Burundi - Manufacturing industry : performance, policies and prospects

Report No. 4708-BU Burundi F Manufacturing Industry: Performance, Policies and Prospects May 15, 1984 Eastern Africa Projects Department Industrial Development and Finance Division FOR OFFICIAL USE ONLY Document of the World Bank This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  Currency Equivalents Currency Unit The Burundi Franc (FBu) Exchange Rates Through February 1973 - US$1.00 = FBu 87.5 March 1973 - May 2, 1976 - US$1.00 = FBu 78.75 May 3, 1976 - November 23, 1983 - US$1.00 = FBu 90.0 Since November 1983 - SDR1.00 = FBu 122.7 US$1.00 = FBu 118.4 Fiscal year January 1 - December 31  FOR OFFICIAL USE ONLY PREFACE This report is an assessment of the performance of the manufacturing sector and of the program envisaged by the Government to accelerate the pace of industrialization in Burundi. It is based on the findings of a mission which visited Burundi in March 1983. The members of the mission were: Messrs. Chuong N. Phung (Mission Chief), Henk Koppen (Industrial Economist, Consultant), Fernan Ibanez (Industrial Projects Specialist, Consultant) and Ms. Marie-1161ane Desgranges (Statistics, Consultant). The work of the mission was greatly facilitated by the support provided by all institutions involved in industrial development in Burundi. The report deals essentially with the current situation and medium-term outlook for manufacturing during the Fourth Plan period (1983-87). A key factor behind the performance and competitiveness of the sector is the Government's industrial policies. Accordingly, the report puts the emphasis on the administrative and policy framework with a view to suggesting changes for a sound development of the sector. An earlier draft of this report was discussed with the Government in February 1984. The comments received were incorporated in this final version. This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.   BURUNDI Manufacturing Industry: Performance, Policies and Prospects Table of Content Page No. Glossary ............................... * .................. ...... SUMMARY and CONCLUSIONS ................. ....................... i - viii I. THE SETTING Economic Structure ...... ..........* 1 Major Constraints .. ...................... ...... 2 Experience with Industrial Planning ........................ 3 Recent Economic Developments ........... ................... 5 Investment Climate .... 0.... * ................. ............... 7 II. PERFORMANCE AND ISSUES IN THE MANUFACTURING SECTOR Industrial Structure ..............*............. * ..* .....* .. 8 Recent Sector Developments ....... ................ 13 Execution of the Third Development Plan .................... 14 Issues in the Manufacturing Sector ............... O.*......... 18 (a) Unused Capacity . .......................... 18 (b) Dependence on Imports ... ........ .......... 18 (c) Role of Government in Manufacturing ................. 19 (d) Sector Effectiveness ................... ....... . .. .. 20 III. POLICY FRAMEWORK AND INDUSTRIAL FINANCE A. Industrial Policies * * * * * * *...... ,, .... 22 1. Description and Implementation ............ 22 Import Licensing ...... .. . ............. 22 Import Tariffs . .......... ...... , , 23 Price Control , ................... ....... 23 - Investment Code . ............... ...... 24 - Foreign Exchange Regulations ................... 26 - Company Tax .............. .. ...... , , 26 Wage and Salary Policies ........... ......... 26 Exchange Rate Policy ...... ... .. .*..* .......... 30 Export Tax ..................................t. 31 2. Impact of Industrial Policies .... ....................... 31  -2- Page No. B. Industrial Finance ............. ............. 34 1. Description of the Financial Sector . . 34 2. Interest Rates ............................... 35 3. Credit Regulation and Availability of Finance for Industry ........................ 36 C. Industrial Promotion Efforts ........................... 38 1. Past Efforts and Present Situation .............. 38 2. Issues and Recommendations . . ....................... 39 IV. REGIONAL ARRANGEMENTS A. Existing Arrangements...... 40 1. The Economic Community of the Great Lakes Countries ................................. . 40 -The Executive Secretary ........................... 40 - BDEGL ........ .. 41 - Trade and Customs Cooperation Agreement . . 42 2. Other Regional Arrangements .. .42 - The Kagera Basin Organization (KBO) ............ 42 - The Preferential Trade Area (PTA) ....... ....... 43 B. Effectiveness of Existing Arrangements ....... .... 43 V. PROSPECTS AND STRATEGY FOR INDUSTRIAL DEVELOPMENT A. The Fourth Development Plan (1983-87) ............. 49 The Proposed Investment Program ............. .... 50 B. Elements of a Medium-Term Industrial Strategy .... .-... 55 1. Role of Industry ...........55 2. Improvement of Existing Firms' Level of Operation... 56 3. Role of Government ................................. 57 4. Promotion of Small-Scale Manufacturing Enterprises.. 58 5* Industrial Policies . . ................. 59 Import Restrictions .................. . ..... . 59 Incentives for Employment ........................ 60 Investment Code .............. ............ 60 Price Control .............. ................ 