Regional Externalities

Regional Externalities Wim Heijman (Editor) Regional Externalities With 74 Figures and 50 Tables 123 Professor Dr...

Author: Wim Heijman

142 downloads 1561 Views 3MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Regional Externalities

Wim Heijman (Editor)

Regional Externalities With 74 Figures and 50 Tables

123

Professor Dr. Wim Heijman Wageningen University Hollandseweg 1 6706 KN Wageningen The Netherlands [email protected]

COST COST is an intergovernmental European framework for international cooperation between nationally funded research activities. COST creates scientific networks and enables scientists to collaborate in a wide spectrum of activities in research and technology. COST activities are administered by the COST Office.

Library of Congress Control Number: 2006938717

ISBN 978-3-540-35483-3 Springer Berlin Heidelberg New York This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable to prosecution under the German Copyright Law. Springer is part of Springer Science+Business Media springer.com © Springer-Verlag Berlin Heidelberg 2007 The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Production: LE-TEX Jelonek, Schmidt & V¨ ockler GbR, Leipzig Cover-design: WMXDesign GmbH, Heidelberg SPIN 11781219

88/3100YL - 5 4 3 2 1 0

Printed on acid-free paper

Acknowledgements

Editing a book like this needs the support of quite a number of people. Without being exhaustive, a number of them deserve to be named. Above all, I would like to thank my co-authors for their loyal cooperation in writing and rewriting their own contributions and refereeing colleagues’ articles. Furthermore, this book would not have been written without the EU COST programme, which provided not only the financial contribution for publication but also the stimulating scientific environment. In this regard, my sincere thanks goes to Jordi Suriñach, the chairman of the COST Action A17 (Small and Medium Enterprises, Economic Development and Regional Convergence in Europe); Berta Ballart, his secretary; and David Gronbaek, the stimulating COST officer for Social Sciences and Humanities, who has superbly coordinated our activities. I would also gratefully like to acknowledge Henk Folmer’s, Peter Nijkamp’s and Ayda Eraydin’s active support for the book. Finally, I would like to thank Annelies Coppelmans for carrying out the difficult and time consuming task of preparing the camera-ready manuscript and Katharina Wetzel-Vandai from Springer for the pleasant and effective cooperation in the complex process of publishing this book. Wageningen, January 2007

Wim Heijman

Contents

Acknowledgements ................................................................................... V 1

Regional Externalities: an Introduction ......................................... 1 Wim Heijman Literature ............................................................................................ 7

Part I Transport 2

Modelling Transport in an Interregional General Equilibrium Model with Externalities .......................................... Morten Marott Larsen, Bjarne Madsen and Chris Jensen-Butler 2.1 Introduction ........................................................................... 2.2 Positive Externalities............................................................. 2.2.1 Pecuniary and Technological Externalities .............. 2.3 Proximity, Income and Externalities in Denmark ................. 2.3.1 Estimating Urban and Pecuniary Externality Effects: the Two Approaches ................................... 2.3.2 Factors Creating Differences in Wage Levels, Not Involving Externality Effects ............................ 2.3.3 The Data ................................................................... 2.3.4 The Effects of Urban Externalities ........................... 2.3.5 Pecuniary Externalities: Wage Differentials and Commuting Distance ......................................... 2.4 The LINE-Model – Modelling Externalities and Transport.. 2.4.1 LINE – an Interregional General Equilibrium Model for Modelling Redistribution of Productivity Changes ........................................... 2.4.2 Direct Wage and Price Effects in LINE ................... 2.4.3 LINE and SAM-K: Configuration............................

11 12 13 14 17 17 18 18 20 22 23

24 26 29

VIII

Contents

2.5

Modelling the Redistribution of the Relative Productivity Decline Associated with Road Pricing – the Case of Denmark............................................................................ 2.5.1 Road Pricing and Changes in Transport Costs ......... 2.5.2 Effects of Transport Costs on Prices and Costs, With and Without Externalities ................................ 2.5.3 Effects on Real Economic Activity from Changes in Cost and Prices, With and Without Externalities.............................................................. 2.6 Conclusions ........................................................................... References ........................................................................................ 3

4

External Effects in Road Traffic: the Pigou-Knight Model and its Extension to Situations With Endogenous Speed Choice and Heterogeneous Traffic................................................ Jan Rouwendal 3.1 Introduction ........................................................................... 3.2 The Pigou-Knight Model of Traffic Congestion................... 3.3 Speed Choice and the Determination of the Travel Cost Function................................................................................. 3.3.1 Introduction .............................................................. 3.3.2 Speed Choice............................................................ 3.3.3 The Steady State....................................................... 3.3.4 A Specific Case ........................................................ 3.4 The Pigou-Knight Model with Heterogeneous Drivers......... 3.4.1 Introduction .............................................................. 3.4.2 Optimal Congestion Tolling ..................................... 3.4.3 Discussion ................................................................ 3.5 A Numerical Example........................................................... 3.6 Conclusion............................................................................. References ........................................................................................ On Traffic Congestion Models à la Mohring and Harwitz......... Pierre v. Mouche, Willem Pijnappel and Jan Rouwendal 4.1 Introduction ........................................................................... 4.2 Setting ................................................................................... 4.3 Real World Interpretation...................................................... 4.4 Existence Results .................................................................. 4.5 The Self Financing Result of Mohring and Harwitz ............. 4.6 Suggestions for Further Research ......................................... 4.7 Conclusions ...........................................................................

30 31 33

35 37 44

47 48 48 51 51 52 53 56 57 57 61 63 64 68 68 71 71 72 75 77 81 85 86

Contents

IX

Appendix: a Closer Look at the Setting............................................ 87 References ........................................................................................ 89 5

Local Collectors Versus Major Infrastructural Works .............. 91 Catharinus F. Jaarsma and Wim Heijman 5.1 Introduction ........................................................................... 92 5.2 The Importance of Local Connections .................................. 93 5.3 Current Policy Around Local Connections ........................... 96 5.4 Some Practical Experiences: the Disappearance of Local Connections ............................................................ 99 5.5 Conclusions, Recommendations and Discussion ................ 103 References ...................................................................................... 104

Part II Clusters and Product Chains 6

7

Regional Differentiation and Location of Industrial Capacity in the Slovak Republic.................................................................. Jana Gašparíková, Edita Nemcová and Michal Páleník 6.1 Introduction ......................................................................... 6.2 Industrial Development in the Slovak Republic with Special Impact on Technology Intensity ..................... 6.3 Regional Differentiation in Employment ............................ 6.4 Survey Study ....................................................................... 6.4.1 Explanation of Variables ........................................ 6.4.2 Regional Differentiation......................................... 6.4.3 Results .................................................................... 6.4.4 Consequences Resulting from the Survey .............. 6.5 Conclusion........................................................................... 6.6 Methodological Notes ......................................................... References ...................................................................................... Automobile Sector in the Slovak Republic: Current Situation and Future Prospects .................................... Daneš Brzica 7.1 Introduction ......................................................................... 7.2 Development of Automobile Production Capacities........... 7.3 Preconditions for Cluster Emergence.................................. 7.4 Automobile/Automobile Parts and Components Sector and Emergence of Automobile Clusters.............................. 7.5 Model Describing Relation Between Automobile and Parts/Components Production ......................................

109 109 111 114 117 118 119 121 125 127 128 129 131 131 133 135 137 143

X

Contents

7.6 Conclusions ......................................................................... 146 References ...................................................................................... 146 8

9

10

IT Market and E-Commerce in Transition Economy: Network Externalities .................................................................. Vytautas Snieška, Regina Virvilaitø, Vaida Kvainauskaitø, Bronius Neverauskas, Rimantas Gatautis, Aistø Dovalienø 8.1 Introduction ......................................................................... 8.2 The Structure of Network Externalities............................... 8.3 Network Externalities and Transaction Cost Minimization ....................................................................... 8.3.1 Transaction Costs in the Market Without Intermediaries......................................................... 8.3.2 Transaction Costs in the Market with Intermediaries................................................. 8.3.3 Transaction Costs and E-Commerce ...................... 8.3.4 E-Commerce and Intermediation Services ............. 8.3.5 The Impact of E-Commerce on the Co-Ordination Costs............................................... 8.4 IT Market and E-commerce Growth Effects: Lithuanian Case................................................................... 8.5 Research Methodology........................................................ 8.6 Results ................................................................................. 8.7 Conclusions ......................................................................... References ...................................................................................... International Outsourcing in the Netherlands........................... Kees Burger and Rein Haagsma 9.1 Introduction ......................................................................... 9.2 Vertical Specialization and Outsourcing in the Dutch Economy......................................................... 9.3 Outsourcing, Productivity, and Employment ...................... 9.4 Regional Aspects of Outsourcing........................................ 9.5 Summary and Discussion .................................................... References ...................................................................................... Regional Externalities and Clusters: a Dutch Network Case-Study..................................................................................... Roel Rutten and Frans Boekema 10.1 Introduction ......................................................................... 10.2 The “KIC- Project”: a DUTCH Regional Case-Study ........ 10.3 Creation of (Inter-Firm) Knowledge ...................................

149

149 150 152 153 153 154 155 158 161 164 167 169 170 173 174 175 182 187 194 196 197 198 201 202

Contents

10.4 10.5 10.6 10.7 10.8 10.9

A Variety of Disciplines...................................................... Tacit and Codified Knowledge............................................ Some Preliminary Results ................................................... Interfirm Relations .............................................................. The Regional Impact ........................................................... The Spatial Relevance of Knowledge Transfer: Proximity and Distance ....................................................... 10.10 Theoretical Implications...................................................... 10.11 Competitiveness and Knowledge ........................................ 10.12 Knowledge Exchange and Trust.......................................... 10.13 Regional Partners and Knowledge Exchange ..................... 10.14 Knowledge Creation in Networks Explained ...................... 10.15 Space Matters (More than Ever?)........................................ 10.16 Conclusion........................................................................... References ...................................................................................... 11

Spatial Dimension of Externalities and the Coase Theorem: Implications for Co-existence of Transgenic Crops .................. Volker Beckmann and Justus Wesseler 11.1 Introduction ......................................................................... 11.2 Assessing the Problem of Co-existence .............................. 11.2.1 A Simple Model ..................................................... 11.2.2 Co-Existence, Economic Damage and Technical Measures ......................................... 11.2.3 Liability Rights and Distribution of Costs and Benefits............................................................ 11.3 Co-Existence: A Coasian View........................................... 11.3.1 Efficient Allocation ................................................ 11.3.2 Spatial Implications ................................................ 11.3.3 Distributional Implications..................................... 11.4 Conclusions ......................................................................... Acknowledgements ........................................................................ References ......................................................................................

XI

204 205 206 207 209 210 212 213 213 214 215 217 218 219 223 224 225 227 229 231 234 235 236 239 239 240 241

Part III Regional Policy 12

Abatement of Commuting’s Negative Externalities by Regional Investment in Houses and Buildings...................... 245 Wim Heijman and Johan van Ophem 12.1 Introduction and Research Problem .................................... 246 12.2 Investment in Houses and Residential Mobility.................. 246

XII

Contents

12.3 A Model of Spatial Mobility ............................................... 12.4 Testing the Model ............................................................... 12.5 Policy Implication and Conclusion ..................................... References ...................................................................................... 13

14

Risk as an Externality in Quantitative and Marginal Approaches ........................................................... Václav Beran and Petr Dlask 13.1 Introduction ......................................................................... 13.2 The Cobb-Douglas Equation............................................... 13.3 Generalisation to Virtual Leading Moments in Virtual Matrix Avirtual ....................................................... 13.4 Externality of Management Decisions and an Alternative to Economic Intuition .................................. 13.5 Generalisation of Process P................................................. 13.6 Application for Regional Development in the Czech Republic: Findings ......................................... 13.7 Possible Strategies for Future Change................................. 13.8 Summary of Strategies ........................................................ 13.9 Conclusion........................................................................... References ...................................................................................... Macro Policies and Regional Impacts in Norway...................... Steinar Johansen 14.1 Introduction ......................................................................... 14.2 Relevant Policy Sectors....................................................... 14.2.1 Macroeconomic Policies ........................................ 14.2.2 19 Sector Policies ................................................... 14.2.3 Government Policies in Seven Selected Regions ... 14.3 Methods for Calculating, Comparing and Ranging Impacts........................................................... 14.3.1 Methods for Calculating Impacts ........................... 14.3.2 Methods for Comparing Impacts............................ 14.3.3 Methods for Ranging Impacts ................................ 14.4 Impacts of Policies: Most Important Sectors ...................... 14.4.1 The Government Re-Distributes Resources Regionally ............................................. 14.4.2 Macroeconomic Strategies Influence Regional Development .......................................................... 14.4.3 Important Other Sector Policies ............................. 14.5 Concluding Remarks ........................................................... Literature and References...............................................................

249 251 252 253 255 255 257 259 261 262 264 273 274 284 285 287 288 289 290 291 292 292 293 296 298 300 300 300 301 304 305

Contents

15

16

The Economics of Tree-planting for Carbon Mitigation: A Global Assessment .................................................................... Pablo C. Benítez, Ian McCallum, Michael Obersteiner and Yoshiki Yamagata 15.1 Introduction ......................................................................... 15.2 Carbon Sequestration Costs ................................................ 15.3 Data ..................................................................................... 15.3.1 Land Available for AR-Projects ............................. 15.3.2 Carbon Uptake........................................................ 15.3.3 Risk-Adjusted Discount Rates................................ 15.3.4 Prices ...................................................................... 15.4 Results ................................................................................. 15.4.1 Geographic Distribution of Carbon Sequestration Costs ................................................ 15.4.2 Global Carbon Supply ............................................ 15.4.3 Long-term Carbon Supply and Policy Implications .......................................... 15.5 Conclusions ......................................................................... References ...................................................................................... Positive Spillovers of Energy Policies on Natural Areas in Poland: an AGE Analysis ........................................................ Adriana M. Ignaciuk 16.1 Introduction ......................................................................... 16.2 Model Specification ............................................................ 16.3 Data ..................................................................................... 16.4 Scenarios ............................................................................. 16.5 Results and Discussion........................................................ 16.5.1 General Results....................................................... 16.5.2 Production .............................................................. 16.5.3 Prices ...................................................................... 16.5.4 Land Use................................................................. 16.6 Conclusions ......................................................................... References ......................................................................................

XIII

307

307 309 311 311 312 313 315 315 315 316 317 318 319 323 324 326 329 330 330 330 333 334 336 337 338

The Authors........................................................................................... 341

1 Regional Externalities: an Introduction

Wim Heijman Wageningen University, The Netherlands, E-mail: [email protected]

Generally speaking, externalities occur when a decision causes uncompensated costs or benefits to individuals or groups other than the person(s) making the decision. Examples of negative externalities are numerous in the area of the environment and natural resources. Some negative externalities result because a particular type of manufacturing technology is used (e.g. water and air pollution caused by industry). Other negative externalities occur because of the transportation system (e.g. air pollution caused by intensive car traffic). Though positive externalities draw less attention than negative externalities, their existence is obvious, for example, beekeepers who provide unpaid pollination services for nearby fruit growers or the positive network effects of a telephone system. The more people who own a telephone, the more useful the device is for each owner (Boardman et al., 2001). From a social planner’s perspective, the existence of externalities results in an economic process outcome that is not socially optimal because marginal costs of the product involved do not equal its price. This implies that, in a well functioning market economy, negative externalities cause too much of a product to be produced, whereas positive externalities cause too little of a product to be produced. In order to internalise externalities and create a Pareto optimal situation, government may tax/subsidise externalities by way of a Pigovian tax or (dis)encourage production in another way, for example, by environmental regulations or by the implementation of a system of tradable rights (Tietenberg, 2001). Theoretically, the general principle on which government policy with respect to externalities should be based is the internalisation of externalities in order to create a Pareto efficient situation. The attempt to internalise externalities by the government is, in general, problematic. On the one hand, trying to reach Pareto optimality through regulations is hampered by the lack of information on the monetary valua-

2

Wim Heijman

tion of the externalities. On the other hand, installing a market system of tradable rights may not be preferred because of high transaction costs connected to it (Coase, 1960). Furthermore, in a system of tradable rights, there is still the necessity to determine a maximum emission level which in itself is a political decision that cannot be left to the market. Despite the problems mentioned, it is generally accepted (and it cannot be avoided) that government plays a role in the internalisation of (regional) externalities. Regional externalities are a specific subset of externalities that can be defined as externalities where space plays a dominant role. This class of externalities can be divided into three categories: (1) externalities related to mobility and transport; (2) external economies of scale and cluster effects, and (3) spatial environmental externalities. This book offers examples of the above mentioned categories. With respect to mobility and transport, a number of externalities can be mentioned (Rietveld and Bruinsma, 1998). These are normally related to the construction of infra-structural works like roads, railroads, canals and ferries. On the positive side, we find the so-called network effects caused by the construction of an infrastructural element like a road which results in decreased use of and therefore a reduction of congestion of other infrastructure. On the negative side, we find regional compartmentalisation in densely populated areas resulting from the construction of a major infrastructural element like a highway or a railroad, or the closure of transportation facilities like ferries. As a result of comparmentalisation, certain regions that are not well connected to the infra-structural main grid may become isolated. Furthermore, more infra-structure may contribute to the degradation of nature and landscape and contribute to more mobility resulting in higher emissions of hazardous gasses like greenhouse gasses (such as CO2) and gasses that cause acidification (NOx and SOx) and a higher morbidity connected with traffic accidents. External economies of scale and cluster effects are generated when firms are located in each other’s vicinity. Starting with Marshall (1890) and Weber (1909), this idea has already had a long history and has been revitalised in more recent years in the framework of the New Economic Geography (Fujita and Krugman, 2004). The general idea is that spatial clustering of firms induces a reduction of costs per unit produced. This cost reduction may be caused by a geographic concentration of firms of the same kind causing external economies of scale, or by a geographical cluster of relating cooperating firms aiming at cost reduction and joint innovations of products and production processes (Porter, 2000). The consequence of this is that economic growth is a process that is spatially concentrated in centres that can be called growth poles, clusters,


3

cores, or agglomerations. This idea of spatially concentrated growth has been especially neglected by mainstream general economists, which is the reason why regional and general economics have developed along different lines (Krugman, 1995, 1997). It has been argued that telecommunication technology will diminish the advantages of geographical proximity of, in particular, small, specialised firms. However, this is probably not true. High knowledge-intensive activities, like the development of new products, not only need face-to-face contact but also “face to product” contact. Therefore, only standardised products may become footloose (Desrochers, 1998; Heijman et al., 2004). The last category of regional externalities, spatial environmental externalities, refers to externalities linked to ecosystems, bio-diversity, and genetic manipulation. Growing ecosystems such as forests are environmentally beneficial because they can absorb CO2 and as such partly compensate for the emission of this greenhouse gas in the industrialised part of the world (IPCC, 2000). Bio-diversity is another aspect of ecosystems that is currently drawing attention for two reasons. Firstly, biodiversity may be a positive externality in itself because it contributes to the preservation of some rare species, which is generally considered to be a positive policy goal. Secondly, bio-diversity contributes to the beauty and variety of landscapes generating economic benefits based on eco-tourism. If these positive externalities are not internalised, the lack of internalisation will result in large environmental and social costs (Hardie et al., 2004). Another type of externality can result from genetic modification. Producers of transgenic crops generate a negative externality for producers of transgenic free crops through potential pollen flow. Moreover, transgenic crops may influence the genomes of their wild relatives, which may influence the stability of vital eco-systems (Wesseler, 2005). This book is organised into three parts. The first part deals with externalities connected with transport and mobility. The second part includes contributions covering regional clusters and product chains. Finally, Part 3 covers issues of regional policy dealing with the internalisation of regional externalities. Part I of the book, “Transport”, contains four chapters (Chapters 2-5). In Chapter 2 by Morten Marott Larsen, Bjarne Madsen, and Chris JensenButler, the regional impacts of road pricing on the use of cars are analysed, taking into account the externality effects of transportation on wages and productivity. In the paper, the direct impacts of changes in transport costs on the level of wages and productivity (the direct externality effects) have been estimated. The direct and derived impacts of road pricing have been analysed using AKF’s local economic model LINE and include the impacts on regional production, income and employment. LINE is an interre-

4

Wim Heijman

gional general equilibrium model that uses an interregional social accounting matrix (SAM-K) and a regional transport satellite account as the basis for modelling. Additionally, data from a GIS-system on transport costs have been included to estimate the demand for transport commodities and increase in transport demand and costs due to road pricing. In Chapter 3, Jan Rouwendal deals with traffic congestion, which is an important example of an externality. Economic analysis of this phenomenon dates back to the first half of the twentieth century with the seminal contributions of Pigou and Knight. The Pigou-Knight model remains an important tool for the analysis of congestion. It is the purpose of this contribution to provide a brief discussion of this canonical model and to introduce two extensions. The first extension concerns the derivation of the travel cost curve, which is considered as an exogenously given relationship in the standard analysis, from a theory about speed choice. The second deals with the incorporation of heterogeneous traffic into the model. In Chapter 4, Pierre van Mouche, Jan Rouwendal and Willem Pijnappel present a further analysis of traffic congestion. In their contribution they deal with the so-called Mohring and Harwitz congestion model. Public roads are usually provided by the government and can be used freely by everyone. The capital invested in public roads is typically provided by public funds and the welfare losses associated with raising these funds can be substantial. If the external congestion effect is internalized by means of a Pigouvian toll, two birds are killed with one stone: an externality is removed and the need to use public funds decreases. Indeed, Mohring and Harwitz established in their congestion model that, under appropriate conditions, the revenues from congestion tolling are exactly equal to the cost of providing the optimal capacity of the road. Their result is nowadays considered one of the cornerstones of transportation economics. In the final chapter of Part 1, Chapter 5, Rinus Jaarsma and Wim Heijman analyse accessibility of the rural area in densely populated countries like the Netherlands. They state that the construction of major infra-structural works like motorways, highways, and high speed railways, and the reduction of ferry services lead to disruption in the present local road network. The fragmentation of rural roads and paths unintentionally affects the opportunities to drive, cycle or walk in the countryside. A fragmentation of rural regions may appear, especially for non-motorized users of the rural infrastructure, because of the reduced opportunities to cross major roads, railways, rivers, and canals. These negative externalities of the main infra-structural grid are often neglected in cost benefit analyses of large scale infra-structural works. The chapter describes the problem for the Netherlands and discusses ways to incorporate the externalities observed in the decision making process of major infra-structural works.


5

Part II of the book, “Clusters and product chains”, contains 6 chapters. Chapter 6 by Jana Gašparíková, Edita Nemcová and Michal Pálenik and Chapter 7 by Danes Brzica describe the regional differentiation of industrial capacity in the Slovak Republic and the Slovak car sector, respectively. Over the last years, regional growth dynamics in the Slovak Republic have changed rapidly. The latest statistical surveys show permanent disproportional concentration of industrial capacities. Chapter 6 focuses on the explanation of this phenomenon, concentrating on the electro-technical sector, the machinery sector, and chemical industries. Chapter 7 addresses the problem of persisting technological gaps vis-àvis the USA and Japan. The Slovak Republic, similar to other EU members, shows a relatively low ability to transform knowledge from basic research into new products. This has had a negative impact on employment and growth. The emergence of clusters can assist in making qualitative functional changes to the Slovak business sector. Chapter 7 deals with this process. It addresses the emergence of the Slovak automobile cluster(s) as a main topic. Chapter 8 by Vytautas Snieska, Regina Virvilaite, Vaida Kvainauskaite, Bronius Neverauskas, Rimantas Gatautis and Aiste Dovaliene presents the main aspects of interaction between network externalities, IT market, and e-commerce development under conditions of a transition economy. In the literature, it has often been found that network externalities cause a decrease of what is generally called transaction costs (the costs incurred in gathering information and controlling and coordinating transactions). With respect to the decrease of transaction costs, this paper shows that network externalities are extremely relevant for the Lithuanian IT market. In Chapter 9 Kees Burger and Rein Haagsma address an important aspect of globalization, i.e. the strong growth of international trade. What is striking in their analysis is that the trade in final products is becoming more and more dominated by the trade in parts and components. Because of new production techniques, better means of transportation, and the worldwide breakdown of tariff and non-tariff barriers, industrial firms are increasingly able to outsource part of their production process to cheaper producers or locations in foreign countries. Although this internationalisation of the production process is a universal phenomenon, it seems particularly relevant for the Netherlands, on which the chapter focuses. In Chapter 10, Frans Boekema and Roel Rutten contribute to the expanding body of literature on knowledge, learning, innovation, clusters, networks and space. Most publications concentrate on the knowledgebased character of today’s Western economies. In other words, learning and knowledge are the key to innovation and improving competitiveness. At the same time, firms depend on collaboration in networks to access

6

Wim Heijman

knowledge beyond their control. In fact, it is argued that these networks have an important spatial dimension. Finally, Chapter 11 by Volker Beckmann and Justus Wesseler concentrates on the spatial externalities of transgenic crops. Adopters of transgenic crops produce a negative externality through potential pollen flow for producers of transgenic free crops. Producers of transgenic free crops produce a negative externality for growers of transgenic crops if they are required to keep a minimum distance from transgenic free crops. This paper examines the current debate on the co-existence of transgenic and transgenic free crops from the perspective of Roland Coase’s influential paper “The Problem of Social Cost” (1960). Part III, “Regional Policy”, contains five papers. Chapter 12 by Wim Heijman and Johan van Ophem focuses on regional investment and domestic residential mobility. In the last several decades, commuting distances between work and home have substantially increased in all developed countries. However, from a societal and environmental point of view, jobs and dwellings should be brought closer together through residential mobility. This idea then raises the question: which type of public investment influences residential mobility of households and reduces commuting. In Chapter 13, Václav Beran and Petr Dlask present a model that aims at explaining the disparities in regional economic growth in the Czech Republic and their relevance for regional economic policy. The proposed specific regional development strategies are based on a compact model containing only a few variables. With the help of the model, the development of 13 regions in the Czech Republic is simulated. The model results show that innovations are the driving force behind regional development. In Chapter 14 Steinar Johansen describes and analyses macro policies and regional impacts in Norway. The impacts of sector policy measures on regional development can be positive or negative, but are normally not planned. Since these impacts are not intended, the impacts on regional policies are externalities of sector polities. This chapter focuses on a broad set of Norwegian fiscal and non-fiscal sector policies and discusses their impacts on regional development. Chapter 15 by Pablo César Benítez, Ian McCallum, Michael Obersteiner, and Yoshiki Yamagata provides a framework for identifying least-cost sites for carbon sequestration through tree-planting and deriving carbon cost curves at a global level in a scenario of limited information. Special attention is given to country risk considerations and the sensitivity to spatial datasets. Their model results show that most least-cost carbon uptake projects are located in Africa, South America and Asia. By comparing


7

emissions reductions through tree-planting with the emission abatement scenarios of integrated assessment models (RICE-99) for a 100-year time span, they find that global carbon uptake of planted forests could represent between 5% to 25% of the emissions reduction targets of relevant climate change mitigation scenarios. In Chapter 16, the final chapter of this book, Adriana Ignaciuk focuses on the role of biomass as the major provider of renewable energy in Poland. She sketches the impact of different energy policies on the development of bio-electricity and the connected land cover. Especially multi-product crops may increase the potential for bio-energy and have a positive impact on natural areas simultaneously.

Literature Boardman A.E., D.H. Greenberg, A.R. Vining, D.L. Weimer (2001). Cost-Benefit Analysis: Concepts and Practice. Prentice Hall, Upper Saddle River (NJ). Coase R.H. (1960). The Problem of Social Cost. Journal of Law and Economics, 3, pp. 1-44. Desrochers P. (1998). A geographical perspective on Austrian economics. The Quarterly Journal of Austrian Economics 1 (2), pp. 63-83. Fujita M. and P. Krugman (2004). The new economic geography: Past, Present and the future. Papers in Regional Science, 83, pp. 139-164. Hardie I.W., P.J. Parks, and G. Cornelis van Kooten (2004). Land use decisions and policy at the intensive and extensive margins. In: T. Tietenberg and H. Folmer, The International Yearbook of Environmental and Resource Economics 2004/2005: A survey of current issues. Edward Elgar, Cheltenham, UK. Heijman W.J.M. and A. Leen (2004). On Austrian regional economics. Papers in Regional Science, 83, pp. 487-493. Krugman (1995, 1997). Development, geography and economic theory. The MIT Press, Cambridge, MA. IPPC, 2000. Land Use, Land Use Change, and Forestry. Special Report Intergovernmental Panel on Climate Change (IPCC), Cambridge University Press, Cambridge, UK. Marshall A. (1890, 1922). Principles of economics. Macmillan, London. Porter M.E. (2000). Locations, Clusters, and Company Strategy. In: Gordon L. Clark, Maryann P. Feldman, and Meric S. Gertler (eds.), The Oxford Handbook of Economic Geography. Oxford University Press, New York. Rietveld P. and F. Bruinsma (1998). Is Transport Infrastructure Effective? Transport Infrastructure and Accessibility: Impacts on the Space Economy. Springer, Berlin. Weber A. (1909, 1962). Theory of the location of Industries. Oxford University Press, New York.

8

Wim Heijman

Tietenberg T. (2001). The implementation and evolution of emissions trading. Ashgate, Aldershot (UK). Wesseler J. (ed.) (2005), Environmental Costs and Benefits of Transgenic Crops. Wageningen UR Frontis Series Vol. 7, Springer, Dordrecht.

Part I Transport

2 Modelling Transport in an Interregional General Equilibrium Model with Externalities

Morten Marott Larsen1, Bjarne Madsen2 and Chris Jensen-Butler3 1 2 3

Institute of Local Government Studies – Denmark, AKF, E-mail: [email protected] Institute of Local Government Studies – Denmark, AKF, E-mail: [email protected] School of Economics and Finance, University of St. Andrews, Scotland, UK, E-mail: [email protected]

Abstract. In this chapter the regional economic impacts of road pricing on cars are analysed taking into account externality effects from transportation on wages and productivity. The direct impacts from changes in transport costs on level of wages and productivity (the direct externality effects) have been estimated. The direct and derived economic impacts of road pricing have been evaluated using the sub-regional economic model LINE. The direct effects on level of wages and productivity have been included in the model together with all the direct effects on commodity prices from road pricing. The total impacts of road pricing have been subdivided into three components: 1) The wage effects of reducing income net of commuting by increasing transport costs with introduction of road pricing, 2) the labour contraction effect arising from increasing wages through increases in commuting costs and 3) the negative productivity effects of introducing road pricing. In total the impacts of road pricing are substantial. Regions with high level of average commuting costs (suburban areas in Greater Copenhagen) suffer most, whereas the centre of Copenhagen suffers least because of short commuting distances. In rural areas impacts are on or just below the average because of low levels of road pricing. Keywords: Road pricing, Externalities, Transport costs, Regional economic model.

12

Morten Marott Larsen, Bjarne Madsen and Chris Jensen-Butler

2.1 Introduction It is customary to model transport in an interregional general equilibrium model without including the effects of both positive and negative externalities. In such a simplified model focus is usually directed at the price effects of direct changes in transport costs as the transport system is modified. Negative externalities such as congestion and environmental damage can be incorporated either as a pre-model extension where transport costs include impacts from congestion, or as a post-model where environmental damage is modelled as a function of level of economic activity and as a set of emission coefficients. This is normally straightforward, as the pre- and post models are often linked in a loosely coupled system. In such a system, by definition feedback effects from the economic model to the pre-model (from the economic system to congestion, for example) and from the postmodel to the economic model (from the environmental system to the economic system, for example) are not incorporated. The inclusion of positive externalities is a more difficult task because it usually involves integration of the externality into consumption and production behaviour. This is because positive externalities normally have both impacts on economic activity and are influenced by the level of economic activity. This means that transport system improvements increase concentration (both number and density of firms) which in turn increases their productivity. On the other hand, increasing productivity reduces prices, increases competitiveness and thereby export and economic activity. These feedback mechanisms imply that modelling positive externalities should involve the application of an interregional general equilibrium model with productivity as a fully integrated sub-model. In the following a theoretical approach to inclusion of positive externalities is examined. This is followed by an examination of the effects of transport system changes on economic activity, through an econometric study of Danish regions. In order to illustrate the distributional impacts on economic activity arising from positive externalities, an interregional general equilibrium model for Denmark is presented and results from a study using this model of the spillover and feedback effects of positive externalities are examined.

2 Modelling Transport in an Interregional General Equilibrium Model

13

2.2 Positive Externalities There is a considerable and growing interest in the effects of externalities on economic activity in space (Goodchild et al 2000, Anselin 2003a, 2003b, Fingleton 2003, Fingleton et al 2005, Ottaviano & Thisse 2004, Baldwin & Martin 2004). Externalities are inherently spatial, they are only to be found at certain locations, they are subject to distance decay and their effects are related to spatial densities of activity. Agglomeration economies are positive externalities associated with spatial concentration of economic activity, resulting in lower marginal and average costs and increases in productivity. These are scale economies which do not apply at the level of the individual firm, but at the level of the industry. It is therefore possible to retain the assumption of perfect competition whilst analysing the effects of externalities. It is usual to distinguish between locational economies, which arise from agglomeration of firms within the same industry and agglomeration or urban economies, which arise when firms in different industries agglomerate in the same (urban) area. Externalities can be best understood by examining the case of locational economies in relation to the industry supply curve. A constant cost industry has a horizontal supply curve, which is derived from the minimum point on the long-run average cost (LRAC) curve of (identical) small firms in the industry. As the industry expands, new firms are added and the supply curve remains horizontal. If, however, there are scale economies which are external to the firm but internal to the industry, then the arrival of each additional firm will mean that the minimum point on the firms’ LRAC curves will be lower and the industry LRAC and marginal cost curves slope down to the right, as shown in Figure 2.1. The supply curve of a declining cost industry becomes the LRAC curve rather than the marginal cost curve, as when price equals marginal cost, firms will make a loss. A further consequence is that an increase in demand (from D to D1 ) for this type of industry will result in both lower prices and increased output, as shown in Figure 2.1, which is a prime reason for policymakers’ interest in external scale economies. Urbanisation economies are related to the size of the local economy and in particular on positive effects, such as knowledge spillovers, between different industries. As noted in the literature, (Glaeser et al 1992, Engelstoft et al 2006) both localisation and urbanisation economies arise for a number of reasons including labour pooling, scale economies for intermediate inputs, development of ancillary trades and knowledge spillover effects. The microeconomic foundations of these economies are discussed in Duranton & Puga (2004) The theory of

14


economies of agglomeration builds upon the assumption that downward sloping industry supply curves is a spatial (regional or urban) phenomenon. Thus, spatial proximity magnifies these effects and in many cases is a condition for their operation.

£

D

D1

S1 =AC = MC

p

AC=S2

p1

MC

q

q1

Q (industry)

Fig. 2.1 Industry supply curves for a constant cost industry (S1) and a declining cost industry (S2).

2.2.1 Pecuniary and Technological Externalities It is important to distinguish between the two types of externality, pecuniary and technological (Scitovsky 1952). These differences are illustrated here using the case of a change in the transport system which involves changes in transport costs. Pecuniary externalities arising from changes in transport costs have their origins in both the commodity and factor markets. Technological externalities have their origins in direct interactions, outside the market, between producers (p) and consumers (c), giving in all four different combinations (p-p, c-c, p-c and c-p). Pecuniary externalities are illustrated in relation to the labour market in Figure 2.2.


15

W/P STC STC’ c

a e i d

f

S g h

b j DTC

o

l1 l2

L

Fig. 2.2 Pecuniary externalities associated with transport system improvements.

Demand (DTC) for and supply (STC) of labour are the standard curves which cross at equilibrium, c. Real wages both with and without transport costs, (TC) are shown in Figure 2.2. Seen from the viewpoint of the producer and the place of production, the demand for labour includes transport costs, DTC. There are two supply curves for labour, STC which show labour supply in relation to the real wage (W/P) including transport costs (from the point of view of place of production) and S which shows labour supply in relation to the real wage (W/P from a place of residence point of view), net of transport costs from place of residence point of view. It can be assumed, that S is fixed in the long run, whereas STC, which includes transport costs, will shift with changes in these costs. In the case of transport system improvements the STC curve shifts to the right (STC’). This leads to a fall in equilibrium wage from a to e, a change in total wage bill from acl1o to egl2o and a change in expenditure on transport for commuting from acdb to eghf. How much demand for transport increases depends on the price elasticity of demand for and the supply of labour. The pecuniary externality effect for the producer is acge and for labour fhdb.

16


W/P

STC

g

e a f b

S

c

h d DTC

O

l1

l2

L

Fig. 2.3 Technological externalities associated with transport system improvements.

Corresponding externalities arising from changes in the transport system appear in the market for commodities, which can be illustrated in a figure similar to Figure 2.2. Equilibrium in the commodity market is established in market prices, including the cost of transport of commodities. The demand curve for commodities intersects with the supply curve at market prices. From the viewpoint of the producer, supply is determined in basic prices, net of transport costs. Supply in market prices is determined by the use of an adding-on principle. A transport system improvement will reduce transport costs and will shift the supply curve in market prices to the right. Equilibrium prices fall, which affects the real wage (W/P). Thus, a change in transport costs will affect both W and P as shown on the vertical axis of Figure 2.2. However, it is reasonable to assume that for any given decrease in transport costs, W will fall more rapidly than P because commuting costs are more important for labour than transport costs for commodity prices. In addition, the labour force is inherently regional, whilst commodities typically are produced in other regions or abroad, implying that the effect of local transport system improvements are greater for W than for P.


17

In the case of technological externalities, shown in Figure 2.3, demand for labour at a given real wage increases because of non-market interaction effects, including knowledge spillover effects (Audretsch & Feldman, 2000), arising from transactions undertaken by other agents. In this case of technological change, the demand curve for labour including transport costs (DTC) shifts to the right, creating a new labour market equilibrium, as shown in Figure 2.3, moving from c to g. This means that demand for labour increases from l1 to l2 and the real wage increases from a to e. The wage bill increases from acl1o to egl2o. The change in total expenditure on transport is shown by cghd.

2.3 Proximity, Income and Externalities in Denmark The previous section outlined a priori expectations concerning the relationship between changes in the transport system and externality effects in relation to the labour market. First, there will be a positive relationship between changes in transport costs and the level of real wages in the case of changes in accessibility in the labour and commodity markets, this being the effect of pecuniary externalities (Figure 2.2). Second, there is a negative relationship between changes in transport costs and changes in the level of real wages in the case of changes in accessibility to urban centres (Figure 2.3). This is the effect of urban (technological) externalities. The analysis also revealed that it is important to distinguish between the direct effects of externalities (which appear at the place of production and in the unit of production) and the end user effects, which appear in the institutions such as households, including the rest of the world, and at the place of residence of these institutions. In this section, the results of an econometric study of real wages and externality effects in the Danish urban system are presented. In Section 5 an analysis of the redistribution of these effects to end users is presented, using the subregional model LINE, which is presented in Section 4. 2.3.1 Estimating Urban and Pecuniary Externality Effects: the Two Approaches Two different approaches to analysis of the relationships between proximity, incomes and externalities are dealt with in the following. In order to estimate the effects of urban externalities the key spatial unit is the place of production, as the type of externality is production-production. In the case of pecuniary externalities, the spatial units of interest are both place of

18


production and place of residence because demand and supply in the labour market involves these two components. The first approach is based upon the hypothesis that there is a positive externality affecting production associated with proximity to an urban centre and that the externality is subject to distance decay. Distance is measured from place of production to an urban centre. The externality is on the production side and producers are able to pay higher wages because workers are more productive near the urban centre because of knowledge spillovers. Firms located near the centre also benefit from the externality. Different measures of distance to centres are employed, these being described below. The second approach is based upon the hypothesis that commuting distance is positively related to wage levels. Commuting distance is from place of residence to place of production which can be interpreted as a workplace disamenity and therefore workers require a higher wage if they have a longer commuting distance. 2.3.2 Factors Creating Differences in Wage Levels, Not Involving Externality Effects In addition to distance, factors such as gender, age, education, and industry are often significant. Research confirms in the Danish case relationships between wage levels and age, education and gender(Albæk et al. 1999, Trigg and Madden 1995, and Berndt, 1991). If industry enters into the explanation, different sectors have different proportions of factor inputs (capital and labour) and may have better opportunities to exploit proximity advantages (cluster effects). The fact that wage levels are normally higher in the private sector (Berndt 1991), indicates that political decisions concerning production levels in the public sector are not based upon the wage equals the value of the marginal product principle. 2.3.3 The Data The Social Accounting Matrix for Danish Municipalities (SAM-K) is the main data source in this study, and a complete description of the data is to be found in Madsen et al. (2001b) and Madsen & Jensen-Butler (2005). Two main sub-sets of SAM-K are used, the data having its origins in register-based individual level data, which are then grouped by variable. Data used to estimate the urban externality effects has the following structure. The dependent variable is mean value of wages and salaries per person defined in relation to the categories used to group the individual values of the independent variables. Grouped data was used as data at individual


19

level was not available because of confidentiality requirements. The independent variables are principally category variables, representing, for each of the 275 municipalities, where place of production is located, grouped data. These data comprise age (3 groups), gender (2), educational qualification (5), industry (132), year (4) (to remove the effect of inflation on incomes). This gives 1,034,000 cells for each of the 4 years (275 x 3 x 2 x 5 x 132), in all 4,356.000 cells. 20.8% of these cells have non zero content with an average of 9.9 employees per observation In addition, there is a variable containing unemployment percentages for each age by sex by qualification by year for each municipality category (275 x 3 x 2 x 5 = 8.250 for each of the 4 years 33,000 possible different values). Finally, there is a distance variable, based upon location of the municipality, which has 275 possible different values. These distances are expressed as monetary values. Three different distance measures are used: i) to the urban centre in the (statistically defined) Danish labour market areas, ii) to the capital city Copenhagen, and iii) to the nearest of the five large university towns: Copenhagen, Århus, Odense, Aalborg, and Esbjerg. The dependent variable is mean value of wages and salaries per person (full-time equivalent) for each of the 66.000 cells for each of the 4 years. Data used to estimate the pecuniary externality effects have the following structure. The dependent variable is mean value of wages and salaries per person defined in relation to the categories used to define the values of the independent variables. Grouped data were used as data at individual level were not available. The independent variables are principally category variables, representing, for each combination of the 275 municipalities where place of production is located and each of the municipalities where place of residence is located. In the data set used to estimate pecuniary externalities, it should be noted that basic data relates to groups of individual data by place of residence. These data comprise age (3 groups), sex (2), qualification (5), year (4) (to remove the effect of inflation on incomes). This gives 2,268,750 cells for each of the 4 years (275 x 275 x 3 x 2 x 5), in all 9,075,000 cells over the 4 years. This contrasts with data used to estimate urban externalities, where the basic data refers to individuals grouped by industry and there is no information concerning place of residence in this data set. In addition, there is a variable containing unemployment percentages by place of residence for each age by sex by qualification by year category (275 x 3 x 2 x 5 = 8,250 cells for each of the 4 years. Finally, there is a distance variable, based upon the 275 x 275 intermunicipality distance matrix. The dependent variable is mean value of wages and salaries per person (Full-time equivalent) for each of the 2,268,750 observations per year (= 275 x 275 x 3 x 2 x 5).

20


The number of observations is of course much smaller than the number of cells. Especially in the data set on pecuniary externalities the majority of cells are empty and a number of observations have been eliminated for different reasons. The following observations have been eliminated from the data set used for the analysis of urban externalities: If place of work is abroad, not available, or located on the small island of Christiansø, and if age is below 15 or above 59. Furthermore, some extreme observations are removed, if the average wage is above 2 million DKK, and if total employment is under 5 working days in the year concerned. This all reduces the number of observations to 277,237 per year in the 4-year-period with an average of 9.0 employees per observation and a standard deviation of about 35 employees. About 1 per cent of the observations contain more than 100 employees. The highest number of employees in an observation is around 2,200. Similar reductions have been made for the data set used to analyse pecuniary externalities. The total number of observations here is 205,680 per year with an average of 9.9 employees per observation and a larger standard deviation of about 89 employees. Different measures of distance are applied in the regressions, but the source is the same and distances are measured at municipal level. The distance from one municipality to another is the number of kilometres between the main post office in every pair of municipalities. An intramunicipality distance is calculated including elements such as size and shape of the municipality. When crossing water, kilometres are not an appropriate measure because there would usually be higher costs involved. Therefore, the kilometres are transformed into Danish kroner using the assumption that one kilometre on land equals one krone. When crossing water the price of a ferry ticket is applied instead of kilometres. The distances are calculated with base in 1996. The distances could be defined in other ways taking other factors into account such as congestion, speed limits, time values, etc., but more accurate distances are not used here. 2.3.4 The Effects of Urban Externalities First, a real wage model with gender, education, sector, and unemployment as explanatory variables is set up and compared with a model which also includes municipalities as fixed effects. An F-test cannot reject the hypothesis that the constant terms are all equal and therefore fixed effects are 1

1

See appendix B, I).


21

left out. In the fixed effects model with time dummies, unemployment is insignificant which corresponds to the results in Albæk et al. (1999). Three measures of the centre are applied now in a model with gender, time, education, sector, and unemployment as explanatory variables. Fundamentally, more than one measure could enter into the model, but here it is assumed that there is only one type of centre in which one positive externality is present. Logarithmic transformation of all three measures describes the data better. Distance to the commuting centre is the worst measure compared with the two others because the adjusted R2 is smallest comparing the three models which only differ with respect to the distance measure used. This might be because a definition of 35 commuting centres is chosen where some of the centres are small islands where no positive externality in production would be expected. The second best measure is the distance to Copenhagen. Given the geography of Denmark one would expect that it could be difficult to identify some distances because of the many belts, and straits. Table 2.A.1 in appendix A contains the variable »distance5« this being the best of the three proposed measures. »Distance5« is the distance to one of the five university towns: Copenhagen, Århus, Odense, Aalborg, and Esbjerg. Whether or not to interpret it as distance to an economic centre or distance to a university city is a matter of choice. The test described regarding heteroscedasticity rejects the hypothesis that there is no heteroscedasticity in all three models with the explanatory variables of Table 2.A.1 in appendix A. The results of the FGLS regression are preferred to WLS and OLS because the t-value of ȕ is numerically smaller in the heteroscedasticity test. Because both the average wage and distances are in logarithmic form the estimated parameters are elasticities. However, the elasticity is small at -0.04. If an improvement in infrastructure could be interpreted as a shorter distance then a 10% improvement in infrastructure would result in a 0.4% higher wage. In the context of this study it is also a welfare gain because the higher wages are due to a positive externality. Even though the elasticity is small the total welfare gain is worth calculating. In 1999 1.5 million workers had a place of work outside the 5 centres and their average wage was 260,000 DKK. If all distances outside the centres were reduced by 10% there would on average per worker be a welfare gain of 1040 DKK. The total welfare gain would be 1.5 million workers times 1040 DKK; a total of 1.560 billion DKK. ($1 = ca 6 DKK) 2

2

See appendix C, I).

22


When dealing with infrastructure investment the positive externality associated with a centre is not the only benefit. Therefore the total welfare gain of the positive externality should be calculated as one of the benefits. When comparing the size of the estimates in the regression the most important contributions to the average wages are gender, education, certain sectors, and age. As mentioned the fixed effects model is rejected, which means that regional unemployment has a small but significant estimated parameter. An interpretation could be that higher unemployment lowers wages because of competition for vacancies. 2.3.5 Pecuniary Externalities: Wage Differentials and Commuting Distance The second hypothesis is that commuting distance affects wages. To test this, a regression analysis on data set 2 is carried out. Data set 2 has information about both place of residence and place of work. Both place of residence and place of work could be used as a fixed effect, but when comparing adjusted R2, place of work is chosen. However, an F-test does not support treating the municipalities as fixed effects and because of that they are abandoned. The same problems concerning grouped data are present in this regression. The test described rejects the hypothesis that there is no heteroscedasticity in all three models with the explanatory variables of Table 2.A.2 in appendix A. Again, the results of the FGLS regression are preferred to WLS and OLS because the t-value of ȕ is numerically smaller in the heteroscedasticity test. Gender, education, age, year, unemployment by place of residence and commuting distances are all significant in the model using FGLS, WLS, and OLS. Estimates and standard deviations are presented in appendix A, Table 2.A.2. Comparing the estimated parameter with the regression using data set 1, the estimated parameters of gender have increased by 17% (using FGLS) and other changes have also occurred. The estimated parameter of unemployment is still small, though it has increased. An explanation could be that unemployment by place of residence is used and in the regression using the first data set unemployment by place of work is used. Commuting distance has a positive effect on wages. The estimated parameter is around 0.03 (using FGLS), which means that if commuting distance doubles the average wage would increase by 3%. 3

4

3 4

See appendix B, II). See appendix C, II).


23

2.4 The LINE-Model – Modelling Externalities and Transport Transport system changes have a direct impact on the costs of transportation, either by reducing transport costs as in the case of transport system improvements or by increasing transport costs as in the case of taxes on transport activities such as road pricing. Changes in transport costs have in turn direct effects on commodity prices and income: Transport cost changes influence directly the prices of commodities, because transport cost is a gross margin added to commodity trade. Transport costs are an addition to the price when shopping for commodities or when consuming as a tourist. In both cases the direct changes in transport costs are added to the price of the commodity after transportation to the buyer. Changes in transport costs also have a direct impact on disposable income net of commuting costs. In addition, as discussed above, transport costs also have an influence on wage levels and productivity through the effects of pecuniary and technological externalities. These changes in transport costs have in turn direct effects on commodity prices and income. The effects derived from externalities add to the direct effects on the regional economy of changes in transport costs. These direct effects on commodity prices and income, which now include the effects derived from externalities, lead to a number of derived effects on the regional economy. The distribution and magnitude of the total effects are, however, not the same as the distribution and magnitude of the direct effects (including the externality effects). The direct effects on commodity prices and income are redistributed through the interregional markets for commodities and production factors, through intra- and interregional trade, shopping and tourism, and through commuting and disposable income, assuming that price changes are transferred directly to the consumer. When prices and income change, the end user reacts by adjusting demand, which influences real economic activity. Therefore the derived effects (being the indirect and induced effects) should be added to the direct effects (including the externality effects) in order to estimate the total effects on regional economic activity. In order to estimate these total effects, it is necessary to construct an interregional/subregional general equilibrium model. A subregional general equilibrium model includes the effects of changes in commodity prices and income arising from transport cost changes, including the externality effects, and also includes the real economic reactions to changes in prices and income. The reason for using an interregional general equilibrium

24


model – and not just a national general equilibrium model- is twofold. First, the regional results are of importance. Second, modelling of the impacts on prices of changes in transport costs is a function of the regional pattern of interaction in the economy. A national model without a spatial dimension would not capture the changes in costs and prices derived from the spatial pattern of economic interaction. In this section, an interregional/subregional general equilibrium model for Denmark, LINE, is briefly described. First, the real circle in LINE is presented which includes modelling the impacts of commodity price changes and changes in wages on real economic activity. This is followed by a presentation of the cost-price circle, which includes modelling of the changes in regional commodity prices and income. Finally, the inclusion of both pecuniary and technological externalities in LINE is considered. 2.4.1 LINE – an Interregional General Equilibrium Model for Modelling Redistribution of Productivity Changes The full model and its equations are described in detail in Madsen et al. (2001a) and Madsen & Jensen-Butler (2004). The data used in the model, together with the interregional SAM, are described in Madsen & JensenButler (2005) and Madsen et al (2001b). LINE is based upon two interrelated circles: a real Keynesian circuit and a dual cost-price circuit. In the Keynesian circuit the well known effects from demand on supply and income and from income to demand are included. In the cost-price circuit the spillover and feedback effects of cost and price changes using an adding-up principle based upon the assumption of perfect competition are modelled. Figure 2.4 shows the general model structure, depicting the real circle employed in LINE. The horizontal dimension is spatial: place of production (P), place of residence (R) and place of commodity market (S). Production activity is related to place of production. Factor rewards and income to institutions are related to place of residence and demand for commodities is assigned to place of commodity market. The vertical dimension follows with its three-fold division the general structure of a SAM model. Production is related to activities (J); factor incomes are related to factors of production with labour classified by gender, age and education (G) and type of household (H), commodities are related to the supply and demand for commodities (I). The real circuit corresponds to a straightforward Keynesian model and moves clockwise in Figure 2.4. Starting in the upper left corner (PJ), pro-


25

5

duction generates factor incomes in basic prices, including the part of income used to pay commuting costs. This factor income is transformed from sectors (J) to gender, age and educational groups (G) and households (H) and from place of production (P) to place of residence (R) through a commuting model. Employment follows the same path. Employment and unemployment are determined at place of residence (R). In addition to other adjustments, taxes are deducted from factor income and transfers added, giving disposable income, which by definition is related to place of residence (RH) . Disposable income is the basis for determination of private consumption in market prices, by place of residence (RH). Private consumption is divided into tourism (domestic and international) and local private consumption and assigned to place of commodity market (SI) using a shopping model for local private consumption and a travel model for domestic tourism. Private consumption, together with intermediate consumption, public consumption and investments constitute the total local demand for commodities in market prices (SI). The market price variables are transformed into basic prices through a use matrix, including information on the commodity composition of demand and commodity tax rates and trade margin shares. In this transformation from market prices to basic prices commodity taxes and trade margins are subtracted. Local demand is met by imports from other regions and abroad in addition to local production (SI). Through a trade model exports to other regions and production for the region itself is determined (PI). Adding export abroad, gross output by commodity is determined (PI). Through a reverse make matrix the cycle returns to production by sector (PJ). Economic activity in the real circle is affected by changes in prices and wages: wages and productivity affect prices of the local production (PJ), which through relative changes in local competitiveness affects exports (PI) and imports (SI) which in turn affects private consumption through changes in real disposable income (RH). The anticlockwise cost/price circuit shown in Figure 2.5 corresponds to this dual problem. In the costprice circle, production and demand are calculated in current prices, which in turn are transformed into relevant price indices. In the upper left corner production in current prices (in basic prices) is determined by costs (intermediate consumption, value added and indirect taxes, net in relation 6

5

Factor income includes modelling different income concepts such as gross value added, GDP at factor cost and primary income. 6 For simplicity in figures 4 and 5 modelling interactions inside the institutional sector between the household and governmental sectors, including taxes and subsidies, are not shown.

26


to production - PJ). Through a make matrix, sector prices by sector are transformed into sector prices by commodity (PI). These are then transformed from place of production to place of demand (SI) and further into market prices through inclusion of retailing and wholesaling costs and indirect taxes. This transformation takes place using a reverse use matrix. Commodities for intermediate consumption enter into the next step in the production chain, determining prices of production and these prices are spread further in a round-by-round distribution process. Finally, private consumption is transformed from place of commodity market (SI) to place of residence in market prices (RI). 2.4.2 Direct Wage and Price Effects in LINE As shown above, wages and prices will change as transport costs change. There are two direct effects, one which is truly direct, and one which arises through creation or reinforcement of externalities which in turn has an impact on prices. In Figure 2.5 the two types of direct effect on wages and prices are shown. The true direct effect is shown by an ellipse with dark shading, whilst the effects which operate through externalities are shown by an ellipse with lighter shading. The first effect enters directly into the price circle as an addition to costs and prices: In Figure 2.5 an increase in commuting costs implies that disposable income is reduced directly. In addition, commodity prices increase due to changes in transport costs for regional and interregional trade and also for changes in transport costs for shopping and tourism. The second effect works through externalities, both technological and pecuniary, which influence equilibrium wages and prices. In 2.5 a change in transport costs changes the size of the labour market and thereby the equilibrium wage. Changes in wages lead to changes in prices of production and in prices of commodities, following the logic of the cost-price circle. A change in transport costs will also affect the level of urban externalities and thereby wages and prices. This in turn changes the prices of commodities following the logic of the cost-price circle.


Place of production (P)

Place of residence (R)

27

Commodity market place (S)

Productivity

Sectors (J)

Wages / Prices

Production (Basic prices) (PJ)

factors of production (G)

Commuting

Disposable income (RG)

Shopping/ Tourism

Commodities (I)

Prices

Intra- & Interregional trade

Export to abroad (PI)

Demand (Market prices) (SI)

Import from abroad (SI) Pri ce

Fig. 2.4. The real circle in LINE – the impacts of price and wage changes on economic activity and behaviour.

28


Place of production (P)

Place of residence (R)

¨ transport costs

Sectors (J)

Urban externalities

Production (Basic prices) (PJ)

Pecuniary externalities

Factors of production (G)

Commuting

Disposable income (RG)

Shopping /Tourism

Commodities (I)

transport costs

Intra- & Interregional trade

Export to abroad (PI)

Demand (Market prices) (SI)

Import from abroad (SI)

Fig.2.5 The cost-price circle in LINE. The impacts of transport cost changes on prices and wages.

It should be noted that this discussion and illustration of the effect of externalities is limited to the consequences of changes in transport costs for wages, as shown in Figure 2.5. Similar effects could be illustrated in relation to the commodity market, which have not yet been included in LINE. The indirect and induced (derived) effects of wage and price changes depend in the medium term on exchange rate regimes. Foreign export and import prices enter into the determination of real foreign exports and imports at the regional level. If the economy is based upon fixed exchange rates, it matters whether or not a change in real wages leads to changes in


29

nominal wages or changes in prices or a combination of these two. If wages change, assuming sticky prices, then competitiveness of domestic production is unchanged and the wage change will only have a minor impact on economic activity. On the other hand, if prices change, assuming sticky wages, then competitiveness of domestic production and international exports and imports will react with correspondingly more substantial effects on economic activity. If the economy is based upon floating exchange rates, then changes in competitiveness due to changing prices will be more moderate as price changes will tend to be neutralised by exchange rate fluctuations. 2.4.3 LINE and SAM-K: Configuration The data used in LINE is SAM-K, which is an interregional SAM. SAM-K can in principle be established at a very low level of spatial disaggregation, the level of the municipality, with 133 sectors and 2,850 commodities. For both theoretical and practical reasons SAM-K and LINE have been aggregated and reconfigured. Aggregate relations are, in general, more stable and robust. Also, working at an aggregate level makes fewer demands on computer capacity. The first question to be decided is which axes both in terms of geography and social accounting (SAM) should be used, and which should be aggregated away. The second question concerns the choice of aggregation of axes in the detailed database. In this version of LINE the model configuration is the following. Sectors (J): 21 sectors aggregated from the 133 sectors used in the national accounts. Factors (G): 7 age, 2 gender and 5 qualification groups. Households (H): 4 types, based upon household composition Needs: For private consumption and governmental individual consumption 13 components, aggregated from the 72 components in the detailed national accounts. For governmental consumption, 8 groups. For gross fixed capital formation, 10 components. Commodities (I): 27 commodities, aggregated from 131 commodities used in the national accounts. For SAM-K and LINE sectoral and commodity aggregations have been defined in relation to the transport focus in the analysis.

30


Regions: 277 municipalities, including one state-owned island and one unit for extraregional activities, this being the lowest level of spatial disaggregation. Regions are defined either as place of production, place of residence or as place of commodity market. In this version of LINE the 277 municipalities have been aggregated into 16 regional units including one unit for extra-regional activities (see Figure 2.6). It should be noted that the econometric analysis of proximity, income and externalities in Section 3 has been undertaken at a more detailed geographical level (municipalities) than the aggregation used in this modelling exercise, where counties have been used.

2.5 Modelling the Redistribution of the Relative Productivity Decline Associated with Road Pricing – the Case of Denmark The main aim of this part of the analysis is to examine the consequences of including externalities in an analysis of the effects of a change in transport costs on regional economic activity. To demonstrate these effects, the hypothetical case of the introduction of road pricing in Denmark is examined. First, the road pricing scheme is presented, followed by a description of the way in which transport costs are affected. Second, the effects of transport costs on prices and costs with and without externalities are presented. Here the results of the estimation of urban and pecuniary externality effects have been applied. The total effects on real economic activity arising from changes in costs and prices divided into direct price effects and externality effects are estimated in Section 5.3. Finally, the methodological advantages of using an interregional general equilibrium model LINE, to analyse the national and regional impacts of road pricing – with and without externalities – are examined.


31

org öteb To G

G

Bornholm Nordjyllands

Viborg Århus

F

Ringkøbing

Frederiksborg Vejle

E

Copenhagen Vestsjællands Roskilde

Ribe

Copenhagen M Frederiksberg M

To Malmø

A

B Fyns Storstrøms

D

C

To

Ger man y

Sønderjyllands

Fig. 2.6 Danish regions Note: Danish regions are counties and two municipalities with county status, Copenhagen M and Frederiksberg M. Three fixed links: A: Oresund, to Malmo, Sweden, B: Great Belt, C: Femer Belt to Germany. Four ferry routes: D : Spodsbjerg-Taars, E: Odden-Aarhus (Mols Line), F: OddenEbeltoft (Mols Line), G: Frederikshavn-Gothenburg (Sweden). Other very local and international ferry routes are not shown.

2.5.1 Road Pricing and Changes in Transport Costs The design of road pricing systems relates to general issues concerning transport policy and technical constraints and possibilities, in concrete institutional and cultural contexts. Road pricing systems are discussed in JensenButler et al. (2005) and Madsen et al (2005). In this study it is assumed that a GPS-based/vehicle metered road pricing system is introduced throughout Denmark permitting precise identification of the location of a vehicle and thereby its road use, related in turn to toll level for the road. A range of different toll levels could be chosen, depending on type of vehicle, time of day, type of road, location, level of congestion etc. Here a simple assumption is made that tolls are set for cars only at DKK 0.6 per kilometre in urban areas

32


and DKK 0.3 per kilometre in rural areas, on a 24 hour basis and irrespective of which type of road, car or level of congestion is involved. Changes in transport costs are calculated using a) an interregional satellite account for transport used to determine levels of transport costs and b) exogenously given interregional transport costs, based on a digital road map, Vejnet DK, used to calculate changes in transport costs. The data in the interregional satellite accounts are estimated in four steps: a1) Taking the national make and use tables, national transport activity is determined by i) transport mode ii) subdivided by transport costs related to intermediate consumption (by sector) and to private consumption (by component) and iii) by external (transport firm based) and internal (own transport, within a non-transport producing firm or a household) costs. Six different modes of transport activity are used in the interregional satellite accounts, four for passenger transport and two for goods. Passenger transport is divided into car, rail, air and other and freight is divided into lorry and rail. a2) In the second step national transport activity related to passenger transport is subdivided (using data from the National Travel Survey) by trip purpose: i) commuting ii) shopping, iii) tourism iv) business travel and v) recreation. a3) National transport activity is then divided by origin and destination using data on intra and inter-regional trade (freight and business trip transport activity) and interregional shopping, tourism and commuting (personal trip transport activity) a4) Regional transport activity is then corrected (using regionalised National Travel Survey data) to ensure that the data reflect regional transport activity by mode. Changes in interregional transport activity for car transportation are estimated using the digital road map Vejnet DK. Transport costs in Vejnet DK are based upon both time and distance where the generalised cost has been calculated as time costs plus distance costs. Also included are costs (tickets, tolls) of travelling by ferry and using fixed links. In addition, costs are calculated both with and without road pricing. The calculations are based on assumptions shown in Table 2.1.


33

Table 2.1 Maximum speeds, distance and time costs, road pricing tariffs (DKK) (DKK 6=ca $1).

Car

Lorry

Motorway

110 km/t

80 km/h

Non-urban highway

80 km/t

70 km/h

Urban

50 km/t or local restrictions taken from VejnetDK

Max 50 km/h or local restrictions if under 50 km/hour

Distance cost per kilometre

1.82 DKK

2.60 DKK

Time cost pr. Hour

0.75

2.78 DKK.

Road pricing – Urban

0.60

-

Road pricing – Rural

0.30

-

The estimation of level of transport activity by region, mode, purpose and by type of consumption described above reflects a basic assumption that data on transport activity obtained from National (transport satellite) Accounts used in a top down procedure, are superior to data on transport activity obtained from different statistical sources, used in a bottom-up approach. In this chapter, only results involving road pricing for private cars is presented. Table 2.2 shows the consequences of introducing road pricing for costs of transport between four representative type of region. Copenhagen and Frederiksberg are the core of the Metropolitan region, Vestsjœlland is the semi-rural hinterland of Greater Copenhagen, Sønderjylland is in the semirural periphery of Jutland and Aarhus is a free standing city in Jutland. The full table has 15 x 15 regions. It is assumed that all ferry routes and fixed links will continue with unchanged ticket prices. Transport costs in general increase from 2% to 13% outside the main cities of Copenhagen and Aarhus, least in the interregional links where use of rural roads is important or where a significant part of the journey uses ferries. In Copenhagen and Aarhus transport costs decline as road pricing results in a reduction of congestion, and thus, transport costs. 2.5.2 Effects of Transport Costs on Prices and Costs, With and Without Externalities In this section the results of the changes in transport costs on costs and prices are presented. First, the results of a calculation without externality effects are shown in Table 2.3a. Then the consequences of including pecuniary externalities (Table 2.3b) and urban externalities (Table 2.3c) are

34


presented. Results for four representative regions out of the 16 and the total changes are shown. Table 2.2 Changes in total transport costs (index: 1.00 =unchanged) for transport between Danish regions after road pricing for cars.

(%)

CF

VS

SJ

AH

Copenhagen & Frederiksberg (CF)

0.91

1.06

1.07

1.01

Vestsjællands (VS)

1.06

1.12

1.08

1.00

Sønderjylland (SJ)

1.07

1.08

1.15

1.10

Aarhus (AH)

1.01

1.00

1.10

0.97

The analyses begin in the interaction components of the cost-price circle shown in Figure 2.5. In Tables 2.3a-c columns 1-10 reflect the cost price circle (Figure 2.5) moving anti-clockwise. Columns 1-3 show the impacts on production by sector and at the place of production (Cell PJ in Figure 2.5). Then impacts on commodity prices by commodity and by place of production are presented in columns 4-5 (cell PI in Figure 2.5). Continuing, columns 6-9 show the impacts by commodity and by place of commodity market (cell SI). Finally column 10 shows the impact on private consumption by commodity and by place of residence (cell RH in Figure 2.5). Starting with the demand for commodities at the place of commodity market the prices of commodities increase in total by 0.05% (column 8), distributed regionally as indicated in the table. Prices increase most in rural areas of the country, whilst prices actually decline in Greater Copenhagen because of declining congestion. Given the point at which the analysis commences, the presentation follows the cost-price circle: private consumption at the place of commodity market (column 9) and private consumption at the place of residence (column 10). Price increases are still moderate at the place of commodity market (at national level 0.04%), but markedly higher at the place of residence (at national level 0.42). This reflects the fact, that impact from cost increases on cars in trade is limited, because most transport on intra- and interregional trade typically is by lorry. But for shopping the transport cost increases from road pricing on cars are much higher, because private cars are used much more frequently for this purpose.


