accepts, and will take appropriate steps to safeguard this right’. In the same vein, the European Council concludes that ‘regaining full employment not only involves focusing on more jobs, but also on better jobs. Increased efforts should be made to promote a good working environment for all, including equal opportunities for the disabled, gender equality, good and flexible work organisation permitting better reconciliation of working and personal life, lifelong learning, health and safety at work, employee involvement and diversity in working life’. And, according to Article 152 of the Treaty of Amsterdam, ‘a high level of human health protection shall be ensured in the definition and implementation of all Community policies and activities’. Briefly, then, the insights exist. But they are not implemented. Given the present world rates of unemployment, underemployment and overemployment, there is an urgent need for integrative approaches to: • minimise unemployment and underemployment • minimise overemployment • promote ‘the good job’ • humanise workforce restructuring. By bringing together researchers from all over the world and from a range of different disciplines, this book makes an important contribution to our understanding of the complex range of issues involved in combating the problems caused by unemployment and underemployment.
Lennart Levi Karolinska Institute, Sweden
vii
viii
Introduction The objective of this book is to present important new findings from leading researchers who have examined the health impact of unemployment and underemployment on the individual and the community. Importantly, this book brings together research findings from Europe, Australia, Asia and the United States, and across the fields of psychology, medicine, economics, sociology, occupational health and organisational development. There is growing interest in employment research, due to the recognition that unemployment and unstable employment can result in negative health effects. In a world work environment where there is an increase in temporary and contract work (so-called ‘flexible [precarious] employment’), where working part-time is more common, and where people may be expected to retrain for new careers several times over, workers are increasingly subjected to new work-related pressures. These changes in the work environment bring financial and social costs to both the individual and society, and research is needed to measure these costs and identify methods of reducing them. This book addresses this growing interest in employment and unemployment research, by compiling papers based on presentations by leading researchers at the Second International Expert Conference organised by the Scientific Committee Unemployment and Health of the International Commission on Occupational Health (ICOH) on ‘Occupational Transitions: Unemployment, Underemployment and Health’, held in Adelaide, Australia in December, 2001. The activities of this body bring together researchers working specifically on occupational health issues related to employment with those focusing on the health effects of being out of work in a society centred on paid employment. Thus the ICOH Scientific Committee tries to bridge these two areas that are traditionally separated and to reduce the gap between unemployment research and occupational health. The Adelaide conference was the second international conference. The first one, held in Paris in 1998, focused on interventions with unemployed people. Additional contributions have been invited from other leading researchers who were unable to attend the conference. The main aim of the book is to outline current theory and findings about the factors responsible for the generally detrimental health effects of being out of work or in unstable employment. In this context, the mechanisms underlying the ix
Unemployment and Health ■ Introduction
benefits of secure employment are also discussed. Second, the book explores the interventions that may help to limit the negative health effects of employment status, including the implementation of corporate and government policies, and the empowerment of employees. Finally, the book examines corporate responsibilities in regard to employment and the dismissal process. A range of issues related to these themes is examined in the book, including the relationship between unemployment and poor mental health, the relative impact of unemployment on men and women, and the nature and reasons for the impact of unemployment and insecure employment on health and wellbeing. The book also covers retraining strategies for people who have been retrenched, public health strategies for managing the health of the unemployed, and the constructive role that industry and business can play in managing retrenchment and reintegration strategies for the longterm unemployed. An outstanding feature of this book is that it provides a rare insight into some of the latest, most comprehensive research on employment and underemployment, by drawing together the work of leading investigators from around the world and across various disciplines. Therefore, this book provides a compilation of the different perspectives on employment research, and may be regarded as a valuable core text for those studying and working in the area of employment and unemployment from a range of subject areas.
Thomas Kieselbach University of Bremen, Federal Republic of Germany
Anthony H. Winefield University of South Australia, Australia
Editors
x
SECTION
Health Effects of Unemployment
1
Chapter
1
Explanations for Deteriorating Wellbeing in Unemployed People: Specific Unemployment Theories and Beyond Peter A. Creed and Dee Bartrum
There is consistent evidence that unemployment is detrimental to the wellbeing of the individuals involved (for a recent review, see Murphy & Athanasou, 1999). What is less clear is the explanation for the differences in wellbeing found between those in employment and those who are unemployed. This chapter will review a number of theories that have contributed to our understanding of the causes of poor wellbeing in unemployed people. Our initial focus will be on theories that have been specifically formulated to account for this decline in unemployed individuals’ wellbeing. We will then examine theoretical approaches from the wider organisational stress literature and make suggestions as to how these approaches might be profitably applied to future research within the unemployment domain.
Moderator Variables Before presenting the various theoretical models, it is important to highlight two issues. The first is that a range of individual and situational moderator variables have been identified that influence the individual’s response to unemployment. These moderator variables (third variables that influence the relationship between two other variables; Baron & Kenny, 1986) include age, gender, length of unemployment, ethnic and racial origins, family unemployment, local levels of unemployment, social class, attribution of cause of job loss, prevailing community views of unemployment, the nature of the welfare system, personality variables, values and beliefs (for reviews see Feather, 1990; Fryer & Payne, 1986; Winefield, 1995). This long list, which is by no means exhaustive, is also likely to contain overlapping variables, which may distort the impact of any individual variable. 1
Unemployment and Health ■ Health Effects
Developing a taxonomy of the more salient moderator variables would allow these aspects to be included in theoretical models of wellbeing in the unemployed, and allow for a clearer focus on the dynamic relationship between unemployment and wellbeing.
Drift Versus Social Hypothesis The second issue to be addressed is the drift versus social causation hypothesis. While all theoretical models make the assumption that there is a relationship between being unemployed and deteriorating wellbeing, some authors have observed that individuals may be predisposed to become or remain unemployed because of existing personal characteristics. The competing hypotheses are known as ‘drift’ (poor psychological wellbeing leads one to be more loosely connected to the labour market), and ‘social causation’ (being loosely connected to the labour market leads one to develop poor psychological health; Dooley, Catalano, & Hough, 1992). While much of the evidence from longitudinal studies has indicated a predominant role for social causation (for a recent review, see Winefield, 1995), this issue has not been resolved (Hammarström & Janlert, 1997). While it seems probable that the effects of unemployment and poverty contribute significantly to the decline in wellbeing so commonly observed in the unemployed, it is likely that individuals with preexisting mental and/or physical health problems are more at risk of not entering or reentering the workforce, or losing a job if they have one. It is well accepted, for example, that individuals with lower levels of education and fewer skills are more at risk of unemployment, more likely to lose the work they have and more susceptible to intermittent periods in the labour force (Machin, 1989). Chronically unemployed individuals have been found to differ from the general population in general and reading ability (Creed & Poulton, 1999). General ability and literacy levels have been found to predispose unemployed youth to longer periods of unemployment (Creed, 1999), and some studies have found that unemployed school leavers have poorer levels of wellbeing while still at school (e.g., Creed, Muller, & Patton, 2003), although this latter outcome is not universally found (Schaufeli, 1997). Any theoretical account, to be helpful in explaining deteriorating wellbeing in the unemployed, needs to be able to incorporate these person variables.
The Latent Deprivation Model To date, the most influential account for the deterioration in wellbeing experienced by unemployed people is provided by Jahoda’s (1981, 1982, 1987) latent deprivation model. Jahoda argued that employment, as a social institution, provides both manifest (associated with income) and latent (associated with meeting psychological needs) benefits to the individual. She argued that people primarily work to obtain manifest benefits, but while employed also benefit from having a range of human needs met, 2
Unemployment Theories and Beyond
in particular the five basic human needs for time structure, social contact, sharing of common goals, status and activity. While Jahoda indicated that other social institutions, such as the family and leisure, could provide these latent benefits to some extent, there are no contemporary institutions that do so in combination with the important manifest function of earning one’s living. Being excluded from the institution of employment, that is, being unemployed, leads to a deprivation of both the manifest and latent benefits, but it is the loss of the latent benefits that impacts negatively on psychological wellbeing. Researchers have examined the latent benefits of employment individually and reported their effect on psychological wellbeing. Jahoda (1982) argued that the most important of the latent benefits was time structure, and that the loss of this benefit was ‘experienced as a heavy psychological burden’ (p. 23). Several authors have reported that unemployed people have less structured and purposeful time-use than the employed (Jackson, 1999; Wanberg, Griffiths, & Gavin, 1997). Others have linked lower levels of structured and purposeful time-use to lower levels of self-esteem, more depression and higher levels of psychological distress (Bond & Feather, 1988; Evans & Haworth, 1991; Hepworth, 1980; Rowley & Feather, 1987; Ullah, 1990). In a large-scale longitudinal youth study, Winefield, Tiggemann and Winefield (1992) concluded that time structure played a buffering role in mediating the negative effects of unemployment. In relation to activity, unemployed people have been found to have lower levels than general (Underlid, 1996) or employed samples (Waters & Moore, 2002). Activity has also been associated with wellbeing in the unemployed. Haworth and Ducker (1991) found unemployed people who spent their time meaningfully (e.g., by doing household chores, socialising or in active leisure, rather than idling or by passive leisure) to have better psychological wellbeing. Consistent with this, Winefield et al. (1992) found that structured or socially gregarious activity was better for psychological wellbeing in unemployed youth. Evans and Haworth (1991) divided their sample into two groups based on their level of activity. Those with higher levels of activity had more access to the other latent benefits. These authors concluded that activity played an important mediating influence in gaining access to the other latent benefits. Social contact has been shown to have a positive effect on psychological wellbeing (Haworth & Ducker, 1991; Henwood & Miles, 1987) and depression (Bolton & Oatley, 1987). Kilpatrick and Trew (1985) found that an unemployed sample with high levels of social contact had superior wellbeing compared to a sample of unemployed with few social contacts. Hammer (1993) found that social support from a close social network of family and friends moderated some of the negative effects of unemployment on wellbeing. This author also found that social isolation was linked to low self-esteem. Unemployed people have also been found to be involved 3
Unemployment and Health ■ Health Effects
in fewer social activities than those who are not unemployed (Underlid, 1996), and to have less social support from close relations and authority figures when compared with an employed group (Jackson, 1999). There has been little empirical work examining the relationship between collective purpose and wellbeing. Jahoda (1982) argued that, outside of the nuclear family, it was employment that allowed people to feel a part of society, and that the loss of this benefit was related to lower wellbeing. For unemployed people, Haworth and Paterson (1995) found that collective purpose was significantly related to wellbeing if gained either through work or leisure activities. Also, feelings of having a collective purpose or common goals have been found to be associated with wellbeing in unemployed people (Evans & Haworth, 1991; Haworth & Ducker, 1991). Jahoda (1982) argued that one’s status was often defined by one’s job. Donovan and Oddy (1982) found that older workers who became unemployed lost their work-related identity, while school-leavers entering the workforce were faced with the prospect of establishing one. Status has also been shown to be related to psychological wellbeing. Evans and Haworth (1991) found that employment status was significantly correlated with wellbeing independent of activity, while Haworth and Paterson (1995) found leisure-time status to have a strong relationship with psychological wellbeing. Creed and Macintyre (2001) found status to be the most important contributor to wellbeing, followed by time structure and collective purpose. Rather than focus on individual latent benefits, some studies have attempted to test Jahoda’s hypothesis directly. These attempts have used scales developed specifically to measure all five latent benefits of employment. For example, Brief, Konovsky, Goodwin and Link (1995) devised an instrument to assess global deprivation of the latent functions of employment, which they labelled the Experiential Deprivation Scale. However, by far the most widely used scale for measuring the latent benefits of employment has been the Access to Categories of Experience (ACE) Scale. This scale was originally devised by Miles (1983) and later revised and shortened by Evans (1986), with input from Miles. The revised ACE Scale has been widely used in studies that have examined associations between the latent benefits of employment and psychological wellbeing with samples of unemployed youth (Evans & Banks, 1992; Miles & Howard, 1984 [original scale]), unemployed adults (Evans, 1986; Henwood & Miles, 1987; Miles, 1983 [original scale]; Waters & Moore, 2002), employed youth (Creed et al., 2003; Evans & Haworth, 1991; Haworth & Ducker, 1991; Miles & Howard, 1984), underemployed adults (Creed & Machin, 2002), employed adults (Haworth & Hill, 1992; Haworth & Paterson, 1995; Waters & Moore, 2002) and full-time students (Creed et al., 2003; Miles & Howard, 1984). These studies have variously examined the wellbeing correlates of the ACE subscales and the ACE total score. 4
Unemployment Theories and Beyond
However, a recent study (Creed & Machin, 2003) has drawn into question the usefulness of the ACE Scale. This study found a better fit for a four-factor structure of the ACE, rather than a five-factor one as intended by the original developers. The recommendation from this study is that the ACE Scale should only be considered as a global measure, rather than a measure that can be interpreted along the five latent benefit dimensions. However, even when used as a global measure, the ACE Scale cannot be considered to represent all five latent benefits of employment, as no factor for time structure emerged. The findings from this study cast doubt on much of the previous research that has used the ACE to measure the latent benefits, both where the ACE subscales have been examined, and where the ACE total scores have been used. The lack of a sound psychometric tool to measure Jahoda’s five latent functions of employment has meant that it has been quite difficult to test Jahoda’s hypotheses regarding the latent benefits. Muller, Creed, Waters and Machin (2005) have presented preliminary data for a newly developed scale (the Latent and Manifest Benefits [LAMB] Scale) that they propose will measure the latent and manifest benefits of employment. However, the full psychometric properties of this scale are yet to be published. One study has directly tested Jahoda’s hypothesis by using a range of individual scales taken from the general literature to measure the five latent functions (Creed & Macintyre, 2001). This study showed that all measures of the latent benefits, of time structure (r = –.47), activity (–.33), social contact (–.34), collective purpose (–.39) and status (–.56) were moderately correlated with psychological distress (as measured by the 12-item version of the General Health Questionnaire; Goldberg, 1972). This was also the case for the manifest benefit (operationalised as financial strain; r = .61). Further, when psychological distress was regressed on the latent and manifest benefits, these variables were able to account for 52% of the variance. Financial strain was identified as the most important predictor of psychological distress (uniquely contributing 17%), followed by the latent benefits of status (8%), time structure (5%) and collective purpose (2%). This study found no differences between unemployed males and unemployed females in levels of latent benefits, manifest benefits or psychological distress (although this latter finding needs to be interpreted with some caution as the males in the sample were significantly older than the females). The evidence from this study is that all latent benefits were associated with wellbeing, but that only three of the five were important individual predictors of psychological health. Psychological wellbeing was most strongly correlated with financial strain, and financial strain emerged as the most important predictor of wellbeing. The findings from these studies that have used psychometric tools to measure the latent benefits, either individually or in concert, generally support the notions that the unemployed have less access to the latent benefits than their employed counterparts, and that the 5
Unemployment and Health ■ Health Effects
level of access to the latent benefits is related to wellbeing, with the more access to the latent benefits the better the wellbeing. Jahoda has suggested that any job is preferable to the state of joblessness, as all jobs inevitably provide access to the latent benefits of employment. However, several studies have questioned this particular proposition by contrasting so-called ‘good’ and ‘bad’ jobs and concluding that individuals in bad jobs experience many of the negative effects that the unemployed do (O’Brien & Feather, 1990; Winefield, Tiggemann, & Goldney, 1988; Winefield et al., 1992). Other studies have found that the latent benefits can be accessed from sources other than employment (Creed et al., 2003; Evans & Banks, 1992; Miles & Howard, 1984), and that different categories of unemployed (e.g., unemployed with no paid employment vs. unemployed with some paid employment) differ on levels of latent benefits (Creed & Machin, 2002). Issues that remain unresolved here include the extent to which the latent benefits can be met through other institutional structures, the types of latent benefits that can be met through different structures, and the implications for the development of interventions for unemployed people. While Jahoda argued that all latent benefits were important, it was her opinion that time structure was the most important. The empirical evidence to date does not support this. For example, as just described, Creed and Macintyre (2001) found status to be the best individual latent benefit predictor of psychological distress. These same authors did not find that all latent benefits contributed equally to predicting wellbeing. If all five latent benefits do not make similar contributions then the question arises: Which ones are important, for which groups, and under what conditions? Further, the unemployed have not been found to differ from the employed on all latent benefits (e.g., Evans, 1986). Does this imply that a more parsimonious model is possible, or is it the case, as Warr (1999) suggests, based on studies of people in employment, that a more comprehensive range of latent benefits needs to be considered? Warr outlined 10 aspects of the environment, many of which overlap with Jahoda’s latent benefits, that he proposed affect wellbeing in the workplace, and would be absent when an individual lost employment. These environmental features are opportunity for skill use, control, externally generated goals, variety, environmental clarity, availability of money, physical security, opportunity for interpersonal contact, valued social position and supportive supervision. Warr also suggests a nonlinear relationship between these environmental features and wellbeing, a hypothesis that has not been tested with unemployed people using these environmental features or the latent benefits. The studies to date have largely examined the latent benefits (either individually or globally) in isolation, and have assumed their relationship with wellbeing to be linear (although see Warr, 1987). No studies have 6
Unemployment Theories and Beyond
examined the interaction effects among the latent benefits, or possibly more importantly, between the latent and manifest benefits of employment. Further, no studies have tested for mediation effects. For example, it is possible that the loss of manifest benefits has direct effects on wellbeing as well as indirect effects by influencing access to the latent benefits. It is also possible that there will be direct as well as indirect effects among the latent variables. For example, it is likely that having reduced status might also affect levels of activity. Lastly here, there is no evidence indicating how unemployed people react to the loss of latent benefits over time. There is some evidence that when people lose their jobs they respond to the loss of manifest benefits by immediately reducing spending on luxury items, but as economic deprivation persists they also have to reduce spending on essential items (Waters & Moore, 2002). There has been no analysis as to how perceptions of access to the latent benefits change over time and how these changes impact on wellbeing. An important omission in the studies to date has been the failure to test for differential effects for men and women on the latent benefits. Jahoda’s model grew out of empirical work that was conducted when men were the primary income earners, and women were less likely to be members of the labour market. This is no longer the case. Several authors have suggested that the experience of unemployment is more stressful for males (e.g., Lahelma, 1992; Muller, Hicks, & Winocur, 1993; Warr, Jackson, & Banks, 1982), although the reverse has also been found (Warr, Banks, & Ullah, 1985; Warr & Payne, 1983), and other reports have found no differences (Feather & O’Brien, 1986a, 1986b; Winefield & Tiggemann, 1985). The implication is that gender may be interacting with other personal or situational variables to influence outcomes. Lastly, the generalisation that all paid employment is beneficial and that the latent benefits will prove beneficial to all individuals fails to acknowledge that individuals perceive situational information biased by individual differences. The latent deprivation model largely ignores the notion that people have different reactions to the same situations, and that they evaluate environmental conditions based on their unique expectations, values, previous experiences and personality (DeNeve & Cooper, 1998; Ezzy, 1993). Several authors (Creed & Evans, 2002; Creed, Muller, & Machin, 2001; Payne, 1988) have argued that to provide a more coherent explanation of wellbeing the model must give consideration to the individual differences that people bring to the situation, in the same way that other models attempt to (e.g., see Diener, Suh, Lucas, & Smith, 1999). The individual disposition most studied in this context is the personality trait of neuroticism. Levels of neuroticism have been found to moderate the experience of unemployment, with wellbeing levels being influenced by the way an individual experiences anxiety. High levels of neuroticism have been associated with elevated levels of psychological distress (Creed, Machin, & 7
Unemployment and Health ■ Health Effects
Hicks, 1996; Payne, 1988; Schaufeli, 1992) and lowered levels of selfesteem (Creed et al., 1996). Payne (1988), for example, found neuroticism to be strongly correlated with psychological wellbeing and to account for a significant amount of the variance in the prediction of wellbeing in a group of unemployed males. It would be expected, for example, that an unemployed person with elevated neuroticism scores would view his or her unemployment situation as being more hostile, frightening and demanding than someone else with lower levels of trait neuroticism. Personality has been identified as a strong and consistent predictor of wellbeing across a wide range of populations, with neuroticism particularly seen to play an important role (Diener et al., 1999). Despite this evidence, personality variables have typically not been examined when investigating the negative effects associated with unemployment, and temperament does not feature strongly in Jahoda’s or other unemployment-specific explanations for the decline in wellbeing of the unemployed.
The Agency Restriction Model Despite the intuitive appeal of the latent deprivation model it has been criticised by a number of authors (Creed et al., 2001; Fryer, 1992, 1995; Fryer & Payne, 1986; Haworth, 1997). These authors argue that too much emphasis has been placed on environmental and social factors at the expense of internal regulatory processes. Fryer has proposed an agency restriction model as an alternative to the latent deprivation approach. Fryer considers individuals to be ‘socially embedded agents who are actively striving for purposeful determination, attempting to make sense of, initiate, influence and cope with events in line with personal values, goals, expectations of the future in a context of cultural norm, traditions and past experience’ (Fryer, 1995, p. 270). He suggests that the main negative consequence of unemployment is not the loss of the latent benefits, but rather the loss of the manifest benefits (loss of income). Fryer contends that ‘… unemployment generally results in psychologically corrosive experienced poverty’ (p. 270), and it is this experience of poverty that severs the individual from a meaningful future and leads to a reduction in psychological health. Fryer acknowledges the role of the latent benefits (Fryer & Payne, 1984), but considers these insufficient to fully explain the deterioration in wellbeing experienced by the unemployed. Fryer and McKenna (1987) have provided support for this model. These authors compared two small groups of unemployed men who were laid off from their factory jobs. One group had been made temporarily redundant, while the other had been permanently laid off. Contrary to the predictions of the latent deprivation model, both groups were not equally deprived of the latent benefits, and were not equally distressed. Those who were temporarily laid off had arranged active and productive lives and appeared to be psychologically healthy, which was not the case for the 8
Unemployment Theories and Beyond
indefinitely retrenched group. The authors concluded that the temporarily laid-off men were able to obtain the latent benefits outside of the work environment, were more optimistic about the future and would eventually return to work for the manifest, rather than the latent, benefits. Other authors have also demonstrated that financial hardship plays a substantial role in the lives of unemployed people. Jackson (1999) found that unemployed individuals reported more financial stress than employed or student samples. Kokko and Pulkkinen (1997), in a study comparing unemployed people with employed, found that the unemployed experienced more financial strain and were more distressed. Dew, Bromet and Penkower (1992) found an association between financial difficulties and depression in a sample of unemployed women, and a similar link has been reported for unemployed people in general (Vinokur, Price, & Caplan, 1991). Using a large-scale Irish national database, Whelan (1992) examined both subjective experiences of financial strain and objective material deprivation in a large cohort of unemployed individuals, and concluded that poverty, as construed by these two broad variables, played a substantial role in mediating the effects of unemployment for both the individual involved and the person’s family. Creed and Macintyre (2001) found stronger correlations between financial strain and psychological distress than between any of the latent benefits and distress, and found financial strain a much more important predictor of wellbeing than the latent benefits. However, despite evidence demonstrating the important effects of poverty on the unemployed, using both direct and proxy measurements (e.g., financial strain), few studies have directly tested the personal agency theory with unemployed samples. The roles played by the important variables of goals, future orientation and planning, for example, which have been stressed in this theory, have not been examined individually. Nor have these variables been tested in the context of other more general theories, such as social cognitive theory (cf. Bandura, 2000), which make specific claims regarding the explication and measurement of personal agency (for one example of this see Millman & Latham, 2001).
The Financial Distress/Shame Model It has recently been suggested (Rantakeisa, Starrin, & Hagquist, 1999) that the level of financial distress experienced by unemployed people, in combination with the shame associated with the joblessness, can account for the negative psychological and health effects associated with unemployment. People are seen to strive to interact and generate affinity with others. They accomplish this largely by meeting accepted common standards and ideals, of which an important social convention is to be productively employed. Paid employment has a long history of being an honourable activity with both financial and moral dimensions. It allows individuals to meet their financial obligations, as well as to meet societal expectations and standards. When 9
Unemployment and Health ■ Health Effects
these goals cannot be met, for example by becoming unemployed, the individual feels humiliated, degraded and shamed in the eyes of others. In this way, shame is viewed as the product of strained social interactions (Scheff, 1990). Unemployment then not only produces financial deprivation and restricts people from engaging in a socially acceptable lifestyle, it also deprives them of the important social role of worker. Relying on support from others, whether it is from social security or family, does not have the same social value as living off one’s own effort and earnings. Early reports in the descriptive literature associated shame with unemployment (e.g., Bakke, 1933; Eisenberg & Lazarsfeld, 1938), while more recent empirical evidence has identified a link between shame and psychological distress. Eales (1989), using a small-scale qualitative methodology, found that 25% of the unemployed males in the sample reported shame, and that shame was correlated with financial distress and the presence of negative moods. However, in this study no meaningful association was found between shame and level of social intimacy, which might be expected from the theory. Rantakeisa et al. (1999) reported on a large-scale survey of unemployed young people. These authors found that 42% of their sample reported shaming experiences, and that shaming was associated with gender (higher incidence in males) and length of unemployment (more shaming the longer unemployed). Shaming was also associated with poorer mental wellbeing, physical health, reduced spare-time activities and reduced contact with friends. Shaming and financial distress interaction effects were found for wellbeing, activity and social support, with the most severe wellbeing deficits found in young males with the highest financial strain and the most shaming experiences. In a recent study, Creed and Muller (in press), using a psychometric approach, did find differences in levels of shame between unemployed and employed individuals. However, they found no association between shame and length of unemployment, which is predicted by this model, as increasing periods of unemployment are likely to create greater discrepancies between actual situation and societal expectations. More importantly, this study found only a weak association between shame and wellbeing in the unemployed sample, and identified no role for shame in predicting wellbeing in this sample. The most important variables predicting wellbeing were financial strain, social support and the personality variable of neuroticism, implicating latent benefits and other internal processes. It may be that moderating variables, such as community unemployment rates, cultural values or attribution for unemployment, may help explain these negative results for the shame model, and need to be examined empirically. Shame might also be examined in the context of self-regulatory models, where, for example, it might influence outcome expectations or efficacy. Thus far, we have highlighted theoretical approaches specifically developed to account for psychological wellbeing in the unemployed. Several 10
Unemployment Theories and Beyond
concerns have been raised about the restrictive focus of these theories and the inconsistency of these theories to the models applied within the area of organisational stress or wellbeing. Compared to the organisational theories of stress, unemployment theories have limited focus on person variables (such as optimism, locus of control, perceived control, goals, values and beliefs), temperament (such as neuroticism) and self-regulatory aspects (such as appraisal, emotion, future expectations and coping). One aspect in particular, control, has been identified as an influential factor in the experience of psychological wellbeing in the employed. For example, Warr’s (1987) vitamin model focuses on the importance of perceived control within the work context (opportunity for personal control), as it allows the individual to make adjustments in his or her environment to offset aversive consequences of work stressors. Control has also been recognised as a moderator in the experience of job stress and wellbeing, particularly within Karasek’s (1979) job demands–control theory. The fundamental proposition of Karasek’s model is that, ‘although excessive work demands may clearly be associated with higher levels of psychological strain and psychological health outcomes, the impact of these demands may be offset by the perception that one has control over important aspects of the work environment’ (Cooper, Dewe, & O’Driscoll, 2000, p. 135). Control has been well researched within the organisational stress domain but has yet to be examined widely in the area of unemployment (but see Morrison, O’Connor, Morrison, & Hill, 2001).
The Demands–Control Model The central prediction of the demands–control model is that the most adverse reactions produced by psychological strain occur when psychological demands are high and decision latitude is low. Psychological demands have been defined within the demands–control model as the psychological stressors present in the work environment, while decision latitude is defined as the potential control that a worker has over his or her tasks and his or her behaviour during the working day. The principal driving force of the demands–control model is the existence of an interactive effect of demand and control, where control moderates the effects of demands on the outcome. Johnson and Hall (1988) have also extended Karasek’s demands–control model to include social support as a moderator of the connection between job demands and control with strain. To bridge the gap between unemployment and organisational stress models, Creed and Bartrum (2005) tested the efficacy of integrating the latent deprivation, financial strain and demands–control model in predicting wellbeing in an unemployed sample. The main addition to the variables from the latent deprivation model (latent benefits and manifest benefits) was that of control. The results indicated that control contributed to variance in wellbeing over and above that contributed by the latent and 11
Unemployment and Health ■ Health Effects
manifest benefits, and that control moderated the effect of one of the latent benefits (activity) and the manifest benefit (financial strain) on psychological wellbeing. Control also mediated the relationship between financial strain, time structure, collective purpose, status and wellbeing. The significant interaction effects indicated that the relationship between activity and wellbeing, and between financial strain and wellbeing, were conditional upon the level of perceived control. Specifically, when activity was low, wellbeing was lower when control was also low; and when financial strain was high, wellbeing was lower again when control was also low. For the mediation effect, the evidence was that the latent and manifest benefits affected wellbeing directly as well as by influencing the level of control perceived. The integration of Jahoda’s latent and manifest functions (demands) with the demands–control model provides a useful framework to investigate the psychological wellbeing of unemployed. The demands–control model provides the opportunity to investigate the contribution of perceived control to the deprivation model and the moderating effects that perceived control has on the demands experienced by the unemployed and their psychological wellbeing. However, while the demands–control model may provide a parsimonious approach to the investigation of psychological wellbeing in the employed and unemployed, it has been criticised by a number of authors (see, e.g., de Jonge & Kompier, 1997; Kasl, 1996; Kristensen, 1995; Warr, 1994) for this simplicity. One of the primary concerns about the demands–control model, mirroring the concerns expressed in this chapter about the specific unemployment theories, is the lack of focus on person variables. Furthermore, the demands–control and unemployment theories do not directly address the self-regulatory mechanisms (including coping) and the affective experiences of the individual. Researchers within the organisational psychology realm are recognising the importance of investigating the emotional experiences of the employed (see Fisher & Ashkanasy, 2000), and their ability to cope. For example, one of the five indicators of mental health associated with job loss within Warr’s (1987) vitamin model is the individual’s capacity to cope (competence). Theories that provide the opportunity to integrate and test the impact of self-regulatory mechanisms, emotional experiences and coping ability on psychological wellbeing include cybernetic and appraisal theories.
Cybernetic and Appraisal Theories Cybernetics concerns the functioning of self-regulating systems (Ashby, 1966; Edwards, 1992; Wiener, 1948). These approaches focus on the discrepancy between the individual’s perceived and desired state, when the individual considers the discrepancy to be important. At the core of cybernetic theories is the negative feedback loop, which acts to minimise discrepancies between environmental characteristics and relevant reference 12
Unemployment Theories and Beyond
criteria. The four key aspects of a cybernetic model are reference goals or standards, an input sensor function (which perceives the environment and sends a signal to the comparator), a comparator (which assesses the perceived environment against a relevant reference criterion) and an output function (which attempts to change the environment to reduce or remove a discrepancy). Several theoretical frameworks within the occupational stress area are viewed as falling within the cybernetic framework. These theories include role stress theory (Kahn, Wolfe, Quinn, Snoeck, & Rosenthal, 1964), stress cycle (McGrath, 1976), facet model (Beehr & Newman, 1978), person–environment fit theory (French, Caplan, & Harrison, 1982), Cummings and Cooper’s stress theory (1979) and Lazarus’ transactional model (Lazarus, 1966; Lazarus & Folkman, 1984). Edwards (1992), who highlighted differences across these various frameworks, proposed a comprehensive, multidimensional cybernetic theory that addressed these inconsistencies. The constructs of this theory include desires, importance (degree to which the discrepancy is viewed as central to the individual’s overall wellbeing), the physical and social environment, personal characteristics, cognitive construction of reality and social information. Edwards (1992) proposed that coping, defined as any attempt to reduce the negative impact of stress on wellbeing, is activated when damage occurs to an individual’s psychological and/or physical wellbeing. Coping strategies are seen to influence stress indirectly by altering those aspects that produce a stressful discrepancy (such as changing perceptions, desires and importance), and to improve wellbeing directly by managing environmental stressors. When discrepancies are experienced, the result is a decline in psychological and/or physical wellbeing. Edwards viewed discrepancies as interrelated, both hierarchically and at the same level, ‘such that changes in one discrepancy may resolve or exacerbate others’ (p. 256). Thus, Edwards conceived of the stress and coping process as a dynamic system comprised of several interrelated feedback loops. These negative feedback loops identify discrepancies that may exist between an individual’s perceptions and desires. When discrepancies are perceived, they result in damage to an individual’s wellbeing, and consequently activate coping. A change in one discrepancy has the ability to influence the magnitude and/or importance of other discrepancies. As an applied example of this theory, Latack, Kinicki and Prussia (1995) recently proposed a process model specifically for coping with job loss. These authors integrated coping theory (Lazarus & Folkman, 1984) with the control theory framework (the founding framework of cybernetic theory; Carver & Scheier, 1982; Edwards, 1992). The key variables included in this model, in addition to those that are common to cybernetic theories (such as reference criterion, comparator, output function and input function), were appraisal, coping goals, coping resources, coping efficacy and coping strategies. However, while Latack et al. (1995) outlined
13
Unemployment and Health ■ Health Effects
the propositions for their process model of coping with job loss, no empirical research supporting the model has yet been provided. Cybernetic theories acknowledge the importance of person variables, environmental variables, the perceptions of the individual and the effect of coping on the individual’s wellbeing. The negative feedback loop recognises the continuing process that occurs for the individual and the complex nature of maintaining a balance between the individual’s desired state and perceived state. Appraisal theories (Roseman & Smith, 2001; Scherer, Schorr, & Johnstone, 2001) build further on cybernetic theories by addressing the emotional reaction and state of the individual. Appraisal theories in their simplest form propose that emotions are elicited by evaluations (appraisals) of events and situations. Appraisal itself is viewed as performing a central role in shaping and organising the individual’s emotional reaction. Appraisal theories take into account the salience of person and environment variables, coping and the contribution that these factors have on the appraisals and emotions experienced by the individual. Lazarus (1991) acknowledged the importance of emotions, thereby adapting the transactional theory of stress and coping to incorporate the emotional experience of the individual. He labelled this approach the cognitive–motivational–relational theory of stress and wellbeing. However, to date very little research has examined the perceptions and emotions of the employed and unemployed within an appraisal theory framework and the impact of these variables on wellbeing.
Summary and Conclusions In this chapter we have outlined the most widely used theoretical approaches specifically developed to account for the deteriorating wellbeing observed when individuals become unemployed. We have attempted to use the empirical evidence accrued to date to highlight the strengths and shortcomings associated with these approaches. Further, we have briefly examined some of the more promising theories from the occupational literature that are likely to assist with more comprehensive explanations for this decline in wellbeing. Central to these occupational theories is the dynamic, appraised relationship transacted between the individual and the environment. This dynamic relationship is influenced by both person and contextual variables, many of which have not been adequately accounted for in the specific unemployment theories. Unemployment theories have contributed greatly to our understanding of the experiences of individuals under stressful conditions. However, these theories can be further developed by integrating the variables and processes that have been identified as salient for the unemployed with the transactional models (such as cybernetic and appraisal theories). Developing improved causal explanations for declining wellbeing in unemployed people is important because of their implications for interventions and policy-making. Training, individual 14
Unemployment Theories and Beyond
counselling and community education programs are all influenced by the explanatory models that are available, and policy-makers rely on these explanations when they make decisions that directly and indirectly affect the lives of unemployed people.
References Ashby, W.R. (1966). An introduction to cybernetics. New York: Wiley. Bakke, E.W. (1933). The unemployed man. London: Nisbet. Bandura, A. (2000). Social cognitive theory: An agentic perspective. Annual Review of Psychology, 52, 1–26. Baron, R., & Kenny, D. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. Beehr, T.A., & Newman, J.E. (1978). Job stress, employee health, and organisational effectiveness: A facet analysis, model and literature review. Personnel Psychology, 31, 665–699. Bolton, W., & Oatley, K. (1987). A longitudinal study of social support and depression in unemployed men. Psychological Medicine, 17, 453–460. Bond, M.J., & Feather, N.T. (1988). Some correlates of structure and purpose in the use of time. Journal of Personality and Social Psychology, 55, 321–329. Brief, A.P., Konovsky, M.A., Goodwin, R., & Link, K. (1995). Inferring the meaning of work from the effects of unemployment. Journal of Applied Social Psychology, 25, 693–711. Carver, C., & Scheier, M.F. (1982). Control theory: A useful conceptual framework for personality — social, clinical and health psychology. Psychological Bulletin, 92, 111–135. Cooper, C.L., Dewe, P.J., & O’Driscoll, M P. (2000). Organisational stress: A review and critique of theory, research and applications. London: Sage. Creed, P.A. (1999). Predisposing factors and consequences of occupational status for long-term unemployed youth: A longitudinal study. Journal of Adolescence, 22, 81–93. Creed, P.A., & Bartrum, D. (2005). Personal control as a mediator and moderator between life strains and psychological wellbeing in the unemployed. Manuscript submitted for publication. Creed, P.A., & Evans, B. (2002). Personality, wellbeing and deprivation theory. Personality and Individual Differences, 33, 1045–1054. Creed, P.A., & Machin, M.A. (2003). Multidimensional properties of the Access to Categories of Experience scale. European Journal of Psychological Assessment, 19, 85–91. Creed, P.A., & Machin, M.A. (2002). Access to the latent benefits of employment for unemployed and underemployed individuals. Psychological Reports, 90, 1208–1210. Creed, P.A., Machin, M.A., & Hicks, R. (1996). Neuroticism and mental health outcomes for long-term unemployed youth attending occupational skills training programs. Personality and Individual Differences, 21, 537–544. Creed, P.A., & Macintyre, S.R. (2001). The relative effects of deprivation of the latent and manifest benefits of employment on the wellbeing of unemployed people. Journal of Occupational Health Psychology, 6, 324–331. Creed, P.A., & Muller, J. (in press). Psychological distress in the labour market: Shame or deprivation? Australian Journal of Psychology.
15
Unemployment and Health ■ Health Effects Creed, P.A., Muller, J., & Machin, M.A. (2001). The role of satisfaction with occupational status, neuroticism, financial strain and categories of experience in predicting mental health in the unemployed. Personality and Individual Differences, 30, 435–447. Creed, P.A., Muller, J., & Patton, W. (2003). Leaving high school: The influence and consequences for psychological well-being and career-related confidence. Journal of Adolescence, 26, 295–311. Creed, P.A. & Poulton, M. (1999). Ability levels, mental health and attitudes-to-work of the chronic unemployed. South Pacific Journal of Psychology, 10, 1–14. Cummings, T.G., & Cooper, C.L. (1979). Cybernetic framework for studying occupational stress. Human Relations, 32, 395–418. de Jonge, J., & Kompier, A.J.M. (1997). A critical examination of the demand– control–support model from a work psychological perspective. International Journal of Stress Management, 4, 235–258. DeNeve, K.M., & Cooper, H. (1998). The happy personality: A meta-analysis of 137 personality traits and subjective well-being. Psychological Bulletin, 124(2), 197–230. Dew, M.A., Bromet, E.J., & Penkower, L. (1992). Mental health effects of job loss in women. Psychological Medicine, 22, 751–764. Diener, E., Suh, E., Lucas, R., & Smith, H. (1999). Subjective well-being: Three decades of progress. Psychological Bulletin, 125, 276–302. Donovan, A., & Oddy, M. (1982). Psychological aspects of unemployment: An investigation into the emotional and social adjustment of school leavers. Journal of Adolescence, 5, 15–30. Dooley, D., Catalano, R., & Hough, R. (1992). Unemployment and alcohol disorder in 1910 and 1990: Drift versus social causation. Journal of Occupational and Organizational Psychology, 65, 277–290. Eales, M.J. (1989). Shame among unemployed men. Social Science and Medicine, 28, 783–789. Edwards, J.R. (1992). A cybernetic theory of stress, coping, and well-being in organisations. Academy of Management Review, 17, 238–274. Eisenberg, P., & Lazarsfeld, P.F. (1938). The psychological effects of unemployment. Psychological Bulletin, 35, 358–390. Evans, S.T. (1986). Variations in activity and psychological well-being in employed young adults. Unpublished doctoral dissertation, University of Manchester, UK. Evans, S.T., & Banks, M.H. (1992). Latent functions of employment: Variations according to employment status and labor market. In C.H.A. Varhaar & L.G. Jansma (Eds.), On the mysteries of unemployment (pp. 281–295). Dordrecht, the Netherlands: Kluwer Academic. Evans, S.T., & Haworth, J.T. (1991). Variations in personal activity, access to categories of experience and psychological well-being in unemployed young adults. Leisure Studies, 10, 249–264. Ezzy, D. (1993). Unemployment and mental health: A critical review. Social Science Medicine, 37, 41–52. Feather, N.T., & O’Brien, G.E. (1986a). A longitudinal analysis of the effects of different patterns of employment and unemployment on school-leavers. British Journal of Psychology, 77, 459–479. Feather, N.T., & O’Brien, G.E. (1986b). A longitudinal study of the effects of employment and unemployment on school leavers. Journal of Occupational Psychology, 59, 121–144. Feather, N.T. (1990). The psychological impact of unemployment. New York: Springer–Verlag. 16
Unemployment Theories and Beyond Fisher, C.D., & Ashkanasy, N.M. (2000). The emerging role of emotions in work life: An introduction. Journal of Organizational Behavior, 21, 123–129. French, J.R.P., Caplan, R.D., & Harrison, R.V. (1982). The mechanisms of job stress and strain. London: Wiley. Fryer, D.M. (1992). Psychological or material deprivation: Why does unemployment have mental health consequences? In E. McLaughlin (Ed.), Understanding unemployment (pp. 103–125). London: Routledge. Fryer, D.M. (1995). Benefit agency? Labour market disadvantage, deprivation and mental health. The Psychologist, June, 265–272. Fryer, D.M., & McKenna, S.P. (1987). The laying off of hands: Unemployment and the experience of time. In S. Fineman (Ed.). Unemployment: Personal and social consequences (pp. 47–73). London: Tavistock. Fryer, D.M., & Payne, R.L. (1984). Pro-active behaviour in unemployment: Findings and implications. Leisure Studies, 3, 273–295. Fryer, D.M., & Payne, R.L. (1986). Being unemployed: A review of the literature on the psychological experience of unemployment. In C.L. Cooper & I. Robertson (Eds.), International review of industrial and organisational psychology (pp. 235–278). Chichester, UK: John Wiley. Goldberg, D.P. (1972). The detection of psychiatric illness by questionnaire. London: Oxford University Press. Hammarström, A., & Janlert, U. (1997). Nervous and depressive symptoms in a longitudinal study of youth unemployment: Selection or exposure? Journal of Adolescence, 20, 293–305. Hammer, T. (1993). Unemployment and mental health among young people. A longitudinal study. Journal of Adolescence, 16, 407–420. Haworth, J.T. (1997). Work, leisure and well-being. London: Routledge. Haworth, J.T., & Ducker, J. (1991). Psychological wellbeing and access to categories of experience in unemployed young adults. Leisure Studies, 10, 265–274. Haworth, J., & Hill, S. (1992). Work, leisure and psychological well-being in a sample of young adults. Journal of Community and Applied Social Psychology, 2, 147–160. Haworth, J.T., & Paterson, F. (1995). Access to categories of experience and mental health in a sample of managers. Journal of Applied Social Psychology, 25, 712–724. Henwood, F., & Miles, I. (1987). The experience of unemployment and the sexual divisions of labor. In D. Fryer & P.O. Ullah (Eds.), Unemployed people: Social and psychological perspectives (pp. 94–110). Milton Keynes, UK: Open University Press. Hepworth, S.J. (1980). Moderating factors of the psychological impact of unemployment. Journal of Occupational Psychology, 53, 139–145. Jackson, T. (1999). Differences in psychosocial experiences of employed, unemployed, and student samples of young adults. Journal of Psychology, 133, 49–60. Jahoda, M. (1981). Work, employment and unemployment: Values, theories and approaches in social research. American Psychologist, 36, 1983–1991. Jahoda, M. (1982). Employment and unemployment: A social-psychological analysis. London: Cambridge University Press. Jahoda, M. (1987). Unemployed men at work. In D. Fryer & P. Ullah (Eds.), Unemployed people (pp. 1–73). Milton Keynes-Philadelphia, PA: Open University Press. Johnson, J., & Hall, E. (1988). Job strain, workplace social support and cardiovascular disease: A cross sectional study of a random sample of the working population. American Journal of Public Health, 78, 1336–1342. Kahn, R.L., Wolfe, D.M., Quinn, R.P., Snoeck, J.D., & Rosenthal, R.A. (1964). Organisational stress: Studies in role conflict and ambiguity. New York: Wiley.
17
Unemployment and Health ■ Health Effects Karasek, R. (1979). Job demands, job decision latitude and mental strain: Implications for job redesign. Administrative Science Quarterly, 24, 285–308. Kilpatrick, R., & Trew, K. (1985). Life styles and psychological wellbeing among unemployed men in Northern Ireland. Journal of Occupational Psychology, 58, 207–216. Kasl, S.V. (1996). The influence of the work environment on cardiovascular health: A historical, conceptual, and methodological perspective. Journal of Occupational Health Psychology, 1, 42–56. Kokko, K., & Pulkkinen, L. (1997). Economical and psychological well-being of the unemployed. Psykologia, 32, 349–359. Kristensen, T.S. (1995). The demand–control–support model: Methodological challenges for future research. Stress Medicine, 11, 17–26. Lahelma, E. (1992). Unemployment and mental well-being: Elaboration of the relationship. International Journal of Health Sciences, 22, 261–274. Latack, J.C., Kinicki, A.J., & Prussia, G.E. (1995). An integrative process model of coping with job loss. Academy of Management Review, 20, 311–342. Lazarus, R.S. (1966). Psychological stress and the coping process. New York: McGrawHill. Lazarus, R.S. (1991). Emotion and adaptation. New York: Oxford University Press. Lazarus, R.S., & Folkman, S. (1984). Stress, coping and adaptation. New York: Springer–Verlag. Machin, M.A. (1989). The role of general cognitive ability in employment. Discussion paper. Brisbane, Australia: Department Employment Education and Training. McGrath, J.E. (1976). Stress and behaviour in organisations. In M.D. Dunnette (Ed.), Handbook of industrial and organisational psychology (pp. 1351–1395). Chicago: Rand McNally. Miles, I. (1983). Adaptation to unemployment. Brighton, UK: Science Policy Research Unit, University of Sussex. Miles, I., & Howard, J. (1984). A study of youth employment and unemployment. Occasional Paper. Brighton, UK: Science Policy Research Unit, University of Sussex. Millman, Z., & Latham, G. (2001). Increasing reemployment through training in verbal self-guidance. In M. Erez & U. Kleinbeck (Eds.), Work motivation in the context of a globalizing economy (pp. 87–97). New York: Lawrence Erlbaum. Morrison, T.G., O’Connor, W.E., Morrison, M.A., & Hill, S.A. (2001). Determinants of psychological well-being among unemployed women and men. Psychology and Education: An Interdisciplinary Journal, 38, 34–41. Muller, J., Creed, P.A., Waters, L., & Machin, M.A. (2005). The development and preliminary testing of a scale to measure the latent and manifest benefits of employment. European Journal of Psychological Assessment, 21, 191–198. Muller, J., Hicks, R., & Winocur, S. (1993). The effects of employment and unemployment on psychological well-being in Australian clerical workers: Gender differences. Australian Journal of Psychology, 45, 103–108. Murphy, G.C., & Athanasou, J. (1999). The effect of unemployment on mental health. Journal of Occupational and Organizational Psychology, 72, 83–99. O’Brien, G., & Feather, N. (1990). The relative effects of unemployment and quality of employment on the affect, work values, and personal control of adolescents. Journal of Occupational Psychology, 63, 151–165. Payne, R. (1988). A longitudinal study of the psychological well-being of unemployed men and the mediating effect of neuroticism. Human Relations, 41, 119–138. Rantakeisa, U., Starrin, B., & Hagquist, C. (1999). Financial hardship and shame: A tentative model to understand the social and health effects of unemployment. British Journal of Social Work, 29, 877–901. 18
Unemployment Theories and Beyond Roseman, I.J., & Smith, C.A. (2001). Appraisal theory: Overview, assumptions, varieties, controversies. In K.R. Scherer, A. Schorr, & T. Johnstone (Eds.), Appraisal processes in emotion: Theory, methods, research (pp. 3–19). New York: Oxford University Press. Rowley, K.M., & Feather, N.T. (1987). The impact of unemployment in relation to age and length of unemployment. Journal of Occupational Psychology, 60, 323–332. Schaufeli, W.B. (1992). Unemployment and mental health in well- and poorly-educated school-leavers. In C.H.A. Verhaar et al. (Eds.), On the mysteries of unemployment (pp. 253–271). Dordrecht, the Netherlands: Kluwer Academic. Schaufeli, W.B. (1997). Youth unemployment and mental health: Some Dutch findings. Journal of Adolescence, 20, 281–292. Scheff, T.J. (1990). Microsociology: Discourse, emotion and social structure. Chicago: University of Chicago Press. Scherer, K.R., Schorr, A., & Johnstone, T. (Eds.). (2001). Appraisal processes in emotion: Theory, methods and research. New York: Oxford University Press. Ullah, P. (1990). The association between income, financial strain and psychological wellbeing among unemployed youths. Journal of Occupational Psychology, 63, 317–330. Underlid, K. (1996). Activity during unemployment and mental health. Scandinavian Journal of Psychology, 37, 269–281. Vinokur, A.D., Price, R.H., & Caplan, R.D. (1991). From field experiments to program implementation. American Journal of Community Psychology, 19, 543–562. Wanberg, C.R., Griffiths, R.F., & Gavin, M.R. (1997). Time structure and unemployment: A longitudinal investigation. Journal of Occupational and Organizational Psychology, 70, 75–95. Warr, P.B. (1987). Work, unemployment and mental health. Oxford, UK: Clarendon Press. Warr, P. (1994). A conceptual framework for the study of work and mental health. Work and Stress, 8, 84–97. Warr, P.B. (1999). Well being and the workplace. In D. Kahneman, E. Diener, & N. Schwarz (Eds.), Well-being: The foundations of hedonic psychology (pp. 312–412). New York: Russell Sage Foundation. Warr, P.B., Banks, M.H., & Ullah, P. (1985). The experience of unemployment among black and white urban teenagers. British Journal of Psychology, 76, 75–87. Warr, P.B., Jackson, P., & Banks, M. (1982). Duration of unemployment and psychological well-being in young men and women. Current Psychological Research, 2, 207–214. Warr, P.B., & Payne, R.L. (1983). Social class and reported changes in behaviour after job loss. Journal of Applied Social Psychology, 13, 206–222. Waters, L.E., & Moore, K.A. (2002). Reducing latent deprivation during unemployment: The role of meaningful leisure activity. Journal of Occupational and Organizational Psychology, 75, 15–32. Whelan, C.T. (1992). The role of income, life style deprivation and financial strain in mediating the impact of unemployment on psychological distress: Evidence from the Republic of Ireland. Journal of Occupational and Organizational Psychology, 65, 331–344. Wiener, N. (1948). Cybernetics: Control and communication in the animal and the machine. New York: Wiley. Winefield, A.H. (1995). Unemployment: Its psychological costs. In C.L. Cooper & I.T. Robertson (Eds.), International review of industrial and organizational psychology (pp. 169–212). Chichester, UK: Wiley. Winefield, A.H., & Tiggemann, M. (1985). Psychological correlates of employment and unemployment: Effects, predisposing factors and sex differences. Journal of Occupational Psychology, 58, 229–242. 19
Unemployment and Health ■ Health Effects Winefield, A.H., Tiggemann, M., & Goldney, R.D. (1988). Psychological concomitants of satisfactory employment and unemployment in young people. Social Psychiatry and Psychiatric Epidemiology, 23, 149–157. Winefield, A.H., Tiggemann, M., & Winefield, H.R. (1992). Spare time use and psychological well-being in employed and unemployed young people. Journal of Occupational and Organizational Psychology, 65, 307–313.
20
Chapter
2
Insecurity, the Restructuring of Unemployment and Mental Health David Fryer
Unemployment puts mental health at risk. While there have been grounds for concern about the social, physical and mental health consequences of unemployment for at least 200 years (Burnett, 1994; Garraty, 1978; Keyssar, 1986), this has been beyond reasonable doubt since at least the 1930s (Bakke, 1933; Eisenberg & Lazarsfeld, 1938; Jahoda, 1938/1987; Lazarsfeld-Jahoda & Zeisel, 1933; Pilgrim Trust, 1938). It was confirmed by substantial empirical research and scholarship in the 1980s and is currently being reiterated by sophisticated reviews and metareviews. Research clarifying the relationship between unemployment and mental health has been carried out in many ‘first world’ industrialised countries, including Australia (e.g., Feather, 1992, Winefield, Tiggemann & Winefield, 1993); Austria (e.g., Kirchler, 1985; Moser & Paul 1999; Belgium (e.g., van Heeringen & Vanderplasschen, 1999); Canada (Goldenberg & Kline, 1997; Finland (e.g., Lahelma & Kangas, 1989; Virtanen, 1993); The Republic of Ireland (e.g., Whelan, 1992); Germany (e.g., Kronauer, Vogel, & Gerlach, 1993); Italy (e.g., Gatti, 1935; Pugliese, 1993); the Netherlands (e.g., Engbersen, Schuyt, Timmer, & Van Waarden, 1993; Verhaar, de Klauer, de Goede, van Ophem, & de Vries, 1996); New Zealand (e.g., Barnett, Howden-Chapman, & Smith, 1995; Drewery, 1998); Norway (e.g., Claussen & Bertran, 1999; Westin, 1990; Ytterdahl, 1999); Spain (e.g., Varela Novo, 1999); Sweden (e.g., Brenner, Petterson, Arnetz, & Levi, 1989; Janlert & Hammarström, 1992); the United Kingdom (see Fryer & Ullah, 1987; Smith, 1987; Warr, 1987); and the United States (e.g., Dooley, Catalano, & Hough, 1992; Hamilton, Hoffman, Broman, & Rauma, 1993). The early community-based studies remain exemplars of empirical research excellence with their painstaking attention to detail in context, ingenious triangulation of qualitative and quantitative research methods, integrity of fieldwork practice and boldly innovative conceptualisation 21
Unemployment and Health ■ Health Effects
(Jahoda, 1938/1987; Lazarsfeld-Jahoda & Zeisel, 1933). Some recent community-based studies bear comparison with these classic studies (Bostyn & Wight, 1987; Pappas, 1989; Wight, 1993), but in general, recent research has been dominated by decontextualised quantitative cross-sectional and longitudinal survey research, using scales with proven reliability and validity to measure mental health operationalised in a variety of ways. This research played an important function in the development of the field. Well-designed longitudinal studies tracked large, carefully matched samples of people in and out of paid jobs, from school or employment to unemployment, and from unemployment to employment. In study after study, groups that became unemployed during the course of the research exhibited deterioration in mean mental health compared with continuously employed groups. Some of the most persuasive longitudinal quantitative studies were done with young people. Typically these studies measured the mental health of large groups of young people in school and followed them out into the labour market, periodically measuring the mental health of those who got jobs and those who did not, and comparing group mean scores cross-sectionally and longitudinally. In study after study, groups of unemployed youngsters were demonstrated to have poorer mental health than their employed peers but statistically significant differences were seldom found between the scores of the same groups when at school (for a relatively rare exception, see Hammarström, 1994). That is, overwhelmingly poor mental health was shown to be the consequence rather than the cause of the labour market transition. In brief, such studies provided powerful evidence that unemployment causes, rather than merely results from, poor psychological health. Many such studies have been done but three research programs have been particularly influential. The first was the 1980s research program based at the Social and Applied Psychology Unit in Sheffield, England (see Warr, 1987, for an overview or, for more detail, see the following authors in various permutations: Banks & Jackson, 1982; Fryer & Ullah, 1987; Payne, Warr, & Hartley, 1984; Stafford, Jackson, & Banks, 1980). The second program was based at Flinders University in South Australia (see Feather, 1992; O’Brien, 1986). The third was based at the University of Adelaide (see Winefield et al., 1993). Recently, sophisticated metareviews have reiterated that social causation is involved (Moser & Paul, 1999; Murphy & Athanasou, 1999). To give one example, after a metareview of 16 longitudinal studies using valid and reliable measures and published in the last 10 years in an English language scientific journal, Murphy and Athanasou (1999) concluded that ‘the results from the 16 longitudinal studies suggest that unemployment has reliable (negative) effects on mental health’. They also reported that the ‘effect size information … suggests that moving from unemployment to 22
Insecurity, the Restructuring of Unemployment and Mental Health
employment not only produces a reliable change to mental health, but a change that is “practically significant”’. It was never remotely plausible that mass unemployment was caused by mass, sometimes organisationally confined, epidemics of mental illness. Nevertheless, this has not discouraged some politicians and commentators from suggesting that the relationship between unemployment and poor mental health is best explained by people with poorer mental health being more likely to become and remain unemployed (‘individual drift’ or ‘selection’). This has the merit, from their perspective, of blaming unemployed people for their own distress and absolving governments and businesses of responsibility for the health consequences of unemployment. Of course, in some individual cases, preexisting mental health problems predispose people to loss of, and/or failure to get, a job. However, even in these cases social causation processes may be hard to discount. For example, during economic recession some employers ratchet up the criteria they use to select employees, with the result that people are excluded from employment on mental and physical health grounds at one stage of the economic cycle who would have been employed in other economic circumstances (Catalano & Kennedy, 1998). Additionally, some who become unemployed because of preexisting mental health problems have those prior mental health problems exacerbated or compounded through social causation processes. Finally here, some whose poor mental health renders them difficult to employ are mentally unhealthy because of prior pathogenic labour market experiences. Therefore, mental health in relation to unemployment and reemployment has to be understood in the context of dynamic labour market experience over time. Other conceptual and methodological concerns suggesting that the dichotomy between social causation and individual drift may be false are discussed more fully in Fryer (1997) and Winefield and Fryer (1996). In subconclusion, the research described here, and elsewhere, has persuaded most researchers in the field that unemployment has mental health consequences that are negative, widespread and sometimes severe.
The Size of the Unemployment Problem The number of people at risk of negative psychological effects of unemployment is vast, though usually underestimated. Until recently, unemployment was officially measured in the United Kingdom in terms of the number of people in receipt of unemployment benefit, the so-called ‘claimant count’. Counted this way, there were 940,500 ‘claimants’ in the United Kingdom in September 2002. However, this way of measuring unemployment is flawed for our purposes because large numbers of changes in the eligibility criteria for claiming unemployment benefit (called ‘job seekers’ allowance’ at the time of writing) have been made by UK governments over the years for political and administrative reasons. Increasingly strict criteria have 23
Unemployment and Health ■ Health Effects
excluded many people from state financial support and also have had the politically convenient result of reducing the published headline counts of the number of people unemployed. Changes in the number of people who ceased to be recorded as unemployed have not always been the same as changes in the numbers of people who ceased to be unemployed. For these and other reasons, a survey definition that counts people as unemployed if they have not undertaken any paid employment in the survey reference week, have looked for a job within the previous four weeks and are available to start one within two weeks, which is the International Labour Organization (ILO) definition, is preferred by many. Using this yardstick there were 1,541,000 people unemployed in the United Kingdom in July–September 2002. However, given that the psychological consequences of unemployment include depression and demoralisation, it might be expected that some unemployed people might stop actively looking for employment after a period of time, yet might still be appropriately counted as unemployed. Accordingly, another way of indicating how many people are unemployed (the Broad Labour Force Survey) counts all those who say they want a job and are available to start within two weeks but who have not necessarily been actively looking for a job (they think there are no jobs to look for, that even if they found a job they would not be offered it, and so on). This way of counting gives a UK unemployed count of 2,170,000 for the same period. However, this may still greatly underestimate the number of people looking for employment. If we also count the full-time equivalence of those on employment and training schemes and those who are involuntarily parttime, the UK figure climbs to 4,217,700 people, the so-called slack labour force (Bivand, 1999). So far, this is to refer only to the unemployment situation in the United Kingdom. In 1996, the ILO calculated unemployment to stand at 34 million in the countries of the OECD (Organisation for Economic Cooperation and Development). Worldwide, the ILO calculated unemployment to be 184.7 million at the end of 2004 (ILO, 2005). That said, in the United Kingdom, all measures of unemployment have been falling in recent years. Even the slack labour force measure has fallen from 16% of the economically active UK population in June–August 1997 to 13.3% currently. On the world stage, in the 11 countries of the ‘Eurozone’ of the Economic Community, unemployment was predicted by the OECD to fall from 10.4% in 1999 to 8.9% in 2001, while unemployment in OECD countries was predicted to fall from an average of 9.4% in 1999 to 8.3% by the end of 2001 (Atkinson, 1999). Given the negative consequences of unemployment on mental health discussed above, these apparent reductions in unemployment are surely good news for public health.
24
Insecurity, the Restructuring of Unemployment and Mental Health
The ‘Flexible Labour Market’ Taking the 1960s as a reference period, the UK labour market has traditionally been seen as dominated by male employment in permanent, secure, full-time, unionised jobs in large companies in manufacturing industries. This was always an imperfect characterisation, of course, but captures something of the mainstream, some might say ‘malestream’, view. Certainly, there is now wide consensus that a shift is well underway from male to female, permanent to temporary, full-time to part-time, unionised to deregulated, manufacturing to service employment (Leadbeater & Martin, 1998; Royal Society of Arts, 1997). Over the last twenty years or so, millions of full-time jobs have been lost and millions of part-time jobs have been created (Ford, 1995). However, nearly three quarters of the jobs created since 1992 have not only been part-time, but also temporary, and short-term and fixed contract employment is growing (RSA, 1997). In at least some periods, nearly 90% of the overall increase in employment was accounted for by temporary, insecure employment (The Trades Union Council, 1995). In the decade from 1975, the average period men stayed with an employer fell from just over 8 years to 61/2 years and it has fallen since to an all-time low (Leadbeater & Martin, 1998). In 1992, 9.3% of workers in the European Union were on temporary contracts, with more than three out of five of such jobs being in low-skilled occupations (De Grip, Hoevenberg, & Willems, 1997). In the decade from 1984, the number of men working part-time in the United Kingdom doubled (Cooper, 1998); however, there were still five times as many UK women working part-time as men. About 15% of the European workforce was employed part-time in 1991 and the percentage has been increasing since then. However, there is massive variation between countries: over 30% of employment in the Netherlands is part-time, as opposed to 3.4% in Greece, for example. Fifty-six per cent of part-time work in the European Union is done by low-skilled (and low-paid) workers and twice as many woman as men are employed part-time (De Grip et al., 1997). In 1996, the Scottish Low Pay Unit surveyed and analysed the 6142 vacancies advertised in 28 Jobcentres. Just over 40% of the jobs were part-time, with 20% offering less than 16 hours per week, and just over 15% temporary. Three quarters of the jobs advertised paid £4 per hour or less with nearly a third of the vacancies paying below the National Insurance threshold. As the authors of the report commented, in these circumstances ‘paid work will only move large numbers of people from unwaged poverty to waged poverty’. The trend looks set to continue. In May 1999, at its annual meeting in Paris, the OECD recommended member countries make working practices more flexible: abolish employment protection policies, ensure that minimum wages for young people are not prohibitively high, switch from 25
Unemployment and Health ■ Health Effects
unemployment to in-work benefits (Coyle, 1999). Part-time work, on-call contracts, fixed-term contracts, seasonal work, agency work, working from home, teleworking, freelancing, self-employment and informal work are all on the increase (Delsen, 1991). In Europe the term ‘flexible work force’ is widely used to refer to this cocktail of employment contracts. However, many other terms, some of them more disparaging, seem to be used as variants: ‘churning of the labour market’, ‘disguised unemployment’, ‘underemployment’, ‘contingent workforce’, ‘careers of labour market disadvantage’, ‘atypical employment’, ‘precarious employment’, ‘nonstandard employment’ (Beukema & Valkenburg, 1999; Cooper, 1998; Crompton, 1999; Delsen, 1991; Dooley & Catalano, 1999; Fryer, 1997). In summary, ‘the ongoing restructuring and globalization of the economy has changed the nature of work. Workers accustomed to well paid, full-time, secure jobs are increasingly falling not into unemployment but into less desirable forms of work’ (Dooley & Catalano, 1999, p. 18).
Mental Health Consequences of ‘Less Desirable Forms of Work’ What are the psychological consequences of flexible labour market participation and how do they compare with those of unemployment? In a recent Joseph Rowntree Foundation study (Burchell et al., 1999), research respondents reported that, over the last 5 years, their working hours had increased, and they were having to work more quickly and with less adequate staffing levels. Experienced work pressure was associated with poor psychological wellbeing and more tense family relations, particularly when supervisory support was weak. Interestingly, while working hours have generally been falling in the European Union, those in Ireland and the United Kingdom have actually increased, with employees in Britain now working the longest hours in Europe. The flexible labour market has probably developed further and faster in the United Kingdom than in any other European country and, depressingly, according to Cooper (1998) ‘an ISR (1995) survey, of 400 companies in 17 countries employing over 8 million workers throughout Europe, found that over the last ten years, the UK’s employee satisfaction level dropped from 64% in 1985 to 53% in 1995, the biggest drop of any European country’. According to Dooley and Catalano (1999), in the United States ‘underemployment (involuntary part-time or poverty wage) both appears increasingly common and seems to have health effects more like those of unemployment than adequate employment’. As long ago as 1938, it was recognised that ‘just having a job itself is not as important as having a feeling of economic security. Those who are economically insecure, employed or unemployed, have a low morale’ (Eisenberg & Lazarsfeld, 1938, p. 361; see also Cobb & Kasl, 1977, for 26
Insecurity, the Restructuring of Unemployment and Mental Health
classic psychophysiological research in this area). Advances in clarifying phenomenological and conceptual aspects of job insecurity came with Hartley et al. (1991), who explicated the notion of ‘job insecurity’ as the ‘fundamental and involuntary change from a belief that one’s position in the employing organisation is safe, to a belief that it is not’ (Hartley et al., 1991, p. viii). Studies by the authors in Israel, the Netherlands and the UK showed that job insecurity is associated with experienced powerlessness, impaired mental health (depression and reported psychosomatic symptoms), reduced job satisfaction, reduced organisational commitment, reduced trust in management, resistance to change and poorer industrial relations. Burchell et al. (1999) also report the results of secondary analysis of data from relevant sources, in-depth interviews and a survey of employees at senior management, line management and shop-floor level in 20 organisations involved in a variety of work: manufacturing, transport, retail, education, health care, and so on. The researchers found that job insecurity has gradually increased over the last 30 years for both men and women in both full-time and part-time employment and is now higher than at any time since 1945. While job insecurity has traditionally been accepted as problematic for blue-collar workers, professional workers (who in 1987 were the most secure workers) were found to be the most job insecure 10 years later. Job insecurity was found to be associated with poorer psychological wellbeing (as measured by the GHQ-12; Goldberg, 1972), poorer family relations and organisational demotivation. Worryingly, both physical and mental wellbeing were found to continue to deteriorate with continuing job insecurity. Isaksson (personal communication, 1999), in a study of repeated downsizing in a large retail company, found that the mental health, as measured by the GHQ-12 again, of those who were still employed, but insecurely so, was as bad as that of those who had been made unemployed (see also Burchell, 1994; Dekker & Schaufeli, 1995). Platt, Pavis and Akram (1999) conducted a systematic and comprehensive review of research between 1993 and 1998. They concluded that job insecurity (often as a result of privatisation and restructuring) is associated with deteriorating physical and mental health (see Armstrong-Stassen, 1993; Ferrie, Shipley, Marmot, Stansfield, & Smith, 1995, 1998; Heaney, Israel, & House, 1994; Kinnunen & Natti, 1994; Nelson, Cooper, & Jackson, 1995; Vahtera, Kivimaki, & Pentti, 1997). Previously unemployed people who have become reemployed are disproportionately at risk of job insecurity, as reemployed people are vulnerable to being made unemployed again due to last-in-first-out practices (Daniel, 1974, 1990). However, researchers are increasingly emphasising that job insecurity is not only, or necessarily, about fear of job loss but may also be about perceived risk of a deteriorating position within an organisation, for example, loss of status within the company or reduced opportunity for promotion (Burchell et al., 1999). It may even be ‘the pervasive fear 27
Unemployment and Health ■ Health Effects
of unknown threats’, as Hallier (2000) puts it, writing of ‘security abeyance’. This is ‘a chronic state of general concern about the individual’s security which emerges from specific organisational changes possessing ambiguous meaning and contractual salience for the continuation of the job’, but not associated with job loss. Little is yet known about the mental health consequences of such states. Further, evaluation of mental health consequences of participation in employment and training schemes, a vital part of the flexible labour market, reveals a complicated picture. Stafford (1982), in a cross-sectional study, found that trainees’ mental health was not significantly worse than that of an employed comparison group and was significantly better than the mental health of an unemployed comparison group. Branthwaite and Garcia (1985), on the other hand, found their trainees were significantly more depressed than an employed comparison group and not significantly less depressed than an unemployed comparison group. Davies (1992) revealed a multifaceted experience of ‘community program’, with participants positive about some aspects and negative about others. The nature of the labour market context, quality of training and job placement, and degree of experienced compulsion are probably critical. Virtanen (1993) found that unemployed people, of any age or either sex, who were forced into employment through workfare-type programs used primary care facilities (e.g., visits to the doctor) more frequently, while their primary care visits decreased if they became unemployed again. Reemployment in poor quality jobs can be as psychologically corrosive as unemployment, and may even be more so. It has been shown that reemployment is likely to be at a lower level (Daniel, 1974, 1990). Kaufman (1982) reported that 20% of his reemployed professionals were underemployed (in terms of salary, type of work and skill use) and less than 50% said that their lives had returned to normal after reemployment, while Fineman (1987, p. 269) found ‘those reemployed in jobs which they felt to be inadequate were experiencing more stress and even poorer self-esteem, than they had during their period of unemployment’. More recently, Graetz (1993, p. 722) concluded: the health consequences of employment and unemployment are directly contingent on the quality of work. This means that the benefits of employment are confined to those, albeit a majority, who manage to find a satisfying job. In contrast, those who do not end up in a satisfying job — approximately one in every five workers — report the highest levels of health disorders. At the same time, the adverse consequences of job loss are confined to those who were satisfied with their former jobs.
Winefield et al. (1993) followed 3000 young Australians over a 10-year period and found that those youngsters who had taken jobs with which they were dissatisfied were indistinguishable in terms of mean mental health scores from unemployed youngsters. Both dissatisfied employed and unemployed young people had poorer mean mental health than satisfied 28
Insecurity, the Restructuring of Unemployment and Mental Health
employed young people. More recently, Leana and Feldman (1995, p. 1398) report that ‘unsatisfactorily reemployed workers reported significantly higher levels of anxiety and psychological distress than those who were satisfactorily reemployed; they also reported significantly lower levels of life satisfaction than those who were unemployed’. Finally, there is a vast literature on the impact of organisational or employment stressors on psychological strain (see Fryer & Winefield, 1998; Karasek & Theorell, 1990).
Some Implications Most recent research on unemployment and mental health has proceeded by contrasting the better mean mental health of employed people with the worse mean mental health of unemployed people in both cross-sectional and longitudinal designs. In the past the mental health of employed groups has been reliably better than that of otherwise comparable unemployed groups. As Murphy and Athanasou (1999) noted, ‘because the majority of those in work are satisfied, the overall effect of employment is higher well-being’ than that of the minority in unemployment. However, should the mental health of the majority of those in employment deteriorate even relatively moderately, or that of a sizeable minority deteriorate more dramatically, the lower mean mental health of samples of employed people would obliterate the formerly largely reliable finding that unemployed respondents have poorer mean mental health. The evidence reviewed in this chapter about the flexible labour market suggests that this trend is well underway. Note that, if the trend were to continue, it would weaken the research strategy of demonstrating that unemployment causes negative mental health consequences by showing that, on average, mental health deteriorates as groups of people move from employment to unemployment, not because unemployment had become any the less psychologically damaging, but because employment had become more psychologically damaging. There are many implications of this way of thinking. First, if research through cross-sectional and longitudinal comparison of the mental health of matched employed and unemployed groups is to continue to be informative it will become necessary to establish mental health benchmarks against which the mean mental health of each group can independently be gauged. As any particular benchmark could be regarded as sociohistorically arbitrary, this is likely to be very problematic. Second, the realisation that comparison of labour market transition outcomes is unsatisfactory in certain economic circumstances should alert researchers to the fact that it has always been unsatisfactory, relying upon ideologically problematic assumptions about the salutogenic nature of employment. Rather than relying on discovering differences in outcomes, researchers must increasingly try to establish the social causation of mental health problems by both unemployment and employment by uncovering 29
Unemployment and Health ■ Health Effects
the detailed multilevel processes through which social causation occurs. This is likely to require ecologically contextualised qualitative research. Third, the assumption that threats to the mental health of unemployed people cease when they become employed is no longer tenable. For many, mental health is at risk whether employed, quasi-employed, underemployed or unemployed. Radical rethinking is required of the alleged dichotomy between employment and unemployment. The problems of unemployment have been widely seen in this field as the problems of the absence of employment. Consideration of the pathogenic nature of flexible labour market participation may be a stimulus to reconceptualisation. For many unemployed people the search for a job and the hoops through which they have to jump to receive income in the meantime is a job in itself. This work under punitive state surveillance and income sanction is low-status, insecure, part-time and poorly paid, carried out in poor working conditions with virtually no negotiating rights or scope for collective action and with a high risk of occupational strain. Much is said these days about employment being a route out of ‘social exclusion’. Ironically, the social inclusion of flexible labour market participation might be anything but salutogenic for many and, indeed, unemployment might itself be usefully regarded not so much as exclusion from the labour market as another form of involuntary insecure pathogenic participation in the flexible labour market.
Author Note This chapter has been published in German as: Fryer, D., (2000). Strukturwandel der Arbeitslosigkeit und psychische Gesundheit. In H.G. Zilian & J.F. (Eds.), Soziale Sicherheit und Strukturwandel der Arbeitslosigkeit (pp. 240–256). Munich, Germany: Rainer Hampp Verlag.
Acknowledgments The support of ESRC grant ROOO 231798 during some of the preparatory work for this paper is acknowledged with gratitude.
References Armstrong-Stassen, M. (1993). Production workers’ reactions to a plant closing: The role of transfer, stress, and support. Anxiety, Stress and Coping, 6, 201–214. Atkinson, M. (1999, November 17). Don’t stow the life jackets yet. The Guardian, p. 31. Bakke, E.W. (1933). The unemployed man. London: Nisbet. Banks, M.H., & Jackson, P.R. (1982). Unemployment and risk of minor psychiatric morbidity in young people: Cross-sectional and longitudinal evidence. Psychological Medicine, 12, 789–798. Barnett, P., Howden-Chapman, P., & Smith, A. (1995). Unemployment, work and health: Opportunities for healthy public policy. The New Zealand Medical Journal, 108, 138–140. 30
Insecurity, the Restructuring of Unemployment and Mental Health Beukema, L., & Valkenburg, B. (1999). Atypical employment in the Netherlands. In J. Lind & I.H. Moller (Eds.), Inclusion and exclusion: Unemployment and nonstandard employment in Europe (pp. 111–129). Aldershot, UK: Ashgate. Bivand, P. (1999). Unemployment and benefits. Working Brief, 109, 21–24. Bostyn, A.-M., & Wight, D. (1987). Inside a community: Values associated with money and time. In S. Fineman (Ed.), Unemployment: Personal and social consequences (pp. 138–154). London: Tavistock. Branthwaite, A., & Garcia, S. (1985). Depression in the young unemployed and those on youth opportunities schemes. British Journal of Medical Psychology, 58, 67–74. Brenner, S.-O., Petterson, I.-L., Arnetz, B., & Levi, L. (1989). Stress reactions to unemployment among Swedish blue-collar workers. In B. Starrin, P.-G. Svensson, & Wintersberger, H. (Eds.), Unemployment, poverty and quality of working life (pp. 101–114). Berlin, Germany: Sigma. Burchell, B.J. (1994). Who is affected by unemployment? Job insecurity and labour market influences on psychological health. In D. Gallie, C. Marsh, & C. Vogler (Eds.), Social change and the experience of unemployment (pp. 188–212). Oxford, UK: Oxford University Press. Burchell, B.J., Day, D., Hudson, M., Ladipo, D., Mankelow, R. Nolan, J.P., et al. (1999). Job insecurity and work intensification: Flexibility and the changing boundaries of work. York, UK: Joseph Rowntree Foundation. Burnett, J. (1994). Idle hands: The experience of unemployment, 1790–1990. London: Routledge. Catalano, R., & Kennedy, J. (1998). The effects of unemployment on disability caseloads in California. Journal of Community and Applied Social Psychology, 8, 137–144. Claussen, B., & Bertran, J. (1999). The International Commission on Occupational Health Working Group ‘Unemployment and Health’, Paris 25–26 September 1998. International Archives of Occupational and Environmental Health, 72(Suppl.), 1–48. Cobb, S., & Kasl, S.V. (1977). Termination: The consequences of job loss. Cincinnati, OH: US Department of Health, Education and Welfare. Cooper, C.L. (1998). The changing nature of work (1). Community, Work and Family, 1, 3. Coyle, D. (1999, May 26). OECD urges reforms to stem unemployment. The Independent, p. 23. Crompton, R. (1999). Non-standard employment, social exclusion, and the household: Evidence from Britain. In J. Lind & I.H. Moller (Eds.), Inclusion and exclusion: Unemployment and non-standard employment in Europe. Aldershot, UK: Ashgate. Daniel, W.W. (1974). A national survey of the unemployed. London: Political and Economic Planning Institute. Daniel, W.W. (1990). The unemployed flow. London: PSI. Davies, J.B. (1992). The experience of community programme, unemployment and employment: Mental health and individual differences. Unpublished doctoral thesis, the University of Sheffield, UK De Grip, A., Hoevenberg, J., & Willems, E. (1997). Atypical employment in the European Union. International Labour Review, 136, 49–71. Dekker, S., & Schaufeli, W. (1995). The effects of job insecurity on psychological health and withdrawal: A longitudinal study. Australian Psychologist, 30, 57–63. Delsen, L. (1991). Atypical employment relations and government policy in Europe. Labour, 7, 73–91. Dooley, D., & Catalano, R. (1999). Unemployment, disguised unemployment and health: The US case. International Archives of Occupational and Environmental Health, 72(Suppl.), 16–19. 31
Unemployment and Health ■ Health Effects Dooley, D., Catalano, R., & Hough, R. (1992). Unemployment and alcohol disorder in 1910 and 1990: Drift versus social causation. Journal of Occupational and Organisational Psychology, 65, 277–290. Drewery, W. (1998). Unemployment: What kind of problem is it? Journal of Community and Applied Social Psychology, 8, 101–108. Eisenberg, P., & Lazarsfeld, P.F (1938). The psychological effects of unemployment. Psychological Bulletin, 35, 258–390. Engbersen, G., Schuyt, K., Timmer, J., & Van Waarden, F. (1993). Cultures of unemployment: A comparative look at long-term unemployment and urban poverty. Oxford, UK: Westview Press. Feather, N.T. (1992). The psychological impact of unemployment. New York: Springer Verlag. Ferrie, J.E., Shipley, M.J., Marmot, M.G., Stansfield, S., & Smith, G.D. (1995). Health effects of anticipation of job change and non-employment: Longitudinal data from the Whitehall II study. British Medical Journal, 311, 1264–1269. Ferrie, J.E., Shipley, M.J., Marmot, M.G., Stansfield, S., & Smith, G.D. (1998). The health effects of major organisational change and job insecurity. Social Science and Medicine, 46, 243–254. Fineman, S. (1987). Back to employment: Wounds and wisdoms. In D. Fryer & P. Ullah (Eds.), Unemployed people: Social and psychological perspectives (pp. 268–284). Milton Keynes, UK: Open University Press. Ford, J. (1995). Middle England: In debt and insecure. Poverty, 92, 11–14. Fryer, D. (1997). International perspectives on youth unemployment and mental health: Some central issues. Journal of Adolescence, 20, 333–342. Fryer, D., & Ullah, P. (1987). Unemployed people: Social and psychological perspectives. Milton Keynes, UK: Open University Press. Fryer, D., & Winefield, A.H. (1998). Employment stress and unemployment distress as two varieties of labour market induced psychological strain: An explanatory framework. Australian Journal of Social Research, 5, 3–18. Garraty, J.A. (1978). Unemployment in history: Economic thought and public policy. New York: Harper and Row. Gatti, A. (1935). Prima relazione sulla efficienza lavorativa dei disoccupati. Italian Archives of Psychology, xiii, 67–91. Goldberg, D.P. (1972). The detection of psychiatric illness by questionnaire. Oxford: Oxford University Press. Goldenberg, S., & Kline, T. (1997). Canadian white-collar workers’ views of their experience with downsizing or job loss. Psychological Reports, 80, 707–719. Graetz, B. (1993). Health consequences of employment and unemployment: Longitudinal evidence for young men and women. Social Science and Medicine, 35, 715–724. Hallier, J. (2000). Security abeyance: Coping with the erosion of job conditions and treatment. British Journal of Management, 11, 71–89. Hamilton, V.L., Hoffman, W.S., Broman, C.L., & Rauma, D. (1993). Unemployment, distress, and coping: A panel study of autoworkers. Journal of Personality and Social Psychology, 65, 234–247. Hammarström, A. (1994). Health consequences of youth unemployment. Public Health, 108, 403–412. Hartley, J.F., Jacobson, D., Klandermans, B., van Vuuren, T., Greenhalgh, L., & Sutton, R. (1991). Job insecurity: Coping with jobs at risk. London: Sage. Heaney, C.A., Israel, B.A., & House, J.S.(1994). Chronic job insecurity among automobile workers: Effects on job satisfaction and health. Social Science and Medicine, 38, 1431–1437. 32
Insecurity, the Restructuring of Unemployment and Mental Health International Labour Organization. (2005). Global employment trends. Brief, February 2005. Retrieved December 19, 2005, from http://www.ilo.org/public/english/ employment/strat/stratprod.htm Jahoda, M. (1938/1987). Unemployed men at work. In D. Fryer & P. Ullah (Eds.), Unemployed people: Social and psychological perspectives (pp. 1–73). Milton Keynes, UK: Open University Press. Janlert, U., & Hammarström, A. (1992). Alcohol consumption among unemployed youths: Results from a prospective study. British Journal of Addiction, 87, 703–714. Karasek, R.A., & Theorell, T. (1990). Healthy work: Stress, productivity and the reconstruction of working life. New York: Basic Books. Kaufman, H.G. (1982). Professionals in search of work: Coping with the stress of job loss and underemployment. New York: Wiley. Keyssar, A. (1986). Out of work: The first century of unemployment in Massachusetts. Cambridge, UK: Cambridge University Press. Kinnunen, U., & Natti, J. (1994). Job insecurity in Finland: Antecedents and consequences. European Work and Organisational Psychologist, 4, 297–321. Kirchler, E. (1985). Job loss and mood. Journal of Economic Psychology, 6, 9–25. Kronauer, M., Vogel, B., & Gerlach, F. (1993). Im Schatten der Arbeitsgesellschaft: Arbeitslose und die Dynamik sozialer Ausgrenzung. Frankfurt, Germany: Campus Verlag. Lahelma, E., & Kangas, R. (1989). Unemployment, re-employment and psychic well being in Finland. In B. Starrin, P.-G. Svensson, & H. Wintersberger (Eds.), Unemployment, poverty and quality of working life (pp. 135–163). Berlin, Germany: Sigma. Lazarsfeld-Jahoda, M., & Zeisel, H. (1933). Die Arbeitslosen von Marienthal. Psychologische Monographien, 5, 1933. [An English language translation appeared in 1971 as Jahoda, M., Lazarsfeld, P.F., & Zeisel, H. (1971). Marienthal: The sociography of an unemployed community. Chicago: Aldine Atherton.] Leana, C.R., & Feldman, D.C. (1995). Finding new jobs after a plant closing: Antecedents and outcomes of the occurrence and quality of reemployment. Human Relations, 48, 1381–1401. Leadbeater, C., & Martin, S. (1998). The employee mutual: Combining flexibility with security in the new world of work. London: Demos/Reed. Moser, K., & Paul, K. (1999, September). Arbeitslosigkeit und seelische Gesundheit, Metaanalysen. Paper presented at the conference Arbeitslosigkeit: Wege aus der Krise, University of Erlangen-Nurnberg, Germany. Murphy, G.C., & Athanasou, T.A. (1999). The effect of unemployment on mental health. Journal of Occupational and Organizational Psychology, 72, 83–99. Nelson, A., Cooper, C.L., & Jackson, P.R. (1995). Uncertainty amidst change: The impact of privatisation on employee job satisfaction and well-being. Journal of Occupational and Organizational Psychology, 68, 57–71. O’Brien, G.E. (1986). Psychology of work and unemployment. Chichester, UK: Wiley. Pappas, G. (1989). The magic city: Unemployment in a working class community. Ithaca, NY: Cornell University Press. Payne, R.L., Warr, P.B., & Hartley, J. (1984). Social class and psychological ill-health during unemployment. Sociology of Health and Illness, 6, 152–174. Pilgrim Trust. (1938). Men without work. Cambridge, UK: Cambridge University Press. Platt, S., Pavis, S., & Akram, G. (1999). Changing labour market conditions and health: A systematic literature review (1993–1998). Final report of Project 0203 Employment and Health to the European Foundation for the Improvement of Living and Working Conditions. 33
Unemployment and Health ■ Health Effects Pugliese, E. (1993). The Europe of the unemployed. In M. Kronauer (Ed.), Unemployment in Western Europe, Special Issue of the International Journal of Political Economy, 23, 3–120. Royal Society of Arts. (1997). Key views on the future of work. Redefining work discussion paper. Royal Society of Arts Journal, June, 7–13. Smith, R. (1987). Unemployment and health: A disaster and a challenge. Oxford, UK: Oxford University Press. Stafford, E.M. (1982). The impact of the Youth Opportunities Programme on young people’s employment prospects and psychological well-being. British Journal of Guidance and Counselling, 10, 12–21. Stafford, E.M., Jackson, P.R., & Banks, M.H. (1980). Employment, work involvement and mental health in less qualified young people. Journal of Occupational Psychology, 53, 291–304. Trades Union Council, The. (1995). Britain divided: Insecurity at work. London: TUC. Vahtera, J., Kivimaki, M., & Pentti, J. (1997). Effect of organisational downsizing on health of employees. The Lancet, 350, 1124–1128. van Heeringen, K., & Vanderplasschen, W. (1999). Unemployment and suicidal behaviour in perspective. International Archives of Occupational and Environmental Health, 72(Suppl.), 42–45. Varela Novo, M. (1999). Unemployment and mental health in Galicia, Spain. International Archives of Occupational and Environmental Health, 72(Suppl.), 14–15. Verhaar, C.H.A., de Klauer, P.M., de Goede, M.P.M., van Ophem, J.A.C., & de Vries, A. (Eds.). (1996). On the challenges of unemployment in a regional Europe. Aldershot, UK: Avebury. Virtanen, P. (1993). Unemployment, re-employment and the use of primary health care services. Scandinavian Journal of Primary Health Care, 11, 228–233. Warr, P.B. (1987). Work, unemployment and mental health. Oxford, UK: Clarendon Press. Westin, S. (1990). The structure of a factory closure: Individual responses to job-loss and unemployment in a 10-year controlled follow-up study. Social Science and Medicine, 31, 1301–1311. Whelan, C.T. (1992). The role of income, life-style deprivation and financial strain in mediating the impact of unemployment on psychological distress: Evidence from the Republic of Ireland. Journal of Occupational and Organisational Psychology, 65, 331–344. Wight, D. (1993). Workers not wasters: Masculine respectability, consumption and employment in central Scotland. Edinburgh, UK: Edinburgh University Press. Winefield, A.H., & Fryer, D. (1996). Some emerging threats to the validity of research on unemployment and mental health. The Australian Journal of Social Research, 12, 115–134. Winefield, A.H., Tiggemann, M., & Winefield, H.R. (1993). Growing up with unemployment: A longitudinal study of its psychological impact. London: Routledge. Ytterdahl, T. (1999). Routine health check-ups of unemployed in Norway. International Archives of Occupational and Environmental Health, 72(Suppl.), 38–39.