61 Future Studies .............. ................ 61  -3- List of Text Tables Table 1: Main Industrial Products, 1949 and 1960. Table 2: Structure and Growth of GDP, 1978-82. Table 3: Main Characteristics of the Manufacturing Sector, 1979-80. Table 4: Economic and Financial Characteristics of Selected Manufacturing Enterprises, 1981. Table 5: Level of Formal Education in Manufacturing, 1980. Table 6: Execution of Industrial Projects under the Third Plan. Table 7: Financing of Industrial Investments, 1978-1982. Table 8: Value Added per Worker in Manufacturing: Selected Countries. Table 9: Benefits Granted Under the Investment Code, 1980-82. Table 10: Minimum Salary by Level of Skills as of May, 1982. Table 11: Monthly Salary of Skilled Labor by Sector, 1980. Table 12: Changes in Minimum Wage, 1977-82. Table 13: Average Monthly Earnings per Employee in Manufacturing in Selected Countries, 1980. Table 14: Exchange Rate Movements of Burundi Franc vis-&-vis currencies of Burundi's Major Trading Partners, 1977-82. Table 15: Some Results of a Survey of Enterprises, 1978. Table 16: Credit Outstanding by Branch of Activity, 1978-82. Table 17: Share of Selected African Countries in Burundi's External Trade, 1978-82. Table 18: Composition of Burundi's Recorded Manufactured Exports, 1977-82. Table 19: Tariff Rates on Selected Manufactured Goods in Burundi and Rwanda. Table 20: Investment and Employment Creation in Manufacturing During the Third and Fourth Plans. Table 21: Largest Industrial Projects of the Fourth Plan. ANNEX 1 : Note on Measurement of Protection in Burundi. STATISTICAL APPENDIX   GLOSSARY BBA Banque Belgo-Africaine BanCoBu Banque Commerciale du Burundi BCB Banque de Credit de Bujumbura BDEGL Banque de Developpement des Etats des Grands Lacs BRB Banque de la Republique du Burundi (Central Bank) BNDE Banque Nationale pour le D6veloppement Economique (Development Bank o:E Burundi) BRD Banque Rwandaise de Digveloppement (Development Bank of Rwanda) BRARUDI Brasserie de Burundi BBC Brown Boveri and Company CADEBU Caisse d'Epargne de Burundi CAMOFI Caisse de Mobilisation et de Financement CPI Centre de Promotion Industrielle CEPGL Communaut6 Economique des Pays des Grands Lacs COTEBU Complexe Textile du Burundi EAC East African Community FPE Fonds de Promotion Economique INSS Institut National de Securite Sociale IRAZ Institut de Recherche Agricole et Zootechnique KBO Kagera Basin Organization NIC National Investment Committee CPP Centre des Cheques Postaux OCIBU Office des Cultures Industrielles de Burundi EGL Energie des Grands Lacs PTA Preferential Trade Area SBF Societe Burundaise de Financement SOCABU Societe d'assurances du Burundi SOFIDE Societe Financiare de Developpement (Development Bank of Zaire) SOMEBU Societe Mixte d'Etudes du Burundi TIC Technical Investment Committee VERRUNDI Verrerie du Burundi   SUMMARY AND CONCLUSIONS i. Burundi is about to embark on the! execution of its Fourth Development Plan (1983-87) which accords priority to the development of manufacturing industries. With a diminishing supply of arable land per capita and little mineral resources, the Government feels it has to accelerate the pace of industrialization to provide employment for the rapidly growing population. Thus, the Burundian planners propose that FBu 18 billion (US$200 million), representing 18% of the country's GDP in 1981, be invested in manufacturing over the next five years and project that, by 1987, 6,300 new jobs would be created in the sector. According to the Plan, industry should be able to expand at a real rate of 17% per year. Meeting these goals constitutes a formidable challenge for a sector which employed about 7,000 people and had a total value added of US$85 million and exports of less than US$2.5 million in 1982. ii. This report assesses the situation and prospects of manufacturing in Burundi. In assessing performance, the report puts the emphasis on the impact of industrial policies which have received little attention so far. Yet, these influence the use of existing resources, affect the relative price of capital and labor, and determine the sector's effectiveness. A main conclusion of the report is that changes in the policy framework are needed to increase the contribution of manufacturing to the Government's development objectives. Looking at the future, it also appears that Burundi will face a difficult financial situation. Resources available for new investments are not likely to be as abundant as in the recent past. Thus, another recurrent theme of this report is that in view of the high level of industrial investment in the past and the still limited absorptive capacity of the sector, the Government should concentrate efforts on consolidating past achievements. Existing firms operate, on average, at less than 50% capacity, with some as low as 15-20%. Improving the operation of existing activities should therefore have the highest priority. New investments should be de-emphasized and every effort should be made to improve their quality. Although the long-term prospects of the Burundian economy depend to a large extent on a strong and efficient industry, the physical and human constraints faced by the sector also imply that the pace of industrial development should not be forced. The Setting iii. Burundi is a small, densely populated and land-locked country where over 4 million people live on a hilly territory of about 27,800 km2. GNP per capita, estimated at US$255 in 1982 is among the lowest in the world. The population is overwhelmingly rural (95%) and derives its livelihood from subsistence agriculture and coffee cultivation. Despite the predominance of the rural economy (60% of GDP and 95% of exports), Burundi has a relatively developed services sector and a small but active manufacturing sector which the Government is eager to promote as agricultural development is limited by soil detWerioration and erosion.  - ii - In its efforts to accelerate industrial growth, the Government _ces many constraints: (a) the quality of the industrial labor force is low and there is an acute shortage of managers; (b) Burundi's small market and the low purchasing power of its population limit investment opportunities; and (c) Burundi's land-locked position and difficult transport conditions increase prices of imported inputs and often result in supply interruption, forcing firms to hold large stocks of inputs and spare parts, thus increasing costs. v. Despite this difficult environment, the country does offer certain advantages for industrial development and it has recently attracted some foreign investors of Asian origin from neighboring countries. Basic infrastructure (electric power and water) is available in the major urban centers at reasonable costs and a telephone system connects the capital with the United States, Europe and neighboring countries. Burundi's land-locked position constitutes a natural protection for import substitution projects. Finally, domestic finance is readily available to manufacturers through financial intermediaries engaged in term lending to industry. Government Policies vi. The main characteristic of Burundi's incentives system is the blanket protection it provides to local manufacturers: (a) The import licensing system protects the local manufacturer against foreign competition. All imports require a license from the Central Bank and this is not granted if there is a local manufacturer able to satisfy the domestic market. If local demand is not met by local production, an import license is issued but only to the extent necessary to fill the gap; and (b) The industrial licensing system protects the existing manufacturer against excessive local competition. A license issued by the Ministry of Commerce and Industry is required to establish an industrial firm and is generally not granted if an existing enterprise is able to supply the domestic market. vii. In such a protective environment, the Government controls prices to prevent producers and traders from making excessive profits. All goods require price approval by the Ministry of Commerce and Industry. Prices are set on a "cost-plus" basis, with manufactures receiving a net profit  - iii - margin of 10% to 20%. Gross wholesale and retail mark-ups are also set for imported products and vary between 15% and 30%. The measure has probably had some success in reducing profit margins in a few cases. However, as the instructions given to manufacturers do not specify at what level of capacity utilization the prices should be calculated, the system actually allows firms operating at very low capacity to pass on all the costs and ensures them a safe and adequate return. viii. As a means of protecting domestic industry, tariffs are of little relevance as long as imports of competing products are strictly controlled or prohibited. They, nevertheless, serve to alleviate the impact of any remaining foreign competition as imports competing with domestic products (i.e., those which are allowed in Burundi because the local manufacturer cannot satisfy the domestic demand) have high import duties, generally in the range of 50%-150%. In addition, the tariff structure of low duties on imports for industry and for machinery and eq[uipment tends to make import dependent processes relatively attractive and results in a bias against the intensive use of labor. ix. The impact of the Investment Code on investment decisions is probably overrated. The financial incentives offered are not in- considerable, but there is some doubt as to how real they are, given the monopolistic market structure and the "cost-plus" price control system. Equally, the benefit of a tax holiday may be more apparent than real in view of the relatively short period (five years) for which it is actually granted. Interest rates, on the other hand, are low and generally negative in real terms. x. With regard to the exchange rate, the sharp appreciation of the US dollar, to which the Burundi Franc was pegged until recently, had an unsteadying effect on industry over the last two to three years. Imports became cheaper but the effective banning of competing imports did not encourage Burundi's import substituting industries to pass along to consumers the lower prices occasioned by exchange rate fluctuations. These industries until very recently thus reaped the benefits of appreciation. Exporting industries had no such opportunity or protection against exchange rate changes. The substantial devaluation of the Burundi Franc (30%) and the decision to link it to the SDR taken last November 1983 reduced this possibility of super profit for import substituting industries. Although manufactured exports from Burundi are currently insignificant, the new and more realistic exchange rate should in the long run help shift resources toward export activities.  - iv - xi. The Government fixes minimum wages and salaries for all levels of skill in the public and private sectors. Currently, the minimum levels are US$30/month for unskilled labor, US$51/month for skilled personnel and US$253/month for professionals. While unskilled workers generally receive the minimum wage, qualified labor is generally paid more, particularly in the private sector. Burundian monthly earnings are higher than in some other African countries (e.g. Malawi and Mauritius) and considerably higher than in Asian countries. In addition, employers are required to pay a tax on the wage bill. This payroll tax is progressive, starting at 5% if a firm's total wage bill is up to US$33,800 and rising to 30% for wage and salary payments in excess of US$59,100. Performance of the Manufacturing Sector xii. During the period 1978-82, manufacturing expanded at an average annual rate of 13% in real terms to reach 8% of GDP in 1982. This high growth yielded little structural change. Production remained typical of the first generation of import substitution, with exports limited to processed coffee, tea, cotton and a few other items. Capacity utilization was low because of the difficulty of finding equipment adapted to the small size of the domestic market and because of high tariff protection in neighboring countries. Many firms have the capacity to export (some were even designed to serve the Rwanda and Eastern Zaire markets) but produce only for the small domestic market because the present incentives system does not encourage them to look for foreign markets. xiii. Except for a few firms, the manufacturing sector is inefficient. Despite the natural protection and high tariffs, many firms need prohi- bition of competing products to operate. Their production costs are high, in many cases exceeding the cif price of comparable imports, but the quality of the products is generally well below international standards. However, the sector is profitable, with profits before tax averaging 20% of sales. This is due to the monopolistic situation enjoyed by most firms and the "cost-plus" system of price control. xiv. With the exception of coffee, cotton and tea processing, manufacturing industries are based on imported raw materials. Imports for industry (raw materials, spare parts and equipment) accounted for an estimated 35% of total merchandise imports in 1981. Available data do not permit a calculation of the net foreign exchange savings to the country resulting from local manufacturing production, but these are probably small and are produced at a high cost of domestic resources. The past indiscriminate import substitution strategy should not continue as Burundi's already delicate balance of payments could further deteriorate and require the establishment of mechanisms to allocate scarce foreign exchange. Burundi needs to be more selective in the choice of import substitution projects and make a special effort to promote exports.  - v Government's Industrial Investments xv. Since the mid-1970s, the Government has emerged as a dominant force in the development of manufacturing, participating in a number of joint ventures with foreign interests and establishing public enterprises to manage large industrial operations that the private sector had avoided because of their complex management and large financial requirements. Most of these public enterprises have been created without an adequate financial structure and/or sufficient qualified personnel and are now a drain on the budget. The projects they manage are also often of doubtful viability. In view of these disappointing results, the Fourth Plan's stated objective is to give the private sector a greater role in manufacturing. It further recommends to make existing state-run manufacturing operations more efficient, to sell off enterprises to private investors where possible, and to phase out those that are unlikely to become viable. xvi. Despite the Plan's industrial strategy, the proposed industrial investment program shows that the Government's presence in manufacturing would remain strong. Only 23% of the projected total investment is private, while 38% will be either fully or majority-owned by the Government. The remaining 39% has a public participation ranging from 25% to 50%. Equally disturbing is the fact that six projects in which Government has an important participation, have a total investment cost of FBu 13.6 billion (US$151 million, or 60% of the proposed investment program in industry), despite the fact that Burundi does not have the management capabilities to operate large industrial concerns. To be credible and attract foreign donors interested in the development of Burundi, the Plan should present a consistent picture and propose an investment program which reflects its objectives. xvii. The proposed investment program should be carefully re-examined and tailored to better reflect