Uganda - An agenda for trade liberalization

26414 May 1990 UGANDA An Agenda for Trade Liberalization I ... UGANDA AN AGENDA FOR TRADE LIBERALIZATION Interne.tional Monetary Fund JUL 2 <>1992 Trade Policy Division Country Economics Department The World Bank May 1990 」胆……d一―一ーーーーー国■■園国国ー国ーー UGANDA: AN AGENDA FOR TRADE LIBERALIZATION May 1990 This report was prepared under the Trade Expansion Program of the UNDP/World Bank and is based on the work of two missions to Uganda. The report was discussed with representatives of various levels of the Uganda Government in May 1990. The study team was led by Kazi M. Matin (World Bank, Trade Policy Division), and included Sandy Cuthbertson, David Vincent (consultants), and Jaber Ehdai (World Bank, Public Economics Division) TABLE OF C 0 NTENTS EXECUTIVE SUMMARY 1 CHAPTER 1 INTRODUCTION 1-1 1.1 The Main Questions 1-2 1.2 The Approach 1-4 CHAPTER 2 MACROECONOMIC POLICIES AND THE INCENTIVE STRUCTURE 2-1 2.1 The Early Years 2-2 2.2 Adjustment in the Early 1980s 2-8 2.3 Stabilization and Liberalization in the Late 1980s 2-10 Real Exchange Rate 2-11 Devaluation and Inflation 2-13 Liberalization and Capital Flight 2-19 CHAPTER 3 REVENUE IMPLICATIONS OF TAXES 3-1 3.1 Tax Effort 3-1 3.2 Structure of Indirect Taxes 3-5 Excise Duties 3-7 Sales Taxes 3-8 Import Duties 3-9 CHAPTER 4 ELEMENTS OF THE TRADE AND PAYMENTS REGIMES 4-1 4.1 The Exchange Rate Regime 4-2 4.2 Foreign Exchange Allocation and Import Licensing 4-4 Import Controls 4-7 Open General Licensing 4-9 Special Import Program I 4-14 Special Import Program II 4-16 Imports Without Officially Allocated Foreign Exchange 4-19 4.3 Import Composition Under Liberalization 4-21 Import Composition Under "No-Forex" 4-21 Overall Composition of Imports 4-24 4.4 The Tariff Regime 4-26 The Tariff Structure 4-28 Tariff Exemptions 4-29 4.5 The Export Regime 4-31 Export Taxation 4-32 Foreign Exchange Retention 4-33 Export Licensing 4-35 Export Marketing by State Enterprises 4-36 CHAPTER 5 THE INCENTIVES STRUCTURE GENERATED BY THE TRADE 5-1 AND PAYMENTS REGIME 5.1 The Exchange Rate and the Incentives Structure 5-2 5.2 The Agricultural Sector 5-5 The Coffee Sector 5-6 Other Traditional Exports 5-16 Nontraditional Crops 5-19 5.3 The Manufacturing Sector 5-20 Liberalization of Imports 5-20 Protection to Manufacturers 5-22 Viable Industries 5-28 CHAPTER 6 PROPOSALS FOR REFORM 6-1 6.1 Export Reform 6-1 Taxation of Coffee 6-2 Linkage with World Prices 6-3 6.2 Import Reform 6-5 6.3 Domestic Tax Reform 6-7 6.4 Appropriate Macroeconomic Policies 6-8 6.5 Exchange Rate Reform 6-10 Market-Determined Official System 6-10 6.6 Legalization of the Parallel Market 6-12 Uganda's Options 6-14 CHAPTER 7 SHORT-TERM EFFECTS OF PROPOSED REFROM 7-1 7.1 Effects on Agricultural Exports 7-2 Coffee and Other Traditional Agricultural Exports 7-2 Nontraditional Agricultural Exports 7-5 7.2 Effect on Manufacturing Output 7-7 7.3 Impact on Government Revenues 7-13 ANNEX I TABLES ANNEX II THE PROBABLE EFFECTS OF REFORM IN SUBSECTORS TEXT TABLES AND FIGURES CHAPTER 2 Table 2-1. Inflation-Related Variables in Uganda, 1966-88 Table 2-2. Exports, Current Account and Exchange Rate Overvaluation in Uganda, 1965-89 Table 2-3. Trend and Variability of Real Exchange Rates in Uganda, 1970-79 to 1989 Figure 2-1. Annual Inflation and M2 Growth, 1970-88 Figure 2-2. Quarterly Inflation and M2 Growth Figure 2-3. Official Exchange Rate and Middle-Income CPI, 1986 to mid-1989 CHAPTER 3 Table 3-1. Tax-GDP Ratio and Tax Structure in Uganda, 1970/71-1988/89 Table 3-2. Elasticity of Major Taxes in Uganda, 1970-87 Table 3-3. Tax-GDP Ratios and Tax Shares in Selected Low-Income Sub-Saharan Africa Countries, 1987 Table 3-4. Structure of Ugandan Indirect Taxation, 1987/88 and 1988/89 Table 3-5. Duty Collections by Tariff Rates in Uganda CHAPTER 4 Table 4-1. Total Ugandan Imports, 1986/87 to mid-1989 Table 4-2. Imports of Inputs Under OGL, January - December 1988 Table 4-3. Cost of Raising Sales Tax Revenues Using OGL Imports Table 4-4. Percentage Difference in Cost of No-Forex Imports Relative to SIP- II Imports Table 4-5. End-Use Composition of No-Forex Imports, 1988-89 Table 4-6. End-Use Import Composition in Uganda After Liberalization, 1989 Table 4-7. Import Composition in Uganda, 1980 and 1989 Table 4-8. Statutory Ad Valorem Tariff Rates in Uganda Figure 4-1. Official and Parallel Exchange Rate and Middle-Income CPI, 1986 to mid-1989 Figure 4-2. Official and Parallel Exchange Rate and Other Consumer Goods, 1986 to mid-1989 CHAPTER 5 Table 5-1. Policy-Induced Distortions in Coffee Prices, 1984-89 Table 5-2. Policy-Induced Distortions in Cotton and Tea Prices, 1984-89 Table 5-3. Relative Producer Prices of Traditional Ugandan Exports Table 5-4. Consumer Imports Table 5-5. Liberalization of Competing Imports, Selected Items, 1988 and 1989 Table 5-6. Raw Material and Intermediate Imports, Excluding Fuel, 1987-89 Table 5-7. Subsidies from Open General Licensing Table 5-8. Firms Importing Inputs Under Various Schemes Table 5-9. Nominal and Effective Rates of Protection, 1985, Output Changes, 1985-88, and Tariff and Licensing Coverage, 1989 Figure 5-1. Annual Exchange Rate Premium, 1970-88 Figure 5-2. Monthly Exchange Rate Premium, 1986-88 Figure 5-3. Price Wedges for Robusta Coffee, June 1989 Figure 5-4. Total Export Tax Rates on Traditional Crops, 1984-89 CHAPTER 7 Table 7-1. Nontraditional Exports Under the Foreign Exchange Retention Scheme, 1988 and 1989 Table 7-2. Destination of Nontraditional Exports Table 7-3. Structure of the Ugandan Manufacturing Sector Table 7-4. Effect on Revenue of Proposed Reforms ANNEX I, TABLES Table A-1. Tariff Rates by Section Headings Table A-2. Imports by User Firms Under Special Import Program I, December 1988 to April 1989 Table A-3. Imports by User Firms Under Special Import Program I, May to August 1989 Table A-4. Imports by User Firms Under Special Import Program II, June to August 1989 Table A-5. Imports by User Firms Under No-Forex, January to August 1989 Table A-6. Imports by User Firms Under No-Forex, January to June 1988 Table A-7. Imports by User Firms Under No-Forex, July to December 1988 Table A-8. Sales Tax Rates on Major Imports, 1988/89 Table A-9. Sales Tax Rates on Domestic Products, 1988/89) Table A-10. Sales Tax Base, Revenues, and Effective Tax Rates, 1988/89 EXECUTIVE SUMMARY Over the last three years, Uganda has implemented a substantial program of stabilization and structural reform. Many of the reforms relate to changes in the trade, payments, and tax regimes that affect production incentives for both exports and import-competing output. The Government of Uganda requested the UNDP/World Bank Trade Expansion Program (TEP) to undertake an assessment of the incentive structure emanating from the trade, payments and tax regimes, with a view to designing a further program of trade liberalization. This report attempts such assessment of the recent policy changes and their impact on incentives in order to address two broad questions: * Does the current incentive regime encourage an economically efficient choice of production activities? * On the basis of Uganda's experience so far, what further liberalization is required to enhance efficiency? Main Findings The persistence of macroeconomic disequilibrium is the major factor undermining trade performance in Uganda and inhibiting the adoption of appropriate trade policies. An overvalued exchange rate and import and exchange controls (reflected in the parallel foreign exchange market) create bias and instability in the incentive framework facing Uganda's tradables sector. These problems go back as far as the 1970s. - 2 - The country is highly vulnerable to fiscal and monetary shocks. Uganda's total tax revenue as a percentage of GDP is among the lowest in Sub-Saharan Africa. Direct taxation contributes only 11 percent of revenue. Indirect tax provide most of the revenue, but has an extremely narrow base. Dependence on export taxation (both implicit or explicit) has, on average, been quite high, though this has diminished with the collapse in world coffee prices. Collections from import duties, at 5-6 percent of import value, are the lowest among developing countries. Tax administration remains weak. Inefficiency in crop-financing mechanisms, especially for coffee, continue to undermine monetary control. The stabilization efforts since 1987 have had some success. Only in recent months have macroeconomic conditions improved considerably, but the situation remains fragile with respect to both fiscal performance and efficiency in crop financing. Further trade liberalization is thus circumscribed by the need to ensure that reform measures do not have a harmful effect on either the fiscal deficit or the money supply. Fortunately, this does not imply that liberalization should be postponed; instead it argues for a set of reforms that complement stabilization policies. Uganda has succeeded in reducing some of its large policy-induced distortions in the trade and exchange rate regimes after nearly three years of stabilization and trade liberalization efforts. Nevertheless, a significant bias against exports remains. The current policy environment makes production of exports much less profitable than production of import substitutes, with effective assistance varying substantially across import substitutes (for example, in manufacturing, production is more profitable - 3 - for selected firms or industries under the so-called Open General License scheme than for other firms). Much needs to be done before the incentives structure can guide an efficient choice of production activities in the tradables sector--in both export and import-competing sectors. The exchange rate regime, based on administrative allocation of foreign exchange and an overvalued official exchange rate, is a major source of distortion. The parallel exchange rate, applicable to imports without officially allocated foreign exchange, is effectively the marginal rate of exchange. This means that at the official exchange rate, exports are implicitly taxed and imports are subsidized at a rate approximately equal to the exchange rate premium in the parallel market. Since the rate of premium has fluctuated considerably, the tax/subsidy rate has also been unstable. Though there has been relatively more stability since July 1988, the exchange rate premium has moved considerably: 200 percent in September-November 1988, 130 percent in January 1989, 225 percent in August 1989, and around 60 percent in April 1990. The implicit exchange-rate-based tax on exports affects only coffee, since all noncoffee exporters are permitted to retain their foreign currency export receipts. Coffee exports are subject to an explicit tax as well; the government receives any surplus of export sales revenue (converted at the official exchange rate) over the sum of the producer price plus marketing margins. The total export tax (implicit and explicit) on coffee averages around 80 percent, with the implicit tax dominating the total. Notwithstanding the recent reductions world price and in the exchange rate premium, there has not been a significant reduction in total taxation of coffee exports as of April 1990. The level of total export taxation of coffee remains substantially higher in Uganda than in neighboring coffee-producing countries. This discourages coffee production and the question is whether higher production could be exported. In 1989/90 Uganda exported the largest volume of coffee in a long time, thereby calling into question the general pessimism about export markets. With more participants in Uganda's coffee exports (cooperative unions, private sector) and more favorable producer prices, higher coffee exports are feasible. Explicit and implicit export tax on Uganda's noncoffee exports have been eliminated since March 1989. As a result, they have grown significantly though non-coffee exports continue to suffer from three other disadvantages. First, the bureaucratic impediments to exporting are staggering. To obtain an export license an exporter has to complete at least sixteen documents and to obtain approval from five to eight agencies. Second, to receive the export proceeds at the parallel market exchange rate, using the export retention scheme, exporters must not only get payment from their foreign buyers, but they must use the payments to import goods and sell them in Uganda. This process takes around nine months. Third, producing noncoffee exports remains less profitable than producing goods for the domestic market. This is mainly because import- competing output is protected while noncoffee exports are not. Producers of noncoffee agricultural exports buy imported inputs at prices that reflect the parallel market exchange rate (notwithstanding actual imports by intermediaries at the official exchange rate). They also export their output at world prices, with export receipts effectively converted at - 5 - market or parallel exchange rates. Hence, noncoffee exports receive no assistance. On the other hand, manufactured output for the domestic market continues to receive substantial protection. Most of Uganda's imports of final goods, generally consumer items, are imported at the parallel market exchange rate and are subject to tariffs. Since tariffs are assessed on import value converted at the official exchange rate, though imports are imported at the parallel rate, nominal tariff protection for domestic output competing with such imports is lower than that implied by the actual rate of tariff. However, effective protection is very high because most inputs used in manufacturing are imported directly by firms at the overvalued official exchange rate under various official schemes. Firms using such inputs receive an implicit subsidy. In addition, effective protection differs because the size of the subsidy on inputs differs according to the official scheme used to import. The subsidy was greatest for a pre-selected set of firms under the so- called Open General License scheme. This prompted other firms to undertake 'rent-seeking' or deploy resources to obtain access. In addition, between June and October 1989, when three different official schemes were operating, and two-tier exchange rate for imports prevailed, effective protection for the same industrial firm could vary between 192 percent and 272 percent according to the scheme used (assuming a 20-percent tariff on output, 20 percent value added, 70 percent of inputs imported, and an exchange rate premium of 150 percent). If, however, that firm used the parallel market to import inputs, effective protection would be only 100 percent. For different tariff rates on outputs, the dispersion - 6 - will be still greater. Given the decline in the exchange rate premium and elimination of the two-tier official exchange rate for imports in recent months, there has been a reduction in dispersion of effective protection, but considerable dispersion still remains for import-competing output (see chapter 5). Furthermore, among import-competing products, the trade regime discriminates against the production of intermediate inputs relative to the production of final consumer goods. Intermediates are imported at a more overvalued exchange rate than final goods and are generally exempt from import duty or are imported at a very low rate. After July 1989 most imported inputs were subject to 10-percent duty. This policy has been reversed since then for capital goods. Discriminatory policies lower the efficiency of resource use in the tradables sector and reduce national income and living standards for any given investment or production capacity. This is because all sectors of the economy--whether manufacturing, agriculture, or services--are linked, so that policies directed at one activity affect others as well. The interdependencies are of two types. First are the direct links, through sales and purchases of inputs, of one sector's demands for inputs to another sector's outputs. These interindustry links are weak in Uganda. Second are the less obvious but more important indirect links between the economic performance of different sectors that arise because sectors compete with one another for scarce resources such as labor, capital, land, and foreign exchange. If one activity, through special government incentives, is encouraged to grow more than others, it will make a greater claim on those resources than it would otherwise do, making it more - 7 - difficult for other activities that also need the same resources to prosper. Thus special treatment or protection of one activity or a set of activities that enhances its growth prospects necessarily involves a tax on other domestic activities not so assisted. So it is not sufficient to justify special and differential incentives to some subsectors on the grounds that they have a special contribution to make to the economy. The resulting tax on others must also be justified. When policy-induced incentives are equalized across all tradables, no such tax or subsidy exists. The objective of trade liberalization is to move towards such equality so that the incentives structure encourages an efficient choice of production activities. Recommendations Macroeconomic and Complementary Policies Restoring and maintaining macroeconomic stability is essential in order to implement and sustain an appropriate set of production incentives. For this purpose, it will be necessary to restrain government expenditure, mobilize more revenue, and increase the efficiency of crop financing. Greater fiscal and monetary restraint will make it easier to ensure a more appropriate official real exchange rate and thus to remove Uganda's primary source of distortion in incentives. * Efforts to unify the exchange rate regime must continue. As long as monetary and fiscal restraints are in place and inflows of concessional capital are sustained, continuous adjustment in the official exchange rate will reduce the exchange rate premium. However, the current - 8 - policy of adjusting the nominal exchange rate so as only to offset differential inflation will not be sufficient to reduce the exchange rate premium further. To achieve that, the nominal devaluations must exceed the inflation differential, even if by small margins, and this should be part of the "formula" for adjustment. In addition, to depoliticize these devaluations, a quasi-automatic mechanism such as a crawling peg should be established. * The parallel market for foreign exchange should be "legalized" by means of licensed foreign exchange bureaus permitted to buy and sell foreign currency at the market rate. This change is expected to encourage a shift of economic activity from the informal or unrecorded sector to the official and recorded sector, thereby broadening the tax base and indirectly increasing revenue. However, the temptation to directly tax the foreign exchange transactions of the bureaus should be avoided. * This legalization would also permit noncoffee exporters to sell their export receipts to the bureaus instead of the current practice of importing and selling imports to obtain their export receipts. * Equally critical is the need for complementary policies affecting parastatals, infrastructure, and the financial sector in order to facilitate--or at least to avoid inhibiting--the economy's response to the liberalization-induced changes in incentives. Information * A mechanism for disseminating information on existing policies and policy changes must be established, since the current system of one- shot newspaper announcements is clearly inadequate. This would - 9 - involve, at a minimum, recording, documenting, and compiling rules, regulations, and tax rates at regular intervals. Preparing a tariff handbook, a tax manual, and a publication documenting the schemes and procedures for imports and exports (e.g., licensing process, Special Import Program, retention scheme, OGL) is the first step to making trade policy effective. Without them, the private sector will remain unaware of the actual production incentives facing them and thus fail to respond adequately to policy changes. * A customs-based database on trade flows is urgently required to properly monitor various facets of exports and imports. There has been no effect to establish such a system. Yet, this is essential not only for dependable analysis of trade policy issues by the government and the private sector, but also for monitoring the adequacy of revenue collection from imports. Computerization of trade flows information could be coupled with computerization of the customs assessment procedures to yield both revenue and data. Export Policy * Producer prices of coffee need to be linked to world prices so that the two move together. This could be done by substituting an ad valorem export tax for the current residual export tax. * Differential quality-based producer prices for coffee should be instituted, with the differentials linked to world price differentials. This would generate incentives to improve quality. * Overall taxation of coffee exports should be reduced by paying the farmers a larger share of the world price. While the fiscal - 10 - imperatives are important, the current emphasis on tax revenue from coffee must give way to better incentives for coffee growers. With current producer prices substantially lower than farmers' average cost of production, the medium-to-long-run sustainability of coffee production in Uganda is threatened. * Until the exchange rate system is unified, all noncoffee exporters should continue to be allowed to retain all their foreign exchange earnings. * Noncoffee exporters eligible to retain foreign currency receipts should be permitted to sell their foreign currency earnings to potential importers either through the new foreign exchange bureaus (if the parallel market is legalized as suggested) or through other mechanisms (by making retention import licenses officially transferable). The requirement that retained foreign exchange be used for imports by exporters themselves not only delays the firms' receipt of export proceeds, but also makes the profitability of exports contingent on the profits from imports. This inhibits export-specialization by firms. * Administrative impediments to noncoffee exports must be removed quickly. This should consist of simplifying the staggering documentation requirements for export licensing. The cost in terms of time and money to process a license is currently the major obstacle to an export response to other price incentives. * A duty-drawback or duty-exemption scheme should be introduced now, initially in the simplified form of a flat-rate "refund" based on a notion of average duties paid by similar producers. Such a scheme is needed because as the exchange rate premium falls, import duties on - 11 - inputs, especially inputs for noncoffee agricultural exports, will become binding and thus constitute a tax on exports. Import Policy * As long as the official exchange rate is overvalued, different import schemes with different eligibility requirements accentuate the dispersion in incentives. Official allocation of foreign exchange for all nonpetroleum and nonproject imports should be conducted through one import program only. This should, however, be genuinely more open and more general than previous programs. The single, all-encompassing import program (e.g., an Open Import Program) should be based on a small "negative" list (any good not listed can be imported), and, if necessary, with global allocation targets for inputs. All allocations should be on a genuine first-come, first-served basis. Most important, the program should be a sustained and permanent feature of Uganda's import regime, irrespective of the ebb and flow in the release of donor financing for imports. * The so-called Open General License (OGL) scheme should be eliminated. This scheme is neither open nor general; instead, it is discriminatory and subject to misuse by firms. In fact, OGL firms seem to import more than is warranted by their recorded production, under OGL allocations. Recently they have also been important beneficiaries under the new SIP- III. There is increasing evidence of rent-seeking activities by existing and potential OGL firms to maximize access to scarce foreign exchange. Non-OGL firms are thus strongly discouraged. If OGL is eliminated, existing OGL firms should be able to compete with other - 12 - firms to obtain foreign exchange in any proposed Open Import Program (0IP). * There could, however, be a prespecified global-allocation target (e.g., a predetermined share of imports) for all imported inputs in the OIP. All manufactures would compete for this allocation on a first-come, first-served basis. Or alternatively, in the transition OGL could be subsumed in the OIP as a subcomponent. Three-way global allocation for inputs of OGL firms, for inputs of all others and the balance foreign exchange in OIP for all other imports. Over time, the first two could be merged into one allocation for all inputs in a single import scheme for allocating all donor funds. * Tariff rates should be rationalized by making the structure simpler and the rates relatively more uniform. The highest tariff rate should be reduced from 350 percent to 50 percent. Given current coverage, this will have little adverse effect on production or import duty collection. The lowest rate should be raised to 10 percent. * Luxury imports should be taxed through a high sales tax instead of high tariffs. A high tariff on luxuries, by making domestic production profitable, encourages domestic production of the product whose consumption the high tariff seeks to discourage. A sales tax on both imports and domestic production of such luxuries is preferable, as it will discourage both production and consumption of luxuries. * Exemptions from import duties should be substantially reduced if not eliminated. This means that there should be no duty exemptions for inputs, for either intermediates or capital goods. To promote investment, it is preferable to provide other fiscal incentives (e.g., - 13 - depreciation allowances, tax holidays) rather than duty exemptions on capital goods, since the latter will discourage domestic production of even simple machinery, and distort factor-proportions in production. Indirect Tax Policy Over the long-run, revenue should be derived mainly from a consumption tax. However, given the current limited administrative capacity of the tax machinery, it is essential to ensure that the changes are simple enough to avoid transitional declines in revenue. Most of the proposed changes herein can be implemented by existing administrative capacity. * Revenue collection from import duties should be increased. The current collection level of 5-6 percent of import value is the lowest in the developing world. This is not because of low tariffs, since the modal statutory tariff in Uganda is 30 percent. Elimination of exemptions will help, but customs administration has to be streamlined and customs database has to be developed. * A temporary surtax could be levied immediately on all imports (duty- inclusive price) and on domestic output of import substitutes. This would increase revenue from current dutiable imports, while customs administration is being streamlined. This surtax should be low, say 5 percent, and would be removed once the exchange rate was reasonably unified. This surtax should make up for some of the erosion in revenue collection resulting from the divergence between the official and parallel exchange rates. - 14 - * Imports and domestic output should be taxed at the same sales tax rate. All anomalies in this context should be removed. * Domestic output that is exported should be exempt from any sales tax. * The sales tax rates should be reduced from the twelve existing rates (lying between 10 and 120 percent) to three ad valorem rates: a low basic rate for all manufactured goods (e.g., 20 percent) and two high rates for luxury and "super" luxury items. A change in current sales tax rates to two rates of 20 percent and 70 percent is found to be revenue-neutral. * The current suspensive or "ring" system of sales taxes, used to avoid a cascading of sales tax, should be replaced with a "tax credit" system, which should contribute to better revenue collection. The lead time for implementing such a system may be longer than for all other proposed tax measures, and its introduction should await significant improvements in tax administration. * Excise duties should ideally not be expanded to cover goods other than alcoholic beverages and cigarettes. However, because of their current importance to general revenue, excise taxes on soft drinks and soap could be continued. However, recent additions to the excise duty list should be removed and those goods should be subject only to a sales tax. * Excise rates could also be rationalized to a simplified two-rate structure. A shift to 60 percent for alcoholic beverages and high- priced cigarettes and to 30 percent for soft drinks, soap, and low- priced cigarettes is found to be revenue-neutral. - 15 - * Exemptions from sales and excise duties provided to armed forces personnel should be removed. * The Commercial Transactions Levy, applied to the transactions value of a number of services, should be maintained. * To improve tax information and administration, a simple accounting book should be designed and published yearly for use as a bookkeeping system by all taxpayers. The completed book would be returned to the tax authority at the beginning of next fiscal year, to be used for assessment and verification of the payment of taxes. * To achieve all of these reforms, sustained improvements in tax administration, especially customs administration, remain critical and should be the prime target of revenue mobilization effort. IIIIIII111 11II 1 Ilmil CHAPTER 1 INTRODUCTION In the 1960s, Uganda's economy was one of the most promising in Sub-Saharan Africa. Apart from fertile land and favorable climatic and soil conditions that made rapid agricultural growth possible, it had a relatively well-developed manufacturing sector and an effective transportation system. Between 1963 and 1970, the average annual growth of GDP was close to 6 percent, inflation was low, and the current account was frequently in surplus. Exports were diverse and the trade regime was relatively liberal. However, after 1970, the Ugandan economy began to experience severe difficulties reflecting the political strife and economic mismanagement that accompanied the Amin regime. Direct controls on the economy proliferated. Real GDP declined by about 20 percent during the 1970s, and inflation rates soared to more than 100 percent in some years. Recorded trade declined, and by 1980 exports had fallen to less than half their 1970 level. The manufacturing sector came to a virtual standstill. Recorded agricultural output fell, but the emergence of parallel markets in foodgrains and foreign exchange sustained substantial unrecorded productive activities, especially in coffee and food crops. A program of stabilization and trade liberalization initiated under Obote in early 1981 led to some recovery. But with rising political instability since late 1984, public expenditures escalated, inflation accelerated, and the exchange rate again became grossly overvalued. Both import controls and export taxation increased. GDP declined in real terms. 1-2 When the National Resistance Movement (NRM) government assumed power in January 1986, the economic situation was extremely difficult. Efforts at reconstruction and rehabilitation as well as stabilization and liberalization have been underway since May 1987. These efforts, accompanied by growing political stability, have raised the level of economic activity, although the recent collapse in world coffee prices has called for further macroeconomic adjustments. 1.1. The Main Questions Political stability, external economic factors, and government policies have generally determined Uganda's economic performance. Growth in national income over the next decade will no doubt depend on new investment, but in the short to medium run it will depend mainly on fuller and more efficient use of existing capacity. Economic policy can enhance or reduce the benefits from a given investment or production capacity. The wrong sectoral mix of output and investment reduces economic benefits, and the larger the size of investment or the sectors affected by wrong or inappropriate policies, the greater the potential loss. Exports play a crucial role in Uganda's economic growth because of the heavy dependence on imports both to utilize existing capacity and to create new productive capacity. Thus, expansion of foreign trade is necessary if national income is to grow to its potential level. Resource allocation and capacity utilization within the tradables sector and between tradables and nontradables are important determinants of the overall volume and composition of Uganda's exports and imports. 