35

Looking at intermediate consumption, price increases are again moderate (column 2) reflecting the fact that most transportation related to intermediate consumption is by lorry, both in trade and in shopping (intermediate consumption from the wholesaler to the place of production). Therefore the impacts on the gross output deflator (column 3) and in turn on foreign export (column 5) are low. The regional distribution of changes in costs and prices are similar, for each column/regional economic variable, reflecting the fact that the direct impact of changes in transport costs reflect decreasing costs and prices in urban areas and increases in rural areas. Examining the impacts from externalities, these are generated in a completely different way (see Table 2.3b and 2.3c): Here the impacts originate from changes in the Gross Value Added deflator (column 1) and then commence through the cost-price circle (column 2-10). Here the consequences for the price of production of all type of commodities are influenced through the wage impact. GVA In the case of pecuniary externalities (Table 2.3b, column 1) the GVA-deflator increases with 0.21% at the national level but only with 0.03% in the case of urban externalities (Table 2.3c, column 1). The regional pattern follows the one from changes in transport costs (Table 2.2). As a consequence all prices change accordingly. In this case there is no difference between the impacts on private consumption at the place of commodity market and the place of residence. 2.5.3 Effects on Real Economic Activity from Changes in Cost and Prices, With and Without Externalities Real economic activity is influenced by the changes in prices. In addition, the way the revenues from road pricing is recycled, for example either through reduction in taxation or through increases in public consumption, will influence the level of economic activity. In this study only the gross impacts of cost and price changes are presented as the focus of the study is the multiplier effects of inclusion of externalities in the analysis. For a treatment of the public sector see Madsen et al (2005). In this section the results of changes in real economic activity are described. First, the results of a model calculation without externality effects are shown in Table 2.4a. This is followed by an analysis of the consequences of inclusion of pecuniary externalities (Table 2.4b) and then of urban externalities (Table 2.4c). In Tables 2.4a-2.4c columns 1-10 reflect the real circle (see 2.4) moving clockwise. Columns 1-2 show the impacts on disposable income by place of residence (cell RH in Figure 2.4). In column 3 the impacts on private consumption by commodity and by place of residence (cell RI in Figure 2.4) are shown. In columns 4-7 the impacts on

36


private consumption, local demand, foreign imports and demand for domestic production by place of commodity market (cell SI in Figure 2.4) are presented. Continuing in column 8 the impacts on foreign exports by commodity and by place of production are shown (cell PI in Figure 2.4). Finally, columns 9-10 show the impacts on gross output and GDP at factor prices by place of production (cell PJ in Figure 2.4). The first step, where externality effects are not included, examines the impacts on disposable income and private consumption (Table 2.4a, column 1-4). Disposable income in current prices is reduced in average by 0.21% (column 1) because of the income reduction arising from road pricing on commuting. Second, real disposable income is reduced, because prices of private consumption increase. In total, real disposable income is reduced by 0.96% (column 2), which in turn reduces private consumption by 0.96% both at place of residence (column 3) and at place of commodity market (column 4). Demand is reduced, but proportionally less because the real impacts on other components of demand (intermediate consumption) is smaller or even unchanged (public consumption and investments). Exports to abroad are reduced because of increases in relative export prices (column 8). From this follows that gross output and GDP at factor prices are reduced (columns 9 and 10). From a regional perspective Copenhagen and Frederiksberg are scarcely affected whereas Sonderjylland and other parts of Jutland face high price increases resulting in a reduction in private consumption and exports. When including the externality effects (see Tables 2.4b and 2.4c) there are two consequences, which have opposite signs: First, increases in income arising from wage increases related to road pricing (wage compensation), which increase private consumption and demand. Second, wage increases raise export prices and domestic prices which means that exports to abroad decline and imports from abroad increase, reducing domestic production. The second effect, based upon declining competitiveness dominates. The reduction in production will be even greater when including the externality effects. The next step is to include pecuniary and technological externalities in the direct effects. The impact of these affects real wages, as shown above. The effects going from real wages to GVA are presented in column (1) of Table 2.4b (pecuniary externalities) and Table 2.4c (urban externalities). The effects are transmitted through the economic system in the same way as described in relation to Table 2.4a. Table 2.5 shows the real economic consequences of both types of direct transport cost changes. The table shows the employment effect, which is closely related to changes in production, export, demand private consumption and disposable income. All these real effects have been derived from


37

changes in costs and prices which influence demand: the total effect is a reduction in employment of 8423 when externalities are not included and 12654 when they are. Externalities add therefore approximately 50% to the pure direct effect on commodity prices and income.

2.6 Conclusions In the study, an analysis on the impacts of changing transportation by introducing road pricing for cars in Denmark is presented. In the analysis, the impacts from externalities have been included, generating substantially higher gross effects of changing transport costs. The direct effects of changing transport costs on the labour market (pecuniary externalities or labour market enlargement effects) and on productivity (urban externalities through positive technological spillovers) are presented. The results of an econometric study in Denmark are presented. The derived effects were modelled using an interregional general equilibrium for Denmark, LINE. The analysis shows that due to reduction in congestion the Greater Copenhagen area benefits through reduction in the transport costs, whereas rural areas suffer, because of long commuting distances. The total effects on employment are a reduction of approximately 8000 including the conventional commodity price and income reducing effects from road pricing and this reduction is approximately 4000 greater if externality effects are included. This demonstrates that conventional analysis is insufficient, because real impacts are underestimated. There is a need to include externality effects in analysis of transport system changes.

38

Table 2.3a Cost and price changes for production, demand, export, import and private consumption - without externalities (2) Intermediate consumption

(PJ) 0.00 0.00 0.00 0.00 0.00

(PJ) 0.00 0.08 0.10 0.05 0.05

(3) Gross output

(4) Gross output

(5) Foreign export

(PJ) 0.02 0.11 0.12 0.06 0.08

(PI) 0.01 0.10 0.10 0.04 0.06

(PI) 0.04 0.25 0.18 0.11 0.14

(6) Demand: Domestic production (SI) -0.02 0.10 0.13 0.05 0.06

(7) Foreign Import

(8) Demand

(SI) 0.00 -0.01 -0.01 0.01 0.00

(SI) -0.01 0.08 0.10 0.04 0.05

(9) Private consumption Place of market place (SI) 0.01 0.06 0.08 0.03 0.04

(10) Private consumption Place of residence (RH) -0.13 0.60 0.77 -0.04 0.42

CF: Copenhagen & Frederiksberg, VS: Vestsjælland, SJ: Sønderjylland, AH: Aarhus.

Table 2.3b Cost and price changes for production, demand, export, import and private consumption labour market negotiations (Pecuniary externalities)

CF VS SJ AH Total

(1) Supply: Gross Value Added

(2) Intermediate consumption

(3) Gross output

(4) Gross output

(5) Foreign export

(6) Demand: Domestic production

(7) Foreign Import

(8) Demand

(PJ) -0.15 0.37 0.46 -0.05 0.21

(PJ) 0.01 0.19 0.23 0.05 0.15

(PJ) -0.05 0.35 0.42 0.04 0.24

(PI) -0.06 0.32 0.38 0.01 0.21

(PI) -0.01 0.47 0.40 0.06 0.31

(SI) -0.02 0.27 0.33 0.03 0.20

(SI) -0.01 -0.01 -0.01 -0.01 0.00

(SI) -0.02 0.22 0.28 0.03 0.16


(9) Private consumption Place of market place (SI) -0.02 0.19 0.24 0.01 0.15

(10) Private consumption Place of residence (RH) 0.00 0.19 0.24 0.02 0.15


CF VS SJ AH Total

(1) Supply: Gross Value Added

Table 2.3c Cost and price changes for production, demand, export, import and private consumption Urban agglomeration (urban externalities)

CF VS SJ AH Total

-0.16 0.14 0.17 0.02 0.03

(2) Intermediate consumption

(PJ) -0.04 0.06 0.08 0.03 0.03

(3) Gross output

(4) Gross output

(5) Foreign export


(7) Foreign Import

(PJ) -0.09 0.15 0.17 0.05 0.06

(PI) -0.09 0.11 0.13 0.04 0.04

(PI) 0.08 -0.28 -0.09 0.04 -0.03

(SI) -0.08 0.08 0.10 0.03 0.03

(SI) 0.00 0.00 0.00 0.00 0.00

(8) Demand

(SI) -0.07 0.06 0.09 0.03 0.03

(9) Private consumption Place of market place (SI) -0.04 0.04 0.06 0.02 0.02

(10) Private consumption Place of residence (RH) -0.03 0.03 0.05 0.02 0.02


Table 2.4a Consequences for demand, production and income with road pricing – without externalities

CF VS SJ AH Total

(1) Disposable income (current prices)

(2) Real disposable income

(RH) 0.01 -0.29 -0.40 0.00 -0.21

(RH) 0.18 -1.38 -2.00 0.06 -0.96

(3) Private consumption Place of residence (RI) 0.18 -1.38 -2.01 0.06 -0.96

(4) Private consumption Place of demand (SI) -0.05 -1.34 -1.95 0.00 -0.96


(5) Demand

(6) Foreign imports

(SI) -0.05 -0.56 -0.68 -0.05 -0.36

(SI) -0.05 -0.53 -0.55 -0.03 -0.31


(SI) -0.05 -0.56 -0.67 -0.06 -0.37

(8) Foreign export

(9) Gross output

(10) GDP at factor prices

(PI) -0.03 -0.36 -0.24 -0.15 -0.19

(PJ) -0.10 -0.49 -0.55 -0.12 -0.33

(PJ) -0.08 -0.54 -0.64 -0.08 -0.35


(1) Supply: Gross Value Added (PJ)

39

40

Table 2.4b Consequences for demand, production and income with road pricing labour market negotiations (pecuniary externalities) (2) Real disposable income (RH) 0.01 -0.03 -0.02 -0.03 -0.03

(3) Private consumption Place of residence (RI) 0.01 -0.03 -0.01 -0.03 -0.03

(4) Private consumption Place of demand (SI) -0.01 -0.03 -0.01 -0.03 -0.03

(5) Demand

(6) Foreign imports


(8) Foreign export

(9) Gross output


(SI) 0.00 -0.13 -0.13 -0.03 -0.09

(SI) 0.04 -0.11 -0.11 0.02 -0.05

(SI) -0.01 -0.14 -0.16 -0.04 -0.10

(PI) 0.01 -0.82 -0.63 -0.09 -0.51

(PJ) -0.02 -0.23 -0.24 -0.07 -0.17

(PJ) 0.00 -0.16 -0.18 -0.03 -0.12


Table 2.4c Consequences for demand, production and income with road pricing urban agglomeration (technological externalities)

CF VS SJ AH Total

(1) Disposable income (current prices) (RH) -0.06 0.03 0.05 -0.01 0.01

(2) Real disposable income (RH) -0.01 -0.02 -0.02 -0.03 -0.02

(3) Private consumption Place of residence (RI) -0.01 -0.02 -0.02 -0.03 -0.02

(4) Private consumption Place of demand (SI) -0.01 -0.02 -0.03 -0.03 -0.02


(5) Demand

(SI) 0.01 -0.08 -0.09 -0.03 -0.04

(6) Foreign imports


(8) Foreign export

(9) Gross output


(SI) 0.02 -0.07 -0.06 0.02 -0.02

(SI) 0.01 -0.08 -0.10 -0.04 -0.05

(PI) 0.02 -0.19 -0.19 -0.09 -0.12

(PJ) 0.02 -0.16 -0.16 -0.07 -0.08

(PJ) 0.02 -0.10 -0.11 -0.03 -0.05


CF VS SJ AH Total

(1) Disposable income (current prices) (RH) 0.00 0.15 0.22 -0.01 0.11

Table 2.5 Impacts on employment at place of production (2)

(3)

(4)

(5)

(6)

(7)

(8)

Labor market negotiation

Urban agglomeration

With externality

Without externality

Labor market negotiation

Urban agglomeration

With externality

Number

Number

Number

Number

- pct -

CF

-374

-20

59

-335

-0.09

-0.01

0.01

-0.08

VS

-565

-167

-103

-835

-0.44

-0.13

-0.08

-0.65

SJ

-623

-211

-127

-961

-0.50

-0.17

-0.10

-0.77

AH

-259

-94

-87

-440

-0.08

-0.03

-0.03

-0.14

-8423

-2926

-1298

-12654

-0.30

-0.11

-0.05

-0.46

Total


- pct -

- pct -

- pct -


(1) Without externality

41

42


Appendix 2.A Table 2.A1 Wage differentials by place of production. Estimated parameters and standard errors. Dependent variable: ln(average wage)

FGLS

WLS

OLS

Intercept

12.04955 (0.00262)

12.07617 (0.00221)

12.0701 (0.00354)

Gender

0.17744 (0.00056923) 0.19182 (0.00042372) 0.15730 (0.00080864)

Year=1999

0.07181 (0.00086704) 0.06812 (0.00063209) 0.06648 (0.00124)

Year=1998

0.05261 (0.00083882) 0.05011 (0.00059852) 0.04885 (0.00121)

Year=1997

0.02129 (0.00078997) 0.01956 (0.00055930)

Long further and higher education 0.44529 (0.00116)

0.02116 (0.00114)

0.48476 (0.00089247) 0.38434 (0.00153)

Middle-range further and higher education.

0.32206 (0.00094904) 0.31477 (0.00069157) 0.29656 (0.00132)

Short further and hi. edu.

0.19079 (0.00102 )

Skilled worker

0.10002 (0.00065551) 0.10167 (0.00045153) 0.08367 (0.00098931)

Public and personal services

0.03522 (0.00158)

-0.00264 (0.00148)

0.08941 (0.00202)

Financial intermediation

0.19359 (0.00167)

0.19702 (0.00155)

0.19622 (0.00212)

Transport and communic.

0.09228 (0.00182)

0.11263 (0.00161)

0.08528 (0.00239)

Wholesale and retail trade

0.05271 (0.00164)

0.07839 (0.00150)

0.04573 (0.00213)

Construction

0.01555 (0.00207)

-0.00886 (0.00162)

0.03066 (0.00321)

Electricity, gas, & water.

0.15069 (0.00296)

0.10845 (0.00271)

0.17680 (0.00381)

Manufacturing

0.10046 (0.00158)

0.08810 (0.00148)

0.10430 (0.00201)

Age 45-59

0.26275 (0.00071966) 0.28820 (0.00053279) 0.22788 (0.00101)

Age 30-44

0.20693 (0.00070084) 0.22936 (0.00051082)

ln(unemployment)

-0.02047(0.00081802) -0.02993(0.00061965) -0.02309 (0.00117)

ln(distance5)

-0.03893(0.00027692) -0.04252(0.00017878) -0.03165 (0.00045022)

* Adjusted R2

-

0.19089 (0.00084233) 0.16910 (0.00135)

-

0.17211 (0.00099141)

0.215


43

Table 2.A2 Wage differentials and commuting distances. Estimated parameters and standard errors. Dependent variable: ln(average wage)

FGLS

WLS

OLS

Intercept

11.99731 (0.00198)

11.96282 (0.00120)

11.95378 (0.00310)

Gender

0.20739 (0.00058077) 0.20313 (0.00032394) 0.18550 (0.00089959)

Year=1999

0.04847 (0.00088965) 0.6484 (0.00050635)

Year=1998

0.03423 (0.00086104) 0.04703(0.00048466 ) 0.03470 (0.00133)

Year=1997

0.01399 (0.00081145) 0.01898 (0.00045713) 0.01323 (0.00125)

Long further and higher education.

0.46169 (0.00109)

Middle-range further and higher education.

0.29248 (0.00090276) 0.29033 (0.00055063) 0.30186 (0.00138)

Short further and higher education.

0.17716 (0.00101)

Skilled worker

0.10374 (0.00070002) 0.09785 (0.00036837) 0.10002 (0.00111)

Age 45-59

0.30252 (0.00074048) 0.27874 (0.00042973) 0.31246 (0.00112)

Age 30-44

0.24254 (0.00071004) 0.22502 (0.00041411) 0.24134 (0.00106)

ln(unemployment)

-0.07263(0.00082992) -0.03924(0.00047434) -0.06847 (0.00130)

ln(distance)

0.03071 (0.00022535) 0.03067 (0.00016048)

0.04807 (0.00137)

0.49915 (0.00071247) 0.45697 (0.00158)

0.17132 (0.00068469) 0.19049 (0.00150)

0.04145 (0.00036696)

* Adjusted R2 0.2626 Note Table 2.A1 and Table 2.A2: All variables are significant at 10% level. If all dummy variables are zero the representation is: Year=1996, basic education, and age 15-29 years old.

Appendix 2.B F-test for fixed effects F-test: I) F-test in a model with wage differentials by place of production (data set 1) and explanatory variables gender, education, sector, age, and unemployment: As F (4,4) = 1 the hypothesis that all constant terms are equal could not be rejected.

44


II) F-test in a model with wage differentials (data set 2) and explanatory variables gender, education, age, unemployment, and commuting distance: As F (4,4) = 1 the hypothesis that all constant terms are equal could not be rejected. Appendix 2.C I) A model with wage differentials by place of production (data set 1) and explanatory variables gender, time, education, sector, age, unemployment, and distance is examined. The squared residuals of the weighted regression (u2) are estimated on the size of the group (N): The results are: FGLS

WLS

OLS

$

0.01863

0.0060891

-0.00079502

t-value of $

26.89

36.53

-40.9

All of the t-values are significant and therefore a group error component could be present. II) A model with wage differentials and explanatory variables gender, time, education, age, unemployment, and distance is examined. The test described above gives the following results: FGLS

WLS

OLS

$

0.00461

0.00788

-0.00016861

t-value of $

9.84

329.21

-14.96

All of the t-values are significant and therefore a group error component could be present.

References Albæk, K; R. Asplund, S. Blomshag, E. Barth, B.R. Gudmundsson, V. Karlsson and E.S. Madsen (1999): Dimension of the Wage-Unemployment Relationship in the Nordic Countries: Wage Flexibility without Wage Curves. Discussion Paper 99-24, University of Copenhagen, Institute of Economics. Anselin L (2003a) Spatial externalities International Regional Science Review 26, 2 , 147-152. Anselin L (2003b) Spatial Externalities, Spatial Multipliers and Spatial Econometrics, International Regional Science Review 26 (2), 2003: 153-166. Audretsch DB, Feldman MP (2004) Knowledge spillovers and the geography of innovation In: Henderson JV, Thisse JF (eds) Handbook of regional and urban economics. Volume 4 Cities and geography (Elsevier, Amsterdam) 2713-2740


45

Baldwin RE, Martin P (2004) Agglomeration and regional growth In: Henderson JV, Thisse JF (eds) Handbook of regional and urban economics. Volume 4 Cities and geography (Elsevier, Amsterdam) 2671-2712 Berndt, Ernst R. (1991): The Practice of Econometrics: Classic and Contemporary (Addison – Wesley, New York). Duranton G, Puga D (2004) Micro-foundations of urban agglomeration economies In: Henderson JV, Thisse JF (eds) Handbook of regional and urban economics. Volume 4 Cities and geography (Elsevier, Amsterdam) 2063-2118. Engelstoft S, Jensen-Butler CN, Smith I, Winther L (2006) Industrial clusters in Denmark. Theory and empirical evidence. To be published in Papers in Regional Science in 2006. Fingleton B (2003) Externalities, Economic Geography and Spatial Econometrics: Conceptual And Modeling Developments, International Regional Science Review, 26, 2 197-207. Fingleton B, Igliori D C and Moore B (2005) Cluster Dynamics: New Evidence and Projections for Computing Services in Great Britain , Journal of Regional Science 45 283-311. Glaeser EL, Kallal HD, Scheinkman JA, Shleifer A (1992) Growth in cities Journal of Political Economy vol 100, no. 6 1126-1152. Goodchild M, Anselin L, Applebaum R, Harthorn B (2000) Towards spatially integrated social science International Regional Science Review23, 139-159 Jensen-Butler CN, Larsen MM, Madsen B, Nielsen OA, Sloth B (eds) (2005) Road pricing, traffic regulation and the environment, Springer Verlag, Berlin (forthcoming in 2007). Madsen B; Jensen-Butler CN, Dam PU (2001a): The LINE model. (AKF Forlaget, Copenhagen). Madsen B; Jensen-Butler CN, Dam PU (2001b): A Social Accounting Matrix for Danish Municipalities (SAM-K). (AKF Forlaget, Copenhagen). Madsen B; Jensen-Butler CN (2004): Theoretical and operational issues in subregional modelling, illustrated the development and application of the LINE, Economic Modelling 21, 471-508. Madsen B, Jensen-Butler CN (2005) Spatial accounting methods and the construction of spatial social accounting matrices Economics Systems research, 17:2, 187-210. Madsen B, Jensen-Butler CN, Kronback J, Leleur S (2005): A systems approach to modelling the regional economic effects of road pricing. In: Jensen-Butler CN, Larsen MM, Madsen B, Nielsen OA, Sloth B (eds) Road pricing, traffic regulation and the environment, Springer Verlag, Berlin (forthcoming, 2006). Ottaviano G, Thisse JF (2004) Agglomeration and economic geography In: Henderson JV, Thisse JF (eds) Handbook of regional and urban economics. Volume 4 Cities and geography (Elsevier, Amsterdam) 2560-2608. Scitovsky T (1952) Two concepts of external economies Economic Journal, LXII, 54-67 Trigg AB, Madden M (1995) A CGE solution to the household rigidity problem in extended input-output models In: Hewings GJD, Madden M (eds.) Social and demographic accounting. (Cambridge University Press, UK) 145-163.

3 External Effects in Road Traffic: the Pigou-Knight Model and its Extension to Situations With Endogenous Speed Choice and Heterogeneous Traffic*

Jan Rouwendal Free University, Amsterdam, The Netherlands, E-mail: [email protected]

Abstract. Traffic congestion is an important example of an external effect. Economic analysis of this phenomenon dates back to the first half of the twentieth century with the seminal contributions of Pigou and Knight. The Pigou-Knight model still remains an important tool for the analysis of congestion. It is the purpose of this contribution to provide a brief discussion of this canonical model and to introduce two extensions. The first concerns the derivation of the travel cost curve, which is considered as an exogenously given relationship in the standard analysis, from a theory about speed choice. The second concerns the incorporation of heterogeneous traffic into the model. The results of the analysis are illustrated by a numerical example. Key Words : External effects, Traffic congestion, Pigou-Knight model.

__________________________ *The author thanks Morten Marott Larsen for comments on an earlier version. The usual disclaimer applies.

48

Jan Rouwendal

3.1 Introduction Traffic congestion is an important example of an external effect. One driver’s presence on the road decreases the speed of the other vehicles and therefore imposes a cost on traffic that he doesn’t have to pay when road use is free of charge. The implication is that there is a discrepancy between the driver’s private cost benefit analysis, which ignores the congestion he causes, and a social cost benefit analysis, which incorporates it. Economic analysis of this phenomenon dates back to the first half of the twentieth century with the seminal contributions of Pigou (1920) and Knight (1924). Although other models have been constructed, Vickrey’s bottleneck model (see Arnott, de Palma and Lindsey, 1993) being the main alternative, the Pigou-Knight model remains an important tool for the analysis of congestion. It is the purpose of this contribution to provide a brief discussion of this canonical model and to provide two extensions. The first of these concerns the derivation of the travel cost curve. In the standard analysis, this curve is considered an exogenously given relationship. Here we will derive it from a theory about speed choice. The second extension concerns the incorporation of heterogeneous traffic into the model.

3.2 The Pigou-Knight Model of Traffic Congestion The Pigou-Knight model of road congestion concerns the demand for travel on a homogeneous road segment. The demand for trips per time unit can be expressed as a function of the cost associated with these trips by means of a demand function. The cost of a trip is usually measured as the time cost, i.e. as the amount of time needed to complete the trip times the value of time. The time involved in traveling down the road segment is a function of the volume of demand. If there are more vehicles on the road, speed decreases and travel time goes up. In market equilibrium the time cost of the marginal driver is exactly equal to the benefits he derives from the trip. The market equilibrium differs from the social optimum. The reason is that the marginal driver adds to the travel time of all other drivers and he does not take into account this cost when he decides to make the trip. If he did, he would realize that the social cost of his trip exceeds the cost that he bears himself and in some cases this would induce him to stay home. This means that in a market equilibrium there is too much traffic on the road.

3 External Effects in Road Traffic

49

The social optimum can be reached when all drivers are charged for the cost they impose on others. This means that a toll has to be introduced that correct for the discrepancy between private and social cost. The optimal value of this toll is equal to the difference between the marginal (social) cost of road use and the average (private) cost. This conclusion can be reached on the basis of a simple mathematical model. The inverse demand function p gives the individual benefit of the marginal driver when the traffic flow equals f: p p( f ) (1) This function is decreasing. The travel time function c describes the individual time cost involved in making a trip as a function of the traffic flow: c c( f ) (2) This function is non-decreasing. As long as there is no congestion (when the flow of traffic is small) it is flat, but when there is congestion, it is increasing. Consider first the market equilibrium in this model. In equilibrium the flow is such that the private benefit of a trip is equal to its time cost: p c (3) If there are other private costs involved in making a trip, they should be added to c. For instance, if drivers were forced to pay a toll W before being enabled to make a trip, the condition for a market equilibrium would be: p c( f ) W (3’) In order to consider the social optimum in this model, we define the social surplus SS as the sum of consumer surplus CS and toll revenues. The latter are equal to the product of traffic flow and the value of the toll. Consumer surplus is equal to the area below the demand curve, but above the market price: f

CS

³ p( x)dx (c( f ) W ) f ,

(4)

0

where x denotes the demand for trips. Adding the toll revenues gives the social surplus: f

SS

³ p( x)dx c( f ) f

(5)

0

The absence of the toll revenues from the social surplus is a consequence of the fact that the toll revenues are redistributed over the population. They are therefore costs, but also benefits. Even though toll payment and received revenue may be different per person, they cancel when summed over the total population.

50

Jan Rouwendal

The social planner maximizes the value of this social surplus by determining the optimal value for f. The first order condition for a maximum is:

§ wSS ¨¨ © wf

· wc ¸¸ p( f ) c( f ) f wf ¹

0

(6)

In market equilibrium without tolling (3) holds, and we can therefore conclude that the market equilibrium will not coincide with a social optimum if travel cost is increasing in traffic flow, i.e. if there is congestion. However, the toll can be used as an instrument to reach the social optimum. Comparison of (3’) with (6) makes clear that the first order condition will be fulfilled if:

W

f

wc wf

(7)

The right-hand-side of this equation is the product of the traffic flow and the marginal travel cost. It can be interpreted as the additional travel cost the last driver imposes on all others, i.e. as the external congestion effect. We conclude that in the social optimum the toll serves to equalize private travel cost with the social cost of travel. The toll can therefore be said to ‘internalize’ the external congestion effect. The analysis just given can be illustrated with a familiar diagram, given here as Figure 3.1.1 The Pigou-Knight model is a static model. It is commonly interpreted as referring to a uniform road segment on which homogeneous traffic proceeds at constant speed. In what follows we will consider two extensions of this basic model. The first concerns the determination of the travel cost function c(f). This function is usually treated as an exogenously given ‘technical’ relationship comparable to a production function. However, it seems probable – to say the least – that driver behavior is involved when increased traffic density leads to lower speed. In the next section we will see how a simple theory of speed choice can be used to derive the travel cost function endogenously. A second extension concerns the heterogeneity of traffic.

1 The values indicated on the axes originate from the numerical example discussed in Section 5. They are of no importance for the purpose of the present section.


Social optimum

0.35 0.3 c o s t (e u ro s )

51

Market equilibrium

0.25 Average cost Marginal social cost Demand

0.2 0.15 0.1 0.05 0 0

1000

2000

3000

4000

traffic flow (cars/hour) Fig. 3.1. The standard version of the Pigou-Knight model.

3.3 Speed Choice and the Determination of the Travel Cost Function

3.3.1 Introduction The relation between traffic flow and travel cost occupies a central place in the Pigou-Knight model. It is usually treated as a technical relationship, comparable to the cost function derived from the production function in the analysis of the firm. However, it seems reasonable that the relation between traffic flow and cost is determined by driver behavior. Speed decreases when traffic density increases because drivers interact with each other. The presence of other vehicles on the road constrains the possibilities for speed determination. In order to avoid accidents, one sometimes has to choose a lower speed than would otherwise be preferred. In this section we will consider a simple way of modeling such behavior.

52

Jan Rouwendal

It will be clear from the outset that if time costs were the only relevant aspect of travel cost, drivers would prefer to proceed at the highest possible speed. In practice drivers usually choose a lower speed and there is of course a good reason for such behavior: driving at a high speed is more dangerous and this is especially the case when there are other vehicles on the road. Actual driver behavior can be interpreted as the consequence of trading off the desire to minimize travel time against that of keeping accident risks at a moderate level. One way of formalizing this idea is to assume that individual travel costs depend not only on travel time (or its inverse, speed), but also on a measure of traffic density. A particular simple measure is the distance between a vehicle and the nearest vehicle in front of it. This indicates the way in which we will proceed. Even though in what follows attention is concentrated on the relationship between accident risk and headway distance, in practice many other variables may influence the distance between cars. Driver’s knowledge of their reaction times, their state of mind at the moment of driving and numerous other variables may influence this variable as well. However, for the purpose of the analysis of the present chapter we take the other factors as given. 3.3.2 Speed Choice Consider a driver who has to determine his speed. The instantaneous costs k of driving are a function of the driver’s speed s and the (nose-to-tail) distance to the nearest vehicle in front of him į: k k ( s, G ) (8) with k a differentiable function defined for nonnegative values of s and į. The cost k refers to the cost at one point in space and that integration over the road traveled has to take place in order to arrive at the total cost of a trip. The specification of the cost function given in (8) is in some respects a restrictive one. For instance, it includes only the driver’s own speed and not that of other drivers as an argument. This excludes the possibility that speed differences (which are probably an important element of accident risk) determine speed choice in a direct way. Note, however, that speed differences play a role in the model indirectly since they imply a changing distance to the leading vehicle and this induces the driver to change his speed. The function k should be convex in s for every possible value of į and it should have a minimum in s, possibly for s=0, for every possible value of


53

G. At any moment the driver is assumed to take the value of į as given and speed is chosen instantaneously so as to minimize the costs:

arg min k ( s, G )

s(G )

(9) where s(į) denotes the optimal (i.e. cost minimizing) speed, which is a function of the headway distance. If the optimal speed is positive, s(į) is the solution of the first order condition

wk ws

0

(10)

Since the function k is convex in s, the second-order conditions for a minimum are also fulfilled. We can derive the optimal (=minimal) cost as a function of the distance to the leader by substituting (9) into (8).2 Equation (9) defines a relationship between speed and headway distance. It seems natural to expect that s(į) is non-decreasing and concave in the distance to the leader į and has a limit s*, to be referred to as the free flow speed, for įĺ. On the basis of (10) it can be shown that:

ws (G ) wG

w 2k ws wG 2 w k wG 2

(11)

For the desired properties to be present it is necessary that the righthand-side is positive, increasing in į and approaching the value 0 for įĺ. The concave curve in Figure 3.2 gives an example of such a speedheadway relationship. 3.3.3 The Steady State In order to investigate the consequences of this speed choice model for the Pigou-Knight type of analysis, we now consider a stationary state on a homogeneous road segment. A stationary state is here defined by the following properties: (i) every t seconds a car enters the road and (ii) each car has a constant speed s on the whole road segment. Since the headway dis2

The myopic behavior implied by (5) is consistent with minimization of total travel cost ³cxdx if (a) an individual driver cannot influence the density rx and (b) there are no binding constraints this minimization due to, for instance, limits on engine capacity (acceleration) and braking power. Condition (a) is not satisfied in our car following version of the speed choice theory, see the discussion following equation (13) below. Condition (b) seems often satisfied in actual driving situations and will be ignored.

54

Jan Rouwendal

tance determines speed, the headway distance must be constant in a steady state. Moreover, speed must be equal for all drivers, since speed differences would imply changing headway distances for some drivers and therefore non-constant speeds. The stationary state is compatible with the model of speed choice behavior of individual drivers that has been developed above if the headway distance plus the length of the car is equal to st: G P st (12) In the above we have referred to speed, cost et cetera as instantaneous units referring to a particular location. In order to derive the cost of a trip on the road segment we should integrate over the total length of the road segment. Because the segment is (by assumption) homogeneous, this is easy to do: all that is needed is multiplication of the localized variables by the length of the road segment. For convenience we take the road segment to be of unit length, which implies that we can use our instantaneous variables without further modification as referring to trips as well. The flow of traffic on the road segment equals the inverse of t:

f

1 t

(13)

Using equation (11), we find:

s

(P G ) f

(14) In the steady state drivers choose their speed optimally, so (9) holds also, so that we have two equations, (9) and (14), in two unknowns, s and G. Figure 3.2 illustrates. The straight line in the figure is equation (14) with f equal to 3000. Its slope is proportional to f.3 The other line pictures the headway-speed relationship (9), which is assumed to be concave.4 The two lines cross each other at two points if the flow is not too large. The equilibrium with the lowest speed corresponds to a state of hypercongestion and has been shown by Verhoef (2002)5 to be dynamically unstable as a stationary state. Attention may therefore be concentrated on the equilibrium with the highest speed. It can be readily inferred from the picture (and confirmed by formal analysis) that larger flows can only be accommodated at lower speeds. There is It is not equal to f because of the different units used on the two axes: headway distance is measured in meters, speed in kilometers per hour. 4 I.e. speed is a concave function of headway distance. Figure 3.1 pictures the inverse function, which is convex. 5 Verhoef (2002) introduces the relation between headway distance and speed as a car following theory, without deriving it from speed choice behavior. 3


55

a unique maximum flow that corresponds with the situation in which the straight line (14) just touches the headway-speed relation (13). This maximum flow can be interpreted as the capacity of the road. This derivation makes clear that capacity is not only determined by the technological characteristics of the road, but also by driver behavior.6 Equations (9) and (14) determine the combination of speed and headway distance that allow a given flow F to use the road segment. These will be denoted as s*(f) and G*(f), respectively. Since travel time is the inverse of speed, the relation between flow and travel time can easily be derived from this relation. Multiplication by the value of time then gives travel time cost per driver, which are often used in the standard version of the Pigou-Knight model as the relevant travel cost.

160 140 s p e e d (k m /h r)

120 100

speed choice function flow equal to 3000 cars/hr

80 60 40 20 0 0

20

40

60

Fig 3.2. Speed and flow in a stationary state.

It must, however, be noted that in the present setting travel time cost is not the appropriate concept of travel cost. As was pointed out in the introduction to this section, the speed choice model presupposes a trade-off between time cost and (expected) accident cost and both play a role in the cost function k in (8). The appropriate formulation of the cost function can

6

We will return to this point in the next section.

56

Jan Rouwendal

be found by substituting the relevant values of speed and headway distance into this cost function:7 c( f ) k s * ( f ), G * ( f ) (15) We will now consider a particular specification of this cost function in order to investigate the difference between its value and travel time cost. 3.3.4 A Specific Case An intuitively appealing specification for the cost function is:8

k ( s, G )

vot b(G ) s n s

(16)

where vot and n are positive. b(G) is a function of traffic density which is assumed to be positive and increasing. This equation states that the costs are determined by travel time and safety considerations. Travel time cost is proportional to the inverse of speed. The parameter vot is the value of time and the first term on the right hand side of (16) is therefore equal to the cost of travel time. The cost associated with safety is assumed to be a power function of speed. It seems reasonable to assume that n is larger than one, that is, that the cost of safety increases more than proportionally with speed. One argument for this assumption is that the damage done by accidents is related to the kinetic energy of the vehicles involved, which is a quadratic function of speed. The multiplicative term b(.) implies that the cost associated with safety is, for a given speed, increasing with traffic density. The relation between speed and density that follows from (9) is:

s (G )

§ vot · ¨¨ ¸¸ © nb(G ) ¹

1 /( n 1)

(17)

It is easy to verify that this speed choice function has the required properties if b(į) is positive for all nonnegative densities. We can rewrite (17) as:

b(G )

1 vot n s (G ) n 1

(18)

Note that we should integrate k over the length of the road segment. Since the road segment is homogeneous this is particularly simple: k should be multiplied by the length of the road segment, which was assumed to be equal to 1. 8 I am indebted to Robin Lindsey for suggesting this generalization of my original specification, which had n=2. 7


57

and substitute this equation in the cost function (16) to find:

k min (G )

§ 1 · vot ¨1 ¸ © n ¹ s (G )

(19)

This equation says that the true travel cost exceeds the travel time cost by a fraction 1/n. For instance, if accident costs are quadratic in speed, then the appropriate travel cost value is 150% of the value of travel time. The traditional approach therefore underestimates the cost of travel time (and congestion) in this case by one third. Another interpretation is that equation (19) says that the appropriate value of travel time is higher than the ‘pure’ value of time because of the accident risk involved in traveling. If the value of accident risk is incorporated into the value of travel time, the right-hand-side can still be interpreted as the value of travel, and the difference with the conventional approach appears less fundamental. The reason behind the ambiguity implied by the possibility of two different interpretations is that it is not always clear whether empirical estimates of travel time incorporate the accident risk associated with traveling or should be interpreted as a measure of the ‘pure’ value of time.

3.4 The Pigou-Knight Model with Heterogeneous Drivers 3.4.1 Introduction Although in economic analyses drivers are often assumed to be homogeneous, it seems unlikely that in reality all drivers have the same values of the parameters of their utility functions. Note that in the model developed in the previous sections these parameters also include those of the carfollowing relationship. Verhoef et al. (1999) and Rouwendal et al.(2002) have studied the consequences of traffic heterogeneity in models where there are two groups of vehicles that differ in desired free flow speeds.9 In these papers a simplified relation between speed and headway distance was assumed: drivers always used their desired free flow speed unless the minimum critical distance to the leader was reached and speed equals that of the leader if the headway distance equals its minimum value. In what follows we use the more realistic headway-speed relation based on car fol-

9

These differences may be caused by the preferences of the drivers or by the characteristics of the vehicles they use or by a combination of the two.

58

Jan Rouwendal

lowing theory derived earlier in this chapter and study the stationary state properties of the model.10 Assume that there are two groups of vehicles with different headwayspeed relationships. Figure 3.2 illustrates this for a case in which for any value of the headway distance, drivers of group 1 choose a lower speed than drivers of group 2, but situations in which the two curves cross each other are also admitted. Both groups use the same road and overtaking is impossible. We want to use the model with two groups in order to study stationary state traffic with heterogeneous drivers. In order to define a stationary state in the present situation, we have to revise our definition of such a state. The reason is that we can’t have a constant time interval t between subsequent cars that enter the road and the same constant speed for all cars. The stationary state will therefore now be defined as referring to a situation in which all cars have identical constant speed. The headway distance will therefore differ. This means also that the time interval between cars will differ and let ti refer to the time interval before entrance of a car of type i (i=1,2). Vehicles of one type drive closer to their leaders than those of the other type. If the headway distance of the two types at the stationary state speed is substantial, stationary state traffic will take the form of single driving vehicles with a large (in a relative sense) headway distance and platoons consisting of a vehicle with a large headway distance followed by one or more vehicles with a shorter headway distance. In order to complete the description of the stationary state, we have to determine the length and frequency of the platoons. In order to do so, we have to specify the mechanism by which cars enter the road. This mechanism should be compatible with a stationary state. This means that the headway distance should be that corresponding to the stationary state speed. Since the cars are of different types this means that for one type this headway distance will be larger than for the other. Since the speed of both types of vehicles is equal, this means that the time that passes before a vehicle of type 1 enters must be different from the time that passes before a vehicle of type 2 enters. We assume that there are fixed probabilities ʌi that the next car will be of type i=1,2. If the vehicles belonging to the first group have the shortest headway distance, there is a

10

The discussion in the present subsection assumes that the chosen speed will always be increasing in the headway distance, while approaching it s free flow value asymptotically. It is easy to consider also situations in which free flow speed is reached for a finite headway distance, as in eq. 20.


59

probability ʌ 2=(1-ʌ1) that a vehicle of type 2 that has just entered the road drives single and a probability ʌ1n(1-ʌ1) of a platoon of length n.11

160 140 s p e e d (k m /h r)

120 100 group 1 group 2

80 60 40 20 0 0

50

100

headway distance (meters) Fig. 3.3. Headway-speed relationships of two groups.

Let us now consider the stationary state. In such a state all vehicles have the same speed, but the composition of the flow is determined by the stochastic mechanism just described. It implies that the total number of vehicles of each type that enter during a unit of time is also a random variable, as is the total flow.12 Let f1 be the expected flow of vehicles of group 1 and f2 the expected flow of vehicles of group 2. For given values of these flows the equilibrium speed and headway distances can be determined by the following three relations: Other mechanisms may also be defined. Deterministic ones (e.g. k1 cars of type 1 are always followed by k2 cars of type 2, k1,k2>0) and mechanisms in which the probability that the next car is of type 1 depends on the type of its leader are alternative possibilities. The mechanism used here can be interpreted as resulting from a heterogeneous population of drivers who each take their decision to enter the road independently of each other. 12 Note that the flows of the two types of vehicles are dependent upon each other: if one knows the flow of one type, the flow of the other is also determined. However, the total number of vehicles is not a deterministic variable. 11

60

Jan Rouwendal

· f1 f2 (G 1 P1 ) (G 2 P 2 ) ¸¸ f1 f 2 ¹ © f1 f 2 s1 (G 1 ) s s 2 (G 2 ) s §

f1 f 2 ¨¨

s (20) (21)

(22) The first of these equations says that the product of the total flow and the average road distance occupied by a car should be equal to the common steady state speed s. We can simplify (20) and substitute the inverses of the headway speed relationships (21) and (22) into that equation: ( f 1 P1 f 2 P 2 ) f 1 s11 ( s ) f 2 s 21 ( s ) s (23) The function s1-1 gives the headway distance for vehicles of type 1 that corresponds with speed s, and s2-1 has an analogous interpretation. These two inverse functions are defined for nonnegative speeds; they are convex, equal to 0 when s=0 and increasing in s. Figure 3.4 illustrates the steady state equilibriums by picturing the speed as a function of the average headway distance between cars on the road. It is assumed that 50% of the drivers belong to group 1 and the other 50% to group 2. Note that the maximum speed is determined by the group that prefers to drive slowest. On the basis of (20) we can also draw a straight line that gives the combinations of average headway distance and speed for which a given (expected) flow is realized. Just as was the case in Figure 3.2, there are usually two equilibriums. The one with the lowest speed can be considered as hypercongested. It has been shown13 that the hypercongested stationary state is dynamically unstable, just as in the case of homogeneous traffic. Attention can therefore be confined to the non-hypercongeste1d case. It may be noted that in the case of heterogeneous traffic, it is impossible to identify the capacity of the road as a unique number of vehicles that may pass through it per unit of time. There may be various combinations of flows f1 and f2 for which the line picturing the left-hand-side of (23) just touches that picturing the right-hand-side. For any given value of f1 that does not exceed the capacity of the road for homogeneous traffic of type 1, one may derive flow f2 at which the capacity of the road reached. Except in special cases, the total number of vehicles will be different depending on the chosen value for f1. The speed of the non-hypercongested steady state speed will in general also depend on the composition of traffic and we may write: s s ( f1 , f 2 ) (24)

13

In an appendix to this paper that is available upon request from the author.


61

160 140 s p e e d (k m /h r)

120 speed as a function of av. headway dist.

100 80

flow equal to 3000 cars/hr

60 40 20 0 0

20

40

60

headway distance (meters) Fig. 3.4. Steady state equilibrium with two groups of vehicles.

for this speed. If cost function (8) is used, generalized travel cost is inversely proportional to this steady state speed and therefore also dependent on the composition of the total traffic flow. It should be noted that (24) implies that there is in general no unique relation between steady state speed and total traffic flow f1+f2. The exception occurs if: s s ( f1 f 2 ) , (25) but it will be shown below that this specification cannot be valid in the model developed here. Finally, we note that it can be shown that steady state speed is a decreasing function of f1 and f2.14 3.4.2 Optimal Congestion Tolling The standard Pigou-Knight model of static road congestion assumes homogeneous traffic, but is often tacitly considered being a good approximation to actual situation with heterogeneous traffic. The steady-state model

14

See eq. 30 below.

62

Jan Rouwendal

with heterogeneous traffic developed in the previous subsection allows a formal analysis of this conjecture. We maximize the social surplus, i.e. the sum of the consumer’s surpluses of the drivers of the two types and the toll revenues under the constraints that a user’s equilibrium should be realized. We denote the inverse demand function as D, the number of trips as M, the price of a trip (which equals the sum of travel cost and toll) as p and the toll as W and add the appropriate suffixes. The Lagrangian is: f1

f2

³ D1 (M1 )dM1 ³ D2 (M 2 )dM 2 c1 f1 c2 f 2

L

0

0

(26)

K1 p1 c1 W 1 K 2 p 2 c 2 W 2 with the K’s denoting Lagrange multipliers. The first order conditions lead to the following expressions for the optimal tolls:

wc1 wc f2 2 wf 1 wf 1

W1

f1

W2

wc wc f1 1 f 2 2 wf 2 wf 2

(27)

This shows that the optimal tolls for the two groups are different unless wci wf1 wci wf 2 for i=1,2. In order to see whether this is the case, we use cost function (16), which leads to the following expression for the optimal cost (cf. eq. 19):

vot i § 1· ¨1 ¸ n 1 © n ¹ s( f1 , f 2 )

ki

i 1,2.

(28)

Differentiation shows that:

wci wf j

§ 1 · vot ws ¨1 ¸ 2 i © n ¹ s wf j

i, j

1,2.

(29)

In order to find the partial derivative of the steady state speed with respect to each flow we use the differential of (23) to compute ws / wf j . The result is:

ws wf j

Pj G j 1 1 1 f1 f2 ds1 dG 1 ds 2 dG 2

(30)


63

The numerator of the right-hand-side is the distance occupied by a vehicle of type j in the steady state. It is different for the two types of vehicles and this shows that for the model developed here the steady state relation between flows and speed cannot be described by (25). Hence there is no unique relation between speed and total traffic flow in the model with heterogeneous drivers. Returning now to the optimal tolls, we observe, using (25), that the fact that wci wf j depends on j implies that wci wf1 z wci wf 2 . This means that the optimal tolls for the two types are in general different (cf. eq. 27). We conclude that in the model with heterogeneous drivers developed in this chapter, first best tolling requires different treatment of the two types. Uniform tolling is a second best solution. The model with two types of vehicles developed in this and the previous subsection generalizes the Pigou-Knight model to situations in which traffic is heterogeneous. One important reason for developing this model was to investigate the validity of the common practice to use the Pigou Knight approach as an approximation to situations with heterogeneous traffic. We conclude that in one important respect the approximation is invalid: with heterogeneous traffic uniform tolling can in general only be a second best measure. 3.4.3 Discussion It is easy to generalize the model of the previous two subsections to situations in which there is an arbitrary number of vehicle types. Such a generalization seems to be needed in order to capture the diversity of driving styles and vehicles characteristics that interfere with them in actual situations. Newell (2002) has recently stressed the empirical relevance of steady state traffic as would be described by such a model. He proposed a simplified car following theory “wherein, if an n-th vehicle is following an (n-1)th vehicle on a homogeneous highway, the time-space trajectory of the n-th vehicle is essentially the same as the (n-1)th vehicle except for a translation in space and time.” (p. 195, abstract), but where the headway distances between the vehicles may be different. After exposing this (essentially steady state) theory he motivates its empirical relevance by a number of references. The conclusion reached above that optimal congestion tolls should vary with vehicle types implies that the Arnott-Kraus (1998) condition for the feasibility of marginal cost pricing by means of uniform tolling are violated in the model developed above. The reason is that the different types of vehicles contribute to congestion in different ways. The essential differ-

64

Jan Rouwendal

ence is the amount of road space (Pj +Gj ) they occupy. The effect of differences in the length of space occupied by the vehicles themselves can in principle be captured by using a passenger car equivalent (pce) instead of the number of cars.15 However, differences in the preferred headway distances at given speeds are much more difficult to incorporate since they appear to be related to driver characteristics that are difficult to observe. Moreover, these headway distances will probably vary with driving conditions and for this reason a differentiation of the tolls on the basis of observable characteristics (such as vehicle length Pj) will not be able to solve the problem. The conclusion that the uniform congestion tax is sub-optimal if drivers are heterogeneous is perhaps not too surprising. After all, the basic Pigouvian insight suggests that drivers who contribute differently to congestion should be taxed differently. However, there are several reasons why is seems, nevertheless, important to stress the point. The first is that heterogeneity among the drivers is usually ignored in discussions about Pigouvian congestions tolls. The second is that Arnott and Kraus (1998) have shown that there are indeed situations in which heterogeneous actors could be treated as if they are identical. This state of affairs makes it relevant to point out whether heterogeneous drivers satisfy the conditions formulated by these authors and if not, what the relevant differences between the drivers are and what their relation is to the optimal tolls.

3.5 A Numerical Example In order to illustrate the analysis of the previous sections, a small simulation model16 has been constructed. We have adopted cost function (16) with n=2, vot=7.5 (which is a reasonable estimate for the Netherlands) and we specified b(G ) as:

bG exp(6 G .25 ) 0.000002

(31) This specification was chosen on the basis of some experimentation with exponential functions. If the cost minimizing speed is chosen, the relationship between headway distance and speed is the one shown in Fig15

Under severe congestion it may be the case that headway distances are more or less equal for all types of vehicles, so that differences in space occupied are almost completely determined by the length of the vehicles, which would justify a passenger car equivalent rule based on observable characteristics. However, such a rule is unlikely to be valid in general. 16 See Verhoef and Rouwendal (2004) for a more elaborate simulation model.


65

ure 3.2. The constant term in b determines the free flow speed, i.e. the speed chosen when the headway distance is infinitely large. This speed is just above 120 km/hr. The exponent in the first term is a power function. For larger values than 0.25 the relationship between headway distance and speed becomes S-shaped. This is not necessarily unrealistic (an extreme case is the relationship between speed and headway distance assumed in Rouwendal et al., 2002), but it can lead to unfamiliar shapes of the speed flow diagram. The speed-flow diagram corresponding with the chosen specification of the cost function is given in Figure 3.5. It has the familiar shape, even though the maximum flow is on the high side. At a speed of 75 km/hr a flow of more than 3600 cars is realized, implying that every second a car enters (and another on leaves) the road segment that we study. The (head-to-nose) distance between the cars is more than 15 meters, which does not seem to be unreasonably small. As was stressed in the previous sections, the speed flow diagram is the result of optimizing behavior and not a ‘technical’ relationship. One characteristic of this analysis is that it involves not only time costs, but also (perceived) accident costs, and this implies that the optimal toll is higher than that suggested by the conventional analysis. The optimal toll as determined on the basis of time costs alone is the difference between average and marginal costs shown in Figure 3.1. The optimal toll that follows from the analysis of the previous sections is higher and the two are compared in Figure 3.6. The difference between the two tolls is a factor 1+(1/n), which implies for n=2, as was chosen here, that the toll with endogenous speed choice is 50% higher than that suggested by the conventional analysis. The situation with heterogeneous traffic has been simulated by distinguishing two groups on the basis of the value of time, leaving all other parameters of the model unchanged. It was assumed that group 1 drivers had a vot equal to 5, whereas drivers in group 2 had a vot equal to 10. The implied relationships between headway distance and speed are shown in Figure 3.3. Drivers with the higher vot choose a higher speed at every value of the headway distance. In particular, their free flow speed is approximately 135 km/hr, whereas that of drivers with the low vot is 107 km/hr.

66

Jan Rouwendal speed flow curve 4000

flow (cars/hr)

3500 3000 2500 2000

speed flow curve

1500 1000 500 0 0

50

100

150

speed (km/hr) Fig. 3.5. The speed flow diagram implied by the chosen cost function.

In a steady state, all drivers must have the same speed and this means that the speed in such a state can never be higher than the free flow speed of the group with the low vot. The relationship between the average headway distance and speed in steady states when 50% of the drivers belong to each group (shown in Figure 3.4) lies below the analogous relationship for a homogeneous group of drivers with the average vot of 7.5 (shown in Figure 3.2). A consequence of this effect is that the capacity of the road, measured as the maximum flow compatible with the speed-flow diagram, is lower with the two groups of drivers with different vots than that of the homogeneous group with the average vot. Whether this is a general phenomenon or a peculiarity of the specifications used here is a topic for further study.

3 External Effects in Road Traffic 0.4 0.35

toll (euros)

0.3

optimal toll on the basis of time cost

0.25 0.2

optimal toll in the speed choice model

0.15 0.1 0.05 0 2000

2500

3000

3500

4000

flow (cars/hr) Fig. 3.6. Optimal toll in the conventional analysis and with endogenous speed choice.

0.16 0.14

to ll (e u ro s )

0.12 0.1 group 2 group 1

0.08 0.06 0.04 0.02 0 0

1000

2000 flow (cars/hr)

Fig. 3.7. Optimal tolls for the two groups of drivers.

3000

4000

67

68

Jan Rouwendal

Optimal tolling requires different treatment of the different groups. It has been shown in the previous section that the difference in the optimal toll is related to the amount of road space occupied by a vehicle. At every speed, this amount is highest for the drivers with the lowest vot, and they have to pay more. Intuitively, one might interpret the toll as a price for occupying road space, and the driver who takes more space has to pay more. Since all parameters except vot are assumed to be equal, the required different treatment of both groups may not be easy to implement.

3.6 Conclusion In this chapter the classic Pigou-Knight model of traffic congestion has been extended to cases in which speed is chosen by the drivers on the basis of cost minimization and to situations with heterogeneous drivers. In a steady state with homogeneous traffic this model provides a generalization of the conventional Pigou-Knight analysis that works with an exogenously given generalized travel cost function. In contrast with the conventional model, the integration between speed choice and car following behavior proposed in the present chapter implies that the speed flow relation is endogenous. Moreover, the value of travel time should incorporate the cost of accident risk. Another generalization was realized by considering situations with heterogeneous traffic. In the model with different groups of drivers a uniform congestion toll is in general sub-optimal. First best optimality requires differentiation of the toll on the basis of the road space occupied by the vehicles, which equals the sum of the vehicle length and the ‘nose-to-tail’ distance with the leader. The latter is determined by the speed-choice/carfollowing behavior. The result implies that groups with different headwayspeed relationships should in general be tolled differently in order to achieve first best optimality. This result was formally derived for a model with just two groups, but can be generalized to a model with an arbitrary number of groups.

References Arnott, R. and M. Kraus (1998) When Are Anonymous Congestion Charges Consistent with Marginal Cost Pricing? Journal of Public Economics, 67, 45-64. Arnott, R., A. de Palma and R. Lindsey (1993) A Structural Model of Peak-Period Congestion: A Traffic Bottleneck with Elastic Demand. American Economic Review, 83, 161-178.


69

Knight, F. (1924) Some Fallacies in the Interpretation of Social Costs. Quarterly Journal of Economics, 38, 582-606. Newell, S.F. (2002) A Simplified Car-Following Theory: A Lower Order Model. Transportation Research B, 36, 195-205. Pigou, A.C. (1920) The Economics of Welfare. London, MacMillan. Rouwendal, J., E. Verhoef, P. Rietveld and B. Zwart (2002) A Stochastic Model of Speed Differences, Journal of Transport Economics and Policy, 36, 407445. Verhoef, E.T. (2002) An Integrated Dynamic Model of Road Traffic Congestion Based on Simple Car-Following Theory: Exploring Hypercongestion Journal of Urban Economics 49 505-542. Verhoef, E.T., J. Rouwendal and P. Rietveld (1999) Congestion Caused by Speed Differences. Journal of Urban Economics, 45, 533-556. Verhoef, E.T. and J. Rouwendal (2004) A Structural Model of Traffic Congestion: Endogenizing Speed Choice, Traffic Safety and Time Losses, Journal of Urban Economics, 56, 408-432.

4 On Traffic Congestion Models à la Mohring and Harwitz

Pierre v. Mouche1, Willem Pijnappel2 and Jan Rouwendal3 1

Wageningen University, The Netherlands, E-mail: [email protected] The Netherlands, E-mail: [email protected] 3 Free Univerity, The Netherlands, E-mail: [email protected] 2

Abstract. As far as we know, the problem of existence of a maximiser of the welfare function in traffic congestion models à la Mohring and Harwitz has never been posed. In this chapter we present a mathematically rigorous existence result and take the opportunity to provide a similar variant of their selffinancing result. Key words: Transportation economics, Traffic congestion, Self financing result of Mohring and Harwitz, Existence of maximisers.

4.1 Introduction The self financing result derived by Mohring and Harwitz (1962) is generally considered to be of fundamental importance in transportation economics. It is derived in the context of a model that concerns a single transport link. The model deals with welfare maximizing combinations of toll and capacity. Under special conditions, such combinations have a self financing property: the toll revenues are exactly equal to the cost of the transport capacity. Although even extensions of the model with more than one transport link have been investigated, we will in the present chapter reconsider the original model. The reason is that, to the best of our knowledge, even this simple model has nowhere in the literature been well-defined, let alone been given a mathematically rigorous treatment: in the existing literature, the domain of various functions that are used in the model is not defined or only vaguely specified. It is therefore not surprising that somewhat subtle prob-

72

Pierre van Mouche, Willem Pijnappel, Jan Rouwendal

lems like the existence of welfare maximizing combinations of toll and capacity have not been posed. Mathematically, this concerns the problem of maximizing an implicitly defined real valued function on a possibly noncompact subset of 52. It is the main purpose of this article to shed some light on this issue. The book of Mohring and Harwitz (1962) gave rise to a number of subsequent publications, most of them concentrating on the self-financing result. That result was derived under the following conditions: there is a single link, drivers have identical values of time, toll and capacity are policy instruments, costs of capacity are linear and the congestion function is homogeneous of degree zero. In Keeler and Small (1977) and Kraus (1981) the validity of the latter two assumptions is studied. The following publications deal with the self financing result holds under more general circumstances: Strotz (1964) considered drivers with different values of time, Arnott and Kraus (1995) and Yang and Meng (2002) multiple links, Oum and Zhang (1990) the discrete character of capacity and Small (1999) markets without perfect competition. Our article is organized as follows: Section 2 presents our notion of a traffic congestion model à la Mohring and Harwitz; in the appendix we take a closer look at its setting. In Section 3 we discuss the real world interpretation of the model. In Section 4 we identify a class of such models for which the welfare function has a maximiser (Corollary 1). In Section 5 we present (by means of Corollary 3) a formulation of the self-financing result. Finally, Section 6 contains suggestions for further research. Section 7 concludes.

4.2 Setting Definition 1 below gives a formal definition of our understanding of a traffic congestion model à la Mohring and Harwitz. In Definition 2 we define the welfare function W * in such a model in terms of an implicitly defined function Q . Notation: 5+ : {x ± 5 | x 0} and 5++ : {x ± 5 | x > 0}. Definition 1 A traffic congestion model à la Mohring and Harwitz is given by x A function, to be called travel demand function, Q : 5++ 5,


73

with characteristics Q is continuous, Q is strictly decreasing on (0, p0 ] and Q

0 on [ p0 , f) where p0 ! 0.

x A function, to be called congestion function, T : I 5+ 5, with I a non-empty interval of 5+, to be called the domain of capacity. The values of this function are denoted as T (c, q) , and it has the following characteristics: T>0, i.e. T is positive, T is continuous in q, T is strictly increasing in q, the function T(.,0) is constant, say T0. x Constants

K ! 0, U ! 0, to be referred to as the value of time and the fuel price, respectively, such that

J : p0 U KT0 ! 0 . x A function, to be called the cost of capacity function, K:I5 (with I as above) with characteristics: K is strictly increasing,

K t 0. Ⴘ Note that Q t 0 and that T t T0 ! 0 . We refer to q as the number of trips, to c as capacity and to t as the toll. Moreover, we refer to p0 as the reservation price for trips, to T0 as free flow travel time and to J as the reservation toll. Given a traffic congestion model à la Mohring and Harwitz, the equation

Q( U KT (c, q) t )

q,

(1)

74


in q t 0 , has a unique solution for each c I and t [0, J ] . (The reason for restricting here the values of t to [0, J ] is provided in the appendix.) Indeed: the function Mt,c : 5+ 5 (where t ± 5+ and c I ) defined by

M t ,c q : QU KT c, q t q is strictly decreasing and therefore takes on the value 0 for (at most) a single q. Since M t , c is continuous, lim q of M t , c (q ) f and M t , c (0) t 0 , the desired result follows. We denote the unique solution of (1) as Q (t , c) . It defines a function

Q* : [0, J ] u I o 5 , to be called equilibrium travel demand function. Hence, for all t [0, J ] and c I , Q (t , c) is the unique zero of the function M t , c . We notice that the function Q was defined without invoking the cost of capacity function

K . Looking back at the proof of the well-definedness of the function Q we see that also the following conclusions can be drawn:

Q (t , c) ! 0 (0 d t J , c I ),

(2)

Q* (J , c )

(3)

0 (c I ).

Indeed: for (2) we observe that now

M t , c (0)

Q( U KT0 t )

M J , c (0)

Q( U KT0 J )

Q( p0 J t ) ! 0 and for (3) that Q( p0 )

0.

The next definition uses a function 7* : [0,J] I 5, to be called trip cost function, which is defined as 7*(t, c) = ȡ + K T (c, Q*(t, c)) + t. Definition 2 Given a traffic congestion model à la Mohring and Harwitz, we define the function W* : [0,J] I 5, to be referred to as the welfare function, by

W t, c :

p0

³ 7 t ,c Q p dp tQ t , c K c . Ⴘ


75

The maximization problem with which we are concerned is that of the function W*. We would like to prove the existence of a maximiser for W by means of Weierstrass’ theorem, according to which every continuous function on a non-empty compact subset of 5n has a maximiser. However, W is not necessarily continuous and its domain may be, depending on I , not compact. Nevertheless, Theorem 1 will demonstrate that existence results can be obtained under reasonable conditions. In the literature one can also encounter a different setting that does not define the welfare functions in terms of Q and Q , as in Definition 2, but in terms of P and Q , where P : J 5, with J : Q0, U 0 @ , is the inverse function of the bijection Q : 0, p0 @ o J . We note that J is an interval of the form J >0, J with J 0, f @. We will, par abus de langage, refer to P as the inverse travel demand function. Proposition 1 is relevant in this context; its proof can be found in the appendix. Proposition 1 For all t [0, J ] and c I we have:

W (t , c )

³

³

Q (t , c ) 0

Q (t , c ) 0

p0

³ 7 (t , c)Q( p) dp tQ

(t , c ) K ( c )

P( x) dx P(Q (t , c))Q (t , c) tQ (t , c) K (c)

P( x) dx ( U KT (c, Q (t , c)))Q (t , c) K (c). Ⴘ

4.3 Real World Interpretation We now give the real world interpretation of the model: a policy maker plans to make a toll road available that can be used by cars (travelling in one direction) driven by identical agents. The problem for this policy maker is: which pair of toll and capacity maximizes welfare. For given capacity c and toll t the price p of a trip equals:

p U KT ( c , q ) t . Here, q is the total number of trips, U the fuel price, K the value of time and T (c, q ) the time needed to complete a trip. This time therefore does not only depend on the capacity c of the link, but also on the total number of trips q .

76


We take as given a function Q , the so-called travel demand function, that gives the relation between the price p and the number of trips q : q Q( p) . Moreover, we make the (neoclassical conventional) assumption that an equilibrium will be obtained, here in the sense that the following equation holds:

Q( U KT (c, q) t ) q. Briefly: given t and c, a number of trips Q (t , c) (being the solution q of this equation) will be made for a price 7* t , c : U KT c, Q * (t , c) t

per trip. Welfare now will be W t, c :

p0

³7*(t ,c) Q p dp tQ t , c K c ,

which equals net consumer surplus at 7*(t, c), plus toll revenues minus the cost of providing capacity.1 Sometimes one is satisfied with this interpretation of the model, but one may well ask further questions. For instance: what do we mean by ‘capacity’ and when and how will the number of trips Q (t , c) be made? We will not consider these questions further, even though we feel that a closer investigation of the microscopic foundations of this model could be enlightening. The mathematical properties that we attributed in Section 2 to the various elements of the model guarantee that the equilibrium travel demand function Q * and the welfare function W * are well-defined. Notice that we didn’t assume that T is decreasing in c . Indeed, this property is not needed for the derivation of our results. Given a traffic congestion model à la Mohring and Harwitz one can ask which (t , c) , i.e. which combination of toll and capacity, maximizes the welfare function W . But for the purpose of the present article, the first question is not which combinations ‘do the job’ but if, and under what conditions, such combinations exist.

1

This welfare measure, based on consumer's surplus, is popular among neo-classical economists.


77

4.4 Existence Results Lemma 1 If there exist c c, c cc I for which

K (ccc) K (cc) ! ³

Q(U ) 0

P( x) dx J Q( U ) ,

then the maximisers of the function W are the same as those of the (re-

@

stricted) function W* & [0,J] I f,c '' . Ⴘ Proof. - Note that c c ccc . Since , by Lemma 4(7), 0 d Q Q ( U ) the

second expression for W (t , c) in Proposition 1 implies that Q(U )

W (t , c) d ³ 0

P( x) dx 0 J Q( U ) K (c) . For c t c cc this, using (3)

gives:

W (t , c) d ³

Q(U ) 0

P( x) dx J Q( U ) K (ccc) K (cc) W (J , cc).

Hence, W (t , c) W (J , c c) for all t [0, J ] and c t c cc . This implies the desired result. Q.E.D. Theorem 1 Suppose the welfare function W is continuous. 1. If I is compact, then W has a maximiser. 2. If I is left-closed,2 and there exist c c, c cc I for which

K (ccc) K (cc) ! ³

Q( U ) 0

P ( x) dx J Q( U ),

then W has a maximiser.Ⴘ Proof. - 1. The function W is now continuous and its domain is a nonempty compact subset of 52. Weierstrass’ theorem guarantees that this function has a maximiser. 2. The function W* & [0,J] (I (-, c ' ' ]) is also continuous. Since I is left-closed, the domain of that function is (non-empty and) closed and bounded and therefore compact. Weierstrass’ theorem guarantees that this function has a maximiser. Making use of Lemma 1, we see that W also has a maximiser. Q.E.D.

2

To avoid misunderstanding:

[3, 8) is left-closed, but (f, 7] is not left-closed.

78


Note that the last condition in 2 of Theorem 1 is satisfied if K is unbounded. Theorem 1 motivates an examination of the continuity properties of W . Lemma 2 If Q is continuous, then W is continuous if and only if K is so.Ⴘ Proof. - Define the function [0, J ] u I o 5 by Q (t , c )

(t , c ) 6 ³ 0

P( x) dx P(Q (t , c))Q* (t , c) tQ (t , c) .

If Q is continuous, then so is this function. Since this function is nothing

else than the function (t , c) 6 W (t , c) K (c) , the desired result follows. Q.E.D. And lemma 2 in its turn motivates an examination of the continuity of Q .

Proposition 2 Q* is continuous in t and strictly decreasing in t. Ⴘ Proof. - Fix c I . Denote, for t [0, J ], g (t ) : Q (t , c) .Note that, using (11), g (t ) J . By Lemma 4(6), we have for all t [0, J ] 5, t P ( g (t )) U KT (c, g (t )). So, with f : J o 5 defined by f (q) : P(q) U KT (c, q) we have t f ( g (t )). Because the function f is continuous and strictly decreasing due to the properties of P and T(c, .), the inverse h of f is well-defined and also continuous and strictly decreasing. So for all t [0, J ] we have g (t ) h(t ). Thus g is continuous and strictly decreasing, as desired. Q.E.D.

Proposition 3 gives sufficient conditions for Q * being continuous. Its proof makes use of the implicit function theorem and is somewhat demanding since we have to take into account border points in order to reach the desired result. It may be noticed that the proof we offer does not use the inverse demand function P. Proposition 3 If x the function Q is continuously differentiable, x there exists an open interval I ' of 5 that contains I and an open interval Z of 5 that contains 5+, such that T can be extended to a positive


79

~

continuously totally differentiable function T : I ' u Z o 5 that is strictly increasing in q , then the function Q* : (0,J) I D 5 is continuously totally differentiable and the function Q* : [0,J] I 5 is continuous. Ⴘ Proof. - We prove this theorem by proving that for every

(t , c ) [0, J ] u I there is an open neighbourhood V of (t , c ) in 52 and a continuously totally differentiable function < : V o 5 , such that

Q

< on V ([0, J ] u I ) . Fix (t , c ) [0, J ] u I . Put q : Q t , c . Define the function 1 ~ M : Z u ( U , f) u I c o 5 by 2 ~ ~ M q; t , c : Q p KT c, q t q . ~ For each (t , c) ( 12 U , f) u I c , the function M (; t , c) has a unique zero,

say Q

(t , c) . Indeed, this function is strictly decreasing and therefore has at most one zero. And since the function is negative for large enough q , positive for q 0 and continuous, there is exactly one zero. Since

~ M (q; t , c) M t , c (q) (q 5 , t [0, J ], c I ) it follows that Q

Q on [0, J ] u I . In particular: ~ M (q , t , c ) 0. Noting among other things that Q is continuously differentiable, we see that ~ the function M is continuously totally differentiable. Since Q is decreas~ ing, it follows that Qc d 0 and since T is increasing in q we have ~ wT (c , q ) wT (c , q ) t 0 . wq wq Hence

~ wM (q ; t , c ) wq

~ wT ~

Qc( U KT (c , q ) t )K (c , q ) 1 0. wq

The implicit function theorem (see, for instance, Duistermaat and Kolk (2004)) now guarantees that there is an open neighbourhood U of q in 5 and

an

open

neighbourhood

V

of

(t , c )

in

52

with

80


U u V Z u (( 12 U , f) u I c) , such that there is a unique mapping < : V o U for which ~ M (< (t , c); t , c) 0 ((t , c) V ). Since we already know that

~ M (Q

(t , c); t , c)

0 ((t , c) V ) , ~

and for (t , c) V , Q * *(t , c ) is the unique zero in Z of M (, t , c) , it fol-

Q

on V. Since Q

Q on [0, J ] u I it follows that Q on V ([0, J ] u I ) , as desired. Q.E.D.

lows that
0. 2. 7* J , c

p0 c I .

