/0301 Tax Policy0~~~~~~~in DevelopingColuntriesedited byJavad KhazadShi AAnwar ShahA World Bad posium

I /)Tax Policy in Developing CountriesA World Bank Symposium

Tax Policy in Developing Countriesedited byJavad Khalilzadeh-ShiraziAnwar ShahThe World BankWashington, D.C.

01991 The Intemaional Bank for Reconstructionand Development / THE WORLD BANK1818 H Street, N.W., Washington, D.C. 20433, U.S.A.All rights reservedManufactured in the United States of AmericaFirst printing December 1991The findings, interpretations, and conclusions expressed in this study areentirely those of the authors and should not be attributed in any manner to theWorld Bank, to its affiliated organizations, or to members of its Board ofExecutive Directors or the countries they represent.Because of the informality of this series and to make the publication availablewith the least possible delay, the manuscript has not been edited as fully aswould be the case with a more formal document, and the World Bank acceptsno responsibility for errors.The material in this publication is copyrighted. Requests for permission toreproduce portions of it should be sent to Director, Publicatons Department, atthe address shown in the copyright notice above. The World Bank encouragesdissemination of its work and will nomlally give permission promptly and,when the reproduction is for noncommercial purposes, without asking a fee.Permission to photocopy portions for classroom use is not required, althoughnotification of such use having been made will be appreciated.The complete backlist of publications from the World Bank is shown in theannual Indev ofPublications, which contains an alphabetical title list andindexes of subjects, authors, and countries and regions. The latest edition isavailable free of charge from the Publications Sales Unit at the address in thecopyright notice or from Publications, World Bank, 66, avenue d'Iena, 75116Paris, France.AcknowledgmentsThe editors thank John Holsen, Johannes Linn, and Shankar Acharya fortheir support and three anonymous referees for their comments. A teamled by Ann Bhalla and including Lorrie Crutchfield, Nancy Barret, PeggyPender, Carlina Jones, and Leo Oteyza provided excellent secretarialsupport for this volume. The editors very much regret that a number ofimportant papers presented at the conference could not be included inthis volume because of space considerations.Library of Congress Cataloging-in-Publication DataTax policy in developing countries / edited by javad Khalilzadeh-Shirzai and Anwar Shah.p. cm. - (World Bank symposium)Includes bibliographical references.ISBN 0-8213-1990-61. Taxation-Developing countries-Congresses. I. Khalilzadeh-Shirzai, Javad. II. Shah, Anwar. m. Series.1IJ2351.7.T37 1991 9143997336.2'009172'4-dc20 CEP

ForewordThe 1980s witnessed a restructuring of tax systems in form of tariffs and sales taxes and increasingly soughtmany industrial countries. Major elements of these tax to reduce tariffs and replace turnover type sales taxeschanges included attempts at broadening the base of by value added sales taxes. This volume presents apersonal and corporate income and sales taxes by review of this experience as well as a discussion ofcurtailing tax preferences and exemptions (or replac- emerging tax policy issues in developing countries. Iing the latter by tax credits), decelerating previously hope tax policy officials, academics, and students ofaccelerated capital consumption allowances, reducing public finance in developing countries find this volumeboth the number of brackets and rates for income taxes, useful in their work.and, in some cases, introducing a value added tax.Developing countries also almost simultaneouslyadopted tax reform as a key element in their economic Lawrence H. Summerspolicy reform programs. These countries, however, Vice President, Development Economics andunderstandably placed a greater emphasis on the re- Chief Economist, World Bankv

ContentsForeword vContributors xiOpening Remarks xiiiIntroduction and Overview XVJavad Khalilzadeh-Shirazi and Anwar ShahExperience with Tax Reform xvSelected Tax Policy Issues for the 1990s xxConcluding Comments xxiiiReferences xxivPIARr I. TAx REFORM EXPEuENCES1. Tax Reform in Colombia: Process and Results 1Charles McLure, Jr. and George ZodrowEpisodes of Tax Reform 2Marginal Effective Tax Rates in Colombia 9The Distribution of Income and Tax Reform 16Administrative Simplification 17Revenue Performance 19Conclusion 19Notes 21References 212. Tax Reform in Malawi 23Zmarak Shalizi and Wayne ThirskMalawi's Economy and Tax System prior to 1985 23Tax Study of 1985 and Reform Proposals 26Implementation of the Tax Reform Proposals 34Some Lessons from Tax Reform in Malawi 35Notes 36References 373. Tax Administration and Tax Reform: Reflections on Experience 38RichardM. BirdApproaches to Tax Administration and Tax Reform 38Three Aspects of Tax Technology 40Tax Administration and Tax Reform in Latin America 43Some Possible Lessons for Tax Reform-Mongers 48Notes 50References 53ufi

vtit Contents4. Lessons from Tax Reform: An Overview 57Wayne ThirskIntroduction 57The Main Concerns of Tax Policy 58Trends in Tax Reform 62Some Lessons from Tax Reform 64Conclusion 69Notes 70References 70PART IL DESIGN OF INDIREC TAxES5. Design of the Value Added Tax: Lessons from Experience 72Sijbren CnossenThe Characteristics of Value Added Taxes 73Tax Coverage 76Tax Base 78Rate Structure 80Lessons 83Notes 84References 856. The Coordinated Reform of Tariffs and Indirect Taxes 86Pradeep MitraTariff and Tax Policy 86Tax and Tariff Instruments 87The Design of Taxes cum Tariffs 89The Reform of Taxes cum Tariffs 94Conclusions 96Notes 97References 99Pr m. TAxAION OF FOREIGN INvETmeT7. Taxation of International Income by a Capital-Importing Country:The Perspective of Thailand 100Cbad Leecbor andjack MintzIncentives 100Tax Policy Issues 101Tax Regimes of Thailand and Capital-Exporting Countries 101Impact of Taxation on the Financing and Investment Decisions of Multinationals 105Policy Options from the Perspective of Thailand 113Appendix: The Derivation of the Technical Results 117Notes 122References 1238. Taxation and Foreign Direct Investment 125Anwar Shah andJoel SlemrodReview of the Empirical Literature 125Unique Problems and Advantages of Studying Foreign Direct Investment in Mexico 127Taxation of Foreign Investment Income in Mexico 128Some Theory and the Empirical Model 130The Data 131Empirical Estimation and Results 133Policy Implications 135

Contents {xAppendix: The Data 135Notes 136References 137PARr 1V. TAXATION OF AGRICULTURAL LAND AMD F ANaAL JsTrUnoNs9. Prospects for Agricultural Land Taxation in Developing Countries 139Jonathan SkinnerHistorical Patterns of Land Tax Use 140Theoretical Aspects of Land Taxation 141Case Studies: Bangladesh, Argentina, and Uruguay 146What are the Lessons for Tax Reform? 149Notes 151References 15210. Taxation of Financial Assets in Developing Countries 153Christophe ChamleyFiscal Instruments 154Impact of Taxation on Financial Deepening 155Measurements of Revenues 157Efficiency Cost of Taxation 159Conclusion: Financial Taxation and Development 163Notes 163References 164PA V. TAX INCIDENce ANALYs11. The Redistributive Impact of Taxation in Developing Countries 166AnwarShah andJohn WballeyTax Incidence Analysis for Developed Countries 166Previous Tax Incidence Studies of Developing Countries 171Nontax Policy Elements in Developing Countries and Tax Incidence Analysis 173Pitfalls in Applying Developed Country Incidence Analyses to Developing Countries 174Import Licensing, Foreign Exchange Rationing, Quotas, and Incidence Analysis of TradeTaxes (Tariffs) 175Price Controls, Black Market Premiums, White Market Queuing Costs, and the Analysisof Sales and Excise Taxes 177Tax Evasion and the Incidence of Income Taxes 178Rural-Urban Migration Effects and the Incidence of Income and Payroll Taxes 179Credit Rationing, Foreign and State Ownership, and the Incidence of the CorporateIncome Tax 179Some Policy Implications 181Notes 183References 18512. A General Equilibrium Analysis of the Tax Burden and Institutional Distortionsin the Philippines 188Ramon L. ClareteStructure of the Model 188Calibrating the Model and Its Variants 191Incidence of Philippine Taxes 193Concluding Remarks 198Notes 199References 199

x ContentsPART VI. UsE OF QuANmTTATIVE Toois IN TAX PoucY ANALYsIs13. Applying Tax Policy Models in Country Economic Work: Bangladesh, China,and India 201Henrik Dabl and Pradeep MitraBangladesh 201China 207India 209Implementing Tax Policy Models 211Conclusions 213Notes 214References 21514. Tax-Benefit Models for Developing Countries: Lessons from Developed Countries 216Anthony B. Atkinson and Francois BourguignonTax-Benefit Models in Industrial Countries 217The Relevance of Tax-benefit Models in Developing Countries 221Conclusion 224Notes 225References 225PART VILL TAX POLICY AND ECONOMIC GROwTH15. Taxes, Outward Orientation, and Growth Performance in the Republic of Korea 227Irene Trela andJohn WhalleyBackground 227Growth Performance and Korean Policy Regimes 228Tax Policy during the Growth Process 231The General Equilibrium Model Applied to Korea's Tax System 236Results 241Conclusion 242Notes 243References 244PART VIIL PERSPECTIVES ON TAX REFOiM AND AGENDA FOR FuT RESEARCH16. Roundtable Discussions 246Amaresh Bagchi 246Richard Musgrave 247CharlesE. McLure, Jr. 249Nicholas Stern 250John Whalley 251Eduardo Wiesner 251References 252