1-3 The government can influence resource allocation by altering the policy environment that determines incentives for private and quasi-private economic agents. This report assesses the incentive regime facing Uganda's external or tradables sector and attempts to answer, however crudely, the following questions: * Do existing policies encourage an economically efficient choice of activities within the two major producing sectors of agriculture and industry? * Does the present structure of incentives imply an overall bias in favor of import-competing output or of exports? * How do macroeconomic policies affect the incentives regime or constrain the pace of reform? * What should be the direction of reform to facilitate a more efficient allocation and utilization of resources? Conducting a detailed and sufficiently rigorous analysis of all of these questions is a difficult task in any country. In Uganda, the scarcity of data and past research posed even greater problems. Since Uganda has no customs-based trade data for any of the last three years, the data on volume and composition of exports and imports used in this report are based on payments and receipts monitored by the Bank of Uganda. The coverage of such data is inadequate. Also lacking are recent census data on agriculture or manufacturing activities and input-output tables. The only way to ascertain the input structure of products of firms would be to undertake a large-scale sample survey of enterprises. This was clearly beyond the scope of this study. In any case, this would have provided data for analyzing the protective effects only of tariffs. To examine the effects of the more prevalent nontariff barriers, information on domestic 1-4 and world prices would be required. The Department of Statistics survey provided some recent information on output, sales, and employment but no data on cost structure and input usage. Thus, summary indicators of incentives such as actual effective rates of protection could not be estimated. The challenge for this report was to piece together information from a variety of sources to analyze the issues and answer the questions in the best possible manner. Inevitably, the picture that emerges is incomplete. Where gaps remain, the nature of the problem has been identified. But for most issues, the report finds helpful answers. 1.2. The Approach The main policy components of the incentives structure directly affecting product markets are the following: * the real exchange rate; * import licensing/foreign exchange allocation; * export licensing/export marketing system; * import and export taxes/subsidies; * domestic taxes/subsidies; and, * domestic price controls. This report examines changes in these policy components in Uganda in 1988 and 1989 to get a feel for the liberalization process that has been underway. The current status of these policies is examined in chapter 4 and the resultant incentives for production in agriculture and industry and in export and import-competing sectors are assessed in chapter 5. 1-5 To evaluate the appropriateness of Uganda's current policy-induced incentives structure, it has to be compared with a "norm." The norm obtained from economic theory argues for transparency in policy and an appropriate and unified exchange rate. In other words, it argues for an incentive regime based more on prices than on quantitative controls, wherein the amount of foreign currency units (domestic currency units) earned or paid per unit of domestic currency (foreign currency) is equalized for all external transactions, for both current and potential economic activities. Failure to equalize will imply that some activities are subsidized while others are taxed. While changes in the policy components cited above alter incentives directly, macroeconomic policies have had just as important a role in Uganda in determining of the transparency, bias, and stability of the incentive regime (chapter 2). Inconsistent exchange rate and fiscal and monetary policies have led to the imposition of exchange controls and the growth of a large quasi-legal parallel market in foreign exchange. As a result, Uganda's international transactions are conducted at two real exchange rates, with the parallel rate being the marginal rate of exchange. Exports at the official rate (e.g., coffee) are subject to an implicit exchange-rate-based tax while imports under official foreign exchange allocation schemes receive an implicit subsidy, with the rate of tax/subsidy being close to the exchange rate premium in the parallel market.1 The parallel rate is highly sensitive to expectations, making the 1 Between June and October 1989, importers could buy official foreign exchange at two different prices (U.Sh. 200 per dollar under OGL and Special Import Program I, and U.Sh. 400 per dollar under Special Import Program II. 1-6 exchange rate premium unstable and difficult to manage under conditions of macroeconomic instability. Restoring macroeconomic stability and unifying the exchange rate are critically dependent on improved fiscal performance and more efficient crop financing. Controlling public expenditure is the key to improved fiscal performance. Increasing tax revenue is probably just as important --if not more--given Uganda's low tax to GDP ratio relative to that of comparable countries. Revenue from direct taxes and import duties is extremely low. Uganda's dependence on the explicit export tax on coffee has been high in the past, though this has declined with the recent collapse in coffee prices. The current revenue situation and the potential for increasing government revenue, especially from domestic indirect taxes, are discussed in chapter 3. Given the current narrow tax base, most future revenue increases must come from shifts in transactions from the unofficial sector to the official recorded sector of the economy and from growth in overall output rather than from higher tax rates. The directions for reform are based on the current state of the incentives structure and the characteristics of the new regime that would encourage a more efficient choice of production activities. This is sketched out in chapter 6. The short-term implications of an immediate shift to the desired regime are explored in chapter 7, with a view to ascertaining a feasible set of policy changes. In the absence of data, this chapter uses a "historical" approach that draws on the actual impact of liberalization measures in 1988 and 1989 on the economy to assess qualitatively the likely short-run effects of proposed reforms. CHAPTER 2 MACROECONOMIC POLICIES AND THE INCENTIVE STRUCTURE Macroeconomic policy has long been a key determinant of the incentives structure for Uganda's external sector. In the 1960s, a relatively liberal trade and payments policy went hand in hand with macroeconomic stability. The 1970s witnessed growing macroeconomic instability and balance of payments pressures. Political uncertainties and government extravagance led to rising inflation because of increasing dependence on inflationary financing of expenditures. Fiscal and monetary policies inconsistent with a fixed exchange rate led to substantial and sustained real exchange rate appreciation and declining exports. Increasing overvaluation of the official exchange rate necessitated import and foreign exchange controls to restrain the current account deficit to levels consistent with available reserves and external financing. Administered producer prices for export crops were not sufficiently adjusted for inflation. Thus, agricultural exports were taxed twice: once by the overvalued exchange rate and once by falling real prices for producers. The policy-induced incentives structure discriminated against tradables and exports. An increasing share of international transactions began to be conducted through illegal or unofficial channels, with some estimates putting this share at half of all exports and imports. A thriving parallel black market (called the "kibanda" market) in foreign exchange emerged, and its exchange rate rose to ten times the official rate (an exchange rate premium of 900 percent) by the end of the 1970s. Price controls and rising domestic tax rates encouraged most economic activity 2-2 into the underground ("magendo") sector, thereby reducing the tax base. Uganda's tax effort deteriorated rapidly not only because of this reduction in the tax base, but also because of overvaluation of the exchange rate and deterioration of tax administration. However, the presence of this parallel economy did reduce the output costs of Uganda's policy-induced distortions. In the 1980s, Uganda made several attempts at stabilization and liberalization of the import regime, all of which were underpinned by substantial aid flows. Fiscal and monetary restraint coupled with several large depreciations of the official exchange rate between mid-1981 and mid- 1984 reduced the exchange rate premium in the parallel market to around 83 percent by December 1984. This success at unifying the exchange rate proved to be short-lived, however, as money financing of unplanned increases in recurrent government expenditure raised domestic demand and inflation. By early 1987, the parallel market exchange rate premium was back to its all-time high of 900 percent. Producer prices for export crops, which had been adjusted upwards substantially since 1981, were again eroded by inflation, and after some recovery in export volume in 1981-84, exports fell again. Import and exchange controls that had been lifted between 1981 and 1984 were restored and made more restrictive between 1985 and 1987. Another round of stabilization and liberalization was initiated after May 1987. 2.1 The Early Years Macroeconomic stability, a relatively liberal trade regime, and political stability were the hallmarks of Uganda from 1963 to 1970. This 2-3 was also a period of rapid economic growth and rising incomes.1 Successful exploitation of agricultural potential proved to be the main engine of growth. Uganda was not only a large producer of food crops, but also an exporter of considerable quantities of coffee and cotton as well as tea and tobacco. Over most of this period, the macroeconomic situation remained sound: the fiscal deficit rarely exceeded 2.5 percent of GDP and inflation did not exceed 10 percent (see table 2-1). Except for some "essential" consumer items, imports were free from licensing or other quantitative controls. Exporters were required to surrender their foreign currency earnings to commercial banks at the prevailing official exchange rate, but the official rate was not overvalued. Commercial banks were allowed to hold those foreign currency balances for allocation to importers since there were no large scarcity-rents to be earned from such access. Moderate tariffs were the primary restriction on imports for the early part of this period, although selective restrictions on imports were gradually imposed to protect domestic industry. The embryonic industrial sector grew faster than the rest of the economy in the late 1960s. Although the government established several industries as joint ventures, the industrial sector was predominantly privately owned. With growing domestic demand, both local and foreign entrepreneurs (many of Asian origin) established industries to produce a wide range of consumer and intermediate goods. Despite rising levels of government intervention (protection, fiscal concessions, and loan 1 By 1970 Uganda's GNP per capita ($512 in 1980 prices) was the fourth highest in Eastern and Southern Africa, following Zimbabwe, Mauritius, and Zambia. 2-4 Table 2-1. Inflation-Related Variables in Uganda, 1966-88 Z Fiscal Change Change in Change deficit in money domestic in total Inflation as Z of supply credit domestic Year rate GDP a (M2) to govt.b credit b 1966 4.4 -1.0 -- -- -- 1967 -2.1 2.0 -- 2.1 6.2 1968 2.2 1.8 10.5 2.5 12.5 1969 2.1 2.5 7.2 3.0 18.6 1970 12.5 4.4 18.8 19.8 21.4 1971 11.1 7.0 -1.1 21.4 23.4 1972 5.0 8.0 29.1 30.5 34.0 1973 12.7 6.6 34.8 39.6 49.4 1974 21.1 9.5 33.5 28.3 40.8 1975 52.3 5.5 21.1 12.0 15.0 1976 13.7 5.0 32.7 27.8 32.4 1977 102.0 2.3 19.3 11.4 26.5 1978 19.3 0.3 25.5 24.1 28.4 1979 80.2 3.9 47.9 23.1 25.3 1980 54.6 3.1 34.6 41.0 59.9 1981 108.7 4.7 87.3 89.1 119.0 1982 49.3 3.6 11.4 15.0 44.0 1983 24.1 2.2 41.3 24.2 57.3 1984 42.7 3.2 113.7 59.3 64.1 1985 132.4 5.9 125.5 81.9 148.3 1986 168.5 5.9 171.5 22.4 92.4 1987 238.1 4.8 155.7 14.4 79.0 1988 183.6 6.6 116.7 45.8 88.5 a. Excluding grants. b. As percentage of M2 in the previous year. Source: World Bank database. financing), many industrial enterprises were efficient users of resources, and even exported manufactured goods, especially cotton fabrics and animal feed to Kenya and Tanzania. Total exports, mainly primary commodities, and imports each grew at about 3.5 percent a year. Except for two years during this period, Uganda maintained a current account surplus. 2-5 In the 1970s, political repression, internal war, and gross mismanagement of the economy reduced Uganda from one of the most prosperous and promising countries in Sub-Saharan Africa to one of the poorest. As public expenditure rose dramatically, the fiscal deficit rose from 2.5 percent of GDP in the 1960s to 9.5 percent in 1974 (see table 2-1). Most of the increase was financed through money creation. As dependence on inflation taxes grew so did the domestic rate of inflation, rising from an average of 3.5 percent a year in 1963-70 to 35 percent in 1971-80, and 108.7 percent in 1981. Price controls proliferated to restrain the inflationary impact on essential goods, forcing economic activity into the underground economy. Production of export crops fell as taxation and economic insecurity made it a riskier venture. The subsistence agricultural sector maintained its steady growth, providing food for farmers and their families and for the thriving and lucrative underground markets as well. With a fixed nominal exchange rate, a higher rate of inflation domestically than in trading partner countries makes imports less expensive relative to domestic goods and exports uncompetitive at given world prices. For Uganda, the extent of this adverse shift in relative prices is evident from the ratio of its price level to that of its trading partners, which rose from 1 in 1970 to 18 in 1980. A comparison of Uganda's official exchange rate with two indicators of the appropriate rate (the rate that represents a realistic scarcity value of foreign exchange) highlights the extent of overvaluation (table 2-2). The massive degree of overvaluation is reflected in the large 2-6 Table 2-2. Exports, Current Account, and Exchange Rate Overvaluation in Uganda, 1965-89 Exports goods Volume Exchange rate Current and non- of New UG shilling per US dollar account factor coffee balance services exports Official Parallel Estimated ($ m) ($ m) (000 tons) rate rate PPP Year (1) (2) (3) (4) (5) (6) 1965 10.1 302.1 -- -- -- -- 1966 11.2 330.1 -- 0.07 0.08 0.070 1967 -10.1 351.7 -- 0.07 0.08 0.068 1968 -3.1 365.8 184 0.07 0.08 0.070 1969 -5.4 380.7 160 0.07 0.08 0.071 1970 20.3 425.3 190 0.07 0.10 0.080 1971 -85.8 391.8 185 0.07 0.11 0.088 1972 16.4 381.7 176 0.07 0.13 0.092 1973 43.0 339.9 178 0.07 0.21 0.102 1974 -24.1 328.3 150 0.07 0.27 0.121 1975 -56.1 200.6 141 0.07 0.53 0.181 1976 43.2 290.7 135 0.08 0.73 0.207 1977 68.1 270.7 155 0.08 0.63 0.413 1978 -137.4 333.2 140 0.08 0.79 0.483 1979 39.5 413.2 120 0.08 0.68 0.857 1980 -83.2 329.4 110 0.08 0.76 1.305 1981 31.9 273.4 130 0.17 2.08 2.734 1982 -15.9 382.2 165 0.94 2.66 4.098 1983 10.3 379.1 144 1.54 3.18 5.115 1984 -138.1 451.5 134 3.60 4.29 7.307 1985 -32.3 374.1 150 6.72 18.38 16.986 1986 4.7 424.9 144 14.00 65.28 45.021 1987 -129.9 374.6 150 42.84 173.96 150.713 1988 -162.9 305.2 148 106.14 412.08 425.385 1989 242.27 PPP is purchasing power parity. Source: Columns 1 and 2: World Bank database; column 3: Background to Budget: Uganda (various issues); columns 4 and 5: Bank of Uganda; and column 6: Computed using foreign inflation weighted for twenty Ugandan trade partners and middle-income CPI of Uganda. 2-7 differences between the official rate and both the parallel market rate and the purchasing-power-parity-based exchange rate.2 This means that the domestic currency receipts from exports were a fraction of what they would have been had the official exchange rate been adjusted for the differential in inflation. In addition to this implicit taxation of exports, which became very large in the late 1970s, the price paid to producers of export crops was a small fraction of the export value over that period converted at the overvalued official exchange rate--65 percent in 1980 compared to 90 percent and sometimes even more than 100 percent in the early 1960s. The real price (in terms of a consumption basket) received by growers declined drastically. This erosion of price incentives for exporters manifested itself in declining export volume and value (in current dollars) throughout the 1970s. While exports fell, demand for imports rose. To restrain imports, given the overvaluation of the domestic currency and expansionary financial policies, the government resorted to import licensing and foreign exchange controls.3 Also, beginning with the Export and Import Corporation Act in 2 Note that if the nominal rate indicated by purchasing power parity had prevailed, it would have implied a constant real exchange rate. This constant rate may or may not have represented the equilibrium real exchange rate for Uganda because the latter depends also on fundamentals (e.g., terms of trade, capital flows). 3 Faced with an incipient external deficit that cannot be financed, the authorities could choose one or a combination of the following: (a) given sufficient flexibility in prices, to keep the exchange rate unchanged while encouraging necessary price and income changes through deflationary fiscal and monetary policy; (b) to keep exchange rate unchanged but to alter the relative prices of exports and imports through subsidies and tariffs; (c) to depreciate the exchange rate sufficiently; or (d) to impose direct import and exchange controls to restrict imports to a given level--the method chosen by Uganda. 2-8 1970, restrictions on imports proliferated, marking a major departure in policy. Advance import deposits were instituted in 1971 (and abolished in June 1972). Commercial bank holdings of foreign currency balances were restricted, and by 1978 all foreign exchange was administratively allocated. New import and export monopolies were established, and parastatal marketing agencies increased their domination of both internal marketing and exports. In 1972, the government began a series of large-scale nationalization and confiscation of manufacturing firms, especially those of Asian-Ugandans. State control over industrial assets, imports, and foreign exchange came increasingly to be used to guide total imports to levels commensurate with available foreign exchange. Price controls, taxes, and tariffs escalated as policy inconsistencies became more acute. 2.2 Adjustment in the Early 1980s Between 1981 and 1984, Uganda sought to eliminate currency overvaluation, to stabilize the economy by reducing fiscal deficit and domestic credit creation, and to liberalize trade by rationalizing import restrictions and lowering taxation of exports. The shift from a fixed to a predetermined but adjustable exchange rate regime was initiated through a maxi-devaluation in June 1981. This was followed in 1982 by a period of substantial fiscal and monetary restraint and by the introduction of an official two-tier exchange rate system that included an auction mechanism. The first tier (window I) consisted of an "administered" exchange rate, applicable to export receipts from coffee, cotton, tea and tobacco, official loans and grants, petroleum imports, project-aid imports, and debt-service payments. 2-9 All other import transactions were covered by an exchange rate determined by foreign exchange auction (window II). Importers were expected to bid for their foreign exchange needs on the basis of a pre- announced level of auctionable foreign currency, leading to substantial liberalization of imports. The auction rate depreciated rapidly and continuously for two years. Though there was a large difference in the two rates, the window I rate was depreciated faster than the auction rate, until the two rates were merged in mid-1984. In 1984, the official exchange rate (Ug. shilling/US dollar) was around 45 times its 1980 level. At the same time, there was a substantial narrowing of the differential between the parallel rate and the official (window I) rate, largely because of continued fiscal and monetary restraint over these years. The exchange rate premium fell from 900 percent in 1980, to 180 percent in 1982 and to around 50 percent in 1984. These exchange rate adjustments were accompanied by large upward adjustments in producer prices for export crops. Coffee and cotton producer prices were raised fivefold in 1981, and further increases in the following years led to a minimum producer price for coffee in 1984 of about 30 times the 1980 level. This increase in producer prices, together with the reforms in fiscal and exchange rate policy, contributed to substantial reduction in export taxation and liberalization of imports in those years. Unfortunately, these improvements in the trade policy environment did not endure. Fiscal and monetary expansion in late 1984, 1985, and 1986 undermined Uganda's economy, and a military coup followed by a civil war further devastated its infrastructure and brought the economy to a standstill. 2-10 In January 1986, the National Resistance Movement (NRM) assumed power. Although security conditions began to improve, economic decline continued. In January 1987, the new government sought to mobilize external financial support to arrest the economic decline. The first phase of this recovery effort was initiated in May 1987 with the announcement of a number of strong policy initiatives, taken with the support of the International Monetary Fund (IMF) and the World Bank. As in the past, the greatest challenge facing Uganda in its efforts to liberalize the trade and payments regime is the restoration of macroeconomic stability. 2.3 Stabilization and Liberalization in the Late 1980s The interaction between macroeconomic stabilization and trade liberalization is crucial to success in both areas. Without stable macroeconomic policy, the real exchange rate will be differentiated and variable, conveying confusing price signals to private investors and producers of tradables, the key economic agents in any trade reform program. Thus, successful macroeconomic stabilization is critical to successful trade liberalization. At the same time, however, it is important to ensure that any conflict of trade reform with stabilization is minimized and any complementarity between the two maximized. For example, lowering the explicit export tax sharply is likely to increase Uganda's fiscal deficit. Raising producer prices for coffee, given the inefficient crop-financing system, is likely to increase the money supply. Devaluation of the official exchange rate could have a cost-push impact on some domestic prices. Also, liberalization could facilitate capital flight, thereby 2-11 aggravating the balance of payments problem. On the other hand, devaluation of the official exchange rate is likely to improve the fiscal balance, given the large inflow of external financing. Import liberalization can be complementary to macroeconomic stabilization to the extent that it dampens inflation by increasing the availability of essential consumer goods and inputs. Real Exchange Rate A stable and realistic real exchange rate is critical for the tradables sector of any country--and is particularly so for Uganda, given its history. However, the existence of both an official foreign exchange market and a quasi-legal parallel foreign exchange market in an environment of unstable macroeconomic policies makes it difficult to ensure a single stable and realistic real exchange rate. The relationship between fiscal/monetary policies and the economy generally depends on the nature of the dual or multiple exchange rate system that exists. If all nominal rates are fixed or predetermined, this relationship will be the same as that in a unified fixed exchange rate regime: expansionary fiscal and monetary policies inconsistent with the fixed exchange rate will result in domestic inflation higher than world inflation, and thus to overvaluation of the exchange rate or its real appreciation and to loss of reserves. If, however, as in Uganda, a fixed or predetermined official exchange rate coexists with a free-floating parallel market rate, expansionary financial policies will result in inflation, appreciation of the official real exchange rate, and falling reserves, but the parallel market real exchange rate may either depreciate or appreciate, even if 2-12 there is nominal depreciation of the parallel rate. This occurs because the extent of nominal depreciation in the parallel rate may be more or less than the rise in the domestic price level or inflation. If the parallel foreign exchange market adjusts faster than the goods market, any depreciation of the nominal parallel exchange rate on impact is likely to exceed the rise in price of domestic goods.4 Thus, the parallel market real exchange rate would depreciate in the short run. This means that it is possible for the official real exchange rate to appreciate even as the parallel real rate depreciates. However, this potential for opposite-direction movement in the two real rates in response to monetary expansion would be reduced if such expansion were accompanied by sufficient official devaluation.5 Subsequently, however, the parallel real rate would appreciate as adjustment in the goods market is completed and prices rise more even, although the new adjusted level of the real parallel rate may not reach its pre-shock level. Thus, even a one-shot money supply shock, with no change in expectations, can make real rates fluctuate for some time. In addition, since changes in the parallel rate affect the price level, this too implies a high degree of price variability. 4 Asset markets, like the foreign exchange market, always respond more rapidly than the goods market because goods transactions take longer to complete than a parallel market foreign exchange transaction. Thus, prices in parallel foreign exchange markets will change quickly in response to any change in macroeconomic policy. 5 Such negative correlation was evident in Uganda in the 1970s in the two (annual) real exchange rates. In the 1980s, frequent adjustments in the official nominal rate prevented this even in the face of expansionary policies. 2-13 Over the 1980s, Uganda has had reasonable success in initiating real depreciations of the official rate and reducing exchange rate premiums, especially between 1981 and 1984 and again between 1987 and 1989. But this effort was never sustained. Unstable financial policies led to cycles of appropriate real exchange rates followed by overvalued ones, and low exchange rate premiums followed by high premiums. Both real exchange rates have been highly unstable (table 2-3), although the parallel rate has been less unstable than the official rate and instability in both has declined recently. Most of this instability arose from the "stop-go" monetary and fiscal policy of the 1980s. Econometric estimation of a parallel real exchange rate equation shows that 40 percent of the variance is attributable to terms of trade changes and nearly one-third to monetary instability. Devaluation and Inflation Devaluation is clearly necessary if monetary or fiscal expansion is inconsistent with the exchange rate or if real stocks warrant a change in the real exchange rate. In Uganda, there is considerable fear of the inflationary implications of such devaluation of the official exchange rate. In part this fear stems from the fact that devaluations have been undertaken to overcome the erosion in the real exchange rate caused by rapid domestic inflation. The high visibility of the price effects of devaluation on certain commodities (e.g., petroleum products) has also contributed to this perception. Generally, two transmission mechanisms for inflation are relevant in Uganda. Expansionary fiscal and monetary policies and the cost-push effect of important domestic food price increases. Expansionary fiscal and 2-14 Table 2-3. Trend and Variability of Real Exchange Rates in Uganda, 1970-79 to 1989 Parallel RER Official RER Inflation Average Coefficient Average Coefficient coefficient Period (1980=100) of variation (1980=100) of variation of variation Year-to-year 1970-79 82.0 45.7 0.1 5.7 1980-86 71.6 30.0 19.6 169.3 1985-88 51.3 14.7 42.4 92.4 Month- to-month 1985 23.8 16.5 95.6 1986 22.9 44.6 102.0 1987 13.3 54.7 129.4 1988 16.1 26.8 115.4 1989 a 0.1 0.14 89.3 RER is real exchange rate. a. January to August 1989. Source: Computed from IFS tapes using 20 trade partner weights for annual real exchange rates and 7 partner weights for monthly rates (correlation coefficient between the two foreign price indices is 0.98). monetary policies have been the dominant autonomous factor in Uganda's inflation.6 Figures 2-1 and 2-2 show quite clearly that what happens to inflation in Uganda depends largely upon what happens to monetary growth. 6 This was very obvious in the 1970s. Uganda maintained a fixed exchange rate for the whole decade, and yet the inflation rate rose from 5 percent in 1972 to 102 percent in 1977 and 109 percent in 1981. Similarly between July 1987 and June 1988, although there was no devaluation of the official exchange rate, the inflation rate rose sharply, again because of expansion of the money supply. 2-15 Figure 2-1. Annual Inflation and M2 Growth, 1970-88 (percentages) 240 - 220 - 200 - 180 - 1 60 - 140 - 120 - 100 - 80 - 60 - 40 - 20 -20 . 1970 1973 1976 1979 1982 1985 1988 0 INFLATION + M2 GROWTH Figure 2-2. Quarterly Inflation and M2 Growth (percentages) 100 - 90 - 80 - 70 - 60 - 50- 40 - 30 - 20 - 10 DEC 83 DEC 84 DEC 85 DEC 86 DEC 87 DEC 88 0 INFLATION + M2 GROWTH 2-16 Of course, the monetary and fiscal balance can be adversely affected by a devaluation. If the government is a net buyer of foreign exchange, as was Uganda in 1987/88, devaluation will raise the government's deficit; if it is a net seller, devaluation will reduce the deficit. For Uganda, increased donor contributions for imports and balance of payments support, as well as the collapse in world coffee prices, has changed this over the last two years. The Ugandan government is now a net seller of foreign exchange. Devaluation raises local currency counterpart funds; in addition, indirect tax revenue responds positively to large depreciations since the official rate is the valuation base for import and sales taxes (even though the parallel rate is reflected in domestic prices--see chapter 3). Thus devaluations do not have an adverse impact on the fiscal balance, and some estimates even suggest a substantial positive impact.7 Thus, notwithstanding a 45-fold depreciation of the official exchange rate, annual inflation averaged only 40 percent between 1981 and 1984. During this period, expenditures were reduced and fiscal revenue increased substantially as a result of increases in official transactions (exports, imports, and domestic market sales) and thus in the revenue valuation base. The fiscal deficit fell from an average of 4.5 percent in 1980-81 to 2.7 percent in 1982-84, while the actual decline in money creation was even greater. Similarly, despite devaluations in July and December 1988, the inflation rate was much lower between July and February 7 In the 1989/90 budget, one estimate shows that the government's foreign exchange receipts exceeded its payments by about US $300 million. 2-17 1989 than in 1987/88, when there was no devaluation. (Figure 2-3 shows the absence of a clear association between official devaluation and inflation.)8 The second inflationary mechanism in Uganda is the cost-push effect of import prices and domestic food prices, although its effect is not nearly as strong as that of monetary expansion. An exchange rate depreciation contributes to an increase in the domestic prices of imports. Since the share of imports in Uganda's consumption basket is high, the effect on consumer imports would be correspondingly high (an IMF estimate suggests that a 10-percent devaluation can lead to a 2.4 percent rise in consumer prices). Also, an increase in the domestic prices of imported inputs could raise production costs and so increase the prices of domestically produced consumer goods. Both these effects would have a once-and-for-all cost-push impact on the domestic price level. This impact is likely to be very muted in Uganda, however, because the parallel market exchange rate determines the domestic price level of most final goods with the exception of petroleum.9 With respect to the consumption basket, between 75 and 90 percent of consumer imports come in at the parallel market rate of exchange--called "no-forex" imports. Thus, devaluation of the official rate has often had only a limited impact on the market prices of domestically produced goods. 8 Econometric evidence confirms the dominant role of monetary growth in Ugandan inflation over the long term (1970-88) and for the shorter period using quarterly data (1985-89). 9 Between 1981 and 1984, there was a 45-fold (i.e., 4400Z) depreciation of the official exchange rate (in local currency terms), and a 30-fold rise in producer prices of export crops, but the annual average inflation rate over the same period did not exceed 40 percent. 2-18 Figure 2-3. Official Exchange Rate and Middle-Income CPI, 1986 to mid-1989 (percentage change) 350 - 300- 250 - 200 - 150 - 100 - 50 Jan. 1986 Jan. 1987 Jan. 1988 Jan. 1989 Jul. 1989 0 OFFICIAL + CPI MIDDLE INCOME 2-19 The prices of domestic imported inputs bought under OGL or the Special Import Program (SIP) will rise, but the market price of their output changes little because it already reflects the parallel rate.10 However, to the extent that the parallel market rate moves--even if less than proportionately--with the devaluation, or the government's fiscal/monetary stance generates expectations of such a movement, the effect is inflationary. As overvaluation of the official rate increases, expectations of its devaluation often lead to depreciation of the parallel exchange rate without any change in the official rate. Thus the greater the inconsistence between monetary/fiscal policy and the level of the official rate, the greater the depreciation of the parallel rate and the greater the rate of inflation in expectation of an official devaluation. Liberalization and Capital Flight It is sometimes argued that liberalization of the trade and payments regime will lead to capital outflow and that capital flight can be prevented by quantitative controls. Yet evidence from a number of countries confirms that capital controls are highly porous whenever it is profitable to transfer capital abroad. Nowhere is this likely to be more true than in Uganda, where the parallel market has served as a conduit for capital flight for most of the 1970s. But with increased liberalization, growing economic activity, and 10 For example, the July 1988 devaluation from 60 new U.Sh./dollar to 150 U.Sh./dollar, implying a devaluation of 150 percent in local currency terms, raised the market price of domestically produced cigarettes and beer by only 14 and 42 percent, respectively. 