3. 7 (t , c) ! p0 (t ! J , c I ) . 4. p0 J d 7 (t , c ) p0 (0 d t J , c I ) .

88


5. Q D 7

Q . P D Q .

6. 7

7. 0 d Q Q ( p 0 J ) .Ⴘ Proof.

T t T0 and U KT (c, q ) t t U KT0 p0 J ! 0 . 2, 3. They follow from (3) and T (c, 0) T0 (c I ) . –

1.

Observe

that

t 5 .

Therefore

4. The first inequality holds because of 1. With respect to the second inequality: it follows from (2) that Q (t , c) ! 0 . Using (11), we conclude

Q P t , c Q U KT c, Q t , c t 7 (t , c) p0 .

! 0 .This implies

5. This is nothing else than (11). 6. From (9) and 2, 4. 7. From 4, 5 and because Q is strictly decreasing on ( 0, p0 ] . Q.E.D. Proof of Proposition 1: ‘First equality’: the desired result follows from the identity

³

p0

a

Q ( p ) dp

³

Q(a)

0

P ( x) dx aQ (a ) (a ! 0)

and Lemma 4(5,6). ‘Second equality’: follows from Lemma 4(6). Q.E.D. Finally a remark concerning the possible values of t in the welfare function W . These are numbers from [0, J ] . It would have been perfectly possible to define W in a natural way also for t ! J . However, since we know how W would look like for t ! J and allowing t ! J makes the presentation more cumbersome, we assume from the outset that 0 d t d J . Let us explain this in greater detail. We notice in the first place that for every c I and t ! J the equation (1) in q t 0 has a unique solution. (Indeed, since M t , c (0)

Q( p0 J t ) , it follows from t ! J that M t , c (0) 0

and this implies the desired result since this equation has at most one zero.) The definition of Q is therefore also meaningful for values of t that are larger than J . In that situation Q (t , c)

0 (t t J , c I ) . Since the definition of 7 also remains meaningful for such t , this is also true for W . For


89

values of t larger than or equal to J we understand very well the function

W : for all t t J and c I the formula W * (t , c) K (c) would hold. Indeed: for t ! J , 7 t , c ! p 0 , Q (t , c ) 0 and Q 0 on [ p0 , f) . With this domain of the function W we would therefore have results like the following: W has a maximiser if and only if W* & [0,J] u I (i.e. the restriction of W to [0, J ] u I ) has a maximiser.

References Arnott, R. and Kraus, M., Financing Capacity in the Bottleneck Model, Journal of Urban Economics, 38, 1995, pages 272–290. Duistermaat, J. and Kolk, J., Multidimensional Real Analysis I: Differentiation, Cambridge Studies in Advanced Mathematics, 2004. Keeler, T. and Small, K., Optimal Peak Load Pricing, Investment and Service Levels on Urban Expressways, Journal of Political Economy, 85, 1977, pages 1–25. Kraus, M., Scale Economies Analysis for Urban Highway Networks, Journal of Urban Economics, 9, 1981, pages 1–22. Mohring, H. and Harwitz, M., Highway Benefits: An Analytical Framework, Northwestern University Press, Evanston, 1962. Oum, T. and Zhang, Y., Airport Pricing: Congestion Tolls Lumpy Investment and Cost Recovery, Journal of Public Economics, 48, 1990, pages 353–374. Small, K., Economies of Scale and Self-financing Rules with Non-competitive Factor Markets, Journal of Public Economics, 74, 1999, pages 431–450. Strotz, R., Margolis, J. (editor), The Public Economy of Urban Communities, Urban Transportation Parables, Johns Hopkins University Press, Baltimore, 1964, pages 127–169. Verhoef, E. and Rouwendal, J., Pricing, Capacity Choice and Financing in Transportation Networks, Journal of Regional Science, 44, 2004, pages 405– 435. Yang, H. and Meng, Q., A Note on Highway Pricing and Capacity Choice in a Road Network under a Build-operate-transfer Scheme, Transportation Research Part A, 2002, pages 659–663.

5 Local Collectors Versus Major Infrastructural Works

Catharinus F. Jaarsma and Wim Heijman Wageningen University, The Netherlands, E-mail: [email protected], [email protected]

Abstract. Cohesion between city and surrounding area is essential in the metropolitan landscape. However, this relationship is under strain. Firstly, the infrastructures, which are difficult to cross, are concentrated in exactly the transition zones of city and countryside, where – certainly for cyclists and pedestrians – the surrounding area is often difficult to reach from a residential area and vice-versa. Secondly, road users in the surrounding area are confronted with the loss of local connections because railways have been cleared of intersections or because new main infrastructure has been built. Current government policy considers this situation, but, in actuality, the focus is still too one-sided concentrating on construction costs and (railway) safety. A modification of this policy and a certain protection of local connections (including provisions for pedestrians and cyclists) within the framework of town and country planning are urgently needed. Because of the current developments, the metropolitan landscape is in danger of being divided into small ‘compartments’, which will only be connected by long indirect routes. A more integral approach and establishment of close-knit structures in Regional and Provincial Traffic and Transportation Planning Schemes can offer guarantees for more cohesion in the lowest categories of the road and path network. Key Words: Minor rural roads, Secondary roads, Railways, Rural traffic, Slow traffic, metropolitan landscape, Recreational networks.

92

Catharinus Jaarsma and Wim Heijman

5.1 Introduction In large parts of densely populated, industrialised western countries, the traditional borders between city and country are blurred. Also in the Netherlands, nearly all of its rural area lies within an urban sphere of influence. The cohesion between city and surrounding area is represented in the term “metropolitan landscape”. Characteristic of this is the essential role of the surrounding area for recreational (outlet) possibilities from the residential environment for city dwellers and village inhabitants. The recreational possibilities consist of, not in the very least, with short, daily excursions, mostly by foot or bicycle. Two conditions have to be met for the metropolitan landscape to fulfil its recreational role. Firstly, in order to be able to reach the surrounding area from the urban residential environment at all, enough connections are needed between both parts of the metropolitan landscape. This includes connections for cyclists and pedestrians. Secondly, once arrived in ‘his backyard’ (the surrounding area), the urban dweller wants to pleasantly spend (a part of) his free time there. In order to accomplish this second condition, the presence of a cohesive close-knit net of local connections is a conditio sine qua non. By local connections we mean the whole of “quiet” lower order roads (minor rural roads, with a maximum speed of 60 or, very exceptionally, 80 km/h and open for, at least, all slow traffic) and all the paved or unpaved pathways (open for, at least, cyclists and/or pedestrians). Both conditions for the operation of the surrounding area as outlet area are under strain. For quite a while now, there has been a barrier effect because of main infrastructures around cities (Van der Voet and Haak, 1989), but increasing traffic intensity and the construction of ring roads, even near villages, is intensifying this effect. But also in the surrounding area, the cohesiveness within the network of local connections is under strain, especially in places where they cross existing main infrastructure (the regional and national main roads and motorways, railways, rivers and canals) and with the construction of a new main infrastructure (motorways and high speed railways). Examples of this are the removal of existing level crossings on grounds of railway safety and the non-replacement of existing roads which are transected by a newly constructed motorway. In the most serious form, this can lead to “compartmentalisation” of the surrounding area, where one compartment is only accessible from the other via large indirect routes, along less attractive, busy connections. This infrastructural barrier effect forms an important – be it still unsatisfactorily recognized – ‘negative externality related to mobility and transport.’


93

The aim of this contribution is to illustrate the importance of local connections as a cohesive fabric in a metropolitan landscape and to illustrate, in real terms, the problems which occur with them. These problems are reinforced by the traditional sectoral approach. We argue, therefore, for an adapted policy (see also Jaarsma et al., 2005), in which more protection is offered to local connections through an integral approach than is currently the case. In earlier articles (Jaarsma en Heijman, 2004; Jaarsma et al., 2004) we have indicated possibilities to objectively come to such an integral assessment of interests. This chapter is organised as follows. First, we examine the importance of local connections (sec. 2). Afterwards, the current policy is outlined (sec. 3). Then, using examples taken from actual situations, we will discuss how the fabric of cohesive connections threatens to crumble despite protective policy intentions (sec. 4). The chapter will end with conclusions, recommendations and a short discussion (sec. 5).

5.2 The Importance of Local Connections The importance of local connections is best illustrated with examples of its use. Unfortunately, systematic data on this subject are not being collected. Incidental measurements of rural roads mostly consist of only motorised traffic. The available counts show that the numbers are often quite large in relation to the residence function of these roads: on average at least a few hundred cars per day. Numbers of a few thousand also occur. Because of the mixed composition of the traffic on these narrow roads and the frequently high speeds of car traffic, pedestrians and cyclists, in particular, get into problems. This group would certainly profit from a cohesive network of local connections that are less heavily trafficked and in which sufficient links are found with low(er) numbers of motor vehicles. The use of local connections can also be approached by looking at the activities that the Dutch undertake in their free time. Therefore, we focus on those activities which are normally carried out in the fresh air and in the countryside. Research from the Dutch national statistical office, ‘Centraal Bureau voor de Statistiek’ (CBS), among others, shows that almost everyone (98% of the population) takes a stroll or makes day trips. In 2002 77% of the population occasionally took ‘a walk for enjoyment’, 68% went for a bike ride, 15% did some physical exercise or went running and 5% went ‘horseback riding outside’. However, there are a range of activities that make up these general categories. Walking is comprised of ‘walking for

94


enjoyment’, ‘wandering around’, ‘browsing about’, ‘taking the dog out for a short walk’, ‘taking a walk with the pram’, ‘taking a stroll’, ‘getting some fresh air’, ‘getting some physical exercise’, ‘working out’, or ‘going for a walk during the day, the weekend or on holiday’. A similar differentiation is made with cycling and equestrian sports. It is a diversity of activities with different meanings, backgrounds, motives and behaviours. For many of these activities, an individual does not travel far: see Figure 5.1. Most of the day trips take place close to home (40% up to 5 km). People doing physical exercise/joggers/runners as well as many walkers and many people who ride horses stay within a distance of 5 km (respectively: 63%, 46% and 45%). This figure illustrates the importance of the metropolitan landscape, in which (small) daily outings can be made from the residential environment of the city dweller or villager into the direct environment.

Fig. 5.1. Recreational day trips (2 hours and longer) in the Netherlands to the most important activity and the distance travelled (in percents). Source: ‘Nederlands Research Instituut voor Recreatie en Toerisme’ (NRIT), Day Recreation in the Netherlands 2002/2003 (NRIT, Dagrecreatie in Nederland 2002/2003).

Figure 5.2 shows the surroundings where these activities are carried out. For most of the activities, about 20% are carried out in a city/village. In relation to Figure 5.1, it may be assumed that activities in, for example, agricultural areas and woods, are also often pursued close to home. It can be concluded that the place of residence, if this is now a city or a little village centre, and its direct surroundings are the most important for everyday forms of recreation. This counts all the more if we consider that the results refer to trips of 2 hours and longer. In the framework of Dutch governmental policy, ‘Green In and Around the City’, (in Dutch: ‘Groen in en om de Stad’) zones of 10 km around the cities are looked at. We emphasise, however, that not only city dwellers but also inhabitants of centres in rural areas need outlet possibilities. These certainly are not – as a matter of course – present in agrarian areas. Furthermore, the surroundings of other user places of origin are also important, such as the surroundings of overnight recreational accommodations, stalls, public transportation stops and parking places/transfer points.


95

Fig.5.2. Recreational day trips (2 hours and longer) in the Netherlands to the most important activity and surroundings (in percents). Source: ‘Nederlands Research Instituut voor Recreatie en Toerisme’ (NRIT), Day Recreation in the Netherlands 2002/2003 (NRIT, Dagrecreatie in Nederland 2002/2003).

The dense net of local connections could tempt someone to think that “with so many links, one won’t be missed.” However, this is incorrect: every removal of a crossing has consequences; it leads to barrier formation and forced detours no matter if it is walking, cycling or riding. If the detours become too large for pedestrians and cyclists, people may begin using their cars. Another factor which plays a role in recreational activities is the reduced possibility for taking walks and bicycle rides, where one will be continually forced to use the same road. In the most serious case, things can lead to the complete or partial functional loss of areas and the local facilities which are already there. This illustrates that the loss of “only a simple little link” can actually have serious consequences. The seriousness and the extent of the consequences are connected with the speed of the road user, the mesh of the network and the distance to and the quality of the alternative crossing. In Table 5.1, an estimate has been made of the duration of a detour in a number of situations. Table 5.1. Estimated duration for a detour along the parallel road in connection with the means of transportation and the distance between the crossings across the barrier.

Distance Between Crossings [m] 500 1000 2000 4000

Detour [minutes] By foot (4 By horse km/h) (7,5 km/h) 7,5 4 15 8 30 16 60 32

By bicycle (12 km/h) 2,5 5 10 20

By car (60 km/h) 0,5 1 2 4

Source: www.recreatieenoverwegen.nl (2004)

As is to be expected, pedestrians are the most vulnerable group. It should be emphasised that this quantitative approach is not sufficient: functional demands are also made on the roads and paths. For walks, at least two crossings are always needed. In order to facilitate a one hour stroll for

96


walkers, the crossings cannot be more than 1 to 1.5 km from each other maximum. Furthermore, it shall be clear that the ‘parallel roads’ alongside the main connection, which have been assumed in Table 5.1, are in actuality not always present and certainly do not always meet the demands for safety and attractiveness. An additional effect of the reduction in the number of crossings is the concentration of traffic on the roads with the remaining crossings. An increase, particularly in the motorised traffic, makes these connections less attractive for recreational use. Objective and/or subjective traffic safety can also be an issue, certainly for slow traffic participants.

5.3 Current Policy Around Local Connections How does Dutch governmental policy view local connections in the metropolitan landscape? We will discuss this with regard to the ‘Policy Document on Mobility’ (‘Nota Mobiliteit, anon., 2004a), which, such as presented in different policy documents, has to be considered in connection with the ‘Policy Document on Land Use’ (‘Nota Ruimte’, anon., 2004b), the ‘Agenda-Paper on the Vital Countryside’ (Agenda Vitaal Platteland, anon., 2004c), the ‘Long Term Programme on Habitat Defragmentation’ (anon., 2004d) and the ‘2nd Framework Policy Document on Rail Safety’ (‘2e Kadernota Railveiligheid’, anon., 2004e). All these documents are national policy documents, so focusing on a national scale. Further, it should be noted that the last three documents mentioned are dealing with specific sectoral problems: rural development, habitat fragmentation and railway safety respectively. The ‘Policy Document on Mobility’ is the national traffic and transportation plan for the Netherlands. This policy document elaborates on the ‘Policy Document on Land Use’ (see below) with regard to traffic and transportation policy. The goal of this policy document, “Getting the Netherlands Moving Again” is strongly economically oriented. However, consideration is given to improving the quality of the living environment, but this is limited to emission reductions and the defragmentation of the natural habitats of plants and animals. An important statement in relation to our subject is “different road networks and means of transportation have to be looked at together”, while “possible solutions also have to be fully analysed”. The ‘Policy Document on Mobility’ is especially focused on the major road network, but also states that improvements to the connections in the underlying network can also be considered. It can be assumed that the provincial secondary road network is primarily being considered here.


97

However, this brings with it the risk of rat-run traffic (cut-through traffic avoiding congestion on main roads) on the local connections. That would be contrary to the other role of these capillaries, namely as part of a network of safe bicycle routes. On this issue, the ‘Policy Document on Mobility’ states that “The government, along with the decentralized governments, is responsible for the maintenance, repair and improvement of bicycle routes, where these are cut across by road or railroad track. Where possible, the central government shall take into consideration bicycle facilities when constructing infrastructural works.” The ‘Policy Document on Land Use’ (the Dutch national report for physical planning) is aimed at “the spatial contribution to a strong economy, a safe and liveable society and an attractive country”. In the context of our subject, it is important to note that the existing greenery often does not satisfy the current quality demands. The ‘Policy Document on Land Use’ recognises three problems: (1) a multiple fragmentation by barriers, such as railways and (motor)ways, (2) insufficient connections between greenery in the city and surroundings and (3) the surroundings themselves are only moderately accessible for recreational use. Recognition of this problem offers contact points for preservation and development of local connections, but, on the other hand, other objectives of the physical planning policy (strengthening the competitive position and promoting powerful cities) form a threat. In the ‘Agenda-Paper for a Vital Countryside; A Vision’ the changed perspective of the countryside is sketched, now that a fundamental change is coming about: the transition from a rural area as physical space for food production to the countryside as consumption space in which central values such as authenticity, naturalness and quality play an important role. It is being proposed that the countryside has essential functions for its inhabitants and for the urban dwellers, specifically for “living, working and enjoying”. Therefore, the cabinet wants to broaden the possibilities to experience the countryside. Remarkably, new possibilities are especially being considered while not a word has been mentioned about maintaining the existing possibilities and about measures which are necessary to better utilize the existing strengths. The ‘Long Term Programme on Habitat Defragmentation’ is aimed at the maintenance and improvement of the ecosystems. This is important to our topic because plant and animal variation is also important for people because they “can enjoy nature and recreate there”. Under certain circumstances, ecologically focused defragmentation measures can also be suitable for walkers and cyclists and vice-versa. Consequently, fragmentation as a problem for people comes into consideration in this policy piece, namely: the importance of the “prevention, removal and softening of barri-

98


ers”, both in the living and vacation environment, in order to be able to walk and cycle “in closely knit but open nature and rural areas”. It has been confirmed that much of our nature and rural areas have become more poorly accessible in the last century. In the ‘2nd Framework Policy Document on Rail Safety’, a new foundation has been laid for the period 2004 to 2010 for the goal around railway safety. The goal includes pursuing permanent improvement of railway safety for a socially acceptable cost, in which, if possible, quantitative goals are set. Nine themes have been set out for this purpose. In the framework of our subject, theme 3: the safety of the risk bearers, is interesting because of level crossing safety. The five year average of the number of victims from level crossing accidents in the Netherlands has declined from more than 40 per year in the first half of the nineties to 30 in 2002 and 2003; the target figure for 2010 is 24 or less. The decline in the number of victims in the past period is attributed to the removal or better safety measures of 950 level crossings in the period 1999 through 2003, in the framework of the ‘Programme for the Improvement of Safety at Level Crossings’. From analysis, it turns out that a safety device with lights, bells, and half level-crossing barriers is safer than a safety device with a St. Andrew’s cross, lights and bells by a factor of 10. The ‘2nd Framework Policy Document on Rail Safety’ names a number of situations in which street level crossings are not (no longer) allowed. It has also been determined that improvement of an unprotected level crossing can consist of more than just removal, whereby the recreational function can also play a role. Luckily, there is, at least on paper, consideration for the special position of slow traffic: “The closing off and combining of level crossings can, especially for slow traffic, have the effect that the railway track becomes a difficult barrier to cross. In addition to the longer route which has to be taken in order to pass by the railway, slow traffic will also be confronted with the additional risk of the road traffic. Using a risk analysis or an area oriented approach, we should see if the level crossings can be maintained for slow traffic with a low risk profile.” From the previous information, a picture emerges of a national government that is aware of the importance of the surrounding area and its network of local connections within the context of the metropolitan landscape. However, the policy resolutions formulated in the policy documents sometimes show contradictions and/or seem to insufficiently consider the factual developments which are at odds with the intended policy. Examples of potential contradictions are the use of the underlying network as proposed in the ‘Policy Document on Mobility’ and promoting powerful cities in the ‘Policy Document on Land Use’. Factual developments such as a decreasing accessibility of our nature and rural areas are mentioned in


99

both ‘Long Term Programme on Habitat Defragmentation’ and ‘2nd Framework Policy Document on Rail Safety’, however without any proposal for a solution. The sectoral character of these documents may play a role here, because the complexity of the problems calls for an integral approach. To illustrate the decreasing accessibility of rural areas by actual developments related to transportation infrastructure, we present in the following chapter two examples.

5.4 Some Practical Experiences: the Disappearance of Local Connections Construction of new major infrastructure, both motorways and high speed train lines, is always accompanied with new transections of the rural area. This also cuts through local connections, which in practice are never completely restored. An example of such an abrupt intervention in the landscape is the construction of the A73 motorway near Venlo. This example is representative for other areas with such major construction works. The barrier formation sometimes displays itself as an insidiously appearing deterioration of the existing situation. This is the case with our second example: the implementation of policy around railway safety for the Brabant city triangle area in Middle Brabant. The second example is representative for all rural areas traversed by railways and therefore submit to the policy of the ‘2nd Framework Policy Document on Rail Safety’. Our first example (Jaarsma et al., 2004) is concerned with the construction of the A73 in Middle Limburg (Figure 5.3). The A73 in Middle Limburg cuts through the eastern Maas Valley, bordered by the Maas River and the steep edge of the Maas Terrace near the German border. In the area, there are residential areas and commercial zones from Venlo/Tegelen and Roermond, the centres Belfeld, Reuver, Beesel, Swalmen to the north of Roermond and Linne and Maasbracht to the south of these. In addition to agrarian use, the rural area also has country estates, woods and east-west running stream valleys. The rural area and certainly the green wedges between the residential areas are heavily used by cyclists and walkers from the area and from elsewhere. Among other things, there is a bicycle junction network, walking routes/hiking trails, places to stay and stables. Even without the A73, the area has north-south oriented corridors because of the parallel running Maas, the old national trunk road (N271) and railway line Venlo-Roermond-Echt-Sittard. Except for the construction of the A73 (42 km in total, of which 35 km between Echt and

100


the Maas near Venlo on the eastern bank of the Maas), the construction/extension of the A74, N280 and N293 are being carried out and the level crossings are being reorganised in this area.

Fig. 5.3. Construction A73/A74 in Middle Limburg (left) and consequences for the local connections on the right Maas bank between Reuver and Venlo/Tegelen (Jaarsma et al., 2004).

The most important effects for the metropolitan landscape on the east bank are: x Division of the urban area Venlo/Tegelen: in Tegelen, attractive, crossing possibilities for slow traffic, which are less heavily trafficked, are lacking over a distance of 3200 m. x By planning this road directly to the east from Belfeld, Reuver, Beesel, Linne and Maasbracht, these places lose their less heavily trafficked outlet possibilities in an easterly direction. x By planning this road directly west from Swalmen, this place loses its less heavily trafficked outlet in the direction the Maas. x Roermond loses its less heavily trafficked outlet possibilities on the north and east side. The Pieter Path, a LAW (long distance footpath), runs for about 2 km through a motorway landscape because of the N280


101

and the A73. Only near the A73 tunnel in southeast Roermond is there no barrier formation. x The mapped out road (35 km) crosses 71 cyclable roads and paths. They are being replaced by 13 viaducts, 2 tunnels and 19 crossings at ground level near a sunken A73. From the remaining 34 crossings, 4 are specifically for cyclists/pedestrians and 13 are relatively lightly trafficked. The remaining 17 are important connecting roads and/or connections to the A73. 18 of the 47 level crossings on the railway line VenloRoermond-Echt are being considered for removal; 9 from these in connection with the A73. A number of them have already been removed. x Limitation of a number of crossings means a concentration of mixed traffic flows on the remaining crossings. Considering crossings usable for slow traffic without access and exit ramps, the distance between these crossings varies from about 350 m to 2300 m, locally increasing to 4500m. The crossing of the corridor itself via construction and corresponding connections demands distances which vary from ca. 350 m to 2 km. Use of a wild animal crossing by pedestrians can avoid the 4500m mesh and the roundabout routes via construction and connections of 2000m. x Green wedges, country estates between Tegelen and Belfeld, Belfeld and Reuver and Beesel and Swalmen, will be fragmented into unreachable snippets. x Damage to ecological values in the green wedges will be mitigated through roughly 45 eco-passages, among other things. Only one of these will be made suitable for recreational use: the passage from the MaasSchwalm-Nette-Route (part of the LAW network). This example illustrates the strong compartmentalisation of the area. Over large distances, north-south oriented compartments of less than 500 m to compartments of 1000 m wide form between the Maas, railway line and the motorway (Teunissen, 2003). This has large consequences for the inhabitants of the area, for the recreational use and for the slow traffic. We have to fear that a new motto is coming into being: “Want to get out for a bit? Get the car!” Our second example illustrates the effects of railway safety policy implementation for the accessibility of the metropolitan landscape in the Brabant city triangle area. This area is referred to as Middle Brabant. It has been commented on in the ‘Policy Document on Land Use’ (‘Nota Ruimte’) as (1) a so called robust connection because it lies in the EHS (=national ecological network), (2) National Landscape ‘The Green Forest’ (‘het Groenewoud’) and (3) an important outlet area for the Brabant city triangle area. On the other hand, it is criss-crossed by roads, railway lines

102


and water courses of national and regional importance. In regard to the railway, the conversion of the line Boxtel-Eindhoven from a 2 to a 4 track trajectory plays a role; tunnelling from the area of Best and making the above ground track crossing free by removing the level crossings and constructing a limited number of tunnels/viaducts. The main roads in this area are (A2) or will gradually be extended (N65) to complete motorways. Figure 5.4 shows the fragmentation between Best and Boxtel caused by the A2 and the railway line.

Fig. 5.4. Barrier formation by the A2 and railway line; the bicycle/walkers/horse riding tunnel halfway between Best and Boxtel under the railway line does not continue under the A2; these routes pass over the A2 via the busy rat run Boxtel-Liempde or go parallel with the ring road of Best.


103

Overpasses will be constructed, in part, on local roads. Such a barrier exists between Best and Boxtel and is 4.5 km (about 1 hour by foot). The environment there is defragmented by an ecoduct which is inaccessible for recreational use. Meshes of 2.5 km between crossings on the motorways occur frequently, often because of busy roads and access and exit ramps. Level crossings are also being discussed in other parts of Middle Brabant. ProRail wants to make its lines free of level crossings. This can mean the removal of level crossings which will be replaced by a smaller number of tunnels and viaducts. Both of these examples are symptomatic of the prevailing approach: a strong sectoral approach (such as a one-sided focus on ecology or railway safety) and/or tendencies towards a minimalisation of construction costs. The result is often a strong limitation on the number of possible short trips, especially for cyclists and pedestrians, because of the reduction of the number of crossings with Fremdkörper (foreign bodies) in the metropolitan landscape. However, it seems repercussions for other interests have not been considered.

5.5 Conclusions, Recommendations and Discussion From the policy discussion in chapter 3, it turns out that the danger of barrier formation, fragmentation and compartmentalisation are being pointed out in the ‘Policy Document on Land Use’. Experience (chapter 4) shows that the metropolitan landscape will become even more compartmentalised without giving concrete form to the proposed policy and sweeping changes in the present policy. Actual developments occur to the detriment of both the green living space and the necessary small scale infrastructure of roads and paths, the local connections, which are used for short daily trips in the green space. Discrepancy between the policy intentions and the policy implementation is being reinforced by one-sided sectoral approaches and by focusing on one aspect, such as construction costs or (railway)safety. In the construction of the major infrastructure, the urgency to focus on the underlying road network is lacking. With regard to the area of policy dealing with defragmentation, no explicit connection is being made between the interests of the regional inhabitants and their outlet possibilities and the interests of the environment with regard to ‘robust connections’ (EHS). Although in the ‘Long Term Programme for a Vital Countryside’ the opening of nature counts as an important condition for the commitment of government revenues for the realisation of environmental aims, the number of

104


eco-ducts that have been opened for walkers/ hikers and/or cyclists so far can be counted on one hand… There are certainly possibilities for a wellfounded, complete assessment, aimed at infrastructure: a cost-benefit analysis (such as the Research project Economical Impacts of Infrastructure (OEEI)-method; see CPB, 2000 and Van der Straaten et al., 2003); a multi-criteria analysis (such as the Dutch national Information and Technology Centre for Transport and Infrastructure (CROW) maintenance method; see Jaarsma en Webster, 1998) and a risk analysis (for example for railway safety; see anon. 2004e). All of these methods have specific advantages and disadvantages. It can be useful to apply them simultaneously (Jaarsma et al., 2004). When the processes, which have been pointed out in the first paragraph, are not stopped, this will have negative consequences for the quality of the residential and living environment. This continuing compartmentalisation threatens our society more and more – and threatens more and more activities – with becoming even more dependent on cars. The goal in the ‘Policy Document on Mobility’, “Getting the Netherlands Moving Again”, is not getting any closer with the rising compartmentalisation. Dependence on a car is little desired for life in the metropolitan landscape. Better would be a policy, in which the local connections as crucial factor for reachability and accessibility of the surroundings get the (economic) appreciation and the (administrative) protection (within the framework of town and country planning) that they deserve. On this point, the provinces could take the initiative for an integrated approach on a regional scale with their physical planning and with their transportation planning by demanding the cohesion of the network of local connections.

References Anonymous, 2004a. Policy Document on Mobility. Toward a Reliable and Predictable Accessibility. Ministry of Transportation and Water Management/Housing, Planning and Environment, The Hague (Nota Mobiliteit. Naar een betrouwbare en voorspelbare bereikbaarheid. Ministerie van V&W/VROM, Den Haag). (160 p). Anonymous, 2004b. Policy Document on Land Use. Space for Development. Ministry of Housing, Planning and Environment/Agriculture, Nature Conservation and Fisheries/Transportation and Water Management/Economic Affairs, The Hague. (Nota Ruimte. Ruimte voor ontwikkeling. Ministerie van VROM/LNV/ V&W/EZ, Den Haag.) (216 p). Anonymous, 2004c. Agenda-Paper for a Vital Countryside: Anticipating Change. A View. Ministry of Agriculture, Nature Conservation and Fisheries, The


105

Hague. (Agenda voor een Vitaal Platteland: inspelen op veranderingen. Visie. Ministerie van LNV, Den Haag). (64 p). Anonymous, 2004d. Long Term Programme on Habitat Defragmentation. Ministry of Transportation and Water Management /Agriculture, Nature Conservation and Fisheries/Housing, Planning and Environment. The Hague. (Meerjarenprogramma Ontsnippering. Ministerie van V&W/LNV/VROM, Den Haag). (127 p). Anonymous, 2004e. Safety on the Railway, ‘2nd Framework Policy Document on Rail Safety. Ministry of Transportation and Water Management. The Hague. (Veiligheid op de rails. Tweede kadernota voor de veiligheid van het railvervoer in Nederland. Ministerie van V&W, Den Haag). Central Planning Office, 2000. Evaluation of the Infrastructure Project: Guidelines for a Cost-Benefit Analysis. SDU, The Hague, 2000. (Centraal Planbureau (CPB), 2000. Evaluatie van infrastructuurprojecten: leidraad voor kostenbatenanalyse. SDU, Den Haag, 2000). Jaarsma, C.F., W.J.M. Heijman en J.L.M. van der Voet, 2004. The Necessity of Local Connections. (De noodzaak van lokale verbindingen). Verkeerskunde 55, nr 10: 32-39. Jaarsma, C.F., W.J.M. Heijman en J.L.M. van der Voet, 2005. Local Connections: Worth the Maintenance. (Lokale verbindingen: het handhaven waard). CROW Verkeerskundige werkdagen 13 : paper 4.2 on CD-ROM (11 pp). Jaarsma, C.F. en M.J. Webster, 1998. Local Connections and Main Infrastructure. Removal or Maintenance? (Lokale verbindingen en hoofdinfrastructuur. Opheffen of handhaven?) CROW-publicatie 127, Ede. ISBN 90 6628 2746 (102 pp). Straaten, J.W. van der, W.J.M. Heijman en J.A. Fanoy, 2003. A Method for the Cost-Benefit Analysis of Infraprojects. (Een methodiek voor de kostenbatenanalyse van infraprojecten). Wegen 77 nr 6: 10-13. Teunissen, L., 2003. Barrier Formation on the Motorway A73. (Barrièrevorming bij autosnelweg A73) (Venlo-Echt). Rapport Fietsersbond, Utrecht. Voet, J.L.M. van der en M.Th. Haak, 1989. Towards a Recreationally Accessible Countryside? (Op weg naar een recreatief toegankelijk landelijk gebied?) Landinrichting 29 nr 5: 31-40.

Part II Clusters and Product Chains

6 Regional Differentiation and Location of Industrial Capacity in the Slovak Republic

Jana Gašparíková, Edita Nemcová and Michal Páleník Slovak Academy of Sciences, Slovak Republic, E-mail: [email protected], [email protected], [email protected]

Abstract. Regional disparities resulting, above all, from lasting disproportional location of industrial capacities, have shaped Slovakia’s development in recent years. The goal of the chapter is to map regional employment differentiation in the SR, according to high-tech and low-tech manufacturing sectors. To achieve this we worked with recent regional statistical data. We concentrated on this question: Does there exist a substantial relationship between location of high-tech manufacturing and employment growth in selected regions. According to our survey, the hightech sector did not prove to have strong impact on total employment. However, considering the changing structure of manufacturing in the future, it will represent one of the most decisive issues concerning competitive growth. Key Words: Regional disparities, Structure of manufacturing sectors, Slovak regional employment.

6.1 Introduction Regional differentiation is a very important issue, as this influences the entire character of socio-economic processes in the Slovak Republic. After the revolution in 1989, the command economy was changed to a more open economy. This turnover in Slovakia’s economy brought turbulent changes being both positive and negative. One of these mainly negative changes concerned not only abrupt regional differentiation in the Slovak Republic, but also very immense increases in regional disparities.

110

Jana Gašparíková, Edita Nemcová and Michal Páleník

The question is if the regional differentiation of the Slovak Republic is influenced by industrial agglomerations and industrial districts, or by other factors. Since the work of Alfred Marshall (Marshall 1920), the studies on differing industrial clusters tried to define dimensions, which help in understanding the contexts in which different firms operate and what is behind the industrial concentration of different clusters and industrial districts. Porter (Porter 1990) defined cluster externalities using a different term: competitive advantage. He explained a firm’s competitive advantage in certain areas as a result of the interaction of factor conditions, demand conditions, related and supporting industries or firm strategy, and structure or rivalry. We will concentrate our ideas on Porter’s (Porter 1990) thinking, based on related and supporting industries. Several types of industries are the influential backbone of Slovakia’s new economic development, but on the other hand, there are also fundamental factors occurring and shaping regional economic disparities. What is necessary is to take into account the change that is observable after the transition period in our economy, where there is recorded an interesting shift from industries supporting low technologies to industries supporting medium-low technologies (this problem is addressed in the second part of this chapter). Several authors have, within recent years, reported the importance of high-tech1 branches for economic development (Rausch 1998). There is a wide range of reasons for the significance of these branches, as was quoted in several studies. The first one is that high-tech firms innovate more and such firms tend to gain market share, create new product markets, and use available resources more productively. The second reason is that high-tech firms are associated with high value-added production and success in foreign markets, which generally helps to support higher remuneration to the people they employ. And last but not least, industrial R&D performed by high-tech industries has other spill-over effects. It is also connected with

1

There is no universally accepted definition of “high-tech”, nor is there a standard list of industries considered to be high-tech (Hatzichronoglou 1997). The OECD prepared a classification consisting of two lists: one for manufacturing industries (the sector approach) and another for manufactured goods (the product approach), which was elaborated as a supplement and was more appropriate for analysing international trade. The following four groups in the manufacturing industry were identified according to technology intensity: 1. high-technology 2. medium-high-technology 3. medium-low-technology 4. low-technology Due to available NACE 2 digit level data uses, article, the Eurostat and OECD breakdown of the manufacturing industry according to the technology intensity [EC 2002, EUROSTAT 2004], which is specified in methodological notes.

6 Regional Differentiation and Location of Industrial Capacity

111

the shift in type of employment towards high tech employment in industry, and a decline of employment in heavy industry.

6.2 Industrial Development in the Slovak Republic with Special Impact on Technology Intensity The current structure of Slovak industry took roots in the past when the economy of this predominantly agriculturally oriented region of the former communist Czechoslovakia, was industrialized according to CMEA requirements. As a result, the comparison of the industrial profile of the SR with more and less advanced EU countries indicates an excess of heavy industry. After 1989, the branch structure of the Slovak economy changed fundamentally. At the beginning of the transformation, industry had to adjust their production to market demand and to adapt to the redirection of exports and imports from the CMEA to EU markets2. This period saw the biggest decrease in Slovak industrial production and employment. Along with its declining GDP share of 25.1% in 1995, 23.6% in 2003, the branch structure of manufacturing as the decisive part of Slovak industry changed considerably. In recent years, the structure of Slovak industry has begun to show the first signs of convergence with those of developed European countries. In spite of this, the share of labour and energyintensive branches is significantly higher, while the share of more sophisticated production has remained relatively low. From 1993 to 2003 the share of the transport equipment industry has changed most significantly. Its share of total manufacturing production rose from 4.6% in 1993 to 26.0% in 2003. Due to an expanding automotive industry (Brzica 2003), this sector has been the largest and most important manufacturing branch in recent years. While the automotive industry is not a high-tech industry, it is a major driver of new technologies and the diffusion of innovation. Its close re2 There was a strong share of industries featuring a lower degree of procession and high raw material, energy and transport intensity. Therefore, a shift from predominantly resourcebased manufacturing towards less intensive labour and energy branches was attempted. Structural changes were concentrated on a few important industries, which depended on CMEA markets to a large extent, such as mechanical engineering (above all, arms production), transport equipment, the metal industry, food industry. At the beginning of transition, the high above-average share of coke and refineries was most prominent, together with a considerable surplus in mechanical engineering. Being an important place for arms production in the CMEA, the problem of modernizing the mechanical engineering sector was aggravated by the problem of conversion. The majority of substantial problems of less developed regions with high unemployment rates could be derived from the conversion.

112


BE

DE

EL

high-tech

ES

3,5

FR

PT

SK

39,7 12,9

5,8

40

37,3 13

4,5

5,6

9,4

28,4

38,9

46

55,6

EU 15

7,9

38,1 9,7

100 90 80 70 60 50 40 30 20 10 0

61,8

lationship to other manufacturing branches (chemicals, plastics, electrical and electronic parts, etc.), contributes to rapid diffusion of new technologies (European Competitiveness Report 2004). Compared with the EU average, the Slovak Republic has a smaller proportion of people working in the high-tech and medium high-tech sectors of manufacturing and services [Baláž 2004]. In the Slovak Republic only 0.8% of the total employment sector was employed in high-tech manufacturing. Compared to total manufacturing in the EU however, the high-tech and medium high-tech industries did marginally better in terms of employment growth (Eurostat 2003). Comparing the share of high-tech and low-tech manufacturing on total employment in the SR, with selected more or less developed EU countries in 2000 gives an interesting picture (see Figure 6.1).

UK

low-tech

Fig. 6.1 Share of high-tech and low-tech manufacturing in 2000. Source: EUROSTAT. The share of high-tech manufacturing in the SR, in 2000, was lower than the EU average, as well as in more developed EU countries. On the other hand, compared with less developed EU countries such as Greece, Spain and Portugal, the Slovak Republic achieved better results.

From 1995 to 2003, the percentage of the share of total manufacturing employment in the four groups in the SR, is presented in Figure 6.2. There are no significant changes in the employment share of more technology intensive industries from 1995 to 2003. Their share of total manufacturing employment dropped from 39.3% in 1995 to 37.4% in 2003. The share of low-tech industries has been shrinking in favour of medium-lowtech industries. In spite of this, more than 60% of employees in manufacturing were still employed in medium-low-tech and low-tech branches. Compared with the growing production share of technology, for more intensive industries (see Figure 6.3), this may be connected with increasing labour productivity in high-tech and medium-high-tech industries.


113

50 40 30 20 10 0 high-tech

medium-hightech 1995

medium-low-tech

low-tech

2003

Fig. 6.2 The share of employment according to technology intensity in SR (manufacturing = 100%). Source: Statistical Office of the SR.

45 40 35 30 25 20 15 10 5 0 high-tech

medium-high-tech 1995

mediumlow-tech

low-tech

2003

Fig. 6.3 The share of industrial production according to technology intensity in SR. Source: Statistical Office of the SR.

During this entire period, the share of high-tech industries has increased very modestly. The most important reason for the most significant expansion of medium-high-tech industries has been the development of the automotive industry. However, the technology level of Slovak industry with a more than 50% share of medium-low-tech and low-tech industries has still remained insufficient. The comparison of labour productivity with the EU average in 2000, also reflects the underdeveloped technological level of Slovak manufacturing (see Figure 6.4)


114

80 70 60 50 40 30 20 10 0 high-tech

medium-high-tech medium-low-tech EU-15

low-tech

Slovak Republic

Fig 6.4 Labour productivity according to technology intensity in 2000. Source: EUROSTAT.

Labour productivity in all four groups of manufacturing in the Slovak Republic was significantly below the EU average. The technological level of Slovak manufacturing is not only lower compared to that of old EU member countries (it is one tenth of the standard of the EU-North countries – Belgium, France, Germany and Great Britain) but also – according to some authors [Outrata 2002] – to that of other new EU member countries. These unfavourable conditions can be attributed to low research and development expenditures in the long term. These explanations help us to see not only the differences among hightech sectors and low-tech sectors in different countries in the EU, but also in the Slovak Republic. In a way, understanding the situation in the Slovak Republic also requires concentrating on the regional differentiation influencing the recent allocation of industrial enterprises and employment in industry.

6.3 Regional Differentiation in Employment This transitional shift is influenced also by human resources, which are characterised by good educational levels, but also by insufficient professional skills and knowledge reflecting new market conditions and investor demands. There exist strong differences among qualified structural characteristics of human resources, not only between Bratislava and other territories, but also among other individual territories of the Slovak Republic.


115

Table 6.1 Selected Indicators in Industry in 2002 (SR) Source: Regions of Slovakia, VEDA, Publishing House of the Slovak Academy of Sciences, Bratislava, 2004, p.85.

Regions

Turnover in Labor Productivity Current Turnover Prices

Bratislava Trnava Trenþín Nitra ħilina Banská Bystrica Prešov Košice Total

463 327 91 067 115 038 85 205 117 315 90 380 64 462 138 342 1 165 135

4 471 113 1 952 337 1 237 847 1 511 428 1 767 147 1 316 648 1 182 885 2 140999 22 104 191

Average Average of Monthly for Number Wage per Employees Employee 103 627 19 074 46 645 13 941 92 934 12 828 56374 12 468 66 387 13 469 68 644 12 656 54 496 11 430 64 616 15 668 553 721 14 303

Regional differentiation is also deeply influenced by social mobility in the Slovak Republic. Social mobility is not very strong for different reasons - especially as a consequence of typical traditional Central European life styles, and conservative reality of the market. Certain flexibility in mobility is recorded in localities, where large cities and towns are located. Regarding what concerns the negatives of regional differentiation we need to speak about a total decrease of demographic vitality in the Slovak Republic and an increase of regions with an insufficient demographic reality. These regions also slow down their social and economic development and dynamics. This is typical with an insufficient labour market, where especially in small towns this tendency is visible. The labour market will be crucial for young people, who move to bigger cities. They do not want to stay in underdeveloped regions for the simple reason they will not find job possibilities and good conditions there. All those negative side effects will influence regional disparities in the future as well. In these regions, it is also difficult to support the development of SMEs and their supportive programs, because there is not very sufficient infrastructure. Concerning the positives of regional differentiations, we must admit there exists much industrial differentiation and its large territorial and regional diffusion (see Table 6.1). The 1970s experienced strong differentiation among regions. The concentration of economic activities and consequently population had an expressly selective character. Industrial regions and cities resulted from this concentration. Regions were polarized according to urbanization centrelines and tracks. Non-urbanized spaces were developed as marginalized regions, and especially neighbouring zones situated in the south and north and in some other regions of the east. In that time, an integrated waistline of problematic regions was formed, where the situation gradually deterio-

116


rated. On the other pole of development were located such towns as Bratislava, Košice, Banská Bystrica and Trnava. When analysing the influence of regional differentiation we must face the question of regional difference in GDP [FalĢan – Pašiak 2004], especially after 1989, because its definition starts to be important in terms of regional structure, innovation activity, regional vicinity and manpower training. The foundation and activity of new entrepreneurial subjects, physical and legal entities, had different dynamics among the regions. This fact influences the difference in individual regions. The regions with the highest poverty rates are probably Kežmarok, Sabinov, Vranov nad TopĐou, Gelnica, Spišská Nová Ves, Snina, and the Košice- suburbs located in the territory of eastern Slovakia; Revúca and the other four regions - Rimavská Sobota, VeĐký Krtíš, Zlaté Moravce and Žarnovica, are located in the southern part of central Slovakia. The employment situation on the territory of Slovakia was characterized as conjectural unemployment [Gleave – Palmer 1980]. This unemployment is the result of an insufficient demand for manpower on the territory of Slovakia. The difference in unemployment on the NUTS 3 level reaches more than 25 percentage points in that case. The lowest rate of unemployment is in the Bratislava region (from 2.7 to 3.6%) and the highest is in the Rimavská Sobota region (28.6%) and Revúca (28.4%). This marked difference in regional employment appeared between 1997 and 2002. The average rate of unemployment, with the exception of the year 2002, which was influenced by the extraordinary use of public works on the labour market, proves that Slovakia is in a state of economic recession. As was mentioned above concerning GDP, Bratislava has the best position, especially from the point of view of regional differentiation. The other regions are far behind the capital city of Slovakia. GDP created in this region (Bratislava region) was 22 708 million SKK on PPS, two times larger than the average in Slovakia. GDP creation in the Trenþín region is also above the average level. The other regions fall below the median for Slovakia, which is 48.1% of the median for the European Union in 2000 (see Table 6.2). To understand this special situation in Slovakia, there are various ways of explaining those large differences. After 2000, very strong regional differentiation took place. According to our regional economists, the situation in the Slovak Republic is the following: Regions belonging to category A represent developed regions with a multi-sector economic platform. We can distinguish a total of 10 subregions and the agglomeration of Bratislava. These regions are home to 24.89% of the population in Slovakia,


117

which represents a decrease of about 7 points in comparison with 1997 (31.63%). Table 6.2 Regional GDP in PPS per capita in regions – from 1996 to 1999 as a share of the EU average. Source: Regions of Slovakia. VEDA, Publishing House of the Slovak Academy of Sciences, Bratislava 2004, p. 85.

Region Bratislava Trnava Trenþín Nitra Žilina Banská Bystrica Prešov Košice Slovakia

1996 92 51 44 37 38 42 30 44 46

1997 100 50 44 39 39 43 31 44 48

1998 99 50 44 40 40 43 31 47 48

1999 95 52 44 40 40 43 31 47 48

Regions belonging to category B represent economically stabilized regions, with conditions for their future development in supposition of their selective promotion of progressive sectors. In this category we can find 23 subregions, and 25.46% of the population, which represents a decline of about 3 points in comparison with 1997(29.73 %). Regions belonging to category C are stagnant regions with one-sided economic potential, which are not able to absorb potential manpower, and which results in increasing unemployment. In this category we can find 10 subregions and 10.15% of the population. This is a decrease of 6 points in comparison with 1997 (17.25%). Regions belonging to category D represent economically depressed regions, which are influenced by structural crises or which lack sufficiently developed infrastructures. In this category we can distinguish 29 subregions, and 35.23% of the population, which represents an increase of about 10 points in comparison with 1997 (25.66%).

6.4 Survey Study The following survey study intends to be pioneering, in that we wish to explain the different allocation of the high-tech sector industry and the low-tech sector industry, and consequently employment in those sectors with regard to special relations among those two branches (if there is a real relation between employment in high tech sectors in different regions and consequently if this relation concerns services). Industrial regional differ-

118


entiation was not investigated in such a dimension before, and from this point of view we embarked on a new and hypothetical investigation. The database we worked with has several limitations. Most of the data came from the Statistical Office of the Slovak Republic, and some data (on unemployment) came from the National Employment Institute of the Slovak Republic. All of the data was quarterly from the first quarter of 1997 to the third quarter of 2004, and were separate for each region (8 regions)3. The data from the Statistical Office was based on data from companies with 20 or more employees. The main problem with this data is the fact that companies were put into regions based on the location of headquarters, and not based on the location of each production department. This leads to extremely high employment figures in the region of Bratislava, where most of the headquarters are located. The goal of this study was to explain employment in several sectors of the economy by employment in the high-tech sector. To achieve this and due to the available data, we had to use several new variables. The definitions of the high-tech and low-tech sectors are found in the appendix. Both of them are part of industry. 6.4.1 Explanation of Variables The first variables were total employment in all sectors of the economy. The second auxiliary variable: the number of total employment minus high tech employment was introduced. The third variable deals with employment in services, and was calculated as total employment minus employment in industries. The fourth variable is employment in high-tech industry in the neighboring regions (emp_high_tech_neighbour), and employment in high tech industry of the region plus half the employment in high-tech 3

All of the estimations in this work were done by least squares method. This method was chosen due to the status of available data. Data are spatial time series, which consequently cause technical problems with spatial methods. The other factor discriminating against spatial methods is the number of regions, which is only eight, while the time series is 31 units (quarters) long. Also, equation results show very small impact of neighbouring regions by rook matrices. For these reasons, we chose a simple linear, ordinary, low squares model for a time series, with separate variables for each region. Explanatory variables mainly concern employment in a given region and sector, or employment in neighbouring regions. If regions showed similar statistical behaviour, their separate variables were merged into one variable, thus appearing as one region. If a region’s variable is not appearing, it is not statistically significant (p-value bigger than 9 per cent, no variables with p-values between 1 and 9 per cent appeared during estimations).


119

industry in neighboring regions (emp_high_tech_plus_half_neighbors). These variables proved to be explanatory. As can be seen from Figures 6.5, 6.6, 6.7, low-tech and high-tech employment enjoyed rapid development over recent years (from 1997). While high-tech development was different among regions, low-tech employment simultaneously decreased in all regions. For display purposes, only four regions are displayed. Non displayed regions have development similar to thick regions in the figures. 400000 375000 350000 325000 300000 275000

Bratislava (1)

250000

Trnava (2) Trenþín (3) Nitra (4) Žilina (5)

225000 200000 175000 150000

Banská Bystrica (6) Prešov (7) Košice (8)

125000 100000 75000 50000 25000 0

Fig. 6.5 Total employment.

6.4.2 Regional Differentiation The differentiation of variable coefficients appeared in most of the 8 regions in Slovakia. Region No. 1, Bratislava, is widely considered the most developed region and therefore behaves differently from the rest of Slovakia. Another region is region No. 8 - it is Košice. The city of Košice leads the way in some production activities and services. Trnava, region No. 2, is also interesting. It is the only region bordering with Bratislava, and therefore is sometimes considered its suburb. Some of the statistical results of Trnava are somewhat surprising. Other regions are Trenþín (3), Nitra (4), Žilina (5), Banská Bystrica (6) and Prešov (7). The behavior of these regions cannot be described as easily as the others.

120


45000 42500 40000 37500 35000 32500 30000 27500 25000 22500 20000 17500 15000 12500

Bratislava (1) Trnava (2) Trenþín (3) Nitra (4) Žilina (5) Banská Bystrica (6) Prešov (7) Košice (8)

10000 7500 5000 2500 0

Fig. 6.6 High-tech employment.

65000 60000 55000 50000 45000 40000 35000 30000 25000 20000 15000 10000 5000 0

Fig. 6.7 Low-tech employment.

Bratislava (1) Trnava (2) Trenþín (3) Nitra (4) Žilina (5) Banská Bystrica (6) Prešov (7) Košice (8)


121

6.4.3 Results Due to this regional differentiation, some coefficients of the model differ between regions. Overall, Bratislava (1) has very positive coefficients; Kosice (8) also has positive coefficients, and Trnava (2) has both positive and negative coefficients. Other regions differ from time to time. In the analysis below, the number following variable names represents which region the value is applicable (see Table 6.3) . Table 6.3 explains total employment minus employment in high-tech industries by employment in high-tech industry

Model Formula: emp_minus_high_tech ~ emp_high_tech_region1 emp_high_tech_plus_pol_neighbors2345678 Estimate

+

emp_high_tech_region8

Std. Error 6865.01 0.3653 0.398645 0.133937

t value

+

Pr(>|t|)

(Intercept) 63224.3 9.209644 0 emp_high_tech_region1 13.109379 35.886604 0 emp_high_tech_region8 3.509512 8.8036 0 emp_high_tech_plus_half_neigh1.116165 8.333504 0 bors2345678 Residual standard error: 20613 on 244 degrees of freedom, Number of observations: 248 Root MSE: 20613 Multiple R-Squared: 0.93 Adjusted R-Squared: 0.93

Table 6.3 explains total employment minus employment in high-tech industries by employment in high-tech industry. As can be seen, the rise of employment in high-tech in region 1 leads to an increase in total employment by 13 people and in region 8 by 4.6 people. In other regions, the coefficient is 1.1. Also if high-tech employment in a neighbouring region rises by 2 people, employment in that region rises by 1.1 people. By this simple equation, 93% of the total employment is accounted for. Table 6.4 and Figure 6.8 explain employment in the low-tech sector by employment in the high-tech sector. Overall, this equation caused many problems and there was no possibility of grouping regions together. The direct impact of high-tech employment on low-tech employment was 0.8 in Bratislava (1), 0.9 in Košice (8), 0.5 in Trenþín (3), 0.2 in Žilina (5). In Trnava (2) the coefficient is -0.5. This negative value is influenced by the special status of Trnava as the only neighbouring region of Bratislava (1). The impact in Nitra (4), Banská Bystrica (6) and Prešov (7) proved to be zero statistically. Also the rise in high-tech employment in neighbouring regions by 1 results in the rise of low-tech employment in the region by 0.3. This equation accounts for 90% of low-tech employment (see Table 6.4).

122


Table 6.4 Explains employment in the low-tech sector by employment in the high-tech sector.

Model Formula: emp_low_tech ~ emp_high_tech_region1 + emp_high_tech_region2 + emp_high_tech_region3 + emp_high_tech_region5 + emp_high_tech_region8 + emp_high_tech_neighbors2345678 Estimate Std. Error t value Pr(>|t|) (Intercept) 13177. 689.235 19.118385 0*** emp_high_tech_region1 0.805041 0.038561 20.877097 0*** emp_high_tech_region2 -0.494412 0.034277 -14.423824 0*** emp_high_tech_region3 0.507725 0.019117 26.559119 0*** emp_high_tech_region5 0.169992 0.023905 7.111108 0*** emp_high_tech_region8 0.918574 0.049554 18.536857 0*** emp_high_tech_neighbors2345678 0.267512 0.009953 26.87702 0*** Residual standard error: 2505 on 241 degrees of freedom, Number of observations: 248 Root MSE: 2505.93441 Multiple R-Squared: 0.90 Adjusted R-Squared: 0.90

Table 6.5 explains total employment by high tech employment. As can be seen, if high-tech employment increases by 1 person, total employment increases by 11 people in Bratislava (1). In Košice (8) it is 2.3 and in Banská Bystrica (6) it is 4.3. In Trenþín (3) and Žilina (5) it is 0.9. Other regions did not prove to be statistically different from zero. Employment in neighbouring regions did not prove to be non-zero. Overall, this equation accounts for 96% of total employment. Table 6.6 explains employment in services. The situation is similar in Bratislava, where the increase of high-tech employment by 1 person leads to service employment increases of 10 people. In all other regions the situation is different. The increase of employment in high-tech brings a decrease of service employment by 2 people. However this leads to an increase of 0.4 in neighbouring regions. This situation is partially influenced by the fact that the data only deals with enterprises of 20 or more employees; services are mainly small enterprises and do not appear in this data. Apart from this drawback, this equation accounts for 95% of service employment changes. As was shown, all equations explain more than 90% of employment in the sectors. A quick summary of the model is in the last table (see Table 6.7). Figure 6.8 represents location of output from Table 6.4.


123

Fig. 6.8 Influence of high tech sector employment on low tech sector employment. Table 6.5 Explains total employment by high tech employment.

Model Formula: emp_total ~ emp_high_tech_region1 + emp_high_tech_region8 + emp_high_tech_35 + emp_high_tech_region6 Estimate Std. Error t value Pr(>|t|) (Intercept) 121660.00 1491.516 81.56804 0*** Emp_high_tech_region1 11.374179 0.142737 79.686269 0*** Emp_high_tech_region8 2.292354 0.241426 9.495068 0*** Emp_high_tech_35 0.876082 0.096623 9.067054 0*** Emp_high_tech_region6 4.31093 0.233832 18.435988 0*** Residual standard error: 14770.089507 on 243 degrees of freedom, Number of observations: 248 Root MSE: 14770 Multiple R-Squared: 0.96 Adjusted R-Squared: 0.96

124


Figure 6.9 represents location of output from Table 6.5.

Fig. 6.9 Influence of high tech sector employment on total tech sector employment.

Table 6.6 explains employment in services.

Model Formula: emp_services ~ emp_high_tech_region1 emp_high_tech_neighbors2345678 Estimate

+

emp_high_tech_2345678

Std. Error

t

+

value

(Intercept) 104314 4509.9879 23.1297 Pr(>|t|) Emp_high_tech_region1 9.76424 0.255764 38.177 0*** Emp_high_tech_2345678 -2.051 0.209975 -9.77 0*** Emp_high_tech_neighbors2345678 0.377736 0.056464 6.69 0*** Residual standard error: 17131.449333 on 244 degrees of freedom, Number of observations: 248, Root MSE: 17131 Multiple R-Squared: 0.95 Adjusted R-Squared: 0.95

Overall, this model simply describes employment in Slovakia by employment in the high-tech sector. This study has proved that the high-tech sector needs to be considered when modelling employment development in various sectors. Figure 6.11 shows the differentiation of employment by sector and region in all regions (including the Bratislava region). With the addition of Bratislava, the graph changes rapidly by high employment in services in Bratislava, which is caused by administrative factors.


125

Table 6.7 quick summary of the final model. Summary emp_minus_high_tech emp_low_tech Emp_total emp_services

N 248 248 248 248

DF 244 241 243 244

MSE 424923365 6279707 218155544 293486556

RMSE 20613.67 2505.93 14770.09 17131.45

R2 .929397 .904124 .964769 .950561

Adj R2 .928529 .901737 .964189 .949954

Figure 6.10 shows employment differentiation in four regions (Bratislava (1), Trnava (2), Prešov (7) and Košice (8)). As can be seen, the largest differences are in employment in services. Low-tech employment is almost the same, while high-tech employment slightly differs.

High-tech 1,5 Bratislava(1)

1 Trnava(2)

0,5

Prešov(7) Košice(8)

0

Services

Low-tech

Fig. 6.10 Employment differentiation by sectors and regions.

6.4.4 Consequences Resulting from the Survey One of the very important factors influencing the situation in employment is migration. When speaking about migration, 928 000 people commuted to work in 1991.The growth of population in large centres also represented a new tendency in finding new jobs. Only 13% of economically active people commuted from 77 centres. On the other hand, 60% of economically active people commuted from rural centres in 1991. In some of these regions the number of commuters was more than 70%. The migration to the Czech Republic was very high. The opposite was true for the eastern part of Slovakia (2 000 people yearly).


126

The adverse balance of commuters in the central part of Slovakia was approximately half of that balance for the eastern part of Slovakia and twice as much as that of the western part of Slovakia [FalĢan – Pašiak 2004]. Furthermore, migration mostly influenced the employment situation in Bratislava’s neighbouring regions. The following graph illustrates this statement, especially for Trnava. We see also an important connection between the high tech sector, the low-tech sector and services. In one of the studies (Zajac explicitly concentrates on this problem), the shift to services and sectors of sophisticated technologies in manufacturing, is obvious from different employment trends in different sectors. For highly developed economies, it is observable that employment in services is growing, especially in different segments: social, financial, insurance and entrepreneurial services. Jobs gradually increased in sectors of high technologies, especially in the 1980s, but their gradual movement has a very cyclical character. A certain revival in the second half of the 90s cannot hide the reality that these jobs form only a very small part of a certain unity, and that it is not possible to expect this substantial contribution to the growth of total employment. These jobs had a negative influence on total employment in the entrepreneurial sector from 1980 to 1995. On the other hand, they substantially promoted employment growth in social and private services [Zajac 2004].

7,6 2,1

2,1

1,2

-1,3 -3,4

B. By st ric a

-5,8

N itr a

-4,6

Tr na va

Br at is la va

10 8 6 4 2 0 -2 -4 -6 -8

Growth rate

Fig. 6.11 Growth Rate of Employment in Manufacturing in %.


127

Employment in different industrial enterprises was influenced by a decrease in the number of employees. There is a very insufficient structure for unemployed people, especially in terms of education and length of unemployment. The unemployment rate also influences regional differentiation. The differences in these regions are possible to solve, especially through infrastructure construction in given regions and through the strengthening of local institutions.

6.5 Conclusion The Slovak Republic, as a constituent part of the European Union, must participate in fulfilling of the Lisbon strategy. One of the most wishful priorities of this strategy is the strengthening of the competitiveness of the European economy by 2010, by which time we expect transformation to be transformed into an information society. New modern technologies will play an important role in this development. According to our survey, most significant was the impact of high-tech in Bratislava, where the highest point capital, qualified labour force, perspective industrial branches and constructed infrastructure (especially traffic connections with EU industrial centres) are located. Domestic capital is concentrated in development of medium-high-tech and mediumlow-tech, especially in accord with its investment possibilities. Foreign investors also concentrate on this field of interest, because high-tech units are located in their mother countries and they do not want to invest this capital as risk capital. It is also an explanation concerning why in Bratislava and other cities, there is not a massive application of high-tech industrial branches. Changes in the field of medium-high-tech employment and mediumlow-tech employment, influence low-tech employment in the Slovak Republic. This phenomenon is linked with gradual employment upgrading in the field of services, which especially absorbs the labour forces dissolved from manufacturing industry. In reality, services cannot possibly grow as fast as changes in the structure of employment in accord with conditions concerning the labour force in manufacturing. The strongest tendencies here are manifested in the Bratislava region, but in other regions the correlation is not so evident. In spite of the fact that high-tech did not have a very strong impact on total employment, it indisputably represents one of the most decisive issues of competitive growth. This growth must be the most important factor shaping the form of future industrial policy. It is important to create the appropriate realities, so that

128


all of the above-mentioned factors (infrastructure, qualified labour force, capital) are potentially developed in all regions in the Slovak Republic, and potentially contribute to the equalization of all regional differences Our survey has shown large regional disparities in the SR, and their very slow equalisation process in all regions in recent years. In spite of very slow, but observable shifts in development of medium-low-tech and mediumhigh-tech industries, there is no observable significant shift in employment in these industries and services. From this point of view the industrial policy in the SR cannot assume that concentration on development of successful regions automatically means development in less advanced regions (in accordance with the thesis that successful regions will gradually upgrade the development in less developed regions). In reality, successful regions as for instance, Bratislava, have not accelerated the economic growth in less developed regions in Slovakia. The causes of this not very positive trend are also influenced by other very important factors like technical and traffic infrastructure, and regional structure of the qualified workforce etc.

6.6 Methodological Notes High-tech manufacturing x NACE 30 manufacturers of office machinery and computers x 32 manufacturers of radio, television and communication equipment and apparatus x 33 manufacturers of medical precision and optical instruments, watches and clocks Medium-high-tech manufacturing x 24 manufacturers of chemicals and chemical products x 29 manufacturers of machinery and equipment n.e.c. x 31 manufacturers of electrical machinery and apparatus n.e.c. x 34 manufacturers of motor vehicles, trailers and semi-trailers x 35 manufacturers of other transport equipment Medium-low-tech manufacturing 23 manufacturers of coke, refined petroleum products and nuclear fuel 25 manufacturers of rubber and plastic products 26 manufacturers of other non-metallic mineral products 27 manufacturers of basic metals 28 manufacturers of fabricated metal products 35.1. shipbuilding

x x x x x x


129

Low-tech manufacturing x 15 manufacturers of food products and beverages x 16 manufacturers of tobacco products x 17 manufacturers of textiles x 18 manufacturers of wearing apparel x 19 for tanning and dressing of leather x 20 manufacturers of wood and of products of wood x 21 manufacturers of pulp, paper and paper products x 22 for publishing, printing and reproduction x 36 manufacturers of furniture; manufacturing n.e.c.

References Baláž, V. (2003): Knowledge Intensive Business Services in a Transition Economy. Journal of Economics, Vol. 51, 2003, No. 4, pp. 475 – 488. Brzica, D. (2003): Budovanie väzieb medzi veĐkými a malými podnikmi v podmienkach regionálnej heterogenity: príklad SR. [Working Paper No. 18] Institute for Forecasting of SAS, Bratislava. Commission of the European Communities (2004): European Competitiveness Report 2004. Commission staff working document. European Communities, Luxemburg. European Communities (2002): Science and Technology in Europe. Statistical pocketbook, Data 1991 – 2001. Office for Official Publications of the European Communities, Luxemburg. EUROSTAT (2004): Statistics in Focus. Science and Technology. Theme 9 – 2/2004. EUROSTAT (2003): Share of Employment in Knowledge-intensive Services in the Acceding Countries still below EU Average. Eurostat news release, STAT/03/127. FalĢan, Lubomír - Pašiak, Ján (eds.)(2004): Regionálny rozvoj Slovenska. Východiská a súþasný stav, Interlingua, Bratislava. Geyer, Anton at all (2003): The Future of Manufacturing in Europe 2015 – 2020: The Challenge for Sustainability. Scenario Report. European Commission Joint Research Centre, Brussels, 2003. Gleave, D. – Palmer, D. (1980): Spatial Variations in Unemployment Problems: a Typology. Papers of the Regional Science Association, 44, p. 57 – 71. Hatzichronoglou, Thomas (1997): Revision of the High-Technology Sector and Product Classification. STI Working Papers 1997/2. OECD, Paris. IN the Spotlight, High-tech: A Product, a Process or Both? In: IN Context, Vol. 1, 2000, pp. 1 – 4. Kask, Christopher - Sieber, Edward (2002): Productivity Growth in ‘High-tech’ Manufacturing Industries. In: Monthly Labor Review, Vol. 125 (2002), No. 3, pp. 16 – 31.

130


Luker, William Jr. – Lyons, Donald (1997): Employment shifts in hightechnology industries,” Monthly Labor Review, Vol. 120 (1997), No 6 pp. 12–25. Marshall, A. (1920): Principles in Economics. The Macmillan Press LTD, London. Ministry of Economy of SR (2000): Elaboration of the Principles of the European Union´s Industrial Policy for Conditions of the Slovak Republic (Strategy for Slovak industry development for the 21st century. Ministry of Economy of SR, Bratislava. OECD (1997): OECD Science and Technology Indicators, No 2: R&D, Innovation and Competitiveness. OECD, Paris, 1997, pp. 58-61. Outrata, R. (2002) Konkurenþná schopnosĢ ekonomiky, in: Vízia vývoja Slovenskej republiky do roku 2020. Institute of Forecasting, SAS, Bratislava. Porter, M. E. (1990): The Competitive Advantage of Nations. The Macmillan Press LTD, London. Rausch, L M. (1998): High-tech industries drive global economic activity. Issue Brief. Washington, DC, Division of Science Resources Studies, National Science Foundation. Slovak Chamber of Trade and Commerce (2005): Základné makroekonomikcé proporcie vývoja ekonomiky Slovenska v roku 2005 z pohĎadu podnikov. Bratislava. Smith, Keith (2000): What is the “Knowledge Economy”? Knowledge intensive industries and distributed knowledge bases.1 STEP Group Oslo. Statistical Office of the SR (2003): Statistical Report on Basic Developmental Tendencies in the SR 2002, Statistical Office of the SR Bratislava. Statistical Office of the SR (1999): Yearbook of Industry of the SR 1999, Statistical Office of the SR, Bratislava. Statistical Office of the SR (2000): Yearbook of Industry of the SR 2000, Statistical Office of the SR, Bratislava. Statistical Office of the SR (2001): Yearbook of Industry of the SR 2001, Statistical Office of the SR, Bratislava. Statistical Office of the SR (2002): Yearbook of Industry of the SR 2002, Statistical Office of the SR, Bratislava. Statistical Office of the SR (2003): Yearbook of Industry of the SR 2003, Statistical Office of the SR, Bratislava 2003. Statistical Office of the SR (2004): Yearbook of Industry of the SR 2004, Statistical Office of the SR, Bratislava. VEDA (2004): Regions of Slovakia. VEDA, Publishing House of the Slovak Academy of Sciences, Bratislava. Zajac, Štefan (2004): Some Problems of Employment Effect of Innovation, Journal of Economics, Vol. 52 (2004), No. 1 p.74 – 90.

