UNDP-WORLDBAW 22139 TRADE EXPANSION PROGRAM April 1989 OCCASIONAL PAPER4 TRADE RESTRICTIONS AND REFORMS BY DEVELOPING COUNTRIES IN THE 1980s Nadav Halevi This occasional pper is a productor t joirt UNDPMord Bauk Trde ExVpension Pmgra which potvidos echical and policy admce b courtries inng b reon theibade remes. The views twrtaied herein are those o the a=us aid do not nessarily reflect thse of the Lkited Nations Delopnert Pmgran or the World Bank. FILE COPY TRADE RESTRICTIONS AND REFORMS BY DEVELOPING COUNTRIES IN THE 1980s Nadav Halevi Hebrew University of Jeruselem April 1989 Trade Policy Division The World Bank Washington, DC ABSTRACT Despite various trade liberalization attempts by developing countries in the 1960s and 1970s, most developing countries still had restrictive trade regimes in the early 1980s. By the end of the 1970s many developing countries were also in serious economic difficulty: growth rates were low, balance of payments problems were becoming unmanageable, and inflation and unemployment rates were high. Recognizing that serious economic restructuring was a prerequisite to increasing economic growth, many of these countries undertook reform programs. Trade reform was a component of most restructuring programs because of the contribution that trade liberalization can make to achievement of balance of payments viability and economic growth. Countries had imposed a wide variety of restrictions on foreign trade, some for balance of payments reasons and some for protection. A classification of trade restrictiveness in forty developing countries found that for each component examined, at least 40 percent of countries had highly restrictive policies. Two-thirds of the countries had a high overall level of antiexport bias. Protection, whether through quantitative restrictions or tariffs, was high in 85 percent of the cases. Concern over adjustment costs affected trade reform programs. Temporary reductions in output and employment can arise when productive activities suddenly face competition from previously excluded imports. Over time, output and employment will rise as firms improve efficiency or switch to other lines of production, but in the short run adversely affected firms may oppose reductions in protection. Similarly, there may be a short-term deterioration in the current account balance, and government revenues may decline. The fiscal implications of liberalization carry considerable weight, particularly in small economies with a relatively large trade sector and without a sophisticated tax system. Thus, because of concern about possible adjustment costs, most trade reform programs during the 1980s followed a policy of gradualism or multistage liberalization. This sequencing process starts with a reduction in impediments to exports, proceeds to a replacement of quantitative restrictions on imports with tariffs, and only at a later stage undertakes significant reductions in protection levels and differentials. Export incentives are used to reduce antiexport bias remaining during the process of reform. For a sample of twenty-four developing countries that had undertaken trade reform during the 1980s and for which sufficient data were available, more progress was made in reducing impediments to exports than in other elements of liberalization. Many countries adopted more flexible exchange rate policies and export incentives to reduce antiexport bias. There was some minor switching of forms of protection from quantitative restrictions to tariffs, as well as some tariff reform (reducing maximum rates, limiting the number of tariff classes, decreasing the number of exemptions). Actual lowering of protection levels, however, has been very slight. The study reinforced the finding of a World Bank study on earlier liberalization efforts that trade reform is a remarkably drawn-out process, outlasting government administrations and sometimes lasting as long as a generation. None of the countries that had had a highly restrictive trade regime in the early 1980s had moved to a very liberal regime by 1988. Yet many had made substantial progress, and others seem committed to a gradual process of liberalization. In many cases progress was slow (and sometimes was even reversed) because of four main factors: lack of commitment, administrative problems, faulty exchange rate and other macroeconomic policies, and balance of payments problems. TABLE OF CONTENTS Page No. I. Introduction 1 II. Trade Restrictions 2 A. Types of Trade Restrictions 2 B. Restrictiveness of Trade Regimes 4 III. Issues in Trade Liberalization 5 A. Effects of Trade Liberalization 5 B. Gradualism and Sequencing 8 1. Removal of Impediments to Exports 9 2. Replacement of Quantitative Restrictions on Imports 11 3. Reduction in Protection Differentials 12 4. Conclusions on Timing and Sequencing 14 IV. Trade Reform in the 1980s 16 A. Reforms Initiated 16 B. Progress in Trade Liberalization 17 C. Conclusions on Causes of Delay and Reversal 19 Bibliography 20 1. INTRODUCTION The world has become increasing trade conscious in the last forty years. In the 1950s the industrialized countries of Western Europe began a process of dismantling their trade barriers that culminated in the European Economic Community. In the 1960s and 1970s many developing countries attempted trade liberalization, with varying degrees of success.1 But most developing countries, including some that had failed in earlier attempts at liberalization, still had restrictive trade regimes in the early 1980s. By the end of the 1970s many developing countries were in serious economic difficulty; growth rates were very low, and balance of payments problems, often arising from worsening terms of trade or overwhelming foreign debt, were becoming unmanageable. In addition, many countries suffered from high inflation rates or severe unemployment. Recognizing that serious economic restructuring to improve the long-run balance of payments was a prerequisite to attaining economic growth, many of these countries undertook reform programs, usually in conjunction with financial support for this purpose from the World Bank. The restructuring programs usually included elements of trade liberalization, although the restructuring was in many cases more-far reaching. Trade liberalization, while not a panacea, helps make the price mechanism a more efficient allocative instrument and may help reduce macroeconomic instability. International competitiveness and the attainment of balance of payments viability are two sides of the same coin. I am indebted to Vinod Thomas and Jaime de Melo for helpful comments in the design of this paper and earlier drafts, to Donald B. Keesing for improvements in the classification of restrictiveness, and to Marinela Dado, Shai Peled, and Zohar Haimovitz for aid in collecting and analyzing the data. I remain, however, solely responsible for the opinions and conclusions presented. 