2-20 rising confidence, the parallel market seems to have become a source of net capital inflow. For most of the 1970s and the early 1980s, only contraband current and capital account transactions were channeled through this market. Receipts from smuggled exports constituted the dominant source of foreign exchange in the parallel market in the 1970s. Currently, this is less the case, but in the absence of any substantive information on the size and source of supply of foreign currency in this market, one can only speculate on its composition and quantity. Foreign currency receipts of resident expatriates is clearly one source. So is return of flight capital. From all indications, remittances from Ugandans working abroad probably dominate the current foreign exchange supply to the parallel market.11 Foreign currency receipts from overinvoicing of imports is another possibility, although given the customs inspections, one would expect it to be limited. If we assume 20 percent overinvoicing, which is a high estimate, such receipts would be US$20 million at most--less than one-fifth of no-forex imports. Unofficial coffee sales are probably not a large source of parallel market foreign exchange. Deliveries of coffee to the marketing board have been rising rapidly every year since 1987 although coffee acreage did not increase immediately in 1987. So either the volume of smuggled coffee exports has fallen or a large amount of coffee has been kept in storage. It also seems likely that unrecorded cross-border trade in noncoffee products has at least not increased and probably even declined 11 This is not to imply that there is no capital outflow, but only that the net capital flow is inward through the parallel market. 2-21 following the mid-1987 change allowing "official" exporters to retain their foreign currency export earnings (see chapter 7). Since 1986 the parallel market has acquired a quasi-legal status because the government has permitted licensing of imports not funded by official foreign exchange. Such importers claim to be using their "own" foreign exchange, ostensibly the earnings of family or friends working abroad. In practice, however, the foreign exchange is purchased from the parallel market. The value of such licensed imports probably indicates the lower bound on the level of parallel market transactions. Changes in the supply of foreign currency in this market undoubtedly depend on changes in the parallel exchange rate relative to the official rate and changes in private sector confidence in the Ugandan economy. Both have been on the rise, as has private sector supply of parallel market foreign exchange. No-forex import levels have risen despite a decline in smuggled exports, which suggests that there is a net capital inflow through the parallel market that is adding to Uganda's import capacity. CHAPTER 3 REVENUE IMPLICATIONS OF TAXES Uganda's macroeconomic problems stem in large measure from the money financing of the private sector and the government. The highly inefficient system of coffee crop-financing is a major factor behind private sector credit creation, while Central Bank financing of that part of the fiscal deficit that cannot be financed through external grants and loans is responsible on the government side. Both have been highly inflationary. 3.1 Tax Effort Uganda's tax-GDP ratio has been both low and unstable: deteriorating during the 1970s, recovering in the early 1980s, then falling again (see table 3-1). All major taxes have exhibited a similar pattern. The primary causes have been fluctuation and decline in the terms of trade (especially the decline in the world price of coffee) and in the real producer price and world price of coffee, appreciation of the official exchange rate, and deterioration in tax administration. Although the domestic market prices of imports and competing domestic output reflect changes in the parallel market rate, imports and domestic output are valued for tax purposes using the official exchange rate, which has lagged behind inflation. This valuation practice reduces the base for indirect taxes and undermines the responsiveness of indirect tax revenue to the general price level. In addition, the tax administration's lack of auditing capacity permits manufacturers to under- 3-2 Table 3-1. Tax-GDP Ratio and Tax Structure in Uganda, 1970/71-1988189 (Z of GDP in market prices) Sales tax and excise Tax- Tax on imports Tax on exports duty on Fiscal GDP Income Customs Sales domestic year ratio tax duty tax Total Coffee production 1970/71 12.6 2.25 2.72 0.93 2.59 2.23 3.02 1971/72 12.8 3.08 2.57 1.14 2.47 2.03 3.23 1972/73 10.0 2.03 1.35 0.93 2.75 2.61 2.69 1973/74 7.8 1.35 1.20 0.59 2.24 2.11 1.93 1974/75 10.2 0.93 1.41 1.55 3.64 3.53 2.42 1975/76 9.0 0.78 1.31 1.25 3.63 3.62 1.86 1976/77 7.9 0.71 0.54 1.16 3.84 3.84 1.54 1977/78 9.7 0.73 0.75 1.00 5.81 5.76 1.23 1978/79 3.4 0.41 0.39 0.50 1.33 1.33 0.70 1970/80 3.5 0.42 0.35 0.52 1.22 1.22 0.60 1980/81 1.6 0.28 0.32 0.29 0.07 0.07 0.42 1981/82 7.9 0.74 1.69 1.61 2.21 2.21 1.43 1982/83 10.3 0.49 1.27 1.28 3.18 3.18 1.38 1983/84 11.2 0.79 1.18 1.48 5.20 5.20 1.19 1984/85 9.3 0.56 0.82 1.28 5.48 5.37 0.85 1985/86 6.5 0.36 0.40 0.57 4.39 4.34 0.62 1986/87 4.5 0.51 0.53 0.57 1.81 1.81 0.88 1987/88 5.6 0.47 0.55 0.58 1.62 1.62 1.85 1988/89 5.2 0.57 0.93 0.63 0.70 0.69 2.03 Source: Ministry of Finance; Government Financial Statistics; Background to Budget, 1989/90; Ministry of Economic Planning; IMF. report even the domestic component of their ex-factory prices, which otherwise should have moved with changes in the domestic price level. Given the dominance of indirect taxes in government revenue, this has resulted in low inflation-responsiveness of total tax revenue. Estimates of tax elasticities during 1970-87 indicate that all major taxes 3-3 were inelastic with respect to the general price level, although the elasticity of all taxes with respect to real GDP was around one (table 3-2). In the face of rapid inflation, this low responsiveness of the overall tax system to inflation explains a large part of the decline and instability of Uganda's tax effort. Table 3-2. Elasticity of Major Taxes in Uganda, 1970-87 With respect to Real GDP Type of tax GDP deflator Tax revenue net of export taxes 1.02 0.56 Income tax 0.96 0.74 Import tax (sales tax and customs duty) 1.06 0.37 Domestic sales tax and excise duty 0.98 0.69 Commercial transaction levy 1.03 1.03 Uganda's tax-GDP ratio is the lowest among low-income Sub-Saharan African countries (table 3-3). This is particularly evident with respect to import taxes and direct taxes. Uganda also has the lowest share of direct taxes in total revenue: 11 percent versus 30 percent for Sub-Saharan Africa as a whole. In part, this is due to the large size of Uganda's agriculture sector, which constitutes 74 percent of GDP (twice the average of other Sub-Saharan countries) and, as in those countries, is not taxed directly. Second, Uganda has the narrowest personal income and business tax base. At more than six times per capita GDP, Uganda's income threshold for individual income taxes is higher than in most countries (it is less than two times per capita GDP in Togo and Zambia, for example). This high threshold 3-4 Table 3-3. Tax-GDP Ratios and Tax Shares in Selected Low-Income Sub-Saharan Africa Countries, 1987 Percentage of Total Tax Revenue GNP per Tax on Tax on Tax- capita Direct domestic international GDP Country (US $) taxes transactions trade ratio Sub-Saharan Africa 252 29.8 35.2 36.4 17.5 Countries a Zaire 150 36.3 24.4 39.4 13.9 Malawi 160 43.3 36.0 20.5 18.5 Tanzania 180 27.2 63.8 9.1 15.5 Burkina Faso 190 27.9 32.9 43.9 13.7 Mali 210 14.2 54.6 31.3 13.6 Zambia 250 24.2 41.9 33.9 23.7 Togo 290 54.0 13.7 41.5 24.7 Kenya 330 34.1 44.3 21.5 18.5 Lesotho 370 12.4 11.7 75.8 19.7 Ghana 390 24.0 28.4 47.5 13.0 a. Other low-income Sub-Saharan Africa countries were excluded due to lack of data for 1987. Source: World Development Report 1989, pp. 164 and 186. automatically exempts a large number of individuals from personal income tax. Company income tax rates are high at 50 percent (although this is not much higher than in most Sub-Saharan Africa countries), but company tax collection is low because activities in the underground economy are widespread. Third, personal and company income constitute the entire direct tax base since there are no property taxes. Finally, Uganda has significant weaknesses in its tax administration.1 For these reasons, the potential for raising direct taxes appears to be considerable. 1 Recurrent expenditures of the Ministry of Finance accounted for over 30 percent of total tax revenue collected in 1988/89. 3-5 3.2 Structure of Indirect Taxes Domestic indirect taxes and international trade taxes provide a higher share of revenue in Uganda than in neighboring countries, although customs duty collections are much lower. Uganda's indirect taxes include excise duties, sales taxes, customs duties, export taxes (mainly on coffee), and taxes on services (commercial transactions levy). The excise tax is applied ex-factory on a number of locally manufactured products. A sales tax is levied on the excise duty-inclusive ex-factory value of domestically manufactured goods and on the customs duty-inclusive c.i.f. value of imports. Imported inputs are exempt from sales tax. Customs duty is levied on the c.i.f. value of imports, and the export tax is applied to coffee exported through the Coffee Marketing Board as a residual (after the producer price and marketing margins are paid to farmers, coffee processors, and the coffee board). Taxes on international trade have dominated the total tax effort (in most years accounting for at least half of tax revenue and as much as four-fifths in coffee boom years), and coffee export taxes have dominated trade taxes. Fluctuations in coffee export prices have thus been an important source of instability in total tax revenue and pose a major challenge to any reform efforts. Customs duties have not provided more than 14 percent of revenue for most of the 1980s. Though the modal duty rate is around 30 percent, duty collections have not exceeded 6-7 percent of total import value. (Table 3-4 reports the share of each indirect tax in total tax collections for 1987/88 and 1988/89.) The most noticeable feature of Uganda's indirect tax collection is the narrowness of the base. Over 95 percent of excise duties and 80 3-6 Table 3-4. Structure of Ugandan Indirect Taxation, 1987/88 and 1988/89 Millions of Percent share of new Ugandan shillings total tax revenue Type of taxes 1987/88 1988/89 1987/88 1988/89 Total tax revenue 18,320 43,885 100.0 100.0 Total indirect taxes 16,016 38,096 87.4 86.8 Taxes on domestic production 6,524 19,211 35.6 43.8 Sales tax 4,301 12,263 23.5 27.9 Excise duty 1,711 4,786 9.3 10.9 Commercial transactions levy 375 942 2.0 2.2 Other minor taxes 138 1,219 0.8 2.8 Taxes on international trade 9,492 18,885 51.8 43.0 Imports Customs duty 1,866 7,775 10.2 17.7 Sales tax 2,351 5,271 12.8 12.0 Export Tax 5,275 5,839 28.8 13.3 Source: Ministry of Finance, Uganda. percent of sales tax revenue were collected from the domestic production of three items: beer, cigarettes, and soft drinks. Their effective tax rates (excise plus sales) are 214 percent, 215 percent, and 70 percent, respectively. The overall effective rate of domestic indirect tax averages 48 percent. The effective rates for beer and cigarettes seem excessive, even though their output growth appears to be unaffected. Recently, excise duties have been lowered because of this concern. In part, the high rates are offset by the subsidy on inputs under the Open General License system, but they also reflect the low levels of total industrial output in other sectors and the compulsion to generate revenue. With respect to trade taxes, a similar narrowness of base is evident: motor vehicles generate 60 3-7 percent of sales tax revenue from imports, petroleum products and motor vehicles provide around 70 percent of the revenue from customs duty, and coffee provides 99 percent of export taxes. The scope for increasing revenue by raising rates on these items is clearly limited. However, other changes could increase revenue from existing transactions. In the medium term, increases in domestic indirect tax revenue will have to come from expansion of output through improved capacity utilization and from shifts in economic activity from the underground economy to the official or recorded sector. Import liberalization and the liberalization of the private sector will contribute to expanding the tax base. Tax policy changes should also aim to encourage such expansion. Excise Duties Currently, all excise duties are ad valorem ranging from 5 percent to 90 percent. The highly differentiated rate structure creates problems for tax administration. Excise duties are collected monthly by the Customs and Excise Department. The tax liability per unit of output is computed by applying the relevant excise duty rate to the ex-factory price, which is defined as cost of production plus a markup (normally 10-20 percent). In effect, the excise tax is similar to the sales tax, which is also imposed on the ex- factory price--a similarity that calls into question the need for excise taxes. In terms of the long-run reform of the system, only sales taxes at the factory level could replace excise taxes altogether. Because of a general lack of proper bookkeeping, auditing of production costs by the tax authorities is impossible. The most tax 3-8 collectors and tax examiners can do is to compare production costs per unit within an industry. As a result, manufacturers are encouraged to show lower than actual production costs, thereby lowering not only their excise and sales tax levies but their corporate income taxes as well (which are computed as a percentage of the markup, itself defined as a percentage of the production cost). In addition, producers who are also wholesalers of their manufactured products can easily divert a large proportion of their profit from the production stage to the wholesale level, which is not subject to any kind of tax. Sales Taxes The sales tax is applied to all domestic and imported manufactured goods. Exemptions to sales taxes on domestic output apply to capital goods, re-exports, basic foodstuffs, inputs, and a number of items for the armed forces.2 The exemptions for the armed forces hide the extent of the subsidy provided to the armed forces personnel--if that is what the exemptions are designed to do. They also encourage leakage and misuse of exemptions. The sales tax is a single-stage tax collected from manufacturers and importers. Cascading is avoided through the use of the suspensive or "ring" system, a system selected for its administrative simplicity. Manufacturers that have registered for the sales tax receive an Authorized 2 Beer, wines, spirits, cigarettes, tobacco, soap, detergents, toothpaste, brushes, matches, and canned foods. In 1988/89, sales taxes of U.Sh. 418.5 million on domestically manufactured beer, cigarettes, soft drinks and soap were lost due to these exemptions. Furthermore, the subsidy is greater for beer than for toothpaste, for example, thereby subsidizing beer drinkers more than toothpaste consumers. 3-9 Trade Certificate that allows them to import or procure locally the needed inputs free of sales tax. No one knows whether registered manufacturers divert sales-tax-free inputs to nonregistered manufacturers. There are twelve different ad valorem sales tax rates, ranging from 10 percent to 120 percent. This highly differentiated rate structure poses administrative problems and creates distortions in resource allocation. Nor are the rationales for some of the rates convincing on either equity or efficiency grounds. For example, the maximum sales tax rates are 50 percent on motor vehicles, 120 percent on soft drinks, 80 percent on wheat flour, and 10 percent on exercise books but 40 percent on the paper used in their production. The mission was told that unified sales tax rates are now applied to both imported and domestically manufactured goods. In 1988/89, differential rates were applied for some goods: for example, sales tax rates on domestically produced shoes, cooking oil, wheat flour, tubes, and tires were lower than rates on competing imports. Import Duties Uganda's revenue from customs duties has averaged around 14 percent of total tax revenue over the last two years, substantially lower than in most Sub-Saharan African countries. Petroleum products contribute around half of that amount, although they constitute only 10 percent of imports by value. Motor vehicles contribute another two-fifths. The average effective duty rate (duty collections as a percentage of import value) is around 6 percent even though the modal statutory tariff rate is 40 percent--in part because Uganda's tariff schedule is riddled with exemptions (see chapter 4). Indeed, before the budget of 1989/90, 3-10 half of all imports came in free of duty (table 3-5). But even if duty- free imports are excluded, the effective duty rate is still only 12 percent, considerably lower than in other countries in Sub-Saharan Africa. There is thus an urgent need to structure customs administration and to improve customs-based trade data to increase duty collection and monitor trade-flows. Table 3-5. Duty Collections by Tariff Rates in Uganda Import value Duty collection Tariff band (Z of total) (Z of total) 0 48.5 0.0 5 10.4 8.6 10 12.5 21.0 15 0.7 1.7 20 13.8 37.9 25 0.2 0.3 30 9.0 19.0 40 3.0 7.0 50 0.5 1.8 60 0.0 0.0 70 0.2 1.0 80 0.0 0.0 100 0.02 0.3 150 0.04 0.7 200 0.02 0.5 Source: Customs and Excise Department, Ministry of Finance, Uganda. Currently, duty rates above 50 percent provide about 2.5 percent of total duty collections, those at 5 and 10 percent account for 30 percent, and those at 20-40 percent contribute more than 60 percent of import duty revenue (table 3-5). The high rates are clearly being evaded, so eliminating them will not reduce tariff revenue very substantially. 3-11 Indeed, if existing import value now coming in at tariffs greater than 50 percent were taxed at a tariff rate of 50 percent, tariff revenue would be higher than at present. In addition, eliminating unwarranted exemptions would lead to a rise in tariff revenue, and eliminating even half of all exemptions would result in a substantial revenue increase. IIIIIIII1111 1 .. - Ino ll ill CHAPTER 4 ELEMENTS OF THE TRADE AND PAYMENTS REGIMES Incentives for resource allocation in the external sector are largely determined by the trade and payments regime, although domestic taxes and regulations also play a role. The major elements of Uganda's trade and payment regime are the exchange rate system, import policy (quantitative restrictions and tariffs), and export policy (licensing and implicit/explicit taxes). Documenting policies and changes in policies is not an easy matter in Uganda. No government publication compiles rules and regulations governing import and export policy, including licensing restrictions. New regulations are advertised in the press, but they are never collected in a single, published source. So entrepreneurs and traders are frequently unaware of the regulations and of incentives provided by government. Only tariff rates are available in one place in a printed Tariff Schedule, but the last version was printed in 1983. This schedule, annotated with hand- written changes based on budget speeches, continues to serve as the major source of information on the tariff structure for the Customs and Excise Administration. Even this document is not available to all customs officials and is rarely accessible to importers. Thus, a major problem with Uganda's trade and payments regimes is the dissemination of information about them. The first task of reform should be to institute documentation of the licensing system, the tariff system, and the domestic tax system, at regular intervals (say six months) and to make this information widely available to all. 4-2 4.1 The Exchange Rate Regime With both an official and a parallel foreign exchange market, the Ugandan economy faces two, often very different, exchange rates. Coffee exports, government imports, debt-servicing, and oil imports are transacted at the official exchange rate. All other imports and noncoffee exports are de facto subject to the parallel rate. The existence of two rates raises important questions: which is the relevant marginal rate for importers and exporters or which rate represents the opportunity cost of foreign exchange to Ugandans? For the incentives system, this is a critical question. The marginal rate of exchange for imports appears to be the parallel rate. In Uganda, as in countries in which the parallel market is an important source of merchandise imports, the domestic price of imported goods often reflects the parallel market exchange rate. Changes in the nominal parallel exchange rate track changes in the consumer price index (CPI) much more closely than do changes in the official exchange rate (figure 4-1). This relationship holds for various components of the CPI, including other consumer goods (see figures 4-2), though it is much weaker for fuel and food. Over a longer period, the correlation between the annual inflation rate and the annual rate of depreciation in the parallel rate is also greater than with the official rate. A quick survey by mission members of a small number of consumer goods in the Kampala market that were being imported at both official and parallel market exchange rates confirmed domestic currency prices considerably higher than would be obtained by converting world dollar prices at the official exchange rate. Domestic producers also appear to price goods to compete with no-forex imports, especially for consumer 4-3 Figure 4-1. Official and Parallel Exchange Rates and Middle-Income CPI, 1986 to mid-1989 (January 1986 = 100) 2.8 - 2.6 - 2.4 - 2.2 - 2 1.8 1.6 0.8 0.6 0.4 - 0.2 J86 J87 J88 J89 C OFFICIAL + PARALLEL 0 CPI MIDDLE INCOME Note: To read percentages off the Y axis, multiply each unit by 100 (e.g., 0.9 = 900). Figure 4-2. Official and Parallel Exchange Rates and Other Consumer Goods, 1986 to mid-1989 (January 1986 = 100) 2 1.8 1 .7 1.6 1.5 - 1.4 3 1.3 1.2 1.1 0.9 0.8 0.7 - 0.6 - 0.5 - 0.4 - 0.3 - 0.2 - 0.1 -a J86 J87 J88 J89 3 OFFICIAL + PARALLEL o OTHER CONSUMER GDS Note: To read percentages off the Y axis, multiply each unit by 100 (e.g., 0.9 = 900). 4-4 goods. And, for products subject to implicit or explicit controls on the factory price (based on the official exchange rate), distributors and retailers were clearly reaping (and maybe also sharing with producers) the rents by pricing goods closer to levels implied by the parallel market rate. The marginal rate for exporters is less straightforward. Ninety- five percent of official export (i.e., coffee) receipts are surrendered at the official rate. These exporters receive a fixed producer price that is not always adjusted for changes in the exchange rate. By contrast, all "official" noncoffee exports, which have been rising rapidly from a small base, are effectively subject to the parallel rate by virtue of the foreign exchange retention scheme. Unrecorded coffee exports are no doubt subject to the parallel rate, although given the costs of smuggling, that is clearly not what exporters receive. In view of the obvious overvaluation of the Ugandan shilling, the marginal rate for exports is taken to be the parallel rate for the purpose of this analysis. 4.2 Foreign Exchange Allocation and Import Licensing Given the overvaluation of the official exchange rate, Uganda's foreign exchange allocation and import licensing system has been the dominant instrument for restricting official imports. The government's concern about liberalizing this regime relate to worries about exhaustion of foreign exchange reserves, inappropriate use of scarce foreign exchange, and the effect of import competition on domestic industry. The foreign exchange allocation mechanism has gone through several changes since 1982. Until 1987, foreign exchange for nonproject and 4-5 nonpetroleum purchases was allocated by a high-powered ministerial committee headed by the Prime Minister and comprising the ministers of finance, planning, industry, commerce, agriculture, energy, and transport. The committee formulated a monthly foreign exchange budget on the basis of Bank of Uganda projections of export receipts. "First-charge"1 expenditures were netted out, and then the balance was allocated to the private sector on the basis of a set of priorities established by the ministries. In the last two years, first-charge allocations have virtually exhausted available export earnings, so this process of allocating foreign exchange has virtually ceased. Instead, foreign exchange, obtained mainly from external grants and loans, is now allocated by a smaller committee (chaired by the Bank of Uganda, with representatives of the ministries of industry, planning, commerce, finance, and agriculture). Allocations are approved more frequently now under several official import schemes. Official foreign exchange allocations finance only around two- thirds of all nonproject and nonpetroleum imports. The balance is financed by importers' "own" foreign exchange, which is obtained mainly from the parallel market. These no-forex imports have averaged around US$100 million over 1987/88 and 1988/89 (see table 4-1), and their share of nonproject and nonpetroleum imports has increased significantly over earlier years. 1 First-charge allocations relate to debt service and foreign office expenses, and payments for oil imports, other government imports, and the imports of parastatals and export marketing boards. Since March 1989, only coffee export receipts have been available to the Bank of Uganda for such allocations. This is unlikely to make much difference, since noncoffee exports rarely exceed 5 percent of total exports. 4-6 Table 4-1. Total Ugandan Imports, 1986/87 to mid-1989 (millions of U.S. dollars) July-Dec. Jan.-June July-Dec. Jan.-June Item 1986/87 1987 1988 1988 1989 Total 494.6 336.8 316.7 313.6 338.8 Project imports 1814.5 114.3 114.6 130.2 126.1 Petroleum imports (cif) 83.0 31.3 37.7 37.91 37.1 Nonproject and nonpetroleum imports 247.0 189.9 183.4 145.53 175.6 With officially allocated 109.1 90.7 46.7 80.87 78.3 foreign exchange a Without officially allocated 70.5 48.5 57.6 39.10 54.5 foreign exchange (no-forex) b Other imports c 57.4 50.8 59.1 45.5 42.8 Share of total imports (%) 52.0 56.6 51.8 48.4 51.8 a. Allocations made by the Bank of Uganda to firms licensed by the ministry of commerce using foreign exchange not committed to aid-financed development projects or to petroleum imports. Only a small proportion of this foreign exchange came from the country's own export receipts; the rest was financed by donor import-support programs. b. Imports using importers lown' foreign exchange, which is generally obtained from the parallel market. c. Imports under suppliers' credit, under Kenya compensation funds, under government-to-government barter, and the like, and imports using own export proceeds under a dual license (back-to-back export-import license) system. Source: Bank of Uganda (data is on payment basis). Although total official-allocation imports have not grown, their level remains substantially higher than 1986/87 levels. External donor financing has become increasingly important in funding the chief import schemes: Open General License (OGL) and Special Import Programs I and II (SIP-I and SIP-II) until October 1989, and SIP-III and OGL schemes since then. 4-7 Import Controls Once a firm has obtained access to foreign exchange (generally from the Bank of Uganda), it applies to the ministry of commerce for an import license. The license is required to open a letter of credit. The quantity and composition of imports are controlled mainly through the foreign exchange allocation process. The licensing system seeks mainly to ensure the applicant's legal status as a trader/manufacturer. It also seeks to assess merchandise price and quality data with the help of a preshipment inspection firm, the Soci6t6 General de Surveillance (SGS) to prevent importers from overinvoicing, an issue of considerable concern to the authorities.2 Except for petroleum and project imports, both of which are imported by the government, Ugandan imports currently come in under three basic arrangements:3 * Official allocation arrangements, under which the importer seeks foreign exchange allocations from the Bank of Uganda under the OGL or SIP-III schemes. Between January and October 1989, OGL, SIP-I, and SIP-II were the official schemes. 2 Applications for import licenses had to be accompanied by the following four documents: (1) proforma invoice, indicating the nature of the product, its quantity, and its unit price; (2) income tax certificate for established firms or advance tax assessment for new businesses; (3) certificate of business registration from the ministry of justice; and (4) trading license from local authority. The income tax certificate is no longer required. 3 This excludes imports under government-to-government barter and suppliers credit, which, although important, are not amenable to useful "liberalization." 4-8 * The export retention scheme, under which exporters retain foreign currency receipts from noncoffee exports and use them to import eligible items (the eligibility list is the same as for SIP). * The "no-forex-required" arrangement, under which no foreign exchange is officially allocated, but licensed importers use their "own" foreign exchange (usually obtained on the parallel market). The first arrangement is clearly subject to official decisions about the quantity and composition of imports. The third, because licensing is automatic and follows importation, is subject to little or no government control. The export-retention arrangement falls between the two extremes, but is more liberal than OGL and SIP: export earnings are retained by the exporter, and control over export composition is similar to control under SIP, but less stringent and there is no official control over the volume of "retention" imports. The severity of the government's exchange control varies. Under the no-forex arrangement, there is no restriction on volume or composition; any item not banned can be imported in any amount. The volumes of export- retention imports is determined by the volume of exports. It is subject to a positive "list" of imports, but there are no official priorities or targets to which importers must adhere. Greater restrictions apply for the official schemes, among which SIP-II was the most liberal. The new SIP- III, which came into existence in 1990, appears to be more restrictive than SIP-II, having a larger negative list and using "notional" subsector allocation targets like those used in SIP-I. The stringency of the licensing process also varies. The appropriateness of prices and the quality of imports are checked under all 4-9 official schemes, but no-forex and retention scheme imports are not subject to such checks. Similarly, letters of credit are not necessary for no- forex but are needed for all others.4 The official schemes are part of Uganda's attempts to liberalize official imports. Although substantial devaluations and donor funding permitted their creation, expansionary financial policies and inadequate adjustment of the official exchange rate set practical limits on the degree of import liberalization possible. Demand has always exceeded supply quickly, except in the case of SIP-II, which was priced higher (a more depreciated official rate)5 and required importers to have positive bank balances. The temporary nature of SIP makes this inevitable, as importers bring their future import demands forward whenever an announcement is made that SIP funds are available. Open General Licensing The Open General Licensing (OGL) system initiated in May 1987 was fully working by January 1988. Although an OGL system is generally viewed as an important step toward import liberalization, Uganda's OGL system is not typical.6 It is neither open nor general. OGL foreign exchange is 4 Recently this requirement was relaxed for SIP, for which the consignment value is $5,000 or lower. 5 The exchange rates were U.Sh. 200/dollar (official) and U.Sh. 400/dollar (SIP-II) in September 1989; in November, they were U.Sh. 370/dollar and U.Sh. 400/dollar. 6 The features of a standard OGL scheme are as follows: any importer is eligible to import goods permitted under the scheme. Goods selected for inclusion would generally be outputs of(formerly) highly restricted/ protected import-substituting industries or inputs into (formerly) lightly protected industries. Gradually more and more goods are brought into the OGL system, until most allocations are determined mainly by the exchange rate rather than administrative decision. 