7 Automobile Sector in the Slovak Republic: Current Situation and Future Prospects

Daneš Brzica Slovak Academy of Sciences, Slovak Republic, E-mail: daneš[email protected]

Abstract. The European Union pays attention to the problem of persisting technological gap vis-à-vis USA and Japan. The Slovak Republic, similar to other EU members, shows lower ability to transform knowledge from basic research into new products and this has had negative impact on employment and growth. Emergence of clusters can assist to qualitative changes in functioning of the Slovak business sector, which has been for many years characterized by large-scale production. This situation changed during the 1990s. A transition process, generated by the Velvet revolution in 1989 and the liberalization of Slovak markets, initiated a restructuring process among Slovak enterprises coupled with an emergence of small domestic firms and foreign direct investment (FDI). The chapter reflects this process and focuses on the main element of it – automobile sector. It addresses emergence of Slovak automobile cluster(s) as a main topic. Keywords: Automobile sector, Industrial cluster, Competitiveness, the Slovak Republic

7.1 Introduction The European Union (EU) deals with the problem of persisting technological gap vis-à-vis its rivals - USA and Japan. The Slovak Republic also shows lower ability to transform knowledge from basic research into new products and this fact has negative impact on both employment and growth. An emergence of industrial and knowledge clusters1 can assist to 1

Besides numerous papers on various aspects of clusters (e.g., OECD, 2000, Cooke, 1999), there are interesting papers on cooperation in automobile sector

132

Daneš Brzica

fill the gap through some qualitative changes in functioning of the Slovak business sector, which has been for many years characterized by largescale and rather low value-added production. A transition process, fostered by the revolution in 1989, and liberalization of Slovak markets, has initiated restructuring process in Slovak large enterprises, increase in number of small and medium-sized enterprises (SMEs) as well as inflows of foreign direct investment (FDI). This chapter focuses on one of the main areas of Slovak industrial production – automobile sector. It addresses dynamic development of automobile sector as well as possible emergence of Slovak automobile cluster(s) as the main topic. We discuss the increased potential for development of automobile cluster(s) expected from recent Slovakia’s membership in the EU and radical policy reforms. As a result, higher inflow of foreign companies and more dynamic domestic industry, due to a favorable environment for foreign investors and domestic firms, give an opportunity to emergence of automobile sector’s clustering effects. The chapter provides, after this introductory part, in the part 1 an overview of general situation in this sector in Slovakia. In the second part, preconditions for cluster emergence are mentioned. The description of automobile/automobile parts and components sector development of the sector towards clusters is provided in the part 3. Part 4 presents a simple model describing the relation between production of supplies to Volkswagen from domestic suppliers and production of VW cars. Finally, concluding part provides some general remarks to the topic and policy recommendations. Our hypothesis in the model is that there exists a direct relation between the amount of car produced in the country and the volume of production of domestic car suppliers. The results obtained from the estimation show that more car produced by VW means also more supplies from domestic firms to the automobile producer (more than 30 percent of the whole car components come to Volkswagen from domestic suppliers).

(see, e.g., Dicken, 1992). Feldman, for example, provides information on knowledge complementarity (Feldman, 1994). For research on strategic technology partnering see, e.g., Hagedoorn (1993), for technological cooperation, networks and evolution of technology organizations see, e.g., Håkansson (1989) and Hales (2001). Forming regional innovative areas is mentioned in Hansen (1992). The broader situation regarding cooperation between small and medium-sized enterprises and large firms in the Slovak Republic can be found in Nemcova and Gasparikova (2004).

7 Automobile Sector in the Slovak Republic

133

To summarize, transformation has generated in Slovakia conditions for modernization of corporate sector. However, the process of forming modern cooperation systems (like e.g., clusters and networks) has not reached yet the intensity of firm cooperation required, which would have lead to creation of sufficient local/regional competitiveness. Nowadays, as Slovakia is becoming more and more integrated in the Union’s structures, firms based in the country only try to form more advanced cooperative links (an embryonic forms of clusters) in order to cope with global competitive pressures. As the European countries are increasingly competing in offering attractive environment for firms, e.g., with tax systems, there is also an urgent need to follow wise economic policy.

7.2 Development of Automobile Production Capacities The importance of the automobile sector in the Central and Eastern European countries (CEECs) seems to be increasing. According to K. Gerhardt (2004), this sector is able to improve the countries´ balance of trade, but it will also serve as a vehicle for technological development in other segments. New manufacturing technologies and systems and the training of skilled workers and managers will have a positive impact. Other investments in CEECs are likely to bring new dynamism to the sector. Table 7.1: Automobile producers based in countries neighboring with Slovakia. Source: author, based on data from CTK, TASR and SITA press agencies, 2004a,b

Producer Skoda (VW) TPCA(a) VW(b) Daewoo-FSO (GM) Opel (GM) Fiat Fiat Audi (VW) Suzuki-Fiat

Country Czech Republic Czech Republic Poland Poland Poland Poland Poland Hungary Hungary

Location Mlada Boleslav Kolin Poznan Warsaw Gliwice Bielsko-Biala Tychy Györ Esztergom

a) The decision of a French-Japanese consortium consisting of Toyota and PSA Peugeot Citroën (TPCA) to build a factory in the Czech Republic for producing a compact car surely fosters already important Czech car industry. In total, € 1.5 bn will be invested for the production of a projected 300,000 units. b) VW has announced plans for producing a new delivery vehicle in Poland; plant capacity will reach 150,000 by 2005 (Gerhardt, 2004).

134

Daneš Brzica

As shown in the Table 7.1, the Slovak automobile sector has numerous competitors in the countries of Central Europe. In addition to this, there are huge automobile clusters in Germany and Austria. Their regional proximity implies that they also represent a competitive threat to Slovak local producers. However, some complementing nature of such production and joint ownership in some cases (e.g., in the case of Volkswagen) coupled with temporal competitive advantage of lower wages still give some advantages to Slovak location. Unlike more general Central European picture presented in Table 7.1, Table 7.2 shows present and planned capacities expected to be in operation before 2006 in Slovakia. As predicted by theory, big transnational corporations have gradually been attracting more and more component suppliers helping thus create clusters, which seems to be the case also here. Table 7.2: Automobile producers based in Slovakia. Source: author, based on data from TASR and SITA press agencies, 2004a

Producer

Size of Invest- Expected jobs ment (in mln €)

VW PSA Peugeot Citroen(a) Kia Ford

1000 700 700 400

Location

Bratislava, 10 000 Martin 3 000 Trnava 3 000 Zilina 3 000 Kechnec

a) PSA is building a € 700 mln car plant in Slovakia, in which some 300,000 cars are to be manufactured starting in 2006 (Gerhardt, 2004).

In the Table 7.3, an estimate of the impact of automobile producers´ investment on the growth of Slovak economy is provided. Except for Kia´s 2006 expected value, all other periods´ impact of investments is positive. Table 7.3: Impact of automobile producers´ investment on growth of Slovak economy (in %). Source: Zdruzenie automobiloveho priemyslu SR, 2004.

Producer/Year Volkswagen PSA Kia Ford Total

2003 0.249 0.033 0.000 0.000 0.282

2004 0.000 0.277 0.071 0.000 0.348

2005 0.000 0.000 0.214 0.039 0.253

2006 0.180 0.450 -0.005 0.070 0.695

2007 0.021 0.100 0.123 0.119 0.363

2008 0.000 0.100 0.159 0.121 0.380

2009 0.000 0.000 0.000 0.010 0.010

Slovakia has been continuing in its focus on development of automobile sector. Using investment incentives for foreign investors and radical tax and other economic reforms have helped to skyrocket FDI inflows mostly aimed at automobile sector (automobile industry plus automobile-related


135

production and service). Table 7.4 shows statistics covering recent production and employment development in this sector compared to general industrial production and employment. Table 7.4: Position of automobile industry in SR compared to general industry. Source: Zdruzenie automobiloveho priemyslu SR, 2004a

Indicator/Year Industrial production of the SR (revenues in bn Sk) Of which: Aut. ind. – producers

1999

2000

2001

2002

2003

685.4 105.2

785.0 130.2

888.5 150.6

947.0 185.0

1095.7 278.8

Of which: VW

69.0

85.4

88.8

109.6

n.a.

Aut. ind. – sales and services

26.4

26.0

34.5

34.1

36.0

500.4 35.6

485.8 35.9

494.1 44.4

497.2 50.2

501.5 n.a.

Number of employees in industry (000 persons) Of which: Aut. ind. – producers Of which: VW Aut. ind. - sales and services

6.5

7.1

7.5

9.1

n.a.

10.0

10.0

10.0

10.0

10.5

a) n.a. = data are not available

As presented in the Table 7.4, automobile production has more than doubled between 1999 and 2003. Contracts, announced estimated capacities and investment activity in production facilities and infrastructures indicate that in the coming years the automobile production will increase making thus Slovakia one of the main automobile producers in Europe measured by production per capita. It is estimated that by 2006 60,000 jobs will exist in automobile industry. Its share on total industry is estimated to be 30 percent in 2006 (compared to an estimated 26 percent in 2004).

7.3 Preconditions for Cluster Emergence Due to its common past linked to Czechoslovakia, known for its automobile sector, Slovakia has become soon after 1989 an important car producer. The main firm, VW Slovakia, has become since then a huge automobile producer in the country. German automobile producers already enjoy a strong position not only here but also in other CEECs. (Gerhardt, 2004). The parts suppliers are sure to expand their presence in these coun-

136

Daneš Brzica

tries.2 Manufacturing in the CEECs is attractive as they can serve as a base for supplying other candidates for EU succession (Croatia, Bulgaria, and Romania) as well as other countries (Russia, Belarus, and Ukraine). These countries represent major potential for growth, albeit initially limited to minis and compacts. Furthermore, a number of countries in Southeastern and Eastern Europe already feature in the strategic plans of carmakers and parts producers as possible production locations (Gerhardt, 2004). For an emergence of clusters, initial cooperation corporate linkages have to be further intensified and supplemented by various public institutions (e.g., universities or research institutes). From the point of view of potential cluster formation important for the Slovak economy are the present automobile-related FDI inflows.3 One issue is clear today and it is the required complex/developed basis of knowledge concentration and social capital4, which only can guarantee further cluster development together with potential increase in size and scope of high value-added activities (services and components) being supplied to the European automobile sector. Higher investments into research and education can help follow this trend in the future. However, to what extent clustering in the next future can assist the Slovak economy in the development of other related/unrelated industries and businesses with even higher scientific research content and more substantial profit generation remains the question. Many observations show that if there is no harmonization in cooperation among individual SMEs, then decline in competitiveness occurs. This seems to be true also for emerging automobile cluster(s) of SMEs. In such harmonization, firms can have problems with ICT-related costs and other costs arising with a SME´s integration into a cluster. However, clusters can enforce their interests, cooperate with universities, obtain domestic subsidies and assistance from EU and individual firms can spin-off unneeded activities and thus increase their specialization.5 Sharing knowledge based 2

In spring 2003, an IKB survey of some 200 of these companies revealed that CEECs would be the main target of their foreign investment plans in coming years (Gerhardt, 2004). 3 The initial structure of the economy as well as the degree and forms of FDI inflows are important for both the adjustment process related to the emerging clusters and future country’s focus on higher value added industrial production. 4 Social capital can be understood simply as a network of social relations characterized by norms of trust and reciprocity. The core of social capital is the quality of social relations (see Bullen, 1999). Social capital represents degree of social cohesion, which exists in communities (De Rossi, 2004). 5 By this an origin of new spin-off firms have been supported as well as cooperation with other large scale enterprises or clusters within the framework of domestic as well as cross-border cooperation.


137

on rapid exchange of information via internet for communication between firms and cost reduction for the communication represents cost savings in the whole cluster. Building more complex relations with automobile producers or systems suppliers (see in the scheme 7.1 a structure of them), who assist firms in the cluster, improve product, and gain better quality information and services, helps to form better possibilities to compete on the market. Cooperating SMEs in automobile cluster can make bigger investments; have better image; communicate more intensively and jointly conduct development, marketing or education. Improvement of accessibility and use of production sources and from this resulting production effects or joint presentation are other possible advantageous. Among barriers for forming automobile clusters in Slovakia are especially general legal and financial conditions for automobile productionoriented SMEs, old fashion view of industrial policy, lack of businesspersons’ confidence in institutions or absence of informal networks among businesspersons. Weak points of clusters can be low level of information about possibilities of corporate clustering; problem of management of huge number of firms; issue of selection firms suitable for cluster activities; geographical allocation and transport costs; low level of experience with this new organizational form; mutual distrust between firms as well as a distrust of firms towards a cluster. In the following part, we present a hypothesis (tested here on the case of Slovak economy) that as automobile production in a country increases so are increasing the production volumes of domestic parts/components suppliers.

7.4 Automobile/Automobile Parts and Components Sector and Emergence of Automobile Clusters Emergence of automobile cluster(s) is in the case of Slovakia related to the development of passenger automobile production.6 Production of these cars dynamically increases. Whereas in 1995 19 688 passenger cars were produced in Slovakia, in 2000 it was 180 803 and in 2003 already 281 160. From 2007 onwards, the three main producers (Kia, PSA and VW) together plan to produce 800 000 cars a year. This would make Slovakia the country with the highest car production per capita in the world. The main

6

Compared to other EU countries, the Slovak Republic offers relatively cheap and qualified labor force and its firms attract clients due to the quality of assembled products, but also the quality of components supplied.

138

Daneš Brzica

activity is formed by production of passenger cars and production of components. Because production of automobiles requires timely and continuous supplies of numerous parts and components, the prospective subindustries and firms, which should have become able and ready, from qualitative perspective, for supplies for automobile industry had been identified already in the past by the Ministry of Economy of the Slovak Republic. It concerned numerous sub-industries,7 where domestic suppliers had a chance to be successful if they will invest into modernization8 oriented on production of particular supplies for selected types/categories of vehicles. This initiative has no real consequence for the development of automobile industry in Slovakia. However, it provided signals that changes in automobile sector emerge. For example, supplying firms in this process were advised to take into account expected demand from automobile industry, where a gradual change from part suppliers to systemic suppliers has taken place. Systems supplier9 to an automobile firm is a financially strong partner with its own research and development capacities. Therefore, it is able to cooperate in joint production of a car. Such suppliers have developed diversified network of own suppliers and supply whole subsets ready for final assemble. The scheme 7.1 shows various categories of subjects in a cluster (or in a supplier system with a dominant firm). It is expected that in Slovakia gradually evolve all forms of suppliers mentioned. Policy priority in long-term perspective is to have more suppliers with high value-added production.

7

It is for example, pressing of parts from metal plates, production of tires and parts from rubber, production of pressure casts from aluminum, production of goods, pressing of plastics, production of electric accessories, production of tools or production of soundproof materials. The system of zero-stored (just-in-time) supplies requires very precise coordination, logistics and production. Logistics guarantees supplies of more than one hundred parts from various European countries, their efficient warehousing and storage (World Bank, no date). 8 Investments have to be oriented on purchase of high technologies and on increase of capacities, because requirements of firms in automobile industry for supply volumes exceed capacity potentials of most Slovak firms. 9 An example is the firm SAS Automotive established in Bratislava in 2000. It provides firm VW Slovakia with completely assembled cockpits. Modules consist of panels, electronic components, air condition, airbags, gear levers and pedals. Module must be assembled without defects and supplied directly on the production line of a particular car within two hours from the reception of an order.


139

Scheme 7.1: Potential cluster structure. Source: Author, based on Rentmeister (1999)a

Automobile producer (PRODUCER)

MAIN SUPPLIER Subcontractor to main supplier

Direct supplier to producer Direct supplier to producer Direct supplier to producer Subcontractor to main supplier

Systems supplier (e.g. door modules, front end) Small engineering service firms (specialized services, e.g., design, calculation, rapid prototyping) Software service firms (e.g., CAD systems) Consultants (e.g., project management, simultaneous engineering, TQM) Integrated engineering service firms (Broad range of services) Parts and components suppliers (e.g., plastic parts, sheet metal parts)

a) Specification of cooperation forms is just illustrative. In reality, there are different modes of cooperation between producers and their suppliers. In some cases there are several main suppliers, in others only one.

The Slovak firms’ competitiveness in this sector is based on its production structure, its high quality and increase in the number of strategic investors. However, demand changes on local and global market can substantially affect the volume of supplies both cars and components. The volume of component supplies has been gradually increasing, as shown in the Table 7.5. Table 7.5: Development in component supplies for automobile industry (in bn Sk). Source: Zdruzenie automobiloveho priemyslu SR, 2004.

Indicator/Year Total supply of components Of which for: ŠKODA AUTO, a.s. VW Other producers

1997 17.2

1998 26.1

1999 31.2

2000 35.9

2001 41.7

2002 53.3

2003 104.8

4.4

5.0

3.9

3.2

2.9

2.5

2.4

7.3 5.5

11.3 9.8

14.5 12.8

17.5 15.2

19.9 18.9

28.0 22.8

64.7 37.7

The economies of time, scope and scale probably will lead to the situation that suppliers will produce not only for firms located in Slovakia but also for firms in abroad. Their success will depend on competitiveness, be-

140

Daneš Brzica

cause, as mentioned above, important automobile clusters are located in many European countries (including close neighbors - the Czech Republic and Austria). So far, the results reached on cluster formation have been mixed.10 On the one hand, rapid structural adjustments have been made, but on the other information network, social capital or transport infrastructure have not been developed enough to be able to attract FDI of the second type (i.e., highly innovative firms) capable to form innovative clusters. The distribution of the Slovak production facilities shown in Table 7.6 and picture 1 indicates in what areas/regions potential clusters can emerge. Many investors operate in Povazi (North-west part of Slovakia). For example, in Trencin there is located a firm, Leoni Autokabel Slowakia, which should start production of cables for BMW carmaker in Ilava in September 2004. Numerous other smaller investors have already indicated their interest to follow the main automobile investors. However, as shown in the picture 1, the automobile part suppliers are located in several parts of the country and there are only three main automobile production locations, which may represent up to three potential clusters. A look into history shows that, initially, each supplier had supplied most of the components or parts only to one main producer (VW Slovakia). The company often used also its own foreign suppliers. Nowadays, however, delivering of goods and services to more producers located in Slovakia or neighboring countries by Slovak-based firms means higher economies of time, scale and scope. Supplies from home-located car part suppliers come to several automobile producers. As is seen from Table 7.7, such automobile producers are Skoda (the Czech Republic), VW (Slovakia), Audi, Ford, Seat, Tatra (the Czech Republic), Karosa (the Czech Republic), Volvo, Daewoo, Dacia, Porsche, Mercedes, Auto VAZ, GM, Citroën, Renault and Fiat. With more and more producers being located in Slovakia, also more supplying firms start to operate here. Such process helps to form elementary automobile cluster structures, which can further expand and develop into complex automobile cluster(s) by involvement of various public institutions and infrastructure. Automobile clusters cannot be created by state intervention. However, public sector can stimulate such initiatives and assist in fostering private activities by providing automobile firms with financial and organizational support. General Slovak situation is not good in this respect, because 10 Until now, there have not been clusters developed in Slovakia, but such possibility exists as more firms penetrate the Slovak automobile production market and cooperate.


141

SMEs lack financial and social sources, cooperate less and specialize only to a limited extent. Therefore, a cluster support policy should have been implemented, in which shared responsibility of various (public and private) subjects have been guaranteed. There should be an increasing awareness of the business and government subjects about the concept of clusters and networks. To form framework conditions suitable for cluster development is critical and possible, but it requires also new forms of state assistance. However, firms remain the key components in this process and main inter-mediator for integration of local clusters into international clusters. Cooperation among ministries responsible for measures oriented for cluster support is critical and should be coordinated. However, Slovakia so far has neither specific policy nor structure for support in forming clusters whether in automobile sector or in general. Regarding the EU membership and increasing competition, the Slovak automobile firms should form clusters or more intensively participate in the existing important ones. Without changes in their existing passive approach to cooperation, their position can worsen. They can finally become only suppliers of simple, low value added parts. In abroad, automobile and other firms operate in specialized clusters of SMEs, which are supported by governmental and regional programs. Governmental agencies of the neighboring countries, besides providing general conditions of export competitiveness, selectively support also individual firms, especially in the area of information and communication technologies (ICTs). Their policies are based on the philosophy that, in the long-term perspective, only dynamic firms survive and that the only possibility for SMEs is their participation in networks and clusters. Therefore, from policy perspective, it is necessary to analyze also existing precondition for a complex development of automobile and other clusters and basic policy and infrastructural frameworks needed for its successful development within the EU context. The following part 7.4 discusses a model showing relation between automobile and parts/components production within the Slovak automobile sector.

142

Daneš Brzica

Table 7.6: Slovak cities with high concentration of production for automobile sector. Source: author.

City Banska Stiavnica Bratislava

Cadca Dolne Vestenice Dolny Kubin

Kosice Krasnany Kysucke Nove Mesto Lazy pod Makytou Malacky Michalovce Modra Nitra Nove Mesto nad Vahom Podbrezova Povazska Bystrica Predmier Presov Puchov Snina Trencin Trnava Velký Meder Vlkanova Vrable Ziar nad Hronom

Automobile sector firms AKUTRADE, s.r.o., Banska Stiavnica MOLPIR, s.r.o., Bratislava; PRESSKAM, s.r.o., Bratislava; Volkswagen Slovakia, a.s., Bratislava; Technicke sklo, a.s. Bratislava; Brose Bratislava, s.r.o; Johnson Controls International, s.r.o., Bratislava; BAZ – Inalfa, a.s, a.s. Bratislava AVC, a.s., Cadca CONTITECH Sealing Systems Slovakia, s.r.o., Dolne Vestenice METALSINT, a.s., Dolny Kubin GLACIER TRIBOMETAL Slovakia, a.s. Dolny Kubin MIBA SLOVAKIA, s.r.o., Dolny Kubin U. S. Steel Kosice, s.r.o. KRASPLAST, s.r.o. Krasnany KLF-ZVL, a.s., Kysucke Nove Mesto VS-Mont, s.r.o., Lazy pod Makytou TOWER AUTOMOTIVE, a.s. Malacky Avin, s.r.o., Puchov KNOTT, s.r.o., Modra Sluzba VD Nitra; PLASTIKA, a.s., Nitra; VW ELEKTROSYSTEMY, s.r.o., Nitra. MAGNA-SLOVTECA, s.r.o., Nove Mesto nad Vahom MARBO SLOVAKIA, s.r.o., Nove Mesto nad Vahom Zeleziarne Podbrezova, a.s., Podbrezova ADTOOL, s.r.o. Povazska Bystrica; PSL, a.s., Povazska Bystrica RUBENA SLOVAKIA, a.s. Predmier FRAGOKOV Presov, vyrobne druzstvo; VAP, a.s., Presov; MATADOR, a.s., Puchov Vihorlat s.r.o. Snina, Jas-Elmont, Snina Leoni Autokabel Slowakia ZF Sachs Slovakia, a.s., Trnava EDSCHA Slovakia Cabrio-Dachsysteme, Velky Meder KÜSTER autom. technika, s.r.o, Vlkanova SVEC a SPOL, s.r.o., Vrable PAL-INALFA, a.s., Vrable ZSNP, a.s., zavod ZLIEVAREN, Ziar nad Hronom


143

Fig. 7.1: Locations of some foreign investors in Slovakia. Source: Author, based on SARIO (2004) Note: Ellipses show locations of main automobile producers. Table 7.7: Supplies of automobile parts and components – Slovak producers. Source: Zdruzenie automobiloveho priemyslu SR, 2004. Turnover (in mln Sk)) Part

Supplies for:

Plastics, insulating material

VW, Skoda, Tatra, Karosa, Ford

Forgings, castings, sheets, tubes

Skoda, AutoVAZ, Dacia, Fiat

2 577.0

3 478.9

3 731.3

4073.8

Cable bundles, lighting, switches

VW, Skoda, Ford, Porsche, Audi, Mercedes

9 228.0

13 002.7

14 357.5

17676.6

Seats, steering wheels, hand brakes

VW, Skoda, Ford, Audi

4 932.0

7 008.3

14 675.5

35476.8

2000

Tools, tooling, SPM, jigs

VW, Skoda, Audi

Bearings, machining, surface treat.

VW, Skoda, Tatra, Daewoo, GM, Citroën

Axles, gears, clutches, brakes

VW, Skoda, Peugeot, Tatra, Volvo, Karosa

2001 214.0

387.4

2002 285.7

2003 691.8

88.0

73.5

101.0

470.0

1 528.0

1 346.4

1 803.5

6968.5

13 051.0

12 179.2

12 764.2

12191.5 12127.1

Rubber, tyres, rubber-metal prod

VW, Skoda, Audi, Ford, Seat

1 724.0

1 486.0

2 048.2

Pressed parts, springs

VW, Skoda, Mercedes

1 136.0

1 353.3

1 047.1

5084.9

Bowdens, filters, rear view mirrors

VW, Skoda, GM, Volvo, Audi

901.0

1 098.6

2 412.4

9108.8

Other

Skoda, Tatra, Daewoo, Renault, Ford

Total

491.0

311.4

96.3

908.1

35 870.0

41 725.7

53 322.7

104777.7

7.5 Model Describing Relation Between Automobile and Parts/Components Production According to our hypothesis, we estimate dependency of parts and components supplies from domestic suppliers for Slovak-based automobile pro-

144

Daneš Brzica

ducer Volkswagen Slovakia (VW).11 For this purpose, we have created a simple model showing relation between the production of automobiles by VW and the volume of component supplies for this company. Intuitively it is possible to see some relation but using a model can make our analysis more rigorous. The output shows the results of fitting a linear regression model to describe the relationship between DOD VW and one independent variable (PROD VW). The model takes the following form: DOD VW = Į +ȕPROD VW

(1)

where DOD VW represents supplies of components and parts from domestic producer to VW Slovakia for the period 1999-2002 (in bn SKK) and PROD VW means the number of personal automobiles produced by VW Slovakia for the same period. Estimated parameters of the model are Į a ȕ. Data were obtained from statistics provided by ZAP SR. The equation of the fitted model is: DOD VW = -10.0197 + 0.33999*PROD VW

(2)

Tables 7.8 and 7.9 show some results obtained from this model. Table 7.8: Regression analysisa

Parameter

Estimate

Constant PROD VW

-10.0197 0.33999

Stand. Error 4.46085 0.0499115

TStatistics -2.24613 6.81187

P-value 0.1538 0.0209

Lower limit(a) -29.2132 0.125239

Upper limit(a) 9.17385 0.554742

a) This shows the confidence intervals at the level of 95% for coefficients in the model.

R-squared = 95.8679 percent R-squared (adjusted for d.f.) = 93.8018 percent Standard Error of Est. = 1.44209 Mean absolute error = 0.908511 Durbin-Watson statistic = 2.27705

11

In the case of both automobile producer and part and component supplier, we consider them as domestic if the product is produced in Slovakia. It is not therefore important, what is the ownership structure of the firm, which owns this production plant.


145

Table 7.9: Analysis of Variance

Source Model Residual Total (Corr.)

Sum of Squares 96.4974 4.15923 100.657

Df 1 2 3

Mean Square 96.4974 2.07962

F-Ratio 46.40

P-Value 0.0209

Since the P-value in the ANOVA table is less than 0.05, there is a statistically significant relationship between the variables at the 95% confidence level. The R-squared statistics indicates that the model as fitted explains 95.8679% of the variability in DOD. Durbin-Watson (DW) value shows that there is probably not any serious autocorrelation in the residuals. Testing the model on Mallows Cp statistics as well as the test according to the highest R2 suggests that the model was properly selected. Based on our model (1)12 we expect that under the similar conditions of cooperation intensity the sector of parts and components for VW company should reach the volume of about 102 bn SKK in 2006 (under condition of expected planned capacity of the company). It is important to note that we have tested another model: DOD Total = Į +ȕPROD Total

(3)

with data covering longer period with production and supplies including the whole automobile sector. However, inability to identify what part of domestic firms´ supplies is exported to foreign-based companies has lead us to reformulate the model focusing on VW company only (for which data exist also for incoming supplies). Thus original estimate of about 19% of domestic supplies for VW does not support corporate information from the company’s management that the share is about 40 percent (for Touareg model even more).13 Because there is substantial part of production generated by VW, it is possible to make this substitution of total production and supplies (model (3)) by VW production and supplies (model (1)). With new investments in real production (especially PSA and Kia), such substitution will not be so adequate.

12

The regression analysis has shown that the model explains sufficiently variability in supplies of components and parts from domestic producers (DOD). 13 According to corporate management, this high share of domestic component supplies has stimulated the decision of the French car company PSA Peugeot Citroen to establish its plant in Slovakia.

146

Daneš Brzica

7.6 Conclusions Recent economic policy changes probably have had and will have many positive effects on the Slovak economy and its automobile sector. Economic efficiency is likely to improve if the reform trends continue. In an international context, a focus on automobile sector has received increased attention by the Slovak government. Nevertheless, tough competition in this sector and increased international dependency of the Slovak economy has also brought about problems regarding automobile sector. The possibilities for small open country, as Slovakia is, to pursue an independent economic policies have been reduced. Therefore, thinking ahead and flexible adaptation have to be principles used by both business and public sectors in Slovakia. The chapter has shown some overview of how the automobile sector in the Slovak Republic has evolved from traditional but marginal automobile producer country into the major one. In addition, based on a simple model, showing relation between the production of automobiles by Slovak-based automobile producer Volkswagen Slovakia (VW) and the volume of component supplies for this company, we estimated dependency of parts and components supplies from domestic suppliers for VW. It was demonstrated that there exists significant relationship between the two variables. Based on our model we expect that under the similar conditions of cooperation intensity the sector of parts and components for VW company should reach the volume of about 102 bn SKK in 2006 (under the condition of expected planned capacity of the company).

References Automotive (2001): “Set up in Slovakia,” Automotive News Europe, Vol. 6, Issue 2, p. 12. Bullen, P. (1999): “What is social capital?” (Based on extracts from “Social Capital: Family Support Services and neighbourhood and Community Centres in NSW”, Paul Bullen and Jenny Onyx, April 1999). http://www.mapl.com.au/socialcapital/soccap1.htm. Cooke, P. (1999): “The German Biotechnology Sector, the Public Policy Impact and Regional Clustering: An Assessment. Report to the U. K. Department of Trade and Industry, Cardiff: Centre for Advanced Studies. De Rossi, L. C. (Ed.) (2004): “What is social capital?” http://www.masternewmedia.org/2004/05/06/what_is_social_capital.htm Dicken, P. (1992): “Global Shift: The Internationalization of Economic Activity.” Paul Chapman Publishing Ltd., London.


147

Feldman, M. P. (1994): “Knowledge Complementarity and Innovation,” Small Business Economics 6(5), pp. 363–372. Frimmer, S. (2003): “Firma, ktora dodavala Fordu, prepusta,” SME, May 20, 2003, p. 6. Gerhardt, K. (2004): “The Automotive Industry in Eastern Central Europe,” Economics and Research Department, IKB Deutsche Industriebank AG, Düsseldorf, Apríl 2004. Hagedoorn, J. (1993): “Understanding the Rationale of Strategic Technology Partnering: Interorganizational Modes of Cooperation and Sectoral Differences”, Strategic Management Journal, 14, pp. 371–385. Håkansson, H. (1989): “Corporate Technological Behaviour: Co-operation and Networks.” London: Routledge. Hales, M. (2001): “Birds were dinosaurs once – The diversity and evolution of research and technology organizations. Final report. Centrim, University of Brighton, January 2001. Hansen, N. M. (1992): “Competition, Trust, and Reciprocity in the Development of Innovative Regional Milieux”, Papers in Regional Science 71, pp. 95–106. MH SR (no date): “Analyza vyvoja v odvetvi strojarskeho priemyslu za roky 1996-1999.” Ministerstvo hospodarstva SR, Bratislava. MH SR (no date): “Program transferu technologii.” Ministerstvo hospodarstva SR. OECD (2000): “Innovative Clusters: Drivers of National Innovation Systems.” OECD: Paris. Nemcova, E., Gasparikova, J. (2004): Industrial and Regional Development in Slovakia. In: Bruno S. Sergi, William T. Bagatelas (Eds.) The Slovak Economy and EU Membership. Iura Edition, Bratislava, pp. 173–178. Rentmeister, B. (1999): “Wissensintensive Dienstleistungen in der Automobilentwicklung (Knowledge-intensive Services in Car Development),” Working Paper SFB 403 AB-99-27, Institut fuer Wirtschafts- und Sozialgeographie, Johann-Wolfgang-Goethe University, Frankfurt am Main. Riives, J., Otto, T., Olt, M. (2002): “Business-aid networking in production”, Proceedings of the 3 rd International Conference Industrial Engineering – New Challenges to SME, 25 – 27 April 2002, Tallinn: DAAAM National Estonia, pp. 249-252. World Bank (no date): “Slovenska republika – pristupovy proces do EU: Prehlad rozvojovej politiky.” SARIO (Slovak Investment and Trade Development Agency), (no date) “Sector analysis – Automobiles, trailers and semitrailers production in the Slovak Republic,” SARIO (Slovak Investment and Trade Development Agency), (2004): “Slovakia: The Most Attractive Investment Location in Central Europe.” http://www.mobileeurope.net/samba/documents/SAMBA_4_SarioPresentatio n_040429.pdf IRC (no date): “Automotive industry in the Slovak Republic”. Document on the web-site: http://www.irc-slovakia.sk/tt/tool/index-en.shtml.

8 IT Market and E-Commerce in Transition Economy: Network Externalities

Vytautas Snieška1, Regina Virvilaitơ, Vaida Kvainauskaitơ, Bronius Neverauskas, Rimantas Gatautis, Aistơ Dovalienơ 1

Kaunas University of Technology, Lithuania, E-mail: [email protected]

Abstract. The article presents the main aspects of interaction between network externalities, IT market, and e-commerce development under conditions of a transition economy. In literature it has often been found that network externalities cause a decrease of what is generally called transaction costs (costs of gathering information, controlling, coordinating transactions). With respect to the decrease of transaction costs, this chapter shows that network externalities are extremely relevant for the Lithuanian IT market. Keywords: Network Externalities, IT Market, Transition Economy, Transaction Economy

8.1 Introduction As the economy becomes more interconnected, more products in computing, consumer electronics, and telecommunications industries exhibit network externalities (Yoffie, 1997). Extensive literature in economics has examined the strategic and welfare implications of network externalities (Economides, 1989; Farrell and Saloner, 1986; Katz and Shapiro, 1985). A consistent finding in the literature is that network externalities alter customer behaviour (e.g., before adopting the product, it is rational for people to wait for others to do it) and have important implications on a firm’s strategy. Previous studies have explored the implications of network externalities on (1) customer behaviour and market structure (Frels, Shervani, and Srivastava, 2003; Goldenberg, Libai, and Muller, 2002; Shankar and Bayus, 2003); (2) product-related decisions such as pre-announcements (Nagard-Assayag and Manceau, 2001), timing of product introductions

150

V. Snieška et al.

(Padmanabhan, Rajiv, and Srinivasan, 1997), product differentiation (Esser and Leruth, 1989), and (3) market entry (Gupta, Jain, and Sawhney, 1999; Xie and Sirbu, 1995). Network externalities have been found to be important in the context of e-commerce development (Shapiro and Varian, 1998; Liebowitz and Margolis, 1994; Gulati, 1999). The strategic network perspective has been fostered by the economics of electronic networks where the cost of gathering information, controlling, and coordinating transactions with other economic actors, has been significantly reduced (Malone et al., 1989). Transacting in e-commerce industries may have significant effects on transaction costs. Direct and indirect costs may decrease as a result of an increasing frequency of transactions (due to open standards), reduction in uncertainty (by providing more transaction-specific information), and reduction in asset specificity (lower site specificity) (Amit and Zott, 2001). The aim of the article is to evaluate the influence of IT market and ecommerce on network externalities in transition economy.

8.2 The Structure of Network Externalities To discuss the advances of this age of technology, the economist has invoked a concept of externality. Externality has been defined as a change in benefit, or surplus that an agent derives from a product when a number of other agents consuming the same kind of product changes (Bailey, 1995). M.M. Halgren and A.K. McAdams (1995) assume, that this type of economic goods that fall between pure private goods and pure public goods are referred to as goods with externalities. An unintended “spillover” of any good is called an externality (Baumol and Oates, 1975). If the spillover is positive, then it is a positive externality or a benefit; if the spillover is negative, then, it is a negative externality or a cost to society. In some cases, the positive economic spillover may actually be of more benefit than the intended benefit of the product to its original creator, as it often occurs with research (Halgren and McAdams, 1995). Externalities are endemic to e-commerce and because of specific business experience problems of such industries and different characteristics of those that have business of more ordinary commodities; externalities are named as network externalities. Regarding the increasing importance of network externalities in the economy, extensive literature in economics has examined the strategic and welfare implications of network externalities (Farrell and Saloner, 1985; Katz and Shapiro, 1985).

8 Market and E-Commerce in Transition Economy

151

The concept of network externality has been applied in the literature on standards, in which the primary concern is the choice of a correct standard (Farrell and Saloner, 1985, Liebowitz and Margolis, 1994). M.L. Katz and C. Shapiro (1985) consider two types of positive network externalities. Firstly, they consider direct network externalities – those generated “through a direct physical effect of the number of purchasers on the quality of the product.” Their example of direct externality is the number of homes attached to a telephone network. Secondly, they consider indirect network externalities or “indirect effects” such as complementary goods being more plentiful and lower in price as the number of users of the good increases. Their example here is better software as the number of computers of a particular type increases. Furthermore, they consider another source of indirect network externality, the availability of post-purchase service for durable goods such as automobiles. In a similar vein, J. Farrell and G. Saloner (1985) observe that “There may be direct ‘network externality’ in the sense that one consumer’s value for a good increases when another consumer has a compatible good, as in the case of telephones or personal computer software. There may be a market-mediated effect, as when a complementary good (spare parts, servicing, software, etc.) becomes cheaper and more readily available the greater the extent of the (compatible) market”. Although economists observe a distinction between direct and indirect externalities, this distinction does not figure into the existing theoretical analyses. In the theoretical treatments, both types of network externalities are assumed to have the same consequences: direct and indirect interactions alike are embodied in payoff functions, regardless of their source. J. Farrell and G. Saloner (1985), for example, postulate a benefit function Bj (S,Y), in which j denotes the firm, Y denotes the firm’s choice of technology and S denotes the size of the network (number of firms choosing Y). M.L. Katz and C. Shapiro (1985) specify that the net benefits of a consumer are v (x1+x2)-p, where x1 and x2 are the size of the network in time period one and two, and p is the price of a unit of the technology. In our opinion, direct and indirect network externalities are fundamentally different and should not be modelled as equivalent. From the modern perspective, the early twentieth-century debate on the nature of externalities may appear rather unusual due to the nonmathematical apparatus that it used. However, this unusual apparatus did ultimately manage to distinguish between technological and pecuniary externalities. A consistent finding in the literature is that network externalities alter customer behaviour (e.g., before adopting the product, it is rational for

152

V. Snieška et al.

people to wait for others to adopt the product) and have important implications for transaction cost economics.

8.3 Network Externalities and Transaction Cost Minimization As the economy becomes more interconnected, more products in computing, consumer electronics, and telecommunications industries exhibit network externalities (Yoffie, 1997). The use of IT has considerably changed the costs of information gathering, as well as controlling and coordinating of market transactions. Electronic market places that electronically connect buyers and sellers through a central database reduce transaction costs for both buyers and sellers. Transactions are transferred from internally coordinated activities to market exchanges (Malone et al., 1989). Value chains are disintegrated by outsourcing activities (Evans and Wurster, 1997), and distribution channels disintermediate by eliminating intermediaries or reintermediate by existing intermediaries migrating to the electronic market places as market makers (Amit and Zott, 2001). The importance of information-based resources and capabilities increase within e-business firms and accessing of such resources through partnering and resource sharing agreements becomes more viable (Amit and Zott, 2001). These resource sharing organizational forms are commonly denoted as strategic networks. Sociologists have focused on network structures in terms of density and centrality (Freeman, 1979); strategic management has been concerned with trust (Lorenzoni and Lipparini, 1999) as well as resources and capabilities (Gulati, 1999); and economists have studied network effects such as indirect network externalities (Gupta et al., 1999) and direct externalities (Shapiro and Varian, 1999). The idea of network externalities is based largely on the general impression that there are a large and increasing number of activities in which costs or benefits rise or fall as the number of participants increases. And this impression seems to apply particularly to new, high-tech industries, for instance, e-commerce. Transacting in e-commerce industries (over the Internet) may have significant effects on transaction costs. Direct and indirect costs may decrease as a result of an increasing frequency of transactions (due to open standards), a reduction in uncertainty (by providing more transactionspecific information), and a reduction in asset specificity (lower site specificity) (e.g., Amit and Zott, 2001).


153

Transaction costs are the costs, which emerge when commodities or services are exchanged, but they are not related to the production of a commodity or a service. According to some authors, electronic commerce can reduce transaction costs due to lower: 1) search costs (Bakos, 1997), 2) co-ordination costs (Malone, et al., 1989), and 3) payment processing costs (Sirbu and Tyger, 1995). If using e-commerce reduces transaction costs to the point where they become lower than the costs related to transaction between the external and turn into the internal firm costs, an organisational change occurs: market transactions turn into internal firm transactions (transactions Æ intratransactions). In the opposite case, the number of market transactions is higher than that of transactions inside a firm. Since no detailed study of the effect of e-commerce on the transaction costs has been conducted, a deeper research can help to explain the factors, which can impact the change of transaction cost. 8.3.1 Transaction Costs in the Market Without Intermediaries The most common transaction costs emerge in direct transaction between supplier and consumer. Supplier, consumer or both can cause transaction costs. As a result of the transaction cost effect, supply and demand curves shift to the left due to higher transaction costs. Since market is nonfrictional, to conclude a transaction one of the parties has to pay an extra amount. Supplier can have the best commodity produced at lowest cost but not to be able to sell it because potential consumers are not aware of it. Therefore, to make consumers enter a transaction, supplier may have to advertise the commodity or to develop a web page. All costs related to disseminating the information to increase trade volume are transaction costs, which are covered by supplier. On the other hand, consumers may want to purchase the commodity, however, they do not know whether it exists, whether anyone sells it and at what price. Before purchasing consumers have to search for information, to communicate and negotiate with potential suppliers. These costs covered by consumers are also transaction costs. 8.3.2 Transaction Costs in the Market with Intermediaries A third party, i.e. a mediator can also assume transaction costs. It is a firm that markets a commodity without producing or consuming it. Intermediaries compete with other firms that can be involved in selling similar commodities or services. Intermediaries can be in a better position to reduce transaction costs than suppliers or consumers. Since intermediaries participate in a number of repetitive transactions, they develop a network of rela-

154

V. Snieška et al.

tionships and accumulate experience, which allows them to reduce transaction costs. Besides, intermediaries can invest into technologies that require high fixed costs but allow reducing marginal costs of other transactions. High volumes of transactions enable intermediaries to absorb fixed costs. Although intermediaries reduce transaction costs, it is not clear what their role and importance are. In the literature it is argued that a mediator is a participant of the transaction which co-ordinates transactions between consumer and supplier, however, in more complex definitions mediators’ roles are not defined in an equally detailed way. Often intermediaries’ role depends on the context since they perform different roles in different transactions. Mediator’s role is characterised as changeable. Analysis of the intermediaries is often ignored in the literature (Bailey, 1998). Intermediation and disintermediation define inclusion or elimination of certain elements between supplier and consumer. Disintermediation occurs when a mediator is eliminated from the transaction. Intermediation emerges when a mediator appears. The process of disintermediation does not necessarily mean that the intermediation level in the value chain changes from n to n-1. For example, disintermediation occurs when market is shifting from the two-level to one-level intermediaries. However, a rather commonly held view that e-commerce causes decrease in intermediation services and stimulates disintermediation is unjustified (Gatautis, Neverauskas, Snieška, 2002). 8.3.3 Transaction Costs and E-Commerce Transaction costs include costs that are necessary for market transaction to happen. As it has been mentioned above, these costs do not modify the commodity but help to assess the market. In physical markets these costs may be related to property, window case inventory, advertising, etc. Meanwhile, in e-commerce, these costs are limited to advertising costs because retailers do not need to present the products physically to show the range they are offering. However, in this case other fixed costs emerge. They are necessary for e-commerce to cover market transactions happening in the virtual environment, e.g., the costs of maintaining a virtual store, Internet use costs, etc. Fixed costs related to creating an on-line presentation can be lower than those incurred when organising physical presentation. In the latter case, to be able to attract customers one needs to choose a certain location with relevant infrastructure. Meanwhile Internet server can be located in a very remote place and even be shared with other suppliers and intermediaries. ‘See and feel’ buying is different in e-commerce; however, costs related to


155

creating an analogous atmosphere are much lower. Lower market entry costs stimulate entry of the new players and competition, which results in consumers being able to choose from a greater number of suppliers and commodities. When prices of physically sold commodities and of those sold on-line are identical, profit margins are different. When the price falls, while profitability remains the same, sales have a tendency to increase. Transactions in physical market are limited by territorial location. In the on-line space, suppliers and intermediaries can expect to attract also those consumers, whose choice of products is limited due to high costs of searching for physical goods. When markets shift from physical to virtual environment, this can result in the intense competition and bigger choice. Therefore, those who decide to operate in the virtual environment will have to do so under conditions of more intensive competition because in this case firms compete in the broader market (both physical and virtual). In a physical market, a certain correlation between sales volumes and geographical location is observed (Figure 8.1). This dependence illustrates Hotelling’s (1929) so-called ‘length city’. Physical and electronic markets are different because e-commerce transactions do not depend on physical location. In the e-market, distribution takes place along the dark line showed in the figure. Those retailers who prefer on-line presentation can compete in the bigger geographical space. Also, Internet makes it possible to enter the market for those suppliers, whose market entry costs are too high and the scope of the geographical market is too narrow. 8.3.4 E-Commerce and Intermediation Services Sarkar, Butler and Steinfield (1995) argue that transactions can be grouped according to the functions performed by a mediator. The co-ordinating role of the mediator in the exchange process is not the most important. The mediator performs various functions in the transaction process, and with emergence of e-commerce these functions change to a greater or less extent. Transaction cost theory is often used to explain the effect of the emarkets on intermediation services. Transaction is defined as ‘an exchange of commodities or services between sellers and buyers’ (Williamson, 1979). The firms’ attempt to minimise transaction costs by reducing coordination costs. It allows discontinuing trade with external firms and hierarchical organisations. Sarkar’s (1995) model is based on this theory. It is believed that Internet can be used to make pressure on hierarchical firms and that it will stimulate emergence of the new market players. This is es-

156

V. Snieška et al.

pecially true in case of intermediation. It could be explained by a simplified model of the transaction costs and mediator’s functions (Figure 8.2). Sales Physical market E-market

Business location Fig. 8.1 The relationship between business location and sales.

In the market where mediator services are used, Tp-c is bigger than Tpi+Ti-c. If e-commerce reduces transaction costs to zero, the situation changes, i.e., Tp-c gets smaller than Tp-i+Ti-c. Thus, producer can discontinue using mediator’s services. These conditions are based on two theoretical assumptions: Internet accessibility makes transaction costs equal to zero, and transactions are very small (micro-transactions). However, taking into account these assumptions, Sarkar (1995) notes that: x Transactions are different, and this leads to a greater number of possible forms of intermediation; x The role of intermediaries as co-ordinators includes a lot of different functions. In case of several different types of transactions (Figure 8.3) it is likely that not all operations are affected by e-commerce in the same way. Therefore, it can be argued that not all operations reach the same minimum of transaction costs. Minimisation of transaction costs is related to the four types of intermediation: x The first possibility is a process-taking place without intermediation (disintermediation). However, this process is limited due to various reasons (lack of knowledge, social and cultural characteristics of intermediaries, etc.). x Re-intermediation implies performing already existing functions, only they are redefined on-line. Although these functions exist, other intermediaries can perform them in virtual markets.


157

x Internet causes increase in the number of virtual intermediaries (cybermediaries) that present new possibilities in the virtual values chains. They can be characterised as extra-intermediaries. x Finally, information technologies supplement direct market: producers can expand in already existing markets by exploiting the IT advantages. For example, Dell.com is shifting from call centres to e-commerce. Mediator Ti-c

Tp-i

Producer

Consumer Tp-c

Tp-c – transaction costs between producer and consumer, Tp-i – transaction costs between producer and mediator, Ti-c – transaction costs between mediator and consumer. Fig. 8.2. Transaction costs and the relationship between producer, consumer and mediator (Sarkar, 1995).

Sarkar (1995) argues that the co-ordinating role of the intermediary in the transaction process involves a set of functions. These functions differ; therefore, some of them are not affected by any electronic services enabled by IT. Traditional intermediaries perform many services and functions. Meanwhile the value of the virtual intermediaries lies in the fact that they facilitate operations in the e-market, which results in the increasing number of players joining this market. In addition, virtual intermediaries generate network effects. Business in virtual environment is directly related to reduction in transaction costs (Malone, et al., 1989). This enables producers to revise their commodities and services. With increase in the number of functions, virtual market players expand their activities and in this way increase the number of market players. E-market enables higher market accessibility, which increases competition and gives intermediaries new specialised functions. Integration of e-services with products and services helps to create new transactions, which serve consumer society (Buxmann, Gebauer, 1998).

158

V. Snieška et al.

Mediator Ti-c

Tp-i

Producer

Consumer Tp-c

Fig. 8.3. Differential minimisation of transaction costs (Sarkar, 1995).

8.3.5 The Impact of E-Commerce on the Co-Ordination Costs When the firm can choose out of many market partners, production efficiency and competition helps it to keep costs low, but it has to cover relatively high co-ordination costs (Malone et al., 1989). On the other hand, agreements with single partners (hierarchies) reduce the firms’ choice. However, close relationship decreases search costs, eliminating the need to collect and analyse information about potential partners. Therefore, decision to buy or to produce is regarded as an alternative between production and co-ordination costs. Referring to Coase (1937), Williamson (1979) emphasises that the costs of the hierarchical transactions are lower than those of the market transactions. According to Malone et al. (1989), while reducing co-ordination costs, at the same time e-commerce makes it easier to shift from hierarchical agreements and internal production to market agreements when the firm has several business partners. For example, in the literature, it is disagreed whether the progress that reduces the costs of searching for buyers, such as e-commerce, should increase the number of suppliers to choose from, especially in the markets for differentiated products (Bakos, 1991). Thus, decrease in co-ordination costs related to e-commerce, all things being equal, should enable most of the firms to increase the number of their suppliers. It would seem that increase in the number of suppliers and choices is beneficial for firms. However, the number of business partners and choices also depend on organisational and technological factors such as affiliation costs, search costs and other costs that can be relatively called co-ordination costs. Potential business partners offer possibilities that differ by such characteristics as price or parameters of the product. In any case, communica-


159

tion with every partner results in co-ordination costs. When collaborating with a number of partners, an organisation can choose the one that gives the offer value according to the criteria set by organisation. An optimal number of the partners could be determined by comparing the coordinating costs with the benefits, which would be received from a broader choice of partners. This comparison is shown in Figure 8.4 (Bakos, Brynjolfsson 1999). As it has been mentioned, it is believed that e-commerce reduces coordination costs inside the firm itself. By illustrating that lower coordination costs are related to the bigger number of business partners, Figure 8.5 shows the impact of lower co-ordination costs on the TC curve. A bigger number of partners emerges due to the decreasing marginal coordination costs with every new partner, which emerge even despite high costs of implementation of e-commerce, e.g. implementation of electronic data transmission (EDT). When the firm has several outsourcing partners, relationship with business partners becomes especially important (Brynjolfsson 1991, Johnston, Lawrence 1988). It is obvious that despite higher trust on the market, firms are inclined to be dependent on the lower number of partners. Dependence on the larger number of partners is related to the risk that they will behave opportunistically or do some harm to the firm (Klein, Langenohl, 1995). As pointed out by Clemons, Reddi and Row (1993), ecommerce can reduce specific investment into co-ordinating costs and, at the same time, provide the firm with better opportunities to monitor partners’ activities. As a result, the risk of encountering an opportunistic behaviour of the partners when several highly interdependent firms negotiate can be highly reduced. Clemons, Reddi and Row (1993) argue that this not only makes it easier to use outsourcing, but also enables the firm to collaborate with a smaller number of partners. IT-based relationships with a few partners also result in e-commerce increasing the total co-ordination costs. It determines the steepness of the coordination costs curve in Figure 8.4. In this case, the effect would be opposite than that shown in Figure 8.5. For example, if implementation of ecommerce to communicate with every business partner requires higher fixed technological and organizational investments, then, the firm, seeking to reduce co-ordination costs, will do business only with a few partners. Similar effect will be achieved if the size of the firm’s investment into e-integration depends on the number of partners. In those cases when new partners cannot replace them, the firm will create selection costs that will determine the number of business partners in a given period.

160

V. Snieška et al.

Costs

TC

CC

PFC

TC – total costs CC – coordinating costs PFC – poor fit costs

Number of partners

Fig. 8.4. Optimal number of business partners (Q).

Costs

TC CC TC1

Change in coordination costs

CC1

PFC

Number of partners,Q

Fig. 8.5. The optimal number of business partners after reduction of co-ordination costs.


161

8.4 IT Market and E-commerce Growth Effects: Lithuanian Case In 2003, when Lithuanian GDP increased by 8 %, foreign experts recognized the Lithuanian economy as the fastest growing economy in Central and Eastern Europe. One of the most important growth criterions of Lithuanian economy was the information technologies sector, which has been developing faster than the overall economy by nearly 3 times. Rapidly growing GDP has influenced home user and small business market, therefore, sales of computer hardware in this segment have been and are still developing dynamically. A second tendency is that the bulk of Lithuanian IT companies are specialized by putting much more investment into competency of their workers. Their target is to establish themselves within IT solutions and services market. Thirdly, Lithuanian IT companies are starting to operate in the countries of neighbouring region by establishing their daughter companies and exporting technological solutions. Market turnover, computerization level, number of consumers, frequency of Internet use, etc. characterize the changes in the Lithuanian IT market and e-commerce. In 2002, the Lithuanian IT and telecommunications market reached 1.035 billion LTL (about 300 million €), and increased by 8 to 10 % in 2003 (Figure 8.6). The turnover of telecommunications market extended to 2.1 billion LTL (about 610 million €) in 2003. The sector of mobile telecommunications increased by 11 % in 2003. In 2002, the penetration of mobile telecommunications services reached about 47 % of all Lithuanian population, and, at the end of 2003, it amounted to already 62 %. The gross IT market increased by 8 % in 2003, and reached 528 million LTL (153 million €). The overall Lithuanian IT export increased by 8,7 % in 2003, and the export of the largest companies increased from 40 to 80 %. In 2002, the export of IT services amounted to 35 million €. The key markets of the Lithuanian IT export have been Germany, Russia, Belarus, Finland, Denmark, the USA, and Latvia. According to the data at the beginning of 2004, 695.7 thousand people in Lithuania used Internet at least once a month, or it composed 26,5 % of population of the age of 15-74. It is 31 % more than in the same period of 2003 (533 thousand or 20.3 %) and 17 % more than in autumn 2003 (593.3 thousand or 22.6 %). In the period of January – February of 2004, 577.6 thousand people (22.2 %) used Internet at least once a week. It is 31 % more than during 2003 (443.7 thousand or 16.9 %) (7 Figure). 28.1 % of respondents used Internet at least once in a half of a year (in winter of 2003, this rate was

162

V. Snieška et al.

22 %, in autumn – 24.7 %). The more active users of Internet in Lithuania are young people. 43.2 % from about 737.7 thousand people, who used Intenet during the past half-year, was from the age group of 15-24, 25.1 % 25-34, 24 % - 35-49, and 7.8 % - 50-74. 350

10

300

8

250

6

200

4

IT market, mln. €

150

2

Change in GDP, %

100

0 -2 -4 1999

2000

2001

2002

2003

Fig. 8.6. IT market development and changes in GDP in 1999-2003.

It should be noted that such a rapid growth of Internet users (31 %) exists in the same time with the growth of GDP by 9 % (Figure 8.6). It indicates the evident benefit for IT market and e-commerce in market changeover. By the data of Statistical Department, in the 3rd quarter of 2003, onefifths of Lithuanian households (20 %) had a PC at home (in 2002 – 12 %) (Figure 8.7). The provision with computers, mobile phones and the use of Internet had a direct dependence on the household’s income. Among the households with the monthly income of more than 1000 LTL, 47 % owned personal computer and 23 % had Internet. The provision with PCs among the Lithuanian households demonstrated a notably rapid growth in the past three years. In 1996, only one of one hundred households had a PC at home; in the 3rd quarter of 2003 – every fifth household was “computerized”. Such rapid growth of IT market has a direct influence on the development of e-commerce. Although, recently, the basis for e-commerce has consisted of Internet orders, this area is growing quite fast. By this time, there have been about 100 networks of Internet commerce in Lithuania, most of which started their activities in 1999-2000. The positive influence on the development of e-commerce has caused the increase of computerization level, spread of Internet, expansion of e-bank services and rapid speed of Internet advertisement growth. (In 2003 the market of Internet advertisement increased to 3.7 million LTL or 1.5 times more vs. 2002).


163

As a reaction to rapid IT development, intentions of business companies to develop e-commerce have been emphasized. According to the Infobalt association research, the interest in e-commerce opportunities among the Lithuanian companies has been notably growing (see Table 8.1) Even 25.4 % of the Lithuanian enterprises have been preparing to the use of e-commerce for business expansion. It shows that enterprises understand the importance of electronic trade. Comparing the attitude to Internet and separate elements of e-commerce among the Lithuanian enterprises, we can assume that the majority (74.6 %) absolutely agrees that communication over the Internet would be more intensive in the future; other 14.2 % agrees partially. About a half of the Lithuanian enterprises (48.2 %) partially agree, that Internet raises many concerns related to data safety, 13.5 % absolutely agrees with that. Approximately one fourth of the Lithuanian enterprises (24.7 %) partially agree that the purchase of goods or services by Internet is unreliable, and 41.3 % does not have any definite opinion on it. The greatest disapproval was on the statement that “internet rather aggravates than eases work”: almost half of respondents (47.2 %) partially disagree and almost one-third (28.3 %) absolutely disagrees with this statement. 25 20 Internet users (at least once per week), % of the population Households having PC, % of the households

%

15 10 5 0 2001

2002

2003

2004

Fig. 8.7 Internet users and the computerization level in Lithuania in year 2001-2004.

With the increase of the Internet users number, Lithuanian IT business is investing more and more into e-commerce – companies are starting to believe that e-business will pay back. This is encouraged by positive examples of computer wholesalers whose 60 % of trade is done through ewarehouses. Distribution of software by the Internet is also done quite successfully.

164

V. Snieška et al.

Services of online order collection and provision for companies of various fields are becoming more and more popular - companies specialised in construction, transportation or publishing find their orders online by paying a regular subscriber’s fee. Table 8.1. Intentions of Lithuanian enterprises to develop e-commerce.

Will create/develop local network Will develop use of Internet Will begin to use Internet for communication Will create/develop Internet sites Will begin/develop Internet sales Will begin to use/develop Internet advertisement

Certainly begins in the next 6 months 8.7

Probably begins in the next 6 months 17

13.8 8.6

25.7 18.6

14.0

19.6

2.4

13.2

9.4

25.7

According to the information of Statistics Department, 6.7% of Lithuanian companies indicated that they sold goods (services) by Internet, 9.6% - purchased them online in the first half of 2003. The analysis shows that rapid growth of IT and telecommunication markets constitute the background for e-commerce development in Lithuania. The possibility to do shopping without rush, lower prices and possibility to order products 24 hours a day, motivates enterprises to use ecommerce in their performance. Such positive attitude on e-commerce sets the basis for network externalities development.

8.5 Research Methodology Trying to assess conditions and assumptions of network externalities that emerged from the rapid growth of IT and e-commerce market in Lithuania, empirical research has been made using the method of expert assessment. Research objective – to assess network externalities emerging from the IT infrastructure development and e-commerce spread that determined minimization of transaction costs. Research hypothesis H1 – network externalities are impacted by the minimized costs of IT infrastructure and e-commerce development. Fifty (50) people - executives and managers of the companies that had been establishing IT - were chosen to be the experts. The expert selection


165

was based on the assumption that IT companies making e-commerce decisions in various companies were well informed about the opportunities of IT use in e-commerce and its benefit to the companies. The direct expert survey was carried out in the form of a questionnaire that was delivered directly to expert by asking them to fill it in. This method is more efficient than the post survey because the probability of getting answers is bigger (the response rate is higher). The questionnaire comprised two parts: 1) expert opinion about general influence of ecommerce on the possibilities to minimize transaction costs (experts were asked to evaluate the significance of this factor under 10 point scale) (Table 8.2); and 2) expert opinion on the opportunities of e-commerce to minimize transaction costs of the company main activities (for example, incoming logistics, production – technological operations, etc.) (Table 8.3) and in the area of inter-organizational contacts (for example, intermediary selection, setting up the optimal number of partners) (Table 8.4). Table 8.2 How do you evaluate the e-commerce on the possibilities to minimize transaction costs?

Very low 1 2

3

4

5

6

7

8

Very high 9

10

Table 8.3 How do you evaluate the opportunities of e-commerce to minimize transaction costs of the company main activities?

Very low 1

2

3

4

5

Very high 7 8 9

6

10

Incoming logistics Production – technological operations Outgoing logistics Marketing and sales Table 8.4 How do you evaluate the opportunities of e-commerce to minimize transaction costs in the area of inter-organizational contacts?

Very low 1 Intermediates selection Settle of optimal number of partners

2

3

4

5

6

Very high 7 8 9

10

166

V. Snieška et al.

Analysing the opportunities of e-commerce use, the experts gave 10 points for the most important factor, and, accordingly, 9, 8, …, 1 points to the others. When several factors had the same importance, the same point (of importance) was given to them. The significance level of 0.05 was chosen in evaluating the significance of various factors. Unanimity of experts’ opinions was assessed by Kendal concordance coefficient W:

12S m (n 3 n)

W

2

(1)

When experts gave the same rates to several factors, the Kendal concordance coefficient was calculated as follows:

12S

W

m

(2)

m (n n) - m¦ Tj 2

3

j 1

Q

Tj

¦ (t

3 jq

t jq )

(3)

t jq 1

where

S - sum of square deviation; n – number of factor groups; m – number of experts; tjq – number of the same ratings in j row. The Kendal concordance coefficient may vary from 0 to 1. When W=1, all experts gave the same factor ratings, when W=0, we can assert that experts’ opinions differed. When W t 0.6, it means that the experts’ opinion was unanimous, and the results of expert assessment were reliable. The reliability of questionnaire was assessed by Cronbach alfa coefficient:

Į where

Nr 1 (N 1)r

(4)

N – number of experts; r – inter-item correlation.

In order to choose the correlation coefficient, the Kolmogorov – Smirnov test was carried out. It tested the hypothesis of the normal distribution. In case of normal distribution, Pearson, Kendall and Spearmen correlation coefficients were used; in case the distribution was not normal, Kendal and Spearmen correlation coefficients were used. All calculations were made by using the SPSS software.


167

Pearson correlation coefficient was calculated as follows: n

¦(x

i

x )( yi y )

i |1

r|

(5)

n ªn 2 ºª 2º ( x x ) ¦ ¦ i « » « ( yi y ) » ¬ i |1 ¼ ¬ i |1 ¼

where dx and dy – difference between every value and corresponding arithmetical average. Spearmen correlation coefficient rs was calculated as follows:

rs

1

where

6 ¦ d i2 ; n( n 2 1)

¦d

2 i

(6)

- sum of comparison of differences between the

ranged values. n – number of paired comparison.

8.6 Results After assessing the unanimity of experts’ opinion, the Kendal concordance coefficient was W=0.62. Where the Kendal concordance coefficient was higher than 0.6, we can suggest that the experts’ opinion was unanimous. The reliability of questionnaire was assessed by Cronbach alfa coefficient, which was 0.78 and ensured that the questionnaire was reliable. The correlation between experts’ opinion on e-commerce significance and experts’ opinion on the opportunities of e-commerce to minimize transaction costs in the company main activities was assessed. The level of significance was calculated by Smirnov-Kolmogorov test (see Table 8.5). Since the level of significance was higher than the settled level (0.148>0.05 0.10>0.05 0.146>0.05 0.023>0.05 0.10>0.05), the distribution was normal. In order to assess relations between variables of the research, the Pearson, Kendall and Spearmen correlation coefficients were calculated (see Table 8.6). If the value of correlation coefficient was less than 0.5 – relation was weak; if the value varied from 0.5 to 0.7, the relation was moderate; if the value varied from 0.7 to 1, the relation was strong. Summing up the results, we can suggest that two areas of the company activities – production-technological operations and outgoing logistics - had a strong correlation with development of e-commerce infrastructure.

168

V. Snieška et al.

Table 8.5 The results of Kolmogorov-Smirnov test (I).

E-commerce Incoming significance logistics

Production – Outgoing Marketing technological logistics and sales operations (I) Assessment of e-commerce use opportunities by minimizing transaction costs in key areas of activity Mean 1.87 1.33 1.60 1.60 1.67 Standard de0.92 0.49 0.63 0.51 0.49 viation Kolmogorov1.142 1.624 1.144 1.491 1.624 Smirnov coefficient Level of sig0.148 0.10 0.146 0.023 0.10 nificance Table 8.6 Pearson, Kendall and Spearmen correlation coefficients (I).

Incoming logistics

Production – Outgoing Marketing technological logistics and sales perations (I) Assessment of1 e-commerce use opportunities by minimizing transaction costs in key areas of activity Pearson corre0.586 0.888 0.800 0.693 lation coefficient Kendall corre0.571 0.882 0.775 0.671 lation coefficient Spearmen correlation coeffi0.601 0.932 0.816 0.707 cient

Determining the significance of intermediary selection and the settling of the optimal number of partners for the reduction of transaction costs, according to experts’ opinion, the correlation between IT based ecommerce valuation and intermediary selection as well as the optimal number of partners were assessed (see Table 8.7). Since the level of significance was higher than the settled level (0.148>0.05 0.166>0.05 0.023>0.05) the distribution was normal. Pearson, Kendall and Spearmen coefficients are presented in Table 8.8. Summing up the results we can suggest that the correlation between ecommerce and selection of intermediary was strong, meanwhile, the correlation between e-commerce and the settlement of optimal number of partners was moderate.


169

Table 8.7 The results of Kolmogorov-Smirnov test (II).

E-commerce significance

Intermediates selection

Settle of optimal number of partners (II) Assessment of e-commerce use opportunities by minimizing costs in the area of inter-organizational relations. Mean 1.87 2.33 2.40 Standard deviation 0.92 0.72 0.51 Kolmogorov-Smirnov coef1.142 1.116 1.491 ficient Level of significance 0.148 0.166 0.023 Table 8.8 Pearson, Kendall and Spearmen correlation coefficients (II).

Intermediates Settle of optimal number selection of partners (II) Assessment of e-commerce use opportunities by minimizing costs in the area of interorganizational relations. Pearson correlation coefficient 0.895 0.689 Kendall correlation coefficient 0.745 0.605 Spearmen correlation coefficient

0.802

0.596

8.7 Conclusions 1. Although economic analysis of network externalities shows clear possibility to compute costs and benefits that the user of network derives from additional person using the same network, it is quite complicated to distinguish and describe tangible effects of network externalities in Lithuania. 2. Network externalities alter customer behaviour (e.g., before adopting the product, it is rational for people to wait for others to adopt the product) and have important implications for transaction cost economics. E-commerce serves as a catalyst for such changes in customer behaviour. 3. The analysis of the conditions and perspectives of network externalities of e-commerce growth in Lithuania should be analyzed and evaluated indirectly, because the comprehensive analysis of network externalities effects is limited by: 1) monopolistic competition in telecommunication markets; 2) scarce quantity of e-

170

V. Snieška et al.

4.

5.

6.

7.

commerce services; 3) lack of information about advantages of ecommerce. E-commerce enables firms to reduce the co-ordination costs and, as a result, to increase the number of partners. A firm can minimize transactions costs by using such types of intermediaries as direct market, immediate market, new (virtual) intermediaries and re-intermediaries. The research results revealed that network externalities were determined by some factors of the company transaction costs connected to production – technological operations, incoming and outgoing logistics, marketing and sales, intermediary selection and setting up the optimal number of partners, whose efficiency was partly influenced by e-commerce growth in Lithuania. Estimated correlation coefficients showed that the use of ecommerce spread generated positive network externalities and minimized transaction costs especially in such areas as production – technological operations, outgoing logistics, intermediary selection. Estimated Pearson, Kendall and Spearmen correlation coefficients evaluated the relation of e-commerce growth with network externalities over transaction cost minimization that varied from 0.571 (incoming logistics) to 0.932 (production – technological operations); the estimated level of significance done by using Kolmogorov-Smirnov test demonstrated the normal distribution of experts’ assessment.