34
Chapter
3
Justice Concerns and Mental Health During Unemployment Claudia Dalbert
This chapter analyses the effect of unemployment on coping and mental health from a justice psychology perspective. The just world hypothesis serves as theoretical background and is briefly described. Hypotheses about the impact of the Belief in a Just World (BJW) on mental health, and problem- and emotion-focused coping are portrayed, and empirical support is summarised. In sum, evidence so far underlines that BJW impacts on emotion-focused coping and thus serves as a buffer for the unemployed individual’s mental health. BJW’s influence on efforts to become reemployed, however, requires systematic research. The psychological consequences of unemployment are well known: unemployment is a major risk factor for mental and physical ill health. Unemployed individuals are at risk of developing depression, and the longer the period of unemployment, the greater the psychological stress (e.g., Axelsson & Ejlertsson, 2002; Hammarström & Janlert, 2002; Martella & Maass, 2000; see for a review, Häfner, 1990). The term ‘structural unemployment’ describes unemployment arising from societal factors such as factory closures or significant shifts in the political landscape as, for example, took place in East Germany after unification and in Central Europe after the transformation of those countries’ political systems (e.g., Frese, 1994). Falling victim to unemployment — and especially to structural unemployment — implies suffering an unjust fate, a fate that is highly aversive but generally not self-inflicted, and therefore not deserved. Unfairness seems to be a prime theme for unemployed individuals. Indeed, one study found that individuals who were either unemployed or threatened by the possibility of unemployment rated their fate as significantly more unfair than victims of other serious fates such as giving birth to a premature or disabled child (Dalbert, 1997). Thus, unemployed individuals can undoubtedly be seen as victims of an unjust fate. But although victims 35
Unemployment and Health ■ Health Effects
often evaluate their unemployment as unfair, justice theories have rarely been applied in psychological unemployment research (Kieselbach, 1995). Feeling as though one has been treated unfairly constitutes a serious threat to an individual’s mental health and psychological functioning (Dalbert, 2001). As a consequence, the psychological consequences of unemployment may also be partly due to the associated experience of unfairness. Individuals cope differently with perceived unfairness. Thus, investigating differences in coping from a justice psychology perspective may deepen our understanding of the psychological consequences of unemployment. A starting point for such a perspective is provided by Lerner’s ‘just world hypothesis’ (e.g., Lerner, 1980). This theoretical approach is described in the following section. Following this, its implications for understanding the psychological consequences of unemployment are discussed and empirical evidence is summarised.
The Just World Hypothesis According to Lerner’s (1965, 1980) just world hypothesis, people are motivated to believe that the world is basically a fair place where people get what they deserve and deserve what they get. It is well documented that BJW varies in strength from individual to individual (Dalbert, 1999; Rubin & Peplau, 1975) and these individual differences explain a variety of human behaviour. Just world research has identified two functions of the BJW (Dalbert, 2001), which are important here. First, BJW endows individuals with the confidence that they will be treated fairly by others and will not fall victim to unforeseeable disaster. Thus, for example, people high in BJW feel less threatened by demands (Tomaka & Blascovich, 1994) and are more likely to achieve than people low in BJW (Dalbert & Maes, 2002). Second, because of these adaptive functions, people try to protect their BJW. Being confronted with an injustice that has been inflicted upon either oneself or others threatens the belief that justice prevails in the world. Individuals high in BJW therefore try to restore justice, either objectively, by modifying their environment, or psychologically, by modifying their response to it. When they experience unfairness that they do not believe can be resolved objectively, they try to assimilate this experience to their BJW. This can be done by justifying the experienced unfairness as being at least partly self-inflicted by the individuals or groups concerned (Bulman & Wortman, 1977; Comer & Laird, 1975; Lupfer, Doan, & Houston, 1998; for reviews, see Dalbert, 1996; Furnham, 2003; Furnham & Procter, 1989; Lerner & Miller, 1978), by playing down the unfairness of the experience (Dalbert, 1996; Lipkus & Siegler, 1993) and by avoiding self-focused rumination (Dalbert, 1997). Thus, BJW provides a conceptual framework that helps individuals to interpret the events of their personal life in a meaningful way. This assimilation of unfairness to one’s BJW is often highly adaptive; it enhances 36
Justice Concerns and Mental Health During Unemployment
wellbeing (e.g., Dalbert, 1998; Hafer & Correy, 1999) and reduces feelings of anger (Dalbert, 2002). In the eyes of the unemployed, unemployment is usually seen as a highly unfair experience of loss (Dalbert, 1997). Thus, to maintain psychological equilibrium, unemployed individuals should be motivated to protect their BJW, and, as a consequence, BJW must also impact on their mental health and psychological functioning. Three aspects of this relationship between BJW and mental health among the unemployed are outlined below: (1) BJW and problem-focused coping with unemployment; (2) empirical evidence of the relationship between BJW and the mental health of the unemployed; and (3) BJW and cognitive coping with unemployment.
Belief in a Just World and Problem-Focused Coping With Unemployment Since BJW may be viewed as some kind of positive illusion, it may be suspected of diminishing problem-focused coping. Just world research does not support such expectations, however. Questionnaire studies indicate that BJW acts as a buffer encouraging problem orientation. People high in BJW have been found to be low in procrastination (Ferrari & Emmons, 1994), self-defeating behaviour (Schill, Beyler, & Morales, 1992), wishful thinking and forgetfulness (Rim, 1986), but high in problem-orientation and action-planning (Rim, 1986). However, it should be noted that none of these studies involved victims coping with unemployment. Just world theory states that individuals want to be able to trust in the fairness of others. This is the personal contract: I am obliged to behave fairly and, as a result, I am able to trust in a just world. This trust in fairness has at least two aspects. First, individuals high in BJW are able to place more trust in others. They are less suspicious of others (Zuckerman & Gerbasi, 1977), less cynical about motives underlying others’ prosocial behaviour (Furnham, 1995) and they expect to be treated fairly by others, for example, by being confronted only with fair tasks (Tomaka & Blascovich, 1994). Overall, this trust in fairness has several consequences. People high in BJW are more likely to invest in their future and to direct their behaviour towards the fulfilment of long-term goals because they are confident that their investments will be fairly rewarded (Hafer, 2000; Otto & Dalbert, in press). Second, individuals high in BJW feel less threatened and less distressed by the demands of others, because they are confident that they will be treated fairly (Tomaka & Blascovich, 1994). Third, individuals high in BJW achieve more than individuals low in BJW (Dalbert & Maes, 2002; Tomaka & Blascovich, 1994). Until now, the impact of the second function on the coping efforts of the unemployed individual has not been investigated. However, it can be hypothesised that trust in being treated fairly by others should increase the investment of the unemployed in their future employment opportunities. In 37
Unemployment and Health ■ Health Effects
this vein, Lerner (2002) reports a positive relationship between BJW and efforts to find another job among employed respondents who were about to lose their jobs; however, this adaptive relationship turned into a negative one several weeks later. A study about the vocational goals of adolescent school leavers seems to support the notion of an adaptive relationship between BJW and investments in one’s vocational future (Dette, Stöber, & Dalbert, in press). The estimated probability of success is a central goal dimension that leads to goal-commitment (Hollenbeck & Klein, 1987; Metz-Göckel & Leffelsend, 2001), and goal-commitment can be seen as an important precondition of goal-attainment. Dette et al. (in press) focused on the expectancy of goal-attainment and tested the relationship between BJW and the belief in successful goal-attainment. Four hundred school students were asked about their two most important vocational goals and then to rate the likely success of these goals. As expected, students high in BJW believed more strongly that they could successfully attain their vocational goals. Moreover, this relationship was true even after controlling for global personality dimensions. Students who were better prepared for vocational training (e.g., had gathered information about schools offering vocational training), displayed more conscientiousness and less neuroticism, and were more likely to believe in successful goal-attainment than their less well-prepared counterparts. In addition to these three factors, BJW positively impacted on the expectancy of goal-attainment. Furthermore, other evidence suggests that reemployed individuals with a high BJW should feel less anxious and less threatened than those with a low BJW during the probationary period of a new job and that they should cope more successfully with the demands of their new job (see Dalbert & Maes, 2002; Tomaka & Blascovitch, 1994). In sum, BJW can be seen as a resource which enables the unemployed to cope more successfully either with the task of becoming reemployed, or, in the case of school-leavers, with the task of gaining a first job. A high BJW may enable unemployed individuals to persist in goal-attaining behaviour, that is, to invest in their occupational goals, and to get a better start in their new job. In contrast, unemployed individuals low in BJW are less likely to believe that they can successfully attain their occupational goals, are more likely to experience anxiety as a result of new job demands and are less likely to achieve. However, again, none of these hypotheses have so far been tested with unemployed individuals.
Belief in a Just World and the Mental Health of Unemployed Individuals Based on the second function of BJW — to provide a framework that helps to interpret events in one’s life in a meaningful way — the main focus of research on BJW and unemployment was the relationship between BJW and mental health. According to justice motive theory, BJW 38
Justice Concerns and Mental Health During Unemployment
serves as a psychological buffer for victims of unjust fates. Individuals high in BJW are more likely to assimilate the experienced unfairness to their BJW and, in so doing, maintain their mental health. Meanwhile, a large number of studies have supported the hypothesised positive relationship between BJW and various mental health dimensions (for a review, see Dalbert, 2001). As Dalbert (1998) has shown, BJW is even more strongly correlated with the mental health of victims than with that of nonvictims. This pattern of results supports the buffer hypothesis, that BJW is especially important in situations of unfairness (Dalbert, 2002). However, sometimes BJW is related to mental health for both victim and nonvictim groups alike, especially to more trait-like dimensions of mental health, such as life satisfaction. This may occur in part because most people experience unfairness from time to time. Consequently, BJW may be able to buffer the mental health of the majority of individuals. Thus, the buffering effect of BJW is not restricted to current victims only. Some of the existing victim studies have been conducted with samples of unemployed individuals or workers threatened by unemployment. In a questionnaire study of workers facing unemployment following a plant closure, Lerner and Somers (1992) found a positive relationship between two broad factors: wellbeing and personal beliefs, including BJW. Thus, they concluded that workers high in BJW were better off. By way of contrast, however, Benson and Ritter (1990) found unemployed adults high in BJW to be more depressed than those low in BJW. In an East German study, the aim of which was to assess individuals threatened by job insecurity, Dalbert (1993) compared the relationship between BJW and mental health in nonvictims (female university students) and victims (female teachers threatened by job insecurity). Overall, the study revealed that women high in BJW were more satisfied with their life, showed a better habitual mood level, and exhibited more positive and less negative mood states at the time of the assessment. With the exception of negative mood, which correlated with BJW only in the victim sample, the relationship between BJW and mental health was the same for victims and nonvictims. In this study as in further studies described below, the association between BJW and negative dimensions of subjective wellbeing often forms an exception by revealing no direct relationship. It can only be speculated why these exceptions occur. One possible explanation is that the effect of BJW on negative affect, in particular, is mediated by specific coping reactions, which leads to a nonsignificant direct relationship between BJW and negative affect. Sallay (in press) investigated BJW and mental health of university students graduating from university and compared those who were already employed with those who were still seeking a job. It was consistently shown that a strong BJW was associated with better mental health, consisting of a more positive attitude towards life, more joy in life, higher self39
Unemployment and Health ■ Health Effects
esteem, together with fewer somatic complaints and less depressive mood. Most importantly, these adaptive relationships were equally true for employed university leavers and job-seekers. Personal worries, on the other hand, did not show any relationship with BJW. In two questionnaire studies with three groups of Slovak adolescents (students, short-term and long-term unemployed), Dzuka and Dalbert (2002) explored whether BJW would act as a buffer between unemployment and mental health (life satisfaction, personal worries, positive affect, negative affect, self-esteem). BJW showed a positive relationship with life satisfaction, positive affect and self-esteem, and this was true after controlling for demographic (gender, unemployment status) and situational (subjective financial situation) predictors, and for additional personality dimensions such as social desirability, extraversion and neuroticism. These adaptive relationships were consistently observed for students and for all unemployed adolescents. However, the results for long-term unemployed participants and the negative mental health dimensions were mixed. For the long-term unemployed adolescents, a positive (maladaptive) relationship between personal worries and BJW was revealed in the first study; in the other, a negative (adaptive) relationship with negative affect was observed. With the exception of the negative mental health dimensions in the long-term unemployed group in the first study, both studies support the hypothesis that BJW buffers the unemployed individuals’ mental health. The results of these two Slovakian studies (Dzuka & Dalbert, 2002) support the notion that BJW’s meaning and impact may change over time. Following suggestions originating from earlier research (Furnham & Procter, 1989; Lerner & Miller, 1978), recent investigations have shown that it is necessary to distinguish the belief in a personal just world from the belief in a general just world (Dalbert, 1999; Lipkus, Dalbert, & Siegler, 1996). The personal BJW reflects the belief that events in one’s life are just; the general BJW reflects the belief that, basically, the world is a just place. It has been shown that the two kinds of BJW are significantly correlated. However, individuals tend to endorse the belief in a personal just world more strongly than the belief in a general just world, and personal BJW may be more important in predicting mental health and wellbeing (Lipkus, Dalbert, & Siegler, 1996; Dalbert, 1999). According to this differentiation, this pattern was observed in the two Slovakian studies described above, with the exception of the long-term unemployed group in the first one. In this first study (Dzuka & Dalbert, 2002), personal and general BJW were not correlated within the long-term unemployed group, and nor did their means significantly differ. A closer look at the general BJW in this first study showed that the items of this scale did not form a homogeneous scale within the long-term unemployed group. Thus, these odd scale characteristics may explain the unusual maladaptive relationship between general BJW and personal worries for the 40
Justice Concerns and Mental Health During Unemployment
long-term unemployed individuals in this study. However, in the second study, general BJW turned out to be a buffer for the mental health (was negatively associated with negative affect) of those adolescents who were long-term unemployed; simultaneously, general BJW showed the expected relationships towards personal BJW. In sum, the unexpected maladaptive relationship between BJW and personal worries may be seen as a samplespecific exception. Nevertheless, the results of the first study raise the question of whether BJW changes the meaning and impact of unemployment over time. After long-term unemployment (e.g., 12 months or more), individuals’ BJW may be undermined by the enduring reality of unemployment, and they may tend to alter their BJW in light of their burdening experiences. For example, Cubela Adoric (in press) revealed that Croatian long-term unemployed young adults (on average, unemployed about 8 years) believed less strongly in BJW than members of an employed comparison group. Or to state it differently, BJW’s buffering effect may typically occur during the early stages of unemployment. Generally speaking, most studies indicate an adaptive relationship between BJW and mental health (Dalbert, 1993; Dzuka & Dalbert, 2002; Lerner & Somers, 1992; Sallay, in press). However, as previously mentioned, a maladaptive relationship was observed by Benson and Ritter (1990). Moreover, the adaptive relationship seems to be more pronounced for positive mental health dimensions, such as life satisfaction, positive affect and self-esteem than for negative dimensions, such as negative affect or personal worries. Also, the relationship seems to be more consistent for the personal than for the general BJW. Doubts exist about whether the adaptive relationship holds true also for long-term unemployed individuals. Hence, findings concerning the long-term unemployed need to be examined in further detail.
Belief in a Just World and Emotion-Focused Coping With Unemployment Overall, existing research supports the notion that BJW may buffer the unemployed individual’s mental health against the effects of unemployment. Moreover, I contend that BJW can be seen as an adaptive resource that fosters the individual’s successful coping with unemployment. BJW and emotion-focused coping with unemployment were initially investigated in a study of 54 female unemployed East German blue-collar workers (e.g., Dalbert, 1997). This study was based on the premise that coping efforts mediate the impact of BJW on mental health. There are three reasons why BJW may be seen as a coping resource for the unemployed. First, in comparison with victims of other critical life events, unemployed individuals typically rate their fate of being unemployed as highly unjust (e.g., Dalbert, 1997), and injustice experiences are most typically accompanied by feelings of anger (e.g., Scherer, 1997; Smith & Ellsworth, 41
Unemployment and Health ■ Health Effects
1985). Subjects high in BJW aim to reduce their experience of injustice by minimising self-experienced injustice. It has been found that the more subjects believe in a just world the less they feel they are victims of injustice (Lipkus & Siegler, 1993) and the more they minimise the injustice of self-experienced critical life events (Dalbert, 1996). Thus, the adaptive effect of BJW on unemployed individuals’ mental health may be mediated by the tendency to minimise experienced injustice. The expected negative relationship between BJW and ratings of injustice was observed in Dalbert’s (1997) study of unemployed East German bluecollar workers. The more these women believed in a just world, the less they rated their fate of being unemployed as unfair. However, this injustice rating was not significantly correlated with any mental health dimension (Dalbert, 1996). Cubela Adoric (in press) compared long-term unemployed young Croatian adults with employed adults of the same age. In particular, she investigated the relationship between BJW, injustice experiences and discontent. Her study showed again the expected negative relationship between BJW and injustice ratings. Furthermore, she revealed that it is not so much the number of critical experiences per se — for example, the number of rejected job applications — that seems to shape BJW. Rather, the relationship between BJW and injustice experiences seems to be more important. Individuals who feel that they are regularly treated unjustly at their workplace or that they are often treated unfairly during their job search reveal a diminished BJW. Finally, the more often the individuals felt treated unjustly — either during job search or at the workplace — the more angry and disappointed they were. In sum, the studies of both Cubela Adoric (in press) and Dalbert (1996, 1997) demonstrated consistent negative relationships between BJW and injustice ratings. However, the relationship between injustice experiences and wellbeing was less consistent. Second, BJW provides unemployed individuals with a conceptual framework that guides the interpretation of misfortune. One way to reestablish justice is to justify the experience as being at least partly self-inflicted, because a self-inflicted misfortune is perceived as less unfair than one that is inflicted by others or external events (Smith & Ellsworth, 1985). Thus, BJW should increase the tendency for internal attributions of one’s harm in order to maintain one’s BJW and this, in turn, should protect one’s mental health. In addition to defending BJW, such internal attributions should enhance feelings of control over future life events. Because of both these hypothesised effects, internal attributions can be viewed as adaptive coping reactions. However, findings concerning the relationship between internal attribution and mental health are mixed: only nine out of 39 studies reviewed by Dalbert (2001) showed the expected adaptive relationship between internal attributions and mental health, while 14 evinced a maladaptive association and 16, the majority, demonstrated nonsignificant relations.
42
Justice Concerns and Mental Health During Unemployment
However, several studies have supported the notion that BJW leads to internal attributions for misfortune (e.g., Hafer & Correy, 1999; KiecoltGlaser & Williams, 1987), and the same has been observed for unemployed individuals (Dalbert, 1996, 2001). Not surprisingly, however, unemployed individuals tend to attribute their unemployment largely to external forces such as political changes. Only about 8% of the women in Dalbert’s (1996, 1997) study showed a tendency to attribute their unemployment partly to their own behaviour. These behavioural attributions were correlated significantly with the BJW, but not with any other variable. Thus, women high in BJW were slightly more likely to see their own behaviour as having contributed to their unemployment in some way; but overall, behavioural attributions were rare and did not seem to be an important part of the coping process. Third, unemployed individuals often find themselves in a state of despair, ruminating ‘Why me?’ and facing existential doubts. Brooding over the ‘Why me?’ question appears to be typical for the victims of critical life events. It constitutes the opposite of the belief that the world is structured and just, which leads to minimising the experienced injustice and enhances internal attributions of the misfortune. Being unable to answer ‘Why me?’ satisfactorily can thus be assumed to present a serious threat to the victim’s BJW, which may explain why asking this question is maladaptive. It is well documented for critical life events other than becoming unemployed that posing this question has a negative effect on victims’ mental and physical health (e.g., Bliemeister, Frey, Aschenbach, & Köller, 1992; Burgess & Holmstrom, 1979; Rogner, Frey, & Havemann, 1987; Silver, Boon, & Stones, 1983; Witenberg, Blanchard, Suls, Tennen, McCoy, & McGoldrick, 1983). The ‘Why me?’ question can be expected to be especially distressing for unemployed individuals, because they are likely to have difficulties finding a satisfactory answer. Making sense of the loss by reinterpreting the fate as self-inflicted may be a less reasonable strategy for victims of structural unemployment. In such cases, there is usually a socially shared interpretation that external factors, such as management or politics, are responsible for job cuts. Moreover, a study with Swedish unemployed individuals (Grossi, 1999) revealed that emotion-focused coping, consisting of selfblame, wishful thinking and impaired health behaviour, increased symptoms of anxiety and depression. It can therefore be expected that individuals high in BJW will avoid the ‘Why me?’ question, a course of action which should be possible at the beginning of the period of unemployment, at least. However, facing existential doubts in the form of ‘Why me?’ and, at the same time, feeling the need to believe in a just world can be expected to be even more threatening to mental health. The results of Dalbert’s study (1996, 1997) underline the importance of the ‘Why me?’ question for unemployed individuals. The majority of the unemployed women in this study (81%) had indeed asked themselves this 43
Unemployment and Health ■ Health Effects
question, but only one third of those individuals had found an answer. Women who did not ask ‘Why me?’ endorsed BJW more strongly than those who did. The relationship between BJW and the ‘Why me?’ question was even stronger for those who had been unemployed for 6 months or less. The negative correlation was significant for these respondents (about 23% variance in common), but for those who had been unemployed for longer than 6 months, the relationship was not significant. In sum, in line with the hypotheses, unemployed women high in BJW were found to avoid asking ‘Why me?’ especially during the first 6 months of unemployment. Finding an answer was independent of the extent of their BJW, however. Dalbert (1997, 2001) used moderated regression to analyse the effects of BJW and coping on the mental health of unemployed female blue-collar workers in the same sample mentioned above. Mental health was indicated by depressive mood state (Hautzinger & Bailer, 1993; Radloff, 1977). The interactions between BJW and asking ‘Why me?’ on the one hand, and finding an answer on the other, were both significant, explaining 19% of the variance in depression in addition to the variance already explained by other variables, such as length of unemployment. Negative (adaptive) relationships between BJW and depression were observed for women who did not ask ‘Why me?’ in particular, as well as for those women who did ponder this question and found an answer. Thus, if these two groups of women were high in BJW, they were less depressed. In contrast, a positive (maladaptive) relationship between BJW and depression was observed for those who ruminated on the ‘Why me?’ question without coming to an answer. Finally, the interaction between the length of unemployment and finding an answer was also significant (an additional 7% of variance) and revealed that finding a solution to the ‘Why me?’ question was more protective for the long-term unemployed. Thus, the longer the women had been unemployed, the more finding an answer reduced their levels of depression. Note that what kinds of answer were found to the ‘Why me?’ question did not shape the result pattern. To understand the women’s coping efforts further, these results should be examined in more detail. Four coping patterns could be distinguished. Let us start with the women who did not ask the ‘Why me?’ question. Women who neither asked this question nor believed in a just world were expected to be most severely depressed. The mental approach of these women can best be described as fatalistic: they did not believe in a just world, and unemployment was no incentive for them to start looking for meaning in their life. The study found that this mental approach was associated with depression. These women were very likely to show symptoms of depression, even in the early stages of unemployment. This fits with the hopelessness explanation of depression (e.g., Abramson, Metalsky, & Alloy, 1986).
44
Justice Concerns and Mental Health During Unemployment
It is important to distinguish fatalism (low BJW, ‘Why me?’ absent) from the adaptive BJW-protection strategy (high BJW, ‘Why me?’ absent). This second group of women protected their BJW by avoiding rumination about ‘Why me?’ Women taking this approach were highly successful in maintaining their mental health. They were not in danger of developing depressive symptoms. The only difference between the harmful fatalistic approach and the healthy BJW-protection strategy was the level of BJW. Thus, the BJW of these unemployed women appeared to be an important resource, enabling them to avoid rumination and to protect their mental health. A detailed scrutiny of those who did ask ‘Why me?’ further supports the importance of finding meaning in times of existential threat and for the important role of BJW in this process. The sample contained women who had a strong BJW, but who also ruminated on ‘Why me?’ These women were used to believing in a structured and just world, but their unemployment made them question these beliefs. This mental situation (high BJW, ‘Why me?’ present) is characterised by existential doubts and could, in the long run, result in a depressive mood state if the existential doubts are not resolved satisfactorily. Fortunately, such existential doubts in this sample were rare, at least in the early phases of unemployment. In the first 6 months of their unemployment, women high in BJW were less likely to ask ‘Why me?’ than those low in BJW. Existential doubts can be distinguished from a fourth mental state, which can be described as a search for meaning (low BJW, ‘Why me?’ present). Women engaged in a search for meaning did not believe in a just world, but unemployment seemed to stimulate them to look for meaning in their life by asking the ‘Why me?’ question. This kind of approach did not seem to increase the chances of depression. Indeed, it can be speculated that if these women were to succeed in finding an answer (and perhaps construct a belief in a just world), this search for meaning might be adaptive. Here again, it was shown that BJW was the decisive factor. To summarise, the implications and consequences of the ‘Why me?’ question differed, depending on whether the questioner believed in a just world or not. Note that unemployed women low in BJW exhibited more variation in mental health than women high in BJW. This again demonstrates the buffering capacity of BJW. While coping efforts could influence the unemployed women’s mental health to a limited degree, BJW was found to have a far greater impact. It is crucial to be able to give meaning to one’s life (Davis, Nolen-Hoeksema, & Larson, 1998), and BJW seems to be a major resource in this coping attempt. In sum, the BJW’s adaptive effect on the unemployed women’s mental health was mediated by the tendency of subjects high in BJW to avoid selffocused rumination. The more the blue-collar workers endorsed BJW, the less likely they were to ruminate about the ‘Why me?’ question, and this was especially true for persons who were unemployed for less than 7 months (Dalbert, 1997, 2001). 45
Unemployment and Health ■ Health Effects
Discussion The evidence reviewed above suggests that a justice psychological perspective may deepen our understanding of the psychological consequences of unemployment. Unemployed individuals typically evaluate their fate as being highly unfair, and coping with this injustice is a challenge to them. In this chapter, studies were summarised that give evidence to the notion that BJW is an adaptive resource in this coping process. Unemployed individuals who believe in a just world seem to be more likely to protect their mental health; in the studies reviewed, unemployed individuals with a strong BJW had higher self-esteem, were more satisfied with their life, and experienced more positive and less negative mood states than their counterparts with a weak BJW. This adaptive impact on mental health may be mediated by coping efforts. The findings suggested that unemployed workers high in BJW are more likely to avoid the ‘Why me?’ question, thus averting depressive symptoms. Note that BJW’s adaptive effect on mental health seems to be relatively short-lived, being evident only in the first 6 months of unemployment in our research. If unemployed subjects are not successful in regaining employment after a few months, another process seems to take place. A strong BJW can no longer protect against destructive rumination. After some months, unemployed individuals may still believe in a structured and just world, but their continued unemployment may make them start wondering whether the world really is a just place. Such existential doubts could, in the long run, undermine mental health. Moreover, in the long run, unemployment may shatter the belief in a just world itself. Individuals may start to accommodate their world beliefs to the enduring reality. Coming to an individually satisfying end of the search for meaning seems to be essential for the long-term unemployed individuals. One way to reestablish one’s BJW when confronted with long-term unemployment may be to narrow the boundaries of one’s personal just world, such that the belief in a personal just world may no longer include the occupational domain, but may focus on other domains, in particular, family life. It should be mentioned that the evidence that has been summarised in this chapter is correlational and cross-sectional. Therefore, inferences about underlying causal processes can only be tentative. However, the fact that the data reviewed in this article were gathered in different labour markets (Croatia, East Germany, Hungary, Slovakia and Canada), and from both adults and adolescents, suggests that the results can be generalised to a variety of populations; it also tends to support the causal inferences that have been set forth. Moreover, some of these studies controlled for alternative causal explanations, for example, global personality dimensions such as neuroticism, or the financial situation. Future research should apply a longitudinal and multidimensional approach, allowing conclusions about possible causes to be strengthened. Studies on BJW and unemploy46
Justice Concerns and Mental Health During Unemployment
ment so far have investigated emotion-focused coping. Clearly, problemfocused coping that results in becoming reemployed is an important coping strategy for unemployed individuals. Future studies should investigate more closely the interrelationship between emotion- and problem-focused coping from a justice psychological point of view.
Conclusion Overall, the evidence described in this chapter supports interpretations that identify beliefs in a just world as explaining emotion-focused coping behaviour and mental health during unemployment. The unemployed individuals struggle with their trust in fairness by others. They tend to protect this trust (BJW), which enables them to sustain their mental health and to invest in their future. The longer the unemployment endures the more the people seem to be unable to find meaning in their fate. These observations have implications for the psychological guidance of unemployed individuals. Future investigations on the psychological consequences of unemployment should take these justice concerns into account.
References Abramson, L.Y., Metalsky, G.L. & Alloy, L.B. (1986). The hopelessness theory of depression: Does the research test the theory? In L.Y. Abramson (Ed.), Social cognition and clinical psychology: A synthesis (pp. 33–65). New York: Guilford Press. Axelsson, L., & Ejlertsson, G. (2002). Self-reported health, self-esteem and social support among young unemployed people: A population-based study. International Journal of Social Welfare, 11, 111–119. Benson, D.E., & Ritter, C. (1990). Belief in a just world, job loss, and depression. Sociological Focus, 23, 49–63. Bliemeister, J., Frey, D., Aschenbach, G., & Köller, O. (1992). Zum Zusammenhang zwischen psychosozialen Merkmalen und dem Gesundheitszustand HIV-Infizierter: Eine interdisziplinäre Querschnittstudie [About the relationship between psychosocial descriptors and the health state of HIV-infected: An interdisciplinary cross sectional study]. Zeitschrift für Klinische Psychologie, 21, 182–196. Bulman, R.J., & Wortman, C.B. (1977). Attributions of blame and coping in the ‘real world’: Severe accident victims react to their lot. Journal of Personality and Social Psychology, 35, 351–363. Burgess, A.W., & Holmstrom, L.L. (1979). Adaptive strategies and recovery from rape. American Journal of Psychiatry, 136, 1278–1282. Comer, R., & Laird, J.D. (1975). Choosing to suffer as a consequence of expecting to suffer: Why do people do it? Journal of Personality and Social Psychology, 32, 92–101. Cubela Adoric, V. (in press). Belief in a just world and young adults’ way of coping with unemployment and the job search. In C. Dalbert & H. Sallay (Eds.), The justice motive in adolescence and young adulthood: Origins and consequences. London: Routledge. Dalbert, C. (1993). Gefährdung des Wohlbefindens durch Arbeitsplatzunsicherheit: Eine Analyse der Einflußfaktoren Selbstwert und Gerechte-Welt-Glaube [Well-being threatened by job-insecurity: Analyzing the effects of self worth and belief in a just world]. Zeitschrift für Gesundheitspsychologie, 1, 235–253. 47
Unemployment and Health ■ Health Effects Dalbert, C. (1996). Über den Umgang mit Ungerechtigkeit: Eine psychologische Analyse (Dealing with injustice: A psychological analysis). Bern, Germany: Huber. Dalbert, C. (1997). Coping with an unjust fate: The case of structural unemployment. Social Justice Research, 10, 175–189. Dalbert, C. (1998). Belief in a just world, well-being, and coping with a unjust fate. In L. Montada & M.J. Lerner (Eds.), Responses to victimizations and belief in a just world (pp. 87–105). New York: Plenum Press. Dalbert, C. (1999). The world is more just for me than generally: About the Personal Belief in a Just World Scale’s validity. Social Justice Research, 12, 79–98. Dalbert, C. (2001). The justice motive as a personal resource: Dealing with challenges and critical life events. New York: Plenum Press. Dalbert, C. (2002). Beliefs in a Just World as a buffer against anger. Social Justice Research, 15, 123–145. Dalbert, C., & Maes, J. (2002). Belief in a Just World as personal resource in school. In M. Ross & D.T. Miller (Eds.), The justice motive in everyday life (pp. 365–381). Cambridge, UK: Cambridge University Press. Davis, C.G., Nolen-Hoeksema, S., & Larson, J. (1998). Making sense of loss and benefiting from the experience: Two construals of meaning. Journal of Personality and Social Psychology, 75, 561–574. Dette, D.E., Stöber, J., & Dalbert, C. (in press). Belief in a Just World and adolescents’ vocational and social goals. In C. Dalbert & H. Sallay (Eds.), The justice motive in adolescence and young adulthood: Origins and consequences. London: Routledge. Dzuka, J., & Dalbert, C. (2002). Mental health and personality of Slovak unemployed adolescents: The impact of Belief in a Just World. Journal of Applied Social Psychology, 32, 732–757. Ferrari, J.R., & Emmons, R.A. (1994). Procrastination as revenge: Do people report using delays as a strategy for vengeance? Personality and Individual Differences, 17, 539–544. Frese, M. (1994). Psychische Folgen von Arbeitslosigkeit in den fünf neuen Bundesländern: Ergebnisse einer Längsschnittstudie [Psychic consequences of unemployement in the five new federal states]. In L. Montada (Ed.), Arbeitslosigkeit und soziale Gerechtigkeit (pp. 193–213). Frankfurt, Germany: Campus. Furnham, A. (1995). The just world, charitable giving and attitudes to disability. Personality & Individual Differences, 19, 577–583. Furnham, A. (2003). Belief in a just world: Research progress over the past decade. Personality and Individual Differences, 34, 795–817. Furnham, A., & Procter, E. (1989). Belief in a just world: Review and critique of the individual difference literature. British Journal of Social Psychology, 28, 365–384. Grossi, G. (1999). Coping and emotional distress in a sample of Swedish unemployed. Scandinavian Journal of Psychology, 40, 157–165. Hafer, C.L. (2000). Investment in long-term goals and commitment to just means drive the need to believe in a just world. Personality and Social Psychology Bulletin, 26, 1059–1073. Hafer, C.L., & Correy, B.L. (1999). Mediators of the relation of beliefs in a just world and emotional responses to negative outcomes. Social Justice Research, 12, 189–204. Häfner, H. (1990). Arbeitslosigkeit: Ursache von Krankheit und Sterberisiken [Unemployment: Cause of disease and mortality risks]. Zeitschrift für Klinische Psychologie, 1, 1–17. Hammarström, A., & Janlert, U. (2002). Early unemployment can contribute to adult health problems: Results from a longitudinal study of school leavers. Journal of Epidemiology and Community Health, 56, 624–630. 48
Justice Concerns and Mental Health During Unemployment Hautzinger, M., & Bailer, M. (1993). Allgemeine Depressionsskala (ADS) [General Depression Scale]. Weinheim, Germany: Beltz-Test. Hollenbeck, J.R., & Klein, H.J. (1987). Goal commitment and the goal-setting process: Problems, prospects, and proposals for future research. Journal of Applied Psychology, 72, 212–20. Kiecolt-Glaser, J.K., & Williams, D.A. (1987). Self-blame, compliance, and distress among burn patients. Journal of Personality and Social Psychology, 53, 187–193. Kieselbach, T. (1995). Arbeitslosigkeit und Ungerechtigkeitserleben: Perspektiven eines zukünftigen Umgangs mit beweglichen Transitionsprozessen [Unemployment and experience of unfairness: Perspectives for future dealing with transition processes]. Berichte aus dem Zentrum für Gerechtigkeitsforschung an der Universität Potsdam, Band 2, Eröffnungssymposium, S.86–109. Lerner, M.J. (1965). Evaluation of performance as a function of performer’s reward and attractiveness. Journal of Personality and Social Psychology, 1, 355–360. Lerner, M.J. (1980). The Belief in a Just World: A fundamental delusion. New York: Plenum. Lerner, M.J. (2002). Pursuing the justice motive. In M. Ross & D.T. Miller (Eds.), The justice motive in everyday life. Cambridge, UK: Cambridge University Press. Lerner, M.J., & Miller, D.T. (1978). Just world research and the attribution process: Looking back and ahead. Psychological Bulletin, 85, 1030–1051. Lerner, M.J., & Somers, D.G. (1992). Employees’ reactions to an anticipated plant closure: The influence of positive illusions. In L. Montada & S.H. Filipp (Eds.), Life crises and experiences of loss in adulthood (pp. 229–254). Hillsdale, NJ: LEA. Lipkus, I.M., Dalbert, C., & Siegler, I.C. (1996). The importance of distinguishing the Belief in a Just World for self versus for others: Implications for psychological wellbeing. Personality and Social Psychology Bulletin, 22, 666–677. Lipkus, I.M., & Siegler, I.C. (1993). The Belief In A Just World and perceptions of discrimination. Journal of Psychology, 127, 465–474. Lupfer, M.B., Doan, K., & Houston, D.A. (1998). Explaining unfair and fair outcomes: The therapeutic value of attributional analysis. British Journal of Social Psychology, 37, 495–511. Martella, D., & Maass, A. (2000). Unemployment and life satisfaction: The moderating role of time structure and collectivism. Journal of Applied Social Psychology, 30, 1095–1108. Metz-Göckel, H., & Leffelsend, S. (2001). Motivationale Determinanten der Zielbindung im Studium [Motivational determinants of goal commitment at university]. Zeitschrift für Psychologie, 209, 153–73. Otto, K., & Dalbert, C. (in press). Belief in a just world as a resource for different types of young prisoners. In C. Dalbert & H. Sallay (Eds.), The justice motive in adolescence and young adulthood: Origins and consequences. London: Routledge. Radloff, L.S. (1977). The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401. Rim, Y. (1986). Coping-Strategien, der Glaube an eine gerechte Welt, Konservatismus, Werteinstellungen und das Konfluenz-Modell [Coping styles, Belief In A Just World, conservatism, value orientation, and the confluence model]. Schweizerische Zeitschrift für Psychologie, 45, 17–27. Rogner, O., Frey, D., & Havemann, D. (1987). Der Genesungsverlauf von Unfallpatienten aus kognitionspsychologischer Sicht [The course of convalescence of victims of accidents from a cognitive psychology perspective]. Zeitschrift für Klinische Psychologie, 16, 11–28. Rubin, Z., & Peplau, L.A. (1975). Who believes in a just world? Journal of Social Issues, 31(3), 65–89. 49
Unemployment and Health ■ Health Effects Sallay, H. (in press). Entering the job market: Belief in a Just World, fairness and wellbeing of graduating students. In C. Dalbert & H. Sallay (Eds.), The justice motive in adolescence and young adulthood: Origins and consequences. London: Routledge. Scherer, K.R. (1997). The role of culture in emotion-antecedent appraisal. Journal of Personality and Social Psychology, 73, 902–922. Schill, T., Beyler, J., & Morales, J. (1992). The role of just world belief and anger issues in self-defeating personality. Psychological Reports, 70, 595–598. Silver, R.L., Boon, C., & Stones, M.H. (1983). Searching for meaning in misfortune, making sense of incest. Journal of Social Issues, 39, 81–102. Smith, C.A., & Ellsworth, P.C. (1985). Patterns of cognitive appraisal in emotion. Journal of Personality and Social Psychology, 48, 813–838. Tomaka, J., & Blascovich, J. (1994). Effects of justice beliefs on cognitive, psychological, and behavioral responses to potential stress. Journal of Personality and Social Psychology, 67, 732–740. Witenberg, S.H., Blanchard, E.B., Suls, J., Tennen, H., McCoy, G., & McGoldrick, M.D. (1983). Perceptions of control and causality as predictors of compliance and coping in hemodialysis. Basic and Applied Social Psychology, 4, 319–336. Zuckerman, M., & Gerbasi, K.C. (1977). Belief in a just world and trust. Journal of Research in Personality, 11, 306–317.
50
Chapter
4
Quantitative Reviews in Psychological Unemployment Research: An Overview Karsten Ingmar Paul and Klaus Moser
As early as 1911, Rowntree and Lasker observed that the effects of unemployment were ‘demoralising’ and ‘deleterious … on a man’s character’ (p. 309). If we take this statement as the beginning, psychological unemployment research is nearly 100 years old. During this century, psychologists, sociologists and researchers from other disciplines have accumulated vast amounts of knowledge concerning the psychological aspects of unemployment. However, this knowledge is spread over thousands of journal articles, books, dissertations, masters theses, diploma theses, unpublished institute reports and other reports. As a result, it is difficult for any single person, even the best expert, to review the information. We can conclude, therefore, that unemployment researchers are confronted with the same problem that concerns all social and medical scientific disciplines today, which is that ‘the individual scientist is … overloaded with scientific information and [can] no longer keep up with and assimilate all the information being produced’ (Garvey & Griffith, 1971, p. 350). As a result, literature reviews are of growing importance within the scientific process, and there is increasing scrutiny of the methodological quality of these papers (cf. Cooper & Hedges, 1994). Within unemployment research, there are several reviews that summarise and integrate the results of nearly a century of scientific efforts concerning the psychological effects of being involuntarily out of work (e.g., Catalano, 1991; Dooley, Fielding, & Levi, 1996; Eisenberg & Lazarsfeld, 1938; Ezzy, 1993; Frese & Mohr, 1978; Fryer & Payne, 1986; Hammarström, 1994; Kasl, Rodriguez, & Lasch, 1998; Lennon, 1999; McKee-Ryan & Kinicki, 2002; Olafsson & Svensson, 1986; Schultz-Gambard & Balz, 1988; Warr, Jackson, & Banks, 1988; Winefield, 1995). Most of these reviews have been written in the traditional narrative form, however, and, while they have provided valuable insights into the psychological effects of unemployment, 51
Unemployment and Health ■ Health Effects
they have also used varying standards of problem formulation, literature search, data extraction and evaluation, data analysis, and interpretation and presentation of results (cf. Cooper & Hedges, 1994). Importantly, they have not always provided information on the size of the effects under consideration, which makes it difficult to gauge the strength and practical importance of the findings they have presented. Meta-analyses are based on a quantitative technique with well-formulated standards and guidelines. They can be described as ‘quantitative summaries of research domains that describe the typical strength of the effect or phenomenon, its variability, its statistical significance, and the nature of the moderator variables from which one can predict the relative strength of the effect or phenomenon’ (Rosenthal, 1995, p. 183). Such quantitative summaries are rare in unemployment research. The purpose of the present article is to give a short overview of those that do exist in this research field. Prior to the appearance of the first meta-analyses of unemployment research (e.g., Foster, 1991), several researchers (e.g., Catalano, 1991; Dooley & Catalano, 1980; Frese & Mohr, 1978; Hammarström, 1994; O’Brien, 1986; Platt, 1985; Warr, 1985; Warr, Jackson, & Banks, 1988; Warr & Parry, 1982) attempted to introduce a systematic approach to the review of research in this field, by presenting important design characteristics and/or the main results of the reviewed studies in tabular form. However, these authors stopped short of integrating the findings from the primary studies by computing an overall mean effect size. Fryer and Payne (1986) took the next step and gave an estimation of the strength of the negative effects of unemployment on mental health by computing the median of exemplary results from selected primary studies. The estimation of the correlation between unemployment and distress was r = .34 (the median from five correlations reported by Banks and Jackson, 1982). It was not until 1991 that researchers started to use modern meta-analytic techniques in summarising and integrating research results in the field of psychological unemployment research. Foster (1991) began by undertaking an ‘exploratory meta-analysis’, in which he aimed to ‘take a quick and dirty look at what a cross-study effect size might be’ (p. 159). He integrated 22 effect sizes from 10 recently published primary studies at the time and computed an average (unweighted) effect size of d = .19. This is equivalent to a correlation of r = .09,1 which is surprisingly small and seems to contradict the conclusions of earlier narrative reviewers, namely that being without work has a damaging effect on mental health. However, Foster’s (1991) effect size estimation may have been artificially reduced by his methods of operationalising unemployment and mental health. While the usual design in unemployment research consists of comparisons between groups of employed and unemployed persons, Foster’s (1991) meta-analysis also included studies that correlated job insecurity among employed persons with distress levels, as well as studies that compared spouses of employed and 52
Quantitative Reviews in Psychological Unemployment Research
unemployed persons with regard to their levels of mental health. Furthermore, Foster (1991) included infrequently used measures of psychological health, such as number of ‘bed days’ (p. 160). It is possible that these methods of operationalising unemployment and mental health may have resulted in an underestimation of the true effect size.2 Temple, Fillmore, Hartka, Johnstone, Leino and Motoyoshi (1991) reviewed the effects of employment status on alcohol consumption, an aspect of behaviour that is relevant to psychological unemployment research because of its connection with public health. The authors included 12 prospective longitudinal studies in their meta-analysis and reported some unexpected results. For example, they found that after controlling for baseline alcohol consumption and education, there was no significant effect of status change (becoming unemployed or becoming employed) on subsequent alcohol consumption. They also reported that continuous unemployment was significantly associated with alcohol consumption in two subgroups (older men and younger women), with continuously unemployed persons drinking less than continuously employed persons. This finding seems to contradict common assumptions regarding the effect of unemployment on drinking habits. However it should be noted that what Temple et al. (1991) termed ‘unemployment’ would be more accurately termed ‘nonemployment’, as it included both unemployment and being out of the labour force. Furthermore, it is not clear what percentage of students, homemakers, retirees and other persons out of the labour force were included within these nonemployed groups. As a result, it is difficult to know how relevant the studies that were summarised by Temple et al. (1991) really are for unemployment research. Another problem with this meta-analysis is the use of b coefficients from multiple regression analyses as measures of effect size. These b coefficients are difficult to interpret within a meta-analysis as they are influenced by the other variables included in the respective regression analysis and therefore cannot be generalised to contexts in which the additional variables are different. Thus, they are not an ideal means to reach the main goal of meta-analysis, which is to give a generalisable estimation of an effect size. Murphy and Athanasou’s (1999) review of the effects of unemployment on mental health was especially concerned with problems of causality. These authors included 16 longitudinal studies from 1986 to 1995, nine of which could be used in the meta-analysis. The authors concluded in favour of direct causation of mental health problems by unemployment: ‘job loss provokes psychological distress, rather than the reverse’ (p. 89). They presented effect sizes of d = 0.36 for moving from employment to unemployment and of d = 0.54 for moving in the opposite direction. This means that losing a job was associated with deterioration in mental health, whereas finding a new job was associated with a strong improvement in mental health. These longitudinal effect sizes appear to provide strong support for 53
Unemployment and Health ■ Health Effects
the causation hypothesis, rather than for the alternative selection hypothesis, which proposes that the association between unemployment and reduced mental health arises because persons with preexisting mental health problems become unemployed. However, causation and selection may not be mutually exclusive processes, and it is possible that both contribute to the accumulation of mental health effects among the unemployed. Unfortunately, Murphy and Athanasou (1999) did not test the selection hypothesis using meta-analysis. In addition to computing overall effect sizes, Murphy and Athanasou (1999) conducted moderator analyses for nationality (Anglo-Saxon vs. European), age (young vs. adult), gender (male vs. mixed gender) and type of measurement procedure (General Health Questionnaire vs. other instrument). No significant moderator effects were found, but this may be the result of the very small number of primary studies in the Murphy and Athanasou (1999) meta-analysis. Kanfer, Wanberg and Kantrowitz (2001) summarised 73 primary studies of the antecedents and outcomes of job search. They reported a mean correlation (corrected for artifacts) of r = .28 between the job search behaviour a person exhibits (measured as intensity or effort) and the number of job offers a person gets. The correlation between job search behaviour and employment status (i.e., getting a job or not) was r = .21 and the mean correlation between job search behaviour and the duration of the job search was r = –.14. This suggests that job seekers who engaged in more job search behaviour were more likely to obtain new employment. However, these correlations are only of small to medium size, which suggests that the influence of individual behaviour on labour market outcomes is limited and that success in searching for a job depends to a large amount on other factors. In regard to the potential antecedents of job search behaviour, Kanfer et al. (2001) found that the influence of demographic variables is only weak, with correlations ranging from r = .05 to r = .12 for age (younger people reporting more search behaviour than older people), education (welleducated people reporting more search behaviour than others), gender (males reporting more search behaviour than women) and for ethnicity (Caucasian people reporting more search behaviour than others). On the other hand, the authors reported stronger correlations between job search and personality variables, especially extraversion (r = .46), conscientiousness (r = .38), openness to experience (r = .27) and self esteem (r = .25; persons with high scores in these variables reporting more search behaviour than persons with low scores). Job search behaviour was also influenced by employment commitment (r = .29), and by contextual variables such as financial need (r = .21) and social support (r = .24). On the one hand, the strong correlations that were found between psychological variables and job search behaviour indicate that this behaviour may be susceptible to psy54
Quantitative Reviews in Psychological Unemployment Research
chological interventions. On the other hand, however, the weak to medium correlations between job search behaviour and reemployment success suggest that interventions that merely target job search behaviour (without also addressing other important variables, such as labour market conditions) might meet with only limited success. Paul and Moser’s meta-analytic project (2004), which is still in progress and will be reported elsewhere in detail, is concerned with the effects of unemployment on mental health. In this study we have identified and incorporated more than 160 cross-sectional and longitudinal primary studies that matched our inclusion criteria. Because of the very large database, this meta-analysis will yield stable and highly reliable estimations of the influence of unemployment on mental health and will be highly representative of the field of psychological unemployment research. According to Paul and Moser, the cross-sectional association between unemployment and mental health is highly significant. This statement is true for very different indicators of mental health, such as depression, anxiety, somatic complaints, mixed symptoms of distress, subjective wellbeing and self-esteem. The size of this effect varies considerably across the different variables. It is small or very small for anxiety and somatic complaints, and it is of small-to-medium size for depression, mixed symptoms of distress, self-esteem and wellbeing. Paul and Moser were able to identify significant moderator effects for gender, with males showing stronger effect sizes than females, and for geographic location, with British studies showing stronger effects than studies from other parts of the world. As expected, the study also found that the long-term unemployed felt worse off than the short-term unemployed. Employment commitment also had a moderating effect, as it positively correlated with distress symptoms in unemployed people, but negatively correlated for employed people. There was also a tendency for studies of persons from blue-collar occupations to yield larger effect sizes than studies of persons from white collar or professional occupations. Ethnicity and age seem to have no moderating influence on the relationship between unemployment and mental health. Paul and Moser’s results are similar to those of Murphy and Athanasou (1999) in that both research teams found a significant deterioration in mental health when persons change from employment to unemployment and a significant improvement when unemployed people found new employment. The positive effect of finding new employment was stronger than the negative effect of losing a job. On the other hand, according to Paul and Moser, selection effects on the way into and out of unemployment can also be found. They reported that persons who will lose their job in the future show more symptoms of distress when still in employment than continuously employed persons. Unemployed persons who subsequently find employment have better mental health than continuously unemployed persons. These selection 55
Unemployment and Health ■ Health Effects
effects are significant but small, indicating that they may be of only minor importance in explaining the cross-sectional correlation between unemployment and distress. The impact of unemployment on distress is considerably stronger than the impact of distress on unemployment. A comparison of the results of the quantitative reviews revealed more commonalities than discrepancies. All authors agreed that there is a significant relationship between unemployment and mental health. However, the magnitude of the estimated effect sizes varied considerably, ranging from a small effect size estimation in the case of Foster (1991) to a medium-tolarge effect size estimation in the case of Fryer and Payne (1986). However, closer inspection revealed that both extreme values could be the result of the reviews integrating very small numbers of primary studies and employing unweighted means or simple medians. The meta-analyses of Paul and Moser (2004) and Murphy and Athanasou (1999) integrated larger numbers of primary studies and used a technique that is recommended by statistical authorities (Hedges & Olkin, 1985; Hunter & Schmidt, 1990; Rosenthal, 1995) for purposes of research synthesis, namely sample size–weighted means. Both Murphy and Athanasou (1999) and Paul and Moser (2004) found very similar effects of small to medium size for the relationship between unemployment and mental health. This effect size for the relationship between unemployment and mental health can be compared with results from other research fields. For example, the correlation between mental health and employment is stronger than the correlation between smoking and lung cancer, which has been reported to be r = .17 (Ondersma, Chaffin, & Berliner, 1999, as cited in Dallam, Glaeves, Cepeda-Benito, Silberg, Kraemer, & Spiegel, 2001). The effect of aspirin use on the rate of heart attacks is only r = .03, but this effect size translates into nearly half as many heart attacks in the aspirinuser group as in the nonuser group (Steering Committee of the Physicians’ Health Study Research Group, 1988, as cited in Rosenthal, 1994). This demonstrates that small effect sizes can be of great practical significance. In cases of a mass phenomenon like unemployment, a small to medium effect size means millions of suffering people. Some meta-analytic results are in agreement with the conclusions drawn by narrative reviewers, for example the moderating effect of length of unemployment reported by Paul and Moser (2004). However, others are not. For example, since Eisenberg and Lazarsfeld (1938), there has been strong agreement in the research field that people in middle age suffer most from unemployment, while young and older people suffer less. However, Murphy and Athanasou (1999) and Paul and Moser (2004) were not able to identify a moderating effect of age. It might also be useful to consider the different moderator concepts that have been used within unemployment research. Some researchers (e.g., Warr & Jackson, 1984; Winefield, 1995) seem to have conceptualised moderator 56
Quantitative Reviews in Psychological Unemployment Research
effects as a difference between two groups of unemployed persons (e.g., unemployed adults and unemployed youth). However, others such as Paul and Moser (2004) have conceptualised moderator effects as the differences in the effect size between two groups of persons (e.g., the difference in mental health between adult unemployed persons and adult employed persons compared with the difference between young unemployed persons and young employed persons). These different ways of conceptualising the moderator concept in the literature should be kept in mind when comparing results concerning moderator effects in unemployment research. It can be seen that applying quantitative review techniques to psychological unemployment research leads to new questions. For example, why is the effect size that goes along with job loss considerably smaller than the effect size that goes along with finding new employment? Why are the negative effects of unemployment found in Great Britain stronger than those found in continental Europe, Australia and the United States? How can the larger effect sizes for men compared to women be explained? Do unemployed men suffer more than unemployed women or is the experience of men and women similar? Finally, if job search behaviour is only weakly related to reemployment success, and the effect of demographic characteristics (race, age, gender and education) on reemployment is almost negligible (Kanfer et al., 2001), what else is it that determines reemployment success among the unemployed? In summary, we can conclude that applying quantitative methods of research synthesis to the field of unemployment research leads to interesting answers, as well as to new possibilities for research.
Endnotes 1
2
The effect sizes r and d can be easily converted to each other. Within the normal range of effect sizes that can be found within psychology, d is about twice as large as r. The formulas (taken from Rosenthal, 1994) are: d2 d = 2r / √1 – r2; r = (Rosenthal, 1994). d2 + 4 This short critique shows one of the advantages of meta-analysis: The reviewer has to carefully describe every step she or he does within the review process, giving the reader the possibility of identifying methodological problems and questionable decisions. Narrative reviewing may make it harder to detect problems and errors.
√
References Banks, M.H., & Jackson, P.R. (1982). Unemployment and risk of minor psychiatric disorder in young people: Cross-sectional and longitudinal evidence. Psychological Medicine, 12, 789–798. Catalano, R.A. (1991). The health effects of economic insecurity. American Journal of Public Health, 81, 1148–1152. Cooper, H., & Hedges, L.V. (1994). Research synthesis as a scientific enterprise. In H. Cooper & L.V. Hedges (Eds.), The handbook of research synthesis (pp. 3–14). New York: Russell Sage Foundation. 57
Unemployment and Health ■ Health Effects Dallam, S.J., Glaeves, D.H., Cepeda-Benito, A., Silberg, J.L., Kraemer, H.C., & Spiegel, D. (2001). The effects of child sexual abuse: Comment on Rind, Tromovitch, and Bauserman (1998). Psychological Bulletin, 127, 715–733 Dooley, D., Fielding, J., & Levi, L. (1996). Health and unemployment. Annual Review of Public Health, 17, 449–465. Dooley, D., & Catalano, R. (1980). Economic change as a cause of behavioral disorder. Psychological Bulletin, 87, 450–468. Eisenberg, P., & Lazarsfeld, P.F. (1938). The psychological effects of unemployment. Psychological Bulletin, 35, 358–390. Ezzy, D. (1993). Unemployment and mental health: A critical review. Social Science and Medicine, 37, 41–52. Foster, C. (1991). ‘Because no man hath hired us’: A study of the impact of unemployment on mental health through the analysis of linear structural relationships. Unpublished doctoral dissertation, Brandeis University, MA. Frese, M., & Mohr, G. (1978). Die psychopathologischen Folgen des Entzugs von Arbeit: Der Fall Arbeitslosigkeit [The psychopathological consequences of the loss of work. The case of unemployment]. In M. Frese, S. Greif, & N.B. Brenner (Eds.), Industrielle Psychopathologie (pp. 282–320). Bern, Germany: Hans Huber. Fryer, D., & Payne, R. (1986). Being unemployed: A review of the literature on the psychological experience of unemployment. In C.L. Cooper & I.T. Robertson (Eds.), International review of industrial and organizational psychology (pp. 235–278). Chichester, UK: Wiley. Garvey, W.D., & Griffith, B.C. (1971). Scientific communication: Its role in the conduct of research and the creation of knowledge. American Psychologist, 26, 349–361. Hammarström, A. (1994). Health consequences of youth unemployment: Review from a gender perspective. Social Science and Medicine, 38, 699–709. Hedges, L.V., & Olkin, I. (1985). Statistical methods for meta-analysis. Orlando, FL: Academic Press. Hunter, J.E., & Schmidt, F.L. (1990). Methods of meta-analysis: Correcting error and bias in research findings. New York: Sage. Kanfer, R., Wanberg, C.R., & Kantrowitz, T.M. (2001). Job search and employment: A personality-motivational analysis and meta-analytic review. Journal of Applied Psychology, 86, 837–855. Kasl, S.V., Rodriguez, E., & Lasch, K.E. (1998). The impact of unemployment on health and well-being. In B.P. Dohrenwend (Ed.), Adversity, stress, and psychopathology (pp. 111–131). Oxford, UK: Oxford University Press. Lennon, M.C. (1999). Work and unemployment as stressors. In A.V. Horwitz & T.L. Scheid (Eds.), A handbook for the study of mental health (pp. 284–294). Cambridge, UK: Cambridge University Press. McKee-Ryan, F.M., & Kinicki, A.J. (2002). Coping with job loss: A life-facet perspective. In C.L. Cooper & I.T. Robertson (Eds.), International review of industrial and organizational psychology (vol. 17; pp. 1–29). Chichester, UK: Wiley. Murphy, G.C., & Athanasou, J.A. (1999). The effect of unemployment on mental health. Journal of Occupational and Organizational Psychology, 72, 83–99. O’Brien, G.E. (1986). Psychology of work and unemployment. Chichester, UK: Wiley. Olafsson, O., & Svensson, P.G. (1986). Unemployment-related lifestyle changes and health disturbances in adolescents and children in the western countries. Social Science and Medicine, 22, 1105–1113. Ondersma, S.J., Chaffin, M, & Berliner (1999). Comments on Rind et al. meta-analysis controversy. APSAC Advisor, 12, 2–5.
58
Quantitative Reviews in Psychological Unemployment Research Paul, K., & Moser, K. (2004). Unemployment and mental health: Meta-analyses. Manuscript in preparation. Platt, S. (1985). Suicidal behaviour and unemployment: A literature review. In G. Westcott, P.G. Svensson, & H.F.K. Zöllner (Eds.), Health policy implications of unemployment (pp. 87–132). Copenhagen, Denmark: World Health Organization. Rosenthal, R. (1994). Parametric measures of effect size. In H. Cooper & L.V. Hedges (Eds.), The handbook of research synthesis (pp. 231–244). New York: Russell Sage Foundation. Rosenthal, R. (1995). Writing meta-analytic reviews. Psychological Bulletin, 118, 183–192. Rowntree, B.S., & Lasker, B. (1911). Unemployment: A social study. London: Macmillan. Schultz-Gambard, J., & Balz, H.J. (1988). Schicksal arbeitslos: Ein Überblick über die Ergebnisse angewandt-sozialpsychologisxcher Forschung zu den Folgen und Einflußfaktoren bei Arbeitslosigkeit [Living as an unemployed person: A review of results of applied social psychological research concerning consequences and influencing factors associated with unemployment]. Gruppendynamik, 19, 239–273. Steering Committee of the Physicians’ Health Study Research Group. (1988). Preliminary report: Findings from the aspirin component of the ongoing physicians’ health study. New England Journal of Medicine, 318, 262–264. Temple, M.T., Fillmore, K.M., Hartka, E., Johnstone, B., Leino, E.V., & Motoyoshi, M. (1991). A meta-analysis of change in marital and employment status as predictors of alcohol consumption on a typical occasion. British Journal of Addiction, 86, 1269–1281. Warr, P. (1985). Twelve questions about unemployment and health. In R. Bryan, R. Finnegan, & G. Dallie (Eds.), New approaches to economic life (pp. 302–318). Manchester, UK: Manchester University Press. Warr, P.B., & Jackson, P. (1984). Men without jobs: Some correlates of age and length of unemployment. Journal of Occupational Psychology, 57, 77–85. Warr, P.B., Jackson, P., & Banks, M. (1988). Unemployment and mental health: Some British studies. Journal of Social Issues, 44, 47–68. Warr, P.B., & Parry, G. (1982). Paid employment and women’s psychological well-being. Psychological Bulletin, 91, 498–516. Winefield, A.H. (1995). Unemployment: Its psychological costs. In C.L. Cooper & I.T. Robertson (Eds.), International review of industrial and organizational psychology (vol. 10; pp. 169–212). Chichester, UK: Wiley.
59
Unemployment and Health ■ Health Effects
60
Chapter
5
Unemployment, Secure Employment and Insecure Employment: Differences in Self-Reported Ill Health Bengt Starrin and Staffan Janson
The dichotomy between work and unemployment has long been the dominating principle for division of labour-market status, and it has frequently been used to study health and welfare. The overall conclusion of metaanalyses has been that unemployment has a modest negative effect on mental health (Hallsten, 1997). However, it has been argued that the changes that have taken place in the labour market during the last 10 to 15 years might have led to the emergence of new categories among the employed, which supplement the traditional dichotomy between the employed and the unemployed. This is because with increasing flexibility and globalisation in the labour market, permanent employment is increasingly being replaced by time-restricted employment. It has been argued that these changes might increase the risk for psychological ill health of different kinds (Fryer, 2000; Hallsten & Isaksson, 2000). Time-limited employment represented the fastest growing form of employment during the 1990s, not only in Sweden but also in other parts of the western world. There is, however, great variation among the types of fixed-term contracts offered to employees. Those employed on special projects usually enjoy good working conditions, including good salaries and prospects of personal development. However, conditions are very different for employees on demand who are given work at short notice (socalled casual employees), and for whom the psychosocial working conditions — for example, to control and influence the work situation — are often unsatisfactory (Aronsson, 1999). In a survey on working conditions within the European Union (EU) that comprised 21,703 workers, it was found that 10% of employees had fixedterm contracts and 2% had temporary agency contracts.1 However, of those who had been employed for less than a year only 49% had open-ended 61
Unemployment and Health ■ Health Effects
contracts. The survey also showed that those who were temporarily employed more often had intense, fast-paced jobs than those who occupied permanent jobs (Third European Survey on Working Conditions, 2000). One of the consequences for employees of the current flexible labour market is that jobs have become more insecure because of fear of job loss. Studies have shown that job insecurity among employees increases the risk of ill health. Wichert (2002) found in her review of the literature on the relationships between job insecurity and decreased psychological wellbeing, such as anxiety and depression, that there is consistent international evidence for the detrimental effects of the experience of job insecurity on psychological ill health. Kinnunen and Natti (1994) studied the relationship between job insecurity and psychosomatic symptoms. Their results showed that those who experienced a high level of insecurity, defined as a threat of being fired, were also more troubled by psychosomatic symptoms than those who experienced a low level of insecurity. Burchell (1994) found that job insecurity was associated with poorer psychological wellbeing. In their review of research in the field, Platt, Pavis and Akram (1999) concluded that job insecurity may lead to worse self-rated physical health and an increase in some clinical symptoms. Extensive research has shown that, regardless of how secure the job is, certain physical and psychosocial conditions, as well as the salary, have a significant effect on people’s health. Concerning the psychosocial environment, research has indicated that jobs which allow freedom and promote personal development are more advantageous for health than are restrained and monotonous jobs (Griffin, Fuhrer, Stansfeld, & Marmot, 2002; Karasek, 1998; Karasek & Theorell, 1990). Surveys have shown that since the economic recession which affected Sweden during the early 1990s the psychological health of much of the population has deteriorated (Åberg, Strandh, Nordenmark, & Bolinder, 1997; National Public Health Report, 2001). Fatigue, amnesia and minor psychological symptoms such as anxiety have increased both among those who work and among the population as a whole (National Public Health Report, 2001). The pace of jobs has increased, as has work-related stress. So-called ‘hectic’ jobs (e.g., with high work pressure, high workload) became more numerous during the 1990s, especially among women. Anxiety and fatigue have increased substantially, not only among the gainfully employed, but also among the unemployed. Both working environments and wellbeing have deteriorated, especially among municipal and county employees. This is particularly true for teachers, and employees in nursing and health-care (National Public Health Report, 2001). Åberg et al. (1997) showed that anxiety levels were relatively low during the economic boom of the late 1980s but rose during the early 1990s in both the employed and unemployed equally. 62
Unemployment, Secure Employment and Insecure Employment
The issue that is relevant for the work–unemployment dichotomy is the prevalence of psychosocial ill health, not only among the unemployed, but also among the employed, depending upon the perceived quality of their working conditions. In a longitudinal study, Graetz (1993) examined the effects of unemployment and work on psychosocial health. The conclusions were that the health consequences of employment and unemployment may be directly contingent upon quality of work. Levels of health risk were highest among dissatisfied workers and lowest among satisfied workers. Between these two extremes were (1) employed people who were neither satisfied nor dissatisfied with their jobs; (2) the unemployed (irrespective of how long they had been unemployed); and (3) students. These results suggest that what happens in the workplace may have a stronger impact on a person’s health than does mere success in finding and keeping a job. Burchell’s (1994) conclusion from his survey on the psychosocial health effects of one’s labour market position was that depression in unemployed men might not be reduced by reemployment in an insecure, rather than a secure, job. In light of the social changes that have occurred during the last 10 to 15 years, such as increased globalisation of the economy and increased flexibility of the labour market, there are reasons to reconsider the usefulness of the employment–unemployment dichotomy (Fryer, 2000). Benach, Benavides, Platt, Diez-Roux and Muntaner (2000) argue that future research needs to move away from investigations that compare the health and wellbeing in unemployed and employed persons towards an analysis of unemployed and underemployed workers compared with workers with stable jobs. Future research should consider the impact of different kinds of working conditions and social environments on the relationship between flexible employment and health (Benach et al., 2000). In the present study, special attention was paid to psychosocial aspects of the work environment as well as to employment conditions. However, economic dimensions of work and unemployment were also considered, as economic conditions also have an impact on mental health and wellbeing. Several studies show that economic stress covaries with psychiatric ill health in both the employed (Pearlin & Lieberman, 1979) and the unemployed, alike (see, e.g., Forsberg & Kalander-Blomqvist, 2000; Kessler, Turner & House, 1987; Starrin, Rantakeisu, Vinokur, Price, & Caplan, 1996; Whelan, 1992). The objective of the present study was to compare the mental health of five groups. Four of these comprised people employed in jobs with varying degrees of security and perceived psychosocial work environments (‘secure–empowered’, ‘secure–disempowered’, ‘insecure–empowered’ and ‘insecure–disempowered’ jobs), and the fifth group was of unemployed people.
63
Unemployment and Health ■ Health Effects
Method This study was originally based on a survey involving 16 municipalities in central Sweden. It was conducted during the period March to May 2000. The sample consisted of 11,900 people aged 18 to 79, and was stratified such that 700 people were randomly selected from each municipality, except for the principal town, from where 1400 people were included. The response rate among those aged between 18 and 64 (i.e., of working age) was 70% (n = 5572). From this group, five subgroups were selected and categorised according to the description below.
Instruments Employment and Unemployment The four employment groups were constructed as follows: Respondents were asked ‘How worried are you about losing your job within the next year?’ Response alternatives were: very worried, fairly worried, not very worried and not worried at all. Two groups were formed based on the responses to this question. Respondents who answered not worried at all were classified as having secure jobs. Those who answered very worried or fairly worried were classified as having insecure jobs. The employment types ‘empowered’ and ‘disempowered’ were formed according to responses concerning the psychosocial work environment. Respondents were presented with the following statements: ‘I can decide my own working pace’, ‘I can take a break in my job when I want to’, ‘I can influence how my job is to be carried out’ and ‘I have the possibility of developing and learning new things’. Response alternatives were strongly agree, partly agree, partly disagree and strongly disagree. By using principal components analysis, a weighted measure was established and scores were dichotomised to form two categories of respondent: those in ‘empowered’ jobs (high scores) and those in ‘disempowered’ jobs (low scores). Combining the security and empowerment dimensions allowed employed respondents to be categorised further, as ‘secure–empowered’, ‘secure–disempowered’, ‘insecure–empowered’ and ‘insecure–disempowered.’ A fifth group of respondent comprised those who were unemployed, either without any kind of employment at all, or employed through some kind of labour market policy measure. The group under study consisted finally of 2245 persons aged 18 to 64. As can be seen in Table 1, approximately 35% had a secure–empowered job and 31% a secure–disempowered job. The smallest group comprised those who had secure–disempowered jobs (8%). Fourteen per cent had insecure–disempowered jobs and 12% were unemployed. Operationalisation of Ill Health Chronic psychosomatic problems. In the survey, respondents were asked whether during the last three months they had been troubled by, among other symptoms, stomach-ache, headache (not migraine) or sleeping 64
Unemployment, Secure Employment and Insecure Employment Table 1 Employment Category by Gender, Age and Education Employment Category Secure– empowered
Secure– disempowered
Insecure– empowered
Insecuredisempowered
Unemployed
%
n
%
n
%
n
%
n
%
n
Total
35
796
31
705
8
169
14
313
12
280
Gender Women Men
27 44
306 488
39 24
432 265
6 9
70 98
17 11
189 123
11 13
126 148
Age (in years) 18–24 25–34 35–44 45–54 55–64
19 35 37 38 36
29 191 185 228 163
34 28 30 33 33
54 149 153 198 151
5 8 9 8 5
8 45 43 49 24
22 14 14 14 11
34 74 72 83 50
20 15 10 8 15
32 81 51 47 69
Education Preupper secondary Upper secondary Upper secondary Postupper secondary Postupper secondary
32 29 34 46 51
143 232 150 142 125
26 31 32 34 37
118 248 140 105 91
7 9 7 7 5
32 73 32 20 12
14 20 14 7 3
64 159 60 21 8
21 12 13 6 4
95 100 57 19 9
problems. The response alternatives were: 1 = Have not been bothered, 2 = Occasionally, 3 = A number of times and 4 = Practically all the time. To be categorised as having one or more chronic psychosomatic problems, respondents had to have been troubled practically all the time by at least one symptom. Psychiatric ill health. To assess psychiatric ill health the 12-item version of the General Health Questionnaire was used (GHQ-12; Goldberg & Williams, 1988). Scoring for each item could range from 0 to 3 to give a possible total of 36. A sum above 18 was used to determine psychiatric ill health (e.g., Aronsson, Dallner, & Lundh, 2000). Depression. Depression was measured with the following question: ‘During the last three months, have you been troubled by depression?’ The response alternatives (1 = Have not been bothered, 2 = Occasionally, 3 = A number of times and 4 = Practically all the time) were combined and used as indicator of depression. Anxiety. Anxiety was measured with the following question: ‘During the last three months, have you been troubled by anxiety?’ The response alternatives were 1 = Have not been bothered, 2 = Occasionally, 3 = A number of times and 4 = Practically all the time. The categories A number of times and Practically all the time were combined and used as an indicator of anxiety. 65
Unemployment and Health ■ Health Effects
Social Background Variables In the survey, gender, age, education and type of household were used as social background variables. Age. Five categories were formed for age: (1) 18 to 24 years; (2) 25 to 34 years; (3) 35 to 44 years; (4) 45 to 54 years; and (5) 55 to 64 years. Education. Education was categorised as follows: (1) preupper secondary school, not more than 9 years of schooling; (2) upper secondary school, 10 to 11 years; (3) upper secondary school, 12 years; (4) postupper secondary school, 13 to 15 years; and (5) postupper secondary school, more than 15 years. Household. Household arrangements were classified according to presence/absence of a partner, and presence/absence of a child under 18 years of age. Accordingly, they were: (1) single without children under 18 years; (2) single with children under 18 years; (3) couple without children under 18 years; and (4) couple with children under 18 years. Economic Stress A measure of economic stress was formed from responses to several questions. First, respondents were asked to list how frequently during the past 3 months they had found it difficult to pay their monthly costs. Response alternatives were: 1 = No, 2 = Yes, occasionally, 3 = Yes, every two months and 4 = Yes, every month. Second, respondents were asked to indicate how often during the previous 3 months they had been required to take extra measures to raise money for their monthly costs by: (1) borrowing money from a bank; (2) borrowing money from a private person; (3) pawning their possessions; (4) selling their possessions; and (5) applying for social welfare. Response alternatives for each item were: 1 = Several times, 2 = Once and 3 = Never. Third, respondents were asked about their ability to meet normal household expenditure. They were asked to indicate whether, during the previous 3 months, lack of money had: (1) limited them in their purchase of food; (2) forced them to refrain from buying clothes; and (3) forced them to refrain from taking a holiday. Response alternatives for each question were: 1 = Several times, 2 = Once and 3 = Never. Finally, respondents were asked about their ability to meet unforeseen expenses. The question was: ‘If you suddenly ended up in an unforeseen situation where you had to raise kr18,000 (equivalent to €2000), would you be able to?’ The response alternatives were: 1 = Yes and 2 = No. A weighted measure of economic stress was created by summing individual responses to each item (Cronbach’s alpha = 0.79). Scores from the resulting variable, economic stress, were classified into three groups (greater, medium and lesser degree of economic stress).
66
Unemployment, Secure Employment and Insecure Employment
Results Social Background Conditions and Health Concerning the connections between social background conditions on the one hand, and depression, anxiety and psychiatric ill health on the other, chi-square analyses showed that there were significant relations to gender (p < .001), education (p < .05 for depression and psychiatric ill health and p < .01 for anxiety) and type of household (p < .001; Table 2). The presence of chronic psychosomatic problems was significantly related to gender (p < .01) and education (p < .05). Women showed more depression, anxiety and psychiatric ill health than men, while single people living with children under 18 years of age scored higher on these measures than did participants living in other household types. No clear pattern of relationships between education and the different symptoms was found. The prevalence of psychiatric ill health was
Table 2 Depression, Anxiety, Psychiatric Ill Health and Psychosomatic Problems by Gender, Age, Education and Household Type Depression
Total Gender Women Men Age 18–24 25–34 35–44 45–54 55–64 Education Preupper secondary school Upper secondary school Upper secondary school Postupper secondary school Postupper secondary school Household type Single, no children under 18 Single, children under 18 Couple, no children under 18 Couple, children under 18
Anxiety
Psychiatric ill health
Psychosomatic problems
%
n
%
n
%
n
%
n
14
298
18
407
7
136
5
115
17 9 ***
190 103
22 14 ***
244 159
8 5 ***
80 56
7 4 **
77 38
10 14 14 15 12 ns
15 74 67 88 54
21 19 21 17 15 ns
33 102 104 103 65
8 8 7 7 3 ns
11 38 34 38 15
3 5 5 6 6 ns
4 25 25 34 27
12 16 9 15 14 *
53 127 40 43 35
16 22 15 18 17 **
70 178 63 55 41
4 8 6 7 9 *
15 60 22 20 19
7 6 3 3 6 *
30 49 12 9 15
18 30 11 12 ***
49 35 101 99
22 40 16 17 ***
59 47 142 140
9 14 5 6 ***
25 15 44 47
6 9 5 5 ns
17 10 42 39
Note: * p < .05; ** p < .01; *** p < .001; ns = not significant.