ContributorsAnthonyAtkinson Robert ConradProfessor, London School of Economics, London, Professor, Institute for Policy Science and PublicU.K. Affairs, Duke University, Durham, NCAmaresh Bagchi Henrik DahlDirector, National Institute of Public Finance and Simulation Planning Corporation, DenmarkPolicy, New Delhi, India Dennis de TrayRoy W. BahL, Jr. Senior Economic Advisor, Office of the ViceProfessor, Georgia State University, University Plaza, President, Development Economics and ChiefAdanta, Georgia Economist, The World BankRichard Bird Dono Iskander DjojosubrotoProfessor, Department of Economics, University of Head, Budget, Credit and State Finance Agency,Toronto, Toronto, Ontario, Canada Ministry of Finance, Jakarta, IndonesiaFrancoisj. Bourguignon Vinod DubeyProfessor, Delta/Ecole Normale Superieure, Paris, Former Director, Economic Advisory Staff, TheFrance World BankAvishay Braverman Harry GrubertPresident, Ben Gurion University of the Negev, Intemational Economist, Office of the Tax Analysis,Beer-sheba, Israel U.S. Treasury, Washington, D.C.Kenan Bulutoglu John HolsenConsultant, Public Economics Division, Country Special Advisor, Office of the Senior Vice President,Economics Department, The World Bank Policy, Research and Extemal Affairs, The WorldChristophe Chamley BankProfessor, Department of Economics Boston JaVad hailzadch-ShiraziUniversity, Boston, Massachusetts Division Chief, Country Department IV (India), AsiaSheetal Chand Regional Office, The World BankAdvisor, Fiscal Affairs Department, International Chad LeechorMonetary Fund Senior Fiscal Economnist, Country Departrnent 4,G.ILR. Chipande Africa Regional Office, The World BankSenior Deputy Secretary, Ministry of Finance, charles McLure, Jr.Lilongwe, Malawi Senior Fellow, Hoover Institution, StanfordKwang Choi University, Stanford, CaliforniaProfessor, Department of Economics, Hankuk Jack MintzUniversity of Foreign Studies, Seoul, Korea Professor, Department of Economics, University ofRamon L Clarete Toronto, Toronto, Ontario, CanadaProfessor, School of Economics, University of the Pradeep MitraPhilippines, Diliman, Quezon City, Philippines Lead Economist, Country Department 1, AsiaSijbren Cnossen Regional Office, The World BankProfessor, Erasmus University, Rotterdam, The Richard A. MusgraveNetherlands Professor, Department of Economics, University ofCaliforniia, Santa Cruz, Californiaxi

Sarfraz Qureshi Alan TaitJoint Director, Pakistan Institute of Development Deputy Director, Fiscal Affairs Department,Economics, Islamabad, Pakistan Intemational Monetary FundI1 Sakong Wl-fied ThalwitzVisiting Fellow, Institute for Intemational Economics, Senior Vice-President, Policy, Research and ExternalWashington, D.C. Affairs, The World BankGunter Schramm Wayne ThirskAdviser, Office of the Director, Industry and Energy Professor, Department of Economics, University ofDepartment, The World Bank Waterloo, Waterloo, Ontario, CanadaAnwar Shah Irene TrelaSenior Economist, Public Economics Division, The Research Associate, Department of Economics,World Bank University of Westem Ontario, London, CanadaZmarak Shaizi JohnWhalleyChief Administrative Officer/Lead Economist, Office Director, Centre for Intemational Economicof the Senior Vice President, Policy, Research and Relations, University of Westem Ontario, London,External Affairs, The World Bank CanadaJonathan Skinner Eduardo WiesnerProfessor, Economics Department, University of Director, Mision para la Decentralizacion y lasVirginia, Charlottesville, Virginia Finanzas, Republic of Colombia, Bogota, ColombiaJoel Slemrod OktayYenlProfessor, School of Business, University of Chief, Resident Mission, The World Bank, NewMichigan, Ann Arbor, Michigan Delhi, IndiaNicholas Stern George ZodrowProfessor, Economics Department, London School of Professor, Department of Economics, RiceEconomics, London, England University, Houston, TexasEmil SunleyDirector of Tax Analysis, Deloitte and Touche,Washington, D.C.*The affiliations of the contributors are those as of November 1, 1991.xli

Opening RemarksWilfried ThalwitzA s the papers in this conference make clear, taxa- toward fiscal issues, in response to the growing recog-tion has become a vital component of the devel- nition that inappropriate and unsustainable expendi-opment effort. Indeed, without tax systems that func- ture and revenue policies are, in many instances, thetion well, governments cannot provide even basic major cause of disappointing economic performance.infrastructure and social services. The role that public It is also now recognized that a flawed tax structure isfinance plays in development featured prominently in often a contributing factor to economic inefficiency.public policy discussions during the turbulent 1980s- Initially, the Bank's policy dialogue on fiscal adjust-a time when many developing countries experienced ment focused on public expenditure issues, that is, onsignificant macroeconomic imbalances and a slow- the level and composition of recurrent and investmentdown in economic growth. These problems were in spending and the operation of public enterprises.part caused by external factors, such as drastic changes When the external funding that had previously helpedintheirterms of trade and high interest rates on external finance high levels of public expenditure dropped,loans. Many countries saw their GDP drop 10 percent in countries had to cut public spending and reorder theirthe span of a few years as a result of their changing priorities. Now, however, we have reached a stageterms of trade. These severe strains have revealed the where the belt tightening approach to fiscal reform isinherent brittleness of some of the structures of public no longer sufficient. Populations are growing, thefinance systems and underscored the need for funda- physical infrastructure is inadequate to support a revi-mental reforms. talized private sector, and social services must be pro-In the early phases of reform, "stabilization" policy vided, particularly for the needy. There is indeed a limitdominated the discussion. Our experience with stabi- to how far and how fast public expenditures can be cut.lization programs was that they required financial and Therefore, adjustment programs are becoming moredesign assistance from international agencies to concerned with mobilizing revenue through improvedsmooth the transition to a stable economic environ- taxation and better pricing of public services.ment. But we also realized that macroeconomic stabil- We must focus on tax reform for several reasons.ity could only be sustained when structural reforms First, structural reforms cannot pay off fully without anenable a country to use its available resources effi- improved public infrastructure, which is necessary tociently. So stabilization policy packages have had to promote the private sector as an engine of growth. Butinclude structural measures designed to reduce both public infrastructure cannot be improved without anthe distortions that retard growth and the inefficiencies equitable and efficient means of mobilizing revenue.in resource allocation. Second, reform measures often cause short-term dis-The World Bank took the lead in financing structural ruptions in the economy, such as a temporary increaseadjustment. It began by encouraging reforms in trade in unemployment. Appropriate fiscal measures wouldpolicy. Since then, the emphasis has gradually shifted prevent these problems from alienating the populationxUii

Opening Remarks xivfrom the reform program. Third, many countries are about a third of those of European and North Americanbeginning to see that the tax system has a role to play countries. Increasing this yield is justified in order toin providing a safety net for the poor. Finally, the tax fulfill the objectives I have mentioned, but any pro-system should be providing appropriate incentives to posal to do so should be carefully studied to ensure thatprotect the environment. Eastern Europe provides a it will be politically feasible. Also, if tax revenues aresobering example of the careless use of resources that raised, the instruments and the rates used must becould have been prevented through fiscal measures for carefully devised to mininize any disincentive effects.environmental protection. With appropriate pricing The experiences that developing countries have hadand taxation, it should be possible to generate addi- with tax reform suggest that broadening the base andtional revenues and to contain environmental damage eliminating the taxes that have significant distortionaryas well. Tax systems clearly need to be reformed if effects will be vital ingredients of any tax system that isgovernments are to pursue growth, equity, and envi- expected to boost growth and equity.ronmental protection. By providing an opportunity to exchange views on anIn particular, reformers need to look closely at the important subject, this conference is fulfilling an impor-structure of taxation-at the level, composition, design, tant function of the policy and research complex of theand implementation of taxes and charges. The total World Bank. The subject is of great concern to the Bankyield of the tax system in many developing countries is and all its member countries.