1 The results of a major World Bank study of the liberalization experiences of nineteen developing countries are forthcoming in a series of volumes entitled Liberalizing Foreign Trade. Preliminary findings have been presented in Papageorgiou, Michaely, and Choksi (1987) and Hichaely (1988). - 2 - Freer trade2 affects two other components of growth: (1) domestic saving and investment and foreign investment increase as a result of a more favorable climate for economic activity, and (2) with foreign investment, prospects for technological innovation improve. Empirical studies give ample evidence that open, outward-looking developing economies have better growth records than more restrictive ones.3 For this reason, trade reform has been stressed in many restructuring programs. This study presents a brief survey of the trade restriction systems common in developing countries and the attempts at trade reform undertaken in the 1980s. Section II discusses the main types of trade restrictions and classifies levels of restrictiveness for a sample of forty countries. Section III considers problems of liberalization and their implications for desirable sequencing of reform. Section IV surveys the trade reforms initiated by the forty countries in the sample and the implementation so far achieved by a subsample of twenty-four countries for which enough time has elapsed and sufficient data are available for reaching some conclusions. II. TRADE RESTRICTIONS A. Types of Trade Restrictions In an ideal world countries would interfere minimally with free trade. The rate of exchange would be sufficient for comparing the prices of tradables and nontradables and would not discriminate between different types of tradable goods and services, between different types of exports and imports, or between exports and imports. Thus allocation of resources would be along the lines of comparative advantage. In practice governments apply a wide variety of restrictions to foreign trade. Sometimes restrictions are imposed for balance of payments 2 The term 'freer trade, rather than 'free trade" is used because complete free trade is neither theoretically justifiable (exceptions can always be found) nor practically feasible. 3 See, for example, Little, Scitovsky, and Scott (1976); Bhagvati (1978); Krueger (1978); Donges (1976); and World Bank (1987). reasons and are designed as substitutes for exchange rate adjustments. Their main purpose is to discriminate between tradables and nontradables, although restrictions are often designed to create intratradable discrimination between exports and import substitutes and between different products. Frequently, measures adopted for balance of payments reasons create a high degree of distorting discrimination. Intentionally discriminatory restrictions -- protection measures -- may have some economic justification, such as correcting for other internal market distortions, improving the terms of trade, and most frequently, promoting infant industries. Although justifiable cases can be made, however, trade intervention is hardly ever the best corrective measure. Protection tends to be self-perpetuating: these infants do not grow up; instead, they create powerful vested interests that resist reform.4 In discussing trade restrictions, a useful distinction is that between administrative restrictions and intervention via the price mechanism. Administrative restrictions include prohibition of particular exports or imports, and quantitative restrictions applied through licensing arrangements or foreign exchange controls and rationing. Often less obvious but equivalent measures are used, such as labeling requirements, origin and content rules, and state trading or publicly sanctioned monopolies. Price intervention includes taxes (and subsidies) on traded goods, the most common of which is the tariff. Often, various nontariff taxes are used with equivalent effect, such as special levies, reevaluation of imports for tax purposes, and import deposit requirements. Taxes on trade are among the easiest to collect, so their use may be based more on revenue considerations than on their effects on trade. Quantitative restrictions are a more insidious form of protection than tariffs because they are nontransparent. Even the public official who approves them rarely knows just how high a level of protection is being granted. Tariffs are more transparent, but not completely so. Since imported inputs usually are taxed less than imported final goods, the level of effective protection is not apparent. In terms of allocative 4 For a succinct recent summary of the issue of protection, see Corden (1987). -4 efficiency, the degree of dispersion of effective rates is often much more harmful than the average protection level. B. Restrictiveness of Trade Regimes Details of consultations between the World Bank and developing countries on loans for economic restructuring provided information on trade practices for forty countries in the early and mid-1980s. Although this is obviously a selective sample, the large number of countries included and their broad geographic distribution make possible some general observations on the intensity and types of trade