4-10 available at the official rate but is limited to imports of industrial inputs and spares. Moreover, not all industrial sectors are eligible and not all firms in a chosen industrial sector or subsector can avail themselves of OGL imports. Four criteria were the stated basis for choosing the industrial subsectors. They were as follows: * The product produced by those subsectors should be an essential input for other priority industries or an essential mass consumption good. * The subsectors should help generate substantial tax revenue. * Production in the subsectors should require few imports to ensure a high multiplier effect. * The subsectors should facilitate or support export production. Eight industrial sectors were chosen in 1988: soap, soft drinks, cigarettes, mattresses, beer, textiles, crown seals, and cement. This choice does not seem balanced in terms of the expressed criteria; it supports production of consumer goods (in six out of eight sectors) but not production of inputs for these sectors.7 The only input-producing sectors selected were crown seals and cement. Nor was there support for export production, especially since agricultural inputs were excluded. If anything, imported inputs under OGL were higher than the average for manufacturing as a whole. No implicit criteria of economic efficiency or viability are apparent from the choice of sectors. In fact, products such as fishnets, gunny sacks, and hessian cloth were found to be uneconomical in a 1986 study (Industrial Sector Memo), yet two firms that produce them were selected for OGL. The choice of firms within sectors was based on 7 As table 4-2 shows, 96.5 percent of OGL allocations went to firms producing consumer goods. 4-11 Table 4-2. Imports of Inputs Under OGL, January - December 1988 Utili- Share of No. zation public Industrial of Allocation Sectoral rate sector subsector firms ($ million) share (Z) (Z) (Z) Soap 5 12.58 45.0 80.8 0.05 Soft drinks 3 4.42 15.8 89.0 2.40 Cigarettes 1 3.12 11.2 55.7 0.00 Mattresses 4 2.97 10.6 60.6 0.00 Beer 2 2.55 9.1 32.5 0.36 Textiles 6 1.31 4.7 98.1 0.87 Crown seals 1 0.80 2.9 95.3 0.00 Cement 2 0.19 0.7 15.8 Total 24 27.74 100.0 83.8 14.00 Source: Bank of Uganda. judgments about their management ability and their access to working capital or credit, although why this preselection was necessary is not obvious.8 In terms of performance, US$55 million was available under the OGL, but total license approvals amounted to US$27.74 million for 1988 (twenty-four firms availed themselves of OGL allocations) and US$12.6 million was approved in the first eight months of 1989. About 94 percent of imports were raw materials and intermediate inputs; the rest were spares (see table 4-2 for actual OGL allocations by industry in 1988). Despite slow allocations and imports, four subsectors met OGL's limited expectations and objectives for 1988 (soap, soft drinks, cigarettes, and 8 Initially, twenty-two firms were eligible, another three were added later. 4-12 beer). Their output rose faster than that of other manufacturing subsectors, and domestic prices for products of those industries fell.9 Increased output raised indirect tax revenues, an important OGL objective. It is a moot point whether enough revenue was earned to justify "subsidizing" inputs so highly, through an overvalued exchange rate. It appears (see table 4-3) that for soap, mattresses and textiles, the government could have increased its receipts even more, or at least by as much, by selling the donor-provided foreign exchange to firms at a more appropriate exchange rate. The OGL is administered by the Bank of Uganda (BOU). Approval of OGL applications has typically taken three to four weeks, and efforts are under way to reduce that time. OGL procedures make excessive demands on BOU's scarce administrative resources and importers' financial resources.10 When the firms' annual demand for imported inputs is being estimated as the basis for allocations, BOU tries to ensure that the items are legitimate 9 Output in 1988 rose 14.5 percent (for soap), 12.1 percent (soft drinks), 14.1 percent (cigarettes), and 27 percent (beer). Only soap and beer output grew faster in 1987. 10 Until March 1989, importers had to supply evidence of the availability of local currency funds for importing at the time of application. A fee of 1 percent is paid to BOU for the SGS, whose valuation service is used to assess invoice prices. Once allocations were approved, the applicant's bank account was automatically debited for 100 percent of local coverage under BOU rules. All documentation was returned to the applicant's commercial bank so that the OGL applicant could submit them to the Ministry of Commerce (MOC) for an import license at a fee of 1 percent of the value of imports. The process of transferring documents (from applicant to BOU to commercial bank to MOC and back to the commercial bank) involved delays of three to four months. After obtaining the license, further delays occurred. Recently BOU has relaxed the rule of full local coverage, leaving the extent of coverage to the discretion of commercial banks. 4-13 Table 4-3. Cost of Raising Sales Tax Revenues Using DCL Imports (1) (2) (3) (4) (5) (6) (7) Cost of Sales rev. Cost of OOL foreign Official rate collected foreign Revenue utilization exchange Value of of tax on from OCL ex to forgone Sector Firm 1988 to firms sales saIes firms govt by govt (SUS 000) (mill. sh.) (mill. sh.) () (mill. sh.) (mill. sh.) (mill. sh.) Soft drinks Jubillee ice & soda 43.8 4.8 7.1 100.0 10.0 13.1 3.1 Kampala bottlers 537.9 58.6 215.4 100.0 275.9 161.4 -114.5 Lake Vic bottling 2,106.0 229.6 1,516.8 100.0 919.4 631.8 -287.6 Cigarettes B.A.T. Uganda 3,012.1 328.3 2,486.7 50.0 5,548.0 903.6 -4,644.4 Beer Uganda breweries 2,115.3 230.6 1,945.0 60.0 3,855.1 634.6 -3,220.5 Nile breweries 157.9 17.2 395.7 60.0 650.3 47.4 -602.9 Textiles Mulco textiles 72.4 7.9 154.4 20.0 24.9 21.7 -32.0 Uganda blanket manufacturers 112.5 12.3 62.9 20.0 11.1 33.8 22.7 Nyanza textiles 213.0 23.2 1,267.9 20.0 241.5 63.9 -177.6 Uganda bag A hessian 224.5 24.5 92.4 20.0 15.9 67.4 51.5 Uganda fishnet manufacturers 107.0 11.7 225.7 20.0 0.0 32.1 32.1 Soap Mukwano industries 11,824.4 1,288.9 2,525.4 3.0 98.4 3,547.3 3,448.9 Cement UCI Hima 33.0 3.6 171.1 25.0 42.8 9.9 -32.9 Mattresses Vitofoam 1,540.4 167.9 157.4 30.0 41.4 462.1 420.7 Column (1) is data on 0CL usage provided by the Bank of Uganda. Column (2) is the cost to firms of 0CL foreign exchange calculated at 109 new Uganda shillings per US dollar. Column (3) is value of sales taken from Department of Statistics Survey forms. Column (4) is taken from official rates of sales tax schedules. Column (5) is taken from Department of Statistics Survey forms. Column (6) is Column (1) x (409-109). Source: Calculated from manufacturing records and information supplied by the research department of the Bank of Uganda. inputs for the firms' eligible output and that their import and output volumes are reasonably consistent--that input-output coefficients are acceptable. Despite significant efforts on BOU's part, the mission found that several firms were importing inputs at levels that were much higher than was warranted by their output volumes. 4-14 The number of eligible firms in the eight subsectors grew from twenty-five in the first year to thirty-eight in the second. In late 1989, the number of OGL sectors increased, with the addition of pharmaceuticals, sugar processing, aluminum products, roofing iron, matches, and packaging products. It is not obvious what criteria were used to choose the new sectors. The mission learned that manufacturers preferred OGL over SIP-I allocations, partially because they were not confident about continued access to SIP-I. The costs of obtaining a SIP-I allocation are viewed as higher (for reasons discussed below) even if the exchange rate is the same. Despite the arbitrary nature of OGL's selection process, firms deploy considerable resources to be selected, and this rent-seeking imposes cost on the economy. In fact, the expansion of OGL may have given inappropriate investment signals to manufacturers. Special Import Program I The Special Import Program (SIP, or SIP-I as it came to be called) was more open and general than OGL in terms of access and eligible imports, but had the same applicable exchange rate as OGL. Given the higher transactions cost of access to SIP however, the cost of SIP-I imports was probably higher than for OGL imports. Moreover, SIP tends to be viewed by importers as a temporary scheme since there have now been three SIPs, each beginning and ending with a transient inflow of donor funds. SIP-I was launched at the government's initiative in December 1988, largely in response to the slow approval and use of import allocations under the OGL. The initial objective of SIP-I was to accelerate the flow of imported goods by using a donor line of credit and 4-15 to mop up liquidity in the banking system.11 By reducing administrative review of applications and permitting a wide range of import items (a large "positive" list) SIP-I became much more open and general than "Open General Licensing." SIP-I contributed significantly more to Uganda's import liberalization in the first six months (December 1988-May 1989) than did the OGL in its first eighteen months--even in terms of the volume and diversity of imported inputs for industry. Weekly allocation targets were specified in advance for each of the seven categories on the positive list.12 This was quite restrictive. All types of applicants (firms, traders, distributors, agriculturalists) were eligible on a first-come, first-served basis within each category until the allocation target was met. The mission could not ascertain whether this quicker allocation method was truly applied on a "first-come, first-served" basis. Since the exchange rate was overvalued, targets were quickly met and applicants were often arbitrarily rejected.13 Importers' transaction costs to minimize the possibility of rejection were substantial. In addition, the supply of foreign exchange to SIP-I was residual, and since 11 Under SIP importers had to provide bank statements indicating that they had positive balances in their own accounts that were at least equivalent to the value of the licenses sought. 12 Agriculture, industry (inputs only of raw materials and machinery), transport equipment, building or construction materials, education and sports, health, and consumer goods. 13 Between December 1988 and April 1989; allocations exceeded targets for industrial inputs/machinery (28.1Z vs. 20Z), transport (23.4Z vs. 15Z), and consumer goods (12.3Z vs 1OZ), Agricultural input targets were half met and the building materials target was met 93 percent. 4-16 overall demand for SIP-I quickly exceeded available foreign exchange, arbitrary rejections increased.14 The scheme's only restrictive feature was that applicants had to have a positive balance in their commercial bank accounts. This also made SIP-I costlier than OGL. Because of both transactions and financial costs, the rents available from access to each dollar of SIP-I foreign exchange were lower than the rents available from OGL. Special Import Program II In June 1989, Uganda launched SIP-II to allocate donor funds for imports at a more depreciated exchange rate than the OGL or SIP-I rate (U.Sh. 400/dollar instead of U.Sh. 200/dollar). Unlike the OGL, SIP-II was motivated by the need to obtain more revenue by selling donor funds to importers at a higher exchange rate. All imports (including consumer luxuries) were eligible,15 with no prespecified allocation or beneficiary targets. There was no check on importers' tax payments, but the prices, quality, and quantity of imports continued to be scrutinized at the time of licensing. In fact, SIP-II was a hybrid, similar to no-forex in that it used a short negative list for its imports and required no tax verification, but close to SIP-I in terms of scrutinizing invoices for 14 Letters of credit for SIP imports were opened at a monthly rate of $14.6 million in the first two months of its existence (December 1988 to January 1989). As aid funds dwindled, the rate of SIP imports declined to $5.8 million a month in the next three months (February to April 1989) and to $4.6 million in the next four (May to August 1989). By September, when the TEP Mission visited Uganda, SIP was not viewed by the private sector as a source of foreign exchange for imports. 15 Only second-hand clothes and reconditioned vehicles were not allowed. 4-17 prices, quantity, and quality of imports and requiring that a letter of credit be used to pay foreign suppliers. The usage rate for SIP-II licenses (40 percent) was lower than for SIP-I (96 percent): of the US$52 million in SIP-II licenses issued between June and August 1989, only US$22 million was used by September 1989. Does this suggest a fall in import demand? Probably not, since there was no major shift to a contractionary policy. If anything, domestic credit was expanding too quickly. However, strict enforcement of the requirement that importers have a positive bank balance reduced demand somewhat. Imports under no-forex were proceeding at the same average monthly rate (US$9 million a month) over this period, while the rate of SIP-II imports did not exceed US$7 million, though foreign exchange under no-forex was available at about a 50-percent premium over the SIP-II rate. The mission was repeatedly told by traders and manufacturers importing under both SIP-II and no-forex that "other" costs of importing were higher under SIP-II than under no-forex. For large purchases a discount was usually available on the quoted rate for foreign exchange on the parallel market. The quoted rate applied mainly to small cash transactions (less than US$1,000-$1,500). The discount on the quoted rate appeared to be 10-20 percent, depending on circumstances and the size of the transaction. The mission also learned from importers who used no-forex that they generally underinvoiced the dollar value of their imports by about 15-20 percent to lower their duty payments. Although many believe that no-forex importers pay no duties, most importers conceded that underinvoicing by margins greater than this attracted too much scrutiny by customs, which offset the cost savings through delay and side payments. Finally, the mission learned that virtually all no-forex imports could be 4-18 processed without letters of credit--that is, that bank drafts could be used to pay the supplier. Apparently letters of credit raise the cost of imports by 3-5 percent, because of the high cost of bank charges and amendments.16 So the cost advantage of SIP-II over no-forex may not be as large as the difference in exchange rates implies. In fact, the cost difference for high-tax items may even be negligible, as is shown in table 4-4. Table 4-4. Percentage Difference in Cost of No-Forex Imports Relative to SIP-II Imports Exchange rate t = 50 t = 50 t = 40 t = 30 t = 10 premium (Z) st = 70 st = 40 st = 20 st = 20 st = 0 50 -13 5 10 12 24 75 10 15 23 26 43 100 18 25 36 39 62 150 35 46 61 67 101 200 52 66 86 94 140 Note: t is the duty rate and st is the sales tax on imports. The cost- difference is estimated using the formula given below: 0.85 Eb + tm + st (1 + ti) - 1 (1 + tm)(1 + st) eo m t where eb and eo are the quoted parallel and official exchange rates. 16 Bank charges consisted of 1 percent of the LC's value to open the LC; 1 percent of the LC's value for each amendment; and costs from delays if SGS hold up shipments for inspection or for clarification from supplies. 4-19 A cost comparison on the basis of a 15-percent discount on the quoted parallel rate and 15-percent underinvoicing of the import value under no-forex suggests that an exchange rate premium of 50 percent may imply a difference in the landed cost of SIP-II and no-forex imports of not more than 10 percent if the tariff is 40 percent and the sales tax is 20 percent. The difference is even negative if they are 50 and 70 percent, respectively. Given other costs, this is a negligible difference. Only for inputs exempt from the sales tax and on which the duty is low is the difference large. Given the institutional factors underlying the official system for SIP-II, the exchange rate for SIP-II appears to be the closest to an exchange rate for which demand for foreign exchange does not exceed supply. Experience with SIP-II suggests that the official exchange rate, if set at an appropriate level, can be used to reduce excess demand for official imports in Uganda. Imports Without Officially Allocated Foreign Exchange Nearly a third of Uganda's imports (excluding petroleum and project imports) are not funded by officially allocated foreign exchange (called no foreign exchange or no-forex imports). The value of no-forex imports rose from US$70 million in 1986/87 to an average of US$100 million in the next two fiscal years, increasing Uganda's import capacity. Ostensibly, no-forex imports are funded by a firm's own foreign exchange, but it is common knowledge in Kampala that this "own" foreign exchange is purchased from the parallel market. Despite what seems to be wide de facto acceptance of a de jure illegal market, there was some evidence that such illegality constrains the more established enterprises. By law, possession 4-20 or purchase of foreign exchange from unofficial channels is illegal. Technically, importers who claim their "own" sources of foreign exchange but in effect buy it on the parallel market can be prosecuted. This means that the higher cost of such exchange cannot be shown for tax or other purposes. Of course, to the extent that tax authorities undertake no audits, this cannot be a significant constraint for most firms. Importers using no-forex foreign currency are required to apply to the ministry of commerce for a no-forex import license--generally after the no-forex imports have arrived. Letters of credit are not necessary, there is virtually no official scrutiny of the invoices to check on the prices or quality of merchandise, and tax payments need not be verified. The ministry of commerce uses the licensing process mainly to record the type of goods being imported on the parallel market; it exercises no control over the composition of these imports.17 The no-forex import scheme is clearly the most liberal. Import allocation is market-determined. Import demand unmet by the official market spills over into the no-forex market, albeit at a substantially depreciated exchange rate. A no-forex importer is a marginal importer, so the cost of no-forex imports determines their domestic market price.18 17 Some implicit restrictions probably apply. Although no official "ban" on imports is known to exist, it is difficult to explain why some high- demand consumer items (such as beer and cigarettes) are not imported under either official schemes or the no-forex scheme. 18 A no-forex importer's domestic currency price of one dollar of imports (PM) is given by PM = ep + [t + S(1+t)] eo, where e and eo are the parallel and official exchange rates, t is the tariff, and s is the sales tax. 4-21 4.3 Import Composition Under Liberalization In discussions of import liberalization, concern is often expressed about the potential for "inappropriate" composition of "market- determined" imports. In principle, under a liberalized regime imports would go to activities that earn the highest financial returns. Conceivably, those products or activities may not produce the maximum social rates of return. Two reasons are generally given for expecting such a socially undesirable market allocation of imports. First, Uganda's financial system cannot provide enough credit for productive enterprises, so import traders involved in short-term speculation are likely to be attracted to imports that produce a rapid turnover, large profits, and high liquidity. The more volatile the domestic market, the greater the impact of a certain composition of imports. For example, if consumer goods are profitable, they are likely to represent a bigger share of market-determined imports than are socially desirable capital goods and inputs, which are critical to increasing production. Second, when income distribution is highly skewed, demand for luxuries may be excessive, so even among consumer goods, luxury goods will dominate more essential, or "incentive," consumer goods. Import Composition Under "No-Forex* One reason to examine the composition of Ugandan imports under the liberal no-forex scheme is that theoretically it is determined solely by private-sector preferences. About 42 percent of no-forex imports were consumer goods in 1988 and 1989 (see table 4-5)--a higher share than for 4-22 SIP imports and foreign exchange retention imports.19 In interpreting this evidence, however, three facts should be kept in mind. Table 4-5. End-Use Composition of No-forex Imports, 1988-89 (percentages) Jan.- July- Jan.- July- June Dec. June Aug. End-use categories 1988 1988 1989 1989 Consumer goods 36.2 50.0 37.6 54.3 Automobiles/sedan cars 9.6 20.9 4.7 4.9 Electronics 3.7 3.4 4.3 3.4 Food 7.3 3.6 8.3 34.4 Other goods 15.5 18.1 20.3 11.6 Intermediate goods 36.2 18.5 25.3 22.5 Manufacturing Capital goods 32.6 35.5 37.1 23.2 Transportation equipment 21.5 22.8 20.5 11.7 Total a 100.0 100.0 100.0 100.0 Total import value ($000) 57,597 39,042 54,520 17,426 a. This excludes about 10 percent of imports that could not be classified because of inadequate information (e.g., assorted goods). First, Uganda's official foreign exchange allocation system repressed imports of consumer goods more than capital and intermediate goods and luxuries more than essentials--so the scarcity premium on those items was high. The no-forex market was effectively residual, so demand 19 The share is much higher in the second half of each calendar year (see table 4-5). About 43 percent of consumer goods in 1988, and 21 percent in 1989, were luxuries. 4-23 unmet elsewhere spilled over there--the more restricted the item, the greater the spillover. Items with such a high unmet demand could be expected to have a higher share in no-forex than in official imports. Second, not all consumer items were of low social priority. Considering Uganda's rundown industrial capacity, domestic production of essential consumer items (incentive goods) was lower than in other countries with similar income levels. The social rate of return for such imports is likely to be high because their supply is critical to sustaining reasonable consumption and motivating economic activity. A large share of no-forex consumer imports was for food and other essentials, for example, which was consistent with government priorities.20 In 1989, these consumer goods comprised 33 percent of total no-forex imports and three-quarters of no-forex consumer imports. Luxury consumer goods--such as refrigerators, cassette recorders, cookers, electronic household appliances, and sedan cars--constituted only 9 percent of no-forex imports and 21 percent of no- forex consumer imports. (These shares were considerably higher in 1988, presumably to build up stocks that had been run down since 1985.) Third, capital goods and intermediate inputs averaged 58 percent of no-forex, even though no-forex importing firms paid a substantially higher local currency price than official importers. Buses, trucks, mini- vans, pick-ups, bicycles, and spares constituted nearly two-thirds of no- forex capital-goods imports and one-fifth of total no-forex imports. Considering the importance of rebuilding Uganda's transport sector, market allocation in this direction was appropriate. 20 Edible oil, table salt, and baby food were among the food items. Other items were second-hand clothing, soap, matches, shoes, blankets, notebooks, and textiles. 4-24 These data on the composition of non-forex imports cannot be used alone to confirm or deny government concern. The important question is whether no-forex and official import liberalization have distorted the overall mix of imports. Overall Composition of Imports Among official schemes, foreign exchange retention and SIP-II are most liberal in terms of allowing market determination of the composition of imports. About 36 percent of retention scheme imports and 28 percent of SIP-II imports were consumer goods, compared with only 15 percent for SIP- I. Luxuries (such as sedan cars and other consumer durables) as a share of consumer imports differ among schemes: none in the case of retention, but about one-ninth of SIP-II consumer imports (see table 4-6). This is not surprising. Explicit allocation targets sought to restrict the share of consumer goods under SIP-I, encouraging inputs for production and restricting consumer "luxuries" the most. Retention scheme imports are also subject to explicit restrictions on luxury imports. There were no imports of luxuries under either SIP-I or retention schemes. Despite no-forex imports and increasing liberalization of official imports, the share of consumer goods in total nonproject and nonpetroleum imports remains below 30 percent. This compares favorably with the share that prevailed in Uganda in 1980, when all nonproject and nonpetroleum imports were subject to stringent official allocation and licensing (see table 4-7). Recent liberalization of imports has not crowded out production inputs or distorted the composition of imports. If anything, raw materials, intermediate inputs, and capital goods--by volume and as a share of nonproject and nonpetroleum imports--were greater in 1989 than in Table 4-6. End-Use Import Composition in Uganda After Liberalization, 1989 (percentage) Special Export Special All three Imports No-forex retention Import market Program I scheme scheme Program II schemes Dec. 1988 - All Jan-Aug Jan-Aug June-Aug combined Aug. 1989 schemes (1) (2) (3) (4) (5) (8) Consumer goods (X) 41.7 35.7 27.9 36.2 14.8 29.4 Automobiles/Sedan cars 4.7 0.0 0.5 2.8 0.0 1.8 Electronics 4.1 0.0 3.0 2.9 0.0 2.9 Food 14.7 22.8 10.2 12.8 3.8 9.3 Other goods 18.2 12.8 14.1 17.9 11.2 15.4 Intermediate goods 24.8 67.2 27.2 33.5 28.5 31.8 L^ Capital goods 33.7 7.1 45.0 30.3 568.7 40.2 Transport equipment 21.7 1.1 24.5 16.0 25.2 19.0 Total imports value (US 3 mill.) 72 13 22 107 65 172 Note: Includes nonproject and nonpetroleum imports only. Source: Computed from Bank of Uganda and Ministry of Commerce data. 4-26 1980.21 And if project and petroleum imports are added, consumer goods as a share of total imports did not exceed 20 percent in 1989.22 Table 4-7. Import Composition in Uganda, 1980 and 1989 (end-use) Year Consumer goods Intermediate goods Capital goods 1980 46.3 11.3 42.4 1989 29.4 30.6 40.0 Source: For 1980 end-use classification, K. Krumm, Exchange Auctions: Review of Experience, (CPD No. 1985-22), table 7, which is based on data from earlier World Bank reports on Uganda. 4.4 The Tariff Regime Tariffs raise prices so they generally create a wedge between exports and imports and among different imports. A uniform tariff rate usually creates a price wedge only between exports and imports. Generally, if imports are subject to binding quantitative restrictions, tariffs have no allocative function as they do not influence prices. Under those conditions, tariff increases serve mainly to convert importers' scarcity rent to government revenue. 21 The picture does not change if data for 1979 or 1981 are used. Unfortunately, a breakdown of data on consumer goods is not available for those years. 22 This compares favorably with other low-income Sub-Saharan countries that strictly control imports. 4-27 Despite quantitative restrictions on official imports, liberalization of no-forex imports has given tariffs more power to influence prices and protection in Uganda. This occurs because the marginal importers' cost of importing already incorporates scarcity rents because foreign exchange is purchased at the parallel market exchange rate. Tariffs have a price-raising effect, but because of the valuation method for tariffs, the price-raising effect is lower than the statutory tariff rate. To establish the tariff base, all imports are converted into local currency at the official rate of exchange regardless of the exchange rate actually used for the import transaction. This means that importers under different schemes pay different ad valorem tariffs for the same item. On an import item with a statutory ad valorem tariff of 40 percent, for example, a no-forex importer effectively pays only about 13 percent ad valorem on his actual duty-exclusive import price, compared with 40 percent under OGL and SIP-I, and 20 percent under SIP-II.23 This lowers the differential in the landed cost of imports arising from the multiple exchange rate regime. The tariff regime is not transparent. Because the parallel exchange rate sets the domestic import-parity price, lower ad valorem tariffs for no-forex transactions imply lower nominal protection. The tariff schedule will play a more significant role as the official exchange rate moves toward an equilibrium or parallel rate. 23 This estimate assumes the following exchange rates for import transactions under different schemes: for OGL and SIP-I, an official rate of U.Sh. 200/dollar; for SIP-II, U.Sh. 400/dollar; and for no- forex, U.Sh. 600/dollar. 4-28 The Tariff Structure Information on Uganda's statutory tariff comes from the printed tariff schedule with handwritten amendments and comments about administration and exemptions. It was difficult to ascertain how much of the handbook is relevant now. Many mid-level customs officials and importers work at a disadvantage because of the short supply of copies of the handbook, none of which are updated.24 Uganda's tariff handbook is a relatively simple document. Virtually all duties are ad valorem. For most imports, the range is small --between 10 and 40 percent. The duty on most goods is 10, 20, 30, or 40 percent; the modal rate is 40 percent. Only on a handful of goods is the duty more than 100 percent. The highest tariff rate (see table 4-8) is 350 percent (for beer). The tariffs on motor vehicles and liquor are about 200 percent. Few items are duty free on the official tariff schedule. The minimum statutory duty of zero applies to infant food, malt, diesel fuel and kerosene, salt, insecticides, fertilizer, medical equipment, and the like. Ad hoc exemptions raise the number of duty-free goods. Raw materials, intermediate inputs, and machinery came in duty-free as exemptions until the recent 1989/90 budget, which raised the minimum duty on raw materials and imported inputs to 10 percent.25 Machinery and equipment are also exempt. 24 Customs recently prepared a version of the Harmonized Tariff System for Uganda but at the time of the mission it had been neither finally approved nor printed. 25 But ad hoc exemptions to that budget directive had already been made in September 1989. 4-29 Table 4-8. Statutory Ad Valorem Tariff Rates in Uganda (percentages) Section Modal Modal High Outlier No. Description rate range outliers items I Live animals and animal products 30 10-30 None II Vegetable products 40 10-50 None III Fats and oils 10 5-40 None IV Prepared foodstuffs 30 10-50 250 Cigarettes V Mineral products 10 10-30 None VI Chemical products 10 10-20 100 Cosmetics VII Artificial resins and plastic 10 10-20 50 IX Wood and wood articles 20 10-30 50 Bond paper XI Textiles and textile articles 30 30-40 None XIII Articles of stone, cement, plaster 10 10-20 None XIV Pearls and precious stones 30 10-50 None XV Base metals and articles thereof 10 10-20 None XVI Machinery and mechanical appliances 0 0-20 None XVII Vehicles, vessels 20 0-30 100-200 Motor vehicles (over 2000 cc) XVIII Optical/photographic materials 10 0-30 None Note: Details are provided in Annex I. Source: Uganda's tariff schedule and 1987/88 budget. The tariff structure is escalated, with duties on inputs at 10 percent and duties on final goods in the 40-50 percent range. The average level of tariffs is low, but effective protection varies greatly--and to the extent that variations in effective protection arise from multiple exchange rates, the tariff structure is far from transparent. Tariff Exemptions This lack of transparency is exacerbated by the large number of exemptions listed in Schedule 3 of the Customs and Excise Tariff Handbook. Some exemptions, imposed by donor conditions, cover imports for government projects and project aid; others apply to visitors, diplomatic personnel, 4-30 and the like. Conditional exemptions--which depend on the person or authority doing the importing rather than on the good itself--probably have only minor harmful effects.26 Moreover, the distinctions are clearcut and require little administrative discretion--except for motor vehicles. There are also exemptions on imported goods used in government contracts. Officials say that this eliminates the need for two offsetting book transactions--that without the exemption both contract costs and government revenues would be that much higher. These exemptions may reduce administrative duplication, but they also affect the allocation of resources. For example, local producers who might meet some of the supplying contractor's needs do not receive tariff protection on their output and yet must pay duties on their inputs. The magnitude of this problem could not be ascertained, but considering the government's predominance as an investor, it may be useful to ensure that local producers are not handicapped by a system geared to reducing the administrative burden. Much more important in terms of revenues and resource allocation are the general exemptions on imports of raw materials and some capital goods. General exemptions apply to the following goods: aircraft and 26 Bodies such as the East African Development Bank and the Public Corporations and Boards (dealing with electricity, airlines, posts, water, and railways) can all import their requirements free of duty. As in most countries, diplomatic and first-arrivals privileges apply. Other exemptions apply to educational institutions, the Red Cross, religious bodies, disabled persons, and the like, as in many other countries. 4-31 aircraft parts27 and fuels, containers and pallets, film, fish, printed matter, samples, protective clothing, sewing machines for industrial use, timber for mining companies, packing materials, and bottles and jars. Many of these exemptions help the industries using these goods but discourage production of those goods in Uganda. Under the 1987 budget provisions, exemptions also apply to new capital equipment. Importers get an exemption certificate from the ministry of finance which specifies the importer's name, the goods to be imported, and the period of exemption. The July 1989 budget withdraw exemptions on raw materials and intermediate inputs, which were made subject to a 10-percent import duty. There is some evidence that ad hoc exemptions are being granted again at the discretion of the tax administration. This undermines the credibility of policy announcements and should be avoided. 