Introduction and OverviewJavad Khalilzadeh-Shirazi and Anwar Shahl ecent experience with growth-oriented adjust- The conference was organized around two core\R ment programs for developing countries indi- areas: (a) the findings of a Bank research project on taxcates that tax reform is an essential component of any reform experience in individual developing countriescomprehensive strategy for structural adjustment and and an overview of tax administration, tax reform, andthe resumption of growth (see Chhibber and the general lessons from the reform experience; andKhalilzadeh-Shirazi 1988). The existing tax systems of (b) a number of specific tax policy issues related to themany developing countries are distortionary and con- research conducted in, or sponsored by, the Bank'stribute to a host of economic problems, including pro- Public Economics Division. The conference concludedduction inefficiency, capital flight, and fiscal and with a round table discussion by a panel of experts onbalance of payments disequilibria. Furthermore, there the Bank's future research agenda on tax policy andis a growing recognition that fiscal imbalances cannot related topics. There was not enough time, however,be addressed simply by curtailing expenditures. In to cover other important issues, such as municipalorder to obtain a deeper understanding of what consti- taxation and intergovernmental fiscal relations (seetutes successful tax reform and to enhance the World Shah 1991a, 1991b). The present volume contains theBank's ability to assist member countries in this area, revised versions of the papers presented at the confer-the Bank's Public Economics Division has been con- ence. This overview highlights the main themes of theducting a program of research on this topic over the papers and the conclusions of the discussions.past few years.One of the principal research projects of this pro- Experience with Tax Reformgram has been devoted to examining the experiencesthat developing countries have had with tax reform. Participants discussed the experience with tax re-The intention was to collect information on past suc- form by examining the background to the reforms andcesses and failures so as to provide guidance for coun- the lessons gained from them.tries that are facing similar sets of circumstances andare embarking on tax reform. With the completion of Background to Reform Processthis project, as well as research on a number of otherimportant tax policy issues, a conference was held in The conference devoted considerable time to a diag-Washington, D.C., on March 28-30, 1990, to discuss, nosis of existing tax structures and to a review of thedisseminate, and evaluate the findings of this research. broad themes emerging from the tax reform movementThe conference brought together a number of leading that has swept the developing world in recent years.tax policy specialists from both developed and devel-oping countries to discuss the lessons from tax reform FRAmEWORK FOR ANALYSIS. The Ramsey rule calling forexperiences in developing countries, selected aspects a highly differentiated structure of taxation, by varyingof tax policy, and a future research agenda in this the tax rate inversely with the elasticity of demand andimportant policy area. supply, has endured in the optimal tax literature (seexv

x:Y Introduction and OtAenewAtkinson and Stiglitz 1976; Diamond and Mirrlees 1971; treatment distorts individual choices with respect to theRamsey 1927; and Stem 1976, 1982, 1987). As Thirsk form of income, the asset ownership, the businesssuggests in chapter 4, putting this rule into operation organization, the sector of investment activity, and theleads to an intractably large number of rates, which time profile of investment. Thus business decisions arewould be difficult to calculate and infeasible to admin- not guided by economic considerations alone. Taxister effectively (see also Deaton 1987; Feldstein 1978; considerations may be playing a significant part inand Slemrod 1990). Slemrod has argued that the opti- these decisions, and the resulting allocation of re-mal tax theory can serve as a guide to designing "opti- sources might not be consistent with least-cost outputmal tax systems' only if one considers the technology choices for the economy as a whole.of tax collection, that is, the feasibility of tax instru- Tax expenditures (forgone revenues) are widelyments, the cost of tax administration, and compliance. used to advance wide-ranging but sometimes conflict-The response to these difficulties has been quite prag- ing tax policy objectives, such as promoting industrialmatic. In recent tax reform episodes around the world, development, savings, investment, employment, andthe emphasis has been on fairly uniform taxation exports (see Boadway and Shah in Shah forthcomingthrough an explicit recognition that gains in economic a). Given the limited tax bases, poor compliance, andefficiency, horizontal equity, and administrative sim- enforcement in many developing countries, tax expen-plicity that stem from uniform taxation outweigh any ditures are often ill-suited to achieving individual pol-vertical equity losses. icy objectives. These incentives confer windfall gainson some activities at substantial cost to the treasuryDIAGNOSTICS. A common theme of the review of tax without inducing commensurate behavioral responsesreform experiences was that tax structures in most (for empirically derived benefit-cost ratios for variousdeveloping countries are complex (difficult to admin- tax incentives in developing countries, see Bernsteinister and comply with), inelastic (nonresponsive to and Shah in Shah forthcoming b; and Shah and Baffesgrowth and discretionary policy measures), inefficient in Shah forthcoming a). Because of these poorly con-(raise little revenue but introduce serious economic ceived tax preferences, along with widespread taxdistortions), inequitable (treat individuals and busi- evasion and avoidance, many economic activities,nesses in similar circumstances differently) and, quite even in the formal sectors, go untaxed while the restsimply, unfair (tax administration and enforcement are are taxed at inefficiendy high levels. Taxable activitiesselective and skewed in favor of those with the where- are subject to a multitude of rates, some of them quitewithal to defeat the systemn). There is a heavy reliance onerous, established on narrow bases. At such heavyon taxes on international trade, which undermines the levels of taxation, after-tax rates of return are oftenlong-term international competitiveness of developing below the opportunity cost of funds. Furthermore, highcountries. User charges and taxes on income, property, rates encourage tax avoidance and compromise theand wealth contribute only a small proportion of total fairness of the tax system. In view of these perverserevenues. Agricultural incomes, fringe benefits, and, in incentives and the resulting tax evasion, the inadequatesome countries, public sector wages, are not taxed. resources of tax administrations are often stretched toTaxes on wealth, bequests, land, and property exist in the limit. Poor tax compliance introduces considerabletheory but have been rendered ineffective by design inequity into the current structure and also makes theproblems or the lack of interest in administration, or tax system an inefficient and ineffective instrument ofboth. Personal and corporate income taxes are levied public policy.on narrow bases at high rates. Sales taxes are levied ina cascading manner, thereby imposing tax pyramiding DIRECrIONs OF REFORM. As already mentioned, a large(gross price inclusive of tax is taxed as the commod- number of developing countries have undertaken taxity passes through various production and distribu- reform in recent years. Their successes and failures cantion channels) and, in some instances, more than provide guidelines for countries in similar circum-100 percent full forward shifting (final sales price in- stances that may now or in the future attempt to reformclusive of the tax rises more than the amount of the their tax systems. The countries that have embarked ontax). reform differ in the nature, substance, procedure, con-The existing tax structures impose varying levels of text, and timing of their tax reform. For example, Co-taxation, depending on the form of income, the type of lombia (McLure and Zodrow this volume), theassets, the size and legal status of the businesses, and Republic of Korea (Choi 1990), and Turkey (Bulutogluthe kind of business activity. As a result, both the and Thirsk 1991) each had a long drawn-out period ofaverage effective tax rate (tax as a percentage of in- tax reform, whereas other countries such as Indonesiacome) and the marginal effective tax rate (the tax (Asher 1990) and Malawi (Shalizi and Thirsk this vol-wedge on the after-tax rate of retumn) vary substantially ume) carried out major changes in their tax system in aacross assets and sectors, thereby creating an uneven single episode implemented over a short period ofplaying field for economic agents. Such differential time.