restrictiveness. Using this information, an attempt was made to classify each country's level of trade restrictiveness into one of three categories, 'high,6 'medium,* or 'low,' based on six specific components: restrictions (administrative or taxes) on exports; restrictions on imports of inputs for exports; quantitative restrictions on imports of goods not produced domestically (that is, restrictions for balance of payments purposes rather than protection); protective quantitative restrictions (that is, on goods that compete directly with domestic production); protective tariff levels; and tariff dispersion. An overall restrictiveness category was added, reflecting the degree of antiexport bias created by the various restrictions. For example, although South Korea was classified as having a high level of quantitative restriction protection, its special export promotion measures prevented overall antiexport bias. Table 1 summarizes the classifications for the forty countries, showing the percentage of countries in each category of restrictiveness. In some cases, such as foreign exchange control or export licensing systems, classification was based on quantitative data. Usually, however, partial quantitative information and impressions regarding the importance of the restrictions formed the basis for classification. Thus the classification is highly subjective. Nonetheless, while there may be room for disagreement on a particular item or country, the overall picture on trade restrictiveness is probably fairly accurate because of the care in selecting indicators. - 5 - Table 1: Distribution of Forty-Country Sample by Intensity of Trade Restrictiveness in the Early to Mid-1980s (percentages) Intensity of Restrictions Restrictions High Medium Low On exports 40.0 35.0 25.0 On imports for export 42.5 52.5 5.0 Nonprotective quantitative restrictions 55.0 25.0 20.0 Protective quantitative restrictions 70.0 20.0 10.0 Tariff level 52.5 47.5 0.0 Tariff dispersion 45.0 50.0 5.0 Overall antiexport bias 65.0 30.0 5.0 Only two countries (Chile and South Korea) had a low level of antiexport bias, while two-thirds of the countries had a high level. For each item examined at least 40 percent of countries were classified as having highly restrictive policies. Protection, whether by quantitative restrictions or tariffs, was high in 85 percent of the cases. Latin America had a somewhat larger percentage of highly restrictive practices than other regions. Somewhat surprisingly, however, a high level of protection was found even in many countries with a low level of industrialization. This finding may reflect the use of quantitative restrictions for balance of payments purposes, with the resulting protection being incidental. III. ISSUES IN TRADE LIBERALIZATION A. Effects of Trade Liberalization In theory the effects of trade liberalization can be studied under ceteris paribus assumptions. The main adjustment costs that may require attention are unemployment, balance of payments difficulties, income distribution effects, and fiscal losses. - 6 - Temporary reductions in output and employment can arise when productive activities suddenly face competition from previously excluded imports. Over time many firms can adjust, either by improving efficiency or by switching into lines of production in which they are more competitive, but in the short-run some unemployment can be expected.5 More rational relative prices can encourage increased output and employment in other activities. This is particularly the case for exports, if liberalization reduces antiexport bias and removes other impediments to exports. Thus on the macro level net unemployment need not be large; in the longer run the net effect on employment can be expansive. The main opponents of reduced protection, however, are firms concerned with immediate employment and the output effects on the micro level. The long- and short-run balance of payments effects may differ and cannot be analyzed without regard to exchange rate policy. Tariff reductions by themselves lower the effective exchange rate for imports, leading to increased imports. Exports may expand after removal of impediments, but their longer term development (as well as that of imports) will depend on what happens to the real exchange rate. Since exports will probably respond more slowly than imports, the short-term effects of liberalization usually include a deterioration in the current account balance. Any real structural change will lead to changes in income distribution. Obviously, industries hit by new competition will suffer, and even if, as expected, the net effects on national welfare are positive, the redistributive effects between urban and rural areas and between old and new industries may be large enough to have social and political consequences that policymakers cannot ignore.6 How the reduction in import tariffs will affect tariff revenue depends on the price elasticity of demand for imports. Thus, either 5 The adjustment experience of firms in the Southern Cone countries of Latin America is documented in Corbo and de Melo (1985). These studies clearly indicate that efficiency improvements often involve the use of less labor in existing activities. 