4.5 The Export Regime Uganda is virtually a single-export economy now, with coffee the major export. Noncoffee exports represented no more than 3 to 5 percent of exports between 1984/85 and 1987/88. Yet in 1970/71 Uganda had a more diversified export structure than many other Sub-Saharan countries. The basic elements of Uganda's export regime--export taxes, foreign exchange retention, and export licensing--are discussed below. 27 Most countries have exemptions or zero duties for aircraft and aircraft parts. Uganda is unlikely to have its own aircraft industry, so the implications of this exemption are unimportant. 4-32 Export Taxation As of March 1989, only coffee remains subject to an export tax--in the form of a residual tax. The Agricultural Policy Committee, based on advice from the Agricultural Secretariat, sets producer prices and the marketing margins for processors and the Coffee Marketing Board, which has an export monopoly. Coffee is procured from farmers at the producer price by primary societies, which receive a commission, including transport costs and interest on working capital, from cooperative unions. Processing margins for cooperative unions and private processors are determined on a cost plus basis. The margin for the coffee board is determined on a pure cost basis, including overhead. The board thus has no incentive to reduce costs or maximize sales. The coffee board surrenders its foreign exchange export earnings to the Bank of Uganda, which converts the dollar receipts from coffee into Ugandan shillings at the prevailing official exchange rate and pays the board the producer price plus a marketing margin for each unit of coffee exported. The Bank of Uganda transfers the balance to the government exchequer as export tax revenue. Producer prices and marketing margins are announced in advance, so any fluctuation in export price is absorbed by the government through the explicit or residual export tax. Similarly, changes in exchange rate or in world prices do not necessarily affect producer prices. This "residual" mechanism of taxation is designed to shield producers from unstable world prices but in the process it blocks out world market information to producers. 4-33 Foreign Exchange Retention The foreign exchange retention scheme is currently the most significant offset to the implicit tax on exporters that would result if they had to surrender their earnings at the official exchange rate. Initiated in early 1988, the scheme originally applied to exporters of nontraditional items. Since March 1989, it was extended to all exports except coffee.28 Under the current scheme, noncoffee exporters are allowed to retain all of their foreign currency export receipts for purposes of importing inputs and other essential items (the list of currently eligible import items is the same as that under the SIP). This means that exporters can import items and sell them in the domestic market at prices that reflect the parallel market exchange rate. If successful in doing so, exporters have effectively converted their foreign currency export earnings at the parallel market exchange rate.29 Until recently, to ensure that retained foreign currency was used for imports, exporters were required to apply for a "dual license." An 28 The mission learned from several sources that the retention scheme for tea is not functioning properly. Available time did not permit further investigation. All tea exports, including those by private firms, appear to be channelled through Uganda Tea authority. Notwithstanding the retention and the announced privatization of export trade, there has been no private exports independent of the Tea Authority. 29 In some quarters in Uganda, the retention scheme is not seen in the above manner. Instead it is viewed as part of government's allocation system. 4-34 import license equivalent to the full value of exports was issued at the same time as the export license. This unnecessary restriction compelled exportc-s to arrange for imports simultaneously with exports. Since March 1989, eligible exporters have had the option of either using the proceeds through a dual license or holding their foreign currencies in a special non-interest-bearing account to use for their own imports at a later date. Foreign currency proceeds cannot be sold or transferred to other importers. However, given the credit constraint, this new option is not that helpful. Exporters have not embraced the option, notwithstanding its greater flexibility, because delaying imports means that receipts from exports are also delayed. This raises interest cost, or if there is a working capital constraint (as there is), it prevents continued exports. Thus, this increased flexibility is not very helpful given the current credit situation. The system thus remains inefficient on two counts. To maximize profits, exporters must succeed not only in exporting at a good world price but also in importing the items most profitable for the domestic market. Failure to import successfully undermines their total and marginal returns on exports. This discourages an efficient division of labor between exporters and importers. Even if exporters are successful in maximizing their returns per unit of exports, these receipts are available to them only after a long lag (say nine to ten months) under both mechanisms. So long as the working capital constraint is binding, this will undermine exports. 4-35 Export Licensing Even if exports are provided with a favorable exchange rate, growth in exports will be slow, as long as administrative impediments to export remain. All exporters are currently required to obtain an export license from the ministry of commerce before they can export. Obtaining an export license is an exceedingly bureaucratic process, with processing and documentation requirements that are staggering. At least sixteen documents are required from eight different institutions, apart from export and import invoices. For dual licensing, total fees exceed 4 percent of export value. Assuming all these documents are in place, the ministry of commerce processes an export license form for approval. Processing this form entails comparing the price quoted on the pro forma invoice with a reserve price, authenticating the form (requiring two authorized signatories of the ministry), stamping the form, and endorsing the form with a machine to indicate the value of the export license. Next, the exporter hand carries an approved export license to a commercial bank for onward forwarding to the Bank of Uganda together with all the documents. At the Bank of Uganda, officials scrutinize the export license and accompanying documents to ensure that the values on the export license and on other pertinent documents tally; the exporter has no outstanding licenses; the export price quoted is above the reserve price; the export is backed by letter of credit or, if it is for fulfilling a preshipment, evidence that such funds were received by the Bank of Uganda (Money Market Minutes); and that import fees have been paid in case of dual licensing. If all these documents are available and in order, the Bank of Uganda processes Form E for approval. This involves authenticating the 4-36 form (two authorized signatories), endorsing a C.D.3 form, stamping the forms, recording particulars of the form in a register, and separating and dispatching approval documents to the exporter's commercial bank or handing them over to the exporter. The point of all this processing is unclear. After all, a dual license or a "back-to-back" letter of credit ensures that export receipts are used. If the purpose of an export license is to monitor and record export transactions, a one-step licensing process that can be undertaken at any border outpost or local authority office should be sufficient. There is no technical reason why this function cannot be carried out in one institution/department. All stages in the processing of export forms ought to be handled in the Bank of Uganda exchange control department, as was the case in the past. In addition to substantially reducing the time required, this measure could also minimize the problem of loss of documents. No special credit facilities exist for exporters, nor do any other promotional schemes. Neither is there a duty drawback or temporary admissions scheme, probably because all inputs, whether for production of exports or for import-competing output, were exempted from duty until the budget of July 1989. Export Marketing by State Enterprises The parastatal monopoly on export marketing expanded in the 1970s. In the last two years, private participation has expanded somewhat, but much remains to be done. A parastatal export monopoly is maintained in the case of coffee, although some cooperative unions are expected to enter into export marketing. Similar monopolies existed for cotton and tea, but 4-37 private participation in tea and cotton exports is now permitted under ministerial directives. The same is true for produce marketing. Many of these directives are not consistent with parliamentary acts that remain on statute books,30 and current evidence suggests that private exports are still restrained by these institutions. Any private exports that are undertaken appear to be conducted under the "umbrella" of the marketing boards. For example, exporters of fresh produce have had to obtain approval of the Produce Marketing Board before they could get an export license. We could not ascertain whether this suggests poor private response to the export retention scheme and to the relaxation of monopoly controls, or whether it shows that other de facto restrictions continue to inhibit exports. That laws still on the statute books run counter to many of the ministerial directives creates confusion and uncertainty among both government officials and private business. What is more important, it permits those opposed to the changes to easily frustrate their implementation. This appears to be particularly relevant to the efforts at eliminating export monopoly. 30 For example, the Exchange Control Act (Cap. 103 at 24/70 and 10/74), Produce Marketing Act (3/7/70), the Link Marketing (Cap. 234), The Uganda Tea Authority decree (8/74), and the Hides and Skins Act (Cap. 183) suggest a situation contrary to announced changes. CHAPTER 5 THE INCENTIVES STRUCTURE GENERATED BY THE TRADE AND PAYMENTS REGIME One can evaluate the incentive patterns emanating from Uganda's trade and payments regime only against certain norms. Generally, incentives should be transparent and there should be an appropriate unified real exchange rate--that is, a transparent trade and payments regime based more on prices than on controls, in which the amount of foreign currency units earned or paid per unit of domestic currency is equalized for all external transactions. Evaluating departures from such equalization requires taking into account not only the appropriateness of the real exchange rate but also the operation of taxes, subsidies, and quantitative restrictions on tradable goods.1 To the extent that this prescription is based on static analysis that fails to capture the evolution of comparative advantage over time, it should be modified. The subtleties of such modification is not critical for Uganda, given its current trade and payments system. The questions that concern us here are: how far does Uganda's incentives structure deviate from the norm? Does the structure of incentives discourage production of tradables over nontradables and, within tradables, exports over import-substitutes? One way to assess this is to derive a single summary measure of incentives by calculating the effective rate of assistance or protection 1 Under such a regime, the incentive to earn a dollar of foreign exchange through exports is equated at the margin with the incentive to save a dollar through import substitution. 5-2 for each economic activity. We have not tried this because of data problems, but have provided other indicators of the level and dispersion in policy-induced incentives. These indicators presented below shows a clear antiexport bias in the trade and exchange rate system for both coffee and noncoffee exports. 5.1 The Exchange Rate and the Incentives Structure Movements in the real exchange rate or in the price of nontradables relative to tradables underlie changes in the overall incentive to produce tradables (both exports and import substitutes). Other things being equal, appreciation of the real exchange rate (a fall in the relative price of tradables) reduces the incentive to produce both exports and import substitutes. Devaluations, to the extent that they depreciate the real exchange rate, improve the incentives for all tradables. In Uganda, where the official exchange rate applies to 97 percent of current exports and the parallel market rate determines domestic prices of imports, devaluation of the official rate, if it reduces the exchange-rate premium in the parallel market, also alters the incentives among tradables. When different international transactions are subject to different exchange rates, as happens in Uganda, one must ask what the "relevant" real exchange rate is for producers.2 2 This is an important question, because in the short run the parallel and official real rates can and have moved not only to different degrees but often in opposite directions. 5-3 The real exchange rate computed at parallel market rates is more appropriate as an indicator of the relative price of nontradables to tradables but its quasi-legal status probably inhibits its use for investment decisions. Moreover, the instability of both real exchange rates has encouraged speculative trading rather than production. This is because when the parallel rate is the marginal rate of exchange, so the use of an official exchange rate implicitly taxes or subsidizes traded goods that pass through official channels. Exporters who surrender foreign currency receipts at the less depreciated official rate are effectively taxed and importers importing at that lower rate are subsidized. The rate of tax or subsidy is close to the exchange rate premium in the parallel foreign exchange market--that is, the difference between the parallel and official rates expressed as a fraction of the official exchange rate. The rate of implicit tax or subsidy (the exchange rate premium) varies greatly from year to year (see figure 5-1). The rate was highest in the late 1970s (900 percent), but fell in the early 1980s in response to significant fiscal restraint and exchange rate reform. It climbed back toward the highs of the 1970s in early 1987, but has fallen since.3 Premium levels have not exceeded 225 percent since July 1988, but they have fluctuated considerably from month to month (see figure 5-2). The present nonunified exchange rate regime (excluding explicit taxes and tariffs) is thus a source of significant dispersion and instability in incentives for production across and within industries. 3 Even for a given premium rate, the magnitude of the rate of this tax and subsidy will not be the same if the relevant official exchange rate differs among exports or among imports (SIP-II imports differed from OGL and SIP-I imports between June and October 1989). 5-4 Figure 5-1. Annual Exchange Rate Premium, 1970-88 (percentages) 1.00 0.90 - 0.80 0.70 - 0.60 - 0.50 - 0.40 - 0.30 - 0.20 - 0.10 - 0.00II 1970 1973 1976 1979 1982 1985 1988 Note: To read percentages off the Y axis, multiply each unit by 1,000 (e.g., 0.9 = 900). Figure 5-2. Monthly Exchange Rate Premium, 1986-88 (percentages) 1- 0.9 - 0.8 - 0.7 - 0.6 0.5 0.4 0.3 0.2 0.1 J86 A86 M87 087 M88 D88 J89 Note: To read percentages off the Y axis, multiply each unit by 1,000 (e.g., 0.9 = 900). 5-5 5.2 The Agricultural Sector Agriculture is the largest sector and has considerable potential for growth because of Uganda's significant comparative advantage in it. It accounts for 75 percent of GDP, provides 97 percent of the country's export earnings, and is an important source of revenue. Import policies have had little effect on output in this sector, unlike that in the manufacturing sector. Foreign exchange allocation restrictions raise the domestic market price of agricultural inputs to levels that reflect parallel market exchange rates. Farmers do not import inputs directly, so they do not get the subsidy that industrial firms using imported inputs get, even when agricultural inputs are imported at the official exchange rate under SIP. Traders and middlemen collect the scarcity rents, and growers pay input prices that reflect the world price at parallel exchange rate. So effective assistance to noncoffee agricultural exports, both traditional and nontraditional, is negligible. It is certainly lower than that received by manufacturers under OGL and SIP. Parastatal monopolies, overvalued exchange rates, and explicit export taxes distort the profitability of coffee relative to other agricultural crops. Coffee, the dominant agricultural sector, has been taxed heavily. However, because of the International Coffee Agreement (discussed in chapter 7) and limited information on nonquota sales, it is difficult to isolate and estimate precisely the efficiency losses associated with such distortionary policies. Nevertheless, now that no Agreement is in effect and current producer prices are substantially below the average cost of production, past levels of taxation cannot be justified 5-6 on efficiency grounds. The following analysis of the effects of government policies on incentives for agriculture examines policies affecting the coffee sector, other traditional export crops (cotton and tea), and nontraditional export crops. The Coffee Sector About 225,000 hectares are planted to coffee. Both robusta and arabica coffees are produced, with robusta comprising 90 percent of the total. All coffee trees are grown by smallholders as a cash crop within mixed farming systems. Coffee has survived better than other crops despite internal strife, high export taxes, and obvious inefficiencies in pricing and distribution through chains of cooperatives and a centralized parastatal with an export monopoly.4 Although coffee withstood taxation, internal conflicts, and insecurity better than plantation crops, there have been no new plantings since the 1960s, and yields are now declining. The Agricultural Secretariat estimates that a 10-percent increase over current output is about all that can be expected from improved crop husbandry, without new plantings. Coffee exports are on average 97 percent of total recorded exports. As a statutory board under the ministry of cooperatives and marketing, the Coffee Marketing Board has a monopoly on coffee exports. With the exception of some arabica coffee, all clean coffee is sold to the 4 Coffee production has fallen, but not by as much as other crops. In 1969, a peak year, 250,000 tons were produced; by 1986, production had declined to 148,000 tons. Coffee deliveries in 1987 were only 64 percent of their peak in the 1970s. 5-7 board and delivered to a central processing/storage unit. Shipments are graded according to Uganda Export Standards and transported by rail to Mombasa or Dar-es-Salam. The bulk of the coffee is sold to international traders who in turn sell it to coffee roasters. The flow of coffee from the farm to the point of export proceeds through several stages, some of which are open to private traders. Under one pathway, primary cooperative societies purchase the coffee beans from farmers, cooperative unions do early-stage processing, and then coffee is sold to the coffee board for screening and bagging. Under an alternative pathway, private buyers can purchase directly from growers for processing --with the same official margins as the cooperatives--for subsequent sale to the coffee board. This alternative route is critical, since it provides farmers with choices and the cooperatives with competition.5 Producer prices are affected by several factors. First, producer prices are announced for the season. These prices are administratively determined on the basis of multiple, often conflicting, objectives with no necessary relationship to world prices. They imply heavy taxation of coffee. Second, until recently (January 1990) there were no price differentials for differences in quality (such as the degree of moisture or impurities in the delivered coffee or the type of coffee).6 5 Late payment to producers in early 1988 led to a large and growing shift in sales to private sector processors over most of 1988. 6 Producer prices were unrelated to quality of deliveries, so the coffee board had no quality control and processors had no incentive to improve quality. Stock built up at the coffee board because of poor quality of the coffee and poor management, aggravating the difficulties arising from low world prices for coffee. 5-8 The interaction of exchange-rate policies with domestic pricing and marketing policies radically alters the incentives facing Ugandan coffee producers. Coffee producers are discouraged by two types of price wedges (see figure 5-3). The first is the explicit export tax. This represents the difference between the price received by the producer and the price obtained by converting the world price to local currency at the official exchange rate after deducting crop marketing margins. This tax is usually deducted when the Bank of Uganda clears the export bills. The second, quantitatively more significant, wedge is the implicit exchange-rate-based tax (discussed in section 4.1). This tax represents the difference between the world price converted to local currency at the official exchange rate and the world price converted to local currency at Figure 5-3. Price Wedges for Robusta Coffee, June 1989 (shillings per kilogram of clean coffee) Producer price received (U.Sh.111) (15.4%) Official export tax (U.Sh. 16) (2.2%) Marketing margins (U.Sh.113) (15.7%) - 1 - Implicit export tax (U.Sh.480) (66.7%, World price at market exchange rate (U.Sh.720) (100%) 5-9 the appropriately valued exchange rate.7 For our purposes, the latter is taken to be the parallel or market exchange rate. The total export tax is measured as the percentage by which the producer price is reduced by the combined operation of the overvalued official exchange rate and the explicit export tax. Between June 1984 and June 1989, that tax fluctuated between 75 and 92 percent of the world price at market exchange rates (see table 5-1). The extent to which these two price wedges might discourage coffee production is dramatically illustrated in figure 5-3. As of June 1989, world prices for clean coffee had dropped by more than 42 percent over 1988. Producers received only 111 new Uganda shillings (U.Sh.) per kilogram (kg) of clean coffee, although the value of the coffee at the market or parallel exchange rate was U.Sh. 720. Marketing margins amounted to only U.Sh. 113; the rest was implicit and explicit export taxes. Even when explicit taxes fell to about 3 percent in June 1989, implicit taxes were about 79 percent. Recent devaluations have lowered this implicit tax but raised the explicit tax with little change in total tax.8 For robusta, the explicit official export tax has fluctuated much more than the total tax. The explicit tax rate ranged from 37 percent in 7 The proceeds from this implicit exchange-rate-based tax on coffee producers are used to finance government imports and imports of petroleum. Thus they provide "hidden" revenue to the government and subsidize consumers of petroleum products. 8 With the official exchange rate at U.Sh. 370/dollar and the parallel market rate at U.Sh. 600/dollar, the exchange rate premium is about 60 percent. 5-10 Table 5-1. Policy-Induced Distortions in Coffee Prices, 1984-89 Item June 84 June 85 May 86 May 87 July 88 June 89 Coffee (robusta) World price (US$/kg.) 2.63 2.85 3.30 2.00 2.10 1.20 Official exchange rate (US$/sh.) 3.00 6.50 14.00 60.00 150.00 200.00 Market exchange rate (US$/sh.) 4.76 16.20 62.75 102.00 480.00 600.00 Export price at market rate (sh.) 12.52 46.17 207.10 204.00 1008.00 720.00 Export price at official rate (sh.) 7.89 18.52 46.20 120.00 315.00 240.00 Implicit export tax 4.63 27.6 160.88 84.00 693.00 480.00 Marketing margins 1.31 3.51 7.56 24.02 74.00 112.96 Official export tax 41.70 9.28 22.90 51.54 129.90 15.94 Producer price received 2.40 5.74 15.76 44.44 111.10 111.10 Producer price if no taxes (sh.) 11.21 42.66 199.51 179.98 934.00 607.04 % reduction in producer price 78.51 86.54 92.11 75.31 88.10 81.70 % implicit export tax 41.31 64.80 80.63 46.67 74.20 79.07 % explicit export tax 37.20 21.74 11.48 26.64 13.91 2.63 X total export tax 78.51 86.54 92.11 75.31 88.10 81.70 Coffee (arabica) World price (US$/kg) 2.97 2.95 4.00 2.30 2.80 1.80 Official exchange rate (US$/sh.) 3.00 6.50 14.00 60.00 150.00 200.00 Market exchange rate (US$/sh.) 4.76 16.20 62.75 102.00 480.00 600.00 Export price at market rate (sh.) 14.14 47.79 251.00 234.60 1344.00 1080.00 Export price at official rate (sh.) 8.91 19.17 56.00 138.00 420.00 360.00 Implicit export tax 5.22 28.4 195.00 96.69 924.00 720.00 Marketing margins 1.02 2.79 6.71 27.30 77.10 105.50 Official export tax 5.26 10.64 29.92 60.70 217.90 129.50 Producer price received 2.64 5.75 19.38 50.00 125.00 125.00 Producer price if no taxes (sh.) 13.12 45.00 244.29 207.30 1266.90 974.50 % reduction in producer price 79.91 87.22 92.07 75.88 90.13 87.17 X implicit export tax 39.85 63.59 79.82 46.60 72.93 73.88 % explicit export tax 40.06 23.63 12.24 29.28 17.20 13.29 X total export tax 79.91 87.22 92.07 75.88 90.13 87.17 Note: Producer prices are for clean kiboku coffee. Source: Agricultural Secretariat, Bank of Uganda. 5-11 1984 to less than 3 percent in 1989 (see table 5-1). As the world price of coffee fell, so did the explicit export tax, and vice versa, since prices paid to producers were not responsive to changes in world prices. The implicit exchange-rate-based tax usually rose when the explicit tax fell, so the total price distortion producers faced remained more or less constant. To ascertain whether the export tax is pitched to drive producer returns to equal marginal returns, one must know the prices of nonquota coffee sales. Nonquota sales are frequently barter transactions, which conceal prices. This may be useful in the presence of an International Coffee Agreement (ICA), but it makes it more difficult for policymaking to know the marginal returns from extra coffee sales. Assuming that two- thirds of coffee exports are sold on the quota market for $2/kg and one- third on the residual market for $1/kg, the quota market premium is about 40 percent of total sales. So an export tax that takes the gains (rents) from the preferred quota market and leaves growers with the returns from the second or residual market is about 40 percent. Only if prices on the residual market fell to 25 cents/kg and the shares remained the same would the quota market premium imply an export tax of 70 percent, a level Uganda has experienced recently. Without information on nonquota sales, the optimal level of export tax is difficult to ascertain, even in the presence of an ICA. Had coffee proceeds been converted at the parallel or market exchange and had the government collected the explicit export tax of 40 percent, then the producer price of coffee would have been between 35 and 52 percentage points above the level actually received. Average production costs exceed current producer prices. Thus any improvement in the 5-12 differential between producer prices and production costs would have a significant positive impact on incentives to expand coffee production.9 How much production would have expanded depends on several factors, including how much more coffee is taxed than other crops, how easily crops are substituted for each other (the cross-elasticity of supply), and what is happening in the world coffee market, particularly whether the ICA is in place (see section 7.1). This raises two questions relating to efficiency losses. First, could Uganda have increased nonquota export sales significantly in the presence of the ICA? On the face of it, the marketing ability of the coffee board and the absence of private participation suggests limited capacity to do so. Yet nonquota market sales have risen substantially, and recently the coffee board exported a total of 3.1 million bags, the highest ever. The critical issue, however, is whether existing institutions should be viewed as givens when assessing pricing policy. Reducing taxes could raise both the volume and the quality of coffee, especially if such reductions were linked to world market differentials for coffee quality. Second, and more importantly, in the absence of an ICA, are the efficiency losses likely to be greater at the same level of export taxes? The answer is generally in the affirmative. It remains a moot point, however, whether the induced increases in coffee production could be exported without institutional changes, such as privatization of coffee exports and rationalization of the coffee board. Notwithstanding such 9 However, the welfare benefits of reducing coffee taxes depend heavily on outcomes in the distorted world coffee market, in particular the future course of the ICA and world coffee quotas. This is examined in chapter 7. 5-13 caveats it can be argued that the need for such institutional changes can be compelling when the output effects of better prices emerge. The cost of inaction on institutions becomes more pronounced. One thing is clear, however. With current taxation of coffee and thus no new plantings, Uganda faces declining coffee production from aging trees that are genetically inferior to the newer trees of competing producers. Clearly, producer prices are currently dictated by compulsions of fiscal and monetary policy. The efficiency cost of this revenue and monetary constraint is likely to be very high in the medium term. Coffee export tax revenue averaged around 20 percent of total revenue in 1988/89 and 1989/90, relatively lower than the 40 percent in 1977 and 1978. Recent devaluation of the official exchange rate reduced the implicit tax revenue but raised the explicit tax since producer prices were kept unchanged. Currently, 35 percent of the f.o.b. value of coffee exports accrue residually as explicit tax revenue, while processing and marketing intermediaries receive 37 percent. Reductions in the latter would give better returns to farmers without lowering tax revenue, and the urgency of doing so should be highlighted. Inefficiency in crop financing of coffee imposes another constraint on reducing taxes. But again crop financing should be made efficient instead of producer prices being depressed. Crop finance refers to funds required by market intermediaries for procurement and does not involve production credit to farmers. Intermediaries require working capital to meet the costs of transporting, processing, or storing coffee. Coffee accounts for over 80 percent of overall crop finance requirements in Uganda, and of this, the coffee board accounts for the dominant share (about 75 percent). Requirements are substantial in absolute terms because 5-14 of the large size of the coffee crop and because the time between a farmer's delivery and receipt of export proceeds is long (six months). Inefficient use of credit caused by inefficient marketing institutions add further to the requirements. Outstanding credit for coffee crop financing during the 1989/90 coffee year averaged around U.Sh. 18 billion (equivalent to US$90 million at the official exchange rate). Crop financing needs were traditionally met by commercial banks, with the coffee board lending to cooperative unions and primary societies. Concerns over delays in payments to farmers, poor liquidity of commercial banks, lending to noncreditworthy cooperative unions, and the commercial banks' apprehension about the deteriorating financial management of the coffee board led the government in December 1988 to direct the Bank of Uganda to intervene as direct lender for crop finance. The present institutional arrangements are predicated on a key policy objective, namely that farmers shall receive cash payment on delivery for all coffee they wish to sell at prices fixed by the government. The Bank of Uganda lends directly to the coffee board and through Uganda Commercial Bank to about 2000 cooperative primary societies. Commercial banks provide crop finance credit to the fifteen cooperative unions and to the private processors. Loans to noncreditworthy unions (about ten out of the fifteen unions) are guaranteed by the Bank of Uganda. Commercial banks charge cooperative unions and private processors commercial interest rates (45-50 percent), within the ceiling on lending rates determined by the Bank of Uganda. The coffee board is charged a lower rate (30 percent) on its borrowing from the Bank of Uganda. The measures instituted in December 1988 resulted in a substantial expansion of crop finance credit, not matched by coffee stocks held in the 5-15 marketing system. This expansion, which represented a dramatic increase in overall credit and money supply, was a major factor behind the rising inflation in the first half of 1989. The new system eliminated farmers' contribution to the financing of crop procurement by substituting cash payment on delivery for irredeemable chits issued by primary buyers. Second, the Bank of Uganda provided an unlimited overdraft facility to the coffee board, encouraging its continued inefficient use of credit. The main reason for the expansion of credit to the coffee board, however, was the government's failure to ensure prompt payment to the coffee board for large quantities of coffee diverted for barter trade. Moreover, the new system caused delays in the realization of export proceeds. For other marketing intermediaries, the credit limits, based on crop finance benchmarks, were strictly enforced. Many cooperative unions and societies, however, failed either to account for the initial advance or to achieve the anticipated turnover of the advance, largely due to the diversion of funds for activities other than coffee procurement. There is thus an urgent need to address a number of fundamental structural problems that are causing inefficient use of credit and expanding credit demand. First, coffee crop financing is heavily dependent on institutional credit. The coffee board, whose demand for crop finance constitutes about 75 percent of overall crop finance requirements, as a statutory board is unable to raise its own funds for such a purpose. Also, the poor financial situation of most cooperatives precludes any significant internal financing of crop financing. Second, inefficiencies in the cooperative and parastatal sector increase the financing required to hold and process stocks. Cooperative unions operate at low capacity utilization and have excessive overheads and weak management. Third, the government 5-16 guarantees to purchase all coffee from farmers. If exports do not expand with procurement, this commitment can be maintained only at the expense of growing coffee stocks and increasing crop finance credit. Fourth, crop finance credit is extended to buyers at all three procurement levels, which puts excess liquidity in the hands of marketing intermediaries and has enabled cooperative unions to divert crop finance funds for other purposes. Finally, Bank of Uganda financing of the coffee board has resulted in inefficient lodging and negotiations of export bills and consequently has increased the coffee board's short-term credit demand. These problems clearly have to be resolved if coffee growing is not to be discouraged. However, in the short run, if producer prices are related to quality differentials and procurement can be limited to good coffee, and exporting will be easier and the lags between procurement and export will shorten. Increasing private participation in export marketing is another quick way of easing the need for crop finance. Other Traditional Exports Government marketing boards have controlled the producer prices not only of coffee but also of other traditional crops--such as cotton, tea, and tobacco. Price distortions for these crops have been similar--a small implicit tax (through an overvalued exchange rate) and a large explicit export tax (through a producer price lower than the world price converted at the official exchange rate and after deducting marketing margins). This is shown in table 5-2. The official export tax for these crops has always been smaller than the one for coffee, while the implicit tax has been similar to that for coffee (see table 5-2). Figure 5-4 compares the total (implicit plus explicit) export taxes levied on traditional crops and coffee for 1984-89. 