67
Unemployment and Health ■ Health Effects
greatest among those with the highest education while for depression and anxiety it was greatest for those with upper secondary school. Women also showed more chronic psychosomatic problems than men. Though the connection to education was significant (p < .05) no clear pattern was found linking education and chronic psychosomatic problems. Work, Unemployment and Ill Health The percentages of respondents from each of the five employment categories (secure–empowered, secure–disempowered, insecure–empowered and insecure–disempowered jobs, and unemployed) who were classified as having depression, anxiety, psychiatric ill health and psychosomatic problems are shown in Table 3. As can be seen, all indicators of ill health were significantly related to job security and empowerment (p < .001). Concerning depression, symptoms were least common in those with secure–empowered jobs, and most common among both the insecure–disempowered employed and the unemployed. A similar pattern was observed for anxiety, except that in this case, anxiety was more common in the insecure–disempowered group than among the unemployed. Concerning both psychiatric and psychosomatic problems, there appeared to be a clear division between the insecure–disempowered employed and the unemployed on the one hand, and the other three employed groups on the other. An obvious gradient can be seen among the four groups representing different combinations of secure–insecure and empowered–disempowered jobs. There is a gradual increase of the prevalence of these troubles of a stepladder model character among the four groups consisting of combinations of secure–insecure — empowered–disempowered jobs. It may be noted that this stepladder-like shape is steeper for men than for women, that is, the gradual differences are greater for men than for women. The prevalence of self-reported depression and anxiety is the greatest among those who have insecure–disempowered jobs. For depression, the rate is comparable with that of the unemployed; and for anxiety, it is somewhat Table 3 Ill Health by Employment Category and Gender Depression
Secure–empowered Secure–disempowered Insecure–empowered Insecure–disempowered Unemployed Note: F = Female; M = Male. ***p < .001.
68
Anxiety
Psychiatric ill health
Psychosomatic problems
F %
M %
F %
M %
F %
M %
F %
M %
11.3 14.3 21.7 26.5 25.2 ***
3.8 8.2 15.3 17.9 18.4 ***
12.2 21.1 24.3 34.6 28.0 ***
5.7 13.1 23.5 30.3 26.2 ***
4.4 6.4 4.6 13.3 13.4 ***
1.5 4.8 5.6 11.7 12.8 ***
3.0 7.5 2.9 11.2 11.8 ***
1.5 3.9 2.1 5.8 8.3 ***
Unemployment, Secure Employment and Insecure Employment
greater. This is valid for both women and men, with the difference that the figures are consistently higher for women. For psychiatric ill health measured with GHQ and chronic psychosomatic problems the patterns are ambiguous. Concerning the prevalence for psychiatric ill health there is a clear stepladder-like pattern for men but not for women. The prevalence of psychiatric ill health for the group insecure–disempowered jobs is on a level with the unemployed. This is valid for both women and men. However, the differences in prevalence of psychiatric ill health between secure–empowered jobs and insecure– empowered jobs are very small for women. The patterns are also ambiguous concerning the prevalence of chronic psychosomatic problems. Among women there are two groups who clearly distinguish themselves from the others, namely the groups insecure–disempowered jobs and unemployed, where the prevalence is similar and clearly the greatest. The prevalence of chronic psychosomatic problems for men is constantly somewhat smaller. It is the greatest among unemployed men, followed by the group insecure–disempowered jobs. It should be noted that the differences between the groups for the reported symptoms are significant and that is valid for both men and women (p < .001). Logistic regression analyses were carried out for the purpose of controlling for the influences of various social conditions. Three models were used. Model A is without adjustment for demographic factor or economic stress. In Model B an adjustment has been made for age, education and type of household, and in Model C for economic stress. Odds ratios are shown separately for women and men. The results are shown in Table 4a and 4b. Overall, the odds ratios shown in Tables 4a and 4b suggest that the effects of employment status on health are stronger for men than for women. For men (Table 4b) the odds ratios increase substantially when adjustment is made for the impact of age, type of household and education (Model B), but they decrease when additional adjustment is made for the impact of economic stress (Model C). This means that adjusting for social background conditions strengthens the relation between type of jobs, unemployment and ill health, while an adjustment for economic stress reduces the strength of the relations. In Model B, the odds ratios are the highest for the unemployed concerning depression, psychiatric ill health and chronic psychosomatic problems. For anxiety, the odds ratio is highest for the insecure–disempowered category. In Model C, where adjustments have also been made for economic stress, it can be stated that there is a strong reduction of the size of the odds ratios, in particular for the unemployed, compared to Model B. For depression and anxiety the relative risk for the unemployed is consistently lower than for the two insecure groups. For psychiatric ill health 69
Unemployment and Health ■ Health Effects Table 4a Ill Health Among Women Related to Employment Category: Logistic Regression Depression Employment category
Secure–empowered Secure–disempowered Insecure–empowered Insecure–disempowered Unemployed
Anxiety
Model
Psychiatric ill health
Model
Psychosomatic problems
Model
Model
A
B
C
A
B
C
A
B
C
A
B
C
1.0 1.3 2.2c 2.8c 2.6c
1.0 1.4 2.1c 2.7c 2.7c
1.0 1.2 2.1c. 2.4c 2.2c
1.0 1.9b 2.3b 3.8c 2.8c
1.0 2.0 b 2.4b 3.3c 2.6c
1.0 1.9 b 2.2 3.0c 2.2b
1.0 1.5 1.1 3.4c 3.4c
1.0 1.7 1.2 3.1b 4.3c
1.0 1.6 1.1 2.6a 3.5b
1.0 2.7b 1.0 4.1c 4.3c
1.0 2.5b 0.9 3.1b 3.3b
1.0 2.7b 0.4 2.9a 3.4b
Note: a = p < .05; b = p < .01; c = p < .001 Models: A = No adjustment; B = Adjusted for age, household type and education; C = Adjusted for age, household type, education and economic stress.
it is lower than the insecure–disempowered group while it is substantially higher for chronic psychosomatic problems than it is for the other groups. For men, economic stress seems to strongly reduce the connection between type of jobs and unemployment on the one hand and depression, anxiety, psychiatric ill health and chronic psychosomatic problems on the other. For women (Table 4a), the patterns in Model A and Model B concerning depression and anxiety differ very little. For psychiatric ill health the odds ratio decreases slightly in Model B for the insecure–disempowered group but increases for the unemployed group. For chronic psychosomatic problems the odds ratios decrease substantially in Model B for the insecure–disempowered group and the unemployed. Regarding economic stress (Model C) the odds ratios decrease for the insecure–disempowered jobs and unemployed for depression, anxiety and psychiatric ill health, meaning that taking economic stress into consideraTable 4b Ill Health Among Men Related to Employment Category: Logistic Regression Depression Employment category
Secure–empowered Secure–disempowered Insecure–empowered Insecure–disempowered Unemployed
Anxiety
Model
Psychiatric ill health
Model
Psychosomatic problems
Model
Model
A
B
C
A
B
C
A
B
C
A
B
C
1.0 2.2a 4.6c 5.5c 5.7c
1.0 2.8b 5.4c 7.0c 7.8c
1.0 2.3a 4.9c 5.5c 3.7c
1.0 2.5c 5.1c 7.2c 5.9c
1.0 3.1c 6.1c 8.7c 8.0c
1.0 3.0c 5.7c 7.0c 4.6c
1.0 3.2b 3.8a 8.5c 9.4c
1.0 3.7b 4.4b 10.3c 13.9c
1.0 3.9b 4.3a 7.8c 7.4c
1.0 3.9b 1.4 4.1b 5.9c
1.0 3.6a 2.0 6.1c 10.9c
1.0 3.4a 1.9. 3.6a 6.7c
Note: c = p < .001; b = p < .01; a = p < .05 Models: A = No adjustment; B = Adjusted for age, household type and education; C = Adjusted for age, household type, education and economic stress.
70
Unemployment, Secure Employment and Insecure Employment
tion decreases the differences in prevalence between these groups and the reference group, secure–empowered jobs. The impact of economic stress on the relation between type of employment and ill health is similar for both women and men and is in line with previous research that shows a strong link between financial hardship and various symptoms of psychological ill health. However, the connection between type of employment and ill health is stronger for men than for women. It seems that a secure and empowered job does not give women the same health gain as it does men. The reasons for that might be affected by conditions beyond the examined categories, such as women’s greater responsibility for household work and relatives.
Discussion The purpose of this study has been to compare indications of ill health among five groups. Four of these comprised people with varying degrees of security and empowerment at work and the fifth group comprised unemployed people. The survey shows that the five groups differ in reported ill health. Concerning differences between men and women, it can be stated that regardless of their employment status (with few exceptions) women were more likely to report symptoms of ill health more frequently than men. However, multivariate analyses showed that men’s wellbeing was more strongly affected by employment status than was women’s. This was found both for simple analyses without adjustment for social background conditions and economic stress, and for analyses which included adjustments for these factors. Although the prevalence of health symptoms was higher for women than for men in most groups studied, the reasonable interpretation seems to be that men’s self-reported health troubles to a great extent are linked to the actual situation in work or unemployment. The finding that women in all employment groups were more likely than men to report ill health is consistent with other research (e.g., Molarius & Janson, 2002). The possible reasons for this are probably complex. It might reflect the greater responsibilities for conditions in domains outside work that women have due to traditional gender roles. The result shows that the prevalence of symptoms among women who have a secure–empowered job is substantially greater than it is for men. However, the link between type of employment and ill health is much stronger for men than for women. The strong health gain that men with a secure–empowered job make is not to be found for women. The results point to differences in the prevalence of ill health for the different employment categories. Those who have a secure–empowered job enjoy the best health. The worst health is to be found among the insecure–disempowered and the unemployed. In some cases (e.g., psychiatric ill health and chronic psychosomatic problems) it is worse for the 71
Unemployment and Health ■ Health Effects
unemployed and in other cases (e.g., anxiety) for those who have an insecure–disempowered job. It can be noted that adjusting for economic stress substantially reduces the relative risk of acquiring ill health. The differences in ill health between the five employment groups are strongly reduced when controlling for economic stress. On the whole the reduction is particularly true when comparing the insecure–disempowered and the unemployed with the secure–empowered. This suggests, in line with previous research, that the economic conditions of the unemployed have a substantial bearing on health. Economic stress constitutes an important factor to be taken into consideration when assessing whether it is unemployment itself or the economic conditions that are important in explaining increasing ill health among the unemployed. Furthermore, economic stress seems to play an important role for those who have a job. In this case too, the finding is in line with previous research (see Pearlin & Lieberman, 1979). The relative risk of suffering ill health is, in some cases as we have noted, even greater among those who have an insecure–disempowered job than for those who are unemployed. This is in line with what has previously been documented, that job insecurity, lack of influence and reduced possibilities for personal development at work increase the risk of ill health. Several authors have argued that the changes that have taken place during the last 10 years have rendered obsolete the traditional dichotomy between work and unemployment when studying the relationship of employment status to health and welfare (Fryer, 2000; Hallsten & Isaksson, 2000). The point of the arguments is that, as a result of these changes, the category of work embraces nuances that have to be taken into consideration in order to make better-suited comparisons between the employed and unemployed. One of the distinctive features of the new labour market is that the working conditions for a growing number of people have become more uncertain. It is a web of different conditions that are thought to form the basis for the changes that have taken place; namely, increased globalisation of the economy, increased international competition, the emergence of new technologies, de-industrialisation and recurring recessions. Privatisation of previously state-owned industries has usually brought with it restructuring and reduction of business operations. Highly specialised and streamlined organisations and companies have, to a greater extent, replaced large-scale production. The outcome of the resulting ‘flexible labour market’ is that the different types of employment have become increasingly insecure. The results of this study support the view that research into health risks and ill health should pay close attention to employment conditions and psychosocial working conditions. While it is generally acknowledged that unemployment has serious consequences for psychosocial health, it appears that there are also some groups in the contemporary workforce who, as a
72
Unemployment, Secure Employment and Insecure Employment
result of insecure or disempowering working conditions, face psychosocial risks that are comparable to those experienced by the unemployed.
Endnote 1
Temporary agency work combines elements of both commercial and employment contracts, and brings a third party – the agency – into what was previously a twoparty relationship. A definition of temporary agency work is one whereby the temporary agency worker is employed by the temporary work agency and is then, via commercial contact, hired out to perform work assignments at the user firm (Storrie, 2002, p. 1).
References Åberg, R., Strandh M., Nordenmark, M., & Bolinder M. (1997). Massarbetslösheten på 90-talet. In J. Vogel & L. Häll (Eds.), Levnadsförhållanden: Välfärd och ojämlikhet i 20-årsperspektiv 1975–1995 (pp. 553–550). Stockholm: Statistiska Centralbyrån. Aronsson, G. (1999). Contingent workers and health and safety. Work, Employment & Society, 13, 439–459. Aronsson, G., Dallner, M., & Lundh, T. (2000). Flexibla inkomster och fasta utgifter: En studie av ekonomisk stress och hälsa bland korttidsanställda [A flexible income and fixed expenses: Economic stress among on call-employees] (Arbete och Hälsa Research Report No. 29, pp. 1–26). Stockholm, Sweden: Arbetslivsinstitutet. Benach, J., Benavides, F.G., Platt, S., Diez-Roux, A., & Muntaner, C. (2000). The healthdamaging potential of new types of flexible employment: A challenge for public health researcher. American Journal of Public Health, 90(8), 1316–1317. Burchell, B. (1994). The effects of labour market position, job insecurity, and unemployment on psychosocial health. In D. Gallie, C. Marsh, & C. Vogler (Eds.), Social change and the experience of unemployment (pp. 188–212). Oxford, UK: Oxford University Press. Fryer, D. (2000). Strukturwandel der Arbeitslosigkeit und psychische Gesundheit [The restructuring of unemployment and mental health]. In H.G. Zilian & J. Flecker (Eds.), Soziale Sicherheit und Strukturwandel der Arbeitslosigkeit (pp. 240–256). Munich, Germany: Rainer Hampp Verlag. Goldberg, D.P., & Williams, P. (1988). A user’s guide to the GHQ. London: NFER, Nelson. Graetz, B. (1993). Health consequences of employment and unemployment: Longitudinal evidence for young men and women. Social Science and Medicine, 36(6), 715–724. Griffin J., Fuhrer R., Stansfeld S., & Marmot, M. (2002). The importance of low control at work and home on depression and anxiety: Do these effects vary by gender and social class? Social Science and Medicine, 54(5), 783–798. Hallsten, L. (1997). Arbetslöshet och psykisk ohälsa 1980-1996: En meta-analys [Unemployment and mental health 1980–1996: A meta-analysis] (Arbete och Hälsa Research Report No. 28, pp. 1–79). Stockholm, Sweden: Arbetslivsinstitutet. Hallsten, L., & Isaksson, K. (2000). Arbetslöshet, osäker anställning och psykisk ohälsa [Unemployment, insecure employment and mental ill-health]. In S. Marklund (Ed.), Arbetsliv och hälsa 2000 (pp. 261–286). Stockholm, Sweden: Arbetarskyddsstyrelsen, Arbetslivsinstitutet. Karasek R. (1998). The demand/control model: A social, emotional, and physiological approach to stress risk and active behaviour development. In Encyclopaedia of occupational health and safety (4th ed.; pp. 34.6–34.14). Geneva, Switzerland: International Labour Office. 73
Unemployment and Health ■ Health Effects Karasek, R., & Theorell, T. (1990). Healthy work. New York: Basic Books. Kessler, R.C., Turner J.B., & House J.S. (1987). Intervening processes in the relationship between unemployment and health. Psychological Medicine, 17, 949–961. Kinnunen, U., & Natti J. (1994). Job insecurity in Finland: Antecedents and consequences. European Work and Organizational Psychologist, 4, 297–321. Molarius, A., & Janson, S. (2002) Self-rated health, chronic diseases and symptoms among middle-aged and elderly men and women. Journal of Clinical Epidemiology, 55, 364–370. National public health report. (2000). Stockholm, Sweden: Socialstyrelsen. Pearlin, L.I., & Lieberman, M.A. (1979). Social sources of emotional distress. In R.G. Simmons (Ed.), Research in community and mental health. Greenwich, CT: JAI Press. Platt, S., Pavis, S., & Akram, G. (1999). Changing labour market conditions and health: A systematic literature review (1993–1998). Dublin, Ireland: European Foundation for the Improvement of Living and Working Conditions. Starrin, B., Rantakeisu, U., Forsberg, E., & Kalander-Blomqvist, M. (2000). Understanding the health consequences of unemployment: The finance/shame model. In T. Kieselbach (Ed.), Youth unemployment and health: A comparison of six European countries. Opladen, Germany: Leske + Budrich. Storrie, D. (2002). Temporary agency work in the European Union, European Foundation for the Improvement of Living and Working Conditions (pp. 1–98). Luxembourg, Belgium: Office for Official Publications of the European Communities. Third European survey on working conditions. (2000). European Foundation for the improvement of living and working conditions. Retrieved October 20, 2001, from http://www.eurofound.ie/publications/files/EF0121EN.pdf Whelan, C.T. (1992). The role of income, life-style deprivation and financial hardship in mediating the impact of unemployment on psychological distress: Evidence from the Republic of Ireland. Journal of Occupational and Organizational Psychology, 65(4), 331–344. Vinokur, A.D., Price, R.H., & Caplan, R.D. (1996). Hard times and hurtful partners: How financial hardship affects depression and relationship satisfaction of unemployed persons and their spouses. Journal of Personality and Social Psychology, 71, 166–179. Wichert, I. (2002). Job insecurity and work intensification. The effects on health and well-being. In B.J. Burchell, D. Ladipo, & F. Wilkinson (Eds.), Job insecurity and work intensification (pp. 92–111). London: Routledge.
74
Chapter
6
The Finances–Shame Model and the Relation Between Unemployment and Health Bengt Starrin and Leif R. Jönsson
For many years the Swedish economy was characterised by growth and low unemployment. However, this trend was broken in the early 1990s. Between 1990 and 1993 half a million jobs were lost, most of them in the industrial sector. Unemployment — including among those on labour market training schemes — thus rose rapidly from approximately 2% to 13% in just 4 years. These years of crisis, 1990 to 1993, had serious repercussions in the private sector. Those who suffered most were migrants, young people and those with a low level of education. Since 1994 there have been severe cutbacks in municipal and county council health and social care. This new unemployment differs from that in previous periods of unemployment in modern Sweden. It is long term and more structural in nature. Although there has been a significant decrease recently, the total unemployment rate is still considerable. The total unemployment rate during 2000 was almost 7%. The first half of the 1990s was also marked by extensive cuts in social welfare. These cuts affected a diversity of welfare payments, including unemployment benefit funds, cash unemployment allowances, education allowances, sick pay, parents’ allowances, child allowances, housing benefits and maintenance advances. This, together with the high and longterm unemployment, meant that new groups were forced to apply for supplementary benefits in order to cope with their living costs. Between 1990 and 1995 the numbers receiving supplementary benefits in the country rose from 517,000 to 720,000, an increase of about 40%. During the same period the costs for supplementary benefits rose by 86% and the benefit months by nearly 30% (SoS-rapport, 1997, p. 5). 75
Unemployment and Health ■ Health Effects
The 1990s crisis placed many people under financial stress. Between 1987 and 1995 the proportion of people in Sweden who indicated that they were often worried about their own financial situation or that of their family in the coming year virtually doubled. Above all, it was those who risked losing their jobs or who were already unemployed who were most worried about their finances. In response to the question ‘What worries you most about the future?’, ‘making ends meet’ was by far the most frequent answer from both young unemployed men and women (Starrin, Rantakeisu, & Hagquist, 1996), and from workers at iron and steel plants who had been given notice (Starrin & Lundberg, 1993). Roughly speaking, six out of 10 indicated that their financial situation worried them most. Modern research has shown that unemployment may create social and health problems (Fryer, 1995; Hammarström, 1996; Janlert, 1991; Jin, Shah, & Svoboda, 1995; Shortt, 1996; Warr 1987; Wilson & Walker, 1993). However, what is characteristic of much of the research that has studied the relation between unemployment on the one hand, and social problems and ill health on the other, is that it is largely survey-based and often lacks sound theoretical underpinnings. Nevertheless, theoretical approaches to this problem do exist. Among the more well-known are Jahoda’s theory on the manifest and latent functions of work (Jahoda, 1982), Warr’s vitamin theory (Warr, 1987), Fryer’s agent theory (Fryer, 1986), Ezzy’s status passage theory (Ezzy, 1993), the theory of control locus (O’Brien, 1986), classic stress theory (Janlert, 1991) and the theory of financial deprivation (Janlert, 1991). Among these, it would seem that stress theory in a more or less developed form underpins interpretations of the relationship between unemployment and ill health. Based upon an analysis of the results from research on unemployment and stress, Kieselbach (1999) has developed an integrative model with the concept of victimisation as the central concept for the understanding of the health consequences of unemployment. Certainly the stress perspectives run in several different directions. However, what unites these strands is that under certain circumstances the individual becomes so pressed that negative stress occurs. Stress is developed by different types of stressors. Negative stress occurs when an individual perceives a situation as threatening and the perception continues to place the individual in a condition of preparedness in which various ways out of the situation, such as flight, adjustment and change, are considered. The negative stress can involve a range of changes, for example, physiological changes (increased heartbeat), behavioural changes, changed perception and changed emotional condition. Negative stress can also result in a condition that is defined as illness. Stress theory employs two central concepts, coping and social support, to explain the reason why some people who are exposed to stressors do not become ill. Exposure to stressors leads to illness if the destructive power of 76
The Finances–Shame Model
the stressors are not counteracted by protective circumstances in the individual’s surroundings (Cassel, 1976). In stress theory there is an assumption that an individual’s experience of the social situation under certain circumstances can lead to the commencement of an additional mental and physiological process which, in its turn, starts a pathological process. Emotions are encompassed in this physical process. However, what is usually not provided is a more detailed description of exactly which emotions are stimulated, how they are activated and what role they play in setting in motion a pathological process.
The Finances–Shame Model In this study we use the finances–shame model as a framework for our analysis. The model has been developed to increase the understanding of why unemployment might be an affliction for some, while others are not affected negatively by it to any marked degree (Jönsson & Starrin, 2000; Rantakeisu, Starrin, & Hagquist, 1999; Starrin, Jönsson, Forsberg, & Rantakeisu, 1998; Starrin, Jönsson, & Rantakeisu, 2001; Starrin, Rantakeisu, Forsberg, & Kalander-Blomqvist, 2000; Starrin, Rantakeisu, & Hagquist, 1997; Starrin et al., 1996). Two research traditions are brought together in the finances–shame model. The financial aspect of the model is linked to research on financial stress and deprivation (Horwitz, 1984), while the social aspects are associated with microsociological research on social bonds (Scheff, 1990), social networks (Berkman & Syme, 1979), self-image and identity (Hayes & Nutman, 1981). The model postulates the significance of two conditions for understanding the relationship between unemployment, ill health and social problems: the level of financial hardship and the extent of shaming experiences. The finances–shame model is illustrated in Figure 1. The model also postulates that it is the combined effect of the financial hardship and shaming that determines the health outcome. More ill health is expected among those who have been exposed to both greater financial hardship and more shaming experiences, while less ill health is expected
Shaming Experiences More
Financial hardship
Fewer
Greater
More ill health
Moderate ill health
Lesser
Moderate ill health
Less ill health
Figure 1 The finances–shame model and the anticipated effects of unemployment on health.
77
Unemployment and Health ■ Health Effects
among those who have been exposed to less financial hardship and fewer shaming experiences. No specific assumptions are made about those in ‘greater–fewer’ and ‘lesser–more’ except that the prevalence of ill health is expected to be moderate. It has been pointed out in the literature that research on unemployment has paid insufficient attention to what is lost through unemployment: on the one hand, money, in the form of wages, and on the other, social prestige, status and respect (Willis, 1986). The finances–shame model takes both these aspects into account. The financial aspect considers the importance money has for people’s everyday lives, their wellbeing and health. Money is important for stabilising and maintaining an established lifestyle and also for social intercourse. But money also gives a sense of control and power (Underlid, 1992). In general, it can be said that research that is concerned with illustrating the relationship between financial conditions on the one hand and ill health on the other has found that it is primarily low financial status, financial stress and poverty that are linked with an increased risk of ill health and have a powerful impact on wellbeing (Horwitz, 1984). Research that has studied the financial aspect of unemployment has shown that there is a link between financial hardship and ill health, primarily mental ill health (Kessler, Turner & House, 1987; Starrin & Lundberg, 1993; Starrin, Rantakeisu & Hagquist, 1996; Underlid, 1992; Vinokur, Price & Caplan, 1996; Whelan, 1994). Studies of the financial aspect of unemployment use various terms to describe the condition where financial problems become so great that they can have social consequences and/or affect health. The terms used are ‘economic stress’ or ‘financial stress’, ‘economic hardship’ or ‘financial hardship’, ‘economic strain’ or ‘financial strain’, and ‘economic vulnerability’, ‘financial vulnerability’ or ‘economic deprivation’. The various terms are generally used without any precise definition of their meaning. The term ‘economic’ or ‘financial stress’ would seem to be used primarily with explicit reference to stress theory. Terms like ‘hardship’ and ‘strain’ are usually given no other meaning apart from general commonsense understanding. Economic or financial stress has been defined in slightly differing ways. Elder and Caspi relate economic stress to the pressure or strain resulting from a tangible loss of income (Elder & Caspi, 1988). Ilstad defines economic stress as a discrepancy between actual income and the income needed, in the sense that the income is less than that needed and that it is difficult to compensate the loss of income with previous savings, loans and credit over a longer period (Ilstad, 1989). It is not satisfactory that the terms are unclear and that they are sometimes used as if they were self-defining. In this study we will use the term ‘financial hardship’ to refer to various degrees of hardship. Financial hardship is taken to refer to the problems that arise in connection with, or 78
The Finances–Shame Model
when a person is faced with, the threat of a shortage of money; and that find expression in the fact that the person may be forced to take extraordinary steps such as pawning or selling possessions, borrowing money or not being able to pay bills on time. The other aspect taken into account in the finances–shame model is the potential sense of shame at being out of work. Apart from the fact that wage labour provides income, it also provides status, prestige and respect to a greater or lesser extent. Historically there is clear evidence that the lower social classes — including the unemployed — may be looked down upon and treated with contempt. George Orwell described this as early as 1937 in The Road to Wigan Pier (Orwell, 1937). Marsden (1975) pointed out over 25 years ago that the view of the unemployed as work-shy parasites has persisted, despite lacking firm basis in fact. The belief that unemployment is the result of the negative personal qualities and disinclination to work of the unemployed themselves has been widespread (Gallie, 1994). Many of the unemployed have had the experience of being ignored or being considered lazy and ignorant (Starrin et al., 1998). How other people see the unemployed individual is of significance for their identity. Studies show that many of the unemployed are concerned about how they are judged and valued by others (Hayes & Nutman, 1981). In encounters with other people, authorities and welfare institutions, the unemployed individual faces a test of their self-respect. Unemployment seems to result in the previously accepted role of student or employee being replaced by a role which has lower prestige and which is seen by many as deviant and shameful (Schlozman & Verba, 1979; Seabrook, 1982; Warr, 1987). In his classic study, Cato Wadel notes that becoming angry and engaging in aggressive argumentation provide a means by which many unemployed preserve their dignity and self respect (Wadel, 1973). Thus the unemployed seem to be aware of their vulnerable position (Warr, 1987). Some authors use the term ‘stigma’ to describe this explicitly. McFayden (1995) summarises her review of the research by maintaining that there is a general perception of stigma among all age groups of unemployed, except for a minority of young people living in areas with high levels of unemployment. Goffman (1963) used the term ‘stigma’ to refer to an attribute that is deeply discrediting (McFayden, 1995). Stigma may be seen as an aspect of status. An individual who is stigmatised is assigned low status. According to Max Weber, status is a matter of social dignity or the lack of it. The status enjoyed by different groups is based on either a specific positive or a specific negative social assessment of their decency and creditability. There is thus a close relationship between status and stigma (Spicker, 1984). People who are stigmatised run a greater risk of losing their social position either because the stigma may restrict their chances of assuming a special role or because the stigma defines lower or subordinate roles. 79
Unemployment and Health ■ Health Effects
Stigmatisation also increases the risk of getting into financial difficulty and becoming impoverished as the stigma in itself may be an obstacle to getting a job and earning money. As a stigmatised group it seems that the unemployed face a real dilemma. The deficits of character attributed to unemployment and poverty, such as laziness and lack of ability, are not easy to disconfirm. If the unemployed have little opportunity to express and demonstrate their capacity to work, they might have little chance to disavow their stigma (Dressner, 1988). Thus stigma may be seen as an aspect of social disgrace — as a sign of demotion, loss of dignity or loss of a good reputation. Low status implies inferiority and the charge of inferiority is an aspect of stigmatisation. In some studies there is a reference to the concept of shame (Eales, 1989; Neckel, 1991; Starrin, 2001; Starrin et al., 1996; Starrin et al., 2000). Shame may be said to be a feeling of inferiority arising from a sense of personal failure, most commonly in the presence of another person who disapproves of one’s action or outlook. It is a result of seeing oneself negatively in the eyes of the other, such as feeling rejected, unworthy or inadequate (Scheff, 1990). As a phenomenon, then, the sense of shame means that a person is seen and acknowledged in a painfully reduced manner. Shame reveals the inner person, the self, and exposes it to the eyes of others. It has been suggested that because shame is potentially so painful, no other feeling may be as central to the development of identity and self-image (Kaufman, 1993). Within sociology, there are two more detailed theoretical accounts of shame that are relevant to the present study. One is associated with Cooley (1902/1922), Retzinger (1991) and Scheff (1990, 2003). The basic idea in this account is that shame, but also pride, have a signal function with respect to the social bond. Both serve as intense signs of the state of the social bonds. Shame is the sign of a threatened, severed or insecure social bond, and pride is the sign of an intact and secure bond. If the shame is overt one shrinks, averts or lowers one’s gaze, casting only furtive glances at the other. If the shame is bypassed one stares, attempting to outface the other (Scheff, 1990). Retzinger (1991) and Scheff (1990, 2003) define shame broadly as a class name for a large family of emotions. It includes many variations, from social discomfort and embarrassment, characterised by weak intensity and transient duration, to humiliation, characterised by powerful intensity and long duration. What these emotions have in common is that they signal threat to the social bond (Retzinger, 1991). The state of the social bonds is continually noted, not just by interpreting what is said, but primarily by how it is said; that is, by the way people behave. Feelings of shame and its opposite, pride, are signals of the quality of the social bonds — whether they are strained or secure.
80
The Finances–Shame Model
The other theoretical account of shame that is of importance for the present chapter deals with the close association that shame has with hierarchical aspects of society, such as low social status, low social class, social subordination and inferiority (Neckel, 1991; Scheff, 1990; Sennet & Cobb, 1972, Sennett, 1999). It has been argued that the typical form for the status-bound sense of shame in modern society is the feeling of inferiority (Neckel, 1991) and that the feeling of shame contains a dynamic force that can create and reproduce inferiority. Disadvantaged groups not only suffer from disrespect from others; in order to cope with everyday life they are forced to adopt the attitudes of the dominant groups and thus see themselves through the disapproving or disdainful eyes of others. This creates a strong sense of self-generated shame (Scheff, 1990). Making other people feel a sense of shame has become an increasingly important element of the exercise of power. An individual who feels shame usually falls into line (Sennett, 1980). In Zawadski and Lazarsfeld’s study of autobiographies by over 50 unemployed people it emerged that feelings of degradation were common. Many authors wrote about their feelings of shame. Most of them were ashamed of their misery (Zawadski & Lazarsfeld, 1935). In his 1980s study Eales (1989) showed that feelings of shame are common among unemployed adult men. There is much to suggest that the negative effects of unemployment on identity, health and wellbeing are related to societal attitudes towards the unemployed as a social category (Breakwell, 1985; Breakwell, Collie, Harrison, & Propper, 1984). In Eales (1989) the experience of shame was associated with mental health problems such as depression and anxiety. A Swedish study has also revealed a link between the perception of contempt felt by a person’s neighbours and associates and mental health, such that the more the unemployed individual experienced shame the more they were troubled by mental ill health (Rantakeisu, Starrin, & Hagquist, 1997). In the present chapter we do not focus on the feeling of shame itself as a specific condition but we treat the relational aspect by examining whether an individual has experienced degradation and contempt. The term we use for this is ‘shaming experiences’. These refer to experiences that resemble what is sometimes termed ‘toxic’ or ‘pathological’ shame; that is, a type of shame that we bear as a result of being repeatedly subjected to humiliation, insults and rejection by other people. It differs from normal shame, which may be considered a natural part of human life (Scheff, 1992). The aim of the study is to examine self-reported health among the unemployed in the light of the finances–shame model (Starrin et al., 2000; Starrin et al., 2001). The assumption is that the greater the financial hardship during unemployment and the more experiences of being shamed when unemployed, the more serious the health consequences will be. The
81
Unemployment and Health ■ Health Effects
lesser the financial hardship and the fewer the shaming experiences, the less serious the effects on health will be.
Method Participants and Selection Participants were drawn from a population of unemployed people involved in labour-market training and workfare programs in 14 municipalities in southern Sweden. The 14 municipalities were selected on the basis of two criteria: (a) the group should include both large, medium-sized and small communities throughout the region; and (b) there should be labour-market training and workfare programs in the municipalities selected. The data were collected in cooperation with the district organisation of LO (the Swedish Confederation of Trade Unions in the region) from 1615 people who were on some form of labour-market training or workfare program in weeks 39 and 40 (September/October) of 1996. There were 1249 responses to the questionnaire, a response rate of 77%. At the time of the survey a total of 6980 individuals (3701 men and 3279 women) in the region were involved in some form of labour-market training or workfare program; thus the population surveyed represented 18% of this group. Measures Participants were asked to complete a questionnaire that included questions about their financial situation, health, family situation, ethnic origin, perception and experience of being unemployed, and their encounters with the social services. There were also questions designed to test the finances–shame model (see Figure 2). Weighted measures were constructed to measure financial hardship and shaming experiences. The factors financial hardship and shaming were formed by weighting together the answers to each question provided by the individuals (see Figure 2). This was achieved by means of a component analysis (principal component, factor scores, regression). The weighted measures for each individual were put in two groups of approximately equal size: ‘greater’ and ‘lesser’ level of financial hardship and a ‘higher’ and ‘lower’ extent of shaming experiences. The finances–shame model was formed by combining financial hardship and shaming experiences. ‘Greater’ and ‘lesser’ was combined with ‘more’ and with ‘fewer’. Depression was measured by the question ‘Have you been troubled during the last three months by … depression?’ Only the response alternative More or less every day was considered in the analysis. Nervousness was measured in the same way with the question ‘Have you been troubled during the last three months by … nervousness?’ Only the response alternative More or less every day was considered in the analysis. Factor deterioration in health was measured by the question, ‘Has unemployment led to
82
The Finances–Shame Model Financial Hardship
Shaming Experiences
Since you have been unemployed, have you been forced to sell possessions to cover your expenses? Since you have been unemployed, have you been forced to pawn possessions to cover your expenses? Since you have been unemployed, have you been forced to borrow money to cover your expenses? Since you have been unemployed, have there been occasions when you have not been able to pay bills on time through lack of money? Response alternatives Yes, several times Yes, on occasion No If you were suddenly to find yourself in a situation where you needed to obtain SEK. 14000 within a week, could you manage it?
Have you felt that others have not paid attention to what you have done or said because you are unemployed? Have you felt that others consider you less intelligent because you do not have a job? Have you felt that others consider you lazy because you do not have a job? Have you experienced that people have been irritated with you because you do not have a job? Have you experienced that others have avoided you because you are unemployed? Since you have been unemployed, have you experienced that others talk disparagingly of the unemployed? Response alternatives Yes, many times Yes, several times Yes, on occasion No, never
Response alternatives Yes, without difficulty Yes, but with difficulty No
Self-Reported Il Health Depression Have you been troubled during the last three months by … Depression Response alternatives Had no trouble On occasion Several times a week More or less every day
Changes in health Has unemployment led to any changes in your general state of health? Response alternatives Yes, my general state of health is better now No, there is no difference Yes, my general state of health is worse now
Nervousness Have you been troubled during the last three months by … Nervousness Response alternatives Had no trouble On occasion Several times a week More or less every day
Figure 2 The finances-shame model.
any changes in your general health?’. In the analysis only the response alternative Yes, my general state of health is worse now was considered. Figure 2 shows how the finances–shame model was devised.
83
Unemployment and Health ■ Health Effects
Collecting and Processing the Data The questionnaire was distributed directly to respondents, who were included on participant lists. They were ticked off when they received, and when they returned, the questionnaire. The questionnaire was distributed and collected in two different ways. In some places it was distributed to respondents who took it home so that they could answer it in peace and quiet, and return it to the project manager a few days later. However, the most common method was to distribute the questionnaire on a particular occasion when the group in question was gathered. The questionnaire was then completed and returned directly. Logistic regression analysis was used to check for influences from various social conditions (multivariate analysis). Logistic regression analysis produces odds ratios; that is, ratios between two odds. In simplified terms the odds ratio may be said to express a kind of relative risk that an event or a state will occur. In the logistic regressions the reference value = 1.0. The Social Background of the Participants Table 1 provides a description of the participants’ social background. As is apparent, there were slightly more men. The largest age group were those aged 30 to 39. The most common household type was ‘single without children’. Thirty-eight per cent of participants had a migrant background. One third were first-generation migrants. Table 1 Social Background of Respondents % Gender Age
Household type
Ethnicity
Length of unemployment
Education
Note: (n = 1249)
84
Female Male < 29 30–39 40–49 50–64 Couple without children Single without children Couple with children Single with children Nonmigrant background First-generation migrant Second-generation migrant 1–12 months 13–24 months 25–36 months 37–48 months > 48 months Elementary school Comprehensive school Vocational school Upper secondary school University
49 51 23 30 23 25 25 32 30 14 62 32 6 27 25 17 18 14 15 22 34 23 7
The Finances–Shame Model Table 2 Financial Hardship During Unemployment Distributed According to Social Background Pawned Borrowed Sold Unable to possessions money possessions pay bills on time
Total Gender Age
Household type
Ethnicity
Length of unemployment
Education
No cash reserves
%
%
%
%
%
Female Male < 29 30–39 40–49 50–64 Couple without children Single without children Couple with children Single with children Nonmigrant background First-generation migrant Second-generation migrant
30 21 38 24 38 38 18 20 34 31 37 22 45 25
68 65 71 80 77 69 44 50 73 70 85 65 72 79
37 30 45 36 43 43 26 26 44 36 46 33 45 39
65 63 67 74 73 66 47 48 69 67 83 63 66 79
69 71 67 72 74 73 56 56 74 69 80 68 70 65
1–12 months 13–24 months 25–36 months 37–48 months > 48 months Elementary school Comprehensive school Vocational school Upper secondary school University
29 28 33 32 29 17 31 28 35 37
73 67 69 64 63 44 73 72 71 74
34 34 40 38 43 24 34 37 43 50
69 61 69 63 63 46 71 69 66 62
71 71 66 65 71 67 77 69 64 64
Almost 75% had been unemployed (open, or on labour-market training or workfare programs) for more than 1 year. Seventy per cent had an educational level below upper secondary school.
Results The Financial Situation of the Participants As shown in Table 2, 30% of participants had been forced to pawn possessions to cover their expenses. Over two thirds were forced to borrow money for the same reason. More than one third had sold possessions, and almost two thirds had been unable to pay their bills on time. Almost 70% of the sample had no cash reserves. This is a much higher figure than the average for the Swedish population as a whole. According to a national study from the National Board of Health and Welfare, 23% of the population were without cash reserves in 1994 (Social Rapport, 1997).
85
Unemployment and Health ■ Health Effects
Comparisons Between Subgroups of the Unemployed Pawning, borrowing money, selling possessions and being unable to pay bills on time were more common among men than women. On the whole, all these were least common in the oldest age group. Not being able to pay bills on time and the lack of cash reserves were most common among single participants with children. Eight out of 10 in this group had no cash reserves. This may be compared with the figure given in the population study from the National Board of Health and Welfare for the group as a whole, which was 48% (Social Rapport, 1997). Borrowing money was most common in the youngest age group. Around 80% of participants aged under 30 had borrowed money while unemployed. Selling possessions was most common among those unemployed with a university degree.
Table 3 Unemployment-Related Shaming Experiences by Social Background. Show Talk Avoidance Consider irritation disparagingly you less about the intelligent unemployed
Total Gender Age
Household type
Ethnicity
Female Male < 29 30–39 40–49 50–64 Couple without children Single without children Couple with children Single with children Nonmigrant background First-generation migrant Second-generation migrant
Length of unemployment 1–12 months 13–24 months 25–36 months 37–48 months > 48 months Education Elementary school Comprehensive school Vocational school Upper secondary school University
86
Consider you lazy
Ignore/ disregard you
%
%
%
%
%
%
74 58 66 67 64 66 51 57 63 66 61 59 68 66
30 73 74 69 79 77 69 72 73 76 75 71 77 82
51 23 36 20 33 39 25 25 28 36 30 22 48 23
57 51 51 49 57 60 37 43 47 59 57 46 61 49
43 54 61 67 62 62 39 50 59 61 59 55 60 68
62 41 45 38 45 53 35 39 41 47 49 36 57 37
60 62 64 57 68 57 61 65 64 56
68 76 77 71 80 73 72 73 76 72
26 28 32 30 36 23 28 28 35 39
49 49 54 49 54 47 51 48 56 51
55 56 60 57 63 48 55 60 61 56
41 38 48 43 48 38 45 43 43 45
The Finances–Shame Model
Nature and Prevalence of Shaming Experiences Table 3 presents the prevalence of reported shaming experiences; that is, the extent to which, as a result of their unemployment, participants perceived themselves to be the target of others’ negative attitudes. Sixty-two per cent reported that others had conveyed irritation about their unemployment, while approximately 75% reported hearing others make disparaging comments about the unemployed. Thirty per cent reported feeling that they were avoided by others because of their unemployment. Just over half thought that others considered them less knowledgeable, and almost 60% that others considered them lazy. More than 40% believed that others did not heed what they said or did. Comparisons Between Subgroups of the Unemployed It may be noted that migrants and, above all, first-generation migrants reported the highest level of shaming experiences. Those who experienced the least shaming were, in general, the oldest age group (50–64 years).
Health As is apparent from Table 4, 14%of the sample were troubled by depression almost daily during the 3 months prior to the survey. For women, household type was significantly related to depression, such that depression was more common among singles than among couples. For men, no significant relationship was found between the different backgrounds factor and depression. Twenty-two per cent of women and 20% of men were affected almost daily by nervousness. For women, nervousness was significantly related to age and education. For men, nervousness was also related to ethnicity. Forty-three per cent of the women and 45% of the men reported that unemployment had resulted in a deterioration of their health. For both women and men, deterioration in health was significantly related to ethnicity and was most common among first-generation migrants.
The Finances–Shame Model and Ill Health As we noted in the introduction, the finances–shame model postulates that ill health during unemployment is a function of both financial hardship and shaming experiences. Figure 3 shows levels of daily health problems, depression and deterioration in health according to levels of financial hardship and shaming experiences and without any adjustments for possible influences from other conditions. We can see that the extent of depression, nervousness and deterioration in health was clearly greatest among those who were both exposed to a greater level of financial hardship and had more experience of being shamed during their unemployment. In the greater–more group, 33% of 87
Unemployment and Health ■ Health Effects Table 4 Depression, Nervousness and Deterioration in Health Due to Unemployment Distributed By Social Background Multiple logistic regression where each individual background factor is controlled with respect to other background factors (percentages in the table are not adjusted for the background factors shown). Depression
Women Total
14
Nervousness
Men
Women
14
22
ns Age
Household type
Ethnicity
< 29 30–39 40–49 50–64
Men
Women
20
43
ns
Men 45
ns
13 13 20 10 ns
11 11 19 14 ns
18 23 20 15 *
11 17 30 22 **
40 47 47 39 ns
35 49 52 40 ns
10 20 11 19 * Nonmigrant background 13 First-generationmigrant 20 Second-generation migrant 11 ns
12 17 9 18 ns 13 15 21 ns
18 25 20 26 ns 20 29 17 ns
16 20 24 22 ns 17 27 11 **
39 47 41 50 ns 33 59 44 ***
37 42 53 54 ns 38 56 31 ***
12 13 21 12 16 ns 21 21 13 22 28 ns
20 23 23 24 16 ns 3 19 9 19 32 ***
16 15 32 18 27 ns 14 16 3 18 14 ***
39 41 40 46 59 ns 38 43 35 52 46 ns
40 43 51 39 54 ns 42 51 40 48 39 ns
Couple without children Single without children Couple with children Single with children
Length of Unemployment 1–12 months 13–24 months 25–36 months 37–48 months > 48 months Education
Deterioriation in health during unemployment
Elementary school Comprehensive school Vocational school Upper secondary school University
13 14 11 16 20 ns 14 26 16 22 33 ns
Note: Level of significance: ***p < .001; ** p < .01; * p < .05; ns = not significant
women and 27% of men had been affected almost daily by depression. In the lesser–fewer group, the corresponding figures are 3% and 2% respectively. In the greater–more group, 45% of women and 35% of men were affected almost daily by nervousness. The corresponding figures in the lesser–fewer group are 5% and 6% respectively. As is also shown in Figure 3, 74% of the women and 64% of the men in the greater–more group suffered a deterioration in health, while the figures for the lesser–fewer group are 23 and 16 respectively. The prevalence of depression, nervousness and deterioration in ill health in the greater–fewer and the lesser–more groups is moderate; for 88
The Finances–Shame Model
example, somewhere between the two others. However, the patterns are ambiguous and point in different directions for women than for men. For women, the higher prevalence of depression and nervousness in the greater–fewer group than in the lesser–more might indicate that financial stress has greater impact than shaming. On the other hand, the prevalence of deterioration in health indicates the reverse; namely, that shaming might have greater impact than financial stress. For men the pattern is different than for women. The prevalence of depression and nervousness is greater in the lesser–more group than in the greater–fewer, indicating that shaming might be of greater importance than financial stress. However, the differences in percentages are in many cases small. Table 5 gives the results of a multiple logistic regression analysis to predict the prevalence of ill health. Adjustments have been made for the effects of the factors age, length of unemployment, household type, education and ethnicity. The odds ratios are given for both women and men. As can be seen, there is considerable support for the basic assumptions in the finances–shame model. A comparison reveals that the odds ratio for the risk of being affected daily by depression, nervousness and deterioration in health was between 10 and 22 times greater for those who were exposed to greater financial
Shaming Experiences More
Fewer
Greater
Financial hardship
Lesser
Figure 3 The finances-shame model, affected daily by depression, affected daily by nervousness and deterioration in health due to unemployment.
89
Unemployment and Health ■ Health Effects Table 5 The Finances–Shame Model and Ill Health Multiple logistic regression, stratified for gender and adjusted for age, length of unemployment, education, household type, and ethnicity with the aid of multiple regression analysis. Type/level of distress
Health outcomes Women
Men
Financial
Shame
Depress.
Nervous
Reduced health
Depress.
Nervous
Reduced health
Greater Greater Lesser Lesser
More Fewer More Fewer
13.8 *** 3.8 ** 3.4 * 1
16.1 *** 5.8 *** 4.0 ** 1
11.9 *** 2.2 ** 2.9 *** 1
21.8*** 4.5 * 7.1 ** 1
10.3 *** 2.7 * 3.8 ** 1
9.6 *** 3.7 *** 5.6 *** 1
Note: Level of significance: *** p < .001; ** p < .01; * p < .05.
hardship and who had more shaming experiences than for those who were exposed to lesser financial hardship and who had fewer shaming experiences. The odds ratio for women for being affected daily by depression was almost 14 times higher and for men 22 times higher in the greater–more group compared to the lesser–fewer group. There was a similar pattern for nervousness and deterioration in health. The odds ratio for being affected on an almost daily basis by nervousness was 16 times higher for women and 10 times higher for men in the greater–more group compared with the lesser–fewer group. The odds radio for deterioration in health was about 10 times higher in the greater–more group compared to the lesser–fewer group. Thus there was a similar pattern across health outcomes. The combination of financial hardship and shaming experiences clearly increased the risk of being troubled by ill health.
Summary and Discussion The aim of the study was to examine self-reported health among unemployed in the light of the finances–shame model. The assumption was that the greater the financial hardship during unemployment and the more experiences of being shamed when unemployed, the more serious the health consequences will be. The lesser the financial hardship and the fewer the shaming experiences, the less serious the effects on health will be. The results of the study lend support to the finances–shame model as it applies to understanding the link between unemployment and ill health. It showed that irrespective of length of unemployment, household type, education, age and ethnicity, ill health was related to the combined effect of two of the conditions we have discussed here. The first of these is the financial hardship that the unemployed are exposed to in varying degrees. The second is the hardship arising from damaged or threatened social relations in which unemployed individuals experience shame as a result of being 90
The Finances–Shame Model
treated (or perceiving themselves to be treated) with contempt. It would seem then, that the stress that occurs during unemployment and that increases the risk of ill health results from a combination of material and social deprivation. Ill health occurs primarily in those unemployed who suffer from a higher level of financial hardship and have more extensive experience of being looked down on in various ways; that is, ‘shamed’, in our terminology. In the cases where the financial hardships are small, or even nonexistent, and the social attitudes from friends and neighbours are supportive rather than condemning, the prevalence of ill health is very limited. The study shows, irrespective of other background factors such as age, length of unemployment, education, household type and ethnicity, strong support for the assumptions of the finances–shame model. But this does not mean that we have been able to establish a causal relation. For this a longitudinal design is needed. In conclusion, the issue which we would like to discuss and speculate on is what might be the deeper social mechanisms that ‘transform’ financial hardship and shame into processes which result in ill health. In general stress theory there is an assumption that the individual’s perception of their social situation can, in certain circumstances, lead to an extraordinary mental and physical mobilisation which, in its turn, can initiate a pathological process. It is possible that the key to a deeper understanding of the social dimension of this stress lies in the role played by emotions. As pointed out by Lazarus (1999), when there is stress there are also emotions, and the reverse often applies, although this is not always the case. The oddity is that two separate literatures have developed, almost as if stress had no bearing on the emotions, and the emotions had no bearing on stress. Therefore there is a need to view stress, emotion and coping as existing in a part–whole relationship, with emotion being the superordinate concept because it includes stress and coping (Lazarus, 1999). In this context, an interesting contribution to a deeper understanding of the role of emotions in the emergence of ill health is that by Tomkins (1979). He maintains that what is usually termed ‘negative stress’ is basically the consequence of ‘backed-up’ emotions. The term ‘backed-up’ is, however, difficult to define. According to Tomkins, every emotion has a free and natural means of expression. However, emotions are subjected to societal cultural control, which means that people learn to suppress their emotions. The suppression of emotions results in what is often experienced is ‘pseudoemotions’ or what Tomkins terms ‘backed-up’ emotions. In this context ‘backed-up’ means that the ‘natural’ emotions are suppressed, delayed or masked. Many of the endocrine changes which, according to Tomkins, are associated with negative stress are the consequence of such backed-up emotions.
91
Unemployment and Health ■ Health Effects
Thus, Tomkins’ claim is that the consequences of a systematic suppression of emotions are profound. Suppression creates backed-up emotions in connection with suppressed anger that can result in endocrinal changes such as raised blood pressure. Much has been written about the psychosocial causes of illness and the role that events in a person’s life — their difficulties and lack of social support — play in the occurrence of physical and mental ill heath. A number of interesting attempts have been made with the object of seeing emotions as a link between body, soul and society (see Freund, 1990). In these attempts the assumption is made that different social conditions influence one another and that this influence may be perceived and embodied emotionally in many different ways. Research on the relation between shame and ill health suggests that experiences of shame may have a powerful, and so far unobserved, significance for the development of heart disease (Scheff, 1990, 1992), depression (Hobilitzelle, 1987), suicide (Shreve & Kunkel, 1991) and attempted suicide (Arcel, 1985). Various emotions have been registered in people who have fallen ill. An examination of the studies shows that the emotions that are most often reported as preceding the occurrence of the symptoms are bitterness, disappointment, depression, anxiety and helplessness (Luborsky, Docherty, & Penick, 1973). Experiences of subordination, inferiority, powerlessness, low social status and poverty might produce emotions like shame and distrust, which in turn might initiate a process that results in ill health by way of, for example, psychoendocrine mechanisms (Wilkinson, 1999). When people become more vulnerable to feelings of shame these feelings are, according to Wilkinson (2002), plausibly one of the most powerful and recurrent sources of the chronic stress that increases people’s vulnerability to a range of infectious and cardiovascular diseases. Wilkinson adds: This provides the most plausible model we have of why low social status, weak social networks, and poor early attachment should be related not only to worse health but also to raised basal cortisol levels and attenuated cortisol responses to experimental stressors (Wilkinson, 2002, pp. 267–268).
Scheff considers shame to be the main emotion, in the sense that emotions such as fear, sorrow, anger and even shame itself only result in pathological processes when they occur in the context of shame; that is, when shame generates further emotional reactions — for example, fear–shame (ashamed of one’s fear), shame–shame (ashamed of one’s shame) and anger–shame (ashamed of one’s anger; Scheff, 1992). In the fear example, shame in itself and anger are suppressed by shame. But not all forms of shame are harmful. It is only, according to Scheff, the unacknowledged shame (i.e., lingering, suppressed shame) that may lead to pathological processes. To summarise, our reasoning thus far implies the assumption of a necessary, although insufficient, condition that would enable us to claim that anxiety or illness of a mental or psychosomatic nature is the combination of emotional activation and emotional suppression; that is, when one 92
The Finances–Shame Model
emotion is activated to suppress another. The main causes of this are disruptions in social relations. The finances–shame model postulates that the combination of a higher level of financial hardship during unemployment and more experiences of shaming by friends and neighbours increases the risk of ill health. We propose that there are primarily two emotions connected with financial hardship. One is fear and the other shame. Living under a high level of financial hardship entails living under a very real threat that one’s income will not be sufficient for the basic necessities, such as housing and food. But financial hardship is probably also, to a considerable degree, associated with a sense of shame. A lack of money is not only a risk in a general sense, it can also create a sense of shame. One is ashamed of one’s financial situation, of not being able to live up to culturally shaped social expectations (Starrin, Forsberg, & Rantakeisu, 1999). But the sense of shame is perhaps most clearly associated with the second part of the finances–shame model and has its origin in how one perceives one is treated by other people, authorities and institutions. Here, quality in both horizontal and vertical relations plays an important role. It is well-known — and research results would seem to be unambiguous on this point — that groups lower down in the social hierarchy run a greater risk of suffering from psychosomatic and mental problems and illnesses. This means from our standpoint that they would also be exposed to circumstances that give rise to the combination activation–suppression in which shame is the emotion that does the suppressing. When this occurs, it does so in a context of shame. In its turn, the context of shame gives an indication of the quality of both the horizontal and vertical relations and ultimately the relation individual–society in the Durkheimian sense. The question that may be raised and on which we shall speculate briefly is what it is in the social position in the social hierarchy, or in the social class position, that enables us to understand the connections. Our suggestion is that the differences can largely be explained by the position of the individual in the social hierarchy. This determines the extent to which the individual (1) has access to resources that they can mobilise to protect themselves from exposure to conditions that may give rise to lingering shame-bound emotional reactions; (2) risks being exposed to conditions which may give rise to lingering shame-bound emotional reactions; and (3) has access to resources which can be used partly for resolving the lingering shame-bound emotional reactions when they have arisen (Starrin et al., 1999). Deprived of a decent income and a decent social status and identity, it seems that the unemployed are consigned by both themselves and society to the bottom of the national hierarchy (Shortt, 1996). Finally, we believe that there is much to suggest that the intermediary link between the conditions in the social environment, characterised by both a financial and social relational vulnerability (finances–shame), and 93
Unemployment and Health ■ Health Effects
psychosomatic reactions is largely a matter of the activation of two emotions, fear and shame. Our suggestion is, further, that this vulnerability gives rise to more serious health effects when these emotions interact. In concrete terms, this means that the emotions of fear and shame are tied to each other in sequences — fear–shame. The basis of this connection is social and this social dimension gives us an indication of the relation between the individual and society in a broad sense. It should, however, be emphasised that this concluding argument is speculative and should be seen as an attempt to understand the visible results that emerge in the finances–shame model. Further research will show whether our ideas are viable.
References Arcel, L.T. (1985). Skam og selvmordsforsog hos en pubertetspige [Shame and suicide attempt of a girl in puberty]. Psyke & Logos, 6, 25–58. Berkman, L.F., & Syme, L.S (1979). Social networks, host resistance and mortality: A nine-year follow-up study of Alameda County residents. American Journal of Epidemiology, 109, 186–204. Breakwell, G.M. (1985). Abusing the unemployed: An invisible injustice. Journal of Moral Education, 14, 56–62. Breakwell, G.M., Collie, A., Harrison, B., & Propper, C. (1984). Attitudes towards the unemployed: Effects of threatened identity. British Journal of Social Psychology, 23, 87–88. Cassel, J. (1976). The contribution of the social environment to host resistance. American Journal of Epidemiology, 104, 107–123. Cooley, C.H. (1902/1922). Human nature and the social order. New York: Scribner’s. Dressner, K.L. (1988). The construction of unemployment as a social problem: Powerlessness, stigmatization and the unemployed. Unpublished doctoral dissertation, University of Californa, Santa Cruz, CA. Eales, M.J. (1989). Shame among unemployed men. Social Science and Medicine, 28, 783–789. Elder, G.H., Jr., & Caspi, A. (1988). Economic stress in lives: Developmental perspective. Journal of Social Issues, 44, 25–45. Ezzy, D. (1993). Unemployment and mental health: A critical review. Social Science and Medicine, 37, 41–52. Freund, P. (1990). The expressive body: A common ground for the sociology of emotions and health and illness. Sociology of Health and Illness, 12, 452–477. Fryer, D. (1986). Employment deprivation and personal agency during unemployment: A critical discussion of Jahoda´s explanation of the psychological effects of unemployment. Social Behaviour, 1(2), 3–23. Fryer, D. (1995). Labour market disadvantage, deprivation, and mental health: Benefit agency? The Psychologist, June, 265–272. Gallie, D. (1994). Are the unemployed an underclass? Some evidence from the social change and economic life initiative. Sociology, 28, 737–757. Goffman E. (1963). Stigma: Notes on the management of spoiled identity. Eaglewood Cliffs, NJ: Prentice Hall. Hammarström, A. (1996). Arbetslöshet och ohälsa: Om ungdomars livsvillkor. [Unemployment and ill-health: On life conditions of youth]. Lund, Sweden: Studentlitteratur.. 94
The Finances–Shame Model Hayes, J., & Nutman, P. (1981). Understanding the unemployed. London: Tavistock. Hobilitzelle, W. (1987). Differentiating and measuring shame and guilt: The relation between shame and depression. In H.B. Lewis (Ed.), The role of shame in symptom formation. Hillsdale, NJ: Lawrence Erlbaum. Horwitz, A.V. (1984). The economy and social pathology. Annual Review of Sociology, 10, 95–119. Ilstad, S. (1989). Okonomisk stress hos familier: Årsaker, virkninger of mestring [Economic stress among families: Causes, effects and copying]. Tidskrift for Norsk Psykologforening, 26, 238–244. Jahoda, M. (1982). Employment and unemployment: A social-psychological analysis. Cambridge, UK: Cambridge University Press. Janlert, U. (1991). Work deprivation and health. Unpublished doctoral dissertation, Karolinska Institutet, Stockholm, Sweden. Jin, R.L., Shah, C.P., & Svoboda, T.J. (1995). The impact of unemployment on health: A review of the evidence. Canadian Medical Association Journal, 153, 529–540. Jönsson, L.R., & Starrin, B. (2000) Ekonomi-skam modellen och reaktioner på arbetslöshet [Finances-shame model and reactions on unemployment]. Socialvetenskaplig tidskrift, 7, 267–284. Kaufman, G. (1993). The psychology of shame: Theory and treatment of shame-based syndroms. London: Routledge & Kegan Paul. Kessler, R.C., Turner J.B., & House J.S. (1987). Intervening processes in the relationship between unemployment and health. Psychological Medicine, 17(4), 949–961. Kieselbach, T. (1999). Unemployment and victimization. [Working paper]. Institute for Psychology of Work, Unemployment and Health, University of Bremen, Germany. Lazarus, R.S. (1999). Stress and emotion: A new synthesis. London: Free Association Books. Luborsky, L., Docherty, J.P., & Penick, S. (1973). Onset conditions for psychosomatic symptoms: A comparative review of immediate observation with retrospective research. Psychosomatic Medicine, 35, 187–204. McFayden, R.G. (1995). Coping with threatened identities: Unemployed people’s self-categorizations. Current Psychology, 14(3), 233–257. Marsden, D. (1975). Workless: Some unemployed men and their families. An exploration of the social contract between society and the worker. Harmondsworth, UK: Pelican Books. Neckel, S. (1991). Status und scham: Zur symbolischen reproduktion sozialer ungleichheit [Status and shame: On the symbolic reproduction of social inequality]. Frankfurt, Germany: Campus. O’Brien, G.E. (1986). Psychology of work and unemployment. Chichester, UK: John Wiley. Orwell, G. (1937). The road to Wigan Pier. London: Victor Gollancz. Rantakeisu, U., Starrin, B., & Hagquist, C. (1997). Unemployment, shame and ill health: An exploratory study. Scandinavian Journal of Social Welfare, 6, 13–23. Rantakeisu, U., Starrin, B., & Hagquist, C. (1999). Financial hardship and shame: A tentative model to understand the social and health effects of unemployment. The British Journal of Social Work, 29(6), 877–901. Retzinger, S.M. (1991). Violent emotions. Shame and rage in marital quarrels. London: Sage. Scheff, T.J. (1990). Microsociology. discourse, emotion, and social structure. Chicago: The University of Chicago Press. Scheff, T.J. (1992). Emotion and illness: Anger, bypassed shame and heart disease. Perspectives on Social Problems, 3, 117–134. 95
Unemployment and Health ■ Health Effects Scheff, T.J. (2003). Shame in self and society. Symbolic Interaction, 26(2), 239–262. Schlozman, K., & Verba, S. (1979). Injury to insult: Unemployment, class and political response. Cambridge, MA: Harvard University Press. Seabrook, J. (1982). Unemployment. London: Quartet Books. Sennett, R., & Cobb, J. (1972). The hidden injuries of class. New York: Vintage. Sennett, R. (1980). Authority. London: Secker & Warburg. Sennett, R. (1999). The corrosion of character. New York: WW Norton. Shortt, S.E.D. (1996). Is unemployment pathogenic? A review of current concepts with lessons for policy planners. International Journal of Health Services, 26(3), 569–583. Shreve, B., & Kunkel, M.A. (1991). Self-psychology, shame, and adolescent suicide: Theoretical and practical considerations. Journal of Councelling & Development, 69, 305–311. Social Rapport [Social Report]. (1997). Stockholm, Sweden: Socialstyrelsen [The National Board of Health and Welfare]. SoS-rapport (1997). 11 röster om socialbidrag [11 opinions about living on socialwelfare]. Stockholm, Sweden: Socialstyrelsen. Spicker, P. (1984). Stigma and social welfare. London: Croom Helm. Starrin, B. (2001). Skammen, självet och den sociala underordningen [Shame, the self and social subordination]. In G. Aronsson & J.C. Karlsson (Eds.), Tillitens ansikten (pp. 48–80). Lund, Sweden: Studentlitteratur. Starrin, B., & Lundberg, B. (1993). Uppsagd från järnverket: En studie av ekonomisk stress och ohälsa [Given notice from ironwork: A study of economic stress and illhealth]. Socialmedicinsk tidskrift, 70(5), 221–228. Starrin B., Rantakeisu U., & Hagquist, C. (1996). Om arbetslöshetens ekonomi och skam [Finances and shame during unemployment]. Socialvetenskaplig tidskrift, 1–2, 91–114. Starrin, B., Rantakeisu, U., & Hagquist, C. (1997). In the wake of the recession: Economic hardship, shame and social erosion. Scandinavian Journal of Work and Environment, 23, 47–54. Starrin, B., Jönsson, L.R., Forsberg, E., Rantakeisu, U. (1998). Varför blir arbetslösheten till en plåga för vissa men inte för andra? [What makes unemployment a torment for some but not for others]. Socialmedicinsk tidskrift, 75(6), 280–290. Starrin, B., Forsberg, E., & Rantakeisu, U. (1999). I arbetslöshetens spår: Ekonomisk stress, skam och ohälsa [In the wake of unemployment: Economic stress, shame and ill health]. In A. Härenstam., U. Lundberg., & E. Lindblad (Eds.), I vanmaktens spår: Om sociala villkor, utsatthet och ohälsa (pp. 21–93). Umeå, Sweden: Borea förslag. Starrin, B., Rantakeisu, U., Forsberg, E., & Kalander-Blomqvist, M. (2000). Understanding the health consequences of unemployment: The finance/shame model. In T. Kieselbach (Ed.), Youth unemployment and health: A comparison of six European countries (pp. 35–42). Opladen, Germany: Leske + Budrich. Starrin, B., Jönsson, L.R., & Rantakeisu, U. (2001). Sense of coherence during unemployment. International Journal of Social Welfare, 10, 107–116. Tomkins, S.S. (1979). Script theory: Differential magnification of affects. In H.E. Howe & R.A. Dienstbier (Eds.), Nebraska symposium on motivation (vol. 26; pp. 201–236). Lincoln, NE: University of Nebraska Press. Underlid, K. (1992). Arbeidslaus [Unemployed]. Oslo, Norway: Det Norske Samlaget. Wadel, C. (1973). Now, whose fault is that! The struggle for self-esteem in the face of chronic unemployment. Toronto, Canada: University of Toronto Press. Warr, P. (1987). Unemployment and mental health. Oxford, UK: Oxford Science Publications. 96
The Finances–Shame Model Whelan, C.T. (1992). The role of income, life-style deprivation and financial hardship in mediating the impact of unemployment on psychological distress: Evidence from the Republic of Ireland. Journal of Occupational and Organizational Psychology, 65, 331–344. Whelan, C.T. (1994). Social class, unemployment and psychological distress. European Sociological Review, 10, 49–61. Wilkinson, R.G. (1999). Income inequality, social cohesion and health: Clarifying the theory. A reply to Muntaner and Lynch. International Journal of Health Services, 29, 525–543. Wilkinson, R.G. (2002), Putting the picture together: Prosperity, redistribution, health, and welfare. In M. Marmot & R.G. Wilkinson (Eds.), Social determinants of health (pp. 256–274). Oxford, UK: Oxford University Press. Willis, P. (1986). Unemployment: The final inequality. British Journal of Sociology and Education, 7, 155–169. Wilson, S.H., & Walker, G.M. (1993). Unemployment and health: A review. Public Health, 107(3), 153–162. Vinokur, A.D., Price, R.H., & Caplan, R.D. (1996). Hard times and hurtful partners: How financial hardship affects depression and relationship satisfaction of unemployed persons and their spouses. Journal of Personality and Social Psychology, 71, 166–179. Zawadski, B., & Lazarsfeld, P. (1935). The psychological consequences of unemployment. Journal of Social Psychology, 6, 224–251.
97
Unemployment and Health ■ Health Effects
98
Chapter
7
Retrenchment and Health Parameters: A Short Report Dimity Pond, Elizabeth Harris, Parker Magin, Amber Sutton, Vanessa Traynor, Kate D’Este and Susan Goode
Background and Aims The association between retrenchment and health has not been well documented in Australia, although there are a number of overseas studies that suggest that retrenchment adversely affects health. Iversen, Sabroe and Damsguard (1989) studied hospital admissions for 887 men before and after shipyard closures at Elsinore, and found that while admissions for accidents and other external causes fell from 23.1 to 13.9 (relative risk falling from 1.33 to 0.46), admissions for circulatory diseases rose from 5.7 to 13.9 per 1000 observed years following the closure (relative risk increasing from 0.8 to 1.6). Morris, Cook and Shaper (1994) found an association between loss of employment and mortality in the British Regional Heart Study. Unemployed or retired men had a relative risk of death of 1.95 after adjusting for confounding variables, and those who retired for reasons other than ill health also had an increased risk of death (relative risk 1.86 after adjusting for confounding variables). Ferrie, Shipley, Marmot, Stansfield and Smith (1995), authors of the now-famous Whitehall study, showed that anticipation of job loss affects self-reported health status even before employment status changes. Avery, Betts, Wittington, Wilson and Reeves (1998) studied a cohort of miners following the 1992 national pitclosure program and found that while unemployed miners were those most at risk of psychological disorder, both employed and unemployed miners were disadvantaged physically and psychologically compared with employed nonminers. Wadsworth, Montgomery and Bartley (1999) demonstrated that unemployment has a persisting effect on the future health and socioeconomic circumstances of individuals, even after reentering employment and remaining employed. 99
Unemployment and Health ■ Health Effects
This chapter describes the physical, psychological and social wellbeing of a cohort of Australian men in the Hunter Valley region of Australia who had been retrenched when a large steelworks closed down. The study hypothesis was that the health status of these men would deteriorate over time following the factory closure, particularly in the unemployed group, despite the different social and economic conditions in Australia. The cohort reported is part of a larger study that also examined the health of spouses and trialled an intervention for general medical practitioners in the region aimed at raising their awareness of the relationship between unemployment and health.
Method A longitudinal cohort study design (with employed controls) was implemented. Data collection was prospective. The study group consisted of former male steelworkers of a large BHP steelworks in Newcastle, a large regional town in New South Wales, Australia. These men were retrenched when the factory closed down in 1999. Members of the three main industrial unions in the region (Australian Workers Union, Australian Manufacturer Workers Union and the Electrical Trades Union) were targeted in an attempt to include a homogeneous group of blue-collar workers and exclude clerical, administrative, management and professional staff. Due to ethical requirements, all recruitment had to occur initially through a third party, with the three unions making an initial call to members to determine whether they were interested in receiving follow-up from the researchers. Following this recruitment phase, participants who agreed were sent an information kit and were followed up by project staff to see if they were interested in participation in the project. Study group participants were interviewed three times over a 12- to 18-month period following the BHP closure. Baseline interviews were conducted between 2 and 6 months after the BHP closure and were followed by 6- and 12-month follow-ups. At each data collection round, participants completed a selfadministered survey and participated in a nurse-administered interview and health check.
Measures In addition to demographic information (age, marital status, and socioeconomic status including home ownership and education level) and measures of physical health (blood pressure, cholesterol level and body mass index [BMI]), the following self-report measures were obtained:
Perceived Stress The Perceived Stress Scale (PSS; Cohen, Kamarck, & Mermelstein, 1983) is a widely used instrument for assessing psychological stress. It captures the degree to which respondents perceive life situations as stressful by measur100
Retrenchment and Health Parameters
ing how unpredictable, uncontrollable and overloaded the respondents consider their lives to have been during the previous month.
Optimism The Life Orientation Test (LOT; Scheier, Carver, & Bridges, 1994) was used to assess individual differences in generalised optimism (the characteristic expectation that good things will happen) versus pessimism.
Social Support The Duke Social Support Index (DSSI; Landerman, George, Campbell, & Blazer, 1989) assesses the social network of the respondent and the support provided by that network using multiple dimensions, including satisfaction with social support (4 items), frequency of social interaction (4 items), perceived social support (7 items), and instrumental support (13 items).
Self-Rated Physical and Mental Health The Medical Outcomes Study Short Form 36 (SF-36; Ware & Sherbourne, 1992) is one of the most widely used self-completion measures of health status. It is a multi-item scale covering a variety of health concepts that relate to both physical and mental health.
Analysis Changes over time in the physical health parameters measured were analysed using repeated measures t tests. Chi-square was used to compare groups by age, follow-up employment status and employment transition, looking at a range of outcomes classified into declined vs stable/improved on each health parameter. Statistical significance was taken to mean that p was < .5. For the purposes of this study, we defined ‘out of the workforce’ as those who stated that they were not looking for work for any reason, including retirement.
Results This paper presents a preliminary analysis of the first (baseline) and third (12-month) data collection rounds for the study group only. Union representatives contacted a total of 646 employees, of whom 408 (68%) agreed to further follow-up by the researchers. Of those followed up, 301 (74%) agreed to receive information about the project. A total of 230 participants (76%) completed a baseline interview, while 166 of these (72%) completed a 12-month follow-up interview. Incomplete 12-month follow-up data were available for a further 9% of participants (n = 21). Age and marital status of participants and nonparticipants were not significantly different, but men who refused participation in the study had worked an average of 2 years less than participants, and were more likely than participants to describe their health as either excellent or poor. The 43 men who were not followed up at 12 months were younger, reported more stress and had better self-reported physical health at baseline. 101
Unemployment and Health ■ Health Effects Table 1 Employment Status at Baseline and Follow-Up Employment status
Baseline
Follow-up
Employed Unemployed Out of workforce*
25% 30% 46%
40% 15% 46%
Note: Compared with baseline: 88% of employed were employed at follow-up 78% of out of the workforce were out of the workforce at follow-up 43% of unemployed were reemployed at follow-up. *Includes 20% retired.
Characteristics of Sample The male study group participants were aged between 22 and 64 years (mean 47 years, SD 9.8), 78% were married or in de facto relationships, and 84% owned their own home. Sixteen per cent were from non-Englishspeaking backgrounds (of whom 36% spoke a language other than English at home) and 22% had no formal professional qualifications. The employment status of the study group participants is reported in Table 1.
Physical Health Parameters Repeated measures t tests showed significant improvements at follow-up in systolic and diastolic blood pressure (t = 3.2, p < .01; t = 2.2, p < .05), cholesterol t = 3.9, p < 0.001), BMI (t = 2.6, p = < .05) and overall physical health (t = –2.3, p < .05), as measured by the physical health subscale of the SF36 (Table 2). No significant changes were reported in stress (general and financial), optimism (the Life Orientation Test), social support (Duke Social Support Index) or the mental health subscale of the SF36. On more detailed examination, the data showed that this population could be divided into those who improved and those who worsened on each of the above parameters (Figures 1–5). In order to explore the associ-
Table 2 Mean, SD and 2-Tailed Significance for Baseline and 12-Month Paired Sample t Tests (Total n = 166) Measure
Systolic BP Diastolic BP Cholesterol BMI Physical health (SF36 subscale)
Baseline
Follow-up
df
p
Mean
SD
Mean
SD
131.0 82.6 5.2 29.1
16.0 8.9 0.9 4.6
127.4 81.0 4.9 28.9
14.7 8.9 0.8 4.7
3.2 2.2 3.9 2.6
160 160 151 161
< .01 < .05 < .001 < .05
4 9.5
10.3
51.0
9.1
–2.3
165
< .05
Note: BP = blood pressure; BMI = body mass index. No significant changes in perceived stress, optimism, social support or mental health.
102
t
Retrenchment and Health Parameters
Figure 1 Diastolic blood pressure: frequency of changes. Mean change –1.5; Median change 0.0; SD 9.02.
ations with membership of these groups, we performed a small number of chi-square tests, dichotomising health outcomes into declined versus stable/improved, and looking at groups by age, follow-up employment status and employment transition.
Figure 2 Systolic blood pressure: frequency of changes. Mean change –3.6; Median change 0.0; SD 14.5.
103
Unemployment and Health ■ Health Effects
Figure 3 Total cholestoral: frequency of changes. Mean change –0.28; Median change –0.25; SD 0.9.
We found a significant association between optimism and follow-up employment status, χ2(2, N = 154) = 9.50, p = .009. The group unemployed at follow-up had lower optimism scores at baseline and follow-up compared to the employed and out-of-the-workforce groups. Men out of
Figure 4 Body mass index: frequency of changes. Mean change –0.26; Median change 0.0; SD 1.3.
104
Retrenchment and Health Parameters
Figure 5 Physical health (SF-36): frequency of changes. Mean change: +1.5; Median change + 2.5; SD 8.3
the workforce at follow-up were more likely to have stable or improved optimism scores than other employed or unemployed groups (Figure 6). There was also a significant association between optimism and employment transition, χ2(2, N = 154) = 6.54, p = .038, with men employed at both time periods more likely to show a decline in optimism than other employment transition groups, although they still maintained the highest mean optimism scores at both time periods (Figure 7).
Figure 6 Optimism by follow-up employment. Employment status at 12-month follow-up by optimism change (declined or not), χ2(2, N = 154) = 9.50, p < .01.
105
Unemployment and Health ■ Health Effects
Figure 7 Optimism by employment transition. Employment transition by optimism change (declined or not), χ2(2, N = 154) = 6.54, p < .05. Note: E-E = Employed at baseline and 12-month follow-up; OOW-OOW = Out of the workforce at baseline and 12-month follow-up; Other = All other employment transitions.
Similarly, there was a significant association between physical health change and employment transition group, χ2(2, N = 166) = 6.35, p = .042. Mean scores showed that all employment transition groups improved in physical health overall. Men out of the workforce at both periods started with poorer physical health than the other groups and improved by a
Figure 8 Physical health by employment transition. Employment transition by physical health change (declined or not), χ2(2, N = 166) = 6.35, p < .05. Note: E-E = Employed at baseline and 12-month follow-up; OOW-OOW = Out of the workforce at baseline and 12-month follow-up; Other = All other employment transitions.
106
Retrenchment and Health Parameters
greater amount, but still had mean scores which were lower than those of the other groups at 12-month follow-up (Figure 8). No other significant associations were found.
Discussion This study failed to support the hypothesis that a group of retrenched factory workers would suffer from deteriorating health parameters over time. In fact, all significant changes in health parameters overall were changes towards improvement. This might be explained by a number of factors. This particular factory closure was planned and supported by significant payouts as well as strong community support for the workers. This may have led to a lower decline in health status than in other studies. In addition, factory work is dangerous and physically arduous. Iversen et al. (1989) showed that hospital admissions for accidents and other external causes may fall after a works closure, while Avery et al. (1998) showed that employed nonminers were better off physically than employed miners. In the same way, health status may improve after retrenchment from a difficult workplace. Nevertheless, the overall trends conceal a large group whose health deteriorated over the 12 months of the study. Subgroup analysis using chisquare tests showed a little about this group, and also some factors associated with health improvement. Those who were unemployed at both time periods showed lower optimism scores at both measurement points than the other groups. This suggests some association between optimism and unemployment, which may of course be mediated by other factors. The association could be used as a marker of poor employment outcomes by professionals working with this group. The group who remained unemployed showed the most improvement in health status, although it was still poorer than other groups at both time points. The improvement may have been due to some adaptation to the status of being unemployed, or because time away from the works meant time away from accidents. The overall poor health status of this group is supported by the literature, which shows a relationship between unemployment and poor health (Harris & Morrow, 2001). This relationship is also worth emphasising to professionals caring for this group. Men out of the workforce at both time periods were more likely to improve in optimism and remain stable or improve in physical health than other employment groups. Clearly, for this group, being out of the workforce benefited health.
Limitations of the Study This was a relatively small study of 166 men who were followed up over a 12-month period following retrenchment from the BHP steelworks in Newcastle, Australia. The study commenced after closure of the works, 107
Unemployment and Health ■ Health Effects
and failed to capture the health status of the participants before closure. This means that any depression of health status due to the threatened closure is not captured in these data. The small number puts limits on the power of statistical testing, in particular for within-group comparisons. Subgroup analysis is only just beginning and needs further work. Because this sample only represented a small fraction of the original cohort of retrenched workers, it is difficult to generalise these findings to the full cohort. Because of the planned nature of the closure, particularly good payouts and the community support provided to these men, it is difficult to generalise these findings to other cohorts in a less supportive context.
Conclusion This study showed an overall improvement in health status of a cohort of retrenched men over a 12-month follow-up period. The experience of the men within the group in both health and employment terms was heterogeneous. Subgroup analysis showed that men out of the workforce throughout the study period improved in optimism and were more likely to remain stable or improve in physical health, while men who were unemployed at both time periods had poorer health and significantly lower optimism than the other groups.
References Avery, A.J., Betts, D.S., Wittington, A., Wilson, S.H., & Reeves, J.P. (1998). The mental and physical health of miners following the national pit closure program: A cross sectional survey using GHQ-12 and SF-36. Journal of Public Health, 112(3), 169–173. Cohen, S., Kamarck, T., & Mermelstein, R. (1983). A global measure of perceived stress. Journal of Health and Social Behavior, 24, 385–396. Ferrie, J. E., Shipley, M.J., Marmot, M.G., Stansfeld, S., & Smith, G.D. (1995). Health effects of anticipation of job change and non-employment: Longitudinal data from the Whitehall II study. British Medical Journal, 311, 1264–1269 Harris, E., & Morrow, M. (2001). Unemployment is a health hazard: The health costs of unemployment. The Economic and Labour Relations Review, 12(1), 18–31. Iversen, I., Sabroe, S., & Damsguard, M.T. (1989). Hospital admissions before and after shipyard closure. British Medical Journal, 299, 1073–1076. Landerman, R., George, L.K., Campbell, R., & Blazer, D.G. (1989). Alternative models of the stress-buffering hypothesis. American Journal of Community Psychology, 17, 625–542. Morris, J.K., Cook, D.G., Shaper, A.G., (1994). Loss of employment and mortality. British Medical Journal, 308, 1135–1139. Scheier, M.F., Carver, C.S., & Bridges, M.W. (1994). Distinguishing optimism from neuroticism (and trait anxiety, self-mastery and self-esteem): A reevaluation of the Life Orientation Test. Journal of Personality and Social Psychology, 67, 1063–1078. Wadsworth, M., Montgomery, S., & Bartley, M. (1999). The persisting effect of unemployment on health and social well-being in men early in working life. Social Science and Medicine, 48, 1491–1499. Ware, J.E., & Sherbourne, C.D. (1992). The MOS 36-item short form survey (SF-36): I. Conceptual framework and item selection. Medical Care, 30, 473–483. 108
Chapter
8
Suicidal Ideation in the Long-Term Unemployed: A Five-Year Longitudinal Study Bjørgulf Claussen
Does unemployment lead to suicide? This question has been much discussed since the end of the ‘golden age’ of employment in 1973 (Short, 1996). After reviewing over 100 studies, Platt (1984) found that suicide and parasuicide were far more common among the unemployed than among the employed. The present study aimed to investigate this question. Two main mechanisms were examined: (a) the selection hypothesis, that people with psychiatric symptoms and increased risk for suicide may be selected into long-term unemployment; and (b) the causal hypothesis, that suicidal tendencies may develop more easily, and be implemented more rapidly, as a result of unemployment. Data are scarce, however, due to a lack of longitudinal studies, and it is difficult to formulate firm hypotheses. Martikainen and colleagues (1990) conducted a 5-year longitudinal study using census data from all Finnish men aged 30 to 54 years. They controlled for health-related selection into unemployment using data on the number of sick days and use of reimbursible medicines during their final year of employment prior to becoming unemployed. Suicide was found to be 92% more common among unemployed men than among the employed. Although the study has been criticised for having inadequately controlled selection, the results nevertheless tend to support the causation hypothesis. The present chapter is based on two follow-up studies based on a community sample of 310 registered long-term unemployed individuals from Grenland, Norway (Claussen, 1999). Participants answered four items concerning suicidal ideation in the General Health Questionnaire (GHQ; Goldberg, 1972), and underwent two clinical examinations. This sample was too small to allow analyses of suicide itself. Suicidal ideation has a strong association with suicide and parasuicide (Brent et al., 109
Unemployment and Health ■ Health Effects
1986), however, and should be taken seriously, both clinically and epidemiologically (Rudd, 1989). Therefore, it was the focus of investigation in the present study. Three main questions were addressed in the present study: (1) How common is suicidal ideation among unemployed people? (2) Are suicidal unemployed subjects selected away from new jobs? (3) How does reemployment influence suicidal ideation? The overall aim was to gather knowledge that would be useful in guiding the development of clinical and preventive measures in primary health care to assist the long-term unemployed.