JavadKhalilzadeh-Sbirazi and Anwar Shah .XvlFaced with mounting deficits and having cut expen- ments against their domestic tax liability (see Shah andditures as far as is prudently possible, particularly on Slemrod this volume; and Slemrod in Shah forthcom-public investment and social spending, a number of ing a).developing countries have decided to restructure their Although the broad directions of reform are remark-tax systems to seek higher revenues or to improve the ably similar, a number of unresolved and controversialrevenue elasticity and buoyancy of the tax structure. issues remain. For example, all recent attempts at taxThe secondary goals of these reform movements have reform have curtailed tax preferences, especially forbeen to (a) eliminate the disincentive effects of onerous investment, but some economists would argue thatlevels of taxation; (b) reduce the economic inefficien- certain tax incentives, such as the investment tax credit,cies induced by the distortionary taxation of assets and are desirable because they lower the user cost of (new)sectors; (c) protect the poorest of the poor from the tax capital and thereby encourage greater capital forma-net; and (d) provide partial relief from the unwelcome tion. (For a fuller discussion of this issue, see Boadwayeffects of inflation. Thus revenue enhancement, eco- and Shah in Shah forthcoming a. The effectiveness ofnomic efficiency, horizontal equity, and simplicity is- such incentives in the presence of market imperfec-sues have dominated the world agenda on tax reform, tions is discussed by Rajagopal and Shah in Shah forth-and other issues such as vertical equity and intema- coming a, b).tional income taxation have received only scant atten- The proper role of progressive income taxes in de-tion. The emphasis on the redistributive role of the tax veloping countries is another intensely debated issue,system is gradually waning-a direct consequence of as is the question of whether the personal income taxthe fact that tax evasion is so pervasive. Although should have fewer tax brackets and rates on account ofprogressivity remains high on the political agenda in simplicity (see Musgrave and Stem, both in chapter 16).theory, often the political will to enforce income tax Some would argue that the proponents of simplicitycompliance is lacking. Vertical equity is increasingly should focus on the defnition of the base, not onbeing perceived as an elusive goal and therefore is whether there is a single income tax rate or four or five.being assigned a lower order of priority in tax reform. Others would point out that there is a tradeoff betweenIn pursuit of revenue enhancement, many countries are simplicity and progressivity. And on the question ofrelying less on narrowly based trade taxes and are replacing income taxes by broadly based consumptionemphasizing broadly based consumption (hybrid (expenditure) or cash-flow taxes, perplexing philo-value added) taxes. To reduce the disincentive effects sophical and transitional issues continue to dominateof taxation, some countries are bringing down the current discussions. Broadly based consumption taxesaverage and marginal effective tax rates by eliminating in their pure form would tax wage income only (seeineffective tax preferences and thereby broadening the Zodrow and McLure 1988) and would exempt capitalbases, while leveling the rates. These measures, how- income apart from rents. The equity implications ofever, compromise vertical equity. As Richard Musgrave such taxes would create considerable controversy.points out (see chapter 16), broadening bases may raise Cash-flow taxes would be simple in design and arethe threshold of taxation and have fewer and lower tax conceptually superior to the existing income taxes,rates, but it does not pay adequate attention to the especially the corporate (equity) income tax, in thedistribution of relative tax burdens across income setting of a closed economy. Developing countriesclasses. Some have attempted to protect the poor by with an open economy may find cash-flow taxes un-exempting or zero rating foods under a value added tax suitable, especially because they would not be credit-(VAT) and by raising the threshold of taxes on personal able under the existing foreign tax credit regimes andincome, urban property, and agricultural land. cannot be used as withholding taxes. Moreover, suchIntersectoral and interasset distortions are also being taxes are considered to be so difficult to implement thatreduced as countries endeavor to create a level playing no country has yet adopted them (except in enclavesfield by eliminating special preferences and by replac- such as mining), although Mexico has recently indi-ing cascading turnover and sales taxes by more neutral cated that it hopes to move gradually toward a cash-value added taxes. There have also been attempts to flow taxation of business incomes.mitigate the unwelcome effects of taxation in the highlyinflationary economies of Latin America through partial Lessonsfor Tax Reformindexation of the tax system. Regional and interna-tional tax competition to attract foreign investment has Tax reform experiences to date offer some importantbeen intense (see Chia and Whalley in Shah forthcom- insight into useful tax policy design and institutionaling a). Countries that provide special incentives for development.foreign direct investment tend to overlook the implica- 1. 7The value added tax should be an instrument oftions of the tax systems of those advanced nations that cboice for most developing countries contemplatingtax their residents on theirworldwide income but allow reformn of their sales taxes. A value added tax canthem to claim credit for taxes paid to foreign govern- provide greater revenue, tax neutrality (economic effi-

xt4ii Introduction and Ovnewciency), and, under certain circumstances and to a ronment, and to discourage public "bads" has welcomelimited extent, vertical equity. That the VAT increases effects in that it discourages such activities and alsorevenue and economic efficiency is well documented. raises additional revenues (see Shah and Larsen forth-Indeed, the VAT has been an unqualified success in this coming; Shah 1990, 1988; and Summers 1991). Thus inregard. Vertical equity is also improved through the devising tax policies to meet economic and social objec-trade component of the VAT, which reduces rents ac- tives, potential gains must be weighed against the rev-cruing to the wealthy recipients of import or foreign enues and potential losses in efficiency that might beexchange licenses. Furthermore, a VAT can assist in associated with these measures. Furthermore, the designimproving the collection of other taxes. This potential of tax measures must be consistent with their objectives.has yet to be exploited by a developing country. The 4. Tax reform must take into account initial condi-VAT has helped raise additional revenues and reduce tions at home and abroad. In reforming their tax sys-the efficiency costs of taxation in Indonesia, Turkey, tems, developing countries are severely constrainedBrazil, Colombia, Mexico, Korea, and Malawi. Of not only by their own institutional settings but also bycourse, several difficulties arise when the vAT is imple- the tax structure in capital-exporting countries. Formented in developing countries. A VAT cannot cover example, the U.S. foreign tax credit regime discourageseconomic activities carried out in the informal sector in the adoption of a cash-flow system of taxation in de-a typical developing country. Also, to keep the poor veloping countries. Moreover, the circumstances inout of the tax net, basic foods and necessities are such countries are usually such that they would expe-usually exempted, which gives rise to administrative rience serious transitional difficulties if the tax systemcomplexities. Interregional trade creates its own spe- were to be redesigned from scratch. Developing coun-cial problems for vAT administration. A value added tax tries must take into account initial conditions at homeis best administered by the central or federal govem- and abroad. Otherwise, the reform effort is likely to fail.ment. And a value added tax is not necessarily superior The impact of tax policy on intemational competitive-to a well-functioning retail sales tax in small, island- ness has not received much attention in tax reformtype economies. analysis, but it appears that developing countries often2. The base of existing taxes should be broadened at engage in wasteful tax competition and do not givethesame time that tax administration reforn is carried adequate thought to the tax regime that a potentialout. Base broadening is compatible with a number of marginal investor faces in his home country. A marginaleconomic objectives. It can increase revenues and im- investor from a country with a worldwide system ofprove the simplicity, neutrality, and equity of the tax taxation can be taxed by the host country at the homesystem. Neutrality increases because base broadening country tax rate without feeling any disincentive ef-usually reduces differential tax treatment among assets fects. Furthermore, the host country needs to adoptand sectors of economic activity by leveling the playing appropriate income attribution rules to circumvent thefield. Vertical equity also increases because tax expen- shifting of income to low-tax countries or to tax havensditures that offer disproportionate levels of benefits to through transfer pricing.the rich are curtailed. Lower and fewer tax rates also 5. The credibility of the tax regime is the key to theenhance neutrality but are not compatible with vertical success of any tax refonn. A stable tax policy environ-equity objectives. Tax administration difficulties con- ment encourages businesses to take a longer-term per-tinue to stand in the way of broadening existing tax spective in their finance and investment decisions.bases and having fewer and lower rates. The record to Making tax changes without giving adequate consider-date for these measures in improving the taxation of ation to transitional arrangements can undermine theincome is not dear. The apparent lack of success in this credibility of a tax regime. Therefore, transitional ar-area is attributable to several factors: selective and lax rangements require much more careful analysis thanenforcement practices, ineffective tax administration in they have hitherto been given in developing countries.part due to political inertia, institutional and political In addition, tax changes must be presented as part of adifficulties associated with bringing agricultural in- long-term strategy to improve the public sector envi-comes into the tax net, and an overall disenchantment ronment for the private sector. The tax regime wouldwith income taxes as revenue instruments in an eva- gain the confidence of business if more attention wassion-pervasive environment. paid to the preparation and analysis of reforms, ad-3. The use of the taxsystemforspecialtaxpreferences vance consultation, providing a reasonable period ofshould be carefully evaluated. Using the system to adjustment prior to implementation, grandfatheringprovide tax incentives (tax expenditures) usually provisions, and the historical consistency of tax reform.causes a serious drain on the national treasury by 6. Coordinated tax reform offers significant advan-conferring windfall gains on existing activities or by tages over isolated piecemeal tinkering with the taxshifting resources to tax-preferred activities (see Shah system. A coordinated reform ensures that individualand Baffes in Shah forthcoming a). But the use of the tax changes will be consistent with the central objec-tax system for corrective purposes, to protect the envi- tives. For example, a reduction in tariffs without a

