What are the major barriers to services trade? To what extent would further liberalisation of trade in services result in increased welfare gains and economic growth? It is critically important for governments to understand this if they are to participate effectively in, and maximise the benefits from GATS negotiations. Empirical evidence is also important to demonstrate the economic effects of market and regulatory changes as an aid to the design of adequate regulatory reform. The papers in this volume, presented at the Quantification Session of the Third OECD Services Experts Meeting, explore fundamental issues for empirical research on trade in services: How can methodologies for measuring the effects of restrictions be sharpened? How can modelling frameworks be improved to increase their relevance for negotiators and trade policy makers? How can the impact of regulatory reform in various services sectors be best captured in empirical models? Ranging from economy-wide empirical assessments to sectoral research in telecommunications services, this volume provides information on available empirical research and highlights the specific data requirements and conceptual challenges for modelling liberalisation of services trade. It also identifies priority areas that will need to be addressed in order to increase the applicability of quantification work to the realities of the services economy and its potential use in a GATS context.
OECD's books, periodicals and statistical databases are now available via www.SourceOECD.org, our online library. This book is available to subscribers to the following SourceOECD theme: Industry, Services and Trade Ask your librarian for more details on how to access OECD books on line, or write to us at
[email protected] www.oecd.org
-:HSTCQE=VUUYW]:
ISBN 92-64-10042-3 22 2003 01 1 P
Quantifying the Benefits of Liberalising Trade in Services
Quantifying the Benefits of Liberalising Trade in Services
« Quantifying the Benefits of Liberalising Trade in Services
Industry Industry Industry Industry Industry Industry Industry Industry Industry Industry Industry Industry
Services Services Services Services Services Services Services Services Services Services Services Services
Trade Trade Trade Trade Trade Trade Trade Trade Trade Trade Trade Trade
Quantifying the Benefits of Liberalising Trade in Services
ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT
ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT
Pursuant to Article 1 of the Convention signed in Paris on 14th December 1960, and which came into force on 30th September 1961, the Organisation for Economic Co-operation and Development (OECD) shall promote policies designed: – to achieve the highest sustainable economic growth and employment and a rising standard of living in member countries, while maintaining financial stability, and thus to contribute to the development of the world economy; – to contribute to sound economic expansion in member as well as non-member countries in the process of economic development; and – to contribute to the expansion of world trade on a multilateral, non-discriminatory basis in accordance with international obligations. The original member countries of the OECD are Austria, Belgium, Canada, Denmark, France, Germany, Greece, Iceland, Ireland, Italy, Luxembourg, the Netherlands, Norway, Portugal, Spain, Sweden, Switzerland, Turkey, the United Kingdom and the United States. The following countries became members subsequently through accession at the dates indicated hereafter: Japan (28th April 1964), Finland (28th January 1969), Australia (7th June 1971), New Zealand (29th May 1973), Mexico (18th May 1994), the Czech Republic (21st December 1995), Hungary (7th May 1996), Poland (22nd November 1996), Korea (2th December 1996) and the Slovak Republic (14th December 2000). The Commission of the European Communities takes part in the work of the OECD (Article 13 of the OECD Convention).
© OECD 2003 Permission to reproduce a portion of this work for non-commercial purposes or classroom use should be obtained through the Centre français d’exploitation du droit de copie (CFC), 20, rue des Grands-Augustins, 75006 Paris, France, tel. (33-1) 44 07 47 70, fax (33-1) 46 34 67 19, for every country except the United States. In the United States permission should be obtained through the Copyright Clearance Center, Customer Service, (508)750-8400, 222 Rosewood Drive, Danvers, MA 01923 USA, or CCC Online: www.copyright.com. All other applications for permission to reproduce or translate all or part of this book should be made to OECD Publications, 2, rue André-Pascal, 75775 Paris Cedex 16, France.
FOREWORD
One session of the Third OECD Services Experts Meeting, organised in conjunction with the World Bank in Paris in March 2002, under the chairmanship of Anders Ahlind, was devoted to the quantification of benefits from services liberalisation. This session brought together trade negotiators, economic modellers and industry experts with a view to exploring the potential use of quantification work in a negotiating context and the applicability of this work to the realities of services sectors. The topics covered ranged from economy-wide investigations relating to trade in services to sector specific empirical research in telecommunication services. This volume reproduces the papers presented at the conference with a view to providing some insights into the different approaches to quantifying the welfare effects of services trade liberalisation. The papers explore fundamental issues for empirical research on trade in services: How can methodologies for measuring the effects of restrictions be sharpened? How can modelling frameworks be improved to increase their relevance for negotiators and trade policy makers? How can the impact of regulatory reform in various services sectors be best captured in empirical models? The volume provides information on available empirical research, highlights the specific data requirements and conceptual challenges for modelling liberalisation of services trade, and identifies areas that deserve priority attention in the future to increase the applicability of quantification work to the realities of services sectors and its potential use in a GATS context. It is hoped that the publication will stimulate further research and help guide empirical efforts in quantifying services barriers and the effects of their removal. We are grateful to everyone who participated in the conference, especially those who prepared papers. Special thanks go to: Aaditya Mattoo and Carsten Fink from the World Bank, Patrick Jomini from the Australian Productivity Commission, Greg McGuire from Economic Insights and Giuseppe Nicoletti from the OECD. This session of the Experts Meeting was planned and implemented by Nora Dihel from the OECD, with valuable guidance from Ken Heydon from the OECD.
3
TABLE OF CONTENTS
Introduction ................................................................................................................................... 8 Chapter 1. The Economy-Wide Effects of Product Market Policies .......................................... 15 Chapter 2. Methodologies for Measuring Restrictions on Trade in Services ............................. 33 Chapter 3. Quantifying the Effects of Liberalising Services: The Experience of the Australian Productivity Commission......................................................................... 63 Chapter 4. Liberalising Basic Telecommunications: Evidence from Developing Countries..... 85 Chapter 5. Quantifying Costs to National Welfare from Barriers to Services Trade a Review of the Literature ....................................................................................................... 113 Chapter 6. Quantification of the Costs to National Welfare of Barriers to Trade in Services: A Scoping Paper ...................................................................................................... 147
Boxes Box 1. Box 2. Box 1.1. Box 2.1. Box 2.2. Box 5.1. Box 6.1. Box 6.2.
Data and modelling requirements for simulating liberalisation of services trade. ........ 9 Recent empirical research at the sectoral level....................................................... ….12 Describing product market regulations across the OECD area. .............................. …19 The effect of different types of restrictions on prices.............................................. …38 Measuring restrictions on trade in services and developing economies.................. …52 Data and modelling requirements for simulating liberalisation of services trade. .. ..116 Methods used for estimating barriers to trade in services. ...................................... ..150 Advances in the weighting of barriers…................................................................... 156
Tables Table 1.1. Table 1.2. Table 1.3. Table 2.1. Table 2.2 Table 2.3. Table 2.4. Table 2.5. Table 2.6. Table 3.1. Table 3.2.
Product market regulatory reform in selected OECD countries ............................ 20 Accounting for differences in employment rates across OECD countries ............ 22 Product market regulation and industry specialisation in the OECD .................... 26 A simple illustrative example of calculating an index........................................... 39 An example of classifying trade restrictions on banking services......................... 41 Results from measuring restrictions on banking services...................................... 44 Price and cost effect measures for developing economies .................................... 47 Price and cost effect measures for developed economies...................................... 48 Sectoral coverage of measures............................................................................... 53 Barriers to services trade ....................................................................................... 68 Tax equivalents of post-Uruguay Round barriers to services trade....................... 69 5
Table 3.3. Table 3.4. Table 3.5. Table 3.6. Table 4.1. Table 4.2. Table 4.3. Table 4.4. Table 6.1. Table 6.2. Table 6.3. Table 6.4. Table 6.5. Table 6.6. Table 6.7. Table 6.8. Table 6.9.
Projected effects on real GDP and income of eliminating services trade barriers................................................................................................................... 70 Contributions of liberalising services to changes in real income .......................... 72 Tax equivalents of post-Uruguay Round barriers in telecommunication services .................................................................................................................. 74 Projected effects on real GNP of liberalising telecommunication markets ........... 75 Effects of individual reforms on mainline penetration and productivity............... 94 Effects of combinations of reforms on mainline penetration and productivity............................................................................................................ 95 Effects of full reform (compared to partial or no reform) on mainline penetration and productivity.................................................................................. 96 Effects of sequencing of reform on mainline penetration and productivity .......... 98 Estimates of services barriers employed in general equilibrium modelling of services trade liberalisation ............................................................................. 152 Ad valorem tariff equivalents: "guesstimates" by 1-digit ISIC services sectors for selected countries ............................................................................... 154 Average gross operating margins of firms listed on national stock exchanges, 1994-96, by sector ............................................................................................... 155 Estimated tariff equivalents in traded services: gravity model based regression method................................................................................................ 157 Price effect measures for banking and telecommunication services ................... 159 New estimates for services barriers ..................................................................... 161 The effect of restrictions on distribution ............................................................. 163 Impact of restrictions on engineering services .................................................... 163 Distribution services: cost impact to establishment, by type of barrier............... 164
Figures Figure 1.1. Figure 1.2. Figure 1.3. Figure 1.4. Figure 1.5. Figure 2.1. Figure 2.2. Figure 2.3. Figure 4.1. Figure 4.2. Figure 4.3.
Regulatory approaches across countries................................................................ 17 Regulation in non-manufacturing industries of OECD countries, 1998................ 18 Employment rates and product market regulations, 1998 ..................................... 21 The contribution of product market liberalisation to changes in the employment rate, 1982-98..................................................................................... 23 The contribution of product market regulation to differences in R&D intensity across countries....................................................................................... 26 An illustration of the results from the trade restrictiveness index ......................... 42 Banking services.................................................................................................... 43 Price effects on banking services........................................................................... 56 Trends in mainline penetration and worker productivity in the telecommunications sector, 1985-99 ..................................................................... 88 Patterns of policy reform in telecommunications .................................................. 89 Example of alternative policy sequences and their effects .................................... 98
6
INTRODUCTION
Abstract. This volume contains the papers presented at the quantification session of the Third Services Experts Meeting. It provides information on available empirical research, highlights the specific data requirements and conceptual challenges for modelling liberalisation of services trade, and identifies areas that deserve priority attention in the future to increase the applicability of quantification work to the realities of services sectors and its potential use in a GATS context. The topics covered range from economywide investigations relating to trade in services to sector specific empirical research in telecommunication services.
INTRODUCTION
To participate effectively in, and maximise the benefits from GATS negotiations, it is critically important that governments know what are the major barriers to services trade and the extent to which further liberalisation of trade in services would result in increased welfare gains and economic growth. They also must have a well-defined domestic reform agenda which supports the opening of services markets. Empirical evidence is again important: to demonstrate the economic effects of market and regulatory changes as an aid to the design of adequate regulatory reform. Against this background, the Quantification Session of the Third Services Experts Meeting held in Paris in March 2002, brought together in one forum, representatives of three communities (GATS negotiators, experts familiar with a selected industry and economic modellers) to discuss possible applicability of quantification work to the realities of services sectors and the potential use of such work in a GATS context. This volume contains the papers presented at the Quantification Session of the Experts Meeting. It provides information on available empirical research and highlights the main questions which need to be addressed in a quantification context. 1. Economy-wide empirical assessments The challenge In principle, measuring the economy-wide impact of trade liberalisation requires a global framework, which captures both the inter-sectoral effects in each economy and the links among countries. General equilibrium modelling is the only technique enabling economy-wide assessments of the impact of trade policy changes. It may also provide some useful insights into broad negotiating modalities (i.e. partial versus complete liberalisation). Therefore, this module focuses on applied general equilibrium studies relating to trade in services. The main objective is to provide some insights on current work on quantifying the effects of services trade liberalisation and to obtain feedback from trade negotiators about the usefulness of this kind of empirical research. While numerous empirical studies concerning the quantification of economic impacts of policies affecting goods trade are available, relatively little work has been done to assess the potential gains from alternative liberalisation scenarios in services. The difficulties arise from the specific data requirements and conceptual challenges arising from the special characteristics of services trade. Box 1 identifies the special data and modelling requirements which should be fulfilled for simulating liberalisation in services sectors.
8
Box 1. Data and modelling requirements for simulating liberalisation of services trade Since many services require proximity between producers and consumers, trade in services includes not only cross-border delivery, but also the movement of suppliers and consumers. The four-part typology of international services transactions adopted by GATS as a basis for multilateral liberalisation negotiations constitutes the generally recognised framework for the analysis of services. It requires separate data for the different modes of services supply as well as the development of a complex modelling framework which considers the different modes of supply, i.e. incorporation of elements which allow the representation of foreign direct investment and movement of natural persons. The restrictions to trade in services typically take the form of non-tariff barriers. Conventional non-tariff instruments of trade policy like quantitative restrictions, price based instruments, licensing or certification requirements, discriminatory access to distribution and communication systems are imposed especially on services providers and classified in the GATS in two main categories relating to market access and national treatment. In addition to the larger spectrum of barriers than in the case of goods, it is necessary to determine whether regulations actually constitute barriers, as one cannot simply equate regulations with restrictions to trade. Further, given that regulations on services are generally designed to serve a range of policy objectives, it might also be relevant to consider whether the regulation is more burdensome than necessary to achieve its policy objective and whether other, equally effective but less trade restrictive, measures might be available. These policy measures are not easy to quantify and require the development of sophisticated measurement methods. Subsequently, the estimates of barriers to trade in services need to be incorporated into a computable general equilibrium model according to the GATS categorisation of market access and national treatment.
The first chapter, the “Economy-wide effects of product market policies” by Giuseppe Nicoletti, discusses recent OECD empirical analysis aimed at shedding light on potential long-run economywide effects of policy reforms in services sectors. It provides quantitative assessments of the impact of regulatory reforms on employment, innovation and productivity performance. The empirical results suggest that regulatory reforms have positive effects not only in product markets, where they tend to increase innovation and productivity, but also for employment rates. These findings show that the reduction of barriers to trade and competition in potentially competitive product markets can be a complement to labour market reforms aimed at increasing longrun employment levels of OECD countries. Also, they indicate that policies aimed at strengthening private governance and increasing competitive pressures improve not only the static allocation of resources but also the dynamic efficiency of OECD economies by positively affecting business R&D spending, technological catch-up and productivity performance. The next two chapters and the background documents deal with methods used for estimating barriers to trade in services as well as the different modelling frameworks employed for simulating services liberalisation. The OECD Trade Directorate has provided a review and discussion of empirical work on measuring and modelling barriers to trade in services. The two documents “Quantification of Costs to National Welfare from Barriers to Services Trade - A Literature Review” and “Quantification of the Costs to National Welfare of Barriers to Trade in Services: A scoping paper”, prepared by Nora Dihel, were included in the background documentation for the meeting. These papers present the methods used for estimating barriers to trade in services (e.g. frequency type measures, direct and indirect price- and quantity- impact measures) as well as the different modelling frameworks employed for simulating services liberalisation (e.g. explicit modelling of FDI versus standard models used to simulate goods trade liberalisation, imperfect competition and increasing returns to scale 9
versus perfect competition and constant returns to scale). The papers outline the main empirical findings concerning the welfare effects derived from the computable general equilibrium literature on services trade liberalisation, relating them to the differences in the model structures and the different estimates of services barriers employed in various studies. Preliminary results Existing studies show that the effects of removing services barriers can be substantial. Generally, economies with initial high protection levels tend to gain most (in terms of percentage gains to GDP). As the values of estimates for services trade barriers were higher for developing countries than for developed countries, it suggests that the potentially major winners would be the developing economies. However, the magnitude of welfare effects is strongly dependent on the accuracy of input data employed (estimates of services barriers) and on the various modelling assumptions. Therefore, results should be taken only as indicative given the difficulties related mainly to poor information on international services transactions and on prevailing barriers to trade in services, as well as to the necessity of developing a different modelling structure than that used for goods trade in order to incorporate the various modes of services supply (i.e. to account for the movement of factors of production). Towards solutions To help address these challenges and make quantification work more useful for policy considerations, two papers were presented at the meeting addressing these critical issues. Estimates of services barriers Amongst all the elements influencing the results of empirical studies into the gains from services trade liberalisation, the estimation of actual services restrictions represents arguably the most critical area. The paper prepared by Greg McGuire, “Methodologies for Measuring Restrictions on Trade in Services”, presents a more detailed analysis of estimates of services barriers employed in general equilibrium models. It indicates the progressive improvement of methods used for estimating services barriers and the remaining limitations. The paper shows that measures of restrictions are valuable for policy makers and negotiators as they illustrate the cost of protection to their economy. It identifies the main methodologies used for measuring restrictions to trade in services, highlighting the difficulties associated with the measurement of regulatory restrictions. It provides a complete overview of existing estimates, discussing the reliability of results and their potential use in negotiating settings. The paper also identifies the different challenges faced by countries in different stages of development. The author concludes that the quality of estimates has been improving progressively both in terms of the range of barriers addressed - which are collected from a growing number of sources- and of measurement techniques employed - which at this stage permit the estimation of price and cost effects of services barriers and begin to determine the correlation between these effects and the individual underlying restrictions. However, in order to sharpen the estimates of services barriers, he examines further directions for improvement in methodology and coverage.
10
Assessment of gains from services liberalisation The assessment of the real benefits of reform are crystallised for negotiators when estimates of services barriers are used in general equilibrium models as they provide insights into the effects and projected real income gains from liberalisation and assist in countering protectionism. The modelling structure employed for simulating services liberalisation is, therefore, a crucial element for generating a realistic measure of the effects of services liberalisation. A general equilibrium model, which supports the analysis of services liberalisation, should include treatment of each of the four modes of supply. The first generation of studies analysing services liberalisation did not model explicitly the different modes of services supply, employing models used to simulate goods trade liberalisation. A number of recent studies explicitly incorporate trade in services through commercial presence, improving substantially the framework for modelling services liberalisation. The paper presented by Patrick Jomini, “Quantifying the Effects of Liberalising Services”, summarises the Australian Productivity Commission’s experience in estimating the effects of aggregate services liberalisation using a modelling framework which explicitly analyses trade in services in the form of cross-border supply and commercial presence. The paper singles out telecommunication services to highlight modelling and data issues in disaggregated analyses. It presents a detailed discussion of results and their decomposition into various contributing factors, highlighting the distinction between the sources of gains from a conventional goods-trade model and a services-trade model. The paper concludes that analysing services trade liberalisation requires progress in areas such as data collection, development of the modelling framework at a modal and sectoral level, and development of methods to incorporate services barriers. 2. Sector-specific focus: Telecommunication services The second module of the Quantification Session focused on empirical research on the effect of regulatory regimes in telecommunication services, given the critical importance of an appropriate regulatory framework in establishing a ‘level playing field’ among operators and protecting the interests of all consumers, as well as the perceived trend towards increased reliance on private investment and competition to foster network expansion and the provision of services Several empirical sectoral studies determine to what extent the removal of barriers to trade in particular services sectors leads to lower prices, improved quality and greater variety if domestic regulatory regimes support the openings of services markets. Box 2 presents some examples of recent empirical research at the sectoral level, with a special focus on telecommunication services. The paper prepared by Carsten Fink and Aaditya Mattoo “Liberalising Basic Telecommunications: Evidence from Developing Countries” deals with issues concerning the general linkages and interactions among different policies affecting telecommunication services. The authors assess how developing economies have fared in profiting from changes in the telecommunications market. They also examine the policy challenges that remain, the design of proper regulatory policies, the sequence of reform, and the embedding of domestic telecommunications in the world market.
11
Box 2. Recent empirical research at the sectoral level Sector-specific general equilibrium studies A number of sector-specific general equilibrium studies have been developed to capture key features of individual services sectors and to highlight modelling and data issues in disaggregated analyses. For example, the sectoral simulations undertaken by the Productivity Commission for telecommunication services [Verikios and Zhang, (2000)] analyse welfare gains under different partial liberalisation scenarios (national treatment and market access liberalisation) as well as under a complete liberalisation scenario for telecommunication services, providing some insights on broad negotiating modalities at the sectoral level. A different sectoral model on telecommunication services, developed by Whalley and Bhattarai (1998), explores the different nature of cross-border liberalisation in network related services, such as telecom services, compared to liberalisation in goods. It gives some indications on the specific effects of sectoral services liberalisation. Studies on the impact of services liberalisation on economic growth A number of econometric studies investigating the impact of liberalisation in specific sectors on economic growth, serve as a timely complement to general equilibrium studies which identify one-time static gains of eliminating or reducing barriers to trade in services. However, as with general equilibrium models, analyses of the economic impact of services liberalisation on economic growth are more sparse than comparable empirical studies on the relation between goods trade liberalisation and long run economic growth. Data constraints and the necessity to consider distinctive, relatively unexplored features of services liberalisation in examining the link between services liberalisation and economic growth make empirical exercises more difficult. Available econometric evidence suggests that openness in services influences long-run growth performance. A study by Mattoo, Rathindran and Subramanian (2001) finds that countries with full open telecommunication and financial markets grow 1.5% faster, on average, than countries closed to trade in services. The study confirms earlier results obtained by Francois and Schuknecht (2000). They found that more open and competitive financial markets increase growth rates by 1.3% to 1.5%. Studies on the effect of regulatory regimes in services sectors on performance Empirical research on the effect of regulatory regimes in services sectors is beginning to emerge. The OECD undertook research to analyse the effects of domestic regulatory regimes on productivity, prices and quality of services in telecommunications, international air passenger transport, electricity supply, and road freight and retail distribution for OECD countries (OECD Economic Studies No.32). The findings generally attribute a positive effect of policy reforms to sector performance. Extending OECD’s research, the Australian Productivity Commission estimated [Doove et al., (2001)] the extent to which regulatory regimes in international air passenger transport, telecommunications and electricity supply have raised prices in a number of OECD and non-OECD economies. The results suggest a positive relationship between the restrictiveness of regulatory regimes and prices in these sectors. The strongest impact is found for international air passenger transport where the system of restrictions on the number of flights between countries and the conditions under which they operate collectively increase airfares by between 3 and 22 per cent. Focusing on developing countries, a World Bank study [Fink et al., (2001)] assesses the impact of alternative policy and regulatory reforms in telecommunication in 12 developing Asian economies. The results generally support the positive contribution of liberal policy to the performance of telecommunication services in the considered economies.
Conclusion All papers prepared for this conference suggest that existing economy-wide models have the potential to assess the effects of multilateral services liberalisation in a realistic manner. Recent theoretical developments and empirical advances have made significant contributions to quantifying the costs of barriers to trade in services. However, there are a number of improvements to be made from both a theoretical and empirical point of view. OECD, in close co-operation with the World Bank, will continue to address these improvements. 12
REFERENCES Bhattarai, K. and J. Whalley (1998), “The Division and Size of Gains from Liberalization of Service Networks”, NBER Working Paper No.6712. Doove, S., O, Gabbitas, D. Nguyen-Hong and J. Owen (2001), “Price Effects of Regulation: Telecommunications, Air Passenger Transport and Electricity Supply”, Productivity Commission Staff Research Paper, AusInfo, Canberra. Fink, C., A. Mattoo and R. Rathindran (2001), “Liberalizing Basic Telecommunications: The Asian Experience”, World Bank Working Paper 2718. Francois, J. and L. Schuknecht (2000), “International Trade in Financial Services, Competition, and Growth Performance”, Centre for International Economic Studies, Discussion Paper No. 6. Mattoo, A., R. Rathindran and A. Subramanian (2001),“Measuring Services Trade Liberalization and its Impact on Trade Growth: An Illustration, World Bank Working Paper 2655. OECD (2001), OECD Economic Studies No. 32, Special Issue on Regulatory Reform, Paris. Verikios, G. and X. Zhang (2000), “Sectoral Impact of Liberalising Trade in Services”, paper presented at the Third Conference on Global Economic Analysis, Melbourne, 27-30 June. Available at: www.monash.edu.au/policy/conf/53Verikios.pdf
13
Chapter 1 THE ECONOMY-WIDE EFFECTS OF PRODUCT MARKET POLICIES
by Giuseppe Nicoletti OECD Economics Department
Abstract. This chapter discusses recent OECD empirical analysis aimed at shedding light on potential long-run economy-wide effects of policy reforms in services sectors. It provides quantitative assessments of the impact of regulatory reforms on employment, innovation and productivity performance. The empirical results suggest that regulatory reforms have positive effects not only in product markets, where they tend to increase innovation and productivity, but also for employment rates.
Introduction and summary* Over the past two decades, many OECD countries have implemented regulatory reforms aimed at promoting competition in product markets. A number of studies have documented the effects of industry-specific reforms on productivity and prices,1 but much less is known about their macroeconomic implications. Yet as reforms spread to an increasing number of industries and involve as well changes in general-purpose regulations (e.g. administrative procedures), their macroeconomic repercussions are likely to be significant. This chapter discusses recent OECD empirical analysis aimed at shedding some light on potential long-run economy-wide effects of policy reforms in product markets.2 It provides quantitative assessments of the impact of regulatory reforms on employment, innovation and productivity performance. Its goal is to present a short summary of the main results obtained in this area. The interested reader is referred to the original sources for more detail on data, empirical methodologies and results. It is important to note at the outset that the main focus of the analysis is on the effects of policies aimed at strengthening private governance (e.g. through privatisation) and liberalising markets where competition is economically viable. To this end, cross-country patterns and time-series developments in regulations that discourage or promote private governance and competition are briefly described by means of summary indicators, and the implications of these regulations for macroeconomic outcomes are explored by means of simple simulations. The empirical results suggest that regulatory reforms have positive effects not only in product markets, where they tend to increase innovation and productivity, but also for employment rates. The finding that economy-wide effects of product market reforms can be significant has two main policy implications:
*
x
The reduction of barriers to trade and competition in potentially competitive product markets can be a complement to labour market reforms aimed at increasing long-run employment levels of OECD countries.
x
Policies aimed at strengthening private governance and increasing competitive pressures improve not only the static allocation of resources but also the dynamic efficiency of OECD economies by positively affecting business R&D spending, technological catchup and productivity performance.
Most of the results reported in the chapter stem from research carried out in the context of the ProductLabour Market Interactions and Growth projects of the Structural Policies Analysis Division at the OECD Economics Department. However, the views expressed are those of the author and should not be held to represent those of the OECD or its member countries.
16
Cross-country patterns of product market regulation in 1998: economy-wide, industry-specific and time-series evidence The OECD has attempted to measure, by means of quantitative regulatory indicators, differences in product market regulation across countries as of 1998 (see Box 1.1 and Annex for details). The guiding criterion for ranking countries has been the impact of regulation on market mechanisms. Specifically, the indicators measure the extent to which entry and other regulations affect product market competition and state control of business sector enterprises affects private governance. Three main types of indicators have been constructed: i) economy-wide indicators, which provide a summary view of the extent to which regulation and state control affect competition and private governance in each country; ii) industry-specific indicators, which provide a more detailed view of the impact of regulation and state control in a number of non-manufacturing industries for which data were available; and iii) time-series indicators, which describe developments in regulation and public ownership over the past two decades in a subset of non-manufacturing countries for which historical information could be found. Figure 1.1 illustrates the respective roles of economic and administrative regulations in shaping economy-wide regulatory environments. Administrative regulation includes reporting, information and application procedures and burdens on start-ups, implied by both economy-wide and sector-specific requirements; economic regulation includes all other domestic regulatory provisions affecting product market competition (such as state control and legal barriers to competition). The United Kingdom remains the least restrictive country on both counts, but economy-wide and/or industry-specific administrative regulations appear to be somewhat heavier in other liberal countries such as Ireland and New Zealand. In 1998, the heaviest administrative regulations were found in France, Italy, Belgium and, to a lesser extent, Japan and Germany.3 Interestingly, countries that have tight economic regulations also tend to impose burdensome administrative procedures on business enterprises.
Administrative regulations
Figure 1.1.
Regulatory approaches across countries
1
Economic and administrative regulation
4
France Belgium
3
Italy
Japan
Germany
Switzerland Spain
Finland
Sweden
2
Ireland
Greece
Netherlands
New Zealand
United States
Denmark
Australia
1
Austria Portugal Norway
Canada United Kingdom Correlation coefficient 0.48 t-statistic 2.42 0 0
1
2
3
4
Economic regulations
1. The scale of indicators is 0-6 from least to most restrictive of market mechanisms. Source: Nicoletti et al. (1999).
17
According to the OECD indicators, regulatory approaches in non-manufacturing industries also varied widely in 1998. The industry-specific indicators are necessarily somewhat imprecise and subject to error, but overall they point to a wide dispersion of industry-specific policies both across and within OECD countries, even when taking into account differences in the technological characteristics of industries. Figure 1.2 illustrates these phenomena by means of a boxplot showing the median level of regulation and the dispersion of regulatory approaches across countries in selected non-manufacturing industry aggregates.4 The juxtaposed boxes reveal how the median level and crosscountry dispersion of regulation varies across industries. The figure suggests that the lowest level of regulatory restrictions and the strongest regulatory convergence across countries is reached in finance and land transport (which, on average, is dominated by road freight), while the highest median level and cross-country variance of restrictions is reported in the air transport, communications and utility industries. It is also noticeable that regulatory restrictions still appear to curb competition in retail distribution and professional services in a number of OECD countries. These data suggest that the scope for pro-competitive reforms appeared still to be very broad until recently in several countries and industries.5 Figure 1.2. Regulation in non-manufacturing industries of OECD countries, 1998 Median and dispersion of regulatory approaches across countries in each industry1 Median and dispersion of product_market_regulation
Scale 0-1 from least to most restrictive
1
0 A.Finance C.Insurance E.Prof.services G.Retail I.Post & tel. B.Land transp. D.Aux.transp. F.Water transp. H.Air transp.
K.Water J.Energy
1. The graph shows, in each industry, the median OECD value of the regulatory indicator (the horizontal line in the box), the third and second quartiles of the cross-country distribution (the area comprised between the edges of each box) and the extreme upper and lower values (the two whiskers extending from the box). Dots identify outlier observations. Table 1.1 shows the historical evolution of regulation in seven non-manufacturing industries (air travel, rail transport, road freight, gas and electricity supply, post and telecommunications) by means of a summary indicator. Over the 1978-98 period, regulatory reform (as measured by both the absolute variation and the percentage decline in the regulatory indicator) was deepest in the United Kingdom, New Zealand, the United States and Australia, while policies changed relatively little in southern European countries, Ireland and Switzerland.6 The evolution of regulatory indicators reveals three main country groupings: the United States, which began regulatory reform at the turn of the 1980s; the United Kingdom, Japan, New Zealand and Canada, which began reforming during the 1980s; and most other countries, which changed regulatory policies over the 1990s.
18
Box 1.1. Describing product market regulations across the OECD area The indicators describing the policy environment in OECD product markets cover both economy-wide and industry-specific regulations which restrict domestic market mechanisms (in potentially competitive environments) and international trade. It is worth stressing that only regulations that have a potential for curbing competition and hindering market mechanisms or international transactions – where competition, market mechanisms and trade are viable – have been included in the regulatory indicators, and these have been constructed in order to rank countries according to an increasing level of unfriendliness to competition. Unavoidably, the construction of the indicators involved a fair amount of discretion, which may affect country rankings and empirical results based on the indicators. All the economy-wide regulatory indicators used in this paper are from Nicoletti et al.,(1999), based on the OECD International Regulation Database (available on line on the OECD Web site). The data on market and industry structure and industry-level product market regulations and the corresponding industry-specific indicators (presented in OECD, 2001a, and Nicoletti et al., 2001) cover most of the energy and marketable service industries (a total of 21 industries and industry aggregates) in (or around) 1998 and, for seven of them, the 197598 period. Depending on the industry, the resulting data set covers barriers to entry, public ownership, price controls, government involvement in business operation, market concentration and vertical integration. In network industries – such as utilities, post and telecommunications and railways – the basic data concerned regulatory and market conditions in different (vertical or horizontal) segments of the industries (e.g. gas production, distribution and supply, or regular and express mail). In manufacturing, the industry-specific regulatory indicators cover only tariff and non-tariff barriers to trade. The industry-specific data were ordered according to a common criterion which reflects the friendliness of regulations, market structures and industry structures to competition (on a cardinal scale from least to most restrictive). Furthermore, the resulting cardinal indicators were rescaled to account for structural differences in * industry characteristics, such as differences in minimum efficiency scale or vertical and horizontal relationships. For each regulatory and market dimension covered in the data set, cross-country indicators at the two-digit (ISIC Rev. 3) industry level were constructed by weighting the indices for lower-digit industries with average ** OECD employment shares. Finally, summary indicators of product market regulation by industry were derived by taking the simple average of the regulatory dimensions covered in each industry. To reconstruct a time series of product market regulation, the data used in this chapter cover regulatory and market developments over the 1975-98 period in seven energy and service industries: gas, electricity, post, telecommunications (mobile and fixed services), passenger air transport, railways (passenger and freight services) and road freight. The coverage of regulatory areas varies across industries. Regulatory barriers to entry are reported for all industries; public ownership is reported for all industries except road freight; vertical integration is documented for gas, electricity and railways; market structure is documented for gas, telecommunications and railways; and price controls are reported for road freight. The aggregate time-series indicator of non-manufacturing regulation was *** constructed by taking a simple average of the summary indicators for the seven industries. Further details about coverage and sources in each of the industries included in the analysis are provided in Nicoletti et al., (2001) and are summarised in the Annex. __________________________________________ *
For instance, indicators for barriers to entry in each industry were rescaled using the OECD average of the frequency of barriers to entry in that industry. As a result, indicators of barriers to entry in structurally competitive industries (such as retail distribution) take by construction a lower range of values than indicators of barriers to entry in industries having natural monopoly elements (such as electricity). **
Aggregation of segments within each industry was made either by simple average (for vertical segments) or with shares in total sales (for horizontal segments). For instance, indicators for postal services were constructed by aggregating indicators for ordinary mail, express mail and parcels using the shares of each of these services in total turnover of the postal services industry. ***
An alternative would have been to aggregate across industries using value added or employment weights. However, value added or employment data at this level of disaggregation are seldom available. More importantly, this would have underevaluated the reform drive in important industries, such as electricity or telecommunications, which have a relatively low weight in GDP but are crucial inputs into overall economic activity. Source: Author.
19
Table 1.1. Product market regulatory reform in selected OECD countries Time series regulatory indicators
1
Scale 0-6 from least to most restrictive
Australia Austria Belgium Canada Denmark Finland France Germany Greece Ireland Italy Japan Netherlands New Zealand Norway Portugal Spain Sweden Switzerland United Kingdom United States
1978
1982
1988
1993
1998
1998-1978
Percentage change
4.5 5.2 5.5 4.2 5.6 5.6 6.0 5.2 5.7 5.7 5.8 5.2 5.3 5.1 5.0 5.9 4.7 4.5 4.5 4.3 4.0
4.5 5.1 5.5 4.2 5.5 5.5 5.9 5.2 5.7 5.7 5.8 5.2 5.5 5.1 5.0 5.9 4.7 4.4 4.5 4.2 3.3
4.2 4.5 5.0 2.8 5.5 4.8 5.7 4.7 5.7 5.1 5.8 3.9 5.5 3.6 4.3 5.4 4.6 4.2 4.5 3.5 2.5
3.3 3.9 4.3 2.6 4.0 4.0 4.7 3.8 5.5 4.8 5.3 3.2 4.1 2.2 3.2 4.9 4.2 3.5 4.4 1.9 2.0
1.6 3.2 3.1 2.4 2.9 2.6 3.9 2.4 5.1 4.0 4.3 2.9 3.0 1.4 2.5 4.1 3.2 2.2 3.9 1.0 1.4
-2.9 -2.0 -2.4 -1.9 -2.7 -3.0 -2.1 -2.8 -0.6 -1.7 -1.5 -2.3 -2.4 -3.7 -2.5 -1.8 -1.5 -2.3 -0.6 -3.3 -2.7
-0.65 -0.39 -0.43 -0.44 -0.48 -0.53 -0.35 -0.54 -0.10 -0.29 -0.25 -0.44 -0.44 -0.73 -0.49 -0.30 -0.31 -0.51 -0.14 -0.76 -0.66
1. Simple averages of indicators for seven industries : gas, electricity, post, telecoms, air transport, railways, road freight. Depending on the industry, the following dimensions have been included: barriers to entry, public ownership, market structure, vertical integration, price controls. Source : Nicoletti and Scarpetta (2001).
The impact of product market regulations on employment Product market regulations that are anti-competitive economy-wide, more specifically in nonmanufacturing industries, are negatively correlated with employment rates in the non-agricultural business sector across OECD countries (Figure 1.3). This points to the output-restraining and wageraising effects of these regulations, which influence negatively labour demand.7 Thus, the procompetitive stance of product market regulations in Australia, the United Kingdom and the United States has been accompanied by a relatively high employment rate in the private sector. On the other hand, relatively stringent state control, barriers to entrepreneurship, and/or barriers to foreign trade and investment in France, Greece, Italy and Norway have coincided with a comparatively low share of the working-age population being employed in the private sector.8
20
Figure 1.3. Employment rates and product market regulations, 1998
Employment rate1 (%) 60 AUT DEU
USA AUS
GBR 55
CHE CAN DNK
50
NLD
PRT
NOR
NZL SWE
ITA
45
BEL FIN
FRA
ESP 40 Correlation coefficient -0.68 t-statistic -3.81
GRC
35 0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
Product market regulation, summary index2
Employment rate1 (%) 60 AUT GBR 55
DEU
USA AUS
CHE CAN DNK
50
NLD NZL
SWE
PRT
NOR
45
ITA
BEL FIN
ESP
FRA
40 GRC
Correlation coefficient -0.53 t-statistic -2.60 35 1
1.5
2
2.5
3
3.5
4
4.5
5
Product market regulation, non-manufacturing industries
1. Non-agricultural business sector. 2. Indicators are scaled 0 to 6 from least to most restrictive of market mechanisms. Source: Nicoletti et al. (2001).
21
5.5 2
Empirical estimates based on a panel of OECD countries over the past two decades confirm the importance of product market regulations for employment performance (Table 1.2). Not surprisingly, differences in labour-market policy settings appear to account for the major part of the cross-country difference in employment rates that can be explained.9 Still, differences in anti-competitive regulation may on average (over the 1982-98 period) explain over one percentage point of the difference in private non-agricultural employment rates between countries. This is equivalent to half of the deviation that can be attributed to the tax wedge (2.5 percentage points) and a third of the deviation that can be explained by employment protection and benefit policies (3.25 percentage points).
Table 1.2. Accounting for differences in employment rates across OECD countries Percentage deviations from OECD average
1
Contribution from: EPL and benefit policies Australia Austria Belgium Canada Denmark Finland France Germany Greece Ireland Italy Japan Netherlands New Zealand Norway Portugal Spain Sweden United Kingdom United States
4.1 -0.1 -2.5 4.9 -3.0 -0.6 -2.3 -2.7 -1.3 3.8 -2.2 1.5 -3.8 3.6 -2.9 -4.6 -3.8 -2.2 6.4 8.0
Tax wedge
Product market regulation
5.0 -1.6 -3.3 2.6 -3.0 -1.8 -2.2 -1.3 0.3 1.9 -1.5 2.3 -3.6 2.9 -0.5 2.0 0.5 -3.6 2.4 3.1
1.0 -0.2 -0.7 1.6 -0.6 -0.2 -1.4 0.0 -2.0 -1.4 -1.9 0.4 -0.8 1.8 0.4 -1.5 -0.1 0.8 2.2 2.9
Other 2
Total
0.9 23.0 -1.8 -1.3 9.7 -0.4 -4.6 17.7 -22.8 -22.7 0.7 26.0 1.0 -12.7 -2.1 -1.6 -12.7 10.1 -3.9 -3.9
11.1 21.1 -8.3 7.7 3.0 -3.0 -10.5 13.7 -25.8 -18.4 -5.0 30.2 -7.2 -4.4 -5.0 -5.7 -16.1 5.0 7.1 10.2
1. Based on parameter estimates from pooled cross-country/time-series regressions covering 20 OECD countries over the period 1982-98. 2. Includes effects of bargaining systems, unionisation, output gaps and other unexplained factors (country-specific effects and residuals). Source: OECD (2001c).
22
In some countries the effect of product market regulations is estimated to be particularly strong. For instance, in Italy, where on average employment rates are 5 percentage points lower than the mean of OECD countries, anti-competitive product market regulations account for about one-third of this gap. Also, comparatively stringent regulations may keep employment rates in France, Greece, Ireland and Portugal 1.5 to 2 percentage points below the OECD mean. Conversely, in the United Kingdom and the United States, product market regulations that promote competition account for about onefourth of their better-than-average employment rate performance (respectively 10 and 7 percentage points above the mean of the OECD countries). Over the past two decades, regulatory reforms have played a significant role in increasing employment in the OECD area (Figure 1.4). This is notably the case for countries where pro-competition policy developments have been particularly extensive. Thus, product market reforms in New Zealand and the United Kingdom are estimated to have added around 2.5 percentage points to their employment rate in the non-agricultural business sector over the 1982-98 period. On the other hand, countries where regulatory reform has made more modest progress have experienced correspondingly smaller employment gains, with Greece, Italy and Spain only adding around 0.5 to 1 percentage point to their employment rate via such reforms.
1
Figure 1.4. The contribution of product market liberalisation to changes in the employment rate, 1982-98
3.0 2.5 2.0 1.5 1.0 0.5
Ita
ly Sp ai n Ire la nd Po rtu ga l Ca na da A us tri a Fr an ce Ja pa n Sw ed Ne en th er la nd s Be lg iu m N or w ay D en m U ar ni k te d St at es G er m an y A us tra lia Fi U nl ni an te d d K in g do N ew m Ze al an d
G
re ec
e
0.0
1. The figure reports the estimated impact on the employment rate in the non-agricultural business sector of pro-competitive regulatory reform in 7 non-manufacturing industries (gas, electricity, post, telecommunications, passenger air transport, railways and road freight). Depending on the industry, changes in the following dimensions have been considered: barriers to entry, public ownership, market structure, vertical integration and price controls (see Nicoletti and Scarpetta, 2001, and Nicoletti et al. 2001, for full regression results). Source : OECD (2001c ).