Method Participants A random sample of all persons who had been registered as unemployed for more than 12 weeks, and who were between the ages of 16 to 63 years in Grenland, Southern Norway, were invited to a routine health examination in 1988 (Claussen, Bjørndal, & Hjort, 1993). A period of 12 weeks, unemployment was regarded as ‘long-term’ in Norway at that time. Of 374 invited persons, 310 (83%) attended, 291 returned for a follow-up in 1990, and 228 answered a postal questionnaire in 1993. The present study was based on data from the 210 participants who provided a complete data set for all three waves of data collection.
Materials Psychometric Tests All examinations included three self-report psychometric tests: Psychological distress. The Hopkins Symptom Checklist (HSCL; Derogatis, Lipman, Rickels, Uhlenhuth, & Covi, 1974) is a 30-item version of the Symptom Checklist-90, and the three subscales are similar to the corresponding ones in the 58-item version (Derogatis et al.,1974). Each item is scored on a 4-point scale, and was used in the present analyses both as a continuous measure of mental distress and as a dichotomous measure with a cut-off point of 1.75. Excessive alcohol use. The Alcohol Use Disorder Identification Test (AUDIT; Claussen & Aasland, 1999) is a 10-item test which taps signs of harmful drinking. Each item uses a 4-point scale, giving a total possible score between 0 and 40.
Suicidal ideation. The GHQ has been frequently used to assess psychological distress among unemployed people (Goldberg & Williams, 1988). The 28-item version used in the present study includes the following four questions about suicidal ideation: (1) Have you recently felt that life isn’t worth living? (2) Have you recently had thoughts about the possibility that you might do away with yourself? (3) Have you recently found 110
Suicidal Ideation in the Long-Term Unemployed
yourself wishing you were dead and away from it all? (4) Have you recently found that the idea of taking your own life kept coming into your mind? Responses are 0 (not at all or definitely not), 1 (no more than usual or I don’t think so), 2 (rather more than usual or has crossed my mind), and 3 (much more than usual or definitely). For dichotomous ‘case-finding’ the cut-off point lies between 1 and 2, that is, a score of 2 or 3 is taken to indicate presence of the symptom in question. Goldney, Winefield, Tiggemann, Winefield and Smith (1989) have summarised the dichotomous answers to the four GHQ items to make an index of suicidal ideation (the Suicidal Ideation Index, or SII), with scores ranging from 0 (no positive answers) to 4 (positive answers on all four items).
Clinical Examinations The 1988 and 1990 examinations included a routine clinical check-up by the author for the purpose of trialing a routine health examination of the long-term unemployed for preventive and therapeutic reasons. This has been followed up in Norway (Ytterdal, 1999), and has recently been trialled in French occupational medicine (Fréal, 1999). The first examination lasted an average of 61 minutes (range 20–105 minutes). Suicidal prevention was an important aim. The psychological tests were the initial starting point for these examinations. One person was clinically judged to be suicidal. Lack of work seemed to be the underlying reason, and the author was able to negotiate with the labour office to procure him a work opportunity post. A total of 26 individuals (16 men and 10 women, 8% of the sample) showed clinically significant signs of suicidal ideation. Four were referred to a psychologist, and the others to their general practitioner. All these individuals were advised to take suicidal ideation seriously according to recommendations issued by the Norwegian Ministry of Health (1993). At the postal questionnaire in 1993, 29 individuals were deemed to be suicidal according to their SII scores. They were telephoned, and given the same advice as above. None of them believed that their suicidal ideations were serious, and none reacted negatively to the call.
Sociodemographic Variables When examining reemployment in relation to health, it is important to take account of potential confounding variables, such as those postulated by theories of human capital (e.g., age and education) and the segmented labour market (e.g., gender, socioeconomic status and social network; Claussen et al., 1993). In the present study, the duration of participants’ education ranged from 7 to 18 years. Socioeconomic status was measured using an index based on occupation (1 = unskilled workers and 5 = high-level salaried employees, self-employed and freelance professions). The index for social network was based on six questions about parents, spouse, children and 111
Unemployment and Health ■ Health Effects Table 1 Differences Between Respondents and Nonrespondents at the 1993 Postal Follow-Up: Sociodemographic and Medical Characteristics Respondents (n = 210)
Nonrespondents (n = 100)
p
Gender (percentage females) Age (average years, range 17–64) Socioeconomic status (range 1–5) Education (range 7–18 years) Social network (range 1–5) Employment Commitment Test
49 33.80 2.32 10.40 10.20 4.32
28 31.70 2.20 9.90 9.40 4.27
.00 .14 .43 .04 .03 .66
Somatic symptoms (% cases) Anxiety (HSCL-30) (% cases) Depression (HSCL-30) (% cases) Alcohol use disorder (% cases) Suicide ideation (range 0–4)
15 17 21 18 0.39
18 17 22 23 0.37
.55 .94 .88 .32 .88
Somatic diagnosis (percentage cases) Psychiatric syndrome (percentage cases) Personality disorder (percentage cases)
31 35 12
29 30 9
.76 .40 .34
Social and medical characteristics 1988
friends, giving possible scores from 0 to 20. Another aspect of people’s behaviour in the labour market is their personal commitment to paid work; this was assessed using the Employment Commitment Test (Warr, Cook, & Wall, 1979). Scores on these variables are shown in Table 1.
Analyses The chi-square statistic and independent groups’ t tests were employed for analyses, according to whether dependent variables were categorical or continuous. Because the distribution of the SII was markedly bimodal it was dichotomised according to the cut-off point, (i.e., 0 = none, and 1 = one or more positive answers to the four suicidal items), and used as a dependent variable in logistic regressions. Predictors included the seven variables described above (age, gender, education, socioeconomic status, social network, living alone and employment commitment), and the continuous HSCL-28 scales for anxiety and depression. The selection hypothesis, that sick persons are selected to long-term unemployment, was tested by examining the odds of reemployment across levels of suicidal ideation, while controlling for other predictors of reemployment. This is not entirely satisfactory because selection by job loss may be the main mechanism for the primary job loss, but the latter is nearly impossible to measure. The causal hypothesis, namely that unemployment leads to health loss, is usually measured as recovery by reemployment on an individual level, as was done here.
112
Suicidal Ideation in the Long-Term Unemployed Table 2 Logistic Regression Predicting Suicidal Ideation in a Community Sample of Long-Term Unemployed in 1988: Influence of Demographic, Social and Psychological Variables (n = 210) Predictors Age Gender Education Socioeconomic status Social network Living alone Employment commitment Anxiety (HSCL-30) (range 1–4) Depression scale (HSCL-30) (range 1–4) Alcohol abuse (range 0–38) Constant
B
p-value
–0.09 0.85 0.05 0.22 –0.02 –0.24 0.12 0.01 0.03 0.01 –12.91
.00 .13 .73 .26 .82 .72 .00 .02 .00 .82 .00
Results As mentioned previously, respondents in the present analyses were the 210 persons who completed all three waves of data collection, that is, 68% of the 310 attendees in 1988 (Table 1). Compared with respondents, nonrespondents were more likely to be men, and to have less education and smaller social networks. They had about the same prevalence of mental disorders, including suicidal ideation, and medically diagnosed diseases were equally prevalent in both groups. Logistic regressions showed that age, employment commitment, anxiety and depression each exerted an independent influence on suicide ideation (Table 2). Bivariate analyses showed that, of participants aged under 30, 18% showed suicidal ideation; this compared with 21% of respondents aged 31 to 49, and only 3% of respondents over 50. In 1988, when all respondents were classified as long-term unemployed, 83% demonstrated no suicidal ideas, 12% had from one to three positive answers and 5% scored positively on all four items (Table 3). Suicidal ideation was a little less common in the whole sample in 1990 than at baseline in 1988, but in 1993 it was the same again as at baseline.
Table 3 Suicidal Ideation in a Community Sample of Long-Term Unemployed: Scores in 1988, 1990 and 1993. Suicide Ideation Index Scores
0
1
2
3
4
n
% 1988 1990 1993
83.0 89.8 83.2
7.5 5.1 5.0
2.0 1.5 5.0
2.4 1.1 2.7
5.1 2.5 4.1
296 275 222
113
Unemployment and Health ■ Health Effects Table 4 Suicidal in a Community Sample of Long-Term Unemployed: Mean SII Scores in 1988 and at Follow-Up in 1990 and 1993 1990 Suicide ideation (SII Scores)
1993
Unemployed (n = 162)
Reemployed (n = 113)
Unemployed (n = 139)
Reemployed (n = 79)
0.38 0.26 —
0.40 0.15 —
0.36 0.22 0.50
0.45 0.16 0.13*
1988 1990 1993 Note: * p < .01.
In 1993 36% of the follow-up sample were employed, 38% of the women and 35% of the men, about the same as at the 3-year follow-up in 1990. There seemed to be a reduction in suicidal ideation by reemployment at the group level, but only 1993 results reached statistical significance (Table 4). Reemployment seemed not to be influenced by former suicidal ideation. Selection to reemployment was tested using logistic regression, with the dichotomised suicide ideation scores as predictors of employment status 5 years later (Table 5). Suicidal ideation did not predict reemployment, but education did. The same results emerged when the continuous SII index was used (figures not shown). Change in suicidal ideation was measured using the score in 1993 as a dependent variable and the score in 1988 as a predictor (Table 6). Reemployment reduced the odds of having suicidal thoughts (the inverse value of eβ) to 25% of those for the still-unemployed. The same magnitude of recovery was found when using the continuous SII index (figures
Table 5 Selection Into Reemployment: Multivariate Predictors of Employment Statusa in 1993 (Logistic Regression) Predictors b
Suicidal ideation 1988 Age Gender Education Socioeconomic status Social network Living alone Employment commitment Anxiety (HSCL-30) Depression (HSCL-30) Alcohol abuse Constant
B
p
0.33 –0.01 –0.29 0.17 0.15 0.10 0.12 0.00 –0.00 –0.01 0.03 2.13
.47 .73 .46 .04 .27 .12 .77 .81 .59 .15 .37 .30
Note: a: 0 = unemployed, 1 = reemployed; b: 0 = suicidal ideation not present, 1 = suicide ideation present.
114
Suicidal Ideation in the Long-Term Unemployed Table 6 Recovery by Reemployment: Multivariate Predictors of Suicidal Ideation in 1993a (Logistic Regression) Predictors b
Employment status 1993 Suicidal ideation 1988a Age Gender Education Socioeconomic status Social network Living alone Employment commitment Anxiety (HSCL-30) Depression (HSCL-30) Alcohol abuse Constant
B
p
–1.39 0.83 –0.01 –0.02 0.01 –0.06 0.01 0.86 0.02 0.00 0.00 0.01 –6.04
.01 .16 .61 .98 .93 .79 .87 .20 .52 .53 .61 .82 .06
Note: a: 0 = suicidal ideation not present, 1 = suicidal ideation present; b: 0 = unemployed, 1 = reemployed.
not shown). Thus, dichotomising the index did not lead to significant loss of information.
Discussion The consequences for Norwegians of being registered at the labour office for three months or more appear to be severe. Only 37% of the participants in the present study were employed 5 years after an initial spell of long-term unemployment in 1988. Only one in eight had been employed during the intervening 4 years. The present sample may be compared to those who are usually called long-term unemployed people in Europe today. Attrition bias may have influenced the present results. The 24% who were lost to the 5-year follow-up were often males (Table 1). Suicidal ideation is more common in men than in women (Vandivort & Locke, 1979), but in our sample the prevalence was similar in both sexes (Claussen, 1998), as has been found elsewhere (Rudd, 1989; Schweitzer, Kalyich, & McLean, 1995). Otherwise, the differences between participants and nonparticipants were small, indicating that the effect of attrition bias was probably minor. The original sample was representative of all registered unemployed in Norway (Claussen et al., 1993). They were relatively young, 52% being 18 to 30 years old, and only 12% being aged 50 or more. Thus, suicide prevention is important because young people are more likely to commit suicide than older people (Rudd, 1989). At all three waves of data collection the participation rate was relatively high, probably because the author had been a general practitioner in the area for 10 years and was well known to the local population. Thus, the present study is one of the few community studies of the unemployed that 115
Unemployment and Health ■ Health Effects
has obtained more personal information from participants than is available from census data. However, since only 210 individuals participated, the extent to which the results can be generalised to the population as a whole is unclear. The General Health Questionnaire is a well-validated measure of mental disorders, and is much used (Goldberg & Williams, 1988). The Suicidal Ideation Index showed good concurrent validity as a scale for degrees of suicidal ideation in the construct validation (Goldney et al., 1989). In the present study, the level of internal consistency was good (Chronbach’s alpha 0.860–0.958 for the three years). A strong correlation with the HSCL depression scale showed good content validity. In an Australian student sample with a mean age of 20 years, Goldney et al. (1989) found a prevalence of 11% suicidal ideation. We found 17% in an older sample of unemployed people, and 21% among those aged 17 to 24 years. These figures indicate that suicidal ideation is more common among unemployed people than in a population sample of young adults. This result has serious implications; first, because unemployed individuals are a known risk group for suicide; and second, because labour authorities place high demands on registered unemployed people in Norway, making them a positive selection of all people without work (Claussen et al., 1993). Thus, the present sample was quite healthy and demonstrated high motivation for work in 1988. The selection hypothesis was not supported by the present study of suicidal ideation. On the other hand, significant recovery upon reemployment was demonstrated (see Table 4 for crude rates and Table 5 for adjusted rates). This suggests that, over the long-term, a lack of paid work may lead to an increase in thoughts of suicide, while reentry into employment may reverse this tendency. This effect remained, even after controlling for anxiety and depression, showing that reemployment may have a direct effect in reducing suicidal ideation, independently of the effects of existing psychological distress (Paykel, 1976). Our results indicate that measures to prevent suicide should be taken by labour market authorities. Following Norwegian Ministry of Health (1993) recommendations, resource centres for suicide prevention were set up in 1994. A telephone helpline was set up for labour office staff to contact when they become aware of suicidal ideation in a client, or when they perceive that suicide is being used as a threat to obtain access to scarce resources such as educational programs. It is well documented that suicide prevention centres assist in the prevention of suicides (Norwegian Ministry of Health, 1993). They are also important in reducing pressure on labour office staff. The study indicates that suicidal ideation among unemployed people is predicted by mental disorders, alcohol abuse and high employment commitment (cf. Claussen, 1998). Reducing psychological distress among the unem116
Suicidal Ideation in the Long-Term Unemployed
ployed may be a means of primary prevention. This is a task for society in general and much is done in western European countries. Sport groups (Fasting & Grønningsæter, 1986) and psychological support programs (Vinokur, van Ryen, Gramlich, & Price, 1991) have documented effects. Educational and vocational programs may also have some beneficial effects, but providing genuine opportunities for reemployment appears to be the best method for reducing suicide ideation in our highly motivated sample.
Conclusion Long-term unemployment appears to lead to suicidal ideation, and may also lead to suicide. It seems to be the dominant reason for the high prevalence of suicidal ideation among unemployed people. Given that preventive measures are possible and effective, it is imperative that they be implemented as soon as possible.
Acknowledgments This study was supported by the Norwegian Research Council for Science and the Humanities, and by the County Labour Authority of Telemark. Thanks are also due to Professor Lars Mehlum, Head of Research at the National Research Centre for Suicide and Suicide Prevention.
References Brent, D.A., Kalas, R., Edelbrock, C., Costello, A.J., Dulcan, M.K., & Covenor, N. (1986). Psychopathology and its relationship to suicidal ideation in childhood and adolescence. Journal of American Academy of Child Psychiatry, 25, 666–673. Claussen, B., Bjørndal, A., & Hjort, P.F. (1993). Health development in a community sample of long-term unemployed. Journal of Epidemiology and Community Health, 47, 14–18. Claussen, B. (1998). Suicidal ideation among long-term unemployed in a five-year follow up. Acta Psychiatrica Scandinavica, 98, 480–486. Claussen, B., & Aasland, O.G. (1999). Alcohol disorders and re-employment in a fiveyear follow up of long-term unemployed. Addiction, 94, 133–138. Claussen, B. (1999). Health and re-employment in a five-year follow up of long-term unemployed. Scandinavian Journal of Social Medicine, 18, 1–7. Derogatis, L.R., Lipman, R.S., Rickels, K.L., Uhlenhuth, E.H., & Covi, L. (1974). The Hopkins Symptom Checklist (HSCL): A measure of primary symptom dimensions. In D. Pichot (Ed.), Psychological measurements in psychopharmacology. Basel, Switzerland: Karger Verlag. Fasting, K., & Grønningsæter, H. (1986). Unemployment, trait anxiety and physical exercise. Scandinavian Journal of Sports Science, 8, 99–103. Fréal, S. (1999). A medical report form concerning an examination at job end. International Archives of Occupational and Environmental Health, 72(Suppl.), 46–48. Goldberg, D.P. (1972). The detection of psychiatric illness by questionnaire. London: Oxford University Press. Goldberg, D., & Williams, P. (1988). A user’s guide to the General Health Questionnaire. Windsor, UK: NEFR-Nelson. 117
Unemployment and Health ■ Health Effects Goldney, R.D., Winefield, A.H., Tiggemann, M., Winefield H.R., & Smith, S. (1989). Suicidal ideation in a young adult population. Acta Psychiatrica Scandinavica, 79, 481–489. Martikainen, P.T. (1990). Unemployment and mortality among Finnish men, 1981–85. British Medical Journal, 301, 407–411. Norwegian Ministry of Health (1993). Nasjonalt program for forebygging av selvmord i Norge [National programme for prevention of suicide in Norway]. Oslo, Norway: Norwegian Ministry of Health. Paykel, E.S. (1976). Life stress, depression and attempted suicide. Journal of Human Stress, 2, 3–12. Platt, S. (1984). Unemployment and suicidal behaviour: A review of the literature. Social Science and Medicine, 19, 93–115. Rudd, M.D. (1989). The prevalence of suicidal ideation among college students. Suicide and Life Threat Behavior, 19, 173–183. Schweitzer, R., Klayich, M., & McLean, J. (1995). Suicide ideation and behaviours among university students in Australia. Australian and New Zealand Journal of Psychiatry, 29, 205–218. Short, S.E.D. (1996). Is unemployment pathogenic? A review of current concepts with lessons for policy planners. International Journal of Health Services, 26, 569–589. Vandivort, D.S., & Locke, B.Z. (1979). Suicide ideation: Its relation to depression, suicide and suicide attempts. Suicide and Life Threat Behavior, 9, 205–218. Vinocur, A.D., van Ryen, M., Gramlich, E.M., & Price, R.H. (1991). Long-term followup and benefit-cost analysis of the jobs program: A preventive intervention for the unemployed. Journal of Applied Psychology, 76, 213–219. Warr, P., Cook, J., & Wall, T. (1979). Scales for the measurement of some work attitudes and aspects of psychological well-being. Journal of Occupational Psychology, 52, 129–148. Ytterdal, T. (1999). Routine health check-ups of long-term unemployed in Norway. International Archives of Occupational and Environmental Health, 72(Suppl.), 38–9.
118
Chapter
9
Quality of Life of the Employed and Unemployed in Turkey: A Comparative Field Study Yucel Demiral, Alp Ergör, Belgin Unal and Semih Semin
Unemployment and employment have been shown to adversely affect a variety of health outcomes, including mortality, morbidity and self-rated health. Research has consistently shown that unemployed people report diminished levels of psychological health and wellbeing in comparison to their employed counterparts (Brenner & Levi, 1987; Hammarström & Janlert, 2000; Graetz, 1993; Kessler, Turner, & House, 1988). However, no clear cause-specific effects have been proven (Ezzy, 1993; Hammarström, Janlert, & Theorell, 1988). It has also been suggested that economic boom and recession periods might play an important role in the association between health and unemployment (Bartley & Owen, 1996). The experience of the unemployed varies considerably, depending on such factors as age, gender, income, social support and length of unemployment. In addition, not all types of employment are preferable to unemployment (Svensson, 1987). Friis, Carter and Edling (1998) have reported that for some groups, such as miners, unemployment could be beneficial because of the cessation of potentially harmful occupational exposures. Moreover, a vast majority of studies have taken place in industrialised countries, in which social rights and social welfare systems are much more robust than in developing countries. In developing countries, the causes and health consequences of unemployment may therefore be quite different from those in industrialised countries. Scarcity of jobs in developing countries may prevent one from working and this, combined with the absence of welldeveloped welfare systems, means that unemployment connotes absolute poverty. On the other hand, in industrialised countries what individuals expect to gain from having a job and/or their past job experiences may play a key role in determining how unemployment is experienced and the health 119
Unemployment and Health ■ Health Effects
effects of unemployment. Therefore, the relationship between unemployment and health should be considered in a country-specific context. Currently, most European countries, as well as developing countries, are facing high unemployment rates. Economic instability has forced governments to encourage more flexible working arrangements, which often means that workers are employed in relatively insecure and/or poorly paid jobs. Underemployment (working at a lower capacity than one is capable of) and overemployment (working excessively long hours) are also becoming much more widespread in both developed and developing countries. Apart from the increase in unemployment rates due to a prolonged economic crisis, underemployment has been a long-standing problem in Turkey (Bulutay, 1995). The decline in agricultural employment during the last 20 years may have helped to bring about an increase in unemployment. Unemployment rates have been ranging from 6.9% to 8.7% in Turkey during the past 10 years. The urban unemployment rate is higher than the rural one. Women’s unemployment rate is higher than men’s, and unemployment is more widespread among youngsters (Bulutay, 1995). This study is an introductory exploration of the effects of unemployment on physical and mental health in Turkey.
Methods Participants The sample was drawn from a pool of unemployed people who registered at the official employment office in Izmir, Turkey, during a 5-day period in April 2000; and from the entire workforce at a paper processing plant in Izmir, one of the large industrial cities in Turkey, with a population of 2.5 million. Only persons who had been diagnosed with a chronic disease or who were disabled or handicapped were excluded from the study. The final sample included all of the unemployed registrants who met the inclusion criteria (n = 201), and 144 out of 151 (95.4%) factory employees, giving a total of 345 participants. Participants’ ages ranged from 17 to 59 years, with a mean age of 28.9 ± 7.2 years. The majority of participants had graduated from primary school (53%). Most were male (84%) and in blue-collar occupational classes (see Table 1). Materials The Short Form-36 (SF-36; Ware & Sherbourne, 1992) health questionnaire was used to assess physical and mental health. This widely used instrument has been shown to be suitable for assessing population health status (Brazier et al., 1992; Dijkers, 1999; Coons, Rao, Keininger, & Hays, 2000; Garratt, Ruta, Abdalla, Buckingham, & Russell, 1993; Jenkinson, 120
Quality of Life in the Employed and Unemployed in Turkey Table 1 Descriptive Statistics of the Workers and Unemployed Workers (n = 144)
N Age groups
Gender Education
Marital status Occupational class
≤ 24 28 25–34 63 35–44 43 ≥ 45 10 Women 18 Men 126 Primary school 71 High school 24 Technical high school 29 University 20 Married 95 Single/divorced/ widowed 49 White-collar 24 Skilled blue-collar 34 Unskilled blue-collar 86
Unemployed (n = 201)
Total (n = 345)
%*
N
%*
N
%*
19.4 43.8 29.9 6.9 12.5 87.5 49.3 16.7 20.1 13.9 66.0 34.0 16.7 23.6 59.7
77 107 14 3 39 162 112 28 49 12 89 112 27 56 118
38.3 53.2 7.0 1.5 19.4 80.6 55.7 13.9 24.4 6.0 44.3 55.7 13.4 27.9 58.7
105 170 57 13 57 288 183 52 78 32 184 161 51 90 204
30.4 49.3 16.5 3.8 16.5 83.5 53.0 15.1 22.6 ˆ9.4 53.3 46.7 14.8 26.1 59.1
Note: * Column percentage.
Layte, Coulter, & Write, 1996; Marmot, 1998; McHorney, Ware, & Raczek, 1993; Miilunpalo, Vuori, Oja, Pasanen, & Urponen, 1997; Tarlov et al., 1989; Testa & Simonson, 1996; Ware & Sherbourne, 1992). The eight subscales are designed to assess physical, emotional, social and role functioning; bodily pain; vitality; and general and mental health perceptions. For most items, respondents are asked to evaluate their own health during the past 4 weeks. Total SF-36 scores range from 0 to 100, with 0 showing the worst possible health and 100 indicating the absence of any impairment. In our study the internal consistency reliability of the subscales was between 0.57 and 0.85. Physical Component Scale scores (PCS) and Mental Component Scale scores (MCS) were further calculated as recommended by Ware, Kosinski and Keller (1994). We used PCS and MCS as continuous variables for the physical and mental health outcomes. Procedure The unemployed participants were interviewed in the employment office and workers were interviewed in the factory. Three occupational classes were identified: white collar, skilled blue collar and unskilled blue collar. Participants were asked if their economic status was fair, average or poor. The unemployed group was asked about the duration of their unemployment and about their employment history (i.e., whether they had previously been employed or were looking for their first job). Workers were asked (1) whether they were satisfied with their current job; and (2) whether they were seeking a different job. Employed and unemployed groups were compared regarding mean physical and mental health scores 121
Unemployment and Health ■ Health Effects Table 2 Mean PCS and MCS Scores of Employed and Unemployed Participants
PCS MCS
Worker n = 136
Unemployed n = 185
t
p
52.6 ± 5.5 46.3 ± 8.6
54.8 ± 5.4 47.3 ± 9.1
3.56 0.98
< .001 .33
using Mann–Whitney U tests. The Kruskal–Wallis test was used to compare mean PCS and MCS scores in financial strain groups and in employed and unemployed participants.
Results Regarding employment status, there was no difference in MCS mean scores between the employed and unemployed groups. On the other hand, the unemployed group had significantly higher PSC mean scores than their employed counterparts (Table 2). As shown in Tables 3 and 4, subjective financial strain in both groups was significantly related to MCS scores, such that participants with financial difficulties obtained significantly lower MCS scores than those who were better off. The data in Tables 3 and 4 also suggest that the negative association between financial strain and MCS scores may have been stronger for unemployed participants than for employed participants. The mean MCS scores of unemployed participants who had been looking for a job for more than 6 months were significantly lower than those who had been looking for a job for less than 6 months (Table 5). Among the workers, those who were looking for a new job obtained significantly lower MCS scores than those who were not (Table 6).
Discussion In this study, the finding that participants who had been unemployed for more than 6 months had lower MCS scores than those who had been unemployed for a shorter time suggests that long-term unemployment may
Table 3 Subjective Financial Strain and Physical and Mental Health of Employed Participants PCS
122
MCS
Financial strain
Mean ± SD
p
Mean ± SD
p
Living comfortably Just about getting by Finding it difficult or very difficult
50.7 ± 6.6 53.7 ± 4.4 51.7 ± 6.1
.061
47.1 ± 10.2 46.8 ± 8.2 43.2 ± 7.0
.045
Quality of Life in the Employed and Unemployed in Turkey Table 4 Subjective Financial Strain and Physical and Mental Health in Unemployed Participants PCS
MCS
Financial strain
Mean ± SD
p
Mean ± SD
p
Living comfortably Just about getting by Finding it difficult/very difficult
56.3 ± 3.0 54.3 ± 5.4 55.5 ± 5.6
.193
55.0 ± 4.7 48.1 ± 8.8 43.8 ± 9.1
< .001
Table 5 Unemployment Duration and PCS, MCS Scores Duration of unemployment
PCS Mean ± SD
Less than 6 months More than 6 months
54.4 ± 5.2 55.8 ± 5.8
p .165
MCS Mean ± SD 48.4 ± 8.9 44.0 ± 8.8
p .003
adversely affect mental health. Additionally, in the group of employed people, those who were trying to find a new job had lower MCS scores, indicating that the negative influence on mental health is due to seeking a job even more than to being unemployed. The relationship between employment status and mental health may be complex and subtle: the effects on health of unemployment may not be unambiguously negative while those of employment may not be unambiguously positive. Poverty and unemployment should be considered together. This may be even more important for developing countries in which social security and other official remedial services are limited, if they exist at all. In our study it was found that there were strong associations between financial strain and MCS scores. This finding supports the view that financial strain might influence the relationship between unemployment and health. In this study, PCS scores were found to be lower in the employed group than in the unemployed group, indicating that employed participants had lower levels of physical health. There may be several explanations for this finding. For example, it is possible that the nature of the work and/or working conditions in the factory
Table 6 PCS, MCS Scores Among Employed Participants Who Are Seeking Versus Not Seeking Alternative Employment PCS Mean ± SD Seeking alternative employment Not seeking alternative employment
52.5 ± 5.6 52.9 ± 5.2
p .825
MCS Mean ± SD 46.9 ± 8.7 42.8 ± 8.2
p .01
123
Unemployment and Health ■ Health Effects
may have undermined physical health among working participants; working participants may have spent less time than their unemployed counterparts in exercising and looking after their health; or, alternatively, either the unemployed or the employed participants may have misreported their physical health. In addition, it is possible that job satisfaction, and reason for job loss play an important role in the relationship between work-life and health. This relationship could be better evaluated using a prospective study design that follows up individuals’ work experience and health status. However, this would need a long follow-up period and a substantial number of participants to start with. Alternatively, this relation could be evaluated retrospectively using the individual records of work and unemployment and health status. However, the scarcity of individual records and information in developing countries makes this difficult. Income, poverty, unemployment, exploitation, social class and working conditions together form an intricate structure. This suggests that interventions expected to reduce unemployment rates might affect public health adversely, if administered without accounting for social context. Work and unemployment should be considered together. Regarding the effects of globalisation on working life, flexible working arrangements are enforced to a greater extent by governments and by international organisations in developing countries than by governments and companies in developed countries. Flexible work often means that individuals lose control over their work. Studies investigating different types of employment and working arrangements might be fruitful in future debates on unemployment, employment and health.
References Bartley, M., & Owen, C. (1996). Relation between socioeconomic status, employment, and health during economic change, 1973–1993. British Medical Journal, 313, 445–449. Brazier, J.E., Harper, R., Jones, N.M.B., O’Cathain, A., Thomas, K.J., Usherwood, T., et al. (1992). Validating the SF-36 health survey questionnaire: New outcome measure for primary care. British Medical Journal, 305, 160–164. Brenner, S.O., & Levi, L. (1987). Long- term unemployment among women in Sweden. Social Science and Medicine, 25, 153–161. Bulutay, T. (1995). Employment, unemployment and wages in Turkey. Ankara, Turkey: International Labour Office. Coons, S.J., Rao, S., Keininger, D.L., & Hays, R.D. (2000). A comparative review of generic QOL instruments. Pharmacoeconomics, 17, 13–35. Dijkers, M. (1999). Measuring quality of life: Methodological issues. American Journal of Physical Medicine & Rehabilitation, 78, 286–300. Dooley, D., Prause, J., & Rowbottom, K.A.H. (2000). Underemployment and depression: Longitudinal relationships. Journal of Health and Social Behavior, 41, 421–436. Ezzy, D. (1993). Unemployment and mental health: A critical review. Social Science and Medicine, 37, 41–52. Friis, L., Carter, N., & Edling, C. (1998). Self-reported health problems among Swedish miners one year after unemployment. Occupational Medicine, 48, 297–301. 124
Quality of Life in the Employed and Unemployed in Turkey Garratt, A.M., Ruta, D.A., Abdalla, M.I., Buckingham, J.K, & Russell, I. (1993). The SF 36 health survey questionnaire: An outcome measure suitable for routine use within the NHS? British Medical Journal, 306, 1440–1444. Goldberg, D., & Williams, P. (1998). A user’s guide to the General Health Questionnaire. Windsor, UK: NEFR-Nelson. Graetz, B. (1993). Health consequences of employment and unemployment: Longitudinal evidence for young men and women. Social Science and Medicine, 36, 715–724. Hammarström, A., & Janlert, U. (2000). Do early unemployment and health status among young men and women affect their chances of later employment? Scandinavian Journal of Public Health, 28, 10–15. Hammarström, A., Janlert, U., & Theorell, T. (1988) Youth unemployment and ill health: Results from a 2-year follow up study. Social Science and Medicine, 26, 1025–1033. Jenkinson, C., Layte, R., Coulter, A., & Write, L. (1996). Evidence for the sensitivity of the SF 36 health status measure to inequalities in health: Results from the Oxford healthy lifestyles survey. Journal of Epidemiology and Community Health, 50, 377–380. Kessler, R.C., Turner, J.B., & House, S. (1988). Effects of unemployment on health in a community survey: Main, modifying, and mediating effects. Journal of Social Issues, 44, 69–85. Marmot, M.G. (1998). Contribution of psychosocial factors to socioeconomic differences in health. The Milbank Quarterly, 76, 403–447. McHorney, C., Ware, J.E., & Raczek, A.E. (1993). The MOS 36-item short form survey (SF-36): II. Psychometric and clinical tests of validity in measuring physical and mental health constructs. Medical Care, 31, 247–263. Miilunpalo, S., Vuori, I., Oja, P., Pasanen, M., & Urponen, H. (1997). Self-rated health status as a health measure: The predictive value of self-reported health status on the use of physician services and on mortality in the working-age population. Journal of Clinical Epidemiology, 50, 517–528. Svensson, P.G. (1987). International social and health policies to prevent ill health in the unemployed: The World Health Organization perspective. Social Science and Medicine, 25, 201–204. Tarlov, A.R., Ware, J.E., Greenfield, S., Nelson, E.C., Perrin, E., & Zubkoff, M. (1989). The Medical Outcomes Study: An application of methods for monitoring the results of medical care. Journal of the American Medical Association, 262, 925–930. Testa, M.A., & Simonson, D.C. (1996). Assessment of quality of life outcomes. The New England Journal of Medicine, 334, 835–840. Ware, J.E., Kosinski, M., & Keller, S.D. (1994). SF–36 physical and mental health summary scales: A user’s manual (5th ed.). Boston: New England Medical Centre. Ware, J.E., & Sherbourne, C.D. (1992). The MOS 36-item short form survey (SF–36): I. Conceptual framework and item selection. Medical Care, 30, 473–483.
125
Unemployment and Health ■ Health Effects
126
Chapter
10
Health and Lifestyle of Reemployed and Unemployed People Following the Japanese Corporate Reorganisation Law Tatsuya Ishitake and Tsunetaka Matoba
Employment in Japan has become unstable due to prolonged economic recession and restructuring. The unemployment rate has increased correspondingly, reaching 5% according to a government report in 2001 (Ministry of Public Management, Home Affairs, Post and Telecommunications, Statistics Bureau & Statistics Center, 2001). Given the continuing recession it seems that unemployment will continue to increase. The rapid increase in unemployment is becoming an important political issue. Individual unemployment is correlated with subsequent ill health. The effect of unemployment on physical and psychological health has been extensively studied in western European countries over the past 20 years. This research has found that unemployment leads to an increase in morbidity and an unhealthy lifestyle (Claussen, 1993; Leino-Arjas, Liira, Mutanen, Malmivaara, & Matikaine, 1999; Weber & Lehnert, 1997; Westin, 1990). According to national statistics, suicide in Japan was running at more than 30,000 people per year over this time (Health and Welfare Statistics Association, 2001). It has been suggested that some suicides are a consequence of unemployment. It is therefore important to understand the effects of unemployment on health in Japan. As well as looking at the unemployed, we should consider workers who experience job insecurity associated with downsizing and restructuring. According to the national survey of lifestyle reference, the proportion of the Japanese population feeling secure in their jobs has gradually decreased over the last 22 years. This increase in job insecurity has occurred alongside an increase in the unemployment rate. Job insecurity here is defined as workers who feel under threat of losing their job, and applies particularly to part-time and
127
Unemployment and Health ■ Health Effects
temporary workers, and workers reemployed following the application of the Japanese Corporate Reorganisation Law, who are probably working under inadequate conditions. A few researchers have reported that job insecurity of part-time workers is associated with adverse health effects (Arnetz et al., 1991; Dooley, Prause, & Ham Rowbottom, 2000; Ferrie et al., 2001) but no study has investigated workers reemployed by the same company after application of the Corporate Reorganisation Law. The present chapter examines the effect on health of both reemployed and unemployed people following the sudden bankruptcy of a large shoemaking company to which the Corporate Reorganisation Law was applied.
Subjects and Methods Study Sample A Japanese shoemaking company was declared bankrupt in April 1998. In July 1998, after company resuscitation was considered, application of the Corporate Reorganisation Law reduced the workforce of the company by 760 people; a voluntary redundancy scheme was offered and accepted by these 760 people. Of the remaining workers, 765 were reemployed by the company under a rehabilitation plan. Our study subjects comprise two groups. Group A consists of the 765 workers who remained at the company. Group B consists of 473 of the unemployed people, who joined a support organisation in the form of a self-help group. Method A questionnaire about health status and lifestyle was sent to all participants in November 1998. The questionnaire included items asking about past history of 10 major diseases and acceptance of treatment for these medical conditions. Second, the questionnaire looked at 17 subjective symptoms divided into two classes; one class consisted of somatic-related symptoms, such as thirst, poor appetite, weight loss, stomach-ache, constipation, palpitation, flushing of the skin and stiff neck or shoulder; the second class consisted of psychological symptoms such as sleep disorder, headache (dull or otherwise), difficulty waking in the morning, proneness to fatigue, tendency to anger, inactive behaviour, depressive mood and reduced libido. Lifestyle was the third aspect of the questionnaire looked at: (1) sleep and rest, including getting regular sound sleep, falling asleep easily, sleeping deeply, waking too easily, waking early in the morning, good awakening, taking naps and taking adequate rest; (2) diet, including having regular mealtimes, having a balanced diet, abstaining from fatty food, eating snacks, dining out, eating ready-to-eat or frozen food and eating with family; (3) smoking or not smoking; (4) drinking alcohol or not drinking; (5) having a regular physical exercise habit or not (answer ‘yes’ or ‘no’).
128
Health and Lifestyle of Reemployed and Unemployed People Table 1 Demographic Characteristics of the Two Subject Groups
Age (mean age; year) Sex ratio (men/women) Past history (%) Subjective symptom (%) Smoking (%) Drinking (%) Exercise (%)
Group A (reemployed: n = 701)
Group B (unemployed: n = 413)
Significant difference
48.3 2.1 29.3 51.6 42.7 56.0 19.2
52.7 1.4 37.2 46.3 31.1 49.0 42.3
p < .001 p < .001 p < .01 p = .08 p < .001 p < .05 p < .001
Note: Significant difference was calculated but not adjusted by sex and age.
A t test was used to compare the mean age between the two groups. The chi-square test was used to compare the effect of past history and subjective symptoms and health between the two groups. Multiple logistic regression analysis was used to adjust for age and sex. A difference was taken to be statistically significant when p was less than .05. These statistical analyses were conducted with the JMP computer program (version 5).
Results Response Rate and Basic Characteristics of Subjects The final sample included 701 people in Group A (91.6% response rate) and 413 people in Group B (87.3% response rate). Table 1 shows the demographic characteristics of these subjects. The age and sex distributions were significantly different. Except for the effects of past history and participation in exercise, the values in Group A were statistically higher than those in Group B. There was no significant difference in mean age between sex within either group (48.2 years for men and 48.5 years for women in Group A; 52.5 years for men and 53.0 years for women in Group B). Prevalence of Diseases The prevalence of diseases (actual or past ones) in the two groups is shown in Figure 1. The prevalence of cerebrovascular disease, hypertension, diabetes mellitus and cancers in Group B was significantly higher than in Group A, after age and sex adjustment. There was also a significant difference between the sexes for anaemia and gastroduodenal ulcers in both groups. Prevalence of Subjective Symptoms Figure 2 shows the prevalence of subjective symptoms in the two groups. The prevalence of somatically related symptoms was generally higher in Group A, with statistically significant differences for proneness to fatigue, stomach-ache, weight loss and palpitations. By contrast, the prevalence of
129
Unemployment and Health ■ Health Effects
Figure 1 The prevalence rate of the past and presence of several diseases between the two subject groups. Significant difference was calculated for adjusted age and sex.
psychological symptoms such as tendency to anger, inactive behaviour and depressive moods was significantly higher in Group B. There were significant differences between the sexes in both groups in the prevalence of dull headache, thirst, constipation, neck or shoulder stiffness, flushing of the skin, and reduced libido. 0
10
20
fatiguability
stomach-ache constipation palpitations stiff neck and/or shoulder flush
(%)
40
**
dull headache sleep disorder waking early in the morning thirsty poor appetite loss in weight
30
Group A Group B
** ** *
decreased libido easy to anger inactive behaviour
* ** **
depressed mood others
** *p < .05, **p < .02
Figure 2 The prevalence rate of subjective symptoms between the two subject groups. Significant difference was calculated for adjusted age and sex.
130
50
Health and Lifestyle of Reemployed and Unemployed People
Men
No drinking
50 No smoking
Reduced salt intake (%)
50
50
* *
50
50
Regular exercise
Not eating between meals
* 50 Observing nutritional balance
50 Regular meals *p < .05, **p < .01
Group A Group B
Figure 3 The rate of good health habit between the two subject groups for men. Significant difference was calculated for adjusted age.
Healthy Lifestyle Habits Figures 3 and 4 show the prevalence of healthy lifestyle habits (about 7 items) for men and women. For men, the prevalence of healthy habits was higher in Group B than in Group A. Significant differences were especially found between the two groups for smoking, nutritional balance and salt intake. Although similar tendencies were found for women in the prevalence of healthy habits, there was no significant difference between women in the two groups.
Discussion The impact of company bankruptcy on health and lifestyle was found to be greater in Group A (reemployed workers) than in Group B (unemployed people). With regard to subjective symptoms, workers in Group A complained about somatic symptoms more than those from Group B. The lifestyle habits of unemployed people were also found to be generally healthier in unemployed than in reemployed people. However, the prevalence of psychological symptoms was higher in the unemployed group. The reemployed workers in our study were likely to be under strain in the process of enterprise restructuring. These factors resulted in an increase of somatic-related complaints and poor health habits. The number of bankruptcies in Japan in 2001 was 19,441, the second highest annual total since World War II. The Corporate Reorganisation Law or the Civil Rehabilitation Law was applied to 804 of these and the numbers are 131
Unemployment and Health ■ Health Effects
Women
No drinking
100 No smoking
Reduced salt intake
50
100
50
30
Regular exercise
Not eating between meals
50
50
Observing nutritional balance
Regular meals
*p < .05, **p < .01
Group A Group B
Figure 4 The rate of good health habit between the two subject groups for women. Significant difference was calculated for adjusted age.
increasing rapidly. It is thought that rehabilitation would be appropriate for reemployed workers because of their poor working conditions. Previous research has suggested that unemployment is related to an increase in unhealthy behaviour, such as increased alcohol and tobacco consumption, worse dietary habits and reduced rates of physical exercise (Claussen, 1993; Wilson & Walker, 1993). However, Leino-Arjas et al. (1999) found a significant reduction in alcohol consumption and more exercise in workers in prolonged unemployment. Similarly, our follow-up survey found that a healthy lifestyle and good health were maintained by people in long-term unemployment (Matoba, Ishitake, & Noguchi, 2003). Health promotion directed at the long-term unemployed might be important in helping them to preserve good health habits. It is also possible that demographic variables such as age, sex, education and cultural background are factors in health behaviour and must be taken into consideration. In contrast to these findings about somatic health, the mental health of unemployed people was correlated with their economic condition. The time of the survey would be important, as an unemployment benefit is paid for 12 months in Japan. Our questionnaire was sent 3 months after the company went bankrupt, so all of the unemployed people in our survey were receiving benefits at that time. We have reported that long-term unemployment with cessation of unemployment benefit is associated with an increase in the prevalence of depressive states (Matoba, Ishitake, & 132
Health and Lifestyle of Reemployed and Unemployed People
Noguchi, 2003). Maintaining financial security, providing proactive healthcare and retraining for reemployment can all reduce the impact of unemployment on health (Dooley, Prause, & Ham Rowbottom, 2000). In conclusion, we believe that some supporting and counselling systems for both mental and somatic conditions should be available to reemployed workers as well as to unemployed people for a period after company restructuring or downsizing.
References Arnetz, B.B., Brenner, S.O., Levi, L., Hjelm, R., Petterson, I.L., Wasserman, J., et al. (1991). Neuroendocrine and immunologic effects of unemployment and job insecurity. Psychotherapy and Psychosomatics, 55, 76–80. Claussen, B. (1993). A clinical follow up of unemployed. I: Lifestyle, diagnosis, treatment and re-employment. Scandinavian Journal of Primary Health Care, 11, 211–218. Dooley, D., Prause, J., & Ham Rowbottom, K.A. (2000). Underemployment and depression: longitudinal relationships. Journal of Health and Social Behaviour, 41, 421–436. Ferrie, J.E., Martikainen, P., Shipley, M.J., Marmot, M.G., Stansfeld, S.A., & Smith, G.D. (2001). Employment status and health after privatisation in white collar civil servants: Prospective cohort study. British Medical Journal, 322, 647–651. Health and Welfare Statistics Association. (2001). Suicide statistics [in Japanese]. Journal of Health and Welfare Statistics, 49, 53–55. Leino-Arjas, P., Liira, J., Mutanen, P., Malmivaara, A., & Matikaine, E. (1999). Predictors and consequences of unemployment among construction workers: Prospective cohort study. British Medical Journal, 319, 600–605. Matoba, T., Ishitake, T., & Noguchi, R. (2003). A 2-year follow-up survey of health and life style in Japanese unemployed persons. International Archives of Occupational and Environmental Health, 76, 302–308. Ministry of Public Management, Home Affairs, Post and Telecommunications, Statistics Bureau & Statistics Center. (2001). Labour Force Survey 2001. Tokyo: Author. Weber, A., & Lehnert, G. (1997). Unemployment and cardiovascular diseases: A causal relationship? International Archives of Occupational and Environmental Health, 70, 153–160. Westin, S. (1990). The structure of a factory closure: Individual responses to job-loss and unemployment in a 10-year controlled follow-up study. Social Science & Medicine, 31, 1301–1311. Wilson, S.H., & Walker, G.M. (1993). Unemployment and health: A review. Public Health, 107, 153–162.
133
Unemployment and Health ■ Health Effects
134
Chapter
11
Do the Health Consequences of Unemployment Differ For Young Men and Women? Anne Hammarström and Urban Janlert
Little research has been devoted to the question of whether or not the health consequences of unemployment differ for men and women. Furthermore, most research about the health consequences of unemployment has focused on men, especially on white middle-aged men in bluecollar occupations (Leane & Feldman, 1991). Thus, theories of and research into unemployment have often not taken women’s perspectives into account. The leading unemployment researchers in the 1970s and 1980s, including some of the few female researchers within the field, have claimed that unemployment is not as harmful for women as it is for men (Jahoda, 1982; Ovesen, 1978). The supposed explanation is that women can compensate for the negative effects of unemployment by returning to their positions as housewives. It has thus been the belief that housework can satisfactorily provide compensation for paid work. Similar opinions are heard even in countries with a high female employment rate, with men continuing to be regarded as the main ‘breadwinner’ of the family. For example, in a Swedish qualitative study about the gendered rehabilitation process among young working-class men and women with work-related disorders, it was shown that paid work was assumed to be less important for women than for men (Ahlgren & Hammarström, 2000). However, one of the few studies within the field that focuses on women shows that their health is greatly affected by unemployment (Davies & Esseveld, 1988). Furthermore, these effects are as serious as they are for men (Hammarström, 1994; Hammarström & Janlert, 1997; Iversen, Andersen, Andersen, Christoffersen, & Keiding, 1987), given that the same criteria are used to define unemployment (Leeflang, Klein-Hesselink, & Spruit, 1992). This suggests that unemployment is as serious a problem for women as for men. 135
Unemployment and Health ■ Health Effects
Because of the lack of research within the field, there is an obvious need to raise new research questions and new perspectives based on women’s experiences; however, gender perspective is rare in unemployment public health research. The aim of this chapter is to analyse whether the health consequences of unemployment differ for young men and women.
Method A prospective study was started in 1981, including all 1083 pupils (506 girls and 577 boys) in the last year of compulsory schooling at the age of 16. The study was carried out in the municipality of Luleå, an industrial town situated in the northern part of Sweden. This town has had high unemployment for many years compared to the national average. The youth unemployment rate in Luleå was 8% during the first 3 years of the study and thereafter fell to 5.5%. The figures were somewhat higher for boys but the gender differences were minimal. Extensive work was carried out in order to maximise the response rate. The total nonparticipation rate in the study (including the original nonparticipation in 1981 and those who died during the 5-year period) was 2%. The cohort was followed up 2 years and then 5 years after leaving compulsory schooling. The investigation tools consisted of a comprehensive selfadministered questionnaire with questions on social background factors and psychological and somatic health problems. The majority of the questionnaire was constructed using questions from the Swedish National Survey on living conditions, the Low Income Study, and the Swedish council for information on alcohol and other drugs (Hammarström, 1986). Additional information about diseases that were experienced during time at school was collected from the school health service. The data about the length of unemployment were measured with a specially constructed battery of questions. If necessary, these were supplemented with an interview or with register data from the county labour board. The measure of unemployment was defined differently and will therefore be described in the results section for each analysis. Nervous complaints were measured as an index of seven questions about restlessness, lack of concentration, uneasiness, palpitation, anguish and other nervous complaints. Depressive symptoms were measured as an index, consisting of three questions about down-heartedness, distress and life dissatisfaction. Every question could be answered on a 4-point scale rating the frequency of the phenomenon — from never to constantly. The two scales were then constructed by summing the answers to the separate questions. The same questions were used at 16 and 21 years of age. Smoking was defined as daily smoking at 21 years of age (among those who did not smoke at the age of 16). A reliability test of the tobacco questionnaire was made on the first occasion in 1981. It indicated that the group gave similar answers after a 2-month period (Hammarström, 1986). 136
Do the Health Consequences of Unemployment Differ For Young Men and Women?
Alcohol consumption was estimated as per annum consumption. The volume of different alcoholic beverages was converted into decilitres of pure alcohol. Beer was considered to contain 4.5% alcohol, wine 10% and spirits 40%. The estimated average alcohol volume for each alcohol type was calculated by multiplying the occasions per year on which alcohol was drunk by the volume of alcohol on each occasion. The resulting product was the approximate per annum alcohol consumption in decilitres of pure alcohol. The study was approved by the ethics committees of Uppsala University and Umeå University, which means that the principles of the declaration of Helsinki apply to the study. The study has been described in more detail elsewhere (Hammarström, 1996; Hammarström & Janlert, 2000; Novo, 2000).
Results Psychological Ill Health When comparing the odds ratios for an increase in nervous complaints or depressive symptoms, Figure 1 shows that the longer the total length of unemployment (measured in weeks of unemployment between 16 and 21 years of age), the higher the odds ratio for increases in symptoms. Regarding nervous complaints, there was a fairly uniform increase with increasing unemployment length. For depressive symptoms, higher odds ratios were found after an unemployment period of 1 year or more. There were no more pronounced gender differences, except concerning depressive symptoms in relation to long-term unemployment, where young women showed the highest increase. Alcohol In Figure 2, the increase in alcohol consumption during a 5-year period was analysed in relation to both different length of unemployment and different
10
Odds ratio
Number of unemployed
500
9
10
Odds ratio
Number of unemployed
500
9
8
400
7
8
400
7
6
300
5
6
300
5
4
200
3
4
200
3
2
100
1
2
100
1
0
0 0
10
20
30 40 50 60 Weeks unemployed
70
80
90
0
0 0
10
20
30 40 50 60 Weeks unemployed
70
80
90
Figure 1 Odds ratio for increase in nervous complaints (left) and depressive symptoms (right) compared to the length of unemployment. O = women, ■ = men. The right y axis represents the number of unemployed persons (x).
137
Unemployment and Health ■ Health Effects
1–20 weeks
Figure 2 Daily smokers in the beginning and at the end of the study and length of unemployment during the period.
consumption patterns at the baseline of the study (the last year of compulsory school). Two groups were defined: those who had a high alcohol consumption and those with a low alcohol consumption at the baseline of the study. The young men who were unemployed long-term increased their alcohol consumption significantly, independent of whether they were low or high consumers in the baseline study. Among women, a different pattern was found. Those who were low consumers at school and later became unemployed did not significantly increase their consumption during the 5-year period. However, the high consumers at the baseline of the study who later became unemployed long-term markedly decreased their alcohol consumption during the 5-year period. After controlling for motherhood and alcohol consumption at the baseline of the study, there was a positive correlation between unemployment and alcohol consumption for women. Tobacco Figure 3 shows the smoking habits among those who did not smoke at the start of the study, in relation to different lengths of unemployment (no unemployment, unemployment for between 1 and 20 weeks, and more than 20 weeks’ unemployment during the 5-year period). The figure shows that there is a significant association between smoking initiation during the period and long-term unemployment. Furthermore, we see that the gender differences among the long-term unemployed increase. 138
Do the Health Consequences of Unemployment Differ For Young Men and Women?
1–20 weeks
Figure 3 Daily smokers at the age of 21 among those who did not smoke at the age of 16 years.
Discussion Methodological Considerations The population can be considered to be representative of young people in middle-range industrial towns in Sweden. The high response rate in this study contributes to the high quality of the data. The questionnaire was qualitatively validated in an interview study described elsewhere (Hammarström, 1996). The reliability was tested by comparisons with another study, which had similar questions about alcohol and tobacco consumption and the same population (Hammarström, 1986). The question of selection or exposure has been debated in relation to research on unemployment and health. In our study, we analysed the differences between 16 and 21 years of age, thus controlling for selection bias. Even so, earlier research regarding tobacco consumption shows that while there is an exposure among women, there is a selection among men. Several studies have shown that although the association between unemployment and ill health/health behaviours is a question of both selection and exposure, the exposure effects remain after control for selection (Claussen, 1994; Hammarström & Janlert, 1997; Winefield, 1995). Results Our results indicate that unemployment appears to decrease the gender gap with regard to psychological symptoms. These findings are in accordance with other studies. For example, in a Dutch study, the health effects of unemployment were similar for men and women, given that the same criteria were used to define unemployment (Leeflang et al., 1992). 139
Unemployment and Health ■ Health Effects
Rantakeisu, Starrin and Hagquist (1997) conducted a cross-sectional analysis of psychological ill health and found that the combination of unemployment and feelings of shame was associated with the same degree of ill health for men and women. In earlier research, we have found that lack of control at school has a similar negative effect on alcohol consumption for young men and women (Hammarström, Janlert, & Theorell, 1988). A hypothesis is that unemployment brings similar uncontrolled conditions for men and women, and thus has similar effects on psychological health. In future research we will go deeper into the question of mediating or moderating factors in the closure of the gender gap in relation to unemployment and psychological ill health. However, with regard to health behaviour our findings were different. The gender gap increased in different directions among the unemployed. In relation to alcohol consumption, a significant increase was found in young men who were unemployed long-term, while a large decrease was found among young women who were unemployed long-term. Gender research has demonstrated how the poorer health behaviour of young unemployed men may be explained by the way gender is constructed among long-term unemployed young men (Connell, 1995). Their dangerous behaviour can be interpreted as a way for them to make themselves men, from a life situation as young, marginalised working-class boys. Their masculinity is shaped in relation to the labour market as a whole, where these young men are alternating between low-income jobs, labour market measures and unemployment. An outstanding feature of their life is violence in different forms, such as traffic accidents, fights, bullying and being arrested. The connection between the construction of this kind of masculinity and ill health, as well as alcohol consumption, is obvious. William Courtenay has also analysed the impact of constructions of masculinity on men’s health (2000). He suggests that health-related behaviours, like other social practices that men and women engage in, are a means of demonstrating femininity and masculinity. Health behaviours are used in daily interactions in the social construction of gender and power. Thus, he proposes that the social practices that undermine men’s health are signifiers of masculinity that men use in the negotiation of power and status. Social position determines which unhealthy behaviours men use to demonstrate their masculinity. Alcohol consumption could be regarded as a signifier of masculinity for marginalised men, who may attempt to compensate for their subordinated position by constructing alternative forms of masculinity, which are frequently dangerous and dominated by behaviours such as alcohol consumption, violence and risk-taking. However, no similar gender theoretical analyses have been performed that analyse the impact of constructions of femininity on women’s health behaviour. Our study shows how marginalised young long-term unem140
Do the Health Consequences of Unemployment Differ For Young Men and Women?
ployed women decrease their alcohol consumption but increase their smoking. The decrease in alcohol consumption among women disappears when controlling for children. However, controlling for children does not affect the alcohol consumption among men. An outstanding feature of unemployed young women’s experience is relationships with others, particularly their boyfriends, relatives and girlfriends. As young unemployed women with a working-class background, their femininity is constructed around relational orientation and responsibility-taking within a context of female subordination, where relational orientation is characteristic for human beings in societal subordinate positions (Hammarström, 1996). Compared with young men in similar situations, the young women tend to be more sensitive to the needs of others and thus tend to break vicious cycles more easily. There is a connection between the construction of this kind of femininity and positive health development, such as decreasing one’s alcohol consumption. How then could the increase in smoking among unemployed young women be explained? The ideal of the thin body is an important part of the construction of the dominant forms of femininity in our society. A thin body may be even more important among unemployed young women, who often gain weight during unemployment as a result of poorer eating habits and physical inactivity (Hammarström, 1996). Cigarette smoking may be used to suppress the appetite in order to lose weight. Besides, earlier research has showed that financial problems are of special importance for smoking among unemployed young women (Novo, Hammarström, & Janlert, 2000). Thus, the increase of the gender gap in relation to smoking could be regarded as a result of the gendered distribution of the economy in society, where marginalised young women have a particularly poor position (Hammarström, 1996). Smoking among women is often regarded as a way of coping with stress, which relates to a poor financial situation in combination with responsibility for children (Graham, 1987).
Conclusion Unemployment seems to decrease the gender gap in psychological ill health, while the gender gap in health behaviour increases. There is a need to move away from the simplistic and essentialist research on gender differences in health towards a more critical gender theoretical analysis of the inconsistencies and complexities in the patterns of gender differences in health.
References Ahlgren, C., & Hammarström, A. (2000). Back to work: Gendered experiences of rehabilitation. Scandinavian Journal of Public Health, 28, 88–94. Claussen, B. (1994). Deprived of work and health? A two years follow up of long term unemployed from Greenland, Norway, 1988–90. Unpublished doctoral dissertation, National Institute of Public Health, Unit for Health Services Research and University of Oslo, Unit for Insurance Medicine, Oslo, Norway. 141
Unemployment and Health ■ Health Effects Connell, R.W. (1995). Masculinities. Cambridge, UK: Polity Press. Courtenay, W.H. (2000). Constructions of masculinity and their influence on men’s wellbeing: A theory of gender and health. Social Science & Medicine, 50, 1385–1401. Davies, K., & Esseveld, J. (1988). To play hop-scotch in the Swedish labour market [in Swedish]. Stockholm, Sweden: Raben och Sjögren. Graham, H. (1987). Women’s smoking and family health. Social Science & Medicine, 25, 47–56. Hammarström, A. (1986). Youth unemployment and ill-health: Results from a two year follow-up study [in Swedish, summary in English; monograph]. Solna–Sundbyberg, Sweden: Karolinska Institute, The National Institute of Psychosocial Factors and Health and the Department of Social Medicine ‘Kronan’. Hammarström, A. (1994). Health consequences of youth unemployment: Review from a gender perspective. Social Science & Medicine, 38, 699–709. Hammarström, A. (1996). Unemployment and bad health: About youth condition [in Swedish]. Lund, Sweden: Studentlitteratur. Hammarström, A., & Janlert, U. (1997). Nervous and depressive symptoms in a longitudinal study of youth unemployment: Selection or exposure? Journal of Adolescent Health, 20, 293–305. Hammarström, A., & Janlert, U. (2000). Does early unemployment and health status among young men and women affect their possibility of later unemployment? Scandinavian Journal of Public Health, 28, 10–15. Hammarström, A., Janlert, U., & Theorell, T. (1988). Youth unemployment and illhealth: Results from a 2-year follow-up study. Social Science & Medicine, 26, 1025–1033. Iversen, L., Andersen, O., Andersen, P.K., Christoffersen, K., & Keiding, N. (1987). Unemployment and mortality in Denmark, 1970–80. British Medical Journal, 295, 879–884. Jahoda, M. (1982). Employment and unemployment. A socialpsychological analysis. Cambridge, UK: Cambridge University Press. Leane, C.R., & Feldman, D.C. (1991). Gender differences in responses to unemployment. Journal of Vocational Behaviour, 38, 65–77. Leeflang, R.L.I., Klein-Hesselink, D.J., & Spruit, I.P. (1992). Health effects of unemployment: II. Men and women. Social Science & Medicine, 34, 351–363. Novo, M. (2000). Young and unemployed: Does trade cycle matter for health? A study of young men and women during times of prosperity and times of recession. Unpublished doctoral dissertation, Umeå University, Umeå, Sweden. Novo, M., Hammarström, A., & Janlert, U. (2000). Smoking habits: A question of trend or unemployment? A comparison of young men and women between boom and recession. Public Health, 114, 460–463. Ovesen, E. (1978). The psychological consequences of unemployment [in Swedish]. Stockholm, Sweden: Raben & Sjögren. Rantakeisu, U., Starrin, B., & Hagquist, C. (1997). Unemployment, shame and ill health: An explorative study. Scandinavian Journal of Social Welfare, 6, 13–23. Winefield, A.H. (1995). Unemployment: Its psychological costs. In C.L. Cooper & I.T. Robertson (Eds.), International review of industrial and organizational psychology (pp. 169–212). Chichester, UK: Wiley.
142
Chapter
12
Mass Lay-Offs and Tolerance for Mental Illness: Racial Differences in the Economy’s Effect on Coerced Treatment Eric Kessell and Ralph Catalano
The fertile intersection of sociology and psychology has spawned many hypotheses, including that tolerance for deviance among minorities moves over time with the status of the local economy. The theoretical underpinnings of the hypothesis have been described at length elsewhere (Blalock, 1967; Kinloch, 1974). In essence, the hypothesis arises from combining frustration–aggression theory (Berkowitz, 1989), and is attributed to Dollard, Doob, Miller, Mowrer and Sears (1939), with the notions of social position often traced to Blumer (1958). The former argues that people feel frustrated when they believe that their realisation of an earned or entitled reward has been thwarted for reasons unrelated to their behaviour. Some of these people supposedly manifest their frustration with aggression aimed at those they view as deviant, socially inferior and unlikely to retaliate. Among the earliest, and probably best known, attempts to test frustration aggression theory was Hovland and Sears’ (1940) study of lynching of African Americans in the United States from 1882 to 1930. The authors assumed that economic contraction would thwart many people by denying them access to income for reasons unrelated to work performance. Some Whites in the South, they theorised, would manifest their frustration with aggression towards African Americans believed to have acted in ways threatening to the White community. Hovland and Sears reported an inverse relationship between the price of cotton and the incidence of lynching. Several researchers have reanalysed the lynching data with increasingly sophisticated methods (e.g., Green, Glaser, & Rich, 1998;
143
Unemployment and Health ■ Health Effects
Hepworth & West, 1988; Mintz, 1946). All report replicating the original findings, although with varying estimates of the strength of the association. Although Hovland and Sears did not explore the implications of their work for understanding racial bigotry, sociologists and other psychologists have done so (e.g., Blalock 1967; Vanneman & Pettigrew, 1972). This work shares the argument that dominant groups acquire economic and other privileges through their relative power. They defend against any feelings of guilt induced by this exaction by believing that contributions that they, or their ancestors, made to society through skills, hard work and virtue entitle them to privilege. This defence logically implies that those with fewer or no privileges have inferior skills and character. Blalock (1967) extended the above argument to suggest that dominant groups will be most dismissive of the abilities and character of others when contracting economies induce frustration in the population. He argued that dominant groups fear, correctly or otherwise, that others want to redistribute privilege to relieve frustration induced by economic contraction. The dominant group’s response to this perceived threat, itself intensified by frustration, includes increasingly aggressive assertion of the notion that differences in ability and character make the status quo legitimate. Because these differences typically correlate highly with race, a contracting economy induces increased expression of racial bigotry. Blalock’s argument has been extended to explain criminal aggression against minorities defined by other than race. Green et al. (1998), for example, hypothesise that hate crimes against homosexuals increase during periods of economic contraction. We propose to apply Blalock’s theory to a less conspicuous but more common phenomenon than overt acts of criminal violence. Specifically, we argue that Blalock’s theory predicts that tolerance for mentally ill AfricanAmerican males declines when mass lay-offs shock the labour market. Local health authorities in the United States and elsewhere routinely coerce into treatment people whose mental illness renders them gravely disabled or a threat to themselves or others. Formal rules guide clinicians in determining whether individuals meet the criteria for coerced treatment, but such rules govern neither the discovery of candidates for coerced treatment nor the decision to present such candidates to health authorities. Family, neighbours, friends, or even strangers, whose lives are inconvenienced, disrupted or threatened by a mentally ill person decide who among the population of ill people comes to the attention of police or emergency personnel. Blalock’s theory implies that intolerance heightened by economic insecurity affects the layman’s perception of what threatens a business, neighbourhood, property or person. Opportunistic political leaders, moreover, may sense or even cultivate such intolerance and influence the policies of law enforcement and clinical institutions in ways that reflect popular 144
Racial Differences in the Economy’s Effect on Coerced Treatment
biases. Blalock’s mechanisms, therefore, could be at work in the processes that lead to coerced treatment of minorities with mental illness. We contribute to the literature by offering a novel approach to detecting the effect, if any, of Blalock’s mechanisms on the coerced treatment of African-American males. We demonstrate the method using data from the state of California, where roughly 1.26 million of the 34 million residents are African-American males. We test the hypothesis that the median level of functioning among African-American males coerced into treatment varies positively with the number of people who lose jobs in mass lay-offs. As described below, clinicians in California score all people coerced into treatment on a scale that measures how well the patient functions regardless of diagnostic category or other circumstances that lead to treatment. We assume an essentially normal distribution of functioning in the population at large. People at the low end of the distribution will be the most likely to experience coerced treatment because they appear most disabled and threatening. If the level of functioning of the candidates presented by the community to clinicians were constant, the median functioning score among those coerced would remain constant even if a contracting economy pushed an increasing number of people beneath that score. If the intolerance mechanism were at work, however, a contracting economy would move the criterion for coercion upward (i.e., better functioning people would be at elevated risk of coerced treatment). The median level of functioning of those coerced, regardless of their number, would rise. We use the number of people who lose jobs in mass lay-offs as our independent variable because this type of job loss best fits the classic frustration–aggression theory. As described below, ‘mass lay-off’ refers to an event in which at least 50 workers lose jobs because an enterprise does not need their labour. Workers displaced by such events can reasonably infer that they have been denied employment and its benefits for reasons unrelated to their behaviour.
Methods Data We used the monthly median of the Global Assessment Scale (GAS) and Global Assessment of Functioning (GAF) scores as our measure of functioning in each of the monthly cohorts of people coerced into treatment. The California State Department of Mental Health, which provided us with the functioning scale data, requires clinicians to assess functioning after admitting a patient to treatment. The GAS measures overall psychosocial functioning, including psychological symptoms, as well as social and occupational functioning (Endicott, Spitzer, Fleiss, & Cohen, 1976). The American Psychiatric Association incorporated the GAF, a slightly modified version of the GAS, into the Diagnostic and Statistical Manual of 145
Unemployment and Health ■ Health Effects
Mental Disorders (DSM-III-R) in 1987 as Axis V, a formal component of diagnosis (American Psychiatric Association, 1987, 1994). GAS and GAF scores range from 1 to 100. The lowest scores (i.e., 1–10) imply ‘persistent danger of severely hurting self or others or persistent inability to maintain minimal personal hygiene’, while the highest (i.e., 90–100) imply ‘superior functioning in a wide range of activities’. Psychiatric inpatients typically have scores between 1 and 40 (Spitzer, Gibbon, Williams, & Endicott, 1994). The mean of monthly median GAF scores for African-American males over our 39-month test period was 30.615 (SD = 1.549), with a range from 30 to 35. The GAF is widely used in both research and clinical practice as a global measure of functioning. Both the GAS and the GAF have demonstrated adequate reliability and validity (Endicott et al., 1976; Goldman, Skodol, & Lave, 1992; Hilsenroth et al., 2000; Patterson & Lee, 1995). Scores typically correlate well with other measures of clinical status (Hilsenroth et al., 2000; Roy-Byrne, Dagadakis, Unutzer, & Ries, 1996). The US Bureau of Labor Statistics, in cooperation with state governments, collects mass lay-off and total lay-off data (United States Bureau of Labor Statistics, 2001). Data describing mass lay-offs and the functioning level of people coerced into treatment were available for the 39 months beginning April, 1995 (first month of mass lay-off data) and ending July, 1998 (last month of coerced treatment data). The mean number of mass lay-offs was 30,937 (SD = 17,019), and monthly values ranged from 11,248 to 83,203. We test the effect of mass lay-offs controlling for the effects of other job losses. We computed other job losses by subtracting mass lay-offs from total job losses. Design Data are consistent with causation when the dependent variable differs, as predicted by theory, from the values expected under the null hypothesis (i.e., under the assumption that the independent variable has no effect). The true experiment derives expected values under the null hypothesis from a randomly assigned control group. We cannot randomly assign individuals to control and test communities in which we manipulate the macro economy. Tests of the effect of mass lay-offs on tolerance for deviance among minorities therefore must use statistical control. Such tests typically assume that the expected value of the dependent variable under the null hypothesis is the mean of its observations. Time series, however, often exhibit trends, cycles and the tendency to remain elevated or depressed after high or low values. These patterns, referred to collectively as autocorrelation, complicate tests because the expected value of autocorrelated series is not their mean (Catalano, Dooley, & Jackson, 1983). Researchers typically respond to the problem of autocorrelation by identifying patterns and expressing them as an effect of earlier values, or 146
Racial Differences in the Economy’s Effect on Coerced Treatment
‘lags’, of the dependent variable itself (Catalano, Hansen, & Hartig, 1999; Catalano & Hartig, 2001; Catalano & McConnell, 1999; McCleary & Hay, 1980). The residuals from a time-series equation with the correct lags exhibit no autocorrelation. The analyst can therefore add the independent variables to the equation to determine if their coefficients differ from zero in the hypothesised direction. Any approach that removes autocorrelation from the dependent variable before testing the effect of the independent variable has the added benefit of avoiding Type I errors due to shared trends and cycles (Catalano & Serxner, 1987). The estimated coefficient for the independent variable is net of autocorrelation shared by the dependent and independent variables. The approach described above leaves open the possibility that an uncontrolled ‘third’ variable without trends and cycles could induce a spurious association between the variables of principal interest. Epidemiologists (Catalano & Serxner, 1987) have suggested controlling this threat by measuring the dependent variable in a comparison population and including the variable among the predictors. This strategy controls for any third variables, whether autocorrelated or not, that affect both the test and comparison population. We included the monthly median GAF scores for non-Hispanic white males as a control in our tests. The strategy described above requires estimating the following equation:
⎛⎜1 − θB q ⎞⎟ ⎠a n n ∇ dYt = ω 0∇ d X1t + ω 1B ∇ d X2t + ω 2B ∇ d X3t + ⎝ t ⎛⎜1 − φB p ⎞⎟ ⎝ ⎠ ∇d is the difference operator that indicates a series was differenced at lag d (i.e., values at month t are subtracted from those at month t–d) to remove secular trends or seasonal cycles (i.e., render the series stationary in its mean). Yt is the median GAF score in California during month t for AfricanAmerican Males. ω0 to ω2 are the effect coefficients for X1t, X2t and X3t. Bn is the ‘backshift operator’ that indicates a variable has been lagged n months. For the w coefficients, Bn implies the association between y at month t and X2 or X3 at month t–n. X1t is the median GAF score in California during month t for nonHispanic white Males. X2t is the number of mass lay-offs (in thousands) in month t. X3t is the number of other job losses (in thousands) in month t. θ is the ‘moving average’ parameter. Moving average parameters express the number of months that perturbations typically last.
147
Unemployment and Health ■ Health Effects
φ is the autoregressive parameter. Autoregressive parameters gauge the number of months required for a series to revert to preperturbation levels after a perturbation ends. Bp and Bq are the backshift operators for the error term and indicate the lags at which the autoregressive and moving average effects begin. at is the error term at month t. Analyses We first regressed the monthly median GAF score for African-American males on the same scores for non-Hispanic White males, and job losses other than mass lay-offs. We then inspected the residuals of the regression for autocorrelation. We used the augmented Dickey–Fuller test (Dickey & Fuller, 1979) to detect secular trends and cycles expressed by the ∇d terms in equation 1. We used the strategy attributed to Box and Jenkins (Box, Jenkins, & Reinsel, 1994) to identify and model autocorrelation (as expressed by the p, q, θ and φ parameters in Equation 1) in the dependent variable. The strategy, Autoregressive, Integrated, Moving Average (i.e., ARIMA) modelling, allows the researcher to fit empirically any of a large family of possible models to a time series. ARIMA models can be thought of as mathematical expressions of various filters through which temporally unpatterned series pass (Catalano et al., 1983). Each filter imposes a unique pattern of autocorrelation. In the Box–Jenkins approach, the researcher uses an iterative model-building process to infer the filter that imposed the autocorrelation observed in a time series. We next added the three X terms to the equation resulting from the above steps. The test specified the n in Bn as zero (i.e., no lag).
Results Table 1 shows the estimated coefficients for the final test equation. The coefficient for non-Hispanic White males indicates, as might be expected, that the median GAF scores for both groups of males tends to move together over time. As indicated by the moving average parameter, median GAF scores for African-American Males coerced into treatment exhibit ‘memory’ in that large and small values tend to persist into the next month. As hypothesised, the median level of functioning among AfricanAmerican males coerced into treatment increased in months when mass lay-offs increased. The strength of the association discovered in this test can be characterised in at least two ways. First, we derived the familiar ‘change in R2’ metric by comparing variance ‘explained’ by equations with and without the mass lay-offs term. The equation without the term accounted for 47% of the variance in median monthly GAF scores of AfricanAmericans coerced into treatment. The equation with the term accounted for 52%.
148
Racial Differences in the Economy’s Effect on Coerced Treatment Table 1 Final Model of Monthly Median Functioning Score for African-American Males Coerced into Treatment in California for 39 Months Beginning April 1995 (Standard Errors in Parentheses) Variable Non-Hispanic white median monthly GAF score Job losses other than those from mass lay-offs, in thousands. Moving average at Lag 1 Mass lay-offs, in thousands
Estimate coefficient .8292**(.0675) .0039 (.0086) –.9656** (.0702) .0288* (.0113)
Note: *p < .05; two-tailed test; **p < .01; two-tailed test.
Second, the coefficients imply that about 34,700 lay-offs from mass events induce a 1-point increase in GAF scores of African-American males in the month of the lay-offs. Mass lay-offs exceeded that number in 9 of the 39 months observed. We also used the methods attributed to Alwan and Roberts (i.e., Alwan & Roberts, 1985; Hillmer 1984) to determine if outliers had distorted the estimation of parameters for GAF scores. The replacement of the GAS with the GAF, for example, could have induced slight level shifts in the time series. The method detected and controlled several outliers but the results of the test did not change. We tested the possibility that the discovered association extended into following months by adding two lags of the mass lay-offs variable to the test equation (i.e., n in Bn set to 1 and 2 in addition to zero). No extension was found. This finding means that any persistence in the effect of mass lay-offs did not exceed that following all perturbations in median GAF scores. The moving average parameter noted above therefore would control such persistence.
Discussion Our test should be considered a demonstration of one approach to testing the intolerance hypothesis. The fact that we had only 39 months of data limits the external validity of the test. We do not know if the coefficients we estimated would be found with longer series that would allow more precise specification of autocorrelation. We believe, however, that the approach is sound and worthy of replication where and when more data describing the functioning of people coerced into mental health treatment become available. Using the median functioning scores among those coerced into treatment, rather than the mean score, allowed us to discriminate between a circumstance in which more people are forced below constant social criteria for coercion and one in which the criteria moved upward (intolerance). Evidence supporting the intolerance mechanism does not exclude the possibility that the stress from a contracting economy also forces more people below some objective criteria. 149
Unemployment and Health ■ Health Effects
Despite their demonstrated clinical reliability and validity (Endicott et al., 1976), GAF/GAS scores could be subject to bias by changes in the tolerance of the clinicians who assign them. For this explanation to be consistent with the data, clinicians would have to have given higher median scores at times of mass lay-offs. We know of no evidence of such economyinduced change in physician behaviour. We think it more likely that physicians would intentionally lower scores to justify commitment if they thought an acutely ill person would be better off in hospital than in a household or community made intolerant by economic contraction. Although frustration–aggression theory can reasonably explain why coerced treatment increases during bad economic times, not all coerced treatment necessarily stems from aggression. Friends, family, caretakers and others close to mentally ill people doubtlessly bring their loved ones in for treatment (Hoge et al., 1998). Although they might not wish to have the person treated against his or her will, friends and family may be unable to provide the support necessary following mass lay-offs or other adverse economic events. We would expect the level of functioning to be lower in those coercions initiated by friends and family (somebody within a support network) than by a stranger or other person from outside the support network. This explanation does not predict the difference in response to mass lay-offs between African-American males and non-Hispanic white males. Mass lay-offs would have to affect the African-American males’ support networks more than those of non-Hispanic white males. The dataset we used did not contain details of the relationship of the person coerced into treatment to those who were associated with the call. Such data often appear in medical records, and extracting them may be feasible for a smaller-scale study, particularly one that entails fewer observations but examines effects over a smaller area. Finding a relationship between stranger- or police-initiated coerced treatment and the economy would support the frustration–aggression mechanism. Friend- or family-initiated treatment could be reasonably attributed to other support-network mechanisms, although frustration aggression could not be ruled out. Describing the context in which someone is coerced into treatment and capturing that context in our models is an important challenge in this line of research and probably extends beyond the sole use of medical records (Morgan, Mallett, Hutchinson, & Leff, 2004). Clearly we could learn much from more detailed data on those who initiated the treatment. Investigators may be able to gather more information when the patient’s friends or family initiated treatment, but data collection challenges increase when only strangers are involved. For stranger-initiated coercions, the most detail we may be able to get is the location from which the patient was brought to the attention of the public health/safety system. Location data potentially inform us about a number of factors that may be related to tolerance towards the patient. Demographic data routinely 150
Racial Differences in the Economy’s Effect on Coerced Treatment
collected by the census can provide a measure of how similar the patient is to those who live in the place where was brought to the attention of the public health/safety system. The census also compiles median income and other economic data at the neighborhood level, but does not collect lay-off, as opposed to unemployment, data. Mass lay-offs may offer a solution to these missing data in that firms may be able to supply the home addresses of affected workers. From these data we could construct a model where the functioning of those coerced into treatment is related to the density of laidoff workers. Not all forms of unemployment necessarily have the same effect on community tolerance. Mass lay-offs (those resulting in a loss of more than 50 jobs) may differ from other lay-offs in a number of ways. First, jobs lost due to mass lay-offs may be perceived differently to the same number of jobs lost to individual or smaller-scale lay-offs. Smaller-scale lay-offs across a number of companies or across a long stretch of time may not reach the same level of collective recognition. Mass lay-offs may be reported by local media and thus be brought to the attention of those who otherwise may not have been aware of them, and thus have a wider psychological impact. Such lay-offs may also signal fundamental weakness in an area’s economic base and occupy a large part of public discourse. We have previously found a link between acute community stress (caused by the events of September 11, 2001) and the number of people coerced into treatment (Catalano, Kessell, McConnell, & Pirkle, 2004). Mass lay-offs may not have the same acute effect as a community disaster, but both may induce greater stress on social support networks than do regular lay-offs. Social support networks can be thought of as informal insurance pools to which individuals contribute when friends and family members encounter difficulties, and from which they expect similar support when they themselves require it (see, e.g., Mitchell & Trickett, 1980). Such systems assume regular, endemic demands distributed relatively evenly and unpredictably among members. Although normal lay-offs are not randomly distributed in a population, they occur at a regular frequency or in cycles allowing social support networks to anticipate and absorb them. Mass lay-offs and disasters, ‘epidemics’ of stressors, overburden the network by unexpectedly increasing the number of demands placed upon it while reducing the capacity of its members to respond. Following a mass lay-off, those who try to draw on support in the community may find that the help they had expected is not available, be these formal programs, such as job training/placement and employment counselling, or informal social support such as financial, housing or food assistance, or emotional support from friends and family. Community resources normally available to those seeking assistance are unexpectedly scarce. 151
Unemployment and Health ■ Health Effects
The psychological impact of mass lay-offs may also affect those who still hold jobs. Job holders may not only find that the needs of their friends, family and neighbours have increased, but also that their own employment may feel less secure. They may fear that if they do become unemployed, the resources that might normally be available to them will have been used up. A mass lay-off could increase the perception that a person’s ‘fair share’ of potential coping resources has been taken away. According to frustration–aggression theory, this loss (or potential loss) of expected coping resources from mass lay-offs will cause some to be aggressive towards people whom they had tolerated earlier. This effect would be observed independent of the number of nonmass lay-offs, the expectation of which are already built into the informal social support network. Furthermore, if there were a community perception that marginalised groups draw more than their ‘fair share’ of coping resources, then the aggression expressed against the marginalised would be greater. Time-series analysis of statewide data may not always be the most efficient means of evaluating the economy’s effect on treated mental illness. Enterprise closings may have stronger effects on a local economy, and we may not expect any but the very largest mass lay-offs (or coincident combination of smaller mass lay-offs) to have a measurable effect on tolerance for mental illness. Since local health systems are responsible for providing psychiatric care, it would be advantageous to be able to anticipate the effects in the health service areas in which mass lay-offs occur. In such cases statewide data would still be useful as a means of controlling for shared trends. Some economic effects undoubtedly have statewide impact. We could evaluate a number of recent macroeconomic policies’ effect on tolerance for mental illness. For example, tariffs imposed in 2001 on steel imported into the United States were widely perceived as having protected the domestic industry and a substantial number of jobs in several states (Becker & Sanger, 2003). These states (and specific regions within them) may have experienced higher stress following reports that the tariffs were to be lifted. We might also expect recent changes in unemployment insurance benefits or tax rates to affect individual assessments of their social safety net or the degree to which they will be able to meet their financial obligations. Other, more regular, economic events may affect tolerance for persons with mental illness. Federal and state tax returns must be filed in mid-April, and many municipalities collect property taxes at the same time each year. Researchers may wish to consider these influences, but designs that control for autocorrelation would seem to frustrate such work given that the effects would be both cyclic and generally occurring. The only way we know of testing such effects would be with cross-national data from regions close in space and social conditions but different in the timing of tax and other
152
Racial Differences in the Economy’s Effect on Coerced Treatment
fiscal events. Such studies might be possible, for example, in Western Europe and at the US–Canadian border. Smaller units of analysis, such as the region or neighborhood, lend themselves to exploring more detailed contextual questions around diagnoses of mental illness. As discussed above, the characteristics of the neighbourhood in which disordered behaviour is observed may affect the likelihood of treatment. For example, a marginalised person living among other marginalised groups may have a different probability of coerced treatment than one living among dominant groups. The challenges to multilevel analysis include the problem of separating the effect of individuals’ selection into their environments from the effect of environment on individuals (Browning, Cagney, & Wen, 2003). Neighbourhood analyses can include more than characteristics of the residential population. Many other neighbourhood attributes are readily available or could be constructed. Gentrification, commercialisation, proximity to mental health services and crime rates could all influence a community’s tolerance. It is important for epidemiologists to look to other fields and to develop their own theoretical constructs before testing for relationships between these factors. The relationship between the economy and the functioning of people coerced into psychiatric treatment reflects the often-unacknowledged observation that illness, both mental and physical, is not necessarily an objective or constant phenomenon. The literature of a number of disciplines discusses the consequences of being labelled ‘ill’ on the individual and society (Estroff, 1981; Goffman, 1963; Gove, 1975; Langer & Abelson, 1974; Mechanic, 1969; Rosenhan, 1973), but the societal implications are not just academic. Knowing the relationship between unemployment and coerced psychiatric treatment can help the public health system anticipate the demand, if not need, for treatment and allocate resources appropriately following mass lay-off. Those who are jobless because of disability are often eligible for forms of public assistance that are funded differently, and often more generously, than the programs available to those unemployed for other reasons. Reducing the level of treated illness may or may not be desirable, depending on the cost of doing so and how society chooses to distribute those costs (Rodriguez, 2001). This discussion does imply that mental health services are not evenly provided in California; African Americans in particular use lower levels of outpatient psychiatric care but higher levels of emergency care than non-Hispanic whites (Hu, Snowden, Jerrell, & Nguyen, 1991). We hope it will contribute to a larger literature on the social construction of illness and permit others to make informed decisions about addressing disparities in mental health.