conditions in Burundi. First, it will be difficult to find the necessary financing. The program is too ambitious: US$200 million in 1981 prices, or the equival(ent of 4% of GDP per year while during the Third Plan (1978-82) less than 2% of GDP was invested in manufacturing annually. Second, the number of projects is too large: 62 compared to 33 executed over the past five years. It is doubtful that existing institutions would be able to handle this many projects, some of which have sophisticated technological and managerial requirements. Third, about 45% of the projects do not yet have feasLbility studies and should not be launched unless their economic and financial viability have been established; among those which have, a few are losing propositions and should be redesigned or dropped. Finally, the average investment cost per job of US$35,500 is very high and will not help solve the country's unemployment problem. It is worth noting that although the average investment cost per job created in manufacturing between 1977 and 1982 was about US$40,000, jobs created in projects promoted by private sector cost US$25,000.  - vi - Elements of a Medium-Term Industrial Strategy xviii. The role of Industry. Prospects for the Burundi economy remain closely linked to the outlook of the agricultural sector. Food production and the output and price of coffee will determine to a large extent the well-being of the Burundian people for the foreseeable future. In that context, an objective of industry should be to support the development and modernization of agriculture. Local production of small agricultural implements, fertilizers, seed drillers, grain crushers should be given high priority. In addition, basic consumer goods and those using local raw materials such as processing of vegetables, preserves, fruit juices, furniture and fixtures in wood and bamboo, pottery, manufacture of rope, cordage and related products from jute and cotton, etc., are types of industries which appear economically viable given the country's resource availability. xix. With this role assigned to industry, a selective import substitution strategy should be pursued. However, because of the large unused capacity in the manufacturing sector Burundi will need to promote exports. In this strategy, priority should also be given to the develop- ment of small and medium enterprises because of their potential as a source of employment creation. A study on the SSE sector is currently being undertaken by the Government with IDA financing and is expected to be completed in late 1984. This study will serve as a basis for the formulation of a financial and technical assistance program to small enterprises. xx. Proposed Action Program. In the short-term, apart from the need to eliminate uneconomic projects from the current investment plan, the most important tasks for Burundi are to (i) improve the country's capacity to prepare industrial projects and (ii) increase the capacity utilization of existing firms. In 1981, the Government created the Industrial Promotion Center (CPI) and entrusted it with the task of preparing small and medium industrial projects. The center has recently become operational with technical assistance from UNIDO, but its mandate is too broad and covers everything from project preparation and assistance to enterprises to industrial research and management of industrial estates. It is recommended that CPI's terms of reference be redesigned: CPI's main task should be to assist private promoters in the preparation of their projects. Specifically, CPI's activities should cover the following areas: (i) Identification and preparation of small and medium industrial projects for private promoters in coordination with banks and Ministries involved in industrial development;  - vii - (ii) Technical advice to entrepreneurs in the selection and procurement of equipment and in the preparation of loan requests to local financial institutions; (iii) Assistance to entrepreneurs in obtaining the necessary administrative clearings and authorizations, and during the implementation of their projects; (iv) Assistance to enterprises in difficulty in collaboration with the Ministry of Commerce and Industry; (v) Exploration of new local and foreign market outlets; and (vi) Promotion, through fairs, visits, seminars etc., of investment opportunities in Burundi. xxi. To improve the capacity utilization of existing firms, the Government may consider taking the following actions: (i) establishing "production and/or employment contracts" with selected firms whereby they will be granted financial and fiscal advantages when achieving an agreed production or employment program; (ii) helping entrepreneurs (through CPI) improve the management of their operations and reduce production costs; (iii) exploring new local and foreign markets and assisting Burundian firms in concluding production arrangements with complementary firms in neighboring countries. Such arrrangements have been successfully tried by a few firms in ZaLire. xxii. In contrast to the actions proposed above, which can be taken rapidly and are quick yielding, some longer term actions are outlined below. (a) The incentives system. The Government should change the present