23
There is still considerable scope for most member countries to expand employment via regulatory reforms in the product market. As an illustrative gauge of this unused potential, the OECD has assessed the employment consequences if countries were to align their overall regulatory stance on that of the United States. This would involve a fundamental overhaul of the extent of state control, of barriers to entrepreneurship and of constraints on external trade and investment in countries with relatively stringent regulations, but these countries would also experience the greatest employment gains. Thus, some southern European countries and Ireland might add as much as 2 to 2.5 percentage points to their employment rate compared with 1998. Smaller, but still noticeable, employment gains could be obtained in countries with less regulated product markets.10 Product market policies, productive efficiency and innovation A well-performing product market is characterised by high rates of multi-factor productivity (MFP) growth, which reflects changes in output that are not embodied in changes in factor inputs.11 Improvements in MFP play a crucial role in the process of economic growth. Recent OECD estimates suggest that, in most countries, MFP growth accounted for between one-third and one-half of the average business sector GDP growth observed over the past two decades (Scarpetta et al., 2000). There are basically three ways in which MFP improvements can be achieved: eliminating slack in the use of resources, adopting the most efficient technologies and increasing innovative effort. The OECD empirical research reported below suggests that, by affecting the incentives to innovate and improve efficiency, regulations that limit product market competition (e.g. by imposing entry or operational restrictions) can have important effects on innovation, technology diffusion and MFP performance. Technological catch-up and multi-factor productivity The main effect of pro-competitive product market regulations is to strengthen the incentives to improve MFP and adopt new technologies. Cross-country empirical evidence at industry level suggests that MFP is positively affected by regulatory environments that favour competition, even after accounting for other potential influences, such as R&D and country- and industry-specific factors (Scarpetta et al., 2002). Pro-competitive regulation appears to improve MFP performance both directly and, especially, by enabling faster catch-up to best practice in countries that are far from the technological frontier. For example, product market reforms that would align the regulatory stance with that of the most liberal OECD country are estimated to reduce in the long run the MFP gap vis-à-vis the leading country by around 10% in high-gap countries such as Greece and Portugal, and by 2-4% in several other continental European countries and Japan. Product market policies can also have repercussions on the turnover of firms, which is an important determinant of economy-wide improvements in MFP. Product market regulations that raise the cost of entry (e.g. administrative burdens) or prevent it altogether (e.g. legal limitations on the number of competitors) tend to lower industry-specific and aggregate entry rates. Recent OECD empirical work based on firm-level data covering both manufacturing and services in ten OECD countries over a decade shows that countries with relatively low administrative barriers and pro-competitive sector-specific regulations typically experienced higher entry rates of small-sized firms (Scarpetta et al., 2002). The effect of regulations on entry is particularly important for productivity performance in industries in which technology is rapidly evolving, such as ICT industries or industries characterised by high ICT adoption. In these industries new entrants play an important role in introducing new generations of technology. Therefore, product- and labour-market regulations that minimise the 24
prospective costs faced by new entrants are likely to create favourable conditions for increasing the contribution of ICT to productivity growth. Innovation Innovation is one of the main sources of technological progress, MFP improvement and, ultimately, economic growth (OECD, 2000a). Economic analysis suggests that product market competition is as important as protection of intellectual property rights in generating the incentives to engage in innovative activity. Therefore, the right policy environment for innovation is one that strikes the right balance between ex ante competitive pressures and ex post protection of intellectual property rights.12 The benefits of competition for R&D mainly derive from the effects of competitive pressures on the incentives of managers to implement cost-reducing innovations, improve product quality or diversify products in order to preserve or acquire market shares. Indeed, empirical evidence shows that strong product market competition is beneficial for innovative activity when intellectual property rights (IPR) are adequately protected.13 For instance, recent empirical research carried out at the OECD showed that economy-wide and industry-specific product market regulations that strengthen private governance and enhance competition have a positive effect on R&D intensity in manufacturing (Nicoletti et al., 2001; Bassanini and Ernst, 2002). Cross-country differences in such regulations explain a good deal of the deviations of manufacturing R&D intensity from the OECD average in member countries in the late 1990s (Figure 1.5). Regulations that promote competition were estimated to explain more than one-third of the excess R&D intensity in the United States, Japan, Germany and Sweden relative to the OECD average and provide a large positive contribution in the United Kingdom, Canada and Ireland. The opposite effect was particularly strong in Italy and Greece, where excessive regulatory restrictions on competition accounted for one-third and two-thirds, respectively, of the shortfall in R&D intensity relative to the OECD average. Regulatory restrictions also provided a large negative contribution to R&D in France and Belgium. These estimates suggest that further liberalisation of product markets might spur innovative activity significantly in several OECD countries. Besides their effects on R&D intensity in individual industries, product market policies can also affect the propensity of a country to concentrate production in innovative industries, for instance by affecting the pace of resource reallocation in the economy. Cross-country evidence on the relationship between regulatory regimes and industry specialisation is very preliminary. However, simple crosscountry correlations suggest that state interference in private governance decisions and regulations that increase the cost of doing business for R&D-intensive firms (e.g. administrative burdens) are associated with specialisation patterns that are unfavourable to innovative industries (Table 1.3).14 If this suggestive evidence is confirmed by further empirical analysis, the implication for policy would be that strengthening competitive forces could also stimulate structural adjustment towards innovative industries, thereby raising overall innovative activity.
25
Figure 1.5.
The contribution of product market regulation to differences in R&D intensity across countries Percentage deviations from OECD average1
Per cent 150 Product market regulation Other factors (2) 100
Total
50
0
-50
Ita ly
re ec e Po rtu ga l
G
n
Sp ain
Sw
Ja pa
ed U en ni te d St at es Fi nl an d Fr an ce G er m an y Be lg iu m A us tri a D en U m ni a te rk d K in gd om No rw ay N eth er la nd s Ire la nd Ca na da
-100
1. Adjusted for industry composition. 2. Includes employment protection legislation, other controls, country-specific effects and unexplained residual. Source: Nicoletti et al. (2001) and Bassanini and Ernst (2002).
Table 1.3. Product market regulation and industry specialisation in the OECD 1
Cross-country correlation coefficients
Aggregate R&D intensity -0.45 *
Product market regulation State control
-0.79 **
Barriers to entrepreneurship
0.19 2
Excessive adm. burdens on corporations
-0.64 **
Barriers to trade
-0.19
1. Cross-country correlations between the component of manufacturing R&D intensity that is due to industry composition and product market policies (see Nicoletti et al., 2001, for full details on the breakdown of R&D intensity in within-sector and industry composition components). *, ** denote significance at the 5% and 1% level, respectively. 2. Difference between burdens imposed on corporations and sole-proprietor enterprises. Source: Nicoletti et al. (2001).
26
ANNEX Industry-specific product market regulation: coverage and sources
Industry
Electricity
Gas manufacture and distribution
ISIC code Revision 3
401
402
Period
Regulatory and market dimensions covered1
1998
P, E, PO, MS, VI
1975-1998
E, PO, VI
1998
P, E, PO, MS, VI
1975-1998
E, PO, MS, VI
Energy
40
1998
E, PO, VI
Water works and supply
41
1998
Electricity, gas and water
40_41
Wholesale trade
Industrial segments covered
Prod., Trans., Dist.
Prod., Trans., Dist.
Countries covered
Main sources2
24-25
OECD
21
OECD, EC, PI, WB
26
OECD, EC, PI, WB
21 25
OECD, EC, PI, WB
E, PO, VI
23
OECD, EC, PI, WB
1998
E, PO, VI
23
OECD, EC, PI, WB
50_51
1998
E, PO
25
OECD
Retail trade
52
1998
E, CBO
28
OECD
Restaurant and hotels
55
1998
E
25
OECD
Railways
601
Road freight
602
1998
P, E, PO, MS, VI
1975-1998
E, PO, MS, VI
Prod., Trans., Dist.
27 Passenger, freight
21
OECD, ECMT
1998
P, E, CBO
27-29
OECD
1975-1998
P, E
21
OECD, ECMT
P, E
27
OECD, ECMT
Land transport
60
1998
Water transport
61
1998
E, CBO
22
APC
1998
E, PO, MS
27
OECD
1975-1998
E, PO
Air transport carriers
Passenger
62
21
OECD, EC
60_62
1998
E
22
OECD, ECMT EC, APC
Supporting services to transport
63
1998
E, PO
21
OECD
Post
641
1998
P, E, PO, VI
Telecoms
642
Communication
Transport
1975-1998
Letter, parcel, express
22-26 21
1998
P, E, PO, MS, VI
1975-1998
E, PO, MS
64
1998
P, E, PO, MS
26
OECD
Financial institutions
65
1998
E, CBO
23
OECD, APC
Insurance
66
1998
P, E
12
OECD
Legal services
7411
1998
E, CBO
22
APC
Accounting services
7412
1998
E, CBO
23
APC
Architectural and engineering services
7421
1998
E, CBO
23
APC
74
1998
E, CBO
22
APC
Professional business services
Note 1: P = Price regulation E = Barriers to entry PO = Public ownership CBO = Constraints to business operation MS = Market structure VI = Vertical integration
Fixed, mobile
Life, general, health
20-29
OECD, EC, UPU
21
Note 2: ECMT = European Conference of Ministers of Transportation EC = European Commission WB = World Bank PI = Privatisation International APC = Australian Productivity Commission UPU = Universal Postal Union
27
OECD
NOTES 1.
See OECD (1997), Chapter IV in OECD (2000a) and the sector-specific papers contained in OECD (2001a). OECD (1997) also contains an early attempt to quantify the economy-wide effects of reforms.
2.
The main sources are Nicoletti and Scarpetta (2001), Nicoletti et al. (2001), Bassanini and Ernst (2002), Scarpetta et al. (2002), Chapter VI in OECD (2001c) and Chapter VII in OECD (2002a).
3.
Some of these countries have implemented significant regulatory reforms since 1998. For instance, see OECD (2001d) for Greece, and OECD (2001b) and Nicoletti (2002) for Italy.
4.
The boxplot shows, in each industry, the median OECD value of the regulatory indicator (the horizontal line in the box), the third and second quartiles of the cross-country distribution (the area comprised between the edges of each box) and the extreme upper and lower values (the two whiskers extending from the box). Dots identify outlier observations.
5.
Even in countries that have widely implemented pro-competitive policies there are industries (e.g. retail distribution in the United Kingdom and professional services in the United States) in which regulation could be brought closer to best practice by further reforms.
6.
For Ireland, this points to a potential conflict between relatively liberal economy-wide policies (e.g. trade and general purpose administrative regulations) and relatively restrictive regulatory policies in some industries. For more details, see OECD (2002b).
7.
See Blanchard (2000) and Nickell (1999).
8.
While France and Italy partly compensated for restrictive domestic policies by extensive openness to international trade and investment, Greece and Norway also appeared to have relatively high barriers to trade and investment according to the OECD indicators.
9.
Labour market policy variables and product market regulations explain only 40% of the variation in employment rates across countries; the remaining 60% of the variation is due to labour market institutions, output gaps and unexplained factors.
10.
These figures tend to underestimate the potential employment gains from product market reforms because they do not take into account the possible indirect effects of these reforms on labour market arrangements (e.g. the effects of enhanced product market competition on the bargaining power of insiders).
11.
MFP represents residual output growth once the direct contribution of changes in the quality and quantity of capital and labour inputs are accounted for. Therefore, MFP estimates involve a number of difficult measurement problems. For instance, it is hard to adjust for quality and compositional changes in labour input and, especially, capital stock. Other potential sources of measurement error are economies of scale and mark-up pricing (see Morrison, 1999).
12.
When the coverage of IPR is too limited in scope or over time, potential innovators expect to be unable to fully recover the value of their sunk R&D investments, because competitors may wipe out innovation rents, for instance by imitating the new products. This is likely to discourage innovative activity. Conversely, when IPR are extensive but ill-designed, competitive pressures can induce firms to adopt innovation strategies aimed at eliminating competitors or pre-empting the entry of new firms.
28
13.
For a recent review, see Ahn (2002).
14.
For instance, there is a significant negative cross-country correlation between industry specialisation in innovative industries and the OECD indicators of excessive administrative burdens on corporations (i.e. the difference between burdens imposed on corporations and sole-proprietor enterprises) and anticompetitive product market regulations (see Nicoletti et al., 2001).
29
REFERENCES
Ahn, S. (2002), “Competition, innovation and productivity growth: a review of theory and evidence”, OECD Economics Department Working Papers, No. 317, Paris. Bassanini, A. and E. Ernst (2002), “Labour market institutions, product market regulation and innovation: cross-country evidence”, OECD Economics Department Working Papers, No. 316. Blanchard, O. (2000), “Rents, product and labor market regulation, and unemployment”, Lecture 2 of The Economics of Unemployment: Shocks, Institutions, and Interactions, Lionel Robbins Lectures, London School of Economics. Morrison, C.J. (1999), Cost Structure and the Measurement of Economic Performance, Kluwer Academic Press, Norwell, Massachusetts. Nickell, S. (1999), “Product markets and labour markets”, Labour Economics, Vol. 6. Nicoletti, G. (2002), “Institutions, Economic Structure and Performance: Is Italy Doomed?”, ISAE Annual Report on Monitoring Italy, April, Rome. Nicoletti, G. A. Bassanini, E. Ernst, S. Jean, P. Santiago and P. Swaim (2001), “Product and labour market interactions in OECD countries”, OECD Economics Department Working Papers, No. 312. Nicoletti, G., and S. Scarpetta (2001), “Interactions between product and labour market regulations: D they affect employment? Evidence From OECD countries”, paper presented at the Banco de Portugal Conference on “Labour Market Institutions and Economic Outcomes”, 3-4 June, Cascais. Nicoletti, G., S. Scarpetta and O. Boylaud (1999), “Summary indicators of product market regulation with an extension to employment protection legislation”, OECD Economics Department Working Papers, No. 226. OECD (1997), OECD Report on Regulatory Reform I-II, OECD, Paris. OECD (2000a), OECD Economic Outlook, No. 67, OECD, Paris. OECD (2001a), OECD Economic Studies No. Special Issue on Regulatory Reform, OECD, Paris. OECD (2001b), Regulatory Reform in Italy, OECD, Paris. OECD (2001c), OECD Economic Outlook, No. 70, OECD, Paris. OECD (2001d), Regulatory Reform in Greece, OECD, Paris. OECD (2002a), OECD Economic Outlook, No. 71, OECD, Paris. OECD (2002b), Regulatory Reform in Ireland, OECD, Paris. 30
Scarpetta, S., A. Bassanini, D. Pilat and P. Schreyer (2000), “Economic Growth in the OECD Area: Recent Trends at the Aggregate and Sectoral Level”, OECD Economics Department Working Papers, No. 248, OECD, Paris. Scarpetta, S., P. Hemmings, T. Tressel and J. Woo (2002), “The role of policy and institutions for productivity and firm dynamics: evidence from micro and industry data”, OECD Economics Department Working Papers, No. 329, Paris.
31
Chapter 2 METHODOLOGIES FOR MEASURING RESTRICTIONS ON TRADE IN SERVICES
by Greg McGuire Economic Insights Pty Ltd, Brisbane, Australia
Abstract. This chapter identifies the main methodologies used for measuring restrictions to trade in services, highlighting the difficulties associated with the measurement of regulatory restrictions. Also, it indicates the progressive improvement of methods used for estimating services barriers and the remaining limitations. The paper provides a complete overview of existing estimates, discussing the reliability of results and their potential use in negotiating settings. In addition, it identifies the different challenges faced by countries in different stages of development.
Introduction* The services sector is the largest and most important sector of an economy. Not only do economies derive the bulk of their employment and income from services, but many services – financial, telecommunications and transport – are vital intermediate inputs for the production of other goods and services. International cross-border trade in services accounts for about 20% of total trade, and, with global integration and technological developments, services are increasingly being traded. The efficiency of this sector is crucial for the efficiency of the overall economy. However, many governments hinder efficiency in the sector and international trade by imposing restrictions on services without understanding the effects. Powerful domestic interest groups influence governments to impose and maintain restrictions. These restrictions impose high costs on societies. The benefits accrue to a few, but the costs are imposed on the whole community. Measuring the costs of restrictions, although difficult, crystallises the costs of protection for governments, reveals the benefits that will accrue from their removal and is an impetus for reform. Measuring restrictions on services is more difficult than measuring restrictions on trade in goods. International trade in goods involves the exchange of a physical product between a producer and consumer, and restrictions on such trade usually take the form of tariffs. The effect of trade restrictions on the price of goods can be measured relatively easily by the amount of the tariff. In contrast, trade in services involves a less tangible exchange between the producer and the consumer, and restrictions usually take the form of government regulation. However, in itself, regulation does not mean a restriction with an adverse impact on economic efficiency. A certain level of government regulation is necessary to meet economic and social objectives, but it is difficult to judge the level of regulation that is “more than necessary” or restrictive and the amount that is “necessary”. However, it is imperative to make this judgement to measure a restriction’s size and its effect on market outcomes or, more specifically, the price of a service. Measures of restrictions are valuable for negotiators. Negotiators can use internationally comparable measures of protection to illustrate the costs of maintaining restrictions to their trading partners. They also realise the costs of protection to their economy, and particularly to consumers in sectors where restrictions significantly raise prices above world prices. However, the real benefits of reform become clear to negotiators when the measures are used in sophisticated general equilibrium models. Results from these models provide insight into the effects and projected real income gains from liberalisation and help to counter the protectionist forces that work against liberalisation.
*
I am grateful for comments by Professor Christopher Findlay of the Australian National University and Ms. Nora Dihel of the OECD.
34
Why measure? International trade in cross-border services represents about USD 1.4 trillion or about 20% of world exports (WTO, 2001). Global economic integration, technological developments and liberalisation have led to a continuing expansion of the volume and range of traded services. International transactions, which in earlier times would have been impossible or prohibitively expensive, have become commonplace because of the ease with which people move and communicate across international borders. For example, the fall in the cost of international air services in recent decades has made tradable many tourist services that were previously only available to domestic consumers. Restrictions on trade in services impose costs, usually in the form of higher prices for businesses and consumers. Restrictions limit domestic and international competition, decrease efficiency and permit incumbent services suppliers to charge prices above those in a competitive market. Estimating the extent to which restrictions increase prices shows the benefits of removing restrictions for consumers, policy makers and trade negotiators. Robust measures of restrictions on trade in services and the use of these measurers in general equilibrium models represent a powerful tool for analysing the effects and project the gains of liberalisation. There is a significant amount of methodological thinking still required on modelling services trade liberalisation, but a number of studies have projected the benefits – Benjamin and Diao (1998; 2000), Brown et al. (1996), Chadha (2000), Chadha et al. (2000), Dee and Hanslow (2000), DFAT (1999), Hertel et al. (1999), Robinson et al. (1999). The results are similar in a number of respects: x
There are always substantial global real income gains from services trade liberalisation. In many studies, the gains in terms of real income are similar to or greater than liberalisation of trade in agriculture and manufacturing combined.
x
Developing economies gain more than developed economies. Economies with higher restrictions, mainly developing economies, reap the greatest benefits from liberalisation.
x
Liberalisation of trade in services has powerful impacts on agriculture and manufacturing through inter-sectoral linkages in an economy. Services are essential inputs and an increase in the supply of more efficient services produces substantial productivity gains for other sectors (McGuire, 2002).
The projected gains from services trade liberalisation provide an impetus for reform, but the results are dependent on robust and accurate measures of the effect of restrictions on trade in services. Measures of the costs of services protection are also useful in multilateral and regional trade negotiations. Services are becoming increasingly prominent in trading agreements. WTO members have settled on negotiating guidelines and are currently tabling proposals on the structure and content of the new round of negotiations on services. New regional trading agreements now cover services and existing regional trading agreements are being extended to services. Robust and comparable cross-country measures of the costs of restrictions on trade in services are useful for negotiators in light of the increasing importance of services trade in international forums. Negotiators can use measures of restrictions to illustrate to their trading partners the costs of maintaining restrictions. Negotiators (and policy makers) also realise the costs of protection for their
35
economy, and more particularly for consumers in those sectors that purchase services above world prices. It is crucial for trade negotiators and policy makers to understand the nature of services, trade in services and, most importantly, the effects of restrictions on trade in services. Such understanding feeds into the development of negotiating priorities so as to maximise the net benefits and reduce unintended consequences of liberalisation in an area where there is little empirical work. The downside risk of getting it “wrong” is significant given the importance of the services sector (Dee, 2001). What to measure? Researchers measure the effect of restrictions on market outcomes – profit margins or price-cost margins, prices and costs. This provides the necessary input for general equilibrium modellers to simulate the global welfare gains of removing restrictions. Measuring restrictions and their effects raises the vexed question of defining a restriction. Restrictions can be any type of measure that limits the most efficient supply of a service to a consumer. Restriction can be imposed on services, such as limits on their distribution, or on a services supplier, such as a limit on the number of firms in a given services market. The effect of restrictions can be partially or fully reflected in the price of a service. Regulatory or non-regulatory measures can impose restrictions. Regulatory restrictions are those imposed by government that also affect competition, such as unduly stringent licensing requirements. Non-regulatory measures are usually private-sector practices that restrict effective competition in a market, such as the exclusive buyer-seller networks that exist in some distribution markets for services. In some instances, regulation is necessary to correct the effects of non-regulatory restrictions and ensure effective competition (see below). There are also non-regulatory and regulatory restrictions on trade in services, which have the effect of limiting the free flow of international trade. Such restrictions are often only thought of as being discriminatory or treating foreign services suppliers less favourably than domestic ones. However, restrictions can be non-discriminatory. Non-discriminatory restrictions treat domestic and foreign services suppliers equally but are “more than necessary” to ensure the quality of a service. For example, a restriction that prohibits the entry of foreign services suppliers into a market is discriminatory. A restriction that prohibits the entry of all services suppliers is non-discriminatory. The prevalence of non-discriminatory measures that affect trade in services makes it difficult to differentiate measures with legitimate regulatory objectives from those that are intentionally or unintentionally restrictive. Most discriminatory measures aim to restrict trade, but some have legitimate regulatory objectives. Governments regulate when markets fail to produce optimal economic outcomes and when it is necessary to address environmental, equity and other social objectives that are unlikely to be met by market forces (Coghlan, 2000). Regulations that meet such objectives, using the most efficient regulatory instruments, can be considered to achieve optimal economic outcomes or, for measurement purposes, an “optimal level” of regulation. Above that optimal level, too much regulation can be restrictive, stifling competition, restricting trade and increasing prices of services for the consumer. However, lower levels of regulation in some sectors do not necessarily mean better regulation. In some sectors, regulation is essential to foster competition and produce optimal economic outcomes. In network sectors, such as telecommunications, elements of the system possess natural monopoly characteristics and these 36
services need to be regulated to facilitate access to networks for new operators, limit market power of large incumbents and ensure that efficient prices are charged. While too much regulation can be restrictive, too little regulation can also have adverse economic effects. Too little regulation can create externalities by being less stringent than necessary to achieve the desired objective. For example, governments regulate financial services to ensure stability, professional services to ensure quality and transport services to ensure safety. Kalirajan (2000) argues that too little regulation can also be a restriction. For example, inadequate protection of intellectual property rights can deter foreign services suppliers from establishing a presence in an economy. From a measurement perspective, the “more than necessary” or restrictive component of regulation needs to be determined. In practice, this is achieved by measuring the difference between the level of regulation that is “necessary” (or optimal) and the total level of regulation. This difference can be considered a restriction on trade in services. This approach is practical, given the many constraints that come with international comparisons of regulation, but a number of qualifications need to be made. There are always difficulties involved in judging an optimal level of regulation for use in empirical work. Among other things, judgements need to be made about the suitability of the instruments chosen to meet the desired objectives and the appropriateness of the levels at which the instruments are set. International assessments of regulatory regimes are even harder, as governments often pursue different policy objectives, and economies’ needs, priorities, values and circumstances differ (Doove et al., 2001). Linking levels of regulation to market outcomes Restrictions on trade in services affect the economic performance of services suppliers (profit margins, price-cost margins). Too much regulation protects incumbent services suppliers from competition, allowing them to increase prices and expand their price-cost margins. Many economies impose many and varied restrictions on the entry of foreign services suppliers by prohibiting entry, limiting the number of firms or imposing strict licensing requirements. These restrictions protect domestic services suppliers from international competition, allowing them to charge prices above those they would have to charge in the absence of restrictions (Box 2.1). Many restrictions are thought to increase prices, but some increase the costs of services suppliers (Dee, 2001). Most economies regulate to ensure that services are provided at or above a certain standard. This may have the effect of prohibiting potential or existing services suppliers from operating efficiently and may thus push up business costs. The higher costs may be partially or fully reflected in price increases when services suppliers can pass these costs on to consumers in the form of higher prices. Some restrictions increase both price and cost of services. These restrictions limit competition and require some type of standards to be met. Some licensing requirements for services suppliers have these characteristics. Regulatory restrictions are one of many factors that affect the price-cost margins of services suppliers. Non-regulatory aspects of services markets, such as the economic cycle, substitutability of products, competition, market structure and many others, are determinants of price-cost margins. An analysis of the effects of restrictions on price-cost margins needs to take these factors into account and separate the regulatory from the non-regulatory influences (see Box 2.1).
37
Box 2.1. The effect of different types of restrictions on prices Restrictions on trade in services tend to limit the entry or the operations after entry of services suppliers in a market. Many of these restrictions affect the price of services. The effect of different types of restrictions that only affect prices is depicted in the figure below. Ds is the demand for a service in an economy. For simplicity, the services market is competitive, the only effects on prices are restrictions and supply curves are horizontal. Foreign services suppliers are also assumed to be more efficient at supplying the service than domestic ones. When regulation is at an optimal level, Po is the lowest price at which a service can be supplied to ensure its quality, that is, there are no unnecessary discriminatory or non-discriminatory restrictions. Po can also be viewed as the world price. Prices start at Pm and fall as restrictions are removed. Pm is the price at which there is a monopoly supplier and government regulation prohibits new entrants. Under these conditions, the consumer pays the highest price for a service. Pd is the price at which there is free and open domestic competition among domestic services suppliers, but no international competition. Pf is the price of a service when there are restrictions on foreign services suppliers. Foreign services suppliers are able to enter and operate in the market but have more onerous restrictions than domestic ones. Domestic services suppliers continue to supply services in the market but at the price set by the more efficient foreign services supplier. Pn is the price of a service with non-discriminatory restrictions – applied equally to foreign and domestic services suppliers – which are “more restrictive than necessary” to ensure the quality of the service. The imposition of different types of anti-competitive restrictions on domestic and foreign services suppliers raises the price of services in the domestic market above Po. Too much regulation can have adverse effects on prices, but too little regulation can have adverse effects on quality. Pi is the price at which there is insufficient regulation. The price of the service is lower than at Po but the reliability or quality of a service is poor. Many restrictions are thought to be price-increasing, but some restrictions are cost-increasing or can be a combination of price-increasing and cost-increasing. Pure cost effects, although not illustrated, will increase the costs of a services supplier. Price and cost effects will increase the prices and costs. The extent of the increase in prices is dependent on the ability of services suppliers to pass on these increased costs in the form of higher prices to consumers. Source: Adapted from Findlay and Warren (2000).
Methodologies for measuring restrictions There are a number of methodologies for measuring restrictions on trade in services that aim to capture the effect of restrictions on trade in services. The methodologies can be divided into those that: x
Measure the level of restrictions on services, converting qualitative information about restrictions into comparable quantitative information.
x
Measure the effect of restrictions on price-cost margins of services suppliers, by measuring the effect of restrictions on prices and/or costs of services suppliers.
While measuring the effect of restrictions on price-cost margins is the objective, some studies use measures of the levels of restrictions to estimate the effects on price-cost margins. An understanding of both methodologies is useful. The methodologies for measuring restrictions are somewhat simpler to understand and interpret. Methodologies for measuring the level of restrictions Restrictions on trade in services are typically measured using an index. An index is a system of scores and weights that converts qualitative information about restrictions into quantitative information based on the number and severity of restrictions. A services sector in an economy with 38
fewer and less severe restrictions is more liberal than an economy with a greater number of restrictions. Steps for measuring the level of restrictions Although research on measuring restrictions has been undertaken at different times by different researchers, the methodologies all use an index measure. The methodology for all studies can be generally summarised as follows. x
Step 1: Collect information on restrictions. Information on restrictions or regulation is collected from sources such as the WTO or may already be available, for example from the General Agreement on Trade in Services (GATS) schedules.
x
Step 2: Classify information on restrictions. Similar restrictions are grouped or classified together so that their relative level of restrictiveness can be compared across economies.
x
Step 3: Develop an index. An index is developed from commonly classified restrictions on the basis of their level of restrictiveness or openness. An index uses scores and weights to convert qualitative information on restrictions into quantitative information. Scores are assigned to restrictions on the basis of a judgement as to their stringency. The more stringent the restriction, the more restrictive or less open the service. The score for an index typically ranges from 0 to 1.
x
Step 4: Calculate the index. The index is calculated for restrictions in an economy’s services sector.
Table 2.1 provides a simple illustrative example of calculating an index for banking licences, one type of restriction on banking services. Where an economy issues no new banking licences and prohibits the entry of all new banks, the economy is assigned the most restricted score of 1 under a restrictiveness index or the most liberal score of 0 under an openness index. Where banking licences are granted subject to the applicant meeting prudential requirements, the economy is assigned the least restricted score of 0 under a restrictiveness index or the most liberal score of 1 under an openness index. Common restrictions on services include restrictions on the licensing of services suppliers, direct investment in services suppliers, business activities and the movement of natural persons. Table 2.1. A simple illustrative example of calculating an index
Weight for type of restriction 0.25
Score for an openness index 0.00
Score for a restrictiveness index 1.00
Restrictions on banking licences
0.50
0.50
No foreign banking licences available
1.00
0.00
Banking licences granted subject to meeting prudential requirements
No new banking licences available
Source: McGuire and Schuele (2000).
The body of restrictions that apply in a services sector are then weighted on the basis of a judgement as to their relative restrictiveness or openness. For example, restrictions on banking licences are weighted more heavily than other restrictions such as on the temporary movement of natural persons. Some researchers use a different weighting methodology. For example, Mattoo (1998) uses a trade-weighted system for GATS commitments. A commitment made against the commercial
39
presence mode of supply is weighted more heavily than a commitment made against consumption abroad. The majority of studies construct an index according to which the score for a services sector ranges from 0 to 1 (Table 2.1). The accuracy of the results is more important than the range or scale. Some studies use other scales: Claessens and Glaessner (1998), McGuire (1998), McGuire and Schuele (1999) and Warren (2000a) use a scale of 1 to 5; Mattoo et al. (2001) use a scale of 1 to 9. Evolution of methodologies for measuring the level of restrictions Hoekman (1995) pioneered the use of an index to measure restrictions on trade in services. He developed a three-category weighting method for assessing the openness of GATS commitments. Generally, the more commitments made by a WTO member, the more liberal the economy. The GATS schedules of all WTO members were examined and a score was assigned to each commitment. A score of 1 was assigned where a member made a “bound” commitment; a score of 0.5 where a commitment was specified in a GATS schedule; and a score of 0.0 where an “unbound” commitment was made. The higher the score the more open the economy. A number of other researchers have used the same or a similar methodology to measure restrictions on trade in services (Kemp, 2000; Marko, 1998; Mattoo, 1998; and Mattoo et al., 2001). Two major weaknesses of the Hoekman methodology were recognised by Hoekman (1995), Hardin and Holmes (1997) and many others. First, Hoekman interpreted “no commitments” as meaning that restrictions exist and applied the most restricted score. However, in many cases an economy had no restrictions but, for whatever reason, made no GATS commitments. Further, some economies may not make commitments simply because they do not have the relevant sectors or activities. For example, few economies produce space transport services. Second, all specified commitments in a GATS schedule are assigned an equal weight in the index. No account is taken of the likely differences in the economic impact of commitments. An economy that lists a market access restriction in the form of a 49% limit on foreign ownership is assigned the same score as an economy that lists a screening process for foreign direct investment (FDI). These weaknesses provided a stimulus for researchers to collect the best available information on restrictions and score restrictions according to their level of restrictiveness or openness. Studies by Claessens and Glaessner (1998), Kalirajan (2000), Mattoo et al. (2001), McGuire (1998), McGuire and Schuele (2000), McGuire et al. (2000) and Nguyen-Hong (2000) collected information on restrictions for services sectors from the Asia-Pacific Economic Cooperation (APEC), the International Monetary Fund (IMF), the World Bank, the WTO, the OECD, the United States Trade Representative (USTR) and many other sources. Common types of restrictions in services sectors were identified and an index based on the level of restrictiveness or openness was developed and then calculated. The extent of the collection of information on restrictions and the sophistication of the index are the major differences among these studies. Trade restrictiveness index The most sophisticated methodology to date is that originally developed by McGuire and Schuele (2000) for banking services and then applied by Kalirajan (2000), McGuire et al. (2000) and NguyenHong (2000) for other services sectors. These studies compile a comprehensive database on restrictions from various sources and develop a detailed index for different types of restrictions. The methodology warrants further discussion. 40
These studies classified restrictions in two ways. The first is by whether a restriction applies to: x
Establishment: the ability of services suppliers to establish physical outlets in an economy and supply services through those outlets.
x
Ongoing operations: the operations of a services supplier after it has entered the market.
The second way a restriction can be classified is by whether it is: x
Non-discriminatory: restricting domestic and foreign services suppliers equally.
x
Discriminatory: restricting only foreign services suppliers.
This two-by-two classification is similar to that used in the GATS schedules of commitments. Restrictions on establishment (or commercial presence) include those affecting services delivered via FDI. Restrictions on ongoing operations can affect services delivered by cross-border supply, consumption abroad or the presence of natural persons (other modes of supply recognised under the GATS). Non-discriminatory restrictions are similar to the GATS’ limitations on market access, and discriminatory restrictions are similar to limitations on national treatment. Table 2.2 provides an example of how trade restrictions on banking services are classified. Table 2.2. An example of classifying trade restrictions on banking services Establishment Mode of supply: commercial presence Nondiscriminatory
The number of banking licences is restricted.
Ongoing operations Modes of supply: cross-border, consumption abroad, movement of natural persons Banks are restricted in the manner in which they can raise funds.
Discriminatory
The number of foreign banking licences is restricted.
Foreign banks are restricted in the manner in which they can raise funds.
Source: McGuire (2000).
The distinction between different types of restrictions is important in a general equilibriummodelling context (Dee and Hanslow, 2000). The reason for the distinction between restrictions on establishment and on ongoing operations is so that the former can ideally be modelled as “taxes” on the movement of capital and the latter modelled as “taxes” on the output of services suppliers. The classification also makes possible an analysis of the effects of removing certain types of restrictions, such as the gains from removing non-discriminatory restrictions on establishment. An index score is calculated separately for domestic and foreign services suppliers. A foreign index is calculated to measure all restrictions that hinder foreign firms from entering and operating in an economy. It covers both discriminatory and non-discriminatory restrictions. A domestic index represents restrictions applied to domestic firms and covers non-discriminatory restrictions. The difference between the foreign and domestic index scores is a measure of the discrimination against foreigners. Figure 2.1 provides an illustration of the results from the trade restrictiveness index.
41
Figure 2.1. An illustration of the results from the trade restrictiveness index Score 1.00
Trade restrictiveness index The restrictiveness index measures the number and severity of restrictions on trade in services for foreign and domestic service suppliers. The foreign and domestic indices include restrictions on establishment and ongoing operations. Index scores generally range from 0 to 1. The higher the score the more restrictive an economy.
0.90
0.80
0.70
0.60
0.50
0.40
0.30
0.20
Foreign index A measure discriminatory discriminatory foreign service foreign index domestic index.
Discrimination A measure of restrictions that only apply to foreign service suppliers.
of
nonand restrictions on suppliers. The includes the
Domestic index A measure of discriminatory restrictions all service suppliers.
0.10
nonon
0.00
Economy X
Source: McGuire (2000).
In calculating a score for an overall economy, it was not determined which restrictions might be justified to enhance the efficiency of a services sector and which might not. In general, trade restrictions, by reducing competition in a services market, will reduce that market’s efficiency. However, regulation that limits competition is sometimes necessary to deal with “market failure” and to meet particular social objectives. No assessment was made of the merits or otherwise of the restrictions covered by the trade restrictiveness index or the “optimal level” of regulation. As mentioned above, it is extremely difficult to assess the merits of regulation for economies with different regulatory objectives and structures. Figure 2.2 shows the results of the trade restrictiveness index for banking services in low- and middle-income economies (LMIEs) (or developing economies) and the average for high-income economies (HIEs) (or developed economies). The focus is on developing economies because they are more restrictive and have more variability than developed economies. World Bank (2001) provides the groupings for LMIEs and HIEs.
42
1,2
Figure 2.2. Banking services Score 0.70 Foreign index Domestic index
0.60
0.50
Score
0.40
0.30
0.20
0.10
Average HIEs
Venezuela
Average LMIEs
Economy
Uruguay
Turkey
Thailand
South Africa
Philippines
Mexico
Malaysia
South Korea
Indonesia
India
Colombia
Chile
Brazil
Argentina
0.00
1. Scores range from 0 to 1. The higher the score the more restrictive and economy. 2. Based on available information of restrictions in place as at 31 December 1997. Source: McGuire and Schuele (2000).
In this grouping, Brazil, India, Indonesia, Malaysia and the Philippines have the most restricted markets for banking services. These economies are all characterised by very tight entry controls and restrictions on business operations. Generally, they limit new foreign bank entry, strictly limit foreign equity participation and prohibit banks from expanding their existing operations. Chile, Korea, Thailand, Turkey and Uruguay are moderately restrictive. These economies have at least one significant restriction that limits foreign access to their markets. They restrict licensing or foreign equity participation in domestic banks or impose restrictions on their operations such as the opening of new outlets or street branches. Argentina, Colombia, South Africa and Venezuela are the least restricted and have fewer restrictions on licensing and foreign equity participation. Brazil, India and Indonesia have the most discriminatory restrictions against foreigners for banking services as measured by the large difference between the foreign and domestic index scores. Developed economies are far less restrictive and discriminatory than developing economies. The average trade restrictiveness index scores for developed economies are considerably lower than those for developing economies. McGuire (2002) discusses the results for other services sectors. Productivity Commission (2001) provides trade restrictiveness indices for a number of services sectors (such as telecommunication services, banking services, maritime transport, distribution and professional services) and has greater coverage in terms of both developing and developed countries. Results from different methodologies A comparison of the results of different studies measuring restrictions on banking services generally shows similar results (Table 2.3). Where necessary, the results from similar studies are
43
converted to restrictiveness index scores ranging from 0 to 1 and ranked from most to least restricted. The results provide a “check” on the consistency of results in different studies. As would be expected, the level of the scores varies between studies with different methodologies but the ranking of the economies is generally similar. For all studies, Asian and South American economies are generally found to be more restrictive than European and North American economies. Table 2.3. Results from measuring restrictions on banking services
1
Score and rank Economy
Malaysia India Indonesia Philippines Brazil Uruguay Korea Chile Thailand Turkey Singapore Colombia Japan South Africa Venezuela Mexico Australia Hong Kong, China Switzerland Canada 4 EU
McGuire and Schuele (2000)
2
Mattoo et al. (2001)
Mattoo (1998)
Claessens and Glaessner (1998), McGuire (1998) and McGuire and Schuele 3 (1999)
Score
Rank
Score
Rank
Score
Rank
Score
Rank
0.65 0.60 0.55 0.53 0.51 0.46 0.43 0.40 0.39 0.37 0.37 0.23 0.19 0.19 0.17 0.17 0.12 0.09 0.08 0.07 0.07
1 2 3 4 5 6 7 8 9 10 10 11 12 12 13 13 14 15 16 17 17
0.44 0.44 0.22 0.44 0.44 0.56 0.44 0.22 0.44 0.11 0.11 0.22 nc 0.11 0.56 0.33 0.11 0.11 0.11 0.11 0.11
2 2 4 2 2 1 2 4 2 5 5 4 na 5 1 3 5 5 5 5 5
0.84 0.80 0.72 0.76 0.80 0.80 0.80 0.80 0.92 0.40 0.64 0.80 0.16 0.20 0.80 0.60 0.27 0.37 0.16 0.36 0.36
2 3 5 4 3 3 3 3 1 8 6 3 13 12 3 7 11 9 13 10 10
0.52 0.55 0.36 0.33 nc nc 0.41 0.44 0.36 nc 0.50 0.35 0.34 nc nc 0.38 0.16 0.05 nc 0.20 nc
2 1 7 10 na na 5 4 7 na 3 8 9 na na 6 12 15 na 11 na
Argentina 0.07 17 0.11 5 0.16 13 nc na United States 0.06 18 0.11 5 0.36 10 0.10 13 New Zealand 0.06 18 0.11 5 0.16 13 0.08 14 nc – not calculated. na – not applicable. 1. The restrictiveness index scores range from 0 to 1. The higher the score the more restrictive the economy. Claessens and Glaessner (1998), Mattoo (1998), Mattoo et al. (2001) and McGuire (1998) have been converted from openness indices to restrictiveness indices and normalised to a score of 1 for comparison. 2. Mattoo et al. (2001) covers financial services and these results are used as a proxy for banking services. 3. McGuire (1998) and McGuire and Schuele (1999) extend the economy coverage of Claessens and Glaessner (1998). 4.The score for the 15 EU member states was the same in each study. Source: Claessens and Glaessner (1998), Mattoo (1998), Mattoo et al. (2001), McGuire (1998) and McGuire and Schuele (1999).
44
Methodologies for measuring the effect of restrictions There are generally three methodologies for measuring the effect of restrictions on the prices and/or costs of services. These are: x
Converting an index measure to a tariff equivalent. This methodology converts results from an index to tariff equivalents based on an arbitrary judgement about the conversion factors.
x
Indirect methodology. This methodology determines a benchmark price for a service and attributes part or all of a price above the benchmark price to the effect of restrictions.
x
Direct methodology. This methodology directly measures the effects of restrictions, as measured by an index, on the price-cost margins of services suppliers.
Conversion of an index on restrictions to a tariff equivalent Hoekman (1995) uses an index as a starting point for estimating tariff-equivalent measures of the relative degree of restrictions on trade in services. He arbitrarily defines a set of benchmark “guesstimates” of tariff equivalents for each sector to reflect an economy that is highly restricted with respect to market access. A value of 200% is chosen for the most restricted sectors such as postal and telecommunications services, while values between 20% and 50% are assigned to more open sectors such as tourism and education services. Each economy and sector is then assigned a value related to that benchmark. For example, the benchmark for postal services is set at 200; if an economy has an index score of 50% for postal services, its tariff equivalent for that sector is 100. For economies that make no commitments for postal services, the index score is 0 and the tariff equivalent for the sector is 200. Indirect methodology Indirect methodologies determine a benchmark price and attribute part or all of a divergence from the benchmark price to restrictions. The researchers that produced these methodologies recognise that prices above the benchmark may be a result of factors other than restrictions such as market size, market structure and the business cycle. Francois and Hoekman (1999) use the size of gross operating margins and a gravity model approach to estimate the effect of restrictions on gross operating margins. Under this approach, a benchmark is chosen and compared with the gross operating margins of services suppliers. The higher gross operating margins are suggestive of higher restrictions. A scaling factor is used to calculate a tariff equivalent. For the ten services sectors studied, the results generally show high margins in finance and hotels and lower margins in distribution and consulting services. Under the gravity model approach, the “natural level” of bilateral services trade without restrictions is estimated and compared with the “actual level” of services trade with restrictions. The extent to which the actual level of services trade is below the natural level or “free trade” benchmark is attributed to the effect of restrictions. Hong Kong (China) and Singapore are used as free trade benchmarks. It was found that tariff equivalents for services trade are moderate for business and financial services and high for construction services.