5-17 Table 5-2. Policy-Induced Distortions in Cotton and Tea Prices, 1984-89 Item June 84 June 85 May 86 May 87 July 88 June 89 Cotton World price (US$/kg) 1.94 1.76 1.21 1.45 1.59 1.68 Official exchange rate (US$/sh.) 300.00 650.00 1400.00 60.00 150.00 200.00 Market exchange rate (US$/sh.) 476.00 1620.00 6275.00 102.00 480.00 600.00 Export price at market rate (sh.) 923.44 2851.20 7592.75 147.59 763.20 1007.40 Export price at official rate (sh.) 582.00 1144.00 1694.00 86.82 238.50 335.80 Implicit export tax 341.44 1707.20 5898.75 60.77 524.70 671.60 Marketing margins 197.61 462.15 816.92 21.70 85.83 199.48 Official export tax 15.16 4.92 -353.69 6.66 -93.48 -263.68 Producer price received 369.23 676.93 1230.77 58.46 246.15 400.00 Producer price if no taxes (sh.) 725.83 2389.05 6775.83 125.89 677.37 807.92 X reduction in producer price 79.13 71.67 81.84 53.56 63.66 50.49 Z implicit export tax 47.04 71.46 87.06 48.27 77.46 83.13 X explicit export tax 2.09 0.21 -5.22 5.29 -13.80 -32.64 Z total export tax 49.13 71.67 81.84 53.56 63.66 50.49 Tea World price (US$/kg) 1.80 2.00 1.40 1.20 1.20 1.37 Official exchange rate (US$/sh.) 300.00 650.00 1400.00 60.00 150.00 200.00 Market exchange rate (US$/sh.) 476.00 1620.00 6275.00 102.00 480.00 600.00 Export price at market rate (sh.) 856.80 3240.00 8785.00 122.40 576.00 822.00 Export price at official rate (sh.) 540.00 1300.00 1960.00 72.00 180.00 274.00 Implicit export tax 316.80 1940.00 6825.00 50.40 396.00 548.00 Marketing margins 293.50 735.60 1240.00 37.02 92.90 165.00 Official export tax 21.50 64.70 20.00 9.98 -12.90 -66.00 Producer price received 225.00 500.00 700.00 25.00 100.00 175.00 Producer price if no taxes (sh.) 563.30 2504.40 7545.00 85.38 483.10 657.00 X reduction in producer price 60.06 80.04 90.72 70.70 79.30 73.36 Z implicit export tax 56.24 77.46 90.46 59.03 81.97 83.41 X explicit export tax 3.82 2.57 0.27 11.69 -2.67 -10.05 Z total export tax 60.06 80.04 90.72 70.72 79.30 73.36 Source: Agricultural Secretariat, Bank of Uganda. 5-18 Figure 5-4. Total Export Tax Rates on Traditional Crops, 1984-89 100.0 90.0 Ct4oeearaoa 80.0 e **. *c4tion 1984 1985 1986 1987 1988 1989 Traditional crops dominate Ugandan exports and agriculture, so both agriculture and the export sector have been heavily taxed by these arrangements. As a result, export growth has probably been curtailed and fewer resources have been engaged in agriculture than would otherwise have been the case, despite other constraints. The incentives system favors industry. Since March 1989, the foreign exchange retention scheme has been extended to cotton, tea, and tobacco and explicit export duties have been removed. The price of these crops relative to coffee rose substantially (see table 5-3). For example, between June 1988 and June 1989, the producer price ratio of cotton to coffee doubled and the ratios for tea and cocoa to coffee tripled. Coffee growing relative to other crops came to be discouraged even more than before. 5-19 Table 5-3. Relative Producer Prices of Traditional Ugandan Exports Before March 1989 After March 1989 Price ratio Price ratio relative to relative to Export Price coffee Price coffee Coffee (robusta) 111.1 1.0 111.1 1.0 Cotton 400.0 3.6 807.9 7.3 Tea 175.0 1.6 657.0 5.9 Cocoa 225.0 2.0 627.0 5.6 Note: Computed by using June 1988 and June 1989 prices received by producers. Source: Agricultural Secretariat of the Bank of Uganda. Nevertheless it is not clear why producer prices have to be set for traditional crops that are subject to export retention and private sector participation in export marketing and domestic procurement. Their prices should be determined by the market. Nontraditional Crops Exporters of nontraditional crops such as maize, pineapple, beans, ginger, sim sim, sesame, and other vegetables and fruit have not been subject to producer price controls nor to the requirement of surrendering export proceeds at official exchange rates since 1988. In short, they have not been taxed for the last two years, making the relative price of nontraditional export crops to other export crops more favorable. While such taxation has been minimal, the disincentives arising from administrative impediments have remained substantial. Notwithstanding recent directives, these crops cannot be procured and exported without permission of the Produce Marketing Board and several other government agencies. 5-20 5.3 The Manufacturing Sector The manufacturing sector produces import-competing output, mainly consumer goods. Incentives in this sector are affected by the exchange- rate system, tariffs, taxes, and price controls. Most imports of consumer goods are channeled through the no-forex scheme at parallel market exchange rates, but firms have access to imported inputs mostly at the overvalued official exchange rate (paying prices which are a fraction of the border price equivalent), which amounts to a significant subsidy to manufacturers able to secure such inputs. Other costs of obtaining allocations at the official rate are not uniform among official import schemes, however, which makes such subsidies difficult to estimate (see evidence below). So significant protection exists, but the allocation system for foreign exchange creates considerable variation in such protection between and within industries. Liberalization of Imports A unified exchange rate and greater liberalization of imports can be expected to reduce overall protection and to substantially reduce, if not eliminate, the policy-induced differentials in price and thus in incentives between and within industries. In principle, under a liberal import regime foreign currency is priced in such a way that it goes to import items that earn the highest financial return.10 With consumer goods heavily restricted under the 10 If the domestic market is distorted, however, these import items may not be those that earn the highest social rates of return. But this problem can be solved with a domestic tax-subsidy system. 5-21 official allocation system, the most liberal import arrangements--no-forex and foreign exchange retention--became the dominant sources of consumer imports (see table 5-4), accounting for about 90 percent of such imports in 1987 and 1988 and 70 percent in 1989.11 Except for noncompeting luxury goods such as sedan cars and household electrical and electronic appliances, imports competed with domestic production (see table 5-5) of such consumer items as textile fabrics, matches, soap, blankets, mattresses, fishnets, shoes, batteries and dry cells, paints, and so on. Table 5-4. Consumer Imports (millions of U.S. dollars) Official Export a No- SIP-I SIP-II Others allocations retention forex Total 1987 -- -- 4.6 4.6 -- 20.0 24.6 1988 -- -- 0.5 0.5 5.0 40.4 45.9 1989 b 9.7 6.1 -- 15.8 4.6 31.0 51.4 a. First eight months only. b. Data relate to licenses issued, not actual use of licenses. Source: Bank of Uganda and Ministry of Commerce. 11 The decline in this share was due largely to liberalization of the official allocation system. The absolute dollar value of consumer imports under no-forex and retention was higher in 1989 (an annualized value of $68.5 million). 5-22 Table 5-5. Liberalization of Competing Imports, Selected Items, 1988 and 1989 US $ 000 Percentage Item 1988 1989 Change Batteries/dry cells 1,072 2,340 118 Bicycle tires/tubes 673 1,498 122 Other tires/tubes 1,824 5,061 178 Cement 3,097 7,599 145 Fishnets 5 60 1,200 Mattresses 142 56 -61 Paint 611 411 -33 Paper 100 1,524 1,424 Shoes 554 902 63 Soft drinks 35 0 -100 Steel products 2,987 4,683 57 Sugar 5 8,600 18,000 Fabrics 2,029 4,362 115 Twines and cords 15 252 1,533 Clothing/garments 3,598 4,191 17 Matches 1,361 1,959 44 Blankets 226 637 182 Soaps 1,367 67 -95 Protection to Manufacturers Import liberalization was expected to reduce the dispersion in nominal protection in the final goods industry. However, the impact of liberalization on the level and dispersion of effective protection was not as favorable as expected because firms had differential degrees of access to imported inputs at overvalued official exchange rates. The total dollar value of imports of raw materials and intermediate goods increased considerably in 1988 and 1989 (see table 5-4). That value was 28 percent higher in the first eight months of 1989 than it was for all 12 months of 1988.12 No-forex provided only about 28 percent of 12 At an annualized rate it was 90 percent higher than in 1988. But except for OGL, these imports were not all for inputs for industry. The increase in imported inputs for industry was lower. 5-23 Uganda's total imports of raw materials and intermediate inputs in 1989, compared with 43 percent in 1988, because of a 60-percent increase in imports under OGL, SIP-I, and SIP-II. Import liberalization clearly relaxed the constraints on manufacturing output. Two-thirds of all imported inputs, and an even higher share of industrial inputs, came in at less than the parallel exchange rate (see table 5-6). OGL and SIP-I provided more than half these imports. Firms that get OGL and SIP-I foreign exchange for inputs receive rents or subsidies of different magnitudes. OGL firms are preselected, so they have lower transaction costs than firms that import inputs under SIP-I. Inventory costs are also likely to be lower under OGL, for which access to foreign exchange is more certain because firms are not competing with each other. If the rent or subsidy under OGL is taken as the difference between Table 5-8. Raw Material and Intermediate Imports, Excluding Fuel, 1987-89 (by import schemes, in USS millions) Total Export No- Grand Year OGL SIP-I SIP-II Other official retention forex total 1987 -- -- -- 12.0 12.0 -- 4.0 16.0 1988 23.2 -- -- -- 23.2 4.5 21.0 48.7 1989 a 12.6 18.5 6.0 -- 37.1 7.4 17.7 62.2 b a. First eight months only. b. Agricultural inputs imported under USAID's Trade Promotion Credit do not show up because data were unavailable. 5-24 the official and parallel market exchange rates, that amount is higher than the level of subsidy for SIP-I imported inputs.13 Table 5-7 measures the sudsidy to production of OGL firms based on production data for 1988. The subsidy is calculated as the difference between the cost of imports at the market rate of exchange (which represents the scarcity value of the opportunity cost of foreign exchange to Uganda) and the official rate of exchange. In 1988, for every U.S. dollar's worth of imports under OGL, a firm's production could be subsidized by about U.Sh. 300. The cost of purchasing a dollar of foreign exchange at the market rate was about U.Sh. 409, but at the average official exchange rate, only U.Sh. 109. As a percentage of production, the rate of subsidy received by specific firms was high (see table 5-7). Estimates for two firms show that the import subsidy under OGL exceeded the value of production recorded that year. This could be due to lags between the purchase of inputs and the production of output, but is more likely attributable to the firms' resale of imported inputs, given the scarcity of imported inputs and the large parallel market premium. Conditions are favorable for resale trade in imported inputs between OGL and non-OGL firms. It is difficult to prevent "hoarding" or "leakage" when rents are high. 13 The OGL subsidy in table 5-7, computed on the basis of the difference between the official and quoted parallel rate, is overestimated to the extent that producers could buy parallel market foreign exchange at a discount on the quoted rate (see chapter 4). But the relative difference in subsidies between OGL and SIP-I users will hold. 5-25 Table 5-7: Subsidies from Open General Licensing a (1) (2) (3) (4) (5) Import Rate of Ratio of OGL use subsidy subsidy imported Value of Jan.-Dec. under on pro- inputs to production Sector Firm 1988 b OGL C duction d production * 1988 ($US 000) (mill. sh.) (percent) (percent) (mill. sh.) Soft drinks Jubilee ice and soda 43.8 13.1 59.2 80.7 22.2 Kampala bottlers 537.9 161.4 33.0 45.0 489.3 Lake Vic bottling 2106.0 631.8 25.9 35/3 244.5 Cigarettes B.A.T. Uganda 3012.1 903.6 10.6 14.5 8510.9 Beer Uganda breweries 2115.3 634.6 108.9 148.5 582.7 Nile breweries 157.9 47.4 5.3 7.2 898.5 Textiles Mulco textiles 72.4 21.7 5.4 7.4 400.8 Uganda blanket manufacturers 112.5 33.8 80.7 110.1 41.8 Nyanza textiles 213.0 63.9 3.9 5.3 1646.8 Uganda bag & hessian 224.5 67.4 49.2 67.1 136.9 Uganda fishnet manufacturers 107.0 32.1 12.9 17.5 249.5 Soap Mukwano industries 11824.4 3547.3 135.0 184.0 2628.4 Cement UCI Hima 33.0 9.9 5.1 6.9 195.0 Mattresses Vitafoam 1540.4 482.1 177.8 242.1 260.2 a. For 1988 an average official exchange rate of U.Sh. 109/dollar was assumed. The average market rate in 1989 was assumed to be U.Sh. 409/dollar. b. Data provided by the Bank of Uganda. c. Calculated as column (1) x 409 - col. (1) x 109 = col. (1) x 300. d. Calculated as col. (2)/ col. (5) x 100. e. Calculated as [col. (1) x 409] / Icol. (5) x 1000]. f. Data from Department of Statistics Survey Sheets. Source: Calculated from information supplied by the research department of the Bank of Uganda and manufacturing records. 5-26 More recently, OGL firms' ability to misuse their favored access to imported inputs has increased, because they were permitted to seek SIP funds as well. Between February and April 1990, OGL firms availed themselves of one-third of the imports of inputs allocated under SIP-III. This additional access adds to the discrimination that non-OGL firms face under the current import regime. SIP-I inputs were also imported at the official exchange rate, but the costs of importing under SIP-I were not the same as under OGL. Industries and firms did not have the same assurance of sustained, repeated licensing of inputs. They had to compete not only with other firms but with traders seeking foreign exchange for agricultural inputs, consumer goods, transportation equipment, and so on. And they had to maintain positive bank balances, which meant more queuing time, transaction costs, and uncertainty. Members of the Uganda Manufacturers Association indicated a strong desire to be selected for the OGL rather than a SIP-I facility, no doubt for these reasons. Some firms indicated that plans to modernize or rehabilitate their plants would depend on their being selected for OGL. Liberalization of imported inputs affected different manufacturing industries differently. First, importers of industrial inputs were mainly the actual users (firms), so they obtained rents from both 0GL and SIP-I foreign exchange. Rents from the two sources of inputs differed because of other costs (e.g., queuing, uncertainty, financial). Second, liberalization may have led certain firms and industries to expand to levels that are unsustainable without the subsidy. Third, most imported manufacturing inputs came in at the official exchange rate, free of duty, which discriminated against domestic input production. 5-27 There is some concern that without this implicit subsidy or assistance, it would not be sufficiently profitable to produce. This is not true. Several Ugandan firms found it profitable to consistently produce in 1988 and 1989 with inputs imported under no-forex.14 Many have combined inputs imported under no-forex with inputs imported under SIP-I and SIP-II (see table 5-8 and annex tables). Table 5-8. Firms Importing Inputs Under Various Schemes No. of Firms also using other schemes firms SIP-I SIP-II No-forex OGL OGL 26 7 3 2 - No-forex 7 5 3 - 2 SIP-I 43 - 14 5 7 SIP-I 25 14 - 3 3 Source: Computed from firm-level import data from MPED, Uganda. Without estimates of effective protection one can undertake only a qualitative assessment of relative levels of assistance in manufacturing. The level and dispersion of protection has risen since 1986 and 1987. In 1989, for example, with an average parallel exchange rate of U.Sh. 561/per dollar and an official rate of U.Sh. 221/dollar, the average manufacturer importing 70 percent of its recurrent inputs received an effective subsidy of 43 percent on total inputs. So a firm that used to enjoy protection 14 In 1988, before SIP-I and SIP-II, about eighteen firms used no-forex to import inputs and machinery. 5-28 through quantitative restrictions and now faces competition from no-forex priced imports is probably experiencing lower nominal protection on output but still enjoying substantial effective protection. Take, for example, a firm that produces $100 worth of world-priced output using $80 worth of border-priced inputs. A 20-percent nominal protection of output implies a selling price of $120 for the firm. A 43-percent effective subsidy on OGL input implies an input-cost of $45.60 instead of $80. So, the domestic market value of the firm's output exceeds its input costs by $74.40. In short, the current protected value added is 272 percent more than the value added at world prices for OGL firms. Under SIP-I, effective protection would be lower than under OGL because of other costs of obtaining SIP-I foreign exchange. Similarly, under SIP-II, which existed between June and October 1989, the effective subsidy on inputs would have been 23 percent, with effective protection of 192 percent. With no-forex inputs, effective protection for this firm would have been 20 percent. Thus, effective protection varies widely even though overall import liberalization has reduced the dispersion in nominal protection of output. (Table 5-9 shows the dispersion in protection for different shares of value added.) Viable Industries Does existing protection sustain economically unviable industries? Will its reduction lead to many firm closures? To assess this, we draw on an earlier report, the 1986 World Bank Uganda Industrial Sector Memo, for data. It surveyed and analyzed forty-four firms and products. Half of the firms had negative nominal protection--that is, the prices they received were below world market prices--mainly because of the price controls that 5-29 Table 5-9. Nominal and Effective Rates of Protection, 1985 Output Changes, 1985-88, and Tariff and Licensing Coverage, 1989 Nominal Effective Output Source of protection protection Change Tariff imported Product (1985) (1985) (1985-88) (1989) input (Z) Batteries -32 -99 -33 na? 20 SIP-I Beer -48 -55 -74 250 350 OGL Bicycle tires 57 59 na 15 SIP-I Bicycle tubes 103 a -10 15 SIP-I Biscuits 66 210 50 - Cables 92 -30 20 SIP-I Conductors 93 175 -30 20 SIP-I Cement -66 -43 22 50 OGL Fabric dyed -88 -95 30 OGL printed -81 -99 30 OGL unbleached -73 30 OGL Fishnet 26 60 -140 -20 20 OGL Hessian 91 (a) 30 OGL Mattresses -23 -100 300 30 OGL Paint gloss -8 (a) 80 SIP-I/No-forex emulsion -45 -20 80 SIP-I/No-forex Paper sacks 14 20 SIP-I bond 15 -60 duplicating 15 cover 17 Plywood 112 -7 30 SIP-I Shoes -25 30 SIP-I/No-forex leather -32 SIP-I/No-forex plastic -5 -56 SIP-I/No-forex rubber -34 -74 SIP-I/No-forex Softdrinks 48 10 -40 200 120 OGL Steel rolled -35 10 SIP-I/No-forex sheets - -112 -66 10 SIP-I/No-forex Sugar 85 500 30 SIP-I/No-forex Twines/cords 53 120 20 SIP-I Source: Columns 1 and 2 from Uganda Industrial Sector Memo, World Bank (1986); Column 3 from Background to the Budget, Department of Statistics, Table 54; the rest from budget speeches, tariff schedules, and the like. 5-30 operated in the mid-1980s. Affected products included textiles, leather and rubber shoes, batteries, paint, cement, wheat flour, and beer. Products that had positive nominal protection then included plastic shoes, gunny sacks, rolled steel, fishnet, bicycle tires, twines and cords, mattresses, sugar, biscuits, and soft drinks. Except for fishnet, rolled steel, and gunny sacks, all products (whether effective protection was positive or negative) were economically viable or had a domestic resource cost of less than one. But the study concluded that the production of hessian cloth, cables of some dimensions, paper sacks (used for cement and sugar), plywood, and cotton-seed oil was not economical in Uganda because at world prices the value of inputs exceeded the value of output, leaving nothing with which to pay labor and capital. They were produced only because of government assistance. Clearly, there have been many policy changes since the 1986 study, and its estimates of effective protection levels and dispersion are of historical value only. However, the study's conclusions on the lack of economic viability of some products holds unless there have been major shifts in relative world relative prices or in domestic production technology for those products. As of end-1989, price controls have been removed on most goods, except for salt, soap, and hoes. So current nominal protection of fourteen subsectors (see table 5-9) must have risen and is currently positive because of the tariffs. In this particular sample, only rolled steel was found to be economically unviable. Effective protection is currently higher than at the time of the study because of preferential OGL access to foreign exchange at overvalued rates for imported inputs. To the extent that OGL imports were available at subsidized rates to nine subsectors, five of which previously had price 5-31 controls, their effective protection is likely to have risen substantially since 1986. The same is true of several other enterprises that received inputs under SIP-I to produce batteries, bicycle tires and tubes, cables, conductors, plywood, and paper bags. The level of subsidy was no doubt lower than under OGL, but access to inputs under both OGL and SIP-I supported the production of some products that had been found to be unviable in the earlier study (including cables, conductors, fishnet, hessian cloth, gunny sacks, plywood, and paper bags). The subsidy implicitly derived from the direct import of inputs under OGL and SIP-I may have been responsible for rejuvenating production of at least eight or so products that might not be economically viable without that subsidy. These products are likely to be hurt by reductions in the exchange rate premium and thus in the subsidy in the near future. Considering the scarcity of foreign exchange, its allocation to firms and products with negative value added at world prices is clearly too costly for the Ugandan economy to continue indefinitely. IIm CHAPTER 6 PROPOSALS FOR REFORM The main objective of trade liberalization in Uganda is to make the incentive system for tradables--especially exports--more transparent, stable, and neutral. Policy reform should aim to do that. This will require changes not only in commercial policy but in the supportive macroeconomic policies needed to ensure realistic exchange rates. Also important will be the removal of administrative impediments to exports, of which there are many. Complementary measures such as the rationalization of parastatal organizatinos, privatization of public enterprises, private participation in export marketing, and the development of th!e financial sector, are also necessary to enhance the supply response to reform. 6.1 Export Reform Moving toward the desired incentive system through export reform will involve a more liberal foreign exchange retention scheme for noncoffee exports (pending a unified exchange rate), lower coffee export taxes, and greater automaticity in export licensing. Foreign exchange retention of 100 percent for all noncoffee exporters should be continued, but most important, exporters should be permitted to sell their foreign currency to importers at the market rate of exchange. Various mechanisms could be devised if the parallel market is not "legalized." The simplest scheme is for the Bank of Uganda and the Ministry of Commerce to issue transferable import licenses against export proceeds, allowing original exporters to sign away their title to another bona fide importer. 6-2 Taxation of Coffee With no International Coffee Agreement (ICA) in effect and no monopoly power in the world market, an export tax on coffee is unwarranted. Efficiency considerations argue against it and for depreciating the official exchange rate and ensuing that the producer price of coffee is close to the export parity price. The wisdom of an action like this for a commodity such as coffee is often questioned on two grounds: constraints on additional exports and the government need for revenue. Pessimism about Uganda's ability to export additional coffee stems from Uganda's past experience. Under the ICA, exports to quota markets were restricted, and the coffee board failed to sell enough coffee to nonquota markets to exhaust its coffee stocks. Stocks built up largely through the coffee board's mismanagement and poor quality control and the country's failure to harness the entrepreneurial energies of private traders in exporting coffee. Still, nonquota sales, especially barter sales, expanded rapidly in 1988/89. And now, with no ICA, additional exports to quota markets are no longer restricted. There is concern that increased supplies will lead to a decline in terms of trade, but this is unlikely. Uganda is a small supplier (supplying only 3.5 percent of the world coffee market and 12 percent of the robusta market), and with the recent collapse in world coffee prices, the likelihood of a large increase in world output is slight. Moreover, private firms are now being allowed to participate in the coffee export trade. The operational efficiency of marketing intermediaries has improved, and the coffee board's operations are being rationalized. Uganda's costs compared to its competitors costs are favorable. Pessimism based on past experience should not prevent Uganda 6-3 from trying to expand its world market share. More important, if coffee export taxes are not reduced, Uganda's coffee output could decline in the next few years. Can Uganda afford such a decline? Despite this need to reduce coffee export taxes, the government's need for revenue must also be considered. Explicit export taxes have contributed between 40 and 60 percent of total government revenues (more than 5 percent of GDP) most years, except during the slump in world coffee prices in 1988/89, when coffee export taxes accounted for only 13 percent of revenues (0.7 percent of GDP). Implicit exchange-rate-based taxes, by contrast, have always been substantial. As of June 1989, more than two- fifths of the world coffee price was being extracted in this form.1 Under current budget conditions, further revenue losses could undermine reform. Moreover, problems in administering taxes on income and profit from the coffee sector (because coffee growers are widely dispersed) may justify export taxes as a substitute for direct taxes. Linkage with World Prices Whatever their level, export taxes should be in a form that restores the link between changes in domestic producer prices and changes in the world price of coffee. With a marketing board, one way is to set the rate of export tax as a fixed percentage of the world export price for Ugandan coffee. Another way is to devise a formula for linking world prices to producer prices so that producer prices reflect medium-term trends rather than short-run fluctuations. 1 With a 50-percent premium on the parallel market, an implicit tax of 30 percent would continue in the short run. 6-4 These alternatives are often resisted on the grounds that the producer price of coffee should be stable and should cover the costs of production. This may require a fixed producer price that can be altered independently of world market conditions. In many ways these arguments are academic in Uganda. Current producer prices do not cover production costs, and real producer prices have been unstable. Delinking producer prices from world prices prevents the flow of information from the world market to the growers. This is often justified on the grounds that changes in the short-term price mislead growers--as if the government had better information than farmers on long-term prospects, which is often not true, as the experience of marketing boards around the world confirms. As for risk, the outcome of thousands of decentralized farmers' decisions based on their individual assessments may be safer than one all-our-eggs-in-one-basket pricing decision from centralized government. Similarly, basing producer prices on the cost of production makes less sense economically than letting prices equal opportunity costs (the cost of purchasing coffee on the world market). The marginal cost of growing coffee should equal the marginal benefit. Farmers' planting, husbandry, and cropping decisions should reflect the earning potential for Uganda implicit in world prices. These decisions are different if farm prices are based on production costs.2 Restoring the link between producer 2 Generally, when prices look good, farmers think about growing more, and when they look bad they think about growing less--and more of something else. When producer prices guarantee that production costs are covered, the thinking is different. 6-5 prices and world prices also leads to different producer prices for different qualities of coffee, and so will encourage Ugandan farmers to produce more high-quality coffee. 6.2 Import Reform Official imports should be further liberalized and high tariff rates should be reduced. The government is concerned about lowering tariffs, but high tariffs can be replaced with a high sales tax on both imports and domestic output. The official allocation of foreign exchange should continue to be liberalized. A single Open Import Program (OIP) would be the best way to achieve this. The OIP should be a genuine open general licensing scheme in the traditional sense. There should be a short negative list (all products not listed may be imported) with OIP allocations at an increasingly appropriate official exchange rate. It should be managed as a permanent feature of the import regime but with minimum direct control. The issue is how best to move from the present situation to a single OIP. The rationale for a closed allocation scheme such as the current OGL scheme is extremely weak. As the Ugandan economy recovers, preselection of certain industries and, within them, certain firms is unjustifiable. In 1987 and 1988, when donor support was not as strong as it is today, OGL provided some assurance of foreign exchange funding for industrial inputs. Though SIP-I, SIP-II, and SIP-III were much more open and general, they were managed in a way that made their continued availability somewhat uncertain. A SIP was set up with each line of credit, but was closed when another line became available. While three 6-6 different SIPs came and went, the same OGL has continued notwithstanding the ebb and flow of donor disbursements. Thus SIP is viewed by importers as a temporary "window" of foreign exchange that opens and closes without notice. There are several ways in which the current official import schemes can merge into OIP, one is to set up the OIP in such a way that part of it works like the current OGL and the other part like the SIP. For example, the OIP would be set up for all official imports, but with a clearly defined foreign exchange allocation for OGL firms, another global allocation for non-OGL firms, and agricultural inputs, and the rest for all other imports. Over time, the two allocations for inputs can be collapsed into one global allocation for all inputs for all firms, leaving the rest for all other imports. Another way is to set up OIP in parallel with OGL and after a transition period, merge the two. A third option is to focus on unifying the exchange rate without worrying about merging OGL and OIP. Tariffs are not currently the main source of trade distortions but they do affect the domestic price of imports at the margin. Thus, to a slight degree, they protect import-competing output and are likely to become a more important source of protection the less the official and parallel rates diverge. Rationalizing tariffs now would minimize disruptions in production. Tariff reform should satisfy certain criteria. For one thing, tariffs should be simple and easy to administer, given the difficult task of monitoring trade flows through Uganda's many entry and exit points. There should be few rates and no differentiation according to end-use. Tariffs should be relatively uniform--certainly more so than they are now. Under the new rate structure, ad hoc exemptions from import duty should be avoided. 