favadKhalilzadeh-Shirazi and Anwar Shah xixcorresponding increase in other taxes, generally of a country in question. In low- to middle-income coun-value added type, can increase the fiscal deficit and tries-such as Colombia, Malawi, Turkey, and Indone-exacerbate macroeconomic difficulties. Furthermore, sia-broader income taxes are not likely to produceto improve economic performance in general, tax re- large revenue gains and therefore the VAT is expectedform should be closely integrated with structural ad- to be the mainstay of the revenue-raising effort. Injustrnent measures. newly industrialized countries such as Korea, however,broader bases offer considerable potential for increas-Poltical Economty of Tax Reform ing revenue. In a country with a federal system ofgovernment, the powers of taxation are delegatedTax reform is a sensitive and difficult process. The among the levels of govermnent in a way that typicallypayoff tends to be of a long-term nature and therefore constrains tax reform choices. In India, for example, ait is difficult to get politicians to commnit themselves to full-fledged union VAT would meet with state-levela comprehensive reform. Few developing countries are opposition because of a concem that a federal VATlikely to give tax reform initiatives serious consider- would not leave much room for state and local salesation until they are faced with a fiscal crisis. In theory, taxes. In Pakistan, the octroi tax, which is a tax oncomprehensive reforms are more desirable because a intermunicipal trade, could not be repealed because ittax system is better able to meet revenue, efficiency, is a significant source of local revenue.equity, growth, and simplicity objectives in such a Tax policy advice must also give due attention toframework. In practice, since the gains from compre- current administrative practices and what potentialhensive reform become visible only in the medium to there may be for improvement. Experience suggestslong term, it is a challenge to assemble a political that compartmentalizing public policy in various de-quorum to carry through a wholesale reform. Often, the partments (or even various branches of the samepragmatic course is to strive for incremental reforms in departments) limits tax reform options. None of thea consistent manner over time. Historical consistency, countries reviewed by the conference handle tax andalthough desirable, is difficult to achieve. Consider the transfer options simultaneously. The range of choicescase of Colombia, where a net wealth tax was consid- is restricted to altemative tax instruments, and directered an important progressive element of taxation in expenditure options are excluded.the 1974, 1986, and 1988 reform episodes. In 1989, The political and civil service elite in the countryhowever, it was repealed. Also consider what mright must assume the 'ownership" of the proposals if thehappen if the tax rates on income are reduced at the reforns are to succeed. The chances for success alsosame time that the base is broadened. If tax preferences increase when local experts participate in the design ofare later restored to appease special interests, the initial the reform because they are better judges of the polit-reform effortwould have contributed to a deterioration ical pulse of the country. The success of tax reform inof the tax structure for in the final analysis the lower Colombia and Malawi can be attributed in part to therates would be applicable to still narrower bases (see trained core of local experts who worked closely withThirsk chapter 4). Broader bases and lower rates can foreign advisers.erode the tax structure further wherever tax evasion is The way to increase compliance is to make sure notprevalent. only that the people are consulted on the reform pro-Tax changes create winners and losers. Also, each posals themselves but also that they are given a cleartax change introduces some efficiency and vertical idea about how the money will be spent. The authori-equity tradeoffs that must be recognized and appropri- ties in Malawi consulted widely with taxpayers toately addressed. Canada, for example, introduced re- gauge their reaction to income tax changes beforefundable tax credits to counteract the regressivity of the finalizing their proposals. This helped to ensure that theVAT. Developing countries deal with regressivity final reforms were acceptable to a majority of thethrough exemptions. It is important to identify gains population that would be affected by them.and losses by income class, by geographic region, and Whatever choices may be made on the path to re-by political affiliation so that the long-run viability and form, it helps to have a coherent plan in place beforesustainability of reform measures can be objectively implementation begins. Also, tax reforms must remainevaluated. Short-term tax expenditures designed to flexible so that they can respond to changing economicmeet nonrevenue objectives should be avoided since and social conditions.they create strong political constituencies wedded tothese measures. A comprehensive reform offers some Selected Tax Policy Issues for the 1990spossibility of balancing the gains and losses of variousgroups, which usually is not the case in piecemeal The conference debated a number of issues that arereform. expected to dominate tax policy discussions in theCountry experiences suggest that tax reform pro- 1990s. These include tax administration, the design ofposals must consider the institutional features of the indirect taxes, the taxation offoreign investment, fnan-

Introduction and Ovenlewcial taxation, resource taxation, the incidence of taxes, nues without causing major economic distortions. Thustax policy and economic growth, and the quantitative it is particularly important to coordinate the reform oftools for tax policy analysis. tariffs and domestic indirect taxes and to design anappropriate value added tax.TaxAdminisration Mitra argues in chapter 6 that tariff reform should notbe carried out in isolation from the reform of otherAs already mentioned, tax administration plays a indirect taxes if the potential losses in public revenuesvital role in the success or failure of any attempt to arising from tariff reductions are to be offset and mac-reform taxes. Unfortunately, with the notable excep- roeconomic difficulties kept at bay. A coordinated re-tion of the studies by Deaton (1987) and Slemrod form of these taxes would combine reductions in tariffs(1990), the existing public finance literature does not with an offsetting or, preferably, revenue-enhancingpay adequate attention to tax administration issues. upward adjustment in the sales tax orvAT, which wouldFrom the experiences of tax reform in Latin American apply equally to both domestic production and im-countries, Richard Bird (chapter 3) develops some ports. The protection function would be served bybasic rules for tax administration reform. He argues that customs duties and the revenue objective by the salestax structure and administration are interdependent tax or VAT. Not only could revenue neutrality be pre-and therefore that they must be considered together. served in this scheme, but the rate structure could beDeveloping policy recommendations on administra- raised so as to meet the demand for any assistance fortive reform requires closer and, preferably, quantitative trade liberalization that may become necessary.analysis of many aspects of administration such as the As noted earlier, the vAT is the most pervasive featureinternal incentive structure and the operating costs of of tax reform in many developing countries. From thethe tax system. In the 1980s Latin American countries lessons learned about VAT design, Cnossen in chaptermoved away from progressive personal income taxes 5 gleans advice for developing countries that are con-and toward VATs. Bird interprets this change as a clear templating a vAT-type tax. First, he condudes that pre-recognition of the administrative dimension of tax re- retail VATS cause such significant distortion andform. These countries appear to have recognized that administrative complexities that they are not worthprogressive income taxes are difficult to administer. adopting. He argues for the use of a scale factor (say,Bird advocates simplicity as the fundamental rule in size of turnover), supplemented by administrative cri-tax reform and proposes a tax reform package that teria relating to trader type and employment, to ex-takes into account the interdependency of tax structure clude small firms from VAT coverage. Second, alland tax reform. The main measures he proposes are to services (except health care, education, social welfare,eliminate unproductive taxes; keep differential rates to banking, and insurance) should be included in thea minimum, whether in commodity taxes or, to reduce base. Third, rates should be differentiated as little astax arbitrage, in the effective rates of income taxes; draft possible, although to protect the poor it may be neces-the law clearly and communicate it effectively to both sary to reduce rates for food, essential consumer items,administrators and tax payers; and focus on collecting drugs, electricity and fuel, newspapers and books, andrevenue and not on using the tax system to achieve public transportation. If luxuries are to be taxed at anonfiscal ends. Bird concludes that modest research higher rate, then this should be done by special excises(and action) on alternative administrative arrange- rather than by a higher VAT. Fourth, the VAT is an idealments is more likely to lead to a more or less fair and tax for large, integrated economies with sophisticatedefficient tax system for most developing countries than production and distribution channels. It is less suitablewould the application of either traditional reform pre- for small, island-type economies that have a narrowscriptions, such as comprehensive income taxation, or manufacturing base and that depend heavily on cross-of the latest optimal tax theorem. The discussion indi- border trade.cated that although the institutional aspects of tax ad-ministration are reasonably well understood, the Taxation of Foreign Investmenteconomic dimensions based on the theoretical insightsneed further research. Moreover, the data problems in Attitudes toward foreign investment in developingthis area impede the development of sound economic countries have changed considerably in recent years.advice. For example, the marginal administrative costs Such investments used to be seen as an instrument ofof various tax measures or inspection or compliance- foreign domination and control and were thereforeinducing actions are usually not known. treated with suspicion. This perception is now chang-ing, and developing countries have come to recognizeThe Design oflndirect Taxes that foreign capital can provide positive economicgains, particularly through technology transfers andIn most developing countries, the design of indirect access to world markets. As a result, some developingtaxes affects the ability of governments to raise reve- countries have begun to compete in the provision of