45
Direct methodology The most sophisticated methodologies are those that directly estimate the effect of restrictions on the price and/or cost of services suppliers (Kalirajan, 2000; Kalirajan et al., 2000; Nguyen-Hong, 2000; Warren, 2000b).1 An econometric model is developed which includes all the relevant determinants of economic performance of services firms in that services sector – industry- and economy-wide influences – plus a measure of trade restrictions, as measured by an index. The econometric model is then used to estimate the determinants of economic performance in that services sector. Wherever possible, the components of the trade restrictiveness index are entered separately so that the econometrics reveal something about the relative effects of different types of restrictions. Price and cost measures are calculated from the results. Depending on the performance measure chosen, the results provide an indication of the extent to which restrictions are priceincreasing, cost-increasing or a combination of both. Tables 2.4 and 2.5 show the results for price- and cost-effect measures for developing and developed economies. They show the extent to which restrictions, as measured by the trade restrictiveness index, increase the price and/or cost of services. The price- and cost-effect measures are up to 150% for developing economies and up to 32% for developed economies. Restrictions have a greater effect on prices in banking, telecommunications and engineering than other sectors. The price-effect measures on most foreign services suppliers are up to 150%. These measures show the extent to which all restrictions on foreigners – discriminatory and non-discriminatory – increase the price of services. In banking, telecommunications and engineering services, India, Indonesia, Malaysia and the Philippines generally have extremely high price effects for foreign suppliers. The European Union, Hong Kong (China), New Zealand and United States generally have low price effects. The price-effect measures on most domestic services suppliers are also up to 150%. These measures show the extent to which non-discriminatory restrictions on foreign and domestic services suppliers increase the price of services. In banking and telecommunications services, Chile, India, Indonesia and Malaysia generally have high price effects. Australia, the European Union, New Zealand and United States generally have low price effects. Restrictions on establishment contribute the most to increasing the price and cost of services. These are mainly restrictions on market access which include restrictions on the licensing of new firms, FDI and requirements for foreigners to enter the market through a specific type of legal entity. Foreign and domestic cost-effect measures for distribution and engineering services are significantly lower than the price effects. These restrictions increase the costs of services suppliers but by a lesser extent than the price effect.
46
5.34 45.56 34.00 18.35 55.08 49.33 60.61 13.40 47.36 14.90 36.73 33.06 31.54 40.34 13.44 33.27
Banking 3.81 5.68 1.68 24.27 >150.003 138.41 16.08 14.43 72.85 20.89 8.43 55.12 33.53 11.92 14.94 38.14
Tele-communications
Price
Foreign
na na na na na 10.22 11.98 14.17 na 3.73 na na na na na 10.03
Engineering1 na na 1.32 na na 3.66 8.23 na na 0.47 na na na na na 3.42
Distribution2
Cost
0.00 0.87 23.20 3.54 3.54 5.35 22.11 0.00 10.99 0.00 14.93 0.00 3.54 11.00 0.00 6.60
Banking
>150.003 70.70 6.73 6.24 21.43 13.77 4.30 29.90 19.59 7.61 9.57 23.98
3.81 3.81 1.68 10.55
Telecommunications
Price
Domestic
na na 1.92 na na 0.00 3.97 na na 0.00 na na na na na 1.47
Distribution
2
Cost
na na na na na 3.23 5.28 1.95 na 0.69 na na na na na 2.79
1
Engineering
1.. Nguyen-Hong (2000) calculated price and cost effects for one professional service – engineering. Insufficient data were available to calculate price and cost effects for accountancy, architectural and legal services. 2. These cost effects are for restrictions on establishment. 3. These economies have significantly large price effects which are greater than 150% but are capped at 150%. Source: Kalirajan (2000), Kalirajan et al. (2000), Nguyen-Hong (2000) and Warren (2000b).
na – not available. Insufficient data are available to estimate a price and/or cost effect for these economies.
Argentina Brazil Chile Colombia India Indonesia Malaysia Mexico Philippines South Africa Korea Thailand Turkey Uruguay Venezuela Average
Economy
Percentage
Table 2.4. Price and cost effect measures for developing economies
9.31 5.32 5.32 5.34 5.32 5.32 5.32 5.32 5.32 6.91 5.32 5.32 15.26 5.32 5.32 4.69 5.32 31.45 5.32 5.32 5.95 5.32 4.75 7.11
Banking
1
0.31 0.85 1.31 3.37 0.20 0.00 1.43 0.32 4.52 1.26 2.67 1.00 0.26 1.05 0.20 0.27 6.25 2.72 3.93 0.65 1.23 0.00 0.20 1.48
2.82 14.54 0.52 5.31 1.14 2.28 0.92 10.17 na 5.06 na na 6.57 na 3.67 na na 5.04 8.73 6.76 na 2.54 7.38 5.21
Telecommunications Engineering
Price
Foreign
0.57 na 4.87 3.09 na na 5.16 na 0.25 0.06 2.70 na 2.26 na 2.73 0.77 na 0.03 na na 5.24 2.76 2.26 2.34
Distribution
Cost 2
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 2.65 0.00 0.00 10.03 0.00 0.00 0.00 0.00 8.15 0.00 0.00 0.00 0.00 0.00 0.91
Banking
0.31 0.85 0.65 1.07 0.20 0.00 0.34 0.32 2.56 1.26 1.46 1.00 0.26 1.05 0.20 0.27 3.80 2.10 2.03 0.65 1.23 0.00 0.20 0.95
Telecommunications
Price
Domestic
0.00 na 6.69 0.98 na na 7.10 na 0.00 0.00 0.00 na 6.79 na 0.00 0.00 na 0.00 na na 8.32 0.00 0.00 2.13
Distribution
2
1
2.09 6.78 0.66 2.68 0.69 0.73 0.69 2.93 na 2.34 na na 2.24 na 5.25 na na 0.78 3.86 0.74 na 1.39 3.79 2.35
Engineering
Cost
1. Nguyen-Hong (2000) calculated price and cost effects for one professional service – engineering. Insufficient data were available to calculate price and cost effects for accountancy, architectural and legal services. 2. These cost effects are for restrictions on establishment. Source: Kalirajan (2000), Kalirajan et al. (2000), Nguyen-Hong (2000) and Warren (2000b).
na – not available. Insufficient data is available to estimate a price and/or cost effect for these economies.
Australia Austria Belgium Canada Denmark Finland France Germany Greece Hong Kong (China) Ireland Italy Japan Luxembourg Netherlands New Zealand Portugal Singapore Spain Sweden Switzerland United Kingdom United States Average
Economy
Percentages
Table 2.5. Price and cost effect measures for developed economies
How can the measures be improved? Research on measures of restrictions and on measures of the effect of restrictions on price-cost margins has contributed significantly to economic thinking on measuring these types of barriers. This research is a starting point from which to improve these measures, particularly in terms of methodology and coverage. Improvements in methodology Methodology for measuring the effect of restrictions on trade in services is in its infancy, but research has progressed since the pioneering work of Hoekman (1995). More comprehensive internationally comparable information on restrictions and financial data has also contributed significantly to methodologies for measuring the effect of restrictions on the prices and costs of services. Before efforts to improve these methodologies are made, it is useful to look at some desirable attributes of such a methodology. It should: x
Collect comprehensive information about all known restrictions on trade in services.
x
Use an index to convert qualitative information about restrictions into internationally comparable quantitative information.
x
Estimate the effect of restrictions, as measured by an index, on the price-cost margins of services suppliers after taking account of other influences on price-cost margins.
An advantage is that this methodology produces estimates of the effects of restrictions that are explicitly linked to characteristics of the restrictions themselves, rather than being generated from an “unexplained residual” (Dee, 2001). However, this methodology can be improved in a number of areas, and some of the improvements are often a function of available information on restrictions and financial data. Chapter 6 in this volume and Dee (2001) also discuss some of the methodological improvements. The major difficulty in estimating the effect of restrictions on trade in services is determining the “optimal level” of regulation. This is used as a basis for measuring the level of a restriction or the difference between the total level of regulation and the optimal level. This requires choosing “appropriate” benchmarks or setting up criteria for the attributes of “best practice” regulation. Progress has been made in discussing some of the characteristics of best practice regulation for services sectors, but these characteristics have not being converted to empirical benchmarks for different economies with different regulatory objectives and structures (Productivity Commission and Australian National University, 2000). Different types of restrictions can affect prices, cost or a combination of the two. Kalirajan et al. (2000) and Warren (2000b) attributed the total effect of all restrictions to price increases. Kalirajan (2000) and Nguyen-Hong (2000) estimated some of the effects of restrictions that increase prices and some that increase costs. Separating the different effects of restrictions on prices and costs is difficult, but important, as they are modelled differently. The removal of restrictions that increase prices are modelled as improvements in allocative efficiency and the removal of restrictions that increase costs are modelled as productivity gains. This has implications for the results obtained from general equilibrium models (Dee, 2001). 49
The methodology uses a system of scores and weights to measure restrictions based on a judgement about the relative stringency of restrictions. This system is sometimes criticised as being “arbitrary”, but it is an essential component of measuring restrictions. Restrictions on trade in services, by their nature, are qualitative. Estimating the effect of restrictions requires converting qualitative information into quantitative or numerical values. The only way to convert this information is to use an index methodology, as it takes into account the number of restrictions and their restrictive effect. A pure frequency ratio only counts the number of restrictions. In spite of the criticism of arbitrariness, Table 2.3 shows that results are very similar when different index methodologies are used. In most of the current methodologies, the indexes are constructed after reviewing the literature on the economic impact of different types of restrictions. Restrictions are rigorously assigned a score, but the weights are more arbitrarily chosen and can have a significant effect on the overall result. (This being said, the weights are applied uniformly across all economies in the sample so as to give internationally comparable results.) Subject to adequate in-sample variation and large enough sample sizes, econometric techniques can be used to determine the precise weights of different classes of restrictions. In the econometrics, the overall index score is included in the equation as a variable. The overall index score can be disaggregated into its components and entered separately to reveal the effects of different types of restrictions. Thus, the econometrics provides non-judgemental and precise data-driven estimates of the weights (Dee, 2001). The methodology could be further disaggregated for modes of supply. There are various reasons for disaggregating for modes of supply, most importantly to assess the precise impact of restrictions in a services sector on different modes of supply. The measures assess all restrictions on a services sector, but some restrictions may have a smaller impact than others because of the possibility of switching between modes of supply. Take for example a restriction on FDI in a domestic services supplier. FDI may be the preferred and most efficient mode of supply for a service, but other modes are also available. Under the current methodology, a restriction on FDI would have a significant effect on prices and/or cost. Switching to the next best alternative mode of supply, which is, say, free of restrictions may be slightly less efficient, but the service can still be provided. Therefore, the cost of a restriction through one mode of supply needs to be weighed against the benefit of supply through another mode. As technological developments increasingly offer many ways for a service to be supplied, such an alternative is likely to become more prevalent. There may also be other benefits to disaggregating for modes of supply. General equilibrium modellers currently recognise that different types of restrictions - establishment and ongoing operations – need to be modelled differently (see above). As the sophistication of these models increases, it is quite possible that restrictions may need to be modelled differently for each mode of supply. Disaggregation may also stress to negotiators the importance of reducing restrictions via modes of supply and increasing their consistency with the GATS. Services negotiations are likely to place increasing emphasis on modes of supply and the restrictions imposed on the them (WTO, 2000). International trade data are only available for one mode of supply, cross-border trade, and statisticians are endeavouring to collect data for the other modes. This will make it possible to compare the results of measures of the effect of restrictions with individual modes of supply. This may then enable negotiators to focus on removing restrictions on those modes of supply through which the majority of trade occurs for a particular service.
50
Improvements in coverage Coverage of the measures of services sector and economies varies. Table 2.6 matches sectoral services studies on measures and measures of the effect of restrictions with the WTO Sectoral Classification of Services (WTO, 1991). There is better coverage of the services sectors for measures of restrictions than for measures of their effects. This may be because the methodology for the indexes is reasonably simple and easy to calculate. For both measures of restrictions and measures of their effects, Hoekman (1995), Hardin and Holmes (1997) and Francois and Hoekman (1999) dominate the sectoral coverage. Hoekman’s coverage is extensive, but, by using the GATS schedules, he relies on incomplete information about restrictions. The coverage of Hardin and Holmes also appears to be extensive, but only cover restrictions on FDI for APEC member economies. Francois and Hoekman have good coverage of sectors but the methodology infers the effect of restrictions from the size of gross operating margins. Most studies, including the more methodologically rigorous ones, tend to focus on the more important services sectors – telecommunications, financial and transport. Researchers have a good knowledge of the issues and results in these sectors, and this can be used to improve the measures. Researchers also need to turn their attention to sectors where there are little or no rigorous measures – construction and engineering, education, environment, health, tourism and travel and recreational services. Extensive coverage of economies is also lacking in a number of studies. Most studies cover 20-40 economies and those with more extensive coverage tend to rely on incomplete information to calculate an index of the effect of restrictions on price-cost margins. Almost all studies cover developed economies, but numbers decrease significantly for developing economies (Box 2.2). Timeseries data on the effects of restrictions would be useful for analysing the effects of restrictions over time, although this would involve collecting information about restrictions on services trade for more than one time period, which is a large task. Many of these coverage issues can be overcome by use of a comprehensive and “easy-to-use” database of restrictions on trade in services. Researchers to date have independently collected and compiled information on restrictions or used information that is already available such as the GATS schedules. Collection of information on restrictions is extremely time-consuming and a co-ordinated effort on identification, collection and storage of such information would pay large dividends for all trade practitioners. There are currently many tariff databases for trade in goods but no comprehensive database for trade in services. In trade in goods, the UNCTAD TRAINS database is a comprehensive information system that provides data for each tariff line on tariffs, para-tariff and non-tariff measures as well as import flows by origin for more than 100 countries. There is no comparable database for services. UNCTAD, the OECD, the World Bank and the Australian Productivity Commission have data on restrictions but there is no co-ordinated effort to centralise this information and keep it regularly updated.
51
Box 2.2. Measuring restrictions on trade in services and developing economies Most studies measuring restrictions and their effects cover developed economies. The coverage of measures decreases significantly as methodologies are applied to developing economies. This is unfortunate because general equilibrium modelling shows that developing economies would experience major structural changes and gain the most from liberalisation (Dee and Hanslow, 2000). Robust measures of restrictions would help developing economies to understand the precise nature and extent of the effects and benefits from liberalisation. However, there are a number of difficulties in applying the methodologies for measuring restrictions to developing economies. The methodologies fundamentally rely on information about restrictions. Developing economies tend to have little or no regulation in some services sectors or “sweeping” laws that are applied on a case-by-case basis. Restrictions on the entry of one foreign services supplier may be different from those imposed on another seeking to enter the same market. Information for developed economies is obtained from domestic and international reports on trade policy. Such detailed information does not exist for developing economies. The GATS, while incomplete because of the positive listing approach, illustrates this information gap. High-income economies scheduled about 45% of their services while low-income economies scheduled about 12% (Hoekman, 1995). Comparable international financial data on services suppliers are also difficult to obtain for developing economies. The Worldscope database is often used to measure the effect of restrictions on price-cost margins (Disclosure, 2000). The data are obtained from publicly listed companies on domestic stock exchanges. In developing economies, poor corporate governance and reporting requirements produce poor financial data and sample sizes that are insufficient for estimating the effect of restrictions on price-cost margins. Some methodologies also rely on measuring the level of restrictions or of “too much” regulation and then estimating the effects of removing restrictions. However, many developing economies have “too little” regulation, most of which is poorly structured and implemented. For example, a contributing factor to the 1997 Asian financial crisis was poor prudential regulation. Affected economies were required to improve their institutional prudential framework and standards as part of the IMF Stand-by Credit Facilities (Fischer, 1998). These limitations may weaken the results and may require taking a slightly different approach. Information gaps might be filled by seeking feedback on restrictions and their effects from foreign services suppliers operating in a developing economy market. Internationally comparable survey information could be an alternative to collecting information on actual restrictions. An assessment of the level of regulation in developing economies may also be needed before estimating the effect of restrictions. Source: Author.
52
Table 2.6. Sectoral coverage of measures WTO sectoral classification of services 1. Business services Professional services
Measures of restrictions x Hoekman (1995) x Hardin and Holmes (1997) x Colecchia (2000)3
1,2
Measures of the effect of restrictions x Hoekman (1995) x Francois and Hoekman (1999) 4
x Nguyen-Hong (2000)
4
x Nguyen-Hong (2000) Computer and related services Research and development services Real estate services Rental/leasing services without operators Other business services 2. Communication services Postal services
x Hoekman (1995) x Hardin and Holmes (1997)
Courier services
x Hardin and Holmes (1997) x Hardin and Holmes (1997) x Warren (2000a) x Doove et al. (2001) x Mattoo (2001)
Telecommunication services
Audio-visual services
x Hoekman (1995)
x Warren (2000b) x Doove et al. (2001)
x Hardin and Holmes (1997)
Other communication services 3. Construction and engineering services
x Hoekman (1995) x Hardin and Holmes (1997)
x Hoekman (1995) x Francois and Hoekman (1999)
x Hoekman (1995) x Hardin and Holmes (1997) x Kalirajan (2000)
x Hoekman (1995) x Kalirajan (2000)
Construction work for buildings Construction work for civil engineering Installation and assembly work Building completion and finishing work Other construction and engineering services 4. Distribution services
Commission agents’ services Wholesale trade services
x Francois and Hoekman (1999)
Retailing services
x Francois and Hoekman (1999)
Franchising Other distribution services 5. Education services
x Hoekman (1995) x Hardin and Holmes (1997) x Kemp (2000)
Primary education services Secondary education services Higher education services Adult education services Other education services
53
x Hoekman (1995)
Table 2.6. Sectoral coverage of measures (cont.) WTO sectoral classification of services 6. Environmental services
Measures of restrictions
Measures of the effect of restrictions
x Hoekman (1995) x Hardin and Holmes (1997)
x Hoekman (1995)
x x x x x
x Hoekman (1995) x Francois and Hoekman (1999)
Sewage services Refuse disposal services Sanitation and similar services Other environmental services 7. Financial services
All insurance and insurance-related services
Hoekman (1995) 3 Mattoo (1998) 3 Mattoo (2001) 1 Hardin and Holmes (1997) Claessens and Glaessner (1998), McGuire 2 (1998), McGuire and Schuele (1999) c
Banking and other financial services
x Mattoo (1998) x Hardin and Holmes (1997) x Claessens and Glaessner (1998), McGuire
3
x Kalirajan et al. (2000)
2
(1998), McGuire and Schuele (1999)
x Mattoo (1998) c x McGuire and Schuele (2000) c
8. Health services
x Hoekman (1995) x Hardin and Holmes (1997)
x Hoekman (1995) x Francois and Hoekman (1999)
x Hoekman (1995) x Hardin and Holmes (1997)
x Hoekman (1995)
Hospital services Other human health services Social services Other 9. Tourism and travel services
x Francois and Hoekman (1999)
Hotels and restaurants Travel agencies and tour operators services Tourist guides services Other tourism and travel services 10. Recreational services
x Hoekman (1995) x Hardin and Holmes (1997)
Entertainment services News agency services Libraries, archives, museums and other cultural services Sporting and other recreational services Other recreational services
54
x Hoekman (1995) x Francois and Hoekman (1999)
Table 2.6. Sectoral coverage of measures (cont.) WTO sectoral classification of services 11. Transport services Maritime transport services
Measures of restrictions x Hoekman (1995) x Hardin and Holmes (1997) x McGuire et al. (2000)
Measures of the effect of restrictions x Hoekman (1995) x Francois and Hoekman (1999)
Internal waterways transport services Air transport services Space transport services Rail transport services Road transport services Pipeline transport services Services auxiliary to all modes of transport Other transport services 12. Other services 1.Some studies only partially cover the services covered by the WTO Sectoral Classifications List. 2.Hardin and Holmes (1997) only measure restrictions on FDI. 3.Colecchia (2000) only covers accountancy services. 4. Nguyen-Hong (2000) covers accountancy, architectural, legal and engineering services. Source: Author.
Measures of restrictions on trade in services and multilateral trade negotiations Negotiation of the GATS was one of the major achievements of the Uruguay Round and established a structure and framework of rules for global trade in services and liberalisation commitments. However, little actual liberalisation was achieved in the Uruguay Round negotiations, with many countries’ commitments often representing the regulatory status quo or, in some cases, less than the status quo (OECD, 2001). A significant hindrance to advancing services trade liberalisation in the Uruguay Round is likely to have been a lack of information on actual restrictions, their effects and the benefits of their removal. Subsequent research on measures of restrictions and their effects contributes to providing the necessary analysis for negotiators to set priorities in areas likely to achieve the greatest benefits from liberalisation. Hoekman (1999) identifies the lack of information on restrictions as a major weakness of the GATS and suggests that the goal of the new round of negotiations should be to increase transparency. The information collected and compiled to date by different researchers provides a valuable resource on types of restrictions. Many studies compile information on restrictions beyond the GATS, thus moving from a positive list to a negative list of restrictions. Information on restrictions can also be used to develop formula, cluster or sectoral negotiating approaches. Measures on restrictions and their effects help in setting negotiating priorities for economies, services sectors and certain types of restrictions. The greatest benefits from liberalising services will come from reforming the economies and sectors with the highest levels of restrictions (Dee and Hanslow, 2000). The effect of restrictions in developing economies is significantly higher than in developed economies (Figure 2.3). In banking, the average price effect for all restrictions is 33% for developing economies and 7% for developed economies. In telecommunications, the average price effect for all restrictions is 38% for developing economies and 2% for developed economies (McGuire, 2002).
55
Figure 2.3. Price effects on banking services
1
Percentage 100 Foreign price effect 90
Domestic price effect
80 70
Score
60 50 40 30 20 10
HIEs Average
Venezuela
LMIEs Average
Economy
Uruguay
Turkey
Thailand
South Korea
South Africa
Philippines
Mexico
Malaysia
Indonesia
India
Colombia
Chile
Brazil
Argentina
0
1.The chart shows the effect of restrictions on the price or net interest margins of banking services. Source: McGuire and Schuele (2000).
Measures of restrictions and their effects generally show that banking and telecommunications are more restrictive than other sectors. Liberalising these sectors is likely to produce greater gains than in other sectors as they are essential inputs for other sectors and their liberalisation will produce substantial productivity gains for those sectors. In banking and telecommunications, the results show the precise effects of negotiating for the removal of discriminatory and non-discriminatory restrictions on establishment or restrictions on ongoing operations1 For example, Malaysia’s discriminatory and non-discriminatory restrictions (that is, the total foreign price effect) are estimated to raise the price of banking services (or net interest margins) by 60% over what they would be in the absence of these restrictions (Kalirajan et al., 2000). Indonesia’s non-discriminatory restrictions on establishment (that is, the domestic price effect for establishment) is estimated to raise the price of telecommunications services by 30% (Warren, 2000b). The measures can also show the effect of removing specific restrictions, such as restrictions on FDI. For example, the Philippines’ restriction on FDI (that is, the foreign price effect for direct investment) is estimated to raise the price of banking services by 11% (Kalirajan et al., 2000). The sequencing of reform is not specifically discussed, but Dee et al. (2000) suggest that the best strategy for liberalisation may be to negotiate gradual reductions in all types of restrictions simultaneously. Non-discriminatory restrictions on all services suppliers should be reduced or eliminated before removing discriminatory restrictions on foreign services suppliers. Reducing nondiscriminatory restrictions on all services suppliers together is a better approach than reducing discriminatory restrictions on foreign services suppliers alone. Reducing discriminatory restrictions on foreigners alone can have a negative impact on the level of services supplied by domestic firms. This will result in lower prices and higher total sales, but domestic services suppliers will end up with a smaller share of the services sector. However, if restrictions that affect foreign and domestic services suppliers equally are reduced, all services suppliers will have the same opportunities to increase the amount of services they supply in an expanding market. 56
Conclusion Methodologies for measuring the effects of restrictions on trade in services are in their infancy, but considerable methodological progress has been made in a reasonably short period of time. Measures of restrictions are useful tools for governments, trade negotiators, policy makers and consumers. They show the cost of maintaining restrictions on the largest sector of an economy and provide empirical estimates of the benefits of liberalisation in services sectors. There are a number of methodologies for measuring the effects of restrictions. Some desirable characteristics of a methodology for measuring restrictions are the collection of comprehensive information on restrictions, the conversion of this qualitative information into internationally comparable quantitative results and econometrically estimating the effect of restrictions on price-cost margins. This directly links restrictions to their effects on price-cost margins and can determine whether restrictions increase prices, increase costs or a combination of the two. Work can be done to sharpen the measures and improve coverage. The major difficulty in estimating the effect of restrictions on trade in services is determining the “optimal level” of regulation against which to measure the amount of restrictive regulation. This requires choosing “appropriate” benchmarks or establishing criteria for the attributes of “best practice” regulation. However, it is difficult to convert these qualitative criteria into empirical benchmarks. Research also needs to continue on analysing the precise effects of restrictions on price-cost margins and possible further disaggregation of estimates by modes of supply. Coverage of sectors and economies is better for developed than for developing economies. Improving coverage depends heavily on improving data on restrictions, which is extremely timeconsuming. A number of international organisations are collecting such data, and co-ordinating these efforts is likely to have large payoffs for trade practitioners. Extending the coverage to developing economies is particularly important. They have the most to gain from liberalisation of services trade, and reliable measures of the effect of restrictions can help them to recognise the benefits and understand the structural changes that are likely to take place if they liberalise their economies. These estimates are valuable tools for negotiators. They can be used to set negotiating priorities in certain sectors and for certain types of restrictions. Negotiators can demonstrate to their trading partners the benefits of liberalisation in terms of reductions in prices and costs while also understanding the benefits that will accrue to their own economies. A better understanding of these estimated benefits will help overcome the forces of protection and advance trade liberalisation in the new round of services negotiations.
57
NOTES 1.
Boylaud and Nicoletti (2000), Gonenc and Nicoletti. (2000) and Steiner (2000) use a similar methodology which predominantly covers the effects of domestic regulation on prices for OECD economies. Doove et al. (2001) extended this coverage to non-OECD economies.
2.
The sequence of removing certain restrictions has certain implications. Dee and Hanslow (2000) note that some approaches to partial liberalisation can worsen disparities in protection and real income by moving further away from a world free of distortions.
58
REFERENCES
Australian Productivity Commission (2001), “Measures of Restrictions on Trade in Services Database”. Available at: www.pc.gov.au/research/memoranda/servicesrestriction/index.html Australian Productivity Commission and Australian National University (2000), Achieving Better Regulation of Services, Conference Proceedings, AusInfo, Canberra. Benjamin, N. and X. Diao (1998), “Liberalising Services Trade in APEC: A General Equilibrium Analysis with Imperfect Competition”, The Economic Implications of Liberalising APEC Tariff and Non-Tariff Barriers to Trade, United States International Trade Commission. Benjamin, N. and X. Diao (2000), “Liberalising Services Trade in APEC: A General Equilibrium Analysis with Imperfect Competition”, Pacific Economic Review, Vol. 5:1, pp. 49-75. Boylaud, O. and G. Nicoletti (2000), “Regulation, Market Structure and Performance in Telecommunications”, Economics Department Working Paper No. 237, OECD, Paris. Brown, D., A. Deardorff and R. Stern (1996), “Modelling Multilateral Liberalisation in Services”, Asia-Pacific Economic Review, Vol. 2, pp. 21-34. Chadha, R. (2000), “GATS and the Developing Countries: A Case Study of India”, in Robert M. Stern (ed.), Services in the International Economy: Measurement and Modelling, Sectoral and Country Studies, and Issues in the WTO Services Negotiations, University of Michigan Press, Ann Arbor, Michigan. Chadha, R., D. Brown, A. Deardorff and A. Stern (2000), “Computational Analysis of the Impact on India of the Uruguay Round and the Forthcoming WTO Trade Negotiations”, Discussion Paper No. 459, School of Public Policy, University of Michigan. Claessens, S. and t. Glaessner (1998), “Internationalisation of Financial Services in Asia”, World Bank Discussion Paper, World Bank, Washington, DC. Coghlan, P. (2000), “The Principles of Good Regulation”, in Productivity Commission and Australian National University, Achieving Better Regulation of Services, Conference Proceedings, AusInfo, Canberra. Colecchia, A. (2000), “Measuring Barriers to Market Access for Services: A Pilot Study on Accountancy Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York. Dee, P. (2001), “Trade in Services”, paper prepared for the conference, Impacts of Trade Liberalisation Agreements on Latin America and the Caribbean, Inter-American Development Bank, Washington, DC, 5-6 November.
59
Dee, P. and K. Hanslow (2000), “Multilateral Liberalisation of Services Trade”, Productivity Commission Staff Research Paper, Ausinfo, Canberra. www.pc.gov.au/research/staffres/multilatlib/index.html (accessed 15 October 2001). DFAT (Department of Foreign Affairs and Trade) (1999), Global Trade Reform: Maintaining Momentum, DFAT, Canberra. Doove, S., O, Gabbitas, D. Nguyen-Hong and J. Owen (2001), “Price Effects of Regulation: Telecommunications, Air Passenger Transport and Electricity Supply”, Productivity Commission Staff Research Paper, AusInfo, Canberra. Disclosure (2000), Global Researcher – Worldscope Database, Disclosure, United States. Findlay, C. and T. Warren (eds.) (2000), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York. Fischer, S. (First Deputy Managing Director of the International Monetary Fund) (1998), “The Asian Crisis: A View from the IMF”, address to the Midwinter Conference of the Bankers’ Association for Foreign Trade, Washington, DC, 22 January. Francois, J. and B. Hoekman (1999), “Market Access in the Services Sectors”, unpublished paper. Gonenc, R. and G. Nicoletti (2000), “Regulation, Market Structure and Performance in Air Passenger Transportation”, Economics Department Working Papers No. 254, OECD, Paris. Hardin, A. and L. Holmes (1997), “Services Trade and Foreign Direct Investment”, Industry Commission Staff Research Paper, AGPS, Canberra. Hoekman, B. (1995), “Assessing the General Agreement on Trade in Services”, in W. Martin and L. Winters (eds.), The Uruguay Round and the Developing Economies, pp. 327-364, World Bank, Washington, DC. Hoekman, B. (1999), “The Next Round of Services Negotiations: Identifying Priorities and Options”, paper presented at the Federal Reserve Bank of St. Louis conference on Multilateral Trade Negotiations: Issues for the Millennium Round, 21-22 October (revised 4 November 1999). Hertel, T., J. Francois and W. Martin (1999), “Agriculture and Non-agricultural Liberalisation in the Millennium Round”, paper presented at the Global Conference on Agriculture and the New Trade Agenda from a Development Perspective – Interests and Options in the WTO 2000 Negotiations, sponsored by the World Bank and WTO, Geneva, 1-2 October. Kalirajan, K. (2000), “Restrictions on Trade in Distribution Services”, Productivity Commission Staff Research Paper, Ausinfo, Canberra. Available at: www.pc.gov.au/research/staffres/rotids/index.html (accessed 15 October 2001). Kalirajan, K. G. McGuire, D. Nguyen-Hong and M. Schuele (2000), “The Price Impact of Restrictions on Banking Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York.
60
Kemp, S. (2000), “Trade in Education Services and the Impact of Barriers to Trade”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York. Marko, M. (1998), “An Evaluation of the Basic Telecommunications Services Agreement”, CIES Policy Discussion Paper 98/09, Centre for International Economic Studies, University of Adelaide. Mattoo, A. (1998), “Financial Services and the World Trade Organization – Liberalization Commitments of the Developing and Transition Economies”, World Bank Discussion Paper, World Bank, Washington, DC. Mattoo, A., R. Rathindran and A. Subramanian (2001), “Measuring Services Trade Liberalization and Its Impact on Economic Growth: An Illustration”, World Bank Research Paper, World Bank, Washington, DC. McGuire, G. (1998), “Australia’s Restrictions on Trade in Financial Services”, Productivity Commission Staff Research Paper, AusInfo, Canberra. McGuire, G. (2000), “Measuring and Modelling Restrictions on Trade in Services”, note for the OECD Trade Committee Working Party meeting, OECD, Paris, 18-19 September. McGuire, G. (2002), “How Important are Restrictions on Trade in Services?”, paper presented at the UNCTAD Workshop on Market Access, New York, 8-9 January. McGuire, G. and M. Schuele (1999), “Restrictions on Trade in Financial Services for APEC Member Economies”, paper presented at the APEC Business Advisory Council Meeting, Tokyo, 21-23 May. McGuire, G. and M. Schuele (2000), “Restrictiveness of International Trade in Banking Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York. McGuire, G., M. Schuele, and T. Smith, T. (2000), “Restrictiveness of International Trade in Maritime Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York. Nguyen-Hong, D. (2000), “Restrictions on Trade in Professional Services”, Productivity Commission Staff Research Paper, Ausinfo, Canberra. Available at: www.pc.gov.au/research/staffres/rotips/index.html (accessed 15 October 2001). OECD (2001), Trade in Services – Negotiating Issues and Approaches, OECD, Paris. Robinson, S., Z. Wang and W. Martin (1999), “Capturing the Implications of Services Trade Liberalisation”, paper presented at the Second Annual Conference on Global Economic Analysis, Ebberuk, Denmark, 20-22 June. Steiner, F. (2000), “Regulation, Industry Structure and Performance in the Electricity Supply Industry” Economics Department Working Paper No. 238, OECD, Paris.
61
Warren, T. (2000a), “The Identification of Impediments to Trade and Investment in Telecommunications Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York. Warren, T. (2000b), “The Impact on Output of Impediments to Trade and Investment in Telecommunications Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York. World Bank (2001), World Development Indicators Database on CD-ROM, World Bank, Washington, DC. World Trade Organization (1991), Services Sectoral Classification List, note by the Secretariat, MTN.GNS/W/120, WTO, Geneva, 10 July. World Trade Organization (2000), Market Access: Unfinished Business – Post Uruguay Round Inventory and Issues, WTO, Geneva. World Trade Organization (2001), Annual Report 2001, WTO, Geneva.
62
Chapter 3 QUANTIFYING THE EFFECTS OF LIBERALISING SERVICES: THE EXPERIENCE OF THE AUSTRALIAN PRODUCTIVITY COMMISSION
by Patrick Jomini, George Verikios and Xiao-Guang Zhang Australian Productivity Commission
Abstract. This chapter summarises the Australian Productivity Commission’s experience in estimating the effects of aggregate services liberalisation using a modelling framework which explicitly analyses trade in services in the form of cross-border supply and commercial presence. The paper singles out telecommunication services to highlight modelling and data issues in disaggregated analyses. It presents a detailed discussion of results and their decomposition into various contributing factors, highlighting the distinction between the sources of gains from a conventional goods-trade model and a services-trade model. The paper concludes that analysing services trade liberalisation requires progress in areas such as data collection, development of the modelling framework at a modal and sectoral level, and development of methods to incorporate services barriers.
Introduction The Australian Productivity Commission has analysed the effects of liberalisation using two global models of world trade that account explicitly for the activities of foreign affiliates and for crossborder trade. The models account for barriers to trade in the form of market access barriers and derogations from national treatment. The experience gathered in the modelling process has led researchers at the Commission to identify future modelling and data needs to improve the estimation of the effects of liberalisation. Liberalising services has potentially large impacts on global, regional and sectoral incomes and production patterns. Simulation results indicate that economies that liberalise their services sectors may benefit through reductions in the costs of providing the services to consumers and industries, as new entrants and foreign affiliates increase their share in a services sector. The services sector’s share of economic activity has grown considerably over recent decades. Today it accounts for a majority of gross product and employment in developed economies (around 80% in Australia). In 2000, world exports of services increased by 6% to USD 1 435 billion (WTO, 2001). Since the General Agreement on Trade in Services (GATS) in 1995, restrictions on multilateral trade have been reduced in a few services sectors. In most countries, however, barriers to trade in services are still significant, suggesting that potentially large gains may be expected from further liberalisation. Services trade negotiations are part of the built-in agenda of WTO negotiations, i.e. they do not depend on the launching of a full round of negotiations. In 1999, a report by the WTO Council for Trade in Services noted that there had been no comprehensive empirical study of the effects of multilateral liberalisation of services trade (Secretariat of the Council for Trade in Services, 1999). Since then, some research on this issue (e.g. Hertel, 1999; Markusen et al., 1999; McKibbin, 1999) has led to estimates of the effects of some aspects of liberalising trade in services. Quantifying the effects of liberalising trade in services requires two steps:
x
Estimating the effects of trade barriers or restrictions on the relevant economic variables (price, cost, price-cost margin or output quantity) of services firms. Recent theoretical developments have led to a number of estimates for various sectors. Further developments in theory and econometric techniques will lead to improved estimates.
x
Inserting these estimates into a computable general equilibrium (CGE) framework to estimate the impacts of liberalisation on individual economies and globally.
The views expressed in this chapter are those of the authors and do not necessarily reflect those of the Productivity Commission.
64
Incorporating services trade barriers in existing models of world trade requires developments in data and model theory. These requirements include: An accurate representation of the economics of services sectors and trade, including the various forms of delivery. Consistent estimates of foreign direct investment (FDI) to represent commercial presence and cross-border trade at the bilateral and sectoral levels, two of the modes of international services delivery. In a collaborative project, researchers from the Australian Productivity Commission and the Australian National University estimated trade barriers in the form of impacts on the prices of several services, including banking services (Kalirajan et al., 2000), maritime services (McGuire and Schuele, 2000), professional services (Nguyen-Hong, 2000), distribution services (Kalirajan 2000), and, most recently, air transport, telecommunications and electricity (Doove et al., 2001). The Productivity Commission has also built the FTAP model to analyse the sectoral, regional and global impacts of liberalising services. The FTAP model accounts explicitly for two of the four modes of service delivery recognised in the GATS agreement: commercial presence and cross-border supply.1 Modelling commercial presence requires incorporating FDI into the framework of analysis. The FTAP model was initially developed as a three-sector (primary, manufacturing, services), 19-economy model.2 Dee and Hanslow (2000) used the FTAP model to analyse the global and distributional impacts of liberalising trade in all services. Verikios and Zhang (2001a) modified the original model to address modelling issues that arise from more detailed sectoral analysis. In this analysis, the authors used the FTAP2 model to quantify the effects of liberalising telecommunications and financial services This analysis is based on an eightsector database which identifies separately the two main sectors of interest.3 Australian Productivity Commission’s experience in estimating the effects of liberalising services. The first part concentrates on the main characteristics of the FTAP model and the results of the aggregate analysis presented in Dee and Hanslow (2000). The second part concentrates on the disaggregated analysis presented in Verikios and Zhang (2001a), using the FTAP2 model. Telecommunication services are singled out to highlight modelling and data issues in disaggregated analysis. Both analyses are used to illustrate the types of mechanisms at work and the results obtained. In both analyses, barrier estimates are assumed to exist in a form that allows them to be incorporated in these types of analyses.4 In the final section, directions for future research are identified. Liberalising services Dee and Hanslow (2000) use a three-sector, 19-region CGE5 model (FTAP), with international trade and investment flows, to quantify the effects of removing barriers to trade in aggregate services. This initial work provides preliminary estimates of the benefits to individual economies, and to the world as a whole, from eliminating the barriers to services trade that remain after full implementation of the Uruguay Round. To do so, it uses a model that accounts explicitly for the activities of offshore affiliates through treatment of FDI and cross-border trade, two of the vehicles by which services are delivered internationally. The study finds that complete liberalisation of trade in services leads to annual gains of the order of USD 133 billion or 0.45% of annual world gross national product (GNP). The gains vary among economies: the largest gains are projected for the economies with the highest barriers to trade in services, while economies with low barriers are projected to experience smaller gains or, in some 65
cases, losses. The following sections detail the model structure, database and results of Dee and Hanslow (2000). The FTAP model FTAP is developed from GTAP (Hertel, 1997) by adding bilateral stocks of FDI to the database and modelling the behaviour of firms investing abroad. The treatment of FDI follows earlier work by Petri (1997). FTAP also incorporates increasing returns to scale and large-group monopolistic competition in all sectors. This follows Francois et al. (1995), among others, who adopted this treatment for manufacturing and resource sectors; Brown et al. (1995); and Markusen et al. (1999), who used similar treatments for services. Finally, FTAP makes provision for capital accumulation and international borrowing and lending, using a treatment of international (portfolio) capital mobility developed by McDougall (1993) and recently incorporated into GTAP by Verikios and Hanslow (1999). FTAP is implemented using the GEMPACK software suite (Harrison and Pearson, 1996). Its structure is documented fully in Hanslow et al. (1999). Theoretical structure FTAP takes the standard GTAP framework as a description of the location of economic activity, and disaggregates by ownership. For example, each industry located in Australia includes Australianowned firms, along with affiliates of US, European and Japanese multinationals. Each of these firm types is modelled as making an independent choice of inputs to production, according to standard GTAP theory. Each firm type has its own sales structure. On the purchasing side, agents in each economy make choices among the products of each firm type, distinguished by both ownership and location of production, and then among the individual firms of a given type. Firms purchase intermediate inputs and investment goods, while households and governments choose among final goods and services. Agents are assumed to choose first among products from domestic or foreign locations. They then choose among products from particular foreign locations and among firm types in a particular location. Finally, they choose among products from the individual firms of a particular ownership and location. With firm-level product differentiation, agents benefit from having more firms to choose among, because it is more likely that they can find a product suited to their particular needs. The first two choices, among domestic and foreign locations, are identical to the choices in the original GTAP model. Supply of capital The supply of FDI is determined by the same imperfect transformation among types of wealth as in Petri (1997). Investors in each economy first divide their wealth between “bonds” (which can be thought of as any instrument of portfolio investment), real physical capital, and land and natural resources in their country of residence. A bond is a bond, irrespective of who issues it, implying perfect international arbitrage of rates of return on bonds. However, physical capital in different locations is seen as different things. Investors next choose the sector in which they invest. They then choose whether to invest at home or overseas in their chosen sector. Finally, they choose among different overseas regions in which to invest. Petri’s model assumed that total wealth in each region was fixed. In FTAP, while regional stocks of land and natural resources are fixed (and held solely by each region’s residents), regional capital 66
stocks can accumulate over time, and net bond holdings of each region can adjust to help finance the accumulation of domestic and foreign capital by each region’s investors. The treatment of capital accumulation follows McDougall (1993). With this treatment of capital accumulation, FTAP provides a long-run snapshot view of the impact of trade liberalisation, ten years after it has occurred. To the extent that liberalisation leads to changes in regional incomes and savings, this will be reflected in changes to the capital stocks that investors in each region will have been able to accumulate. To finance capital investment, investors in each region are not restricted to their own savings pool. They may also issue bonds to help with that investment, but only according to their own preferences about capital versus bond holding, and only according to the willingness of others to accept the additional bonds. Model database The starting point for FTAP’s database is an updated version of the GTAP database, following a simulation in which barriers yet to be eliminated under the Uruguay Round are removed (see Verikios and Hanslow, 1999). FDI stocks and returns The treatment of FDI requires the addition of data on bilateral stocks of FDI, and on activity levels and cost and sales structures of FDI firms. The methods used to estimate the stocks are similar to those of Petri. APEC (1995) and United Nations (1994) provided limited data on FDI stocks by source, destination and sector.6 The database reflects stylised facts about FDI. For example, Europe and the United States are the main sources of and destinations for FDI. Japan is much more important as a source than as a destination. The OECD area provides 87% of outward FDI and receives 73% of inward FDI. The detailed data show that 80% of FDI from Asia (excluding Japan) remains in Asia, and that there are strong bilateral ties between neighbouring countries (Australia-New Zealand, United States-Canada). Finally, about 20% of FDI is in the primary sector, and about 40% each in the secondary and tertiary sectors. The FDI stock data were used to generate estimates of the output levels of FDI firms. Capital income flows (that is, returns to capital) were estimated by multiplying the FDI stocks by rates of return obtained from accounting information in the Worldscope Global Equity Database (Welsh and Strzelecki, 2000; Disclosure, 1999). The detailed cost and sales structures of FDI firms were assumed to be the same as for locally owned firms, and were obtained by pro-rating the GTAP database. Barrier estimates Barriers to services trade discriminate not only against foreign firms, they may also affect domestic firms. The FTAP model distinguishes (Table 3.1): x
Barriers to market access which do not discriminate between the activities of domestic and foreign entrants, and derogations from national treatment which typically treat foreign firms less leniently than domestic firms.
x
Barriers to establishment from barriers to ongoing operation. 67
Table 3.1. Barriers to services trade
Restrictions on market access
Restrictions on national treatment
Barriers to establishment
Tax on capital applied to domestics and affiliates
Tax on capital applied to foreign affiliates only
Barriers to ongoing operations
Tax on output applied to domestics and affiliates
Tax on output applied to foreign affiliates only
Source: Based on Hanslow et al. (1999).