6-7 The top rates should be lower and the lowest rate higher. A ceiling of 50 percent, although arbitrary, is supported by two arguments. First, it reduces the level and dispersion of protection without being too disruptive: rates higher than 50 percent are mainly for noncompeting imports. Second, eliminating high tariff categories would affect only 2.5 percent of total import revenues in 1988/89. The lowest duty rate, 10 percent for intermediate inputs (already announced in the budget for 1989/90), should be maintained. The sixteen current rates will be replaced by five: 10, 20, 30, 40, and 50 percent. This will move the tariff structure toward more simplicity and uniformity. 6.3 Domestic Tax Reform In the medium to long term, the tax system should move toward a consumption-tax-based system. A manufacturer-level value added tax should become the main revenue instrument. Until then, domestic indirect taxes should be changed to facilitate the shift to the new system. Manufacturer- level sales taxes should be the focus of current reform. The sales tax should be simplified to a three-rate structure: a basic low rate for all imports and domestic manufacturing output and two high rates for luxuries and "super" luxuries. All manufacturing products, even new items that are not produced currently, should be taxable at the basic low rate. Rates of 20 percent and 70 percent are found to be revenue-neutral on current tax bases. These rates should be levied on both imports and domestic output, so that they have no protective impact. Some existing nonuniformities in this respect (see chapter 3) should be eliminated. Exports of domestic 6-8 output should be exempt from sales tax. A system of tax credits for sales taxes paid on inputs should replace the current "ring" system as soon as possible. Ideally, since both excise and sales tax are levied ex-factory, there should be only one tax. Excise duties on all goods except alcoholic beverages, cigarettes, soft drinks, and soap, should be eliminated and replaced by a sales tax. Given current levels of fiscal dependence on excise taxes from these four items, however, they appear necessary during the transition to a predominantly sales-tax-based system, but additional excise taxes are best avoided. The excise duty structure should be simplified to one or two rates. Rates of 60 and 30 percent are found to be revenue-neutral. The higher rate should apply on goods with low elasticity, such as alcoholic drinks and high-quality cigarettes. The lower rate should apply on low-priced cigarettes, soap, and soft drinks. Exemptions for the armed forces from sales tax and excise duty should be eliminated, to prevent revenue loss. To meet Uganda's urgent revenue needs, a surtax of 5 percent on the duty-inclusive price of all imports (official and no-forex) and on the ex-factory price of all domestic manufacturing output (except exports) is proposed. This will not increase protection if applied uniformly. 6.4 Appropriate Macroeconomic Policies Consistency and stability in exchange rate, fiscal, and monetary policies are a prerequisite to making Uganda's resource allocations and production more efficient. Nominal devaluations fail to improve the relative prices of tradables if fiscal and monetary policies are 6-9 inconsistent. Short-term macroeconomic goals should be to reduce the exchange-rate premium on the parallel market and move toward more realistic official real exchange rates and then to maintain stability in the lower premium and the real exchange rate. Unification of the official and parallel market rates should proceed quickly. Given the heavy donor financing of imports, official devaluations do not hurt the fiscal balance. And the inflationary impact of devaluations on the official exchange rate ought to be limited since most prices already reflect parallel market rates. The quasi-legal parallel market is highly sensitive to expectations, so these objectives require great improvements in macroeconomic management. Generally, under a predetermined unified exchange-rate system, nominal devaluations or crawling-peg adjustments of the official rate can be used to sustain appropriate real exchange rates in the face of expansionary financial policies. But the nominal parallel market exchange rate cannot itself be a policy instrument. It is an endogenous rate determined mainly by such factors as terms-of-trade shocks, fiscal monetary policy, asset preferences or substitutability, and of course, news and expectations. Also, nominal devaluations of the official rate have little impact on the exchange-rate premium unless supported by macroeconomic policies, as Uganda's experience in May 1987 and July 1988 6-10 confirms.3 Maintaining monetary and fiscal restraint4 will greatly help to reduce the exchange-rate premium and stabilize the official real exchange rate at a realistic level. 6.5 Exchange Rate Reform As the exchange rates merge and the macroeconomic environment stabilizes, a mechanism for adjusting the official exchange rate must be set in place. This would help reduce overvaluation of the official rate because of delayed adjustments and would impart some stability to the premium on the parallel market. One option is to allow the official exchange rate to be market-determined and adjustments to be automatic. The other is to improve and liberalize the system for allocating foreign exchange and simultaneously to institute a quasi-automatic mechanism for maintaining a realistic value for the real official exchange rate. Market-Determined Official System Instead of administratively determining the official exchange rate, the government can allow market demand and supply to determine the equilibrium rate. The debate about fixed versus flexible or market-based 3 Econometric estimates of an equation for Uganda's parallel real exchange rate show that changes in the nominal official exchange rate have only a minor impact on the parallel real exchange rate. The level of the official rate relative to a notional equilibrium rate has an expectations-based effect on the parallel rate. 4 Under the World Bank and IMF adjustment programs, Uganda expects to sharply decelerate monetary growth over the next two years. Success in this endeavor will help restore stability and transparency to the incentive system. 6-11 exchange-rate systems is unresolved, but the market-based system has certain advantages. During reform, for example, if the appropriate rate is not known, a market system removes the need to select a rate. Equally important, it depoliticizes the exchange rate, and adjustment is quick and rapid. And setting reserve levels instead of a price permits a more liberal trade policy, although a stable, consistent macroeconomic policy is a prerequisite to a successful market-determined system. Market-determined exchange rate systems generally take one of two basic forms: an interbank float (as in Gambia, Philippines, Uruguay, and Zaire) or central bank auctions (as in Bolivia, Ghana, Guinea, Jamaica, Nigeria, Somalia and Zambia). Uganda experimented with a foreign-exchange auction between 1982 and 1984, then abandoned it. Parallel to it was an administered system in which the exchange rate was rapidly depreciated to merge with the auction rate. Export receipts and donor balance of payments support were the source of supply. Under fiscal and monetary restraint (in 1982 and 1983) the auction performed well. However, unplanned increases in recurrent expenditures and heavy money financing of the deficit brought large, rapid depreciations of the auction rate, which led to restrictions on transactions to dampen demand. Ultimately, the system was abandoned. Uganda's current macroeconomic situation cannot support such a system. A market system for determining Uganda's official exchange rate is likely to remain risky until structural fiscal problems are under control, crop financing is more efficient, and export receipts increase substantially. Currently, program aid would be the only source of foreign exchange for such a system. Donor financing is conditional and often tied to both products and countries. An auction system or an interbank float under those conditions would quickly degenerate into segmented markets for 6-12 different types of products, maybe even with different auction rates. Monetary or fiscal laxity or any uncertainty about donor commitments could generate rapid depreciation of the official auction or float rate. This would inevitably lead to new restrictions on the market mechanism in an effort to offset those undesirable effects. Such restrictions could seriously undermine the credibility of the reform package and thus block the producer response to liberalization. The proposal here is therefore to adopt some form of a crawling- peg system. The official exchange rate should be adjusted frequently, not only to compensate for Uganda's inflation being higher than its trading partners, but also to reduce the degree of overvaluation in the real exchange rate. In addition, one way to depoliticize devaluations is to undertake them gradually rather than in large discrete jumps as Uganda has been doing. 6.6 Legalization of the Parallel Market A market-determined official exchange rate system probably should not be introduced immediately, but the market-based parallel exchange market can be legalized. Currently, all transactions on the parallel market and all holdings of foreign currency balances therefrom are illegal under the Bank of Uganda's foreign exchange regulations. So the actual level of such transactions or balances is unknown. The amount of no-forex imports provides a basis for estimates, however, since most of them are financed by foreign exchange from the parallel market.5 Clearly, 5 Note that a no-forex importer can legally finance imports only with foreign exchange earnings provided by friends or family working and living abroad. 6-13 substantial transaction balances are held, some of them no doubt for speculative purposes. Under current macroeconomic conditions, it is not surprising that residents prefer to hold foreign rather than domestic currency. But holdings for asset diversification or portfolio preferences are more likely to be kept abroad in interest-bearing deposits. The illegality of foreign currency transactions in the parallel market creates a risk premium in the supply price of foreign exchange. This may not be large given the government's laissez-faire approach to this market, but there is evidence that some firms' apprehension about that illegal status lowers demand for foreign exchange from the parallel market. The current market-clearing rate in the parallel market would change in a similar legal market--legalization could raise or lower the parallel market exchange rate. More important, the current rate may not even be a market- clearing rate, given the problems of disseminating information about an illegal market. The quoted rate or the rate at which parallel market foreign exchange is sold to current buyers may be lower than the amount other buyers would be willing to pay if they could be reached. So, even a free foreign exchange market may fail to allocate resources efficiently. Legalizing Uganda's parallel market will clearly generate efficiency gains. Access to foreign exchange will be widened and potential investors will be more willing to base decisions on this rate. More important, legalization of the parallel market will lead to "legalization" of other parts of the economy, which should improve indirect revenues. Not only will all parallel market transactions be officially recorded, but firms and enterprises that currently remain outside the official sector because of their dependence on the parallel foreign exchange market will have an incentive to legalize their positions. This should bring much more 6-14 economic activity into the tax net, increasing the economy's tax base and revenues. Different countries have legalized the parallel foreign exchange market in different ways and at different points in the reform process. Some countries (Egypt 1974, Peru, 1978, Yemen 1981) initiated legalization even though the official foreign exchange system remained administered and the macroeconomic situation was precarious. Others (such as Ghana) delayed legalization until reasonable macroeconomic stability was restored and the official foreign exchange system had been substantially liberalized. Some countries (Egypt, Mexico, Peru, Yemen) made it legal for residents to hold interest-bearing foreign currency deposits in the banking system as long as the deposits were used for specified purposes. Others such as Ghana licensed foreign exchange bureaus outside the banking system to hold foreign currency and exchange it at any exchange rate they chose. The difference in approaches is of critical importance if legalization is undertaken while the macroeconomic situation is still unsettled. Foreign exchange bureaus facilitate residents' transactions in parallel market foreign exchange; allowing foreign currency deposits facilitates both transactions and financial investment in foreign currencies. Uganda's Options Uganda's choices about the form and timing of legalization should be influenced by its economic situation. Ideally, legalization would follow stabilization and the successful liberalization of the official market for foreign exchange, so that there is a single exchange rate for current account transactions. Then legalization makes the parallel market superfluous and leads to a unified exchange market. With earlier 6-15 legalization, there is a risk of institutionalizing the existing dual exchange market, thereby reducing the pressure to make the official system of allocation more efficient. But exchange rate unification in Uganda is feasible only in the medium term because of short-term revenue constraints. And one way to raise revenue is to expand the narrow tax base quickly. So to bring more economic activity into the official sector and thus into the tax net, Uganda should consider legalizing the parallel market even before it unifies current account transactions. In the present macroeconomic situation, the form of legalization will make a crucial difference. The economic impact of legalization will depend on the extent and source of increases in officially recorded foreign currency balances. The increase could come from current holdings of parallel market balances or the creation of new balances as foreign currency is substituted for domestic currency balances. How much recorded balances increase will depend on the private sector's confidence in legalization, the new regulations governing the holding and use of these balances, and how much larger a financial return one can get on foreign currency holdings than from domestic assets. Countries that have legalized residents' holding of interest- bearing foreign currency balances in an unstable macroeconomic environment have faced many problems.6 Large-scale shifts from domestic to foreign currency balances intensify foreign exchange pressures and reduce the 6 Foreign currency deposits as a share of total financial sector liabilities to the domestic private sector rose from 30 percent in 1980 to 50 percent in 1989 in Peru; from 5 percent in 1980 to 47 percent in 1985 in Yemen; and from 28 percent to 47 percent in Egypt. 6-16 effectiveness of exchange rate and monetary policy. In addition, the banking sector becomes highly unstable if its interest payments and a growing share of liabilities are denominated in foreign exchange, while bank assets are held mainly in domestic currency--and the lender of last resort (the central bank) cannot be expected to function easily with heavy foreign currency liabilities. The other option, perhaps better suited to Uganda, is to legalize transaction balances in foreign exchange only through nonbank dealers, following Ghana's model of foreign exchange bureaus. The main features of this approach would be the following: * Foreign exchange bureaus (outside the banking system) would be licensed by the Bank of Uganda (for a lump-sum annual license fee) to buy and sell major foreign currencies at any market exchange rate. Their current transaction rates would have to be displayed. * Any individual or firm could buy or sell foreign exchange from or to these bureaus for any purpose. But officially allocated foreign exchange could not be sold to bureaus. Noncoffee exporters would be permitted to sell their retained receipts to the bureau and no-forex importers to buy foreign currency. * The bureau would report monthly the aggregate volume of purchases and sales of each currency traded. No details of individual transactions or customers' names would be required by either the bureaus or the Bank of Uganda. * The bureau would be permitted to hold balances in foreign currency but only as non-interest-bearing bank deposits. There should be no attempt to tax the foreign exchange balances or the transactions themselves. This would make such transactions less profitable and discourage the desired shift from illegal to legal foreign exchange balances. CHAPTER 7 SHORT-TERM EFFECTS OF PROPOSED REFORM What happens when countries liberalize their trade and exchange rate regimes? Good things, including increased output, more competitiveness, greater flexibility, the removal of policy-induced biases, and gains from specialization and trade between countries as well as between individuals and households. The economic performance of countries with a relatively liberal and neutral trade regime has been better than that of those without. Indeed, it is often the poor performance of countries with closed regimes that finally convinces them to liberalize trade. But short-run transitional problems are possible. Predicting what will happen if certain policy reforms are introduced is hazardous. Generally, it can be done in two ways. In a synthetic or quantitative approach one might construct a model of Uganda's economy depicting the links between sectors to assess how output, trade, and macroeconomic balances would respond to the price changes that the more liberal trade and exchange rate policies would stimulate. In what might be called the historical or qualitative approach, one would observe the policy changes Uganda has made since 1987 to open up its trade regime and the effects they have had, as a basis, however limited, for calculating the short-term effects of further rounds of liberalization. We chose the latter approach only because data on how the economy operates and on the links between competing activities are unavailable in any form. No input-output table exists, nor are there any census data on manufacturing or agriculture. At best, we can sketch directions of change and possible short-term problems to determine if anything is likely to worsen Uganda's economic situation and undermine reform. 7-2 In this chapter we examine the potential short-term effects of liberalization on agricultural exports, manufacturing output, and government revenues. 7.1 Effects on Agricultural Exports A program of policy reform that removes both price and nonprice obstacles to exports should have a favorable effect on Uganda's exports. The monopoly of parastatal marketing agents and administrative impediments to exports are the main obstacles to the success of trade, pricing, and exchange-rate reform. Coffee and Other Traditional Agricultural Exports Uganda's coffee exports are large, so even a small increase is likely to have a substantial impact on Uganda's total exports and income. Substantially reducing export taxes should lead to a significant short-run rise in output even if elasticity of supply is low.1 The point is whether incremental increases in output can be exported successfully and that depends on how open or restrictive world market conditions are for Uganda and whether domestic institutional arrangements permit Uganda to exploit whatever export potential exists on the world market. 1 Variations in picking, pruning, and other aspects of crop husbandry are the main source of supply elasticity in the short run. Bushes in Uganda are old. The average age is 40 years, and 40 percent of bushes need to be replanted. There is currently no incentive to replant them. 7-3 On the external market, what happens depends partly on whether the International Coffee Agreement (ICA) is revived. When the ICA was in effect, Uganda and other major coffee producers received a premium price. To secure this premium, a country had to promise to limit production. The disadvantage of the ICA to Uganda was that its exports were limited to 3.5 percent of ICA markets, substantially below the country's market share of 5.7 percent in 1969-71. This quota proved inadequate to absorb production, based on the evidence of rising stocks of the coffee marketing board in 1987 and 1988. Small members like Uganda could probably get away with sales on second-order nonquota or residual markets, and it helped if these sales were made at disguised prices so that other members of ICA could not accuse the seller of selling below the ICA price. Uganda used barter transactions to sell on the residual market without revealing the price.2 With no ICA in effect, the world market for Uganda is limited only by Uganda's ability to produce and to find buyers. Uganda's robusta coffee is generally regarded as high in quality, although grading and delivery problems may have hurt that reputation in recent years. Uganda has received higher prices than have C8te d'Ivoire, Indonesia, and Zaire, and its cost of production is lower than in several African countries. World consumption of coffee is likely to increase annually by 1.5 percent. To expand output at a faster rate, Uganda would have to aggressively increase its share of both quota and nonquota markets. Its current small share in world exports is a positive factor. Increasing its coffee exports by 50 percent will increase its world share by only 1.5 2 As it happens, non-ICA members are often Eastern bloc countries in which foreign exchange is also rationed. In 1988-89 about 30 percent of the coffee crop was sold on the nonquota market. 7-4 percentage points. Recent success in exporting 3.1 million bags of coffee, which is the highest since 1986, gives cause for optimism. So reducing the export tax and ensuring that producers prices move in line with world prices are likely to provide important efficiency gains. Equally important is establishing a supportive institutional structure capable of fully exploiting Uganda's export potential. Uganda's ability to export any increases in coffee output currently depends wholly on the coffee board, whose inefficient management of both procurement and exporting hinders its capacity to market coffee aggressively. Mechanisms for monitoring export prices and repatriating export proceeds remain notoriously slow. Both slow shipments of procured coffee and long delays in payments from buyers plague the coffee board. This inefficiency increases the inflationary impact of Uganda's system of coffee crop financing.3 An improved pricing policy will have a positive effect only if the coffee board becomes more efficient at procuring high- quality coffee, recovering foreign currency proceeds from its exports, and finding new foreign buyers. More importantly, cooperative unions and private sector participation in the export of Ugandan coffee can make a major difference. Competition will encourage efficiency in penetrating export markets, and Uganda will be less dependent on public resources for external marketing. Other traditional exports such as tea and cotton also require less parastatal control and elimination of export monopoly. Increased private- sector participation in marketing must be complemented with supportive investments in processing and quality control if there is to be a 3 Some estimates suggest an inflation-elasticity of 0.4 for coffee-crop financing. See discussion in chapter 5. 7-5 substantial export response to the elimination of implicit and explicit export taxes. Nontraditional Agricultural Exports Since the introduction of the foreign exchange retention scheme for noncoffee exports in 1988, recorded or licensed exports of nontraditional items have been increasing steadily. For the first half of 1989/90, it was US$28 million, implying an annualized value of US$56 million. This is a significant improvement over exports in 1988/89, which amounted to US$31.6 million. Hides and skins and horticultural products (beans, fruits, vegetables) account for more than 80 percent of total nontraditional exports under the "dual-license" export retention scheme (see table 7-1). Table 7-1. Nontraditional Exports Under the Foreign Exchange Retention Scheme, 1988 and 1989 (percentage of nontraditional exports) Jan.- July Jan.- July- June Dec. June Dec. 1988 1988 1989 1989 Composition (Z) Hides and skins 2.9 35.4 38.7 29.8 Timber 40.3 11.8 3.3 2.0 Cereals (maize, millet) 20.9 16.3 2.6 9.5 Beans -- 6.1 37.7 25.5 Fruits 6.5 6.4 2.1 12.6 Vegetables/spices 11.7 6.6 4.2 6.6 Fish -- 5.0 7.9 6.5 Total nontraditional exports 9.4 18.1 13.5 28.0 ($ mill.) Note: Tea is excluded because it is a traditional export. Source: Calculated on the basis of dual licenses issued by Uganda's ministry of commerce. 7-6 There are some who argue that the issuance of a license does not imply utilization. While this is theoretically correct, it is difficult to understand why exporters would go to so much trouble and expense to obtain licenses, and then not use them. The absence of customs-based export data makes this contention difficult to resolve definitively but highlights the urgency of such data for policy. The destination of traditional exports given in table 7-2 confirms that poor infrastructure notwithstanding, Uganda has been able to sell more than a quarter of its nontraditional exports to European markets. Table 7-2. Destination of Nontraditional Exports (US$millions) Destination Value Africa 30.9 Europe (UK excluded) 16.0 UK 10.0 Middle East 5.0 Asia 3.5 Others 3.6 Total 69.0 Source: Computed from Ministry of Commerce data on "dual licenses." 7-7 The involvement of regular exporters in exports of nontraditional products under the export retention scheme appears to be rising. Between January and December 1988, only one firm exported more than once; between July 1988 and June 1989, only three firms exported more than once and two exported three times.4 Currently, there are seven such firms. This is no mean achievement given the administrative impediments to the export of nontraditional items. Even the retention scheme involves the blocking of working capital for at least nine to ten months, as explained in chapter 4. Removal of administrative impediments and liberalization of the export retention scheme could easily lead to nontraditional exports considerably greater than US$56 million per year. 7.2 Effect on Manufacturing Output In the absence of systematic data on the manufacturing sector, predictions about the effect of reform on this sector must be viewed with caution. Output in the industrial sector was greatly affected by political strife and economic insecurity in the 1970s and between 1984 and 1986. In the last two years, overall performance has improved significantly. The manufacturing sector has experienced extremely high rates of growth, albeit from a low base (see last column of table 7-3). Better security, higher domestic demand, and increased imports of inputs have been key factors in 4 The frequency of export involvement is calculated on the basis of the number of "dual licenses" a particular exporting firm received in those two periods. 7-8 Table 7-3. Structure of the Ugandan Manufacturing Sector Output Share of Share of growth No. of No. of manufacturing manufacturing 1987-89 establishments employees employment value added (Jan-May) Industry (1988) (1988) (percent) (percent) (percent) Food processing Meat and dairy 5 1.7 16.1 Grain milling 11 4.3 50.8 Bakeries 7 1.4 45.3 Sugar and Jaggery 4 1.8 354.9 Coffee roasting 3 0.2 32.4 Coffee processing + 8.6 21.4 Tea processing 1+ 1.4 18.5 Other food processing 3 0.3 18.0 Animal feed 5 1.0 16.5 Subtotal 39 3.508 18 20.7 568.1 Tobacco and beverages Beer and spirits 5 6.6 34.0 Soft drinks 6 5.4 144.8 Cigarettes 1 14.0 12.3 Subtotal 12 2,727 14 26.1 45.2 Textiles and clothing Textiles 4 12.0 13.4 Textile products 3 3.1 1.3 Garments 4 2.2 241.1 Subtotal 11 7,517 39 16.3 27.9 Leather and footwear Subtotal 6 424 2 2.3 51.1 Timber, paper, etc. Sawmilling and timber 2 3.2 22.1 Furniture, foam products 5 2.9 113.7 Paper and printing 10 2.9 113.9 Subtotal 17 1.158 a 9.0 65.4 Chemicals, paint and soap Chemicals 1 0.3 13.2 Paint 3 0.5 118.2 Medicines 2 0.5 18.1 Soap 8 11.0 61.0 Subtotal 14 131 1 12.3 60.1 7-9 Table 7-3. Structure of the Ugandan Manufacturing Sector (cont.) Output Share of Share of growth No. of No. of manufacturing manufacturing 1987-89 establishments employees employment value added (Jan-May) Industry (1988) (1988) (percent) (percent) (percent) Bricks and cement Bricks, tiles, etc. 6 2.2 14.8 Cement 2 2.1 42.4 Subtotal 8 1,454 8 4.3 28.0 Iron and steel 4 1.5 52.2 Structural steel 4 2.3 49.4 Steel products 5 1.5 27.5 Subtotal 13 1,853 10 5.3 1.4 Miscellaneous Vehicle parts and accessories 5 0.9 45.1 Plastic products a 0.8 23.0 Electrical products 2 1.1 36.7 Miscellaneous products 5 1.0 149.7 Subtotal 15 518 2 3.7 59.5 Source: Estimates of the numbers of establishments and the output growth rate are from Statistical Bulletin No. IP12, Index of Industrial Production, Statistics Department, Ministry of Planning and Economic Development, September 1989. Estimates of employees by sector are drawn from table 66 in Background to the Budget, 1989-90, Ministry of Planning and Economic Development, July 1989. Although these estimates do not refer to all manufacturing establishments, they are considered sufficiently representative to illustrate the distribution of manufacturing employment across the various activities. that overall growth. Improvements in them will continue to be a dominant factor in determining manufacturing output. The question here concerns the short-run impact of proposed reform. Eliminating OGL and using SIP at a more appropriate exchange rate for all official allocations would no doubt reduce the rent or subsidy now available on industrial inputs. But as the gap between official and parallel rates narrows, the ad valorem tariffs and thus domestic prices of final imports will rise, providing more tariff protection than before. 7-10 The fact that high tariffs (higher than 50 percent) will be reduced is not very significant in terms of adverse impact, since they provide little assistance to current production. In addition, there are few imports at that rate. Raising the lowest tariff rate to 10 percent and eliminating duty exemptions will effectively increase assistance to intermediate inputs. Elimination of OGL will also increase the relative protection of inputs since most imported industrial inputs are effectively subsidized by the official exchange rate. Growth in manufacturing output has been uneven, however, partly because of different levels of demand growth, but more because of supply deficiencies. Shortages of imported inputs, inadequately maintained machinery, and inadequate credit or liquidity affected different firms differently. The dominance of parastatal firms with poor managerial ability and uncertain ownership has exacerbated uneven development as has the government's policy of selectively supporting a few firms with large subsidies for imported inputs, especially through the OGL (see section 5.3 in chapter 5). OGL has influenced the structure of the sector by encouraging different rates of growth in the subsectors. Significant structural changes between 1983 and 1988 are evident from the data on output and value added (see table 7-3). In 1988, tobacco and beverages accounted for 26 percent of value added, compared with 7 percent in 1985. Cigarettes, soft drinks, and beer, the main beneficiaries of OGL, were also growth industries. Chemicals' share went from 7 percent to 12 percent (soap was an OGL sector). Among the non-OGL industries, shares show an increase only for clothing and garments. At 20 percent, the share for food processing has not changed since 1985, but all other shares have fallen. Textiles' 7-11 share fell despite the OGL, largely because of other constraints. The difficult policy question is how much of current industrial growth is the result of the artificial, probably unsustainable, encouragement of subsidized imported inputs.5 Removing or reducing the OGL subsidy will reduce growth of the OGL industries. But no fewer than seventy-five firms have imported inputs at higher rates than OGL rates (nearly 32 of them at either SIP-II or no-forex rates), and produced at a profit in 1989, so output declines in OGL sectors are not likely to be precipitous. If earlier studies of economic viability are valid today, only fishnets, gunny sacks, and hessian cloth are expected to cease production. Among industries receiving SIP subsidies on inputs, only rolled steel and plywood output are likely to be so affected. Tariff reductions are likely to affect only a few industries, since most tariffs over 50 percent cover noncompeting imports. Paint is probably the only industry affected. The duty on beer is 350 percent, but there are currently no recorded imports of it, probably because it is on an implicit negative list of imports. A quick survey 6f the output performance of sixteen products that account for about 46 percent of the sector's value added (see Annex II), suggests that import liberalization has had a limited negative impact on the output growth of non-OGL firms or products. This may be partly because of competition from unrecorded imports that existed even before liberalization and even more because of such factors as the condition of machinery, managerial ability, uncertain ownership, infrastructural 5 Many people interviewed by the mission cited the age and poor condition of capital equipment as a major constraint. 