JavadKbalilzadeh-Shirazl and Anwar Shab xxitax incentives to attract foreign capital. In many in- The tax sensitivity of 'Di in developing countries hasstances, however, such incentives boil down to a trans- not been examined empirically in past studies. Evenfer of resources from the host developing country to the relevant empirical literature on advanced nationsforeign treasuries without any special benefit to foreign does not capture the home country tax regime. Further-investors. Thus the taxation of multinationals by a more, the disincentive to invest caused by the taxdeveloping country cannot be assessed in isolation system is usually implicitly measured by an average taxfrom the tax regime of the home country, from tax rate, whereas the incentive to undertake new invest-havens or conduit countries, or from transfer pricing ment depends on the effective marginal tax rate, whichpractices. These factors will have bearing on the tax can deviate substantially from an average tax rate con-sensitivity of foreign direct investment (FD). cept. Shah and Slemrod (see chapter 8) have devisedThe tax sensitivity of FDI has important policy impli- an empirical model to study the relevance of host andcations. If, on one hand, FDi is not responsive to taxa- home country tax regimes to FDI using data on LJ.S.tion, it may be an appropriate target for taxation by the multinational transfers and reinvestments in Mexico.host country, which can raise additional revenues with- The model distinguishes FDi financed by transfers fromout sacrificing any economic benefits that FDI pro- that financed by retained earnings, and it incorporatesduces. If, on the otherhand, the volume ofFDI responds tax and nontax factors for both the host and the homenegatively to taxation, then the host country must trade country, including host country risk factors and theoff the revenue gains of increased taxation against the credit status of multinationals. Both the marginal andeconomic costs of discouraging PDi. the average effective tax rates are incorporated into theThe relevance of host and home country tax regimes analysis. The authors conclude that FDI in Mexico isto FDI transfers and reinvestments is the subject of sensitive to tax regimes in Mexico and in the Unitedconsiderable theoretical controversy. According to the States, to the credit status of multinationals, to countryold view, both tax regimes matter-the home country credit ratings, and to the regulatory environment. Theytax system is relevant even if a subsidiary finances its also find that developing countries in which the degreeinvestment by reinvesting eamnings or by raising local of FDi penetration is large need not worry about pro-debt. This is because its financing and investment de- viding special tax incentives forforeign investment, butcisions affect tax liability at home with respect to the they should ensure instead that their tax system isdistribution of dividends. In the more recent view, in competitive with the home tax regime of a marginalthe case of FDI financed by local debt or reinvested investor who has access to foreign tax credits againstearnings, the home country tax rate is irrelevant. domestic tax liabilities.In chapter 7, Leechor and Mintz challenge the newview. They argue that home country taxes influence the Resource Taxationuser cost of capital even when retained earnings areused at the margin. They also find that foreign firms in Many developing countries bring agricultural in-a typical host country would face substantial variations come into the tax net indirectly, by means of distortingin the user cost of capital because of factors such as (a) taxes on agricultural exports, marketing boards, andthe country where capital is owned; (b) the type of overvalued exchange rates. The possibility ofreplacingorganization (because branches are often subject to these with a nondistorting land tax is discussed byaccrual taxation and subsidiaries subject to exemption Skinner in chapter 9. He examines in some detail theor deferral); (c) the rate of remittance (the higher the advantages and disadvantages of the land tax, both inrate, the higher the weight of home country taxes in theory and in practice, in selected developing coun-determining the user cost of capital); (d) financial pol- tries. He concludes that a land tax is not necessarily aicy (since real interest rates and applicable withholding superior alternative to export taxes for federal govern-tax rates vary across countries, the debt-equity ratio and ment revenues, because it is too inflexible to deal withthe country where the debt is raised have iznportant instability in agricultural incomes and with adrninistra-implications for the user cost of capital); and (e) the net tive and political difficulties. Progressive tax rates onforeign tax-credit position. The authors' calculations landholdings are nearly impossible to administer, hefor Thailand indicate that effective tax rates there vary argues, citing the example of Bangladesh, where thewith the source of funds, the type of organization, and top marginal rate on the wealthiest farmer's land isthe rate of remittance. They also argue that the policy nearly ffty times the minimum rate, although in realityoptions of the host country are usually constrained by there is little or no evidence that rich farmers pay morethe tax rules in capital-exporting countries and by the than three times the minimum rate.strategic behavior of mulinationals. Leechor and Mintz According to Skinner, the record to date suggests thatconclude that, given the international mobility of cap- land taxes have not been effective in attaining nonrev-ital, global tax neutrality is possible only through a enue goals such as (a) transferring resources from thecomprehensive multilateral agreement on the coordi- agricultural to the nonagricultural sectors; (b) discour-nation of capital income taxes. aging inefficient or speculative land use; (c) assisting in

xxii Introduction and Ovenrewland reform; and (d) promoting environmentally sound The Distributional Impact of Taxationland management. A land tax, however, is a suitableinstrument for local government financing because it The incidence of various taxes has been the subjectwould be seen as a benefit tax or simply as a user of considerable debate. The importance of this issue incharge for local public services. tax policy discussions cannot be overemphasized. Theissue is addressed by Shah and Whalley in chapter 1Financial Taxation and by Clarete in chapter 12. Shah and Whalley arguethat, despite decades of research and some obviousMany economnists believe that any strategy for pitfalls, tax incidence analysis for developing countriesgrowth must devote attention to developing financial continues to be based on the same shifting assumptionsmarkets. Most developing countries consider their that are used for developed countries. Taxes are as-banking, insurance, and finance sectors to be lightly sumed to be shifted forward to consumers or backwardtaxed. Chamley argues in chapter 10, however, that the onto factor incomes, in accordance with tax incidencefinancial sectorin many developing countries is heavily work on developed countries ranging from that oftaxed if one looks at both explicit and implicit taxes. Bowley and Stamp to that of Pechman and Okner. ButImplicit taxes include seigniorage, reserve require- the nontax policies and regulatory environments ofments, lending targets at nonmarket rates (earning developing countries are quite different from those ofbelow-market rates), and interest ceilings combined developed countries, with features such as higher pro-with inflation. These taxes are never reported as tax tection, rationed foreign exchange, price controls,revenues in standard national accounts but yield reve- black markets, and credit rationing. According to Shahnues far in excess of traditional taxes. Inflation, in and Whalley, all these features can greatly complicateparticular, is often overlooked as a source of tax revenue. and even obscure the incidence effects of taxes inUsing a partial equilibrium framework, Chamley ar- developing countries. In the case of several taxes,gues that most of the effective taxation of financial when such features are taken into account signs mayinstitutions falls on deposits. Although the revenue be reversed and estimates of incidence substantiallyfrom the taxation of financial assets is difficult to mea- changed from those that would be produced by con-sure because of the complexity of the instruments, their ventional thinking. The authors present calculations forefficiency cost is very large when the rate of taxation is Pakistan on the incidence of selected taxes to substan-greater than 40 percent. At lower rates, say, less than tiate this newer view of tax incidence.20 percent, the efficiency cost is smaller in relation to Clarete uses a general equilibrium framework torevenues. The removal of onerous levels of taxation analyze the interactions between the institutional dis-stimulates financial intermediation, provided such a tortionstypicallyfoundinadevelopingcountryandthemove is seen as a permanent policy change. The results incidence of selected taxes. The institutional distortionsvary depending on the initial conditions (tax rates, level he covers are quantitative import restrictions, the Har-of development, and inflation rate) and the credibility ris-Todaro effects, and foreign exchange rationing. Heof the tax regime in each country. In countries that have finds that the incidence of various taxes is sensitive todeveloped financial markets to a relatively high level the type of institutional distortion in the model. Forand in those that have experienced an annual inflation example, excise taxes are regressive in the presence ofrate in excess of 100 percent, the supply of financial quantitative import restrictions and Harris-Todaroassets is highly responsive to tax changes, provided the labor market distortions but are progressive in thepolicy change is seen as credible. As countries develop presence of foreign exchange rationing alone. Valuea large and sophisticated menu of financial assets, such added taxes range from being almost proportional, ifas those in Thailand and Indonesia, greater possibilities foreign exchange rationing is present, to being slightlyfor substitution emerge and the efficiency costs of progressive, if quantitative import restrictions or thefinancial taxation rise. Most countries in Latin America Harris-Todaro labor market distortions are featured inand Southeast Asia, along with Ghana, Zaire, Uganda, the model.and Somalia in Africa, have either high inflation orsophisticated financial systems. In these countries, the Quantitative Toolsfor Tax PoUcy Analysisreduction in the level of financial assets in the formalsector that is associated with implicit taxes is thought A quantitative evaluation of the impact of changes into outweigh the revenue gains from such taxation. The tax structures is essential for both economic and polit-impact of taxation is estimated to be significantly ical economy reasons. Along with the chapters onweaker in countries with inflation rates of 60 percent incidence analysis, two others present frameworks foror lower. Tanzania, Nigeria, and Zambia are cited as evaluating reform proposals according to their impact onexamples of this weaker association between taxation factor use by varous sectors, aggregate employment,and the accumulation of financial assets. pnces, government revenues, and income distribution.