incentives system and lend more weight to the concept of economic efficiency. In terms of industrial policy, this entails moving away from the present indiscriminate protection of industry, exposing existing firms to foreign competition, and adopting economic criteria for evaluating large public industrial projects. Although a shock treatment has been applied with success by some countries which opted to expose their manufacturing sector to foreign competition, in the case of Burundi a gradual approach is recommended. As a first step, the Government could consider to gradually relax quantitative restrictions. This measure should be discussed with  - viii - manufacturers and implemented over an agreed period, say five years, so that an increasing proportion of domestic production is exposed to foreign competition every year. As a second step, the present tariff structure, which is uneven, should be replaced by a more uniform structure to establish a more neutral incentives system for all manufacturing firms. This would, however, require a comprehensive review of Burundi's existing tariff system. The broad policy reforms recommended above would introduce an element of competition in the domestic scene. Control over prices could be temporarily exercised on a no-objection basis, and gradually released as the trade policy reforms are implemented. This would be a less costly and more effective means of keeping prices down. (b) Export Promotion. The reform discussed above combined with the recent devaluation should go a long way in removing the anti-export bias of the present incentives system and put exporters on a more equal footing relative to import substituting industries. However, the Government can also act on two other fronts to encourage exports. On the international level, the Government should negotiate with its CEPGL (Communaute Economique des Pays des Grands Lacs) partners a reduction of tariff barriers to regain access to this regional market which was the traditional market for its industrial firms. On the national level, the Government should revive the drawback system which apparently exists in the legislation but is not known to local manufacturers. As a temporary measure to stimulate exports, the Government could also consider introducing an export subsidy based on net foreign exchange earnings and financed out of the budget. Although such a subsidy would have to be substantial to have an impact, its budgetary cost is likeky to be small. (c) Employment Policy. To encourage employment, the Government may consider the following changes: (i) increase tariffs on capital goods and abolish exemptions on them; (ii) raise real interest rates on credits for equipment and capital goods to positive levels; and (iii) abolish the payroll tax which is increasing labor costs. The revenue loss for the budget would be marginal as the tax itself is small (about US$1.1 million a year on average during 1979-81, or less than 1% of the budget), while 45% of it could be recuperated through higher profit tax (industrial firms are subject to a profit tax of 45% and, by scrapping the payroll tax, their taxable profits will increase by the amount of the payroll tax). In addition, the revenue loss would be largely compensated by the extra revenue brought by the proposed increase in duties on capital goods. xxiii. Further studies. The implementation of part of this action program will require detailed data and information that this report cannot provide. It is, therefore, recommended that the following studies be undertaken: (a) comprehensive tariff study; (b) study on the export potential of existing firms; and (c) study on the effective demand for small agricultural implements, and the feasibility of producing such equipment in Burundi at prices competitive with imports.  I. THE SETTING Economic Structure 1.01 With a GNP per capita of US$255 in 1982 (World Bank Atlas) Burundi is classified among the 25 Least Developed Countries in the world. Life expectancy is 45 years. Average daily per capita caloric intake is slightly below minimum levels. The population has been growing at a rate of 2.1% per year over the last decade and now about 4.4 million people live in a territory of 27,800 km2, making Burundi one of the most densely populated countries in Africa (150 inhabitants/km2); ninety five percent is rural and derives its livelihood from subsistence agriculture and coffee cultivation. Burundi's 200,000 urban residents mainly live in the capital city (Bujumbura), the most important business and industrial center. 1.02 Most of the economic activity in the country depends on the size of the coffee crop and the price it commands on the world markets. Cotton and tea are also grown for export, but they are still small foreign exchange earners, accounting for less than 10% of export receipts. Manufacturing exports are negligible. The most important food crops are beans, potatoes, sorghum, maize, and manioc, the main staple of the population. 1.03 Productivity in the rural sector is low because of poor soils, irregular rainfall and traditional cultivation methods, but adoption of modern agricultural techniques is hampered by the absence of practical training and the weight of tradition. However, what is most worrisome about the Burundian situation is that agricultural development tends to be more and more constrained by land availability and by the rapid degradation of agricultural soils due to high population pressure. 