References Acs, Z. J. (1999). The New American Evolution. Z. J. Acs, (Eds.), Are Small Firms Important? Their Role and Impact. Norwell: Kluwer Academic Publishers. Amit, R. and Zott, C. (2001). Value Creation in E-Business. Strategic Management Journal, 22, pp. 493-520. Bailey, J.P. (1995). The Economics of Internet Interconnection Agreements. In L.W. McKnight and J. Bailey (Eds.), Internet Economics (pp. 155 – 168). Cambridge: the MIT Press. Bakos, J. Y. (1991). A Strategic Analysis of Electronic Market places. MS Quarterly, September, pp. 295-309. Bakos, Y. (1997). The emerging role of electronic marketplaces on the Internet. Communication of ACM, 41 (8). Bakos, Y., Brynjolfsson, E. (1999). Bundling information goods: Pricing, profits, and efficiency. Management Science, 45 (12).


171

Baumol, A. and Oates, K. (1975). The Theory of Environmental Policy. New York: Prentice-Hall. Bresnahan, T.F. and Trajtenberg, M. (1995). General Purpose Technologies: Engines of Growth. Journal of Econometrics, 65, pp. 83 – 108. Burn, J., Marshall, P., Barnett, M. (2002). E-business Strategies for Virtual Organizations. Oxford: Butterworth Heinemann. Church, J. and Gandal, N. (1993). Complementary Network Externalities and Technological Adoption. International Journal of Industrial Organization, Vol. 11, pp. 239-260. Economides, N. (1989). Desirability of Compatibility in the Absence of Network externalities, The American Economic Review, 79 (5), pp. 1165–1181. Esser, P. and Leruth, L. (1989). Marketing Compatible, yet Differentiated Products. International Journal of Research in Marketing, 5 (4), pp. 251–70. Evans, P.B. and Wurster, T.S. (1997). Strategy and the Economics of Information. Harvard Business Review, September-October, pp. 71-82. Jorgenson, D.W. and Stiroh, K.J. (1999). Information Technology and Growth. American Economic Review, 89(2), pp. 109-115. Farrell, J. and Saloner, G. (1985) Standardization, Compatibility and Innovation. Rand Journal, Vol. 16, pp. 70-83. Frels, J.K., Shervani, T. and Srivastava, R. (2003). The Integrated Networks Model: Explaining Resource Allocations in Network Markets, Journal of Marketing, 67 (January), pp. 29–45. Gatautis, R., Neverauskas, B., Snieška, V. (2002). Interneto Ƴtaka sandoriǐ kaštams ir tarpininkavimo paslaugoms. Inžinierinơ ekonomika, No.3(29). pp. 89-95. Goldenberg, J., Libai, B. and Muller, E. (2002). Is the Bandwagon Rolling? The Chilling Effect of Network Externalities on New Product Growth, working paper, Jerusalem School of Business Administration, Hebrew University of Jerusalem. Gulati, R. (1999). Network Location and Learning: The Influence of Network Resources and Firm Capabilities on Alliance Formation. Strategic Management Journal, 20, 5, pp. 397-420. Gupta, S., Jain, D.S. and Sawhney, M.S. (1999). Modeling the Evolution of Markets with Indirect Network Externalities: An Application to Digital Television. Marketing Science, 18 (3), pp. 396–416. Hallgren, M.M. and McAdams, A.K. (1995). The Economic Efficiency of Internet Public Goods. In L.W. McKnight and J. Bailey (Eds.), Internet Economics (pp. 455 478). Cambridge: the MIT Press. Helpman, E. and Trajtenberg, M. (1998). A Time to Sow and a time to Reap: Growth Based on general Purpose Technologies. In E. Helpman (Eds.), General Purpose Technologies and Economic Growth. Cambridge: MIT Press. Katz, M. L. and Shapiro, C. (1985). Network Externalities, Competition, and Compatibility. American Economic Review, Vol. 75, pp. 424-440. Kelly, K. (1998). New Rules for the New Economy: 10 Radical Strategies for a Connected World. New York: Penguin.

172

V. Snieška et al.

Klein, S., Langenohl, T. (1995). Electronic Markets: An Introduction, in Schmidt, B. (Ed.), Information and Communications Technologies in Tourism. New York: Springer-Verlag, pp. 262-270. Liebowitz, S. J. and Margolis, S. E. (1994). Market Processes and The Selection of Standards. Working paper. Lorenzoni, G. and Lipparini, A. (1999). The Leveraging of Interfirm Relationships as a Distinctive Organizational Capability: A Longitudinal Study. Strategic Management Journal, 20, 4, pp. 317-338. Malone, T., Yates, J., Benjamin, R. (1989). The logic of electronic markets. Harvard Business Review, May-June, pp.166-172. Nagard-Assayag, E. and Manceau, D. (2001). Modeling the Impact of Product Pre-Announcements in the Context of Indirect Network Externalities. International Journal of Research in Marketing, 18 (3), pp. 203–219. Padmanabhan, V., Rajiv, S. and Srinivasan, K. (1997). New Products, Upgrades, and New Releases: A Rationale for Sequential Product Introduction. Journal of Marketing Research, 34 (November), pp. 456–72. Rayport, J.F., Jaworski, B. (2000). E-commerce. New York: McGraw-Hill. Rosenberg, M. (1998). Chemical Engineering as a General Purpose Technology. E. Helpman (Eds.), General Purpose Technologies and Economic Growth. Cambridge: MIT Press. Rusteika, M. (2001). Lietuviškasis naujosios ekonomikos supratimas. Vilnius. Sarkar, M.B., Butler, B., Steinfield, C. (1995). Intermediaries and Cybermediaries: A Continuing Role for Mediating Players in the Electronic Marketplace, Journal of Computer Mediated Communication (JCMC), Vol. 1, No.3. Shankar, V. and Bayus, B. L. (2003). Network Effects and Competition: An Empirical Analysis of the Home Video Game Industry. Strategic Management Journal, 24 (4), pp. 375–84. Shapiro, C. and Varian, H.R. (1998). Information Rules. Boston: Harvard Business School Press. Wigand, R.T. (1997). Electronic Commerce: Definition, Theory and Context. The Information Society, p. 1-16. Williamson, O.E. (1979). Markets and hierarchies: Some elementary considerations”, American Economic Review, Vol. 63, pp. 316—325. Yoffie, D. B. (1997). Competing in the Age of Digital Convergence. Boston: Harvard Business School Press. Xie, J. and Sirbu, M. (1995). Price Competition and Compatibility in the Presence of Positive Demand Externalities. Management Science, 41 (May), pp. 909– 926.

9 International Outsourcing in the Netherlands

Kees Burger and Rein Haagsma Wageningen University, The Netherlands, E-mail: [email protected], [email protected]

Abstract. An important aspect of globalisation is the strong growth of international trade. What is striking is that the trade in final products is more and more dominated by the trade in parts and components. Because of new production techniques, better means of transportation, and the worldwide breakdown of tariff and non-tariff barriers, industrial firms are increasingly able to outsource part of their production process to cheaper producers or locations in foreign countries. Although this internationalisation of the production process is a universal phenomenon, it seems particularly relevant for the Netherlands. This country is not only one of the world’s most open economies, but is also too small to reap all the scale economies of keeping the entire production chain within the domestic borders. It suggests that the presence and consequences of international outsourcing are especially apparent in the Netherlands. Since outsourcing of production (whether it is international or not) decreases average production costs, in a competitive environment, it leads to lower consumer prices and, therefore, increases economic welfare. Nevertheless, the international debate on outsourcing focuses on its possible negative side-effects. It is argued that outsourcing decreases national employment and particularly harms the labour market position of unskilled workers in Western countries. International outsourcing may also reduce industrial production, and thereby undo the positive externalities that result from a vital industrial sector.In this chapter, we try to quantify the amount of internationally outsourced production by the Dutch industry. Since there are several measurement problems, we apply more than one method and use different statistical data bases. Our focus is on the period 1987-2003. Keywords: International outsourcing, Vertical specialization, Regional outsourcing.

174

Kees Burger and Rein Haagsma

9.1 Introduction The world economy has become increasingly integrated since World War II. New production techniques, better modes of transport and communication, and lower international protection barriers have brought about a significant internationalization of production. More than ever, firms are able to separate their production processes into component stages and to locate these stages in different countries (vertical specialization) or serve foreign markets directly by duplicating entire production processes in these countries (horizontal specialization). With production facilities abroad, either as subsidiaries or through sub-contracting, firms can exploit powerful locational advantages. Moving production to ‘dissimilar countries’ often implies lower labour costs per product and less stringent environmental restrictions and workplace standards. Relocating production to ‘similar countries’ may imply proximity to customers, specialized workers, and contractors. The resulting welfare gains resemble those of international trade and factor mobility. They include lower product prices and greater product variety for consumers, higher business fixed net-investments, and access to specialized knowledge for firms. Despite these gains, the impact of international fragmentation of production is hotly debated in academic and political circles, not only in the U.S. and Canada but also in Western European countries like the Netherlands. Most concerns are expressed about the side-effects that may occur in the sending, highly developed country:1 x Shifting production of unskilled labour intensive products abroad may decrease the employment and relative wage of unskilled workers at home. x Adjustments between and within production sectors may cause capital destruction, loss of expertise, and significant transition unemployment. x The relative decline of manufacturing and growing importance of distribution and services activities may undermine the economy’s potential to increase overall productivity in the long run. x Vital research and development activities may cross the border, eliminating also its positive externalities at home.

1

The implications of internationalization of production are examined from various perspectives in a.o. Feenstra and Hanson (1997), Feenstra (1998), Hummels, Rapoport, and Yi (1998), Venables (1999), Krugman (2000), Jones and Marjit (2001), Kohler (2001), and Egger and Stehrer (2003). For an overview, see the papers gathered in Arndt and Kierzkowski (2001).


175

x Even if nation-wide employment and wage effects are moderate, there may be large disparities on the regional level, especially when labour mobility between regions is relatively low. The article investigates the empirical relevance of international outsourcing in the Netherlands. International outsourcing occurs when some or all stages of the production of a good or service are relocated from the home country. As such it is closely linked to both vertical and horizontal specialization. The vertical aspect of international outsourcing is probably best dealt with by analyzing trade flows and the horizontal aspect by analyzing capital flows (since the construction of complete production lines abroad typically involves direct investments). We will focus on trade flows and only address the role of foreign direct investments in the concluding section. The article pays special attention to the regional dimension of international outsourcing. Although the regional dimension is sometimes acknowledged in the literature, an empirical investigation appears to be absent. We try to fill the gap for the Netherlands with an empirical discussion of the causal relationship between regional developments and international outsourcing and also look at the consequences for regional employment. The plan is as follows. Section 2 examines the production structure of the Dutch economy and tries to identify the production sectors that engage most in international vertical specialization and outsourcing. The manufacturing sector and the transport, storage and communication sector turn out to be the most important. Section 3 focuses on these sectors, and examines some statistical relationships between outsourcing and sector characteristics. Section 4 discusses some regional aspects of international outsourcing. Section 5 gives a summary and briefly compares our main findings with some existing research.

9.2 Vertical Specialization and Outsourcing in the Dutch Economy The story of economic growth is the story of specialization. The traditional notion of specialization is horizontal: firms that produce everything from scratch narrow their range of goods and services. Recent decades show, however, a shift in focus towards vertical specialization. Rather than producing end-products, firms tend to specialize in product parts and components. Production chains are more and more split up into separate stages, thus involving the implicit cooperation of numerous production plants located at home or abroad. As a consequence, firms or subsidiaries of firms

176


increasingly rely on the use of intermediate inputs. In this section, we study the production structure of the Dutch economy and try to identify the production sectors that engage most in international vertical specialization and outsourcing, that is, where intermediate inputs are partly imported from abroad. Table 9.1 presents data on the use of intermediate inputs as a percentage of gross production for several sectors. It is suggested that between 1970 and 2000, almost all production sectors have increased the relative use of intermediate inputs; the only exceptions appear to be agriculture, mining, and financial services. Surprisingly, on the macro level, this upward trend is lost. The Dutch economy as a whole does not seem to exhibit a greater reliance on intermediate inputs between the first and last year of our data sample. The answer to this paradox lies, of course, in the change of the sector shares in total production, with the share of manufacturing falling from 37% in 1970 to 28% in 2000 and the share of financial services rising from 10% to 20%. Thus the sectors with relatively low use of intermediate inputs have gained in importance and the sectors with typically high use of intermediate inputs, notably manufacturing, have declined. Table 9.1 Relative use of intermediate inputs and relative gross production for the main sectors in 1970 and 2000 (in current prices). Source: CBS, The Hague.

Total industries Agriculture, forestry and fishing Mining and quarrying Manufacturing Electricity, gas and water supply Construction Trade, hotels, restaurants and repair Transport, storage and communication Financial and business activities General government Care and other service activities

Intermediates as a Sector share in percentage of total gross progross production duction (as a percentage) 1970 2000 1970 2000 52 51 56 53 6.4 2.9 25 21 0.8 1.6 67 72 37.1 27.8 47 70 2.1 2.4 60 65 9.9 8.8 37 42 12.1 12.8 46 51 6.4 7.4 40 37 9.9 20.4 24 32 8.6 8.1 n.a. 36 6.6 8.6

Looking inside these broad groups, some important sub-sectors stand out.2 Within manufacturing, the strongest increase in intermediate use is in transport equipment, chemical products, and machinery. Within the trade 2

The underlying data are available from the authors.


177

category, the strongest increase is in car repairs and car trade. Within the transport category, the highest growth is in air transport and post and telecommunications. Conversely, the real estate trading and rental sector, which is a dominant sub-sector of the financial services group, shows a decline of the share of intermediate goods. Hence, the longer-term picture is one of increasing specialization of firms, with the consequent increased use of intermediate inputs for the production of final goods and services. Superimposed on this development is the broad structural change of the economy, mainly from manufacturing to services, which tends to stabilize the overall relative use of intermediates. What is the international dimension of this trend towards more specialization? International vertical specialization occurs when a country uses imported intermediate inputs to produce goods and services it later exports. This definition assumes a sequential mode of production with stages located in two or more countries, where one country imports a good from another country, uses that as an input in the production of its own good, and then exports this good to the next country. The significance of international vertical specialization may be indicated by the following index:

I

M int X Y ( M int X ) / 2

where Mint denotes the value of imported intermediates, Y the value of gross production, and X the value of exports (see Hummels, Rapoport, and Yi, 1998). Index I relates the amount of imported intermediates embodied in a country’s exports to the (average) amount of trade. Higher values of I indicate more international vertical specialization. As can be seen, international vertical specialization increases when the fraction of imported intermediates or the fraction of exports (both in terms of gross production) increases. Clearly, I may be calculated for a country but also for separate production sectors. Table 9.2 calculates values of I with respect to several goods and services for the period 1990-2003. Vertical specialization is most prominent in manufacturing. Around 40% of this sector’s total trade consists of, what may be called, vertical trade. Another important sector is transport, storage, and communication with about 20% vertical trade. In both sectors, vertical specialization slightly grows in importance over time.3 This ten3 Data for 2003 are preliminary and therefore less reliable, which might explain the unexpectedly low figure of 0.15.

178


dency is also visible in agriculture and mining. Yet, as before, the growing importance of vertical specialization does not appear at the macro level. Table 9.2 Vertical specialization for the main sectors in 1990, 2000, and 2003, based on measure I (goods and services in current prices). Source: CBS, The Hague (data for 2003 are provisional).


1990

2000

2003

0.18 0.09 0.09 0.38 0.02 0.01 0.06 0.18 0.05 0.01 0.04

0.18 0.14 0.14 0.42 0.02 0.03 0.06 0.22 0.06 0.01 0.04

0.17 0.14 0.14 0.41 0.01 0.04 0.09 0.15 0.06 0.02 0.03

Table 9.3 Vertical specialization in the manufacturing sector in 1990, 2000, and 2003, based on measure I (goods in current prices). Source: Own calculations based on data from CBS, The Hague (data for 2003 are provisional).

Manufacturing Food products, beverages, tobacco Textile and leather products Paper and paper products Publishing and printing Petroleum products Basic chemicals, man-made fibres Chemical products Rubber and plastic products Basic metals Fabricated metal products Machinery and equipment Electrical and optical equipment Transport equipment Other manufacturing

1990 0.36 0.31 0.44 0.41 0.15 0.64 0.45 0.38 0.44 0.45 0.27 0.35 0.34 0.40 0.21

2000 0.39 0.35 0.43 0.44 0.13 0.68 0.48 0.39 0.43 0.44 0.27 0.35 0.37 0.47 0.20

2003 0.38 0.35 0.42 0.41 0.13 0.62 0.48 0.41 0.40 0.45 0.26 0.34 0.33 0.44 0.19

Let us also look inside manufacturing, because this sector increasingly dominates world trade. Table 9.3 shows that (now only with respect to goods) industries with more than average vertical trade are petroleum products, basic chemicals , basic metals, rubber and plastic products, paper


179

products, transport equipment, and textile and leather products. Among these sub-sectors, only basic chemicals and transport equipment show more vertical trade over time. Vertical specialization is closely related to the much debated subject of international outsourcing. The latter is said to occur when one or more stages of the production of a good or service are relocated from the home country. The act of relocation has essentially two effects. First, it raises the use of imported intermediate inputs (except if it only relates to the final production stage). Second, it lowers domestic value added and/or the use of domestic intermediates.4 Hence, international outsourcing tends to raise the ratio:

O

M int VA Dint

M int Y M int

where VA is value added and Dint the value of domestically produced intermediates.5 Note that if firms relocate or duplicate their entire production processes abroad, index O may fall at a sector level, since these firms are likely to be characterized by a high value of O. Therefore, measuring changes in O on a sector level tends to underestimate the significance of outsourcing. Table 9.4 presents values of Ö, again for the period 1990-2003. It appears that outsourcing, as indicated by Ö > 0, especially occurs in agriculture and mining and, since 1995, also in manufacturing, electricity (but see our remark below), and transport. On the macro level, the value of index O is remarkably stable (about 0.17), suggesting that since 1990 the Dutch economy as a whole has not significantly engaged in any international outsourcing.

4

Note the difference between outsourcing on the country level (our focus) and on the firm or sector level. Substituting domestic by foreign intermediates is generally not considered as outsourcing on the firm or sector level, but it is on the country level. 5 Here we use the identity: Y = D int + Mint + VA.

180


Table 9.4 International outsourcing in the main sectors between 1990 and 2003, based on measure Ö (goods and services in current prices), Source: Own calculations based on data from CBS, The Hague (data for 2003 are provisional).


1990-1995

1995-2000

2000-2003

0.00 0.02 0.02 0.00 0.00 0.01 0.01 0.01 0.00 -0.01 0.00

0.01 0.02 0.02 0.07 0.04 -0.01 0.00 0.05 0.00 0.01 0.00

-0.01 0.00 0.00 -0.03 0.04 0.00 0.01 0.02 0.01 -0.01 0.00

Table 9.5 shows that within the manufacturing sector, the picture is very diverse. If we focus on the two sample years 1990 and 2000 (see fourth column in Table 9.5),6 it is seen that the sub-sectors: textile and leather, rubber and plastic, and basic metals indeed seem to increase the relative use of domestic rather than foreign inputs. International outsourcing does seem to be important in the sub-sectors: food, paper products, and notably transport equipment. But note that the performances of the petroleum and basic chemical industries and also the electricity (gas and water) sector are probably biased. These industries share two special characteristics. First, they have the highest labour productivity in the Dutch economy (with very low employment). Opportunities for outsourcing of low-productivity activities thus appear to be scarce. Second, imported intermediates generally consist of raw materials and energy (oil, electricity). The recorded values of intermediate inputs are thus very sensitive to energy price fluctuations, while these may only partially be reflected in end-product prices. All this suggests that the electricity, petroleum, and basic chemical industries are unlikely candidates for outsourcing activities.

6

Our inferences from the period 1990-1995 may be biased by the recession in 1993, and those from the period 1995-2000 by the extraordinary booming years. Also recall the preliminary character of the 2003-figures.


181

Table 9.5 International outsourcing in the manufacturing sector between 1990 and 2003, based on measure Ö (goods in current prices). Source: Own calculations based on data from CBS, The Hague (data for 2003 are provisional).

Manufacturing Food products, beverages, tobacco Textile and leather products Paper and paper products Publishing and printing Petroleum products Basic chemicals, man-made fibres Chemical products Rubber and plastic products Basic metals Fabricated metal products Machinery and equipment Electrical and optical equipment Transport equipment Other manufacturing

1990-1995 0.00 0.04

1995-2000 0.07 0.00

2000-2003 1990-2000 -0.03 0.06 -0.01 0.04

0.01 0.07 0.00 -0.09 -0.06

-0.08 -0.02 0.01 0.78 0.14

-0.06 -0.09 0.00 -1.02 0.00

-0.07 0.05 0.01 0.79 0.08

-0.01 -0.01 -0.09 -0.02 -0.03 0.00

0.03 -0.08 0.02 0.01 0.01 0.02

0.02 -0.08 0.03 -0.03 -0.02 -0.05

0.02 -0.09 -0.06 -0.01 -0.02 0.02

0.03 -0.02

0.04 0.01

-0.07 -0.03

0.08 -0.01

We summarize our major findings up so far. Especially the manufacturing sector and the transport, storage, and communication sector exhibit a high level of international vertical specialization and international outsourcing, which moreover tends to grow over time. Within the manufacturing sector, food, paper products, and especially transport equipment are the most important. Nevertheless, the domestic and international tendencies towards specialization do not appear at the macro level of the Dutch economy. During the selected period, the economy as a whole did neither demand more intermediate inputs nor become more integrated into the world economy in terms of more vertical trade or outsourcing. The answer to this paradox lies in the broad structural change of the economy, mainly from manufacturing to services, which reduces the role of sectors using relatively many intermediates and raises the importance of sectors using relatively few intermediates. Finally, the close correspondence between vertical trade and outsourcing is illustrated in Figure 9.1, which plots indices O and I for the year 2003. It appears that sectors with high vertical specialization also use relatively many imported intermediates. At the bottom end, we find nontradables such as government, care and other service activities and also financial and business activities; at the top end are basic metals, basic

182


chemicals, transport equipment, and textile and leather products. The basic reason for the positive relationship lies in the fact that sectors that import more also tend to export more.7

I - index

2003 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0.00

0.20 0.40 O - index

0.60

Fig. 9.1 The relation between vertical specialization and international outsourcing in 2003, based on measures O and I (data for 2003 are provisional).

9.3 Outsourcing, Productivity, and Employment Above it is suggested that, in the Netherlands, international outsourcing particularly occurs in the manufacturing sector and the transport, storage and communication sector. Restricting our attention to these two production sectors, we will now examine a number of statistical associations between outsourcing on the one hand and certain sector characteristics on the other hand. To begin with, consider that economic growth per capita directly stems from increases in labour productivity. Just as technological innovations typically increase the output per worker by raising the capital-labour ratio, outsourcing increases the output per worker by raising the ratio of imported intermediate inputs to labour. Hence, in this view, outsourcing is just another source for economic growth. Do we observe a positive relationship between outsourcing and labour productivity on the sector level? Assume Mint /X is the same across production sectors, and define a parameter Į = (Mint /X + 1)/2. Hence, Į = 1 if Mint = X and ½ Į < 1 if Mint < X; e.g., Į = 0.78 in manufacturing (2003). Then the relation between O and I can be written as O = ĮI/(1 – ĮI), which is a positively sloped function on [0,1). 7


183

Relative change in Value Added per job 1990-2000

Figure 9.2 sheds some light on this question by presenting data for 13 sub-sectors of manufacturing (excluding petroleum products) and the transport, storage and communication sector. Our index of outsourcing, Ö, is measured along the horizontal axis and the relative change in value added per full-time job along the vertical axis (all in deviation from that of the manufacturing sector (including petroleum products)). Figure 9.2 indeed shows a positive association between outsourcing and labour productivity for the period 1990-2000. The coefficient on outsourcing is statistically significant at the 3 percent level. For the manufacturing sector as a whole, the value of Ö equals 0.06 and the value added per job increased by 0.44 (44 percent).8 Nevertheless, no statistically significant relationship arises from the period 1990-2003, which might be due to the preliminary character of the 2003-statistics.

0.40 0.30 0.20 0.10 0.00 -0.20 -0.10

-0.15

-0.10

-0.05

0.00

0.05

-0.20 Change in O-index 1990-2000

Fig. 9.2 The relation between the change in labour productivity and international outsourcing in the manufacturing sector and the transport sector for the period 1990-2000, based on measure Ö.

Insofar as outsourcing is motivated by cost considerations, it tends to lower labour costs per product or the labour share in value added. Is there a negative relationship between outsourcing and labour costs per product? Figure 9.3 deals with this question in a similar setting as before. Along the vertical axis, we now find the change in the labour share in value added, calculated as the ratio of the gross annual wage and value added per full-time job. Unfortunately, our observation period is limited because the wage statistics are only available since 1995, while the data of 1995 may 8

If we exclude petroleum products, these figures are 0.01 and 0.44, respectively.

184


be biased because of the recession in 1993. Anyhow, contrary to our expectations, Figure 9.3 shows an upward-sloping relation between outsourcing and the labour share in value added for the period 1995-2003. The coefficient on outsourcing is, however, not statistically significant at the 5 percent level (only at 11 percent). A similar exercise for 1995-2000 yields even less conclusive results.

Change in share of wages in Value Added per job

0.20 0.15 0.10 0.05 0.00 -0.05-0.30

-0.20

-0.10

0.00

0.10

0.20

-0.10 -0.15 -0.20 Change in O-index 1995-2003

Fig. 9.3 The relation between the change in the labour share in value added and international outsourcing in the manufacturing sector and the transport sector for the period 1995-2003, based on measure Ö.

It is unlikely that firms are equally able or willing to outsource particular stages of their production processes or to relocate their production plants abroad. The availability of new production techniques, improvements in transport and communication technology, and lower international protection barriers, which has paved the way for the internationalization of production, has probably benefited some production sectors more than others. Maybe most potential benefits accrue to those sectors that used relatively few imported intermediates before the recent wave of technical innovations and trade liberalization. Thus we ask: does a negative relationship exist between outsourcing and the initial level of our index O? Figure 9.4 considers this question, and relates Ö in the period 19902003 to the level of index O in 1990. It indeed shows a negative association between outsourcing and the initial use of imported intermediates. The coefficient on the O-index is statistically significant at the 4 percent level. For the manufacturing sector as a whole, the value of Ö equals 0.03 and


185

the O-index for 1990 is 0.44.9 A similar test for the period 1990-2000 with O for 1990 also gives a negative relation, but not statistically significant (only at 17 percent). Although the data for 2003 are preliminary, these results are in line with the observation by some analysts that outsourcing particularly occurs since the year 2000. 0.30 0.20

O 1990

0.10 0.00 -0.20 -0.10

-0.15

-0.10

-0.05

0.00

0.05

0.10

-0.20 -0.30 -0.40 Change in O-index 1990-2003

Fig. 9.4 The relation between the relative use of imported intermediates in 1990 and international outsourcing during the period 1990-2003 for the manufacturing sector and the transport sector.

Our final question is concerned with the much debated effect on the position of low skilled workers. With Dutch firms facing a higher relative wage for low skilled labour than that found in many other, often less developed countries, the activities that will be outsourced will be those that use a large amount of low skilled labour, such as assembly of components and other repetitive tasks. Moving these activities abroad will reduce the relative demand for low skilled workers at home, thereby affecting their wages and employment. Insofar as wages are rigid and less responsive, which is characteristic for the Dutch labour market, the fall in relative demand will bring about an extra decline in low skilled employment. Note that this effect points at shifts within the sector: it differs from the standard Heckscher-Ohlin effect that applies to trade in final goods and refers to between-sector shifts.10 Do we find a positive relationship between outsourcing and the average skill level of workers on the sector level?

9

If we exclude petroleum products, these figures are 0.02 and 0.40, respectively. Keep in mind that, just as before, it may be that all sectors experience a lower fraction of low skilled workers, but that on the macro-level this fraction is stable. 10

186


To begin with, there is an overall tendency for skill levels to rise in the Dutch industry. Whereas the share of low skilled jobs in manufacturing falls from 40% in 1996 to 39% in 2002, the share of high skilled jobs rises from 15% to 18%. In some sub-sectors these changes are much more outspoken. Unfortunately, examining the possible link with outsourcing is restricted by the degree of detail in the data.11 We are only able to compare the data for six (sub) sectors: manufacturing as a whole and its sub-sectors: food products, textile, paper and publishing, and electrical equipment; and the transport, storage and communication sector. Three levels of jobs can be distinguished: low, middle, and high skilled jobs. The average skill level in a (sub) sector is calculated as 1 times the share of low skilled jobs + 2 times the share of middle skilled jobs + 3 times the share of high skilled jobs. Figure 9.5 relates the change in the O-index between 1995-2003 to the change in average skill level between 1996 and 2002, for the four subsectors of manufacturing and the transport, storage and communication sector (again all in deviation from that of the manufacturing sector).12 The graph confirms the tendency towards higher levels of skills in industries that are more engaged in outsourcing. The average skill level in the sector of manufacturing as a whole rises from 1.75 in 1996 to 1.79 in 2002, a change of 0.04. For the other industries, we find changes of +0.07 in food products, -0.03 in textile, +0.08 in paper and publishing, +0.06 in electrical equipment, and +0.04 in the transport sector.13

In the literature, this phenomenon is dealt with by drawing a distinction between the factor bias and the sector bias of outsourcing. 11 Data on skill levels are only available for some sectors, while these sectors sometimes differ from those for which production and intermediate use data are collected. 12 The value of Ö for the manufacturing sector is 0.04 (or 0.02, if we exclude petroleum products). 13 The graph in Figure 5 is clearly dominated by the fall in average skills in the textile industry. A regression with the change of the fraction of high skilled jobs yields a more balanced picture with an upward slope, but gives no information about any shifts between low and middle skilled jobs. A regression with the change of the fraction of low skilled jobs does not give a clear result and also lacks information about shifts between middle and high skilled jobs.


187

change in mean level of skills

0.10 0.08 0.06 0.04 0.02 0.00 -0.15

-0.1

-0.05

0

0.05

0.1

-0.02 -0.04 change in O-index 1995-2003

Fig. 9.5 The relation between the change in average job skills between 1996 and 2002 and international outsourcing in the period 1995-2003, based on measure Ö.

What can we conclude from these observations? Of course, the inferences from the statistical associations must be treated with some caution: they may be biased by the selection of years, the level of aggregation, the omission of relevant variables, endogeneity, and – perhaps most of all – our measure of outsourcing. Nevertheless, there is some evidence that the sectors that used only few intermediate inputs in 1990, hereafter engaged relatively much in outsourcing and showed relatively high labour productivity and, in certain instances, high job skill levels.

9.4 Regional Aspects of Outsourcing Regional developments can be a cause of outsourcing when regional economic conditions induce firms to seek other sources for the intermediates used in the production process. If such effects are important, they should become visible in a multi-region analysis of the determinants of outsourcing. Regional effects then appear as major determinants in explaining outsourcing. This is one reason for looking at regional effects in this section. Another reason for taking a regional approach is that the developments in a region might well be caused by outsourcing, rather than being a driving force of this. If a region largely depends on some industry that sources out part of the production process, its local economic conditions are probably

188


affected and, in particular, employment may suffer. That is the second issue investigated in this section. Regional data are available for the years 1995-2002. For 40 regions (“COROP gebieden” conforming to NUTS-3 regions) and 38 production sectors the use of intermediate inputs is recorded, as well as gross production and employment. Import and export data are not available at the regional level, so we cannot distinguish imported intermediates from domestic intermediates. This leaves us with two approaches to study the relationship between outsourcing and changes in employment per region. One is to translate the concept of international outsourcing, as defined by the O-index in Section 2, to the regional level. This would beg the question of what is ‘foreign’ to a region in the absence of regional trade data, but our regions are so small that any intermediate inputs easily come from outside the region. Hence, regional outsourcing as defined by the change in the ratio of intermediate use and a sector’s value added comes close to what otherwise would be the change in the O-index. The other approach is to find out what the repercussions of international outsourcing could have been for the regions, on the assumption that production sectors do not behave differently from region to region. What matters then is the weight that a sector has in the region. If the region is heavily dependent on one type of industry, it shares the industry’s fate, and more international outsourcing may lead to less employment in the region. The underlying assumption that sector developments are independent of the region is implicitly confirmed by the regression results below. Starting with the first approach, we define regional outsourcing by sector i in region r as the change in Oireg , r , where

Oireg ,r

Ai ,r Yi ,r Ai ,r

.

Here Yi,r refers to gross production and Ai,r to the use of (both domestic and imported) intermediates by sector i in region r. To see whether regional outsourcing is related to the sector rather than the region, 'Oireg , r is regressed on a set of dummies for the regions (with Amsterdam as default region), the sectors, and the years:

'Oireg , r ,t

D r Ei J t


189

(adding a subscript t for the time dimension). Hence, the change in the index of regional outsourcing by sector i in region r and year t is decomposed into a region-effect (Įr), a sector-effect (ȕi), and a year-effect (Ȗt). The change is calculated from one year to the next. The first regression is for all the sectors aggregated, so for total economic activity (the ȕ-coefficients are therefore dropped). Results are reported in Table 9.A.1 in the Appendix.14 We see that regional outsourcing occurs less over time, but with an exceptional peak in the period 1999/2000 (with 2001/02 as basis, the coefficient is +0.067 in 1995/96 and goes down to +0.024 in 2000/01; the coefficient in 1999/2000 is + 0.070). The only region that shows a significantly stronger tendency is Zeeuws Vlaanderen (DR31), a region dominated by a single chemical-based industry (Dow Chemical). None of the other regions comes out significantly, and the overall development, therefore, has no outspoken regional character. Regional outsourcing on the sector level clearly differs among sectors, but not among regions. In the second regression (see Table 9.A.2 in the Appendix), the ȕ-coefficients are all statistically significant, but the Įcoefficients are not. High values are estimated for food (+0.59), electricity (+0.48), and chemical products (+0.42); somewhat lower values are obtained for electrical equipment (+0.37), petroleum products (+0.36), fabricated metal products (+0.35), post and telecommunications (+0.36), and water and air transport (+0.35); the lowest values are found for textile (+0.29) and land transport (+0.22).15 The differences over the years disappear in this regression, indicating that composition effects (changing weights of the sectors per region) may have caused their significance in the earlier aggregate estimation. In sum, the developments in regional outsourcing are specific for the sector but not for the region, and the extent to which the region is affected depends on its reliance on particular industries. What can be said about the repercussions of international outsourcing for the regions? The regional tendencies towards higher use of intermediate inputs are partly caused by moves towards suppliers located not just outside the region but also outside the country. Although regional import and export data are lacking, we can construct a kind of weighted O-index

14

The Appendix is available from the authors on request. Recall our remark that the performances of the petroleum and basic chemical industries and also the electricity (gas and water) sector are probably biased 15

190


for each region and define international outsourcing by region r as the change in Orint , where

Orint

§ Lir · ¸¸Oi . © r ¹

¦ ¨¨ L i

Here Li,r indicates employment in sector i in region r and Lr total employment in region r. Hence, the Orint -index is a weighted average of the sectors’ O-indices determined at the national level, with weights equal to the sector shares in total regional employment.16 Figure 9.6 shows a map of the Netherlands where, for each of the 40 regions, the relative change of Orint is calculated for the period 1995-2000.17 The darker the colour of a region, the greater is the relative importance of international outsourcing. Coastal areas beat them all: the IJmond region in the middle ( 'Orint 0.22 ), Delfzijl/Eemshaven in the north ( 'Orint

0.11 ), and Zeeuws-Vlaanderen in the south ( 'Orint

0.10 ).

It is of some interest to compare our indices of regional and internaint tional outsourcing, Oireg , r and Or . Since the former depends on the sector, we calculate regional outsourcing for the region as a whole as Orreg Ar /(Yr Ar ), where Ar ( Ai ,r ) refers to the total use of in-

¦ i

termediates in region r and Yr (

¦Y

i,r

) to the corresponding total pro-

i int r

duction. The stronger Orreg and O are related, the more regional outsourcing can be attributed to the international performance of the country on world markets.

16

Using value added as weights gives similar results.

17

More precisely, (Or , 2000 Or ,1995 ) / Or ,1995 .

int

int

int


191

Fig. 9.6 International outsourcing by regions in the Netherlands in the period 1995-2000, based on

'Orint .

Figure 9.7 shows the relationship between Orreg and Orint for 1995. The three regions to the right of the graph are (from top to bottom) Zeeuws Vlaanderen, Delfzijl, and IJmond. Characteristic for these three regions are metal, chemical, and transport industries and major harbour facilities. These regions have more outsourcing abroad than what could be expected on the basis of the ratio of intermediates and value added alone. For the other regions, the relationship takes the form of a more or less straight line, with the Orint -index being about one seventh of the Orreg -index. Its much lower size reflects the moderate importance of imported inputs among the

192


ratio intermediates/value added 1995

intermediate goods. It is striking, however, that the regional differences in the relative use of intermediates are quite closely linked to international developments. 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 0.00

0.20

0.40

0.60

0.80

calculated O-index 1995

Fig. 9.7 The relation between

Orreg (vertical axis) and Orint

(horizontal axis) in 1995.

Above it is indicated that the outsourcing activities of a region are closely linked to the particular production structure of that region. The international competitive environment affects industries and, thereby, regions. The fairly close connection (be it at different levels) between international outsourcing and regional outsourcing is witness to this effect. We now proceed by investigating the relationship between employment and regional outsourcing. Looking at the statistics, we find that employment has increased in all regions, with average growth of employment between 1995 and 2002 being equal to 17%. Yet, there are large differences between the regions, with a maximum of 44% in Flevoland and a minimum of only 1% in Delfzijl. Are these differences in employment growth in any way related to regional outsourcing? Let us first consider this question at the regional level. If we look at the growth rate of employment between 1995 and 2002 and the relative use of intermediates ( Orreg ) in 1995 for corresponding regions, a clear negative relation can be observed, as depicted in Figure 9.8. That is, regions with relatively high use of intermediate inputs show less employment growth. For example, Flevoland has Orreg = 0.78 and a growth rate of 43%, whereas


193

Zuid-West Drenthe has Orreg = 1.28 and a growth rate of only 2%. Looking at regional outsourcing between 1995 and 2002 ( 'Orreg ), however, no significant relationship with regional employment growth exists. Hence, we cannot claim that outsourcing of activities is related to changes in employment at the regional level. 0.500 0.450 0.400 0.350 0.300 0.250 0.200 0.150 0.100 0.050 0.000 0.00

0.05

0.10

0.15

0.20

0.25

O - i nd ex ( r eg ) 19 9 5

Fig. 9.8 The relation between regional employment growth between 1995 and 2002 and 1995.

Orreg in

Is there a connection at the level of the industry in a region? We regress the yearly change in employment in sector i and region r (ǻLi,r) on the corresponding amount of outsourcing ( 'Oireg , r ), combined with region and sector dummies.18 The results show that the coefficient of the amount of outsourcing is insignificant.19 This finding is robust against inclusion of lagged instrument variables, or consideration of 5-year instead of 1-year differences. Hence, there is no immediate link between outsourcing, as measured by 'Oireg , r , and changes in employment. The preceding section mentioned a positive association between international outsourcing and labour productivity. Taking up this lead, we may

18 Observations for 2002 are incomplete, so the regressions are done for 1995/96,…, 2000/01 (with Amsterdam as the default region). 19 The results are available from the authors on request.

194


ask whether outsourcing is accompanied by less employment per unit of output. To this end, we regress the yearly change in the employment/output ratio per region and per sector (ǻ(Li,r/Yi,r)) on the corresponding amount of outsourcing ( 'Oireg , r ), again combined with region and sector dummies. We find a significantly negative coefficient for the outsourcing-variable.20 Outsourcing indeed coincides with less employment per unit of output, but not in a causal way.21 Finally, we summarize our major findings. First of all, there is no evidence that industries engage in outsourcing because of regional conditions. Instead, the outsourcing activities of a region are closely related to the particular composition of the regional production structure. Some regions have more manufacturing and transport activities and, therefore, tend to exhibit more outsourcing. Prominent industries seem to be food, electrical equipment, fabricated metal products, post and telecommunications, and transport by water and air. Changes in regional employment can be largely attributed to changes in the local representation of the various industries. Regions that use relatively few intermediate inputs in 1995 exhibit large increases of employment in later years. No evidence is found, however, that outsourcing activities affect the volume of employment. This holds at both the regional level and the level of industry in a region. Nevertheless, there are indications that outsourcing reduces the number of jobs per unit of output, but this can clearly also be regarded as evidence that outsourcing increases labour productivity.

9.5 Summary and Discussion The domestic and international tendencies towards specialization do not appear at the macro level of the Dutch economy. During the last decades, the economy as a whole did neither demand more intermediate inputs nor become more integrated into the world economy in terms of more vertical specialization or outsourcing. On a sector level, however, we often see the opposite. Especially the manufacturing sector and the transport, storage, and communication sector exhibit a high level of international vertical specialization and international outsourcing, which moreover tends to 20

See the preceding footnote. There may be a degree of simultaneity, however, since the decision of firms to outsource activities may be the same one as to reduce employment. Another regression with the outsourcing-variable of last year as instrument variable generates a non-significantly negative coefficient.

21


195

grow over time. Within manufacturing, food, paper products, and especially transport equipment are the most important in this respect. This picture is only slightly altered when we account for the role of foreign direct investments. Recall that our results on sectors may be biased because we only looked at trade flows. Insofar as international outsourcing implies setting up a subsidiary with a complete production line abroad, the amount of trade is unaffected or may even fall, so that our measures provide an underestimation. Data on net-foreign direct investments (net-FDI) may fill the gap. CPB (2005) observes that net-FDI are especially large in non-manufacturing sectors. A prominent example is financial and business activities (ING, ABN-AMRO, Fortis). Within the manufacturing sector, investments are high in the petroleum and chemical industries, but the capital outflows are almost compensated by direct investments from abroad. SEO (2004) reports that net-FDI are only large in food products and textile. Hence, probably only the textile sector must be added to our list of manufacturing industries that engage most in international outsourcing. For the manufacturing sector and the transport sector, a number of relationships have been examined between outsourcing and specific sector characteristics. Although the outcomes must be treated with some care, it appears that the industries that used only few intermediate inputs in 1990, hereafter engaged relatively much in outsourcing and showed relatively high labour productivity and, in certain instances, high job skill levels. This agrees with the general observation that Dutch industries more and more specialize in high-tech products and import high tech intermediate inputs (see Berenschot, 2004; CPB, 2005). Special attention has been paid to the causal relationship between regional developments and international outsourcing. There is no evidence that industries engage in outsourcing because of regional conditions. Instead, the outsourcing activities of a region are closely related to the particular composition of the regional production structure. Some regions have more manufacturing and transport activities and, therefore, tend to exhibit more outsourcing. Prominent industries seem to be food, electrical equipment, fabricated metal products, post and telecommunications, and transport by water and air. Changes in regional employment can be largely attributed to changes in the local representation of the various industries. Regions that use relatively few intermediate inputs in 1995 exhibit large increases of employment in later years. No evidence is found, however, that outsourcing activities affect the volume of employment. This holds at both the regional level and the level of industry in a region. Nevertheless, there are indications that outsourcing reduces the number of jobs per unit of output, but

196


this can clearly also be regarded as evidence that outsourcing increases labour productivity.

References Arndt, S.W. and H. Kierzkowski (eds.), 2001, Fragmentation: New production patterns in the world economy, Oxford, Oxford University Press. Berenschot, 2004, Aard, omvang en effecten van verplaatsen bedrijfsactiviteiten naar het buitenland, Onderzoeksrapport, Utrecht. CPB (Central Planning Bureau), 2005, Verplaatsing vanuit Nederland: motieven, gevolgen en beleid, CPB Document 76, The Hague. Egger, P. and R. Stehrer, 2003, International outsourcing and the skill-specific wage bill in Eastern Europe, World-Economy, 26, 61-72. Feenstra, R.C., 1998, Integration of trade and disintegration of production in the global economy, Journal of Economic Perspectives, 12, 31-50. Feenstra, R.C. and G.H. Hanson, 1997, Foreign direct investment and relative wages: evidence from Mexico’s maquiladoras, Journal of International Economics, 42, 243-56. Hummels, D., D. Rapoport, and K.M. Yi, 1998, Vertical specialization and the changing nature of world trade, Economic Policy Review (New York: Federal Reserve Bank), June, 79-100. Jones, R.W., and S. Marjit, 2001, The role of international fragmentation in the development process, American Economic Review, Papers and Proceedings, 91, 2, 363-366. Kohler, W., 2001, A specific-factors view on outsourcing, North American Journal of Economics and Finance, 12, 31-53. Krugman, P.R., 2000, Technology, trade and factor prices, Journal of International Economics, 50, 51-71. SEO (Stichting Economisch Onderzoek), 2004, Verplaatsing industrie: hoe erg is het? Onderzoeksrapport, University of Amsterdam, Amsterdam. Venables, A., 1999, Fragmentation and multinational production, European Economic Review, 43, 935-45.

10 Regional Externalities and Clusters: a Dutch Network Case-Study

Roel Rutten1 and Frans Boekema2 1 2

Tilburg University, The Netherlands Radboud University Nijmegen and Tilburg University, The Netherlands, E-mail: [email protected]

Abstract.This chapter contributes to the expanding body of literature on knowledge, learning, innovation, clusters, networks and space. In most publications the main argument is that today’s economy can be best characterized as a knowledge based economy. In other words, learning and knowledge are the key to innovation and improving competitiveness. At the same time, firms depend on collaboration in networks to access knowledge beyond their control. In fact these networks are argued to have an important spatial dimension. One of the main criticisms of these publications is that it may represent some very interesting theoretical views, whereas at the same time the empirical support for these views is very poor. In this chapter we try to answer that criticism as it presents a case study of inter-firm collaboration on innovation in a certain region in the Netherlands. The theoretical part of the chapter is not very elaborate, because we have put an emphasis on the empirical side. In the chapter the relevant literature is only dealt with to a certain extend. We assume that most readers will be more or less familiar with what can be referred to as the established literature. Instead, the chapter focuses on the representation of a case study. In such a way the empirical information will speak for itself and it will be interpreted in the light of the established literature in order to be able to both support as well as to criticize this literature. Keywords: Clusters, Knowledge-based Economy, Innovation, Networks.

198

Roel Rutten and Frans Boekema

In the past decade and nowadays regional networks and clusters, regional competitiveness, regional externalities and regional innovation systems are real topics in regional economic and economic geographical literature. A lot of researchers are focusing their efforts on these kind of subjects and problems. This chapter contributes to the expanding body of literature on knowledge, learning, innovation, clusters, networks, and space. In most publications the main argument of this literature is that today’s economy can best be characterized as a knowledge-based economy. In other words learning and knowledge are the key to innovation and competitiveness. At the same time, firms depend on collaboration in networks to access knowledge beyond their control. In fact these networks are argued to have an important spatial dimension. One of the main criticisms of these publications is that it may present some very interesting theoretical views, whereas at the same time the empirical support for these views is generally poor. In this chapter we try to answer the critical questions by presenting a case study of inter-firm collaboration on innovation in a certain region in the Netherlands. The theoretical part of the study is not very elaborate, because we have put an emphasis on the empirical approach. We assume that most readers will be more or less familiar with what can be referred to as the established literature. Instead, the chapter focuses on the representation of a case-study. In such a way the empirical information will speak for itself and it will be interpreted in the light of the established literature in order to be able to both support as well as to criticize this literature and only briefly introduces this literature below. Instead, this work focuses on the presentation of a case study in order to let the empirical information speak for itself. The empirical information will then be interpreted in the light of the established literature in order to both support and criticize this literature.

10.1 Introduction An expanding body of literature aims to identify success factors for regional economic development in general and clustering, networking and innovation in particular. Most of these factors relate to the firm, the management or the production environment. Unfortunately this literature remains, however, inconclusive on the relevance of such factors. There is no doubt that only very few are self-sustainable in maintaining and developing the (relevant) knowledge base. Almost all firms rely on external sources of information, data and knowledge. The economic geographic literature seems to be rather clear in this respect. If new knowledge is disseminated among proximate actors


199

more efficiently, firms tend to cluster in areas with distinctive knowledge bases. In other words, (geographical) space, more precisely the region may therefore become a success factor in maintaining and improving the knowledge base of firms. Nevertheless, we have to emphasise that the overall literature remains inconclusive as to whether regions matter in the context of the Netherlands of limited size and homogeneity with respect to general business conditions. The central question, as to whether either specialized or diversified local and regional production structures will favor the dissemination of knowledge, remains still without a clear answer. Specialisation implies that knowledge is industry-specific and tends to spill over between firms within the same industry. In fact this can be called the Marshallian hypothesis, as firms in certain geographical areas with production structures specialized towards a particular industry may therefore be more succesfull in economic performance and innovation. If regions have a diversified production structure, the Jacobian diversification hypothesis however would lead to the conclusion that one might expect more innovative firms and a better economic performance. In the literature on the knowledge-based economy, five basic elements can be identified: 1. learning, 2. innovation, 3. networks, 4. knowledge, and 5. space. In this chapter we will not discuss the established literature. One can find such a discussion in many recent publications (e.g., Grabher 1993, Nonaka and Takeuchi 1995, Morgan, 1997, Uzzi, 1997, Storper 1997, Maskell et al 1998, Boekema et al 2000, Rutten 2003, etc.). What matters here is a general picture, a birdseye view of the established literature on the bases of these five distinguished elements. In Figure 1, the LINKS pentagram is presented. The word LINKS is an acronym of the first letters of each of the pentagram’s five elements and it indicates that these elements are linked to each other. There is no doubt that Knowledge is at the core (or beating heart) of the knowledge-based economy. In fact economic development can be explained in terms of the ability of firms, clusters and networks, and regions to create, exchange, and use knowledge to their advantage. Learning, which means a process of transformation through knowledge, leads to innovations. And finally this will create a competitive advantage. In a modern knowledge-based economy, innovations are most durable when they are based on knowledge, in particular tacit knowledge that does not easily slip away to competitors. Consequently, (interactive) learning is a crucial process in the knowledge-based economy since it transforms (tacit) knowledge into innovations. Many examples can be found to illustrate what is meant by this. In many cases the very succesfull story of Silicon Valley has been told in this respect. The established literature also argues that the

200


knowledge-based economy is a network economy, in other words many relations between relevant actors. Networks are important for more than just one reason. Firstly, and this is an important issue, they are a means for firms to access knowledge which is beyond their own control. Secondly, networks are helpful in the innovation process as they allow firms to specialize. In other words, through networks, firms can make new combinations of knowledge from various firms thus making innovation a network effort. Thirdly, and closely connected to the previous two reasons, networks are helpful for learning as firms are exposed to more and richer flows of information. There is no need to stress that the possibilities for flows of innovation and data have expanded very much by the digital revolution of internet. The final element of the knowledge-based economy is space. Space refers both to proximity as well as to distance Space, too, is connected to all the other elements as mentioned before. It is connected to knowledge through a phenomenon that is known as the geography of knowledge. This phenomenon says that to exchange tacit knowledge requires face-to-face communication, which means the need for proximity. In other words, people have to actually meet each other to exchange tacit knowledge. Consequently, distance between them may be a handicap. So, the more tacit the knowledge, the more proximity becomes and advantage with regard to an efficient transfer of information and knowledge exchange. Moreover, knowledge exchange in networks very often works better when the firms in the network are located close to each other and share a common background. One can put it in this way; physical and cultural proximity facilitates the exchange of knowledge. Firms are also more likely to produce innovations in regions where regional competencies and institutions have been developed that support innovation. Finally, learning can result in the development of localized capabilities, the merits of which will benefit firms in regional networks. One can find several excellent cases in a Scandinavian setting by Maskell et al (1998). The above section can, more or less, be considered as a kind of a synopsis of the established literature in this field of research. Within the limited space allowed, this work does two things. First, it discusses the process of knowledge creation and, second, it considers the role of space in this process. In other words, this work highlights some relations of the LINKS pentagram, i.e., the relations between knowledge, networks, and space. In other words, one of the most important questions “Does space matter” could be emphasized once again. There is no doubt that this question is considered to be a very relevant one in spatial sciences.


201

LEARNING

crea tion

regional competencies

and institutions

of innovation

know ledg e

y of raph geog

ti eti mp co

ble ina sta su

ve

pr ox im ity

access to knowledge

Fig. 10.1 LINKS pentagram: Web of relations Space. Source: Roel Rutten (2002).

ing learn

cu ltu ra l

ge nta va ad

INNOVATION spec ializ ati o n and com bina tion

ph ys ica la nd

KNOWLEDGE

pr oc es so f in no va tio n

of

e ledg know

SPACE

rgies syne

ing lop ve e d

d ze ali loc

s itie bil pa a c

NETWORKS

Learning-Innovation-Networks-Knowledge-

10.2 The “KIC- Project”: a DUTCH Regional Case-Study This study focuses on the so-called KIC (Knowledge Industry Clustering) project of Océ, one of the leading firms in the Dutch manufacturing industry. The history starts in 1993, when Océ initiated the KIC project in order to involve suppliers more closely in the product development process. The main objective of Océ was to put more emphasis on the ‘front side’ of the product development process, i.e., research and development, as that is where it creates its added value compared to, for example, competitors like Xerox and Canon. The ‘back side’ of the product development process, i.e., the engineering, could then be outsourced to suppliers. This would allow Océ to make better use of the manufacturing knowledge of its suppliers and thus develop better products. In other words, The KIC project, placed substantially heavier demands on suppliers as they were now asked to solve engineering problems for Océ, while, formerly, they were merely asked to manufacture as efficiently as possible. The benefits for the suppliers in the KIC project lay in the fact that they could upgrade from ‘jobber’ to ‘co-developer,’ or even ‘main supplier.’ In doing so they would learn how to engineer, that is, to compete on knowledge rather than costs (i.e.,

202


efficient production). In sum, the KIC network was a big step away from the traditional arm’s-length buyer-supplier relations as it required the companies involved to work intimately with each other for several years – this means the time it took to complete an engineering project. This also changed substantial dependency relations as Océ became more dependent on the knowledge of its suppliers. In this regard, the KIC project seems to be a good example of the knowledge-based economy at work. The KIC project involved almost 40 suppliers who were organized in approximately 20 clusters. On average, a cluster was composed of two to three suppliers and a representative from Océ. Each cluster was focused on a specific engineering project for Océ. The clusters were operative from 1994 to 1998, when the KIC project ended. Clusters worked for two years on average on their projects, which meant that the first clusters were already dissolved by the time the last clusters were formed. For this study, 14 suppliers in 10 different clusters were interviewed in depth. In addition, 10 representatives of the Océ R&D department, who were involved in the clusters, were interviewed, as well as four representatives of the Océ management and purchasing departments. For each of the 10 clusters involved in this study, the Océ representative and at least one supplier were interviewed.

10.3 Creation of (Inter-Firm) Knowledge This case-study focuses mainly on the engineering phase of the product development process. This phase starts mainly when the functional specifications of a product or module have been determined. The following example might illustrate this; the stapler inside the copier has to be able to apply a certain number of staples per minute and must be easy to reach for users when they want to put a new supply of staples in the machine when old supply is used up. These conditions determine a number of functions and design specifications which are the input for the engineering process. In the KIC project, the engineering was done by teams (or clusters) of suppliers who would work on the engineering of a complete module whereas, formerly, Océ engineers would do most of the engineering themselves and then have individual suppliers work on the final engineering of separate parts. In other words, the KIC approach thus involved the suppliers in the engineering process to a far greater extent and called on their skills and expertise to design something new rather than on their ability to manufacture a given part as efficiently as possible. The idea behind the KIC project was to involve suppliers in that part of the product development process


203

where they can offer most added value because of their specialization. The resulting dependency of Océ on the suppliers is considered a small price well worth paying if the overall quality of the product development process improves. Although Océ formally was in charge of the KIC network, it left most of the operational management to the suppliers in their various clusters. The objective was to let the suppliers take the initiative so as to challenge them to use their skills and expertise in the engineering process. In practice, however, the Océ engineers played a much more prominent role as, in many cases, Océ had overestimated the suppliers’ engineering capabilities. The collaboration between Océ and the suppliers was based on a contract that reflected Océ’s position in the KIC network as a central and dominant but not dominating actor. In fact one might conclude that the technical aspects of the clusters illustrate how the engineering – and, therefore, the knowledge exchange – took place. For reasons of confidentiality, it is not possible to go into detail about the engineering process, but a brief description of the technical objectives of two of the clusters nevertheless gives an impression of what the KIC project set out to do. The aim of the discussion is to show the (technical) complexities of the engineering, as these had implications for the process of knowledge exchange between the companies involved. x Cluster C: TOSUP (toner dosage system). The objective of this project was the functional development of a new toner dosing system for a new generation of color copiers. The new color copier has seven pictureforming units that must each be supplied with its specific toner. The units hold a small supply of toner and must, therefore, constantly be resupplied during the copying. The dosing mechanism is the heart of the system. The purpose of this mechanism is to supply the picture-forming units with a stable amount of toner from the reservoirs. During this dosing, the toner is not to be thermally or mechanically disturbed as this will change the structure of the toner which, in turn, could cause malfunctions in the picture-forming units. x Cluster F: glass transfer cylinder. The glass transfer cylinder is applied in the warm process functions of the copier. Flanged bearings attach the cylinder to a frame in the copier. Inside the glass cylinder, two halogen radiators are placed as a heating device which can produce various temperatures when the machine is turned off. When the machine is operative, rollers are pressed onto the cylinder, and the complete toner image is subsequently transferred to the receiving material (i.e., a sheet of paper). The objective of this cluster was to develop a new, cheaper, and more reliable production technology for glass transfer cylinders with improved specifications for the new generation of color copiers.

204


These clusters in fact emphasize the role of different disciplines within the whole process In the next section we will elaborate on the role and meaning of a multi-disciplinary approach.

10.4 A Variety of Disciplines One might conclude from the above discussion that it is clear that each cluster involved several ‘disciplines.’ In the manufacturing industry, a discipline is not understood in terms of a scientific line of inquiry and teaching but as a set of related activities in the field of Research and Development, engineering, and production. Examples of disciplines are sheet metalworking, electronic engineering, glass technology, and, increasingly, ICT. Whereas suppliers were used to working with a mono-disciplinary approach – i.e., to do a limited amount of re-engineering of an individual part that Océ had designed for them – they now had to adopt a multidisciplinary approach to engineer a complete module based on functional specifications. This involved a substantial degree of knowledge exchange between the suppliers as they had to develop some sort of understanding of each other’s disciplines, or the suppliers would not be able to develop a joint solution to the engineering problem presented to them, which, after all, was the objective of the KIC project. In order to arrive there, each cluster had a lead supplier who was responsible for the management of his cluster. Among others, lead suppliers had to ensure that knowledge exchange took place between the engineers involved in their cluster. In addition, every cluster made some sort of distinction between the management and the engineering levels. Basically, the management level was responsible for the formal, contractual side of the projects whereas the engineers focused on the technical contents. It is important to keep in mind, though, that informal communication often bypassed the formal communication channels. Although the second-tier supplier, i.e., those who were not lead suppliers, were not supposed to frequently communicate with the Océ engineer of their cluster, Océ engineers often found themselves heavily involved in the engineering process. A company’s ability to master a discipline is – at least related to a knowledge perspective – at the heart of its competitive advantage. The more a company knows about its discipline and the better it is able to translate that knowledge into concrete solutions and products, the stronger the competitive position of that company will be. Mastering a discipline thus involves creating a substantial body of tacit knowledge. As one supplier observed, “induction is a specialization that cannot be acquired from


205

reading books” (Derix 1998: 44). The difficulty is clear: working together on engineering requires companies to exchange parts of their respective tacit knowledge. The question is, how did they do so? Formal meetings, according to Nonaka and Takeuchi (1995) are hardly the place for exchanging tacit knowledge. In any case, most knowledge was exchanged on the ‘workfloor,’ that is, in the interaction between the engineers while they were working on the project and the informal day-to-day communication between team members working on the same job. The engineers used various mechanisms to interact with each other, ranging from telephone conversations and electronic data exchange (such as e-mail and facsimile messages) to face-to-face communication where they would often sit together around a prototype of their module to experiment with it and to demonstrate what it could and could not do. The fact that engineers could bypass the formal communication channels and directly communicate with each other resulted from the shortening of communication lines in the KIC project. This, the respondents argued, stimulated the creativity that the engineers needed when they have to solve technical difficulties. The ease with which the engineers could discuss matters with one another outside the formal communication channels indicates that the boundaries between the organizations in the various clusters were not an obstacle to knowledge exchange. This lead respondents to argue that, in the KIC project, more knowledge was exchanged than in earlier cases.

10.5 Tacit and Codified Knowledge Finally, it is important to notice that many companies acknowledged the need for face-to-face communication with regard to knowledge exchange. They argued that electronic communication is ideal to exchange “most of the technical information,” they argued, but in order to exchange “knowledge and ideas it is necessary to meet face to face.” 1 Respondents were not familiar with the terminology of this study, such as tacit and codified knowledge. Yet, their answers speak for themselves. Océ engineers argued that engineering is “a creative process; you have to have your nose on the machine,” and that “you cannot put everything in writing.” Clearly, Océ engineers considered tacit knowledge essential for the success of their projects. Evidence of tacit knowledge being exchanged in the KIC project can 1

Answers from respondents are placed between double quotation marks. However, these are not their exact answers. As the interviews were in Dutch, the original answers have been translated and stylized in order to maintain a readable and academic text.

206


be found in the suppliers’ answers, too. One supplier, for example, argued that “a physical confrontation with the product you are working on is important to understand it, to demonstrate how it works.” The respondents realized that codified knowledge (the technical information) requires different exchange mechanisms than tacit knowledge (the ideas). The suppliers in this study often referred to the latter as “looking in each other’s kitchen.” This shows that engineering is about exchanging and creating tacit knowledge. As much as possible, the companies in the KIC project tried to eliminate barriers that hinder the free flow of knowledge between organizations. They did so, first, by separating engineers from managers in order to allow the engineers to concentrate on the technical side of the projects without managers looking over their shoulders. Secondly, the model of communication ensured that the exchange of (tacit) knowledge between organizations could take place without organizational structure being a problem.

10.6 Some Preliminary Results First we would like to stress that it is beyond the scope of this study to assess the (technical) qualities of the modules engineered in the KIC clusters and compare them to modules in which no suppliers were involved. Therefore, the analysis will have to be based on opinions of the respondents. A little more than half of them (15, or 54 percent) argued that involving suppliers leads to better engineering. A further 11 respondents (or 39 percent) argued that, in the case of the KIC project, this may not have been the case but is certainly possible in future projects. The suppliers, they argued, just needed more time to upgrade their skills. Only two respondents (or seven percent) said that Océ might just as well do the engineering without the suppliers and reach the same (or even better) results. In general, this supports the theoretical argument that involving external knowledge is beneficial to a company’s innovation efforts. Looking at the different categories of respondents, however, there are some important differences. The Océ engineers were relatively skeptical about the role of suppliers in the engineering process. Their skeptical attitude may, of course, be due to the engineering outcomes actually not being better than if Océ had not involved the suppliers. On the other hand, this was a new experience for the Océ engineers as the suppliers now entered into what, till then, had been their exclusive domain. The suppliers, of course, take a different approach to engineering than the Océ engineers. This is why they were invited to participate in the KIC project in the first place. Whereas the Océ engineers


207

look at engineering from a conceptual point of view, i.e., finding the optimal solution, the suppliers look at it from a production perspective. They are used to thinking in terms of how to manufacture something as efficiently as possible, and that is not necessarily the same as looking for the optimal solution from a technical point of view. In other words, the skepticism from the Océ engineers may also have been caused by a different perception of the engineering process. The suppliers overwhelmingly believed in Océ’s new approach to engineering: 71 percent (10 respondents) said that engineering in the KIC project had led to better outcomes while the remaining 29 percent (four respondents) believed it could lead to better results in the future. This outcome, too, is not really surprising since the suppliers were asked to assess their own performance. Still, from their perspective, the KIC project did produce better results because now they had the chance to voice their views from an early stage on. Océ engineers, for example, had never considered the possibilities, limitations, tolerances, etc., of a supplier’s machinery when designing a part. However, the machinery has to manufacture the parts. In other words it became clear that this kind of re-engineering could now be avoided.

10.7 Interfirm Relations If companies do work well together this will certainly have an impact on the knowledge exchange in networks. This boils down to the issue of trust. In the literature, trust is often associated with long-term relationships (cf. Granovetter 1985, 1992, and Williamson 1993, 1999). However, the relations in the KIC project show a peculiar absence of long-termism. In 50 percent of the, for example, no previous relationship existed between the suppliers and the Océ engineer and in only 14 percent of the cases had the suppliers ever worked with each other before (Rutten 2002). Considering the risky and uncertain nature of the KIC project, one would expect Océ to collaborate with established partners and to select suppliers who, also among themselves, have a history of favorable work relations. Williamsonian theory, at least, suggests this (Williamson 1993, 1999). Given the absence of long-term relationships, other mechanisms, apparently, provided of the certainties that companies need to work trustfully with each other. Doing a bad job in KIC, for example, would certainly have compromized a supplier’s position in the wider context of the sector and regional networks it is embedded in. Moreover, even though the KIC contract did not enforce the suppliers to perform well, as it did not provide for sanctions, the suppliers, by their strategies, were committed to making KIC a success. In

208


other words, similar strategies point at similar interests. To pursue their similar interests the companies involved had no other choice than to work trustfully with each other. In other words, the fact that companies depended on each other and that they needed each other’s knowledge to make their projects a success (see the discussion on engineering) seems to have ensured their trustful working together. Ultimately, however, the behavior of companies in a network is largely determined by the way they perceive their relationships. Perceptions have to be used with great caution in scientific analyses. Yet, they do provide valuable background information and coloring. Océ respondents, for example, argued that they tried to make the relations with the suppliers “as open as possible” and that lines of communication where short. Moreover, they said that, on the engineering level, there was absolutely no patron/subordinate-like relation between Océ and the suppliers. Formally, such a relation did exist, but in practice there was little sign of it, according to the Océ engineers. The suppliers’ perception in this respect corresponded with the Océ view. One supplier argued, for example, that “in the clusters, they had the freedom to do what was necessary to achieve the best possible technical result.” This supplier actually says two things: firstly, that he had more or less carte blanche in technical matters, i.e., the engineering part of the KIC project, but, secondly, that Océ had a big say in non-technical matters. The supplier in this case appreciated this situation, as he was commending the way he did business with Océ on “equal terms.” Other suppliers, too, argued that “we had a lot of freedom,” and that “to a large degree, we could make our own decisions.” The answers show that both Océ and supplier respondents perceived the relations in the KIC network as favorable. They also show that both sides appreciated these favorable relations and acted trustfully. The construction of the clusters with their short lines of communication and the absence of rival firms supported the development of trust in the relations. So one might conclude that the conditions for knowledge exchange were therefore favorable and very efficient.