The GTAP model already contains estimates of barriers to trade in agricultural and manufactured goods, and the updated version of this database (Verikios and Hanslow, 1999) has these at their postUruguay levels. However, GTAP does not contain estimates of barriers to services trade. Estimates of barriers to trade in banking services were taken from Kalirajan et al. (2000), and estimates of barriers to trade in telecommunications services were taken from Warren (2000a).7 The rates were taken as indicative of post-Uruguay Round rates since, while the Uruguay Round established the architecture for services trade negotiations, it did not achieve much in the way of services trade liberalisation (Hoekman, 1995). A simple average of the estimates for banking and telecommunications was taken to reflect the regional pattern of barriers. The resulting structure of post-Uruguay Round barriers to trade in services is summarised in Table 3.2. Barriers to establishment are modelled as taxes on capital. Barriers to ongoing operation are modelled as taxes on the output of locally based firms (either domestic or foreign owned), and barriers to cross-border trade imposed by importing regions are modelled as taxes on exports.8 The taxes in columns [2] and [3] in Table 3.2 represent the effects of restrictions on market access (affecting the establishment and ongoing operations, respectively, of domestic firms and foreign affiliates in a region). The taxes on the capital and output of foreign affiliates in columns [4] and [5] are higher than the corresponding taxes on domestic firms, because they account for the effects of market access restrictions and derogations from national treatment. Many economies have significant barriers to trade in services. The only economies where barriers are estimated to be low are New Zealand, Hong Kong (China), Canada, the United States and the European Union.9 Barriers to trade in services have been modelled as tax equivalents that generate rents – a markup of price over cost – rather than as factors that raise costs above what they might otherwise have been (as in Hertel 1999, for example). This decision was based on the way in which the price impacts of barriers to trade in banking and telecommunications services were measured. Kalirajan et al. (2000) measured the effects of trade restrictions on the net interest margins of banks, a direct measure of banks’ mark-up of price over cost. Warren (2000a) measured the effects of trade restrictions on the quantities of telecommunications services delivered, and these were converted to price impacts using an estimate of the elasticity of demand for telecommunications services. Thus, Warren’s estimates did not provide direct evidence of a mark-up of price over cost, but the relative profitability of telecommunications companies in many countries suggests that some element of rent may exist. By contrast, there is evidence that trade restrictions in sectors such as aviation raise costs (Johnson et al., 2000).
68
Table 3.2. Tax equivalents of post-Uruguay Round barriers to services trade Percentage Region
Australia New Zealand Japan Korea Indonesia Malaysia Philippines Singapore Thailand China Hong Kong (China) Chinese Taipei Canada United States Mexico Chile Rest of Cairns4 European Union Rest of the world
Exports1
Market access2
National treatment3
[1]
Output [2]
Capital [3]
Output [4]
Capital [5]
4.8 3.8 4.4 4.6 4.7 4.5 4.8 4.7 4.1 4.1 9.9 4.4 3.5 4.3 5.2 4.4 4.5 4.7 4.9
0.0 0.0 3.6 5.1 13.2 3.6 8.4 3.4 4.7 18.8 1.4 2.9 0.3 0.1 2.2 3.0 1.0 0.10 4.9
0.6 0.4 0.3 1.9 22.7 15.4 7.4 2.4 12.2 123.5 1.4 1.9 0.5 0.0 0.7 14.2 7.2 1.3 39.1
0.7 0.7 4.8 6.8 28.1 10.2 22.7 8.3 13.4 26.4 2.4 4.9 1.7 1.1 5.6 4.1 5.6 1.3 13.9
14.8 4.2 3.0 22.0 68.1 37.6 54.3 24.5 36.5 250.7 5.4 19.2 6.1 3.8 13.0 20.4 19.5 6.5 87.0
1. Cross-border trade. 2. Applies to domestic and foreign firms (non-discriminatory). 3. Applies only to foreign firms (discriminatory). 4. Argentina, Brazil, Colombia and Uruguay. Source: Adapted from Warren (2000b).
One important implication of this treatment of services trade barriers is that gains from liberalising trade in services are likely to be understated, perhaps significantly. If trade restrictions create rents, then the allocative efficiency gains from trade liberalisation are the “triangle” gains associated with putting resources to more efficient use. By contrast, if trade restrictions raise costs, then the gains from trade liberalisation include “rectangle” gains from lower costs, which are equivalent to a more effective use of reallocated resources. Because barriers to services trade appear to be significant, and because they have been modelled as taxes, the rents they generate will be significant. A key issue is whether these rents should be modelled as being retained by incumbent firms, appropriated by governments via taxation, or passed from one country to another by transfer pricing or other mechanisms. In FTAP, the rents on the output and capital of affiliates are assumed to accrue to the region of ownership (home region), once the government in the region in which the products were sold (host region) has taxed them at its general property income tax rate. Thus, liberalisation of services trade could have significant income effects in both home and host regions as the rents are eliminated. The effects of eliminating post-Uruguay Round barriers The results of eliminating the post-Uruguay Round barriers to services trade are comparatively static, showing only the impact of trade liberalisation. During the ten-year adjustment period, many other changes will affect each economy, but they are not taken into account in this analysis. For this 69
reason, the results should not be interpreted as indicating the likely changes that would occur over time in each economy; such results would require taking into account all changes, not just changes in trade barriers. The model results should instead be seen as providing an indication, at some point in time ten years after liberalisation, of how different each economy would be, compared with the alternative situation at the same point in time, had the liberalisation not taken place. Table 3.3. Projected effects on real GDP and income of eliminating services trade barriers Real output GDP
Real income Net national product
%1 Australia New Zealand Japan Korea Indonesia Malaysia Philippines Singapore Thailand China Hong Kong (China) Chinese Taipei Canada United States Mexico Chile Rest of Cairns2 European Union Rest of the world World
%1
0.0 -0.1 0.0 0.1 5.1 0.7 0.4 -1.3 0.2 14.6 1.0 0.2 -0.1 -0.1 0.1 0.4 0.1 0.0 0.8
0.4 0.5 0.2 0.3 1.2 1.5 1.0 -0.4 0.7 14.0 4.5 0.4 0.1 -0.1 0.3 0.3 0.5 -0.1 0.9 0.45
USD millions 2 098 257 4 130 1 886 2 470 1 015 1 236 -247 1 698 90 869 5 896 -142 -499 -1 809 357 330 6 970 -6 169 23 039 133 386
1. Percentage changes from base. 2. Argentina, Brazil, Colombia and Uruguay. Source: FTAP2 model projections.
Table 3.3 shows the effects of eliminating services trade barriers at regional activity levels (as measured by changes in real GDP) and on economic well-being (as measured by the real income accruing to the residents in each economy). The measure of well-being is net national product (NNP), the income accruing to the residents of an economy. This income is equivalent to net domestic product adjusted for the income earned on outward FDI, net of the income repatriated overseas from inward FDI, plus the income from net bond holdings. Real income is projected to increase by the order of USD 130 billion a year worldwide. The majority of this increase accrues to economies with large barriers. Australia is projected to gain about USD 2 billion from global liberalisation of services trade. This is the projected gain in annual income, about ten years after liberalisation has occurred and the associated resource adjustments have taken place. For some economies – the European Union, the United States, Canada, Singapore and Chinese Taipei – the impact of multilateral services trade liberalisation is projected to be negative. For the European Union, the projected loss is USD 6 billion and for the United States it is almost USD 2 billion.
70
Explaining results The changes in net national product can be broken down into a number of influences. Two effects typical of trade liberalisation are the contributions from: x
Changes in allocative efficiency within an economy as it specialises in its industries of comparative advantage. This has a positive impact on income.
x
Changes in terms of trade as the costs of production in an economy fall relative to costs in other economies. The revenue it raises from its exports falls relative to the cost of its imports and may reduce the economy’s income.
Changes in FDI patterns contribute two additional effects. First, FDI can add to (or reduce) the capital stock located in an economy, leading to a positive (or negative) contribution to income through a change in productive capacity. Second, changes in rents that accrue from barriers on affiliate activities can affect national incomes. Column [1] in Table3.4 shows the contribution to changes in real income of changes in an economy’s capital stock. GDP increases when an economy’s capital stock is increased and when its capital is reallocated to sectors that use it more efficiently. Some changes in capital stocks come from FDI, and some come from investment by domestic residents. Column [2] in Table 3.4 shows the contribution of changes in FDI stocks. Column [3] shows the contribution to changes in income of changes in real bond holdings. Columns [2] and [3] indicate how changes in capital stocks are financed. For example, Japan’s capital stock shrinks, partly because of a big increase in outward FDI. It also borrows (a negative contribution from bond holding) to finance this outward FDI. By contrast, China’s increase in capital stock comes from a large increase in inward FDI and from additional borrowing. The United States is projected to have a smaller capital stock than otherwise (a negative contribution), but increases its outward FDI and its lending to other economies. The last two columns of Table 3.4 show the contributions of changes in the rents from barriers to services trade to recipient countries as these barriers are eliminated. What is striking is the loss of rents to the main providers of FDI – Japan, Hong Kong (China), the United States and the European Union. In fact, the loss of rents to the United States is more than sufficient to explain its projected loss of real income in Table 3.3, and the loss of rents in the European Union explains most of its projected loss.
71
Table 3.4. Contributions of liberalising services to changes in real income USD millions
Australia New Zealand Japan Korea Indonesia Malaysia Philippines Singapore Thailand China Hong Kong (China) Chinese Taipei Canada United States Mexico Chile Rest of Cairns EU Rest of the world
Change in capital stocks
Change in FDI stocks
[1]
[2]
Change in real bond Change in rents holding from barriers to establishment [3]
Change in rents from barriers to ongoing operations [5]
[4]
58 -43 -1 030 438 7 158 367 164 -1 071 305 52 164 102
0 5 3 120 -5 -541 -103 -91 -198 -24 -12 649 7 829
4 52 -2 978 39 -4 519 -168 47 -108 -393 -5 776 -621
534 -10 -3 629 51 162 253 70 401 227 4 163 -2 638
-39 6 -5 101 72 368 332 144 1 049 259 8 686 -5 573
312 -747 -5 713 131 202 401 -3 672 15 002
378 34 2 665 -67 -39 -137 1 441 -2 337
-583 1 086 1 708 332 -54 1 800 6 327 3 285
-137 27 -3 057 247 101 450 -2 265 5 427
-286 -52 -3 659 266 56 486 -3 110 6 581
1. Selected contributions only; excludes for example allocative and terms of trade effects; therefore the rows do not add up to the totals reported in the Table 3.3. Source: FTAP2 model projections.
Given the uncertainty about the allocation of existing rents, it is not at all clear that the true impact on the United States and European Union would be as shown in Table 3.4. If barriers to services trade in these economies are understated, then their gains in allocative efficiency will be as well. Thus, the projected net income losses from services trade liberalisation shown in Table 3.3 should be heavily qualified. Similarly, Canada’s overall income loss, which is due primarily to adverse terms of trade effects, should also be qualified, given uncertainty about many key features of the model. Liberalising telecommunications The inclusion of a single services sector in Dee and Hanslow (2000) confines that analysis largely to the global and regional effects of liberalising services. A disaggregated approach provides an understanding of the effects of liberalising specific services industries. This is the approach taken by Verikios and Zhang (2001a) who use the FTAP2 model to analyse the impacts of removing discriminatory and non-discriminatory barriers separately and in combination.10 72
The FTAP2 model The FTAP2 model is a modified version of the original FTAP model. Individual services sectors have unique features that must be reflected in the structure of the model used to analyse them. Three main features of FTAP2 distinguish it from FTAP and are described below. The treatment of wealth accumulation FTAP2 is a comparative static model with no capital or wealth accumulation effects. Although FTAP is also a comparative static model, it has some quasi-dynamic features to capture wealth accumulation effects. Experience with FTAP in a disaggregated setting shows that more work is required to model the interactions between trade, portfolio and FDI flows. Removing the wealth accumulation mechanisms allows for greater concentration on the allocative effects of sectoral liberalisation.11 As a result, estimates using FTAP2 can be considered as a lower bound for the effects of liberalisation. The treatment of commercial presence and international capital mobility FTAP2 assumes that capital is region-specific and perfectly mobile between an economy’s national firms and their affiliates overseas. This implies that the rates of return on capital used by an economy’s own firms and their affiliates will be equalised in equilibrium. This is different from the original FTAP model which assumes firm-specific capital (consistent with the treatment of FDI in Markusen et al., 1999), and no equalisation of rates of return, even between a parent firm and its overseas affiliates. By removing this setting and assuming arbitrage of rates of return on regional capital, FTAP2 allows for a greater degree of capital mobility than FTAP.12 The treatment of cross-border supply of telecommunications FTAP2 recognises a fundamental attribute of cross-border supply of telecommunications which differentiates it from other goods and services. Telecommunications supplied across borders is a type of service very different from domestically supplied telecommunications. For example, international phone calls are not the same product as domestic phone calls.13 Moreover, cross-border telecommunications, as recorded in balance of payment statistics (and the model database), represent only part of international telephone connection services purchased by final users. An international telephone call needs a domestic network to connect to the end user. Therefore, domestic services suppliers are also involved in providing cross-border delivery. As a result, if the demand for crossborder supply of telecommunications increases, the demand for domestic telecommunications will increase as well. To capture these features of cross-border telecommunications delivery, FTAP2 assumes that cross-border supplied and domestic telecommunications are not substitutable in demand. This implies that suppliers of domestic telecommunication services do not compete with cross-border trade suppliers.
73
The effects of liberalising telecommunications Table 3.5 presents the tax equivalents of post-Uruguay Round barriers to telecommunication services. The pattern of barriers is similar to that in Table 3.2, with relatively high barriers in developing economies. As evidenced by the relative closeness of market access and national treatment barriers, market access barriers prevail in telecommunication markets.14 Because cross-border trade is assumed not to be substitutable for domestic services in FTAP2, no export taxes are implemented.15 Table 3.5. Tax equivalents of post-Uruguay Round barriers in telecommunication services Percentages
Market access1
Australia New Zealand Japan Korea Indonesia Malaysia Philippines Singapore Thailand China Hong Kong (China) Chinese Taipei Canada United States Mexico Chile Rest of Cairns3 European Union Rest of the world
National treatment2
Output
Capital
Output
Capital
-
1.4 1.3 1.8 4.8 85.9 3.6
-
-
-
-
2.4 41.1 5.5 32.6
-
2.5 41.0 5.5 21.4 0.8 14.0 50.0 0.6
5.5 32.7 339.4 3.2
14.0 50.0
4.4 77.4 11.4 50.0 2.6 23.1 339.4
-
-
5.8 1.8
0.8 1.5
3.5 4.8
-
-
-
5.6
6.0
-
-
-
-
1.3 0.8 0.2 5.4 1.3 2.7
1.8 1 11.8
3.8
4.1
0.3 16.4
3.6 114.0
0.4 16.6
1.0 101.0
1. Applies to domestic and foreign firms (non-discriminatory). 2. Applies only to foreign firms (discriminatory). 3. Argentina, Brazil, Colombia and Uruguay. -= Nil. Source: Adapted from Warren (2000b).
Table 3.6 shows the effects of liberalising telecommunications in terms of changes in real GNP, for each economy and for the world as a whole. Global gains from liberalising telecommunication services are estimated of the order of 0.1% of GNP, or about USD 24 billion. In telecommunication markets, global gains from liberalisation come mainly from removing market access barriers.
74
Results for each economy are influenced by the initial size of the barriers and the initial market share of foreign affiliates in each economy’s telecommunications sector. Economies with no barriers cannot benefit from reducing their barriers; however, they may benefit from reductions in barriers in other economies, which increase their access to these markets. Low penetration by foreign affiliates is associated with the biggest discriminatory barriers. Removing these barriers does not lead to large gains for the world as a whole.16 Table 3.6. Projected effects on real GNP of liberalising telecommunication markets Percentage change
Australia New Zealand Japan Korea Indonesia Malaysia Philippines Singapore Thailand China Hong Kong (China) Chinese Taipei Canada United States Mexico Chile Rest of Cairns1 EU Rest of the world 2 World
National treatment 0.00 0.00 0.04 0.00 -0.77 0.00 0.22 0.04 -0.01 -0.39 0.64 -0.01 0.00 0.02 0.00 0.00 0.00 0.02 -0.01 0.01
Market access 0.06 0.09 0.04 0.01 0.36 -0.03 -0.01 0.00 -0.35 0.79 0.25 0.02 0.01 0.01 -0.07 0.01 0.01 0.05 0.38 0.09
Complete liberalisation 0.06 0.09 0.04 0.01 0.70 -0.03 0.72 0.02 -0.35 0.81 0.16 0.02 0.01 0.01 -0.06 0.01 0.02 0.05 0.39 0.10
1. Argentina, Brazil, Colombia and Uruguay. 2. Almost entirely composed of non-APEC developing countries. Source: FTAP2 model projections.
Eliminating derogations from national treatment gives investing economies better access to the domestic markets of other economies. Economies with high discriminatory barriers and large inflows of FDI may not benefit from liberalising access for foreigners. This is because, as foreign affiliates expand their market share at the expense of domestic firms, they repatriate the rents that arise from remaining market access barriers. Distribution of gains Table 3.6 shows that major investing economies, such as Japan, the United States and the European Union, benefit from national treatment liberalisation, while economies with high barriers, such as Indonesia and China, do not. Hong Kong (China) benefits because national treatment without additional access will substitute some foreign ownership for domestic ownership, without increasing total access to the market; as China removes its high barriers to market access, Hong Kong (China) increases its investment and related returns to capital. 75
Almost all economies gain from the removal of market access barriers. Economies benefit directly from removing these barriers through lower telecommunication prices for consumers and industries and improved competitiveness relative to other economies. For investing economies, further gains come from freer access to other economies’ domestic markets. In the Philippines and Thailand, remaining high national treatment barriers prevent foreign affiliates from significantly expanding their market share in the domestic telecommunications sector. Although they benefit from cost reductions that improve the competitiveness of their exports, these result in terms of trade losses that are not compensated by increased inward FDI. In the case of Malaysia and Mexico, low initial market access barriers mean that direct benefits from liberalisation are low relative to gains in other economies; they do not gain much in terms of improved competitiveness. Moreover, since these economies are not significant international investors, they do not increase their investments abroad enough to take advantage of liberalisation in other economies. Partial vs. complete liberalisation For economies with high barriers (usually developing economies), the gains from complete liberalisation tend to be greater than the sum of the gains from the two partial liberalisation scenarios. For economies with low barriers (usually developed economies), the gains from complete liberalisation tend to be smaller than the sum of the gains from the two partial liberalisation scenarios. This is because, compared to complete liberalisation, partial liberalisation leaves some barriers, and therefore some rents, in place, thus reducing gains for economies with high barriers and increasing gains for economies with low barriers. In complete liberalisation, these effects are absent. Complete liberalisation leads to gains for all economies except Malaysia, Thailand and Mexico. Table 3.6 shows that the distribution of these gains is close to the distribution of gains from market access liberalisation, confirming that market access liberalisation is the main source of gains from liberalising telecommunications. Conclusion The Australian Productivity Commission has modelled the benefits from liberalising services as a whole and telecommunications separately. This work has been complemented by developing measures of barriers to services trade. Both streams of work conclude that: x
Market access barriers and derogations from national treatment result in significant price increases for the affected services. Whether these price increases result in rents to the owners of factors involved in production or cost increases that result in sub-optimal resource use needs to be further investigated.
x
There are significant global benefits from multilateral reduction of barriers to trade in services. Different economies are at different stages in liberalising their services sectors. Simulation results indicate that economies with large barriers to trade stand to make significant gains by liberalising their services sectors.
Main sources of gains The main purpose of the type of modelling initiated with FTAP and FTAP2 is to provide insights into the mechanism by which liberalisation can improve measures of well-being. These models were 76
developed from a conventional global trade framework originally developed to analyse goods trade. There is an important distinction between the sources of gains from a conventional goods-trade model and a services-trade model. In a conventional goods-trade model, as imported goods are substitutable with domestically produced goods, trade liberalisation forces domestic firms to shift production to those goods in which they have a comparative cost advantage. The economy as a whole benefits from a more efficient allocation of resources. In a model of telecommunications trade such as FTAP2, as services supplied cross-border are not directly substitutable with domestic services, this conventional source of gains from trade is no longer available.17 Instead, the gains from liberalisation come from the movement of factors across borders, especially capital. Liberalisation increases foreign commercial presence, which becomes an increasingly important mode of delivery and a major source of potential benefits to liberalising economies. Increased trade in telecommunications will be delivered by relatively efficient suppliers offering differentiated services. In this context, the gains from liberalising telecommunications come from a more efficient use of global capital through the relocation of affiliates around the world. Liberalising telecommunications is therefore driven by global factor (and especially capital) mobility. The estimates of gains from liberalising trade in services presented in this chapter are probably lower bounds. The FTAP model does not account for the sectoral variation in barriers to trade. Experience with goods-trade models shows that variation in industry support within an economy is a major source of costs. This principle is expected to apply in services trade as well. Conversely, in the FTAP2 model, impacts on portfolio investment flows are omitted. As liberalisation changes returns to factors across economies, this may affect not only FDI but also portfolio investment flows across economies. Given the large size of portfolio flows, even if a small proportion of these flows are affected by improved market access, this could result in significant increases in the amounts of capital flowing to liberalising economies and further gains to liberalising trade in services. Future directions Modelling trade in services is a relatively new development and is needed to inform the debate on services trade liberalisation. Current work on services trade by the Australian Productivity Commission represents only a beginning. From this experience, analysing the liberalisation of services trade requires progress in the following areas: x
Modelling at the industry level is important, as it enables researchers to incorporate the key characteristics of each service type. These characteristics can play an important role in determining the effects of liberalisation. This requires theoretical developments to capture the behaviour of economic agents in these disaggregated sectors.
x
These theoretical developments require the collection of reliable and consistent data on bilateral stocks of FDI at the detailed sectoral level.
x
Better incorporation of technological change. Removing some regulations seems to lead to product innovations and more efficient production techniques. Neither of these effects has really been captured in the FTAP-style frameworks.
x
Developing further the theory behind measuring barriers to services trade. Current studies assume that barriers generate rents, measured as tax equivalents. The underlying assumption is that production is still efficient and restrictions cause only a redistribution 77
of income among economic agents. This type of cost is relatively small, represented by small “Harberger triangles”. However, regulations may cause inefficiencies and waste resources. Such waste (measured in terms of trapezes under demand and supply schedules) could be many times more than the triangles. Future research may identify to what extent regulations lead to this latter type of inefficiencies. These developments will lead to better estimates of gains from the types of liberalisation that have been considered in this chapter. However, some issues that are relevant to broader services trade negotiations require yet other developments in theory and supporting data. These include developing a framework and supporting data to account explicitly for other modes of services trade and accounting for potential barriers to these modes of supply. For example, the FTAP-style frameworks can deal with the impacts of barriers on educational institutions wishing to provide services outside of their home country through FDI. They can also deal with the potential benefits from offering courses over the Internet, a form of cross-border trade. However, they are less well equipped to analyse the impacts of barriers that might prevent students from attending a course in a foreign country.
78
NOTES 1.
“Consumption abroad” (for example, tourism and education) is accounted for in the database, but not identified separately in the database structure. Short-term “movements of natural persons” are partially accounted for.
2.
See Table 3.2 for a list of the economies included.
3.
The sectors are: primary, manufacturing, construction, distribution, telecommunications, finance, other services and ownership of dwellings. Only telecommunication and financial services are liberalised.
4.
This chapter does not address issues pertaining to the estimation of the barriers themselves. The reader is referred to the work by the Productivity Commission mentioned above. At this stage of model development, barriers are mainly expressed in the form of tax equivalents.
5.
A global CGE model accounts for all economic transactions in an economy and the trade flows between economies. It can be thought of as a collection of input-output models connected with diaggregated trade flows, and accounts for the behaviour of consumers, governments and firms in the face of changes in relative prices. To date, most global CGE frameworks do not account for global investment flows in a very satisfactory manner.
6.
The FDI stock data were fleshed out to provide a full bilateral matrix of FDI stocks by source, destination and sector, using RAS methods (Welsh and Strzelecki, 2000; Verikios and Zhang, 2001b).
7.
These are part of a comprehensive set of estimates of barriers to services trade, documented in Findlay and Warren (2000).
8.
These taxes are applied in the exporting region, rather than as “tariffs” in the importing region, so that the rents created by the barriers are retained in the exporting region. While this type of barrier may be relevant for cross-border supply of some services, and therefore appears in the form of barriers for the entire services sector, they may not be relevant to some services, such as telecommunication services.
9.
This statement should be heavily qualified, because the services sector estimates are based only on estimates of barriers to banking and telecommunications.
10.
While this study simulates the impacts of liberalising telecommunication and financial services, the discussion here is limited to liberalising telecommunications.
11.
At the cost of assuming that physical capital is reallocated as a result and portfolio, financial capital flows are assumed to react to changes in financial markets.
12.
This issue goes to the heart of the problem of modelling the role of proprietary information (firmspecific capital) and product differentiation and requires further theoretical development. One solution may be to disaggregate returns to capital into returns to physical capital and returns to knowledge capital.
13.
For this reason, any effect of a change in the price of domestic phone calls would affect the demand for international phone calls through income effects, rather than through direct substitution effects. This is in contrast with what happens in disaggregated goods-trade models for example, where an
79
import is often a close substitute for local production and there are large substitution effects associated with changes in relative prices. 14.
The pattern of estimates is more extreme than in Table 3.2 because they are not averaged with barriers to financial services. Barriers to financial services tend to discriminate against foreigners.
15.
Warren (2000a) indicates also that technological changes are making it difficult for governments to restrict access to services supplied cross-border.
16.
This is partly due to the low incidence of discriminatory barriers in telecommunication markets, but also to a modelling limitation which does not allow small market shares to grow very much. A recent development by Hanslow (2001) should alleviate this limitation.
17.
There are no barriers specific to cross-border telecommunications in FTAP2, though any market access barriers or derogations from national treatment that might affect producers of services supplied cross-border do apply to services supplied via this form of delivery.
80
REFERENCES
APEC (Asia-Pacific Economic Cooperation) (1995), Foreign Direct Investment and APEC Economic Integration, APEC Publication No. 95-EC-01.1, Singapore. Brown, D., A. Deardorff, A. Fox and R. Stern (1995), “Computational Analysis of Goods and Services Liberalisation in the Uruguay Round”, in W. Martin and L. Winters (eds.), The Uruguay Round and the Developing Economies, Cambridge University Press, Cambridge, pp. 365-80. Dee, P. and K. Hanslow (2000), “Multilateral Liberalisation of Services Trade”, Staff Research Paper, Productivity Commission, Ausinfo, Canberra. Disclosure (1999), Global Researcher – Worldscope Database, Disclosure, United States, January. Doove, S., O. Gabbitas, D. Nguyen-Hong and J. Owen (2001), “Price Effects of Regulation: International Air Passenger Transport, Telecommunications and Electricity Supply”, Staff Research Paper, Productivity Commission, Ausinfo, Canberra, October. Findlay, C. and T. Warren (eds.) (2000), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York. Francois, J., B. McDonald and H. Nordstrom (1995), “Assessing the Uruguay Round”, in W. Martin and L. Winters (eds.), The Uruguay Round and the Developing Economies, pp. 117-214, Discussion Paper No. 307, World Bank, Washington, DC. Hanslow, K., T. Phamduc and G. Verikios (1999), “The Structure of the FTAP Model”, Research Memorandum MC-58, Productivity Commission, Canberra. Hanslow, K. (2001), “The Application of CRESH-Based Functional Forms in CGE Models”, Research Memorandum GA-506, Productivity Commission, Canberra. Harrison, J. and K. Pearson (1996), “Computing Solutions for Large General Equilibrium Models using GEMPACK”, Computation Economics, Vol. 9, No. 2, pp. 83-127. Hertel, T. (ed.) (1997), Global Trade Analysis: Modelling and Applications, Cambridge University Press, Cambridge. Hertel, T. (1999), “Potential Gains from Reducing Trade Barriers in Manufacturing, Services and Agriculture”, paper presented at the 24th Annual Economic Policy Conference, Federal Reserve Bank of St. Louis, 21-22 October. Hoekman, B. (1995), “Assessing the General Agreement on Trade in Services”, in W. Martin and L. Winters (eds.), The Uruguay Round and the Developing Economies, World Bank Discussion Paper 307, World Bank, Washington, DC. 81
Johnson, M., T. Gregan, P. Belin and G. Gentle (2000), “Modelling the Impact of Regulatory Reform”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York. Kalirajan, K. (2000), “Restrictions on Trade in Distribution Services”, Productivity Commission Staff Research Paper, Ausinfo, Canberra. Available at www.pc.gov.au/research/staffres/rotids/index.html (accessed 15 October 2001). Kalirajan, K., G. McGuire, D. Nguyen-Hong and M. Schuele (2000), “The Price Impact of Restrictions on Banking Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York. Markusen, J., T. Rutherford and D. Tarr (1999), “Foreign Direct Investment in Services and the Domestic Market For Expertise”, paper presented at the Second Annual Conference on Global Economic Analysis, GI, Avernaes, Denmark, 20-22 June. McDougall, R. (1993), “Incorporating International Capital Mobility into SALTER”, SALTER Working Paper No. 21, Industry Commission, Canberra. McGuire, G. and M. Schuele (2000), “Restrictiveness of International Trade in Banking Services”, paper presented at the GTAP Advisory Board meeting, Centre for Global Trade Analysis, Purdue University, West Lafayette, Indiana, 11 April. McKibbin, W. (1999), “International Capital Flows, Financial Reform and Consequences of Changing Risk Perceptions in APEC Economies”, paper presented at the conference Experiences of Economic Reform within APEC, 12-14 July, Wellington. Nguyen-Hong, D. (2000), “Restrictions on Trade in Professional Services”, Staff Research Paper, Productivity Commission, Ausinfo, Canberra. Petri, P. (1997), “Foreign Direct Investment in a Computable General Equilibrium Framework”, paper presented at the conference, Making APEC Work: Economic Challenges and Policy Alternative, 13-14 March, Keio University, Tokyo. Secretariat of the Council for Trade in Services (1999), “Recent Developments in Services Trade – Overview and Assessment”, Background Note S/C/W/94, Geneva. At: www.wto.org/english/tratop_e/serv_e/w94.doc (accessed 3 June 2000). United Nations (1994), World Investment Directory Latin America and the Caribbean, New York. Verikios, G. and K. Hanslow (1999), “Modelling the Effects of Implementing the Uruguay Round: A comparison using GTAP model under alternative treatments of international capital mobility”, paper presented at the Second Annual Conference on Global Economic Analysis, GI Avernaes, Denmark, 20-22 June. Verikios, G. and X. Zhang (2001a), “Global Gains from Liberalising Trade in Telecommunications and Financial Services”, Staff Research Paper, Productivity Commission, Ausinfo, Melbourne. Verikios, G. and X. Zhang (2001b), “FTAP2: Theory and Data”, Research Memorandum, Productivity Commission, Canberra.
82
Warren, T. (2000a), “The Identification of Impediments to Trade and Investments in Telecommunications Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London. Warren, T. (2000b), “The Impact on Output of Impediments to Trade and Investment in Telecommunications Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London. Welsh, A. and A. Strzelecki (2000), “Estimating Domestic and Foreign Returns to Capital for the FTAP Model”, Research Memorandum MC-60, Productivity Commission, Melbourne. World Trade Organization (1996), “Trade and Foreign Direct Investment”, WTO, Geneva. Available at: www.wto.org/test/english/news_e/pres96_e/pr057_e.htm (accessed 18 May 2001). World Trade Organization (2001), Annual Report 2000 – International Trade Statistics, WTO, Geneva.
83
Chapter 4 LIBERALISING BASIC TELECOMMUNICATIONS: EVIDENCE FROM DEVELOPING COUNTRIES
by Carsten Fink, Aaditya Mattoo and Randeep Rathindran Development Research Group, World Bank
Abstract. This chapter deals with issues concerning the general linkages and interactions among different policies affecting telecommunication services. First, it introduces a new data set on telecommunications policy for 86 developing countries. Second, it analyses the impact of telecommunications policy on the sector’s performance. Three questions are addressed. First, what impact do specific policy changes – relating to ownership, competition and regulation – have on sectoral performance? Second, how is the impact of any one policy change affected by the implementation of complementary reforms? Third, does the sequence in which reforms are implemented affect performance?
Introduction* The dynamism of global telecommunications markets is widely attributed to rapid technological development and an increasingly liberal policy environment. Over the past decade, a large number of African, Asian and Latin American economies have embarked on reform and witnessed a significant expansion of their telecommunications networks and striking improvements in productivity.1 Over the period 1985-99, mainline penetration and productivity in these three regions more than tripled. Neither performance nor policy, however, was uniform within or across the regions. For example, while mainline penetration in Sri Lanka increased more than five-fold, Malawi saw a more modest 40% increase. It is clear where improved performance is due to specific policy choices and where more could have been achieved if policy choices had been different. Telecommunications liberalisation is a complex and relatively new process. Choices have to be made regarding the privatisation of state-owned telecommunications operators, the introduction of competition, the opening of markets to foreign investment and the establishment of pro-competitive regulations. While there is growing consensus that each of these elements is desirable, governments have differed in their willingness to concede control to the market and most have shown a penchant for gradualism. Competition has been introduced, but the number of firms has been fixed by policy; privatisation is often partial and there are limits on foreign participation; “autonomous” regulators have been created but are rarely fully independent. This chapter has a dual purpose. First, it introduces a new data set on telecommunications policy for 86 developing countries (see Appendix 1). Second, it analyses the impact of telecommunications policy on the sector’s performance. Three questions are addressed. First, what impact do specific policy changes – relating to ownership, competition and regulation – have on sectoral performance? Second, how is the impact of any one policy change affected by the implementation of complementary reforms? Third, does the sequence in which reforms are implemented affect performance? Several recent cross-country econometric studies examine the effect of telecommunications reform on sector performance (see Appendix 2).2 Wallsten (1999) explores the effects of privatisation, competition and regulation on several performance indicators, using a panel data set for 30 African and Latin American countries for the period 1984-97. While competition is generally found to have a positive effect on performance, the impact of privatisation is mixed. One weakness of Wallsten’s study is that it approximates the degree of competition in fixed-line telecommunications by the number of mobile operators not owned by the incumbent operator. This is inappropriate because many countries have introduced competition in mobile services while maintaining a monopoly in fixed-line services.
*
The views expressed here are those of the authors and should not be attributed to the World Bank. Special thanks to Simon Evenett for valuable advice and to Christine Zhen-Wei Qiang for providing access to some of the data used. The United Kingdom’s Department for International Development provided financial support for the services trade database used.
86
Ros (1999) examines the effects of privatisation and competition on network expansion and efficiency on the basis of data for 110 countries for the period 1986-95. Using fixed effects estimation, he finds that countries that allowed majority private ownership in their incumbent telecom operator had significantly higher teledensity (mainline penetration) and a higher growth rate in teledensity.3 Allowing a majority private stake in the incumbent was also found to improve efficiency (telephone mainlines per employee). By contrast, competition in at least one fixed-line market segment (local, long distance, international) did not significantly affect mainline penetration, but had a positive effect on efficiency.4 The present empirical investigation improves upon existing studies in several ways. First, not only are individual and interactive effects of policy choices explored, but also whether the sequencing of privatisation and competition affects performance. This dimension of telecommunications reform has not been analysed before. Second, the fact that, aside from policy reforms, improvements in telecommunications performance in recent years were driven by technological advances is explicitly allowed for. Technological progress is captured by an autonomous time trend and the relative importance of autonomous and policy-induced improvements in sector performance is quantified. Third, the study is based on more comprehensive data on policy and regulation than previous studies. The panel spans the period 1985-99 and thus captures a large number of reform initiatives in developing countries in the second half of the 1990s. The competition variable focuses specifically on observed competition in the local market segment, which is here considered to be the most relevant influence on teledensity and productivity. Finally, the estimates control for the problems of serial correlation and panel-level heteroscedasticity, which were not addressed by previous studies (Wallsten, 1999; Ros, 1999). The following discussion first describes the pattern of telecommunications policy and performance in Africa, Asia, Latin America and the Middle East. Next, a conceptual framework for analysing the impact of reforms on performance is presented and builds upon existing literature on the subject. Then the estimation methodology and results are presented. Concluding remarks follow. Telecommunications performance and policy in developing countries Over the period 1985-99, mainline penetration in all developing countries tripled from 2.4 telephone mainlines per 100 population to 7.27 mainlines per 100 population (Figure 4.1.a). The trend in productivity is even more impressive, rising from 27.2 mainlines per worker in 1985 to 91.2 mainlines per worker in 1999 (Figure 4.1.b). There is however, considerable variation in performance across regions. For example, mainline population. Africa experienced a three-fold increase from 1.03 in 1985 to 3.2 in 1999, while Latin America experienced a smaller proportionate increase from (a higher initial) 5.29 in 1985 to 13.4 in 1999.
87
Figure 4.1. Trends in mainline penetration and worker productivity in the telecommunications sector, 1985-99 4.1.a.Mainlines per 100 population (all countries)
4.1.b. Mainlines per worker (all countries)
8
100
7
90 80
6
70 5
60
4
50
3
40 30
2
20
1
10 0
0 1985
1990
1995
1985
1999
1990
1995
1999
Source: International Telecommunications Union.
The pattern of policy reform is equally diverse. In 1985, 8% of the Asian economies in the sample had a privatised incumbent, compared with 2% of African countries and 12% of Latin American countries. However, by 1999, about 42% of the Asian countries, 27% of the African countries and 68% of the Latin American countries in the sample had at least partially privatised their incumbent phone operators (Figure 4.2.a). Among the Middle East and Arab states in the sample only the United Arab Emirates had private ownership of the incumbent over the period 1985-99. Asian countries have been the most reluctant to allow foreign equity participation in their incumbent phone operators. By 1999, only 8% of Asian countries had privatised to foreign telecommunication companies or strategic investors; the corresponding figure was 24% for Africa and 38% for Latin America (Figure 4.2.b). Regarding competition, in 1990, no country in the sample had licensed an operator that competed with the incumbent local services provider. By 1999, two-fifths of Asian and Latin American countries had introduced some form of competition in basic services, while less than one-fifth of African countries had done so (Figure 4.2.c). None of the Middle East or Arab states had licensed a second fixed-line operator over the period 1985-99. In 1985, fewer than 10% of Asian, African and Latin American countries had instituted an autonomous regulator, but by 1999, 50% of Asian countries, 56% of African countries, 30% of Middle Eastern countries and 72% of Latin American countries had done so (Figure 4.2.d).
88
Figure 4.2. Patterns of policy reform in telecommunications
Proportion of countries with foreign ownership in the incumbent operator (by region)
Proportion of countries with privatised incumbent phone operators (by region)
1.00
Asia
1.00
Africa
0.80
Asia
0.80 0.60 0.40
Arab & East
0.20 0.00 1985
1990
1995
Africa
0.60
Latin America
0.40
Latin
0.20
Arab & East
0.00
1999
1990
Figure 4.2.a
1995
1999
Figure 4.2.b
Proportion of countries with competition in local services (by region)
Proportion of countries with an independent regulator (by region)
1.00
1.00
Asia
Asia 0.80
0.80 Africa 0.60 Latin 0.40 Arab & East
0.20
0.60
Africa
0.40
Latin
0.20 Arab Middle
0.00
0.00 1990
1995
1985
1999
Figure 4.2.c
1990
1995
1999
Figure 4.2.d
Source: International Telecommunications Union, Worldbank Database on Telecommunications Policy.