153
Unemployment and Health ■ Health Effects
References Alwan, L., & Roberts, H. (1985). Time series modeling for statistical process control. Journal of Business and Economic Statistics, 6, 87–95. American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders (3rd ed.). Washington, DC: Author. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Becker, E., & Sanger, D.E. (2003, December 2). President in a political vise over steel tariff decision. The New York Times, A: 20. Berkowitz, L. (1989). Frustration aggression hypothesis: Examination and reformulation. Psychological Bulletin, 106, 59–73. Blalock, H. (1967). Toward a theory of minority group relations. New York: John Wiley and Sons. Blumer, H. (1958). Race prejudice as a sense of group position. Pacific Sociological Review, 1, 3–7. Box, G., Jenkins, G., & Reinsel G. (1994). Time series analysis: Forecasting and control. London: Prentice-Hall. Browning, C.R., Cagney, K.A., & Wen, M. (2003). Explaining variation in health status across space and time: Implications for racial and ethnic disparities in self-rated health. Social Science & Medicine, 57, 1221–1235. Catalano, R., Dooley, D., & Jackson, R. (1983). Selecting a time-series strategy. Psychological Bulletin, 94, 506–523. Catalano, R, Hansen H., & Hartig, T. (1999). The ecological effect of unemployment on the incidence of very low birthweight in Norway and Sweden. Journal of Health and Social Behavior, 40, 422–428. Catalano, R., & Hartig, T. (2001). Communal bereavement and the incidence of very low birthweight in Sweden. Journal of Health and Social Behavior, 42, 333–341. Catalano, R., Kessell, E., McConnell, W., & Pirkle, E. (2004). Psychiatric emergencies following the terrorist attacks of September 11, 2001. Psychiatric Services, 56, 858–862. Catalano, R., & McConnell, W. (1999). Psychiatric emergencies: The check effect revisited. Journal of Health and Social Behavior, 40, 79–86. Catalano, R., & Serxner, S. (1987). Time-series designs of potential interest to epidemiologists. American Journal of Epidemiology, 126, 724–731. Dickey, D., & Fuller, W. (1979). Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Society, 74, 427–31. Dollard, J., Doob, L., Miller, N., Mowrer, O., & Sears, R. (1939). Frustration and aggression. New Haven, CT: Yale University Press. Endicott, J., Spitzer, R.L., Fleiss, J.L., & Cohen, J. (1976). The global assessment scale: A procedure for measuring overall severity of psychiatric disturbance. Archives of General Psychiatry, 33, 766–771. Estroff, S.E. (1981). Making it crazy. Berkeley, CA: University of California Press. Goffmann. E. (1963). Stigma: Notes on the management of spoiled identity. New York: Simon & Schuster. Goldman, H.H., Skodol, A.E., & Lave, T.R. (1992). Revising Axis V for DSM-IV: A review of measures of social functioning. American Journal of Psychiatry, 149, 1148–1156. Gove, W. (Ed). (1975). The labelling of deviance: Evaluating a perspective. Oxford, UK: John Wiley & Sons. Green, D., Glaser, J., & Rich, A. (1998). From lunching to gay bashing: The elusive connection between economic conditions and hate crime. Journal of Personality and Social Psychology, 75, 82–92. 154
Racial Differences in the Economy’s Effect on Coerced Treatment Hepworth, J., & West, S. (1988). Lynchings and the economy: A time-series analysis of Hovland and Sears (1940). Journal of Personality and Social Psychology, 55, 239–247. Hillmer, S. (1984). Monitoring and adjusting forecasts in the presence of additive outliers. Journal of Forecasting, 3, 205–215. Hilsenroth, M.J., Ackerman, S.J., Blagys, M.D., Baumann, B.D., Baity, M.R., Smith, S.R., et al. (2000). Reliability and validity of DSM-IV Axis V. American Journal of Psychiatry, 157, 1858–1863. Hoge, S., Lidz, C., Eisenberg, M., Monahan, J., Bennett, N., Gardner, W., et al. (1998). Family, clinician, and patient perceptions of coercion in mental hospital admission: A comparative study. International Journal of Law and Psychiatry, 21, 131–146. Hovland, C., & Sears, R. (1940). Minor studies of aggression: VI. Correlation of lynchings with economic indices. The Journal of Psychology, 9, 301–310. Hu, T.W., Snowden L.R., Jerrell J.M., & Nguyen, T.D. (1991). Ethnic populations in public mental health: Services choice and level of use. American Journal of Public Health, 81, 1429–1434. Kinloch, G. (1974). The dynamics of race relations: A sociological analysis. New York: McGraw-Hill. Langer, E.J., & Abelson, R.P. (1974). A patient by any other name …: Clinician group differences in labeling bias. Journal of Consulting and Clinical Psychology, 42, 4–9. McCleary, R., & Hay, R. (1980). Applied time-series analysis for the social sciences. Beverly Hills, CA: Sage. Mechanic, D. (1969). Illness and cure. In J. Kosa, A. Antonovsky, & I.K. Zola (Eds.), Poverty and health: A sociological analysis (pp. 191–214). Cambridge, MA: Harvard University Press. Mintz, A. (1946). A re-examination of correlations between lynchings and economic indicators. Journal of Abnormal and Social Psychology, 41, 154–160. Mitchell, R.E., & Trickett, E.J. (1980). Social networks as mediators of social support: An analysis of the effects and determinants of social networks. Community Mental Health Journal, 16, 27–44. Morgan, C., Mallett, R., Hutchinson, G., & Leff, G. (2004) Negative pathways to psychiatric care and ethnicity: The bridge between social science and psychiatry. Social Science & Medicine, 58, 739–752. Patterson, D.A., & Lee, M. (1995). Field trial of the Global Assessment of Functioning Scale — Modified. American Journal of Psychiatry, 152, 1386–1388. Rodriguez, E. (2001). Keeping the unemployed healthy: The effect of means-tested and entitlement benefits in Britain, Germany, and the United States. American Journal of Public Health, 91, 1403–1411. Rosenhan, D.L. (1973). On being sane in insane places. Science, 179, 250–258. Roy-Byrne, P., Dagadakis, C., Unutzer, J., & Ries, R. (1996). Evidence for limited validity of the Revised Global Assessment of Functioning Scale. Psychiatric Services, 47, 864–866. Spitzer, R., Gibbon, M., Williams, J., & Endicott, J. (1994). Global Assessment of Functioning (GAF) Scale. In L. Sederer & B. Dickey (Eds.), Outcomes assessment in clinical practice (pp. 76–78). Baltimore, MD: Williams and Wilkins. United States Bureau of Labor Statistics. (2001). How the government measures unemployment. Washington, DC: Author. Retrieved December 10, 2003, from http://www.bls.gov/cps/cps_htgm.htm Vanneman, R., & Pettigrew, T. (1972). Race and relative deprivation in the urban United States. Race, 13, 461–486.
155
Unemployment and Health ■ Health Effects
156
Chapter
13
Demographic, Occupational and Employer-Related Determinants of Long-Term Unemployment Among Danish Employees Thomas Lund and Merete Labriola
Although Scandinavia experienced a severe increase in unemployment rates during the late 1990s, this trend had a marginal effect on the Danish labour market. Instead, rates declined in Denmark during the 1990s and were relatively low compared both to the situation in the 1980s as well as to the prevailing situation in the rest of Scandinavia. However, between 2002 and 2003 the Danish unemployment rate increased, especially with regards to long-term unemployment (LTU): in March 2002, unemployment was about 140,000 people (corresponding to 4.9% of the labour force), which was the lowest level for more than 25 years. By June 2003, however, unemployment in Denmark had increased by just over 30,000 people, corresponding to an additional 1.2% of the labour force. This increase once again made unemployment prevention and the reemployment of unemployed people the focus of political attention. The Danish Government recently launched a national plan of action that focuses on reintegrating the ‘weaker’ unemployed into working life as well as restructuring labour market policy. This is to be done by ensuring that there are broad competencies among individuals in the workforce, thereby ensuring flexibility and easier reemployment (Danish Government, 2003). At the same time the demographic trend towards an ageing population in Denmark calls for action to include marginalised sections of the workforce: the forecasts predict that there will be more older, and fewer younger, people in the population, along with increasing problems in recruiting qualified labour. Alongside the launch of the Danish national action plan for employment, similar plans have been launched to reduce rates of sick leave and 157
Unemployment and Health ■ Health Effects
workers compensation, and to promote the safe return to work following workers’ compensation leave. The revitalisation of the unemployment issue also calls for attention to the expected future increase in unemploymentrelated problems. The task of alleviating the psychological and social problems that attend LTU is bound to increase with the rise in unemployment rates that has been identified in several studies (e.g., Berth, Forster, & Brahler, 2003; Borg & Burr, 1997; Breslin & Mustard, 2003; Byrne, Agerbo, Eaton, & Mortgensen, 2004; Iversen & Sabroe, 1988; Kieselbach, 2003; Kivimäki et al., 2003; Marmot 1999; Sermet & Khlat, 2004). In order to focus initiatives to increase reemployment and limit the negative individual and social consequences of rising unemployment, an understanding of the basic epidemiology of LTU is needed. The aim of this study is to identify risk groups for LTU on the basis of such individual demographic characteristics as gender, age, education and occupational group, as well as such employment and employer-related characteristics as employment sector, employer size and employer ownership. The study focuses on the situation preceding the rise in unemployment in Denmark that occurred after 2002.
Method Participants The study is based on the database DWECS/DREAM. This is a coupling of the Danish Work Environment Cohorts Study (DWECS; Burr, Bjorner, Kristensen, Tuchsen, & Bach, 2003) with the national register on social transfer payments (DREAM). DWECS features a random sample of 5357 employees aged 18–64, who were interviewed in 2000 regarding job, work environment exposures, age, gender, education and employer characteristics. The cohort was followed up in the DREAM register during the period January, 2001 to June, 2003. DREAM is a register based on data from the Danish ministries of Employment, Social Affairs, and Education. DREAM contains information on all social transfer payments for all citizens in Denmark since mid-1991, including not only unemployment benefits, but also workers compensation payments, education grants and pensions. The type of social transfer payment is reported per week for each person. The DREAM database includes records for approximately 3.2 million people and is updated every 3 months. The register is supplemented with information on ethnic background, marital status, town of residence, unemployment fund membership, immigration, transition to old-age pension and mortality. The weekly information on transfer payments is registered if a person has received any kind of transfer payment for more than one day. The present study analyses determinants of LTU, using data gathered with the baseline DWECS questionnaire (determinants) and unemployment data from DREAM among the 5357 persons categorised as employees for at least 2 months prior to baseline. 158
Determinants of Long-Term Unemployment
Measures Outcome: Long-Term Unemployment The outcome, LTU, was defined as having unemployment benefits for at least 8 weeks during the follow-up period from January, 2001 to June, 2003. Determinants Demographic characteristics. Data were obtained on employees’ gender and baseline age group (18–29 years, 30–39 years, 40–49 years, 50–59 years, 60+ years). Education was divided into five categories of postschool education: higher education longer than 4 years, higher education 3–4 years, higher education shorter than 3 years, vocational training and none. The composition of the groups is shown in Table 2. Employer characteristics: Employer size was measured according to number of employees at the participant’s workplace. Categories were: 1–4, 5–9, 10–19, 20–49, 50–99, 100–499 and 500 or more employees. Employer ownership was defined according to whether the workplace belonged to a private or government-owned company and was divided into four categories of ownership: private, state, county and municipality. Occupational group: the population was divided into 12 occupational groups covering a range of unskilled, skilled and professional jobs. The categories are shown in Table 4. Employment sector: participants were categorised according to 11 employment sectors. An overview of the categories can be found in Table 5.
Analysis Logistic regression methods were used to analyse associations between baseline determinants and LTU. The SAS procedure PROC GENMOD (SAS version 8.02) was used to perform the logistic regression analyses. Risk is expressed in Prevalence Proportion Ratios (PPR; Skov, Deddens, Petersen, & Endahl, 1998). A PPR of 1.50 for unemployment among exposure Group A versus exposure Group B, means that Group A has a 50% excess risk of unemployment compared to Group B. In Tables 2 through 5 the presented PPR values are unadjusted for other determinants. This is due to the aim of this study, which was to identify risk populations and jobs rather than causal relations.
Results During the final week of follow-up, 1035 persons (19.3%) received one or more types of social transfer payment. Of these, the largest group (25.5%) received unemployment-related benefits, while 16.1% and 15.7% received education grants and early retirement pensions, respectively, and 11.2% received workers’ compensation payments. Table 1 shows the distribution of transfer payments. Sixteen (1.5%) of the sample had died. During the follow-up period (i.e., January, 2001 to June, 2003 inclusive), 611 persons (12%) experienced one or more periods of LTU. During 159
Unemployment and Health ■ Health Effects Table I Distribution of Social Transfer Payments the Week Prior to Follow-Up (N = 5357) Social transfer payment
N
%
Unemployment (related) benefits Sickness absence compensation Maternity related sickness absence compensation Supported employment Leave (parental, educational) Government grants, education Disability retirement pension Rehabilitation Early retirement pension Old age pension Emigrated Dead
264 116 101 51 5 167 31 57 160 24 43 16
25.5 11.2 9.8 4.9 0.5 16.1 3.0 5.5 15.5 2.3 4.2 1.5
1035
100.0
Total
the same period 4472 persons (88%) experienced no LTU or periods of unemployment lasting less than 8 weeks. People who during follow-up died, emigrated or permanently retired (274 persons, 5.1%) were excluded from further analysis. Tables 2 through 5 are based on those 611 people experiencing LTU and the 4472 experiencing no or shorter periods of unemployment, totalling 5083 people. Table 2 shows that the effect of gender on LTU was not significant. With regards to age, however, all age groups apart from those aged 60 and over were significantly less likely to have experienced LTU than those aged 18–29. There was a statistically significant association between education and LTU: those having no secondary education had a 1.8 higher prevalence of LTU than those with 4 or more years of higher education (Table 2).
Table 2 Demographic Determinants of LTU (N = 5083) Determinant p, determinant Gender
.12
Age
< .00
Education
< .00
Level
n
PPR
95% CI
Female Male 60+ 50–59 40–49 30–39 18–29 None Vocational training Higher education < 3 yrs Higher education 3 ≤ 4 yrs Higher education 4+ yrs
2488 2595 85 1029 1377 1529 1063 1066 2131 516 863 491
1.12 1.00 0.60 0.74 0.62 0.74 1.00 1.82 1.27 1.28 0.80 1.00
0.97–1.31
.12
0.30–1.17 0.60–0.92 0.51–0.92 0.61–0.90
.13 < .00 < .00 < .00
1.34–2.47 0.94–1.71 0.89–1.84 0.56–1.15
< .00 .12 .18 .24
Note: Determinants mutually unadjusted.
160
p, level
Determinants of Long-Term Unemployment Table 3 Employer Characteristics and LTU (N = 5083)
p, determinant
Level
n
PPR
95% CI
p, level
Employer ownership
< .00
0.89 0.63 1.50 1 2.60 2.64 2.19 1.90 1.26 1.56 1
.53 .08 < .00
< .00
1059 391 139 447 380 490 715 952 745 1120 608
0.63–1.27 0.40–1.06 1.11–2.03
Employer size
Municipal County Private State 1–4 5–9 10–19 20–49 50–99 100–499 500+
1.79–3.77 1.85–3.77 1.54–3.10 1.35–2.68 0.86–1.85 1.10–2.21
< .00 < .00 < .00 < .00 .23 .02
Determinant
Note: Determinants mutually unadjusted
Concerning employer characteristics, there were no significant differences in prevalence of LTU between the three groups of government employees (i.e., state, county or municipality). However, those employed by a private company were 50% more likely to be long-term unemployed than those employed by the government. There was a strong association between employer size and LTU, that is, the smaller the employer, the higher the risk of LTU. Those working for an employer with less than 10 employees had more than 2.5 times the risk of LTU than did those working for an employer with more than 500 employees (Table 3). Table 4 shows PPR values for LTU in 12 different occupational groups. Each group is compared to the remaining 11 groups. The results show that people in healthcare jobs, managers, computer professionals, technicians and designers, primary and secondary school teachers, professionals, and
Table 4 Job Group and LTU; PPR for Each Job Group Versus All Others (N = 5033) Job group Others Cleaners, janitors, kitchen staff Unskilled workers and labourers Sales people Skilled workers and craftsmen Kindergarten teachers and day-care jobs Clerks and bookkeepers Professionals Primary and secondary school teachers Computer professionals, technicians, designers Managers Health care jobs
N
PPR
95% CI
p value
1370 205 402 329 357 370 588 237 272 193 451 259
1.67 1.46 1.45 1.11 1.04 0.78 0.77 0.63 0.57 0.56 0.46 0.41
1.43–1.95 1.07–2.00 1.15–1.83 0.83–1.48 0.78–1.39 0.57–1.08 0.59–1.00 0.40–0.98 0.37–0.89 0.89–0.95 0.32–0.68 0.24–0.70
< .00 .02 < .00 .49 .79 .14 .05 .04 .01 .03 < .00 < .00
Note: 50 people excluded due to missing values on job group variable. Job groups with significantly higher (95% level) prevalence of LTU printed in bold. Job groups with significantly lower (95% level) prevalence of LTU printed in bold italics.
161
Unemployment and Health ■ Health Effects Table 5 Sector and LTU; PPR for Each Sector Versus All Others (N = 4732) Sector Construction Graphic industries Services Trade Agriculture Manufacturing Transport Private administration Education and research Health care and social work Public administration
N
PPR
95% CI
p value
239 76 399 231 120 916 500 439 454 936 422
1.68 1.68 1.40 1.30 1.25 1.06 0.94 1.02 0.82 0.74 0.54
1.28–2.20 1.06–2.67 1.10–1.78 0.95–1.78 0.83–1.96 0.88–1.29 0.72–1.22 0.78–1.33 0.62–1.10 0.60–0.92 0.37–0.77
< .00 .03 < .00 .11 .27 .54 .63 .88 .18 < .00 < .00
Note: 351 people excluded due to missing values on sector variable. Sectors with significantly higher (95% level) prevalence of LTU printed in bold. Sectors with significantly lower (95% level) prevalence of LTU printed in bold italics.
clerks and bookkeepers (borderline significant) all had a lower than average prevalence of developing LTU during the study period. Excess prevalence of LTU was found among cleaners, janitors and kitchen staff, unskilled workers and ‘others’ (Table 4). Finally, Table 5 shows the results of analysis for associations between 11 employment sectors and LTU. Each of the 11 sectors was compared to the remaining 10 sectors. Those working in public administration had an LTU prevalence that was half the average prevalence in the remaining sectors. Those working in healthcare and social work also had a below average prevalence. Employees in the construction sector, the graphic industries (PPR = 1.68) and the service sector all had an above average prevalence of LTU.
Discussion During follow-up, 611 persons (12% of the sample) experienced at least one period of 8 weeks or longer in receipt of unemployment benefits. Higher prevalence of LTU was associated with being younger, having a low educational level, being employed in a private company (rather than the public sector), and being employed by a small (as opposed to a large) employer. Cleaners, janitors and kitchen staff and unskilled workers had above average rates of LTU. People in healthcare jobs, managers, computer professionals, technicians and designers, primary and secondary school teachers, professionals, and clerks and bookkeepers had a lower LTU prevalence than average. The construction sector, the graphic industries and the service sector all had an above average prevalence, whereas the healthcare and social work sectors, and the public administration sector, had a lower prevalence. 162
Determinants of Long-Term Unemployment
Before discussing the implications of these findings, certain limitations of the study should be noted. First, the information on exposure and outcome was collected at two points in time, especially with regards to those experiencing LTU in the latter part of the follow-up period. For these people, it means that the researchers had no knowledge of processes and events taking place during follow-up: important life events, changes in work environment or lifestyle occurring after the baseline interviews, which could have influenced the results, are not taken into account. Second, the exposure assessment was based on point estimates: the duration of exposures was not measured. The design also has several strengths, however. First, the study was based on a large representative cohort of employees in Denmark. Second, it was a prospective study over 2.5 years, with a temporal relationship between exposures and outcome allowing causal interpretation. Third, baseline measures were questionnaire-based, whereas the outcome was established using register data: this eliminates possible ‘common method variance’ and the related positive bias (Spector, 1987; Williams, Cote, & Buckley, 1989). The study to some extent confirms previous Danish and international results concerning associations between sociodemography and LTU. With regards to education, results from the Danish National Institute of Social Research showed an increased risk of ‘marginalisation’ (a combined measure indicating transition from employment to permanent exclusion from the workforce) the lower the level of education (Ploug & Søndergaard, 1999). The present study indicates a similar trend. The Employment in Britain Survey (Gallie, White, Cheng, & Tomlinson, 1998) investigated the association between employer size and unemployment spells. There was no association in the British study that could support the strong association found in the present study. However, the British study used broader categories of employer size than were used in this study, probably related to the differences in mean size of corporations in the two countries. Also, previous periods of economic recession in Denmark have shown smaller workplaces to be more vulnerable, which may also be related to the finding of increased risk of LTU among employees in private companies. Thus, there is a link between size and ownership, with smaller workplaces often being privately owned. The effect of employer ownership (state vs. private) is probably not a direct effect but rather mediated by a different set of job types with different characteristics and conditions. However, the finding in this study is highly significant and is an indicator of where to focus future attention and intervention. With regards to occupational group and sector, the present study showed a change in the situation in Denmark compared with that of a decade ago. In a study of the period 1990 to 1994, Hussain and Geerdsen identified manufacturing as the greatest source of marginalised persons 163
Unemployment and Health ■ Health Effects
(Hussain & Geerdsen, 1998). We could not confirm this finding: in our study, people employed in the manufacturing sector had only an average risk of making the transition from employment to LTU. Employees in the graphic sector, on the other hand, had an above average risk, but this sector constitutes a relatively small part (1.6%) of the Danish workforce. The main contributing sectors were construction and services, which together accounted for 13.5% of the study population. The construction sector is usually affected in times of recession, due to the decline in building activities that follows an economic downturn. The above-average risk found in the service sector may have been due to cutbacks in national financial subsidies to this sector. These subsidies had been introduced in the 1990s to decrease unemployment. The findings of excess risk in the service sector were also reflected in the findings of excess LTU between sectors, as cleaners, janitors and kitchen staff (often employed in the service sector) had above average risk. Unskilled workers also had an above average risk, which is in accordance with earlier Danish and international studies identifying low qualification level jobs as those most vulnerable in times of financial cutback (Hussain & Geerdsen, 1998; Gallie et al., 1998). Overall, the present study documents the need to adopt age-, education-, and job-specific perspectives when realising the proposed policy intention on decreasing, and alleviating the effects of, LTU, in relation to the recent unemployment developments on the Danish labour market.
References Berth, H., Forster, P., & Brahler, E. (2003). Unemployment, job insecurity and their consequences for health in a sample of young adults. Gesundheitswesen, 65, 555–560. Borg V., & Burr H. (1997). Work environment and health of Danish employees 1990–95. Copenhagen, Denmark: National Institute of Occupational Health. Breslin, F.C., & Mustard, C. (2003). Factors influencing the impact of unemployment on mental health among young and older adults in a longitudinal, population-based survey. Scandinavian Journal of Work Environmental Health, 29, 5–14. Burr, H., Bjorner, J.B., Kristensen, T.S., Tuchsen, F., & Bach, E. (2003). Trends in the Danish work environment in 1990–2000 and their associations with labor-force changes. Scandinavian Journal of Work Environmental Health, 29, 270–279. Byrne, M., Agerbo, E., Eaton, W.W., & Mortensen, P.B. (2004). Parental socio-economic status and risk of first admission with schizophrenia: A Danish national register based study. Social Psychiatry and Psychiatric Epidemiology, 39, 87–96. Danish Government. (2003). NAP 2003: The Government: Denmark’s National Action Plan for Employment 2003. Copenhagen, Denmark: Danish Ministry of Employment. Gallie, D., White, M., Cheng, Y., & Tomlinson, M. (1998). Restructuring the employment relationship. Oxford, UK: Clarendon Press. Hussain, M.A., & Geerdsen, L.P. (1998). Erhvervene og de langtidsledige — marginalisering og integration på arbejdsmarkedet [Occupations and the long-term unemployed — marginalisation and integration in the labour market]. Socialforskningsinstituttet, 98, 1–84.
164
Determinants of Long-Term Unemployment Iversen L., & Sabroe S. (1988). Psychological well-being among unemployed and employed people after a company closedown: A longitudinal study. Journal of Social Issues, 44, 141–152. Kieselbach T. (2003). Long-term unemployment among young people: The risk of social exclusion. American Journal of Community Psychology, 32(1/2), 69–76. Kivimäki, M., Vahtera, J., Virtanen, M., Elovainio, M., Pentti, J., Ferrie, J.E. (2003). Temporary employment and risk of overall and cause-specific mortality. American Journal of Epidemiology, 158, 663–668. Marmot, M.G. (1999). Job insecurity in a broader social and health context. In J.E. Ferrie, M.G. Marmot, J. Griffiths, et al. (Eds.), Labour market changes and job insecurity: A challenge for social welfare and health promotion (pp. 78–105). World European series. No. 81. Copenhagen: WHO Regional Publications. Ploug, N., & Søndergaard, J. (1999). The future of the welfare society. Copenhagen, Denmark: The National Danish Institute of Social Research. Sermet, C., & Khlat, M. (2004). Health of the unemployed in France: Literature review. Revue Epidemiologique de la Santé Publique, 52, 465–474. Skov, T., Deddens, J., Petersen, M.R., & Endahl, L. (1998). Prevalence proportion ratios: Estimation and hypothesis testing. International Journal of Epidemiology, 27, 91–95. Spector, P.E. (1987). Method variance as an artefact in self-reported affect and perceptions of work: Myth or significant problem? Journal of Applied Psychology, 72, 438–443. Williams, L.J., Cote, J.A., Buckley, M.R. (1989). Lack of method variance in self-reported affect and perceptions at work: Reality or artefact Journal of Applied Psychology, 74, 462–468.
165
Unemployment and Health ■ Health Effects
166
Chapter
14
Organisational Citizenship Behaviours in Relation to Job Status, Job Insecurity, Organisation Commitment and Identification, Job Satisfaction and Work Values N.T. Feather and Katrin A. Rauter
There is now a large body of literature concerned with organisational citizenship behaviours (OCBs). These are behaviours that help the organisation but may not be directly or explicitly recognised in the organisation’s formal reward system (for recent reviews, see Organ & Ryan, 1995; Podsakoff, Mackenzie, Paine, & Bachrach, 2000). For example, a person would be performing organisational citizenship behaviour when he or she stays late to finish work when not specifically asked to do so, or goes out of his or her way to help a co-worker who is having difficulty when that is not part of the role requirement of the job. The main aim of the present study was to investigate OCBs in relation to permanent or temporary job status (Beard & Edwards, 1995; De Witte, 1999), comparing permanent employees with contingent employees who were on a contract. We were also interested in whether the permanent/ contract distinction would be associated with differences in affective commitment to the organisation, organisational identification, job satisfaction and work values. The study extends the investigation of OCBs by focusing on the effects of job status (permanent vs. contract) and also by setting hypotheses within a motivational framework that relates OCBs to a person’s expectations concerning valued goals. Job Status and OCBs Studies that compare OCBs for workers employed on short-term contracts with OCBs for permanent employees are few in number and provide con167
Unemployment and Health ■ Health Effects
tradictory results (e.g., Pearce, 1993; van Dyne & Ang, 1998). Van Dyne and Ang (1998), in a study conducted in Singapore with professional workers from a bank and a hospital, found that contingent workers engaged in fewer OCBs and had lower affective commitment to their organisations. They argued that there would be less pressure for contingent workers, who receive fewer tangible and intangible benefits from their employing organisation, to perform OCBs when the market is one in which there are severe shortages of labour and when their choice of contingent job status would be more likely to be voluntary. However, we would argue that, in labour markets where organisations are downsizing or where there is an oversupply of jobs in a particular area, workers would be more likely to enter contingent work arrangements involuntarily. Contingent workers on short-term contracts might then perform more OCBs in the expectation that their willingness to do so would enhance their image as valued employees, thereby increasing their chances of being made permanent within their organisation (see also van Dyne & Ang, 1998, p. 701). Permanently employed workers would not have this pressure and the OCBs that they performed would be related to other goals. Involuntary contingent workers might also face restrictions in their opportunities to satisfy work values such as security, skill utilisation, variety and control. They might then compensate by performing OCBs that would allow them to achieve goals relating to these values, depending on the importance of these work values for self. For example, involuntary contingent workers who felt that their skills were underutilised and who felt that they had little control or influence over their role-related duties might perform OCBs in order to express and demonstrate their skills and capacity to exert control (see also Parker, Griffin, Sprigg, & Wall, 2002, for an exchange theory analysis). Similarly, we might expect that OCBs among involuntary contingent workers would be positively related to the insecurity that is part of their temporary job status (Beard & Edwards, 1995; De Witte, 1999). In this case, the OCBs would represent attempts to strive towards a more secure, permanent job, and they would be more evident the more a person perceived security to be an important goal. In summary, we propose that it is important to consider the motivational functions of OCBs for workers with involuntary, contingent job status, taking account of the goal structure of individuals and the expectations they hold about whether OCBs will be instrumental to attaining goals that are important to them. We make no specific predictions about relations between OCBS, feelings of insecurity and opportunities for skill utilisation, control, and variety for permanent employees. However, our data analysis will compare these relations for employees with temporary versus permanent job status. OCBs and Other Variables A secondary aim of the study was to investigate affective commitment to the organisation, organisational identity and job satisfaction in relation to job 168
Organisational Citizenship Behaviours
status and OCBs. Theories that refer to social exchange agreements, psychological contracts and social identity are relevant to this part of our study (e.g., Ashforth & Mael, 1989; Hogg & Terry, 2001; Parker et al., 2002; Rousseau, 1995; van Dyne & Ang, 1998). We expected to find more evidence of affective commitment and organisational identity among permanent employees who had secure jobs when compared with involuntary contingent workers. We also expected to find more evidence of positive relations between OCBs and affective commitment and identity when job status was permanent. The review of studies by Podsakoff et al. (2000) showed that these sorts of variables had significant positive relations with citizenship behaviours. Organisational Setting Our study involved a sample of teachers from public schools in Victoria, Australia, who either had permanent employment within the state school system or were contingent workers hired on a short-term contract. The teachers were involuntary contingent workers on fixed-term contracts who, given the choice, would prefer permanent employment as teachers. When our study was in the final planning stage, discussions were taking place between the Victorian Department of Education and the Australian Education Union that led to an agreement in March, 2000, that almost all contract positions would be converted to permanent positions by the end of the school year. Thus, it may be assumed that the contract teachers in the present sample could reasonably expect to become permanently employed as teachers in the future, providing they met performance criteria. However, they were still under contract and we would expect them to feel less secure in their jobs when compared with the permanently employed teachers. Hypotheses We tested the following main hypotheses: H1. Contract teachers whose status is involuntary will perform more OCBs when compared with permanently employed teachers. They will also report stronger feelings of insecurity and fewer opportunities to use their skills, to exercise influence and control, and to achieve variety in the workplace. H2. Permanently employed teachers will report stronger affective or emotional commitment to their school and stronger identification with their school when compared with involuntary contract teachers. H3. OCBs will be positively related to reported feelings of insecurity and negatively related to reported opportunities for skill utilisation, influence and control, and variety for the involuntary contract teachers. H4. OCBs will be positively related to higher emotional commitment to their school and with higher identification with their school for the permanently employed teachers. 169
Unemployment and Health ■ Health Effects
Method Participants and Procedure There were 154 participants from Victorian public schools, of whom 101 were permanently employed and 53 were employed on short-term contracts. There were 74 female teachers, 26 male teachers, and 1 teacher who did not specify gender in the permanently employed group. The mean age of the permanent teachers was 41.75 years (SD = 9.62) and they had been employed for 11.14 years on the average (SD = 3.12). The contract group comprised 44 female teachers and 9 male teachers. The mean age of these contract teachers was 28.67 years (SD = 6.39) and they had been employed for 2.06 years on the average (SD = 4.30). The percentage of contract teachers who indicated that they would prefer permanent employment as a teacher rather than contract work was 96%. These teachers also gave very strong endorsements on 5-point rating scales to the difficulty of finding permanent work and having little other choice of employment as important reasons for pursuing contract work. Other reasons such as associating contract work with more variety and freedom, having more control over one’s life in contract work, and the need to balance commitments to work and family were either perceived to have no role or to have only a weak role for this sample. Distribution of questionnaires was arranged through the Victorian branch of the Australian Education Union (AEU) with their support. The questionnaire was mailed with reply-paid envelopes to 370 AEU members and 154 were returned, generating a response rate of 41.6%.
Questionnaire We included a number of variables in the questionnaire, of which the following are most relevant to our analysis. Job insecurity. Job insecurity was measured by the following three items: ‘I am worried about having to leave my job before I would like to’, ‘I am not concerned about leaving my job in the near future’ and ‘I feel uneasy about losing my job in the near future’. Participants answered each item by using a rating scale numbered from 1 (strongly agree), through 3 (neither agree nor disagree) to 5 (strongly disagree). Responses to the second item were reverse coded. The internal reliability of the scale (Cronbach’s α) was .86. Total scores could range from 3 to 15 with higher scores indicating more perceived job insecurity. Organisational commitment. This variable was measured by using an 8item affective commitment scale from the organisational commitment questionnaire developed by Allen and Meyer (1990). Items in the scale were adapted so as to refer to the teacher’s emotional commitment to the school. Examples of items are ‘I would be happy to spend the rest of my career with this school’, ‘I enjoy discussing my school with people outside it’ and 170
Organisational Citizenship Behaviours
‘I do not feel emotionally attached to this school’. Each item was answered by using a 1–7 rating scale numbered from 1 (strongly disagree), through 4 (neither agree nor disagree) to 7 (strongly agree). Four negatively worded items were reverse coded. Total scores could range from 8 to 56 with higher scores indicating more organisational commitment (α = .85). Organisational identification. We used the 6-item organisational identification measure described by Mael and Ashforth (1992), adapting it for the school context. Examples of items are ‘When someone criticises this school, it feels like a personal insult’, ‘When I talk about this school I usually say “we” rather than “they”’ and ‘This school’s successes are my successes’. Each item was answered by using a 1–5 rating scale numbered from 1 (strongly agree), through 3 (neither agree nor disagree) to 5 (strongly disagree). Scores on the eight items were subsequently reversecoded. Total scores could range from 6 to 30 with higher scores indicating stronger organisational identity (α = .85). Job satisfaction. We used five items from the Job Diagnostic Survey (Hackman & Oldham, 1975) to measure general job satisfaction. Examples of items are ‘Generally speaking, I am very satisfied with this job’, ‘I frequently think of quitting this job’ and ‘Most people in this position are very satisfied with the job’. Participants used a 1–7 rating scale numbered from 1 (strongly disagree), through 4 (neither agree nor disagree) to 7 (strongly agree). Scores on two negatively worded items were subsequently reverse coded. Total scores could range from 5 to 35 with higher scores indicating more job satisfaction (α = .76). Work values. We used three 1–5 rating scales designed to measure the extent to which the permanent and contract teachers perceived themselves to have opportunities to satisfy work values concerned with influence or control, variety and skill utilisation in their current job. These scales were previously used in studies of employment and unemployment (Feather & O’Brien, 1986). The influence scale involved five items concerned with deciding about the design or layout of the workplace, speed of work, new skills learned, organisation of work and choice of co-workers. Answers were coded from 1 (no influence), through 3 (some influence) to 5 (a great deal of influence) for each item. Total influence scores could range from 5 to 25 (α = .79).The variety scale involved three items concerned with the provision of change or variety in the current job, change in location, and frequency of interaction with other people. Answers were coded from 1 (not at all), through 3 (some) to 5 (a very great deal) for each item. Total variety scores could range from 3 to 15 (α = .81). The skill-utilisation scale involved four items concerned with opportunities for learning new skills on the job, using abilities, using education and experience, and working in the way you think best. Answers were coded from 1 (not at all), through 3 (some) to 5 (a very great deal) for each item. Total skill-utilisation scores could range from 4 to 20 (α = .88). 171
Unemployment and Health ■ Health Effects Table 1 Mean Scores and Standard Deviations (SDs) for Permanent and Contract Teachers and Results of t Tests Mean scores Variable
Mid-point Permanent teachers
Job insecurity Organisational commitment Organisational identification Job satisfaction Influence Variety Skill utilisation Organizational citizenship behaviours (OCBs)
SD
Contract teachers
SD
df
t
9 32 18 20 15 9 12
5.76 38.95 21.73 24.18 15.91 8.91 15.13
2.82 9.63 4.87 6.05 4.15 2.59 3.18
10.75 38.23 21.96 25.52 14.74 8.19 14.58
3.73 10.37 4.46 4.58 3.60 2.43 3.33
152 151 151 151 150 148 152
–9.32*** .43 –.29 –1.40 1.74† 1.66† .99
30
36.60
4.30
39.35
4.16
148
–3.76***
Note: †p < .10; ***p < .001 (two-tailed test). Degrees of freedom varied because of minor missing cases.
Organisational citizenship behaviours (OCBs). Citizenship behaviours were measured by adapting a scale devised by Wittig-Berman and Lang (1990). There were 10 items in the scale. Examples of items are ‘I take work home or stay late to finish up work, even if not specifically asked to do so’, ‘I postpone my vacation or day off, in spite of personal inconvenience, to meet the needs of my school’, ‘I talk about work during lunch’ and ‘I go out of my way to help a co-worker who is having difficulty in his or her job’. Each item was answered by using a 1–5 rating scale numbered from 1 (never), through 3 (sometimes) to 5 (always). Two negatively worded items were reverse coded. Total scores could range from 10 to 50 with higher scores indicating more OCBs (α = .70).
Results Differences Between Groups Table 1 presents the mean scores and standard deviations (SDs) for the measured variables for the permanent and contract teachers and the results of t tests. These results show that, as hypothesised, the contract teachers reported more job insecurity when compared with the permanently employed teachers and they also reported a greater frequency of OCBs. The permanent teachers tended to score higher on opportunities for influence and variety in their job (ps < .10). None of the other differences were statistically significant. Correlations Between Variables and OCBs for Total Sample Table 2 presents the zero-order intercorrelations between variables for all teachers. These results show statistically significant positive relations between OCBs and job insecurity, organisational commitment, organi172
7.48 38.70 21.81 24.63 15.50 8.65 14.94 37.53 1.77 37.25 1.34
M
3.95 9.87 4.72 5.62 4.00 2.55 3.23 4.43 .42 10.64 .48
SD
(.86) .01 .07 –.02 –.19* –.10 –.18* .20* .06 –.36*** .60***
1 (.85) .57*** .54*** .36*** .26*** .43*** .15† .09 .02 –.04
2
(.85) .31*** .32*** .20* .27*** .18* –.06 –.17* .02
3
(.76) .45*** .40*** .51*** .11 .04 –.26*** .11
4
(.79) .50*** .63*** –.05 –.07 .00 –.14†
5
(.81) .67*** .14† .02 –.06 –.13†
6
(.88) .07 .03 –.11 –.08
7
(.70) .05 –.26*** .30***
8
–.24** .10
9
10
–.59***
Note: †p < .10; *p < .05; ** p < .01; ***p < .001. N = 154 for the total sample with variations due to minor missing cases. Alpha coefficients are in parentheses. Gender was coded 1 = male, 2 = female. Job status was coded 1 = permanent, 2 = contract, OCBs = organisational citizenship behaviours. Significance tests for the correlations are two-tailed.
1 Job insecurity 2 Organisational commitment 3 Organisational identification 4 Job satisfaction 5 Influence 6 Variety 7 Skill utilisation 8 OCBs 9 Gender 10 Age 11 Job status
Variable
Means, Standard Deviations (SDs), and Correlations Between Variables for All Teachers
Table 2
Organisational Citizenship Behaviours
173
Unemployment and Health ■ Health Effects Table 3 Partial Correlations Between Job Status, OCBs and Other Variables After Controlling for Age Differences
Job insecurity Organisational commitment Organisational identification Job satisfaction Influence Variety Skill utilisation OCBs Gender
Job status
OCBs
.53*** .00 –.06 –.04 –.20* –.23** –.20* .18* –.02
.09 .15† .13 .06 –.07 .12 .03 — –.01
Note: †p < .10; *p < .05; ** p < .01; ***p < .001. Degrees of freedom = 140. Job status was coded 1 = permanent, 2 = contract, OCBs = organisational citizenship behaviours. Significance tests for the correlations are two-tailed.
sational identification and variety for the combined sample of teachers. In addition, they show that job insecurity was negatively related to opportunities for both influence and skill utilisation. Statistically significant positive correlations were obtained between measures of organisational commitment, organisational identification, job satisfaction, influence, variety and skill utilisation. Older teachers had lower scores on job insecurity, organisational identification, job satisfaction and OCBs. Contract teachers had higher scores on job insecurity and OCBs and lower scores on influence. They were also younger. We also examined the correlations between job status, OCBs and the other variables after controlling for age differences by using partial correlation. Table 3 presents these partial correlations. These results show that there were significant negative correlations between job status and opportunities for influence, variety and skill utilisation, that is, fewer opportunities for the contract teachers. Significant positive correlations were obtained between job status and both OCBs and job insecurity, that is, more of each variable for the contract teachers. The results also show that the correlations between OCBs and job insecurity and opportunities for variety were no longer statistically significant when age differences were controlled. The positive correlation between OCBs and organisational commitment was unaltered. Correlations for Separate Groups and Moderated Regression Analysis Table 4 presents the zero-order correlations between variables for the contract teachers (above diagonal) and the permanent teachers (below diagonal). We do not report the partial correlations after controlling for age differences within each group because this statistical control had virtually no effect on the correlations reported in Table 4.
174
.— .04 .06 –.05 –.13 .03 –.06 –.10 –.07 .02
1
.04 .— .65*** .60*** .39*** .33*** .48*** .29** .02 –.03
2 .10 .43*** .— .33*** .39*** .23* .26** .24* –.07 –.15
3 –.28* .45*** .28* .— .47*** .44*** .51*** .14 .04 –.26**
4 –.13 .30* .19 .51*** .— .45*** .59*** .09 –.12 –.09
5 –.09 .14 .17 .37** .57*** .— .68*** .22* .04 –.12
6 –.32* .33* .29* .56*** .71*** .65*** .— .28*** .11 –.21*
7 .24† –.07 .04 –.05 –.23† .13 –.23† .— .08 –.14
8 .12 .25† –.05 .00 .12 .01 –.12 –.15 .— –.30**
9
β
2.29* –1.73† 1.96*
2.76** 0.39
t
Step 1 Job status Org. commit Step 2 Job status Org. commit. Job status x org. commit.
Variable
Analysis 2
.30 .64 –.51
.30 .16
β
3.83*** 2.75** –2.18*
3.81*** 2.06*
t
β
Step 1 Job status .30 Influence –.01 Step 2 Job status .28 Influence .42 Job status x influence –.45
Variable
Analysis 3
3.55*** 1.75† –1.89†
3.76*** –0.08
t
β Step 1 Job status .31 Skill util. .10 Step 2 Job status .29 Skill util. .77 Job status x skill util. –.70
Variable
Analysis 4
Note: †p < .10; *p < .05; ** p < .01; ***p < .001. For Analysis 1, R2 = .09 for Step 1 (p = .001); ∆R2 = .02 for Step 2 (p = .05). For Analysis 2, R2 = .11 for Step 1 (p = .0002); ∆R2 = .03 for Step 2 (p = .03). For Analysis 3, R2 = .09 for Step 1 (p = .001); ∆R2 = .02 for Step 2 (p = .02). For Analysis 4, R2 = .10 for Step 1 (p = .0005); ∆R2 = .05 for Step 2 (p = .003). Job status was coded 1 = permanent, 2 = contract.
Step 1 Job status .27 Job insecurity .04 Step 2 Job status .23 Job insecurity –.53 Job status x job insecurity .61
Variable
Analysis 1
Results of Hierarchical Regression Analyses for Variables Predicting OCBs
Table 5
t
3.81*** 3.33** –3.06**
3.88*** 1.30
–.09 .10 –.34* –.16 –.14 –.33* –.19 –.06 .04 .—
10
Note: †p < .10; *p < .05; **p < .01; ***p < .001. N = 53 for contract teachers (above diagonal); N = 101 for permanent teachers (below diagonal). Variations from these Ns occurred due to minor missing cases. Gender was coded 1 = male, 2 = female, OCBs = organisational citizenship behaviours. Significance tests for the correlations are two-tailed.
1 Job insecurity 2 Organisational commitment 3 Organisational identification 4 Job satisfaction 5 Influence 6 Variety 7 Skill utilisation 8 OCBs 9 Gender 10 Age
Variable
Correlations Between Variables for Contract Teachers (Above Diagonal) and for Permanent Teachers (Below Diagonal)
Table 4
Organisational Citizenship Behaviours
175
Unemployment and Health ■ Health Effects
We tested the differences between the correlations between OCBs and each of the seven measured variables in Table 4 by using moderated regression analysis. We entered job status (permanent vs. contract) and each variable at the first step of the analysis, and the product of job status and the respective variable at the second step, after centring that variable about its mean. Results at the first step of each analysis provide information about main effects on OCBs for both job status and the variable that was entered. Results at the second step provide information about the interaction of these two variables. The results of these multiple regression analyses showed the statistically significant main effect of job status on OCBs that has already been described (i.e., more OCBs for the contract teachers), and significant main effects for organisational commitment, organisational identification and variety (more OCBs with higher scores in each variable, see Table 2). Of more interest are the statistically significant interaction effects that were obtained. These interaction effects occurred for four variables (job insecurity, organisational commitment, influence and skill utilisation) and the results of the respective regression analyses are reported in Table 5. The form of these interactions can be described by comparing the respective correlations for the contract and permanent teachers. The correlation between OCBs and job insecurity was positive for the contract teachers, r(49) = .24, p < .10, but negative for the permanent teachers, r(97) = –.10, ns; t = 2.00, p < .05. The OCBs/organisational commitment correlation was negative for the contract teachers, r(49) = –.07, but positive for the permanent teachers, r(96) = .29, p < .0l; t = –2.08, p < .05. The OCBs/influence correlation was negative for the contract teachers, r(49) = –.23, p < .10, but positive for the permanent teachers, r(95) = .09, ns; t = –1.88, p < .10. The OCBs/skill-utilisation correlation was negative for the contract teachers, r(49) = –.23, p < .10, but positive for the permanent teachers, r(97) = .28, p < .01; t = –2.97, p < .01. Mediation Analyses We also conducted analyses to investigate whether the variables listed in Table 1 mediated relations between job status and OCBS, following the procedures described by Baron and Kenny (1986). We controlled for age differences at each step. There was no evidence of mediation from these analyses. In particular, the relation between job status and OCBs was not significantly reduced when each variable was included with job status in the regression equation.
Discussion These results show that, as predicted, the contract teachers reported more OCBS, stronger feelings of insecurity and less opportunity to satisfy work values (influence, skill utilisation and variety) when compared with the per176
Organisational Citizenship Behaviours
manently employed teachers. The findings for the OCBs contrast with those of van Dyne and Ang (1998), who found that contingent workers engaged in less OCBs than regular employees. However, we studied involuntary contingent workers whereas van Dyne and Ang studied workers whose contingent job status was more likely to be voluntary. We would expect the contract teachers to undertake extrarole behaviours and voluntary duties that would help their schools and improve their prospects of obtaining permanent employment as teachers, a goal that was clearly important to them. Consistent with hypothesis, we also found that citizenship behaviours for the contract teachers were positively related to feelings of insecurity and negatively related to opportunities for influence and skill utilisation. The corresponding correlations for feelings of insecurity and for influence were nonsignificant for the permanently employed teachers but in their case there were statistically significant positive correlations between OCBs and opportunities for skill utilisation and variety. The implication is that contract teachers who perceived themselves as feeling more insecure in their job, who felt that they had little influence or control over their rolerelated duties, and who also felt that their skills were underutilised, perceived OCBs as a way of achieving valued goals relating to security, influence and the use of skills, and also as a way of presenting a self-image that would increase their likelihood of achieving a permanent job. In contrast, the permanently employed teachers had secure jobs, more responsibilities (e.g., in key decision-making), and engaged in tasks that required more experience as teachers. The OCBs would have a different function for them. Those permanently employed teachers who perceived themselves as having more rather than less opportunity for skill utilisation and variety in their work situation were willing to extend these opportunities further by performing more OCBS, perhaps in this way expressing and fulfilling work values that were important to them or helping to achieve other goals such as obtaining promotion. We found no evidence to support the hypothesis that the permanently employed teachers would show more affective commitment to their school and would report stronger identification with their school when compared with the contract teachers. Nor was job status linked to job satisfaction. However, as predicted, OCBs were positively related to levels of affective commitment and identification for the permanently employed teachers. The corresponding correlations were nonsignificant for the contract teachers. Thus, holding a permanent and secure teaching job with its associated work characteristics provided a context in which citizenship behaviours were linked to both commitment and identification. Future studies of citizenship behaviours should include measures of expectations relating to changes in job status and measures of each person’s goal structure for both permanent and contract employees and for employ177
Unemployment and Health ■ Health Effects
ees at different levels of appointment within organisations. These studies should take account of the labour market and work characteristics and they should sample a range of occupations in order to test the degree to which effects generalise. They should use behavioural observations as well as self-reports so as to minimise common method variance and they should also include controlled longitudinal studies (e.g., Parker et al., 2002) to achieve a better understanding of cause–effect relations. The present study would have been improved had we been able to include measures of teachers’ expectations, had we been able to use measures that went beyond verbal report (e.g., supervisor ratings) and had we been able to sample different types of occupation. At the same time, our study investigated a wide range of variables that included work values and the results usefully add to the relatively sparse literature on permanent versus contract employment. Finally, we framed our study within a functional account of citizenship behaviours. In future research there could be more specific attempts to relate the study design to motivational theory. For example, the expectation variable along with other variables relating to the goal structure of activities could be integrated into a motivational analysis of OCBs based on expectancy–value theory (Feather, 1982), taking account of the beliefs that people hold about the likelihood that OCBs will lead to positive outcomes as well as the perceived attractiveness or value of those outcomes. Such an analysis of expectations and the subjective value of possible outcomes would recognise the functional significance of OCBs for employees who differ in their status within the organisational structure.
Note This chapter was originally published in 2004 in the Journal of Occupational and Organizational Psychology, Volume 77, pp. 81–94. Reproduced with the kind permission of the British Psychological Society.
References Allen, N.J., & Meyer, J.P. (1990). The measurement and antecedents of affective, continuance and normative commitment to the organization. Journal of Occupational Psychology, 63, 1–18. Ashforth, B.E., & Mael, F. (1989). Social identity theory and the organization. Academy of Management Review, 14, 20–39. Baron, R.M., & Kenny, D.A. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. Beard, K.M., & Edwards, J.R. (1995). Employees at risk: Contingent work and the psychological experience of contingent workers. In C.L. Cooper & D.M. Rousseau (Eds.), Trends in organizational behavior (pp. 109–126). Chichester, UK: Wiley. De Witte, H. (1999). Job insecurity and psychological well-being: Review of the literature and exploration of some unresolved issues. European Journal of Work and Organizational Psychology, 8, 155–177.
178
Organisational Citizenship Behaviours Feather, N.T. (1982). Expectations and actions: Expectancy-value models in psychology. Hillsdale, NJ: Erlbaum. Feather, N.T., & O’Brien, G.E. (1986). A longitudinal study of the effects of employment and unemployment on school-leavers. Journal of Occupational Psychology, 59, 121–144. Hackman, R.J., & Oldham, G. (1975). Development of the job diagnostic survey. Journal of Applied Psychology, 60, 159–170. Hogg, M., & Terry, D.J. (Eds.). (2001). Social identity processes in organizational contexts. Philadelphia, PA: Psychology Press. Mael, F.A., & Ashforth, B.E. (1992). Alumni and their alma mater: A partial test of the reformulated model of organizational identification. Journal of Organizational Behavior, 13, 103–123. Organ, D.W., & Ryan, K. (1995). A meta-analytic review of attitudinal and dispositional predictors of organizational citizenship behavior. Personnel Psychology, 48, 775–802. Parker, S.K., Griffin, M.A., Sprigg, C.A., & Wall, T.B. (2002). Effect of temporary contracts on perceived work characteristics and job satisfaction: A longitudinal study. Personnel Psychology, 55, 689–790. Pearce, J.L. (1993). Toward an organizational behavior of contract laborers: Their psychological involvement and effects on employee co-workers. Academy of Management Journal, 36, 1082–1096. Podsakoff, P.M., MacKenzie, S.B., Paine, J.B., & Bachrach, D.G. (2000). Organizational citizenship behaviors: A critical review of the theoretical and empirical literature and suggestions for future research. Journal of Management, 26, 513–563. Rousseau, D.M. (1995). Psychological contracts in organizations: Understanding written and unwritten agreements. Thousand Oaks, CA: Sage. van Dyne, L., & Ang, S. (1998). Organizational citizenship behavior of contingent workers in Singapore. Academy of Management Journal, 41, 692–703. Wittig-Berman, U., & Lang, D. (1990). Organizational commitment and its outcomes: Differing effects of value commitment and continuance commitment on stress reactions, alienation and organization-serving behaviors. Work and Stress, 4, 167–177.
179
Unemployment and Health ■ Health Effects
180
SECTION
Interventions to Limit the Health Effects of Unemployment: Activation Policies and Empowerment
2
Chapter
15
Active Labour Market Programs for Young Long-Term Unemployed: Psychological Impact of Participation in a Recent Program Anthony H. Winefield and Edgar Carson
Active labour market programs have been introduced in major OECD countries at different times since World War II, including, for example, Sweden in the late 1950s, Germany in the late 1960s, the United Kingdom, the United States and Australia in the 1970s, and France in the early 1980s. They have become increasingly important throughout the OECD over the last three decades, with interest in this area stimulated by deteriorating labour market conditions since the early 1970s. From the mid-1970s onwards, a range of programs have been tried at some point in most countries — job-search training, specific training programs (including apprentice training), wage subsidies to private employers and public sector job creation. During the past decade there has been an increasing emphasis on the responsibility of job-seekers to undertake skills acquisition and job search. Entitlement to income support for unemployed people has become increasingly dependent on compliance with individual activity agreements in accordance with a principle of ‘mutual obligation’ (Martin, 1998). The most common arguments for targeting or giving priority to young people who are disadvantaged in the labour market are about equity and access. The success of a program is generally defined by its effectiveness in getting unemployed people into jobs. Little or no consideration is given to the quality of the job. International comparisons, summarised by the OECD, show on balance that: (a) Job search assistance is usually the least costly active labour market program and evaluations from several countries show consistently positive outcomes for this type of program (Meyer, 1995). 181
Unemployment and Health ■ Interventions
(b) Training programs for young people were generally ineffective (Friedlander, Greenberg, & Robins, 1997; Martin, 1998). (c) Private sector wage subsidies show large ‘deadweight’ effects (activities that would have occurred in the absence of the program) and ‘displacement’ effects (subsidised workers displacing unsubsidised workers). Data from Australia, Belgium, Ireland and the Netherlands suggest that combined deadweight and substitution effects amount to 90%. However, wage subsidy programs do offer other benefits such as a redistribution of job opportunities for highly disadvantaged job-seekers and may be justified on equity grounds. (d) Direct public sector job creation represents a higher rate of expenditure in OECD countries than wage subsidies (14% compared to 11%); however, there is considerable debate about whether such jobs represent a disguised form of heavily subsidised employment even though this strategy is acknowledged as important for equity reasons, particularly for young people. While there was a trend away from this type of program in the 1990s, the OECD has observed that ‘it is making its comeback in some OECD countries as part of a principle of “mutual obligation” on the unemployed in return for benefits’ (Martin, 1998, p. 21). The current policy mix varies from country to country and there is never a stable set of current programs to evaluate (Martin, 1998). Moreover, in practice, most countries employ a number of combination programs that aim to integrate elements of the above ‘ideal types’. In the United States, policy emphasises selective labour market programs for the most disadvantaged young people, supplemented by links between learning in school and at work (e.g., Job Corps). In Europe, it is common for subsidised employment to be available to all unemployed youth. The United Kingdom emphasises work-based training for unemployed young people, supplemented by an expansion of full-time vocational training at upper secondary school level (e.g., New Deal for Young People). Swedish policy combines job creation in the public sector with employer-based training and work experience. In Australia, the emphasis has been on skills acquisition through a mix of job-search training, on-thejob skills training and intensive personal assistance.
Some Current Schemes: Employment Outcomes In recent years, schemes that combine features of the four kinds of program listed above have been introduced in many OECD countries. In the United Kingdom, for example, the major program for young people is the New Deal for Young People. It represents a combination program that offers job search training, on-the-job skills training and wage subsidy for employers. Of 392,000 participants in the first two years of its operation up to the end of 1999, some 133,000 participants entered jobs lasting more than 13 182
Active Labour Market Programs
weeks (Hasluck, 2000). Nearly 75% of the vacancies to which New Deal participants were recruited are claimed to represent expansion of the workforce, rather than replacing workers in existing posts; however, employers also said that most of the vacancies (69%) would have existed in the absence of the New Deal (Hales, Collins, Hasluck, & Woodland, 2000). While the New Deal for Young People in the United Kingdom has apparently reduced unemployment, including long-term unemployment, a high proportion of those who found jobs left them within 13 weeks (Hasluck, 2000). More generally, in the United Kingdom it has been found that almost 25% of unemployed claimants who found work applied for benefits again within 3 months, with 39% describing their employment experience as a series of casual jobs and repeated spells of unemployment (Bottomley, McKay, & Walker, 1997; Campbell, Sanderson, & Walton, 1998). Youth programs in the US are at present a subset of the implementation of the Workforce Investment education Act (WIA), which provides for multiple supports including counselling, social skills training, shelter and food and aims to address the long-term employment of youth between the ages of 14 and 21 who live in high-poverty communities. Of three major strands, the best known is the continuation and refinement of the Job Corps program, which was established in 1964 and is the nation’s largest program designed to move youth from poverty to work through focusing on vocational job training and health care. Another main strand is a Youth Opportunity Grants program, which also emphasises a youth development approach to youth employment. Finally, a School-to-Work strand provides seed money to state and local partnerships to develop applied learning and project-based learning and a variety of work experience programs. The Workforce Investment Act has provision for a range of evaluation mechanisms but at present only preliminary process evaluations have been undertaken. Arguably the most dramatic development in ‘client-centred’ combination schemes has been in Australia over the last 5 to 10 years. The new employment services market in Australia, centred on the creation of the Job Network, has been in place just long enough for researchers to start producing preliminary assessments and evaluations. The change has been so dramatic, however, that there are many aspects of the system yet to be scrutinised. The main changes included replacement of the Commonwealth Employment Service (CES), previously the single public provider of jobmatching services, with a new service agency named Centrelink (DEETYA, 1996). As well as delivering income-support payments, including the ‘common youth allowance’, Centrelink assesses the level of assistance jobseekers are entitled to receive from contracted Job Network agencies, and administers the activity test for unemployed clients. The other major aspect of change in employment services involved the abolition of a range of labour market programs delivered under the previous Labor Government’s Working Nation Policy and a shift to a highly targeted strategy delivered 183
Unemployment and Health ■ Interventions
via Job Network agencies. This involved a shift from program-based funding with an emphasis on services provided to job-seekers towards a contestable market in employment services, implementing a competitive model of outcome-based funding with an emphasis on rewarding providers for employment outcomes (DEETYA, 1996). Over the last 5 to 10 years, following the abolition of the Australian ‘Working Nation’ labour market programs in conjunction with the creation of the Job Network, there have been fewer dedicated programs for young job-seekers under 20 years old and no provision for income maintenance for young job-seekers less than 17 years old. Although there are still programs that provide assistance to disadvantaged young job-seekers who are under 18 years old and at risk of homelessness, the major program for young people has been Work for the Dole (WFD), introduced in 1997.
Labour Market Programs and Self-Esteem An assumption underlying most active labour market programs in OECD countries is that they will not only increase employment of participants during or immediately after participation in the program, but also increase employability of participants in the longer term by improving their work acculturation and increasing their self-esteem. Self-esteem has been measured in many psychological studies of unemployment, with mixed results. Some, but not all, have found it to be affected by changes in employment status. Shamir (1986a), for example, in a study of highly educated Israeli adults concluded that: ‘… depressive affect, morale and anxiety are affected by employment status, but, contrary to conventional wisdom, self-esteem is neither sensitive to employment status nor to changes in employment status’ (p. 61). On the other hand, some studies of young people have found that self-esteem is significantly higher in the employed than in the unemployed (Gurney, 1980; Patton & Noller, 1984; Prause & Dooley, 1997) and sensitive to changes in employment status. In their 10-year longitudinal study, Winefield, Tiggemann, Winefield and Goldney (1993) found that in most years, although the employed groups had significantly higher self-esteem than the unemployed groups, similar differences had been observed when the groups were still at school. In only one year were the groups matched on self-esteem while still at school, and in that year there were no ongoing differences in self-esteem associated with employment status. Prause and Dooley (1997), despite showing that self-esteem was associated with employment status in young people, nevertheless caution: ‘Present knowledge of the consequences of self-esteem and the techniques for modifying it do not justify translating the present finding into an urgent call for changing economic policy in the service of raising self-esteem’ (p. 257). Indeed, more generally in the US, National Evaluation of Welfare-toWork Strategies undertaken by Manpower Demonstration Research 184
Active Labour Market Programs
Corporation for the Department of Health and Human Services reports that data from 20 programs were used to define subgroups based on risk of poor psychosocial wellbeing (depression, mastery, preference for work, health or emotional barriers to work). With the exception of depression, there was little relationship between these measures and impact of program (Michalopoulous, Schwartz, & Adams-Ciardullo, 2000). The study reported below was commissioned by the Australian government specifically to evaluate the effect on the WFD program on self-esteem, work attitudes and other measures of psychological wellbeing (Winefield, 1999).
Psychological Evaluation of the Australian Work for the Dole (WFD) Program In 1997 the Australian Government, through the Department of Employment Education, Training and Youth Affairs (DEETYA), nominated three main levels of assistance for job-seekers, namely Job Matching, Job Search Training and Intensive Assistance as well as a New Enterprise Incentive Scheme for approved clients aiming to become self-employed. Intensive Assistance involved case management for the long-term unemployed or those at risk of long-term unemployment, with Job Network agencies being able to determine whatever service they felt was appropriate, including counselling, work experience or wage subsidies. A critical issue is that entitlement to income maintenance for unemployed people has become increasingly dependent on compliance with individual activity agreements. In accordance with a principle of ‘mutual obligation’, unemployed job-seekers in receipt of income support must, in addition to actively seeking work, undertake another activity to improve their competitiveness in the labour market and/or contribute to the local community, including participation in the WFD program. The WFD initiative was aimed to: ‘foster work habits and attitudes, improve participants’ self-esteem; and contribute to local communities by the establishment of projects of value to the community’ (DEETYA, 1997, p. 5). In other words, the initiative embodied the notion of mutual obligation: that it is fair and equitable to expect people to contribute to the community in return for the community’s support (i.e., unemployment benefits, or ‘the dole’). It is based on the premise that because many young people who are studying or working must make do with less or little more than those on unemployment allowances, many Australians believe that it is fair that they be asked to give something in return. Community groups (‘sponsors’) were invited to propose ‘quality’ projects, defined as projects that provide: (1) a variety of work experience; (2) challenging and interesting activities for participants; (3) an opportunity for participants to apply their skills and aptitude; (4) responsibility for the participant; (5) social interaction within the workplace; and (6) pride and satisfaction in results. 185
Unemployment and Health ■ Interventions
At least 70 such projects were to be selected over the first year, with at least 50 located in rural and regional areas of high unemployment. The government provided AU$12.4 million (about US$6.4 million) over the 12month pilot project period to place 10,000 people in these projects. The anticipated benefits to the community were to provide communities both with quality projects considered to be of value to the community and with the opportunity to demonstrate their concern for those who have been unable to find employment. The anticipated benefits to the participants, on the other hand, were: (1) to assist unemployed people in combating the drift to despondency and despair; (2) to foster appropriate work habits in young people; and (3) to improve participants’ self-esteem. Although the scheme targeted young people (18–24 years of age) who had been unemployed for more than 12 months, it was not restricted to this group. The general eligibility criteria were: aged 18 years and over, registered as unemployed for at least 6 months, not in case management and not likely to be referred in 6 months, and on full Newstart (i.e., unemployment) Allowance at the time of commencement. However, in addition to satisfying the general eligibility criteria, projects were also required to include at least 80% of participants who were 18 to 24 years old. Also, although the scheme provided for both voluntary and compulsory referrals, priority was given to young people who had been unemployed for 12 months. Participation was not compulsory for job-seekers over the age of 24, although they were free to volunteer. In return for participation, job-seekers continued to receive the basic government allowance (provided they showed they were still looking for work by making two employer contacts per fortnight) and, in addition, received a supplement of AU$20 per fortnight (about US$15). They were required to work for either 24 hours per fortnight (if aged 18–20) or 30 hours per fortnight (if aged 21 or over). The initial stage of the evaluation was aimed to identify the best way to administer the scheme and the extent to which the scheme should be expanded beyond the pilot stage. Focus groups were to be conducted with unemployed young people before projects began to gather information about work habits, motivation and self-esteem among the target group. The focus groups were to be followed by case studies of specific projects, involving interviews with participants, providers and others who were involved. The second stage was to involve a more rigorous assessment of the effectiveness of the pilot projects. The evaluation reported here forms a part of second-stage evaluation. It should be stressed that the study was envisaged as a pilot study only, to be followed by a much larger investigation. The main purposes of a pilot study are to refine the methodology for the main study, as well as to gather data. One of us (AHW) was commissioned by DEETYA to advise on the conduct of the study and to analyse and interpret the results. 186
Active Labour Market Programs
Program evaluation has been defined as ‘social research used to judge a program’s success’ (Dooley, 1995, p. 296.). It can be divided into the following five phases: (1) needs assessment (describing people’s needs in order to set program goals); (2) program monitoring (checking the program to see how well it conforms to plan); (3) program impact (testing whether the intervention has the expected effect); (4) efficiency (weighing the program’s success by its costs); and (5) utilisation (the extent to which others use the evaluation). The present study was restricted to some aspects of phase (3): evaluating the psychological impact of participation in the WFD Pilot Program. Specifically, the study focused on the following expected benefits: (a) improved self-esteem; (b) improved psychological wellbeing; (c) improved work habits and attitudes; and (d) improved social attitudes (e.g., acceptance of the concept of mutual obligation). Program evaluation, like all forms of nonexperimental research, has to contend with several threats to internal validity (Cook & Campbell, 1979). Probably the most serious of these is that posed by the failure to achieve randomisation — a feature of a true experiment. The full use of experimental method with random assignment is confined largely to the US, and broad evaluations of youth programs indicate limited, even perverse, effects on crime rates, drug use and teenage pregnancy, although all of these findings are subject to considerable uncertainty (Orr, Bloom, Bell, Lin, Cave, & Doolittle, 1994; Quint, Polit, Bos, & Cave, 1994). Outside the United States, there is limited use of full experimental design. In the United Kingdom, for example, current evaluations of the New Deal programs involve random assignment to experimental and control groups in pilot programs but at this stage of its operation, only process evaluation and preliminary outcome figures are available. The research design used in this evaluation is known as an ‘interrupted time series design’. Such designs involve obtaining baseline measures (observations made before the treatment) and measures taken after the treatment. They are a form of within-subject design and are commonly used in program evaluation. They are generally regarded as inferior to ‘the untreated control group design with pretest and posttest’ because the lack of an untreated control group allows the possibility that any observed change could be due to historical factors, rather than to the treatment. Unfortunately, the inclusion of an untreated control group was not deemed to be feasible in the present case, as is the case in the majority of labour market program evaluations throughout the OECD, which typically use a quasi-experimental design (Martin, 1998).
187
Unemployment and Health ■ Interventions
Method Participants The longitudinal methodology employed involved contacting participants on two occasions, first, before they had commenced work (Time 1), and second, near the expected completion of the 6 months spent working for the dole (Time 2), although for some participants the second interview took place up to 9 months later. Data are reported from 149 of the 156 Time 1 participants who were contacted between early January and early May (90 men and 59 women). Gender was not recorded for seven of them. Of these, 107 were able to be recontacted at Time 2 (59 men and 48 women). Ages ranged from 18 to 24 years. Measures Self-esteem. The most widely used self-esteem inventory in the youth unemployment literature is the Rosenberg Self-Esteem Inventory (RSE). It is a 10item measure of global self-esteem that was specifically designed for use with young people (Rosenberg, 1965). It has excellent psychometric properties (reliability and validity) and was used in the 10-year longitudinal study of youth unemployment by Winefield et al. (1993), as well as in similar Australian longitudinal studies reported by Gurney (1980) and by Patton and Noller (1984) and recent longitudinal studies from the United Kingdom (Meân Patterson, 1997) and the United States (Prause & Dooley, 1997). Negative mood. A 7-item scale, devised by Tiggemann and Winefield (1984), was used in the Winefield et al. (1993) study, as well as some recent studies carried out in North America, and was found to have very good reliability and validity. For example, Winefield, Winefield, Tiggemann and Goldney (1991) reported an internal reliability coefficient of .67, and Winefield et al. (1993) demonstrated that it possessed very good convergent and discriminant validity, correlating highly with depressive affect and GHQ, and discriminating between students or satisfied employed individuals and unemployed or dissatisfied employed individuals. Psychological distress. The 12-item version of Goldberg’s General Health Questionnaire (GHQ; Goldberg, 1972) has been widely used in studies of youth unemployment (and other occupational studies) in both the United Kingdom (Banks & Jackson, 1982) and in Australia (Winefield et al., 1993). Its psychometric properties have been well documented (Banks, Clegg, Jackson, Kemp, Stafford, & Wall, 1980; Winefield, Goldney, Winefield, & Tiggemann, 1989). Two scoring methods are possible: binary and Likert. The former is recommended for the identification of possible ‘cases’, the latter is recommended as a measure of psychological distress and would be more appropriate for the present study. Using Likert scoring, Banks et al. (1980) reported internal reliability coefficents 188
Active Labour Market Programs
of .82 in employed and school-leaver samples, and .90 in an unemployed sample. Winefield et al. (1989) reported coefficients of .84 in a sample of N = 1013 young South Australians (aged from 19–21 years). Social alienation. A 9-item version of Srole’s (1956) 5-item scale for measuring anomia or social alienation, developed by Dodder and Astle (1980), was included. In the longitudinal study by Winefield et al. (1993) it was the single measure that distinguished the unemployed from the dissatisfied employed (the unemployed showed significantly greater anomia). Work involvement. This construct is similar to, yet different from, the Protestant work ethic (PWE). It refers to the importance to the individual of having a job, or, as Shamir (1986b) puts it, ‘a normative belief in the value of work in one’s life’ as compared with PWE, which refers to ‘a belief in the value of hard work and frugality … as a defence against sloth, sensuality, sexual temptation and religious doubt’ (p. 27). Shamir found that work involvement, but not PWE (measured by a shortened version of the Mirels and Garret [1971] scale), moderated the relationship between employment status and psychological state. For this reason it was felt that work involvement may be more appropriate than PWE and it was assessed using a 6-item scale, based on that proposed by Warr, Cook and Wall (1979). The scale was used in the longitudinal study by Winefield et al. (1993) and was shown to possess good internal reliability. Warr et al. (1979) report internal reliability coefficients of .64 and similar coefficients were reported by Winefield et al. (1993). Attitudes to WFD. A series of questions was developed to assess participants’ overall attitudes to working for the dole. It includes attitudes concerning the social justice aspects, whether the scheme should be voluntary, and whether the participant believes participation will improve his/her chances of obtaining a satisfactory permanent job. Background questions. At time 1 demographic information was obtained, including the name, age and sex of the participant, how long the participant had been continuously unemployed, and whether the participant had been a willing volunteer for the scheme or a compulsory participant, the kind of job the participant aspired to, as well as the employer and the nature of the work involved. Open question. All interviewed participants were invited to offer additional comments. Procedure Two psychologists were contracted by DEETYA to administer the interview using computer assisted telephone interviewing. Both psychologists were male, one aged about 25, the other in his late 40s. The same people undertook the data collection for both waves of the study although no 189
Unemployment and Health ■ Interventions
attempt was made to use the same interviewer for both interviews with a given respondent. To maximise responses, at least six attempts were made to contact possible participants at a variety of times (both in and out of business hours) for each wave of the study. Calls were made from about 10 a.m. to 8 p.m. Participants were given the option of completing the interview at a time convenient to them.
Results An alpha level of .05 was used for all statistical tests. Demographic Characteristics of the Sample Altogether, 156 people were interviewed at time 1. However, gender was recorded for only 149 of them (90 men and 59 women). Of these, a few (25) were interviewed during January/February 1998, in the pilot phase of the survey, and the remainder from late March to early May, in the main study. There were no significant differences between the two groups, so their results are combined in the analyses. Of the 149 time 1 interviewees for whom gender was recorded, 107 were interviewed at time 2 (‘stayers’) and 42 were not interviewed at time 2 (‘drop-outs’). Analyses are reported below comparing the stayers and the drop-outs in order to ascertain whether there was evidence of attrition bias. Only eight of the 156 participants interviewed at time 1 reported that a language other than English was spoken at home when they were children. Consequently no analyses were conducted on this demographic variable. Stayers Versus Drop-Outs To determine whether there was any evidence of selective attrition (attrition bias) that might have threatened the validity of the conclusions, a series of analyses was performed comparing the 107 who were interviewed at both times (‘stayers’) with the 42 who were interviewed only at time 1 (‘drop-outs’). The first set of analyses compared the two groups on the demographic variables age, gender and educational level. The mean ages of the two groups were very similar (21.97 years for the stayers and 21.07 years for the drop-outs) and did not differ significantly. There was a tendency for men rather than women to drop out (31 of the 42 drop-outs were men, whereas 59 of the 107 stayers were men). The association was statistically significant at the .05 level on a Fisher Exact Probability Test. The second set of analyses compared the two groups on the time 1 measures on all the psychological variables. There were no statistically significant differences on any of them. It therefore seems reasonable to conclude that the results were not seriously compromised by selective attrition.
190
Active Labour Market Programs
Time 1 — Time 2 Comparisons Of participants interviewed at time 1, 107 could be recontacted at time 2. Of these only 85 (47 men and 38 women) had taken part in WFD, while 22 had not taken part (12 men and 10 women). For most of the participants the second interview took place near the completion of the 6 months spent working for the dole, although for a few, the second interview occurred up to 9 months later. Initially, it seemed like a good idea to treat the 22 nonparticipants as a no-treatment control group; however, it emerged that seven had not taken part because they had obtained jobs and one had taken up full-time study. After excluding them, the remaining 14 were too few to be a useful control group (particularly after dividing them into gender subgroups and recognising that data would be missing for some). No further analyses were conducted on this group. Table 1 shows the educational level and gender of the participants (excluding missing data and low frequency categories). Year 10 is when students reach 15 years, the minimum school-leaving age. There was a significant association between educational level and gender, χ2(2, N = 77) = 6.88, p < .05, with the women having achieved a higher level of education than the men. Unfortunately, this confounding means that any reported gender effects should be treated with caution. They could be due to differences in educational level. Most of the results reported below were based on two-way repeatedmeasures multivariate analyses of variance (MANOVAs) with gender (male/female) and coercion (voluntary/coerced) as between-subject factors, and time (time 1/time 2) as a within-subject factor. The four groups were fairly well balanced with 20 male and 17 female voluntary participants, and 27 male and 21 female coerced participants. These analyses demonstrated whether there were any significant overall differences between groups (e.g., between men and women, or between difference voluntary and coerced participants) as well as any significant overall differences between time 1 and time 2, or any significant interaction effects between time and group. A significant overall difference between time 1 and time 2 could reasonably be attributed to experience of WFD (although the absence of a control
Table 1 Number of Time 2 Participants by Educational Level and Gender Minimum Educational Level Didn’t complete Year 10 Completed Year 10 or 11 Completed Year 12 Total
Men
Women
Total
7 26 8 41
5 14 17 36
12 40 25 77
191
Unemployment and Health ■ Interventions Table 2 Mean Self-Esteem Scores for Voluntary and Coerced Participants (Standard Deviations in Parentheses) Gender
Voluntary/coerced
Time 1
Time 2
Men
Voluntary (n = 20) Coerced (n = 24)
33.35 (4.72) 32.00 (4.67)
34.50 (4.61) 33.13 (4.16)
Women
Voluntary (n = 16) Coerced (n = 18)
31.06 (4.73) 30.50 (4.55)
31.38 (3.83) 29.83 (4.42)
group allows for other possible interpretations). A significant interaction effect would imply that WFD affected the different groups differently. Psychological Measures Self-esteem. Table 2 shows the mean self-esteem scores at time 1 and time 2 for six subgroups. There was a significant effect for gender, F(1, 74) = 7.55, p < .01, with the men showing higher self-esteem than the women. There were no other significant effects. Negative mood. Table 3 shows the mean negative mood scores at time 1 and time 2 for voluntary and coerced participants. There was a significant effect for time, F(1, 78) = 14.20, p < .001, and a significant Time × Gender interaction, F(1, 78) = 11.59, p < .001, arising because the women (but not the men) showed a marked improvement from time 1 to time 2. Psychological distress (GHQ-12). The internal consistency of the GHQ-12 was acceptable at both times. The value of Cronbach’s alpha coefficient was .80 at time 1. Table 4 shows the mean scores for the groups at the two times (the higher the score, the greater the psychological distress). The effect of gender was significant, F(1, 76) = 5.68, p < .05, with the women showing greater psychological distress than the men. There was also a significant effect for time, F(1, 76) = 11.19, p < .001, with an overall decline in distress from time 1 to time 2.