Javad Kbalilzadeh-Shirazi and Anwar Shah xxduiIn chapter 13 Dahl and Mitra draw on the work done its tax system underwent a major transformation duringby the World Bank to analyze taxes in Bangladesh, the early phase of its dramatic growth. In chapter 15,India, and China and illustrate that applied general Trela and Whalley examine this question using anequilibrium analysis has been useful in addressing a applied general equilibrium model developed earlierwide variety of tax policy questions. The Bangladesh to assess the significance of intersectoral resourcemodel, for example, combines revenue and incidence transfers for Korean economic growth. The Korean taxeffects in a single measure to rank various sectors with system, they point out, has evolved over the years fromrespect to the efficiency-cum-equity cost of raising one that raised small amounts of revenue from narrowrevenue. The China model is used to evaluate whether bases to a broadly based system that yields substantialit is appropriate to recommend to socialist economies revenue. The tax system has been continuouslythat they should adopt broadly uniform tax rates over adapted to serve broader policy objectives (for exam-a large number of sectors, while the India model exam- ple, investment and export promotion). Rebates ofines the coordinated reform of tariffs and indirect taxes. direct and indirect taxes on exports and investment taxDahl and Mitra also discuss the resources required for credits and tax holidays have been liberally used incarrying out such analyses. They argue that establish- pursuit of specific policies. More recently, tax policying a consistent data set is the costliest aspect of mod- has shifted toward neutrality.eling and that computing and software costs are small Trela and Whalley also discuss the significance ofby comparison. They suggest that cost considerations tax-induced intersectoral resource transfers for Koreanmust be weighed against the substantial gains that growth. The modeling results indicate that tax policymodel analyses make possible, notably, the consis- contributes only modestly to Korean economic growth.tency of recommendations and the sound policy deci- It accounted for less than a tenth of the growth duringsions concerning structural reforn. 1962-82, although it did contribute to about 3 percentAtkinson and Bourguignon (see chapter 14) present of export growth. The results described in chapter 15a simple spreadsheet framework, termed the "tax-ben- must be treated as tentative for the model used takesefit" model, which can handle redistributive calcula- only a partial view of the Korean growth process andtions while abstracting from any behavioral responses. does not explicitly take into account savings, invest-This framework uses microeconomic data extracted ment, and the accumulation of human capital. Thefrom household surveys or comparable sources. It ap- authors nevertheless expect the model, once expandedplies to each household in the sample the official in this direction, to reconfirm the conclusion that thecalculation rules for the various taxes and benefits to main factors underlying Korean growth in recent dec-which it may be entitled and derives the resulting ades lie outside of tax policy.distribution of net incomes. The model also providesother important characteristics of the redistributive sys- Conduding Commentstem, such as the effective marginal tax rates that house-holds may be facing, the distribution of individual gains In the concluding session of the conference, a roundand losses associated with a specific reform measure, table discussion reflected upon areas of future researchand any changes in government revenues. The disin- into taxation and the Bank's role in such work. Thecentive effects of the system are incorporated simply Bank, it was noted, is ideally suited to fostering collab-by modifying the household budget constraint in an ad orative research among analysts from developed andhoc manner. The authors summarize the features of the developing countries. A number of topics were singledtax benefit models that are used in the United Kingdom out for such research: the joint analysis and reform ofand France and illustrate how they are applied. They tax and expenditure systems (fiscal reform as a whole);also discuss the possible modifications of these models the fiscal federalism dimensions of tax reform; globalfor use in developing countries and the main features issues of taxation, particularly those related to tax com-of a potential framework for Brazil. petition among developing countries trying to attractforeign direct investment, and to environmental pro-Tax Poltcy and Economic Growtb tection such as carbon tax schemes for controllingemissions of greenhouse gases and their implicationsThere is little disagreement that in theory taxes have for developing countries; the dynamic analysis of tax-some irnpact on economic growth. Other things being ation, in view of the improved techniques for construct-equal, countries with a low tax rate are expected to ing plausible positive models of dynamic economiesgrow faster than those with a high tax rate. Negative that incorporate expectations, learning by doing, andtaxes (incentives and subsidies) stimulate growth. Em- imperfect competition; and the institutional and eco-pirical evidence to this effect is sparse, however, espe- nomic aspects of tax administration.cially for developing countries. Korea provides an The importance of taxadmninistration was repeatedlyinteresting opportunity to look into this question since noted throughout the conference. It is widely recog-

xxiv Introduction and Oteniewnized that the success of tax policy changes depends Deaton, Angus. 1987. "Econometric Issues for Tax Design inheavily on the administrative ability (and, of course, the Developing Countries." In David Newbery and Nicholaspolitical will) to collect revenues through fair and effi- H. Stem, eds., The Theory of Taxation for Developingcient enforcement. All recent efforts at tax reform Countries New York: Oxford University Press.clearly point to simplicity of tax design and ease of Diamond, Peter A., and James A. Mirrlees. 1971. 'Optimaladministration as the fundamental criteria for choosing Taxation and Public Production, Part I: Production Effi-among alternate reform proposals. Unfortunately, no ciency,' and "Part II: Tax Rules." Amercan Economicreliable framework has yet been developed for evalu- Review 61(March and June):8-27 and 261-78.ating the administrative efficiency of alternate policies Feldstein, Martin. 1978. "The Welfare Cost of Capital Incomeor for estimating the relative orders of magnitude of Taxation." Journal of Political Economy 86(April, Partmarginal administrative costs of various administrative 2):S29-S51.measures. Tax administration reform is often contem- Ramnsey, Frank P. 1927. "A Contribution to the Theory ofplated without adequate knowledge of potenitial out- Taxation." EconomicJournal 37(March):47-61.put (the real tax base), the nature of administrative cost Shah, Anwar and Bjom Larsen. Forthcoming. "Carbon Taxes,functions (retums to scale, discontinuities, jointness of The Greenhouse Effect and Developing Countries.' Inproduction, and so on), and compliance reaction WorldDevelopmentReport 1992. New York: Oxford Uni-curves (for example, in an environment where tax versity Press.evasion is pervasive, higher penalties create increased Shah, Anwar. 1988. "Optimal Pricing of Extemalities forincentives for corruption and therefore reduced com- Environmental Protection." In Roy Chamberlain Brown,pliance). Thus additional theoretical and empirical re- ed., Quantityand QualityinEconomics, vol.2. NewYork:search on the efficiency of tax administration can University Press of America.provide better insight into the appropriate path of tax _ _. 1990. "Optimal Pricing of Traffic Extemalities: The-reform. In particular, such understanding could ensure ory and Measurement." InternationalJournal of Trans-that the simplicity criterion in the design of tax reform port Economics 17(1): 3-19.is put in proper perspective and is not overemphasized . 1991a. Perspectives on the Design of Intergovern-at the expense of other policy objectives, including mental Fiscal Relations. PRE Working Paper 726. Worldequity. Bank, Washington, D.C.____. 1991b. Tbe New Fiscal Federalism in BraziL WorldReferences Bank Discussion Paper 124. Washington, D.C.____., ed. Forthcoming a. FiscalIncentivesforlnvestmentThe word wprocessed" describes informally reproduced works in Developing Countries. Washington, D.C.: World Bank.that may not be commonly available through libraries. ., ed. Forthcoming b. Taxes, IncentimesandProductionin Developing Countries Washington, D.C.: World Bank.Asher, Mukul. 1990. "Reforming the Tax System in Indone- Slemrod, Joel. 1990. "Optimal Taxation and Optimal Taxsia." Public Economics Division, World Bank. Processed. Systems." Journal of Economic Perspectmes 4(1):157-78.Atkinson, Anthony B., and Joseph E. Stiglitz. 1976. "The Stem, Nicholas H. 1987. "The Theory of Optimal CommodityDesign of Tax Structure: Direct versus Indirect Taxation," and Income Taxation," In David Newbery and Nicholas H.Journal of Public Economics 6(July-August):55-75. Stem, eds., The Theory of Taxation for Developing Coun-Auerbach, A., and Lawrence J. Kotlikoff. 1989. 'Investment tries New York: Oxford University Press.versus Savings Incentives: The Size of the Bang for the _ _. 1982. "Optimum Taxation with Errors in Adminis-Buck and the Potential for Self-Financing Business Tax tration."JournalofPublic Economics 17(March):181-212.Cuts." In Lawrence J. Kotlikoff, ed., What Determines _ _. 1976. 'On the Specification of Models of OptimumSavings Cambridge, Mass.: MIT Press. Income Taxation."JournalofPublicEconomics6(July-Au-Bulutoglu, Kenan, and Wayne Thirsk. 1991. "Tax Reform in gust):123-62.Turkey." Public Economics Division, World Bank, Wash- Summers, Lawrence H. 1991. 'The Case for Corrective Taxa-ington, D.C. Processed. tion," Address to the National Tax Association and TaxChhibber, Ajay, and Javad Khalilzadeh-Shirazi. 1988. Public Institute ofAmerica Spring Symposium on New DirectionsFinances in Adjustment Programs. PRE Working Paper inTaxPolicyforthe 1990s. Washington, D.C. National Tax128. World Bank, Washington, D.C. Journalvol. XLIV (September): 289-292.Choi, Kwang. 1990. "Tax Policy and Tax Reform in Korea." Zodrow, George L, and Charles E. McLure,Jr. 1988. Implement-Public Economics Division, World Bank, Washington, D.C. ing Direct Consumption Taxes in Developng Countries PREProcessed. Working Paper 131. World Bank, Washington, D.C.