Conceptual framework The objective of this study is to seek a relationship between these diverse patterns of policy and performance. Three dimensions of policy are relevant: change of ownership, introduction of competition and strengthened regulation. Performance itself has two dimensions: internal efficiency within firms and allocative efficiency in the market. To generate testable hypotheses, the conceptual discussion in this section is linked to two proxy variables. The proxy for internal efficiency is labour productivity, measured by the number of mainlines per employee, and the proxy for allocative efficiency is the aggregate output, measured by the number of main lines. It is clear that each of these proxies is imperfect. For instance, internal efficiency is better measured by total factor productivity, and output may be a deceptive measure of allocative efficiency because, for example, there might be excessive expansion of the network. Nevertheless, these are the two measures that can be computed most easily with available data and the smallest measurement error. Privatisation involves the transfer from public to private hands of the ownership of productive assets, with the right to take allocative decisions and entitlement to the residual profit flows. Earlier analyses emphasised the impact of the resulting change in objectives from maximisation of social welfare to maximisation of profit (see Shapiro and Willig, 1990). The implication was that with a 89
concentrated market structure, public ownership was more likely to promote allocative efficiency than private ownership, where the temptation would be to restrict output to maximise profits. More recent analyses of the impact of a change in ownership have focused on the change in the incentives for the firm’s management (e.g. Levy and Spiller, 1996). Changes in performance are attributed to changes in the principal-agent relationship between the firm’s management (the agent) and either private shareholders or the government or ultimately the general public (the alternative principals). Private ownership is likely to lead to greater internal efficiency for a variety of reasons, ranging from lower costs of monitoring, more precise and measurable targets and greater flexibility to devise incentive contracts. In some ways, the traditional and more recent analyses are complementary. It would seem likely that a change in ownership from public to private (or foreign) hands would improve internal efficiency.5 The presumption of a positive impact on the chosen proxy, labour productivity, is even greater because public enterprises may seek to meet social or political objectives by creating excessive employment. The impact on the measure of allocative efficiency, the number of mainlines, is less obvious. Increased internal efficiency owing to privatisation would favour expansion, but the greater emphasis on private profitability may dampen the effect. However, the impact may still be positive if the public provider is resource-constrained in a way that the private (or foreign) provider is not, e.g. because the latter has better access to the capital market. Therefore: Hypothesis 1: Privatisation leads to an increase in labour productivity. There is a weaker presumption that it will lead to an increase in the number of mainlines. The results of increased competition would seem to be relatively straightforward, as it promotes both allocative efficiency and internal efficiency (Vickers and Yarrow, 1988). Firms, private or public, must produce efficiently in order to survive, and there is less scope for monopolistic restraint on output.6 There is, however, a twist. In some cases, public monopolies have sought to expand networks through a system of cross-subsidisation, using revenue from urban areas or international calls, for example, to extend services to poorer areas or consumers. The introduction of competition may threaten these arrangements. This possibility introduces an element of ambiguity to the relationship between increased competition and the expansion in the number of mainlines. On balance: Hypothesis 2: The introduction of competition will lead to an increase in productivity. There is a weaker presumption that it will lead to an increase in the number of mainlines. The impact of individual policy changes may alter when they are implemented in conjunction with other policy changes. Consider first the interaction between privatisation and competition. If a public monopoly is privatised, the introduction of competition helps eliminate the remaining scope for managerial slack and the monopolistic incentive to restrict output (Armstrong et al., 1994). At the same time, privatisation of a public monopoly renders the introduction of competition more credible by eliminating the government’s incentive to favour the public provider (Fershtman, 1989).7 One would therefore expect the interaction of privatisation and competition to have a positive impact on both internal and allocative efficiency, subject to the qualifications noted above. Thus: Hypothesis 3: The interaction of privatisation and competition will lead to an increase in productivity and the number of mainlines. The most critical complementary policy change is in the regulatory framework. In the case of basic telecommunications, regulation can play at least two roles (Laffont et al., 1998a; 1998b). First, if for any reason the market structure is not competitive, regulation of behaviour in the output market 90
(e.g. by fixing consumer prices) can help stimulate a more competitive outcome. In this sense, regulation can function as an imperfect substitute for competition when a public monopoly is privatised. Second, since the incumbent operator invariably controls access to essential facilities, i.e. the network, regulation of the terms of access to the network for entrants is necessary to create competition. Effective regulation of interconnection must, therefore, be seen as a precondition for the emergence of meaningful competition. For these reasons, the interaction of effective regulation with both privatisation and the introduction of competition can be expected to have a positive effect on performance. There is, however, one qualification. There is invariably a conflict between the regulatory objectives of ensuring competitive outcomes and access at any one point of time, on the one hand, and creating adequate incentives for cost-reduction and network expansion over time, on the other. Consider a simple example. A regulatory mechanism that sets prices equal to, say, average costs at every point in time encourages allocative efficiency but eliminates the firm’s incentives to reduce costs. Conversely, a regulatory mechanism that sets prices for a certain length of time allows firms to reap the benefits of, and hence provides incentives for, cost-reductions, but at the expense of allocative efficiency. Therefore, the relationship between regulation and performance is more complex, and requires a more detailed analysis of the nature of regulation than is possible on the basis of available data. Nevertheless, assuming that existing regulatory arrangements generally strike an appropriate balance between the two objectives: Hypothesis 4: The interaction of regulation with privatisation and competition leads to an increase in labour productivity and the number of mainlines. Finally, consider the implications of alternative sequences of reform involving, in particular, privatisation and competition. There are several reasons why it may matter if privatisation precedes the introduction of competition, essentially because the conditions of “competition” may be affected. First, the importance of location-specific sunk costs in basic telecommunications suggests that allowing one provider privileged access may have lasting consequences (Bos and Nett, 1990). Sunk costs matter because they have commitment value and can be used strategically by those who are allowed to enter the market first. The commitment value is stronger the more slowly capital depreciates and the more specific it is to the firm. In general, if one firm is allowed to enter the market early, it may accumulate a quantity of “capital” sufficient to limit, or modify the conditions of, entry of other firms.8 Because of the importance of sunk costs, sequential entry can produce very different results from simultaneous entry. When one firm enters first, the market outcome is not necessarily worse than when all firms enter at the same time, but, for several reasons, it may be. First, if entry is costly, the incumbent may be able to deter entry so that the outcome is a much more concentrated market structure.9 Second, the first-mover advantage may be conferred on an inferior (national) supplier who may nevertheless use it to establish a position of market dominance. How lasting such a position is depends on the degree of cost or quality advantage of more efficient firms.10 A second reason that sequences matter has to do with political economy. Allowing privileged access creates vested interests which may resist further reform or seek to dilute its impact. South Africa provides an example (Lamont, 2001). Private shareholders in the incumbent (national and foreign) successfully lobbied to reduce the number of entrants the government was planning to allow from two to one.11 Finally, sequences matter because of the implied changes in the regulatory environment. The prospects of new entry differ, depending on whether privatisation follows or precedes the introduction of competition. In the former case, the incumbent is a relatively inefficient public operator and the regulator is well informed about the cost structure. In the latter case, the 91
incumbent is a relatively efficient private operator and the regulator is less well-informed about the cost structure. It could be argued that new entry is easier to achieve in the former situation. While there are good reasons to believe that the sequence matters, it is not easy to predict the impact of different sequences. First, differences in internal efficiency may not persist once each of the sequences is complete. Thus, delaying the introduction of competition would allow the privatised monopoly a period of slack, but once competition is introduced, the incumbent would be forced to improve performance rapidly, so that there is no reason to assume continued differences in levels of productivity. As far as allocative efficiency (or its present proxy, mainlines) is concerned, allowing entry sequentially rather than simultaneously could lead to an inferior outcome if sunk costs are so high that new entry is blocked, with the monopolist incumbent producing an output lower than the output produced by, say, two firms that enter simultaneously. This would not necessarily be the case, because in some cases strategic behaviour by the incumbent could lead to a large expansion of output.12 The implications of alternative sequences is therefore an interesting empirical question. Hypothesis 5: Alternative sequences of reform do not have any impact on internal efficiency but matter for allocative efficiency. In particular, the number of mainlines created will be lower if privatisation takes place before the introduction of competition, rather than after or at the same time. Econometric investigation In this section, the above hypotheses are econometrically tested using the data on 86 developing economies in Africa, Asia and Latin America for the period 1985-99 (see Appendix 1).13 One limitation of an econometric investigation is the fact that available measures of policy do not capture the multiple dimensions of a complex reform process. For example, the mere existence of additional licences in a particular services segment is an imperfect indicator of effective competition, let alone the contestability of markets. Similarly, while the existence of a separate regulatory agency is likely to be a useful indicator of a government’s overall willingness to commit to pro-competitive regulatory principles, a regulator can be ineffective if key regulatory responsibilities (e.g. interconnection) fall outside its mandate. Moreover, the overall credibility of a government’s reform programme is not adequately captured by the policy proxies but is likely to exert an important influence on investment decisions, particularly foreign investment. These reservations notwithstanding, an econometric investigation has obvious attractions, especially in comparison to the existing case study evidence on the impact of policy reforms. A country’s level of development can be controlled for. For example, competition in a low-income country like India may not lead to the same level of mainline penetration as in a middle-income country like Malaysia. Country size by population is also controlled for. Countries with large populations tend to have lower teledensity than others. In the panel regression, country fixed effects can capture economy-specific idiosyncrasies that typically complicate cross-country comparisons. In addition, econometric estimates allow for assessing the relative importance of alternative policy reforms and, in some cases, their interaction.
92
The following specifications are assumed for the model:
y
i, t
D P G year C .J X .E H i ,t , i 1,2,...., N ; t 1,2,....T , i i, t i, t
where yi,t is the natural logarithm of the performance indicator (teledensity, mainlines per employee) in country “i” at time “t” LVWKHFRQVWDQWWHUP i is a country-specific dummy variable WKDW LV LQWHQGHG WR FDSWXUH WLPH LQYDULDQW FRXQWU\ IL[HG HIIHFWV LV WKH FRHIILFLHQW RQ D OLQHDU WLPH trend, Ci,t is a matrix of control variables, Xi,t represents the telecommunications policy variables (privatisation, competition, regulation), with the vector of corresponding coefficients of interest, N the number of countries (86 in this case), and T, the number of time series observations (15 in this case) per country.14 Country-wise heteroscedasticity, i.e. that the variance of the error term differs across countries, is allowed for. In addition, account is taken of the existence of first-order autoregressive serial correlation in the errors, but a common autocorrelation parameter across panels is assumed.15 The latter assumption is justified by the fact that the ¶s themselves do not vary across countries.16 The heteroscedasticity and autocorrelation corrections make the estimation far more efficient than an ordinary fixed-effects panel estimation. The model is estimated using Kmenta’s cross-sectionally heteroscedastic and time-wise autocorrelated (CHTA) approach.17 One concern is that the policy proxies may actually be endogenous variables in a regression on telecommunications performance. For example, privatisation may be prompted by the poor service record of the public monopoly. Ros (1999) uses an instrumental variable approach to correct for the potential endogeneity of his policy measures. However, his instruments for the endogenous variables are the same controls used in the second stage estimation. Since these controls are, by definition, correlated with the dependent variables (teledensity, efficiency), the value of this approach is unclear. In any case, his estimation results using the instrumental variables approach do not change from the simple fixed-effects estimates. In the light of this result and given the lack of obvious instruments, it was decided to proceed with a simple one-stage estimation. Effects of individual reforms on performance Table 4.1 presents the results of the first investigation of the effect of individual reforms on mainline penetration and productivity. The dependent variables are the number of mainlines per 100 inhabitants and the number of mainlines per worker (both in natural logs). As control variables, a time trend is used to capture reductions in switching and network costs due to technological progress, GDP per capita and population (both in natural logs). Mainline penetration is expected to be higher in developing countries with higher per-capita GDP, and lower in developing countries with larger populations. In the first model specification, the policy proxies are a dummy variable that takes the value 1 if an incumbent has been partially or wholly privatised and zero otherwise and a dummy variable that equals 1 if there is competition for local services and zero if local services are provided by a monopoly.18 It was argued above that privatisation and the introduction of competition are likely to lead to an increase in labour productivity, and (less strongly) to an increase in the number of mainlines (hypotheses 1 and 2). The empirical estimates in Table 4.1, column 1, suggest that both privatisation and competition significantly increase mainline penetration.19 The coefficient of the privatisation and competition dummy variables are positively significant at the 1% level. All controls and the time trend have the expected sign and are statistically significant at the 1% level. The results with regard to
93
labour productivity (Table 4.1, column 2) suggest that both privatisation and competition significantly boost productivity. Again, all controls and the time trend work as expected.20 Table 4.1. Effects of individual reforms on mainline penetration and productivity Dependent variable Time trend Natural log of per-capita GDP Natural log of population Dummy variable for privatisation Dummy variable for competition in basic services Wald Chi-squared (k-1) AR(1) coefficient Number of observations
Natural log of mainlines per 100 population .091*** (21.02) .294*** (8.27) -1.268*** (7.91) .076*** (4.58) .049*** (2.66) 60 818.28 .67 1 200
Natural log of mainlines per employee .094*** (19.82) .189*** (4.07) -.594*** (3.61) .176*** (7.68) .091*** (3.24) 15 385.82 .54 1 085
Note: All specifications estimated by feasible generalised least squares. “*”, “**” and “***” indicate statistical significance at the 10%, 5% and 1% levels, respectively. The bracketed figures are GLS-corrected z-statistics. Country fixed effects and the intercept are not reported. Source: Authors.
Whether the effects of privatisation and competition differ in the presence of a good regulatory framework were also tested (hypothesis 4). Accordingly, both dummy variables for privatisation and competition were interacted with a dummy variable equalling 1 if a separate regulatory agency exists and zero otherwise. As mentioned before, this is a crude measure of the quality of regulation and the results should therefore be interpreted with caution. Table 4.2, columns 1 and 3, present an estimated coefficient on the interaction terms. As above, both privatisation and competition – confined to observations that exhibit a good regulatory framework – impact positively on teledensity and productivity. Does the interaction of privatisation and competition matter? To capture the interdependence between privatisation and competition, another model was estimated which also includes a two-way interaction term. The interaction of these two policy choices is expected to impact positively on both mainline penetration and labour productivity (hypothesis 3). The findings with regard to teledensity (Table 4.2, column 2) confirm this hypothesis: the coefficients on privatisation and the interaction of privatisation and competition are both positive and statistically significant at the 1% level. Interestingly, competition was not found to be statistically significant in this model. This suggests that the beneficial effects of competition primarily occur through its interaction with privatisation. The same holds for labour productivity (Table 4.2, column 4): privatisation and the interaction of privatisation and competition are statistically significant at the 1% level, whereas competition is not statistically significant.
94
Table 4.2. Effects of combinations of reforms on mainline penetration and productivity Dependent variable
Time trend Natural log of per-capita GDP Natural log of population
Natural log of mainlines per 100 population
[1]
[2]
[3]
[4]
.09*** (21.24) .30*** (8.26) -1.28*** (8.08)
.09*** (21.15) .29*** (7.95) -1.28*** (8.03) .06*** (3.76) -.002 (.09) .09*** (2.64)
.10*** (20.33) .21*** (4.46) -.66*** (4.01)
.09*** (19.99) .18*** (3.97) -.60*** (3.66) .16*** (6.86) .00004 (0.00) .15*** (2.99)
Dummy variable for privatisation Dummy variable for competition in basic services Privatisation*competition Competition*regulation Privatisation*regulation Wald Chi-squared (k-1) AR(1) coefficient Number of observations
Natural log of mainlines per employee
.06*** (2.81) .07*** (4.19) 60 779.79 .67 1 200
61,296.67 .67 1 200
.12*** (3.83) .12*** (5.02) 15 225.99 .54 1 085
15 528.57 .54 1 085
Note: All specifications estimated by feasible generalised least squares. “*”, “**” and “***” indicate statistical significance at the 10%, 5% and 1% levels, respectively. The bracketed figures are GLS-corrected z-statistics. Country fixed effects and the intercept are not reported. Source: Authors.
Whether the presence of an independent regulator reinforces these results was also tested. The estimates with regard to teledensity (Table 4.3, column 1) suggest otherwise. Only the coefficient on the interaction of privatisation and regulation is statistically significant. Both the interaction of competition and regulation and the three-way interaction term are statistically insignificant. It should be pointed out, however, that these two variables are highly correlated (Appendix 4), which may partially account for this “adverse” result. The estimates with regard to labour productivity (Table 4.3, column 3) are similar to those presented above: privatisation and the interaction of privatisation and competition (now in the presence of good regulation) are statistically significant, whereas competition is not.
95
Table 4.3. Effects of full reform (compared to partial or no reform) on mainline penetration and productivity Dependent variable
Time trend Natural log of per-capita GDP Natural log of population Competition*regulation Privatisation*regulation Dummy variable for privatisation, competition and regulation – full liberalisation Wald Chi-squared (k-1) AR(1) coefficient Number of observations
Natural log of mainlines per 100 population
Natural log of mainlines per employee
[1]
[2]
[3]
[4]
.09*** (21.25) .30*** (8.33) -1.29*** (8.11) .05 (1.61) .07*** (4.00) .009 (.24)
.09*** (21.55) .29*** (8.02) -1.33*** (8.20)
.10*** (21.03) .21*** (4.44) -.67*** (4.10)
.08*** (3.29)
.10*** (20.40) .21*** (4.53) -.65*** (3.98) -.02 (.29) .11*** (4.66) .17*** (2.43)
61 350.43 .67 1 200
58 194.82 .67 1 200
15 431.61 .53 1 085
15 270.15 .54 1 085
.19*** (5.58)
Note: All specifications estimated by feasible generalised least squares. “*”, “**” and “***” indicate statistical significance at the 10%, 5% and 1% levels, respectively. The bracketed figures are GLS-corrected z-statistics. Country fixed effects and the intercept are not reported. Source: Authors.
How large are the effects of policy reform relative to autonomous increases? In order to quantify the effects of “complete” liberalisation – defined as the introduction of competition, privatisation of the incumbent and the establishment of a separate regulator – a model was used in which the only policy variable is a dummy variable that equals 1 if all three policies are in place and zero otherwise (i.e. the three-way interaction term). This variable was found to be highly significant for both mainlines and productivity (Table 4.3, columns 2 and 4). The estimated coefficients suggest that mainline penetration is 8% higher and productivity is 21% higher in years of complete reform than in years of no or partial reform. It is revealing to compare these magnitudes to the implied growth in teledensity and productivity due to autonomous factors, including technological progress. The estimated coefficients on the linear time trend suggest an increase of approximately 9% a year in both mainline penetration and productivity. Hence, the empirical investigation suggests that the effect of the policy reforms studied here, while significant, played a relatively minor role in the dramatic increase in teledensity and labour productivity. It should be kept in mind, however, that the time trend captures an average effect across all countries and the influence of policy reforms on the diffusion of telecommunications technology is not considered. The latter is beyond the scope of this study and, arguably, would require explicit data on the international diffusion of telecommunications technology. Does the sequence of reform matter? Having found evidence of the beneficial effects of privatisation and competition and of the interaction of the two on performance, the effects of the order in which the two are introduced were investigated. While it is known that the interaction of privatisation and competition results in a significantly higher mainline penetration, the questions was whether the effects are different if privatisation takes place before the introduction of competition, or vice versa. As argued above, mainline penetration is expected to be higher if competition and privatisation are introduced at the
96
same time than if privation precedes the introduction of competition (hypothesis 5). For labour productivity, little difference is expected in the effects of alternative sequences of policy reforms. Simultaneous introduction of policies is defined as reforms in which privatisation and competition are introduced within a one-year time period. Since no country in the sample introduced competition more than one year before privatising the public operators, a possible third sequence, in which competition clearly precedes privatisation, was not observed.21 (However, some countries introduced competition for local services, but as of 1999, had not privatised their state-owned operator.) To test the effects of different sequences, four dummy variables were constructed (see Appendix 5 for details). First, the “simultaneous sequence” (hereafter, SEQSIM) is represented by a variable that takes the value 1 for the year in which privatisation and competition were introduced simultaneously as well as all subsequent years, and zero otherwise. Second, the “privatisation before competition sequence” (hereafter, SEQPC) is represented by a variable that takes the value 1 for the year in which competition was introduced after privatisation as well as all subsequent years, and zero otherwise. Third, the “competition only” (hereafter, SEQC) variable takes the value 1 for all years in which only competition is observed, and zero otherwise. Finally, the “privatisation only” (hereafter, SEQP) variable takes the value 1 for all years in which only privatisation is observed, and zero otherwise. The estimation results are presented in Table 4.4. The coefficients on both SEQPC and SEQSIM are positive and statistically significant at the 1% level. The two coefficients are significantly different from each other, implying that the simultaneous introduction of policy reforms has a greater impact on mainline penetration than the “privatisation before competition” sequence (see Figure 4.3).22 The coefficient on SEQP (which represents years where only privatisation is observed) is significant, whereas the coefficient on SEQC (which represents years where only competition is observed) is not. Hence, introducing competition without privatising the incumbent firm does not have a significant effect on mainline penetration, whereas introducing privatisation before competition leads to higher mainline penetration. Table 4.4, column 2 shows the estimated effects of alternative sequences of privatisation and competition, given the existence of an autonomous regulator.23 The estimated coefficients and significance levels are similar to those discussed above. The existence of an independent regulator neither reinforces nor softens the sequencing effect. Finally, the effect of different sequences on productivity was tested for (Table 4.4, columns 3 and 4). The effects of different sequences were found to be positive and statistically significant at the 1% level, but the coefficients of the two alternative sequences were not significantly different from each other.24 Productivity is significantly higher in years where only privatisation is observed, whereas the effect of competition (without privatisation) is not statistically different from zero. As above, interacting all policy variables with the regulation dummy does not fundamentally change this result.
97
Table 4.4. Effects of sequencing of reform on mainline penetration and productivity Dependent variable
Time trend Natural log of per-capita GDP Natural log of population Dummy variable for privatisation only (SEQP) Dummy variable for competition only (SEQC) Simultaneous introduction of competition and privatisation (SEQSIM)) Privatisation before competition sequence (SEQPC) SEQP (in the presence of an independent regulator) SEQC (in the presence of an independent regulator) SEQSIM (in the presence of an independent regulator) SEQPC (in the presence of an independent regulator) Wald Chi-squared (k-1) AR(1) coefficient Number of observations
Natural log of mainlines per 100 population [1] [2] .09*** .09*** (21.08) (21.37) .29*** .31*** (7.94) (8.74) -1.28*** -1.30*** (8.00) (8.18) .05*** (2.90) -.01 (.45) .22*** (5.13) .12*** (4.21) .05*** (2.71) .05 (1.01) .22*** (5.10) .08*** (2.78) 61 037.59 62 208.69 .67 .67 1 200 1 200
Natural log of mainlines per worker [3] [4] .09*** .10*** (19.89) (20.51) .19*** .20*** (4.04) (4.36) -.59*** -.64*** (3.55) (3.94) .14*** (5.82) -.007 (.11) .39*** (6.17) .29*** (7.10) .10*** (4.12) .01 (.15) .38*** (6.08) .25*** (6.09) 15574.19 15583.57 .54 .54 1 085 1 085
Note: All specifications estimated by feasible generalised least squares. “*”, “**” and “***” indicate statistical significance at the 10%, 5% and 1% levels, respectively. The bracketed figures are GLS-corrected z-statistics. Country fixed effects and the intercept are not reported. Source: Authors.
Figure 4.3. Example of alternative policy sequences and their effects
Teledensity Simultaneous competition + privatisation Competition after privatisation
O
Only privatisation
Competition after privatisation
Simultaneous competition + privatisation
98
Time
Conclusion This chapter presents an analysis of the impact of policy reform in basic telecommunications on sectoral performance in 86 developing countries in Africa, Asia and Latin America over the period 1985-99. While most countries experienced substantial increases in teledensity and sectoral productivity – in part driven by fast technological progress in telecommunications – the approach to policy reform differed markedly across regions and countries. Most governments have been unwilling to commit to complete liberalisation immediately, preferring instead a gradual reform process, encompassing the privatisation of state-owned operators, the introduction of competition and the establishment of independent regulation. The econometric evidence presented in this study may provide some guidance on possible priorities for telecommunications reform. First, complete liberalisation pays off. Ceteris paribus, teledensity is 8% higher and labour productivity 21% higher in years that saw privatised incumbents, additional competitors and separate regulators than in years with no or only partial reform. Second, both privatisation and competition improve performance and the latter reinforces the former. Third, sequences matter. Introducing competition after privatising incumbent operators leads to fewer mainlines per population than simultaneous introduction of the two policies. This suggests that delays in introducing competition – for example owing to market exclusivity guarantees granted to newly privatised entities – may adversely affect performance even after competition is introduced. An interesting supplemental finding is that, in the past 15 years, improvements in telecommunications performance not attributable to the policy variables considered here have outweighed the impact of policy. According to the study’s crude quantification, autonomous developments accounted for an average increase of 9% a year in both teledensity and productivity. One possible explanation is the rapid pace of technological progress in telecommunications. Another is increased public investment in this sector. A deeper exploration of these issues was beyond the scope of this study, but is a priority for future research. Two questions seem particularly important. What kind of policies support technological diffusion? What role does foreign investment play in transferring modern telecommunications technology to developing countries? More research is also necessary to verify and refine the other findings presented in this study. Improved data would make it possible to analyse several issues that have not been addressed here. How much is to be gained from eliminating all barriers to entry when some competition has already been allowed? How great are the gains from eliminating all barriers to foreign investment when some is already permitted? How significant are the benefits of making commitments under regional and multilateral trade agreements with regard to present and future policy? It will become possible to respond to these questions when more detailed data become available and when more observations for the period following the implementation of policy changes and multilateral commitments are available.
99
Annex 1. The ITU-World Bank Database on Telecommunications Policy
The telecommunications reform process is now advanced enough to have produced the data needed to analyse the implications of alternative policy choices. While the International Telecommunications Union (ITU) has a comprehensive database on performance indicators, there did not exist until now any worldwide database containing detailed time-series information on telecommunications policy. The ITU and the World Bank have recently created a database on telecommunications policy and regulation. The database covers 86 developing countries in Africa, Asia and Latin America. The policy data are drawn from a variety of sources, including responses by governments to an ITU questionnaire, information from World Bank programmes in various developing countries, World Bank Aide-memoires, the Tradeport and International Trade Administration (ITA) databases of the US Department of Commerce, country reports of the Economist Intelligence Unit (EIU), and direct queries to national regulators and telecommunications operators across the world. The data cover various aspects of policy and market structure in fixed-line and mobile telecommunications including, inter alia, information about corporatisation of the incumbent public telephony operator, the share of private equity, the share of foreign equity, the market structure in local, domestic long distance, and international services, and the year an independent regulator was instituted.25 Assumptions made in the creation of the database and sample selection Observed policy changes Data on variables such as private equity and competition are recorded based on observed private equity shares, or observed entry and commencement of services. There usually exists a substantial time lag between the announcement of a policy and an observed result. For example, suppose a government would like to introduce competition. First, it has to pass a new law, which has to be ratified by its parliament. Decisions also need to be made on how many operators to admit, in what regions, and so on. This is followed by the auctioning of the licences, the bidding process which takes time to settle. Even after licences are awarded, there still is a time lag before the licensee(s) enter the market and effectively commence(s) service provision. It seems best to consider a market competitive at the point when a second operator begins providing basic services since this is the least ambiguous criteria. For instance, using the date of issue of licences as an indicator of when competition began can be misleading as licences are sometimes withdrawn or revoked when governments change. Similar considerations arise for the privatisation of a state-owned network operator, with a long time lag (of at least one to two years) between the government’s announcement of its desire to privatise, and the completion of the sale of equity. Timing of policy changes The panel data are on an annual basis but it is sometimes difficult to assign a particular policy to a particular year. For example, if the second operator in Nigeria only commenced services in 100
November 1996, the starting year of effective competition was taken to be 1997 rather than 1996. As a rule, any entry relatively late in a given year was taken as effective from the following year. This approach seemed appropriate because the main concern was to link policy changes in a particular year to the performance variables compiled by the ITU. Similarly, if the sale of a public enterprise was completed relatively late in the year, the privatisation was recorded as effective from the following calendar year. Entry and geographical market segmentation Sometimes, a country has more than one telephone operator, but each has a monopoly in its respective region. For example, Bangladesh has two basic network operators – the incumbent Bangladesh Telephone and Telegraph Board (BTTB), which provides services in the urban areas, and the Bangladesh Rural Telephone Authority (BRTA), licensed in 1990, which provides basic services in rural areas. Similarly, in Argentina, ENTel was separated into two companies in 1990, Telecom Argentina, which provides services in the north, and Telefonica de Argentina in the south. Since the markets are geographically segmented, it was judged appropriate to consider each country as having a monopoly in basic services. Privatisation in limited segments In some cases, the domestic long-distance or the international long-distance segment is separated from the local services segment and then privatised. In this study, only privatisation of local service providers was taken into account.
101
To explore the effects of privatisation, competition and regulation on mainline penetration, payphone penetration, connection capacity and local call prices.
To examine the effects of privatisation and competition on network expansion and efficiency.
Ros (1999)
Objective
Wallsten (1999)
Study
110 countries (including developed countries), 1986-95.
Time period, regional focus and sample 30 African and Latin American countries 1984-97.
Ordinary fixed effects and fixed effects with instrumental variable correction.
Ordinary fixed-effects panel estimation.
Estimation technique
2. No evidence of privatisation leading to higher growth of mainline penetration in countries with annual per-capita income below USD 10 000.
3. Interaction of privatisation and regulation positively correlates with connection capacity and mitigates negative effect of privatisation on mainline penetration. 1. Countries with majority privatised public telephony operators have higher mainline penetration, and to a lesser degree, higher growth in mainline penetration.
2. Privatising an incumbent negatively correlated with mainline penetration and connection capacity.
1. Competition significantly correlated with increased mainline penetration, connection capacity, payphone penetration and a decrease in a local calling prices.
Results
2. Large sample makes fixed effects appropriate.
1. Use instrumental variables to correct for endogeneity.
1. Analyses the costs associated with granting a privatised incumbent an exclusivity period.
Strengths
Annex 2. Review of Empirical Literature on Telecommunications Policy and Performance
3. No corrections for complications in panel error structure.
2. Takes competition and privatisation as policy variables and ignores the effect of an independent regulator.
1. Sample period does not include developing country liberalisation of the late 1990s.
2. No correction for complications in the panel error structure.
1. Weak measure of competition, i.e. use of the number of mobile operators not owned by the incumbent captures spurious correlation.
Problems with analysis
al.
.
Source: Authors
Li and Xu (2000)
Fink et (2001)
Boylaud and Nicoletti (2000)
To explore the relationship between privatisation, competition, regulatory autonomy and interconnection policies on fixed and mobile capacity, profitability and local calling prices.
To investigate the effects of entry liberalisation and privatisation on productivity, prices and quality of service in long-distance (domestic and international) and mobile services. To ascertain the impact of privatisation, competition and regulation on mainline penetration, network quality and productivity.
50 countries over the 1990s.
12 East and South Asian economies 1985-99.
23 OECD countries 1991-97.
Ordinary fixed effects and Arellano-Bond (1991) dynamic panel estimation technique.
Ordinary fixed effects panel estimation.
Fixed effects, robust regressions and random effects.
3. Autonomous regulator has a negative impact, and competition no impact on mainline penetration. The interaction of competition and interconnection has a strongly negative impact on mainline penetration.
2. In model with interactions, only the interaction of privatisation and a good interconnection policy has a significantly positive effect on mainline penetration.
2. Countries that privatise, introduce competition and establish an independent regulator see much higher levels of mainline penetration, network digitalisation and productivity than others. 1. In no interactions model, privatisation is significantly positively associated with mainline penetration.
1. Interaction of privatisation and competition significantly increases mainline penetration.
1. The prospect of competition (measured by time remaining until liberalisation) has a strong positive effect on productivity, quality of services and a strong negative effect on prices.
2. Use of GMM technique to correct for endogeneity.
1. Use of sophisticated interconnection and competition policy indices from pyramid research rather than policy dummies.
2. Uses various techniques to check robustness of estimations. 1. Useful evidence that policy interactions matter, rather than individual policy effects.
1. Exhaustive study of OECD countries regulatory system and reform agendas.
2. Counterintuitive results.
1. Use of information on mobile competition in measuring competition in fixed-line services.
2. No corrections for complications in panel error structure.
1. Sample too small to make inferences about other developing countries.
2. No corrections for complications in panel error structure.
1. No analysis of developing countries.
Annex 3. Choice of Estimation Technique
Estimating a model containing time-series cross-section (TSCS26) data typically implies a complicated regression error structure that involves serial and/or contemporaneous correlation and heteroscedasticity. Models that feature these kinds of non-spherical disturbances are usually estimated by feasible generalised least squares27 (FGLS). A model that involves contemporaneous error correlations, serial error correlation and group-level heteroscedasticity is estimated by researchers using Park’s FGLS method. It is worth noting the criticism of Beck and Katz (1995) on panel data estimation by the Park’s FGLS method. Beck and Katz propose using OLS panel-corrected standard errors (PCSE) estimation, rather than GLS. Based on Monte Carlo simulations, they infer that GLS estimates that correct for contemporaneous28 correlation and panel-specific29 serial correlation produce standard errors that lead to extreme overconfidence, with variability often underestimated by 50% or more. A second genre of TSCS models features errors that are serially correlated, and group-wise heteroscedastic, but not contemporaneously correlated. These models are typically estimated using Kmenta’s cross-sectionally heteroscedastic and time-wise autocorrelated (CHTA) technique, which is also an FGLS procedure. CHTA first transforms the data to eliminate serial correlation in the errors, and then transforms the transformed data to correct for group-wise heteroscedasticity using panelweighted least squares (PWLS). Using Monte Carlo evidence, Beck and Katz (1996) criticise this approach, arguing that, although CHTA does not produce dramatically incorrect estimates or standard errors, its PWLS component is no more efficient than OLS, and further, that it is better to model dynamics using a lagged dependent variable, rather than an autoregressive process for the error. In this study, it was decided to estimate the model using Kmenta’s CHTA approach, assuming a common autocorrelation parameter across countries. Since neither contemporaneous correlations30 nor country-specific serial correlation is assumed, criticisms regarding the inaccurate computation of standard errors mentioned earlier do not apply. While a lagged dependent variable could have been used,31 which Beck and Katz suggest is a better way to capture dynamics, its estimation typically requires the use of instruments if there is serial correlation in the error. Kiviet (1995) has shown that estimation of dynamic panel data models using instrumental variables leads to poor finite sample efficiency. Moreover, it is hard to find good instruments.
104
1.00 .3121 (0.00) .2972 (0.00) .3911 (0.00) .2478 (0.00) .6321 (0.00) .3082 (0.00)
Privatisation
Source: Authors.
Note: Numbers in brackets indicate p-values
Privatisation*competition *regulation
Privatisation*regulation
Competition*regulation
Privatisation*competition
Regulation
Privatisation Competition
Variable
.2285 (0.00) .8626 (0.00) .7812 (0.00) .3008 (0.00) .6798 (0.00)
1.00
Competition
.2007 (0.00) .3453 (0.00) .6163 (0.00) .3005 (0.00)
1.00
Regulation
.6824 (0.00) .3656 (0.00) .7881 (0.00)
1.00
Privatisation* competition
.4135 (0.00) .8701 (0.00)
1.00
Competition* regulation
Annex 4. Partial Correlations between Various Reforms
.4876 (0.00)
1.00
Privatisation* regulation
1.00
Privatisation* competition* regulation
Annex 5. Construction of Sequencing Dummy Variables
The table below illustrates the construction of the sequencing dummies for Malaysia and El Salvador. Malaysia privatised its incumbent Telkom Malaysia in 1990. Competition in basic services was only introduced in 1996, so that Malaysia followed the “privatisation before competition” sequence. On the other hand, El Salvador introduced competition in 1997, but privatised only in 1998, so that El Salvador followed the “simultaneous” sequence.
0 0 0 0 0 1 1 1 1 1 1 0 0 0 0
Simultaneou s sequence (SEQSIM) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Privatisation before competition sequence (SEQPC) 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 1 1
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Country
Year
Only competition observed (SEQC)
Only privatisation observed (SEQP)
Malaysia Malaysia Malaysia Malaysia Malaysia Malaysia Malaysia Malaysia Malaysia Malaysia Malaysia Malaysia Malaysia Malaysia Malaysia
1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
El Salvador El Salvador El Salvador El Salvador El Salvador El Salvador El Salvador El Salvador El Salvador El Salvador El Salvador El Salvador El Salvador El Salvador El Salvador
1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Source: Authors.
106
NOTES 1.
Substantial reform has also taken place in Eastern Europe. However, this study focuses on developing countries where network development was much more limited.
2.
There are also studies that examine the link between telephone density (or teledensity) and economic development. For example, Jipp (1963) first brought to light the strong correlation between teledensity and the level of GDP per capita. Other studies look at the relationship between telecommunications liberalisation and macroeconomic performance, e.g. Mattoo et al. (2001).
3.
The positive contribution of privatisation to the growth in teledensity is not statistically significant for countries with a per-capita GDP below USD 10 000.
4.
Boylaud and Nicoletti (2000) provide additional econometric evidence of the impact of liberalisation and privatisation on productivity, prices and quality of long distance and mobile services, focusing on 23 OECD countries over the period 1991-97. Their findings suggest a generally favourable impact of policy reforms on productivity, quality and prices in the trunk (domestic long distance), international and mobile segments. It is not clear, however, to what degree these results apply to developing countries, most of which have had to implement reforms in situations where telecommunications networks are poorly developed.
5.
Foreign ownership may also be associated with the transfer of improved technology.
6.
Competition also makes it easier to monitor managerial performance, e.g. by diluting the management’s monopoly on information.
7.
De Fraja (1991) arrives at an opposite conclusion. In a theoretical model of Cournot oligopoly, it is shown that the continued presence of a welfare-maximising public firm can impose added competitive pressure on private firms.
8.
Capital need not necessarily take a physical form. A firm may be able to develop a clientele through advertising and promotional campaigns and pre-empt demand. The more imperfect consumers’ information and the more important the costs of switching suppliers, the greater the clientele effect. Consumers are often reluctant to switch telecommunications suppliers even when new entrants offer better terms. Each of these forms of “capital accumulation” enhance the first-mover advantages and allow established firms to restrict or prevent competition.
9.
In situations of network externalities, entry may also be deterred through the choice of a standard that is incompatible with that of potential entrants.
10.
Two qualifications to this argument are important. First, entry by the more efficient firm may take place through acquisition, thereby circumventing some of the problems of first-mover advantage. This would require no asymmetry of information about the value of assets and no direct costs of transferring assets. Second, incumbents may learn by doing: the experience acquired by the established firms during the previous period reduces their current costs, which enhances their competitiveness, and discourages others from entering. This form of entry deterrence may well promote welfare.
11.
While the political economic implications of sequencing are emphasised here, there are also important strategic considerations. For instance, Perotti (1995) argues that one reason that there is partial
107
privatisation is the government’s inability to commit credibly to non-interference after the transfer of ownership takes place. 12.
For instance, the aggregate output in a Stackelberg oligopoly equilibrium, where one firm has a firstmover advantage, need not be lower than in a Cournot equilibrium, where all firms make output decisions simultaneously (Tirole, 1988).
13.
Liberia, Seychelles and Cuba had to be omitted for lack of GDP data. Some small island nations, for example, Vanuatu and Western Samoa, where country size is a constraint on having more than one operator were also omitted.
14.
A Hausman test for fixed versus random effects rejected the hypothesis of random effects, and therefore country-specific differences are captured by the intercept rather than the error term.
15.
A preliminary examination for group-wise heteroscedasticity used the likelihood ratio test. The model was first estimated with only heteroscedasticity and no autocorrelation using iterated generalised least squares (GLS), then the same model was estimated with neither heteroscedasticity, nor autocorrelation, and the likelihood in both cases was compared. In models without autocorrelation, GLS estimates are equivalent to maximum likelihood estimates. A likelihood ratio test of the variances in the two models turned out a F2(74) statistic of 848.71, which strongly rejected the null hypothesis of no group-wise heteroscedasticity. Economically, the reason for the presence of heteroscedasticity is somewhat unclear. Why should the variance of shocks to mainlines differ across countries? It could be because of differing government initiatives on mainline expansion under different regimes, so that countries with a more volatile political environment, or unstable and frequently changing governments have a higher variance in the level of mainlines per capita than others owing to differing government initiatives on mainline expansion. Another hypothesis is that the richer developing countries can more easily than poorer countries overcome natural and geographical obstacles (for example terrain) in laying down the network. Countries also differ in their capacity to adapt to technology shocks and this could be an additional source of variances across countries.
16.
As Beck and Katz (1995) admit, the assumption of a common autocorrelation parameter across panels is unlikely to cause FGLS estimates to estimate variability inaccurately, as it necessitates the calculation of only one additional unknown parameter (the autocorrelation coefficient). Appendix 3 provides more detail on the choice of estimation technique.
17.
Contemporaneous correlations across panels cannot be assumed, as the estimation technique would require as many time-series observations as there are panels to satisfy matrix invertibility conditions during estimation. In the present case, there are only 15 time-series observations per country for 86 countries. Since country-specific correlation is not modelled, the criticism of Beck and Katz (1995) regarding the inaccurate computation of standard errors does not apply.
18.
Regressions were also run with a dummy variable for corporatisation of the incumbent. The coefficient on this variable was consistently insignificant.
19.
A similar model in which the privatisation measure was replaced by a dummy variable that takes the value 1 if foreign equity was observed and zero otherwise was also estimated. The results are similar, which is not surprising given that most privatisation takes place through the sale of strategic equity to foreign investors. Indeed, the partial correlation between the privatisation and foreign equity dummy variables exceeds 0.8.
20.
The findings for labour productivity mirror the results in Ros (1999). However, Ros only finds privatisation to exert a significant impact on mainline penetration. A similar fixed-effects ordinary least squares (OLS) regression was run using this study’s data for the years 1986-95, as in Ros’s
108
specification. The impact of both competition and privatisation are still significantly positive. The most plausible explanation for this result is that the present estimation sample only consists of developing countries, where initial network conditions were weaker and subsequent growth faster. By contrast, most of the countries that introduced competition in Ros’s estimation samples were developed countries that already had a well-developed telecommunications network. Moreover, the different findings may also be due to the different control variables and different specifications of this study’s policy proxies. 21.
As an alternative, a variable was created that represented a “competition before privatisation” sequence, allowing for situations in which competition was introduced before privatisation, even if the gap between the two was only a few months. The estimation results were similar to the ones presented here. The simultaneous sequence was chosen, however, since it is unlikely that there are significant sequencing effects from policies that are introduced within a short time of one another.
22.
H0: SEQSIM (.22) = SEQPC (.12); F2(1) = 3.70; Prob > F2 = .0544.
23.
Care was taken to exclude observations where autonomous regulation was introduced only after privatisation and competition. This led to the exclusion of the Bahamas, Chile, the Dominican Republic and Surinam from the regression sample. Had observations on these countries been included, it would have had the effect of the regulatory variable disrupting a previously chosen sequence and making it start afresh.
24.
H0: SEQSIM (.39) = SEQPC (.29); F2(1) = 1.95; Prob > F2 = .163.
25.
As detailed time-series information on mobile service providers is not yet available, this study focuses on the fixed-line segment.
26.
The term “time-series cross-section data” does not mean the same thing as “panel data”. The latter typically has a few repeated observations on a large number of sampled units. The terms “panel”, “group” and “country” are used interchangeably. For a good exposition on panel data analysis see Hsiao (1986).
27.
Essentially, a feasible generalised least squares procedure first estimates the model by ordinary least squares (OLS), and uses the OLS residuals to estimate serial correlations, if any, in the error. These estimated serial correlations are then used to transform the model into one with serially independent errors. The transformed model is then estimated by OLS, and the residuals are used to estimate the error variance-covariance matrix that contains the estimated contemporaneous correlations. The estimated contemporaneous error correlations and variances are then used to transform the model yet again into one with no contemporaneous correlations and no heteroscedasticity, which can be easily and accurately estimated by OLS.
28.
Suppose there are T time-series observations in each of the N panels/groups. Each element of the matrix of contemporaneous covariances is estimated, on average, using 2T/N observations. If the ratio of T/N is close to 1, then contemporaneous covariances are calculated using about two observations, which is problematic as their accuracy would be highly questionable.
29.
If group-specific autocorrelation (modelled by a first-order autoregressive process – AR1) processes are assumed, then the necessary computation of N extra autocorrelation coefficients (one for each of the N groups), based on only T time series observations per group, is likely to cause more serious underestimations of variability. It is widely accepted that autoregressive parameters estimated in samples of 30 or fewer time-series observations are inaccurate and downward biased. See, for example, Nickell (1981).
109
30.
Contemporaneous correlations across panels are not assumed as the estimation technique would require as many time series observations as there are panels to satisfy matrix invertibility conditions during estimation. This study is based on only 15 time–series observations per country for 86 countries.
31.
Another interesting estimation technique that might have been used is the Arellano-Bond (1991) procedure for dynamic panel data estimation (or panel estimation with a lagged dependent variable) as it could help account for any endogeneity in the explanatory variables. This technique uses a generalised method of moments (GMM) estimation procedure and features variables in first differences with lagged values of explanatory variables acting as instruments. However, lagged values make good instruments only if there is no second-order serial correlation in the error term of the first differenced regression. In the present case, second-order serial correlation in the error was present, and this would have made the Arellano-Bond technique inappropriate and the statistical inferences obtained erroneous.