Table 3 Mean Negative Mood Scores for Voluntary and Coerced Participants (Standard Deviations In Parentheses)
192
Gender
Voluntary/coerced
Time 1
Time 2
Men
Voluntary (n = 20) Coerced (n = 25)
12.40 (3.05) 12.48 (2.80)
12.35 (3.23) 12.16 (3.18)
Women
Voluntary (n = 17) Coerced (n = 20)
14.41 (3.78) 14.95 (3.41)
11.94 (3.11) 12.95 (3.36)
Active Labour Market Programs Table 4 Mean Scores on GHQ for Voluntary and Coerced Participants Gender
Voluntary/coerced
Time 1
Time 2
Men
Voluntary (n = 20) Coerced (n = 25)
7.85 (3.67) 10.16 (5.21)
7.05 (2.61) 8.60 (5.09)
Women
Voluntary (n = 16) Coerced (n = 19)
12.31 (5.80) 11.11 (4.52)
9.00 (3.01) 9.68 (3.50)
Social alienation. The reliability (Cronbach’s alpha) coefficients for the Srole scale were .68 at both time 1 and time 2. These are acceptable, although rather low. There were no significant effects of group or time on social alienation. Work involvement. There were no statistically significant effects for work involvement. Attitudes to WFD. This 8-item scale proved to have good internal reliability (Cronbach’s alpha = .81). High scores reflected a favourable attitude to the scheme. Scores could range from a minimum of 8 to a maximum of 32. As the mean scores in Table 5 show, the overall reactions to the scheme were highly favourable at both time 1 and at time 2. There was a significant effect for coercion, F(1, 67) = 6.14, p < .05, with the voluntary participants expressing stronger support for WFD than the coerced participants. There was also a significant effect for time, F(1, 67) = 4.53, p < .05, with an overall decline in expressed support from time 1 to time 2. To shed further light on the Attitude to WFD Scale, an exploratory factor analysis was carried out. The extraction method used was principal component analysis using Varimax rotation with Kaiser normalisation. For each analysis a rotation converged in three iterations. This analysis revealed two factors (components). The eight items are shown in Table 6. The items loading highest on Factor 1 at both times were items 4 and 8 and the items loading highest on Factor 2 at both times were items 1 and 5. It is clear from these loadings that Factor 1 is measuring the usefulness of the scheme (in terms of improving the chances of getting a job) and that Factor 2 is measuring the fairness of the scheme. Table 6 shows the mean responses for the men and women for each item at time 1 and time 2. The two negative items (6 and 7) were reverse scored, so that for all items, the higher the score, the more favourable was the attitude to the scheme. The only items giving rise to statistically significant time effects were item 6 (‘I regard working for the dole as a waste of my time’) and item 8 (‘Working for the dole will improve the chances of unemployed people of getting a regular job’). For item 6, there was a significant Time × Gender interaction, F(1, 78) = 9.21, p < .01. This interaction arose because the women, but not 193
Unemployment and Health ■ Interventions Table 5 Mean Scores on Attitude to Work for Dole Scale for Voluntary and Coerced Participants Gender
Voluntary/coerced
Time 1
Time 2
Men
Voluntary (n = 20) Coerced (n = 22)
28.65 (2.87) 26.64 (4.81)
27.55 (2.56) 26.55 (3.96)
Women
Voluntary (n = 14) Coerced (n = 15)
27.64 (3.03) 25.73 (3.08)
26.93 (2.95) 23.67 (5.56)
the men, showed a significant increase in the belief that working for the dole was a waste of their time. For item 8, there was a significant effect of time, F(1, 78) = 5.35, p < .05. This suggested an overall decline in the belief that working for the dole is likely to improve unemployed people’s chances of getting a regular job. Detailed analyses comparing coerced and voluntary participants did not reveal any significant time effects beyond those reported above, although, not surprisingly, the coerced participants were generally less favourable to the scheme than the voluntary participants.
Table 6 Mean Scores on Attitude to Work for the Dole Questions for Men and Women at Time 1 and Time 2 (Standard Deviations in Parentheses) Men (max. n = 47) Time 1
Time 2
Time 1
Time 2
1. It’s only fair to expect unemployed people to have to work for the dole
3.14 (0.80)
3.05 (0.98)
3.00 (0.75)
3.00 (0.82)
2. Working for the dole is good for unemployed people
3.50 (0.63)
3.64 (0.65)
3.58 (0.50)
3.60 (0.71)
3. Unemployed people should be willing to work for the dole
3.59 (0.62)
3.50 (0.73)
3.55 (0.55)
3.48 (0.69)
4. Working for the dole will improve my chances of getting a job
3.58 (0.75)
3.40 (0.81)
3.44 (0.74)
3.29 (0.81)
5. Unemployed people should have to work for the dole whether they want to or not
3.00 (0.88)
2.84 (1.02)
2.76 (0.93)
2.70 (0.85)
6. I regard working for the dole as a waste of my time (reverse scored)*
3.42 (0.66)
3.51 (0.70)
3.54 (0.75)
3.12 (0.81)
7. I personally resent being asked to work for the dole (reverse scored)
3.47 (0.66)
3.47 (0.59)
3.46 (0.64)
3.33 (0.70)
8. Working for the dole will improve the chances of unemployed people of getting a regular job*
3.49 (0.76)
3.31 (0.76)
3.38 (0.67)
3.08 (0.83)
Note: *Items giving rise to significant effects.
194
Women (max. n = 38)
Active Labour Market Programs
Open questions. Although many of the open-ended answers were positive, others were critical. For example, some participants thought that the work involved in WFD should match their interests. Others felt that WFD should provide useful training that would improve their chances of obtaining employment.
Discussion The objectives/anticipated benefits of WFD included: assisting unemployed people in combating the drift to despondency and despair, fostering appropriate work habits in young people, and improving participants’ selfesteem. The results of the psychological evaluation reported here provide qualified support for these objectives. However, before considering the implications of the results in detail, it is appropriate to acknowledge the limitations of the study. Of the 156 people interviewed at the beginning, gender was available for only 149. Of these, 107 were interviewed on the second occasion, of whom 22 did not participate in WFD. The reduction in sample size could jeopardise the generality of the findings unless it could be shown that the sample was representative of the population of participants and that the 42 drop-outs did not differ systematically from those who stayed in. Fortunately, there were no strong grounds for questioning the representativeness of the sample. Another limitation was the lack of a control group. Ideally, a matched group of young unemployed people who did not participate in WFD would have been interviewed, as well as the participants. The effect of participation would have been demonstrated by a change in the participants but no corresponding change in the nonparticipants. The 22 nonparticipants were not appropriate for this purpose because eight of them had either obtained work or resumed full-time study. Perhaps not surprisingly there was no evidence of any effect of WFD participation on self-esteem. The research literature suggests that even changes in employment status are not necessarily associated with changes in self-esteem (Shamir, 1986a; Winefield et al., 1993) and participation in WFD can scarcely be regarded as a change in employment status. The effect on participants’ attitudes to the scheme also suggests that participation would hardly be likely to enhance self-esteem. For example, after experiencing the program, participants showed increased scepticism that it would improve the chances of unemployed people getting a job, and the women (although not the men) were more likely to agree that WFD was a waste of their time. Despite its lack of impact on self-esteem, WFD had clear benefits on other measures of psychological wellbeing. For example, there were significant improvements in both negative mood and psychological distress/mental health (as measured by the GHQ). This is an important benefit. Because 195
Unemployment and Health ■ Interventions
there is overwhelming evidence that unemployment has negative effects on mental health, any initiative that reverses these effects is to be welcomed. If WFD improves the morale and mental health of unemployed people, as has been shown in this pilot study, this is a tangible benefit, even if it fails to improve self-esteem or to help people find jobs. Winefield, Tiggemann and Winefield (1992), in their longitudinal study of school-leavers, found that the psychological wellbeing of the unemployed and the ‘dissatisfied’ employed was associated with how they spent their spare time. Specifically, wellbeing was negatively correlated with amount of time spent watching television or doing nothing in particular, but positively correlated with amount of time spent in activities with other people. Given that WFD involved activity with other people, it is not surprising that it should have benefited wellbeing. Although WFD appears to have some ongoing benefits to the psychological wellbeing of participants, these benefits do not involve any changes in self-esteem. It also remains to be seen whether the benefits are lasting. A further wave of data collection after all participants had completed WFD would have been necessary in order to see whether people who had returned to unemployment had benefited from participating. It also needs to be recognised that in the absence of a no-treatment control group (a matched group of unemployed people who were not asked to participate in WFD) the observed benefits cannot be unequivocally attributed to WFD. The attitudes of participants to WFD were not altogether surprising. For example, the volunteers were significantly more favourable to the scheme than the coerced participants. Although there was an overall decline in expressed support for the scheme from Time 1 to Time 2, this decline was due to a decline in the perceived usefulness of WFD in helping young unemployed people get jobs, rather than to a decline in the perceived fairness of the scheme. This suggests that the participants generally supported the notion of mutual obligation. Finally, it had been hoped to evaluate the effectiveness of WFD in improving work habits and work attitudes and the questionnaire included several relevant scales. Unfortunately no useful data were gathered because very few participants, and even fewer employers, responded to the questions. Perhaps the wider implication of the findings of this study for active labour market programs designed for young unemployed people is that, unless they succeed in helping participants master a skill relevant to obtaining a worthwhile job, they are unlikely to enhance self-esteem and thus have long-term psychological (or long-term employment) benefits. Programs that provide young people with an opportunity to engage with others in a purposeful activity appear to have continuing psychological benefits, but whether these benefits are maintained after participation in the program has concluded remains to be seen. Our results confirm those 196
Active Labour Market Programs
of recent studies conducted in Australia and the United States that have shown limited benefits of active labour market programs on the psychological wellbeing of the long-term unemployed (Carson et al., 2003; Michalopoulos et al., 2000).
References Banks, M.H., & Jackson, P.R. (1982). Unemployment and risk of minor psychiatric disorder in young people. Psychological Medicine, 12, 789–798. Banks, M.H., Clegg, C.W., Jackson, P.R., Kemp, N.J., Stafford, E.M., & Wall, T.D. (1980). The use of The General Health Questionnaire as an indicator of mental health in occupational studies. Journal of Occupational Psychology, 53, 187–194. Bottomley, D., McKay, S., & Walker, R. (1997). Unemployment and jobseeking. Department of Social Security Research Report No. 62. London: Department of Social Security. Carson, E., Winefield, A.H., Waters, L., & Kerr, L. (2003). Work for the dole: A pathway to self-esteem and employment commitment, or the road to frustration? Youth Studies Australia, 22, 19–26. Campbell, M., Sanderson, I., & Walton, F. (1998). Local responses to long-term unemployment. York, UK: Rowntree Foundation. Cook, T.D., & Campbell, D.T. (1979). Quasi-experimentation. Chicago: Rand McNally. Department of Education, Employment and Training and Youth Affairs (DEETYA). (1996). Reforming employment assistance: Helping Australians into real jobs. Canberra, Australia: AGPS. Department of Education, Employment and Training and Youth Affairs (DEETYA). (1997). Net impact labour market programs. Evaluation and Monitoring Branch EMB Report 2/97. Canberra, Australia: Author. Dodder, R.A., & Astle, D.J. (1980). A methodological analysis of Srole’s nine-item anomia scale. Multivariate Behavioral Research, 15, 329–334. Dooley, D. (1995). Social research methods. Englewood Cliffs, NJ: Prentice-Hall. Friedlander, D., Greenberg, D.H., & Robins, P.K. (1997) Evaluating government training programs for the economically disadvantaged. Journal of Economic Literature, 34, 1809–1855. Goldberg, D.P. (1972). The detection of psychiatric illness by questionnaire. London: Oxford University Press. Gurney, R.M. (1980). Does unemployment affect the self-esteem of school leavers? Australian Journal of Psychology, 32, 175–182. Hales, J., Collins, D., Hasluck, C., & Woodland, S. (2000). New deals for young people and long-term unemployed: Employer survey. Sheffield, UK: National Centre for Social Research. Hasluck, C. (2000). Early lessons from the evaluation of the New Deal programs. Research Report 49, Employment Service Research and Development Branch. Sheffield, UK: Employment Service. Martin, J.P. (1998). What works among active labour market policies: Evidence from OECD countries’ experiences. Paris: OECD. Meân Patterson, L.J. (1997). Long-term unemployment amongst adolescents. Journal of Adolescence, 20, 261–280. Meyer, B. (1995). Lessons from the US unemployment insurance experiments. Journal of Economic Literature 33, 91–131.
197
Unemployment and Health ■ Interventions Michalopoulous, C., Schwartz, C., & Adams-Ciardullo, D. (2000). What works best for whom: Impacts of 20 welfare-to-work programs by subgroup. New York: Manpower Demonstration Research Corporation. Mirels, H.L., & Garret, J.B. (1971). The Protestant ethic as a personality variable. Journal of Consulting and Clinical Psychology, 36, 40–44. Orr, L., Bloom, H.S., Bell, S.H., Lin, W., Cave, G., & Doolittle, F. (1994). The national JPTA study: Overview, impacts benefits and costs of Title IIA. Bethesda, MD: Abt Associates. Patton, W., & Noller, P. (1984). Unemployment and youth: A longitudinal study. Australian Journal of Psychology, 36, 399–413. Prause, J., & Dooley, D. (1997). Effects of underemployment on school leavers’ selfesteem. Journal of Adolescence, 20, 243–260. Quint, J.C., Polit, D.F., Bos, H., & Cave, G. (1994). New chance: Interim findings on a comprehensive program for disadvantaged young mothers and their children. San Francisco: Manpower Development Research Corporation. Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press. Shamir, B. (1986a). Self-esteem and the psychological impact of unemployment. Social Psychology Quarterly, 49, 61–72. Shamir, B. (1986b). Protestant work ethic, work involvement and the psychological impact of unemployment. Journal of Occupational Behaviour, 7, 25–38. Srole, L. (1956). Social integration and certain corollaries: An explanatory model. American Sociological Review, 21, 709–716. Tiggemann, M., & Winefield, A.H. (1984). The effects of unemployment on the mood, self-esteem, locus of control, and depressive affect of school-leavers. Journal of Occupational Psychology, 57, 33–42. Warr, P.B., Cook, J., & Wall, T.D. (1979). Scales for the measurement of some work attitudes and aspects of psychological well-being. Journal of Occupational Psychology, 52, 129–148. Winefield, A.H. (1999). Measuring the impact of the Work for the Dole pilot projects on work ethic and self-esteem. Report commissioned by DEETYA, Canberra, Australia. Unpublished. Winefield, A.H., Tiggemann, M., & Winefield, H.R. (1992). Spare time use and psychological well-being in employed and unemployed young people. Journal of Occupational and Organizational Psychology, 65, 307–313. Winefield, A.H., Tiggemann, M., Winefield, H.R., & Goldney, R.D. (1993). Growing up with unemployment: A longitudinal study of its psychological impact. London: Routledge. Winefield, A.H., Winefield, H.R., Tiggemann, M., & Goldney, R.D. (1991). A longitudinal study of the psychological effects of unemployment and unsatisfactory employment on young adults. Journal of Applied Psychology, 76, 424–431. Winefield, H.R., Goldney, R.D., Winefield, A.H., & Tiggemann, M. (1989). The General Health Questionnaire: Reliability and validity for Australian youth. Australian and New Zealand Journal of Psychiatry, 23, 53–58.
198
Chapter
16
Explaining the Negative Relationship Between Length of Unemployment and the Willingness to Undertake a Job Training: A Self-Determination Perspective Maarten Vansteenkiste, Hans De Witte and Willy Lens
Within the unemployment literature, several studies point to a correlation between length of unemployment and individuals’ perceptions of financial strain and economic deprivation (e.g., Rowley & Feather, 1987; Warr & Jackson, 1984). The results of longitudinal studies (Brenner & Levi, 1987; Verkleij, 1988; Warr & Jackson, 1987) also suggest that individuals experience a reduction in wellbeing upon first becoming unemployed, but that their wellbeing stabilises at a lower level after a certain period of time. De Witte (1993) has argued that this stabilisation occurs partly because over time individuals adjust to their role as unemployed persons, and partly because, as the length of unemployment increases, they tend to withdraw both psychologically and behaviourally from the labour market. According to De Witte (1993), this withdrawal process might be reflected in several ways, for example, through a reduction of employment commitment or the overall importance people place on finding a job (Warr & Jackson, 1985, 1987), a decline in job search frequency (Miltenburg & Woldringh, 1989; Muffels, Rietveld, & Vriens, 1990; Van Hoye, Janssens, & Peeters, 1988; Vleminckx, 1982), and a decrease in optimism about finding a job (De Goede, 1988; De Valck, 1981; De Witte, 1992; Kroft, Engbersen, Schuyt, & Van Waarden, 1989; Van Hoye et al., 1988). Herein, we argue that a reduction in unemployed people’s willingness to follow a job training (i.e., the motivation to follow an educational program to be trained for a specific job) constitutes another expression of this withdrawal from 199
Unemployment and Health ■ Interventions
the labour market, an outcome that, to the best of our knowledge, has received little attention within the unemployment literature. In the present contribution, we aim to explore in more detail the motivational processes that might underlie this type of behavioural withdrawal from the labour market during long-term unemployment. Using self-determination theory (SDT; Deci & Ryan, 1985, 2000, 2002) as our guiding theoretical framework, we claim that both quantitative (i.e., amotivation) and qualitative (i.e., autonomous motivation) motivational processes (Vansteenkiste, Lens, De Witte, & Feather, 2005) mediate the relationship between length of unemployment and this outcome. Obtaining more insight into the motivational processes that unfold over length of unemployment time is instructive, because it would help us to indicate which variables might best be targeted by intervention programs (e.g., unemployment counselling services), in order to help unemployed people better cope with the negative effects associated with being long-term unemployed. Although SDT has received considerable support in such domains as health behaviour (Williams, Grow, Freedman, Ryan, & Deci, 1996), religion (Ryan, Rigby, & King, 1993), sport (Pelletier, Fortier, Vallerand, & Brière, 2001), exercise (Vansteenkiste, Simons, Soenens, & Lens, 2004) and education (e.g., Ryan & Connell, 1989; see Deci & Ryan, 2000, 2002; Vallerand, 1997, for an overview), only a few studies have applied this theory to the unemployment domain (Vansteenkiste, Lens, Dewitte, De Witte, & Deci, 2004). Therefore, we briefly discuss the most important elements of this theory in the following section and indicate how it can be applied to the study of unemployment.
Self-Determination Theory Autonomous Versus Controlled Motivation Central to SDT is the distinction between autonomous and controlled behavioural regulation. These reflect two different categories or reasons for engaging in a particular behaviour (e.g., job searching). Autonomy involves a sense of volition and full endorsement of the reasons for one’s actions. Because job searching reflects a personal choice for unemployed people who are autonomously motivated, the perceived locus of causality for their actions (Heider, 1958; deCharms, 1968) is internal. Autonomous regulation contains two different motivational subtypes, namely intrinsic motivation and identified regulation (Ryan & Deci, 2000). When intrinsically regulated, people engage in an activity out of curiosity and for its inherent task pleasure (Deci, 1975). Thus, an unemployed person may derive some enjoyment and pleasure from exploring the job market and trying to find a job that meets his or her interests. In such case, job hunting may be experienced as inherently satisfying. Intrinsic motivation stands in contrast to extrinsic motivation, which links performance of an activity to attainment of an outcome that is separated from 200
Duration of Unemployment, Job Training and Self-Determination Theory
the activity itself. But according to SDT, the value and regulation of such extrinsically motivated behaviour can be internalised to a greater or lesser degree (Schafer, 1968); the more that internalisation has occurred, the more it provides the basis for an autonomous type of extrinsic motivation, labelled identification. The process of internalisation refers to incorporating the value of, or the reason for, the particular behaviour (i.e., job searching) into one’s own value system, so that people experience a personal ownership over their behaviour. When unemployed people have internalised the regulation of their job search behaviour, they will identify with the personal importance of job searching and work for their self-selected life goals. In the case of identification, an unemployed person searches for a job because they perceive work as personally meaningful for the development of their talents and personality and for the pursuit of other important life goals. In contrast to autonomous regulation of behaviour, controlled regulation entails engaging in an activity because of a sense of pressure to do so. When controlled, people execute the behaviour in order to comply either with the demands of an external agent (i.e., external regulation), or with an internal demand that is linked to threats of self-imposed guilt or loss of selfesteem (i.e., introjected regulation). An externally regulated job searcher will look for a job because they have financial troubles or are pressured to do so by their parents or partner. Searching for a job because one is ashamed of being jobless is an example of introjected regulation. In both examples of controlled regulation, individuals feel compelled to perform the behaviour, and their actions are characterised by an external perceived locus of causality (deCharms, 1968; Ryan & Deci, 2000). Just as with identification, both types of controlled regulation represent a type of extrinsic motivation. However, in the case of identification, the internalisation process has almost been completed. With introjection, on the other hand, the requested demand or value has been only partially internalised, while in the case of external regulation, it has not been internalised at all (Losier & Koestner, 1999; Koestner, Losier, Vallerand, & Carducci, 1996). A schematic overview of these different self-regulatory styles for job searching can be found in Figure 1. It is assumed that the processes of intrinsic motivation and internalisation can be either enhanced or undermined depending on whether the environment nourishes or impedes satisfaction of basic psychological needs for autonomy (i.e., the desire to feel personal ownership over one’s actions), competence (i.e., the desire to feel effective in what one does and mastering new skills in the process) and relatedness (i.e., the desire to develop satisfying and deeply anchored relationships with others; Deci & Ryan, 2000; Ryan, 1995). In the present study, particular attention was paid to the need for competence (Bandura, 1989; Deci, 1975; White, 1959). The concept of perceived competence has also received considerable attention within other 201
Unemployment and Health ■ Interventions Intentional Motivation
_______________ ____________________________________________________________ Type of
Amotivation
Motivation
External
Introjected
Identified
Intrinsic
Motivation
Motivation
Motivation
Motivation
______________ ______________________________________________________________ Motivational
Discouragement
Force
Helplessness
Expectations Rewards Punishments
Guilt, Shame, Anxiety, Internal
Impersonal
External
Enjoyment,
Commitment
Pleasure,
Compulsion
______________ __________________________ Locus of
Personal Values,
Interest ________________________________
External
Internal
Internal
Causality _______________ __________________________
_________________________________
Controlled Motivations
Autonomous Motivations
________________________________________________ Extrinsic Motivations
Least
------------------------------------------------------------------------------------------------------> Most
Self-Determined
Self-Determined
Figure 1 Schematic relation of five types of motivation according to SDT.
motivational theories, such as self-efficacy theory (Eden & Aviram, 1993; Bandura, 1977, 1989) and expectancy-value theory (Feather, 1990, 1992; Feather & Davenport, 1981), and both have been successfully applied to the unemployment domain on several occasions (but see Deci & Ryan, 2000, for conceptual differences between these competence constructs). SDT claims that perceived competence represents a basic psychological need that, if satisfied, promotes a better quality of motivation, that is, a more autonomous regulation. Research in several domains has confirmed this claim. Experimental studies have shown that negative feedback lowers people’s intrinsic interest in the activity, and that this is because it reduces people’s perceived competence (Elliot et al., 2000; Guay, Boggiano, & Vallerand, 2001; Harackiewicz & Larson, 1986; Reeve & Deci, 1996; Vallerand & Reid, 1984, 1988; Vansteenkiste & Deci, 2003). Several field studies (Vallerand, Fortier, & Guay, 1997; Williams, Wiener, Markakis, Reeve, & Deci, 1994) have similarly indicated that the use of autonomy202
Duration of Unemployment, Job Training and Self-Determination Theory
supportive techniques by parents and teachers increases autonomous motivation by increasing children’s self-perceived competence. Together these studies suggest that perceived competence may mediate the relationship between negative/positive feedback about task performance on the one hand, and the extent to which individuals find tasks enjoyable and/or meaningful on the other. When applied to the unemployment context, they imply that the long-term unemployed may find the job searching process less personally rewarding (intrinsic regulation), and consider work less personally important (identified regulation) than their short-term unemployed counterparts, because they have experienced more unsuccessful job applications (negative feedback) over time, and, as a result, perceive themselves to be less capable of finding a job (reduced perceived competence). Such a view is consistent with the view that the long-term unemployed make internal attributions for their lack of success (Feather & Davenport, 1981; Weiner, 1985). In addition, such a claim is compatible with De Witte’s (1993) observation that one’s pessimism about finding a job increases with length of unemployment, as people become discouraged and demotivated following repeated rejections during application interviews.
Amotivation While autonomous and controlled motivation reflect the quality of job searchers’ motivation, the level of motivation is assessed through the concept of amotivation within SDT. Indeed, both autonomous and controlled types of motivation imply intention and motivation to act, and therefore stand in contrast to amotivation, which reflects a lack of either intention or motivation. Amotivated people tend to feel helpless and futile and they feel unable to change the course of events. They either do not act at all or they act passively — that is, they go through motions of acting without a real intention to act. Amotivation is thought to arise from feelings of powerlessness, which in turn may arise from low self-efficacy beliefs (Bandura, 1977, 1989), low levels of perceived competence (Deci, 1975), and/or a lack of perceived contingency between behaviour and outcomes (e.g., Seligman, 1975). In the case of long-term unemployment, amotivation may be characterised by low levels of job searching, together with feelings of hopelessness about obtaining a job (Ryan & Deci, 2000). In the present study, we predicted that length of unemployment would positively predict amotivation and that perceived competence would mediate this relationship. In other words, long-term unemployed people would feel more discouraged and amotivated towards their job searching — because they perceive themselves to be less effective in finding a job than would short-term unemployed individuals.
203
204
Predicted pattern of relationships.
Figure 2
Duration of Unemployment
-
Perceived Competence
-
+
Amotivation
Autonomous Motivation
-
+ Willingness to Follow a Job Training
Unemployment and Health ■ Interventions
Duration of Unemployment, Job Training and Self-Determination Theory
Willingness to Undertake Job Training As stated earlier, the ultimate aim of the present study was to examine the possible role of motivational processes in explaining the relationship between length of unemployment and psychological withdrawal from the labour market, as characterised by a lack of willingness to undertake job training. The proposed model linking these variables is shown in Figure 2. Because autonomously motivated unemployed people may perceive work as an avenue for self-development, they might consider following a job training as a first step in this process. In the present study therefore, autonomously motivated individuals were expected to be more willing to follow a job training than were individuals who showed controlled motivation. We did not expect such a relationship for controlled job search regulation. Controlled motivation is unlikely to provide people with the necessary energy to engage and persist in a job training, an activity that requires considerable effort. In fact, people have a tendency to resist activities that they feel pressured to pursue. Amotivation was expected to negatively predict people’s willingness to follow a job training. Amotivated unemployed people do not believe that engaging in job search activities might help them to realise their desired goal (of finding a job), because they expect not to have any control over the course of events. In short, the goal of the present study was to examine, on the basis of SDT, whether both a decrease in unemployed people’s level of motivation (i.e., amotivation) but also in the quality of motivation (i.e., autonomous motivation) over length of unemployment time due to a decrease in perceived job search competence could account for a possible decline in the willingness to follow a job training. We first aimed to test each of the separate links displayed in Figure 2 using correlational and regression analyses, and then used SEM to test the full chain of relationships. In conducting the regression analyses, we controlled for unemployed people’s levels of employment commitment (Feather & Davenport, 1981), which is found to be an important predictor studied within the training domain (Mathieu, Tannebaum, & Salas, 1992; Noe, 1986; Sheldon, Turban, Brown, Barrick, & Judge, 2003; Tharenou, 2001) and the unemployment domain (Feather, 1990, 1992).
Method Participants and Procedure The survey was conducted during the period January to May, 2001. A questionnaire package was distributed to unemployed people by government employees who regularly meet with them, and was administered either individually or during group information sessions. Of the 350 questionnaires that were distributed, 273 (78%) were returned. Missing data led to the deletion of a further 39 cases, leaving a pool of 234 participants (67%) for data analyses. 205
Unemployment and Health ■ Interventions
Demographic information indicated the following: 61% of the sample were female; 38% were younger than 26 years old, 42% were between 26 and 40 years old and 19% were older than 41 years old; 42% had been unemployed for less than 6 months, 41% had been unemployed for a period between 6 months and 2 years, and 17% had been unemployed for more than 2 years; 11% of the unemployed people were very low-educated (primary education), 22% were low-educated (Grade 9), 47% were moderately educated (Grade 12) and 20% had enjoyed higher education; 18% lived alone, 12% lived together with her or his children, 21% lived together with their partner, 22% lived together with partner and children and 26% lived together with their parents. Regarding level of unemployment benefits, 16% received no benefits, 29% received less than £375 a month, 21% received between £376 and £625 a month, 32% received between £626 and £875 a month, and 2% received more than £875 a month.
Measures Demographics Participants provided information about their age, gender, education level, unemployment benefit level and the length of time that they had been unemployed. For age, people were asked to tick one of four categories: < 26 years, 26–40 years, 41–55 years and > 55 years. For education level, participants placed themselves in one of 12 categories, 1 = No primary education to 12 = University studies. Monthly unemployment benefit was assessed using 8 categories, 1 = no income, 8 = > £775 a month). For length of unemployment they indicated their agreement with one of the following 8 response categories (0–3 months, 3–6 months, 6–9 months, 9–12 months, 12–24 months, 2–5 years, 5–10 years and > 10 years). In describing the participants at the beginning of the methods section, we collapsed across categories to make it simpler, but when we examined the predicted relations for length of unemployment we used a continuous variable from 1 to 8 corresponding to the 8 successive categories, and we also entered the other background variables as continuous variables in the regression to control for them. Motivation to Undertake Job Training Participants recorded their level of agreement with two items that assessed their motivation and willingness to undertake job training (e.g., ‘I am willing to accept a job for which I need to complete job training of six months’). Response categories ranged from 1 = I do not agree at all to 5 = I completely agree. Internal consistency of this two-item scale was .81. Perceived Competence Participants indicated on a 5-point Likert scale (1 = I do not agree at all to 5 = I completely agree) the extent to which they perceived themselves as competent in looking for or finding a job (e.g., ‘I am good at searching for 206
Duration of Unemployment, Job Training and Self-Determination Theory
a job’; ‘I don’t know how I can possibly find a job’ [reverse scored]). Internal consistency of the 3-item scale was satisfactory (.60) Self-Regulation Questionnaire — Job Searching (SRQ-JS) The development and the initial validation of this questionnaire has been described elsewhere (Vansteenkiste, Lens, et al., 2004). In response to the question ‘Why are you looking for a job?’ participants were asked to indicate their agreement (1 = Completely disagree and 5 = Completely agree) with each of 16 statements that reflected either autonomous or controlled job search regulation (8 items in each case). Of the autonomous regulation items, 4 reflected intrinsic regulation (e.g., ‘I am looking for a job because I find it enjoyable to explore the labour market’), while 4 indicated identified regulation (e.g., ‘I am looking for a job because I consider work as personally important to me’). The 8 items were treated as a single scale (alpha = .85). Alpha for the 8 controlled regulation items was .76. Examples included ‘I am looking for a job because I have financial problems (external regulation) and ‘I am looking for a job because I do not want others to think I am a lazy person’ (introjected job search regulation). Finally, amotivation was assessed with four items (alpha = .80; e.g., ‘I feel I can’t change anything about my unemployment situation’). Employment Commitment Fourteen items assessed the degree to which people were committed to having a job. Participants used a 5-point Likert scale (1 = Completely disagree, 5 = Completely agree) on this widely used questionnaire (e.g., Feather, 1990, 1992; Stafford, Jackson, & Banks, 1980). Example items are ‘It’s better to accept any kind of job than to remain jobless’ and ‘One can only enjoy one’s free time when one has worked for it’. Internal consistency for this scale was .84.
Analyses Simple correlations, multiple hierarchical regression analysis and structural equation modelling were used to test hypotheses. In the regressions, we entered demographic variables (age, gender, length of unemployment, level of education and level of unemployment income) in a first step, allowing us to determine whether length of unemployment would negatively predict the motivation to undertake job training, after other demographic characteristics were controlled. In a second step, psychological variables that significantly correlated with the willingness to follow a job training as an outcome variable were entered. In a final step, employment commitment was entered as an additional predictor to examine whether the obtained effects in step 2 would hold. The proposed model depicted in Figure 2 was tested using LISREL 8 software (Joreksog & Sorbom, 1993). To assess the fit of the model to the observed data, we employed the chi-square statistic, the goodness-of-fit 207
Unemployment and Health ■ Interventions Table 1 Correlations Among Demographic Variables and Subscales 1 1 Length of unemployment 2 Age
—
2
3
.29** –.14* —
3 Level of education 4 Level of unemployment benefits 5 Gender (M = 0; F = 1) 6 Perceived competence 7 Autonomous job search motivation 8 Controlled job search motivation 9 Amotivation 10 Motivation to follow a job training
4
5
–.26**
.08
–.20** –.09 —
6
7
8
–.34** –.18** –.01
.12* –.13*
9
10
.19** –.14*
–.01
–.09
.13*
.28** –.01
–.05
–.26**
–.08
–.12
.05
.09
.07
–.05
.08
–.06
—
–.08
.12
—
–.05 —
.07 .16**
.24** –.21** –.52** —
.24** –.03 —
.26** —
–.10 .19**
.21** .27** .11 –.13* —
index (GFI; Joreskog & Sorbom, 1993), the comparative fit index (CFI; Bentler, 1990), the normative fit index (NFI; Tanaka, 1987), and the root mean square of approximation (RMSEA; Steiger, 1990). A nonsignificant chi-square indicates a good-fitting model (Bollen, 1989). Fit indices above .90 are considered acceptable, as is an RMSEA of 0.05 to 0.08, while an RMSEA of less than 0.05 represents an excellent fit (Browne & Cudeck, 1993). To assess the strength of the relationship between any two individual variables, we calculated maximum likelihood parameter estimates. Maximum-likelihood estimators are the most widely used method for obtaining parameter estimates in LISREL path analyses or structural equation modeling (Bollen, 1989).
Results Preliminary Analyses Correlations As predicted, length of unemployment was negatively correlated with motivation to undertake job training (see Table 1), while educational level was positively correlated with the motivation to undertake a job training. Concerning the motivational variables, length of unemployment was negatively related to perceived competence and autonomous job search motivation, positively correlated with amotivation and unrelated to controlled motivation. Perceived competence was positively correlated with autonomous motivation and negatively related with amotivation. Autonomous motivation positively predicted the motivation to follow a job training, whereas amotivation was a negative predictor. Finally, as 208
Duration of Unemployment, Job Training and Self-Determination Theory
expected, controlled motivation was uncorrelated with the motivation to follow a job training. Then, motivation to undertake job training was regressed onto the five demographic variables, in order to explore whether duration of unemployment would account for unique variance. As predicted, this was the case: length of unemployment time was negatively related to the motivation to undertake job training (β = –.14, p < .05), while none of the other demographic variables were significant predictors. The demographic variables explained 5% of the variance in the motivation to undertake job training. In the second step of the hierarchical regression, the three motivational constructs of SDT were entered in the analyses in order to explore whether they could account for unique variance in the outcomes. This resulted in a significant increase in explained outcome variance in the willingness to undertake job training (∆r2 = .10, p < .01). As predicted, autonomous but not controlled job search motivation was significantly positively related to the willingness to undertake a job training (β = .32, p < .001). Furthermore, amotivation negatively predicted the willingness to follow a job training (β = –.12, p = .05). Finally, in a third step of the regression, employment commitment was added in order to explore whether the effects of the SDT-variables on the outcomes were an artifact. The relationships that reached significance in the previous step of the regression remained unchanged after entering employment commitment. LISREL Analyses We used structural equation modelling to test the full hypothetical mediating model, as depicted in Figure 2. Background variables other than length of unemployment were discarded from this analysis, because we were primarily interested in exploring the relation between length of unemployment and the willingness to follow a job training, and because regression analyses revealed that length of unemployment explained unique outcome variance. Several fit indices indicated an excellent fit: the GFI, the NFI and the CFI were all higher than .90 (.99, .94 and .97 respectively), the RMSEA was .057 and the chi-square statistic was nonsignificant, χ2(7, N = 234) = 8.79, ns. Each hypothesised relationship was supported by a significant parameter estimate. For instance, perceived competence positively predicted autonomous job search motivation (parameter estimate = .24, p < .05). There was one relationship that reached only marginal significance: amotivation marginally negatively predicted people’s willingness to undertake job training (parameter estimate = –.12, p = .06). The full model is depicted in Figure 3.
209
210
-.34**
-.52**
Perceived Competence
Amotivation
Note: + p < .10, * p < .05, ** p < .01.
Results of the structural equation model: χ2(7, N = 234) = 8.79, p > .05; CFI = .97; GFI = .99; NFI = .94, RMSEA = .057.
Figure 3
Duration of Unemployment
.24*
Autonomous Motivation
-.12+
.27**
Willingness to Follow a Job Training
Unemployment and Health ■ Interventions
Duration of Unemployment, Job Training and Self-Determination Theory
Discussion Various unemployment researchers (e.g., De Goede, 1988; De Witte, 1993; Feather, 1990) have suggested that unemployed people tend to withdraw themselves from the labour market over length of unemployment time. In the present contribution, we suggested that a reduced willingness to follow a job training constitutes one manifestation of this withdrawal process. In line with this reasoning, we found that length of unemployment time and the willingness to follow a job training were indeed significantly (but rather weakly) negatively correlated. A second goal of this contribution was to provide more detailed insight into the motivational processes underlying this relationship, using selfdetermination theory (Deci & Ryan, 2000, 2002) as our guiding theoretical framework. The chain of relationships studied is discussed step by step in the following. First, we expected that the longer people remained unemployed the more the quality (i.e., autonomous motivation) as well as the the quantity (i.e., amotivation) of their job search motivation would decline. These predictions were confirmed, suggesting that a double motivational process unfolds over length of unemployment time: first, people start to feel they can not change much about their unemployment situation, their feelings of helplessness increase and, second, their free and volitional engagement in job searching decreases, because job searching is experienced as a less enjoyable and challenging activity (i.e., intrinsic regulation) and because the personal importance of work is reduced (i.e., identified regulation). Second, this raises the question of how these changes in both the quality and the quantity of motivation occur. We reasoned that long-term unemployed people are likely to have more often received negative feedback during more frequent application interviews, so that they feel less effective or competent in their job search which would in turn explain the changes in their motivation. In fact, within SDT, it is claimed that the satisfaction of the need for competence helps unemployed people to (a) maintain their autonomous job search motivation and (b) to buffer against displaying amotivation towards job searching. The results confirmed this hypothesis and are in line with previous laboratory (e.g., Reeve & Deci, 1996; Vallerand & Reid, 1984; Vansteenkiste & Deci, 2003) and field studies (e.g., Williams et al., 1994). The third question concerns the consequences associated with these changes in the quantity and quality of motivation over length of unemployment time. Vansteenkiste, Lens, et al. (2004) reported that searching for a job in an autonomous manner positively predicted self-actualisation and past job search behaviour, while amotivation negatively predicted a variety of wellbeing outcomes. The present contribution adds to this work by showing that maintaining an autonomous job search motivation in the face of failure helps people to remain willing to undertake job training. In 211
Unemployment and Health ■ Interventions
contrast, when people feel discouraged and helpless, their willingness to follow a job training is considerably decreased. Importantly, controlled motivation, that is, being pressured to search for a job, did not predict the motivation to follow a job training. Such a finding is consistent with SDT’s claim that unemployed people are more resistant to engage in job search activities they feel pressured to take part in. Thus, enhancing the pressure on unemployed people to search for a job might not be the most efficient route to effective employment. When the magnitude of the significant relationships is considered, it can be observed that the effect of autonomous motivation was twice as large than the effect of amotivation. This suggests that when we want to understand why unemployed people are less willing to undertake a job training over length of unemployment time, it is more important to consider the reason for their job search motivation (autonomous vs. controlled) than their level of motivation (amotivation).
Limitations Some limitations of this research warrant discussion. Because the findings of this study rely on self-report data, future studies with multiple methods would be useful. In particular, it would be important to have measures of people’s enrolment in job training courses and of the number of rejections after application interviews. Gathering data concerning the latter would be particularly important to explore whether the decrease in perceived competence over length of time would be function of the number of times unemployed people have been rejected during interviews, as assumed in the present study. Further, the data of the present study are cross-sectional, so conclusions cannot confidently be drawn regarding causal relations among variables. Longitudinal studies that follow people over the duration of their unemployment to ascertain how length of time affects both people’s feelings of amotivation and their autonomous job search motivation are needed to confirm the patterns of relationships identified by this study. Indeed, because length of time is a process variable, people’s level and quality of motivation might be a consequence, but also an antecedent, of their unemployment duration. Finally, we primarily focused on the relation between length of unemployment time and the need for competence, but future research might also examine whether people’s need for autonomy and relatedness are also undermined over length of unemployment time. It might well be that people feel less supported by others in their job search over length of unemployment time (relatedness), and that they experience their job searching more as a task that should be accomplished (autonomy) because of the increasing number of financial problems pushing people to search over length of unemployment time.
212
Duration of Unemployment, Job Training and Self-Determination Theory
Practical Relevance and Future Research On a practical level, those results suggest that employment counselling services might try to build the self-confidence and self-efficacy (Bandura, 1977) of unemployed people in their ability to persist in effective job-seeking behaviour. Past research (Caplan, Vinokur, Price, & van Ryn, 1989; Price, van Ryn, & Vinokur, 1992; Vinokur, Price, & Caplan, 1991; Vinokur & Schul, 1997; Vinokur, van Ryn, Gramlich, & Price, 1991) has demonstrated that when people are involved in an intervention program that focuses in part on enhancing people’s confidence in finding a job, these unemployed people obtained higher quality reemployment in terms of earnings and job satisfaction, had greater job-seeking self-efficacy if they were still unemployed and reported lower levels of depressive symptomatology than did their counterparts in the control group. Eden and Aviram (1993) found that a self-efficacy training program enhanced the intention to put effort in search for a job, and increased reemployment among participants who had low job search selfefficacy scores prior to the training program. Although the present research is cross-sectional in nature, we suggest that the present findings might add to this work in three ways. First, an intervention program that focuses on people’s feelings of competence and self-efficacy might increase the possibility that people accept the opportunity to undertake a job training, an outcome that has not been explored yet within such field intervention research. Second, previous intervention research did not examine the motivational processes that might explain why an intervention focused on people’s self-confidence might be so beneficial. The present results indicate that people’s levels of motivation and the quality of their job search motivation might play herein a mediating role. Third, as found by Eden and Aviram (1993), such intervention programs might be particularly useful for long-term unemployed people because they report the greatest lack in job search competence. Thus, it is suggested that the beneficial role of being involved in competence-enhancing intervention programs might be moderated by people’s length of unemployment time.
Conclusion Feather (1990, 1992) suggested that unemployment researchers should try to apply midrange theories to the unemployment domain because such theories might help to illuminate the underlying processes that are associated with becoming and remaining unemployed. In line with this suggestion, we tried to shed light on the processes that can account for a decrease in people’s motivation to undertake job training over length of time, using self-determination theory as our theoretical framework. It appears that this decline is not only function of an increase in people’s lack of motivation to search (amotivation), but also a decrease in people’s volitional job search engagement (autonomous motivation), and that perceived competence plays a mediating role herein. The present contribution reveals one way 213
Unemployment and Health ■ Interventions
that SDT can be meaningfully applied to the unemployment domain, and we would encourage other researchers to examine other aspects of this motivational theory to further understand unemployed people’s experiences and job search behaviour.
Acknowledgments The first author’s contribution was supported by a grant of the fund of Scientific Research, Flanders, Belgium.
References Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84, 191–215. Bandura, A. (1989). Human agency in social cognitive theory. American Psychologist, 44, 1175–1184. Bentler, P. (1990). Comparative fit indices in structural models. Psychological Bulletin, 107, 238–246. Bollen, K.A. (Ed.). (1989). Structural equations with latent variables. New York: Wiley. Brenner, S., & Levi, L. (1987). Vulnerability among long-term unemployed: A longitudinal study of mental health and physical health among Swedish women at different phases of unemployment. In P.S. Schwefel &. H.Z.D. Schwefel (Ed.), Unemployment, social vulnaribility and health in Europe (pp. 239–254). London: Springer. Browne, M.W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K.A. Bollen & J.S. Long (Eds.), Testing structural equation models (pp. 136–162). Newbury Park, CA: Sage. Caplan, R.D., Vinkour, A.D., Price, R.H., & van Ryn, M. (1989). Job seeking, reemployment, and mental health: A randomized field experiment in coping with job loss. Journal of Applied Psychology, 74, 759–769. De Goede, M. (1988). Beleving van niet-werken: Een onderzoek onder werklosen, arbeidsongeschikten en hun partner [The experience of not working. An investigation among unemployed people, disabled and their partners]. Utrecht, Germany: Rijksuniversiteit Utrecht. De Valck, N. (1981). Een enquête over de beleving en de gevolgen van de werkloosheid: Verwachtingen die werklozen hebben omtrent de RVA [An investigation on the experience and the consequences of unemployment and on the expectations that of unemployed people towards unemployment counseling services]. Brussels, Belgium: RVA. De Witte, H. (1993). Psychological consequences of long-term unemployment: Review of the literature. Psychologica Belgica, 33, 1–35. De Witte, H. (1992). Tussen optimisten en teruggetrokkenen: Een empirisch onderzoek naar de het psychosociaal profiel van langdurig werklozen en deelnemers aan de Weer-Werkactie in Vlaanderen [Between optimists and withdrawn: An emperical investigation on psychosocial profiles of unemployed people and participants at the ‘Back-To-Work’ action in Flanders]. Leuven, Belgium: Hoger Instituut voor Arbeid, KULeuven. deCharms, R. (1968). Personal causation: The internal affective determinants of behavior. New York: Academic Press. Deci, E.L. (1975). Intrinsic motivation. New York: Plenum. Deci, E.L., & Ryan, R.M. (1985). Intrinsic motivation and self-determination in human behavior. New York: Plenum.
214
Duration of Unemployment, Job Training and Self-Determination Theory Deci, E.L., & Ryan, R.M. (2000). The ‘what’ and the ‘why’ of goal pursuits: Human needs and the self-determination of behavior. Psychological Inquiry, 11, 227–268. Deci, E.L., & Ryan, R.M. (2002). Handbook of self-determination research. Rochester, NY: The University of Rochester Press. Eden, D., & Aviram, A. (1993). Self-efficacy training to speed reemployment: Helping people to help themselves. Journal of Applied Psychology, 78, 352–360. Elliot, A.J., Faler, J., McGregor, H.A., Campbell, W.K., Sedikides, C., & Harackiewicz, J.M. (2000). Competence valuation as a strategic intrinsic motivation process. Personality and Social Psychology Bulletin, 7, 780–794. Feather, N.T. (1990). The psychological impact of unemployment. New York: SpringerVerlag. Feather, N.T. (1992). Values, valences, expectations, and actions. Journal of Social Issues, 48, 109–124. Feather, N.T., & Davenport, P.R. (1981). Unemployment and depressive affect: A motivational and attributional analysis. Journal of Personality and Social Psychology, 41, 422–436. Guay, F., Boggiano, A.K., & Vallerand, R.J. (2001). Autonomy support, intrinsic motivation, and perceived competence: Conceptual and empirical linkages. Personality and Social Psychology Bulletin, 27, 643–650. Harackiewicz, J.M., & Larsen, J.R., Jr. (1986). Managing motivation: The impact of supervisor feedback on subordinate task interest. Journal of Personality and Social Psychology, 51, 547–556. Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley. Jöreskog, K.G., & Sörbom, D. (1993). LISREL VIII: Structural equation modeling with the SIMPLIS command language. Chicago: Scientific Software International. Koestner, R., Losier, G.F., Vallerand, R.J., & Carducci, D. (1996). Identified and introjected forms of political internalization: Extending self-determination theory. Journal of Personality and Social Psychology, 70, 1025–1036. Kroft, H., Engbersen, G., Schuyt, K., & Van Waarden, F. (1989). Een tijd zonder werk: Een onderzoek naar de levenswereld van langdurig werklozen [A period without work: An investigation of the life-style of long-term unemployed people]. Leiden, the Netherlands: Stenfert Kroese. Losier, G.F., & Koestner, R. (1999). Intrinsic versus identified regulation in distinct political campaigns: The consequences of following politics for pleasure versus personal meaningfulness. Personality and Social Psychology Bulletin, 2, 287–298. Mathieu, J.E., Tannebaum, S.I., & Salas, E. (1992). Influences of individual and situational characteristics on measures of training effectiveness. Academy of Management Journal, 35, 828–847. Miltenburg, T., & Woldringh, C. (1989). Langdurige werkloosheid: Een onderzoek bij jongeren van 18–22 en mannen van 30–44 jaar [Long-term unemployment: Analysis of youth of 18–22 and men of 30–44 years old]. Nijmegen, the Netherlands: Instituut voor Toegepaste Sociale Wetenschappen. Muffels, R., Rietveld, J., & Vriens, M. (1990). Langdurige werklozen op de arbeidsmarkt: Determinanten van arbeidsmarktgedrag en implicaties voor beleid [Long-term unemployed people on the labour market: Determinants of labour market behavior and implications for policy]. Tilburg, the Netherlands: Katholieke Universiteit Brabant. Noe, R.A. (1986). Trainees attributes and attitudes: Neglected influences on training effectiveness. Academy of Management Review, 11, 736–749. Pelletier, L., Fortier, M.S., Vallerand, R.J., & Brière, N.M. (2001). Associations among perceived autonomy support, forms of self-regulation, and persistence: A prospective study. Motivation and Emotion, 25, 279–306. 215
Unemployment and Health ■ Interventions Price, R.H., van Ryn, M., & Vinokur, A.D. (1992). Impact of preventive job search intervention on the likelihood of depression among the unemployed. American Journal of Community Psychology, 19, 806–836. Reeve, J., & Deci, E.L. (1996). Elements of the competitive situation that effect intrinsic motivation. Personality and Social Psychology Bulletin, 22, 24–33. Rowley, K.M., & Feather, N.T. (1987). The impact of unemployment in relation to age and length of unemployment. Journal of Occupational Psychology, 60, 323–332. Ryan, R.M. (1995). Psychological needs and the facilitation of integrative processes. Journal of Personality, 63, 397–427. Ryan, R.M., & Connell, J.P. (1989). Perceived locus of causality and internalization: Examining reasons for acting in two domains. Journal of Personality and Social Psychology, 57, 749–761. Ryan, R.M., & Deci, E.L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development and well-being. American Psychologist, 55, 68–78. Ryan, R.M., Rigby, S., & King, K. (1993). Two types of religious internalization and their relations to religious orientations and mental health. Journal of Personality and Social Psychology, 65, 586–596. Schafer, R. (1968). Aspects of internalization. New York: International Universities Press. Seligman, M.E.P. (1975). Helplessness: On depression, development and death. San Franciso: Freeman. Sheldon, K.M., Turban, D.B., Brown, G.K., Barrick, M.R., & Judge, T.A. (2003). Applying self-determination theory to organizational research. Research in Personnel and Human Resource Management, 22, 357–393. Stafford, E.M., Jackson, P.R., & Banks, M.H. (1980). Employment, work involvement and mental health in less qualified young people. Journal of Occupational Psychology, 53, 291–304. Steiger, J.H. (1990). Structural model evaluation and modification: An interval estimation approach. Multivariate Behavioral Research, 25, 173–180. Tanaka, J. (1987). How big is big enough? Sample size and goodness-of-fit indices in structural equation models with latent variables. Child Development, 58, 134–146. Tharenou, P. (2001). The relationship of training motivation to participation in training and development. Journal of Ocupational and Organizational Psychology, 74, 599–621. Vallerand, R.J. (1997). Toward a hierarchical model of intrinsic and extrinsic motivation. In M.P. Zanna (Ed.), Advances in experimental social psychology (pp. 271–360). San Diego, CA: Academic. Vallerand, R.J., Fortier, M.S., & Guay, F. (1997). Self-determination and persistence in a real-life setting: Toward a motivational model of high-school drop out. Journal of Personality and Social Psychology, 72, 1161–1176. Vallerand, R.J., & Reid, G. (1984). On the causal effects of perceived competence on intrinsic motivation: A test of cognitive evaluation theory. Journal of Sport Psychology, 6, 94–102. Vallerand, R.J., & Reid, G. (1988). On the relative effects of positive and negative verbal feedback on males’ and females’ intrinsic motivation. Canadian Journal of Behavioral Sciences, 20, 239–250. Van Hoye, R., Janssens, E., & Peeters, J. (1988). De determinanten van de herintrede op de arbeidsmarkt [The determinants of re-integration on the labour market]. Antwerpen, Belgium: Centrum voor Sociaal Beleid, UFSIA.
216
Duration of Unemployment, Job Training and Self-Determination Theory Vansteenkiste, M., & Deci, E.L. (2003). Competitively contingent rewards and intrinsic motivation: How can losers remain motivated? Motivation and Emotion, 27, 273–299. Vansteenkiste, M., Lens, W., De Witte, H., & Feather, N.T. (2005). Understanding unemployed people’s job search behavior, unemployment experience and well-being: A comparison of expectancy-value theory and self-determination theory. British Journal of Social Psychology, 44, 269–287. Vansteenkiste, M., Lens, W., Dewitte, S., De Witte, H., & Deci, E.L. (2004). The ‘why’ and ‘why not’ of job search behavior: Their relation to searching, unemployment experience and well-being. European Journal of Social Psychology, 34, 345–363. Vansteenkiste, M., Simons, J., Soenens, B., & Lens, W. (2004). How to become a persevering exerciser? The importance of providing a clear, future intrinsic goal in an autonomy-supportive way. Journal of Sport and Exercise Psychology, 26, 232–249. Verkleij., H. (1988). Langduige werklooshied, werkhervatting en gezondheid. Bevindingen van een 2-jarige follow-up studie [Long-term unemployment, reemployment and health. Findings of a 2-year longitudinal study]. Amsterdam, the Netherlands: Swets & Zeitlinger. Vinokur, A.D., Price, R.H., & Caplan, R.D. (1991). From field experiments to program implementation: Assessing the potential outcomes of an experimental intervention program for unemployed persons. American Journal of Community Psychology, 19, 543–562. Vinokur, A.D., & Schul, Y. (1997). Mastery and inoculation against setbacks as active ingredients in the JOBS intervention for the unemployed. Journal of Consulting and Clinical Psychology, 65, 867–877. Vinokur, A.D., van Ryn, M., Gramlich, E.M., Price, R.H. (1991). Long-term follow-up and benefit–cost analysis of the jobs program: A preventive intervention for the unemployed. Journal of Applied Psychology, 76, 213–219. Vleminckx, A. (1982). Sollicitatiegedrag en arbeidsbemiddeling [Application behavior and counselling]. Kultuurleven, 49, 866–879. Warr, P., & Jackson, P.R. (1984). Men without jobs: Some correlates of age and length of unemployment. Journal of Occupational Psychology, 57, 77–85. Warr, P., & Jackson, P. (1985). Factors influencing the psychological impact of prolonged unemployment and of re-employment. Psychological Medicine, 15, 795–807. Warr, P., & Jackson, P. (1987). Adapting to the unemployed role: A longitudinal investigation. Social Science and Medicine, 25, 1219–1224. Weiner, B. (1985). An attributional theory of achievement motivation and emotion. Psychological Review, 92, 548–573. White, R.W. (1959). Motivation reconsidered: The concept of competence. Psychological Review, 66, 297–333. Williams, G.C., Grow, V.M., Freedman, Z.R., Ryan, R.M., & Deci, E.L. (1996). Motivational predictors of weight-loss and weight loss-maintenance. Journal of Personality and Social Psychology, 70, 115–126. Williams, G.C., Wiener, M.W., Markakis, K.M., Reeve, J., & Deci, E.L. (1994). Medical student’s motivation for internal medicine. Journal of General Internal Medicine, 9, 327–333.
217
Unemployment and Health ■ Interventions
218
Chapter
17
The Role of Limited Duration Contracts in Labour Market Transition Ola Bergström
This chapter is concerned with labour market developments in advanced economies in the late 20th and early 21st centuries, and, in particular, with trends towards the increasing use of limited duration contracts (LDCs). The aim of the chapter is to discuss these trends and the role of LDCs in labour market transition.1 This discussion is timely, given both the current European interest in economic and political integration, and global efforts to create a more flexible and adaptable labour market. In this chapter, an LDC is defined as a contractual relationship between employer and employee that is limited in time. This distinguishes LDCs from permanent or open-ended contracts that are not valid for a predefined period but which instead may be terminated on the request of the employer or the employee under certain conditions. Limited duration contracts also differ from casual employment, which involves workers being employed on an hourly, daily or weekly basis. The proliferation of LDCs during the past decade has intensified the debate regarding the regulation of labour markets in general. While there have been movements towards both regulation and deregulation of LDCs in several countries during this time (Cahuc & Postel-Vinay, 2002), the impact of regulation is not well known. Furthermore, opinions about whether LDCs should be regulated or deregulated vary considerably, partly because of uncertainty about what actually drives their use (OECD, 1996), and partly because of difficulties in determining their effects (Pissarides, 2001). In particular, opinions vary according to the perspective from which consequences are evaluated. In this chapter, I will first outline some of the most common arguments for and against regulation of LDCs. Three perspectives will be described, those of the labour market, the employer, and the employee. Of these three perspectives, the least understood is the employee’s. While it is often assumed that labour market regulation 219
Unemployment and Health ■ Interventions
benefits individual workers, this chapter will argue that some aspects of regulation may actually present problems for workers, and that these problems may be addressed by the use of LDCs. Following the outline of these perspectives, I will draw upon a comparative analysis of LDC use and employment protection legislation in five European countries and in the United States. Finally, I will discuss the impact of employment protection on the use of LDCs, and suggest an alternative way of understanding the role of LDCs in labour markets. Instead of labour market regulation providing workers with security, this chapter argues that it provides workers with a framework for risk calculation. This casts the role of LDCs in the labour market in a different light. It is argued that, from the individual’s point of view, long-term LDCs provide the best insurance for dealing with the effects of economic transformation and employment risk. Thus, LDCs may be regarded as a transitory mechanism in the labour market, enabling workers to move between jobs. With this perspective, LDCs provide a compromise between total job security (which may be unrealistic and unfair to employers) and a total absence of job security (which is unfair to employees).
Regulation and the Labour Market Many influential analysts argue that the increased usage of LDCs in advanced economies during recent decades is indicative of an increasingly insecure and ‘hard-edged’ labour market, with its attendant social costs and market segmentation. From this perspective, LDCs are viewed as unstable and insecure forms of employment that reduce workers’ bargaining power relative to that of the employer. Thus, by creating a pool of marginal labour, these nonstandard forms of employment exacerbate both insider bargaining power in wage negotiations and labour market segmentation (Bentolila & Dolado, 1994).2 Supporting the notion that LDCs reduce workers’ bargaining power is evidence that, in most countries, contingent contracts are overrepresented in vulnerable segments of the labour market, such as women, young people, foreign workers and ethnic groups (Bergström & Storrie, 2003a). These groups are often disadvantaged because of low wages, abuse of working conditions by employers and little or no social security protection (e.g., Meulders & Tytgat, 1989). Regulation of LDCs aims to protect the individual by providing minimum standards for working conditions. It includes measures to prevent LDC abuse, such as limiting both the number of contract renewals and the total duration of successive contracts, as well as specifying objective reasons for contract renewals. The overall aim of these regulations is to prevent temporary employment from becoming a long-term arrangement for LDC workers. There are also rules to prevent contingent employees from being treated less favourably than comparable permanent workers. The underlying assumption is that the open-ended contract is, and 220
Limited Duration Contracts in Labour Market Transition
should be, the norm for all employment relationships, and that any divergence from this norm should be held to a minimum. There are at least two arguments for LDC deregulation. Perhaps the most common is that LDCs facilitate labour market efficiency and flexibility. From this perspective, rigid labour market regulation restricts employers’ control over hiring and remuneration practices, and thereby limits the ability of companies to adapt to competition, globalisation, and economic cycles. If, on the other hand, employers are permitted to negotiate flexible LDCs and to have greater control over wage rates and hiring procedures, there will be greater economic growth, which in turn will lead to greater employment growth. Thus, during a recession, employers will be able to terminate the employment of LDC workers without penalty, while during a boom period they will have the power to negotiate LDCs with new employees. The numbers of LDCs will thus increase during economic upturns and decrease during downturns (OECD, 1996). Another aspect of this argument is that LDCs may be used as an ‘intermediate’ form of employment (a form of probation) allowing firms to screen out unsuccessful job matches. They may also be used to provide certain vulnerable groups (e.g., young people, the long-term unemployed) with access to employment, as in some targeted active labour market programs (OECD, 1996). In such cases, LDCs can be regarded as countercyclical, to the extent that they are used more widely during economic recession than during periods of economic prosperity. From this perspective, regulation of LDCs should be avoided, as they benefit both employers and some groups of employees, and, so the argument goes, if they benefit employers, they benefit society. As Walwei (1998) argues, LDCs may not necessarily have a precarious character, and banning such employment would be absurd and unnatural. Another more radical argument of some social scientists is that LDCs should be interpreted in the light of a new developmental phase in capitalist economies, where globalisation, competition and the use of new technology demand extraordinary changes and restructuring. For example, Rifkin (1995) argues that what we see is the ‘end of work’, where the use of new technology and automatisation leads to greater productivity and less need for labour. This means that there will be continuing unemployment in the labour markets of the advanced economies. Deregulated flexible labour markets, with greater use of LDCs, are regarded as a way to share the burden of underemployment among greater numbers of people (Beck, 1999, 2000). Thus, from this perspective, regulation creates social segregation and conflict rather than eliminating it. However, it is not obvious that deregulation necessarily means increased use of LDCs. Even if, as has often been claimed, the deregulation of temporary contracts in Spain (for example) during the mid-1980s was the dominant reason for increased LDC usage in that country (Esping-Andersen, 2000; 221
Unemployment and Health ■ Interventions
OECD, 1996), the evidence does not strongly support an association between a deregulated labour market and use of LDCs in general. Several empirical studies (OECD, 1993) suggest that, instead, extensive labour market regulation is correlated with greater LDC usage, and vice versa. Robinson (1999), for example, concludes that, while the United States has one of the least regulated labour markets, it also has a relatively low proportion of its workforce in flexible employment. Relatively more regulated labour markets such as those in southern Europe tend to have high levels of temporary employment (Robinson, 1999). Thus, the relationship between regulation and the level of use of LDCs is not evident. Instead, regulation may influence the composition of the contingent labour force. As Esping-Andersen (2000) argues, concerning the impact of employment regulation on unemployment, strong regulation nurtures more stable employment relationships and, therefore, less labour turnover. Deregulating labour markets may diminish security for insiders, but enhance employment opportunities for outsiders, and this may redefine which sections of the workforce or labour market are employed on LDCs. Furthermore, Esping-Andersen claims that, if rigidities do matter, they do so interactively — that is, when they combine with other, more powerful forces such as massive deindustrialisation or changes in the nature of industrial relations. Thus, one cannot regard regulation and LDCs in isolation from each other. They must also be related to broader developments in the economic context, such as unemployment and economic transformation. The relationship between LDC usage and regulation may also depend on what type of regulation one considers. For example, according to Rogowski and Schömann (1996), LDC usage is influenced less by regulation of specific contracts and more by employment protection in general, that is, statutory regulation of open-ended contracts. Rogowski and Schömann (1996) conclude that, thus far, the lesson to be learned from attempts at deregulation is that regulation of fixed-term contracts may be closely linked to dismissal protection. The reform of one part of the employment protection system has repercussions for other parts, which tends to invalidate incremental reforms to increase the flexibility of the institutional structures of the labour market. Thus, we would expect that extensive employment protection would lead employers to hire more workers on LDC contracts. Similarly, in countries where employment protection is more permissive, one would expect that the use of LDCs would be low. The logic of this is that when employment protection is low, employers have no reason not to provide workers with open-ended contracts, as they could be more easily dismissed anyway. Furthermore, to evaluate the impact of labour market regulation on the use of LDCs, it is not enough to consider regulations in law. One must also consider other forms of regulatory institutions, such as co-determination, collective bargaining and the presence or absence of social security systems 222
Limited Duration Contracts in Labour Market Transition
that guarantee a steady income stream (Schmid, 1994). Moreover, a myriad of individual habits, values and preferences, as well as group routines and norms, must be taken into account. Habitual, routine and imitative behaviour is, according to Schmid (1994), the reason why identical formal rules of the game might be played quite differently and different formal rules might end with the same results. To the extent that these norms and values influence individual expectations about the continuation of the employment relationship, they may also impact on how regulatory regimes are actually implemented. In short, the use of LDCs may be influenced by other institutional arrangements than statutory law.
Regulation From the Employers’ Perspective For employers, labour market regulation may be regarded as a disciplinary practice, in that it is directed towards limiting employers’ use of labour in several ways. Regulation of the standard employment contract requires employers to adhere to: (1) administrative procedures for individual dismissal, notice and severance pay; (2) prescriptions about the fairness or unfairness of individual dismissals; and (3) compensation and remediation procedures following unjustified dismissal (OECD, 1999). Furthermore, regulation of LDCs may include restrictions on: (1) the reasons considered permissible for using limited duration contracts; (2) the maximum number of successive contracts; and/or (3) the maximum cumulative duration of successive contracts. When applied to temporary agency work, regulation may include restrictions of the types of work for which this kind of employment is legal, restrictions on the number of renewals, and the maximum cumulative duration of temporary contracts. In some countries, several of these dimensions are used, while other countries place very few limitations on employers’ use of labour (Storrie, 2003). For example, in the case of Spain, it is reasonable to characterise the statutory regulation as extensive. In the OECD ranking, Spain is placed as the 22nd most extensive regulatory regime before countries like Italy, Greece, Portugal and Turkey. The most common form of regulation is to presume that employment contracts are open-ended unless the employer explicitly states otherwise. Contingent contracts are accepted only as an exception to the general rule. In the United States, however, the legal system assumes that all employment contracts may be terminated at will from both parties. Employment protection and more secure jobs are instead regarded as an exception. It is commonly assumed that employers oppose regulation and would rather have the flexibility to hire and fire at will. The costs associated with employment protection are sometimes regarded as a form of taxation on workforce adjustments (OECD, 1999). However, as Esping-Andersen (2000) points out, regulatory frameworks are valued not only by workers, but also by some firms because they foster commitment, trust and cooperation, and facilitate the pursuit of long-term goals in the employment 223
Unemployment and Health ■ Interventions Table 1 Ranking of Employment Protection and Contingent Employment Regulation According to OECD Total employment protection
Total contingent
LDC
TWA
1. US 2. UK 14. Netherlands 16. Sweden 18. Germany 22. Spain
1. US 1. UK 12. Netherlands 13. Sweden 18. Germany 22. Spain
1. US 1. UK 7. Netherlands 15. Sweden 15. Germany 21. Spain
1. US 1. UK 13. Sweden 14. Netherlands 18. Germany 24. Spain
Source: OECD (1999). Note: In the OECD report low ranking means low regulation. Higher ranking means more regulation.
nexus. In a similar vein, Pfeffer (1999) argues that employers’ claims concerning the benefits of labour market deregulation may, in fact, be mistaken. He states that almost all high-performance work practices, such as investment in training, communication and placing people in selfmanaging teams, depend on continuity of employment. Without the expectation of long-term open-ended employment, no company will invest in training. Furthermore, without employment stability, there will not be the trust necessary to practise open management and to share information with employees. In addition, working in teams requires that the team composition is stable enough for people to learn how to work with, and learn from, each other effectively. Furthermore, secure workers may be less likely to resist technological innovation and the reorganisation of working practices (Esping-Andersen, 2000; OECD, 1999). These advantages of employment regulation may encourage firms operating in a deregulated environment to initiate ‘in-house’ regulatory practices (Esping-Andersen, 2000). For example, employers may set up internal policies or guidelines to guarantee a degree of employment security (Guest, 2000), or may provide employees with stock options or similar benefits to retain their loyalty. By way of contrast, formalised regulatory environments may provoke what Esping-Andersen (2000) calls ‘informal flexibilities’, which involve employers devising ways to avoid regulation. For example, in Sweden, which is often regarded as a country with strong employment protection, employers may reach collective agreements with trade unions, which allow the normal rules of employment protection to be circumvented (Bergström & Storrie, 2003b). Furthermore, generous programs to help retrenched employees find jobs elsewhere are implemented (Bergström, 2001). Moreover, employers may try to offset some of the costs of complying with employment protection by negotiating lower wages (OECD, 1999). Thus, regulation could imply both costs and benefits for employers and they may find ways to deal with their flexibility needs independently of the regulatory regime.
224
Limited Duration Contracts in Labour Market Transition
Regulation From the Employee’s Point of View Employment protection is, as stated previously, primarily enforced to provide security and reduce income uncertainty by enhancing job and income security for the individual employee. However, there are also aspects of employment protection and open-ended employment relationships that may have negative effects for individual workers. First, as employment protection is often associated with lower turnover rates and longer employment tenure, workers may spend longer periods in jobs they do not like or for which they are ill-suited (OECD, 1999). Second, openended employment relationships may influence individual expectations concerning income levels. As wages increase with experience in any occupation or skill group, the current job is likely to pay more than a new job (Gregg, Knight, & Wadsworth, 2000). This means that dismissal implies not only a lost job but also a reduced income. The longer a worker has been in the job, the greater this penalty will be if some or all of the gains associated with accumulated experience are lost in the next job. The costs of job loss may therefore be higher among older and more experienced workers, or wherever job loss is a relatively rare event. Third, employment protection affects individual expectations about the continuity of the employment relationship. In most regulatory regimes, employers are required to give advance notice of dismissals or likely future layoffs, thus facilitating job search (OECD, 1999). Studies of downsizing and dismissal programs indicate that employees are often poorly prepared when advance notice is given. The longer they have been employed by the same firm, the more difficult they find it to adjust or to find a new job, and their previous on-the-job training means that their competencies may be limited to those required for that job. The open-ended contract means that employees may also feel an attachment to their employer and be unused to engaging in job-search activities. Even if there have been obvious signals of impending restructuring or lay-offs, employees with open-ended contracts may assume that they will be unaffected. For those workers who are laid off, lack of preparation may increase the risk of long-term unemployment (OECD, 1999). In sum, employment protection and open-ended contracts provide individuals with expectations of a long-lasting employment relationship, but do not provide employees with the skills and preparation required when these expectations cannot be fulfilled. When notice comes, it may often be too late. Fourth, employment protection, in terms of ‘last in, first out’, may inhibit the mobility of workers and their interest in offering themselves to the labour market. As employment protection rewards long-term tenure, individuals relinquish these rewards when they move to a new job. Thus, during periods of economic uncertainty or downturn, protection legislation discourages workers from actively seeking new employment. Thus, there are several problems with employment protection and the norm of an open-ended contract. 225
Unemployment and Health ■ Interventions
As Gregg et al. (2000) assert, the individual’s experience of insecurity is about risk assessment. We may call it employment risk: the chances of losing the current job involuntarily and of securing a new job. From this perspective, it may be argued that, in some circumstances (e.g., in times of economic transformation and uncertainty), various forms of LDC may better serve the needs of the individual. First, since LDCs clearly state when the employment relationship will end, individual expectations of the continuation of the employment relationship is calculable. This means that the individual has incentives to engage in job search activities in order to secure a continuous income. Second, because the LDC encourages transfer to new jobs and employers, it augments the individual’s work experience, which may increase his or her capacity to get another job. Furthermore, switching between jobs means that the individual has a broader network of contacts, which may also be valuable for obtaining new employment. This is particularly so when the temporary job is combined with part-time employment. Individuals may create a work portfolio, which serves the purpose of spreading the risk to employment (Bergström, 2001; Wallulis, 1998). However, when job duration is very short (between 1 and 6 months), the time required or costs of searching for a new employment relationship may be too high. This situation may be taken care of by temporary work agencies that provide employees with job opportunities, by managing their work portfolios as a stockbroker does in financial markets. Thus, from the individual’s perspective, regulation of LDCs may contribute to the individual’s transferability in the labour market, by providing both transparency within the employment relationship, and incentives for employees to take responsibility for planning their future employment.
Comparative Analysis Comparing the use of LDCs, labour market regulation and unemployment statistics in five European countries3 (Spain, Sweden, Germany, the Netherlands and the United Kingdom) and the United States between 1990 and 2000, three general observations can be made: (1) The prevalence of LDCs is higher in those countries that have undergone major changes in unemployment rates during the period. (2) The prevalence of LDCs is high in those countries with relatively strong employment protection. (3) Countries that have undergone major changes in unemployment rates during the period also have the strongest levels of employment protection. During the period studied there are great variations in the level of unemployment in the different countries. However, it is not possible by this comparison to assess whether there is a causal relationship between level of unemployment and the use of contingent employment. On the other hand, comparing the changes in unemployment rates with the level of contingent employment in the 226
Limited Duration Contracts in Labour Market Transition Table 2 Unemployment 1990–2000 1990
1991
1992
1993
1994 1995
1996 1997
1998
1999
2000
Av.
SD
USA 5.6 UK 6.9 Sweden 1.8 Germany 4.8 Spain 15.9 Netherlands 7.5
6.8 8.8 3.3 4.2 16 7
7.5 10.1 5.8 4.6 18.1 5.6
6.9 10.4 9.5 7.9 22.4 6.2
6 9.6 9.8 8.4 23.8 6.8
5.4 4.9 4.5 8.2 7 6.3 9.6 9.9 8.3 8.9 10 9.8 22.2 20.8 18.8 6.3 5.2 4.1
4.2 6.1 7.2 8.8 15.9 3.3
4 5.4 5.9 8.3 14.2 2.8
5.58 7.95 7.26 7.63 19.18 5.61
1.15 1.71 2.77 2.09 3.39 1.58
5.6 8.7 8.8 8.2 22.9 6.9
Source: OECD (1996; 1999).
six countries may offer explanations for the use of contingent employment. Those countries that have undergone major changes in unemployment rates during the period seem to have a higher level of use of contingent employment (Sweden, Spain and Germany) than those with a tighter labour market, where unemployment has been relatively low and stable (Netherlands, the UK and US). Thus, there seems to be a relationship between the use of contingent employment and changes in unemployment rates. It is difficult to determine whether each of these factors influences the use of LDCs alone, or if it is the general employment protection in the labour market that affects both the level of unemployment and the use of LDCs. Unemployment and labour market regulation may affect the use of LDCs in different ways. Increases in unemployment rates, or a more insecure labour market, may increase individuals’ willingness to take contingent jobs, while regulation provides at least minimum security. Employment protection may reduce employers’ willingness to employ individuals on open-ended contracts in times of risk and uncertainty, and thus tend to increase their share of contingent workers. Furthermore, the study emphasises the impact of institutional factors as an explanation for the use of LDCs. On the one hand, in countries where employment protection may be regarded as weak, collective bargaining procedures and legislative practices may compensate for the lack of regulation in statutory law. Thus, what appears as a weak regulation may be relatively strong at a sublegal level. On the other hand, in countries where employment protection may be regarded as strong or extensive, the lack of legal enforcement may contribute to a ‘wild’ or uncontrolled use of LDCs. What appears as strongly regulated may be very difficult to enforce in practice. Thus, the comparison of national legal frameworks needs to be complemented by qualitative analyses of the institutional framework of each particular country. As a result, it is difficult to say which labour market is more flexible: the one with low employment protection or the one with high level of use of LDCs. Labour markets with low levels of employment protection may have as strong a regulation as more strongly regulated regimes on the local level, 227
Unemployment and Health ■ Interventions Table 3 Limited Duration Contracts, Unemployment Rates and Labour Market Regulation Country
LDC intensity1
Unemployment rates2
Employment protection3 Reg. of LDCs4
Spain Sweden Germany Netherlands UK US
Very high High High High Medium Low
Volatile Increasing Increasing Decreasing Decreasing Stable
Strong Strong Medium Medium Weak Weak
Strong Strong Strong Weak Weak Weak
Note: 1 LDC intensity is calculated as the proportion of LDCs in relation to total employment. The comparison of the six countries over the studied period (1990–1999) shows that Spain is the country with the most intensive use of limited duration contracts (32.5%). On the contrary, the US has a relatively low use of LDCs (1.5%). Sweden (14.3%), Germany (and the Netherlands; 13%) would be regarded as relatively high in relation to countries like the UK, which have a more medium position (6.2%) use of LDCs in relation to total employment. Looking at the use of LDCs over the study period 1990–1999 the six countries display a development in different directions. In Sweden, for example, a dramatic increase in the use of LDCs has taken place (31%), while other countries such as the US and the UK have a more stable development. In Spain, a major increase took place during the 1980s and the development has plateaued in recent years. In fact, the LDC in Spain decreased by 10% during the 1990s. 2 Unemployment rates from OECD (1996; 1999). 3 Ranking of employment protection according to OECD (1999). 4 Ranking of LDC regulation according to OECD (1999).
where management adjusts its human resource practices in order to create stability. Similarly, in labour markets with strong employment protection, employers find ways to avoid regulation. The high level of use of LDCs in these countries may be regarded as evidence that there are no problems for employers to provide the numerical flexibility that they need. Thus, the impact of deregulation in these labour markets is difficult to predict. More importantly, the use of LDCs may have important consequences for individuals. On the one hand, it may be regarded as better to hold an open-ended contract in any labour market regime. It provides employees with expectations of a longer employment relationship and thus stable income flows also. On the other hand, for those workers who are exposed to insecurity in times of economic transformation, it may be regarded as better to hold a limited duration contract. In this case, the dismissal is at least calculable. For those who hold an open-ended contract in a weak employment protection regime there is less calculability. It is more difficult to predict when the employment relationship is going to be interrupted and, therefore, more difficult to adjust to future labour market prospects. Thus, the regulation of LDCs provides a framework for calculating the risk of employment from the individuals’ point of view. However, the length of the contracting period may be too short to provide sufficient time to adjust skills and find a new job when the contracting period is over. This feature of LDCs in regulated labour markets may be an advantage for workers in periods of economic transformation when employers need to reduce the number of workers. Limited duration contracts provide the group of individuals who are vulnerable for structural transformation and uncertainty with a calculative framework that enables them to prepare for 228
Limited Duration Contracts in Labour Market Transition
the next stage in their employment. However, it means a shift from collective or employer responsibility to individual responsibility. This is similar to what we find in less regulated labour markets, such as the United States and the United Kingdom. In less regulated labour markets, workers continuously need to calculate and prepare themselves for possible job loss, and the risk of losing the job is always present. A low level of employment protection may stimulate a more active labour force, which in general has a greater capacity to adjust in periods of economic transition. In the regulated labour markets, workers on open-ended contracts tend to postpone their adjustment to labour market changes, trusting the comfort of employment protection, and thereby cause a further need for transformation and renewal at a later stage.
Conclusion: Regulation and Employment Risk The impact of labour market regulation on the use of LDCs is not clear. On the surface there seems to be a relationship between the use of LDCs and the strength of regulation. But the particular dynamics and relationships between different types of regulation are not well known. For example, what is the impact of employment protection regulation and different forms of regulation of LDCs? Is there a causal relationship between legislation and the use of LDCs or are there intermediate factors affecting how labour law is used and interpreted? Furthermore, some forms of deregulation may, due to the institutional framework, have effects that run counter to the aim of labour market flexibility and job stimulation. What is the impact of the institutional framework, the enforcement of laws, and the risks associated with violating the law? It may be argued that policy-makers are mistaken if they assume that extra deregulation automatically translates into extra flexibility. They are equally mistaken if they assume that a given deregulatory practice that works in one country will work similarly in another. As Esping-Andersen (2000) notes, deregulatory policies may, paradoxically, have the perverse effect of strengthening other rigidities. It is difficult to know with certainty what needs to be reformed or deregulated. Thus, the impact of labour market institutions and regulations on the use of LDCs is an open question. However, the analysis shows that the regulation of LDCs may have an important role in facilitating labour market transformation. Previous studies have focused on the use of LDCs as a source for labour market flexibility or an indication of a more insecure and segmented labour market. In contrast to these approaches, I argue that the use of LDCs should be regarded as a transitory mechanism that allows much of the responsibility for, and risk associated with, the employment relationship to be transferred from the employer to the employee. At the same time, LDCs have advantages for employees in that they provide some stability and protection, and this allows employees to plan for future job changes. Such stability and protection may be absent in forms of 229
Unemployment and Health ■ Interventions
employment that have no security at all, such as hourly-paid or casual work. Employment protection encourages employees to expect continuity in the employment relationship but they do not necessarily determine the actual length of the employment relationship. Employees’ expectations are important because they influence their tendency to prepare for and take charge of their employment situation. LDCs encourage employees to protect their interests by thinking about themselves as external to the organisation, and to seek alternative employment before their contract is due to expire. Thus, from the individual’s point of view, limited duration contracts provide a form of insurance for economic transformation and employment risk. However, this presumes that limited duration contracts are relatively long-term (e.g., 3 to 5 years) and not short-term. This suggests a need to rethink the role of limited duration contracts in the labour market. Instead of regarding LDCs as an instrument for employer flexibility or as a symptom of a more insecure labour market, limited duration contracts should be regarded as a model for labour markets in transition. This poses important questions for policy-makers. Instead of trying to avoid the use of limited duration contracts, efforts should be made to make them more stable and of longer duration, providing individuals with a framework to calculate their future. In this way the individual is more prepared to adapt to changes in the labour market.