10.8 The Regional Impact Regional externalities in fact deal with the meaning of regional factors, which means the regional production environment. Working with regional suppliers on engineering matters is a strategic choice for Oce. Océ prefers working with regional suppliers as this makes communication easier. Remember that the KIC project was intended to upgrade regional suppliers in order to become competent partners for future engineering projects. In


209

other words, Océ focuses strongly on regional-based knowledge in the engineering process. Data from Océ’s purchasing department confirm this strategic focus. The total purchasing value of Océ increased from 77 million euros in 1988 to 235 million euros in 1996. This confirms the trend towards involving suppliers more strongly in Océ. Of the purchasing value, 34 percent (or 26 million euro) was allocated to suppliers in the southeastern Netherlands in 1988. This regional share increased to 45 percent (or 105 million euros) in 1996. At the same time, Océ’s purchasing value in the rest of the Netherlands remained fairly constant: almost 21 million euros in 1988 versus 23 million euros in 1996. However, in relative terms, the share of the rest of the Netherlands dropped from 27 percent in 1988 to 10 percent in 1996. The relative shares of Europe and the rest of the world remained constant. In 1988 Europe accounted for 32 percent of Océ’s purchasing value against 35 percent in 1996. The figures for the rest of the world were seven percent in 1988 and 10 percent in 1996 (see Figure 2). These figures show that suppliers in the region (i.e., the southeastern Netherlands) have become significantly more important in recent years, mainly at the expense of suppliers in the rest of the Netherlands. This Figure shows, firstly, that purchasing became more important for Océ and, secondly, that Océ focuses more strongly on its home region. The share of regional suppliers is the only one to show a significant increase (in relative terms). The suppliers, too, predominantly have a regional focus. For the majority of them, the southeastern Netherlands are (one of) the favored area(s) to look for buyers, suppliers, and engineering partners. Nine out of 14 suppliers (or 64 percent) have a high regional orientation, which means that they have few buyers, suppliers, or engineering partners outside the southeastern Netherlands. A further two of the KIC suppliers (or 14 percent) have a moderate regional orientation. This means that, in all, 78 percent of the KIC suppliers involved in this study consider their home region the focal point of their business activities. In short, the available empirical data clearly show that the companies involved in this study have a regional orientation with respect to engineering.

210


250

million euro 200

150

100

w orld

50

Europe Netherlands region

0

1988

1993

1996

Fig. 10.2: Allocation of purchasing per spatial scale; (global/world-international/Europenational/Netherlands-regional/North-Limburg-local/Venlo).

10.9 The Spatial Relevance of Knowledge Transfer: Proximity and Distance An expanding body of literature argues that spatial proximity facilitates the exchange of embedded knowledge through a mechanism that was referred to as ‘the geography of knowledge.’ In order to establish whether this mechanism played a role in the KIC project, it must be determined if the respondents found that spatial proximity facilitated the communication between them. The longest distance between any of the companies involved in the KIC project (not just those involved in this study) was 70–75 kilometers (44–46 miles), which corresponds to about one hour’s driving time. Therefore, it is justified to say that the relations in the KIC network were proximate relations. One would therefore expect to find some evidence in support of the geography-of-knowledge theory. This support comes from Table 10.1, in which the perspectives of the respondents in this study regarding spatial proximity and face-to-face communication are presented. Spatial proximity is an advantage with respect to knowledge exchange


211

when actors experience that the communication between them is easier when they are located close to each other. Table 10.1 presents these experiences for the respondents in this study. It shows, for example, that all of the 14 Océ respondents found that spatial proximity facilitated communication in the KIC project. Of the suppliers involved in this study, Nine out of 14 (or 64 percent) also found that spatial proximity facilitated communication. Only one supplier (or seven percent) did not have this experience. For four suppliers (or 29 percent) it could not be established how they thought spatial proximity affected communication in the KIC clusters. Taken together, this means that 23 out of 28 respondents in this study (or 82 percent) found that spatial proximity facilitated communication. Important as they may have found spatial proximity for communication between them, respondents did not think it was necessary. Only two out of 28 respondents (or seven percent) found that spatial proximity was necessary in engineering projects like KIC, whereas the majority (22 respondents or 79 percent) felt that, if necessary, communication could also be achieved over long distances. Finally, the respondents were asked whether they felt that face-to-face communication was important. Considering that the knowledge that was exchanged in the KIC project contained a significant portion of tacit knowledge, one would expect to find that respondents attached importance to face-to-face communication. This proved to be the case, as 25 out of 28 respondents (or 89 percent) found face-to-face communication to be important in the KIC project. The opinion of three respondents (or 21 percent) on this issue could not be established. From the data presented in this table it shows that, first of all, the respondents found that face-to-face communication was important in the KIC project. This is a significant outcome because, if they had held another opinion, there would be no case to support the geography-ofknowledge theory. Moreover, the fact that respondents found face-to-face communication important indicates that they did indeed exchange tacit knowledge. Secondly, the data in Table 10.1 show that the respondents found that spatial proximity facilitates knowledge exchange, confirming the geography-of-knowledge theory for the KIC project. However, they did not find that spatial proximity was necessary. This supports the assumption that the geography-of-knowledge theory does not dictate that actors must be located close to each other but that, though preferring regional partners, they look for competent partners in the first place. A look at the answers from the respondents during the interviews provides some relevant coloring for these data.

212


Table 10.1: Perspectives on spatial proximity and face-to-face communication.

spatial proximity facilitates comis necessary munication yes no n.a* yes no n.a* Océ N percent Suppliers N percent Total N percent

face-to-face communication is important yes no n.a.

14 100

0 0

0 0

1 7

13 93

0 0

14 100

0 0

0 0

9 64

1 7

4 29

1 7

9 64

4 29

11 79

0 0

3 21

23 1 4 82 4 14 *n.a. means not available.

2 7

22 79

4 14

25 89

0 0

3 11

Respondents associate spatial proximity with short lines of communication and easier meeting opportunities. But, as the data show, they do not put spatial proximity first. One of the suppliers, for example, argued that “you have to find a like-minded partner, in which case proximity is of secondary importance.” With regard to face-to-face communication, the respondents argued that modern electronic communication can never replace it because “you have to look each other in the eye”, and, “you have to ‘taste’ a relationship”. In other words, the suppliers valued the short distances as it made the ‘social aspects’ of the communication easier. It is precisely these social aspects that are crucial to the exchange of tacit knowledge.

10.10 Theoretical Implications Here we will discuss some of the theoretical implications of the KIC project. The idea is to interpret and explain the phenomena observed in the KIC project on the basis of what was referred to as the established literature. Given the limited space, the below discussion will not go into detail with regard to the established literature. Instead, it will highlight relevant issues that can explain the facts and data presented in the previous section.


213

10.11 Competitiveness and Knowledge Competitiveness between firms depends on the combination of internally and externally tacit knowledge in unique (network) competencies, so the literature argues. The previous section showed that the companies involved in the various KIC clusters exchanged their respective tacit knowledge in order to develop new modules. The purpose of the collaboration effort was to create something (the modules) that was beyond the capabilities of the individual firms. Furthermore, it was demonstrated that the work of the clusters cannot be easily copied by other teams of suppliers precisely because it depends heavily on the tacit knowledge of the companies involved. In other words, the suppliers in the respective clusters have developed competencies that are highly specific to their clusters, which were used to create unique modules. The KIC case thus supports the assumption in the literature that companies collaborate in networks in order to develop unique, network-specific competencies. Whether they actually lead to competitive advantage cannot be established within the context of this case study, since it focuses on the collaboration effort itself, not on its consequences.

10.12 Knowledge Exchange and Trust The key element in inter-firm relations is trust and it can be defined as the confidence that actors will work for mutual benefit (Rutten 2002). Obviously, corresponding strategies indicate that the partners involved are committed to the same objectives, as was already discussed in the above. Another important factor in the KIC network was that the companies depended on each other to get the job done. In other words, inter-firm dependency was very strong. In addition, the companies in the various clusters all had complementary skills and knowledge, they were no competitors. Thus, the companies involved, at least, acted as if they trusted each other. They assumed that their network partners could be trusted because of the abovementioned corresponding strategies, common objectives, and complementary skills. What is relevant, then, is that companies are willing to trust each other if certain conditions are met and that these conditions do not involve a history of working together. Whereas, traditionally, theories (e.g., the Williamsonian and Granovetterian approaches) have mainly looked at the past to explain trust, this work argues that the focus should (also) be on the future.

214


10.13 Regional Partners and Knowledge Exchange From the last sections one might conclude that the KIC project is an interesting case to demonstrate the relation between knowledge exchange and spatial proximity, as the literature suggests there is. The findings of the KIC project give strong support for the theory of the ‘geography of knowledge.’ The value of the present study, therefore, is that it clearly demonstrates that space matters for companies that are collaborating on innovation. For many years, science has argued that this is the case but it rarely supported its claims with empirical data on knowledge exchange between companies. The present study shows that companies favor collaboration with regional partners in the engineering phase of the innovation process. In other phases of the innovation process, however, they may come to a different conclusion. It is here that the relevance of the present study for innovation and regional economic theory becomes clear, as it demonstrates that the focus must be on knowledge, not on proximity or space. Proximity is the outcome of a trade-off, knowledge is the substance. Companies ask themselves whether or not it is desirable and possible to involve regional partners in their knowledge-creation effort. In the engineering process, relevant knowledge is specific but not necessarily unique. It is advanced but not necessarily high tech. This implies that the number of potential partners is substantial – though not abundant. In other words, it will most likely be possible to find competent partners nearby. Furthermore, engineering requires face-to-face communication between network partners for a prolonged period of time, which makes collaboration with regional partners desirable from an efficiency perspective. These characteristics – regional availability of relevant knowledge, and the advantages of proximate relations – enable companies to choose regional partners. The answer to the number one question in regional economics, i.e., does space matter? thus depends on a trade-off between the regional availability of knowledge on the one hand and the benefits associated with collaboration with proximate partners on the other hand. As argued, the present study demonstrates that, in the engineering phase, this trade-off is made in favor of proximate relations. Furthering an understanding of the relation between innovation and space requires this trade-off to be explored for other phases in the innovation process as well. We will elaborate on this in the next section.


215

10.14 Knowledge Creation in Networks Explained The literature on innovation is impressive and in the established literature it is argued that the intangible side of the innovation process is the key to firm competitiveness. The present study is no exception given its focus on tacit knowledge. Recognizing that tacit knowledge and tacit knowing are at the heart of the explanation, however, requires a fundamental reconsideration of the nature of knowledge, innovation (i.e., knowledge creation), and space. The concept of tacit knowledge holds that some forms of knowledge have no meaning and cannot exist outside a personal or organizational context. In other words, such knowledge is always embedded in a social context. This, in turn, implies that tacit knowledge is socially constructed knowledge and that, to understand the process of knowledge creation, the focus must be on the social context in which knowledge is created. It is, therefore, perhaps better to speak of embedded knowledge. In the present study, this context was the inter-firm teams, or clusters, at the heart of the KIC project. In ‘The knowledge-creating company,’ Nonaka and Takeuchi (1995) argue that the creation of new knowledge takes place in project teams. They do not elaborate on this but von Krogh et al. (2000) argue that the creation of tacit knowledge takes place in productive work communities that are based on social processes. These work communities, or “micro communities of knowledge,” are small groups of five to seven people2 who maintain a “dense network of relationships” (von Krogh et al. 2000: 14). The creation of tacit knowledge in these micro communities takes place through face-to-face interactions which are facilitated as the team members gradually get to know more about each other’s personalities. These micro communities develop an identity of their own as well as a shared base of tacit knowledge. It takes little imagination to see the KIC clusters as examples of these micro communities. The present study demonstrated that tacit, or embedded, knowledge was actually created in these clusters and it highlighted some of the social issues related to collaboration in small teams. In other words, this is where theory and practice meet. Support for the small-teams approach also comes from other authors such as Johannessen et al. (2001) who argue that the creation of tacit knowledge

2

The number of people in these small groups is irrelevant. What matters is that the emphasis is on small groups as this is an important parallel with the KIC project.

216


takes place in ‘apprenticeship teams,’ and Judge et al. (1997) who speak of ‘goal-directed communities.’3 What does this mean for the state-of-the art concerning the present debate? As far as knowledge creation is concerned, the theories of, for example, Nonaka and Takeuchi (1995) and von Krogh et al. (2000) go a long way in explaining what takes place when individuals in teams create embedded knowledge. The problem, however, is that these theories discuss knowledge creation in intra-firm teams whereas the present study focuses on inter-firm teams. This has few consequences for the process of knowledge creation but it does require one to take a closer look at governance structures. In this light, it is useful to refer to the work of Nooteboom (2000), who argues that, in spite of its shortcomings, some elements of transaction cost economics (Williamson 1993, 1999) are useful as they can be fruitfully employed in a wider theory of coordination in innovation systems. In his view, “one piece of salvage is the notion of specific investments as a cause of dependence” (Nooteboom 2000: 920). Superficially, this appears to be the case in the KIC network as specific investments, arguably, have been made and the companies were mutually dependent on each other. On closer examination, though, the specific investments made in the KIC project were the result of the companies being dependent on each other, not, as Nooteboom argues, the cause. Because the companies in the KIC network followed an innovation strategy, they had to engage in mutual knowledge exchange in order to engineer the desired modules. Their dependence, in other words, followed from strategic considerations, not from transactions. This ‘piece of salvage,’ thus, is a dead end from the perspective of the present study. Coordination seems to be of utmost importance. With regard to different forms of coordination, Nooteboom (2000) argues that transaction cost economics is valuable because it looks into the “redistribution of the ownership of specific investments.” This could be applied to knowledge in so far as the companies in the KIC network did make specific investments in knowledge and Océ became the owner of the fruits of the knowledge creation effort. It should be clear, however, that embedded knowledge cannot be owned because it is socially constructed. Embedded knowledge is, as the concept says, embedded in social relations and no single actor can own embedded knowledge. Neither can an actor own codified knowledge as anyone who has access to codified knowledge can acquire that knowledge 3

The definition of apprenticeship teams and goal-directed communities is irrelevant for the present discussion. It suffices to say that they are very similar to the micro communities of von Krogh et al. (2000).


217

at (virtually) no cost. What Océ owns after the KIC project is not a body of knowledge but the right to use that knowledge. The participants in each cluster ‘own’ the same body of knowledge since this knowledge was constructed in a process of interaction between them. The difference between Océ and the suppliers is that the suppliers have to forbear from using this knowledge. The conceptual framework of transaction cost economics is, in sum, inadequate to explain the process of inter-firm knowledge creation. The focus is on costs and not on the driving forces of learning. Though this is not the place for a lengthy discussion of governance structures, the case study of the KIC project suggests that the emphasis in theory must be on knowledge and on the process of knowledge creation, i.e., learning. This process is far more socially constructed than, for example, transaction cost economics is willing to admit. Given the dependency relations between the companies in the KIC project, the Granovetterian embeddedness approach (Granovetter 1985, 1992) may constitute a more fruitful way to explain why and how companies collaborate on knowledge creation.

10.15 Space Matters (More than Ever?) The key question, then, is about how important is space in this respect? The previous already showed that, in the engineering phase, proximity is the outcome of a trade off between, on the one hand, the fact that knowledge exchange benefits from face-to-face communication and, on the other hand, the question whether or not a company is able to actually involve proximate partners in this process of knowledge exchange. If involving proximate partners in the engineering phase is possible, this certainly seems to be the wisest choice, however, it does not go without saying that this choice will always be made. This certainly puts in perspective the ‘received wisdom’ (Oinas 2000) considering the relation between proximity and learning. It is not as straightforward as some authors have suggested (cf. Angel 2002, Maskell et al. 1998, and Morgan 1997). The focus of scientific inquiry, therefore, should be on knowledge, not on space or proximity. Contrary to what economic geography has subscribed to for many years, space and proximity are not the starting point of the analysis, they are the outcome! Explaining the spatial dimensions of a network, therefore, requires that regional economists look at the spatial dynamics of the processes and activities of that network. Obvious as this may seem, it does shift the level of analysis from ‘space’ to ‘network activities,’ i.e., a lower level of analysis.

218


This also explains why, for example, the learning-region perspective is absent from this work. It belongs to a different level of analysis. The learning region looks for answers on the regional level, such as the role of regional innovation policy and the regional innovation system with regard to regional economic development (cf. Hassink 2001 and Morgan 1997). The present work, in contrast, looks at inter-firm relations first and their spatial dimensions second. In general, this issue of levels of analysis requires more attention in regional economics. Regional externalities are for that reason a topic in future research in spatial science.

10.16 Conclusion One of the objectives of this chapter was, to emphasize the relation between regional externalities and clusters by looking to an empirical Dutch case-study. In the introduction we have paid attention to the ambiguity in literature as far as regional externalities are involved. The literature is not pointing in one clear and distinctive direction. The discussion is on the one hand focused on the specialization hypothesis (Marshall) and on the other hand on the diversification hypothesis (Jacobs). According to the specialization hypothesis of Marshall, knowledge is to a high extend industryspecific and for that reason it tends to be exchanged between firms which are part of one industrial sector. As a result, firms within a particular industry capitalize upon these intra-industry knowledge spillovers. This will be the case more predominantly if and when they are located nearby concentrations of such particular industry. The diversification hypothesis of Jacobs asserts, by contrast, that knowledge developed in one sector of industry, may be successfully applied in other industries. This will cause a diversified rather than a specialized local production structure to be conducive to local economic performance and innovativeness. The hypothesis of Marshall argues that local market power restricts the flows of ideas, whereas the hypothesis of Jacobs advocates fierce local and regional competition, resulting in an increasing flow of innovative ideas. If we consider the Dutch case we will find some authors argue that regional externalities do not affect or discriminate between regions in economic performance and innovation, whereas others tend towards opposite conclusions. From the KIC-case it shows that the hypothesis of Marshall holds; geographical concentration of industries increases local and regional innovativeness in that particular industry. Another objective of the present study is to contribute to the literature on learning, innovation, networks, knowledge, and space by discussing the


219

case study and to interpret its findings in the light of the ‘established literature.’ Inevitably, this approach yields more questions than answers, however, we regard this as a favorable outcome as it places some question marks with regard to the received wisdom from the established literature. As for competitiveness in regional clusters, this work shows that close collaboration on engineering does lead to better results. Whether or not this actually resulted to firm competitiveness cannot be established within the context of the present study. What is important, however, is that this study points out that there is no simple, straightforward relation between learning, competitiveness, and space. Moreover, there are two different questions involved here. In the first place, it is the issue of how inter-firm learning can contribute to competitiveness, or innovation. This, we argue, largely depends on the phase of the innovation process. We feel that, thus far, the different phases of the innovation process have not been sufficiently accounted for in the literature on knowledge creation. The second issue is how spatial proximity can facilitate innovation. This, we argue, is the outcome of a trade off between the desirability to involve external partners and the possibilities that a company has to actually do so. In other words, the focus is not on space but on the content of the inter-firm relationship, knowledge creation in this case. This idea, too, we believe, has had insufficient attention in regional economic literature. Finally we have to emphasize that our presentation of the KIC case in this work has its limitations. We argue, however, that the strength of this work lies in the empirical support it offers for some theoretical considerations, such as the geography of knowledge, and the questions it raises with regard to other considerations. Regional externalities and (innovative) clusters seem to be more than a fruitful topic for more fundamental as well as empirical research in the future. There is no need to emphasize once more that this does mean the urgent need for more multi-disciplinary research, both fundamental as well as empirical. The knowledge-based economy in general and the problem of regional externalities still presents a challenge to the science of regional economics, economic geography and related disciplines.

References Acs, Z.J. (2002), Innovation and the growth of cities, Edward Elgar, Cheltenham. Audretsch, D.B., M.P. Feldman (1996). R&D spilovers and the geography of innovation and production. American Economic Review, 86 (3), 630-640 Angel, D. (2002), Inter-firm collaboration and technology development partnerships within US manufacturing industries, Regional Studies, Vol. 36-4, pp. 333-344.

220


Best, M., (2001), The new competitive advantage. The renewal of American industry. Oxford, Great Britain Boekema, F., H.van Houtum (1995), Regional economic competitiveness; Porter and beyond. In: Beije, P., H. Nuijs (eds.) The Dutch diamond? The usefulness of Porter in analyzing small countries. Siswo,Garant, Leuven-Apeldoorn. Boekema, F., L. Oerlemans, J. Dagevos.(1993), Networking; Risk-reduction in a turbulent environment. In: Beije, P., J. Groenewegen, O. Nuijs (eds.) Networking in Dutch industries. Garant-Siswo, Leuven-Apeldoorn. Boekema, F., A. Nagelkerke (1991) Labor relations and regional development in the Netherlands: A network approach. In: Dietz, F. W. Heijman, D. Shefer (eds.) Locational and labor considerations for regional development. Avebury, Aldershot. Boekema, F., R. Rutten (eds.), (2004), Knowledge, Networks and Proximity: An embeddedness perspective. In: Special Issue: The quest for Spatial Embeddedness- Knowledge, proximity and Capabilities. European Planning Studies, vol. 12, no. 5, pp 603-607 Boekema, F., Morgan, K., Bakkers, S. and Rutten, R. (eds) (2000), Knowledge, innovation and economic growth: The theory and practice of learning regions, Edward Elgar: Cheltenham. Boekema, F.,R.Rutten (eds.) forthcoming, The Learning Region, Past Performance, State-Of-The Art, Future. Edward Elgar, 2006 Braczyk, H.J., Ph. Cooke, M. Heidenreich. (eds.),(1998).Regional Innovation Systems, UCL-Press, London. Granovetter, M. (1985), Economic action and social structure: The problem of embeddedness, American Journal of Sociology, Vol. 91-3, pp. 481-510. Granovetter, M. (1992), Problems of explanation in economic sociology, in Nohria, N. and Eccles, R. (eds), Networks and organizations: Structure, form, and action, Harvard Business School Press: Boston, Ma. Grabher, G. (ed.), (1993), The embedded firm: On the socioeconomics of industrial networks, Routledge: London. Judge, W., Fryxell, G. and Dooley, R. (1997), The new task of R&D management: Creating goal directed communities for innovation, California Management Review, Vol. 39-3, pp. 72-85. Hassink, R. (2001), The learning region: A fuzzy concept or a sound theoretical basis for modern regional innovation policies?, Zeitschrift für Wirtschafstgeographie, Vol. 45-3/4, pp. 219-230. Jacobs, J., (1969), The economy of cities, New York, Random House. Kroch, G. von, Ichijo, K. and Nonaka, I. (2000), Enabling knowledge creation: How to unlock the mystery of tacit knowledge and release the power of innovation, Oxford University Press: Oxford. Lagendijk, A., P. Oinas (eds) (2005), Proximity, Distance and Diversity. Issues on Economic Interaction and Local Development, Aldershot, Ashgate. Marshall, A. (1890), Principles of Economics, 8th ed. Macmillan, London. Maskell, P., Eskelinen, H., Hannibalsson, I., Malmberg, A. and Vatne, E. (eds) (1998), Competitiveness, localised learning and regional development: Specialisation and prosperity in small open economies, Routledge: London.


221

Morgan, K. (1997), The learning region: Institutions, innovation and regional renewal, Regional Studies, Vol. 31-5, pp. 491-503. Morgan, K., C. Nauwelaers, (eds.), (2003), Regional Innovation Strategies, Routledge, London. Nooteboom, B. (2000), Institutions and forms of co-ordination in innovation systems, Organization Studies, Vol. 21-5, pp. 915-939. Oerlemans, L., M. Meeus, F. Boekema (2001), On spatial embeddedness of innovation networks: an exploration of the proximity effect. In; TESG/ Journal of Economic and Social Geography, vol. 92, no 1. Oerlemans, L., M. Meeus, F. Boekema (2001), Firm clustering and innovation: determinants and effects. In; Papers of the Regional Science, Vol. 80, no. 3. Oerlemans, L, M. Meeus, F. Boekema (2001), Innovation and proximity: Theoretical perpectives. In: Green, M., R. Mc. Naughton. (eds.) Industrial Networks and Proximity. Ashgate, Aldershot. Oinas, P. (2000), Distance and learning: Does proximity matter?, in Boekema, F., Morgan, K., Bakkers, S. and Rutten, R. (eds), Knowledge innovation and economic growth: The theory and practice of learning regions, Cheltenham: Edward Elgar, pp. 57,-69. Panne, G. van der, (2004), Entrepreneurship and localized Knowledge Spillovers, Delft. Rutten, R. (2002), The entrepreneurial coalition: Knowledge-based collaboration in a regional manufacturing network, WLP: Nijmegen. Rutten, R., F. Boekema. (2004), A knowledge-based view on innovation in regional networks: the case of the KIC-project. In: H. de Groot et al , Entrepreneurship and Regional Economic Development. A Spatial Perspective, Cheltenham, Edward Elgar, pp. 175-198. Rutten, R., F. Boekema, (2004), The Spatial dimension of inter-firm learning: case study and conceptualization. In; Cooke, Ph. et al. Regional Economies as Knowledge Laboratories, Chaltenham, Edward Elgar, pp. 181-197. Rutten, R. (2003), Knowledge and Innovation in Regional Industry. An Entrepreneurial Coalition. London, Routledge. Storper, M. (1997), The regional world: Territorial development in a global economy, The Guildford Press: London and New York. Uzzi, B. (1997), Social structure and competition in interfirm networks: The paradox of embeddedness, Administrative Science Quarterly, Vol. 42-1, pp. 35-67. Williamson, O. (1993), Calculativeness, trust, and economic organization, Journal of Law & Economics, Vol. 86-1, pp. 453-486. Williamson, O. (1999), Strategy research: Governance and competence perspectives, Strategic Management Journal, Vol. 20, pp. 1087-1108.

11 Spatial Dimension of Externalities and the Coase Theorem: Implications for Co-existence of Transgenic Crops

Volker Beckmann1 and Justus Wesseler2 1 2

Humboldt University of Berlin, Germany, Email: [email protected] Wageningen University, The Netherlands, E-mail: [email protected]

Abstract. Adopters of transgenic crops produce a negative externality for producers of transgenic free crops through potential pollen flow. Producers of transgenic free crops produce a negative externality for growers of transgenic crops if they call for keeping a minimum distance. This chapter examines spatial implications of co-existence of transgenic crops from the perspective of Ronald Coase’s influential paper “The Problem of Social Cost” published in 1960. First, the problem of co-existence will be assessed as a problem of social cost. Second, we discuss the impact of the distribution of different property rights on the adoption of transgenic crops. Third, we show that allocations of property rights result in different spatial agglomeration of transgenic and non-transgenic crops. Keywords: Coase Theorem, Co-existence, Externalities, Property rights, Spatial effects, Transgenic Crops.

224

Volker Beckmann and Justus Wesseler

11.1 Introduction “No form of agriculture should be excluded in the EU.” Many observers see this recent statement by European agricultural commissioner Franz Fischler as a clear signal towards a nearby lifting of the quasi EU moratorium on transgenic crops (or GMs for short) launched in 1998 (European Commission 2003). One of the last obstacles towards lifting the moratorium, however, is the problem of coexistence. How can GM-crops and non-GM-crops coexist? Since the European Environmental Agency published its report on “Genetically modified organisms (GMOs): The significance of gene flow through pollen transfer” (EEA 2002) the debate has focused on the external effects that GM-farmers may cause to non-GM farmers if accidental pollen transfer takes place. While strong supporters of the GM technology argue that the current legislation is sufficient to deal with this problem (e.g. EuropaBio 2003), others demand strict liability rules for GM-farmers and those who distribute GM-crops. Furthermore, elaborated monitoring systems, GM-crop cadastre and other measures should be established according to their view (e.g. Greenpeace & Zukunftsstiftung Landwirtschaft 2003). The discussion on coexistence and private liability for GM-technology, however, is not limited to Europe. There are ongoing debates in the United States, Canada, New Zealand and other countries (see e.g. Smyth, Khachatourians and Phillips 2002; Kershen 2002; Conner 2003). Thus, the governance of the future co-existence of GM-crops, conventional crops and organic crops is becoming a burning issue. This chapter examines the current debate on co-existence from the perspective of Ronald Coase’s influential paper “The Problem of Social Cost” published in 1960 (Coase 1960). Coase was very sceptical about the role of the government for resolving “harmful effects”. He argued, first, that the traditional perception of the problem - making the polluter liable or taxing pollution - is misleading because it ignores the reciprocity of the problem. Second, he stated that if property rights are well defined and the costs of using the market to reallocate property rights are zero or close to zero, the allocation of resources will be independent of the initial distribution of rights. This statement became known as the Coase Theorem (see Cooter 1991; Posner 1993). Third, Coase noticed that if the costs of using the market to reallocate property rights are not close to zero, all institutional alternatives or governance structures must be evaluated in a comparative way, including the “costs involved in operating the various social arrangements” (Coase 1960: 44).

11 Spatial Dimension of Externalities and the Coase Theorem

225

This contribution will basically proceed in the logic of Coase’s paper but will highlight the possible implications of the Coase-Theorem for the governance of co-existence and its implications for the spatial allocation of GM and non-GM crops. First, we characterize the problem of co-existence as a problem of social cost that can be solved institutionally as well as technically. Second, we analyze the impact of different property right structures on the adoption of GM crops and on the value of GM and non-GM production. We will show that under certain assumptions the adoption of GM crops is independent from the allocation of property and liability rights. In this case, technical and managerial solutions may be adopted to solve the problems of co-existence. However, the values of different production systems are highly affected by the allocation of liability rights. Fourth, the implications of these for the spatial allocation of GM and non-GM farms are discussed.

11.2 Assessing the Problem of Co-existence The problem of co-existence is a classical “problem of social costs”. Farmers who plant GM crops may cause negative (or positive) external effects to nonGM or organic farmers by cross contamination through pollen drift or other forms of admixture. The problem is illustrated in Figure 11.1. Let us consider two supply chains that range from seed production over processing to the final consumer. At each stage, possible accidental contaminations across interfaces are possible. The contamination, in principle, can be two sided. GM crops may affect non-GM crops but non-GM crops may also affect GM crops. It is important to note here that the same physical effect, i.e. pollen flow, can have different economic impacts, depending on the institutional setting. The institutional and regulatory setting defines the rules of what is or is not to be labelled as GM and sets the threshold levels for labelling (Smyth and Phillips 2003). The lower the threshold levels the higher the costs of governance and possible economic losses. Therefore, it is not surprising that the definition of the threshold is subject to strong political debate. In the EU the current food-labelling threshold is 1% (EU, 2000)1. However, the European Council agreement on GM Food and Feed proposal established a

1

It should be noted here that for organic farming no threshold has been decided yet. It is usually assumed that the relevant threshold is at the detection level that is currently 0.1 %.

226


Fig. 11.1 Co-existence: the Governance Problem.

0.9% threshold for food and feed. For seeds, the Commission is proposing even lower thresholds (EU, 2003). These thresholds are low and the likelihood of contamination is high. It should be noted that the labelling regime is different in the United States where market actors can voluntarily label food as GMO free (see Crespi and Marette 2003 for an overview of different labelling policies). In this case market participants define the thresholds, which can vary substantially. Thus, the problem of co-existence is also a problem of governing the flow of goods and services and bads and disservices along the supply chain. However, governance structures are not without costs and these costs have to be taken into account when ap-


227

proaching the problem. The specific admixture through pollen drift, the spatial problem, occurs at the stage of seed and plant production. 11.2.1 A Simple Model Think about a region that consists of a number of farms i= 1,…,k, which show similar cropping patterns and share several borderlines and initially grow only one crop. Further, assume a situation similar to the one observed in Europe: only non-GM crops are grown. The regional value of non-GM production VN is then given by

VN

pN QN CN k

VN

¦ vNi i 1

(1a)

k

¦ p

Ni

q Ni cN i

i 1

(1b)

where pN , QN , and CN are the respective price, quantity and cost vectors for non-GM products at the regional level. Further, vNi indicates the farm level value of non-GM production and pNi , qNi , cNi are the respective farm level price, quantity and cost vectors. If all farmers in the region were to shift to the GM-crop variety, e.g. from corn to Bt-corn, the regional value of GM-crop production, VG , is given by:

VG

pG QG CG k

VG

¦ vGi i 1

(2a)

k

¦

pG i qG i cG i

(2b)

i 1

with pG , QG , and CG as the respective price, quantity and costs vectors of GM-crops at the regional level. Again vGi represents the farm level value of GM production and pGi , qGi , cGi are the respective price, quantity and cost vectors for GM crops at the individual farm level. Since it is expected that consumers are willing to pay a price premium for GM-free food, we assume further that the farm gate price of non-GM

228


crops is universally higher than for GM crops2. This assumption is represented by the equation (3):

pNi ! pGi , i 1,..., k

(3)

If the farm gate prices of GM-crops are assumed to be below non-GM crops, GM-crops must allow for sufficient cost reductions or yield increases in order to be attractive to be grown. At least for one farmer the value of GM crop production must exceed the value of non-GM crops, vGi ! vNi . Otherwise GM-crops will not be grown. Figure 11.2 shows the borderline between farms that will adopt GM-crops and those that will not. A farm will only adopt GM-crops if the value exceeds the value of nonGM crops and not otherwise.

Fig. 11.2 Value of production systems and potential technology adoption.

2 However, there is no reason to believe that this should always be the case. It is also possible that the price of GM-food exceeds the price for non-GM products. In the following, however, we will not consider this case.


229

Assume now, that the whole group of k farmers could be divided into two different subgroups. The first group, i 1,! , h , say group A, has a comparative advantage in non-GM crop production, vN i t vG i ; the second group i k h ,! ,k , say group B, has a comparative advantage in GM crop production, vGi ! vNi .3 Farms belonging to group A with i 1,! , h are indicated with the small letter a and farms belonging to group B with i k h ,! ,k with the small letter b. Different regions may show a different population structure with regard to the type of farms. One region may be populated mostly with type A farmers, another region mostly with type B farmers, and a third region may be equally populated by type A and B farmers. If the latter is the case, then the co-existence of both farm types, if it can be established cost free, will be socially preferable compared to the status quo and to the unified adoption of GM-crops, since the value of co-existence in the region, VC, will exceed the value of uniform adoption represented by equation (4): h

VC

¦ vN a a 1

k

¦v

Gb

! VN ,VG

(4)

b k h

11.2.2 Co-Existence, Economic Damage and Technical Measures Equation (4) assumes that there is no co-existence problem. However, if accidental pollen transfer from GM crops to non-GM crops occurs, the non-GM farmer may face the risk that his non-GM crops will be contaminated with pollen from GM crops. If, as a consequence, he cannot sell his product at a price premium, he will face an economic loss or damage, da. The occurrence and magnitude of the economic damages is influenced by a number of factors represented in equation (5a).

3

This does not imply that the alternative non-GM crop will be of the same variety. It includes cases such as Bt-corn in comparison to non-GM spring wheat, e.g.

230


da

0 ° ° ® °( p p )q Ga Na ° Na ¯

if if

D ND QG qNa

D N QG a

qNa

T (5a)

tT

h

D

¦d

(5b)

a

a 1

The occurrence of the damage at the individual non-GM farm, d a , is determined by (1) the quantity of GM-crops grown in the region QG , (2) the diffusion coefficient D N a that indicates the farm and crop specific impact of pollen drifts from GM crops to non-GM crops and (3) the threshold for the good being defined as GM or non-GM T. As it was already argued, the threshold T is an important factor for the occurrence of economic damage. Economic damage occurs only if the fraction of GM crops in non-GM crops exceeds the threshold level. The magnitude of the damage is influenced by (1) the price difference, p N a pGa and the (2) quantity q N a of non-GM products affected. The damage, of course, is zero if the quantity of GM crops or non-GM crops is zero, if the price difference is zero or if the contamination is always below the threshold level. The total damage in the region, D, is the sum of the farm level damages, da. The diffusion coefficient is of specific importance here. This coefficient can be influenced by different technical measures and management practices, i.e. by isolation distances between fields, buffer zones, pollen barriers, crop rotation systems or by genetic use restricted technologies (GURT) such as infertile pollen (e.g. van de Wiel et al., 2005). These management practices are either related to border management or to the spatial and temporal co-ordination of agricultural activities and can be subsumed as fencing activities. However, influencing the diffusion coefficient requires the introduction of different management practices and is connected with additional costs. If we denote mi as the farm-level management practices that are ranked and f i as the farm-level fencing costs of these practices, the following relationships are assumed:

Di

D i ( mi ,mk i )

(6a)

fi

f i (mi , qN i , qGi , QN , QG )

(6b)

k

F

¦ i 1

h

fi

¦ a 1

k

fa

¦

b k h

fb

(6c)


231

The diffusion coefficient at the farm level is influenced by the farm management practices mi but also by the management practices of all other farms. Let us take the example of the buffer zone as one management system and two neighbouring farms. The diffusion coefficient can be reduced if the buffer zone is implemented by a farm that grows non-GM crops and it will be reduced even more if the GM crop farm establishes a buffer zone as well. However, equation (6a) indicates that there is a coordination problem due to the management practices adopted by different farms. The variable costs of establishing the management and fencing systems as in equation (6b) are not only dependent on the management practices of the farmer, mi, but also on the quantity of non-GM and/or GM crops grown on the farm and the quantity of non-GM crops and GM crops grown in the region. To give an example, it makes a difference if a non-GM farm is surrounded by one GM farmer and four non-GM farmers or by five GMfarmers. Finally, the management and fencing costs in the region are the sum of the individual management and fencing costs as indicated by equation (6c). Through coordinated action farmers may reduce damage and/or fencing costs. They can agree on voluntary solutions such as different rotation practices, planting times or buffer zones. These co-ordination activities are not cost free because of transaction costs. Here, we will differentiate between two situations: one, where the transaction costs are prohibitively high and one, where the transaction costs are zero.4 Considering the additional costs discussed above except for the transaction costs equation (4) can now be rewritten: h

VC

¦v i 1

k

Ni

¦v

Gi

i k h

k

k

i 1

i 1

¦ di ¦ fi ! VN , VG

(7)

Now, the regional value of co-existence is the sum of the values of GM and non-GM crops at the farm level minus the sum of damage and/or fencing costs. Equation (7) reflects the sum of the individual decisions. These individual decisions are affected by the distribution of liability rights as shown in the remaining part of the chapter. 11.2.3 Liability Rights and Distribution of Costs and Benefits To incorporate different distributions of property rights in the form of liability rights in the analysis, let us denote vcNi as the farm level coexistence value of non-GM crops and vcGi as the farm level co-existence 4 For an analysis explicitly considering positive transaction costs that are not prohibitively high, consult Beckmann and Wesseler (2005a).

232


value of GM crops. We introduce a superscript Ɛ that indicates if the GMfarmer is liable for the damages he causes and n if he is not. We assume further, first, that there are no additional costs of holding the GM farmer liable and hence there is no uncertainty involved in proving admixture and, second, that transaction costs between GM and non-GM farmers are prohibitively high. Under this setting two different liability systems are discussed. GM farmer not liable If farmers have the unrestricted right to grow GM crops and are not liable, every farmer switching to GM technology will reduce the value of nonGM crops on fields in the neighbourhood due to damages from the GM field. The co-existence value of non-GM farming of farm i, vcNn i , will be reduced if neighbouring farms plant GM crops by the expected damage di and/or by the costs f i of the management and fencing practices that prevent potential damages. The co-existence value of GM farming, however, does not change for farmer i:

vcNn i

vNi d i f i

vcGni

vGi

(8a) (8b) n Gi

n Ni

Farmer i will now choose to plant GM crops, if vc ! vc . The distribution of rights and therefore costs and benefits as indicated by equation (8a) and (8b) can be assumed not only to influence distribution of economic benefits but also technology adaptation and investments in the management and fencing system. Under the circumstances described, a GM farmer has no incentive to invest in management and fencing practices that prevent damages. The non-GM farmer, however, has an incentive to invest in management systems that prevent damages. Cost minimizing behaviour requires that the non-GM farmer introduces management technologies up to the level where the marginal costs of these technologies are equal to the marginal damages. If the damage and/or the management and fencing costs exceed the incremental value of non-GM crops, di fi ! vNi vGi , the farmer will stop non-GM production. Thus, this type of liability rights increases the adoption rate of GM technology. However, as long as the equation does not hold for all farmers in group A, non-GM crop farming will not disappear.


233

This is illustrated in Figure 11.3. The borderline between GM and non GM farmers moves downwards. All farmers that are still to the right of the new borderline will continue planting non-GM crops. Those farmers that now find themselves to the left of the new borderline will switch to GM crops.

Fig. 11.3 Liability rules, values and technology adoption.

GM farmer liable The costs are distributed in a different way if the potential GM-farmer is liable. If the GM farmer causes damages to the non-GM farmer, he has to pay compensation payments cpGi at the rate of the damage. The damage could be caused on more than one farm.5 The compensation payment sets incentives for the GM farmer to undertake managing and fencing practices that reduce the damages. The value of GM farming therefore will be reduced by the compensation payments and the fencing costs. The value of 5

For simplicity we assume that the source of GM-pollen can be clearly identified, a system similar to the German one with total liable adhesion. The quality of our results does not change, if we assume that a group of farmers will be held liable, such as under the Danish system, only the compensation payment per GM farmer will be reduced.

234


non-GM farming will remain the same since the damage is fully compensated by the GM farmer.

vcNA i

vNi

(9a)

vcGA i

vGi cpGi fi

(9b)

If the expected compensation payments for economic damages and/or the fencing costs exceed the value of GM production, cpG i fi ! vG i vN i , GM crops will be prevented from being grown. This situation is illustrated by an upward move of the borderline in figure 11.3. Farmers that were to the left of the borderline before the introduction of liability rules and are to the right of the borderline after the upward move do not plant GM crops. However, they would have done so without the liability risk. In this section we have assumed that transaction costs are prohibitively high and therefore no negotiation and coordination between GM and nonGM farmers takes place. In the next chapter we will analyse the case of zero transaction costs.

11.3 Co-Existence: A Coasian View Economists have two different readings of Coase’s paper “The Problem of Social Costs” which are important to note here. The first reading is that Coase was purely in favour of private bargaining solutions of the problem of social costs. Under the assumption that the “costs of using the price mechanism” are zero or negligible, he argued that private bargaining would lead to efficient outcomes independent of the distribution of property rights.6 This came to be known as the Coase Theorem. The only role of the government is to assign the property rights and there is nothing for the government to add. This point of view is usually labelled the Coasian view (Glaeser, Johnson and Shleifer 2001). The second reading is that Coase advocates a comparative institutional analysis of all possible relevant alternatives taking the costs of operating various social arrangements into account. All organizational alternatives such as markets, firms, laws, and regulations have different benefits and costs and these have to be accounted for. As Coase argued in his Nobel lecture (1992) the introduction

6 We note that compensation payments may change preferences of individuals and result in different forms of allocating goods. The outcome is still efficient (Perman et al., 2003). But, in our case we look at profit maximizing farms.


235

of the comparative institutional view was his main intention (see also Ellickson 1991; Williamson 1995, and Glaeser, Johnson and Shleifer 2001). The following sections will discuss the problem of co-existence from both points of view. However, it will also be argued that both perspectives have their limitations because they ignore the distributional conflicts involved in assigning property rights and establishing governance structures. Further, it is assumed that farmer’s know already whether or not it is profitable to grow either GM or non-GM crops ignoring damage costs. 11.3.1 Efficient Allocation In order to repeat the result of the Coase Theorem for the case of GMcrops, let us first assume that the GM-farmer is perfectly liable for the possible damages he causes. Thus we are considering equation (9b). The GM farmer has to pay compensation equivalent to the damage caused cpGb d a or he has to invest in technologies in order to prevent damages. The value-maximizing amount of GM-crops grown will be determined where the marginal value of growing GM-crops equals the marginal damages. Now, let us assume that the GM-farmer has the unrestricted right to grow GM crops. He is not liable and does not bear any costs of cross contamination. A naïve interpretation would be that the GM farmer now has an incentive to expand GM crops until the marginal value is equal to zero. At this point he will cause a damage of d' . However, if the costs of using the price system are zero, the non-GM farmer will negotiate and be willing to pay for the reduction of damages in order to prevent the GM-farmer from growing GM-plants, cpN a d a ' d a . Thus the willingness to pay for reduced damages creates an opportunity cost for the GM-farmer. If the GM farmer reduces the amount of GM-crops grown he will be compensated by the non-GM farmer. The amount will be reduced until the marginal benefits from planting GM crops are equal to the compensation payments, which are equal to the marginal damage cost. In conclusion, no liability as well as liability of GM farms will result in efficient allocation of GM and non-GM crops. This is the core argument of the Coase Theorem.

236


11.3.2 Spatial Implications The spatial allocation of GM and non-GM crops will be affected by the distribution of liability rights. A farmer will not adopt GM crops if the expected value is less than the expected value of non-GM crops, i.e. vNi vGi ! 0 but he also has to consider the damage and/or fencing costs. The non-GM farmer’s willingness to pay compensation to the GM farmer in order to prevent damages has limits. The first limit is given by the incremental value of growing non-GM corps. If the expected damage exceeds the incremental value of non-GM crops, the non-GM farmer will quit non-GM farming instead of paying compensation. The second limit is given by the costs for a technical solution to the problem. Given these two limits the following three situations are possible:

vN a vGn a ! vGnb vNb d a f a

non-GM farmer compensates GM farmer for not growing (10a) GM crops

vN a vGn a , vGnb vNb ! d a f a

non-GM farmer accepts damages and/or undertakes fenc- (10b) ing

non-GM farmer switches to (10c) GM farming The situation explained in equations 10a, 10b, 10c is summarized in figure 11.4. The horizontal axis indicates the incremental benefits for nonGM farms and the vertical axis the incremental benefits for GM-farms. The 45-degree line is the boundary where possible compensation payments equal incremental benefits. Take a point above the 45-degree line. There the GM farmer could compensate the non-GM farmer for not growing GM and still maintain a profit. Damage and fencing costs are introduced by the vertical line da+fa. Equation (10a) describes the area to the right of the 45degree line and below the dotted line. In this case the GM farmer will become a non-GM farmer and result in spatial agglomeration of non-GM farms. Equation (10b) describes the area above the dotted line and to the right of the vertical line da+fa. In this case there are no spatial agglomeration effects and GM and non-GM farms will coexist. Equation (10c) describes the area above the 45-degree line and to the left of the vertical line da+fa. In this area the incremental benefits from staying non-GM are less than the damages and fencing costs from neighbouring GM farms, and farmers switch to growing GM crops. In this case a spatial agglomeration of GM crops can be observed.

vGnb vNb ! vNa vGn a d a f a


237

If the GM farmer is liable for possible damages, he will only be willing to plant GM-crops as long as the compensation payment cpGi does not exceed the incremental value of GM production or the cost of fencing investments. The farmer’s decision can be illustrated in the following three arrangements7:

Fig. 11.4 Adjustment strategies under no liability regime.

vGA b vNb ! vNa vGA a cpGb f b

GM farmer will compensate non-GM farmer for not (11a) growing non-GM

vGA b vNb ! vN a vGA a ! cpGb f b

GM farmer compensates the non-GM farmer and/or un- (11b) dertakes fencing

vGA b vNb cpGb fb vNa vGA a

GM farmer will switch to (11c) non-GM crops

The situation explained in equations 11a, 11b, 11c is summarized in figure 11.5. Now, the damage and fencing costs are introduced by the horizontal line cpb+fb. Equation (11a) describes the area above the 45-degree 7 Please note, that equations 10a 10b, 10c and 11a, 11b, 11c imply a negative attitude of farmers towards GM crops.

238


line and to the left of the dotted line. In this area the GM farmer will compensate non-GM farmers for not planting non-GM crops and the non-GM farmer will start planting GM crops. This leads to an agglomeration of GM farms. Equation (11b) describes the area to the right of the dotted line and above the cpb+fb line. In this area the GM farmer will compensate the nonGM farmer for the damages and/or invest in fencing but neither will change their crops and GM and non-GM farms will coexist. Equation (11c) describes the area below the 45-degree and cpb+fb line. In this area the GM farmer will switch to non-GM crops as the damage costs are higher than incremental benefits from GM crops resulting in a spatial agglomeration of non-GM farms.

Fig. 11.5 Adjustment strategies under the liability regime.

In both cases, with and without liability, incentives for spatial agglomeration exist. If fencing and damage costs in both cases are the same, the spatial agglomeration will be the same as well. As farms are heterogeneous, it is reasonable to assume that damage and fencing costs differ between farms. Further, the diffusion coefficient D will depend, among others, on the local geography and will result in different damage and fencing costs between farms. Also, the costs of buffer-zones, as one possible fencing mechanism, decrease with farm size (Soregaroli and Wesseler, 2005). This indicates that the liability system can result in different spatial distribution of GM crops.


239

The results presented in equations 10a, 10b, 10c and equations 11a, 11b, and 11c have additional implications for the spatial distribution of transgenic crops to the ones already mentioned. In the case where GM farmers are not liable for cross pollination of neighbouring fields, incentives for non-GM farmers to cooperate and organize GM free zones exist. Collaboration with neighbouring non-GM farmers increases the total area of non-GM crops and reduces the average damage per unit of area as the average distance to fields with GM crops increases. Also, fencing costs decrease. As we assume that transaction costs are zero the results will change with positive transaction costs. The transaction costs will increase with the number of farmers participating in the GM free zone. The higher the transaction costs the smaller the number of participating farmers will be. With the introduction of a liability rule for GM-farmers, incentives change. Now, GM-farmers have an economic incentive to collaborate and organize GM crop zones. The average damage costs per unit of area can be reduced. Each additional unit of land increases the amount of land within the minimum distance to non-GM crops. As a result, agglomeration of land planted with GM crops is further enforced. 11.3.3 Distributional Implications Even though resources are allocated efficiently under the two liability systems, they have distributional implications. In the case where GM farmers are held liable, they have to shoulder additional costs. The compensation payments non-GM farmers receive cover the additional costs. In this case the non-GM farmers will not gain economically. Even though the GMfarmer has to bear additional costs in the form of the compensation payment, he will still be better off than in the case without the availability of GM crops. In the case where the GM farmer is not liable, he receives the full gain from planting GM crops, but the non-GM farmer has to bear costs and his economic situation will be worse than before the introduction of GM crops. Holding GM farmers liable can be justified from a distributional perspective.

11.4 Conclusions Our analysis of externalities and their regulation applied to the special case of GM-crops shows that different spatial agglomerations of production may result. In the case where transaction costs are low, spatial agglomera-

240


tion of GM and non GM farms can be expected. The empirical relevance of this theoretical result is supported by the appearance of GM zones in Germany and non-GM zones in France and Germany, although regulations and low transaction costs might not be the major reason for the development of non-GM zones (Lavelle, 2005; Beckmann and Wesseler, 2005b). However, our results support the argument that liabilities will be important for the adoption of transgenic crops. The model we presented considers a symmetric case where under one scenario the “polluter” has the right and under the other not the right to “pollute”. Under both scenarios spatial agglomeration may occur. Furtan et al. (2005) report similar results using almost the same approach. They show using field data from Canada that economic incentives exist for organic wheat and oilseed rape farmers to form a club. Our specification differs from their model as they only consider the asymmetric case where the property right is with the GM farmer (as provided under the Canadian legislation). The extend of spatial agglomeration largely depends on the heterogeneity among farmers. This is similar to the results found in the literature on spatial pollution (e.g. Goetz and Zilberman, 2000), where the total damage depends on individual farm characteristics. In our case the amount of damage the GM-farmer, the “polluter”, produces, provides incentives for zoning to reduce the damage, while in most spatial pollution models differences in damages are used to improve the efficiency of government control through zonal taxes, zonal permits and/or zonal standards. Our results are also similar to those reported on tradable pollution permits, where polluters (GM –farmers) have the right (possibility) to buy (compensate) from others (non-GM farmers for) additional permits to continue production. In studies on spatial pollution and tradable permits polluters in most cases are assumed to be liable for violation of government regulations. In this sense our contribution is an extension as we consider symmetric responsibilities, although the general argument has already been made by Coase in 1960.

Acknowledgements The second author acknowledges support by the ECOGEN project, funded by the Fifth European Community Framework Programme: Quality of life and management of living resources, under contract QLK5-CT-200201666.


241

References Beckmann, V. 2005. A comment on “The farmer’s value of transgenic crops under ex-ante regulation and ex-post liability” by Soregaroli and Wesseler. In J. Wesseler (ed.): Environmental Costs and Benefits of Transgenic Crops, pp. 183-184. Dordrecht, NL: Kluwer Academic Publishers. Beckmann, V. and J. Wesseler. 2005a. Distributional and allocative affects of the German “Gentechnikgesetz”. Monograph. Humboldt University, Berlin. Beckmann, V. and J. Wesseler. 2005b. Governance of Genetically Modified Crops in the EU. Paper presented at the Workshop “Problems of Polycentric Governance in the Growing EU”, Berlin, Humboldt University, June 15-18, 2005. Coase, R.H. 1960b. “The Problem of Social Cost.” Journal of Law and Economics. 3:1-44. Coase, R.H. 1992. “The Institutional Structure of Production.” American Economic Review. 82:713-719. Conner, D.S. 2003. “Pesticides and Genetic Drift: Alternative Property Rights Scenarios.” Choices.5-7. Cooter, R.D. 1991. “The Coase Theorem.” In J. Eatwell, M. Milgrate, and P. Newman, editors, The New Palgrave: The World of Economics. MacMillan. New York. 51-57. Crespi, J.M. and S. Marette. 2003. “Does Contain” vs. “Does Not Contain”: Does it matter which GMO Label is Used?” European Journal of Law and Economics. 16:327-344. Ellickson, R.C. 1991. Order without Law: How Neighbors Settle Disputes. Harvard University Press. Cambridge. EuropaBio. 2003. Co-existence of GM and non GM crops. Green Biotech Fact Sheet. Brussels. European Commission. 2003. Communication from Mr. Fischler to the Commission: Co-existence of Genetically Modified, Conventional and Organic Crops. Brussels, European Commission. European Environmental Agency (EEA). 2002. Genetically modified organisns (GMOs): The significance of gene flow through pollen transfer. Copenhagen: European Environmental Agency. Europena Union (EU) 2003. Regulation (EC) No 1829/2003 of the European Parliament and of the Council of 22 September 2003 on genetically modified food and feed. Official Journal of the European Union L268/1-L268/23. Europena Union (EU) 2000. Commission Regulation (EC) No 49/2000 of 10 January 2000 amending Council Regulation (EC) No 1139/98 concerning the compulsory indication on the labelling of certain foodstuffs produced from genetically modified organisms of particulars other than those provided for in Directive 79/112/EEC. Official Journal of the European Union L6/13-L6/14. Furtan, W.H., A. Guzel, and A.S. Weseen. 2005. Landscape Clubs: Co-existence of GM and organic crops. Paper presented at the 11th International Congress of the EAAE, 24 – 27 August 2005, Copenhagen.

242


Glaeser, E., S. Johnson, and A. Shleifer. 2001. “Coase versus the Coasians.” Quarterly Journal of Economics. 116:853-900. Goetz, R.U. and D. Zilberman. 2000. The Dynamics of Spatial Pollution: The Case of Phosphorus Runoff from Agricultural Land. Journal of Economic Dynamics & Control 24: 143-163. Greenpeace and Zukunftsstiftung Landwirtschaft. Considerations regarding the Co-existence of GMO, non-GMO and organic farming. Bruessels/Bochum. 2003. Jung, C., K. Krutilla, W.K. Viscusi, and R. Boyd. 1995. “The Coase Theorem in a Rent-Seeking Society.” International Review of Law and Economics. 15:259268. Kershen, D.L. Legal Liability Issues in Agricultural Biotechnology. National AgLaw Center Publications. Univerity of Arkansas, School of Law. 2002. Lavelle, P. 2005. Description and comparison of the French and German legislative frames on co-existence,between genetically modified, conventional and organic crops. M.Sc.-thesis. Environmental Economics and Natural Resources Group, Wageningen University. Perman, R., Y. Ma, J. McGilvray, and M. Common. 2003. Natural Resource and Environmental Economics. Third Edition. Harlow, UK: Pearson Education Ltd. Posner, R. 1993. “Nobel Laureate: Ronald Coase and Methodology.” Journal of Economic Perspectives. 7:195-210. Smyth, S. and P.W.B. Phillips. 2003. “Labeling to Manage Marketing of GM Foods.” TRENDS in Biotechnology. 21:289-393. Smyth, S., G.G. Khachatourians, and P.W.B. Phillips. 2002. “Liabilities and Economics of Transgenetic Crops.” Nature Biotechnology. 20:537-541. Soregaroli, C. and J. Wesseler. 2005. The farmer’s value of transgenic crops under ex-ante regulation and ex-post liability. In J. Wesseler (ed.): Environmental Costs and Benefits of Transgenic Crops, pp. 165-182. Dordrecht, NL: Kluwer Academic Publishers. van de Wiel, C., M. Groot and H. den Nijs. 2005. Gene flow from crops to wild plants and its population ecological consequences in the context of GM-crop biosafety, including some recent experiences from lettuce. In J. Wesseler (ed.): Environmental Costs and Benefits of Transgenic Crops, pp. 97-110. Dordrecht, NL: Kluwer Academic Publishers. Williamson, O.E. 1995. “Some Uneasiness with the Coase Theorem: Comment.” Japan and the World Economy. 7:9-11.

Part III Regional Policy

12 Abatement of Commuting’s Negative Externalities by Regional Investment in Houses and Buildings

Wim Heijman and Johan van Ophem

Wageningen University, The Netherlands, E-mail: [email protected], [email protected]

Abstract. Over the past decades commuting distances have increased substantially in all developed countries. Bringing jobs and dwellings more closely together is desirable from a societal viewpoint. So, instead of commuting, which causes extensive negative externalities like air pollution and traffic congestion, residential mobility should be fostered. The question that then arises is which type of investment influences residential mobility of households and reduces commuting. After an exploration of the relationship between investment in houses and residential mobility, a model is developed that allows for the testing of the hypothesis that, on the one hand, investment in houses in a region favours residential mobility to that specific region and, on the other hand, reduces in-commuting. The hypothesis is tested on data from provinces in the Netherlands for the years 1998 and 1999. The results indicate that the model gives a fairly adequate description of commuting and residential mobility behaviour. The hypothesis is confirmed. Furthermore, it appears that the investments in houses in a province in the Netherlands should be roughly twice as high as the investments in buildings. The policy implications of the results are discussed. Keywords: Commuting, Residential mobility, Investment in buildings; Investment in houses, Externalities

246


12.1 Introduction and Research Problem In general, spatial mobility of households consists of commuting and residential mobility. Commuting distances have, on average, increased substantially in all developed countries. This can be explained by rising incomes and decreasing real costs of mobility in time and/or money caused by improvements in infrastructure. Also changes at the household level, such as the increase in the female labour market participation and rising housing costs induce more commuting. The increase in commuting causes negative externalities like congestion, air pollution, and landscape degradation. Much research has been done on changes in commuting behaviour in the Netherlands in the 1980s and 1990s, see e.g. Rouwendal and Rietveld (1994) and Rouwendal (1998), and the relationship between residential mobility and commuting (van der Vlist, 2001). Dieleman (2001) gives an overview of recent trends in European and North American research for residential mobility at the micro level. However, we analyse the problem at another level, the level of regions, a somewhat more aggregate level of analysis. The aim of this paper is to investigate whether governments, by means of spatial planning in the sense of creating additional housing space by investment in houses, are able to stimulate residential mobility. This can be the case if investment in houses in a particular region favours domestic migration to that specific area. The analysis of the problem is carried out by means of a model (Section 3). The conclusion of the model is that the increase of regional in-commuting is a positive function of investment in buildings (not being houses) and a negative function of investment in houses. The conclusion is tested for the Netherlands in Section 4. For this we make use of regional commuting data at the provincial level. The final section contains the conclusion and a short discussion of the results.

12.2 Investment in Houses and Residential Mobility Official statistics show several forms of investments. For instance, Statistics Netherlands online databank CBS Statline uses five categories of investments: houses, buildings, transportation, machines and miscellaneous. Only one of these types has a direct influence on residential mobility of family households, i.e. investment in houses. One expects migration to a region if investments in houses have been taking place in that region. This can not only be attributed to ‘a supply creates demand’ effect, but also to the relative price effect as well, as is predicted in the classical Alonso

12 Abatement of Commuting’s Negative Externalities

247

model. The role of investment in houses is that it lowers the (relative) prices of houses (building sites) in the area under consideration, ceteris paribus. Of course, the decision to move to another dwelling in another community or municipality is influenced by many factors, the price of the house being one of them, as is shown in the Alonso model. The Alonso model uses neo-classical behavioural assumptions about utility and the budget constraint. It is based on the following rather strong assumptions (see van der Veen, 1999): x employment is concentrated in the city; x the ‘representative consumer (household)’ maximises his utility; x there is an absence of public goods, externalities and agglomeration effects; x the housing market is homogeneous, and only the size of a plot is taken into account by the representative consumer; x there is free choice of residential location, with no intervention by the authorities. Households have a choice between, on the one hand, living close to the city where the employment is located, which implies low transportation costs but high land prices, or, on the other hand, living relatively far away from the city, with high transportation costs to the city but low prices of land for residential purposes. So there is trade-off between the price of living space and the travelling distance to the place of work. Despite the strong assumptions, the argument of the trade-off is still valid today, and both variables are elements of any likelihood function for establishing migration or migration patterns to a particular region. The Alonso model is based on neo-classical economic theory. In present-day research within the field of regional sciences, the limits of this type of analysis are clearly seen. For instance, Van Dijk and Pellenbarg (2000) make use of a broader model as their theoretical framework for a firm’s migration decision. After a discussion of the pros and cons of a distinction between push, pull and keep factors, they distinguish between the following three categories of explanatory variables: firm internal factors, location (site and situation) factors and firm external factors. Firm internal factors include organisational structure, management, organisational goals and financial assets. Firm internal factors can be related to the ‘life cycle’ of the firm. Location factors include size of a lot, occupancy characteristics, accessibility, quality of public space, distance to markets and local government policy (with regard to spatial planning and land-use). Labour market issues, government policy, amount/location/quality of suitable locations available elsewhere and general economic conditions are among

248


the firm external factors. This distinction between three major explanatory or independent variables can also be applied to the (re)location decision of households. A type of decision that is also important for firms as is shown in the following quotation: “Nowadays several firms located in the densely populated Randstad areas of the Netherlands are thinking of moving (parts) of the company to less densely populated areas because they have problems in recruiting personnel. This is due to the fact that the labour market in the Randstad is tight and the house prices are extremely high. Many workers prefer (and can get) a job outside the Randstad where they can buy a much cheaper and nicer house and save a lot of commuting time. As a result firms consider relocating to the less densely populated areas outside the Randstad. Similar changes in other location and external factors since the initial settlement of the firm may also cause relocation” (Van Dijk and Pellenbarg, 2000: 199). This distinction can also be applied to the residential mobility decision of households. As well as life cycle variables like the age of the heads of the household, and the number and age of the children, household internal factors include other organisational structure variables such as the number of breadwinners, the type of job of breadwinner in terms of qualification and duration of the work week. Income and wealth are two other important internal factors. Location factors include the size of the lot, occupancy characteristics, accessibility (by road, by public transport), quality of public space, distance to work, schools and shops, and local government policy on spatial planning and land use. The household external factors are general economic conditions, housing market issues as price of houses, labour market issues and the various kinds of government policy on taxation and subsidies issues (mortgage deduction, rent subsidy etc.), mobility, environment and region. Another, not unimportant part of the external factors for household is the price of mobility. Investment in houses has a positive influence on the decision of households to move, the propensity to be residentially mobile, at all three levels mentioned above (internal, location and external factors). Over the past two decades, many things have changed in the society of the Netherlands with respect to the household internal factors. One of the most profound changes is the withering away of the one breadwinner household, a consequence of the secular rise in the level of education of women. Not only the division of tasks in the dual earner families differs from the one in the traditional households, but the management does as well. The reconciliation of the various time schedules of the people involved is a major task, and


249

much time and attention is devoted to this issue of planning and control of the household. Job types have changed too, with those in the services sector dominating. Services require strong relational attitudes, found more often among women than among men. Home-based working, teleworking or distance working, are more frequently observed than in the past. High real wage rates make leisure and household work scarcer and many households are confronted with time pressure. Investment in houses then creates more opportunities for households, with a variety in characteristics for attaining a higher welfare position. This is even more the case, if account is taken of location factors. The housing market is not a market in which homogeneous goods are traded. So, investment in houses is not to be seen as an investment in the same type of commodity. For many households investment in houses reduces distances to work or school, and, thereby, reduces the time costs. For others, the opportunity to move to a house with the desired quality is realised in this way. Finally, investment in houses affects external factors such as the price of housing and government policies on the reduction of spatial mobility. Investment in houses and prices of houses is negatively related, ceteris paribus. This not only benefits the buyer of a particular house, but also all households in the role of residents. For a household in the role of investor in non-movable property, the opposite may be true. We conclude that regional investment in houses positively influences the decision of households to move to a house in the region concerned. Given the trade-off between residential mobility and commuting, it also leads to less commuting to that region, the so-called in-commuting. In the case of data of individual households, this may be described by a logistic likelihood function, with internal, location and external factors as independent variables in the equation and the decision to move or not as the dependent variable. In the case of data at the regional level, the test procedure is of a different type and more of a classical econometrics one. The hypothesis that regional investment in houses stimulates migration to the region and reduces in-commuting, is tested using the empirical model described in Section 3.

12.3 A Model of Spatial Mobility As described in Section 1, there is a relationship between commuting and residential mobility. The model described in this section contains a number of assumptions reflected in four equations. The first assumption is that the

250


increase of in-commuting in a region 'Ci is a negative function of the increase of the supply of labour 'Ls and a positive function of the increase of demand of labour 'Ld in the region concerned:

'Ci = 'Ci('Ls, 'Ld, …),

w'Ci / w'Ls < 0, w'Ci / w'Ld > 0. (1)

The second assumption is that the increase of the regional supply of labour 'Ls is a positive function of net residential mobility (net immigration) to a region, Mi: 'Ls = 'Ls(Mi,…),

w'Ls / wMi > 0.

(2)

The third assumption is that the increase in demand for labour 'Ld is a positive function of the investment in buildings (other than houses) Ib. The idea behind this is that investment in buildings reflects the expansion of economic activities attracting more staff: 'Ld = 'Ld(Ib,…),

w'Ld / wIb > 0.

(3)

Finally, the fourth assumption is that net immigration in a region Mi is a function of regional investment in houses Ih: 'Mi = 'Mi(Ih,…),

w'Mi / wIh > 0.

(4)

w'Ci / wIb > 0, w'Ci / wIh < 0.

(5)

The conclusion of the model is: 'Ci= 'Ci(Ib, Ih,…),

Where it is a policy aim to avoid an increase in commuting, the investments should be allocated in such a way that, if (5) is a homogeneous function:

w'C i Ib wI b

w'C i Ih. wI h

(6)

In Section 4 we will test the conclusion of the model, equation (5), for the Netherlands.


251

12.4 Testing the Model The hypothesis resulting from the model is that, on the one hand, investment in houses in a region favours domestic migration to that specific region while reducing in-commuting. On the other hand, investment in buildings stimulates in-commuting. We tested this hypothesis for the twelve Dutch provinces, taking the year 1998 for the investments and the difference between in-commuting in 1999 and 1998 for the increase in incommuting (see Table 12.1). Unfortunately the data for commuting across the provincial borders are only available for the two years mentioned. Linear regression with a constant gives the following result (t-values in brackets): 'Ci = 1.1941 - 0.0111 Ih + 0.0203 Ib, (R2 = 0.53, adj. R2 = 0.42). (7) (0.69) (-2.40) (2.88) Because the constant term is not significant, we also carried out the regression procedure without a constant term: 'Ci = -0.0095 Ih + 0.0188 Ib, (-2.44) (2.88)

(R2 = 0.50, adj. R2 = 0.35).

(8)

Except for the constant term in equation (7) the coefficients are significant at the 1% level (one tailed test).1 The number of observations is limited (n=12), but relate to the highly aggregate regional level. The results seem to indicate that the model as specified in Section 3 gives a fairly adequate description of commuting and residential mobility behaviour.

1

Equation (8) is homogeneous of degree 0, therefore equation (6) applies.

252


Table 12.1: Increase of in-commuting, investment in houses and buildings per province. Province

Increase of incommuting 1999-1998 (thousands),

Investment in houses 1998 (millions of Euro’s),

Investment in buildings 1998 (millions of Euro’s),

'Ci

Ih

Ib

Groningen

1

645

311

Friesland

3

854

359

Drenthe

-8

694

272

Overijssel

-1

1539

675

Flevoland

-3

1019

172

Gelderland

7

2280

1294

Utrecht

13

1460

1210

North Holland

16

2864

2013

South Holland

0

4318

2547

Zeeland

2

554

208

North Brabant

2

3337

1749

Limburg

-3

1228

843

12.5 Policy Implication and Conclusion Taking into account equation (6) in Section 3 and (8) in Section 4, in order to prevent a further increase of in-commuting ('Ci = 0), 0.0095 Ih = 0.0188 Ib. This implies that the investment in houses in a Dutch province should roughly be twice as high as the investment in buildings. Investment in houses should be especially stimulated where: 1. there is an increase of in-commuting and 2. the ratio investment in houses to investment in buildings is less than two. If we take a look at Table 12.1, according to these criteria, investment in houses must be especially stimulated in Gelderland, Utrecht, North Holland, and, to a lesser degree, in North Brabant.