1/ 1.04 Despite the predominance of the rural economy (60% of GDP and employing 85% of the active population), Burundi has a relatively developed services sector and a small but fast growing modern sector. Trade has always been important (although not fully documented in official statistics) and reflects to a large extent tlhe historical ties between countries in the region and a certain complemnentarity of their economies. This border trade plays an important role in redistributing food and other goods from surplus to deficit areas to the mutual benefit of the populations concerned. Traditionally, it involved countries of the Economic Community of the Great Lakes (Communaute Economique des pays des Grands Lacs - CEPGL), with Eastern Zaire supplying Burundi and Rwanda with coffee for re-export and other foodstuffs in exchange for manufactured goods, but now includes Tanzania and Uganda as well. Industry has been growing at a fast annual real rate of 13% over the last five years and has now emerged as an important sector, accounting for 8% of GDP at factor cost. Production is typical of the first generation of import substitution: brewing, food processing, soaps, etc. Another sector of importance is public administration, which contributes only 4% to GDP but employs 36% of the total salaried labor force. 1/ For a comprehensive analysis of the problems and prospects of the agricultural sector, see Burundi - Agricultural Sector Memorandum, World Bank.  -2- Major Constraints 1.05 The country has the basic infrastructure necessary for industrialization. Electric power and water are available in the major urban centers (Bujumbura and Gitega) and a telephone system connects the capital with the United States, Europe and neighboring countries. Nevertheless, Burundian manufacturers work in a very difficult environment: (a) There is an acute shortage of skilled personnel in all the sectors of the economy. Training facilities for adults and unschooled primary leavers are inadequate. Seventy five percent of the adult population is illiterate, despite efforts to improve adult education. Enrollment in primary and secondary education (29% and 2.5%, respectively) are among the lowest in the world; (b) Burundi's small market and the low purchasing power of its population limit investment opportunities and domestic demand, although firms manufacturing light consumer goods can and do take advantage of the sizable unrecorded trade between Burundi and its neighbors; (c) Last but not least, Burundi's land-locked position and difficult transport conditions increase the prices of imported inputs and often result in interruption of crucial supplies, forcing firms to hold large stocks of inputs and spare parts. Two main transport corridors currently link Burundi to the Indian Ocean: 2/ the Bujumbura-Kigoma-Dar-es-Salaam central route by lake and rail through Tanzania (about 1,430 km), and the all-road northern route from Bujumbura to Mombasa through Rwanda, Uganda and Kenya (2,025 km). Both routes suffer from serious bottlenecks. On the northern route, some sections are in poor condition and transport is further complicated by cumbersome and uncoordinated administrative procedures in the transmitted countries for customs formalities, border charges, axle loading and vehicle licensing. On the central route, the operating conditions of the railway are inefficient since the breakup of the East African Railways Corporation while two transhipments due to the use of lake and road transports delay traffic and increase losses. These continuing difficulties have often been aggravated by other developments, such as the diversion of Zambian traffic to Dar-es-Salaam between 1975 and 1978, or the Uganda-Tanzania war in 1978-79, which forced Burundi to rely temporarily on special expensive airlifts for much of its imports. Political difficulties and border closures are also frequent. Due to the low level of traffic generated by Burundi (about 150,000 tons) compared to the total traffic using the same routes, the 2/ Burundi does not use routes to the Atlantic Ocean via Zaire or to the south via Zambia because they are as difficult, while the distances and costs involved are greater than for those to the Indian Ocean.  Government can exert little influence over transport decisions made by the countries on which it relies. The Central corridor offers Burundi the least cost alternative and all of its coffee exports is evacuated through that route. It is, however, less reliable and takes longer than the northern route (1'5 to 21 days from Bujumbura to Dar-es-Salaam normally, but the journey can take more than three months, depending on the operating conditions of the Dar-es-Salaam port or the railway). At present, about 40% of Burundi imports still comes through Mombasa. In Burundi, the road network is adequate in length (5,500 km) for the country's needs,but maintenance is difficult and costly due to heavy rains and the country's hilly terrain. Experience with Industrial Planning 1.06 Burundi's first experience