Endnotes 1
2
3
This chapter is based on data collected by an international group of researchers in a EU-financed project called NUEWO — New Understanding of European Work Organization. Marginalisation is generally referred to as the process where particular segments of the workforce become related to particular jobs, modes of work organisation and employment practices to a degree that reduces their objective range of job choice, so that jobs as well as people become marginal (Loveridge & Mok, 1979). This is for example the case if young people, women, foreign workers or ethnic groups are overrepresented in certain types of jobs or contracts. The selection of countries for investigation is based on the fact that there are important differences between these countries in labour market traditions, legislation, industry structure and trade union membership, possibly affecting the use of contingent labour. We also wanted to compare developments and consequences of contingent employment in Europe to those in the United States, where contingent employment has been least regulated. The countries studied cover what Ronen and Shenkar (1985) refer to as the various cultural clusters in the European continent: the Germanic cluster, the Anglo cluster, the Latin European cluster and the Nordic cluster.
Author Note This study relies on an interdisciplinary and international group of researchers from six countries: Spain, Sweden, Germany, the United Kingdom, the Netherlands and the United States. The researchers cooperate in a project (NUEWO) financed within the EU: Key Action Improving the Socio-economic Knowledge Base. 230
Limited Duration Contracts in Labour Market Transition
References Beck, U. (1999). World risk society. Cambridge, UK: Polity Press. Beck, U. (2000). The brave new world of work. Cambridge, UK: Polity Press. Bentolila, S., & Dolado, J.J. (1994). Labour flexibility and wages: Lessons from Spain. Economic policy, 18, 53–100. Bergström, O. (2001). Externalization of employees: Thinking about going somewhere else. International Journal of Human Resource Management, 12, 373–388. Bergström, O., & Storrie, D. (2003a). Contingent employment in Europe and the United States. Cheltenham, UK: Edward Elgar. Bergström, O., & Storrie, D. (2003b, June 19–20) Restructuring in Sweden since the recession of the early 1990s. Paper presented at the European Foundation for the improvement of working life and living conditions, Dublin, Ireland. Cahuc, P., & Postel-Vinay, F. (2002). Temporary jobs, employment protection and labour market performance. Labour Economics, 9, 63–91. Esping-Andersen, G. (2000). Who is harmed by labour market regulations? Quantitative evidence. In G. Esping-Andersen & M. Regini (Eds.), Why deregulate labour markets? (pp. 66–98). Oxford, UK: Oxford University Press. Gregg, P., Knight, G., & Wadsworth, J. (2000). Heaven knows I’m miserable now: Job insecurity in the British labour market. In E. Heery & J. Salmon (Eds.), The insecure workforce (pp. 39–56). London: Routledge. Guest, D. (2000). Management and the insecure workforce: The search for a new psychological contract. In E. Heery & J. Salmon (Eds.), The insecure workforce (pp. 140–154). London: Routledge. Loveridge, R., & Mok, A. (1979). Theories of labour market segmentation: A critique. The Hague, the Netherlands: Martinus Nijhoff Social Sciences Division. Meulders, D., & Tytgat, B. (1989). The emergence of atypical employment in the European Community. In G. Rodgers & J. Rodgers (Eds.), Precarious jobs in labour market regulation: The growth of atypical employment in Western Europe (pp. 179–196). Geneva, Switzerland: International Labour Organisation. Organization for Economic Co-operation and Development (OECD). (1993). Employment outlook. Paris: Author. Organization for Economic Co-operation and Development (OECD). (1996). Employment outlook. Paris: Author. Organization for Economic Co-operation and Development (OECD). (1999). Employment outlook. Paris: Author. Pfeffer, J. (1999). Labor market flexibility: Do companies really know best? Research Paper No. 1592. Stanford, CA: Graduate School of Business, Stanford University. Pissarides, C.A. (2001). Employment protection. Labour Economics, 8, 131–159. Rifkin, J. (1995). The end of work: The decline of the global labor force and the dawn of the post-market era. New York: G.P. Putnam. Robinson, P. (1999). Explaining the relationship between flexible employment and labour market regulation. In A. Felstead & N. Jewson (Eds.), Global trends in flexible labour (pp. 84–94). London: MacMillan Business. Rogowski, R., & Schömann, G. (1996). Legal regulation and flexibility of employment contracts. In G. Schmid, J. O’Reilly, & K. Schömann (Eds.), International handbook of labour market policy and evaluation (pp. 623–51). Aldershot, UK: Edward Elgar. Ronen, S., & Shenkar, O. (1985). Clustering countries on attitudinal dimensions: A review and synthesis. Academy of Management Journal, Sept, 435–454. Schmid, G. (Ed.). (1994). Labour market institutions in Europe: A socioeconomic evaluation of performance. New York: M.E. Sharpe. 231
Unemployment and Health ■ Interventions Storrie, D. (2003). Contingent employment in Europe and the flexibility security trade off. In O. Bergström & D. Storrie (Eds.), Contingent employment in Europe and the United States. Cheltenham, UK: Edward Elgar. Walwei, U. (1998). Flexibility of employment relationships: Possibilities and limits. In T.D. Donley & M. Oppenheimer (Eds.), International review of comparative public policy (Vol. 10). London: JAI Press. Wallulis, J. (1998). The new insecurity: The end of the standard job and family. Albany, NY: State University of New York Press.
232
Chapter
18
Youth Unemployment and the Risk of Social Exclusion in Six European Countries Thomas Kieselbach
The psychological consequences of job loss and unemployment have played an important role in the scientific investigation of the consequences of unemployment. The research that will be described in this chapter has tried to go beyond the classical focus of psychological unemployment research in the last 25 years, first, by linking the psychological consequences of unemployment to processes of social exclusion, and second, by examining the significance of differences in national institutional and cultural contexts for vulnerability to exclusion and for the prevalence of specific types of moderators of the psychological consequences of unemployment.
The Psychology of Unemployment The effects of unemployment can only be understood by taking into account both the social and individual meanings of work. The starting point for any account of social–psychological approaches to the consequences of unemployment must be the work of Marie Jahoda, who, as an author of the Marienthal study in the the early 1930s, was co-responsible for one of the most influential studies in the social sciences (Jahoda, Lazarsfeld, & Zeisel, 2002). Assuming that work represents the most important connection to reality, Jahoda (1979, 1982) tried to define the latent functions of work — as distinct from the manifest function of earning one’s living — in terms of five categories. The primary emphasis was on the significance of structural context. The activity of work incorporates the following psychological dimensions: it imposes a time structure, it compels contacts and shared experiences with others outside the nuclear family, it represents goals and purposes beyond the scope of an individual, it enforces activity, and it provides status and social identity (Jahoda & Rush, 1980). 233
Unemployment and Health ■ Interventions
Fryer and Payne (1984) opened the debate about the importance of individuals’ own agency and the personal meanings they attach to work by focusing on people who apparently did not experience negative consequences but coped relatively well with unemployment (the so-called ‘proactive unemployed’). Their small but highly influential qualitative study was based on 11 unemployed people, who were preselected for their ability to cope well with the experience of job loss and continuing unemployment. While taking account of the resource side of the individual coping process, the study remained focused on the importance of individual agency. Yet, given the enforced nature of exclusion from paid work in our societies, an approach based on the latent functions of work was perhaps a more persuasive social psychological account than one concerned predominantly, as in the agency approach, with the active role of the individual. A new phase in unemployment research was initiated by Warr in the 80s, culminating in his comprehensive review, Work, Unemployment and Mental Health (1987). This readdressed the issue of the characteristics of the social environment that are conducive or detrimental to mental health. Warr emphasised a number of factors in this respect: the opportunity for control, the opportunity for skill use, externally generated goals, variety, environmental clarity, availability of money, physical security (accommodation, food, etc.), opportunity for interpersonal contact and valued social position. The unemployed are likely to experience disadvantage with regard to most of these factors. Their opportunity for control is reduced by repeated failure while looking for a job, the impossibility of influencing employers, and dependence on the social welfare bureaucracy. They experience a decline in opportunities for skill use, and a decrease in externally generated goals due to such factors as their inability to use their vocational qualifications, the fact that only a minority use their leisure time actively, the low level of external demands, the reduction of the range of goals and the decreased significance of previously important goals. Restriction to the family environment and the home, loss of stimulation from the contrast between work and leisure, and reduction in financial possibilities all lead to a reduction in variety. Uncertainty about which behaviour will lead to success in finding a job, and the necessity of presenting oneself in a different light with each application, results in loss of environmental clarity. The clarity of the social environment is further reduced by a style of life ‘characterised by short-term thinking’ (Biermann, Schmerl, & Ziebell, 1985), as a result of the uncertainty regarding any future plans. The reduction of financial means can lead to deterioration of nutrition, and, sometimes, the necessity of moving into a different neighbourhood (‘physical security’), which in turn can reduce the level of support from the unemployed person’s usual social network. Interpersonal contacts can increase in number due to more available time, particularly in the case of young unemployed people. For other groups of the 234
Psychology of Unemployment and Social Exclusion
unemployed, however, there is a greater risk of self-isolation due to financial restrictions or to a sense of shame. The work ethic in our society tends to devalue the unemployed socially. This situation is often reinforced through individualistic ideologies, such as the ‘belief in a just world’ (Lerner, 1974, 1980; see also Dalbert, this volume), which imply that each person gets what he or she deserves. In this way the social position of the unemployed is undermined, their claims on society are deprived of legitimacy, and they themselves are rendered incapable of asserting themselves in conflict situations (Kieselbach, 1987). This weakened social position is supplemented by dependence on state institutions, which put those dependent on them into the role of humble petitioners, giving rise to feelings of worthlessness and helplessness. While research has tended to focus on the general effects of unemployment for psychological wellbeing, a number of studies have recognised that the strength of these effects can vary. The more systematic introduction of a series of moderators and mediators co-determining the individual effects of unemployment has laid the foundation for research into the differential effects of unemployment (Wacker, 1983). This approach is capable of explaining the personal effects of unemployment in a more differentiated way than is often suggested by the relatively naive assumptions underlying the public discourse. The different forms of individual coping with unemployment correspond to a lack of uniformity among the unemployed themselves in terms of both demographic and psychological features (see Kieselbach, 1991). An analysis of the moderators involved in coping highlights the fact that, just as unemployment itself is no uniform event, no uniform reactions can be predicted about the people directly affected by unemployment. Two directions have been taken by researchers in the attempt to develop typologies of the various forms of reactions to unemployment. The first was limited to distinguishing between different types of individual response to unemployment. In the influential interwar Marienthal study, Jahoda et al. (2002) distinguished between the following four types of possible reaction: • brokenness, subdivided into (a) apathy (comprising the inability to properly conduct household affairs; an indifferent, indolent mood; total lack of planning; high alcohol consumption, begging, stealing), and (b) despair (comprising depression, hopelessness, loss of control, abandonment of job search, social comparison with a better past) • resignation (no more expectations from life; relinquishing of planning; lack of future orientation; hopelessness, but still with the ability to manage household affairs) • unbrokenness (highly active, sustaining plans and hopes for the future, continuous attempts to look for work). A more recent approach, derived from a small empirical study in Northern Ireland, developed a typology oriented to the types of activity that people engaged in when unemployed (Kilpatrick & Trew, 1985). The authors 235
Unemployment and Health ■ Interventions
analysed the time use of unemployed men and their respective mental health measured by GHQ mean scores and formulated the following typology: • passive cluster (no, or only passive activities like watching TV: lowest scores for mental health) • domestic cluster (most time at home devoted to domestic chores: mental health only slightly better than in the passive cluster) • social cluster (social activities with others: mental health less affected by unemployment) • active cluster (active leisure time and work-related activities: mental health least affected by unemployment). Warr (1989), criticising any concept of adaptation that merely implies a reduction in psychological morbidity without referring to the extent to which comprehensive individual development is severely hampered by unemployment, distinguishes between ‘resigned’ and ‘constructive’ adaptation. In the case of constructive adaptation the person affected develops interests and activities outside the conventional labour market, spends more time on hobbies, extends his or her social activities and gets involved in voluntary or community organisations, whereas the more widespread form, resigned adaptation, consists of self-isolation and a far-reaching reduction in expectations in many areas of life. A second approach tried to develop a dynamic theory of individual responses, postulating a specific pattern of change as the unemployment spell lengthened. Bakke (1940) formulated a ‘cycle of adaptation’ in unemployed families: momentum stability — unstable equilibrium — disorganisation — experimental readjustment — permanent readjustment. The sequence shock — optimism — pessimism — fatalism, as described by Harrison (1976), has perhaps become the best known formula. The presumption of any such ideal–typical sequence has been criticised on various grounds, in particular because it lacks an adequate empirical basis and creates a stereotyped picture of the behaviour of unemployed people (Kelvin & Jarrett, 1985). The dominant perspective of psychological unemployment research has focused on the health (more specifically mental health) effects of job loss and unemployment. This perspective has been considerably broadened in the last decade by a growing interest in processes leading to social exclusion. While a clear general picture of the relationship between unemployment and health emerged from previous research, the question of key mechanisms linking unemployment and social exclusion had not been dealt with. When excluding mechanisms are discussed, reference is primarily made to the factors that lead to exclusion from the labour market. The processes linking labour market exclusion with social exclusion have not been given priority until now. Research that would allow insights into the more subtle processes underlying the societal inclusion and exclusion of the young unemployed in society clearly requires the systematic incorporation of both monetary and nonmonetary, and objective and subjective factors. 236
Psychology of Unemployment and Social Exclusion
Psychology of Unemployment and Social Exclusion While acquiring considerable political salience among European Union (EU) policy-makers, ‘social exclusion’ was developed as a predominantly scientific concept within French sociology (Paugam, 2000), and subsequently elaborated by US and German sociologists (Kronauer, 2002; Silver, 1994). The increasing interest in the topic can be explained by the trends in the labour market, in particular the persistence of higher levels of structural unemployment than had been experienced prior to the 1980s, and an increasing tendency towards flexibility and insecurity in the labour market. It has been reinforced by the greater policy pressure on people without sufficient coping resources to respond adequately to the changing requirements of the labour market. This reflects changing views about the proper role of the welfare state and an emphasis on activation strategies that seek to place greater responsibility on the individual. An important concern for academic researchers became whether the experience of unemployment may trigger processes of social exclusion that crystallise lines of social stratification, with certain groups of the population coming to be regarded as ‘dispensable’ and outside the scope for assistance (Vogel, 2000). While widely used in the political debate about poverty and unemployment in Europe, social exclusion must be considered as an umbrella concept rather than as a precisely operationalised term. Perhaps the most useful elaboration of the concept is that by Kronauer (1998), which is more comprehensive than many others. In contrast to the use of the term in France (or of the concept of underclass in the USA), Kronauer developed his understanding of social exclusion in the light of the current employment crisis that especially affects low-qualified manufacturing workers. Sustained high unemployment rates have the consequence that more and more people cannot lead a life which fits the societal standards for material and social wellbeing. This analysis of the social consequences of unemployment and poverty requires a terminology that takes into consideration both monetary and nonmonetary aspects of living, and the characteristics of the individual and the society. This broader understanding has also become increasingly important for analysing the situation of young people facing unemployment (Kieselbach, 1997). Correspondingly, in this research project social exclusion is understood as a dynamic, multidimensional process that incorporates economic and social (monetary and nonmonetary) aspects of living, subjective experiences and objective situations, and which depends upon available personal and social resources. The use of the term social exclusion easily evokes the image of a specific situation that stands in contrast to that of social inclusion. But it is better seen as defining the end point of a dimension characterised by greater or less vulnerability to marginalisation. Thus, social exclusion can only be understood by focusing not only on what it means to be excluded, but also on those factors enlarging or diminishing the vulnerability of the individual. For this reason, 237
Unemployment and Health ■ Interventions
e x c l u s i o n
vulnerability inclusion zone
e x c l u s i o n
Figure 1 The exclusion–inclusion paradigm.
we will use principally the term ‘risk of exclusion’. Figure 2 depicts this inclusion–exclusion paradigm. The arrows stand for different hypothetical ‘movements’ of people within this continuum. Kronauer (1998) argues that, while social exclusion is always linked with unemployment, it is only applicable if it is accompanied at least by both a marginal economic position and social isolation. It is then a concept that refers to multidimensional disadvantage. For the purposes of the present research, six dimensions of social disadvantage were investigated: 1. Exclusion from the labour market. This describes the situation of facing external barriers to (re-)enter the labour market combined with resignation regarding one’s employment chances. 2. Economic exclusion. This is usually referred to as poverty and includes financial dependency upon the welfare state or a socially unacceptable income, and the loss of the ability to financially support oneself or one’s family. 3. Institutional exclusion. This can be from the educational system (both schools and institutions of further education and training), institutions dealing with unemployment and poverty, and public and private service institutions (such as banks and insurance agencies). Besides the lack of support both before and during phases of unemployment, the experience of institutional dependency can lead to feelings of shame and passivity, making the effect of state support counterproductive. 4. Exclusion through social isolation. This describes either withdrawal by other members of the person’s social network or their own retreat, leading to a reduction of contacts to only one specific group of people or even to general social isolation. This dimension is closely linked to the next one. 238
Psychology of Unemployment and Social Exclusion
5. Cultural exclusion. This refers to the inability to live according to socially accepted norms and values, with the possible consequence of an identification with deviant norms and behaviours. Stigmatisation and sanctions from the social surroundings are also subsumed within this dimension. 6. Spatial exclusion. This manifests itself in the spatial concentration of people with limited financial means, often coming from a similar social and/or cultural background, and in feelings of isolation due to the lack of an adequate infrastructure within the residential area (e.g., with respect to transport, shops, and so on, but also to cultural events).
The Research and Its Context The Youth Unemployment and Social Exclusion (YUSEDER)1 research project focused on the risks of social exclusion among long-term unemployed youth and their implications for psychological wellbeing. It examined the key mechanisms linking the experience of long-term youth unemployment to various dimensions of social disintegration. In this context, not only the mechanisms exacerbating the stress of unemployment (vulnerability factors) but also the protective mechanisms preventing or reducing the risk of social exclusion have been central to the research (Kieselbach et al., 2000a, 2000b; Kieselbach, van Heeringen, Lemkow, Sokou, & Starrin, 2001). The research was carried out in three northern European countries (Sweden, Belgium, Germany) and in three southern European countries (Spain, Italy, Greece). At the time of the research, youth unemployment was significantly higher in all partner countries than total unemployment rates except in the case of Germany where — due to the dual education system and the associated longer periods of training — the rate of youth unemployment was similar to that of adult unemployment (see Figure 1). However, the German labour market is characterised by a high degree of disparity between the situation of West and East German youth. Young people between the ages of 20 and 24 years are much more severely affected by unemployment in East Germany. There were, however, marked differences between countries in the extent of youth disadvantage. In Belgium, the rates were two to three times higher than the overall rates, with female unemployment rates especially high. An extreme situation can be found in the region Wallonia, where over 40% of all females under 25 are unemployed. Parallel to the increase in total unemployment in the early 90s, in Sweden, youth unemployment increased rapidly and equally strongly for the 16- to 19-year-olds and the 20- to 24-year-olds. After a slight decrease in the mid-90s, the unemployment rate for both age groups rates was about 15%, twice as high as the overall unemployment rate. But, in sharp contrast to the Belgian situation, in Sweden the rates were higher among men than among women (in 1995, 16.6% vs. 14.0%). 239
Unemployment and Health ■ Interventions
45 38.8
40 33.1
35
31
30 23
25
20.9
20.9
overall unempl.
20 15
youth unempl. 9.7 10.3
12.1
10.2
9.5
10
9.6
5 0 n
ai
Sp
ly
a It
e ec re G
en ed
m iu lg Be
Sw
y an m er G
Figure 2 Overall unemployment rates versus youth unemployment rates (> 25 years of age) for six European countries in 1997. Source: Eurostat, 1998.
The highest levels of youth unemployment were to be found in the southern countries. In Greece well over a third of young people between 15 and 19 years of age were unemployed in 1995. Rates among young women were especially high. In 1995 the rate for young women aged 15 to 24 was 37.7%, compared to 19.4% for men. In Italy, the proportion of young unemployed people constituted two thirds of the total number of unemployed people, with lower rates for young men (29% in 1997) than young women (37.7%). The proportion of young people who have been looking for a job for more than 12 months is exceptionally high in this country (approximately two thirds of the young unemployed). Finally, Spain shows the highest rate of youth unemployment. In addition to a continuous rise in the 70s, unemployment figures rose in the late 70s and early 80s to a peak value of nearly 40%. Especially among the 16- to 19-year-olds, the rate sometimes exceeded 50%. To provide an in-depth view of the situation and experience of young unemployed people, 300 interviews were conducted (50 interviews in each country). The comparability of the sample — taking into account different schooling, qualification and welfare support systems within the six countries participating in the project — was ensured through the adoption of common selection criteria. Those interviewed had to be between 20 and 24 years of age2 and officially registered as unemployed for at least 12 months. The distributions by gender and qualification level reflected those existing in the populations of the respective countries. It was decided not to include immigrants in the study as this would have heightened the diversity of the sample. 240
Psychology of Unemployment and Social Exclusion Table 1 Gender in Respective National Study Population Distributed by Region Region A Male Sweden Belgium Germany Greece Italy Spain
11 14 10 7 13
Female
Region B Male
Female
Interviews conducted only in one region 14 9 12 9 22 7 13 14 21 6
16 15 11 16 10
Total
Region Male
Female
20 20 23 17 21 19
29 30 27 33 29 31
120
179
In addition, language barriers might have made it too difficult to conduct the interviews. Due to strong regional disparities in unemployment rates of young people within most countries, partners were asked to select the interviewees from more than one area. With the exception of Sweden, each national study contained samples from two areas reflecting differences in low versus high unemployment rates, urban versus semiurban contexts, and in the case of Italy and Germany, different geographical regions (southern and northern Italy, East and West Germany). The study populations largely coincide with the statistics concerning qualification level of unemployed youths in the respective countries: in the Belgian, German and Swedish studies, it is primarily young people with lower qualification levels that have difficulties in accessing the labour market. In the Greek, Italian and Spanish studies this is also the case for those with a higher educational background.3 It should be stressed, however, that the samples should not be regarded as a representative of the overall population of long-term unemployed young people. The method chosen for carrying out the qualitative interviews with longterm unemployed youth was the ‘problem-focused interview’ that was developed at the University of Bremen in the 1980s and which contains both standardised and narrative elements (Witzel 1987, 1995). The interview schedule contained seven thematic fields based on Kronauers’ (1998) six dimensions of social exclusion, together with a section on psychosocial strains that included questions about feelings of victimisation, shame, financial hardship, and also experiences in the submerged economy. For the southern European countries additional questions were included on the submerged economy to assess the situation in greater depth.
Youth Unemployment and Social Exclusion It is clear from the analyses that the main factors creating a high risk of social exclusion among young people were unemployment and its economic effects. This supports the assumption in earlier social–psychological literature that 241
Unemployment and Health ■ Interventions Table 2 Qualification Level in Respective National Study Population Distributed by Fegion Region A Higher Sweden Belgium Germany Greece Italy Spain Total
11 10 26 15 24
Lower
Region B Higher
Lower
Interviews conducted only in one region 14 10 15 16 15 9 6 17 1 5 25 5 10 10 6
Total Higher
Lower
38 21 25 43 40 33
11 29 25 7 10 17
200
99
work is one of the main mechanisms for overall societal integration. The inability to enter the labour market in the first place must be recognised as a central factor in determining the further development of young people. It might even be justified to say that, in the long run, having versus not having work sets the agenda for integration into or exclusion from society. One of the main reasons for exclusion from the labour market is the absence of qualifications, or the possession of generally low (or notmatching) job qualifications. Without vocational training that meets the (future) requirements of the labour market, lasting inclusion into work seems to be extremely difficult. It is therefore understandable that job qualification schemes are usually regarded as the most effective measures against unemployment and social exclusion. However, the relatively exclusive focus of policy interventions on this type of institutional answer underestimates the importance of other institutions and their potential effects in reducing the vulnerability of adolescents. In particular the research highlighted the significance of earlier school experience in preparing young people for their future job situation and in reducing risks of social exclusion. Here there is clear evidence of the cumulative nature of disadvantage. Absent or low qualification levels are often found among young people who have grown up in a situation of relative poverty, in deprived areas that lead to multiple socialisation deficiencies. Despite these general tendencies in the data, there were also marked divergences between individuals in their risk of social exclusion. Individual cases were assigned to one of three groups according to whether the person faced a high risk of social exclusion, an increased (moderate) risk of social exclusion, or only a low risk of social exclusion. Three dimensions — labour market exclusion, economic exclusion and social isolation — were found to be central dimensions for the risk of exclusion in all countries. The group of long-term unemployed young people at high risk of social exclusion was made up of cases that showed at least three aspects of social exclusion, of which at least two were central dimensions. The group at 242
Psychology of Unemployment and Social Exclusion
Germany Belgium high risk
Spain
increased risk Sweden
low risk
Greece Italy 0%
20%
40%
60%
80%
100%
Figure 3 Types of social exclusion in six countries of the EU (each country N = 50, in %).4
increased risk of social exclusion consisted of those for whom two central but no other criteria applied, or those with not more than one of the central dimensions and any number of noncentral aspects. The group showing a low risk of social exclusion was made up of cases that showed exclusion tendencies in only one area, which was not one of the central dimensions. Such operationalisation provides a balance between the requirement of a common basis for comparison, and the need to take account of national specificities in risk factors. There were marked national differences in the prevalence of these different risk categories. As can be seen in Figure 3, in Belgium and Germany, cases at high risk of social exclusion were most common. In contrast, most of the young people in Sweden and Spain were part of the group with an increased risk of social exclusion. In Greece and Italy, people at a low risk of social exclusion formed the largest group. Type 1: High Risk of Social Exclusion In most of the national studies, a high risk of social exclusion arose if young people experienced simultaneously a high risk of labour market, economic and social isolation. These young people had longer durations of unemployment in all countries compared to other groups. The single most important factor that increased the risk of social exclusion was low qualification. Providing access only to poor and precarious jobs, low qualifications are an obstacle to building a professional career and make it difficult to escape from unemployment. Furthermore, the young people in question came from lower social classes or from families with major social and financial problems. Due to financial and other problems in their family of origin, they were at a disadvantage from childhood onwards. Their qualification level and their self-esteem were low, 243
Unemployment and Health ■ Interventions
they were not sufficiently supported by their social environment and governmental institutions and they tended to be passive, or, in some countries, to engage in problem behaviours, such as drug-taking and socially deviant behaviour. The young people in this group also exhibited a high degree of passivity towards the labour market. They saw few or no chances of finding regular work; often they were no longer actively seeking work or only in a limited way. Besides diminished job search activity, their efforts to improve their qualifications were also significantly lower. A low level of institutional support is a vulnerability factor for all study populations except Sweden, where some members of this group received a relatively high amount of institutional as well as social support. In Belgium, Germany and Sweden, people in this group were also likely to experience cultural exclusion, while in Greece and Italy the risk of spatial exclusion was common. Finally, it was observed in all countries that personality-related factors such as low self-esteem and poor mental health increased the risk of social exclusion. Following are case descriptions and chart profiles of young people at high, increased and low risk of social exclusion. The first example is Andi, a 20-year-old from East Germany. Andi is an example of a person in the high-risk group. He has been without work since he left school at the age of 14. He lives alone in a oneroom flat and has a girlfriend who is expecting his child. Problems with his family and socialisation deficits led to failure in school and, in turn, impeded his integration into the labour market. Other circumstances make him vulnerable to the various dimensions of functional exclusion that can lead to social exclusion. Alcohol and drug consumption and the increased propensity for violence, for example, make it all the more likely he will be socially excluded. The only countervailing forces are a circle of friends and the emotional support of his girlfriend. He experiences illnesses that are both a cause and effect of his exclusion. He has asthma and hallucinations, which mainly result from drug consumption. He also shows signs of nervousness, restlessness and aggressiveness. Furthermore, he has low selfesteem and expresses high insecurity towards his personal future. Type 2: Increased Risk of Social Exclusion The group of young people at increased risk of social exclusion seems to be relatively heterogeneous. In all countries it involves the risk of labour market exclusion. But the other relevant exclusion dimensions for this type vary in each country: they may be institutional exclusion, spatial exclusion (Greece) or economic exclusion (Sweden, Spain). The qualification levels reveal a general lack of educational resources. The risk of social exclusion for this group is mainly counteracted by a high degree of family support (Greece, Italy, Sweden, Spain) or of support from the social environment (Belgium, Germany). Furthermore, general activity is at a high level, which is also likely to be a protective factor. 244
Psychology of Unemployment and Social Exclusion DIMENSIONS OF SOCIAL EXCLUSION
Contributing Factors to Key Mechanisms of Social Exclusion
Antecedent Variables
-
-
no educational training continuous problems in school (lower intellectual ability) parents’ divorce at the age of four
Protective Factors
- emotional support from girlfriend
LM
- good circle of friends
EC
-
qualification deficits
-
self-exclusion
-
financial problems income from criminal activities
-
- intermittent work
IN
-
Vulnerability Factors - heavy alcohol and drug consumption
UNEMPLOYMENT
low institutional support failure in school after the unification
-
social isolation
-
societal isolation awareness of own poverty identification with deviant norms
HEALTH EFFECTS
-
-
-
physical health impairments (asthma caused by drug consumption, hallucinations) mental disturbances (nervousness, restlessness, aggressiveness) low self-esteem worries about future
SI
- health problems (mental disturbances)
- increased propensity towards violence
-
CU
-
-
SP
poor living conditions unsafe residential area
- Low self-esteem
Figure 4 Case chart type ‘High Risk of Social Exclusion’ (Andi, East Germany). Note: LM = Labour market exclusion; EC = Economic exclusion; IN = Institutional exclusion; SI = Exclusion through social isolation; CU = Cultural exclusion; SP = Spatial exclusion.
Strong social connections (friends and family) counteract the risk of social exclusion. The dimension that makes the difference between ‘high’ and ‘increased’ risk of social exclusion is social isolation. The strong link to the family described in the Italian, Greek and Spanish studies, however, is not unambiguously positive. Although the existence of good social networks (especially family, but also friends) reduces the immediate threat of economic exclusion, a feeling of economic dependence on the family undermines people’s capacities for autonomous choice. Olga is an example of the type at increased risk of social exclusion. She lives in West Germany, is 23 years old, lives without a partner and has been unemployed for two years. Problems before being unemployed are deficits in education and a three-time failure in vocational training. Her situation 245
Unemployment and Health ■ Interventions CONTRIBUTING FACTORS to key mechanisms
DIMENSIONS OF SOCIAL EXCLUSION
protective factors
Increased Risk of Social Exclusion
LM -
- no vocational qualification achieved resignation concerning prefered occupation
support from parents
- secure financial and spatial situation -
high orientation towards work
-
institutional support
GER-W12: OLGA
HEALTH EFFECTS
EC
-
-
permanent circle of friends
IN
-
eating due to boredom (gaining in weight) irritability psychosocial strains due to feelings of shame
-vulnerability factors withdrawal on family dependency
UNEMPLOYMENT
-
demotivated towards training
-
low mobility
-
feelings of shame
SI
CU
SP -
-
three time failure in vocational training educational deficits
Antecedent Variables
Figure 5 Case chart type ‘Increased Risk of Social Exclusion’. Note: LM = Labour market exclusion; EC = Economic exclusion; IN = Institutional exclusion; SI = Exclusion through social isolation; CU = Cultural exclusion; SP = Spatial exclusion.
at the time of unemployment is marked by a balance of protective and vulnerability factors. By living with her parents her financial as well as her spatial situation is secured. Furthermore, she can count on help from her circle of friends. Factors which are an obstacle to integration into the labour market are demotivation due to several failures in vocational training schemes and a lack of willingness to be mobile geographically. She 246
Psychology of Unemployment and Social Exclusion
feels that she does not receive sufficient institutional support. Furthermore, security through the family means at the same time family dependence. The psychological consequences she experiences as a result of unemployment are feelings of shame, irritability and boredom. Type 3: Low Risk of Social Exclusion Long-term unemployed youth in this group are at most affected by one dimension of social exclusion and this is not one of the central dimensions. The risk of social exclusion is reduced by a range of protective factors: high qualification, active behaviour at the labour market, secure financial situation, social support, institutional support, high self-esteem and high level of sociocultural activities. Compared with the other two groups, the young people at low risk of social exclusion in all studies are younger, are more highly qualified and have been unemployed for a shorter period of time. They are more actively seeking a job, are in a relatively secure financial situation, and are supported by their social environment. The risk of labour market exclusion is low because unemployment is regarded as a temporary situation, and also as a time for personal development and planning. People at low risk of social exclusion neither feel socially isolated nor economically excluded. In southern Europe they receive sufficient financial support from their families and in northern Europe from state institutions. They are also satisfied with the support they receive from governmental institutions. Moreover, the interviewees are not only actively looking for a job but are also more committed to sociocultural activities than those in the other risk groups. Most of them are continuously occupied with activities of personal interest. They participate actively in associations and organised groups. Different kinds of support (financial, social and institutional) are of major importance for this group. For some interviewees social support enables them to subordinate the search for a job to the maintenance of their current lifestyle, which is characterised by temporary jobs and personal interests like music (Italy). For others, unemployment is closer to a free choice (Belgium). Youth at low risk of exclusion usually have protective personality features. All have a high level of self-esteem and good communication skills. They are able to make decisions, to plan out and implement positive changes in their lives and to cope with new requirements. Some of the young people at low risk of social exclusion depend to a lesser extent on institutional support due to their ability to help themselves (Sweden). Ingo is an example of this type (see case chart in Figure 6). He is 22 years old and has been without work for 1 year. He comes from West Germany. He has completed a vocational training and lives with his parents. He has not been taken on after his apprenticeship, possibly due to losing too many working hours through absenteeism. His situation at the time of unemployment is marked by a number of protective factors. Due to his educational and vocational training he is relatively high qualified. He has clear objectives concerning his career and receives adequate support from the unemployment 247
Unemployment and Health ■ Interventions DIMENSIONS OF SOCIAL EXCLUSION
FACTORS to key mechanisms
Low Risk of Social Exclusion
GER-W6: INGO
HEALTH EFFECTS
protective factors LM
-
high qualification
-
clear personal objectives
EC
adequate support from unempl. office -
support from parents
-
secondary income possibilities
IN
vulnerability factors unrealistic high aspiration level
SI
not willing to make a driving licence
CU UNEMPLOYMENT
SP -
problems in training company (dissatisfaction, many working hours lost through absenteeism)
Antecedent Variables
Figure 6 Case chart type ‘Low Risk of Social Exclusion’. Note: LM = Labour market exclusion; EC = Economic exclusion; IN = Institutional exclusion; SI = Exclusion through social isolation; CU = Cultural exclusion; SP = Spatial exclusion.
office. He is emotionally and financially supported by his family and has good secondary income possibilities. However, his unrealistically high aspiration level, as well as his unwillingness to obtain a driving licence, are barriers to quick integration into the labour market. Nevertheless none of the six dimensions of social exclusion emerge clearly and there is no evidence of health problems. 248
Psychology of Unemployment and Social Exclusion
Youth Unemployment, Social Exclusion and Health Despite the limited number of studies and the quality of empirical evidence (mainly cross-sectional data), the analysis of previous research evidence from the different countries revealed rather homogenous results on the link between youth unemployment and ill health (Kieselbach et al., 2000b). In general, it was reported that, compared to their employed peers, young unemployed people had a distinctly higher risk of health-related problems. This is especially true for mental health and psychosocial problems leading to an increase in depression and a poorer quality of life, but also for objective health indicators, for instance with respect to the higher risk of suicidal behaviour among unemployed youth. This is also reflected in the health behaviour of young unemployed people, especially with regard to alcohol and cigarette consumption. Our research showed that, in addition to the effect of such individual differences, the health of the long-term unemployed youth and the healthrelated effects resulting from unemployment varied considerably between the different countries. In particular, the cultural and social differences between northern and southern European countries had significant implications for health. To a considerable degree, this reflected the implications of such factors for the likelihood that people were in particular categories with respect to social exclusion risks. However, cultural differences also led to some differentiation in the health implications of unemployment within specific social exclusion risk categories. High Risk of Social Exclusion For those at high risk of social exclusion, all studies, except the Greek, reported a number of psychosocial strains directly resulting from unemployment. These effects were mainly passivity, the feeling of lack of opportunities, as well as apathy and a sense of resignation. Furthermore, in nearly all countries, low self-esteem or low self-confidence was observed among youth at high risk of exclusion. However, in the northern European countries, a greater number of psychosocial stress factors were reported compared to the southern European countries. In Belgium, Germany and Sweden, financial stress was considered to be the crucial stress factor. Fear of the future, missing a perspective for life, and feelings of dependency were also common. This indicates that, among people at high risk of social exclusion, the diversity of psychosocial strains is greater in the northern European countries than in Greece, Italy and Spain. In Greece, in general, the health problems of long-term unemployed young people were seen as a preexisting condition and are not considered a result of unemployment. In Spain, psychosocial strain in connection with social problems within the family and circle of friends was particularly salient: if support did not exist or was insufficient, young people were particularly likely to feel lonely, bored, afraid of the future and depressed. In most of the countries, there is a homogeneous picture with respect to the health-related behaviour (tobacco and alcohol consumption, abuse of 249
Unemployment and Health ■ Interventions
drugs and medicine as well as lack of physical activities) of long-term unemployed youth. But Greece and Italy are exceptions, in that no negative effects on health due to unemployment were observed. Increased Risk of Social Exclusion In all six countries the health of those at increased risk of exclusion was considerably better than of youth at high risk of exclusion. Nevertheless, there was substantial variation within the category with respect to the prevalence of strain. Whether or not young people were supported by their parents had a major influence on their health. Young Italians in this situation show hardly any ill-health effects. The only strains reported are a lack of control of their situation, as well as a lack of structure in their everyday life. In the Greek study, the health of unemployed youth within this group was considered to be largely independent of their employment status. In the northern European studies, financial difficulties (Germany, Sweden) or financial insecurity (Belgium) were more important psychosocial stressors than in the southern countries. In all countries, protective factors can be found for youth at increased risk of exclusion, but the specific form they took differed between countries. In Greece, the relative absence of financial difficulties had an important buffer effect. The situation in Italy and Spain is similar: social support especially from friends, but also reduced vulnerability as a result of integration into the family. In northern European studies, there were additional protective factors to those of social integration and economic protection. In Germany and Sweden a higher level of qualification was important. In Belgium, Sweden and Germany the level of personal initiative young people could exercise played an important protective role. In summary, in the southern European studies economic protection (by parents or irregular work) as well as social integration were the crucial factors mediating the risks of strain, whereas in northern Europe, higher levels of qualification and opportunities for personal initiative were also of considerable importance. Low Risk of Social Exclusion In all countries youth in the group with a low risk of social exclusion were in considerably better health than those in groups at higher levels of risk. All countries reported cases that were relatively free from strain. The young people were described as being active, and having high self-esteem and good communication skills. No case of risky health-related behaviour is reported. In the three northern European countries, social support and financial protection were equally important protective factors. In Belgium and Germany, a high level of qualification also reduced risks of strain. Sweden was distinctive in that some young people with a low risk of social exclusion used their unemployment as a time for orientation. Similarly, in Belgium, unemployment was a deliberate choice for some. In 250
Psychology of Unemployment and Social Exclusion
both countries, people in these situations assessed the strain resulting from unemployment as rather low. In the southern European countries, it was above all economic security and the availability of social support that reduced vulnerability to strain.
Societal Variations in Vulnerability and Protective Factors Although the studies in all of the countries show high congruence with respect to the importance of school education and duration of unemployment and to some degree of social origin in determining exclusion risks (and hence vulnerability to mental ill-health), there are marked differences with respect to other factors. In particular there were substantial differences in types of vulnerability and protective factor between the northern and southern European countries. In Greece, Italy and Spain two factors played a much more crucial role in the prevention of social exclusion: the family and the submerged economy. The Role of the Family Both in the northern and southern European countries work plays a central role in the developmental process of young people. The denial of access to the labour market implies above all financial limitations. In Belgium, Germany, Spain and Sweden there is usually access to some type of benefit system. In contrast, in southern European countries young people who have never entered the labour market are in most cases totally excluded from any type of state support. This means that the family has to take on the responsibility of economically supporting the children. Dependence on the family, which partly takes the form of living with the parents, has an important protective role in that it increases the level of social support. Problems of social isolation were much rarer in the southern countries and this was reflected in higher levels of psychological wellbeing. However, forcing young people to be dependent on the parental family also prolongs the juvenile phase of the life cycle. Although family support is an important buffer, this protection may also hinder the development of young people to independent adulthood. This risk is much more important in southern European countries, given that young people are rarely eligible for transfer payments. At the same time, the crucial role of the family means that the welfare of the young unemployed is more heavily conditioned by their milieu of origin. In the southern European country studies, poverty and other social problems in the family of origin play a decisive role in accentuating financial dependency and lack of social support among young unemployed people. While in all studies, a low level of social support contributes to an increased risk of social exclusion, in the Italian, Greek and Spanish studies the emphasis lies on lack of support by the family, whereas in the Belgian,
251
Unemployment and Health ■ Interventions
German and Swedish studies, it is the lack of, or decrease in, local social contacts that is of central importance. The Submerged Economy and Social Exclusion The research carried out in the six countries revealed that the submerged economy plays only a limited role in the lives of long-term unemployed young people in the northern European countries (Borghi & Kieselbach, 2002, 2003). In some cases, resorting to irregular employment is seen as justified as the only answer to severe economic difficulty, and sometimes as a preventative measure to ensure that the situation does not precipitate criminal activity. The submerged economy is also a means to give people a certain financial independence and allows them to maintain their standard of living at an acceptable level. The situation in the three southern European countries is, however, very different. National estimates have pointed to a substantial informal sector in the economies of these countries. Our field research confirms that work in the informal sector is also a much more important part of the experiences of the unemployed in the southern European countries. Thirty-six cases (24%) of involvement in the submerged economy were discovered among the 150 interviewees from central and northern Europe, many of which were only sporadic and of little importance; in southern Europe, on the other hand, 121 cases of long-term unemployment (81%) with experience of irregular work (38 in Spain, 41 in Greece and 42 in Italy) were discovered (out of a total of 150 people interviewed in the three countries), albeit of different kinds and with differing results. The form of involvement of the unemployed in such work, however, could take very different forms. Broadly speaking, it is possible to distinguish between the following patterns: • Permanent irregular employment: this category includes those jobs that take up a significant part of the individual’s time during the course of the day and the week; these are real, permanent jobs of a full-time nature lacking any kind of social security and welfare protection or trade union rights. • Irregular seasonal employment: this category of job is characterised by an intense period of employment (often with longer-than-normal working hours) for a limited period of time. • Casual irregular employment: this category includes all those jobs of a strictly casual nature, often unconnected jobs that last for a very short period and are in no way associated with the education or training of the unemployed. It is clear that the greater presence of irregular employment in the southern European countries reflects the need to soften the financial burden of unemployment on those affected. In those countries where the unemployed receive very little or no financial support at all, there is clearly a much 252
Psychology of Unemployment and Social Exclusion
greater incentive for involvement in the submerged economy (Leonard, 1998). Nevertheless, the economic dimension is not the only one with which to measure the consequences of irregular employment. The studies conducted in Greece, Italy and Spain highlight the ambivalent role played by the involvement of young unemployed people in the submerged economy with respect to the risk of social exclusion. Income earned through irregular work, despite being low, often discontinuous and rarely sufficient for a person to be financially independent, does nonetheless constitute a temporary solution, often a kind of ‘waiting wage’, while seeking a regular employment contract. It partially satisfies the individual’s financial requirements. Moreover, in some local contexts, these ‘shadow’ jobs may constitute the only available alternative to unemployment, given the unsatisfactory nature of employment measures and the paucity of chances to create regular employment, even short-term, low-paid work and work suitable for first-time workers. However, while softening the phenomenon of economic exclusion, irregular employment reveals a more problematic side, especially where it constitutes relatively full-time employment. The studies conducted in the three southern countries show that it often threatens to subtract valuable resources (time, motivation, etc.) from the search for regular work. In the case of those jobs that do not reflect a person’s professional training, education or job aspirations, it can often lead the individual into a professional ‘blind alley’. The studies also highlight the ambivalent nature of irregular employment with regard to the risk of sociocultural exclusion. On the one hand, there is the opportunity to develop new social contacts through work, regardless of its irregular nature, and thus to develop a relational network that enables the individual to escape his or her restrictive family circle and to broaden social experiences that otherwise would be almost impossible to develop within the family. Such working experiences contribute, albeit in a limited way, to a process whereby life skills are further developed. On the other hand, the experiences reported during the interviews show that irregular employment can also be associated with exploitation, the denial of basic rights, and the creation of humiliating, arbitrary relations, which can accentuate sociocultural exclusion. Even in less extreme situations, the fact that such activity often takes place within the family sphere, or within a social milieu similar to the individual’s, and consists of simple, uninteresting tasks, considerably restricts its potential with respect to the development and widening of life skills. With respect to the risk of social isolation, if the expansion of the submerged economy does not lead to criticism of the individuals directly involved (and this does not seem to emerge strongly even in northern European countries, such as Germany, where such a phenomenon is much less frequent), then it allows for a partial degree of social inclusion, a reduced form of citizenship for these unemployed young people, the individual and 253
Unemployment and Health ■ Interventions
social price of which ought to be seen not so much in relation to the present as to the future. We can identify, then, two contrasting scenarios of the impact of irregular employment on the social exclusion of long-term unemployed young people. In the first, irregular employment contributes towards increasing the complexity of an unemployed young person’s social relations and broadening (or keeping) the options available to that person. By softening the economic hardship created by unemployment, irregular work enables an individual to persist in their search for regular employment while not binding them in an exclusive manner: it may even render this search more effective through the social contacts created by irregular employment. In some cases irregular employment may increase an individual’s professional and social skills through the widening of their range of experiences. In the second scenario, irregular work contributes towards reducing this complexity — that is, the range of accessible alternatives — in an almost irreversible manner. It makes sociocultural deficiencies (as well as economic difficulties) increasingly chronic, through a combination of poor quality work, restricted social contacts, and limited provision of cognitive and social resources due to the family and social setting of the work. While in the first case it is clear that the submerged economy can be seen as a form of protection from the risk of social exclusion, in the second it leads towards marginalisation and to increased exposure to the risk of permanent exclusion. In general, this second case emerges in situations where irregular work — taken up as the only available possibility for a young unemployed person in a deprived area — is characterised by jobs of extremely poor quality not only in economic terms, but also with respect to the social and cognitive opportunities they offer. In such a situation, irregular work reinforces (instead of limiting) the disadvantages of the milieu of origin. Hence, depending on the quality of the jobs taken, such involvement can act either as a buffer or as a trap for unemployed young people.
Conclusions Unemployment is a central risk factor for young people, which in the longterm threatens the overall integration of young people into society. This chapter has discussed the potential contribution of unemployment research based on the concept of social exclusion compared to previous perspectives focusing on individual outcomes, analysing stigmatisation effects or focusing on the role of activities as moderators of the unemployment experience. The question arises of whether the concept of social exclusion enables us to conceptualise the individual and social impact of unemployment in a way that significantly extends the present knowledge (Kieselbach & Beelmann, 2003). The dominant approaches of effect-related unemployment research (Wirkungsforschung) try to systematise the unemployment and health relationship and identify a number of moderators (Kieselbach, 254
Psychology of Unemployment and Social Exclusion
1991; Warr, 1987). They deploy a conceptual and methodological approach in which effects tend to remain purely additive. The social exclusion approach, focusing on the key mechanisms increasing or counteracting the process of marginalisation, adds a more structured as well as a more dynamic perspective to social–psychological unemployment research. It offers a more complex and systematic explanation of the effects of unemployment, by moving away from a set of isolated variables and bridging the individual and the social effects of unemployment in a way that is indispensable for a social psychology of unemployment. Social exclusion processes can be considered complex moderators of the health effects of unemployment. The detection of key mechanisms that increase vulnerability to social exclusion will help to identify crucial turning points of the individual biography, susceptible phases where social support will be needed to help people in coping adequately with their life situation. This knowledge is a necessary prerequisite for developing intervention schemes intended to limit the detrimental effects of joblessness. Our research confirmed many of the specific factors contributing to exclusion risks identified in previous research. The most important vulnerability factors that contribute to an increase of the risk of social exclusion for young unemployed people in the long-term are in all countries low qualification, passivity at the labour market, a precarious financial situation, low or missing social support and insufficient or nonexistent institutional support. The most important protective factor for unemployed youth is social support. However, the research also revealed a factor that has been rarely emphasised in earlier psychological studies: the importance of social origin. Social origin can be a protective factor for the youth as well as a decisive vulnerability factor: poverty and other social problems in the family can increase the risk of social exclusion for the youth. This can be interpreted in the sense that the effects of social origin are reinforced by the experience of long-term unemployment of young people. Most crucially, our research showed that it is the specific combination of these vulnerability and protective factors that determines the risk of poor mental health. Moreover, these combinations have been found to vary in a structured way between societies as a result of different institutional frameworks and different cultural traditions, with major implications for the individual experience of unemployment. This is most evident in the striking contrasts between northern and southern European countries. The lack of state welfare provision in the southern countries leads to a much more central role for the family in supporting unemployed people. While integration into social networks is of great importance for youth from northern Europe, in southern Europe the family is more important. Especially due to the high level of family support, the number of youth at high risk of social exclusion in general is lower in southern Europe compared to northern Europe. Increasing individualisation processes in southern European countries, 255
Unemployment and Health ■ Interventions
however, may weaken this buffer effect of family support in the future. Given the very low level of institutional support in these societies, this would create a major need for social policy reform. Another central result of our study is the crucial role of the submerged economy in the southern countries for the psychosocial wellbeing of young people without formal employment. The higher involvement of young people in the southern European countries in irregular work acts at the same time as a psychosocial buffer, moderating the risk of social exclusion, and as a barrier to reentering the primary labour market thus entrapping those involved. This points to the need to develop bridges out of irregular work into the labour market that may reduce the stigmatising impact of the submerged economy. Overall, the normalisation of youth unemployment, the prolongation of the youth period (despite its negative consequences for the maturation needs of young people), and the wide acceptance of the submerged economy strongly influence individual experience of unemployment. In the southern European countries these factors moderate tendencies of blaming oneself, considerably lower the risk of social exclusion and concomitantly reduce illhealth effects associated with long-term unemployment and social exclusion. The necessity to develop more inclusive policies in the EU as it is outlined in many programmatic documents of the European Commission CEC, most specifically in the Employment Guidelines (European Commission, 2001), is dependent on the thorough analysis of the key mechanisms linking adverse life events to the integration or disintegration into society. Employment and the absence of employment must be considered as central factors determining poverty, social exclusion and social integration.
Endnotes 1
2
3
4
256
The consortium of the research project has been coordinated by the University of Bremen (Germany): Prof. Dr Thomas Kieselbach (the German team: Gert Beelmann, Andrea Stitzel, Ute Traiser) and included the following national partners: University of Gent (Belgium): Prof. Dr Kees van Heeringen, Wouter Vanderplasschen, Tine Willems, Gwendolyn Portzky; University of Bologna (Italy): Prof. Dr Michele La Rosa, Dr Vando Borghi, Federico Chicchi, Roberto Rizza; Autonomous University of Barcelona (Spain): Prof. Dr Louis Lemkow, Dr Josep Espluga, Josep Baltierrez; Greek Network of Health Promoting Schools and Institute for Child Health, Athens: Katerina Sokou, Demetra Bayetakou, Valentine Papantoniou, Katerina Christofi; University of Karlstad (Sweden): Prof. Dr Bengt Starrin, Erik Forsberg, Marina Kalander Blomqvist, Dr Ulla Rantakeisu. This selection criterion could not be achieved from all countries. In four national studies (Greece, Italy, Sweden and Belgium), people 25 years of age have also been included. In the Swedish study 49 instead of 50 are included in the analysis: one individual dropped out at a fairly late stage of the analysis when closer inspection showed that a central selection criterion was not filled. The rank order of national studies is given according to the weighted types (high risk = 3, increased risk = 2, low risk = 1).
Psychology of Unemployment and Social Exclusion
References Bakke, E.W. (1940). Citizens without work: A study of the effects of unemployment upon the workers’ social relations and practices. New Haven, CT: Yale University Press. Biermann, I., Schmerl, C., & Ziebell, L. (1985). Leben mit kurzfristigem Denken: Eine Untersuchung zur Situation arbeitsloser Akademikerinnen [Life with short-term thinking: A study on unemployed female academics]. Weinheim, Germany: Beltz. Borghi, V., & Kieselbach, T. (2002). Disoccupazione giovanile e lavoro irregolare: Nord e Sud Europa a confronto. In V. Borghi (Ed.), Vulnerabilita, inclusione sociale e lavoro: Contributi per la comprensione dei processi di esclusione sociale e delle problematiche di policy (pp. 175–186). Milano, Italy: FrancoAngeli. Borghi, V., & Kieselbach, T. (2003, May). The submerged economy as a trap and a buffer: Comparative evidence on long-term youth unemployment and the risk of social exclusion in Southern and Northern Europe. Paper presented at the EUWorkshop ‘Informal/Undeclared Work’, European Commission (DG Research), Brussels, Belgium. European Commission. (2001). The employment guidelines 2002. Brussels, Belgium: CEC. Eurostat. (1998). Basic statistics of the community. Brussels, Belgium: Author. Fryer, D., & Payne, R.L. (1984). Proactive behaviour in unemployment: Findings and implications. Leisure Studies, 3, 273–295. Harrison, R. (1976). The demoralizing experience of prolonged unemployment. Department of Employment Gazette, April, 339–348. Jahoda, M. (1979). The impact of unemployment in the 1930’s and the 1970’s. Bulletin of the British Psychological Society, 32, 309–314. Jahoda, M. (1982). Employment and unemployment. Cambridge, UK: Cambridge University Press. Jahoda, M., Lazarsfeld, P.F., & Zeisel, H. (2002). Marienthal: The sociography of an unemployed community (1933). With a new introduction by Christian Fleck. New Brunswick, NJ: Transaction. Jahoda, M., & Rush, H. (1980). Work, employment, and unemployment (Occasional Paper Series, No. 12). Brighton, UK: SPRU, University of Sussex. Jahoda, M. (1982). Employment and unemployment. Cambridge, UK: Cambridge University Press. Kieselbach, T. (1987). Self-disclosure and help-seeking as determinants of vulnerability: Case studies of unemployed from social–psychiatric services and demands for health and social policy. In D. Schwefel, P.-G. Svensson, & H.F.K. Zöllner (Eds.), Unemployment, social vulnerability, and health in Europe (pp. 281–303). New York: Springer. Kieselbach, T. (1991). Unemployment. In R. Lerner, J. Brooks-Gunn, & A.C. Petersen (Eds.), Encyclopedia of adolescence (pp. 1187–1201). Philadelphia, PA: Garland Publisher. Kieselbach, T. (1997). Individuelle und gesellschaftliche Bewältigung von Arbeitslosigkeit: Perspektiven eines zukünftigen Umganges mit beruflichen Transitionen [Individual and societal coping with unemployment: perspectives of a future coping with transitions in the working life]. In H. Holzhüter, R. Hickel, & T. Kieselbach (Eds.), Arbeit und Arbeitslosigkeit: Die gesellschaftliche Herausforderung unserer Zeit [Employment and unemployment: The societal challenge of our time] (S. 39–64). Bremen, Germany: Kooperation Universität, Arbeiterkammer-Bremen. Kieselbach, T., & Beelmann, G. (2003). Arbeitslosigkeit als Risiko sozialer Ausgrenzung bei Jugendlichen in Europa [Unemployment as a risk for social exclusion among young people in Europe]. Aus Politik und Zeitgeschichte: Beilage zur Wochenzeitung, Das Parlament (B 6–7/2003) v. 3.2.2003. 257
Unemployment and Health ■ Interventions Kieselbach, T., Heeringen, K. van, Lemkow, L., Sokou, K., & Starrin, B. (Eds.). (2001). Living on the edge: A comparative study on long-term youth unemployment and social exclusion in Europe (YUSEDER publications, no. 3, Psychology of Social Inequality; Vol. 11). Opladen, Germany: Leske + Budrich. Kieselbach, T. (Ed.) with K. van Heeringen, M. La Rosa, L. Lemkow, K. Sokou, & B. Starrin. (2000a). Youth unemployment and health: A comparison of six european countries. Psychology of social inequality (Vol. 9). Opladen, Germany: Leske + Budrich. Kieselbach, T. (Ed.) with K. van Heeringen, M. La Rosa, L. Lemkow, K. Sokou, & B. Starrin (2000b). Youth unemployment and social exclusion: A comparison of six European countries (Psychology of Social Inequality; Vol. 10). Opladen, Germany: Leske + Budrich. Kilpatrick, R., & Trew, K. (1985). Life-styles and psychological well-being among unemployed men in Northern Ireland. Journal of Occupational Psychology, 58, 207–216. Kelvin, P., & Jarrett, J.E. (1985). Unemployment: Its social psychological effects. Cambridge, MA: University Press. Kronauer, M. (1998). ‘Social exclusion’ and ‘underclass’: New concepts for the analysis of poverty. In H.-J. Andre (Ed.), Empirical poverty research in a comparative perspective (pp. 51–75). Aldershot, UK: Ashgate. Kronauer, M. (2002). Exklusion: Die Gefährdung des Sozialen im hoch entwickelten Kapitalismus [Exclusion: The precarisation of the social in advanced capitalism]. Frankfurt, Germany: Campus. Leonard, M. (1998). Invisible work, invisible workers: The informal economy in Europe and in the US. London: MacMillan. Lerner, M.J. (1974). Social psychology of justice and interpersonal attraction. In T.L. Huston (Ed.), Foundations of interpersonal attraction (pp. 331–351). New York: Academic Press. Lerner, M.J. (1980). The belief in a just world: A fundamental delusion. New York: Plenum. Paugam, S. (2000). Le salarié de la précarité. Paris: Presses universitaires de France. Silver, H. (1994). Social exclusion and social solidarity: Three paradigms. International Labour Review, 133, 531–578. Vogel, B. (2000). Am Rande der Arbeitsgesellschaft [On the edge of the work society]. Verhaltenstherapie & Psychosoziale Praxis, 32(3), 359–366. Wacker, A. (1983). Differentielle Verarbeitungsformen von Arbeitslosigkeit: Anmerkungen zur aktuellen Diskussion in der Arbeitslosenforschung [Differential coping with unemployment. Annotations to the actual debate in unemployment research]. Prokla, 53, 77–88. Warr, P.B. (1987). Work, unemployment, and mental health. London: Oxford University Press. Warr, P.B. (1989). Individual and community adaptation to unemployment. In Starrin, B., Svensson, P.-G., & Wintersberger, H. (Eds.), Unemployment, poverty, and quality of working life: Some European experiences (pp. 27–44). Berlin, Germany: Edition Sigma. Witzel, A. (1987). Das problemzentrierte Inerview. [The problem-focused interview]. In G. Jüttemann (Ed.), Qualitative Forschung in der Psychologie [Qualitative research in psychology] (pp. 227–256). Heidelberg, Germany: Asanger. Witzel, A. (1995). Auswertung problemzentrierter Interviews: Grundlagen und Erfahrungen [Analysis of problem-focused interviews: Fundamental principles and experiences]. In R. Strobl & A. Böttger (Eds.), Wahre Geschichten? Zu Theorie und Praxis qualitativer Interviews [True stories? Theory and practice of qualitative interviews] (pp. 49–76). Baden-Baden, Germany: Nomos. 258
Chapter
19
Can Volunteering Be a Moderator of the Detrimental Effects of Engagement in (Un)Employment? Jacques C. Metzer
It is well known that unemployment is likely to have deleterious effects on people. Only slightly less well known is the fact that employment can also produce deleterious effects through the phenomenon of work strain (e.g., Karasek & Theorell, 1990), leading Japanese medical researchers to coin two new words: ‘karojisatsu’, meaning ‘ill health due to overwork’, and ‘karoshi’, meaning ‘suicidal death due to overwork’. The converse of the second statement above, namely that employment can also produce positive effects on people, appears to be true (e.g., Levi, 2001). It is not at all surprising that the converse of the opening statement above, that unemployment can also produce positive effects on people, is considered widely to be false. Questions of interest arise regarding the nature of activities engaged in during employment, and resources made available as a result of it. Does unemployment produce the opposite of these? Are buffers possible against the undoubted deleterious effects that can be produced by both employment and unemployment? For a discussion relating to these questions, we should first turn to the nature of work.
Work: Paid and Unpaid Employment is a relatively recent term introduced during or after the Industrial Revolution in Europe. An employer pays a salary or wage to an employee in return for activities completed. The activities engaged in produce obvious economic benefit or gain for the employer. Remuneration, conditions of employment, contractual and legal ramifications and other many standardisations, including hours of work required, long service leave, holidays and sick leave make it appear that employment is a unique 259
Unemployment and Health ■ Interventions
form of human activity. Under this conception, the shutting out of people from this activity through (forced) unemployment is associated with, inter alia, feelings of loss, heightened strain and lowered self-esteem, not to mention the loss of money and feelings of social and economic exclusion. Fryer and Winefield (1998) consider unemployment to be similar to highly stressful employment. The conclusion drawn by most authors is that as forced unemployment is a condition preventing access to the activity known as employment, access is also denied to the positive aspects of employment. Levi (2001) argued that employment for an individual is a source of identity, friends and meaning in life as well as money, with every person deserving access to these sources and therefore access to employment. I would like to take up two points. First, the concept of employment should not be confused with the concept of work. Employment may be understood as paid work in the service of an employer or organisation. Work itself may be understood as a wider category of sustained and purposeful human activity, but not necessarily rewarded by money (only). In this conception of work, volunteering, education and training, caring for dependants and even recreational activities may be regarded as work that is not rewarded directly with a pay packet. In a discussion paper of some breadth of vision, Winefield and his colleagues (Winefield et al., 2002, p. 2) entertained a similar view: We clearly need some different ways of understanding and thinking about work. We are entering a millennium in which work may become a less central part of who and what people are. We need to accommodate better new cultural understanding of personhood and connections, and of meaning and self-fulfilment. The experience of other cultures allows us to broaden, redefine, and reconstrue (e.g., Davidson & Reser, 1996) the nature of work and its relation to life satisfaction and quality of life.
My second point is that if it can be shown that unpaid work may also lead to benefits such as identity, friends, meaning and satisfaction, then perhaps only money remains as the (obvious) difference between paid and unpaid work (Metzer, 1996). Whether the unemployed should be advised to engage in unpaid work may well depend on a demonstration that there are similar benefits to be obtained for the individual.
Unpaid Work Is unpaid work similar to paid work in that it can produce (job) strain and satisfaction? Can unpaid work produce positive or negative effects for the individual? I wish to take the special case of formal volunteering behaviour to address these questions. Formal volunteering is defined as providing a sustained service of one’s own free will to the community without receiving a monetary reward (Noble, 1997). Traditionally, it has been assumed that altruism is the major, if not the only, motivation for engaging in volunteering. However, on the 260
Volunteering as a Moderator of Employment
basis of a large census conducted in Australia by the Australian Bureau of Statistics (1996), Metzer (1996) concluded that the six categories of reasons given by people participating in the census for conducting volunteering could be classified into intrinsically and extrinsically motivated behaviours, of approximately equal distributions in terms of reported frequencies of occurrence. The Volunteer Function Inventory (VFI) was developed by Clary et al. (1998) to measure underlying differences in volunteer motivation. Six subfactors emerged from their work: values, understanding, enhancement, defensive, social adjustment and career advancement. Although Clary et al. (1998) demonstrated reliable differences for different groups of volunteers, questions remain about the reliability of the VFI in predicting long-term volunteering.
Development of Scales to Measure Stress and Satisfaction in Volunteers Metzer (2003) reported results of an earlier investigation in 1999 that involved the creation of a scale to measure both stress (strain) and satisfaction in volunteering work of different natures in 18 organisations that engage volunteers in the state of South Australia. While the scale was subsequently verified and checked for adequate reliability in a variety of other studies (Robertson, 2000; Schmidtke, 2000), there has been little other recent work reported in the literature on the determination of stress and satisfaction in volunteer populations. The Volunteer Stress Scale consists of four items characterised by experiencing anxiety in relation to volunteering, finding difficulties in relaxing, and negative health and stress consequences of volunteering. Cronbach alphas ranged from 0.60 to 0.80 in the above three studies. The Volunteer Satisfaction Scale contains six items, such as ‘The experience of volunteering has been a worthwhile one’ and ‘My volunteer work has been personally fulfilling’, and has received reported alphas of 0.80 to 0.89 in the three studies mentioned. Items on both scales were answered on a 7-point Likert scale. While the theoretical background of the scales’ development was Karasek and Theorell’s (1990) job demand control support (JDC-S) model, of interest to the discussion here is whether this model adequately predicts stress and satisfaction in volunteers. Metzer (2003) reported that three variables were found to predict (greater) satisfaction by volunteers. These were: (a) greater time and effort required for the volunteer work (a demand), (b) adequate organisational resources, good working relationships and communication between staff, adequate assistance and information provided to perform tasks, being provided with feedback on their work, adequate training and feeling in control of job tasks (controls) and (c) feelings of performing interesting and 261
Unemployment and Health ■ Interventions
important work, feeling appreciated and respected by the organisation and others, and feeling ‘connected’ to the organisation (supports). Similarly grouped three variables (demands, controls, supports) were also found to predict the degree of stress (strain) felt by volunteers, except that the controls and support variables worked in the opposite direction (i.e., lower control and lower support predicted higher strain). Time and effort, as well as other organisational demands, increased strain. Both Robertson (2000) and Schmidtke (2000) found very similar data in different volunteer groups. Of interest in all three studies is the predictive power of the 7-item Connectedness Scale (alphas of 0.88 to .90): it alone explained an average of 50% of the variance in the satisfaction experienced by volunteers in their work. Battaglia and Metzer (2000) replicated this finding in different groups of workforce-retired volunteers and found that satisfaction was explained by 55% of the variance in connectedness.
Similarities of Paid and Unpaid Work From the above analysis of data measuring strain and satisfaction, a surprising conclusion may be made: the determinants of levels of stress (strain) and satisfaction in paid and unpaid volunteering work appear to be very similar. While these findings provide only limited evidence of the similarity of a range of volunteering activities to paid work, the implication is that employment or paid work is not a unique human activity with its own particular characteristics, determinants and effects. From a somewhat different perspective, it is known that the mental health of the unemployed is worse than the mental health of employed people (e.g., Winefield, 1995), although contradictory evidence (e.g., Kieselbach & Svensson, 1988; Rodgers, 1991; Schaufeli & van Yperen, 1992) exists as to whether the financial effects of unemployment and their flow-on effects are the main cause. On a parallel track, it is also known that the physical and mental health of volunteers is superior to the health of nonvolunteers (Wilson & Musick, 1999; Young & Glasgow, 1998). From sophisticated reviews of a number of studies, these authors argue persuasively that the cause of better health is the activity of volunteering. It is very tempting to infer from these parallels that paid and unpaid work may cause similar benefits for individuals. Of course, little has been said about deleterious effects of both forms of work, but limited evidence (e.g., Metzer, 2003) suggests that job strain can be experienced by volunteers following a similar path, as predicted by Karasek and Theorell’s (1990) model.
262
Volunteering as a Moderator of Employment
Employee Volunteering A relatively new form of volunteering is known as employee volunteering, defined as a ‘commitment by a commercial organisation to encourage staff to volunteer in the not-for-profit sector’ (Volunteering Australia, 2001). For a company to be a good corporate citizen of this kind it is required to make a considerable investment in both monetary and human resources in order to contribute to the community in which it operates. The European Commission, in a green paper on employment and social affairs (2001), argued that corporate social responsibility by companies should be promoted because: it can be a positive contribution to the strategic goal … [of becoming] the most competitive and dynamic knowledge-based economy in the world, capable of sustainable economic growth with more and better jobs and greater social cohesion (p. 4).
Employee volunteering can contribute to business turnover and community benefits (Tuffrey, 2003), as well as employee benefits (Quirk, 1998). Current estimates are that around 90% of large US corporations and around 30% of UK corporations are engaged in the phenomenon. It is clear that the boundaries between paid and unpaid work are eroded by these policies and activities. The phenomenon of employee volunteering may be considered to be further evidence that the post-Industrial Revolution concept of employment is in need of revision, and expected to eventually surrender itself to history as yet another human activity.
Conclusions and Implications At this point, I could argue that intervention with unemployed people might be to encourage volunteering as a substitute for paid work, because of potentially beneficial individual flow-ons similar to paid work, except remuneration. But this will not be sufficient; there are larger forces that should be martialled to deal with the problem of unemployment. Wide recognition is required that: 1. Unemployment should not be able to prevent a full, purposeful and active lifestyle. 2. Individual meaning and purpose should not be dependent on employment in the narrow traditional sense, but on many human activities. 3. Human activities are complex and often socially interdependent. 4. Unemployment and a traditional view of employment and economic growth lead to a decrease in social capital (Cox, 1995; Putnam, 2002). 5. Increasing social capital improves the quality of individuals’ lives as well as community cohesion and economic wellbeing. 6. Work, whether paid or unpaid, should be nurtured, structured and rewarded not only for immediate economic gain, but also for its potential to build social capital, on the assumption that economic benefits are likely to follow a more cohesive and clever society. 263
Unemployment and Health ■ Interventions
7. Desirable for all people may be a work ‘mix’, in which both paid and unpaid worthwhile and productive activities in a broad sense are rewarded appropriately; the proportions of these activities will vary according to cultural and individual factors, and also variations across a single person’s life span should be expected and encouraged.
References Australian Bureau of Statistics. (1996). Voluntary work 1995 (Cat. 4441.0). Canberra, Australia: Australian Government Printing Service. Battaglia, A.M., & Metzer, J. (2000). Older adults and volunteering: A symbiotic association. Australian Journal on Volunteering, 5, 5–9. Clary, E.G., Snyder, M., Ridge, R., Copeland, J., Stukas, A., Haugen, J., et al. (1998). Understanding and assessing the motivation of volunteers: A functional approach. Journal of Personality and Social Psychology, 74, 1516–1530. Cox, E. (1995). A truly civil society. Sydney, Australia: Australian Broadcasting Corporation. Davidson, G., & Reser, J.P. (1996). Construing and constructs: Personal and cultural? In B.M. Walker, J. Costigan, L.L. Viney, & B. Warren (Eds.), Personal construct theory: A psychology for the future (pp. 105–128). Melbourne, Australia: Australian Psychological Society. European Commission. (2001). Promoting a European framework for corporate social responsibility: Green paper. Industrial relations and industrial change: Employment and social affairs. Luxembourg: Office for Official Publications of the European Communities. Fryer, D.M., & Winefield, A.H. (1998). Employment stress and unemployment distress as two varieties of Labour market induced psychological strain: An explanatory framework. Australian Journal of Social Research, 5, 3–18. Karasek, R.A., & Theorell, T. (1990). Healthy work: Stress, productivity and the reconstruction of working life. New York: Basic Books. Levi, L. (2001, December). Occupational transitions: Unemployment, underemployment and health. Address given at the 2nd Congress of the International Commission on Occupational Health: Scientific Committee Unemployment and Health, Adelaide, Australia. Kieselbach, T., & Svensson, P.G. (1998). Health and social policy responses to unemployment in Europe. Journal of Social Issues, 44, 173–191. Metzer, J.C. (1996) The psychology of volunteering: External or internal rewards? Australian Journal on Volunteering, 1, 20–24. Metzer, J.C. (2003). Volunteering work stress and satisfaction at the turn of the 21st century. In M.F. Dollard, A.H. Winefield, & H.R. Winefield (Eds.), Occupational stress in the service professions (pp. 389–407). London: Taylor & Francis. Noble, J. (1997). Volunteers and paid workers: A collaborative approach. Adelaide, Australia: The Volunteer Centre. Putnam, R.D. (2002). (Ed.). The dynamics of social capital. Oxford, UK: Oxford University Press. Quirk, D. (1998). Corporate volunteering: The potential and way forward. Wellington, New Zealand: Wellington Volunteer Centre. Robertson, M. (2000). Comparison of job strain and job satisfaction between volunteer and paid staff in nursing homes. Unpublished honours thesis, University of South Australia, Australia. 264
Volunteering as a Moderator of Employment Rodgers, B. (1991). Socio-economic status, employment and neurosis. Social Psychiatry and Psychiatric Epidemiology, 26, 104–114. Schaufeli, W.B., & van Yperen, N.W. (1992). Unemployment and psychological distress among graduates: A longitudinal study. Journal of Occupational and Organizational Psychology, 65, 291–305. Schmidtke, B. (2000). A test of the job demand, control, support and person–environment fit models of occupational stress in volunteers. Unpublished honours thesis, University of South Australia, Australia. Tuffrey, M. (2003). Good companies, better employees: How community involvement and good corporate citizenship can enhance employee morale, motivation, commitment and performance. London: The Corporate Citizenship Company. Volunteering Australia. (2001). Corporate volunteering (fact sheet). Canberra, Australia: Australian Department of Family and Community Services. Wilson, J., & Musick, M. (1999). The effects of volunteering on the volunteer. Law and Contemporary Problems, 62, 142–168. Winefield, A.H. (1995). Unemployment: Its psychological costs. In C.L. Cooper & I.T. Robertson (Eds.), International review of industrial and organizational psychology, 10 (pp. 169–212). London: Wiley. Winefield, A.H., Montgomery, B., Gault, U., Muller, J., O’Gorman, J., Reser, J., et al. (2002). The psychology of work and unemployment in Australia today: An Australian Psychological Society discussion paper. Australian Psychologist, 37, 1–9. Young, F.W., & Glasgow, N. (1998). Voluntary social participation and health. Research on Aging, 20, 339–362.
265
Unemployment and Health ■ Interventions
266
Chapter
20
Restructuring and Outplacement in the Netherlands Ellen Heuven, Arnold Bakker and Wilmar Schaufeli
The aim of the present chapter is to analyse which set of integrative, preventative and curative measures employers can take, and have taken, in the Netherlands to secure the wellbeing of dismissed workers, and to support their job search processes. Companies can follow three types of strategies to guide their workers in the current dynamic labour market. They can: 1. empower their employees with appropriate skills, training and working experience in order to secure their position on the labour market 2. treat their employees fairly, so as to reduce the potentially negative psychological consequences of dismissal, and thereby enhance the job search process in terms of speed and quality of reemployment 3. offer their redundant employees guidance and outplacement support, to assist them to cope adequately with their dismissal, both practically (i.e., training labour market skills) and psychologically (i.e., support in the grieving process). These three strategies, employability, justice and outplacement, are the core of our present research proposal. In order to comment on these three central issues, we conducted an extensive qualitative study in the Netherlands (see Appendix A). The project involved a general inventory, interviews with individuals in insecure jobs, individuals who were successfully reemployed following dismissal, case studies from restructuring and outplacement processes, and finally interviews with major stakeholders in these processes. We start by reviewing the labour market situation in the Netherlands, and its existing employment and reemployment interventions. Then we describe proactive and preventative measures that may reduce the psychological damage to employees of restructuring processes. The central issues are justice and employability. This part reflects content interviews with a group of ‘job-insecure’ people and a group of ‘successfully reemployed’ 267
Unemployment and Health ■ Interventions
people. This chapter also draws information from several cases of good practice and from interviews with stakeholders (e.g., human resource managers, labour unions, employers) who have been involved in restructuring and reintegration processes. All these have allowed us to draw some conclusions as to what constitutes good practice for conducting restructuring in a manner that reduces the psychological damage of these operations. Finally, we give a description of the most important findings on the specific topic of outplacement. This part aims to give a description of the Dutch outplacement market, and also describes the specific components and effects of outplacement.
The Economic and Public Context in the Netherlands Development of the Labour Market Between 1994 and 2000, the Netherlands experienced a period of strong economic growth, increased employment and decreased unemployment. This development resulted in a tight labour market, tending towards full employment. Disability became an increasingly important reason for dismissal, while dismissals for economic reasons and collective dismissals became less dominant. Companies continued restructuring during this period, but few employees had to face unemployment, as the possibilities for replacement and outplacement in such a tight labour market were abundant. Following the attacks in the United States on September 11, 2001, however, this situation changed dramatically. Most large companies, such as banks, airlines and information technology companies were faced with large restructuring operations and mass dismissals. However, it is important to note that most of the study that is referred to here was conducted in the context of a tight labour market and strong economic growth. A flexibilisation of the labour market has taken place in the Netherlands during the past decade. The flexibilisation of work that took place in the 1990s has become obvious in the growing number of flexible and part-time contracts, as well as increased internal mobility. Voluntary external mobility has declined over the past decade. In the Netherlands, the new worker is ‘job hopping’ within his or her own enterprise, but is not inclined to change companies often. Another important background variable is the famous Dutch Polder Model (Tjiong, 2005). Since the beginning of the 1980s, consultations between employers, employee representatives and governmental bodies have formed the basis for decision-making on labour market issues in the Netherlands. These so-called tripartite negotiations with respect to socioeconomic problems have become known as the ‘Polder Model’. The 1982 Agreement of Wassenaar, considered one of the pillars of this model, stipulates, among other things, that employees (represented by the unions and the works councils), and employers (represented by the company’s management and employers’ federations) are primarily responsible for the 268
Restructuring and Outplacement in the Netherlands
negotiations on collective labour agreements, restructuring, dismissals and social plans. The Dutch government provides for the legal framework of these negotiations. The Polder Model, which recommends frequent negotiations between unions and employers and that all involved parties strive for satisfactory compromise, has resulted in the development of a historical relationship of trust, in which both parties consider themselves to be partners rather than opponents. Under this model the social partners, comprising employers and employee representatives (unions and the works council), are responsible for collective labour agreements, for wage negotiations, and for jointly finding appropriate solutions for ameliorating the potentially negative social consequences of restructuring and dismissal. The favourable relationships between the social partners have allowed for the creation of a business climate that promotes corporate social responsibility, ethical entrepreneurship, and socially responsible practices in dealing with corporate reorganisation, redundancy and dismissals. However, important differences exist between large companies on the one hand, and small- and medium-sized enterprises on the other, in planning, managing and implementing reorganisations and dismissals. Large companies in the Netherlands generally establish clear scenarios, laid down in collective labour agreements or other regulations, for how to manage restructuring processes and its social implications for the staff involved. For small- and medium-sized companies these arrangements may be set out in a sectoral collective labour agreement, but more often are not defined. Over the past three decades there has been a clear shift in the negotiations between the social partners with regard to securing the position of employees on the labour market. During the 1970s and the beginning of the 1980s, securing lifetime employment within the same company was the main aim of labour unions. In the case of dismissals, the so-called first generation social plans aimed to provide financial compensation for job loss. Starting in the 1980s, a new philosophy and strategy became common in the negotiations between the social partners. This philosophy, incorporated in the second generation social plans, was aimed at using part of the financial compensation for guiding employees in finding another job. Psychological guidance, training of labour market skills and active job search were some of the elements offered by internal mobility centres, temporary employment services and outplacement agencies to achieve this aim of avoiding unemployment. Finally, during the 1990s, a new development was set in motion, in which both companies and unions started to promote a more proactive approach towards restructuring and dismissals. The concept of lifetime employability, which was aimed at facilitating and softening the process of occupational transitions, gained importance in the collective bargaining agreements between the social partners. Although this concept needs further development, fine-tuning and implementation, the
269
Unemployment and Health ■ Interventions
social partners have embraced employability as the most important means to secure the labour market position of Dutch employees.
Public Interventions The so-called SUWI (Structure Implementation Work and Income) Act of January, 2002 has brought about important institutional and conceptual changes in the social security system in the Netherlands. The major focus of this law is to promote work above income. However, all respondents stated that the shift from state care towards corporate and individual responsibility is not easy. In the Netherlands the idea of the caring state is still predominant among both employers and employees. In the last decade, the Dutch government has initiated major policy measures, aimed at a more efficient reintegration of unemployed and reduced public intervention. Active labour market intervention and preventative measures have received more attention while the spending on income maintenance programs has decreased. An important measure in this respect is the ‘New Deal’ policy (‘Sluitende Aanpak’, in Dutch), in which each unemployed person is offered training, a work experience project or a job within 12 months of commencing unemployment. This policy, targeting the newly unemployed, aims to prevent long-term unemployment by means of early intervention. The privatisation of the reintegration activities of the Public Employment Service (PES) and the reorganisation of the social security system are other important developments towards further efficient reintegration of unemployed individuals in the Netherlands. Finally, the Dutch government aims to encourage the reintegration of the unemployed by tightening the criteria and sanctions for people receiving unemployment insurance benefits.