110
REFERENCES
Arellano, M. and S. Bond (1991), “Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations”, Review of Economic Studies, 58, pp. 277-297. Armstrong, M., S. Cowan, and J. Vickers (1994), Regulatory Reform: Economic Analysis and British Experience,. MIT Press, Cambridge, Massachusetts. Beck, N. and J.N. Katz (1995), “What to do (and not to do) with Time-Series Cross-Section Data”, American Political Science Review 89(3), pp. 634-47. Beck, N. and J.N. Katz (1996), “Nuisance vs. Substance: Specifying and Estimating Time-Series Cross-Section Models”, Political Analysis; 6, pp. 1-36. Bös, D. and L. Nett (1990), “Privatization, Price Regulation, and Market Entry An Asymmetric Multistage Duopoly Model”, Journal of Economics (Zeitschrift für Nationalökonomie). Vol. 51(3), pp. 221-257. Boylaud, O. and G. Nicoletti (2000), “Regulation, Market Structure and Performance in Telecommunications”, Economics Department Working Paper, No. 237, OECD. De Fraja, G. (1991), “Efficiency and Privatization in Imperfectly Competitive Industries”, Journal of Industrial Economics, Vol. 39(3), pp.311-321. Economist Intelligence Unit, Various Market Reports, 1990-2000. Fershtman, C. (1989), “The Interdependence between Ownership Status and Market Structure: The Case of Privatization”, Economica, 57, pp. 318-28. Fink, C., A. Mattoo and R. Rathindran (2001), “Liberalizing Basic Telecommunications: The Asian Experience”, HWWA Discussion Paper, No. 163. HWWA-Institut für Wirtschaftsforschung. Hsiao, C. (1986), Analysis of Panel Data, Cambridge University Press, New York. Jipp, A. (1963), “Wealth of Nations and Telephone Density”, Telecommunications Journal, July, pp. 199-201. Kiviet, J. F. (1995), “On Bias, Inconsistency, and Efficiency of Various Estimators in Dynamic Panel Data Models”, Journal of Econometrics; 68, pp. 53-78. Laffont, J-J., P. Rey and J. Tirole (1998a), “Network Competition: I. Overview and Nondiscrimatory Pricing”, The Rand Journal of Economics. Vol. 29(1), pp. 1-37. Laffont, J-J., P. Rey and J. Tirole (1998b), “Network Competition: II. Price Discrimination”, The Rand Journal of Economics. Vol.29(1), pp. 38-56. Lamont, J. (2001), “South Africa U-turn on Telecoms Competition”, Financial Times. 15 August.
111
Levy, B. and P. Spiller (1996), Regulation, Institutions, and Commitment: Comparative Studies of Telecommunications, Cambridge University Press, Cambridge. Li, W. and C. L. Xu (2000), “Liberalization and Performance in the Telecommunications Sector around the World”, mimeo, The World Bank. Mattoo, A., R. Rathindran and A. Subramanian (2001), “Measuring Services Trade Liberalization and its Effect on Economic Growth: An Illustration”, Policy Research Working Paper Series, The World Bank. Nickell, S. (1981), “Biases in Dynamic Models with Fixed Effects”, Econometrica 49, pp. 1417-1426. Perotti, E.C. (1995), “Credible Privatisation”, American Economic Review, 85(4), pp. 847-859. Ros, A.J. (1999), “Does Ownership or Competition Matter? The Effects of Telecommunications reform on Network Expansion and Efficiency”, Journal of Regulatory Economics 15, pp. 65-92. Shapiro, C. and R.D. Willig (1990), “Economic Rationales for the Scope of Privatization”, Discussion Paper No. 41. Woodrow Wilson School of Public and International Affairs, Princeton University, Princeton, New Jersey. Tirole, J. (1988), The Theory of Industrial Organization, MIT Press, Cambridge, Massachusetts. Vickers, J. and G. Yarrow (1988), Privatization: An Economic Analysis, MIT Press, Cambrige, Massachusetts. Wallsten, S.J. (1999), “An Empirical Analysis of Competition, Privatization and Regulation in Africa and Latin America”, Policy Research Working Paper No. 2136, The World Bank.
112
Chapter 5 QUANTIFYING COSTS TO NATIONAL WELFARE FROM BARRIERS TO SERVICES TRADE: A REVIEW OF THE LITERATURE
by Nora Dihel OECD Trade Directorate
Abstract. This chapter provides an overview and discussion of current work on measuring and modelling barriers to trade in services. It examines the main results found in the computable general equilibrium (CGE) literature on services trade liberalisation, indicates the different assumptions made in various studies and differences in the model structures and relates these to the overall patterns of results. The review shows that liberalisation of services trade generates overall welfare gains under all modelling assumptions. The studies indicate that economies with initial high protection levels tend to gain most (in terms of percentage gains as a share of GDP). As the values of estimates for services trade barriers are higher for developing than for developed countries, this suggests that developing economies are potentially the major winners.
Introduction and overview This chapter provides an overview and discussion of current work on measuring and modelling barriers to trade in services. It examines the main results found in the computable general equilibrium (CGE) literature on services trade liberalisation, indicates the different assumptions made in various studies and differences in the model structures and relates these to the overall patterns of results. The chapter first outlines the main empirical findings on the costs to national welfare from barriers to services trade and the gains from liberalisation and analyses these results on the basis of the underlying modelling framework, the estimates of barriers employed and the ways of incorporating these estimates into the model. The review shows that liberalisation of services trade generates overall welfare gains under all modelling assumptions. The studies indicate that economies with initial high protection levels tend to gain most (in terms of percentage gains as a share of GDP). As the values of estimates for services trade barriers are higher for developing than for developed countries, this suggests that developing economies are potentially the major winners. The chapter then presents the advantages, as well as the limitations, of using numerical general equilibrium techniques for quantifying the effects of liberalising trade in services. The quantification of services barriers is unlikely ever to be sufficiently accurate to be used directly in the actual conduct of GATS negotiations. However, by providing an order of magnitude of the costs of services barriers and the corresponding welfare gains from their removal, quantification is a valuable tool for demonstrating what is at stake in the liberalisation of trade in services. In other terms, it may yield useful insights into the opportunity costs of forgoing liberalisation opportunities in the GATS 2000 Round. Finally, recent and ongoing work to overcome the aforementioned limitations is presented, with a view to providing some guidance on possible next steps for future efforts. Even if the data currently available are far from satisfactory, important progress is being made to improve the measurement of services trade and stocks of foreign direct investment (FDI). Moreover, work on measuring restrictions to services trade is beginning to appear and can lead to a more accurate assessment of the scope and significance of actual barriers. This review of the literature suggests that existing economy-wide models may be used to assess the effects of multilateral services liberalisation in a realistic manner. While useful work has been done on incorporating separate analysis of different modes of services delivery into the model, such analysis is strongly dependent on the accuracy of the input data employed. Efforts to make modelling work of greater practical value to the negotiating community will need to focus on generating the required data on a sector-by-sector and country-by-country basis and on identifying how different policies affect the economic performance of firms supplying services.
114
Review of individual publications Measuring the economy-wide impact of trade liberalisation requires a global, general equilibrium framework, which captures both the intersectoral effects in each economy and the links among countries. This enables an assessment of the impact of industry-specific policies on the economy at large. While many general equilibrium studies on the quantification of economic impacts of policies affecting goods trade are available, relatively little work has been done to assess the potential gains from alternative liberalisation scenarios in services. The difficulties arise from poor information on international services transactions and on prevailing barriers to trade in services, as well as from the need to develop a different modelling structure from that used for goods trade in order to incorporate the various modes of services supply (i.e. to account for the movement of factors of production). Box 1 identifies the special data and modelling requirements which should be fulfilled to simulate liberalisation in services sectors, given the special characteristics of services. Existing CGE assessments of services trade liberalisation will be classified according to the modelling requirements shown in Box 1, in order to present trends in the existing empirical literature to date. The following classification summarises pertinent work in this area: Studies which do not explicitly model the different modes of services supply: x
Brown et al. (1996).
x
Chadha (2000).
x
Chadha et al. (2000).
x
Hertel et al. (1999).
x
Benjamin and Diao (1998, 2000).
x
Australian Department of Foreign Affairs and Trade (1999).
x
Robinson et al. (1999).
x
Kawai and Urata (1998).
Studies which explicitly model FDI in services sectors: x
McKibbin and Wilcoxen (1996).
x
Petri (1997).
x
Dee and Hanslow (2000a).
x
Verikios and Zhang (2000).
x
Brown and Stern (1999).
115
Trade
Market access: measures which restrict the contestability of markets. National treatment: policies which discriminate between domestic and foreign suppliers to the advantage of domestic providers.
In addition to the larger spectrum of barriers than in the case of goods, it is necessary to determine whether regulations constitute barriers and to establish whether the incidence of the regulation is more burdensome than necessary to achieve a legitimate policy objective.
-
-
The restrictions to international services transactions typically take the form of non-tariff barriers and are designed to limit not only the access of foreign services, but mostly the access of suppliers or consumers to the domestic market. Conventional non-tariff instruments of trade policy like quantitative restrictions, price-based instruments, licensing or certification requirements, discriminatory access to distribution and communication systems are imposed especially on services providers and classified in the GATS in two main categories relating to:
The special characteristics of services, such as their intangible nature, the high prevalence of regulatory intervention to avoid market failure and achieve non-economic social benefits, the requirement for proximity between producers and consumers and the factor-mobility associated with their trade, determine the nature of restrictions in services trade.
Identification and quantification of barriers to trade in services
Data requirements Data on bilateral flows of services transactions. Separate data for the different modes of supply: Cross-border supply. Consumption abroad. Commercial presence. Movement of natural persons.
To assess whether a regulation unduly impairs market access or alternatively, whether the absence of a particular regulation constitutes a trade barrier (for example, the lack of appropriate essential regulation of facilities in many industries that require access to a network, such as telecommunications, gas, electricity, water and transport), the structure-conduct-performance methodology has been developed. This method assesses the effectiveness of a regulatory regime on the basis of: x The delineation of the key structural elements that are considered most likely to drive the performance of the regime. x The identification of the incentives that alternative elements create for the major actors involved with the regime. x The assessment of the cross-national impact of these incentives on key industry outcomes.
Such indirect methods do not assess the extent to which existing policies contribute to the computed differences (i.e. the extent to which actual policies reduce trade or raise the costs of entry and/or operating post-entry).
Direct methods, which involve: x Elaboration of comprehensive listings of impediments to trade in services on a sectoral basis. x Determination of the restrictiveness of listed barriers. This index represents in quantitative form the nature and extent of restrictions in services sectors. x Estimation of the effects of these restrictions on prices, costs, price- cost margins or quantities and the correlation of the estimated effects with the underlying restrictions, considering also the restrictiveness index. Indirect methods, which assume: x Determining differences between actual and assumed free-trade flows. or x Using price-cost margins for the estimation of services barriers.
The following quantification methods, based on the measurement of non-tariff barriers, extensively used in the case of goods trade, are available:
x x
Box 5.1. Data and modelling requirements for simulating liberalisation of services trade
Characteristics of trade in services The four-part typology of international services transactions adopted by GATS as a basis for multilateral liberalisation negotiations constitutes the generally recognised framework for the analysis of services. It determines the following data and modelling requirements on international services transactions for the analysis of services with the help of CGE models.
Source: Author.
Restrictions
According to the information derived from the qualitative analysis of the barriers, appropriate incorporation of estimates of barriers to trade in services into a CGE model: x As revenue-generating tariff equivalents. x As cost-raising measures. x As rent-creating measures.
Modelling requirements Development of a more complex modelling framework than that considered for analysing liberalisation of trade in goods. Such a modelling structure needs to consider the different modes of supply, i.e. incorporation of elements which allow the representation of FDI, one of the most important modes of services supply.
The main results of the studies listed are provided below, preceded by a brief description of the underlying modelling framework and of the principal characteristics of data inputs (services transactions and estimates for barriers) for each group. In addition, a presentation of results using a standardised format and a bibliography of studies examining the effects of liberalisation policies in services sectors are included in an annex. Studies which do not explicitly model the different modes of services supply In considering services liberalisation, the first group of studies does not explicitly model the different modes of services supply, but implicitly includes the reduction of barriers in all modes of supply. The studies rely on different versions of the GTAP1 (Global Trade Analysis Project) database. Almost all studies in this group use the estimates of tariff equivalents for services barriers constructed by Hoekman (1995) using frequency index methodology. These estimates are calculated on the basis of scheduled commitments under the GATS, using a three-category weighting method. They therefore include, in principle, barriers that are applied to all four modes of supply. However, as GATS schedules do not include all barriers that are in place, these estimates are derived from an incomplete database.2 Furthermore, the weighting methodology applied does not distinguish between barriers in terms of their impact on the economy (i.e. a blanket prohibition on foreign presence will have the same weight as an economic needs test). Only Hertel et al. (1999) improve the quality of services barriers estimates by incorporating the new, quantity-based measures of services barriers calculated by Francois and Hoekman (1999). They estimate tariff equivalents for business and financial services and for construction services by taking the difference between actual and predicted imports as indicating barriers existing in the analysed sectors. The predicted “liberalised” imports are calculated on the basis of a free-trade benchmark.3 The estimates of services trade barriers are incorporated into the models either as revenueraising tariff equivalents or as cost-raising measures. The first assessment of multilateral services trade liberalisation was undertaken by Brown et al. (1996). They simulated a 25% reduction in the services barriers estimated by Hoekman, using different modelling assumptions. Under the assumption of imperfect competition in the services sectors, including economies of scale and product differentiation, they found that all countries gain in the aggregate from the assumed services trade liberalisation, with increases in welfare varying between 0.4% and 2.1% from the base year GDP. The largest welfare gains in absolute terms would accrue to Europe (USD 28.7 billion), the United States (USD 25.8 billion) and Japan (USD 12.5 billion), but in relative terms, Mexico (2.1% increase from GDP base year), Australia-New Zealand (1.9%) and the newly industrialised Asian countries (1.0%) do best. Total trade and output expand for all countries/regions, but the terms of trade vary by small amounts in favour of some and against others. According to the simulation results, the United States, Europe and the newly industrialised Asian countries experience slightly improved terms of trade as a result of liberalisation, while the terms of trade of other countries worsen, though never by much, and never by enough to cause a net reduction in overall welfare, which improves for all countries. Owing to a different initial database on services transactions, as well as different assumptions about liberalisation policies in non-service sectors, somewhat larger effects were observed by Chadha (2000). He employed the same estimates for services barriers and the same modelling structure as Brown et al. to simulate a 25% reduction in tariff equivalents for services. Given the regional focus of his paper, the aggregate impact of the trade liberalisation scenario is reported for 117
India, the rest of South Asia and the ASEAN countries, as well as for developed regions, thereby enabling more extensive comparison between developing and developed countries. The author finds that the liberalisation of services would lead to relatively large gains in economic welfare, as well as in trade in the services sectors themselves. The developing countries would gain more than the developed countries in percentage terms and would encounter higher welfare gains from services trade liberalisation than from goods trade liberalisation. For both groups of countries, however, gains are larger when services liberalisation is accompanied by liberalisation of trade in agriculture and manufactured products. Drawing on similar assumptions, Chadha et al. (2000) then extend the computational analysis of multilateral liberalisation and estimate the worldwide impact in 2005 of the current WTO trade negotiations. For this purpose, the authors project the initial GTAP database from 1995 to 2005 and readjust it by taking into account the results of the Uruguay Round agreement. They assume that the Uruguay Round did not lead to any reduction in barriers to services trade and that liberalisation would only take place in the next round. They consider a one-third reduction of the tariff equivalents of services barriers and find that the potential welfare benefits of liberalising services are five times those of liberalising minerals and manufactures and represent 82% of the total welfare gain of USD 835.6 billion. While the developed countries gain USD 550.8 billion (80% of the USD 687.9 billion total services gain), the relative gains for the developing countries are nonetheless quite large. Additional unilateral liberalisation simulation scenarios are envisaged for India. They show that although unilateral services liberalisation scenarios result in substantial gains, the assumed multilateral liberalisation scenarios result in greater gains in welfare and noticeably higher returns to labour and capital. Hertel et al. (1999) also estimate the impact in 2005 of the WTO negotiations. Compared to Chadha et al., they find lower but still large potential welfare gains of 354.2 billion from liberalising services trade. These gains represent 72% of the total of USD 493 billion due to liberalisation. Reasons for the different results may include the fact that: x
Hertel et al. employ a modelling structure different from that of the previous studies. It assumes perfect competition and ignores additional channels for gains from trade through increased competition, scale and product variety.
x
They use a different set of barrier estimates and incorporate them differently into the model. Contrary to all of the studies mentioned above, they do not model services protection with revenue-raising tariff equivalents; rather, they consider barriers to trade in services as cost-raising components for firms attempting to access the market. They view services liberalisation as increasing the amount of services delivered from a given level of export effort, thereby reducing the effective price of services in the domestic market. They model liberalisation by introducing an import-augmenting technical change component into the model. This component is set at a rate such that the effective price of services imports – considered to be above their free-trade level by the amount of the tariff equivalent estimates – is reduced by 40%.
Benjamin and Diao (1998; 2000) take a similar approach, modelling barriers to trade in services as cost-raising elements. They analyse liberalisation of cross-border trade in services in the APEC Forum using a global applied general equilibrium model with an imperfectly competitive services sector. They assume that services providers face fixed costs and are able to practice price discrimination across countries. They model liberalisation to reduce fixed costs and to remove the market segmentation that allows for price discrimination. In comparing liberalisation of trade in 118
services with general trade liberalisation, they find that the developed APEC members receive the greatest gains from services trade liberalisation, while the developing countries gain greater benefits from goods liberalisation. Several other economy-wide modelling efforts bear mention. They include work conducted under the auspices of the Centre for International Economics in Canberra. Like Hertel et al., it uses the GTAP framework and subsequently the dynamic APG–cubed model. The results are reported in a study published by the Australian Department for Foreign Affairs and Trade (1999). Using the GTAP framework, the study estimates worldwide gains of over USD 400 billion from halving goods and services barriers. With USD 250 billion, services would contribute over 60% of the total gains from liberalisation. The large welfare gains are expected because of the substantial scope for productivity improvements in the services sector. According to the study, gains would be concentrated in the biggest economies, such as the European Union, North America and Japan. Nevertheless, all participant countries would stand to gain, and, as a proportion of GDP, the gains would be spread fairly evenly. The projected gains are not large, because all countries would receive a shock that would favour consumption of domestically produced goods and services. In this area, the modelling significantly understates the dynamic effects of services reforms. Results using the APG-cubed dynamic model show gains to real GDP and welfare from eliminating barriers of similar magnitude to those suggested by GTAP. When these gains are realised, however, depends on the pace at which liberalisation is carried out. The results indicate that removing protection over the five years from 2004 determines additional welfare gains of more than USD 600 billion as early as 2008. Liberalisation over a ten-year time frame delays these gains until around 2013. Robinson et al. (1999) explore the impact of changes in trade in services, including changes in technology and protection, as well as the potential importance of technological externalities that are transmitted through trade in services, intermediate inputs and capital goods. They find that liberalisation in services trade, i.e. a 50% cut in services protection, would improve the welfare of all participating countries, accounting for a worldwide gain of 1.05% from the global base year GDP. Interestingly, the authors find that the welfare gain for the world as a whole from a 50% cut in services sectors protection is five times larger than that from trade liberalisation in non-services sectors. A complete elimination of protection in all manufacturing sectors generates a smaller gain in real GDP than a 50% cut in the level of protection of services sectors. When productivity gains from technology embodied in services imports are also considered, welfare increases substantially. The rise in total factor productivity (TFP) due to liberalisation in the services sector is potentially very significant. Welfare gains more than double and in some cases do even better, especially in developing countries. McKibbin and Wilcoxen (1996) and Kawai and Urata (1998) take similar approaches with respect to the incorporation of TFP. McKibbin and Wilcoxen4 do not model the impact of services/investment liberalisation directly; they analyse how productivity growth in services sectors – as an implicit effect of services liberalisation – increases the return to capital, which then feeds through to the rest of the economy. They conclude that the rise in productivity has large effects on GDP because the services sector is a large share of many economies and has strong links to other sectors of the economy. Higher productivity leads to significant capital inflows, attracted by a higher return to capital in the services sector. The inflow of capital leads to an appreciation of the exchange rate which increases the price of
119
exports and lowers the price of imports, thus worsening the overall trade balance in the short to medium term and lowering output in some sectors. Kawai and Urata use a single-country model to estimate the cost of regulation in the Japanese services industry. Although services are not internationally traded (i.e. no barriers to trade in services have to be removed), the authors look at how deregulation and liberalisation of services sectors improve productivity and prices. Under an exogenously determined increase of 10% in TFP in services sectors, they find that overall price levels fall and real GDP rises by more than 5%. By halving the gap in services TFP between Japan and the United States – which is chosen as a benchmark for measuring productivity – they find that real GDP rises by over 9%. The results from the three preceding studies suggest that TFP is an important factor to consider when evaluating the possible consequences of trade and regulatory reform in services sectors. Studies which explicitly model FDI in services sectors The second group of studies not only model cross-border trade, they also explicitly incorporate trade in services through commercial presence. Explicitly capturing this significant mode of contesting services markets substantially improves the framework for modelling services. Data on international services transactions are taken from the GTAP database (the more recent version 4 with base year 1995) and are supplemented by estimates of FDI stocks and activities of firms calculated by the Australian Productivity Commission. Petri’s (1997) pioneering work on incorporating FDI in services sectors used Hoekman’s (1995) estimates of barriers to trade in services based on frequency index measures, but subsequent studies incorporate improved estimates in the models. These estimates are based on more complete lists of barriers, which provide more detailed qualitative descriptions than those contained in GATS schedules. This allows for a more detailed interpretation and assessment of the restrictiveness of barriers than Hoekman’s methodology. The estimates also provide an indication of the extent to which the underlying barriers influence the economic performance of services firms. They are discussed in more detail when analysing the individual papers below. While a number of studies have attempted to analyse liberalisation of overall FDI using CGE models, Petri is the first to apply this approach to investment in services sectors. He introduces FDI into the model, distinguishing between activities of domestic and foreign-owned firms. Products are differentiated by both country of ownership and place of production, while capital allocation among sectors and between domestic and foreign investments responds to changes in rates of return and to investor preferences. Barriers to FDI are taken from Hoekman and are modelled as a tax on FDI profits. Petri examines only liberalisation of trade in goods and of FDI in services and does not consider liberalisation of other modes of services trade. When barriers to trade in goods and services (i.e. reduction of barriers to FDI in services sectors) are reduced by 50% in all APEC economies on a most favoured nation (MFN) basis, the global welfare gains are about USD 260 billion annually. APEC’s developing economies – the newly industrialising economies (NIEs), ASEAN-4 and China – gain proportionately most, partly because they start out with high initial barriers and partly because they are, as major traders, the greatest beneficiaries of the expansion of trade stimulated by liberalisation. If investment is not liberalised, global gains are diminished by USD 60 billion (or 23%) relative to full liberalisation. With only preferential trade and investment liberalisation among APEC economies, welfare gains are USD 59 billion lower than for MFN liberalisation. In this scenario, FDI from the rest of the world (ROW) is excluded from liberalisation. Since ROW is a
120
principal foreign investor in these countries’ services sectors, the ROW’s lower FDI limits the benefits from services sector liberalisation for the United States, Canada, Australia and New Zealand. Following the work of Petri, Dee and Hanslow (2000a) incorporate into their model a bilateral treatment of FDI (the FTAP model). They also distinguish barriers to commercial presence from those affecting other modes of supply and non-discriminatory barriers to market access from discriminatory restrictions on national treatment. Estimates for barriers to trade in services are taken from Kalirajan et al. (2000), who estimated the impact of barriers to trade in banking services on prices, and from Warren (2000a; 2000b), who estimated the impact of barriers to trade in telecommunication services on quantities of telecommunication services delivered which he subsequently transformed into price impacts. Kalirajan et al. found that non-prudential trade restrictions applied on a discriminatory basis to foreign banks raise the price, or “net interest margin”, of banking services by 5-60%. The price impacts of restrictions on foreign banks are highest for Indonesia, Korea, the Philippines, Malaysia, Chile, Singapore and Thailand. By contrast, Argentina, Australia, Canada, the EU, Hong Kong (China), Switzerland and the United States appear to have relatively low non-prudential regulations for foreign banks. The effect of market access barriers – which apply equally to domestic and foreign banks – on the price of banking services is lower, ranging from 0 to 24%. Warren found that the price effect of restrictions on foreign telecommunication providers is less than 20% for the majority of economies studied. Colombia, Indonesia, the Philippines, South Africa, Thailand and Turkey are the most restrictive, with price effects ranging from 20% to more than 100%. The United States, Denmark, the Netherlands, New Zealand, Australia, Germany and Sweden are the least restrictive. Market access barriers – which apply equally to domestic and foreign suppliers – can raise the price by up to 70%. These barriers are modelled by Dee and Hanslow (2000a) not as cost-raising components, but as tax equivalents which generate rents. The rents on output are modelled as accruing to the selling region, and those on capital are modelled as accruing to the region of ownership, after subtracting the general property income tax, which is collected by the government in the region of location. Thus, a portion of the rent associated with barriers to trade in services is assumed to remain in the region of location in the form of property income tax revenue, while the remainder accrues to the region of ownership. Services liberalisation may therefore have significant income effects in both home and host regions as these rents are gradually eliminated. Dee and Hanslow find welfare gains of about USD 130 billion (or 0.46% of world real income) from liberalising trade in services. These are roughly of the same order of magnitude as those associated with liberalisation of trade in goods and thus generate an overall global welfare gain of USD 260 billion annually. Almost all economies are projected to gain individually from simultaneous goods and services liberalisation. Nevertheless, the impact of these two liberalisation scenarios differs significantly across countries. While economies such as China, Hong Kong (China), Indonesia and Chile record significantly greater gains for services liberalisation than for goods liberalisation, the converse is true for New Zealand, Japan, Korea, Malaysia, Thailand, the EU, Canada and the United States. Contrary to all previous studies, the impact of multilateral services liberalisation is projected to be negative for the EU, Canada, the United States and Chinese Taipei. As the breakdown for welfare shows, these results are explained mainly by the contributions to income of changes in the rents from barriers to services trade as these barriers are eliminated. Critical to determining the overall welfare impact of services liberalisation is the assumption about who receives the impediment rents. As shown above, it is assumed in FTAP that for foreign-owned industries impediment rents 121
accrue mostly to the owning region, with only an income tax contribution going to the region in which the industry is located. Therefore, countries with high outward FDI and low initial impediments lose from the reduction in rents countries in high-barrier in which they invest. This is the case for the United States, Canada, the EU and Chinese Taipei, where the loss of rents offsets positive contributions to welfare. Given the uncertainty about the allocation of existing rents and the aggregation bias, the results should be interpreted with care. Rather, they should be used to identify further research efforts. There is evidence that trade and investment restrictions:
x
May generate rents which do not necessarily accrue to the region of ownership.5
x
May raise costs in sectors such as aviation (Johnson et al., 2000; Tamms, 2000).
x
May both create rents and raise costs.
As a result, the arbitrariness of simulation results is increased by assuming as typical for most services a simple average of estimates in banking and telecommunication services and a rent-creating effect for the region of ownership. Therefore, the authors are re-evaluating their approach to modelling barriers to trade in services. They state: “A topic for further research is to use the next version of the GTAP, which will have more services sector detail, to model barriers to each service separately, thus overcoming the arbitrariness of these assumptions” (i.e. of taking barriers to banking and telecommunication services as indicative for all other services sectors). Also, as estimates of the effects of barriers in other sectors are incorporated into the model, “it will be appropriate to treat some restrictions as cost-raising rather than as rent-creating” (Dee and Hanslow, 2000b). Building on the same model structure and using the same estimates for barriers to trade in services as Dee and Hanslow, Verikios and Zhang (2000) extend the analysis by providing greater sectoral detail. The disaggregation of the single services sector into six sectors is supported by data related to cross-border trade, to flows of FDI and to impediments. This study simulates separately complete multilateral liberalisation of trade in communication services and financial services in a post-Uruguay Round environment and finds positive welfare effects for the world as a whole in both cases. When liberalising trade in communication services, the world as a whole is projected to gain about USD 13 billion or 0.05% in terms of real income; when liberalising trade in finance, insurance and business services, the expected gains are about USD 3.5 billion or a 0.01% rise in real income. The total gain from liberalising both sectors is around USD 16 billion. As for Dee and Hanslow, the distribution of liberalisation effects differs significantly across countries. Most regions are expected to gain from liberalisation, the major winners being China, Japan and Indonesia in the first scenario, and Singapore, Malaysia, Chile and China in the second. Again as for Dee and Hanslow, economies such as the United States, the European Union and Hong Kong (China) are expected to be worse off in terms of real income in both situations. Brown and Stern (1999) employ an approach which draws on the model structure developed by Petri and extended by Dee and Hanslow to simulate liberalisation of trade in services. They model services liberalisation as a reduction of the average fixed costs by an amount equal to the estimates of financial data on gross operating margins calculated by Hoekman (1999) by country and sector. These estimates are taken to be indicative of the relative size of barriers in these sectors. However, no direct correlation between the actual restrictions and the data of estimated margins could be made on 122
the basis of these calculations. The authors find large welfare gains and a rather high dispersion of welfare results under all scenarios and parameter specifications. The far more pronounced welfare effects than in previous applied general equilibrium analysis of trade liberalisation are a consequence of allowing capital reallocation effects to be captured. The capital inflow is correlated with an expansion in output by most or all sectors of the economy. The less mobile the capital, the smaller the welfare effects of liberalisation in terms of absolute levels and dispersion. Concluding remark The overview of the literature shows that liberalisation of services trade generates overall welfare gains under all modelling assumptions. Larger gains are indicated in studies assuming imperfect competition in the services markets, as they capture economies of scale. Studies belonging to the group that explicitly model FDI also capture capital reallocation effects. The empirical exploration also shows that the potential importance of technological externalities transmitted through trade in services is significant, contributing large shares to trade liberalisation gains. As for the regional distribution of welfare effects, the studies indicate that, in general, economies with initially high protection levels tend to gain most (in terms of gains as a percentage of GDP). As the values of estimates for services trade barriers are higher for developing than for developed countries, this suggests that the latter are potentially the major winners from services liberalisation. Advantages and limitations of quantification through general equilibrium models While different models using diverse modelling assumptions and input data can lead to differences in the range and quality of results, they all show that the worldwide effects of services liberalisation are at least of the same order of magnitude as those associated with full liberalisation of trade in goods. Advantages of general equilibrium models The quantification of services barriers is unlikely ever to be sufficiently accurate to be used directly in the actual conduct of GATS negotiations. However, by providing an order of magnitude of the costs of services barriers and the corresponding welfare gains from their removal, quantification is a valuable tool for demonstrating what is at stake in the liberalisation of trade in services. General equilibrium modelling offers two particular benefits: x
It is the only technique enabling an economy-wide assessment of services barriers; this is of particular importance given the linkages between the services sectors and the rest of the economy. In capturing the input-output forward links between services and other sectors of production, modelling helps in assessing the impact of trade policy changes on the economy as a whole. It also provides the best framework for describing the dual function of services: they can either be traded directly or embodied in merchandise trade or in other traded services. In short, computable general equilibrium models give a complete picture of the potential welfare implications, including impacts on consumer and producer welfare, and the mechanisms through which these effects are transmitted, e.g. competition effects, increased efficiency or cost differential effects.
x
It may provide some useful insights into broad negotiating modalities. For example, Dee and Hanslow tentatively conclude that it is difficult to find a Pareto improvement (an outcome where at least some economies gain and none lose) from 123
partial liberalisation when it involves removing only one type of barrier (to market access, national treatment, commercial presence or other modes of services delivery). The authors conclude that a better strategy may be to negotiate gradual reductions in all types of barriers simultaneously. Limitations of general equilibrium models However, the results obtained using general equilibrium models should be taken only as indicative, given the limitations related mainly to data availability for international services transactions and for barriers restricting trade in services. Data on international services transactions Services sector statistics are subject to many weaknesses owing to gaps in coverage, limited sectoral disaggregation, lack of concordance and comparability among countries, and non-availability of data according to each mode of supply. The classification of services sectors is generally not detailed in existing statistics, and differs from that used in the GATS. Therefore, the importance of services transactions in world trade is considered to be quite understated, capturing mainly crossborder transactions and to a very limited extent services traded through other modes of supply (Chang et al., 1999). All these statistical deficiencies are reflected in the databases employed in general equilibrium simulations and affect the results. Data on barriers to trade in services The estimates of impediments to trade in services incorporated into the models have major weaknesses which reflect the difficulty of identifying and quantifying barriers to trade in services. Almost all of the first group of studies use Hoekman’s (1995) estimates, which are based on an incomplete catalogue of existing barriers. Furthermore, the methodology employed for their computation does not take into account the actual impact of different barriers on the economy, with minor impediments receiving the same weight as almost complete denial of access. Other studies try to overcome some of these limitations by incorporating improved estimates, thereby improving the quality and credibility of the results. Such estimates draw on more comprehensive qualitative databases of measures affecting trade in services, which are then used to determine the economic impact of services barriers on prices or quantities, taking into account their individual restrictiveness. However, even if such estimates have recently been calculated for a number of services sectors (see below), only estimates for banking and telecommunications services are included in general equilibrium studies. Weighted averages of these estimates have been considered as indicative for other services sectors and incorporated into models, thereby limiting the accuracy of results. Moreover, there is inevitably a degree of subjectivity in seeking, via the frequency index, to allocate degrees of restrictiveness to different barriers. In addition, all studies fail to assess the relative importance of private trade-restrictive actions and anti-competitive practices, which can have significant implications for market outcomes and can explain costs and price differentials. Other measures, which determine price differences on the basis of profit margins, fail to relate specific, existing barriers to the observed price gaps and to account for how these barriers may contribute to higher prices. This reduces the ability to identify which policies may need to be addressed. 124
Modelling framework A general equilibrium model that supports the analysis of services liberalisation should include treatment of each of the four modes of supply. The first group of studies uses the same model structure as that used for the analysis of goods trade, ignoring the additional possible effects of FDI or movement of natural persons on services trade. In the studies of the second group, notable progress has been made in explicitly modelling FDI as a mode of services supply. Modelling protection Simulation results are conditioned not only by the quality of the estimates of barriers, but also by the modelling of their removal. The approach to the modelling of protection has a strong influence on the overall results of liberalisation through income effects (when barriers are modelled as rentcreating measures) or through allocative efficiency effects (when barriers are modelled as cost-raising elements). In all studies, the decision as to how to model these barriers is to some extent arbitrary and often determined by the way in which the estimates of barriers are measured. Because the barriers to services trade appear to be significant, their impact on the simulation results will also be significant. Therefore, as the analysis of results discussed above shows, protection-modelling deficiencies may accentuate existing insufficiencies in the data, thereby biasing the results. Addressing the limitations Recent theoretical work on the effects of FDI restrictions on services and the development of a modelling framework addressing the separate treatment of trade in services through commercial presence, in addition to cross-border trade in services, represents major progress in improving the practical relevance of general equilibrium models. The potential for exploiting the recent modelling structures is strongly conditioned, however, by complete and accurate data on services flows and on barriers to trade in services. Even if the present state of data availability is far from satisfactory, important progress is being made in improving the measurement of services trade and FDI stocks. Also, work on measuring restrictions to services trade that leads to a more accurate assessment of the scope and significance of actual barriers is beginning to emerge. The main results are summarised below. Data on international services transactions Global Trade Analysis Project (GTAP) Although statistics on trade in services are incomplete, major improvements have recently been made in the international statistics based on work by the International Monetary Fund (IMF), Eurostat, the OECD, the WTO, the UN and UNCTAD (OECD, 2000a). On the basis of these improvements, the treatment of services in the GTAP database – the most frequently used database for multi-regional empirical studies – will be upgraded. According to the 2000 GTAP Advisory Board Report (Hertel, 2000; see also McDougall, 2000), the new services trade data set contained in GTAP version 5: x x
Will draw on the new services trade statistics available from the IMF and Eurostat. Will increase the sectoral classification from eight to 15 services industries, and the number of regions for which data are provided to 65. 125
x x
May contain data for the GATS supply mode consumption abroad (mode 2) separate from the modes cross-border supply (mode 1) and temporary entry of natural persons (mode 4). May contain bilateral trade data for “margin services” (air, maritime and other transport services).
Ongoing work by the Australian Productivity Commission The Australian Productivity Commission has constructed a database for FDI stocks and income flows by home region, host region and sector, enabling the separate modelling of services supply through commercial presence (mode 3) (Hanslow et al., 1999). Measuring and modelling barriers to trade in services Considerable effort is being devoted to improving measurement techniques. As qualitative information on services barriers is a key requirement for assessing their restrictiveness, it is important to examine the present state of qualitative databases. A number of lists of services barriers have been drawn up by various international organisations.6 They provide information on policies affecting different services sectors, with details on the regulatory situation, legal instruments, institutional profiles, regulatory responsibilities, licensing regimes, private-sector participation, market status, etc. Such lists, which increasingly used in modelling work, aim to overcome the limitations of information contained in GATS schedules. In this context, the OECD Trade Committee has conducted various sectoral studies which provide indicative lists of measures for the following sectors: Travel and tourism services (OECD, 2000b). Financial information services (OECD, 1998). Retail trade services (OECD, 1999a). Wholesale services (OECD, 1999b). Accountancy services (OECD, 1997). Air cargo services (OECD, 1999c). Environmental services (2000c). The sectoral studies are based on the GATS schedules of commitments, supplemented by information from other sources. They allow for a more in-depth analysis of barriers in the respective sectors and help facilitate the computation of their restrictive impact on the economy. Attempts are made to assess the restrictiveness of combinations of measures which, taken alone, do not have a strong negative impact (OECD, 1999d). In addition to improvements in the qualitative databases on services barriers, significant progress is being made in measuring the restrictiveness and economic impact of impediments to services trade, for example through work at the Australian Productivity Commission and at the World Bank.
126
Details on recently completed and ongoing work at the Productivity Commission are summarised in OECD (2000d).. As that report indicates, the computing method includes, in a first stage, sophisticated weighting methods for assessing the restrictiveness of different measures (see for example, McGuire and Schuele, 2000; McGuire et al., 2000; and Nguyen-Hong, 2000). The classification and assessment of weights take into account information on types of barriers and their likely relative economic impact. This information is derived from the GATS schedules and from various other qualitative studies. To minimise subjectivity, sensitivity tests are conducted to examine the extent of the variation of the computed index in response to alternative weights.7 Various efforts to improve the explanatory power of models and analyses of the accuracy and plausibility of results have also been carried out (Warren, 2000b). On the basis of qualitative and restrictiveness assessments, an evaluation of the extent to which actual policies raise costs of entry and/or operation post-entry is undertaken in the second stage, using econometric techniques, which include all relevant economic determinants. Within the research project conducted by the Australian Productivity Commission, the impacts of services restrictions on price or quantity have been determined for banking services, maritime services, telecommunication services, distribution services and professional services. The studies on distribution and professional services seek to analyse in more detail the effects of restrictions (i.e. rent-creating or cost-raising) and provide useful information for incorporating these restrictions further into general equilibrium models. This method may yield separate estimates for barriers relating to commercial presence and other modes of supply and for barriers affecting market access and restrictions on national treatment. This distinction is operationalised by assuming that market access restrictions apply in a nondiscriminatory way to incumbents in a particular market and to possible entrants (whether domestic or foreign) and that national treatment restrictions mean discrimination between domestic and foreign suppliers. Although these assumptions do not correspond entirely to the GATS categorisation, they facilitate the incorporation of trade barriers in accordance with the category impediments addressed by the GATS. Recognising the widely felt need for research in this area, the World Bank has launched a work programme on trade in services. A central component of the programme is the creation of a database on measures affecting trade in services. The database will also contain information on trade and investment flows, market structure and performance indicators. It will be a source of information on trade-affecting measures and support work on the quantification of barriers to trade in services. Details are presented in OECD (2000e). In terms of completed research, Fink et al. (2000) is recently released analysis of barriers affecting maritime transport costs. It adopts an econometric approach fairly similar to those described above to assess the restrictiveness of measures affecting maritime transport costs. However, in examining the determinants of those costs, the authors draw upon an improved database, created as part of the World Bank’s services project, which contains a comprehensive data set on public policies and private practices. This facilitates the assessment of the relative importance of both public and private trade-restrictive actions in explaining maritime transport costs. The authors find that private anti-competitive practices have a stronger influence on transport costs than public restrictions. A removal of private anti-competitive practices would generate a 38% reduction in transport costs, while the elimination of restrictive governmental practices would result in an 11% price reduction.
127
In addition, research is beginning to emerge on the identification and rigorous assessment of the extent to which various regulatory structures affect services trade and investment. In particular, work is advancing on the structure-conduct-performance methodology for electricity, with plans to extend the research to a broader set of industries (Findlay and Warren, 2000). It is hoped that the application of such a methodology to a sufficiently large sample of economies will allow for a rigorous assessment of the extent to which various regulatory structures affect services trade and investment. This would improve the information on the effect of actual policies and lead to a fuller assessment of the impediments to services trade. This review of the literature suggests that existing economy-wide models have the potential to assess the effects of multilateral services liberalisation in an increasingly realistic manner. While helpful work incorporating the separate analysis of different modes of services delivery into the model is available, such analysis is strongly dependent on the accuracy of the input data used. Efforts to make modelling work of greater practical value to the negotiating community will need to focus on generating the required data on a sector-by-sector and country-by-country basis and on identifying how different policies affect the economic performance of firms.
128
Annex Key Characteristics of Computable General Equilibrium Studies Examined
Publication Base year/data source Country and sector coverage
Brown et al. (1996) x x x
Data on services barriers
x x
Model structure
x
x Details of policy simulations Simulation results (liberalisation with monopolistic competition, including effects of scale and product variety)
x
Reference year for the database model: 1990. 8 countries/regions: United States, Canada, Mexico, Europe, Japan, Asian newly industrialised economies (Asian NIEs), Australia/New Zealand, other major trading nations, ROW. 29 sectors: 1 agricultural, 21 manufacturing, 1 mining, 5 services (construction, wholesale/trade, transportation, financial services, personal services). Estimates of barriers to trade in services from Hoekman (1995). These barriers to trade in services are considered identical to tariffs on trade in goods, raising domestic price over world price by their estimated tariff equivalents and also raising the revenue of the importing country. Michigan Model of World Production and Trade with varying assumptions about: nature of competition (monopolistic and perfect competition); role of product variety (effects of product variety considered and excluded); role of scale effects (considered and excluded) in the services sectors. Assumptions: full employment, balanced trade, rents and tariff revenues accrue to the importing country, fixed relative wages, fixed labour supply. Services liberalisation: (hypothetical) 25% multilateral reduction in the ad valorem tariff equivalents of services barriers under various assumptions for variety, scale and competition. Welfare gains USD millions
Percentage change in GDP
Terms of trade Percentage change
United States 25 814 0.5 0.2 Canada 3 979 0.7 -0.1 Mexico 4 966 2.1 -0.2 Europe 28 718 0.4 0.1 Japan 12 500 0.4 -0.3 Asian NIEs 4 909 1.0 0.1 Australia/New Zealand 3 846 1.9 -0.4 Other NIEs 8 353 0.7 -0.3 x Results do not change significantly when modelling the liberalisation scenario under various assumptions for services sectors. x All countries gain in aggregate from the assumed liberalisation of trade in services in all scenarios. The maximum increase in welfare is 2% of GDP. Total trade increases for all countries, but the terms of trade move by small amounts in favour of some and against others. x Liberalisation is welfare-enhancing for all countries.