6 For a survey of these aspects of tariff policy, see Baldwin (1982). - 7 - revenue will fall or it will increase and the import side of the balance of payments will worsen. If tariff reductions are replaced by a currency devaluation (as may be the case where temporary tariffs or surcharges were used while devaluation was postponed), tariff revenue will decrease even when total expenditures on imports do not change. Removing export taxes will also reduce revenue. The fiscal implications of liberalization carry considerable weight, particularly in small economies with a relatively large trade sector. For many developing countries without sophisticated tax systems, trade taxes are a major source of government revenue. Fiscal problems are inversely related to the level of development: the fiscal importance of trade taxes decreases at higher levels of development, and substitute taxes are easier to impose. Obviously, the intensity of the adjustment problems that may arise from liberalization will vary from country to country. The amount of adjustment unemployment, and the accompanying effects on income distribution, will be greater where inefficient industries have developed over a long time behind high protective barriers. The difficulties of adjustment and the opposition it sparks may slow the pace of adjustment. Yet countries may use trade liberalization as part of a restructuring process to set a more rational path for future industrialization. Thus, adjustment problems are greater where the level of economic development is further advanced. The negative balance of payments effects of liberalization, and to some extent the temporary employment effects, can be offset, at least theoretically, by suitable exchange rate and macroeconomic policies. But trade liberalization usually is initiated as a response to crisis conditions, which may include severe balance of payments problems, inflation, or unemployment. Consequently, the accompanying economic policies necessary to make liberalization work may not be easily applied. Preliminary findings of a major World Bank project on the trade liberalization experience of nineteen developing countries (hereafter referred to as the World Bank trade liberalization study) show that the biggest single cause for the termination of trade liberalization is balance of payments problems (Papageorgiou, Michaely, and Choksi 1987; Michaely 1988). Sometimes these problems are intensified by external factors (such as rising oil prices or falling prices for other primary products), but - 8 - more often they are due to inadequate demand contraction and faulty real exchange rate management. Trade liberalization will worsen the current account balance unless accompanied by real exchange rate depreciation, yet in many cases the real exchange rate has appreciated following liberalization, sometimes because exchange rate policy was used to reduce inflation. In some cases simultaneous capital account liberalization has caused reverse movements in the nominal exchange rate.7 In other cases failure to implement adequate anti-inflationary policies (for example, by allowing wage increases or failing to cut government fiscal deficits sufficiently) caused domestic prices to rise more than the nominal depreciation. In some cases -- the Argentinean tablita is a prime example -- using an announced exchange rate adjustment as an anti-inflation device led to appreciation of the real exchange rate. Thus many studies have placed the blame for liberalization failures on faulty macroeconomic policy (see Krueger 1978). Implementing liberalization and economic stabilization simultaneously has been quite difficult. Perhaps the most closely studied examples are the Southern Cone countries of South America.8 Their experience, according to Corbo and de Melo (1987), lends support to the videly held view that it is pointless to attempt liberalization prior to stabilization in cases where the rate of inflation is above 25 percent (see Fischer 1986). B. Gradualism and Sequencing9 Since countries differ in their initial conditions, the intensity of their liberalization-induced problems, and their ability to deal with 7 For a discussion of the arguments in favor of deferring capital account liberalization, see Edwards (1986). 8 See, for example, the detailed studies in the World Bank trade liberalization study by de la Cuadra and Hachette, Cavallo and Cottani, and Favaro and Spiller (on Chile, Argentina, and Uruguay, respectively) and the briefer comparison of the three cases in Corbo and de Melo (1987). 9 This section draws heavily on the preliminary findings of the World Bank study on trade liberalization in nineteen developing countries. -9- them, they also differ in the speed of liberalization. Countries attempting liberalization have adopted either a multistage approach or drawn-out gradualism. One somewhat surprising finding of the World Bank study on trade liberalization is how long the process has usually been extended -- from a minimum of several years to as long as a generation. Even in some successful cases of liberalization10 fear of adjustment costsll has led to overcautious gradualism. Gradualism or multistage liberalization implies differentiating between the various components of trade liberalization. In particular it means distinguishing between those aspects with adjustment costs and those without. Considering components separately helps in the evaluation of each component's best position in a sequencing of liberalization, thereby enabling governments to implement first those measures that provide immediate benefits with minimum adjustment costs. The benefits of this approach are efficiency improvements and continuity of the liberalization process. But since the major adjustment costs arise from actually decreasing protection and thereby making domestic producers face international competition, gradualism implies postponement of liberalization's main objectives. Three of the components of liberalization are considered below: removing impediments to exports, relying on the price mechanism rather than administrative intervention, and reducing protection levels and differentials. In general this is the order of the desired sequencing of liberalization, but no definitive time dimension is suggested for each step in the sequence. 1. Removal of Impediments to Exports Given that foreign exchange problems are the most frequent reason for reversing trade liberalizations, measures that lead to quick export 10 Of the nineteen countries studied only eight can be classified as successful: Chile, Greece, Indonesia, Israel, Korea, New Zealand, Spain, and Singapore. 