253

The conclusion is that for domestic migration, the investment in houses is a determining factor. So, for this type of migration our hypothesis (investment in houses in the region stimulates domestic mobility to the region) is confirmed. Further, it is obvious that investments in buildings are connected with an increase in the demand for labour and in-commuting. From this one gets the impression that investment in houses is not done in the same region where the jobs are created. This has led to more commuting with the negative externalities connected to it. So, the analysis indicates that the trade-off between residential mobility and commuting has led to the choice for relatively cheap houses at a relatively long distance from the work place, because of the building policy of the government at the time. To avoid an increase in commuting mobility, policy should be aimed at allocating residential sites near to the location of employment as much as possible. This may sound as a truism, but apparently this was not the case in the years under consideration. The reduction of commuting distances requires a building policy aiming at investment in houses in residential areas closer to the workplace. Of course, this might not always be possible because of physical and technical restrictions on spatial planning. And, as has been discussed in section two, the workplace is for households only one reason, although an important one, for residential mobility. Furthermore, the conclusion that the investment in houses in a Dutch province should roughly be twice as high as the investment in buildings, is based on a limited number of observations due to the lack of appropriate data. However, bringing houses closer to he workplace, fosters the wellbeing of households and, in achieving this, negative externalities can be avoided as much as possible.

References Alonso W., 1960. A Theory of the Urban Land Market. Papers and Proceedings of the Regional Science Association, Vol. 6, 149-157. Dieleman, F.M., 2001. Modelling residential mobility; a review of recent trends in research. Journal of Housing and built environment, 16, 249-265. Dijk J. van and P.H. Pellenbarg, 2000. Firm relocation decisions in the Netherlands: an ordered logit approach. Papers in Regional Science, 79 (2), 191-221. Rouwendal J., 1998. Search Theory, Spatial Labor Markets, and Commuting. Journal of Urban Economics, 43, 1-22.

254


Rouwendal J. and P. Rietveld, 1994. Changes in Commuting Distances of Dutch Households. Urban Studies, 31(9), 1545-1557. Veen A. van der, 1999. Paradise regained: over milieu- en ruimtelijke kwaliteit (Paradise regained: on environmental and spatial quality). Inaugural Address, Twente University. Vlist, A.J. van der , 2001. Residential mobility and commuting. PhD dissertation Vrije Universiteit Amsterdam. www.cbs.nl/statline

13 Risk as an Externality in Quantitative and Marginal Approaches

Václav Beran and Petr Dlask Czech Technical University, Czech Republic, E-mail: [email protected], [email protected].

Abstract. In this chapter a model will be presented that aims at the explanation of disparities in regional economic growth in the Czech Republic and its relevance for regional economic policy. The proposed specific regional development strategies are based on a compact model containing only a few variables. With the help of the model the development of 13 regions in the Czech republic will be simulated. From the computations it appears that innovations are the driving forces behind regional development. However, for sustainable regional growth, it is essential to understand the dynamics of innovative dissemination by networks. This chapter is an attempt to contribute to that end. Keywords: Regional economic disparities, Regional economic policy, Innovations.

13.1 Introduction This chapter is based on the Macroeconomic Theory of the Cobb-Douglas production function and deals with the distribution of income in 13 regions of the Czech Republic in 2002/2003. The dynamics (Rektorys, 1982), (Kubík, Kotek, 1982) of regional growth have changed extensively in recent years (Mandelbrot, 1991). A number of successful regions in the recent past have tended to display some common characteristics. Co-operation and knowledge transfer via various types of networks are especially important. These are the sources of new types of external innovations. However, it is essential to understand the dynamics of innovation dissemination by networks for sustainable

256

Václav Beran and Petr Dlask

innovative growth. There are more general rules that make networks efficient (Beran, J., 1994). First, it is necessary to know the structure of the decision space (Wolfram, 2002), (Raiffa, Schlaifer, 1961), (Sandelin, 1976) for general findings in the regions in order to generate strategies for regions with very different conditions. The theory of the economic behaviour of regions is based on the assumption that x perfect competition exists between different regions, and external influences such as international trade can be neglected, x regions take prices as given, x regional output Y is gross domestic product (GDP) at market prices, x output price of regions is p, x regions have labour L and wages available w, x a region´s rent capital is K at a rate r. We will suppose that there are two general strategies for managing regions. a) industry oriented – maximizing profit, as given in (1), means that output will be measured as Profit = Total Revenue - Total Cost

(1)

b) user oriented – maximizing welfare, as given in (2), will be measured as Welfare = w u L + r u K

(2)

Expression (1) is extended to (3) Profit = [p u F(K, L)] - [w u L + r u K]

(3)

where F(K, L) is the creation model of production Y valued at prices p. The production factors implemented are L and K, their value under local conditions is w (wage rate per hour) and r (rate at which capital is available). The regions have to select the level of K and L that maximizes future profits. The main question is how a firm or even a region chooses for profit (3) an economically suitable K and L. The potential or actual substitution of K by L, and vice versa, is not only a question of rationality, but also a question of volatility in the given solution. This volatility has technical reasons, as even economic or social limits must be taken into account. Calculations in terms of parameterisation may give answers to many real feasibility questions.


257

13.2 The Cobb-Douglas Equation In terms of theory, every existing economic process P operates predominantly on the basis of analytic concepts of description, quantitative or selected qualitative relations. On the level of macroeconomic studies, economic processes are described by production functions. One of most used is the Cobb-Douglas production function (CDF, appeared 1927). However, the first application of a production function was by Knut Wicksell (1851-1926), introduced before 1900 (see in Sandelin, 1976) and modern Charles Jones approach (Jones, 2005). Any production function has to enable interaction between growths of resources and achievable outputs Y. A graphical description of a generalized production function is given in Fig. 13.1. Fig. 13.1 Generalized production function Y = F(K, L) dependent on L (labour) and K (capital), adaptation from (Schumann, Mayer, StrĘbele, 1999). The moral principle of this part of macroeconomic theory is that growth of resources K, L empowers production Y. CDF, very often used in economics, provides enable relatively good results. Charles Cobb´s solution of presenting of a process by means of a production function is based on the equation Y =F(K,L) = AKDL1-D

(4)

where constant A > 0 measures the productivity of the available region technology for L > 0 and K > 0. The production model (Cobb´s model) given in (4) describes the quantitative outputs. These are later defined on the basis of virtual element sets AK (capital) and AL (labour), according to Chapter 5 in (Beran, V., 2002) or (Vlþek, Beran, V., 1984). Changes of quantitative values of (4) are later referred to called as virtual moments, described as AsQ indicate speed of changes, and are calculated as

wY wK

D A K D 1 L1D

(5)

and wY (6) 1 D AK D LD wL where 0 < D < 1. In microeconomic terms, changes in (5) and (6) describe marginal values influencing economic elements (variables). In (4) we have involved labour and capital as decisive input factors. From the microeconomic point of view we use the transcription

258


Y

F (L 2 ,K ) economic output

F (L ,K 2 )

L ISOQUANT for Y

L2 s ubstitution range of is oquant for K

F (L 1 ,K )

L1

F (L ,K 1 )

subs titution range of isoquant for L

K1

K2

K

Fig. 13.1 Generalized production function Y = F(K, L) dependent on L (labour) and K (capital), adaptation from (Schumann, Mayer, StrĘbele, 1999).

MPK # D AK D 1 L1D

(7)

MPL # 1 D AK D LD

(8)

After some algebraic manipulation we have

MPK # D

Y K

MPL # 1 D

(9)

Y L

(10)

in other words, MPL is nothing other than intensity of change in the increase of labour

MPL

F ( K , L 1) - F ( K , L)

(11)


259

and

MPL #

w . p

(12)

A marginal change in the direction of capital will be written as

MPK

F ( K 1, L ) - F ( K , L ) #

r . p

(13)

13.3 Generalisation to Virtual Leading Moments in Virtual Matrix Avirtual The categorisation in the CDF model uses elements grouped in matrix Avirtual. The parameters in the columns create the importance path for further development of model growth in general. The virtual moments on the basis of particular elements are written as the set of vectors of the virtual moments in each matrix A row, described as AQ, AsQ, As

2

, As

Q

3

Q

,…

(14)

We read the first expression as a set of quantities, and the next as a set of quantity changes (for economic parameters applied as productivity, for technical parameters interpreted as production speed), a set of changes of productivity or speed (acceleration), change of acceleration, etc.

ª A1Q « sQ A virtual A = « s12 Q « A1 « «¬ ...

A2Q A2sQ A2s

2

Q

...

A3Q A3sQ A3s

2

Q

...

...º » ...» . ...» » ...»¼

(15)

The relationship to a less restrictive virtual moments notation is given in (Beran, V. at al., 2002). However, there is a close connection to the management and manageability of particular parameters

A

virtual

ª K «A « « wY « wK « w 2Y « 2 ¬ wK

º AL » K AK AL º ª » ª A wY » « » « DAK D 1 L1D # MPK MPL » # « wL » « D 2 1D « 2 w 2Y » «¬ MPK ' MPL'»¼ ¬ (D D ) AK L 2 » wL ¼

AL (1 D ) AK D LD

º » ». (D 2 D ) AK D L(1D ) »¼

(16)

260


Economic output

f(K, L)

f(K, L+1) f K , L 1 f K , L L 1 L

f(K, L)

f K , L 1 f K , L 1

L

L+1

slope MPL

Labour

Fig. 13.2 Graphical interpretation for marginal values of partial segment f(K,L).

As a practical example, we may mention an evaluation of the existing quantitative elements (status quo) by a verbal expert judgement as a virtual matrix

Avirtual

ª AK « high « « strong positive ¬

AL

º » low ». weak positive »¼

(17)

The extension potential for a new model element such as AE (energy) or AInf (information) serves as an example that universalizes and extends scheme (16) and (17). On the basis of the virtual moments we can evaluate (calculate) the second and third line of matrix Avirtual in (15) as indicators of the efficiency of management interventions. In classical microeconomics, marginal values are used. However they are mostly limited to the first level of changes, mentioned in the second line in (15).


261

However, for CDF we will use the adaptation from Fingleton (2003) and write CDF as Y f K, L K D ( AL)1D , which is a more comprehensible expression (Fingleton, Eraydin, 2003). Even capital K is based on investments (defined as quantity) and time dependent capital productivity (second line in (15) interpretation). The Fingleton adaptation (2003) of time dependent capital is Kt = It-1 + (1-d) Kt-1, where d is depreciation. The expression for investments (see also Fig. 13.6) It-1 is given as It-1 = s Yt-2 . The time-lag may play a large role in regional development and the delay may even be one or more years for a significant extensive structural regional investment.

13.4 Externality of Management Decisions and an Alternative to Economic Intuition To clarify the general picture of the problem, it is useful to state that the purpose of these studies is to quantify and test potential development strategies. The formula for output Yt is not the only component that creates the problem structure (read set of process structure P). Decision making and diversity evaluation is one part of the externalities that enable rational behaviour. The memory dependent decision space (D~_ Mem) will later be written in (19) as (D~_ Mem) , where Mem = (dBase, time series, long-term practise, ...)(18) It is not much help that any decision rule influences the problem structure of process P if there are only limited possibilities to develop changes in the particular processes Pi that are based on the set of elements A and causal dependencies K, see (19). The next paragraph presents the application to regional development. For decision making, as a precondition for management practice, it is necessary to have powerful indicators. Decisions are based on a given set of restricting limits. Management is then reduced to a set of feasible solutions. To open a more productive way for decision making it is necessary to formulate a comprehensive decision making environment, valid for a range of time periods and comprehensive organisational dependencies (set of causal connections K).

262


13.5 Generalisation of Process P In CDF given in (4), the increase of output Y is dependent on the increase of inputs labour and capital. However, (4) presents only one possible model based on processes P in (19). For a comprehensive expression of P in (19), P represents an attainable synthesis of controllable elements A and causal dependencies K. Every element A includes description U, time dependence D and quantitative description Q. The causal connection K is a synthesis of elements by a set of intersections V, by the set of construction types ' and the set of time triggers H. Decision making conditioned by long memory data D~~Mem is an important precondition in the selection of efficient steering interventions M (t, P, L). In spite of the complexity of the Cobb-Douglas production function, such a model P forms the basis for developing and defining steering interventions. The sophisticated intention, however desirable, is to create a steering model L (knowledge oriented behaviour synthesis of all included parameters in the process P, e.g. in the Cobb-Douglas function. Let us mention some factors important for economic growth: schooling efficiency, technology development, population growth, etc. The next paragraph focuses on peculiarities of potential growth based on the Cobb-Douglas function. Firstly, we show the statistical givens. Then we will explain the application of risk and sensitivity analysis. In our notation of (19) both analyses will be developed in the context of management or steering interventions M (t, P, L). The complexity of the management models structure is described in (19) as the set of

MN

M iNµM ° ° P ° °° L ® ° ~ ° D ° ~ ° K ¯°

> M t , P, L µD µMem @µK ,

½ ° ° ǹ , K , ǹ U, D ,Q , K V , ǻ, İ , ° °° , ¾ . (19) ° ~ F , dimh µMem databases, long time series,... , ° ° V ~ , ǻ~ ,İ~ ° ¿° ~

~

Desirable models that enable productive design or productive management will be the result. In this sense, it is the low standard of management tools that complicates the development of better and more productive technicaleconomic solutions. When searching intensively for new productive techniques, such as how to protect decisions from unfavourable judgements by reviewers, working with better data or facts, we may be critical about existing methodological


263

conventions. The course and usage of methods practiced for the development of project design is necessarily going to change. On the basis of knowledge resulting from the model, a technician and an economist can use or create steering measures (interventions) freely in the sense of M (t, P, L) in (19), where economic processes Pi are represented by set P and steering processes Li are represented by set L. Existing accompanying methods for describing Pi and Li are used mostly as technical tools for modelling and analysis (simulation, goal parameterisation, optimisation, construction of scenarios, etc.) For further explanation, there is no need to distinguish among particular phases in the description of a real process (P) or steering processes (L). Completion of a model that is able to generate interventions that carry changes from the steering level to the realisation level has been denoted in (19) and carried out by Mi. Tools for the formation of steering interference M (t, P, L), without requirements for determining their robustness or suitability, are as follows: x search for satisfaction of goals by means of solutions regardless of restrictive conditions, x search for solutions on the basis of simulations, x search for an optimal solution, x search for solutions, taking into account time, as desired future solution scenarios. Therefore, it is not appropriate to speak about a single management model. A number of approaches exist, with differing methods and efficiency. Practical modelling of steering intervention for management does not have a uniform character and does not rely on a single theoretical or application pattern. This chapter is in this part an attempt to generalize and extend the theory. The authors consider every technical-economic presentation of reality as a process. The steering models considered represent every abstract description of reality employable for the elaboration of management interventions (19). In this sense, the authors certainly associate themselves with the management school that exploits modelling as a reliable instrument for generating steering proposals. The assumption of homogeneous application fields providing production resources and the application of continuous space and time (still) provides important practical information for decision making. ~ Decision processes D F ~ , dimh are referred F, dim h or D to in (19).

264


Moreover, it will be argued later that a decision is dependent on data with a memory (Mem). Relevant decisions made in the past, influence current decisions. The symbolic notation in (19) respects memory phenomena (time series) as conditioning restrictions on decision making. It is not correct to create an impression that it is feasible to break away from the existing mathematical theoretical and economic rudiments. Even in the modelling of P (in our example CDF) and L (simulation, parameterisation, …), the main model streams practised are simplified models. Decisions may be implemented and propagated (implementation of decision interventions) more carefully, if we are aware of the complexity of the problem.

13.6 Application for Regional Development in the Czech Republic: Findings First of all we will give a concise overview of state of the art in the regions. Relatively rich data sources are available for analysing the economic and social development in the Czech regions. The Statistical Office presents data series in a comprehensive form. Relatively long time data series are available dealing with physical development in the Czech Republic and in the regions. Fig. 13.3 shows differences in land available in the capital city of Prague and in the 13 regions. The land in the regions is used mainly for non-agricultural purposes, and is a prerequisite for generating GDP, nonetheless the production volume given in GDP has different distribution. A relatively small piece of land in the capital generates the highest GDP. Discomfort is compensated by the availability of non-agricultural land for the capital from the surrounding Stĥedoìeský region. The availability of non-agricultural land for development in the capital city region is less adquate. The main flow of new fixed assets is created in the capital and in the surrounding Stĥedoìeský (central) region (Fig. 13.4). Areas with a higher flow of new fixed assets can be found in three of the four regions in the eastern part of the country, in the Jihomoravský, Olomoucký and Moravsko-slezský regions. There is a very similar distribution in the creation of assets such as new buildings and structures. In spite of the relative differences in the creation of assets, there is not much difference between the regions in terms of average monthly gross wages. There is only one exception, the capital city of Prague. Fig. 13.5 compares the vages in agriculture, industry and construction in the regions of the Czech Republic.


Fig. 13.3 Regional urbanised areas (km2).

265

266


Fig.13.4 Fixed assets by head office of the enterprise in regions, (excl. enterprises with 19 employees and less).


267

Table 13.1 provides some more detailed information about the conditions in each region. The low level of unemployment in Prague is an interesting feature. Structural unemployment is much higher in the former heavy industry and mining areas (Ústecký region 17,13 % and Moravskoslezký region 15,89 %). Growth of population is indicated by the birth rate. A low level is indicated in all regions of the Czech Republic. However, the GDP per capita has jumped in Prague to twice the average level. Prague shows different conditions in incomes and living standards. Some more extensive information is given in Table 13.2. Output measured by GDP (Table 13.1) is created in Prague in conditions that use existing tangible and intangible fixed assets (inclusive of buildings and structures). According to the data in Fig. 13.4, there is more emphasis on the creation of fixed assets outside the administrative centre of Prague than inside the centre. In other words there are more assets existing for economic development in the city than in the outside regions. The data in 13.1 and Table 13.2 is available for the last 10 years, and is a practical example of materialisation of (18) in (19). The outputs of the regions given in Table 13.1 and Table 13.2 will be used for microeconomic significance and interpretation. Each region has different conditions for further development (see Fig. 13.3) and other bases of existing technical equipment. Non-agricultural land without forests is an interesting indicator for how far the region is able to develop its industrial and manufacturing potential. Fixed assets located in the regions in the past vary considerably; existing agricultural and non-agricultural land available in the regions differs extensively (Fig. 13.4). Very reliable and clear information is provided by the GDP generated in the regions and even the wages paid to employees (Fig. 13.5). The outputs of the regions are mostly connected with capital assets and labour available for production and wealth. The main point is to be able to compare the regions with each other and to orient future streams of capital and working power.

268


Fig. 13.5 Average monthly gross wages (CZK).


269

270

Beran Vaclav and Dlask Petr

Table 13.3 shows the variety in MPL and MPK through the range of the Czech Republic. Although this data may change over time, the average for Czech Republic is, in the short term, relatively stable. In the context the virtual moment we write

Jihoþeský

PlzeĖský

Karlovarský

Ústecký

Liberecký

Královéhradecký

Pardubický

Vysoþina

Jihomoravský

Olomoucký

Zlínský

Moravskoslezský

MPL =Y/L MPK =Y/K

StĜedoþeský

ýR

Margi nal value

Capital Prague

Table 13.3 Marginal values of labour (MPL) and capital (MPK).

0,6

0,7

0,7

0,8

0,7

0,7

0,7

0,7

0,7

0,7

0,7

0,7

0,7

0,7

5,3

2,7

2,7

4,1

1,3

3,6

4,1

4,2

3,5

3,3

4,1

2,4

3,9

3,5

0,678 3,548

A virtual CzechRepublic

ª K «A « « wY « wK « w 2Y « 2 ¬ wK

ª AK « DAK D 1 L1D « « (D 2 D ) AK D 2 L1D ¬

º AL » » wY » wL » w 2Y » » wL2 ¼

ª AK AL º « » «3,548 0,678 » «¬ MPK ' MPL'»¼

º AL » D D (1 D ) AK L » (D 2 D ) AK D L(1D ) »¼

(20).

Notes: distinguish A as constant in CDF, quantitative description of model elements AK, AL, for A matrix of virtual moments. For further explanations and relevant calculations, see the scheme given in Fig. 13.6. At is interpreted as labour augmenting the productivity of labour factor Lt. However, the main labour flow available is dependent on schooling intensity J and schooling duration W. The other flows in Fig. 13.6 are shaped by population growth rate r. The population growth rate r is relevant to the time lag due schooling. The population from t will be fed into the production process after a lag of 15–20 years. The efficiency of schooling (JW) and population growth rate r are comparable categories. The


271

existing middle range development of Yt for years 1997-2003 is shown in Table 13.4.

(1-d)

J

W

Ht

1

a

At

Kt

Yt

1 s

It

L t- 15 er

Fig. 13.6 Scheme of GDP, creation of the calculation model as Yt.

Table 13.4 shows existing statistical data and data calculated on the basis of existing data. The existing data creates a comparative level of calculation of the production function for the development of future process behaviour. Table 13.4 Input statistical data.

Year

Population Lt

Qualified staff Ht

At

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

10 330 607 10 336 162 10 971 318 10 932 055 10 948 938 10 985 684 11 034 463 11 127 625 10 970 111 10 200 774 10 201 613 10 596 226

4 921 955 4 917 371 4 912 787 4 908 204 4 903 620 4 899 037 4 894 453 4 889 870 4 885 286 4 880 703 4 876 119 4 871 536

0,3151606 0,3976523 0,5083359 0,5747855 0,6094701 0,6590258 0,6767055 0,7188061 0,7740432 0,7998262 0,8451066 0,9217032

Kt [mil. CZK] 217661 339757 463486 520995 534374 554647 550596 594913 638625 643311 677950 750854

D 0,213335 0,287252 0,316010 0,313730 0,299347 0,282625 0,269721 0,276696 0,275834 0,266418 0,265786 0,271991

Yt [mil. CZK] 1 020 278 1 182 784 1 466 681 1 660 649 1 785 130 1 962 482 2 041 353 2 150 058 2 315 255 2 414 669 2 550 739 2 760 586

272


Table 13.6 describes the particular symbols and values used for calculation. A schematic overview of the calculations in the next part of the chapter is shown in Fig. 13.7. CDF is used to predict future behaviour on the basis of aggregated data from the past. Past data (1997-2004) forms the basis for a new calculation describing possible future development (see Table 13.5 and Fig. 13.7). Table 13.5 Calculated results (2006 – 2014).

Year

Population Lt

Qualified staff Ht

At

2006 2007 2008 2009 2010 2011 2012 2013 2014

1 253 855 1 247 715 1 241 150 1 234 168 1 226 776 1 218 981 1 210 793 1 202 220 1 193 270

587 079 583 363 583 856 583 617 580 857 580 025 579 373 578 195 577 022

0,807150 0,840851 0,874552 0,908253 0,941954 0,975656 1,009357 1,043058 1,076759

Kt [mil. CZK] 53139 47864 43115 38839 34988 31521 28398 25586 23053

D

Yt [mil. CZK]

0,227817 0,223529 0,220673 0,219444 0,223124 0,218835 0,217999 0,211632 0,215122

287 852,57 291 574,22 295 942,33 298 710,73 296 243,98 300 810,38 302 423,85 308 989,85 305 887,00

Table 13.6 Values of input parameters.

Inputs Depreciation rate Efficiency of education Length of schooling in years Length of re-education years (new skills) Average savings rate in GDP (1997-2004) Capital share in income GDP (1997-2004)

Symbol d

J years

W s

D

Value 0,1 0,25 15 0,583 0,239875 0,277


273

3500000

GDP [mil. CZK]

3000000

Know n progress of GDP

2500000 2000000 Prognoses

1500000 1000000

KtD At Ht

1D

Yt

500000

Years

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

0

Fig. 13.7 Historical data and calculated prediction of the GDP level as Yt.

13.7 Possible Strategies for Future Change The search for new strategies for future development is associated with the idea presented in formula (19). The calculations in this section are related to the set of tools relevant to L. We present a cut off approach dealing with parameterisation and simulation. A wide range of methods can be used. However, lack of formalisation is mostly a problem. The first robust impact on the output Yt is due to the strength of parameter Į ҏin ҏCDF. Parameter Į regulates the intensity of implementation of capital Kt into the production function output. However, in the notation given in Yt K tD At H t it is interconnected to the balanced involvement of the educated (skilled) labour force Ht . The parameterisation of the involvement of capital in product output Yt is given in Fig. 13.8. Complementary to this parameterisation is the calculation of the simulation of potential changes in this parameter and volatility in time. Fig. 13.9 presents three isolated simulations for instable (risk dependent) factor Į. There are visible differences in the run-in time for a specified range of levels of factor Į. Evaluation on the basis of multiple simulations may mediate the way to sustainable reliable changes and technical-economic solutions. These solutions are the M (L, P, t) composition of changes on the basis of L and processes P in (19). Knowledge about the consequences of the changes and the possible risk enables better management solutions M (x). 1D

274


The main resource for progress is an educated labour force, read Ht. In Fig. 13.10, it depends on H t

Lt a , compared with the graphical 1-J W

scheme of Fig. 13.6. This is the time spent in the educational process, or the re-educational process W parameter, which may raise or decease progress in GDP output. Fig. 13.11 shows three simulations of risk dependences. The rate of increase and stability of growth varies in time and the intensity of time spent in the educational process. A similar situation is population mobility as a total volume Lt and the intensity of transformation into an educated labour force Ht. The intensity of this process is the topic of Fig. 13.12 and Fig. 13.13. The parameterisation of factor J, which describes the intensity (standard) of the educational process is given in Fig. 13.14 and Fig. 13.15.

13.8 Summary of Strategies The calculations in Fig. 13.8 to 13.15 reflect diverse strategies. 1. Efficiency improvement of involved capital (Fig. 13.8 and Fig.13.9). 2. Efficiency improvement of the labour force involved in the process (Fig. 13.10 and Fig.13.11). 3. Intensification of the rate of employment (Fig. 13.12 and Fig. 13.13). The efficiency improvement of the involved capital is calculated according D and is interrelated with the to CDF presented in Yt K t At H t efficiency of the involved labour capital. An increase in the engagement of capital sources leads to a decrease in the labour force. The balance is an important condition for good and harmonised functionality of a real process P modelled by CDF. The efficiency improvement of the labour force is more complicated. In the presented model the standard J of education or re-education plays an important role. The duration of the re-educational process W accelerates a nonlinearly-working improvement in skills and knowledge. Intensification of the rate of employment means the ability to change the rate between population and labour force. This phenomenon concerns the ability to solve problems like long term unemployment, unemployment of school-leavers, the rate of employment of women, the pension age, preretirement regulations and other influencing factors like immigration, invalidity, etc. 1D


275

4 500 000

Level GDP [mil.CZK]

4 000 000 3 500 000 3 000 000 2 500 000 2 000 000 1 500 000 1 000 000

0,4059

500 000

0,2520 0,1488 2014

2010

2012

Year

2008

2006

2004

0

Average D

Fig. 13.8 Parameterisation of factor D in the GDP function Yt

1D

KtD At H t

. The minimal

average D parameter is 0,1488, and the maximal average D parameter is 0,4059, calculation step is 10%.

The development of a productive strategy is important, however difficult the task of combination in time and factual influencing parameters.

276

Beran Vaclav and Dlask Petr 4 500 000

3 500 000 3 000 000 2 500 000 2 000 000 1 500 000 1 000 000

Level GDP [mil.CZK]

4 000 000

500 000 0,4436

0,3533

0,2925

0,2052

0,1790

2012

2014 0,1572

2008

2010

2004

Year

2006

0

D

ue Va l

4 500 000

3 500 000 3 000 000 2 500 000 2 000 000 1 500 000 1 000 000

Level GDP [mil.CZK]

4 000 000

500 000 0,4139

0,3022

0,2503

0,2050

0,1774

2014 0,1375

2012

2008

2010

2004

Year

2006

0

D

ue Val

4 500 000

3 500 000 3 000 000 2 500 000 2 000 000 1 500 000 1 000 000

Level GDP [mil.CZK]

4 000 000

500 000 0,3815

0,3414

0,2741

0,2183

0,1948

2014 0,1608

2012

2008

2010

2004

Year

2006

0

ue Va l

D

Fig. 13.9 Risk parameterisation of factor D in calculation of Yt. The figure above presents three possible risk calculations. The change (risk) in variation of parameter D is ±15% from the average value.


277

4 500 000 Level GDP [mil.CZK]

4 000 000 3 500 000 3 000 000 2 500 000 2 000 000 1 500 000 1 000 000 500 000 0,8541

0,6417

0,4725

2014 0,3445

2012

2010

2008

2006

Year

2004

0

lu Va

e

W

Lt a for calculation of the GDP function. The 1-J W minimal average W parameter is 0,3445, and the maximal average W parameter is 0,9395. The step between calculations is 10%.

Fig. 13.10 Parameterisation of factor W in H t

The productive development of strategies and reliable comparisons of evaluated strategies is possible only on the basis of reliable standard data. It is reasonable to use a fixed standard on the basis of a formalised computer program that enables data and calculation preconditions to be kept in a fixed state. The calculations and recalculations are comparable, and the changes in terms of parameterisations of input parameters are measurable. In other words, vectors and matrices (14), (15), (16), (17) can be constructed.

278



4 000 000 3 500 000 3 000 000 2 500 000 2 000 000 1 500 000 1 000 000

0,5756

2012

0,4691

2010

2012

2010

2014 0,3922

2008

2008

2004

Year

2006

0

0,7395

500 000

Va

e lu

W


4 000 000 3 500 000 3 000 000 2 500 000 2 000 000 1 500 000 1 000 000

0,5529

0,4390

2014 0,2992

2004

Year

2006

0

0,8116

500 000

Va

e lu

W


4 000 000 3 500 000 3 000 000 2 500 000 2 000 000 1 500 000 1 000 000 500 000 0,8013

0,6208

0,4634

2014 0,3711

2012

2010

2008

2004

Year

2006

0

Va

e lu

W

Lt a calculation Yt. The figure above 1-J W presents three possible risk calculations. The risk in parameter W is ±15% from the average value.

Fig. 13.11 Risk parameterisation of factor W in H t

4 500 000 4 000 000 3 500 000 3 000 000 2 500 000 2 000 000 1 500 000 1 000 000 500 000 0

279

0,3328

lu e Va

0,1476 2014

2012

2008

2010

Year

a

0,2250 2006

2004

Level GDP [mil.CZK]


Lt a for calculation of the GDP function. The 1-J W step between calculations is 10%. The minimal average a parameter is 0,1476, and the maximal average a parameter is 0,4026. Fig.13.12 Parameterisation of factor a in H t

280



4 000 000 3 500 000 3 000 000 2 500 000 2 000 000 1 500 000 1 000 000 500 000 0,2973 Va lu 0,3255 e a

0,2317

2014 0,1497

2012

2008

2010

2004

Year

2006

0


4 000 000 3 500 000 3 000 000 2 500 000 2 000 000 1 500 000 1 000 000

0,3940

0,2980

0,2162

2014 0,1545

2012

2008

Year

2010

2004

2006

500 000 0

Va

a

e lu


4 000 000 3 500 000 3 000 000 2 500 000 2 000 000 1 500 000 1 000 000

0,3133

0,2783

0,1885

2014 0,1562

2012

2006

2010

Year

2008

2004

500 000 0

Va

e lu

a

Lt a for calculation Yt. 1-J W The above figure presents three possible risk calculations. The risk in parameter a is ±15% from the average value.

Fig. 13.13 Risk parameterisation of the rate of employment a in H t

4 500 000 4 000 000 3 500 000 3 000 000 2 500 000 2 000 000 1 500 000 1 000 000 500 000 0

281

0,3328

Lt a

Va

lue

J

Fig. 13.14 Parameterisation of factor J in H t

0,1476 2014

2012

2008

2010

Year

0,2250 2006

2004

Level GDP [mil.CZK]


for calculation of the GDP function. The 1-J W minimal average J parameter is 0,1476, and the maximal average a parameter is 0,4026. The step between single calculations is 10%.


4 500 000 4 000 000 3 500 000 3 000 000 2 500 000 2 000 000 1 500 000 1 000 000 500 000 0

0,3131

J lue

Va

2014

2012

2008

2006

0,3253

4 500 000 4 000 000 3 500 000 3 000 000 2 500 000 2 000 000 1 500 000 1 000 000 500 000 0

J lue

Va

0,1535 2014

2012

2010

2006

2008

0,2093

Year

J lue

Va

2012

2008

0,1535 2014

Year

2010

2004

0,3253 0,2093 2006

Level GDP [mil.CZK]

0,1569

4 500 000 4 000 000 3 500 000 3 000 000 2 500 000 2 000 000 1 500 000 1 000 000 500 000 0 2004

Level GDP [mil.CZK]

Year

2010

0,1981

2004

Level GDP [mil.CZK]

282

Lt a for calculation of Yt. The figure 1-J W above presents three possible risk calculations. The risk in parameter J is ±15% from the average value.

Fig. 13.15 Risk parameterisation of sfactor J in H t


283

Fig. 13.16 Application dialog of Cobb-Douglas simulation, options of calculation for a particular region.

The question is how to manage and ensure efficient movement to higher levels of Yt from some starting point Y0. Let us make a model of the efficiency of such a movement created as L in (19). For greater insight, we may create a whole range of simulations on the basis of multiple calculations. More comprehensive information about the transition risk over a period of 5 and 10 years is concentrated in Table 13.7.

284


Table 13.7 Expected trends Yt for simulated data. D

Strategy

D

J

W

Change

Decrease

Increase Decrease

Increase Decrease

Increase Decrease

Increase

2005

nnn

ppp

pp

nn

n

p

n

p

2010

nnn

ppp

pp

nn

n

p

n

p

2014

nnn

ppp

pp

nn

n

p

n

p

13.9 Conclusion The creation of a feasible development strategy for the regions is based on relatively few input variables. Any model is shaped by its structural elements and their causal interconnections. The model used in this article is based on the Cobb-Douglas functions modified and related to Fig. 13.6. A sensitivity calculation of the elements and a study of their risk sensitivity provide interesting information for decisions about appropriate tools and resources for stimulation of further development. Two possible strategies are available: unmixed and mixed. Unmixed strategies have more explanatory potential. Mixed strategies have to be composed according to a utility function. The utility function enables us to evaluate the resources used to give improved parameters. The resources consumed in a time sequence and their utility results are a measure of success, and of the ability to act for future progress. However, the mix can be a very interesting question. The optimal mix may be studied as an optimization task. To obtain some practical solutions, we might use a trial and error method to obtain results for a comparison of various mixed strategies. An optimal strategy is more than an academic exercise. It must also be practical in term of risk acceptance. This introduces more complicated limitations than had been assumed of the beginning of this study in the formulation of an acceptable regional strategy.


285

References Audi M., Šejnoha M., Zeman J., Localisation of inelastic deformation and partition of unity method - one dimensional case. Contributions to Computational and Experimental Investigation of Engineering Materials and Structures. CTU Reports 7 (P. Konvalinka, J. Máca, eds.) 2003, pp 139-160. Beran J., Statistics for Long-Memory Processes. Chapman and Hall, New York 1994. Beran V., Modelling and management 10. University text-book. ýVUT, Praha 1997. (In Czech). Beran V., Modelling and management 20. University text-book. ýVUT, Praha 1999. (In Czech). Beran V. et al., Dynamic Time Charts, Electronic Scheduling of Resources of Technical-Economic Processes. Academia, Praha, 2002. Bayes T., Essay towards Solving a Problem in the Doctrine of Chances, Biometrika (reproduction of 1763 paper) 45 (1958), 293-315. ěeĜicha P., Static and dynamic limit loads of reinforced concrete structures. CTU Prague, Prague, 2000. Beran V., Základy teorie rozhodování, (Foundations of decision theory), ýVUT v Praze 1985. Beran V., Macek D., Programové vybavení Balance Sensitivity, (Software), Praha 1998. Beran V., Macek D., Programové vybavení Fault Cell, (Software), Praha 2000. Beran V., ProstČjovská, Z., Reliability Estimation of Risk Evaluation, In: Advanced Engineering Design AED 2006 [CD-ROM]. Prague: CTU, 2006, vol. 1, s. D2.07-1-D2.07-6. ISBN 80-86059-44-8. Demel J., Graph theory. Academia, Praha, 2002. (In Czech). Fingleton B., Eraydin A, Regional Economic Growth, SMEs and the Wider Europe. England: ASHGATE Publishing Limited, ISBN 0-7546-3613-5, 2003. Jablonský J., Dlouhý M., Models of production unit efficiency evaluation, Professional publishing, Prague, 2004. (In Czech). Jones Charles I., The Shape of Production Functions and the Direction of technical Change, Quarterly Journal of Economics, pp. 517-549, 2005. Kubík S., Kotek Z., Razim M., Hrušák J., Branžovský J., Theory of steering and automation II. SNTL Praha, 1982. (In Czech.). Laguna M., Applying Robust Optimization to Capacity Expansion of One Location in Telecommunications with Demand Uncertainty, in Management Science/ Vol. 44, No. 11, 1998. Mandelbrot B., Die fraktale Geometrie der Natur. Birkhäuser-Verlag, BaselBoston-Berlin, 1991. Mulvey J. M., Vanderbei R. J., Zenios A., Robust optimisation of large scale systems, Oper. Res. 43, 1995. Raiffa H., Schlaifer R., Applied Statistical Decision Theory, Harvard University Press, Cambridge, Mass., 1961.

286


Rektorys K., The Method of Discretization in Time, and Partial Differential Equations. Mathematics and Its Applications, Vol. 4. D.Reidel Publ. Co. Dordrecht 1982. Schumann J., Mayer U., StrĘbele W., GrundzĦge der mikroĘkonomischen Theorie, 1999. Sandelin B., Economy and History, 1976. Tondl L., Evaluation and values (Methodological evaluation dimension). Filosofia AV ýR, Praha1999. (In Czech). Vlþek J., Beran V., Automation and steering systems. SNTL Praha 1984. (In Czech). Wolfram S., A New Kind of Science. Amazon, 2002.

14 Macro Policies and Regional Impacts in Norway

Steinar Johansen Institute of Transport Economics, Norway, E-mail: [email protected]

Abstract. There are many aims for the Government’s policies. Different ministries care for different sector policies. Each ministry administers one (or several) policy sector(s), and normally has a bundle of policy measures to reach their aim(s). Applying a policy measure, directed towards reaching the aim(s) of a policy sector, affects many sectors of the economy, and therefore regional development. The impacts on regional development can be positive or negative, but are normally not intended. Since these impacts are not intended, the impacts on regional policies are externalities of sector polities. In this chapter, I look into a broad set of Norwegian fiscal and non-fiscal sector policies, and discuss their impacts on regional development. The impacts of different policies on regional development are compared and rated according to their significance and magnitude. The chapter illustrates that sector policies, including macro economic policies, influence regional development in different parts of the country differently, and that the non-intended impacts of sector policies are significantly more important for influencing regional development than specific regional and periphery policies. Most significant are the welfare policies, including transfers to local governments and the social security system. Among industrial transfers, agricultural policies have the most significant impacts on regional development. Keywords: Macro policy, Regional development, Sector policy, externalities, Non-intended impacts of sector policies.

This chapter is based on a two-year project, which was published by the Norwegian Government (NOU 2004:2, in Norwegian)

288

Steinar Johansen

14.1 Introduction Several factors influence regional development. One way of distinguishing between these is to separate policy factors from other factors (like prices, costs, market conditions etc). In this section, I look closer into how different sector policies influence regional development. My claim is that the broad sector policies have greater impacts on regional development than the more narrow regional policies. This can be explained partly by the fact that they are quantitatively much larger than the narrow regional policies. I also claim that I can range the different sector policies’ impacts on regional development according to an array of criteria. The section is based on work I have undertaken as a secretary for the Impacts Commission. This Commission was appointed by the Norwegian Government, with a mandate to analyse the sector policies’ impacts on regional development in Norway, and to range these policies according to these impacts. To our knowledge, similar work has not been done anywhere else, at least not with our broad approach. Therefore, experience from previous work could not be applied directly. Our approach had to be ad hoc. The analysis involved several challenges, apart from the main challenge of establishing an overview of different sector policies’ regional and peripheral impacts. These challenges can be summarised in the following questions: a) What policy sectors are relevant to look into? This question involves a screening of different sector policies, in order to distinguish between policies that have important regional impacts and policies that are less relevant for the major question. b) How can we calculate each policy’s impacts for regional and peripheral development? We discussed qualitative as well as quantitative methods for calculating impacts, as well as different indicators, and concluded that the policies are too different to adapt universal methods or indicators for all policies. However, there are certain features of the methods that are universal, of which the most important is to establish a non-factual situation (a situation without the policy) as a basis for comparison. The main requirement for the indicators applied was that they had to be relevant. c) How can we compare the different policies’ impacts for regional and peripheral development? We have discussed several methods for comparing impacts of different policies and concluded that a nonquantitative method is necessary.


289

d) Which policy sectors have the most significant impacts for regional and peripheral development? Since we have compared the policies’ impacts using qualitative methods, the reply to this question will also be of a qualitative nature. Regional (and peripheral) impacts were defined along two dimensions. We have discussed the policies’ impacts for regional growth, industrial development and employment on the one hand (the growth dimension). This is the most relevant perspective compared to the main aims for regional policy in different parts of Europe. On the other hand, distributive measures and provision of welfare services have been very important in Norwegian policies, including regional and peripheral policies. Therefore, we have also discussed the impacts of the sector policies for the regional income distribution, for welfare provision in different parts of the country and for the regional distribution of private and public services (the welfare or distributive dimension). The distributive dimension is relevant especially when discussing the policy impacts in the peripheries, where the economic base is weaker than in the more central parts of the country, and where the public sector’s activities, and more generally public policies, are relatively more important (Johansen 2003).

14.2 Relevant Policy Sectors One might argue that all Government policies influence regional development directly or indirectly. It is relatively easy to see that industrial policies influence industrial, and thus regional, development, and that the Government’s transfers to the local public sector influence local production of public services, and thus regional development. These policy sectors are examples of sectors that influence regional development relatively directly. Some policies do not influence regional development directly. Foreign policies, for instance, are directed towards other countries and do not influence regional development in Norway directly. Indirectly, however, these policies can influence regional development via the Government’s budget balance and the resources allocated for other policies (including polices that influence regional development). The first screening mechanism adopted, was to exclude those sectors that a priori only have these kinds of indirect impacts on regional development. For instance, none of the policies administered by the Ministry of Foreign Affairs were considered relevant for our analysis. The second step of the screening process involved grouping the different types of policies into relevant categories, given that we only looked at

290

Steinar Johansen

central Government policies. There are several ways of categorising policies, of which we chose three, that are overlapping in the sense that one policy sector automatically falls into at least two (normally all three) categories: 1. Means: There are several types of policy means. We have defined three major types, namely (i) money (subsidies, taxes etc), (ii) rules and regulations, and (iii) localisation of public activities and public purchases from the private sector. Within one policy sector, one or a combination of several means might be applied. 2. Purpose: All sector policies are designed to fulfil certain purposes or aims. The purposes might be defined relatively narrowly, but normally (and at least rhetorically), the policies should not be in conflict with general aims. Examples of general, or cross sector, aims are sustainability, efficiency, gender neutrality and regional development. Traditionally, regional (or peripheral) development has been a spinoff from many sector policies. 3. Impacts: Many sector policies have negative (non-desirable) or positive (desirable) impacts on regional development, in addition to their impacts on the sector’s aims. This categorisation of sector policies was used as an aid when step two of the screening process was completed. In order to find the relevant sector policies, that is the policies which are significant for regional and peripheral development, we first looked at the different polices’ purposes (category two) and tried to establish if they have regional or peripheral development as an important aim. Second, we tried to distinguish between sectors with a priori major and minor impacts for regional and peripheral development (category three). Policy sectors which were found to give priority to regional and peripheral development as parts of their purposes, or which were thought to have more than minor impacts on regional and peripheral development, were included into the further analysis. When the sector policies had been defined in this way, we also defined the types of policy means applied within each policy sector (category one). This gave us a heterogeneous set of different, but relevant, policy sectors, which we looked further into. We analysed the regional and peripheral impacts of macro economic policies and of 19 categories of sector policies, and the impacts of all policies in seven different regions. 14.2.1 Macroeconomic Policies We analysed the regional impacts of future macroeconomic development, given some assumptions on international economic development, the activ-


291

ity in the petroleum sector, demographics, the level of the Government’s budget surplus and so on. In addition, we looked further into the regional impacts of a balanced budget change of the public sector’s consumption and of income taxes, and of the impacts of changes in the interest rate. 14.2.2 19 Sector Policies We looked into the regional impacts of 19 different sector policies. They could be further categorised into four main areas of policy, namely x Policies directed (mainly) towards the citizens: transfers to the public, production of public services, labour market policies, policies for higher education, cultural policies and housing policies. x Policies directed (mainly) towards growth and industrial development: agricultural policies, cheap energy policies (mainly directed towards high-energy manufacturing industries), shipbuilding policies, sea transport policies, investment support (national support schemes and support for industries in the peripheries), use of licenses, quotas and restricted access to certain markets, and regionally differentiated tax on labour use. x Infrastructure policies: Building and running infrastructures, and public purchases of transport services in the wide sense. x Localisation of Government agencies: This has impacts mainly at the locality of the Government agency, and the agency can be seen as a base industry for the locality’s economy. The focus of the analysis was to look at the impacts of each of the 19 sector policies. Grouping them into four main categories is done mainly for increasing the oversight of the policies and their impacts. Some of the policies could also be placed in other groups. For instance is there a discussion whether the regionally differentiated (employers’) tax on labour use is directed mainly towards the citizens (reduced employers’ tax implies higher wages), or if it is directed mainly towards growth and industrial development (as a labour subsidy). Cappelen (2003) and Hervik and Rye (2002) discusses this question further. All 19 policy sectors, of course, have specific aims, in the sense that they are directed towards specific sector purposes. This is reflected also in the means applied in each case. We do not discuss whether the mean applied in each case is the optimal mean for reaching a given policy target.

292

Steinar Johansen

14.2.3 Government Policies in Seven Selected Regions In addition, we analysed the impacts of many (most) Government policies in seven regions. The purposes of this part of the analysis were twofold: x We wanted to see to what extent many (most) Government policies together influence regional development. x We wanted to see to what extent regional characteristics are important for the impacts of Government policies. This part of the analysis is not pursued in this section. The question is discussed in some detail in Chapter 7 in NOU 2004:2, and further references to all seven regional analyses are made there. The chapter concludes (not surprisingly) that the sum of Government policies influences regional development significantly in all seven regions, that regional characteristics are important for the impacts of Government policies, and the impacts of Government policies are (relatively) more significant in the periphery than in urban areas.

14.3 Methods for Calculating, Comparing and Ranging Impacts The focus of this section is the regional impacts of sector policies. Norwegian regional policy has, since the War, been built on sector policies (Soderlind 1999). One might say that the Government has had several goals for the sector policies, primarily the sector goal, and secondarily regional policy goals (growth and equity). For many reasons, sector policies have gradually become more focused and aimed solely at the sectors’ problems during the last 20 years or so. A sector policy’s success (or fiasco) is now measured and evaluated according to which extent it satisfies the aim(s) for the policy or not, as opposed to the post-war situation when the policy aims were more complex and diverse, and when it was politically correct to apply more than one goal to each policy measure. One might discuss whether it is possible to achieve several aims applying one policy measure. Johansen’s (1966) view is that this is not possible – you cannot have more aims than policy measures. At the same time, we do know that applying sector policies have impacts, and therefore also regional impacts. The regional impacts can be “positive”, in the sense that they are in thread with the aims for regional policies, or they can be “negative”, in the sense that they oppose aims for regional policies. We have argued that the sector policies are aimed more directly towards the sector’s aims today than some years ago. These single-targeted poli-


293

cies, it might be argued, are more efficient than the multi-targeted policies of previous years, since the measures can be tailor made for each aim, and since they therefore can be used for increasing political efficiency for each sector policy. In many ways, these “new” policies are in thread with some more general processes in society, like globalisation and liberalisation. The deregulation of public sector activities, like welfare production and infrastructure provision, contributes to weakening the public sector’s position as a policy provider as more and more allocation responsibilities are left for the market place. At the same time, organisational reforms within the public sector imply decentralisation and increased responsibilities for lower level bureaucrats. In order to secure political control, the policy measures have to be tailor made for the aims. The success or failure of multi-targeted policies is much more difficult to assess (NOU 2004:2). Regional development is influenced by most sector policies, as we have discussed earlier. The movement from multi to single-target sector policies have implications for regional development. The question is whether single-target sector policies can be co-ordinated in order to achieve “more” regional development. This question can be answered partly by defining what sector policies that have the greatest impacts on regional development. Therefore, we have analysed the regional impacts of sector policies, sector by sector, compared each sector’s impacts, and ranged the sector policies according to their regional impacts. 14.3.1 Methods for Calculating Impacts Regional development S (for instance: employment) can be defined as depending on a vector of n policies (P) and a vector of m other conditions (A): S = F(P, A) Here, F represents the functional relationship between P and A on the one side, and regional development S on the other side. By changing one of the regional measures, the vector P = (p1,…pi,…,pn) changes to Pi’ = (p1,…pi’,…pn). The equation becomes Si’ = F(Pi’, A) By comparing S and Si’ (Si’ – S) we can calculate the impacts of the policy for regional development. This can be done for each of the n policy measures in P, and we will get n impacts (Si’ – S). These n impact situations each represent the impacts for regional development of a marginal change in each of the measures. In principle, this model covers all situations where we look at marginal changes in policy measures. In practice, however, the model cannot be (di-

294

Steinar Johansen

rectly) applied to our situation. Some of the questions that could be arisen regarding this model, are: x What is the variable S? Regional development can certainly be defined in many ways. The variable S, or more correctly, the impacts (Si’ – S), can be a number of things (production, employment, demographics, income etc), which all represent regional development. Can we quantify the impacts along the same (set of) variables? The answer is no. x The vector P represents the different policy measures. Are we able to measure all of them along the same scale, so they can all be a part of the same functional relationship? More precisely, can they all be quantified? The answer is, of course, no. x The functional relationship F, and the exogenous variables A, together represent the framework for calculating regional impacts of policy changes. Are we able to construct such functions, that we are able to foresee the impacts of changes? What about very complex relationships, for instance if the impacts of several measures are connected together? Do we have a theory that can be represented in F( )? x Are we interested in marginal (Si’ – S) or total impacts of a policy measure? Each of them can be equally important. The marginal impacts of a measure can be large, but the size of the measure can be small, or opposite – the marginal impacts of another measure can be small, but the size of the measure can be great. In principle, of course, there is nothing wrong with making an equation like F( ). However, it serves more illustrative than practical purposes. In practice, we will measure impacts in many different ways. We have defined regional development relatively broadly (many variables), and we can measure some impacts quantitatively. Other impacts are we unable to measure along the same scale. This problem is probably not solvable, and therefore we have adopted an array of methods for calculating impacts. 14.3.1.1 Non-Factual Situation

We know what the world looks like today. The world as we know it is a complex function of a number of variables, including policy measures. For all policy sectors, we try to compare today’s world with a non-factual situation – or what we think the world would look like without the policy measure in question. There are several problems connected to constructing a non-factual situation, but these problems have to be addressed when discussing impacts. The non-factual situation is heavily discussed within evaluation theory and methodology, and is also addressed within planning theory and impact assessment theory, see for instance Teigen (2003). Since we are unable to use experiments within social sciences, we have to use


295

other methods. The methods vary from variable to variable, from measure to measure, and from discipline to discipline. 14.3.1.2 Model Simulations

In some cases, model simulations can be applied. These are especially relevant when economics, growth, industrial development, employment and demographics are important variables. Using simulation models implies that we trust the equations and variables, as well as the data and coefficients, in the model. 14.3.1.3 Compare two Situations

In some cases, development within regions with and regions without the actual measures can be compared, and experience be learned from that. It is difficult to distinguish between regional characteristics and impacts from policy measures in this case. This method has, however, been applied for instance when discussing the impacts of a bundle of measures in the Northernmost parts of Norway (Hervik et al 2002), who claim that the method helps in providing significant results. Similar methods, with similar problems, can be used when comparing the situation in an area before and after the measure has been put into effect. Is it the measure, or other factors, that explain development? 14.3.1.4 Questionnaires and Other Types of Questions

In many cases, questionnaires, telephone interviews or other ways of interviewing those who benefit from (or loose from) the measure in hand, can be adopted. Often, the replies become self evident, for instance in the sense that actors receiving benefits answer that the measure is good, while the actors loosing answer that the measure is useless. This method has been adopted in several cases when the Norwegian Regional Support Authority (which has changed names several times the past 15 years) has been evaluated (see NOU 2004:2 or Teigen 2003). 14.3.1.5 In-Depth Interviews

In-depth interviews with experts and/or actors can be used to discuss impacts of policy measures. This method is commonly used within evaluations as well as within planning and impact assessment. It is probably more relevant for discussing processes and probable impacts, than for actually isolating impacts of one measure.

296

Steinar Johansen

14.3.1.6 Non-Empirical (Theoretical) Models

Different social sciences have different theories for how society works, and there are also different schools within each discipline. This implies that there is a wide array of different views on what is important to analyse, as well as on what the impacts of policy measures are. Theoretical considerations and discussions are very important when discussing impacts of policy measures, and the different views illustrate that these are not easy to agree upon. 14.3.2 Methods for Comparing Impacts Since the above differential (Si’ – S) in principle can be adapted for different policy measures, this differential will provide us with the marginal impacts of each of the n measures discussed. Were these differentials comparable, they could naturally be used for comparing impacts of the n measures. However, as I have discussed above, no such luck exists outside the principal equation. We therefore have to look elsewhere for methods that can be used for comparing impacts. In order to compare impacts, it is necessary to present impacts within the same framework, or to make them compatible. 14.3.2.1 Cost-Benefit Analyses

Economists would probably claim that cost-benefit analyses (CBA) provides us with a method that not only gives us the net benefits of different projects and programmes, but also provides us with the means to range them, given that we are able to quantify all costs and benefits. This method could also be applied to policy impacts analyses. However, the problems of quantifying costs and benefits are difficult to overcome, and we therefore have to look elsewhere. 14.3.2.2 Planning and Impact Assessment

From planning theory and environmental impact assessment, we have certain methods that can be applied when comparing impacts. These methods have that in common that they are qualitative, and that there are many considerations and many target variables that have to be discussed at the same time. Hilding-Rydevik et al (2004) is a recent publication that discusses SWOT-analysis, Strategic Environmental Assessment (SEA), Sustainability Assessment (SA) and Territorial Impact Assessment (TIA) as different methods that might be applied. These methods are adapted for instance in


297

the European Spatial Development Perspective (ESDP), in environmental impact assessment and, more generally, in planning. The methods have also been used for policy analyses, for instance within the Norwegian Peripheral Impact Assessment, see AAD (2000). These methods are originally developed for analysing impacts of single events (or policy measures), and they provide us with a multi-targeted analysis for each measure. This analysis can be compared across measures. We made our own multitargeted table of analyses, tailored for comparing regional impacts of sector policies. The following variables (targets) were discussed for each policy measure: x Type of measure: We applied three categories: Money (subsidies, taxes), legal and localisation of public sector or public sector purchases. Each measure was defined in at least one of these categories. If it was a question of money, we tried to quantify how much. x Impacts, tendencies: We applied two broad categories, namely industrial development (including value added and employment), and welfare provision (including accessibility to public services and other equity measures). In the cases we knew something about quantity of the impacts, these were included. x Political control: In a changing world, some measures are left more and more to the market. In addition, some measures are automatic (for instance social securities), while some are controlled by politicians. Whether the politicians can choose to adapt a measure or not, is very important for the control of regional impacts. x Type of region: We defined what type of region the impacts were most significant (Local, Regional, Periphery, Centre). x Marginal view: Here, we discussed whether we would give higher or lower priority to each policy sector in order to achieve more regional development. This part of the analysis includes both potential marginal impacts on regional development and potential impacts (efficiency losses) on sector aims. In quantitative measures, we could call this a cost-benefit analysis. Again, these five variables (targets) represent the regional development perspective, which is not a part of each policy’s primary targets. Some of the measures, like the differentiated employers’ tax on labour and the regional investment incentives have regional aims. Others, like the Government’s transfers to local and regional authorities, and the agricultural policies, have both regional and sector aims. Most policies, like macro-

298

Steinar Johansen

economic policies, tax policies and military defence, however, are targeted at achieving sector aims. Therefore, when we discuss a policy’s (relative) success in regional development, we do not necessarily evaluate the measure’s success at achieving sector political aims. This has to be kept in mind in the discussion. 14.3.3 Methods for Ranging Impacts The five variables (targets) above give us a basis for comparing different policy measures’ impacts on regional development. However, the multitargeted analysis itself does not provide us with the means of ranging the impacts from “small” to “great”, since the variables are not directly comparable. Instead, we have ranged the impacts according to the table below. Table 14.1. Method of ranging impacts

Political control Small Great

Impacts for regional development Small

Great

Important area. Can Strategic area it become strategic? Autonomous area, Less important area with its own argu- to focus on ments

The different policy sectors are put into one of the cells of this table, which uses the multi-targeted analysis in the previous section for ranging the policies according to their impacts for regional development and according to the degree of political control one has over each policy. The table is made relatively simple, in order to stress the major points of ranging the impacts. x Political control By “political control” we mean to what extent politicians are free to decide the contents of each policy sector, when it comes to type of measure and how large the measure can be. In principle, we can say that politicians decide a lot, since they control legal measures as well as fiscal. If this is true, all policy sectors would be to the right in the table. Others would claim that it doesn’t matter to what extent the politicians control a measure, since almost everything is allocated in the market anyway. If this is true, all sectors would be to the left in the table. In the ranging process, we did not take any of these extremists’ views.


299

Instead, we tried to think what means are automatically distributed (for example old age pensions), what means are regulated from international agreements (for example the differentiated employers’ tax), and what means can be decided more freely. x Regional impacts We have distinguished between two categories, small and great. Here, both absolute and marginal impacts are included. Some sector policies would have great impacts because of its size, while others could be very important marginally. In the analysis, we have been more focused on periphery problems and equity, than on growth and industrial development. However, both focuses have been included. The two headings of the table are combined into four cells, which give us four categories for ranging the impacts of sector policies. Each sector policy is placed in one cell only. We have the following combinations: 14.3.3.1 Great Impacts, Great Control

This combination is desirable because the sector policies placed here by definition have important impacts on regional development, while at the same time being politically controllable. 14.3.3.2 Great Impacts, Small Control

This is also an important area, because impacts of the policies placed here are significant. However, for one reason or another (for instance EU competition policies) the political control of these means is small. The question is whether it is possible or desirable to seize political control of these means, in order to use them for achieving better regional development. 14.3.3.3 Small Impacts, Great Control

This is not a very important area for regional development. However, the political control of these policy measures is great. 14.3.3.4 Small Impacts, Small Control

We have named these policy sectors autonomous, in the sense that they are not very important for regional development, and that they are not very controllable.

300

Steinar Johansen

14.4 Impacts of Policies: Most Important Sectors The table in the previous section was applied when ranging different sector policies according to their regional impacts and their political controllability. I will, however, not go through and range all sector policies in detail in this section. NOU 2004:2 presents a comprehensive discussion of the arguments for ranging the sectors the way we did. Instead, I will provide the reader with some of the major points that were made by the Commission, based on the analysis presented in this section. 14.4.1 The Government Re-Distributes Resources Regionally The Norwegian Government is a major re-distributor of resources. Redistribution is made along several lines, of which one is from centre to periphery. Ørbeck (2003) shows that Government spending ranges from 54.000 to 135.000 NOK per inhabitant on the municipal level. This includes all Government spending; transfers to the population and industries, wages paid to Government employees, and transfers to regional and municipal authorities, of which transfers to the people is the most significant (50 % on average). The regions receiving most Government support are located in the periphery. There are two main reasons for this. First, these regions get more money than other regions because they are peripheries. Second, the industrial and demographic structure varies significantly between the centre and the peripheries, and the support schemes give priority to periphery structures. The re-distribution of income from central to periphery parts of Norway illustrates that the Government is a major contributor to regional cohesion in Norway. However, the trend is negative in the sense that the sector policies become more and more single-targeted, while aims for regional development and regional equity become less important. 14.4.2 Macroeconomic Strategies Influence Regional Development Macroeconomic development is important also for regional development. The choice of macroeconomic strategies has regional impacts, mostly due to structural differences between regions. These impacts are normally not important when the strategies for macroeconomic policy are chosen.


301

Increased public expenditure implies more resources to the peripheries if the regional distribution of public expenditure is unaltered. In addition, the Government can control the regional distribution of the public sector. An expansive public budget in this sense therefore gives more degrees of freedom for the Government, which can be used for increasing their activities in, or transferring money to, the peripheries. The opposite will be the case in times of decreased public expenditure. Decreasing taxes (which is also expansive), will on the other hand imply increased private incomes. Private incomes will benefit people living in the central parts of the country more than those living in the peripheries. In addition, increased private consumption is directed more towards private services than towards manufacturing products (see Cappelen og Stambøl 2003), and therefore favours industries located centrally more than those located in the peripheries. It is also more difficult to control the regional distribution of tax cuts than of public expenditure. In times of contractive fiscal policies, the impacts of tax increases will be smaller in the peripheries than in central parts of the country. Public policies directed towards interest rate cuts, which we have had in Norway for several years now, have many similar regional impacts to the regional impacts of tax cuts. This illustrates that the choice of strategy for macroeconomic policies, as well as the instruments chosen, influence regional development differently. It also illustrates that the Government more easily can choose to focus on regional development in the peripheries within some macroeconomic strategies (increased public expenditure or increased taxes) than others. 14.4.3 Important Other Sector Policies Of the 19 specific sector policies that have been thoroughly discussed in NOU 2004:2, I would like to focus on the couple of sectors that the Commission found to be most important for regional development. 14.4.3.1 Transfers to Local and Regional Authorities

Many of the public services that represent the welfare state in Norway are produced by local and regional authorities, who have been delegated this responsibility by the Government. One of the requirements of welfare provision is that each service should at least meet a minimum standard as defined by the Government, irrespective of where in the country it is provided. The Government recognises the differences in local tax bases, and

302

Steinar Johansen

an important source of income for the local and regional public sector is therefore transfers from the Government. Local and regional authorities’ provision of services are therefore partly financed through local taxes, partly through transfers from the Government. Norwegian municipalities are varying in size. In periphery municipalities, private provision of welfare services is very difficult because the market is too small. The Government therefore recognises that municipalities in the peripheries need to provide more services than central municipalities – in other words, the local and regional public sector must be greater in the peripheries than in central areas. One of the most powerful regional policy tools is therefore Government transfers to the local and regional authorities. This has the obvious impact, that welfare services of good quality are accessible all over the country, and irrespective of market size and local tax base. In addition, this is very important for employment, especially in the peripheries. Transfers to local and regional authorities therefore have a double impact on regional development, in addition to being controllable by the authorities. Whether transfers to local and regional authorities are economically efficient tools for regional development or not, is another question (see Borge 2003). A decentralised public sector, with many small municipalities and relatively small regions, might be expensive to run, given national standards1 of welfare provision. Small regions and many municipalities, however, imply public activities in larger parts of the country, which in turn is very important for regional development in the peripheries. There is, in other words, a conflict between regional policy aims and efficiency in welfare provision. 14.4.3.2 Agriculture Policies

There is one private production sector that is relatively evenly distributed among Norwegian regions (outside the cities and towns), and that is agriculture. At the same time, agriculture is the only industry that still is regulated through different policy measures, and where regulation means more for resource allocation than markets and prices. Today’s production structure within agriculture is a function of many years of regulation, to which the farmers seem to have adopted in a ra1

In this section, we have focused on transfers to local and regional authorities. The national standards are, of course, a very important feature of the welfare policy and a crucial factor when discussing regional impacts and political control in this sector. A detailed analysis of more sides of these policies can be found in Berg, red (2003).


303

tional manner. The measures used today can be broadly divided into budgetary support (production and non-production subsidies) and nonbudgetary support (import regulations), which sums up to a total of around NOK 20 billion, of which around NOK 12 billion is budgetary support. Value added within agriculture is, in general, negative, and loosening the regulations will probably result in a deterioration of agriculture all over the country (Vårdal 2003) and increased imports of agricultural products. Agriculture is located in the peripheries, and as shown in Vårdal (2001), the sector is an important employer in many municipalities. Reducing the subsidies would reduce agricultural production and employment, and have significant negative impacts in regions where alternative sources of employment are scarce. Since value added in agriculture is negative, it would probably be more economically efficient to allocate agricultural resources to (any) other sectors of the economy. On the other hand, these sectors will probably be located in central parts of the country. There is, therefore, a conflict between agriculture’s significance as an important production sector in the peripheries and economic efficiency. Another question is the question of political control of agricultural policies. Many international trends, for instance the World Trade Organisation’s (WTO’s) negotiations on a freer trade and the agricultural policies of the EU, influence the control Norwegian authorities have on agricultural policies. We have also seen some changes in agricultural policies, where market in-efficiencies and non-market goods (called multifunctionality, see Vårdal ed 2001) are being used as arguments for keeping the national control of agricultural policies. 14.4.3.3 Employers’ Tax on Labour

Evaluations show that the regionally differentiated employers’ tax on labour has been a very important instrument for promoting employment in the peripheries, both in private and public sectors (Hervik and Rye 2002). The differentiated tax has promoted the use of labour in the peripheries by making labour cheaper, which is a benefit for any use of labour in the peripheries compared to central parts of the country, and especially for labour intensive production (services) in the peripheries. The measure has worked just the way it was supposed to, according to evaluations. This differentiation is no longer allowed under EU (EEA) rules of competition, since it is regarded as a subsidy of production. Therefore, the Government does not have this policy measure as an option anymore.

304

Steinar Johansen

14.4.3.4 Investment Support in the Peripheries

It is not possible to support the use of labour in the peripheries, under EU legislation. However, support can be given to certain investment schemes. Evaluations show that investment support schemes have contributed to increased employment in the peripheries (see NOU 2004:2), and that these schemes must be considered successful given the aim of promoting regional development in an efficient manner. There are two factors that should be pointed out when discussing these measures. Firstly, the total support is relatively small compared to other measures (in excess of NOK 1 billion), and the support has been reduced during the 1990s and the beginning of the 2000s. Secondly, the Norwegian Government’s maximum support limits are significantly lower than the limits that are applied in the EU. In other words, these measures could be increased with efficient impacts in the peripheries and within the limits set out by the EU. 14.4.3.5 Other Policies

I have discussed the most important policy measures, given their regional impacts and possibilities of political control. There are several others that could be mentioned, like infrastructure provision, education, energy policy and so on. Each of them are important in their own way. The report NOU 2004:2 discusses them in more detail, and I would like to refer to this publication for further discussion.

14.5 Concluding Remarks In this chapter, I have pointed out some of the problems of measuring impacts of policies, and comparing these impacts. However, we have been able to draw some conclusions within the framework applied. In order to repeat the relative success of regional policies after the war, there is a need for co-ordinating policy measures. The Government’s coordination has been reduced, even though the Government does have measures that can be applied. One of the questions asked in NOU 2004:2 is therefore whether regional authorities could be given coordinative powers, and be more able to promote regional development in each region than the central Government is. Regionalisation is an interesting issue, which will not be discussed further here. I will only point to the fact that regionalisation is an issue that is very much on the agenda, both in the EU and in Norway, and that the fu-


305

ture structure and number of regions, and the regions’ responsibilities, are issues that are continuously debated in Norway.