Restructuring in the Netherlands The Legal Framework The Netherlands has a system of dismissals that is unique in Europe. The so-called ‘dual dismissal’ system incorporates two different paths for employers to lay off staff. The first involves the ‘quick’ procedure by the cantonal (regional) judge, which brings about expenses for the employer. The cantonal judge decides on the legitimacy and necessity of the dismissal request, as well as on the financial and social compensation that should be offered to the dismissed employee. The second path is the so-called preventative dismissal test by the PES, which is free of charge for employers. In this test, the PES evaluates the necessity for dismissal and the just application of selection criteria. Recently, several changes have been proposed in this system in order to increase the efficiency, uniformity and speed of dismissal procedures. Within the framework of the Dutch Polder Model, the ministerial intervention in reorganisation and dismissal is kept to a minimum, and is left to 270
Restructuring and Outplacement in the Netherlands
the negotiations between the social partners. The state does, however, provide for an elaborate legal framework in which these negotiations take place, and which is aimed at securing the position of especially vulnerable employees and employee representatives. The major aims of the governmental dismissal policy are to protect vulnerable groups on the labour market against illegitimate lay-off, to reduce the access to the social security system by stimulating replacement and outplacement of redundant employees, to ensure that employers provide for justified grounds for dismissing employees, and that objective selection criteria are being applied properly. Furthermore, interactive justice is anchored in Dutch legislation. In the case of collective dismissals, the Dutch legislation provides for regulations that oblige employers to involve and inform employee representatives (unions and works council) in a timely fashion. Within this legal framework social partners have anchored principles with regard to conducting reorganisations and dealing with their personnel consequences in a socially responsible manner. These principles are laid down in collective agreements, social statutes or other sorts of formal arrangements. The collective labour agreements are increasingly aimed at creating possibilities, rights and instruments to assist employees in taking responsibility for their own careers by offering instruments for developing employability. In addition, in the event of a company reorganisation, the trade unions, the works council and the employer usually come to an agreement, laid down in a social plan. This scheme regulates the aspects of a reorganisation (or possibly total closure of the enterprise) relative to labour relations, including the decision on which particular employees are considered to be a surplus to requirements, the actions to be taken in finding them reemployment elsewhere, and the financial compensation that is to be provided to those employees who are actually dismissed. The legal status of a redundancy program varies: it may constitute a unilateral undertaking by the individual employer to the union or works council, or it may form part of the provisions of the relevant collective agreement. No statutory obligation to provide a redundancy program exists in the Netherlands. In the Netherlands, the debate with regard to legislation focuses mainly on possible flexibilisation and simplification of dismissal procedures. Legislation obliges employers to make carefully considered decisions with regard to dismissing employees, which may sometimes be awkward from a managerial perspective, but which are important for preventing decisionmaking based on panic, and for protecting (vulnerable) employees. Some parties, such as employers (and their associations) claim, however, that legislation with regard to restructuring and dismissals should become much more flexible in order to allow companies and employers to respond quickly and efficiently to the current dynamic business environment. For example, companies find that legislation sometimes obstructs them from responding adequately and quickly to situations of crisis, during which 271
Unemployment and Health ■ Interventions
bureaucratic dismissal procedures may endanger the survival of the enterprise. Therefore, employers’ representatives are strongly opposed to any additional European legislation, guidelines or codes of conduct, but rather are in favour of leaving initiatives to the social partners in order to find tailormade solutions for each particular restructuring situation. The unions agree that the current legislation offers a solid and elaborate framework for securing companies’ social responsibility, and that additional European legislation would have no added value in the Netherlands. Only the legal involvement of employee representatives in decision-making over the social consequences of restructuring should be arranged at a European level.
Justice Considerations In considering justice aspects within restructuring and dismissal, three main components are distinguishable: 1. Distributive justice: the perceived fairness of decision outcomes, such as amount of financial compensation. 2. Procedural justice: the structural characteristics of decision-making, including formal policies and procedures. In the context of lay-off decision-making these may include selection criteria. 3. Interactional justice: the interpersonal aspects of procedural justice that refer to the type of interpersonal treatment people receive throughout the lay-off process, including the adequacy with which lay-off decisions are explained and the levels of respect and dignity in communicating with lay-off victims. According to instrumental explanatory models, fair treatment symbolises that lay-off victims receive their share of tangible, concrete outcomes over time. According to the group-value model, fair treatment symbolises respectful and dignified treatment, thereby enhancing lay-off victims’ feelings of selfworth. Several studies have shown that more intangible aspects, such as communication of lay-off decision-making, and a respectful and dignified treatment by the employer, influence victims’ reactions to lay-off in a positive manner, while the amount of severance benefit does not explain a unique variance in their reactions. These findings accord with the core results of our SOCOSE (Social Convey and Sustainable Employability) study, namely that procedural justice, and more specifically, an open, honest, timely and human communication, is of crucial importance to employees in restructuring processes. Good and elaborate communication proves to be crucial both for reducing the stress of job insecurity and dismissal, and for allowing redundant employees to focus more easily on future perspectives. Although the formal, legal and material part of reorganisations seems to be well developed in the Netherlands, the aspect of communication still needs considerable improvement. Managers tend to hide behind formalities and tangible aspects of restructuring in order to avoid emotional confrontations with employees. 272
Restructuring and Outplacement in the Netherlands
In general, the major actors involved in restructuring and dismissals agree that the stakeholder model, in which the interests of all relevant parties are considered, should be adopted rather than the shareholder model. The employers’ associations, on the other hand, promote corporate social responsibility from the perspective of the economic self-interest of the company. That is, fair treatment of employees may reduce legal costs from claims, facilitate implementation of decision-making, and help in retaining a favourable corporate reputation and motivating survivors. The unions in the Netherlands claim that the growing emphasis on shareholder value and stock exchange performance is an important threat to the successful Polder Model. Distributive Justice In the Netherlands, proactive, preventative and curative measures with regard to (collective) dismissals are well observed by companies, on the whole. That is, redundant employees are usually offered financial compensation, training and outplacement or replacement measures to deal adequately with restructuring and dismissal. Financial arrangements should provide security in the transfer to another job, and compensate for practical and psychological costs of the reorganisation. Also, job support of redundant workers by an outplacement agency is common practice in the Netherlands. Offering outplacement is considered a fair and socially responsible way for the former employer to show concern for its dismissed employees. Most interviewees state that support in finding another job should be limited in time, varying between 6 months and 2 years. The arrangements laid down in the social plan generally include additional measures for vulnerable groups such as older workers. In our study, employees aged between 45 and 55 proved a particularly vulnerable group in the context of reorganisations. In general, this group cannot benefit from favourable arrangements for older workers included in the social plan, but have limited chances of easily finding another job. Finally, with regard to proactive measures, the majority of the respondents in our sample were actively offered training, job rotation and career management advice in order to increase their employability. Procedural Justice The majority of the respondents in our study claimed that the settlement of their dismissal was correct but that the process leading to their dismissal was unfair. That is, they experienced high distributive justice and low procedural justice, which clearly had a negative effect on their total experience of the restructuring and dismissal process. Respondents were shown to be more positive at the beginning of the outplacement if their interests had been clearly considered by their former employer, and if they had been given the opportunity to exert influence over the dismissal decision or its personal consequences. Respondents were generally more positive about 273
Unemployment and Health ■ Interventions
the decision if they were offered, for example, the possibility of voluntary repatriation, internal replacement or future job description. Respondents emphasise the importance of actual, real influence as opposed to participation of employees as a mere formality. Sufficient and proper involvement of employees (and/or their representatives) is also an important issue in the negotiations between the social partners. Employers, unions and works councils need to find the right balance between sufficient pace in procedures and clarification of concrete measures on the one hand, and creating support among employees by involving them in decision-making throughout the whole process on the other. Extensive involvement of employee representatives may, in some cases, result in long periods of insecurity about actual outcomes. The strategy followed by unions depends on the specific situation, the nature of the problem and the type of company that is involved. There is a delicate balance between sufficiently involving employees, and thereby creating a much-needed social basis for the restructuring on the one hand, and managing the restructuring in a limited time period so as to reduce insecurity on the other. Clear, transparent rules and procedures, and a proper application of these regulations supervised by an independent commission, prove to be another important aspect of restructuring. In the Netherlands, there is much discussion among the social partners and in the media with regard to objective procedures for selecting leavers and stayers. The so-called LIFO (last-in-first-out) principle generally needs to be observed. While some parties state that these objective criteria (i.e., seniority) are important for protecting vulnerable groups and for explaining decision-making to the workforce, other respondents say that these criteria obstruct a healthy and high-qualitative continuation of the company. Interactional Justice The majority of the insecure and successfully reemployed interviewees underline the importance of communication as the most decisive aspect of justice, crucial to their perception of the restructuring process. Open, honest, clear, fast and personal information proved to be important both for reducing employee distress associated with the dismissal and for helping employees to focus more easily on future career perspectives. Companies in which management showed concern for the emotional shock provoked by the dismissal, and communicated personally with the respondents, both about the reasons for the reorganisation and about the most advantageous way for the employee to leave the company, proved to provoke less frustration and anger in them at the beginning of their outplacement. The exact reasons for decisions with such far-reaching consequences as dismissals should be thoroughly explained to all individuals involved. Early announcement and a continuous flow of information throughout the whole process of restructuring is another essential responsibility of the company. Individuals should be offered a time274
Restructuring and Outplacement in the Netherlands
out period to cope with the shock of their dismissal, and sufficient time to search for another job. Furthermore, companies should inform their employees as quickly as possible and in a personal, face-to-face manner about the consequences of restructuring to each individual. Respondents also indicate the importance of the former employer expressing appreciation in order for the individual involved to focus on the future with sufficient self-belief and self-efficacy. Large companies, in particular, with an extended history of reorganisations often showed a lack of personal, humane communication vis-àvis their employees. Although interactional justice proves to be the most crucial element for employees in coping with an insecure job situation, our study indicated that this aspect may need most improvement. Poor communication within the context of restructuring seems not so much deliberate, but rather the result of a lack of skills on the part of managers to communicate in a clear, open and human manner. Managers are often reluctant to communicate bad news that may touch on feelings of anger or frustration in their employees. Managers therefore prefer to neglect this important part of the restructuring process and focus instead on economic figures, rational processes and tangible elements. Some companies deal with this problem by offering their managers training, either general or specific, for the context of restructuring, in delivering bad news and in dealing with negative emotions of employees. In addition, meeting platforms are sometimes created for managers, where they can ventilate their own frustrations and worries with regard to the reorganisation, and can exchange experiences and information about best practice.
Employability All major stakeholders in the Netherlands labour market agree that a welldeveloped employability policy is the most important strategy to equip employees with the necessary skills and attitudes to cope well with occupational transitions. In order to respond adequately to the shift from lifetime employment to lifetime employability, the majority of large Dutch companies have established an elaborate employability policy, or have implemented regulations to stimulate the employability of their workers. As for the middle and small companies, employability measures still need improvement. Over the past decade, measures to further employability have become more structured and formalised (i.e., laid down in collective agreements), instruments to promote employability have become more advanced, companies have developed a more positive attitude towards employability, and individual employees have acquired increasing responsibility for their own labour market position. Broad-based training, task enrichment, job rotation, horizontal and vertical mobility, career management support and internships are examples of techniques that are used to enhance the employability of employees in the Netherlands. Companies tend to put more emphasis on a proactive, 275
Unemployment and Health ■ Interventions
preventative approach (i.e., furthering employability) as opposed to a curative approach (i.e., financial or other arrangements after dismissal). Collective bargaining often incorporates agreements about training, job rotation, personal development plans and career management facilities. The Dutch government has embraced the concept of employability and has defined a plan of action to stimulate the employability especially of lowskilled workers in the Netherlands. Also, the labour unions in the Netherlands support the concept of a more proactive, versus a curative, approach towards restructuring and dismissals. The concept of lifetime employability, aimed at facilitating and softening the process of occupational transitions, has gained importance in the collective bargaining agreements between the social partners. The labour unions have embraced employability as the most important means of securing the labour market position of Dutch employees. In exchange, an employable workforce provides companies with the flexibility greatly needed in the current dynamic, competitive, global market context. Enhancing employability is generally regarded as a shared responsibility between employers and individual employees. Companies are responsible for adequately preparing their workforce for occupational changes by facilitating and stimulating employability. Employees, on the other hand, should make a continuous effort to learn and to develop their knowledge and skills in order to retain a favourable position on the labour market. Both the unions and the Dutch government are developing new policies with regard to anchoring the responsibility of companies and individuals. For example, in the future, the philosophy of corporate social responsibility might be anchored in the social insurance system by introducing a system of premium differentiation between companies who have put effort in preventing unemployment and others who have failed to do so. Also, there is discussion in the Netherlands with regard to integrating the proper responsibility of individual employees over their labour market position into dismissal legislation. That is, employability might become a future selection criterion or a prerequisite for continuation of the labour contract. The unions in the Netherlands are considering possible sanctions for employees in insecure jobs who have failed to make use of the employability measures offered to them by their employer (i.e., exclusion from outplacement). This is in line with a general tendency of ascribing more responsibility to the individual employee in preventing unemployment and in securing a favourable position on the labour market. Employability is usually defined in terms of objective features, such as level of education, skills, working experience and age. However, our study has clearly shown that ‘softer’, more psychological, features such as selfefficacy and personal wellbeing are extremely important in determining the individual level of employability. Employees who feel confident about their own capacities and skills, and their ability to find another job, clearly have 276
Restructuring and Outplacement in the Netherlands
fewer problems in dealing with job transfers than workers who are less selfconfident. This specific aspect of employability is positively related to having received fair and humane treatment from the employer. Individuals who have been treated in a fair manner by their former employers clearly need less time to cope with grieving over the past, and are able to focus more easily and with more self-confidence on future job prospects. Thus, engaging in fair and humane treatment in the context of reorganisations and dismissals can be an important way for companies to promote the employability of their redundant employees. In our sample, age, years of tenure and frequency of change of employer proved to be most decisive for the level of employability. Older employees, with extended years of tenure with the same employer, clearly felt more insecure about their position in the labour market than younger employees who had frequently changed companies. Not the frequency of occupational transitions, but rather the frequency of changing employers proved to be decisive for the level of employability. Respondents who had been working for the same company for many years had much more difficulty in extricating their skills, knowledge, working methods and identities from their former employer.
Outplacement in the Netherlands The Dutch Outplacement Market The Netherlands can be considered a ‘best practice’ country with regard to outplacement. It represents the country with the highest rates of outplacement in Europe, a long tradition, and a mature and diverse outplacement market, which allows employers to offer redundant workers tailor-made support. Different outplacement and career-management agencies operate in different segments of the market for different target groups using different techniques and strategies. Over the last three decades, outplacement agencies have adapted their strategies to the demand in the market. Beginning as a way to release senior executives from companies in the 1970s, massive restructuring of large companies during the 1980s and 1990s led to the implementation of outplacement strategies for all types of redundant workers, to assist them during their periods of occupational transition. The favourable economic situation in the 1990s and the increasing tightness of the labour market has since motivated outplacement agencies to focus their attention from curative outplacement to preventative forms of career management. In addition, a change in methodology may be observed, in which cookbook-like recipes for dress codes and curriculum vitae layout have changed in favour of a more psychological approach, focused on furthering self-efficacy of the candidate. Outplacement agencies have also become involved at an earlier stage in supporting companies in the restructuring process. The different segments in the Dutch outplacement market and its historical development can be schematised as shown in the following diagram: 277
Unemployment and Health ■ Interventions
Figure 1 Development and segments of the Dutch outplacement market.
Roughly speaking, the favourable economic situation and tight labour market have induced a shift from an instrumental to a psychological approach, from group to individual outplacement, from curative to preventative outplacement, and from senior executive search to outplacement of individuals at lower levels in the organisation. Another recent development in the Netherlands is the arrival of new participants in the outplacement market. That is, several temporary work agencies have created business units in order not to restrict their activities to the in- and throughflow of personnel, but to take care of the outflow for their clients as well. These types of agencies generally focus on an instrumental, pragmatic, group-oriented approach.
Elements Within an Outplacement Intervention Outplacement agencies generally offer similar basic elements in their services to their candidates. A general introduction and intake, establishing a personal profile, psychological guidance in the mourning process and in self-analysis, support in labour market analysis, training in application skills, support in job search and matching, and follow-up care are basic activities that are integrated in all the case studies described here. The agencies place different emphases in terms of group versus individual 278
Restructuring and Outplacement in the Netherlands
support, stimulating self-efficacy versus active support in job search and matching, and psychological guidance versus pragmatic support. Two major elements characterise the basic outplacement procedure: 1. The first aim is to define, with the aid of counselling and guidance, an appropriate and motivating objective for the (dismissed) employee that is realistic and saleable in the labour market. This first part of counselling aims to assist the employee to respond to the questions: ‘Who am I, what am I capable of doing, and what do I want?’ The objective is to capture the individual’s experience, capabilities, personality and motivation, and to match these with current labour market trends. Information is acquired through a series of in-depth interviews and assignments, and a review of the participants’ past accomplishments. Common instruments include career and personal surveys, ability analysis and personality profile questionnaires. 2. The aim of the second part of the counselling is to generate an interview for the outplacement candidate. This can be achieved by training him or her in a variety of job search skills and in the identification of suitable opportunities within the labour market. This part of the counselling aims to assist employees to respond to the questions: ‘Where can I find what I want, and how do I get there?’ Training in networking, curriculum vitae design, verbal presentations and interviews are common elements in this part of the outplacement program. Some outplacement agencies focus on the first part of the process. The common idea among these consultants is that psychological stabilisation and encouragement of the candidate, who may be experiencing considerable distress and turmoil owing to his or her dismissal, will automatically result in finding him or her the right position. Other agencies put greater emphasis on the second, more pragmatic part of the process in focusing on ‘finding a job as best one can, as soon as possible’. There is no single best formula for supporting redundant employees in their transfer to a new job. It is important that the style of the counsellors, the network of the agency and the methods deployed match the specific profile and needs of the target group involved. Some employees benefit most from practical training of basic application skills, while for others, support in psychological stabilisation and furthering self-efficacy prove to be effective. Therefore, tailor-made guidance that matches the specific needs and profile of the individual candidate, in terms of psychological support, and practical aid and training is crucial for the success of the outplacement program. In some cases, support can be restricted to thematic workshops or even omitted in job transfers. Not all dismissed employees need outplacement support for finding another suitable job. However, outplacement could be integrated as a standard in the social plan as a safety net that employees can use on a voluntary basis. 279
Unemployment and Health ■ Interventions
Integrating group elements in the outplacement program can have a positive effect with regard to both task-orientated and emotional support between candidates, and reducing the costs for the employer. Individual elements meet the need for tailor-made guidance and the specific needs and wishes of the candidate. The right balance between group and individual elements depends on the profile of the target group involved on the one hand, and on the budgetary possibilities on the other. In a similar way, it is important to find the right balance between psychological and practical support. Need for psychological guidance in the grieving process, and in supporting candidates to dissociate themselves from the former employer and focus on future possibilities, depends on the background of the individual involved. For a young, highly employable corporate ‘job hopper’, for example, the need for this type of psychological support is rather different from that of a low-skilled production worker who has been working in the same factory for 30 years. The same goes for the need for practical support in acquiring application skills and approaching the labour market. The need for psychological support in reorientation depends both on the individual needs and wishes, and on the demands in the labour market. If there no longer exists a need for the type of skills that individuals have deployed during their career, psychological guidance is needed to support the candidate to reorient him or herself towards other possible jobs. As for the right balance between stimulating self-efficacy and taking over job-hunting activities, once again the profile and background of the individual is decisive. Our case studies show that a shared responsibility in job search and matching between the candidate and the agency seems to be effective in many cases. Therefore, it is important that the specific network of the agency matches the profile of the population involved.
Success and Failure of Outplacement All respondents underline the personal motivation of candidates as one of the most crucial factors in a successful outplacement program. Participation in such programs should therefore preferably be on a voluntary basis. Obliging employees to participate severely reduces the chances of success. Extensive information sessions and the role of labour unions and the works council are important in convincing employees to put their trust in outplacement support. Lack of motivation may result from uncertainty about the dismissal date and about possible internal replacement. Another reason for candidates not to be motivated is the so-called ‘golden cage’. The continuation of pay for a certain period, or a financially rewarding separation arrangement can be more attractive than the salary offered by a potential new employer. People need different types of support to help them find another job. In some of the cases in our study, psychological stabilisation and promotion 280
Restructuring and Outplacement in the Netherlands
of self-efficacy proved to be effective, while other respondents benefited from support in reorientation or skills training. Therefore, tailor-made guidance is crucial for the success of outplacement. The main effects of outplacement that were mentioned in the interviews include the enlargement of employability, knowledge of application skills, awareness of proper skills and abilities, reduction of negative emotions towards the former employer, reduction of feelings of insecurity and a shift in identification with a company towards identification with tasks. Many respondents indicate that without the support of an outplacement agency they probably would not have found another job, or would have accepted an unsuitable job. The lack of active job hunting and networking by the outplacement agency, and lack of tools to evaluate the quality and integrity of agency, were the most important criticisms that respondents put forward. Outplacement may be considered a means by which a company offers to further the employability of dismissed workers. Our study showed that dismissed employees who had been assisted by their former employers to define their skills, qualities and motivations as existing independently of their former working context, had much more success in finding another suitable job. All respondents underlined the importance of employability in the current labour market in which employees are more insecure about staying with the same company until their retirement. Respondents stressed that outplacement is a specific kind of employability that should not only be deployed as a curative instrument, but be made accessible to all employees in a preventative manner. The results of the case studies have raised the question of whether outplacement support is necessary for everyone, or should rather be restricted to vulnerable groups such as low-skilled or older workers. Perhaps the best way to deal with this question is to integrate outplacement as a standard in collective agreements or social plans, as a kind of safety net that workers can use on a voluntary basis. Moreover, the necessary period of outplacement may vary greatly among employees. The support should best provide workers with sufficient support and security after their dismissal, but should not have a negative effect on the motivation for job search (i.e., the ‘golden cage’). A period between 6 months and 2 years (the latter for vulnerable groups) seems most appropriate.
Conclusion Employability, justice and outplacement are three main strategies by which employers can secure the position of their workers on the labour market and ease occupational transitions. In the Netherlands, both employers and employee representatives place great emphasis on preventative, proactive measures to ease job transfers for redundant workers. Specifically, employability has become a primary strategic focus for dealing with increasing job insecurity and the demand for flexibilisation in the labour market. With 281
Unemployment and Health ■ Interventions
regard to justice considerations, a greater emphasis on intangible aspects of the restructuring process (i.e., communication), versus technical, material and formal aspects, is important in dealing with workers in a socially responsible manner, and in reducing the possible negative psychological consequences of reorganisations and dismissals. The processes of employability, fair restructuring and outplacement are closely intertwined. Employees who are treated in a fair manner during the process of reorganisation, and employees who adopt the right skills and attitudes to move flexibly on the labour market clearly experience fewer difficulties in the phase of occupational transition. They need less time to cope with negative feelings towards their employer and can more easily focus on future career perspectives. In considering the social responsibility of companies in the context of reorganisations and dismissals, this seems to be an important issue. Outplacement should not be deployed solely as an instrument to repair the psychological damage and impaired self-confidence of an individual who has been treated in an unfair manner by his or her former employer. Companies who treat their redundant workers in a socially responsible manner can prevent this kind of damage from emerging. In such a case, outplacement can become an opportunity for workers to consider carefully future career steps, particularly with the aid of sufficient time and appropriate instruments and professional support. Outplacement should not be treated as an isolated practice, but rather should form part of an integrative set of instruments deployed by companies to assist their workers in the current dynamic labour market. Outplacement may be considered one instrument among others to further the employability of individual workers. One could state that in a best-case scenario, a welldeveloped employability policy could make the use of outplacement unnecessary. Training, internal mobility, preventative career counselling and personal development plans will generally support individuals in guaranteeing a good position on the labour market and in developing the right flexible attitude needed to easily cope with occupational transitions.
Reference Tjiong, H. (2005). Institutional dynamics in environmental corporatism: The impact of market and technological change on the Dutch Polder Model. Governance-an International Journal of Policy and Administration, 18, 1–33.
282
Restructuring and Outplacement in the Netherlands
Appendix A Summary of Sample Characteristics Description
n
Source of data
Age (years)
Total
Male
Female
Range
Mean
Insecure employees
27
14
13
30–60
44
Successfully reemployed
27
13
14
34–62
48
Case studies
8 Dutch companies that were selected for implementation of best practice outplacement/replacement procedures during company restructuring
Stakeholders
6 human resource managers 12 union and council representatives 3 employer representatives 3 employment services representatives
283
Unemployment and Health ■ Interventions
284
Chapter
21
Unemployment and Activation Policy: The Finnish Experience Simo Mannila
In this chapter, I will describe the joint problems of long-term unemployment and dependency on means-tested cash benefits in Finland in the 1990s. I will also describe some widespread legislative or action program measures to cope with these problems, including new activation legislation launched in 2001. The situation of the target groups of Finnish activation policy is described in light of existing research and some potential results of this approach are indicated.
What is Activation Policy and What Are Its Roots? Activation policy stems from two roots of policy-making: first, active labour market policies, and second, policy interventions towards recipients of means-tested social assistance (i.e., cash benefits). Active labour market policies — with the Swedish tradition dating back to the 1940s as the most famous case — have historically two main aims: the promotion of long-term economic growth, and the provision of support to vulnerable labour market groups (e.g., Sihto, 2001). These policies can be defined operationally in compliance with the Eurostat (2001) classification. However, Finnish public discussion concentrates mostly on the latter aim, nowadays often categorised as special employment policy, aiming to guarantee and increase the employment of vulnerable groups. In practice, this means that a wide range of active labour market measures, such as vocational rehabilitation, training and special employment schemes based on incentives or direct job creation, are available for this purpose (cf. Meager & Evans, 1998). Traditionally, the financial inputs of active labour market policy are significantly lower than those of passive measures. Passive measures include various insurance-based or other unemployment benefits, calculated by the numbers of recipients and funding. The total Finnish expenditure on unemployment was approximately €3.5 billion in 2001, and in the years of peak 285
Unemployment and Health ■ Interventions
expenditure from 1993 to 1995, the sum was €4 billion. Approximately one third of the sum was used for active labour market policy, while two thirds was used to fund unemployment allowance (Prime Minister’s Office, 2001). This bias towards passive, rather than active, labour market policy is similar across all OECD countries (Sihto, 1994, p. 246; cf. Mannila, 2001). Following the recession of the 1990s, the Finnish unemployment rate remained at a substantially higher level than it had been earlier; a new form of structural unemployment emerged. The official unemployment figures released by Statistics Finland showed an unemployment rate of over 15% during the period 1993 to 1995 (see www.stat.fi), and the figures of unemployed jobseekers in the Public Employment Service as shown by statistics of the Finnish Ministry of Labour were considerably higher (see www.mol.fi). The Finnish unemployment rate is usually based on the former figures, which are comparable to the general European Union (EU) statistics. Since those years, the unemployment rate has fallen: it was 9.8% in the year 2000 and 7.2% in November 2001. Today, Finland has almost 80,000 jobseekers who have been unemployed for more than one year, and almost 40,000 jobseekers who have been unemployed for at least two years. The situation has radically changed in the last 10 to 15 years: the first of these two figures was only 3000 in 1990. Within the EU, Finland is very close to the average; the long-term unemployed comprise 2.8% of the labour force. The lowest percentages (below 1%) are in the Netherlands and Denmark, while the highest percentages (above 7%) are in Spain, Italy and Greece (OECD, 2001). Combining the long-term unemployed, the cumulatively unemployed — at least 12 months unemployment during 24 months — and those retired by the unemployment pension scheme, gives a total for Finland of 190,000 unemployed persons in 2001. This is 6.7% of the country’s entire labour force (Nio, 2001). Currently, long-term unemployed jobseekers comprise 29% of all unemployed jobseekers in Finland. This is one of the lowest figures in the EU, although somewhat higher than in Sweden (26%), the United Kingdom (UK) (28%) and Austria (28%), and considerably higher than in Denmark (20%), which has the lowest percentage in the EU. The youth unemployment rate in Finland is very close to the EU average. The unemployment rate of the workforce aged 15 to 24 years is 22% and the unemployment rate of all youth aged 15 to 24 years is 11%; the difference between these two figures is caused mainly by nonparticipation in the labour force due to education or training. The highest youth unemployment rates are in Italy and Ireland, with corresponding figures of 32% to 33%, while the lowest rates are in Austria, the Netherlands and Portugal, with corresponding figures of 3% to 9% (by Eurostat, 2001, criterion; Prime Minister’s Office, 2001).
286
Unemployment and Activation Policy: The Finnish Experience
However, youth unemployment in Finland is rarely long term, and unemployed youth do not constitute a key group of means-tested cashbenefit recipients. The overwhelming majority of the latter group is in the optimal working age group of 25 to 49 years (Stakes, 2001). The risk of long-term unemployment and social exclusion among youth, however, still exists, and it is related to inadequate or nonexistent training. Long-term unemployment and youth unemployment and their impact on family subsistence and public expenditure, as well as the related threat of social exclusion, are considered major social problems in Finland, as can be seen in the National Action Plans of Employment Policy (Finnish Ministry of Labour, 2004) and in the National Action Plan to Combat Poverty and Social Exclusion (Finnish Ministry of Social Affairs and Health, 2001b). Reliance on means-tested cash benefits has similarly become a social problem in western countries. In Finland, benefit dependence, which can be observed with unemployment allowance schemes and means-tested cash benefits, has increased since the mid-1980s; this development could not have taken place earlier because until then both kinds of social protection schemes were well below the European average. Rather than being solely related to inadequate income, dependency on means-tested cash benefits is one component of a complex profile of social deprivation and exclusion. In Finland, for example, this means that only one third of the poor as defined by income criteria and those receiving means-tested cash benefits are the same people (Heikkilä, 1990). On the positive side we may state that, thanks to the income transfers of the welfare state (including means-tested cash benefits), the recession of the 1990s did not lead to increased inequality in income; Finland is still one of the EU countries that has a fairly low level of social stratification by income (Kautto, Heikkilä, Hvinden, Marklund, & Ploug, 1999). While the percentage of households with an income of less than 50% of the median has risen during the 1990s from 2.5% (1990) to 3.6% (1999), it is still very low compared with other EU countries. Interestingly, however, the percentage of households (21%) with difficulties in ‘making ends meet’ is high in Finland. Thus, Finland, together with Greece, Portugal, Spain and Ireland, belongs to some of the most problematic countries in the EU, according to this criterion of subjective poverty (European Community Household Panel data; Prime Minister’s Office, 2001). Reliance on means-tested cash benefits is increasingly related to unemployment. The European link between these phenomena is well illustrated by developments in Finland. Approximately 50% of all means-tested cash benefit recipients are today unemployed, while 10 years ago the corresponding figure was 20%. Some 9% of Finnish households (272,000) and persons (460,000) receive means-tested cash benefits, and the costs today are €400 million, twice as much as 10 years ago. The costs of means-tested cash benefit 287
Unemployment and Health ■ Interventions
provision did not peak until after the recession of the 1990s. In 1997 the costs were approximately €600 million and since then have decreased quite substantially in a rather short period of time. The percentage of those living on the minimum unemployment allowance peaked in 1994, when it covered almost 8% of all labour force, and it is still almost 6% (Prime Minister’s Office, 2001). The probability of receiving mean-tested cash benefits was 10 times higher among the long-term unemployed than among those who were employed throughout the 1990s, independent of economic fluctuation (Kainulainen, 2001). While the unemployment rate has fallen since 1993, until 1996 the probability of re-employment for unemployed jobseekers also fell, and although it has been rising since then, today’s rate of 20% is considerably lower than the pre-recession rates. During the recession and subsequently, high unemployment has been tackled by means of structural economic policies and increased use of traditional active labour market measures. In the latter category, the main types of measures have been either subsidised employment into public (state or municipal) and private sector jobs, as classified by Eurostat (2001) as direct job creation or incentive policy) or labour market training in different forms. The numbers of jobseekers in both types of programs more than doubled during the recession; subsidised employment peaked during 1993 to 1997, with over 60,000 persons yearly employed annually; and labour market training peaked during 1995 to 1998, with more than 40,000 persons channelled annually into training. As the numbers of unemployed jobseekers in those years were between 400,000 and 500,000 according to the stock (i.e., cross-sectional) statistics provided by the Finnish Ministry of Labour (2000), these measures were of key importance in the reduction of unemployment. Today, the levels of both types of measures remain considerably higher than 10 years ago: in 2000, the numbers of jobseekers channelled into subsidised employment and labour market training were approximately 44,000 and 30,000, respectively (Finnish Ministry of Labour, 2001). There are controversial results concerning the effectiveness of subsidised employment and labour market training. The system may have some leaks due to ‘creaming’, that is, selection bias towards those jobseekers who might be employable without any support, and it is quite likely that more tailor-made services are needed to cater for the many different target groups. The problem is not simple, as better targeting requires distinction; in employment services there is an inherent conflict between helping those jobseekers worst off, which may lead to losing the employers’ confidence, and creaming — that is, helping those who, maybe, would manage on their own (cf. Aho, 2001; Hämäläinen, 1999; Nio, 2001; Suikkanen & Linnakangas, 2000). In 1999 the main target group for active employment policies was the long-term unemployed, who received 49% of all assistance measures. The 288
Unemployment and Activation Policy: The Finnish Experience
share apportioned to unemployed youth was 15%, while 7% went to disabled jobseekers, and 29% went to the miscellaneous group of ‘others’, for example, those engaged by means of various projects or otherwise unclassified, but mostly consisting of the above target groups plus immigrants (Finnish Ministry of Labour, 2000). However, today, the standard measures of active employment policy are considered inadequate to tackle the dual problems of unemployment and dependency on means-tested cash benefits, or to integrate vulnerable groups of society such as the long-term unemployed or unemployed youth into working life. This leaves room for the activating policy, which is also motivated economically as a means of reducing the financial burden of social benefits, and promoting labour market flexibility. Both of these outcomes presumably contribute to economic growth (cf. Lind & Hornemann Möller, 2001). Activation policy aims not only to combat long-term unemployment and dependency on means-tested cash benefits, it aims to address the general employment rate, too. In Finland the employment rate fell from 76% in 1990 to a low of 60% in 1994, and is now 68%. This means that policies that aim to integrate older adults and the disabled into the workforce, including the interaction of labour market and pension policies, are relevant to the design of activation policy. Life expectancy in Finland today is some eight years longer than in the 1960s when the bulk of the earningsrelated employment pension system was created. The age limits of premature retirement schemes have been increasing, which makes room for activation policies (Ala-Kauhaluoma, Keskitalo, Lindqvist, & Parpo, 2004; Keskitalo & Mannila, 2004). In short, activation policies constitute the conscious use of ‘stick and carrot’ methods (which threaten withdrawal of benefits for breaches of obligation, while offering additional incentives for compliance) on a crossadministrative basis. In other words, they consist of joint action with employment and social (and health) authorities as the principal partners. They can include (1) legal interventions with regard to service providers (e.g., obligation to offer activation programs) or assistance recipients (e.g., right of access to program participation or obligation for claimants/recipients to accept work or training), (2) financial incentives for employers or recipients (e.g., subsidised work, in-work benefits or reduction of benefits), (3) work integration services (e.g., counselling, job placement, work orientation, education and training, public employment, subsidised work), and/or (4) services to promote social integration (participation in socially useful work, or voluntary work in nongovernment organisations; Hanesch, StelzerOrthofer, & Balzter, 2001, pp. 123–124; cf. Gilbert & van Voorhis, 2001). In principle, there are two main approaches to activation policy, and in practice, both approaches are, to varying degrees, typical of most activation schemes. The approaches are (1) imposing on the target group either 289
Unemployment and Health ■ Interventions
financial pressures or legal obligations to participate in certain policy measures, and (2) emphasising the right to access various schemes or services by which employment and independent living can be supported (e.g., Hvinden, Heikkilä, & Kankare, 2001). There still seems to be rather little research concerning the social impact of activation policy as described above, although the approach is generally accepted in Europe (Heikkilä & Keskitalo, 2001; Weise & Brogaard, 1997; cf. Dröpping, Hvinden, & Vik, 1999).
Who Are Those to be Activated? As stated above, in Finland, the poor, as defined by low income and those receiving means-tested cash benefits, are mainly different people (e.g., Heikkilä, 1990). Means-tested cash-benefit recipients in Finland are persons with less-than-average income who are unable to transform their financial resources into personal or family welfare. Receiving this benefit is related to unemployment. Other dimensions of social deprivation in Finland that are typical of means-tested cash-benefit recipients are low levels of social participation and being single or divorced. The risk of labour market exclusion is to a certain extent generational, and is related to both ill health and inadequate education or training (Heikkilä, 1990; Mannila & Peltoniemi, 1997). In the 1980s, male and female trajectories into ill health and longterm unemployment were quite distinct, and this pattern may still be true; social deprivation was more comprehensive in the case of men. On the basis of a qualitative study it was hypothesised that labour market exclusion and social exclusion from family life seemed to belong to the vicious cycle of the male lifespan, also involving ill health and irregular lifestyle, while for women the labour market failure was counterbalanced by involvement in family (Mannila, 1993; cf. Berghman, 1995). One of the key characteristics in individuals who comprise the target of activation policies is deteriorated health or work capacity. In Finland the work capacity of the long-term unemployed is often questioned, and there is evidence that being unemployed and in retirement via various disability pension schemes are to a certain extent substitutive social and administrative measures, designed to tackle labour market exclusion at individual level (Gould, 2001). Ill health and reliance on means-tested cash benefits are also linked, and in Finland the link is stronger among the single and divorced; this relationship is similar to that between ill health and unemployment (Heikkilä, 1990; Ritakallio, 1991; Tanninen & Julkunen, 1993). It has been stated that means-tested cash-benefit recipients are at risk of mental ill health (Heikkilä, 1990, pp. 166–168). In addition, one study found that the mortality risk for long-term means-tested cash-benefit recipients was 27 times higher than that of the average population of the same age (Tanninen & Julkunen, 1993, pp. 120–124). The overwhelming majority of recipients of 290
Unemployment and Activation Policy: The Finnish Experience
means-tested cash benefits (65%) live today in one-person, mainly male, households. Male, one-person households make up 39% of all recipients. Looking at specific groups of beneficiaries, however, the highest recipiency rate (28%) is among single-parent households (Stakes, 2001). These links may be weaker now that the numbers of long-term unemployed jobseekers and means-tested cash-benefit recipients are higher, which would indicate that the selection into both groups may be weaker than previously when the demand for workforce was higher and the employability of the long-term unemployed jobseekers was lower. However, both long-term unemployment and receipt of means-tested cash benefits are linked to a complex pattern of social deprivation, including deterioration of health. There is a great deal of evidence on the causal links between unemployment and mental wellbeing (Paul & Moser, 2001). The causal mechanisms between unemployment and other forms of morbidity or mortality are, however, still unclear and, to some extent, controversial. In Finland, two main approaches to the relationship can be discerned. First, Martikainen and his colleagues (Jäntti, Martikainen, & Volkonen, 2000; Martikainen, 1990; Martikainen & Valkonen, 1996) have found increased mortality among unemployed Finnish men; in particular, from such causes as alcoholism, accidents, suicides and cardiovascular diseases. The former three causes indicate risky lifestyle or habits, while the link between unemployment and cardiovascular mortality is more difficult to understand. At the macro level, no increased mortality has been found to coincide with increases in unemployment; to the contrary, mortality has decreased (Jäntti et al., 2000). All things considered, these authors emphasise joint selection into unemployment and mortality as the main explanation rather than proposing unemployment as a cause of unemployed jobseekers’ excess mortality. Kortteinen and Tuomikoski (1997) claim to have found the link from long-term unemployment to ill health, using both register and survey data. They examined such health indicators as mental symptoms, medication, sick days, work incapacity and mortality. Qualitative analysis of autobiographies of unemployed people was utilised in order to investigate the mechanisms leading to the deterioration of health in particular cases. The interpretation of the destructive mechanism proposed is based on the idea that an unemployed person needs both material and moral support. Material support is related to the unemployment allowance system (accessibility, coverage, level). For moral support, a key role is given to a characteristic described as social distrust or cynical hostility, which is related to the extent that a person is able to find and utilise social support. This characteristic is strongly dependent on early socialisation, but it changes over time, and economic hardships tend to increase it. In order to understand the mediating role of social distrust and the link between material and moral support it must be remembered that (1) social distrust 291
Unemployment and Health ■ Interventions
is socioeconomically patterned with those higher in social stratification at lower risk, and (2) a person’s economic situation deteriorates with prolonged unemployment. The long-term unemployed are often recruited from the lowest social strata. By implication, therefore, their preexisting levels of social distrust are likely to be high, and being unemployed is likely to increase these levels further. Kortteinen and Tuomikoski (1997) hypothesise that this vicious cycle is responsible for the detrimental effects on health that they found in a subgroup of long-term unemployed. The explanation is similar to that proposed in numerous studies that have found that social support acts as a buffer against the effects of unemployment. It is also in accord with the results of other studies focusing on cynical hostility as a risk factor for poor health (e.g., Christensen et al., 2001; Julkunen, Saarinen, Idänpään-Heikkilä, & Sala, 2000), although scientifically it remains at the level of a hypothesis. In relation to activation policy, the study points to the problematic character of the target group, which is likely to need the implementation of comprehensive, tailor-made measures. Also, benefit cuts are a problematic component of activation policy: for some individuals, they may well strengthen the financial incentives available through employment as planned, but for others, they may simply demonstrate a lack of social and moral support, and may result in a further deterioration of health and wellbeing. Another lesson from the study is that how the activation process is organised influences whether marginalised groups will be socially integrated into the labour market as a result or instead will feel even more excluded and distrustful. Arnkil (2001) states that one-sided activation (i.e., allocating all of the obligations and few of the benefits to the unemployed) is internally contradictory: to have a positive impact on the target groups, the civil service itself must be activated first. The clients of employment and social authorities have little power as clients, but they may have other very active life roles (cf. Borghi & Kieselbach, 2000). We may still look at the relationship of unemployment, receipt of means-tested cash benefits and ill health by comparing healthy with unhealthy people. During the 1990s, the links between ill health and the two forms of social deprivation (unemployment and receipt of meanstested cash benefits) were strong, which was interpreted to reflect the higher level of unemployment. Looking at the main illness categories, those suffering from mental ill health were more often deprived in all dimensions studied (Mannila & Peltoniemi, 1997). However, in spite of increased unemployment in the 1990s, we do not find a worsening state of health at the Finnish national level, which is consistent with results concerning social transfers and their positive impact (e.g., Kautto et al., 1999). However, this result may also be due to the timing of Finnish Living Conditions Surveys (1986 and 1994), that is, they do not capture this phenomenon (Manderbacka, Lahelma, & Rahkonen, 2001). 292
Unemployment and Activation Policy: The Finnish Experience
Health problems have been among the most common individual-level causes of long-term unemployment in Finland, although there has been little research into this topic (cf. Vähätalo, 1991, pp. 79–86). Thus, as stated above, premature retirement and unemployment are substitutive, and for some sections of the long-term unemployed there is a medical way out on the basis of disability or other specific, age-related pension schemes. In Finland the situation of jobseekers aged 50 or more, and the policies adopted towards them, have been rather different from those pertaining to other jobseekers. Like the Netherlands (but unlike Sweden), Finland has largely opted for retirement schemes (which are based on various disability pensions as well as on a scheme based on job record and unemployment) instead of activation measures to solve problems posed by the existence of a surplus labour force aged over 50. Older adults who are nevertheless still under the legal retirement age, with their attendant educational and training disadvantages, have thus been provided with a legitimate social status and stable income maintenance. The average retirement age in Finland is nowadays around 57 or 58 years, although legally it is still 60 to 65 years (Finnish Ministry of Social Affairs and Health, 2001a; Hytti, 1998; ). It has been found that financial incentives exert a modest influence over whether one remains in the labour force (Hakola, 1999). According to the latest Finnish results, the prevalence of subjective work inability in the 50plus age group and the corresponding urge towards premature retirement is very high when compared with the prevalence among the share-retired (Aromaa & Koskinen, 2002). Most premature retirement schemes require a medically certified condition as the criterion for a legitimate exit from the labour force, and the medical profession functions as the main gatekeeper of the pension system. In spite of wide-ranging criticism, it has also been speculated that this policy of medicalising labour market problems has had positive side-effects for public health (Hytti, 1998).1 Results from vocational rehabilitation of the long-term unemployed in Finland indicate two types of positive results that are relevant for activation policy-making. First, there is evidence of increased self-reliance and other positive mental outcomes as the result of rehabilitation measures (although these effects may soon wear out, and it is not easy to transform them into reemployment). Second, the differentiation of the target group towards two opposite directions — employment versus retirement — can also be seen as a positive result of vocational rehabilitation (Mannila & Laisola-Nuotio, 2002). There seems to be no similar review of research into the impact of interventions targeted at Finnish recipients of means-tested cash benefits, although some pilot projects have been carried out. Rehabilitation is increasingly available also to nondisabled jobseekers. In this situation, some writers question the usefulness of the registration of disability in the Finnish public employment service (Suikkanen & Linnakangas, 2000).
293
Unemployment and Health ■ Interventions
The difference between active labour market policy and activation policy concepts can be seen as rather arbitrary from the Scandinavian point of view: all Scandinavian countries have a long-standing tradition of combining ‘work options’ and ‘benefit options’, workfare and welfare. The problem of dependency on means-tested cash benefits is new and strongly dependent on the maturation of social protection schemes. Health concerns also play a different role in each of these examples.
Activation and Rehabilitative Work Experience? In employment administration in Finland, activation policy was preceded by a number of large-scale interventions to solve the problems of the longterm unemployed and unemployed youth, for example, the Employment Act (1987), and the Intervention Programme to Assess the Service Needs of Older Long-Term Unemployed (1996; see Rajavaara, Järvikoski, & Lind, 2000). A number of activating elements were introduced in the mainstream services of the public employment service during the 1990s. This means, for example, that labour market support, the minimum-level means-tested unemployment benefit paid in the case of either very long-term unemployment or no previous working experience (e.g., students and housewives in search for work), was in certain cases made conditional on participation in active labour market measures, while a new, combined subsidy was introduced, enabling higher pay levels to the jobseekers thus placed. Activating elements were also introduced in other benefit systems: the payment of cash rehabilitation benefit in the case of temporary work inability (entering after the period of sickness allowance has expired) was made conditional on a rehabilitation plan; and a new benefit type — rehabilitation allowance — was introduced, which paid at the level of standard income subsistence but with an increase of 10% to 30% during vocational rehabilitation. During the late 1990s Finnish municipalities took an increasing number of initiatives to activate long-term, means-tested cashbenefit recipients by employing them temporarily in municipal jobs. In September 2001, following a heated controversy, legislation was introduced in Finland concerning the activation and rehabilitative work experience of long-term unemployed jobseekers and unemployed youth, both of whom were living mainly on labour market support or meanstested cash benefits. This was the first attempt to tackle the dual problems of unemployment and dependency on means-tested cash benefits in compliance with the European and international activation policy (cf. Gilbert & van Voorhis, 2001; Heikkilä & Keskitalo, 2001). The legislation gained full force in January 2002. The target groups were recipients of labour market support and means-tested cash benefits defined according to various lengths of time, and their definition criteria will not be described here in detail. The idea was that the combined efforts of employment authorities and social and health administration, 294
Unemployment and Activation Policy: The Finnish Experience
together with a new policy measure, would give an added value to the existing services, and the target groups could be helped towards employment and other active solutions. The life situation of the clients was assessed on a tripartite basis (client, employment agent, social worker), and an activation plan to design a pathway towards employment was made. An activation process was considered to have been successful when a person had been employed for 10 months, either by means of special employment schemes or in the open competitive labour market. In the activation plan, the standard services available by the employment administration were considered first, including various employment schemes, training and rehabilitation, which, if necessary, were supported by social and health services. If these services were not considered useful, the client was offered rehabilitative work experience. This activity was provided on a municipal basis, and it was up to the local government to determine the details of the provision. The third sector (the social worker) was also engaged in the provision of the rehabilitative work experience to varying degrees in different municipalities. Rehabilitative work experience could be offered for a period of three to 34 months, covering one to five days a week and four to eight hours per day. In principle, this means that the particular work activity could be tailored according to the client’s situation. The work activity was understood as a form of social or vocational rehabilitation, it is not employment. The client’s benefit structure and level remained the same, with only a small incentive plus travel expenses paid during the work activity. The implementation of the legislation varied by age. For those under 25 years the participation in measures jointly agreed was obligatory and refusal resulted in benefit cuts, while for those aged 25 or more, only activation plans were made and the participation in the measures was voluntary. For the former target group, the sanctions in the case of refusal consisted of stopping labour market support for two to three months or cutting the level of meanstested cash benefits by 20% to 40%. For the latter target group, there were to be no financial sanctions, but the activation process, consisting of interviews and counselling on a tripartite basis, was to be taken up until the employment criterion is fulfilled. The triggering criteria for activation were also different in these two age groups, meaning that the process was to be started much earlier for those under 25 years. It was estimated that the target groups would cover altogether 70,000 persons, both long-term unemployed jobseekers in the job centres living on labour market support or means-tested cash benefits, and unemployed youth not registered in job centres and living on means-tested cash benefits. However, no additional funding was made available to the employment authorities, and that available to social and health administration was modest. The costs of the organisation of rehabilitative work experience obviously exerted an influence on its implementation. 295
Unemployment and Health ■ Interventions
Additional support to the new legislation was given an experimental ‘onestop shop’ program launched in 2002 and covering the 20 largest municipalities in the country. This means that joint offices were established for employment authorities and social and health administrators to improve client service. Significant amounts of additional funding were made available for this program. It was apparent also that the legislation concerning activation and rehabilitative work experience was to be implemented in these offices in the program municipalities, and that the additional funding would, again, have an influence on the implementation. However, it was not quite clear whether this experimental program would support the activation policy, or whether it would merely cause additional workload for the public employment service and social and health administrators. The outcomes and implementation of the legislation were monitored and evaluated during the period 2001 to 2003. The monitoring and evaluation apparatus consisted of several parts. A large representative survey covering 57 municipalities and corresponding job centres gave the main outlines of the new legislation in action during 2002 to 2003. Targeting of the legislation was assessed, and detailed studies of implementation, both from the point of view of employment authorities as well as social and health administration, were carried out in seven pilot municipalities and job centres. Some parts of the study were based on quasi-experimental design. Additional value of cross-administrative cooperation was also assessed. The results achieved were rather controversial and we cannot without any reservation conclude that the new activation policy in Finland has been successful (Ala-Kauhaluoma et al., 2004). However, the reform requires new cooperation of employment and social administration, structures of central government and local governments. The establishment of this new cooperation will require some time, and the follow-up period of the study was short. Medical certification of serious health problems causing inability to work is likely to be one outcome of the activation process for some sections of the target groups. However, according to previous experience, the proportions thus to be channelled out of the workforce and into retirement will remain small (Mannila, Tynkkyen, & Eronen, 1993; Rajavaara et al., 2000). The quality of the partnership between employment authorities and social and health services may be of key importance here (Viitanen, 2000). It will, however, not be an easy task to bring health into the systematic agenda of fighting long-term unemployment. Long-standing experience from vocational rehabilitation, where this kind of partnership has traditionally been considered productive, shows that, so far, the practical solutions have been far from adequate. The local public employment service branches are centrally administered, while social and health administration belong to local government, and this makes an important difference in decision-making and funding. Employment authorities do not focus solely on job search and hard-to296
Unemployment and Activation Policy: The Finnish Experience
place jobseekers, but, increasingly, on employers. At the wider regional level, the key partner of employment authorities consists of administrators of trade and economy, not of health and social affairs. The level of knowledge with the client and his or her problems may be less in job centres than in health and social welfare organisations. The new legislation aims at additional value from cross-administrative and other networking: it remains to be seen whether this value will be found. The implementation of the law was initially delayed, and there is a major geographic variation: different municipalities and public employment service offices showed a very variable interest in activation policy. The activation of unemployed youth has been tackled with a much greater effort than has that of the 25-plus target group. No significant changes in the unemployment figures of the target groups attributable to the functions of the law could be found during the first part of 2002. By the end of 2002, only a small minority of the original target group of 70,000 was served in compliance with the legislation. In the meantime, the target group had grown considerably due to increased unemployment. There seems to have been a great disparity between the officially stated aims of the legislation and its practice during the first year of the implementation.
Some Conclusions Finnish activation policy is well rooted in the Scandinavian tradition of active labour market and employment policy. It is based on a variety of good intentions and is intended to tackle the problems of groups at risk of social marginalisation. The local public employment service and municipal services in all three cases described above have had considerable leeway in the implementation, which means that the results have also been variable. Starting in the late 1980s there is, however, an increased emphasis on combining benefit strategy (negative sanctions and positive incentives) with active labour market measures. It is interesting that the Employment Act of 1988 showed a clear social rights approach, which in later activation policy has been more neglected. The social rights approach has come more clearly to the fore in the 1990s in the anti-legislation policy adopted and is not directly relevant to activation policy (cf. Parkkari, 1996). Activation policy cannot be a narrow and one-way approach; a prerequisite of good activating results is the joint action of employment, social and health services, and their partnership with local nongovernment organisations and the economy (cf. Rajavaara et al., 2000). Since crossadministrative joint action is already a difficult enough target (cf. Hudson, 1993), this kind of conscious production and reproduction of networking and social capital may be even more difficult (Mannila, Valjakka, & AlaKauhaluoma, 2001). One of the underlying weaknesses in activation policies may be their unclear and potentially contradictory targets, quite typical of the 297
Unemployment and Health ■ Interventions
political action of this kind. This may help to explain problems associated with their implementation. In the new changing labour market, activation policy also runs the risk of being a halfway solution. Labour market transitions thus supported may lead into a chain of services, including supported placements, but not into employment in the open competitive labour market. Here we must bear in mind, first, that underemployment and marginal labour market status also contain health risks, although smaller than unemployment (cf. Gallie, Marsh, & Vogler, 1994). Second, the transitional labour market can, however, be seen as an opportunity, not just as a threat, for those excluded from employment, and conscious policy-making efforts should be taken to develop a new social economy (cf. Mannila et al., 2001). At the individual level, active labour market policies for vulnerable jobseekers result in differentiation of the target groups towards two opposite directions, employment and retirement. A prerequisite of the latter solution is medically certified work incapacity, which means detection of serious health problems or a process of medicalisation, where labour market problems are interpreted in the language of health as required by pension schemes. Both results must be seen as feasible and should be recognised as legitimate goals. However, for a subgroup of longterm unemployed there is a risk of increased social exclusion, with its attendant health risks, with no solutions towards either employment or retirement. A similar risk may concern some young people who are under 25 years of age. It is up to the implementation process of the activation policy how these effects can be avoided. The most serious risk, however, is that the new activation policy will not succeed in differentiating the target group at all. The new activation legislation of 2001 sets a strong claim towards developing a new work culture in two very different branches of civil service, and it may take some time before it has been adopted in both the public employment service and social and health administration. A pessimistic view would be that this is just another administrative reform requiring a great deal of effort, whose main aim is to introduce benefit cuts under the guise of a policy that is rhetorically more legitimate. We face here one of the basic problems for the present welfare state, which is inherent in the concept of activation, that is, which part of the policy-making is intended for welfare, and which part is for social control. Social policy-making is difficult in a post-industrial society, where the definition of social exclusion may be controversial and, in any case, must be defined multidimensionally (cf. Hvinden, 2000). In countries like Finland with a universalistic welfare regime (incorporating comprehensive high-level benefits with an extensive employment policy) unemployed people have the lowest risk of poverty and social exclusion. However, there are major differences between countries with similar 298
Unemployment and Activation Policy: The Finnish Experience
regimes: the results depend on how the regime is implemented (Gallie & Paugam, 2000, pp. 351–374).
Endnote 1
For some groups of unemployed jobseekers, with reduced labour market opportunities but working capacity left, a range of special employment measures, including vocational rehabilitation, is available. These measures are also legally accessible to disabled people. Disability is defined here in accordance with the ILO Convention No. 159 and certified medically.
References Aho, S. (2001). Miksi työvoimapoliittisten toimien työllisyysvaikuttavuus on vaatimaton? Keskustelua arviointulosten luotettavuudesta ja tulkinnasta sekä aktiivisen työvoimapolitiikan tavoitteista [Why is the employment impact of active labour market measures modest? Discussion concerning the reliability and interpretation of evaluation results and the goals of the measures]. Työpoliittinen Aikakauskirja, 44, 45–56. Ala-Kauhaluoma, M., Keskitalo, E., Lindqvist, T., & Parpo, A. (2004). Työttömien aktivointi. Kuntouttava työtoiminta -lain sisältö ja vaikuttavuus [The activation of the unemployed. The contents and impact of the legislation on rehabilitative work experience] (Research Reports 141). Helsinki, Finland: STAKES. (With an English summary) Arnkil, R. (2001). Is good practice generative practice? In S. Mannila, M. AlaKauhaluoma, & S. Valjakka (Eds.), Good practice in finding good practice, (pp. 25–42). Working Papers 23. Helsinki, Finland: Rehabilitation Foundation. Aromaa, A., & Koskinen, S. (2002). Terveys ja toimintakyky Suomessa. Terveys 200 — tutkimuksen perustulokset [Finnish health and functional capacity. Basic results from Health 2000 Survey]. Publications of the National Public Health Institute B3/ 2002. Helsinki, Finland: National Public Health Institute. Berghman, J. (1995). Social exclusion in Europe: Policy context and analytical framework. In G. Room (Ed.), Beyond the threshold. London: Policy Press. Borghi, V., & Kieselbach, T. (2000, November). The submerged economy as a trap and a buffer: Comparative evidence on long-term youth unemployment and the risk of social exclusion in southern and northern Europe. Paper presented at the EU Workshop on Employment, Work and Welfare, DG Research of the European Commission, Brussels. Christensen, U., Lund, R., Mogens, T., Damsgaard, T., Iversen, L., Due, P., et al. (2001, June). Cynical hostility: A contributory mechanism in the explanation of social inequality in health. Paper presented at the Nordic Conference in Social Medicine and Public Health, Bergen. Dröpping, J.A., Hvinden, B., & Vik, K. (1999). Activation policies in the Nordic countries. In M. Kautto, M. Heikkilä, B. Hvinden, S. Marklund, & N. Ploug (Eds.), Nordic social policy. Changing welfare state (pp. 133–158). London: Routledge. Eurostat. (2001). European social statistics, labour market policy expenditure and participants. Data 1998. Luxembourg: Author. (Finnish) Employment Act (1987). Työllisyyslaki 13.3.1987/ 275. Retrieved September 1, 2005, from www.finlex.fi/fi/laki/ajantasa/kumotut/1987/198702275 Finnish Ministry of Labour. (2000). Työllisyyskertomus vuodelta 1999 [Report on employment, 1999]. Helsinki, Finland: Author.
299
Unemployment and Health ■ Interventions Finnish Ministry of Labour. (2001). Finnish Labour Review, 3, Statistics. Helsinki, Finland. Finnish Ministry of Labour. (2004). Finland’s national action plan for employment 2003. Retrieved September 1, 2005, from www.mol.fi/mol/en/99_pdf/en/90_publications/nap 2003english.pdf Finnish Ministry of Social Affairs and Health. (2001a). Sosiaali- ja terveyspolitiikan strategiat 2010 [Strategies of social and health policy towards 2010]. Publications 2001: 3. Helsinki, Finland: Author. Finnish Ministry of Social Affairs and Health. (2001b). Köyhyyden ja sosiaalisen syrjäytymisen vastainen toimintasuunnitelma [National action plan to combat poverty and social exclusion]. Memorandum 2001:11. Helsinki, Finland: Author. Gallie, D., Marsh, C., & Vogler, C. (1994). Social change and the experience of unemployment. Oxford: Oxford University Press. Gallie, D., & Paugam, S. (Eds.). (2000). Welfare regimes and the experience of unemployment in Europe. Oxford: Oxford University Press. Gilbert, N., & van Voorhis, R.A. (Eds.). (2001). Activating the unemployed. A comparative appraisal of work-oriented policies. ISSA Series 3. New Brunswick, Canada: Transaction Publishers. Gould, R. (2001). Työttömyys ikääntyneiden työkyvyttömyyseläkekäytännössä 1990luvulla [The role of unemployment in disability pension policy for ageing workers in the 1990]. Finnish Journal of Social Medicine, 38, 172–185. (With an English summary) Hakola, T. (1999). Race for retirement (Research Report 60). Helsinki, Finland: Government Institute for Economic Research. Hanesch, W., Stelzer-Orthofer, C., & Balzter, N. (2001). Activation policies in minimum income schemes. In M. Heikkilä & E. Keskitalo (Eds.), Social assistance in Europe. A comparative study on minimum income in seven European countries, Synthesis report (pp. 122–151). Helsinki, Finland: STAKES. Heikkilä, M. (1990). Köyhyys ja huono-osaisuus hyvinvointiyhteiskunnassa [Poverty and deprivation in a welfare state]. Helsinki, Finland: Finnish National Board of Social Administration, Publications 8. (With an English summary) Heikkilä, M., & Keskitalo, E. (Eds.). (2001). Social assistance in Europe. A comparative study on minimum income in seven European countries, Synthesis report. Helsinki, Finland: STAKES. Hudson, B. (1993). Collaboration in social welfare. A framework for analysis. In M. Hill (Ed.), The policy process: A reader (pp. 362–378). London: Harvester Wheatsheaf. Hvinden, B. (2000). Onko sosiaalipolitiikka omaksunut sosiologian hylkäämät vanhentuneet käsitteet [Have the outdated concepts abandoned in sociology been received by social policy?]. Janus, 8, 134–149. Hvinden, B., Heikkilä, M., & Kankare, I. (2001). Towards activation? The changing relationship between social protection and employment in western Europe. In M. Kautto, J. Fritzell, B. Hvinden, J. Kvist, & H. Uusitalo (Eds.), Nordic welfare states in the European context (pp. 168–197). London: Routledge. Hytti, H. (1998). Varhainen eläkkeelle siirtyminen — Suomen malli [Early retirement — the Finnish model]. Studies in Social Security and Health 32. Helsinki, Finland: The Social Insurance Institution. (With an English summary) Hämäläinen, K. (1999). Aktiivinen työvoimapolitiikka ja työllistyminen avoimille työmarkkinoille [Active labour market policy and employment in non-subsidised labour markets]. Series B 151. Helsinki, Finland: The Research Institute of the Finnish Economy. (With an English summary) Julkunen, J., Saarinen, T., Idänpään-Heikkilä, U., & Sala, R. (2000). Sydänkuntoutuksen tuloksellisuus [Efficacy of inpatient and outpatient cardiac rehabilitation: A ran300
Unemployment and Activation Policy: The Finnish Experience domised, controlled trial] (Research Reports 63). Helsinki, Finland: Rehabilitation Foundation. Jäntti, M., Martikainen, P., & Valkonen, T. (2000). When the welfare state works: Unemployment and mortality in Finland. In G.A. Cornia & R. Paniccià (Eds.), The mortality crisis in transitional economies (pp. 351–369). Oxford: Oxford University Press. Kainulainen, S. (2001, October). Processes of social exclusion. Paper presented at the Annual Conference of the Finnish Society of Social Policy, Rovaniemi. Kautto, M., Heikkilä, M., Hvinden, B., Marklund, S., & Ploug, N. (Eds.). (1999). Nordic social policy. Changing welfare states. London: Routledge. Keskitalo, E., & Mannila, S. (2004). Activation policy — an answer to the problem of long-term unemployment and exclusion. In M. Heikkilä & M. Kautto (Eds.), Welfare in Finland (pp. 103–122). Helsinki, Finland: STAKES. Kortteinen, M., & Tuomikoski, H. (1997). Työtön [Unemployed]. Helsinki, Finland: Hanki ja jää. Lind, J., & Hornemann Möller, I. (2001, August–September). The Danish experience of labour market policy and activation of the unemployed. Paper presented at the Conference of European Sociological Association, Helsinki. Manderbacka, K., Lahelma, E., & Rahkonen, O. (2001). Structural changes and social inequalities in health in Finland. Scandinavian Journal of Public Health, Suppl. 55. Mannila, S. (1993). Työhistoria ja syrjäytyminen [Work career and social exclusion]. Studies in Labour Policy 58. Helsinki, Finland: Finnish Ministry of Labour. (With an English summary) Mannila, S. (2001). Vajaakuntoiset ja työllisyys- ja työvoimapolitiikka: esimerkki politiikkatason arvioinnista [Disabled jobseekers in employment policy: An example of policy-level evaluation]. In A. Järvikoski, K. Härkäpää, & S. Nouko-Juvonen (Eds.), Monia teitä kuntoutuksen arviointiin (pp. 73–96) (Research Reports 69). Helsinki, Finland: Rehabilitation Foundation. (With an English summary) Mannila, S., & Laisola-Nuotio, A. (2002). Pitkäaikaistyöttömien kuntoutuksen vaikuttavuus [The effectiveness of rehabilitation measures for long-term unemployed people]. In A-M. Aalto, H. Hurri, A. Järvikoski, J. Järvisalo, V. Karjalainen, H. Paatero, T. Pohjolainen, P. & Rissanen (Eds.), Kannattaako kuntoutus? Asiantuntijakatsaus eräiden kuntoutusmuotojen vaikuttavuudesta (pp. 57–72) (Research Reports 267). Helsinki, Finland: STAKES. (With an English summary) Mannila, S., & Peltoniemi, J. (1997). Köyhiä ja kipeitä. Terveydentila ja syrjäytyminen 1986-94 [Poor and ill. Ill-health and deprivation in Finland, 1986-94] (Research Reports 57). Helsinki, Finland: Rehabilitation Foundation. (With an English summary) Mannila, S., Tynkkynen, A., & Eronen, M. (1993). Vajaakuntoiset työnhakijoina [Disabled jobseekers and the services of employment authorities] (Research Reports 33). Helsinki, Finland: Rehabilitation Foundation. (With an English summary) Mannila, S., Valjakka, S., & Ala-Kauhaluoma, M. (2001). Crossing the job threshold – in search for good practices. In S. Mannila, M. Ala-Kauhaluoma, & S. Valjakka (Eds.), Good practice in finding good practice (pp. 75–84). Working Papers 23. Helsinki, Finland: Rehabilitation Foundation. Martikainen, P. (1990). Unemployment and mortality among Finnish men. British Medical Journal, 301, 407–411. Martikainen, P., & Valkonen, T. (1996). Excess mortality of unemployed men and women during a period of rapidly increasing unemployment. Lancet, 348, 909–912. Meager, N., & Evans, C. (1998). The evaluation of active labour market measures for the long-term unemployed (Employment and Training Papers 16). Geneva, Switzerland: Employment and Training Department, International Labour Office. Nio, I. (2001). Toimenpiteiden vaikuttavuuden arviointia tilastoseurannan pohjalta – kolme näkökulmaa vaikuttavuuteen [Evaluating the impact of policy measures on the 301
Contributors
About the Contributors Sarah Anderson is a research assistant at the Centre for Applied Psychological Research at the University of South Australia. She is currently working on a longitudinal study that follows school leavers into the workforce. She is also a PhD candidate in psychology at the University of South Australia. Her interests include transition into work and the effects of employment status on health, wellbeing and worker attitudes. Arnold B. Bakker, PhD, is Professor of Positive Organizational Behavior at the Department of Social and Organizational Psychology, Utrecht University, the Netherlands. He graduated at Groningen University, and has published many articles on the Job Demands–Resources model, attitudes, burnout and work engagement. His research interests include flow at work, crossover and organisational performance. Dee Bartrum is a doctoral student in clinical psychology at the School of Applied Psychology, Griffith University, Australia. Her research interests include stress and coping, job change, job insecurity, self-regulation, and child and adolescent development. Ola Bergström, PhD, is Assistant Professor at the Department of Business Administration at the School of Economics and Commercial Law, University of Göteborg, Sweden. He obtained his doctorate at Göteborg University in 1998. His research interests include changes in labour markets, work organisation and forms of control. He has been a coordinator of the EU-financed NUEWO-project (New Understanding of European Work Organization), which is concerned with the changing use of contingent employment in Europe and the United States. His most recent book is Contingent Employment in Europe and the United States, Cheltenham: Edward Elgar, 2003, with D. Storrie. Carolyn Boyd is a research assistant at the University of South Australia’s Centre for Applied Psychological Research. She is also completing a doctoral thesis at Flinders University in South Australia. Her research interests include generativity and adult development and work–life balance.
303
Unemployment and Health ■ About the Contributors
Ed Carson, PhD, is Professor of Social Policy and Dean of Graduate Studies at the University of South Australia. His research interests include employment and labour market programs for disadvantaged job seekers, including mature-age workers, young job seekers and long-term unemployed. Before working as a teacher and researcher at Australian universities, he studied in the United Kingdom (London School of Economics) and worked in labour market programs in the United States (Massachusetts). In Australia he has undertaken employment-related research projects and consultancies for a range of Commonwealth and State government departments. He has also undertaken research on more general aspects of social disadvantage and social policy. His current research includes projects on workforce development in the area of community services and health. Ralph Catalano, PhD, is Professor of Public Health at the University of California, Berkeley. He is Director of the Robert Wood Johnson Health and Society Program at Berkeley, and Director of the Perinatal Outcomes Research Project. He holds a PhD from the Maxwell School of Citizenship at Syracuse University. Catalano’s research has focused on the effects of the macro-economy and of economic policy on health and the demand for health services. He has served as Associate Vice Chancellor at the University of California, Irvine, as well as Vice Mayor of Irvine and a member of the Irvine City Council. He has also been Director of the Division of Health Policy and Management at Berkeley. Bjørgulf Claussen, MD, is Professor of Community Medicine at the University of Oslo, Norway. Following 16 years of work in general practice, he completed his medical thesis in 1994 on the health consequences of unemployment in Norway. Since then he has worked especially with unemployed patients in preventive and clinical medicine. Peter Creed, PhD, is Associate Professor in Organisational Psychology at the School of Applied Psychology, Griffith University, Australia, and is a member of Griffith University’s Service Industry Research Centre. His research interests include the effects of unemployment on wellbeing and confidence, school-to-work transition, career psychology and psychometrics. Claudia Dalbert, PhD, is Professor of Psychology at the Martin Luther University of Halle-Wittenberg, Germany. She received her doctorate from the University of Trier (Germany) in 1987. Leo Montada, the developmental psychologist and justice researcher, was her mentor. Her research focuses on the justice motive theory. Currently she is investigating how people cope with injustice over the life course and how the justice motive develops. Dalbert is author or editor of numerous articles and books about the justice motive including The Justice Motive as a Personal Resource: Dealing With Challenges and Critical Life Events and The Justice Motive in Adolescence and Young Adulthood: Origins and Consequences (edited 304
Unemployment and Health ■ About the Contributors
together with Hedvig Sallay). Her other interest concerns the tolerance of uncertainty and career development. In 2004 she was appointed president of the International Society of Justice Research (ISJR) and will hold this position until 2006. Yucel Demiral, MD, PhD, is Associate Professor of Occupational Health in the School of Medicine at Dokuz Eylül University in Turkey. Alp Ergor, MD, PhD, is Associate Professor of Occupational Health in the School of Medicine at Dokuz Eylül University in Turkey. Norman T. Feather, PhD, is Emeritus Professor of Psychology at the Flinders University of South Australia. He is a graduate of the University of Sydney, the University of New England and the University of Michigan, where he received the Donald G. Marquis Award for his doctoral dissertation. He is a past-president of the Australian Psychological Society and recipient of their Distinguished Scientific Contribution Award. He has published widely in the areas of achievement motivation, expectancy-value theory, attribution theory, gender roles and stereotypes, the psychology of values, social attitudes, the psychological impact of unemployment, deservingness, entitlement, and justice- related behaviour and emotions. He has authored or edited six books, including Expectations and Actions: Expectancy-Value Theories in Psychology; The Psychological Impact of Unemployment; Values in Education and Society; and Values, Achievement and Justice: Studies in the Psychology of Deservingness. David Fryer, PhD, is Reader in Psychology at the University of Stirling in Scotland, where he also coordinates an undergraduate course in community psychology. He has been conducting research into the mental health consequences of unemployment for several years and has recently diversified his research into other aspects of community psychology. Anne Hammarström, PhD, is Professor of Public Health at Umeå University in Sweden and is also a general practitioner. She has more than 23 years of experience in researching unemployment and health from both Swedish and international perspectives. Much of her research has involved longitudinal population-based studies of school-leavers. Elizabeth Harris is Director of the Centre for Health Equity Training and Evaluation (CHETRE) at the University of New South Wales, Australia. She has a long-term interest in development of effective interventions to reduce health inequalities and works primarily in the areas of unemployment and health, primary health care, and health-impact assessment. Ellen Heuven completed her studies of sociology and management of art before working for several years as an occupational health consultant in the Netherlands. Subsequently she worked as a researcher in the field of employment and reintegration for both the University of Utrecht and the 305
Unemployment and Health ■ About the Contributors
Dutch Institute for Applied Sciences (TNO). She currently runs her own company in scientifically based career-management training and is completing her PhD on emotional labour among cabin attendants this year. Tatsuya Ishitake, PhD, is Professor in the Department of Environmental Medicine at the Kurume University School of Medicine. His research interests and expertise lie in the areas of ergonomics; occupational, environmental and public health; and community health. Staffan Janson, MD, is paediatrician and Professor of Public Health at the Karlstad University, Sweden. He is editor of the European Journal of Public Health, and also works for the Swedish Government in the field of child abuse and primary prevention. His main research fields are child abuse, injury prevention, epidemiological surveillance, chronic diseases and longterm illness. Leif R. Jönsson, PhD, obtained his doctorate in social work from Lund University, Sweden. His main areas of research are unemployment and poverty. Eric Kessell is a doctoral student in epidemiology at the University of California, Berkeley’s School of Public Health, where he also earned a Masters of Public Health in epidemiology and biostatistics. His research interests include factors that influence the recognition and treatment of mental illness, including community tolerance, unemployment and neighbourhood ethnic composition. He is also a member of the Shumway Mental Health Outcomes Research Group at the University of California, San Francisco. Thomas Kieselbach, PhD, is a Professor of Psychology and co-founder of the Institute for the Psychology of Work, Unemployment and Health at the University of Bremen. Since 2000 he has been the chairperson of the Scientific Committee: Unemployment and Health of the International Commission on Occupational Health. Since 2003 he has chaired the commission for Psychology of Unemployment of the International Association of Applied Psychology. Merete Labriola trained as an occupational therapist and has a Masters in Public Health. She is a researcher at the National Institute of Occupational Health in Copenhagen. Her research focuses primarily on sickness absence and returning to work, and aims at combining different disciplines in research projects recognising the multifactorial nature of labour market transition mechanisms. Thomas Lund has a PhD in Medical Science and is a researcher at the National Institute of Occupational Health in Copenhagen. His research interests include occupational health, work absence, returning to work and workplace-based interventions. 306
Unemployment and Health ■ About the Contributors
Jacques C. Metzer is Associate Professor in the School of Psychology, University of South Australia, where he was Foundation Head from 1994 to 2004. His research interests include the psychology of volunteering, intrinsic reinforcement, learning theory, and the measurement of job stress and satisfaction and cross-cultural effects. Klaus Moser, PhD, is Professor of Organizational and Social Psychology at the University of Erlangen-Nuernberg, Germany. His research interests include the psychological effects of unemployment; the psychological effects of temporary agency work; studies on commitment, including employment commitment; and the evaluation of training programs for unemployed people. Before coming to Nuernberg he studied psychology and philosophy of sciences at the University of Mannheim, worked as a research assistant at the University of Hohenheim, and as a full professor in work and organisational psychology at the University of Gießen. He also conducts research on personnel selection, performance appraisal, training and development, and online survey research. Karsten Paul is research assistant at the University of Erlangen-Nuernberg, Germany. His research interests include the psychological effects of unemployment, the meaning of work and employment, employment commitment, and work values. Before coming to Nuernberg he undertook German and theatre studies at the University of Munich, and conducted studies of psychology and medicine for psychologists at the University of Gießen. His other research interests are the effects of mood on cognitive processing, social categorisation and the effects of information overload at the workplace. Dimity Pond, PhD, is Professor of General Practice and Head of the Discipline of General Practice at the University of Newcastle, Australia. In that role she teaches undergraduate medical students, and also educates postgraduate general practitioners. She is in clinical general practice two sessions a week. She has a long history of general practice research, in particular focusing on mental health issues and disadvantaged groups. From 2000, she was Chief Investigator on a project looking at the health of a cohort of men retrenched from the BHP Steelworks in Newcastle. The project also investigated the approach general medical practitioners should take to unemployed patients. Katrin A. Rauter was an Honours student in Psychology at Flinders University, Adelaide under the supervision of Norman T. Feather. The study reported in this book was based on her thesis results. Bengt Starrin, PhD, is Professor of Social Work at the Karlstad University, Sweden. His main research fields are the sociology of emotions, social welfare and health. He has published several books in these fields. 307
Unemployment and Health ■ Interventions basis of statistics – Three views on impact evaluation]. Finnish Labour Review, 44(3), 25–46. (With an English summary) OECD. (2001). Employment outlook 2001. Paris: OECD Department of Economics and Statistics. Parkkari, J. (1996). Vammaisten ja vanhusten ihmisoikeudet [The human rights of the aged and disabled] (Report 200). Helsinki, Finland: STAKES. (With an English summary) Paul, K., & Moser, K. (2001). Negatives psychisches Befinden als Wirkung und als Ursache von Arbeitslosigkeit: Ergebnisse einer Metaanalyse [Problems of mental well-being as an impact and cause of unemployment: Results from a meta analysis]. In J. Zempel, J. Bacher, & K. Moser (Eds.), Erwerbslosigkeit. Ursachen, Auswirkungen und Interventionen (pp. 83–110). Opladen, Germany: Leske & Budrich. Prime Minister’s Office. (2001). Työmarkkinoilta syrjäytyminen, tulonjako ja köyhyys [Labour market exclusion, income distribution and poverty]. Economic Council, Publications of Prime Minister’s Offfice, 13. Helsinki, Finland. Rajavaara, M., Järvikoski, A., & Lind, J. (2000). The dilemma of long-term unemployment among older workers. Reflections on a programme evaluation study. In M. Rajavaara (Ed.), Yksilölliset palvelut ja ikääntyneiden pitkäaikaistyöttömyys [Individual services and long-term unemployment among older workers] (pp. 309–332). Studies in Social Security and Health 54. Helsinki, Finland: The Social Insurance Institution. (With an English summary) Ritakallio, V-M. (1991). Köyhyys ei tule yksin [Poverty comes not alone. A study of accumulated welfare deprivation among social assistance recipients] (Research Reports 11). Helsinki, Finland: STAKES. (With an English summary) Sihto, M. (1994). Aktiivinen työvoimapolitiikka. Kehitys Rehnin-Meidnerin mallista OECD:n strategiaksi [Active Manpower Policy. Development from the RehnMeidner Model to the Strategy of OECD]. Tampere, Finland: Tampere University Press. (With an English summary) Sihto, M. (2001). The strategy of an active labour market policy. An analysis of its development in a changing labour market. International Journal of Manpower, 22, 683–706. STAKES. (2001). National statistics on means-tested cash benefits. Unpublished. Suikkanen, A., & Linnakangas, R. (2000). Vajaakuntoisten mahdollisuuksien uusjako? [Redistribution of the options of disabled people?]. In R. Linnakangas, J. Lindh, & A. Järvikoski (Eds.), Työttömyyden ja vajaakuntoisuuden jäljillä (pp. 9–80) (Research Reports 66). Helsinki, Finland: Rehabilitation Foundation. (With an English summary) Tanninen, T., & Julkunen, I. (1993). Elämää säästöliekillä. Tutkimus toimeentulotuen pitkäaikaisasiakkaista pohjoismaiden pääkaupungeissa [Life in frugality. Research into long-term clients of social assistance in Nordic capitals] (Research Reports 32). Helsinki, Finland: STAKES. Vähätalo, K. (1991). Pitkäaikaistyöttömyyden mosaiikki ja työllisyyslaki [The mosaic of long-term unemployment and the Employment Act]. Helsinki, Finland: Finnish Ministry of Labour. Viitanen, M. (2000). Ikääntyneiden pitkäaikaistyöttömien palvelutarveselvityksen toteutus Oulussa [The service needs assessment of older long-term unemployed and its implementation in Oulu]. In M. Rajavaara (Ed.), Yksilölliset palvelut ja ikääntyneiden pitkäaikaistyöttömyys (pp. 285–308). Studies in Social Security and Health 54. Helsinki, Finland: STAKES. (With an English summary) Weise, H., & Brogaard, S. (1997). Aktivering of kontanthjalpsmodtagere. En evaluering of Lov om communal aktivering [Activation of cash benefit recipients. An evaluation study on the implementation of the law on activation in Danish municipalities]. Socialforskningsinstituttet, 21. Copenhagen, Denmark. 302
Unemployment and Health ■ About the Contributors
Maarten Vansteenkiste, PhD, is a postdoctoral researcher at the University of Leuven (Belgium). His fellowship is funded by the Institute for Scientific Research in Flanders. His research interests lie in motivational dynamics, focusing particularly on theoretical and applied perspectives in self-determination theory. He has published extensively in this field. Tony Winefield, PhD, is Professor and Director of the Centre for Applied Psychological Research at the University of South Australia. He obtained his PhD at University College London, and worked at the University of Adelaide for many years before joining the University of South Australia in 1999 as Foundation Professor of Psychology. From 1980 to 1989 he directed a 10-year longitudinal investigation of youth unemployment, culminating in more than 50 publications, including the book: Winefield, A.H., Tiggemann, M., Winefield, H.R., & Goldney, R.D. (1993). Growing Up With Unemployment: A Longitudinal Study of Its Psychological Impact. In 1993 he was made a Fellow of the Australian Psychological Society and in 2000 he gave an invited State of the Art address on the psychology of unemployment at the 27th International Congress of Psychology, Stockholm. He has published extensively on the psychology of unemployment, organisational stress and learned helplessness, and in 2003 he received the Elton Mayo Award from the Australian Psychological Society’s College of Organisational Psychologists for ‘outstanding contributions to industrial/organisational psychology research and teaching’. He is currently Associate Editor of the International Journal of Stress Management.
308
UnemployHealth Cov-final.x:Layout 1
6/6/08
11:37 AM
Page 1
Unemployment and Health provides a rare insight into ground-breaking research from Europe, Australia, Asia and the United States on the health impact of unemployment and underemployment on the individual and the community. Contributions from the fields of psychology, medicine, economics, sociology, occupational health and organisational development outline current theory and findings about the factors responsible for the detrimental health effects of being out of work or in unstable employment. The mechanisms
Unemployment and Health
>
International and Interdisciplinary Perspectives
$
underlying the benefits of secure employment are also discussed, along with interventions that
of corporate and government policies, the empowerment of employees, and management responsibilities in the dismissal process.
+ %
This book is a vital resource for policy-makers, social workers, educators, researchers, students and anyone with an interest in the global effects of unemployment on the human condition.
AUSTRALIAN ACADEMIC PRESS www.australianacademicpress.com.au
Edited by
employment status, including the implementation
Thomas Kieselbach, Anthony H. Winefield, Carolyn Boyd and Sarah Anderson
may help to limit the negative health effects of
$
>
Unemployment and Health International and Interdisciplinary Perspectives Edited by Thomas Kieselbach, Anthony H. Winefield, Carolyn Boyd and Sarah Anderson
+ %