11 Only in the case of Chile were negative unemployment effects substantial, and even there the negative effect in manufacturing was compensated for by increased employment in agriculture (see Michaely 1988). - 10 - expansion are especially desirable. Furthermore, if export expansion increases employment opportunities, liberalizing imports will be easier. These factors have led some economists to adopt the extreme position of advocating a separate, early stage of export promotion prior to import liberalization.12 Host of the World Bank studies in the nineteen-country trade liberalization review express a clear desire for removing antiexport bias but indicate no clear preference for a proexport bias.13 While removing antiexport biases may be a drawn-out process, removal of impediments to exports gives immediate results and can be implemented more readily. Two administrative impediments to exports are quantitative restrictions or outright prohibition of exports and quantitative restrictions on imported inputs used in export production. Quantitative restrictions on exports, which are frequently imposed to guarantee a minimum domestic supply of essential production, are often anachronistic and in most cases can be eliminated at no real cost. Quantitative restrictions on imported inputs are a more complicated issue, because it is not easy to determine with certainty that an imported input is used for export production and not for domestic production. So if import controls continue, some means of distinguishing between uses of inputs is necessary. Quantitative restrictions and equivalent foreign exchange restrictions on imported inputs can be removed and replaced by a tariff drawback system to prevent misuse of inputs. Such a system should be efficient, in the sense of not imposing significant costs on bona fide inputs for exports; however, in most cases the systems used are very inefficient.14 12 Lin (1987) argues that it is the prior concentration on export promotion that differentiates the more successful East Asian experience with liberalization from the less successful Latin American experience. 13 Bhagwati and Krueger (1974) concluded that proexport bias has fewer distortionary effects than antiexport bias. 14 Alternative systems used are bonded warehouses or customs-free industrial zones. - 11 - This important first step in liberalization shows that partial measures do not eliminate all distortions and may create new ones. To the extent that inputs are produced domestically, removing barriers to import such goods reduces their protection and creates additional dispersion in domestic protection rates. In many developing countries the leading primary exports are taxed, often indirectly through sanctioned export monopolies (or Imarketing boards") that pay producers less than world prices for their products. Such monopolies may be justified to ensure quality control or to improve the terms of trade by exploiting monopoly power in foreign markets. Exercise of monopoly power in foreign markets is less widespread than generally believed, however, and even when marketing boards try to use it, they rarely judge the market correctly. The more likely outcome is unduly restricted output (see Sanchez-Ugarthe and Moti 1986). In most cases foreign exchange receipts would rise if differentials between producers prices and foreign prices were eliminated. Nonetheless, where export taxes are a big revenue source, they will not likely be reduced until alternative sources of revenue are found,15 and this may be a complicated, time- consuming process. Consequently, some countries may continue taxing exports while implementing other stages of liberalization. 2. Replacement of Quantitative Restrictions on Imports Import limitation by administrative procedure is popular because it is a relatively quick and easy way to deal with a balance of payments problem; however, the use of an equally prohibitive but more transparent price mechanism is preferred. This holds true for controls on imports that compete with domestic production and on noncompetitive imports, including ,luxuries' (whose production will be encouraged if protection is high enough). The price alternatives to quantitative restrictions are exchange rate adjustment and tariffs. Exchange rate adjustments are made for reasons other than trade liberalization, but they can eliminate the need 15 Partial reduction of taxes may increase tax revenue, depending on the elasticities. - 12 - for those quantitative restrictions that were imposed (or reimposed) as import-reducing alternatives to depreciation.16 In practice, replacing quantitative restrictions with tariffs may proceed in several often overlapping steps and may follow changes in the form of quantitative restrictions from positive lists (whereby imports are prohibited unless specifically allowed) to negative lists (whereby imports are allowed, subject to quotas, unless specifically banned). First, tariffs can be applied to imports that do not compete with domestic production, thereby expropriating quota rents. These tariffs are then raised until the quotas are no longer effective constraints; quotas then may be eliminated (although they are sometimes retained as insurance). At a later stage, and often only gradually, quantitative restrictions on competitive imports may be replaced by equally protective tariffs.17 Uncertainty about the necessary level of protection often leads to higher than necessary duties, or swater in the tariff.* To the extent that changing the form of protection does not cause protection levels to fall, no adjustment costs result from this process. If policymakers are planning rational reductions in protection levels, however, they must switch to visible and quantifiable levels of protection (although this stage is often prolonged unjustifiably). The World Bank trade liberalization study found that sustainable liberalization became more likely the more dramatic the switch in the form of protection, regardless of when the lowering of protection levels began. 3. Reduction in Protection Differentials Trade taxes (or subsidies) involve differentials between tradables and nontradables and among different tradables. The main mechanism for establishing the relative price of tradables and nontradables is the 16 Edwards (1987) gives a number of such examples from recent Latin American experience. 