129
Publication Base year/data source Country and sector coverage
Data on services barriers Model structure Details of policy simulations Simulation results
Chadha (2000) x x
GTAP database version 4; reference year: 1995. 7 regions: India, rest of South Asia (Bangladesh, Buthan, Maldives, Nepal, Pakistan, Sri Lanka), ASEAN-4 (Indonesia, Malaysia, the Philippines, Thailand), NIEs [Hong Kong (China), Singapore, Korea, Chinese Taipei], European Union, Japan, United States, ROW. x 25 sectors: 1 agricultural, 16 manufacturing, 8 services. x Hoekman’s (1995) estimates for services barriers. x Michigan Model of World Production and Trade with imperfect competition, increasing returns to scale and product heterogeneity. x 3 simulation scenarios: bilateral reduction of 25% in average import tariff rates of 1995 on agriculture, mining and manufactured goods sectors, first simulation plus 25% reduction in the ad valorem tariff equivalents for services sectors. Welfare effects from liberalising Welfare effects from liberalising trade in services trade in goods and services Percentage Percentage USD millions USD millions change change India 0.7 2 432 1.4 4 759 Rest of South 0.9 924 3.0 3 153 Asia ASEAN-4 1.8 9 442 2.7 13 869 NIES 1.7 15 250 2.5 22 343 United States 1.0 72 197 1.2 83 713 Japan 0.8 41 820 1.2 59 578 EU 0.8 65 829 1.0 78 919 Liberalisation of services trade leads to relatively large gains in economic welfare, as well as in trade in the services sectors. In percentage terms, developing countries gain more than developed countries. Gains for developed countries are relatively higher under services trade liberalisation than under a scenario of goods trade liberalisation alone.
130
Publication
Chadha et al. (2000)
Base year/data source
x x
Country and sector coverage Data on services barriers Model structure
x x x
Details of policy simulations
Simulation results
GTAP data base version 4; reference year: 1995. Extrapolated 2005 GTAP database incorporating implementation of the Uruguay Round negotiations. 20 countries/regions: 5 developed and 15 developing. 16 sectors of which 5 services. Hoekman’s (1995) estimates for services barriers.
x
Michigan Model of World Production and Trade with imperfect competition, increasing returns to scale and product heterogeneity. x Projections of the database from 1995 to 2005. x Implementation of the Uruguay Round in 2005. x Following scenarios related to the new WTO negotiating round are considered in terms of the adjusted projected 2005 database: o Reduction of bilateral import tariffs on agriculture by 33%. o Reduction of bilateral import tariffs on minerals and manufactured products by 33%. o Reduction of bilateral import tariff equivalents on services sectors by 33%. o Combination of all these scenarios. x Similar unilateral liberalisation scenarios for India. Welfare gains – services Welfare gains – overall liberalisation liberalisation Percentage USD Percentage USD change millions change millions 2.0 550 780 2.4 656 028 Developed economies Australia and New Zealand 2.9 15 115 3.4 17 320 Canada 2.8 20 226 3.1 22 648 EU and EFTA 1.9 210 118 2.3 253 381 Japan 1.9 125 563 2.5 165 669 United States 2.0 179 758 2.2 197 009 2.5 137 085 3.2 179 748 Developing economies India 1.6 6 840 2.7 11 412 India (unilateral liberalisation scenario) 1.3 5 350 2.0 8 376 Sri Lanka 2.8 476 6.0 1 010 Rest of South Asia 1.9 2 186 4.4 5 065 China 1.3 11 812 2.2 20 193 Hong Kong (China) 8.3 10 540 9.6 12 277 Korea 2.7 15 527 3.9 22 018 Singapore 3.8 2 827 5.2 3 897 Indonesia 2.4 6 169 3.1 7 859 Malaysia 1.9 2 301 3.8 4 598 Philippines 3.5 3 065 7.0 6 153 Thailand 4.2 8 733 5.2 10 708 Mexico 3.0 10 699 3.4 12 081 Turkey 3.5 7 390 4.5 9 432 Central European Associates 2.5 9 132 3.1 11 363 Central, South America 2.3 39 388 2.4 41 682 2.1 687 865 2.5 835 776 World total The developed countries’ gain of USD 550.8 billion is 80% of the USD 687.9 billion total services gain; relative gains for the developing countries are nonetheless quite large.
131
Publication
Hertel et al. (1999)
Base year/data source
x x
Country and sector coverage Data on services barriers
x x x x
Model structure
x
Details of policy simulations
x x
Simulation results
GTAP data base version 4; base year: 1995. Projections of the world economy in 2005 when the Uruguay Round is fully implemented. The extrapolated database contains: projections for population, skilled and unskilled labour, investment and capital stock, combined with likely productivity growth rates, predictions about level and composition of GDP in 2005, trade flows, input usage. 19 regional groupings. 22 aggregate sectors. Tariff equivalents for business and financial services and for construction services are taken from Francois and Hoekman (1999). Tariff equivalents for trade and transport services and for government services are taken from Hoekman (1995). GTAP model: a comparative static multi-region model, with perfectly competitive markets and constant returns to scale. Welfare gains arise principally from the reallocation of resources within the economy and the resulting gain in allocative efficiency. In addition, terms of trade effects contribute to welfare results. Liberalisation scenarios: 40% cut in the estimated 2005 level of protection for agriculture, manufactures and services (performed separately and combined). Barriers to trade in services are modelled as cost-raising components for firms attempting to access the market. Liberalisation scenarios Welfare effects (USD millions)
Liberalisation of agriculture (reduction in market price support) Liberalisation of agriculture (reduction in market price support and subsidies) Manufactures and extraction tariffs
58 621 69 319 69 564
Liberalisation of business, finance and 21 604 construction services Liberalisation of trade and transport and 332 565 government services Gains from liberalising services sectors represent 72% of the total USD 493 billion resulting from liberalisation. Liberalisation of trade in services also has powerful impacts on agriculture through inter-sectoral linkages in each economy.
132
Publication Base year/data source Country and sector coverage Data on services barriers Model structure
Details of policy simulations
Simulation results
Benjamin and Diao (2000) x x x x
GTAP version 3; reference year: 1992. 13 regions. 11 sectors (single services sector). Barriers to services modelled as variables which make possible the existence of pricediscriminating opportunities and increased fixed cost; no specific estimates used. x All sectors of production perfectly competitive, except the private services sector in which firms have price-discrimination opportunities across national markets. This makes it possible to capture the cost-raising nature of non-tariff barriers to trade in services, which prevent domestic consumers from cross-border arbitraging. x Removal of bilateral tariffs on goods and self-restricted export taxes on textiles. x Liberalisation of services markets by assuming that oligopoly firms in the services sector switch from their initial price-discriminating strategy to a single pricing behaviour within APEC (market segmentation is removed), and fixed costs of firms operating in the services sector of APEC members are reduced by 10%. x Liberalisation of trade in goods and services. Welfare changes as percentage deviation from the base year Liberalisation of Liberalisation of Liberalisation of trade in goods trade in services trade in goods and services United States 0.91 1.71 2.65 Canada 1.33 1.35 2.71 Mexico/Chile 1.41 0.50 1.87 Australia/New Zealand 2.95 1.10 4.07 Japan 2.89 1.71 4.66 Korea 8.24 1.13 9.47 Singapore 15.24 1.59 16.90 Hong Kong (China) 4.92 0.98 5.99 Chinese Taipei 10.11 0.61 10.80 China 7.42 0.02 7.33 Rest of Southeast Asia 6.85 0.91 7.72 EU 0.19 1.12 1.31 ROW 0.51 0.86 1.38 The large developed economies gain relatively more from services liberalisation than from tariff removal. In the case of tariff removal, the greatest welfare gains are obtained by the economies starting out with the highest levels of tariff protection, namely the developing and newly industrialised economies of East and Southeast Asia.
133
Publication Base year/data source Country and sector coverage Data on services barriers Model structure
Details of policy simulations Simulation results (detailed results are reported for the scenario of halving protection for trade in services)
Australian Department of Foreign Affairs and Trade (1999) x GTAP data base version 4; base year: 1995. x GTAP: 45 regions and 50 sectors. x APG-cubed: 18 regions and 6 sectors. x Modification of Hoekman’s (1995) estimates of services trade barriers. Simulations carried out by the Centre for International Economics using: x The GTAP framework: comparative static CGE, perfectly competitive markets, constant returns to scale. x Asia-Pacific G-cubed framework: dynamic global economy-wide model, with considerable macroeconomic detail, incorporating both real estate sector and financial sector interactions. It accounts: o For the effects of liberalisation on interest rates, exchange rates and international capital movements. o For the effects of different fiscal and monetary responses to liberalisation. o For the adjustment costs associated with the reallocation of labour and capital when trade barriers fall. x Two scenarios: 50% and 100% reduction of distortions in the protection level of services. GTAP results Region
Welfare gains (USD billions) Africa 4.7 ASEAN 4.6 Eastern Europe 7.7 European Union 73.4 Latin America 12.6 North America 63.4 Australia 3.5 China 9.8 India 2.6 Japan 43.5 Korea 3.9 New Zealand 0.7 Other OECD 5.4 Other 15.5 Total 251.2 Total when eliminating barriers 500 Global gains from halving goods and services barriers: Over USD 400 billion from all sectors: x USD 250 billion (62%) from services. x USD 66 billion (16%) from manufactures. x USD 90 billion (22%) from agriculture. Global gains from eliminating all barriers: USD 750 billion. x APG-cubed model results: Removing protection over the five years from 2004 results in additional welfare gains of more than USD 600 billion in 2008. Liberalisation over a ten-year frame delays these gains until approximately 2013. x
Services could generate substantial gains for both developed and developing countries. The biggest winners in dollar terms would be the biggest economies. However, all economies stand to gain and, in proportion to the size of their GDP, some smaller economies stand to benefit substantially.
134
Publication Base year/data source Country and sector coverage
Data on services barriers Model structure
Details of policy simulations
Simulation results
Robinson et al. (1999) GTAP data base version 4; base year: 1995. 10 countries/regions: the United States, EU-15, Japan, other OECD, Asian NIEs, China, ASEAN, South Asia, Latin America, ROW. x 11 sectors of which 6 are services: utilities, construction, trade and transport, private services, public services, housing. Others are agriculture, processed food, products based on natural resources, non-durable consumer goods, intermediate and durable manufactures. x Tariff equivalents for protection in services trade are taken from Brown et al. (1996) and work by Hoekman (1995). x Standard, static multi-region, CGE model, with perfect competition and constant returns to scale. x Assumptions: five primary factors: agricultural land, natural resources, capital, unskilled labour, skilled labour; capital and labour are mobile across sectors but immobile between regions. x Dynamic considerations: the model contains links between trade performance and TFP as an import-embodied technology transfer across regions via trade in intermediate inputs; the regions’ TFP growth is linked with its imports of capital- and technology-intensive products. x Several scenarios which involve: x Technology change in the international transport sector (50% cut in the cost of international trade). x Non-services sector liberalisation (50% cut in food, agriculture and natural resources and 100% cut in manufactures). x Services sector liberalisation (50% cut in all services sectors). x A combination of these. For each of these scenarios, import-embodied technology transfers are assumed via imported durable products and via imported private services and trade and transport as intermediate inputs Welfare gains Welfare gains with technology Percentage from base GDP transfer Percentage from base GDP ManufacServices Total ManufacServices Total tures liberalisa- liberalisatures liberalisa- liberalisaliberalisation tion liberalisation tion tion tion United States 0.03 0.84 1.11 0.02 2.22 2.48 x x
EU-15 0.23 1.22 1.66 0.28 4.10 4.68 Japan 0.78 0.43 1.44 1.05 2.45 3.88 Other OECD -0.30 1.39 1.75 -0.33 2.81 3.17 Asian NIEs 2.06 1.85 4.96 2.45 3.97 7.87 China 0.26 0.34 1.70 0.33 1.55 3.14 ASEAN -0.28 1.29 2.32 -0.25 1.83 2.97 South Asia 0.03 1.13 2.02 -0.06 2.26 3.14 Latin America -0.34 0.98 1.04 -0.42 2.48 2.45 ROW -0.50 2.12 2.22 -0.55 3.57 3.63 World total 0.20 1.05 1.62 0.27 2.99 3.71 x Welfare gains worldwide from a 50% cut in protection in the services sector are five times larger than those from non-services sector trade liberalisation. When import-embodied technology transfers are considered, welfare increases substantially. The growth in TFP induced by liberalisation of services imports is potentially very significant, especially for developing countries. x In spite of potentially significant gains from services liberalisation, there are also potential trade-offs between gains from liberalising trade in services and manufacturing that should provide a basis for negotiations between developing and developed countries: developed countries gain relatively more from increasing exports of services to developing countries, while developing countries potentially gain more from increased access to developed country markets from their exports of manufactures.
135
Publication Base year/data source Country and sector coverage Data on services barriers Model structure Details of policy simulations
Simulation results
McKibbin and Wilcoxen (1996) x Data presumably also cover FDI. x 8 regions. x 12 sectors (1 services sector). x No specific estimates for barriers for trade in services, only exogenously specified changes in TFP are considered. x G-cubed multi-country dynamic model. The following scenarios are considered: x Increases in total factor productivity in the Australian services sector. x Increases in TFP across all services industries globally. The simulated productivity gains may be considered a plausible effect of trade liberalisation in services. They do not quantify the impact of a reduction of barriers to trade in services, but might be interpreted as an important additional factor to consider when evaluating the possible consequences of trade reform in services sectors. As a result of the increase in TFP in services sectors in Australia, the gain in GDP is approximately AUD 16 billion (in 1990 AUD) in 2001. Along with services, which represent a little over 50% of the Australian economy, output and income also rise throughout the economy. The increase in the return to capital in the services sectors determines: x An inflow of foreign financial capital into physical investment in the Australian services sector. x A rise in the return to other domestic assets such as bonds, as funds are relocated to physical investment in the services sector. x An inflow of capital which leads to an appreciation of the exchange rate and worsens the overall trade balance and current account balance. After productivity returns to its baseline level, the trade and current account balances begin to improve. The second scenario assumes a global rise in services productivity, in which case results are similar to those of the first scenario. Only trade balance effects are different because of smaller capital inflows, as there are also opportunities for investment in foreign companies.
136
Publication Base year/data source Country and sector coverage Data on services barriers
Model structure Details of policy simulations
Simulation results
Kawai and Urata (1998) x x x x
162 x 162 input-output table for 1993. One country: Japan. 162 sectors. As services are considered non-tradable, there are no barriers to trade in services; instead it is assumed that regulations in services markets limit competition and generate high prices and low productivity. The restrictiveness of regulations is determined by comparing Japan’s services sector to that of the United States with respect to prices and TFP. For services, TFP is almost always lower in Japan and prices are often more than double prices in the United States. x Single country model for Japan. x The model determines regulatory costs by estimating how improvements in TFP in services industries affect prices, production, and employment throughout the economy. Deregulation in services industries is simulated by reducing the TFP gap between Japan and the United States by 50%. Deregulation Free trade in Deregulation in services tradable and free trade sectors GDP (%) 4.1 2.7 7.5 Investment (%) 13.3 3.2 18.2 Consumption (%) 4.2 2.8 7.5 Welfare per capita (JPY millions) 269 864 196 408 562 181 Services have a major economic impact on all economic sectors, so that improvements in services productivity would significantly improve overall welfare. The study only provides an analysis of what would happen in the event of a broad productivity shock, it does not explore how such a productivity shock might occur.
137
Publication Base year/ data source
Country and sector coverage
Data on services barriers
Model structure Details of policy simulation
Simulation results
Dee and Hanslow (2000a) x
Updated version of the GTAP database version 4 following a simulation which implements the Uruguay Round under an imperfect capital mobility scenario provided by Verikios and Hanslow (1999). x Data on foreign investment were provided by APEC and the United Nations and were processed to provide a full bilateral matrix of FDI. x 18 countries: Australia, Canada, Chile, China, EU, Hong Kong (China), Indonesia, Japan, Korea, Malaysia, Mexico, New Zealand, Philippine, Singapore, Chinese Taipei, Thailand, United States, Rest of Cairns, ROW. x 3 sectors: agriculture, manufactures, services. x Estimates for barriers to trade in services are taken from Kalirajan et al. (2000), who estimated the impact of barriers to trade in banking services on prices, and from Warren (2000b), who estimated the impact of barriers to trade in telecommunication services on quantities of telecommunication services delivered which he subsequently transformed into price impacts. x Barriers to commercial presence are distinguished from those affecting other modes of supply, and non-discriminatory barriers to market access from discriminatory restrictions on national treatment. x Barriers take the form of tax equivalents that generate rents (a mark-up of price over cost). The rents on output have been modelled as accruing to the selling region, and those on capital as accruing to the region of ownership, after subtracting the general property income tax after which is collected by the government in the region of location. Thus, only the portion of rents associated with the general property income tax is assumed to remain in the region of location, while the biggest amount of rents accrues to the region of ownership. x FTAP model: version of the GTAP model with capital mobility and FDI, increasing returns to scale and large-group monopolistic competition in all sectors. Multilateral liberalisation scenarios: x Elimination of tariffs for agricultural and manufactured products. x Elimination of estimates for services barriers. x Combination of both. Partial liberalisation scenario for services: x Removal of barriers to establishment alone. x Removal of barriers to ongoing operation alone. x Removal of restrictions on market access alone. x Removal of restrictions on national treatment alone. GDP (%) Welfare (USD millions) From services Total From services Total liberalisation liberalisation liberalisation liberalisation Australia 0.0 0.2 2 098 4 092 New Zealand -0.1 1.1 257 4 657 Japan 0.0 0.3 4 130 25 094 Korea 0.1 1.6 1 886 10 670 Indonesia 5.1 5.9 2 470 3 921 Malaysia 0.7 4.5 1 015 4 547 Philippines 0.4 5.5 1 236 2 837 Singapore -1.3 -1.5 -247 7 174 Thailand 0.2 2.8 1 698 5 762 China 14.6 18.0 90 869 104 957 Hong Kong (China) 1.0 0.9 5 896 6 812 Chinese Taipei 0.2 3.0 -142 11 517 Canada -0.1 0.0 -499 -1 038 United States -0.1 0.1 -1 809 20 925 Mexico 0.1 0.4 357 274 Chile 0.4 1.1 330 375 Rest of Cairns 0.1 1.3 6 970 19 736 EU 0.0 0.1 -6 169 225 ROW 0.8 0.9 23 039 34 363 World 133 386 266 901
138
Publication Base year/data source Country and sector coverage Data on services barriers Model structure Details of policy simulations Simulation results
Verikios and Zhang (2000) x x x x
Same database as Dee and Hanslow (2000a), but greater sectoral detail provided. 19 regions. 8 sectors, of which 6 services sectors. Estimates for barriers to trade in services taken from Kalirajan et al. (2000) for banking services, and from Warren (2000b) for telecommunication services. x FTAP model (see Dee and Hanslow). x Elimination of barriers to trade in communication services and in financial services, in a postUruguay Round environment. Projected effects of liberalising trade Projected effects of liberalising in communication services trade in financial services Real GDP Real Equivalent Real GDP Real Equivalent (%) income variation (%) income variation (%) (USD (%) (USD millions) millions) Australia 0.00 0.04 112 0.05 0.04 130 New Zealand -0.01 0.03 16 0.03 0.02 10 Japan 0.03 0.04 1 612 0.00 0.02 1 025 Korea 0.02 0.02 89 0.05 0.20 801 Indonesia -0.39 0.70 1 232 -1.22 -0.02 -28 Malaysia 0.02 -0.02 -19 0.14 0.71 581 Philippines 0.12 0.69 436 0.08 -0.16 -98 Singapore -0.14 -0.52 -313 0.76 4.22 2 553 Thailand 1.10 0.32 452 0.03 0.23 334 China 0.46 0.75 4 488 0.17 0.34 2 024 Hong Kong (China) 1.10 -0.87 -880 0.10 0.01 8 Chinese Taipei 0.05 0.07 189 0.13 0.23 601 Canada 0.00 0.01 39 0.02 -0.01 -44 United States 0.00 -0.02 -1 065 0.00 -0.05 -3 245 Mexico 0.01 -0.02 -59 0.01 0.04 89 Chile 0.01 -0.02 -10 0.22 0.43 241 Rest of Cairns -0.01 0.02 215 0.04 0.03 283 EU 0.01 -0.01 -1 008 0.01 -0.05 -3 544 ROW 0.17 0.24 7 133 0.04 0.06 1 733 World 0.05 12 658 0.01 3 454
139
Publication Base year/ data source Country and sector coverage
Data on services barriers
Brown and Stern (1999) x x x x x x
Model structure
x
Details of policy simulation Simulation results
x x x x x
GTAP version 4; base year: 1995. Data on foreign investment provided by the Australian Productivity Commission. 18 countries: Australia, Canada, Chile, China, EU, Hong Kong (China), Indonesia, Japan, Korea, Malaysia, Mexico, New Zealand, Philippine, Singapore, Chinese Taipei, Thailand, United States, rest of Cairns + ROW. 3 sectors: agriculture, manufactures, services. Estimates for barriers to FDI are taken from Hoekman (1999); estimates of financial data on gross operating margins calculated by sector and country. Barriers take the form of increased fixed cost of locating in a host country or the form of a tax on installed capital. Latest version of the Michigan Model of World Production and Trade which incorporates relationships for cross-border services trade and FDI in services sectors. Liberalisation of trade in services: average fixed cost reduced by the margin estimated by Hoekman according to seven scenarios. Welfare effects are large under all parameter specifications. Capital formation plays the most important role in determining welfare effects. Welfare gains and increased economy-wide output in countries which attract physical capital. The less mobile the capital, the smaller the trade and welfare effects of liberalisation.
Source: Author.
140
NOTES 1.
The GTAP database contains detailed bilateral trade, transport and protection data characterising economic linkages among regions, linked to individual country input-output databases which account for intersectoral linkages among the 50 sectors in each of 45 regions. The project is co-ordinated by the Center for Global Trade Analysis which is housed at the Department of Agricultural Economics, Purdue University. The OECD is a member institution of the GTAP Consortium.
2.
For example, McGuire (1998) found that only 38 out of the total of 165 measures affecting financial services in Australia were listed in its GATS schedules.
3.
Hong Kong (China) and Singapore are considered to be the free-trade benchmark.
4.
This study discusses the treatment of FDI in general, not the analysis of services traded via commercial presence. It is described here together with the studies that look at TFP, as the focus of the study is the analysis of productivity growth in services sectors.
5.
For example, in cases where rents are tied to saleable assets (such as land, business licence) the value of rents may be capitalised into the selling price of these assets. In this case, the owner of FDI who purchases these assets, which are necessary for supplying the services, after the restrictions are first imposed, will not receive the rents; rather the owner of these assets will capture the rents through the selling price. See Kalirajan (2000).
6.
WTO Trade Policy Reviews, the World Bank’s Trade in Services Database, the EU Market Access Sectoral and Trade Barriers Database, the National Trade Estimate Reports on Foreign Trade Barriers from the Office of the United States Trade Representative, APEC Individual Action Plans, TradePort Database, ITU World Telecommunication Indicators Database, World Telecommunications Development Report, WTO Sectoral Reports.
7.
Hardin and Holmes (1997) conducted a sensitivity analysis to determine the extent of variation in the FDI restrictiveness index in response to alternative weights across 15 services sectors. Comparing the base case index values with those obtained under scenarios using different sets of weights reveals which types of FDI restrictions countries use more widely for particular services sectors. Furthermore, the authors find that, while varying the weights for the different types of restrictions has a significant impact on the absolute values of FDI restrictiveness indices, the relative index values between countries and sectors remain largely unchanged.
141
REFERENCES
Australian Department of Foreign Affairs and Trade (1999), Global Trade Reform: Maintaining Momentum, Canberra. Benjamin, N. and X. Diao (1998), “Liberalising Services Trade in APEC: A General Equilibrium Analysis with Imperfect Competition”, in The Economic Implications of Liberalising APEC Tariff and Nontariff Barriers to Trade, US International Trade Commission, Washington, DC. Benjamin, N. and X. Diao (2000), “Liberalising Services Trade in APEC: A General Equilibrium Analysis with Imperfect Competition”, Pacific Economic Review 5:1, pp. 49-75. Brown, D. and R. Stern (1999), “Measurement and Modelling of the Economic Effects of Trade and Investment Barriers in Services”, paper prepared for the Coalition of Services Industries (CSI) World Services Congress Atlanta, Georgia, 1-3 November. Brown, D., A. Deardorff and R. Stern (1996), “Modelling Multilateral Liberalisation in Services”, Asia-Pacific Economic Review 2, pp. 21-34. Chadha, R.(2000), “GATS and the Developing Countries: A Case Study of India”, forthcoming in Robert M. Stern (ed.), Services in the International Economy: Measurement and Modeling, Sectoral and Country Studies, and Issues in the WTO Services Negotiations, University of Michigan Press, Ann Arbor, Michigan. Chadha, R., D. Brown, A. Deardorff and R. Stern (2000), “Computational Analysis of the Impact on India of the Uruguay Round and the Forthcoming WTO Trade Negotiations”, Discussion Paper No. 459, School of Public Policy, University of Michigan. Ann Arbor, Michigan. Chang, P., G. Karsenty, A. Mattoo and J. Richterling (1999), “GATS, the Modes of Supply and Statistics on Trade in Services”, Journal of World Trade 33(3), pp. 93-115. Dee, P. and K. Hanslow (2000a), “Multilateral Liberalisation of Services Trade”, Productivity Commission Staff Research Paper, Ausinfo, Canberra. Dee, P. and K. Hanslow (2000b), “Modelling the Liberalisation of Services”, conference paper, Productivity Commission and Australian National University Joint Conference, Australian National University, Canberra, 26-27 June. Findlay, C. and T. Warren (2000), “‘Sins of Commission and Omission’: Measuring Regulatory Impediments to Trade in Services”, paper presented at the Productivity Commission and Australian National University (Joint Conference) on Achieving Better Regulation of Services, Australian National University, Canberra, 26-27 June.
142
Fink, K., A. Mattoo and I. Neagu (2000), “Trade in International Maritime Services: How Much Does Policy Matter?”, World Bank. Francois, J. and B. Hoekman (1999), “Market Access in the Service Sectors”, Tinbergen Institute, unpublished. Hanslow, K., T. Phamduc and G. Verikios (1999), “The Structure of the FTAP Model”, Staff Research Memorandum. Hardin, A. and L. Holmes (1996), “Services Trade and Foreign Direct Investment”, Staff Research Paper, Industry Commission, Australian Government Publishing Services, Canberra. Hertel, T. (2000), “The Global Trade Analysis Project: Issues and Future Directions”, background paper for the GTAP Advisory Board Meeting, Purdue University, 12-14 April. Hertel, T., J. Francois and W. Martin (1999), “Agriculture and Non-agricultural Liberalisation in the Millennium Round”, paper presented at the Global Conference on Agriculture and the New Trade Agenda from a Development Perspective: Interests and Options in the WTO 2000 Negotiations, sponsored by the World Bank and WTO, Geneva, 1-2 October. Hoekman, B. (1995), “Assessing the General Agreement on Trade in Services”, World Bank Discussion Paper No. 307, The World Bank, Washington, D.C. Johnson, M., T. Gregan, P. Belin and G. Gentle (2000), “Modelling the Impact of Regulatory Reform”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York (forthcoming). Kalirajan, K. (2000), “Restrictions on Trade in Distribution Services”, Productivity Commission Staff Research Paper, Ausinfo, Canberra. Kalirajan, K., G. McGuire, D. Nguyen-Hong and M. Schuele (2000), “The Price Impact of Restrictions on Banking Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York (forthcoming). Kawai, H. and S. Urata (1998), “The Cost of Regulation in the Japanese Service Industry”, in The Economic Implications of Liberalising APEC Tariff and Nontariff Barriers to Trade, US International Trade Commission, Washington, DC. McDougall, R. (2000), “International Services Trade Data for CGE Modellers, or, Entropy Methods for Data Reduction”, paper presented at the Third Annual Conference on Global Economic Analysis, Melbourne, Australia, June 27-30. McGuire, G. (1998), “Australia’s Restrictions on Trade in Financial Services”, Staff Research Paper, Productivity Commission, Canberra. McGuire, G. and M. Schuele (2000), “Restrictiveness of International Trade in Banking Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications. Sydney, Routledge (forthcoming).
143
McGuire, G., M. Schuele and T. Smith (2000), “Restrictiveness of International Trade in Maritime Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications. Routledge, London and New York (forthcoming). McKibbin, W. and P. Wilcoxen (1996), “The Role of Services in Modelling the Global Economy”, Asia-Pacific Economic Review 2, pp. 2-13. Nguyen-Hong, D. (2000) “Restrictions on Trade in Professional Services”, Productivity Commission Staff Research Paper, Ausinfo, Canberra. OECD (1997), “Assessing Barriers to Trade in Services: A Pilot Study on Accountancy Services”, TD/TC/WP(97)26, internal working document. OECD (1998), “Assessing Barriers to Trade in Services: Financial Information”, internal working document. OECD (1999a), “Retail Trade Services”, internal working document. OECD (1999b), Wholesale Trade Services”, TD/TC/WP(99)18/FINAL OECD (1999c), “Assessing Barriers to Trade in Services: Air Cargo Services”, TD/TC/WP(99)57/FINAL. OECD (1999d), “Assessing Barriers to Trade in Services: Cross-cutting (“Formula”) Approaches to Multilateral Trade in Services”, TD/TC/WP(99)42/FINAL. OECD (2000a), “Manual on Statistics of International Trade in Services”, internal working document. OECD (2000b), “Assessing Barriers to Trade in Services”, internal working document. OECD (2000c), “Environmental Goods and Services”, internal working document. OECD (2000d), “Measuring and Modelling Restrictions on Trade in Services”, internal working document. OECD (2000e), “An Overview of the World Bank’s Trade in Services Project”, internal working document. Petri, P. A. (1997), “Foreign Direct Investment in a Computable General Equilibrium Framework”, paper prepared for the conference, Making APEC Work: Economic Challenges and Policy Alternatives, March 13-14, Keio University, Tokyo. Australian Productivity Commission (1999), “Trade and Assistance Review 1998-99”, Annual Report Series 1998-99, AusInfo, Canberra. Available at: www.pc.gov.au/research/annrpt/tar9899/index.html Robinson, S., Z. Wang and W. Martin (1999), “Capturing the Implications of Services Trade Liberalisation”, paper presented at the Second Annual Conference on Global Economic Analysis, Ebberuk, Denmark, June 20-22.
144
Tamms, V. (2000), “Frontier Cost Estimates of the Impact of Restrictions on Trade in Air Transport Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York (forthcoming). Verikios, G. and K. Hanslow (1999), “Modelling the Effects of Implementing the Uruguay Round: A Comparison Using the GTAP Model under Alternative Treatments of International Capital Mobility”, presented at Second Annual Conference on Global Economic Analysis, Ebberuk, Denmark, June 20-22. Verikios, G. and X. Zhang (2000), “Sectoral Impact of Liberalising Trade in Services”, paper presented at the Third Conference on Global Economic Analysis, Melbourne, 27-30 June. Available at: www.monash.edu.au/policy/conf/53Verikios.pdf Warren, T. (2000a), “The Identification of Impediments to Trade and Investment in Telecommunications Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York (forthcoming). Warren, T. (2000b), “The Impact on Output of Impediments to Trade and Investment in Telecommunications Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York (forthcoming).
145
Chapter 6 QUANTIFICATION OF THE COSTS TO NATIONAL WELFARE OF BARRIERS TO TRADE IN SERVICES: A SCOPING PAPER
by Nora Dihel OECD Trade Directorate
Abstract. This chapter undertakes a detailed analysis of estimates of services barriers employed in general equilibrium models to indicate the progress made in methods for estimating services barriers and the remaining limitations. It summarises potential research areas related to services barriers and the modelling framework, with a view to identifying possible next steps to improve the accuracy of the assessment of barriers and therefore the relevance of modelling results for the negotiating community.
Introduction Chapter 5 reviews current work on measuring and modelling barriers to trade in services. It outlines the main empirical findings concerning the welfare effects of services liberalisation derived from the computable general equilibrium (CGE) literature and presents the underlying modelling framework, the estimates of barriers employed and the ways of incorporating these estimates into different models. The literature overview showed that liberalisation of services trade generates overall welfare gains under all modelling assumptions. In general, economies with initially high levels of protection tend to gain most (in terms of percentage gains to GDP). As the values of estimated services trade barriers were higher for developing countries than for developed countries, the developing economies are potentially major winners. Chapter 5 presented the advantages, as well as the limitations, of using numerical general equilibrium techniques for quantifying the effects of liberalising trade in services. It showed that general equilibrium modelling is the most appropriate tool for conducting quantitative analyses on multilateral services liberalisation as it is the only technique that allows for an economywide assessment of services barriers. This type of modelling can help to assess the impact of trade policy changes on the economy as a whole and may provide useful insights into broad negotiating modalities. Chapter 5 also showed that the magnitude of welfare effects is strongly dependent on the accuracy of the input data employed (estimates of services barriers) and on the various modelling assumptions, e.g. explicit modelling of foreign direct investment (FDI) versus standard models used to simulate goods trade liberalisation, imperfect competition and increasing returns to scale versus perfect competition and constant returns to scale. The estimation of actual services restrictions represents arguably the most critical area. Therefore, this chapter undertakes a more detailed analysis of estimates of services barriers employed in general equilibrium models to indicate the progress made in methods for estimating services barriers and the remaining limitations. The present chapter suggests that the estimates of services barriers vary on the basis of the data sources and estimation techniques employed. The wide range in the estimated initial extent of policy interventions is reflected in the differing welfare effects from services liberalisation generated by general equilibrium models. The studies that use Hoekman’s “guesstimates” for initial interventions generally report large welfare gains from services trade liberalisation. In contrast, studies that employ the estimates determined on the basis of measures affecting price or quantity tend to generate lower, though still sizeable welfare gains. The quality of estimates of barriers has been improving in terms both of the range of barriers addressed, which are collected from a growing number of sources, and of the measurement techniques employed, which now make it possible to estimate the price and cost effects of services barriers and to begin to determine the correlation between these effects and the individual underlying restrictions.
148
However, it is difficult to determine whether these estimates are realistic, as a number of limitations remain. They are related mainly to the availability of data on existing restrictions and on the effect of these restrictions on economic performance. Major shortcomings are also associated with the weighting assigned to different categories of restrictions. The discussion then turns to the modelling framework to focus on the efforts being made to address major shortcomings and describes the recent improvements to models employed for policy simulations. However, the remaining limitations related to the modelling framework as well as to data availability on barriers suggest that the welfare results generated by the model should be taken as indicative and interpreted with care. The identified limitations of empirical assessments are used to identify the scope for possible further research. The final section of this chapter summarises potential research areas related to services barriers and the modelling framework, with a view to identifying possible next steps to improve the accuracy of the assessment of barriers and therefore the relevance of modelling results for the negotiating community. The analysis of ongoing quantification work suggests that following areas deserve priority attention: x
Data on barriers. Efforts to obtain better estimates of the sizes of the policy interventions with a focus on: Generating comprehensive lists of existing restrictions. Addressing weighting methods. Addressing shortcomings in the price-cost margin approach to studying the effectiveness of protection and the differentiation among restrictions in terms of their economic impact (cost-increasing, rent-creating and/or dual effect).
x
Modelling requirements: Improving the accuracy of models by using flexible functional forms and realistic allocation schemes, and by getting information on the appropriate magnitude of at least some of the parameters and elasticities employed. Exploring dynamic linkages from services liberalisation, in the longer term. Assessing results from a range of models (not only GTAP-based models).
Estimates of services barriers The following discussion first presents the estimates of services barriers that have already been employed in general equilibrium models. Next, new estimates to be incorporated in future general equilibrium models are then noted. Estimates of services barriers employed in general equilibrium models Four sets of estimates have been used as data in general equilibrium models analysing services liberalisation: x
“Guesstimates” calculated by Hoekman (1995). 149
x
Tariff equivalents (financial-based) estimated by Francois and Hoekman (1999).
x
Tariff equivalents (quantity-based) estimated by Francois and Hoekman (1999).
x
Tariff equivalents for banking services estimated by Kalirajan et al. (2000) and for telecommunication services estimated by Warren (2000a; 2000b) (as part of the Australian Productivity Commission services project)
Table 6.1 summarises the main characteristics of the estimation methods used to calculate these four sets and indicates the studies which employed them as data-inputs. An analysis of these estimates as well as the methods used to calculate them is presented below. Box 6.1 identifies the main conceptual aspects of the methods employed for quantifying services barriers. Box 6.1. Methods used for estimating barriers to trade in services The measurement of services barriers is based upon research on the measurement of non-tariff barriers affecting goods trade. General methods for measuring the presence and size of non-tariff barriers classified in frequencytype measures, in price- and in quantity-impact measures [Deardorff and Stern (1997) provide a useful starting point for quantifying barriers to trade in services]. However, these methods need to be adapted to capture the special characteristics of services trade. Frequency-type measures Frequency-type measures comprise frequency ratios and import coverage ratios. Based upon inventory lists of observed non-tariff barriers which apply to particular countries, sectors or categories of trade, they indicate the frequency of occurrence of various types of barriers and the coverage of the barriers with respect to trade or production for individual countries and product groups respectively. Similarly, the restrictiveness index for services measures the number and severity of restrictions on trade in services. In estimating the restrictiveness of services barriers, two issues need to be addressed. Elaboration of comprehensive listings of impediments to trade in services on a sectoral basis An essential preliminary step to measuring the impact of impediments to trade in services is to identify and establish an inventory of such impediments. The GATS provided a starting listing of barriers to trade in services. The scheduled commitments were used to determine the frequency of occurrence of restrictions in different services sectors (Hoekman, 1995; Mattoo, 1998; Classens and Glaessner, 1998). However, the information in the GATS schedules is limited by the positive-listing approach. Therefore, in subsequent studies (mainly those undertaken under the auspices of the Australian Productivity Commission*), the GATS schedules were supplemented with other sources of information on relevant legislation on services barriers. Reports by governments and industry associations as well as material produced by the Asia-Pacific Economic Cooperation (APEC), the OECD, the World Trade Organization (WTO) and the United States Trade Representative have been helpful for identifying additional impediments. Determination of the relative restrictiveness of identified impediments The relative restrictiveness of different measure represents in quantitative form (a score) the nature and extent of restrictions in services sectors. The scores employed in earlier quantification studies did not differentiate among restrictions in terms of their likely economic impact (Hoekman, 1995). In later studies, weights were assigned to the different restrictions according to a judgement about their relative economic cost.** Also, separate scores were calculated for restrictions applied to foreign services providers (foreign index) and to domestic services providers (domestic index). Thus, the index became a sophisticated frequency measure for estimating the restrictiveness of an economy’s trading regime for services based on the number and severity of restrictions. However, frequency measures only approximate the relative degree of restrictiveness of trade barriers. They do not provide any information on the economic impact of barriers on prices, production, consumption and international trade or on the consequences of maintaining or eliminating these barriers.
150
(cont)
Therefore, to quantify the economic impact of services barriers, alternative approaches involving price-impact measures (the impact of non-tariff barriers on domestic prices by comparison with world prices) and quantityimpact measures (which compare an estimate of trade volumes in the absence of non-tariff barriers with actual trade volumes) have been applied. Data availability limited the applicability of these traditional price- and quantity-impact measures to services, so that it was necessary to identify alternative benchmarks against which to compare actual prices or quantities. Price-impact measures The basis of alternative benchmarks was the assumption that a market free from impediments to entry will result in prices which equate marginal costs. In contrast, with restrictions, there will be a wedge between price and marginal cost. Consequently, empirical studies undertaken on the basis of financial-/price-impact measures in services sectors considered the price-cost margin as indicative of the magnitude of barriers (Hoekman, 2000). Quantity-impact measures In the case of quantity-impact measures, while the quantity imported under barriers may be observable – often in a highly aggregated form – there is no other quantity against which to compare it. This requires a satisfactory model of the determinants of trade, as well as data covering a sufficient variety of trading situations, to identify a situation in which trade is at least approximately free. The general approach to measuring the quantity effects of non-tariff barriers involves cross-commodity or cross-country regression models which estimate what trade would have been in the absence of barriers and compare this to trade that actually does occur. Such econometric methods (gravity models of international trade) have been applied for services trade. The difference between actual and predicted imports was taken to be indicative for barriers and was converted into a price or cost effect [Francois and Hoekman (1999)]. Going a step further, these price and quantity effects may be correlated with individual underlying restrictions. All sectoral papers produced at the Australian Productivity Commission on the estimation of the effect of restrictions on the economic performance of services firms broke down the identified price or quantity effects into their various components using econometric techniques. An econometric model which includes relevant determinants of the economic performance of firms, such as industry and economy-wide influences and measures of the trade restrictiveness (represented by the previously calculated frequency index), is developed from economic theory. The econometric method is then used to determine the separate impact of all determinants on prices and costs. Wherever possible, the components of the trade restrictiveness index (the foreign index and domestic index) are “entered” separately, so that the econometrics can reveal the separate effect of different restrictions, thereby substantially improving the practical value of the estimates. The method can also provide an indication of the extent to which restrictions raise price-cost margins, and therefore create economic rents, or raise costs above what they would be in the absence of the restrictions. As trade restrictions may have a simultaneous rent-creating and cost-increasing effect, further research is required to quantify these effects more fully. However, at this stage, the method makes it possible to break down price, costs and quantity impacts into all constituent parts and provides an indirect (and incomplete) indication of the extent to which restriction raise costs or create rents. The final step implies the incorporation of the estimated tariff equivalents into a general equilibrium model according to their rent creating and/or cost increasing effects. * Hardin and Holmes (1997) determined an FDI restrictiveness index across several services sectors. Restrictiveness indexes for specific services sectors have been calculated by McGuire and Schuele (2000) for banking services, McGuire et al. (2000) for maritime services, Kalirajan et al. (2000) for distribution services and Nguyen-Hong (2000) for professional services. ** All previously mentioned studies on the calculation of restrictiveness indexes were undertaken by the research team at the Australian Productivity Commission. See Chapter 5 for more detailed information. Source: adapted from Bosworth et al. (1997); Deardorff and Stern (1997); McGuire (2000); Findlay and Warren (1999).
151
Source: Author.