7-12 deficiencies, and poor access to credit. Output levels on some products were so low that no harmful effects were possible--the only direction to go was up. And by increasing the supply of imported inputs, liberalization relaxed an important constraint on output. These factors explain much of the overall output growth in 1988 and 1989 that has accompanied liberalization. The more important question for the future is whether the decline in output of firms dependent on OGL or SIP (and there will be some decline unless demand rises) will be offset by growth in other firms and sectors. With even more access to foreign exchange and a more favorable policy environment because of reform, such growth is likely. But the output response of the other industries and firms will be constrained by the condition of parastatal enterprises and by the financial sectors' limited ability to provide working capital. Data are poor, but parastatals currently represent about half the value added in the manufacturing sector, employing nearly three-fifths of the labor force in that sector. Their ability to respond to improved incentives is constrained by their financial situation (government transfers are limited), by bureaucratic management practices, and by little government or market pressure to be profitable. Privatization of some of the parastatals and rationalization of the rest will greatly help ensure overall growth in the manufacturing sector. The proposed reforms are expected to have a negligible harmful effect on manufacturing output for five reasons. First, the changes in incentives to produce manufactured goods are not unfavorable, except for a few OGL firms. Second, current levels of capacity utilization are extremely low in most industries. Except for beer, soft drinks, soap, and 7-13 cigarettes, capacity utilization does not exceed 20 percent, and most firms operate at 10 percent of capacity. So the scope for decline is limited, and few or no operating units are likely to close down. Third, as domestic demand and security improve, the overall effect on growth in manufacturing will be more important than any adverse impact of reform on the sector. Fourth, better and more even access to imported inputs will lead to better capacity utilization in most sectors now operating at 10 to 20 percent of capacity. That incremental growth will offset any decline in the OGL and SIP sectors. Fifth, policies to alleviate other constraints (financial sector problems, parastatal management, uncertain ownership, inadequate infrastructure) will be growth-inducing. 7.3 Impact on Government Revenues The proposals for reform will have both direct and indirect effects on government revenues. With no adequate model of the economy, only the direct short-term impact is predictable. The proposals in chapter 6 will affect the economy's tax base, partly because they support growth in the medium term and partly because transactions and economic activity in the underground sector will shift to the official recorded sector. In an economy like Uganda, with very low capacity utilization and a historically strong parallel market economy, increasing the tax base by changing those two factors will provide the most important increases in government revenues. The estimates given below do not take them into account and to that extent underestimate the potential for increased tax revenue. 7-14 Taking account only of the existing tax base, the proposals are expected to increase revenues from import taxes and domestic indirect taxes, but to lower revenues from export taxes (see table 7-4). Explicit export tax revenues averaged about 13 percent of exports in 1988/89. Assuming a 10-percent ad valorem export tax on the f.o.b. value of export receipts means a decline of U.Sh. 1,702 million over 1988/89.6 This will be more than offset by taxes from imports. The proposal to raise or maintain the lowest import tariff to 10 percent and to eliminate exemptions will raise duty collections by U.Sh. 6,285 million on the basis of the 1988/89 import value in local currency. Reducing high tariffs will lower revenue no more than U.Sh. 100 million.7 A surtax of 5 percent on the duty-inclusive price of imports and on ex-factory price of domestic output will provide about U.Sh. 8,475 million in revenues. Excise duties of 60 and 30 percent and the sales taxes of 70 and 20 percent will increase revenues from domestic indirect taxes slightly by U.Sh. 3,480 million and U.Sh. 841 million, respectively, from domestic output and by U.Sh. 3,255 million from imports. 6 Coffee exports will effectively be taxed at a much higher rate if the implicit exchange rate tax is added to this ad valorem tax. But this will decline as exchange rate unification proceeds. 7 Devaluation of the official rate will increase these amounts because of their direct effect on the import and export tax bases. 7-15 Table 7-4. Effect on Revenues of Proposed Reforms (millions of Uganda shillings) Existing With Item (1988/89) reforms Difference Export taxes 5,839 4170 -1668 Import duty 7,775 13,960 6185 Sales tax 17,534 21,269 3,735 Imports 5,271 8526 3255 Domestic output 12,263 12,743 480 Surtax (@ 5Z) 0 8475 8475 Imports 0 7177 7177 Domestic output 0 1298 1298 Excise duty 4,786 5627 841 Total revenue from indirect 34,934 53,501 17,567 taxes Total revenue as X of GDP 4.26 6.34 +2.08 Note: Computed on the basis of current and estimated tax bases. This increase in revenue does not take into account the effect of expanding the tax base. ANNEXES ANNEX I, TABLES - 1 - Table A-1. Tariff Rates Section in Tariff I II III IV Animal and Section Vegetable vegetable fats Prepared characteristic Animal products products and oils foodstuffs Modal range 10-30 10-50 5-40 10-50 Modal rate 30 40 10 30 Degree of 90 50 80 80 concentration Items at free Live animals Barley Tallow Malt Animal somen Hops Infant food Outliers, high None None None Changes to Vegetable oils Cigarettes and printed version from 10 to 5 spirits from 100 to 250 Sales tax All free Flour at 80 Sales taxes of Sales taxes Import duty 30 on lard exceed tariffs at 10 tallow for coffee, tea Other items Duties of 40, vegetables, and free free, and 10 many other 1 tems No. of specific None None None None duties End-use None None None None distinctions - 2 - Table A-1. Tariff Rates (cont.) Section in Tariff V VI VII Artificial Products of resins and Section Mineral chemical plastic characteristic products industries materials Modal range 10-30 10-20 10-20 Modal rate 10 10 10 Degree of 75 70 70 concentration Items at free Diesel and Tannery extracts Natural rubber kerosene, salt insecticides butchery, oil fertilizers for rope and pharmaceuticals spinning Outliers, high Motor and Cosmetics at 100 A few at 30 and 50 aviation spirits Changes Salt from 30 to Soap from 30 to 80 Tires and tubes from free Bar soap free 20 to 15 Sales tax features Sales tax exceeds Paints at 80 with Sales tax of 20 for duty for cement, duty of 30 tires duty of 15 etc. Bar soap free Leather at 40 Petroleum jelly duty at 20 No. of specific duties End-use distinction Oil for rope Products for use Tires for aircraft at free, in the manufacture other at 30 of beverages changed from 20 to 10 - 3 - Table A-1. Tariff Rates (cont.) Section in Tariff IX XI Section Wood and Textiles and characteristic wood articles textile articles Modal range 10-30 30-40 Modal rate 20 30 Degree of 60 70 concentration Items at free Beehive, wood pulp None coffins, printed material Outliers, high Some paper at 60 Changes Bond paper from 10 to 50 Woven fabrics from 30 to 40 cheque books from 30 to free Sales tax features Exercise books 30, duty of 20 Umbrellas, electric blankets, Floor coverings 40, duty of 30 hats, other bedding, Writing paper 40, duty of 30 footwear, outer garments, sacks of jute, sheets, clothing (all with sales tax greater than duty) No. of specific duties End-use distinction - 4 - Table A-1. Tariff Rates (cont.) Section in Tariff XIII XIV XV Articles of stone, Pearls, precious Base metals Section plaster, cement, and semi-precious and characteristic etc. stones articles of base metal Modal rang* 10-20 10-50 10-20 Modal rate 10 30 10 Degree of 70 80 75 concentration Items at free Pharmaceutical Coins Most steel and iron containers ships and sections Hoes, sickles, etc. Outliers, high A few at 30 None Changes None None Sales tax features Sales tax exceeds 50? on jewellery Sales tax exceeds import import duty for duty for padlocks, door worked glass, fittings, filing cabinets, lavatory cisterns, blades, cotton, brake linings, containers, sanitary ware asbestos sheets No. of specific None duties End-use distinctions Articles for agriculture and industry at free - 5 - Table A-1. Tariff Rates (cont.) Section in Tariff XVI XVII XVIII Machinery and Vehicles, aircraft Section mechanical vessels and Optical characteristic appliances associated equipment Photographic Modal range Free to 20 Free to 30 Free to 80 Modal rate Free 20 10 Degree of 50 50 50 concentration Items at free Many Many Many (industrial machines agricultural machinery, office equipment, some parts) Outliers, high None Motor vehicles 200 None Changes None None None Sales tax features Sales tax often 30 Tractor parts 30 Sales tax often in duty 20 duty at 20 excess of duty for Tractors 30 duty free watches, regulators electric meters, photocopiers, glasses End-use Refrigerators Some industrial end- distinctions industrial free users, e.g., glasses Pumps, industrial free free, domestic 20 Agricultural end- users free - 6 - Table A-1. Tariff Rates (cont.) Section Miscellaneous characteristic manufactured articles Works of art Modal range 10-30 Free Modal rate 30 Free Degree of 95 90 concentration Items at free Industrial end-use items All Medical equipment Outliers, high None None Changes None None Sales tax features Powder puffs at 40 duty at 30 Free Typically 30 End-use provisions Some industrial end-use None provisions - 7 - Table A-2. Imports by User Firms Under Special Import Program I, December 1988 to April 1989 (U.S. dollars) Firm's share in Machinery Raw industry Industrial firms and spares materials Total (%) African Ceramics 0 15,783 16,783 0.21 Berger Paints 0 160,938 160,938 2.17 Cable Corp. 0 237,412 237,412 3.20 Chloride U Ltd. 7,025 106,738 113,761 1.53 East African Distilleries 0 68,581 68,581 0.93 International Paints 0 15,555 15,555 0.21 Kampala Chalk Factory 0 50,000 50,000 0.88 Kawompe Paint Factory 0 67,742 87,742 0.91 Knit & Weaving Industry 1,173 0 1,173 0.02 Lake Victoria Bottling 130,268 344,064 474,332 6.40 Mbale Soap Works 43,106 171,970 215,077 2.90 Mukwano Industries 1,660,270 0 1,660,270 22.39 Mulbox Ltd. 0 25,080 25,080 0.34 Nile Breweries 9,155 0 9,155 0.12 Nyamitanga Printing Press 0 19,135 19,135 0.26 Nyanza Textile 0 49,670 49,670 0.67 Oscar Industries 15,817 385,782 401,599 5.42 Picfare Industries 0 38,356 38,356 0.52 Printpak 0 96,328 96,328 1.30 Sembule Steel Mills 0 551,845 551,845 7.44 Ship Toothbrush Factory 30,334 560,083 580,417 7.83 Sugar Corp. of U 644,321 68,893 713,014 9.62 Tororo Steel Works 3,044 245,267 248,310 3.35 Uganda Associated 0 197,098 197,098 2.68 Uganda Baati Ltd. 0 109,139 109,139 1.47 Uganda Bata Shoe Co. 540 19,050 19,591 0.28 Uganda Breweries Ltd. 932,787 0 932,787 12.58 Uganda Brushware Manufacturers 0 31,809 31,809 0.43 Uganda Clays 27,138 0 27,138 0.37 Uganda Grain Milling Co. 0 40,448 88,098 1.19 Uganda Oxygen 0 42,677 42,677 0.58 Uganda Wood Fabrication 65,683 0 865,563 0.88 Westmile Distilling Co. 137,601 0 137,601 1.86 Total imports 375.591 3,659,240 7.415.031 100.00 - 8 - Table A-3. Imports by User Firms Under Special Import Program I, May to August 1989 (U.S. dollars) Firm's Machinery Raw share in Industrial firms and spares materiais Total industry Associated Chemical Industries 0 10,524 10,524 0.39 Berger Paints 0 131,303 131,303 4.87 Cable Corporation 0 88,231 686,231 2.48 Century Bottling 51,870 144,927 196,797 7.31 East African Distilleries 0 140,8623 140,623 6.22 Kakira Sugarworks 4,913 218,610 223,523 8.30 Kanunco Modern Footwear 0 28,000 28,000 1.04 Kiire Sawmill A Plywood 0 24,044 24,044 0.89 Nyanza Textile Industries 40,057 0 40,057 1.49 Oscar Industries 0 80,000 80,000 2.97 Picfare Industries 0 80,000 80,000 2.97 Printpak 10,078 171,632 181,710 6.76 Steel Rolling Mills 0 308,400 308,400 11.45 Sugar Corporation of U 0 60,742 80,742 2.26 Tororo Steel Works 0 50,000 50,000 1.88 UCIL Laboratories 0 60,448 80,448 2.24 Uganda Basti Ltd. 7,500 302,900 310,400 11.62 Uganda Bata Shoe Co. 609 142,807 143,415 5.32 Uganda Clays 5,884 0 5,884 0.22 Uganda Fishnet Manufacturers 0 73,920 73,920 2.74 Uganda Grain Milling 0 11,816 11,816 0.44 Uganda Rayon Textile Manufacturers 0 447,740 447,740 16.62 Uganda Shoe Co. 0 17,955 17,955 0.67 Total imports 120,911 2,572,819 2,693,530 100.00 - 9 - Table A-4. Imports by User Firms Under Special Import Program II, June to August 1989 (U.S. dollars) Firm's Machinery Raw share in Industrial firms and spares materials Total industry Berger Paints 0 14,948 14,948 0.95 Bugishi Industries 337,578 0 337,578 21.52 Cable Corp. 2,485 780 3,265 0.07 Casements Africa Ltd. 16,321 0 3,265 0.21 Chloride Uganda Ltd. 9,374 1,402 10,777 0.69 International Paints 2,387 7,386 7,386 0.47 Lake Victoria Bottling 1,644 16,000 32,414 2.07 Lawson Chemicals 0 3,827 3,827 0.24 Mbale Steel Wire Industry 0 6,000 8,000 0.38 Mukwano Associated Packers 0 2,500 2,500 0.16 Mukwano Industries 741,950 0 741,950 47.29 Mutanywana Maize Mill 0 4,598 4,598 0.29 Oscar Industries 0 9,8 9,828 0.63 Pattex Garment Industry 0 4,995 4,995 0.32 Peacock Paints 0 9,028 9,028 0.58 Picfare Industries 22,750 0 22,750 1.45 Reco Industries 5,833 0 5,833 0.37 Sadolin Paints 0 15,000 15,000 0.96 Sembule Steel Mills 0 118,818 116,818 7.45 Ship Toothbrush Factory 33,229 0 33,229 2.12 Tororo Steel Works 2,905 0 2,905 0.19 Uganda Associated 0 158,931 156,931 10.00 Uganda Breweries 3,674 0 3,647 0.23 Uganda Clays 686 0 686 0.04 Uganda Toothpaste Manufacture 0 4,538 4,536 0.29 Total imports 451,253 1,117,576 1,568,830 100.00 - 10 - Table A-5. Imports by User Industries Under No-Forex, January to August 1989 (U.S. dollars) Firm's Machinery Raw share in Industrial firms and spares materials Total industry (S) African Ceramics Co., Ltd 0 912 912 0.08 Chloride (U) Ltd. 3,432 0 3,432 0.30 Lake Victoria Bottling Co. 24,107 80,313 104,421 9.12 Mukwano Industries (U) Ltd. 250,771 558,028 808,800 70.67 Oscar Industries (U) Ltd. 0 608 608 0.05 Printak (U) Ltd. 12,331 0 12,331 1.08 Ship Toothbrush Factory Ltd. 45,888 188,040 213,926 18.89 Total imports 338,529 807,903 1,144,433 100.00 - 11 - Table A-6. Imports by User Firms Under No-Forex, January to June 1988 (new Ugandan shillings) Machinery Raw Industrial firms and spares materials Total African Ceramics Co., Ltd. 635,218 8,631 643,850 Associated Chemical Industries 743,615 1,129,559 1,873,174 Berger Paints 0 8,466,288 8,466,288 Casements (A) Ltd. 914,027 702,004 1,616,031 Chloride (U) Ltd. 6,047,816 49,245 6,097,061 Mukwano Industries (U) Ltd. 18,518,478 2,958,588 21,477,066 Oscar Industries (U) Ltd. 303,769 0 303,769 Printak (U) Ltd. 0 5,922,382 5,922,382 Sadolin Paints 0 894,778 894,778 Ship Toothbrush Factory Ltd. 3,035,840 7,924,638 10,960,478 Uganda Associated Industries 0 70,346 70,346 West Nile Distilling Co. 1,022,058 0 1,022,058 Total imports 31,220,824 28,126,463 59,347,287 - 12 - Table A-7. Imports by User Firms Under No-Forex, July to December 1988 (new Ugandan shillings) Machinery Raw Industrial firms and spares materials Total African Ceramics Co., Ltd. 0 1,465,019 1,465,019 Associated Chemical Industries 0 497,038 497,038 Chloride (U) Ltd. 581,279 0 581,279 Lake Victoria Bottling 0 1,053,247 1,053,247 Mukwano Industries (U) Ltd. 2,706,388 2,423,808 5,130,196 Sadolin Paints 0 1,997,149 1,997,149 Sembule Steel Mills 545,280 0 545,280 Ship Toothbrush Factory Ltd. 8,259,745 4,785,083 13,044,828 Uganda Clays 0 305,400 305,400 Total imports 12,092,692 12,526,746 24,619,439 - 13 - Table A-8. Sales Tax Rates on Major Imports, 1988/89 (percentage) Sales Sales tax tax Description rates Description rates Citrus fruits Free Butter 30 Medicaments Free Vinilla 30 Fertilizers Free Vegetable oils 30 Electric rails locomotives Free Other sugars 30 Live horses, asses 10 Cocoa paste 30 Meat and edible affa 10 Lubricating preparation 30 Fish 10 Plywood 30 Guts, bladders 10 Agglomerated cork 30 Other live plants 10 Exercise books 30 Manioc, arrowroot 10 Articles of plaster 30 Flours or meals 10 Bolts and nuts 30 Vegetable saps 10 Locks 30 Vegetable products 10 Radiotelegraphic 30 Beet pulp 10 Mattress supports 30 Manufactured tobacco 10 Equipment for parlou 30 Clays 10 Pencils 30 Slag, dross 10 Buckwheat, millet 40 Chemical products 10 Meat extracts 40 Organic 10 Prepared foods 40 Prepared glues 10 Fruit 40 Composite solvents 10 Sauces 40 Artificial resins 10 Perfumery, cosmetics 40 Raw hides, skins 10 Silk 40 Basketwork 10 Woven fabrics of met 40 Waste paper 10 Woven fabrics or fla 40 Picture books 10 Hemp 40 Woven fabrics of cotton 10 Woven pile fabrics 40 Coin 10 Rubberized textiles 40 Tubes and pipes 10 Gloves 40 Nickel 10 Rags 40 Magnesium 10 Footwear 40 Lead 10 Head bands 40 Zinc 10 Parts, fittings 40 Tin 10 Sinks 40 Tugsten 10 Safety glass 40 Swas 10 Travel goods 50 Airplanes 10 Articles of fur 50 Ships 10 Woven fabrics for men 60 Side-arms 10 Woven fabrics for women 50 Worked animals 10 Garments 50 Hand slves 10 Bed-linen 50 Paintings 10 Wigs, false beards 50 Petroleum oils 20 Pearls, stones 50 Glaziers' putty 20 Motor vehicles 50 Other combustibles 20 Other clocks 50 Chemical products 20 Phonographs 50 Rubber tires 20 Cereal flours, wheat 50 Tubes and pipes 20 Wine 50 Centrifuges 20 Lenses 20 Source: Custom and Excise Department, Ministry of Finance. - 14 - Table A-9. Sales Tax Rates on Domestic Products, 1988/89 (percentage) Sales Sales tax tax Description rates Description rates Soap 10 Edible oil* 20 Horch covers 10 Jaggery 20 Coffee 10 Molasses 20 Tea 10 Iron sheets 20 Fencing posts 10 Batteries 20 Curry powder 10 Matches 25 Feeds 10 Cement 25 Toothpaste 10 Biscuits 30 Polythene 10 Shoes* 30 Tubes and tires* 10 Pangas 30 Leather 10 Sweets 30 Textiles* 10 Uganda waragi 30 Hessian bags 10 Paints 30 Welding rod 10 Detergents 30 Bronze bars 10 Plastics 30 Steel bars 10 Timber 30 Welded mesh 10 Boxes and bags 30 Wiring rod 10 Karais 30 Water tanks 10 En-Ware 30 Barbed wire 10 Nails 30 Steel wire 10 Mattresses 30 Chain links 10 Pencils 30 M.S. plates 10 Chalk 30 Manhole covers 10 Rice 40 Maize mill 10 Paper 40 Maize muller 10 Soft drinks 50 Weights 10 Sufurias 50 Pulleys 10 Wheat flour* 60 Cables 10 Wine 80 Brushes 10 Cigarettes 35-90 Beer 90 * Import sales tax rates on these products were higher in 1988/89. Source: Inland Department, Ministry of Finance. - 15 - Table A-10. Sales Tax Base, Revenues, and Effective Tax Rates, 1988/89 (millions of Ugandan shillings) Tax Tax Effective Tax Tax Effective Commodity base yield rate Commodity base yield rate Beer 4,098.0 3,907.9 95 Water tanks 10.11 1.72 17 Cigarettes 5,332.6 3,539.9 86 Pencils 4.69 1.61 34 Soft drink 3,819.1 2027.2 53 Barbed wire 6.06 1.55 28 Wheat flour 1,831.5 1199.4 85 Coffee 6.86 1.22 18 Textiles 1,802.9 502.4 28 En-ware 7.61 1.18 15 Mattresses 453.4 132.9 29 Biscuits 3.53 1.02 29 Timber 353.3 132.6 38 Boxes and bags 10.31 0.97 9 Soap 5,396.7 111.8 2 Brushes 3.80 0.52 14 Iron sheets 513.7 102.7 20 Polythene 3.86 0.39 10 Uganda waragi 265.2 98.2 37 Chalk 1.43 0.34 24 Plastics 223.2 73.1 33 Welding rod 2.98 0.30 10 Cement 223.8 69.0 31 Cables 2.45 0.24 10 Paper 158.0 62.2 40 Steel wire 2.40 0.24 10 Paints 135.5 49.7 37 M.S. plates 2.40 0.24 10 Batteries 159.0 39.9 25 Feeds 1.96 0.20 10 Hessian bags 136.9 35.2 26 Maize mill 1.88 0.19 10 Jaggery 95.9 30.9 32 Ugma (hardware) 1.69 0.17 10 Tea 216.5 28.1 13 Wiring rod 0.87 0.16 19 Nails 86.1 25.6 30 Rice 0.40 0.16 40 Chain links 58.0 17.4 30 Matches 0.50 0.12 25 Molasses 75.1 15.0 20 Manhole covers 0.94 0.09 10 Tubes and tires 36.3 13.1 38 Fencing posts 2.18 0.09 4 Toothpaste 74.4 7.4 10 Curry powder 0.82 0.06 10 Detergents 17.9 6.2 35 Steel bars 0.58 0.06 10 Sufurias 13.9 5.2 37 Horch covers 0.48 0.05 10 Karais 12.4 4.9 40 Bronze bars 0.38 0.04 10 Welded mesh 22.0 3.3 15 Wine 0.04 0.03 80 Leather 24.6 3.2 13 Pulleys 0.19 0.02 10 Edible oil 7.0 2.7 38 Pangas 0.07 0.02 30 Shoes 8.7 2.6 30 Maize huller 0.08 0.01 9 Sweets 6.0 2.1 35 Weights 0.04 0.004 9 Source: Inland Department, Ministry of Finance. ANNEX II THE PROBABLE EFFECTS OF REFORM IN SUBSECTORS This annex provides a qualitative assessment product-by-product of probable changes in protection as a result of the policy changes introduced after 1985. Where possible we have identified the policy change likely to have the greatest effect. This is inherently speculative and unrigorous, but it provides information where there is none. Batteries were reported to have a negative nominal rate of protection and low or negative effective rates in 1985. Between 1985 and 1988, battery output fell by 33 percent. Tariff protection is low at 20 percent, and duty on most inputs appeared to be 10 percent. Battery manufactures did not have access to OGL funds for imported inputs, but batteries were imported under SIP-I, SIP-II, and no-forex. In the 1985 survey, Associated Batteries (now called Choride Uganda), reported a stable number of employees between 1984 and 1988. In 1988, the firm produced fewer batteries than in 1985. What would happen to battery manufacturing in Uganda if the tariff were to become the import barrier that mattered and high tariffs were to be reduced or rationalized? Battery manufacturers would probably benefit. They do not currently benefit from preferred foreign exchange allocation systems, and the tariff on batteries is not high. Output has fallen significantly since 1985, reflecting in part the relatively low incentives this industry gets from the existing protective structure. Output might grow under a unified exchange rate regime. The negative protection on beer prevailing in 1984 has almost certainly been turned around since then. Two of the three breweries, Nile - 2 - Brewers and Uganda Breweries, have benefited from their participation in the OGL scheme since early 1988. Beer imports face a shut-out tariff of 350 percent and there appears to be an implicit ban on its import, so no foreign beer is available in Uganda. Barley and malt, which can be imported free of duty, are the largest components of brewer imports under OGL. What would happen if the tariff were to become the more relevant import barrier and the implicit ban were withdrawn? Beer production has been one of Uganda's fastest growing industries, rising from 15,000,000 liters in 1984 and rising to 21,500,000 liters in 1988 (with a drop to 7,000,000 liters in 1986). It would be stretching a point to link all the changes in output to the changes in protection, but there seems little doubt that access to the OGL scheme has had a lot to do with the recent expansion. So what if the Nile and Uganda breweries' access to sugar, malt, hops, chemicals and bottles through OGL were terminated? One option would be to continue to import those items using foreign exchange purchased on the parallel market, thereby driving up the cost of making beer and its price. At a duty of 50 percent on beer, import competition would be unlikely. If the price of beer rose and demand was inelastic, sales tax revenues would also rise. If tariffs were reduced under rationalization to a maximum duty of 50 percent, import competition would be likely in the short run. But beer is expensive to transport, the production process is simple, and local markets in most countries are usually serviced by local brewers. Put simply, it is usually cheaper to import the malt and chemicals and add the water and bottle the brew in the home country. As long as there are no import barriers apart from tariffs on key inputs, it would be surprising if this did not happen in Uganda. - 3 - Bicycle tires and tubes had high nominal protection according to the 1985 survey. Tires and tubes are made by a single parastatal firm, Dunlop (EA) United. Dunlop does not have access to the OGL scheme. Annual production of tires and tubes was 120,000 and 26,000 in 1985, 37,000 in 1987, and 110,000 in 1988. The duty on tires and tubes is low at 15 percent. Tires and tubes were imported through SIP during 1988. On the face of it, with an import duty of 15 percent and no access to OGL for inputs, it is unlikely that effective protection is very high. Rationalization of tariffs is likely to help in this case. Biscuits production seems to be dominated by the firm Mukesa Foods. Production was at its lowest in 1985 at 33,000 tons; since then, it has recovered to about 140,000 tons in 1988. The firm does not benefit from the OGL scheme, so there has been no big artificial boost to growth from OGL participation, as in some other industries. The biggest contributor to output growth was probably the improvement in domestic security. Access to sugar at concessional duty rates and to malt at duty- free rates may have helped, but we do not have enough information to judge. The 50-percent tariff on biscuits is high enough--particularly since some inputs, such as malt, are duty-free--to provide high effective protection. On these grounds, it would be desirable to lower the duty on the imports. Food products generally have been an expanding sector in Uganda, with the number of firms increasing from 6 to 25 between 1985 and 1988. These positive forces for growth might make up for any negative effects of reduced tariff assistance. Cables and conductors are manufactured by a single, probably parastatal, firm--the Cable Corporation. Output is down from 1985-86 levels but began to recover between 1987 and 1988. Capacity utilization is reportedly about 17 percent. The Cable Corporation has not been helped by OGL assistance. The 20-percent tariff is not high, so there is no reason to expect cable production to be affected by tariff reform. Cement is manufactured by Uganda Cement, which has two plants at Hema and Tao. The mission was told that one of these plants depends greatly on a nearby hydroelectric plant. Cement production at the plant has the advantage of a site well-suited to the production of electricity and deposits of limestone that are easily mined, but the plant itself has fallen into disuse. Again, as we understand it, what happens to output and employment in cement production in Uganda will be influenced less by the tariff than by decisions about infrastructure and refurbishment--including the choice of which site to develop. On the basis of history and engineers' assessments of local characteristics and considering the transport costs of cement, Uganda has a comparative advantage in manufacturing cement, which might be expected to predominate regardless of the tariff change. The tariff rate is 50 percent, but our impression is that exemptions are awarded frequently, particularly when local production cannot meet demand. A lower tariff with fewer exemptions and concessions would be preferable. Fabric production in Uganda has been stable at 182,000 square meters in 1985 and 201,000 square meters in 1988. The section is a large employer. For example, firms producing cotton fabrics employ a total of 6,500 people, which has been stable in recent years. Ownership seems to be important. What would happen to this sector with a tariff change? As in most industries in Uganda, capacity utilization is low, reportedly at about 30 percent. The tariff on output is typically 30 percent. Both garment makers and fabric manufacturers already seem to face extensive competition -5 - from imports of second-hand clothing, a big import item that is regarded as a "commodity" import. These imports come in using foreign exchange obtained on the parallel market. To the extent that this is already important competition for local manufacturers of fabric and garments, adjustment to tariff revisions might not be as difficult as might have been thought. Still, large numbers of people are concentrated in this sector, which, more than most, might warrant thought about transitional measures. Fishnet faces a duty of 20 percent, and Ugandan producers have benefited from OGL help. Production in 1988 was well up from 1985 levels. Tariffs on fishnet are already low, so there is no reason to expect a tariff review to harm this industry. Hessian cloth output fell in 1986 and 1987 but recovered to 1985 levels in 1988. There appears to be only one firm, and employment has been stable at about 750 employees since 1985. The tariff rate on output is 30 percent. The firm has had access to OGL, which would help explain output increases since 1986. We assume that output from this firm is used to pack agricultural exports, so to the extent that OGL assistance is passed through to users in the form of lower prices, this is one case in which OGL might be having a good efficiency outcome. Generally, we would expect the hard part of adjustment for this activity to arise from being weaned from OGL rather than as a result of any tariff change. Mattress production seems to be a big business in Uganda. There are three firms--two that make foam mattresses and one that makes innerspring mattresses. One of the firms that makes foam mattresses has access to OGL and has had large exports over the last two years. That firm is treated grandly and the others are allowed to languish. Output of the innerspring mattress manufacturer has dropped to almost nothing. Vitaform, - 6 - the firm with access to OGL, would face extensive adjustment pressure were OGL provisions to be dismantled. Even so, with a tariff of 30 percent on output, with low value added, and with input duties between 10 and 20 percent, effective protection would probably be high even in a tariff-only regime. The 1982 census shows two paint manufacturers in Uganda, employing a total of 50 people. Output has fallen substantially in recent years. Neither firm has access to OGL imports. A duty of 30 percent appears to have been adjusted upward to 80 percent--while duty on the dyes and pigments used to make paint appear to be about 10 percent. Water-based paint in particular has considerable local advantage in that it is usually more economical to bring in the ingredients and add water locally. Local paint production must of course have suffered with the grim demand conditions in recent years. Broad tariff reform, with maximum duty of 50 percent, would affect paint production, but effective protection would remain high. Local paint production would probably benefit if domestic demand were strong. There were three paper manufacturers in Uganda in 1983, employing about 250 people. Output has plummeted in recent years, reflecting poor demand and the reduced manufacturing base generally. We know little about the paper manufacturing process used in Uganda. If paper is imported as rolls and the process involves low value added (for example, only cutting and sticking), effective protection may well be high. On the other hand, the firms have not been dependent on OGL, except to benefit from derived demand for packing materials and paper as OGL firms expand. Duties on output are generally low, so the industry should not be affected by a general round of tariff cuts. - 7 - Plywood production is dominated by one firm (KIIRA) and output has fallen by about 7 percent in recent years. With no access to OGL and a duty of 30 percent, plywood production seems to be more dependent on demand conditions than on the tariff. Shoes. There are four manufacturers of footwear in Uganda--an increase of one firm in 1988. Even so, the number of people employed has been stable at about 280. Output has fallen by at least 25 percent. We were told, for example, that BATA (the largest company) was hurt by the import of finished footwear under SIP provisions. At 30 percent, the duty would not be affected by most general programs of tariff reform. Soft drinks benefit from the OGL provision. There are now five soft drink firms, up from three in 1986. And the number of people employed increased by 50 percent between 1987 and 1988. At 120 percent, this is one of the few high tariffs, and some, but not all, have access to OGL, so this could be a problem industry. Incentives would be reduced in any further rounds of liberalization, and the number of firms and people involved is substantial. On the bright side, new firms have apparently entered the industry, and not all of them have access to the OGL scheme. Soft drink manufacturing is normally done locally as freight costs are large. We would be surprised, therefore, if the industry did not continue to operate and satisfy local demand, particularly if inputs such as bottles are available. Steel is processed by two firms that make galvanized iron sheets. Steel ingots are also made in Uganda. Output of ingots appears to fluctuate; it is down from high years in 1983 and 1985 but up from 1986 and 1987. Galvanized iron sheet production is well down from 1985 levels. -8- Production of twine and cord is up from 1985. Available statistics do not tell us how many firms or employees are involved. We know that the firms do not have access to OGL arrangements and produce under a tariff of 20 percent. So firms in this industry are unlikely to suffer in a general round of tariff reform.