17 The World Bank trade liberalization study found many cases in which relaxation of quantitative restrictions was accompanied by a lowering of protection levels; the study also found that the distinction between protective and nonprotective quantitative restrictions was rarely made. Another transitional step may be use of tariff-quotas (see Takacs 1989). - 13 - exchange rate. Trade taxes create a multiple effective exchange rate system, even if the nominal exchange rate is uniform. Among tradables, taxes are more rarely imposed (or subsidies given) on services than on comodities. In addition, the protection offered to these categories differs. Because inward-looking policies give higher average protection rates to import substitutes than to exports, drastic reduction of antiexport bias is an objective of trade liberalization. Because import taxes, tariffs, and equivalent instruments are not usually uniform (they are higher on final products than on intermediate goods, and often higher on consumer than on capital goods), the effective protection system is escalated. Although tariffs usually discriminate between types of products, complex tariff structures and interindustry differences in imported input use create a schedule of protection levels often far different and more extensive than originally intended. While exact measures of effective protection rates are extremely useful for developing rational tariff policies, even crude estimates of average import coefficients can indicate whether there are extreme dispersions in protection rates and where the distortions lie. Effective protection is measured as a percentage tax, that is, relative to the existing exchange rate. Consequently, real depreciation of the exchange rate accompanied by reduced tariffs decreases the variance among protection rates and the discrimination between competitive and noncompetitive imports. An increased real effective exchange rate for exports reduces antiexport bias and raises the average price of tradables relative to nontradables without lowering the absolute level of effective protection for import substitutes. While negative adjustment costs are avoided, so too are the efficiency gains imposed on import substitutes by foreign competition. Clearly, liberalization with exchange rate adjustments can attain most of the objectives of liberalization. But if real exchange rate adjustment is not possible, should lowering protective tariffs and reducing antiexport bias be postponed? Where adjustment costs are expected to be low (because the level of industrialization is still low or the economy has great flexibility and industrial mobility), reductions in protection levels may be swift. - 14 - Otherwise gradualism in reducing protective tariffs may be necessary. One way to do this (the concertina method) is to first reduce tariffs in sectors with the highest rates and then gradually to reduce the remaining tariffs from the top down. This method reduces the dispersion of rates and leads to greater efficiency. One important issue in the process of reducing protection differentials concerns export subsidies. Should antiexport bias be allowed to persist when reduction in the protection of import substitutes (or depreciation) is for some reason delayed? Theoretically, the use of export subsidies (preferably uniform rates for value added in foreign exchange) is a desirable, less-than-best solution. But the strong anti-export-subsidy bias prevalent in international organizations (such as the General Agreement on Tariffs and Trade and the International Monetary Fund) makes this temporary solution difficult to implement. 4. Conclusions on Timing and Sequencing Although country differences in initial economic, social, and political conditions make every case special and prevent the establishment of any general rules on timing and sequencing, certain rules of thumb may nonetheless be helpful. Aspects of liberalization that involve only slight adjustment costs can be implemented as fast as technical factors permit. These include removing impediments to exports and simplifying quantitative restrictions on imports, replacing quantitative restrictions on imports with tariffs,18 and rationalizing tariffs, that is, reducing dispersions in nominal and effective protection. These measures should be harmonized with exchange rate adjustments if possible, but they need not be contingent on them. A prerequisite for removal of export taxes may be development of alternative revenue sources. Thus the speed of this phase of liberalization is largely independent of other elements of liberalization. The fiscal impact can be alleviated by replacing quantitative restrictions on noncompetitive imports with nonprotective purchase taxes, thereby 18 In the case of competitive imports, vested interests fearful of the next step in liberalization may force delays. - 15 - transferring quota profits to the government. The most important step in liberalization is reducing antiexport bias, which, however, may entail adjustment costs. Costs and implementation time can be held to a minimum in three instances: 1. In countries at a low stage of industrial development unilateral reductions in protective tariffs will primarily affect future industrial development. 2. Where efficient, nondiscriminatory export subsidies can be applied, they can temporarily reduce antiexport bias without lowering protection levels for import substitutes. 