Part I. Tax Reform Expeniences1Tax Reform in Colombia: Process and ResultsCharles E. McLure, Jr. and George R. ZodrowT his chapter analyzes the process and the results of of personal exemptions and deductions, as well as athe multitude of tax reforms that have occurred in form of income splitting, were introduced in 1960 andColombia since about 1960.1 It examines the evolution modified during the 1970s, but were largely eliminatedof Colombian tax policy and evaluates the major re- in 1986. Inflation adjustment in the measurement offorms in terms of the criteria commonly applied to tax capital income was first rejected, but it was later grad-systems by public finance economists-economic neu- ually enacted to the point that the current system pro-trality (especially with respect to capital investment vides for virtually full indexing. Now a move to aaltematives), distributional equity, administrative and comprehensive balance-sheet approach to inflation ac-compliance simplicity, and revenue performance. The counting is planned for 1992. A tax on net wealth waschapter also examines the extent to which the tax instituted in 1935 and was improved substantially inreform process in Colombia has been influenced by the 1988, but it is currently scheduled for elimination inadvice of foreign tax experts, especially by dhe recom- 1992.mendations of the missions headed by Milton Taylor in The net result of all these changes is a tax system that1965, Richard Musgrave in 1971, and Charles McLure is far superior to the one that prevailed inthe late 1950s.in 1988, and by the work of Richard Bird (1970). An The tax base is significantly broader, and thus scoresunderlying theme is that Colombia has heeded the high marks on both the neutrality and the horizontaladvice of foreign experts; as a result, changes in the equity criteria. A comprehensive system of inflationColombian tax structure have often reflected shifts in adjustment implies that real economic income is mea-the prevailing "conventional wisdom" among public sured much more accurately than would be the casefinance economists regarding the relative importance under an unindexed system. Although inflation index-of the criteria listed above. Thus it should be noted that ing adds complexity, many of the features of recentthis appraisal of the various episodes of Colombian tax reforms have resulted in greater simplicity, as efforts toreform is conditioned by current interpretations of the fine tune income measurement have been scaled backrelative importance of these criteria. in order to design a system that can be administeredThe reforms enacted over this period have often more easily. The vertical equity characteristics of thebeen quite sweeping, and in some important cases, current system are more difficult to characterize. Al-represented dramatic reversals of earlier policy deci- though the reduction of rates, the elimination of the taxsions (sometimes corresponding to equally dramatic on dividends at the individual level, and the eliminationshifts in the conventional wisdom on these issues). For of the tax on net wealth have increased incentives toexample, a wide range of investment incentives were work, save, and invest, they probably also made theintroduced in 1960, but were allowed to expire ten system less progressive than it was thirty years ago.years later. Dividends were first made subject to tax in Certainly the desirability of these changes is open to1953, but then were exempted in 1986. A wide variety debate. Finally, despite several design problems, the1

2 Tax Reform in Colombia: Process and Resultscurrent value added tax is an important revenue raiser National Planning Department at the time and theand is clearly superior to the defective sales tax system Musgrave mission.3 Apart from incentive provisions,it replaced. the basic structure of the Colombian tax system afterThe remainder of the chapter is organized as follows. 1960 consisted of the "income and complementaryThe next section provides a brief chronology of the taxes" (the personal and company income taxes, themost significant features of the tax reforms that have tax on individual net wealth, and an excess profits tax)occurred in Colombia over the past thirty years. It and some relatively minor indirect taxes. The mainfocuses on those changes that provide insights into the features of this system can be summarized as follows.process of tax reform in that country. The chapter thenevaluates several of the more recent reforms in terms INDIVIDUAL INCOME TAXES. Labor income was gener-of the criterion of investment neutrality. Specifically, it ally taxable under the individual income tax, althoughsummarizes the results of calculations of the marginal there were a wide variety of exemptions, includingeffective tax rates on capital income under (a) the annual bonuses, vacation and sick pay, severance pay,pre-1986 law, (b) the 1986 law when fully imple- travel allowances, gambling winnings, and the incomemented, and (c) the 1988 law when fully implemented. of Catholic clergy. Capital income was generally taxedA discussion of the distribution of income in Colombia lightly, because interest on government securities andconsiders the distributional effects of various reforms. limited amounts of other interest and dividends wereThe chapter then examines the effects of these reforms exempt, capital gains were treated generously, andon tax simplification, considers the effects of various taxes on dividends and interest were easily avoided.reforms on government revenues, and draws conclu- Interestingly, the implicit capital income on owner-sions about the process and the results of tax reform in occupied housing (in excess of a small exclusion) wasColombia. subject to tax; although such treatment is commonlyviewed as desirable in principle, few countries attemptEpisodes of Tax Reform to tax this source of income. A somewhat unusual andcomplicating feature of the individual income tax wasAlthough tax reform has been an ongoing process in that personal deductions for medical, educational, andColombia over the past thirty years, it is possible to professional services depended on the income andidentify seven major reforms or reform proposals over number of children of the taxpayer. The taxpaying unitthat period-what we might call episodes of tax re- was the individual, but an element of income splittingform. The principal features of these reforms or reform was allowed because a limited amount of income couldproposals are described in the following subsections. be "ceded" to the lower-earning spouse. Foreignsource income, net of foreign taxes, was taxed at theThe Tax System after the 1960 Reform individual level. The marginal tax rate structure wasquite progressive by current standards, with rates rang-The Colombian tax structure in 1960 was typical of ing from 0.50 to 51 percent.those found in Latin America at the time. In 1957 theUnited Nations Economic Commission for Latin Amer- THE NET WEALTH TAX. The individual income tax wasica had called for a highly interventionist strategy to supplemented by a tax on net wealth. (A tax on the netspur econornic development. In line with this recom- wealth of corporations and simnilar entities was elimi-mendation the Colombian tax system provided a vari- nated in the 1960 reform.) The base of the net wealthety of incentives for investment in favored industries, tax was seriously understated, however, because (a)including industries deemed to be of basic importance there were many exemptions, similar to those grantedto development such as iron and steel. under the income tax, (b) real estate was valued at theThe resulting system of pervasive investment incen- relatively low assessments used for purposes of thetives caused an inefficient allocation of scarce capital local property tax, and (c) depreciable assets wereresources, complicated the tax system, made the valued at historical cost after correcting for deprecia-achievement of horizontal and vertical equity more tion.difficult, and had a high cost in terms of revenue lostper additional dollar of investment obtained.2 As a CoMPANY TAxEs. Company taxation varied accordingresult, most of the incentives were allowed to expire to the organizational form of the business. Corporateten years after the 1960 reform was enacted. Indeed, income was subject to company level taxation at ratesColombia was one of the first of the developing coun- of 12, 24, and 36 percent, and dividends were alsotries to reject the highly interventionist approach and taxed at the individual level.4 Relatively lower com-to reduce dramatically its reliance on investment incen- pany tax rates were applied to the income of limitedtives. This change was consistent with the recommen- partnerships (rates of 4, 8, and 12 percent) and thedations of both the foreign advisers working in the income of ordinary partnerships (3 and 6 percent).

Cbarles E. MdLure, Jr. and George R. Zodrow 3Nonetheless, all such income-both dividends and more influential, the discussion here focuses on itsretained earnings-was imputed to the owners of the recommendations; however, several areas of disagree-firm and thus was also subject to taxation at the indi- ment between the two reports (which were broadlyvidual level. Depreciation allowances were not in- similar in most of their recommendations) are alsodexed for inflation and were calculated assuming noted. In addition, it is important to note that rnanystraight line depreciation over lives that ranged from Colombians assisted in the preparation of the reports;five to twenty years. the two missions thus contributed to the developmentUnder the 1960 reform all Colombian income of of a sizable group of Colombian tax professionals whoforeign-based companies was subject to tax at corpo- would have an impact on the evolution of tax policy inrate rates. Repatriated dividends were also subject to a Colombia for many years thereafter.12 percent withholding tax. The income of Colombianbranches of foreign companies was subject to a 6 INDIVIDUAL INCOME TAXES. The Taylor and Musgravepercent withholding tax; this was raised to 12 percent missions were in agreement in recommending full tax-in 1963. ation of most exempt labor income, coupled with se-vere limits on deductions that were not clearly businessExcEss PROFITS TAX. A tax on excess profits was levied related. The Musgrave nission recommended a stan-at rates ranging from 20 to 56 percent on profits in dard deduction and argued that many exemptions andexcess of 12 percent of net wealth (provided wealth deductions should have "vanishing" provisions, so thatexceeded certain amounts). they would gradually be eliminated with increases intaxpayer income. Besides endorsing the ceding of in-INDIREcrTAxEs. Indirect taxes accounted for less than come from one spouse to another, the mission sug-10 percent of federal revenues, with most such revenue gested that taxation of the imputed rental income fromcoming from stamp taxes and stamped paper sales. owner occupied housing should be eliminated, inBecause the bases for these taxes were chosen primar- which case deductions for home mortgage interest andily for administrative convenience and revenue stabil- property taxes should be disallowed. (In contrast, theity, the impact of the taxes was capricious. There were Taylor mission recommended partial taxation of im-no broadly based indirect taxes, such as a value added puted rent and elimination of the income ceding pro-tax or a national retail sales tax.