McGuire et al. (2000) for banking services; Warren (2000a) for telecommunication services
Research project on services undertaken by the Australian Productivity Commission
Francois and Hoekman (1999)
Francois and Hoekman (1999)
Hoekman
Study
Verikios and Zhang (2000), the FTAP Model
Dee and Hanslow (2000), the FTAP Model
Brown and Stern (1999), New Michigan Model
Brown and Stern (1999), New Michigan Model
Employed in following general equilibrium models Brown et al. (1996) Chadha (2000) Chadha et al. (2000) Australian Department of Foreign Affairs and Trade (1999) Robinson et al. (1999) Petri (1997)
Telecommunications (136 economies)
Banking (35 economies)
Business services and construction services (19 economies/regions)
Cross-sectoral tariff equivalents (recreation, business, construction, consulting, finance, health, hotels, retail, transport, wholesale)
Sectors and number of economies considered Cross sectoral “guesstimates”
x x x x
x x
x x
x
x
WTO Trade Policy Reviews National Trade Estimate Report on Foreign Trade Barriers from the Office of the United States Representatives USTR APEC Individual Action Plans Common List of Barriers prepared by the WTO Financial Leaders Group IMF TradePort site GATS “Telecommunication Reform” produced by the International Telecommunications Union (ITU)
Worldscope Database
GATS schedules
Source for databases
frequency
index
1.Calculation of frequency indexes for different types of restrictions (restrictions imposed on foreign and domestic services suppliers and restrictions on establishment and ongoing operations) 2. Price-based and quantity-based methods to determine the overall price/quantity effect 3. Correlation of price/quantity effects with the underlying restrictions
Multi-step method
Quantity-based method (gravity model)
Price-based method
Unweighted method
Methods used for quantification
Table 6.1. Estimates of services barriers employed in general equilibrium modelling of services trade liberalisation
“Guesstimates” calculated by Hoekman (1995) Most general equilibrium studies analysing services liberalisation employ the estimates of services barriers constructed by Hoekman (1995). These estimates are calculated on the basis of scheduled commitments under the GATS using a three-category weighting methodology. For quantification purposes, values of 1 (most restrictive, no access), 0.5 (specific bound restrictions) and 0 (free access) were allocated to the market access restrictions listed under GATS, and coverage ratios were estimated. These ratios were transformed into tariff equivalents by using a set of benchmark tariff equivalents for individual sectors, which ranged from a value of 200% for sectors considered highly restricted (maritime cabotage, air transport, postal services, voice telecommunication, life insurance) to values of 20-50% for sectors in which market access was less constrained. By multiplying the coverage ratios with the benchmark set of tariffs, Hoekman calculated tariff equivalents by sector and country. The weighted average tariff equivalents for 1-digit ISIC sectors for selected countries (referred to in the general equilibrium models presented) are indicated in Table 6.2. These estimates are derived from a rather incomplete database, as GATS schedules do not include all the barriers in place.1 Furthermore, the applied weighting methodology does not distinguish between barriers in terms of impact on the economy (i.e. a blanket prohibition on foreign presence will have the same weighting as an economic needs test). In spite of these limitations, these tariff equivalents have been extensively used in general equilibrium modelling as they were the only estimates available on the degree of protection in services sectors. These estimates constituted inputs for a number of studies (Brown et al., 1996; Chadha, 2000; Chadha et al., 2000; Australian Department of Foreign Affairs and Trade, 1999; Robinson et al., 1999) which do not explicitly model the different modes of services supply, but also for a study which includes FDI (Petri, 1997). Hoekman’s work represented the starting point for a number of studies that constructed improved restrictiveness indices for different modes of services supply or for specific services sectors, by incorporating information on the type of barriers and their likely economic impact when assigning weights to different barriers. Box 6.2 presents advances in the calculation of the various frequency indexes and openness indicators constructed for different modes of supply and specific services sectors. Tariff equivalents estimated by Francois and Hoekman (1999) – financial-based measures The tariff equivalents calculated by Francois and Hoekman (1999) using financial data on gross operating margins for a number of services sectors and drawing upon the quantity-based approach are also used in general equilibrium models. They developed an approach that uses sectoral data on operating margins in different services industries to assess the tariff equivalent in the respective services sectors. The information was collected from the Worldscope database which contains financial data on all firms listed on stock exchanges throughout the world. By comparing the services margins with manufacturing margins, or with a benchmark (an open country’s average margin), a scaling factor was determined which was then used to compare the prevailing tariff equivalent. The authors noted the impossibility of inferring that high margins are due to high barriers, as factors such as market size (number of firms), business cycle, state of competition, regulatory measures, substitutability of products, fixed costs, etc., determine the ability of firms to generate high profits. However, they concluded that there should be some correlation between data on operating margins and the restrictiveness of barriers, so that these margins are taken as indicative for the relative magnitude of barriers (see Box 6.1). 153
34.4 34.4 35.5 24.9 13.4 35.5 34.8 34.8 32.8 32.5 35.5 35.1 12.4
40.0 12.0 25.0 25.0 12.0 34.0 16.0 10.0 40.0 28.0 40.0 34.0 12.0
12.0 5.0 6.0 19.0 5.0 16.0 24.0 5.0 5.0 5.0 12.0 5.0 5.0 5.0 10.0 32.0
Wholesale and retail distribution 7.4 4.6 9.0 14.6 4.6 21.4 21.3 13.4 13.4 23.4 13.4 8.0 34.4 4.6 10.0 31.5
Construction
Source: Hoekman (1995), pp. 47-54.
Australia Austria Canada Finland Japan Korea Mexico New Zealand Norway Poland Sweden Switzerland Turkey United States EU Hong Kong (China) Chile Singapore China Brazil Argentina India Indonesia Malaysia Philippines Thailand Sri Lanka Pakistan South Africa
Country
182.2 138.8 191.1 142.7 116.9 191.1 190.4 175.8 110.2 189.6 193.9 193.1 118.3
Transport, storage and communication 183.4 98.7 117.7 181.0 142.0 164.9 152.3 181.5 122.2 116.0 184.2 178.1 31.6 111.4 182.0 149.8
Percentage
45.2 35.9 39.8 47.0 36.9 47.2 43.1 36.1 41.7 42.2 51.0 46.4 29.1
Business and financial services 24.8 20.1 25.9 23.8 28.9 36.3 40.9 30.5 25.7 42.6 22.5 27.7 35.4 21.7 27.2 39.0 42.9 33.7 42.0 43.6 42.9 40.6 43.3 34.2 43.3 40.3 43.4 35.4 25.8
Social and personal services 25.4 13.9 40.2 31.7 32.3 40.7 29.8 36.1 24.0 28.1 26.9 32.3 35.9 31.7 23.6 42.9
Table 6.2. Ad valorem tariff equivalents: “guesstimates” by 1-digit ISIC services sectors for selected countries
15.5 22.6 23.8 26.4 25.7 39.3 16.6
21.2
40.8 28.1 34.3 6.0 28.6 11.1 25.1
27.3 31.1
36.6
39.1 30.6 41.8 22.6 18.1 19.6
38.2 36.3
Manufactures
8.4 32.1 22.9 38.4 11.2 38.4 33.3
Agriculture
52.6 39.0
44.0 49.5 41.3 21.6 42.3 22.0 41.3
42.3
16.6 32.9 31.6 28.7 25.8 37.2 26.8
Services
85.4 -
13.3 19.9 46.7 79.9
46.8
19.6 -
17.9 60.1 42.5 28.1
Recreation
35.8 -
81.1 8.6 36.3
56.2
Business services 13.8 51.7 32.1 31.6 41.2 -
38.1 28.9
68.7 45.9 22.9 18.3 40.2 10.6 21.6
20.2
15.3 14.4 19.3 14.2 15.3 25.7 13.8
Construction
Source: Francois and Hoekman (1999), based on Worldscope; Hoekman (2000), p. 38.
Australia Canada EU Japan Korea Mexico New Zealand United States Chile China Indonesia Malaysia Philippines Singapore Chinese Taipei Thailand Other Cairns
Country
-8.8 26.2
67.1 25.3 14.7 7.7 11.1
-
7.0 19.2 22.1 28.6 37.3 -
Consulting
Percentage
60.3 69.8
55.2 34.0 53.6 28.3 53.9 46.3 64.8
56.3
41.0 44.5 51.6 40.5 33.3 57.6
Finance
40.6 29.3
24.3 29.2 -
37.0
2.3 22.3 40.1 -
Health
55.5 64.6
77.5 68.2 38.7 55.8 28.2 74.5
48.5
27.3 67.8 23.7 27.2 49.6 26.9
Hotels
44.2 24.2
21.3 24.4 26.4 11.2 43.9 5.4 21.5
34.6
Retail trade 7.9 12.0 23.6 32.9 26.7 28.4 6.6
Table 6.3. Average gross operating margins of firms listed on national stock exchanges, 1994-96, by sector
25.6 22.9
27.9 25.5 24.8 10.8 40.3 7.9 23.2
27.0
9.1 16.0 19.9 15.6 14.9 25.0 19.7
Wholesale
56.7 52.4
46.8 46.9 45.3 30.7 42.3 28.0 38.9
43.4
Transport/ utilities 36.5 32.6 20.6 31.2 51.0 35.6
Table 6.3 reports the results from this empirical exercise. It shows average operating margins for all firms in manufacturing and services sectors, as well as sectoral data on operating margins. The numbers suggest that services margins are significantly higher than manufacturing margins, generally ranking 10-15 percentage points above those in manufacturing. A comparison across countries indicates that only in Australia, Japan and Korea are operating margins similar and/or relatively low (less than 25%), suggesting that these are countries where services sectors are quite competitive. The opposite is true for China and Thailand, which register the highest margins. In terms of sectoral data on operating margins, margins for hotels and financial services tend to be quite high, while wholesale and retail trade and consulting have margins in the 20% range or less. Box 6.2. Advances in the weighting of barriers The set of weights developed by Hardin and Holmes (1997) to determine the restrictiveness of policies affecting FDI in APEC economies reflects the relative degree of restriction of these barriers. To improve the accuracy of results, they also conducted a sensitivity analysis to determine the extent of variation in the FDI restrictiveness index across 15 services sectors in response to alternative weights. Comparing the base-case index values with those obtained under scenarios using different sets of weights reveals which types of FDI restrictions are used more widely by different countries in particular services sectors. The authors found that, while varying the weights for the different types of restrictions has a significant impact on the absolute values of FDI restrictiveness indices, the relative index values between countries and sectors remain largely unchanged. Summing across weights assigned to existing policies affecting entry and operation post-entry, they obtained an overall FDI restriction index in different services sectors. The results indicate that communication and financial services are most subject to FDI restrictions, while business, distribution, environmental and recreational services are least restricted. The most restrictive countries include Korea, Indonesia, Thailand and China. Various frequency indexes and openness indicators have also been constructed for specific services sectors. Mattoo (1998) analysed market access commitments in financial services, covering direct insurance and banking. The results indicate that Latin America is the most restrictive for direct insurance and Asia the most restrictive for banking services. Classens and Glaessner (1998) constructed openness indicators for financial activities in a number of Asian countries. They found that many countries in the region are more restrictive towards foreign entry than the most liberal economy in the region, Hong Kong (China). Together with McGuire’s study on financial services in Australia (1998), they clearly illustrate the limitations of using information derived only from GATS schedules as these do not correspond to actual regulatory practice. In many instances, the prevailing regulatory stance is not more liberal than the country’s GATS commitment, and in some cases GATS commitments are more liberal than actual policies. In other words, in the latter case the governments made liberalisation precommitments. The sectoral studies undertaken under the auspices of the Australian Productivity Commission identify restrictiveness indices for barriers in specific services sectors. The estimations of trade barriers in services sectors such as banking, maritime services, distribution services, professional services (engineering, legal, architectural, etc.) and telecommunication services introduce further improvements. First, more extensive databases have been drawn upon to overcome some of the limitations of the information in GATS schedules. Second, more sophisticated weighting methods, which take into account information on the types of barriers and their likely economic impact, are utilised to assess the restrictiveness of different measures. Finally, in terms of the restrictions addressed, the authors distinguish between restrictions referring to national treatment and market access, as well as between restrictions related to establishment and to ongoing operations (i.e. postestablishment operations), following the classification adopted under the GATS. Although these indices were not directly incorporated into general equilibrium models, they constituted important elements for determining the effects of restrictions on price-cost margins or quantities, and for correlating these effects with the underlying restrictions (see Box 6.1 for a description of the method). Source: Author.
156
Tariff equivalents estimated by Francois and Hoekman (1999) – quantity-based estimates Hertel (1999) incorporated into the GTAP model the estimates calculated by Francois and Hoekman, who had used a gravity model which they applied to bilateral services trade for the United States and its major trading partners. These estimates of tariff equivalents for business and financial services as well as for construction services are computed by taking the differences between actual and predicted imports as indicative of barriers in the analysed sectors. The predicted “liberalised” imports are calculated on the basis of a free trade benchmark, considered to be Hong Kong (China) and Singapore. Results from the regression based on the gravity model as well as average tariffs on merchandise (for comparison purposes) are reported in Table 6.4. The estimates for business/financial services suggest that although average services barriers are often above applied average tariffs on manufactures, services barriers are not very high in many countries. The barrier estimates are higher for construction services, reflecting the resistance in most countries to allowing foreign construction firms to bid for procurement contracts and the barriers to the movement of semi-skilled workers. Table 6.4. Estimated tariff equivalents in traded services: gravity model based regression method Percentage
Country/region
Average tariff on merchandise
Business/construction
Financial services
North America Western Europe Australia and New Zealand Japan China Chinese Taipei Other NIEs Indonesia Southeast Asia India Other South Asia Brazil Other Latin America Turkey Middle East and North Africa CEECs and Russia South Africa Other Sub-Saharan Africa Rest of world
6.0 6.0 5.0 6.0 18.0 n.a. n.a. 13.0 10.0 30.0 25.0 15.0 12.0 13.0 20.0 10.0 6.0 n.a. n.a.
8.2 8.5 6.9 19.7 18.8 2.6 2.1 6.8 5.0 13.1 20.4 35.7 4.7 20.4 4.0 18.4 15.7 0.3 20.4
9.8 18.3 24.4 29.7 40.9 5.3 10.3 9.6 17.7 61.6 46.3 57.2 26.0 46.3 9.5 51.9 42.1 11.1 46.3
Source: Francois and Hoekman (1999); Hoekman (2000) p. 36.
157
Tariff equivalents for banking services estimated by Kalirajan et al. (2000) and telecommunication services estimated by Warren (2000a) (as part of the Australian Productivity Commission services project) The weighted average of the estimated effect of restrictions on price-cost margins of banking services (Kalirajan et al., 2000) and on costs and quantities of telecommunication services (Warren, 2000a) have been considered as indicative for other services sectors and incorporated into the FTAP model. Effect of restrictions on price-cost margins of banking services Kalirajan et al. estimate the price impact of non-prudential restrictions on the interest margins of banks, calculated as the difference between a bank’s lending rate and its deposit rate or cost of funds. Using a two-stage econometric technique, they separate the effects of bank-specific prudential regulations, such as capital requirements, reserve or liquidity requirements and net non-interest operating expenses, from factors such as interest rate volatility, market structure and the trade restrictiveness index. The authors found that non-prudential trade restrictions applied on a discriminatory basis to foreign banks raise the price, or “net interest margin”, of the different banking services by 5% to 60%. The price impacts of restrictions on foreign banks are highest for Indonesia, the Philippines, Malaysia, Chile, Singapore, Korea and Thailand. By contrast, Argentina, Australia, Canada, the EU, Hong Kong (China), Switzerland and the United States appear to have relatively low non-prudential regulations for foreign banks. The effect of market access barriers, which apply equally to domestic and foreign banks, on the price of banking services is lower, ranging from 0 to 24%. Effect of restrictions on cost and quantity for telecommunications services Warren assessed the effect of impediments to trade and investment in telecommunication services (mobile telephony and fixed network services) drawing upon the quantity-based approach. The estimates for the advanced industrialised countries are relatively low in comparison to the much higher estimates for the newly industrialising countries covered. Warren found that the price effect of restrictions on foreign telecommunication providers was less than 20% for the majority of economies studied. Indonesia, Colombia, the Philippines, South Africa, Thailand and Turkey were found to be the most restrictive, with price effects ranging from 20% to more than 100%. The United States, Denmark, the Netherlands, New Zealand, Australia, Germany and Sweden were found to be the least restrictive. Market access barriers, which apply equally to domestic and foreign suppliers, can raise the price by up to 70%. Tariff equivalents of barriers on domestic and foreign providers in the banking and telecommunication sectors are provided in Table 6.3. Their weighted average was incorporated into FTAP.
158
Banking services Domestic suppliers Foreign suppliers EstablishOn-going Total EstablishOn-going ment operation ment operation Argentina 2.53 2.81 Australia 7.08 2.22 Austria 2.52 2.80 Belgium 2.53 2.80 Brazil 0.87 0.87 35.06 10.50 Canada 2.53 2.81 Chile 15.47 7.73 23.20 22.74 11.26 Colombia 3.54 3.54 6.47 11.88 Denmark 2.52 2.80 Finland 2.52 2.80 France 2.52 2.80 Germany 2.52 2.80 Greece 2.52 2.80 Hong Kong, China 2.65 2.65 1.97 4.94 India 3.54 3.54 28.58 26.50 Indonesia 5.35 5.35 32.91 16.42 Ireland 2.52 2.80 Italy 2.52 2.80 Japan 10.03 10.03 2.05 13.22 Korea 14.93 14.93 18.15 18.58 Luxembourg 2.52 2.80 Malaysia 15.38 6.73 22.11 35.92 24.69 Mexico 10.48 2.92 Netherlands 2.52 2.80 New Zealand 2.52 2.18 Philippines 7.32 3.66 10.99 33.28 14.08 Portugal 2.52 2.80 Singapore 8.15 8.15 10.69 20.76 South Africa 2.64 12.27 Spain 2.52 2.80 Sweden 2.52 2.80 Switzerland 2.54 3.71 Thailand 20.56 12.50 Turkey 3.54 3.54 23.12 8.43 United Kingdom 2.52 2.80 United States 1.95 2.80 Uruguay 11.0 11.0 15.35 24.99 Venezuela 5.35 8.09 Source: Australian Productivity Commission (1999), pp. 31, 33.
Country
5.34 9.31 5.32 5.32 45.56 5.34 34.00 18.35 5.32 5.32 5.32 5.32 5.32 6.91 55.08 49.33 5.32 5.32 15.26 36.73 5.32 60.61 13.40 5.32 4.69 47.36 5.32 31.45 14.90 5.32 5.32 5.95 33.06 31.54 5.32 4.75 40.34 13.44
Total
Establishment 1.60 0.31 0.85 0.22 3.19 0.30 0.40 5.28 0.20 0.34 0.32 1.58 0.65 NE 29.66 1.46 1.00 0.26 1.83 0.65 1.23 0.81 0.20 0.27 1.35 1.28 6.65 1.71 0.65 1.23 15.88 11.20 4.74 4.21
Domestic suppliers On-going operation 2.22 0.44 0.61 0.76 1.28 5.28 -0.98 0.61 NE 41.04 2.47 5.50 5.43 21.43 2.45 0.82 7.12 0.32 14.01 8.40 0.20 2.87 5.37 3.81 0.31 0.85 0.65 3.81 1.07 1.68 10.55 0.20 0.34 0.32 2.56 1.26 NE 70.70 1.46 1.00 0.26 4.30 0.65 6.73 6.24 0.20 0.27 21.43 3.80 2.10 13.77 2.03 0.65 1.23 29.90 19.59 0.20 7.61 9.57
Total 1.60 0.31 0.85 0.87 5.07 1.08 0.40 8.44 0.20 1.43 0.32 1.58 0.65 NE 56.34 2.67 1.00 0.26 3.49 0.65 5.08 3.58 0.20 0.27 19.28 1.35 1.90 6.65 1.71 0.65 1.23 27.09 16.74 4.74 4.21
Establishment
Telecommunication services
Table 6.5. Price effect measures for banking and telecommunication services
Foreign suppliers On-going operation 2.22 0.44 0.61 2.29 1.28 15.83 2.94 0.61 NE 82.08 4.95 11.00 10.85 53.57 4.90 0.82 14.24 2.22 28.03 16.79 0.20 7.18 10.73
3.81 0.31 0.85 1.31 5.68 3.37 1.68 24.27 0.20 1.43 0.32 4.52 1.26 NE 138.42 2.67 1.00 0.06 8.43 0.65 16.08 14.43 0.20 0.27 72.85 6.25 2.72 20.89 3.93 0.65 1.23 55.12 33.53 0.20 11.92 14.94
Total
New estimates of services barriers Recently, new estimates of restrictions on maritime services (Kang, services (Kalirajan, 2000) and on professional services (Nguyen-Hong, 2000) of the Australian Productivity Commission’s services project. Table 6.6 characteristics of the methods used to calculate these new estimates. These below and will be incorporated into a revised version of the FTAP model.
2000), on distribution were calculated as part summarises the main methods are described
Effect of restrictions on trade margins in maritime services Kang (2000) used the policy indicators (frequency indexes developed by McGuire) to estimate the impact of restrictive maritime policies (restrictions on market access and national treatment) on bilateral shipping margins. He found that maritime services restrictions imposed by exporting countries generally appear to have a much greater impact on margins than those imposed by importing countries. As the low-income countries have more restricted markets, the benefits of eliminating restrictions on shipping services appear to be higher in low-income countries.2 Effects of restrictions on price-cost margins in the distribution and engineering sectors The estimates of barriers in distribution and professional (engineering) services are based on the price-impact approach (see Box 6.1). As for all other price-based models, the determinants of the price-cost margin used to break down the price-cost wedge include industry- and firm-specific factors, market structure variables as reflected in the level of concentration in the analysed sector, and the influence of public policies as determined by the restrictiveness index. These papers introduce two improvements over the estimation methods used at the Australian Productivity Commission: x
A separate variable to control for the effect of private-sector practices, such as exclusive buyer-supplier networks, alliances and cartels, is included in the analysis of the price and cost impacts of trade restrictions in distribution services.3 Data limitations on the extent of such practices have precluded their inclusion in the present trade restrictiveness index but their influence could be included in future, when such data become available.
x
Both studies provide tentative explorations of the cost-increasing and rent-creating effect of barriers. Considering whether services barriers increase costs or generate rents is of particular importance in numerical modelling work as the way in which protection is modelled (i.e. the way the estimates are incorporated into the model) significantly influences the welfare results. The greater the extent to which services barriers give rise to resource costs (rather than create rents), the greater the welfare improvement that may result from liberalisation. Issues relating to the determination of the cost-increasing and rent-creating effect of restrictions are discussed in more detail below.
Trade restrictions may have: cost-increasing effects, by restricting potential or existing firms from operating efficiently and thus pushing up business costs; rent-creating effects, by protecting incumbent firms from competition and thus allowing those firms to increase the prices; or simultaneous cost-increasing and rent-creating effects. Therefore, the following problems may arise when applying price-costs margins to determine the impact of barriers.
160
Maritime services 38 economies
Distribution services 38 economies
Professional services 34 economies
Source: Author.
x x
Nguyen-Hong (2000)
x x
Kalirajan (2000)
x x
Kang (2000)
Study, sectors and number of economies considered
WTO Trade Policy Reviews National Trade Estimate Report on Foreign Trade Barriers from the Office of the United States Trade Representative (USTR), which covers information on restrictions on maritime services for most economies OECD study on restrictions on maritime and multimodal trade in selected non-OECD countries APEC Individual Action Plans TradePort Internet site, which covers information on restrictions for most economies
Tradeport Database OECD publications USTR
OECD Inventory of Measures Affecting Trade in Professional Services WTO Questionnaire on Restrictions in the Accountancy Services Sector APEC Directory on Professional Services ILSAC’s Legal Services Country Profiles
x x
x x x
x x x x
x x
x
The GATS
x
Source for databases
Methods used for quantification
The multi-step method described in Table 1A Additional analysis of the cost-increasing and rent-creating effect of restrictions
Table 6.6. New estimates for services barriers
First, price-cost margin data are likely to embody changes in prices and costs other than those created directly by the imposition of restrictions. In the case of rent-creating restrictions, the price-cost margin will reflect the initial increase in prices facilitated by the imposition of restrictions, but could also reflect the subsequent increase in costs resulting from higher profitability. In the case of costincreasing restrictions, price-cost margin data will reflect not only the higher costs created by the restrictions, but also the higher prices which may follow as businesses seek to pass on as much of the higher costs as the market will bear. These “second-round” price increases would dilute the impact of the direct cost increases in the price-cost margin data. Consequently, particular restrictions may have significant effects on prices or costs that would not show up in the price-cost margin data. Second, if restrictions have a dual effect (of raising costs and creating rents), there is a risk that their price- and cost-raising effects may cancel each other out to some extent in studies using pricecost margin data alone. For example, if a country applies rent-creating restrictions, which directly push prices up by 5%, as well as cost-creating restrictions, which directly push costs up by 3%, the total effects of the restrictions would be 8% but the net impact of these restrictions on price-cost margins would be just 2%. These two problems suggest that price-cost margins are unlikely to capture the full effect of restrictions except in limited circumstances. This highlights the need to use additional cost data to estimate the effect of rent-creating and cost-raising restrictions separately and to conceptualise the rent-creating, cost-increasing or dual effect of policy on actual services outcomes once the impediments have been identified. The econometrics suggests that for distribution services, restrictions imposed on both foreign and domestic services suppliers increase costs. For engineering services, barriers to establishment and to ongoing operation imposed on foreign suppliers tend to have a rent-creating effect, while domestic barriers to establishment tend to increase costs. The price and costs impacts of restrictions are provided in Tables 6.7 and 6.8. The breakdown of the cost impact by type of barrier in distribution services is provided in Table 6.9. In addition, ongoing projects at the Australian Productivity Commission are extending work on estimating services barriers to other sectors, such as health and education services. Also, research is beginning to emerge concerning the identification and rigorous assessment of the extent to which various regulatory structures affect services trade and investment. In particular, work is proceeding for the electricity sector, with plans to extend the research to a broader set of industries.
162
Table 6.7. The effect of restrictions on distribution Percentage
Country
Cost impact of foreign barriers to establishment 0.57 4.87 3.09 1.32 5.16 0.25 0.06 3.66 2.70 2.26 8.23 2.73 0.77 0.03 0.47 5.24 2.76 2.26
Australia Belgium Canada Chile France Greece Hong Kong, China Indonesia Ireland Japan Malaysia Netherlands New Zealand Singapore South Africa Switzerland United Kingdom United States
Cost impact of domestic barriers to establishment 6.69 0.98 1.92 7.10 6.79 3.97 8.32 -
Source: Kalirajan (2000), p. 52.
Table 6.8. Impact of restrictions on engineering services
Austria Mexico Malaysia Indonesia Germany Spain United States Sweden Japan Canada Singapore Hong Kong, China South Africa Netherlands Australia United Kingdom Finland Denmark France Belgium
Price impact Foreign barrier to Foreign barrier to establishment ongoing operation 11.1 3.5 13.9 0.2 11.3 0.7 9.9 0.3 4.7 5.5 5.1 3.7 5.1 2.2 5.9 0.9 3.1 3.4 3.1 2.2 4.9 0.2 3.6 1.5 3.5 0.2 3.5 0.2 2.1 0.7 2.3 0.2 1.8 0.5 0.3 0.8 0.3 0.6 0.3 0.2
Source: Nguyen-Hong (2000), p. 63.
163
All foreign barriers 14.5 14.2 12.0 10.2 10.2 8.7 7.4 6.8 6.6 5.3 5.0 5.1 3.7 3.7 2.8 2.5 2.3 1.1 0.9 0.5
Cost impact Domestic barriers to establishment 6.8 1.9 5.3 3.2 2.9 3.9 3.8 0.7 2.2 2.7 0.8 2.3 0.7 5.2 2.1 1.4 0.7 0.7 0.7 0.7
F D F D F D F D F D F D F
Indonesia
0.47 5.24 8.32 2.76 2.26 -
3.66 2.70 2.26 6.79 8.23 3.97 2.73 0.77 0.03
Total impact of barriers to establishment on foreign (F) and domestic (D) suppliers 0.57 4.87 6.69 3.09 0.98 1.32 1.92 5.16 7.10 0.25 0.06
D F D Ireland F D Japan F D Malaysia F D Netherlands F D New Zealand F D Singapore F D South Africa F D Switzerland F D United Kingdom F D United States F D Source: Kalirajan (2000), pp. 53, 54.
Hong Kong, China
Greece
France
Chile
Canada
Belgium
Australia
Country
2.23 2.67 2.25 2.23 2.28 2.21 -
2.17 2.26 2.31 0.12 -
Restrictions on commercial land
2.09 3.73 -
0.07 -
Direct investment
1.16 3.57 0.56 1.70 -
1.09 3.29 1.15 3.47 -
Restrictions on largescale stores
0.48 1.00 1.70 0.32 0.50 1.53 -
0.32 0.49 1.47 0.34 0.98 0.52 1.56 0.03 -
Factors affecting investment
Table 6.9. Distribution services: cost impact to establishment, by type of barrier
0.64 1.04 3.22 0.80 2.27 1.68 5.09 -
0.65 1.96 0.65 1.92 0.69 2.08 -
Factors affecting local establishment
0.48 0.06 -
0.47 0.27
0.45 0.47 0.06 0.03 0.48 0.45 0.03
0.25 0.46 0.49 0.67 0.49 0.03 0.06
Movement of natural persons
Conclusion This overview suggests that the estimates of different services barriers vary depending on the data sources and estimation techniques employed. The wide range in the estimated initial size of policy interventions is reflected in the differing welfare effects from services liberalisation generated by general equilibrium models. The studies that use Hoekman’s “guesstimates” for the initial interventions generally report large welfare gains from services trade liberalisation. By contrast, studies that employ the estimates determined on the basis of price or quantity impact measures tend to generate lower, though still sizeable, welfare gains. The quality of estimates has been improving progressively, in terms both of the range of barriers addressed – which are collected from a growing number of sources, as Table 6.1 and 6.6 indicate – and of the measurement techniques employed, which now make it possible to estimate the price and cost effects of services barriers and to begin to determine the correlation between these effects and the individual underlying restrictions. However, it is difficult to determine whether these estimates are realistic, as a number of limitations remain. These are mainly related to the availability of data on existing restrictions and on their effect on economic performance. There are also major shortcomings with respect to assigning weights for different categories of restrictions. Modelling services liberalisation: recent developments Among the models employed to assess effects from services liberalisation, the studies that explicitly incorporate trade in services through commercial presence considerably improve the modelling framework. By capturing the effects of FDI liberalisation, this approach better represents the objectives of multilateral services liberalisation, in spite of certain limitations. Model structures like those developed by the Australian Productivity Commission – the FTAP model (Dee and Hanslow, 2000) or the New Michigan Model (Brown et al., 2000) – are more appropriate for simulating liberalisation in services sectors than models that do not explicitly consider the different modes of services supply. Ongoing efforts to increase the accuracy of modelling results and to overcome some of these models’ limitations are described below. However, before examining improvements to models explicitly incorporating FDI, the modelling framework developed by Markusen et al. (2000) is described briefly. While this model was not directly used for policy analysis, it proved valuable for incorporating FDI in general equilibrium frameworks (New Michigan Model). It might also provide a useful starting point in the analysis of additional FDI-/services-related aspects such as the potential benefits of specialised knowledge provided by foreign producer services or dynamic effects from FDI liberalisation. Markusen et al. capture the essence of imported services through two channels: the special knowledge needed to produce the foreign services and the imports of specialised intermediate inputs. Their empirical findings support the conventional conclusion of economic theory which states that FDI is beneficial to host economies. They find that liberalisation of producer services, such as managerial services, engineering services, financial services, marketing services and information services, which are important intermediate inputs and are intensive in skilled labour, has a powerful positive impact on the income and welfare of the country receiving the FDI. This impact is much stronger than in traditional competitive models (between 3% and 15% of GDP) because the additional intermediate services firms and the resulting increase in the variety of imported services lead to 165
increased total factor productivity in downstream industries. Also, producer services complement domestic skilled labour as they foster the accumulation of skilled labour. Interestingly, they found that the real wage of domestic skilled labour increases with liberalisation of policies that create barriers for foreign service providers, and the more foreign firms in the domestic market, the more the real wage of domestic skilled workers increases. The reason is that additional foreign firms lower the cost of the intermediate service product in final goods production. The authors also extend the conceptual analysis by incorporating dynamic aspects into the model, mainly to analyse the adjustment costs associated with FDI liberalisation. Recent improvements to models employed for policy simulations The FTAP model The FTAP model is an advanced model employed for estimating effects of multilateral services liberalisation. Drawing upon the structure first developed by Petri (1997), Dee and Hanslow (2000) recently applied the model to assess the economic implications of actual services sector policies. They also constructed the necessary database for FDI flows to support the model structure and incorporate improved estimates of barriers in services sectors. The current FTAP database contains only three (highly aggregated) sectors: primary, secondary and tertiary. With respect to barriers, only estimates for banking and telecommunication services were included in the model. As weighted averages of these estimates have been considered indicative for other services sectors and incorporated into the model, the accuracy of the results is thus limited. This model is currently being extended and improved by creating a database with at least six services sectors (Australian Productivity Commission, 2000). The modelling team is incorporating the recently calculated estimates for barriers in distribution, engineering and maritime services into the FTAP model. In addition to the quality improvement of estimates of barriers, the modelling of their removal has been also reconsidered. Where possible, the rent-creating and/or cost-increasing effect of the restrictions is determined and incorporated into the model. In addition to their efforts to obtain better estimates of the extent of policy interventions and appropriate incorporation into the model (identified as one of the research areas of highest priority, as models tend to be most sensitive to general variations in the barriers), the research team is also seeking to obtain better information on the magnitude of at least some of the elasticities employed. The New Michigan Model Drawing upon the work of Markusen et al. (2000) and Petri (1997), Brown et al. (2000) construct a new version of the Michigan model which takes account of FDI in services sectors. They adopt the same demand structure as Dee and Hanslow (2000) but employ lower values for the elasticities of substitution among foreign and domestic firms, foreign firms from different locations and among various firm types. They also incorporate different estimates for services computed by Hoekman (2000) and model them as cost-increasing components attributable to an increase in fixed costs borne by multinational corporations (MNCs) attempting to establish an enterprise locally. This new study developed on the basis of the Michigan model offers an alternative to the FTAP model which is based on the GTAP structure. It enlarges the available spectrum of models used for simulating liberalisation of services trade and makes it possible to assess the results from a range of models.
166
As in the case of Dee and Hanslow (2000), this is still work in progress, so that the results from an assumed 33% reduction in post-Uruguay Round barriers to trade and investment should be taken as illustrative of the considerable potential benefits that may be realised from trade liberalisation in the presence of FDI. However, contrary to the results reported by Dee and Hanslow, all of the industrialised countries except Japan show welfare gains. Most of the developing Asian countries also show welfare gains. Welfare declines from services liberalisation are mainly associated with the capital losses registered by a country as a consequence of liberalisation. The results from an additional scenario, which assumes an increase in the world’s capital stock in response to an increase in the rate of return, are considerably larger than in Dee and Hanslow. It is obvious that the international allocation of physical capital plays a central role and is far more substantial than consumer distortions in determining the welfare effects of trade liberalisation. Like the FTAP model, the New Michigan Model is subject to further improvement. As a next step, the authors are considering the analysis of the effects of regional negotiating options. Conclusion General equilibrium models are the most appropriate tools for conducting an economic analysis of the magnitude of the costs of services barriers and the corresponding welfare gains from their removal. By capturing the linkages between the services sectors and the rest of the economy and input-output forward links between services and other production sectors, modelling can help to assess the impact of trade policy changes on the economy as a whole and may also provide useful insights into broad negotiating modalities. However, limitations related to the modelling framework and to the availability of data on services barriers suggest that the welfare results generated should be taken as indicative and interpreted with care. The limitations identified in empirical assessments have been used to identify further research efforts. Proposals for further work The foregoing analysis of ongoing quantification work suggest that following areas deserve priority attention in the future. Data on barriers Efforts should continue to obtain better estimates of the size of policy interventions with a focus on the generation of comprehensive lists of existing restrictions, addressing weighting methods and addressing shortcomings of the price-cost margin approach to studying the effectiveness of protection and the differentiation among restrictions in terms of their economic impact (cost-increasing, rentcreating and/or dual effect). Modelling requirements The accuracy of models needs to be improved by using flexible functional forms and realistic allocation schemes, and by getting information on the appropriate magnitude of at least some of the parameters and elasticities employed. Other important areas include the exploration of dynamic linkages from services liberalisation in the longer term and the assessment of results from a range of models (not only GTAP-based models).
167
NOTES 1.
For example, McGuire (1998) found that only 38 measures from the total of 165 existing measures affecting financial services in Australia were listed in the GATS schedules. A number of barriers referring to government monopolies over the provision of certain types of financial services, prudential regulation, restrictions on FDI in banking and insurance, discriminatory government licensing requirements and government guarantees to selected financial service providers are not covered in the GATS schedules (as a result of the positive-listing approach).
2.
Fink et al. (2000) consider that estimating the impact of restrictive policies on bilateral shipping margins is problematic because data on import and export values are not reported by the same statistical entity and because shipping margins vary with unit values of shipped goods. Therefore, they develop an econometric model of liner transport prices for US textile imports to assess the relative importance of restrictions on transport costs. Both components of the shipping price: marginal cost and mark-up are broken down into their potential standard determinants, ranging from distance to technology, and into other potential determinants/trade restrictions such as public policy (cargo reservation schemes), private practices (price-fixing and other co-operative agreements) as well as restrictions on port and auxiliary services. Price-fixing cartels seem to have a stronger influence on prices than cargo reservation policies. A removal of private anti-competitive practices would generate a 38% reduction in transport costs, while the elimination of restrictive governmental practices would result in an 11% price reduction.
3.
Private-sector practices are also considered by Fink et al. (2000) in estimating the impact of restrictions on maritime services.
168
REFERENCES
Australian Department of Foreign Affairs and Trade (1999), Global Trade Reform: Maintaining Momentum, Canberra. Australian Productivity Commission (1999), “Trade and Assistance Review 1998-99”, Annual Report Series 1998-99, AusInfo, Canberra, http://www.pc.gov.au/research/annrpt/tar9899/index.html Australian Productivity Commission (2000), “Themes and Projects for 2000-2001”, July, AusInfo, Canberra. Bosworth, M., C. Findlay, R. Trewin and T. Warren (1997), “Measuring Trade Impediments to Services within APEC”, paper presented at the Symposium on Evaluating APEC Trade Liberalisation: Tariff and Non-tariff barriers, US International Trade Commission, Washington, DC, 11-12 September. Brown, D. and R. Stern (1999), “Measurement and Modelling of the Economic Effects of Trade and Investment Barriers in Services”, paper prepared for the Coalition of Services Industries (CSI) World Services Congress Atlanta, Georgia, 1-3 November. Brown, D., A. Deardorff and R. Stern (1996), “Modelling Multilateral Liberalisation in Services”, Asia-Pacific Economic Review 2, pp. 21-34. Brown, D., A. Deardorff and R. Stern (2000), “CGE Modelling and an Analysis of Multilateral and Regional Negotiating Options”, paper prepared for the Conference: Issues and Options for the Multilateral, Regional and Bilateral Trade Policies of the United States and Japan, University of Michigan, Gerard R. Ford School of Public Policy and Japan Economy Program, Department of Economics, 5-6 October. Chadha, R.(2000), “GATS and the Developing Countries: A Case Study of India”, forthcoming in Robert M. Stern (ed.), Services in the International Economy: Measurement and Modeling, Sectoral and Country Studies, and Issues in the WTO Services Negotiations, University of Michigan Press, Ann Arbor, Michigan. Chadha, R., D. Brown, A. Deardorff and R. Stern (2000), “Computational Analysis of the Impact on India of the Uruguay Round and the Forthcoming WTO Trade Negotiations”, Discussion Paper No. 459, School of Public Policy, University of Michigan. Ann Arbor, Michigan. Classens, S. and T. Glaessner (1998), “Internationalisation of Financial Services in Asia”, World Bank, Washington, DC. Deardorff, A. and R. Stern (1997), “Measurement of Non-tariff Barriers”, OECD Working Papers No. 179. Dee, P. and K. Hanslow (2000), “Multilateral Liberalisation of Services Trade”, Productivity Commission Staff Research Paper, Ausinfo, Canberra.
169
Findlay C. and T. Warren (1999), “How Significant are the Barriers? Measuring Impediments to Trade in Services”, paper presented at the conference Services 2000: New Directions in Services Trade Liberalization, at the University Club, Washington, DC, 1-2 June. Findlay, C. and T. Warren (2000), “‘Sins of Commission and Omission’: Measuring Regulatory Impediments to Trade in Services”, paper presented at the Productivity Commission and Australian National University (Joint Conference) on Achieving Better Regulation of Services, Australian National University, Canberra, 26-27 June. Fink, K., A. Mattoo and I. Neagu (2000), “Trade in International Maritime Services: How Much Does Policy Matter?”, World Bank, Washington, DC. Francois, J. and B. Hoekman (1999), “Market Access in the Service Sectors”, Tinbergen Institute, unpublished. Hardin, A. and L. Holmes (1997), “Services Trade and Foreign Direct Investment”, Staff Research Paper, Industry Commission, Australian Government Publishing Services, Canberra. Hertel, T., J. Francois and W. Martin (1999), “Agriculture and Non-agricultural Liberalisation in the Millennium Round”, paper presented at the Global Conference on Agriculture and the New Trade Agenda from a Development Perspective: Interests and Options in the WTO 2000 Negotiations, sponsored by the World Bank and WTO, Geneva, 1-2 October. Hoekman, B. (1995), “Assessing the General Agreement on Trade in Services”, World Bank Discussion Paper No. 307, World Bank, Washington, DC. Hoekman, B. (2000), “The New Round of Services Negotiations: Identifying Priorities and Options”, Federal Reserve Bank of Saint Louis, Vol. 82, pp. 31-47. Kalirajan, K. (2000), “Restrictions on Trade in Distribution Services”, Productivity Commission Staff Research Paper, Ausinfo, Canberra. Kalirajan, K., G. McGuire, D. Nguyen-Hong and M. Schuele (2000), “The Price Impact of Restrictions on Banking Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York (forthcoming). Kang, J. (2000), “Price Impact of Restrictions on Maritime Transport Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York (forthcoming). Markusen, J., T. Rutherford and D. Tarr (2000), “Foreign Direct Investment in Services and the Domestic Market for Expertise”, Policy Research Paper No. 2413, World Bank, Washington, DC. Mattoo, A. (1998), “Financial Services and the WTO: Liberalization in the Developing and Transition Economies,” paper prepared for presentation at the workshop, Measuring Impediments to Trade in Services, Productivity Commission, Canberra, 30 April-1 May. McGuire, G. (2000), “Measuring and Modelling Restrictions on Trade in Services: Note by the Australian Productivity Commission”, prepared for the OECD. 170
McGuire, G. and M. Schuele (2000), “Restrictiveness of International Trade in Banking Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York (forthcoming). McGuire, G.( 1998), “Australia’s Restrictions on Trade in Financial Services”, Staff Research Paper, Productivity Commission, Canberra. McGuire, G., M. Schuele and T. Smith (2000), “Restrictiveness of International Trade in Maritime Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York (forthcoming). Nguyen-Hong, D. (2000), “Restrictions on Trade in Professional Services”, Productivity Commission Staff Research Paper, Ausinfo, Canberra. Petri, P. A. (1997), “Foreign Direct Investment in a Computable General Equilibrium Framework”, paper prepared for the conference, Making APEC Work: Economic Challenges and Policy Alternatives, March 13-14, Keio University, Tokyo. Robinson, S., Z. Wang and W. Martin (1999), “Capturing the Implications of Services Trade Liberalisation”, paper presented at the Second Annual Conference on Global Economic Analysis, Ebberuk, Denmark, June 20-22. Warren, T. (2000a), “The Identification of Impediments to Trade and Investment in Telecommunications Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York (forthcoming). Warren, T. (2000b), “The Impact on Output of Impediments to Trade and Investment in Telecommunications Services”, in C. Findlay and T. Warren (eds.), Impediments to Trade in Services: Measurement and Policy Implications, Routledge, London and New York (forthcoming).
171
OECD PUBLICATIONS, 2, rue André-Pascal, 75775 PARIS CEDEX 16 PRINTED IN FRANCE (22 2003 01 1 P) ISBN 92-64-10042-3 – No. 52947 2003