Literature and References AAD (2000): Utredningsinstruksen. (Instructions for contents of white papers presented to the Parliament). In Norwegian only. Arbeids- og administrasjonsdepartementet, Oslo. Berg, P O red. (2003): Regionale og distriktspolitiske effekter av statlige tiltak overfor kommunesektoren. (Regional and Periphery Impacts of Policies Directed Towards the Municipalities). In Norwegian only. See www.effektutvalget.dep.no. Borge, L-E (2003): Kommunefinansieringens rolle i regionalpolitikken. (Municipal Financing and Regional Policy). In Norwegian only. See www.effektutvalget.dep.no, in Berg (2003) Cappelen, Å. og L. Stambøl (2003a): Virkninger av å fjerne regionale forskjeller i arbeidsgiveravgiften og noen mulige mottiltak (Impacts of Removing Regional Differences in the Employers’ Tax, and Possible Counter Measures). In Norwegian only. Notater 2003/31, from Statistics Norway Cappelen, Å. og L. Stambøl (2003b): Regionale konsekvenser av makroøkonomiske utviklingstrekk og politikk. (Regional Impacts of Macroeconomic Trends and Macroeconomic Policy Strategies). In Norwegian only. See www.effektutvalget.dep.no. Fingleton, B., A. Eraydin and R. Paci, eds. (2003): Regional Economic Growth, SMEs and the Wider Europe. Ashgate, Aldershot (UK) and Burlington (US). Hervik, A. og M. Rye (2002): En empirisk analyse av differensiert arbeidsgiveravgift som regionalpolitisk virkemiddel (An Empirical Analysis of Regionally Differentiated Employers’ Tax as a Mean for Promoting Regional Development). In Norwegian only. Notat til Effektutvalget www.effektutvalget.dep.no. Hervik, A. et al (2002): Bedre klima for tiltakssonen? (Impacts of a Bundle of Policy Measures in the Northernmost Parts of Norway). In Norwegian only. Rapport 0201, Møreforskning, Molde. Hilding-Rydevik, T. et al (2004):.Tools for Sustainable Regional Development: Experiences and Prospects. Nordregio Report 2004:4, Stockholm Johansen, L (1966): Offentlig økonomikk. (Public Economics). In Norwegian only. Universitetsforlaget, Oslo. Johansen, S (2003): Norwegian Regional Development and the Role of SMEs. In Fingleton et al, eds (2003). NOU 2004:2: Effekter og effektivitet. Effekter av statig innsats for regional utvikling og distriktspolitiske mål. (Effects and Efficiency. Impacts of Government policy on regional development and periphery policies). In Norwegian only. Kommunal- og reigonaldepartementet, Oslo.

306

Steinar Johansen

Soderlind, S. (1999): The Ascent of Regional Policy in Norway, 1945-1980. NIBR Working Paper 1999:128, Oslo Teigen, H. (2003): Effektar av den smale distriktspolitikken. (In Norwegian only. See www.effektutvalget.dep.no. Vårdal, E. ed. (2001): Multifunctionality of Agriculture – Seminar Proceedings. Department of Economics, University of Bergen. Vårdal, E. (2003): Støtten til distriktsjordbruket sett i et samfunnsøkonomisk perspektiv. (Agricultural support in the peripheries and economic efficiency). In Norwegian only. SNF-arbeidsnotat nr. 13/03, Bergen. See also Vårdal, ed (2001), for a discussion of Norwegian Agricultural Policies. Ørbeck, M. (2002): Statlig budsjettpolitikk og regional utvikling. En kvantitativ analyse for Effektutvalget. (Regional Development and Government Budgets). In Norwegian only. Rapport 14/2002, Østlandsforskning, Lillehammer.

15 The Economics of Tree-planting for Carbon Mitigation: A Global Assessment

Pablo C. Benítez1, Ian McCallum2, Michael Obersteiner2 and Yoshiki Yamagata3 1

Department of Economics, University of Victoria, Canada, E-mail: [email protected] International Institute for Applied Systems Analysis, Austria 3 National Institute for Environmental Studies, Japan 2

Abstract. This article provides a framework for identifying least-cost sites for carbon sequestration through tree-planting and deriving carbon cost curves at a global level in a scenario of limited information. Special attention is given to country risk considerations and the sensitivity to spatial datasets. Our model results, illustrated by grid-scale maps, show that most least-cost carbon uptake projects are located in Africa, South America and Asia. By comparing emissions reductions through tree-planting with the emission abatement scenarios of integrated assessment models (RICE-99) for a 100-yr time span, we find that global carbon uptake of planted forests could represent between 5% to 25% of the emissions reduction targets of relevant climate change mitigation scenarios. Keywords: Carbon Sequestration, Carbon Cost Curves, Country Risk Considerations.

15.1 Introduction There is plenty of scientific evidence that increases in anthropogenic emissions of greenhouse gases (GHG) are warming the earth (IPCC, 2001). Latest predictions of the International Panel of Climate Change (IPCC) suggest that by 2100, human-related GHG emissions would cause the global average surface air temperature to increase by 1.4 - 5.8°C, and the average sea level to rise between 8 and 88 cm (IPCC, 2001). This warming would vary be Financial support of the BIOCAP Canada Foundation is gratefully acknowledged.

308

Pablo C. Benítez et al.

tween regions causing diverse impacts on human health, agriculture, forestry, infrastructure, water resources, energy supply, industry and biodiversity. Tropical regions would be more affected by a decrease in agricultural production, while temperate regions would face the expansion of vectorborn diseases like malaria and dengue fever, and would confront higher temperatures as well as more frequent heat waves during summer. Natural systems like coral reefs, mangroves, tropical and boreal forests, polar and alpine ecosystems, and prairie wetlands are at risk of irreversible damage and the loss of vulnerable species (IPCC, 2001). Facing the threats of global warming and the costs of adaptation to be borne by future generations, mitigation measures have been proposed within international agreements like the Kyoto Protocol (UNFCCC, 1998). As a general classification, mitigation is divided into two groups: (1) the reduction of GHG emissions in energy, waste management and industrial sectors, and (2) the enhancement of carbon sinks. For the first group of activities, carbon mitigation cost curves have been estimated based on global energy models (Gritsevskyi and Schrattenholzer, 2003; Sijm et al., 2000), but to a lesser extent, carbon cost curves are available in the sink sector. As an imperative need for finding least-cost mitigation alternatives, we aim to estimate carbon sequestration cost curves through afforestation and reforestation (AR-projects) at a global level and determine sites where these costs are at a minimum. A number of studies have been described in the literature providing estimates of carbon sequestration costs through AR-projects in major countries (e.g. Stavins, 1999 for the US; Xu, 1995 for China; Fearnside, 1995 for Brazil; and de Jong et al., 2000 for Mexico), whereas global studies have been limited and performed at high aggregation levels. For instance, Sohngen and Mendelsohn (2003) have used a timber supply model for estimating cost curves for carbon sequestration globally. For this analysis, they divided the world into few large regions and have not been able to provide details on how such costs are distributed geographically. Contrasting the previous studies and as a new research contribution, we estimate global AR carbon supply curves by using information at a disaggregated level, and scrutinizing the potential tree-planting area so that sequestration costs are estimated at geographically explicit cells of approximately 50 × 50 km. Being aware of the effect of country considerations associated with political, financial and economic risks, we evaluate its influence on the global supply of carbon. For estimating carbon sequestration costs for each cell, we use a uniform method that allows worldwide comparisons. This article is structured as follows. We first describe the model in Section 2. Global datasets and parameters used for our analysis are provided in

15 The Economics of Tree-planting for Carbon Mitagation

309

Section 3. Results are shown and discussed in Section 4. Conclusions are drawn in Section 5.

15.2 Carbon Sequestration Costs The analysis starts by creating a homogenous geographical grid (with a gridcell size of 0.5 degrees) for the whole study area and selecting cells that are suitable for AR-projects, i.e. non-forest areas where tree-planting is viable and will not compromise food security of the region. We then estimate sequestration costs for each cell based on estimates for inter alia biological growth, plantation costs, expected timber and land prices, and carbon storage in products. Finally, we obtain the cumulative sequestration cost-curve by aggregating results, taking into account that AR-projects would occur only in cells where the carbon price exceeds sequestration costs. Besides obtaining the cost-curve, the method identifies the geographic distribution of carbon costs and growth potentials throughout a region. The methodology is fully described in Benítez and Obersteiner (2005) with a short summary of major model equations provided here. The sequestration decisions are made for each cell by considering the profitability of AR-projects vis-à-vis the current agricultural practice, i.e. the net present value of forestry evaluated for an infinite time period including payments for carbon sequestration (Fi) is required to be larger or equal to the net present value of agriculture (Ai):

Fi t Ai .

(1)

Net present value of forestry (f) in cell “i” during one rotation interval is:

fi

- cpi

pwV i i Bi ( 1 ri ) Ri

(2)

where ri is the risk-adjusted discount rate, cpi are planting costs, pwi stumpage timber price, Ri rotation interval, Vi timber volume and Bi present value of the carbon benefits over one rotation. We consider carbon uptake as a positive externality and its benefits are a function of the rate of change of biomass or timber volume over time (van Kooten et al., 1995; Creedy and Wurzbacher, 2001). In a similar fashion, carbon release during harvest is a negative externality and its costs are a function of the amount of biomass or timber volume removed from the forest. In practice, not all the carbon removed from the forest after harvest is immediately released to the atmosphere, however a fraction, ș, is

310


stored for longer time periods in the atmosphere (van Kooten et al., 1995). Storage after harvest could take place in a wide range of products including long-lived products like furniture, structures, construction materials and thick branches, and short-lived products like paper, leaves and thin branches. This broad definition for “forest products” is used throughout this study. Including forest biomass and products, the equation for the present value of carbon benefits is (see Benítez and Obersteiner, 2005): Ri

Bi

pci ¦ t 1

Zi t

1 ri

(3)

pciZi Ri pciTiZi Ri Ri (1 ri ) (1 ri ) Ri

where pci is the carbon price and Zi is the yearly rate of carbon uptake (linear forest growth is considered). The term și depends on the fraction of the forest biomass stored in long-lived products and the decay rates of shortlived products and long-lived products. The net carbon benefits of ARprojects are the ones that provide additional carbon storage in the biosphere as compared with the original land use. This requires subtracting the carbon level in the so-called baseline of the project (IPCC, 2000). In our analysis, we consider that the carbon stored in the baseline represents a fraction bi of the carbon stored in the forest. We call bi the baseline factor. By summing up carbon benefits in biomass and products and subtracting the carbon in the baseline, we get the final expression for total carbon benefits: 1 B pc Z (1 b ) r ª1 (1 r ) Ri º R (1 T )(1 r ) Ri . (4) i

i

i

i

^

i

¬

i

¼

i

i

i

`

By means of equations (2) and (4) we estimate net present value of forestry for one rotation interval (fi); and from this, we obtain net present value for an infinite number of rotations (Fi). Given constant prices and fixed rotation intervals we have: 1 (5) Fi f i ª¬1 (1 r ) Ri º¼ . The net present value of agriculture or land price function is estimated by assuming a two-factor Cobb-Douglas production function. The first factor is suitability for agriculture, Si, and indicates the aptness of the land for agricultural production given its endowments of soil and ecosystem properties. The second is population density, Di, and is a proxy for labor intensity and infrastructure. When output follows a Cobb-Douglas function, net present value of agriculture has the following functional form: Di

Ai Xi Si Di

Ji

(6)


311

where Xi Di and J i are the parameters for the production function. When we set Ai=Fi, we find the minimum carbon price (we define this as the carbon cost) that allows forestry to be as profitable as agriculture:

pci

R Ri Ai ª1- 1 ri i º cpi - pwV i i (1 ri ) ¬ ¼ . Zi (1 bi ) ri 1 ª¬1 (1 ri ) Ri º¼ Ri (1 T i )(1 ri ) Ri

^

`

(7)

By means of equation (7) we estimate for each cell the minimum carbon price that is required for switching from non-forest land to forestry. Carbon supply curves represent the relationship between the carbon price and the amount of carbon sequestered. Equation (7) leads to cost estimation but we also need to predict quantities, which need to be specified for a give time period, T. If there exists a market price of carbon, pc*, the cumulative carbon sequestration that occurs from time zero to time T is: I K b p (8)

CT

¦ ¦ i 1

k 1

Ai ,k 1 bi Ci ,k ,T Ci ,k ,T

i pci pc * . The index k denotes stands within a cell that have different ages.

Cib,k ,T

measures the cumulative carbon sequestration per hectare in the biomass p of stand k of cell i at time T, and Ci ,k ,T measures the cumulative carbon sequestration in forest products. For estimating

Cib,k ,T and Cib,k ,T we

consider that trees are replanted after each harvest.1 This allows to maintain a non-decreasing carbon stock over time (see Figure 15.2) .

15.3 Data 15.3.1 Land Available for AR-Projects For estimating how much area is available for AR-projects, we rely on global land cover datasets provided by the International Geosphere Biosphere Project, IGBP (USGS, 2003). This data is combined with spatial information that identifies world countries (ESRI, 1998), agricultural suitability (Ramankutty et al, 2001), population density (CIESIN, 2000), 1

Note that replanting is the economically preferred alternative. If forestry was the most attractive land use for the first rotation interval, it will be the most attractive land use for the second and subsequent rotation intervals.

312


elevation (GTOPO30, 1996), net primary productivity and carbon stock (Alexandrov et al., 2002). For reasons of uniformity in the analysis, all datasets are converted to a resolution of 0.5 degree. AR-projects would take place on non-forest land where agricultural production is low or unprofitable, since AR-projects can hardly compete on productive agricultural lands with traditional forms of land use and the UNFCCC prescribes that land-use change for carbon benefits should not endanger food security. Given these prescriptions, we selected the following five land cover classes: grasslands, open shrublands, closed shrublands, savannas and croplands.2 In addition, we exclude the following cells from the selected land classes, (1) highly productive land where the indicator of suitability for agriculture is above 50% (this indicator ranges from 0 to 100%); (2) cells where the population density is over 200 hab/km2; (3) cells with elevation more than 3500 m; and (4) cells where there is no net carbon uptake, i.e. the difference between carbon stocks in forest and non-forest is zero. Based on these criteria, the land available for AR-projects sums up 26 million km2. 15.3.2 Carbon Uptake Grid data on carbon uptake is obtained from the spatial databases of Alexandrov et al. (1999, 2002). In order to estimate net carbon uptake, we subtract the carbon in the non-forest scenario (baseline) by considering two components: (1) a site-specific baseline corresponding to the nonforest carbon stock (Alexandrov et al., 1999; Alexandrov et al., 2002); and (2) a regional baseline which subtracts possible afforestation and revegetation trends in a business-as-usual scenario. For the second baseline component we deduct 10% of the carbon sequestration for each cell3. Timber volume is proportional to forest biomass. We use a timber/carbon ratio of 2 m3/tC (Benítez and Obersteiner, 2005). Rotation intervals are 2 These land cover classes are defined as follows: Grasslands are lands with a herbaceous type of cover with tree and shrub cover between 0 and 10%. Open Shrublands are lands with woody vegetation less than two meters tall and shrub cover from 10 to 60%. Savannas are lands with herbaceous and other understory systems, with forest cover between 10 and 30%. Closed Shrublands: are lands with woody vegetation less than two meters tall and shrub cover greater than 60%. Croplands: are lands covered with temporary crops. 3 Afforestation and reforestation worldwide accounts for 4.5 million hectares each year in the absence of carbon payments - baseline scenario (FAO, 2001). If such a rate would continue for a 50-year period, it would represent about 10% of the area suitable for plantations worldwide.


313

considered uniform worldwide with an average value of 30 years corresponding to temperate regions (Nilsson and Schopfhauser, 1995). Note that rotation intervals could be shorter in the tropics and longer in boreal regions. For carbon uptake in products we consider an exponential decay and that 50% of the forest biomass is stored in long-lived products with a half-life time of 20 years and the remaining biomass consisting of short-lived products has a half-life time of one year. Finally, we assume that tree-planting in each cell requires 50 years for completion and that planting occurs at a constant rate as in the forestation scenarios described in Trexler and Haugen (1995). 15.3.3 Risk-Adjusted Discount Rates We use country risk-adjusted discount rates that depend on country-specific aspects like government credibility, corruption, economic stability, inflation, wars and terrorism. It is clear that investors would demand higher returns for forestry projects implemented in countries that have performed poorly in these aspects. There are ways for accounting for risk in investment projects. A commonly applied method is the use of risk-adjusted discount rates or required returns. For employing this technique in our study, the discount rate used for estimating carbon costs (equation 7) needs to be adjusted to risk. Generally, the Capital Asset Pricing Model (CAPM) serves for estimating risk-adjusted discount rates. The CAPM considers market efficiency where the differences between the market return and the risk-free rate are a measure of the price paid for market risk. The fundamental equation of the CAPM is: (9)

r

rf E rm rf

where r is the required return for an asset, rf is the risk-free rate of return, rm is the market rate of return and ȕ (beta) measures the contribution to risk of the investment relative to the market. Extensions of the CAPM have been applied globally (see Bekaert and Harvey, 1995) where expected returns are influenced by both world and country factors. While these CAPM extensions lead the estimation of required returns for different countries, it has limited applicability for worldwide analyses given the absence of equity markets in most developing countries. Considering this factor, Erb et al. (1996a) used an alternative formulation for estimating expected returns in

314


a large number of developing countries, under the assumption that expected returns are a function of risk ratings:

ri

J 0 J 1 ln( RRi ) H i

(10)

where ri is the expected return in country i, RRi is the risk rating of country i, J 0 , J 1 are parameters of the return function, and H i is the error term. The log-linear model has been proposed in order to capture potential nonlinearities when country risk is high. Since risk rating agencies provide data for more than 70% of the world’s countries, this method is applicable worldwide. A number of risk indicators have emerged, which are available for a large number of countries including (Erb et al., 1996b): (1) Institutional Investors: provides country credit ratings (CCR) based on surveys from bankers located worldwide; (2) Moody’s: provides ratings describing the creditworthiness of corporate bonds; (3) Standard and Poor’s (S&P): uses a similar rating system as Moody’s but creates a finer rating; and (4) International Country Risk Guide (ICRG): provides ratings for political, financial and economic risk factors and also calculates a composite index. For our analysis we selected ICRG because it is not limited only to credit risk but compiles political, economic and financial aspects that determine the overall concern for investing in a specific country. We used the ICRG 5-year composite index forecast (PRS, 2004). The parameters of the return function (equation 10) are estimated using Ordinary Least Squares regression based on data for carbon sequestration projects in 8 countries: US (Stavins, 1999), Canada (van Kooten et al., 2002), Argentina (Sedjo, 1999), Brazil (Fearnside, 1995), India (Ravindranath and Somashekhar, 1995), Indonesia (Cacho et al., 2002), Mexico (Masera et al., 1995) and Costa Rica (Nieuwenhuyse et al., 2000). After fitting the regression parameters, we predict the required return for world countries as a function of the ICRG index for each country. When the forecasted return for less risky countries is below 3%, we assign a value of 3% in order to avoid having a rate below a risk-free rate.4 Although the primal data for the regression used a sample of only eight countries, the estimated returns for the others seems reasonable. For example, for the stable economy of Australia we estimated a rate of 3.6%. For China, we have a moderate rate of 7.5%. Chile, as a newly industrialized country, has a rate of 7.4%. Countries under conflict, like Somalia and Liberia, have rates of 33%, reflecting their unattractiveness for private investment. 4

US treasury bills are often used as a reference for risk-free rates. For early 2004, 3-month treasury bills yield about 1%. The average for the last five years is 3.6% (FFC, 2004).


315

15.3.4 Prices Knowing that Brazil has one of the largest areas for afforestation and reforestation, we use it as a reference for prices. We established $800/ha for plantation costs in Brazil, which is within the range provided by Ecosecurities (2002) and Fearnside (1995). For other countries, we correct prices using the price index, which is the ratio between the purchasing power parity (PPP) conversion factor and the official exchange rate in 2001 (World Bank, 2003). Land prices depend on suitability of agriculture and population density following a Cobb-Douglas relationship. For fitting the parameters of the land price function (Ai), we set minimum and maximum bounds, so that the upper bound corresponds to cells where suitability for agriculture and population density are the highest, and the lower bound corresponds to cells where these indicators are the lowest. We assign equal weights for both indicators, so that Įi = Ȗi in equation (6). For Brazil, the higher bound for land prices is set at $2000/ha which resembles sites of good quality in Latin America (de Jong et al., 2000; Benítez et al., 2001). The lower bound is set to $200/ha. Timber stumpage prices across cells are estimated with a similar procedure as the land price. In the absence of a detailed infrastructure map that allows a precise estimation of transportation costs, we consider that timber stumpage prices are dependent on population density. Taking into account that transportation costs are major determinants of timber stumpage prices, we expect that in areas of high population density, transportation costs will be low since distances to markets are small and infrastructure availability is high. The higher bound for timber price in Brazil is $35/m3, based on an export price of $50/m3 (FAO, 2002) and harvesting and transportation costs of $15/m3. The lower bound for timber price is $5/m3 and the values in-between are adjusted linearly with population density. Timber and land prices for countries other than Brazil are estimated using the same price index that was used for plantation costs.

15.4 Results 15.4.1 Geographic Distribution of Carbon Sequestration Costs Having identified the cells that are suitable for AR-projects, we estimate carbon sequestration costs for each of these cells. By mapping these results, we visualize where the least-cost sites for carbon sequestration are located. As shown in Figure 15.1, Africa, Asia and South America are the regions providing the cheaper options for sequestering carbon (for example the Sub-

316


Sahara region, Southeast Brazil and Southeast Asia). Northern regions, although they have very large areas suitable for AR-projects, are less attractive for tree-planting due to their lower rates of carbon uptake.

Fig. 15.1 Geographic Distribution of Carbon Sequestration Costs.

15.4.2 Global Carbon Supply

Cumulative carbon, MtC

We provide price/cumulative carbon quantity relationships through time by estimating carbon accumulation in biomass and products for each cell and aggregating results for the whole globe. Figure 15.2 shows timeprofiles of carbon sequestration under different price scenarios considering a time span of up to 100-years. 50000 40000 $100/tC

30000

$50/tC 20000

$30/tC

10000 0 0

20

40

60

80

100

year Fig. 15.2. Time-profile of carbon sequestration under different carbon prices.


317

15.4.3 Long-term Carbon Supply and Policy Implications For finding the (long-term) policy implications of AR-projects and their role in global warming mitigation, we compare the AR cost curves with the costs of carbon mitigation in the energy sector resulting from the RICE-99 model (Nordhaus and Boyer, 2000). This integrated assessment model of climate change predicts the amount of emission reductions that are needed for reaching a policy target and the costs per ton of carbon associated with such emissions reductions. Since the RICE-99 model deals with time-periods of about a century, we estimate the carbon supply for a 100-year period. Figure 15.3 shows the carbon supply curve with carbon quantities shown in 100-year average carbon sequestration.

carbon price, $/tC

400 300 200 100 0 0

100

200

300

400

500

600

700

100-year average carbon sequestration, MtC/yr Fig.15.3 Carbon supply curve for a 100-year period time span.

The results of this comparison are summarized in Table 15.1 where we show the policy target (first column), the required emission reductions (second column) and the associated average carbon price (third column) according to the different policy alternatives dealt with in the RICE-99 model. Then, based on this carbon price, we estimate how much carbon is sequestered for a 100-year period using Figure 15.3 (fourth column). Results of these analysis allow us to stress two important issues: (1) AR projects are important for global warming mitigation, where its potential carbon sequestration ranges from 6% to 25% of the emissions reduction targets of different policy scenarios; and (2) the relevance of AR strategies increases with increasingly strict policy alternatives. Policy scenarios requiring larger emission abatements would need a larger share of emission reductions through AR-projects than those with smaller abatements.

318


Table 15.1 Comparison of carbon sequestration through AR-projects with emission reductions of RICE-99. Note: The analysis holds for a 100-year period.

Climate Policy Scenario

Average emission reductions of policy scenario – energy sector (MtC/yr)

Average carbon price associated to policy scenario ($/tC)

Carbon sequestration as a fraction of emission reductions in policy scenario (%)

34 312

Average carbon sequestration corresponding to carbon price of the policy scenario (MtC/yr) 51 627

Optimal Limit to 1990 emissions Limit to 2 times CO2 Concentrations Limit temperature rise to 2.5 degrees

800 4400 1400

73

349

25%

3100

207

593

19%

6% 14%

In this chapter we do not present results on the sensitivity of model results with respect to model parameters. However, it is important to mention the conclusions from previous work (Benítez and Obersteiner, 2005): (i) carbon uptake is the most sensitive parameter (ii) land prices have a lower impact on the supply curve but it is difficult to have accurate estimates since ultimately, land prices depend on particular preferences, attitudes of landowners and land market policies, and (iii) carbon prices have a strong influence on the sensitivity where the higher the carbon price is, the more robust the sequestration results are.

15.5 Conclusions This chapter describes a methodology for deriving supply-curves of carbon sequestration through AR-projects by means of a spatially explicit analysis. Major advantages of the method are: (i) there is no need to entirely depend on comprehensive data that are often scarce in developing countries, instead, major parameters are estimated indirectly from more general databases and GIS datasets available worldwide, and a consistent method could be applied over a large region. (ii) Results are obtained for each cell, so that maps with the geographical distribution of carbon costs can be elaborated. This facilitates comparison across countries and identification of least-cost regions for carbon sequestration. (iii) Estimation of sequestration costs takes the entire


319

life-cycle of the sequestered carbon into account, including carbon uptake during the growing phase, carbon emissions during harvest, and residual carbon storage in short and long lived-products. Explicit treatment of the full-cycle helps to avoid problems with carbon accounting, and (iv) it accounts for the impact of country risk in AR investment choice. When we compared emission reductions through AR-projects with the required emission reductions of various climate policy scenarios of the RICE-99 model, we found that AR-projects represent between 5% to 25% of the emissions reduction targets and therefore, it is a policy option that should be exploited to a larger extent. As the majority of least-cost sites are located in Africa, South America and Asia, special attention to capacity building and to strengthening institutions in those regions should take place to facilitate the engagement of AR investors.

References Alexandrov, G.A., Y. Yamagata and T. Oikawa, 1999. Towards a Model for Projecting Net Ecosystem Production of the World Forests. Ecological Modelling 123: 183–191. Alexandrov, G.A., T. Oikawa and Y. Yamagata, 2002. The Scheme for Globalization of a Process-based Model Explaining Gradations in Terrestrial NPP and its Application. Ecological Modelling 148: 293–306. Bekaert, G. and C.R. Harvey, 1995. Time-varying World Market Integration. Journal of Finance 50: 403–444. Benítez P.C. and M. Obersteiner, 2005. Site Identification for Carbon Sequestration in Latin America: A Grid-Based Economic Approach. Journal of Forestry Policy and Economics. Benítez, P.C., R. Olschewski, G.H.J. de Koning and M. López, 2001. Análisis Costo-beneficio de Usos del Suelo y Fijación de Carbono en Sistemas Forestales de Ecuador Noroccidental. Report TÖB TWF-30s of the German Technical Cooperation Agency (GTZ), Eschborn, Germany. Cacho, O.J., G.R. Marshall and M. Milne, 2002. Transaction and Abatement Costs of Carbon-sink Projects: An Analysis Based on Indonesian Agroforestry Systems. In: Conference Proceedings of the Australian New Zealand Society for Ecological Economics, December 2002, Sydney, Australia. Available at http://www.une.edu.au/feb1/economics/carbon/cc06.pdf. CIESIN, 2000. Gridded Population of the World (GPW), Version 2. Center for International Earth Science Information Network (CIESIN), Columbia University; International Food Policy Research Institute (IFPRI) and; World Resources Institute (WRI). CIESIN, Columbia University, Palisades, NY, US. Available at http://sedac.ciesin.columbia.edu/plue/gpw.

320


Creedy, J. and A.D. Wurzbacher, 2001. The Economic Value of Forested Catchment with Timber, Water and Carbon Sequestration Benefits. Ecological Economics 38: 71-83. de Jong, B.H., R. Tipper, G. Montoya-Gómez, 2000. An Economic Analysis of the Potential for Carbon Sequestration by Forests: Evidence from Southern Mexico. Ecological Economics 33: 313–327. Ecosecurities, 2002. Baseline Determination for Plantar: Evaluation of the Emissions Reduction Potential of the Plantar Project. Prototype Carbon Fund of The World Bank, Washington DC. Erb, C.B., C.R. Harvey and T.E. Viskanta, 1996a. Expected Returns and Volatilities in 135 Countries. Portfolio Management, Spring, 46–58. Erb, C.B., C.R. Harvey and T.E. Viskanta, 1996b. Political Risk, Economic Risk and Financial Risk. Working Paper WP9606. Fuqua School of Business, Duke University, Cambridge, US. ESRI, 1998. World Countries 1998. Environmental Systems Research Institute (ESRI). Redlands, California, US. Available at http://www.esri.com. FAO, 2001. Global Forest Resources Assessment 2000. Main Report. United Nations Food and Agriculture Organization (FAO), Rome. FAO, 2002. FAOSTAT Database. Food and Agriculture Organization of the United Nations (FAO), Rome. Available at http://apps.fao.org. Fearnside, P.M., 1995. Global Warming Response in Brazil’s Forest Sector: Comparison of Project-level Costs and Benefits. Biomass and Bioenergy 8(5): 309–322. FFC, 2004. 3-Month Treasury Constant Maturity Rate. The Financial Forecast Center (FFC). Available at http://www.forecasts.org/data/index.htm. Gritsevskyi, A. and L. Schrattenholzer, 2003. Costs of Reducing Carbon Emissions: An Integrated Modeling Framework Approach. Climatic Change 56: 167–184. GTOPO30, 1996. Global Digital Elevation Model (DEM) with a Horizontal Grid Spacing of 30 Arc Seconds. US Geological Survey’s EROS Data Center (EDC). Available at http://edcdaac.usgs.gov/gtopo30/gtopo30.html. IPCC, 2000. Land Use, Land-Use Change, and Forestry. Special Report, Intergovernmental Panel on Climate Change (IPCC), Cambridge University Press, Cambridge, UK. IPCC, 2001. Climate Change 2001: Impacts, Adaptation and Vulnerability. Intergovernmental Panel on Climate Change (IPCC), Cambridge University Press, Cabmbridge, UK. Masera, O.R., M. Bellon and G. Segura, 1995. Forest Management Options for Sequestering Carbon in Mexico. Biomass and Bioenergy 8: 357–367. Nieuwenhuyse, A., H. Hengsdijk, B.A.M. Bouman, R.A. Schipper and H.G.P. Jansen, 2000. Can Forestry be a Competitive Land Use Option? Model Simulations from Humid Tropical Costa Rica. Forest Ecology and Management 137: 23–40. Nilsson, S. and W. Schopfhauser, 1995. The Carbon Sequestration Potential of a Global Afforestation Program. Climatic Change 30: 267–293.


321

Nordhaus, W. D. and J. Boyer, 2000. Warming the World: Economic Models of Global Warming. MIT Press, Cambrige, US. PRS, 2004. The PRS Group International Country Risk Guide. Table T2c: Composite Risk Forecasts. Available at http://prsgroup.com/icrg/icrg.html. Ramankutty, N., J.A. Foley, J. Norman, and K. McSweeney, 2001. The Global Distribution of Cultivable Lands: Current Patterns and Sensitivity to Possible Climate Change. Manuscript in revision, Global Ecology and Biogeography. Available at http://www.sage.wisc.edu/atlas/. Ravindranath, N.H. and B.S. Somashekhar, 1995. Potential and Economics of Forestry Options for Carbon Sequestration in India. Biomass and Bioenergy 8: 323–336. Sedjo, R., 1999. Potential for Carbon Forest Plantations in Marginal Timber Forests: The Case of Patagonia, Argentina. Discussion Paper 99–27, Resources for the Future, Washington D.C. Sijm, J., F. Ormel, J. Martens, S. van Rooijen, M. Voogt, M. van Wees and C. de Zoeten-Dartenset, 2000. Kyoto Mechanisms: The Role of Joint Implementation, the Clean Development Mechanism and Emissions Trading in Reducing Greenhouse Gas Emissions. Energy Research Centre of the Netherlands (ECN), Petten, Netherlands. Sohngen, B. and R. Mendelsohn, 2003. An Optimal Control Model of Forest Carbon Sequestration. American Journal of Agricultural Economics 85: 448457. Stavins, R.N., 1999. The Costs of Carbon Sequestration: A Revealed-preference Approach. The American Economic Review 89: 994–1009. Trexler, M.C. and C. Haugen, 1995. Keeping it Green: Evaluating Tropical Forestry Strategies to Mitigate Global Warming. World Resource Institute, Washington DC. UNFCCC, 1998. Report of the Conference of the Parties on its Third Session, held in Kyoto from 1 to 11 December 1997. FCCC/CP/1997/7/Add.1. United Nations Framework Convention on Climate Change (UNFCCC), Bonn, Germany. USGS, 2003. Global Land Cover Characteristics (GLCC) Data Base, Version 2.0. United States Geological Survey (USGS). Available at http://edcdaac.usgs.gov/glcc/glcc.html. van Kooten, G.C., C.S. Binkley and G. Delcourt, 1995. Effect of Carbon Taxes and Subsidies on Optimal Forest Rotation Age and Supply of Carbon Services. American Journal of Agricultural Economics 77: 365-374. van Kooten, G.C., S.L. Shaikh and P. Suchánek, 2002. Mitigating Climate Change by Planting Trees: The Transaction Costs Trap. Land Economics 78: 559–572. World Bank, 2003. World Development Indicators 2001. Table 5.7. The World Bank, Washington DC. Xu, D., 1995. The Potential For Reducing Atmospheric Carbon by Large-Scale Afforestation in China and Related Cost/Benefit Analysis. Biomass and Bioenergy 8: 337–344.

16 Positive Spillovers of Energy Policies on Natural Areas in Poland: an AGE Analysis*

Adriana M. Ignaciuk Wageningen University, The Netherlands, E-mail: [email protected]

Abstract. Current climate policies in Poland target for an increase in bioelectricity share in total electricity production. In Poland most of the renewable energy comes from biomass (around 90%). Most probably, in the future, biomass will continue to play a dominant role within the renewable energy sources. Except of their primary functions, such as providing CO2 neutral fuels and decreasing dependency on oil, biomass plantations can positively influence the environment. They carry similar functions to natural areas e.g. they can contribute to the improvement of soil and water quality, sequester carbon in the soil and create an environment for many species. However, an often-heard concern is that large-scale biomass plantations might increase pressure on the productive land and might cause a substantial increase of food prices. The aim of this chapter is to investigate the impact of different energy policies, focused on increasing the shares of bioelectricity in the total electricity production, on production and prices of agricultural goods and electricity and on changes in land cover. Key words: Applied General Equilibrium (AGE), Biomass, Energy Policy, Nature Conservation, Renewable Energy

__________________________ * The author would like to thank Arjan Ruijs, Rob Dellink and Hans-Peter Weikard for their comments on the earlier version of this chapter, and Ekko van Ierland for his comments on earlier version of this model. Moreover, the author would like to thank to the anonymous referee for helpful comments.

324

Adriana M. Ignaciuk

16.1 Introduction Current energy policies in Poland target for an increase in bioelectricity share in total electricity production. Nowadays, coal is dominant in the production of electricity in Poland. Around 97% of all electricity generated in the country comes from coal-fired plants that are very inefficient (Nilsson et al., 2006). In 1999 most of the ‘green’ electricity was produced from small hydro plants, but there is not much scope for expansion of this type of electricity in Poland. Other potential sources for electricity production are (i) solar panels, (ii) wind mills and (iii) biomass. Solar energy is relatively expensive compared to other renewable sources. To produce relatively cheap wind energy, the wind parks need to have good geographical conditions. Only in the northern part of Poland there is some development in this field, but both atmospheric conditions and negative community attitude do not encourage further developments. Hence, in the future, it is expected that the biomass is going to play a larger role in the production of green electricity. Considering the fact that in Poland most of the renewable energy comes from biomass (around 90% (GUS, 2002a) mainly in the heat sector), and that there are good socio-economic potential for further biomass production, it is expected that, in the near future, bioelectricity from biomass will continue to play a dominant role within the renewable energy sources. To meet the demand for clean energy, once stringent climate policies take place, large scale biomass plantations are anticipated. Except for their primary function, such as providing CO2 neutral fuels, they reduce the dependence on fossil fuels supply. Biomass plantations can positively influence the environment. They can contribute to the improvement of soil and water quality, sequester carbon in the soil and create an environment for many species (Borjesson, 1999; Tolbert et al., 2002; Londo et al., 2005). Moreover, some of the biomass plantations can carry additional functions e.g. they can be used for recreation purposes. Due to the above mentioned characteristics, we claim that the biomass plantations can provide similar functions to the natural areas. That is why in this chapter we call them semi-nature. However, an often-heard concern is that large-scale biomass plantations might increase pressure on the productive land and might cause a substantial increase of food prices see for instance Azar (2001; 2003), and McCarl and Schneider (2001). In contrast there are claims that current overproduction of food allows for using a part of the agricultural land for

16 Positive Spillovers of Energy Policies

325

other practices, such as biomass plantations, see e.g. Tilman et al. (2002), Trewavas (2002) and Wolf et al. (2003). To reduce the competition between agriculture and biomass for land and to increase biofuel supply multi-product crops can be used. Dornburg (2005) defines it as follows: multi-product crops can be defined as crops that can be split into two or more different parts that are used for different applications. One part of the crop is used directly as energy, i.e. it is used as solid fuel or converted to liquid fuel and the other for material applications. Introducing such systems can influence the changes in land prices and land use allocation. Different types of models exist to study the possible land shift between agriculture and biomass or forestry and its impact on the economy and environment. There are many agricultural models that focus mainly on land shifts, without including any energy systems. Examples of such models include POLYSYS (Torre Ugarte de la and Ray, 2000) and GOAL (WRR, 1992). Walsh et al. (2003) modified the agricultural model POLYSYS to include specific biomass crops (switch grasses, poplar and willow) and provide estimates for changes in annual land use. These models are based on linear techniques. An example of a partial equilibrium model used for determining the allocation of food and biomass crops is the ASM model (McCarl et al., 1993), that accurately describes the agricultural sector in the USA. This model has some successors; one of them is FASOM that enlarges ASM to include the forestry sector (Adams et al., 1996; van Ierland and Lansink, 2003). Another successor is the ASMGHG model that includes emissions of greenhouse gases and mitigation possibilities (Schneider and McCarl, 2003). Different from these models our approach goes beyond agricultural and forestry sectors. In this chapter, the interactions between agricultural sectors and other sectors of the economy are included. Moreover, we include explicitly the electricity market and endogenous CO2 permit prices. Models that focus on the energy supply side are e.g. MARKAL MATTER (Gielen et al., 2001) and LUCEA (Johansson and Azar, 2004). MARKAL MATTER focuses on detailed descriptions of the energy system, and its biomass modules boil down to agricultural and forestry residuals and waste. LUCEA deals with competition between biomass and food crops, using a bottom up approach. It is used to determine food and energy prices in case of stringent climate policies in the USA with exogenous CO2 emission permit prices. Both of these models focus on different energy types; however the interactions between different sectors within the entire economy and the secondary effects of policy implementations are not modeled.

326

Adriana M. Ignaciuk

There are many models that involve a detailed economic analysis of the energy sector, and that are able to provide the secondary effects of shifting the energy production towards renewable energy. However, they often omit biomass resources e.g. Kumbaroglu (2003) McFarland et al., (2004), Babiker (2005) or agricultural multi-product sources e.g. Breuss and Steininger (1998), and Ignaciuk et al., (2005). The aim of this chapter is to investigate the impact of different energy policies, focused on increasing the shares of bioelectricity in the total electricity production, on production and prices of agricultural goods and electricity, and on changes in land cover. In this context, we analyze how these policies might affect the establishment of semi-nature areas. Moreover, we analyze to what extent using the by-product of agriculture and forestry sectors increase the bioelectricity shares, and reduce the pressure on agricultural land. To attain our objective, we develop an applied general equilibrium model (AGE) with special attention to biomass and agricultural crops and different energy systems for a small open economy, with an Armington specification for international trade. Moreover, it distinguishes different land classes to capture differences in productivity. The emissions of the major greenhouse gases CO2, N2O and CH4 are also captured. This chapter is structured as follows. Section 2 presents the model specification. Section 3 describes the data and Section 4 provides the description of scenarios. In the following section, Section 5, the results are presented and discussed. To the end, in Section 6, the conclusions are gathered.

16.2 Model Specification Using a CGE-framework allows us to account fully for the interlinkages between different sectors of the economy. These are relevant, as the agricultural and energy sectors have strong links with the rest of the economy. Moreover, the indirect impacts of environmental policies, which are often ignored but can be highly relevant (Dellink, 2005), are incorporated, ensuring a consistent assessment of the economic costs of environmental policy. The model describes the entire economy, with explicit detail in the representation of production of traditional agricultural and biomass crops1. As in any standard general equilibrium model (CGE) all markets clear, 1

It is an extended version of the model described in Ignaciuk et al. (2005).


327

which means that supply equals demand for all goods through adjusting relative prices (Ginsburgh and Keyzer, 1997). In the model, 35 sectors are distinguished. We consider explicitly both agricultural and biomass sectors. The electricity sector is divided into conventional electricity and bioelectricity, depending on the type of energy source used for its production. We include three primary production factors: labor, capital and land. Four land classes that correspond to the six land classes used in the Polish land classification system (GUS, 2002a), are identified to capture differences in productivity from different land types. Agricultural and biomass crops can grow on three different land use classes z1(very good), z2(good), and z3(poor). Forestry can only grow on the z4 type of land. A representative consumer maximizes utility under the condition that expenditures on consumption goods do not exceed her income. Utility is represented by a nested constant elasticity of substitution (CES) function2: U CES Ci , ELN ;V U (1) in which U is utility, Ci is the consumption of commodities from sector i and ELN CES Ce , Cbe ;V EL where Ce and Cbe are consumption of Electricity and Bioelectricity respectively. Parameters VU and VEL are substitution elasticities. Such specification allows for substitution possibilities between different consumption goods, such as between conventional electricity and bioelectricity. Consumers own production factors and consume produced goods. Labor supply is fixed. The wage rate is fully flexible. The total availability of labor is determined by the initial endowments of the representative consumer. Producers maximize profits subject to the available production technologies. Following Rutherford and Paltsev (2000), production technologies are represented by nested CES functions. Production functions of different commodities have a six-level nesting structure. There is substitution possible between different primary factors and composite inputs, e.g. the electricity nest allows for substitution possibility between conventional electricity and bioelectricity and ELK nest allows for substitution between two composite inputs primary factors and energy. The structure of the production function is presented in Fig. 16.1. For production of all commodities that are CO2 intensive, emission permits are required. Emissions included in this model cover most of the greenhouse gases; CO2, N2O and CH4. Both CH4 and N2O emissions are 2

The CES function Yi

CES(X1,X2;V).

D X U D X U 1

1

2

2

1U

with U = (V-1)/V is written as Yi =

328

Adriana M. Ignaciuk

expressed in CO2 equivalents. Data on emissions is obtained from Sadowski (2001). As CO2 emissions come mostly from fossil fuel combustion they enter the production function in a different place as CH4 and N2O emissions (Fig. 16.1). Environmental policy is implemented by reducing the number of emission permits the government auctions. This way of modeling environmental policy ensures that a cost-effective allocation is achieved (Dellink, 2005). In the model, Poland is considered to be a small open economy. It means that neither domestic prices nor traded quantities change the ‘world market prices’. The international market is assumed to be large enough to absorb any quantities of goods produced in Poland and it can satisfy Polish import demands. In this model, we choose the Armington specification for traded goods, assuming that domestic and foreign goods are imperfect substitutes (Armington, 1969). This allows for a difference in prices between domestically produced goods and their international substitutes. Hence, an increase in domestic prices leads to a shift in demand towards the competitive imports, but only to a limited extent. Similarly, a change in domestic prices will have a limited impact on exports. There will be a demand for export goods even if the domestic price is above the world market price (Dellink, 2005).

Fig. 16.1 Nested CES function.


329

All taxes are collected by the government that uses them to finance public consumption and pay lump-sum transfers to private households. The EU subsidy is an exception and it is paid from external sources, namely EU. There are different subsidy schemes, depending on land cover. The traditional agriculture and biomass sectors are directly subsidized. However, the Forestry sector receives subsidy once it turns agricultural land into forestry. In the model, bioelectricity can be produced by using as fuels the primary agricultural and biomass products or using the by-products that remain from conventional agricultural production. Such by-products are for instance straw produced by the cereals sectors or forestry residuals produced by the Forestry sector. These by-products in the benchmark have a low price, reflecting low demand for biofuels in the benchmark. In the model, the substitution elasticity between traditional biofuels and the fuels that are produced as by-products is very high.

16.3 Data To determine the benchmark equilibrium, a Social Accounting Matrix (SAM) for Poland is specified. For this purpose, we adopted the most recent, at the time of writing, available GTAP data (for 1997) (Dimaranan and McDougall, 2002). In the SAM, agricultural and biomass data are disaggregated based on the FEPFARM model built by Mueller (1995), using FAO country land use data for Poland. The FEPFARM model provides the shares of production costs. Data on land use pattern and emissions are obtained from Polish statistics (GUS, 2002b; 2002a). Data on agricultural and biomass residuals are taken from Gradziuk (2001) and Dornburg et al. (2005). We specify the substitution elasticities between different production inputs in the production functions, based on literature surveys and experts’ opinions. Estimates of substitution elasticities between capital, labor and energy, are estimated by Kemfert (1998), Rutherford and Paltsev (2000), Kiuila (2000), and Dellink (2005)3.

3

The full data set used in the model can be obtained from author.

330

Adriana M. Ignaciuk

16.4 Scenarios Polish policy makers set two goals concerning an increase of the bioelectricity share into total electricity production: 7.5% by 2010 and 14% by 2020. We present two policy scenarios aimed at increasing the bioelectricity share and at reducing CO2 emissions. Both of these scenarios are analyzed in a unilateral setting, which means that only Poland undertakes the energy policies, and RoW continues its business as usual. The first scenario, Scenario S, considers a reduction of emission permits by 10% and adoption of bioelectricity subsidy in a single-product setting. Since Poland has already fulfilled its Kyoto obligations, further emission reductions can be beneficial once Poland can trade its emission rights. Scenario M adopts the same rate of emission permits reduction (10%) and subsidy on bioelectricity rate, however, the analysis focuses on the multiproduct setting. In the single-product setting the biomass used for bioelectricity production comes as the primary biomass product. In the multi-product setting, also the by-product, of mainly cereals and forestry, are used. As mentioned the by-products of cereals and forestry are straw and forestry residuals respectively.

16.5 Results and Discussion This section presents the results of the policy analysis for all scenarios. In section 5.1, we discuss the general results, including the impact of the scenarios on bioelectricity share, utility and prices of emission permits. Sections 5.2 and 5.3 focus on impact on production and prices of different commodities, respectively. Subsection 5.4, analyses the changes in land allocation. 16.5.1 General Results Fig. 16.2 presents the influence of the implementation of CO2 emission permit reduction combined with a biomass subsidy scheme on the share of bioelectricity in electricity production. The results show clear differences between the bioelectricity shares for the single-product and the multiproduct settings. Noticeably, for every level of bioelectricity subsidy, in the multi-product setting there are higher shares of bioelectricity than in the single-product setting. This is not surprising, considering the fact that in multi-product setting bioelectricity producers obtain cheaper biofuels.


331

The first policy goal of 7.5% bioelectricity share using the singleproduct option is reached with around 22% subsidy on bioelectricity. The same goal in the multi-product setting is reached with around 20% subsidy. The second goal of 14% shares is reached with around 31% subsidy rate in single-product setting and by utilizing by product the same goal can be reached with a 4% lower subsidy. Welfare costs of these policies tend to be virtually the same (see Fig 16.3). However, it may seem puzzling that the utility level increases with the size of the subsidy rate. One explanation of this is that in a second best world (with many distortional taxes) bioelectricity subsidy covers some of the welfare loses that society pays once emission are reduced by 10%.

Fig. 16.2 Bioelectricity share for single-product (S) and multi-product (M) scenarios for different levels of bioelectricity subsidy.

332

Adriana M. Ignaciuk

Utilit y

0,989

0,988

Index (BM=1)

0,987

0,986

Scenario S Scenario M

0,985

0,984

0,983

0,982 0

10%

20%

30%

40%

50%

Policy level - bioelectricity subsidy

Fig 16.3 Utility change for single-product (S) and multi-product (M) scenarios for different levels of bioelectricity subsidy.

This phenomenon can be also explained by the fact that subsidizing bioelectricity, provides more ‘clean’ energy that can substitute the dirty conventional one. Producers and consumers can switch their demand towards CO2 neutral fuels and reduce the demand for emission permits. Moreover they can import cheaper ‘dirty’ goods from the RoW. Cheaper emission permits (Fig. 16.4) influence positively the utility. From Figures 16.2-16.4 we observe that the share of the bioelectricity, utility and price levels change in a non-linear manner. Small changes in emission reduction triggers small changes in bioelectricity shares, utility level and price of emission permits. More stringent environmental policies affect bioelectricity shares, utility level and price of emission permits substantially more.


333

Fig. 16.4 Emission permit price for single-product (S) and multi-product (M) scenarios for different levels of bioelectricity subsidy.

16.5.2 Production Table 16.1 comprises the results of production changes for the two scenarios and for two different bioelectricity subsidy levels: 10% and 40%. The economy adapts to these reductions by switching towards (i) ‘clean’ energy and (ii) ‘clean’ production. In both scenarios there is a clear increase in bioelectricity production, considering 10% emission permit reduction and a 10% subsidy. It increases by 198% in Scenario S and 305% in Scenario M. Since the Bioelectricity sector is very small compared to the Electricity sector, to meet the energy demand of the total economy, this sector has to grow considerably. Labor and capital released from the declining Electricity sector are used to intensify the production of the Bioelectricity sector. The ‘clean’ sectors such as e.g. sectors producing rape, willow or hemp increase their production substantially, since there is a high demand for biofuels. In multi-product setting scenario, those changes are larger than in single-product setting. This difference is caused by the ability of producing more biofuels without additional costs for traditional agricultural production in multi-product setting. Moreover, since the by-products are cheap, the Bioelectricity sector demands them in large quantities. Using these new fuels it can grow and substitute even more conventional electricity. Some agricultural sectors decrease their production; however it is a very small reduction, the largest change is the three percent decline of Other Agriculture sector, at the 10% biomass subsidy level.

334

Adriana M. Ignaciuk

It might seem surprising that most of the agricultural, biomass and forestry goods increase their production. This can be explained, however, by the fact that those sectors can intensify their production by substituting land for other production factors that become available due to the production losses in the industrial, energy and services sectors. In both scenarios (S and M) the dirty sectors decrease their production substantially (see Table 16.1). In the multi-product setting, there are slightly smaller losses in production of the ‘dirty’ sectors, except in the electricity sector. This can be explained by the fact that using by-products, most of the agricultural and biomass sectors increases their production without using additional production factors. 16.5.3 Prices The policies adopted in the model induce price changes; the AGE framework allows an analysis of relative prices, but the absolute price level is undetermined (this is solved by choosing the Consumer Price Index as numéraire). Generally, the prices of dirty goods e.g. conventional electricity, for which the production costs increase substantially given that they have to pay for emission permits, go up compared to prices of clean goods. The impact of emission reduction policies on price level for a selection of goods is presented in Table 16.2. We can observe an increase of agricultural commodity prices. However, this increase is much lower than in other studies, at most 2%, if the emission permit price rises to around 5 Euro per ton of CO2 (for both scenarios at 10% bioelectricity subsidy rate). Generally, the price level of land increase for all types of land (Table 16.3). Such increase is caused by two factors. First, the EU subsidies cause a distortion and increase the income of farmers without increasing the productivity of land. Second, in the multi-product setting, the productivity of land increases without compromising any other factors. More stringent policies induce higher land prices.


335

Table 16.1 Changes in the production in selected sectors for all scenarios at 10% and 40% bioelectricity subsidy rate (% change compared to benchmark).

Other Agriculture Rape Willow Hemp Wheat Other Cereals Food & animals Forestry Coal Oil Gas Petrochemicals Electricity Bioelectricity Industry Services

Singleproduct 10%

Multiproduct 10%

-3 5 220 17 -2 -1 -2 0 -9 -17 -14 -15 -6 198 -1 -1

-3 7 310 24 -2 -1 -2 0 -9 -17 -14 -15 -7 305 -1 -1

Singleproduct 40% 2 95 3730 299 -1 12 -1 13 -9 -16 -14 -14 -20 3214 -1 -1

Multiproduct 40% 3 120 4578 381 -1 16 -1 16 -10 -15 -14 -13 -23 4160 -1 -1

Table 16. 2 Changes in prices of selected commodities for all scenarios at 10% and 40% bioelectricity subsidy rate.

Singleproduct 10% Other Agriculture Rape Willow Hemp Wheat Other Cereals Forestry Electricity Bioelectricity

2% 0% -1% 0% 0% 1% 0% 3% -9%

Multiproduct 10%

Singleproduct 40%

Multiproduct 40%

2% 0% -1% 0% 0% 1% 0% 3% -12%

1% 0% -1% 0% 0% 1% 2% 3% -29%

1% 0% -2% 0% 0% 1% 2% 3% -31%

336

Adriana M. Ignaciuk

Table 16.3 Changes in prices of land for all scenarios at 10% and 40% bioelectricity subsidy rate.

Land type z1 Land type z2 Land type z3 Land type z4

Singleproduct 20% -4% 4% 1% 18%

Multiproduct 20% -5% 6% 3% 31%

Singleproduct 40% 69% 38% 40% 154%

Multiproduct 40% 91% 52% 55% 181%

16.5.4 Land Use 16.4 presents the land use allocation of all crops. In the single-product scenario there is limited land reallocation, the multi-product scenario show larger changes in sown area. In 16.4, we can observe that the acreage of biomass (willow, hemp and forestry) hardly increase in Scenario S for 10% emission reduction and 10% bioelectricity subsidy rate; however for 40% subsidy rate, it increases considerably to the amount of 380 000 ha. In Scenario M, we observe immediate change in the size of semi natural area. For 10% emission reduction and 10% bioelectricity subsidy rate it increases by 1 700 ha and for 40% subsidy rate by 540 000 ha. This large increase is caused mainly by converting some of the agricultural land into forestry, thanks to the EU subsidy and the fact that Forestry sector produces also a cheap by product used as fuel in bioelectricity. This increase in acreage of semi natural areas is caused mainly by increased demand for clean fuels. Hence, the proposed policies target the reduction in CO2 emission as well as increase of nature areas.


337

Table 16.4 Land use (in 1000 ha) for all scenarios at 10% and 40% bioelectricity subsidy rate.

Other Agriculture

Rape

Willow

Hemp

Wheat

Other Cereals

Forestry

BM

Singleproduct 10%

Multiproduct 10%

Singleproduct 40%

Multiproduct 40%

z1 z2 z3 z1 z2 z3 z1 z2 z3 z1 z2

102,4 1839,5 1051,6 0,0 349,4 87,3 0,0 0,0 0,5 0,0 0,0

101,9 1829,2 1046,0 0,0 353,2 88,2 0,0 0,0 0,7 0,0 0,0

101,5 1817,0 1040,4 0,0 376,5 94,2 0,0 0,0 2,2 0,0 0,0

90,3 1655,7 946,0 0,0 602,0 150,3 0,0 0,0 20,0 0,0 0,0

87,5 1608,8 918,4 0,0 652,7 162,8 0,0 0,0 24,1 0,0 0,0

z3 z1 z2 z3 z1 z2 z3 z4

0,1 87,4 1570,1 897,7 218,6 3894,5 2301,1 8769,0

0,1 87,5 1570,7 898,2 219,0 3900,4 2305,0 8769,0

0,2 86,9 1555,1 890,5 220,0 3904,9 2310,8 8769,0

0,5 74,3 1363,0 778,8 212,4 3864,7 2282,1 9129,2

0,6 71,4 1311,9 749,0 209,9 3828,3 2258,7 9285,0

16.6 Conclusions In this chapter we present a general equilibrium model to investigate the effects of climate policies on biomass systems and their influence on economy and the resulting land allocation. Before discussing the results, we would like to mention some of the caveats of the model. This is necessary because the results of the model depend crucially on the assumptions made in the model. First, a dynamic model would be able to show the transition path toward a “biomass economy”. Second, if we include the positive impact of increased environmental quality on welfare, we could calculate the efficient level of environmental policy and determine the optimal mix of agricultural and biomass production. Third, some of the substitution elasticities used in this model were adopted from other studies and calibrated for the conditions of other European countries. A full set of substitution elasticities calibrated for Poland could to some extent influence the results. Despite of theses limitations, we would like to highlight some interesting results. The results

338

Adriana M. Ignaciuk

show that Polish policy targets of increasing the bioelectricity shares can be fulfilled with modest emission reduction rates and bioelectricity subsidy levels. Moreover, we can conclude that multi-product crops can substantially increase the potential for bioelectricity and at the same time reduce the pressure on productive land. With proposed energy policies (emission permits in combination with bioelectricity subsidy) sectors producing agricultural, biomass and forestry commodities increase their production. There are several explanatory factors of this phenomenon. First, those sectors can intensify their production by substituting land for other production factors that become available due to the production losses in the industrial and energy sectors. Second, due to EU subsidies, production of land intensive sectors becomes more profitable. Moreover, using multi-products crops brings additional benefits to many agricultural and biomass producers, they can benefit from having higher output per unit of production factors. Energy policies, that were discussed, not only have a positive impact on emission reduction, but also on establishment of semi-natural areas. The positive externalities of these policies are visible in reclaiming productive land for nature and using ‘clean’ biomass crops instead of more polluting traditional crops. Especially when using multi-product crops, we observe an increase of acreages of biomass and forestry plantations. At current prices, bioelectricity is not economically interesting. The benefits that are brought by multi-product crops are the reduced prices of bio fuels itself and of bioelectricity. Thus, the costs of climate policy can be substantially reduced and the policy goals set for bioelectricity use can be achieved with less effort. However to reap all these opportunities, environmental policies are needed.

References Adams, D. M., R. J. Alig, J. M. Callaway, and B. M. McCarl, 1996, The Forest and Agricultural Sector Optimization Model (FASOM): Model Structure and Policy Application, Forest Service Report PNW-RP-495 (US Department of Agriculture, Washington, D.C.). Armington, P., 1969, A Theory of Demand for Products Distinguished by Place of Production, (IMF Staff Papers 16, Washington). Azar, C. (Ed.). 2003, Emerging Scarcities - Bioenergy-Food Competition in a Carbon Constrained World, (Resources for the future Inc. John Hopkins University Press (forthcoming). Babiker, M. H., 2005, Climate change policy, market structure, and carbon leakage, Journal of International Economics 65, 421-445.


339

Borjesson, P., 1999, Environmental Effects of Energy Crop Cultivation in Sweden - I Identification and Quantification, Biomass and Bioenergy 16, 137-154. Breuss, F., and K. Steininger, 1998, Biomass Energy Use to Reduce Climate Change: A General Equilibrium Analysis for Austria, Journal of Policy Modeling 20, 513-535. Dellink, R., 2005, Modelling the Costs of Environmental Policy - A Dynamic Applied General Equilibrium Assessment, (Edward Elgar, Cheltenham, Northhampton, MA). Dimaranan, B. V., and R. A. McDougall (Eds.), 2002, Global Trade, Assistance and Production: The GTAP 5 Data Base, (Department of Agricultural Economics. Purdue University, Center for Global Trade Analysis). Dornburg, V., G. Termeer, and A. Faaij, 2005, Economic and greenhouse gas emission analysis of bioenergy production using multi-product crops—case studies for the Netherlands and Poland, Biomass and Bioenergy 28, 454-474. Gielen, D., M.A.P.C. de Feber, A.J.M. Bos, and T. Gerlagh, 2001, Biomass for Energy or Materials? A Western European Systems Engineering Perspective, Energy Policy 29, 291-302. Ginsburgh, V., and M. Keyzer, 1997, The Structure of Applied General Equilibrium Models, (The MIT Press, Cambridge, London). Gradziuk, P., 2001, Produkcja i Kierunki Wykorzystania Slomy (Straw Use and Production), in A. Grzybek, P. Gradziuk, and K. Kowalczyk, eds, Sloma Energetyczne Paliwo (Straw - Energy Fuel), (Wies Jutra, Warszawa (Warsaw)). GUS, 2002a, Ochrona Srodowiska 2002 (Environmental Protection), (Glowny Urzad Statystyczny (Polish Statistics), Warsaw). GUS, 2002b, Skup i Ceny Produktow Rolnych w 2001 Roku (Quantities and Prices of Agricultural Commodities in 2001), (Glowny Urzad Statystyczny (Polish Statistics), Warsaw). Ignaciuk, A., A. Ruijs, and E. C. van Ierland, 2005, Can Climate Policies Contribute to Nature Conservation? On Biomass and Land Use in an AGE Approach, MWP 15 (Wageningen University, Wageningen). Johansson, D. J. A., and C. Azar, 2004, A Scenario Based Analysis of Land Competition Between Food and Bioenergy Production in the US, submitted to Climatic Change. Kemfert, C., 1998, Estimated substitution elasticities of a nested CES production function approach for Germany, Energy Economics 20, 249-264. Kiuila, O., 2000, Badanie Kosztow Dostosowywania Polskiej Gospodarki do Drugiego Protokolu Siarkowego. Model Rownowagi Ogolnej. (Costs of Adapting Poland's economy to the Second Sulfur Protocol. A General Equilibrium Model), Wydzial Ekonomii, (Warsaw University, Warsaw) Kumbaroglu, G. S., 2003, Environmental taxation and economic effects: a computable general equilibrium analysis for Turkey, Journal of Policy Modeling 25, 795-810. Londo, M., J. Dekker, and W. ter Keurs, 2005, Willow short-rotation coppice for energy and breeding birds: an exploration of potentials in relation to management, Biomass and Bioenergy 28, 281-293.

340

Adriana M. Ignaciuk

McCarl, B. A., C. C. Chang, J. D. Atwood, and W. I. Nayda, 1993, Documentation of ASM: The U.S. Agricultural Sector Model, (A&M University, Texas). McCarl, B. A., and U. A. Schneider, 2001, Greenhouse Gas Mitigation in U.S. Agriculture and Forestry, Science 294, 2481-2482. McFarland, J. R., J. M. Reilly, and H. J. Herzog, 2004, Representing energy technologies in top-down economic models using bottom-up information, Energy Economics 26, 685-707. Mueller, A. M., 1995, Fepfarm - mixed-integer farm simulation model of farms in the Tarnovsky Gory Pilot Project - Poland, (Personal E-mail (November 2003), Nilsson, L. J., M. Pisarek, J. Buriak, A. Oniszk-Poplawska, P. Bucko, K. Ericsson, and L. Jaworski, Energy policy and the role of bioenergy in Poland, Energy Policy In Press, Corrected Proof. Energy Policy 34 (15), 2263-2278 (2006). Rutherford, T. F., and S. V. Paltsev, 2000, GTAP-Energy in GAMS: The Dataset and Static Model, (University of Colorado, Boulder). Sadowski, M., A. Olecka, A. Romanczuk, M. Radwan-Rohrenschef, and A. Sienkiewicz, 2001, Third National Communication to the Conference of the Parties to the United Nations Framework Convention of Climate Change, (Institute of Environmental Protection, Warsaw), 109. Schneider, U. A., and B. A. McCarl, 2003, Economic Potential of Biomass Based Fuels for Greenhouse Gas Emission Mitigation, Environmental & Resource Economics 24, 291-312. Tilman, D., K. G. Cassman, P. A. Matson, R. Naylor, and S. Polasky, 2002, Agricultural Sustainability and Intensive Production Practicies, Nature 418, 671-677. Tolbert, V. R., J. Todd, D. E., L. K. Mann, C. M. Jawdy, D. A. Mays, R. Malik, W. Bandaranayake, A. Houston, D. Tyler, and D. E. Pettry, 2002, Changes in soil quality and below-ground carbon storage with conversion of traditional agricultural crop lands to bioenergy crop production, Environmental Pollution 116, S97-S106. Torre Ugarte de la, D., and D. E. Ray, 2000, Biomass and Bioenergy Applications of the POLYSYS Modeling Framework, Biomass and Bioenergy 18, 291-308. Trewavas, A., 2002, Malthus Foiled Again and Again, Nature 418, 668-670. van Ierland, E. C., and A. O. Lansink, 2003, Editorial Introduction: Economics of Sustainable Energy and Agriculture, Environmental & Resource Economics 24, 287-289. Walsh, M. E., D. G. de la Torre Ugarte, H. Shapouri, and S. P. Slinsky, 2003, Bioenergy Crop Production in the United States: Potential Quantities, Land Use Changes, and Economic Impacts on the Agricultural Sector, Environmental & Resource Economics 24, 313-333. Wolf, J., P. S. Bindraban, J. C. Luijten, and L. M. Vleeshouwers, 2003, Exploratory study on the land area required for global food supply and the potential global production of bioenergy, Agricultural Systems 76, 841-861. WRR, 1992, Ground for Choices. Four Perspectives for the Rural Areas in the European Community, 42 (Netherlands Scientific Council for Government Policy, the Hague), 144.

The Authors

Volker Beckmann: Humboldt University of Berlin, Germany Pablo C. Benítez: Department of Economics, University of Victoria, Canada Václav Beran: Czech Technical University, Czech Republic Frans Boekema: Radboud University Nijmegen and Tilburg University, The Netherlands Danes Brzica: Slovak Academy of Sciences, Slovak Republic Kees Burger: Wageningen University, The Netherlands Ian McCallum: International Institute for Applied Systems Analysis, Austria Petr Dlask: Czech Technical University, Czech Republic Aistø Dovalienø: Kaunas University of Technology, Lithuania Jana Gašparíková: Slovak Academy of Sciences, Slovak Republic Rimantas Gatautis: Kaunas University of Technology, Lithuania Rein Haagsma: Wageningen University, The Netherlands Wim Heijman: Wageningen University, The Netherlands Adriana Ignaciuk: Wageningen University, The Netherlands Catharinus F. Jaarsma: Wageningen University, The Netherlands Chris Jensen-Butler: School of Economics and Finance, University of St. Andrews, Scotland, UK

Steinar Johansen: Institute of Transport Economics, Norway Vaida Kvainauskaitø: Kaunas University of Technology, Lithuania Morten Marott Larsen: Institute of Local Government Studies, Denmark, AKF Bjarne Madsen: Institute of Local Government Studies, Denmark, AKF Pierre v. Mouche: Wageningen University, The Netherlands Edita Nemcová: Slovak Academy of Sciences, Slovak Republic Bronius Neverauskas: Kaunas University of Technology, Lithuania Michael Obersteiner: International Institute for Applied Systems Analysis, Austria Johan van Ophem: Wageningen University, The Netherlands Michal Páleník: Slovak Academy of Sciences, Slovak Republic Willem Pijnappel: The Netherlands Jan Rouwendal: Free University, Amsterdam, The Netherlands Roel Rutten: Tilburg University, The Netherlands Vytautas Snieška: Kaunas University of Technology, Lithuania Regina Virvilaitø: Kaunas University of Technology, Lithuania Justus Wesseler: Wageningen University, The Netherlands Yoshiki Yamagata: National Institute for Environmental Studies, Japan

Regional Externalities

Knowledge Externalities, Innovation Clusters and Regional Development

Regional Airports

Regional Integration

Regional Oceanography: Introduction

Regional Integration: Choosing Plutocracy

The Regional City

Regional Monetary Integration

Handbook of Regional and Urban Economics, Vol. 1: Regional Economics

Local and Regional Development

Regional Oceanography: An Introduction

Essentials of Regional Anesthesia

Complications of Regional Anesthesia

Textbook of Regional Anesthesia

American Regional Cuisine

Urban and Regional Planning

American Regional Cuisine

The Regional City

American Regional Cuisine

American Regional Cuisine

Atlas of Regional Anesthesia

Regional Financial Cooperation

World Regional Geography

Regional Conflict Management

Regional Monetary Policy

Complications of Regional Anesthesia

World Regional Geography

Regional geology of Africa

Essentials of Regional Anesthesia

Assessing Regional Integration in Africa: Rationalizing Regional Economic Communities

Rethinking Regional Innovation and Change: Path Dependency or Regional Breakthrough

Regional Externalities

Knowledge Externalities, Innovation Clusters and Regional Development

Regional Airports

Regional Integration

Regional Oceanography: Introduction

Regional Integration: Choosing Plutocracy

The Regional City

Regional Monetary Integration

Handbook of Regional and Urban Economics, Vol. 1: Regional Economics

Local and Regional Development

Regional Oceanography: An Introduction

Essentials of Regional Anesthesia

Complications of Regional Anesthesia

Textbook of Regional Anesthesia

American Regional Cuisine

Urban and Regional Planning

American Regional Cuisine

The Regional City

American Regional Cuisine

American Regional Cuisine

Atlas of Regional Anesthesia

Regional Financial Cooperation

World Regional Geography

Regional Conflict Management

Regional Monetary Policy

Complications of Regional Anesthesia

World Regional Geography

Regional geology of Africa

Essentials of Regional Anesthesia

Assessing Regional Integration in Africa: Rationalizing Regional Economic Communities

Rethinking Regional Innovation and Change: Path Dependency or Regional Breakthrough

Recommend Documents