3. A combination of tariff reductions and exchange rate adjustments lowers antiexport bias. But real devaluations depend on the careful coordination of monetary, fiscal, and wage policy with exchange rate management. Circumstances will determine whether this is possible. In addition, the ability to coordinate macropolicy and the severity of existing problems will determine whether tariff and quota reductions should be used to lower real protection levels in order to force existing industries to become more productive. If none of the above cases apply, only unilateral reduction in protection levels may be possible. This will incur greater adjustment costs and will likely produce pressure to draw out the reform process, which has two major drawbacks. It allows the adversely affected parties to organize resistance to reform, and it leaves room for doubt about the sustainability of reform. Crucial to reform success is a convincing preannouncement. Once economic agents believe the government's commitment to the process, they will adjust in the desired direction. - 16 - IV. TRADE REFORM IN THE 1980s A. Reforms Initiated Our information on reforms initiated is biased because our sample was restricted to countries that agreed to undertake at least some reform of their trade regime. But there was considerable variety in the reform programs agreed to: some encompassed wide-ranging transformation of the entire trade regime, whereas some merely eased slightly the restrictions on exports or on imported inputs for exports. A subjective classification by intensity of reform agreed upon was attempted, with each agreed-upon reduction in restrictions classified as 'significant,' less significant,' or anegligible.8 As in the earlier classification of countries according to restrictiveness of trade regimes, the general picture should not be affected by disagreement over specific items. The distribution is shown in Table 2. Table 2. Distribution of Forty-Country Sample by Extent of Agreed-Upon Reductions in Restrictions (percentages) Extent of Reduction of Restrictiveness Restrictions Significant Less Significant Negligible On exports 37.5 37.5 25.0 On imports for export 42.5 37.5 20.0 Nonprotective quantitative restrictions 32.5 42.5 25.0 Protective quantiative restrictions 32.5 40.0 27.5 Tariff level 17.5 52.5 30.0 Tariff dispersion 17.5 60.0 22.5 Overall antiexport bias 42.5 32.5 25.0 Clearly, initial agreement on the desirability (or feasibility) of reducing impediments to exports and switching forms of protection was greater than agreement on an actual reduction of protection levels. In fact, of the thirty-four countries classified as having a high level of - 17 - protection (quantitative restrictions or tariffs), only twelve (35 percent) agreed to significant reductions in protection levels. Some countries (for example, Jamaica, Mexico, and Turkey) did agree, however, to adopt programs for a drastic, if gradual, dismantling of highly restrictive trade regimes. B. Progress in Trade Liberalization More important than agreeing on reforms is actually implementing them. In the case of many of the forty countries in the sample not enough time had elapsed since initial reforms were undertaken or not enough information was available to allow any judgement on the extent of liberalization achieved so far. Enough information was available, however, for a subsample of twenty-four countries to provide some basis for comments on implementation of reform measures. The experience of the developing countries in the 1980s reinforces the conclusion of the earlier World Bank trade liberalization study that liberalization is indeed a drawn-out process: none of the countries that had a highly restrictive regime in the early or mid-1980s had moved to a very liberal regime by 1988.19 Yet many had made substantial progress, and others seem to be committed to a gradual process of liberalization. The most significant improvements have been in the reduction of impediments to exports: licensing was abolished and prohibitions were removed in a number of countries, export taxes were lowered, and improved systems were introduced to reduce the disparity between international and producer prices. What is even more significant, many countries have adopted more flexible exchange rate adjustment and export incentives to reduce antiexport bias.20 Many countries have eased quantitative restrictions and foreign exchange controls for imported inputs for exports. Many have attempted to improve duty drawback systems, but these systems have not usually been 19 Chile, which was relatively liberal at the start of the period, further reduced its uniform tariff level, while Korea, which had little antiexport bias to begin with, effectively dismantled most of its protection. 20 Some countries, for example Senegal and COte d'Ivoire, used subsidies on export value added as substitutes for exchange rate adjustment. - 18 - efficient. Duty exemption systems, which were tried in some cases, have been more successful. So, too has been the use of bonded warehouses or customs-free zones, but those schemes have usually been limited in scope. Switching from quantitative restrictions on noncompetitive imports to tariffs has generally been slow, but several countries (for example, Jamaica, Senegal, and Turkey) have made substantial progress, and most others have made at least partial progress. Switching the form of protection from quantitative restrictions to tariffs has been generally slower and less significant progress has been achieved, although some countries have made considerable progress (for example, Madagascar and Mauritius). Most countries have adopted tariff reform programs, but actual changes have been limited. Progress has frequently been made in reducing maimun tariff rates, limiting the number of tariff classes, imposing a (low) minimum tariff, and decreasing the number of tariff exemptions. Actual lowering of protection levels has generally been slight, however, and most systems still retain an escalated tariff, that is, higher tariffs on consumer goods than on capital goods, with the lowest rates (and exemptions) on intermediate goods and raw materials. Thus, although tariff dispersion has usually been reduced, a discriminatory effective protection system remains. The net result has been some reduction in antiexport bias, but mainly as a result of exchange rate adjustment, export incentives, and reduced impediments to exports, and very little as a result of reduced protection levels for domestic production. Substantial progress has been made in overall reform in a number of countries, including Ghana, Jamaica, Mauritius, Mexico, Morocco, and Senegal, but often at a slower pace than anticipated. In some other countries such as Colombia, C6te d'Ivoire, Madagascar, the Philippines, and Thailand progress achieved so far has been more modest and slower than expected, while in others such as Bangladesh, Kenya, and Togo progress has been very slight. Turkey is an example of a country in which the limited liberalization achieved so far belies the broad extent of reform adopted and liberalization undertaken. There are a number of examples of outright failures. Guyana and Yugoslavia both agreed to some trade liberalization but have not - 19 - implemented any liberalization measures, presumably for lack of ideological commitment. Zambia made substantial reforms between 1984 and 1987, but these were reversed because of balance of payments problems. Malawi, too, reversed in 1986 some reforms adopted earlier. And there are numerous cases of at least partial backsliding. C. Conclusions Regarding Causes of Delay and Reversal Four main factors influence the degree of success achieved by the majority of countries implementing more liberal trade policies: administrative problems, accompanying macroeconomic policies, fundamental structural problems, and commitment to reform. Administrative complications, such as changes in legislation, new classification systems, and studies of the protection system itself, frequently caused delays. In general, reform programs were too optimistic about the amount of time needed for implementation. Other problems were lack of administrative talent, too much bureaucracy, and difficulties in implementing accompanying macroeconomic policies. The most frequent cause for delay in liberalization was implementation of appropriate macroeconomic policies. Inflation, lags in depreciating nominal exchange rates, and terms of trade reversals (especially for Zambia) were the main problems. The main reason for reversal of liberalization was balance of payments problems. Fundamental economic reorganizations and export diversification take time. Appropriate macroeconomic policies help, but terms of trade reversals can present insuperable obstacles. The ability to finance temporary external imbalances is also crucial; otherwise, long-term balance of payments viability may be incompatible with short-term liquidity, especially for countries that are beginning from a precarious position. Political commitment is essential for continuous and sustainable trade liberalization. The process of liberalization itself is longer than the lifespan of many political regimes. If a program is adopted because of external pressure, without real commitment, implementation will not take place or will be reversed. Opposition from sectors fearing negative consequences from liberalization can postpone implementation indefinitely. - 20 - BIBLIOGRAPHY Bhagwati, J. 1978. Foreign Trade Regimes and Economic Development: Anatomy and Consequences of Exchange Control Regimes. Cambridge: Bellinger. Bhagwati, J., and A.0. Krueger. 1974. Trade Strategies for Economic Development: The Asian Experience. New York and Manilat National Bureau of Economic Research and Asian Development Bank. Baldwin, R.E. 1982. 'The Political Economy of Protectionism.' In J. Bhagwati, ed., Import Competition and*Response. Chicago: University of Chicago Press. Corbo, V. 1987. "Lessons from the Southern Cone Policy Reforms.w Research Observer 2 (July). Corbo, V., and J. de Melo. 1985. 'An Overview and Sumary: Liberalization with Stabilization in the Southern Cone of Latin America.' World Development 13 (August). Corden, W.M. 1987. Protection and Liberalization: A Review of the Analytical Issues. Occasional Paper No. 54. Washington, D.C.: International Monetary Fund. Donges, J. 1976. 'A Comparative Study of Industrialization Policies in Fifteen Semi-Industrial Countries." Weltwirtschaftliches Archiv. Edwards, S. 1986. "The Order of Liberalization of the Current and Capital Accounts.' In A.M. Choksi and D. Papageorgiou, eds., Economic Liberalization in Developing Countries. Oxford: Basil Blackwell. Edwards, S. 1987. 'Exchange Controls, Devaluation and Real Exchange Rates: The Latin American Experience.* Paper presented at conference in Bogota. Fischer, S. 1986. 'Issues in Medium-Term Macroeconomic Adjustment." World Bank Economic Observer 1(2): 163-82. Guttman, L. 1986. 'Coefficients of Polotonicity and Monotonicity.' Encyclopedia of Statistical Sciences. Vol. 7. New York: Wiley. Khan, M.S., P. Montiel, and N.U. Haque. 1986. Adjustment with Growth: Relating the Analytical Approaches of the World Bank and the IMF. World Bank Discussion Paper. Washington, D.C.: World Bank. Krueger, A.0. 1978. Foreign Trade Regimes and Economic Development: Liberalization Attempts and Consequences. Cambridge: Ballinger. Lin, C. 1987. "East Asia and Latin America as Contrasting Models.' IMF Working Paper. Washington, D.C.: International Monetary Fund. - 21 - Little, I., T. Scitovsky, and M. Scott. 1976. Industry and Trade in Some Developing Countries: A Comparative Study. London: Oxford University Press. Michaely, M. 1988. 'Trade Liberalization Policies: Lessons of Experience.' Paper presented at the conference in Sao Paulo, Brazil, for a New Policy Toward Foreign Trade in Brazil, April 1988. Washington, D.C.: World Bank. Papageorgion, D., M. Michaely, and A.M. Choksi. 1987. IThe Phasing of a Trade Liberalization Policy: Preliminary Evidence.' World Bank Discussion Paper. Washington, D.C.: World Bank. Sanchez-Ugarthe, F., and J.R. Moti. 1986. 'Are Export Duties Optimal in Developing Countries: Some Supply Side Considerations.' Departmental Memorandum, International Monetary Fund (DM/86/40). Takacs, W. 1989. 'Alternative Transitional Measures for Liberalizing Quantitative Restrictionst Implications for Developing Countries.' UNDP-World Bank Trade Expansion Program Occasional Paper No. 3. Washington, D.C.: World Bank, Trade Policy Division. World Bank. 1987. World Development Report. Washington, D.C.: World Bank.