with planning for industrial development dates back to 1950 when Belgium decided to plan the development of the Great Lakes region (Burundi, Rwanda and Eastern Zaire) and elaborated a 10-year Economic and Social Development Plan (1951-60) for Burundi and Rwanda which were administered as one territory. In industry, the Plan put the emphasis on the processing of agricultural products to supply the region in such basic goods as soap, palm oil, flour, sugar, cigarettes, etc. Investments were to be made by the private sector, with the role of the Government limited to providing the basic infrastructure. Because Burundi was thought to be poor in mineral resources, the Plan concentrated on its territory most of the industrial projects, while Rwanda received the bulk of resources allocated to mining exploration and development. 1.07 It was during that period that the ERuzizi hydro-electric power station was built at the outlet of Lake Kivu on the border between Rwanda and Zaire. Until recently, this station was the only source of electricity for Burundi. With energy readily available at reasonable cost, industrial firms, mainly from Belgium, established themselves in Burundi to serve the relatively large regional market of Burundi-Rwanda and Eastern Zaire. They were mainly concentrated in the food processing subsector, although textile and construction materials also attracted a few investors. Today, these firms still constitute an important part of Burundi's manufacturing sector. 1.08 The early 1960s were marked by important political changes, notably Burundi's accession to independence and the breakup of the economic union formed with Rwanda under the Belgian Administration. These developments adversely affected the performance of Burundi's manufacturing sector as most firms were created to produce for the Great Lakes region and lost the Rwanda and Eastern Zaire markets.  - 4 - Table 1: Main Industrial Products, 1949 and 1960 1949 1960 Food Processing Cotton Oil 28 tons 1,500 tons Oil Cake 100 tons 5,500 tons Beer - 227,000 hl Milk 2,000 1/day 2,500 1/day Soft Drinks - 24,000 hl Textile Cotton Ginning 1,325 tons 10,310 tons Blankets - 498,000 blankets Metal Products Cooking Utensils - 338 tons Boilers - 1,600 tons Nails - 356 tons Construction Materials Tiles, Flagstones - 18,000 m2 Fibrocement Products 632 tons Cement (grinding) - 2,164 tons Lime 1,703 tons 493 tons Chemical Products Soap 620 tons 2,987 tons Oxygen - 3,280 m3/month Source: Potentiel de D6veloppement Industriel a partir des Ressources Naturelles Dans les Pays le Moins Developpes: Burundi. UNIDO, February 1982. 1.09 Since independence, Burundi has implemented three 5-year Development Plans. All of them accorded a relatively high importance to industrial development. The First Plan covering the period 1968-72 gave priority to increasing capacity utilization of existing enterprises and emphasized the need to regain access to neighboring countries' markets. Measures were introduced to facilitate industrial exports, including a drawback system whereby duties paid on imported inputs would be reimbursed to exporting firms. In this strategy, creation of new enterprises was not encouraged and no special advantage was given to them. The Plan generally relied on the initiative of the private sector, which was to find in BNDE (Banque Nationale de Developpement Economique), the newly created development bank, the necessary resources to finance its investment needs.  The results were, however, disappointing. Over the plan period, only FBu 60 million were invested in industry per year (less than US$700,000/year). Burundian firms did not succeed in penetrating neighboring markets and their capacity utilizaLtion remained at about the same level as in 1967, i.e., less than 40%. 1.10 The Second Plan (1973-77) took a new orientation. It recommended the creation of new enterprises that use local raw materials and produce for the local market, and assigned to the Government a more active role in the industrial sector. Directed and led by the public sector, industrial investment increased rapidly during the period 1973-77, and averaged about FBu 255 million (US$2.8 million) per year. This period also saw a strengthening of the planning machinery with the effective use of foreign technical assistance. Industrial data started to be collected and partial surveys of industrial enterprises were conducted to better understand the sector and monitor the execution of projects. 1.11 The Third Plan (1978-82) pursued essentially the same policies, objectives and strategy as the Second Plan. The Government continued to play a leading role in the sector, accounting for more than two-thirds of the FBu 8.7 billion (US$97 million) invested in industry between 1977 and 1982. On the institutional side, the Industry Division of the Planning Ministry was reinforced and the investment code was revised to inter alia extend its application to public enterprises, many of which have been created to manage the Government-sponsored industrial investments. However, it soon appeared that most of these public enterprises have been created without an adequate financial structure and/or sufficient qualified staff to operate efficiently. The projects they are supposed to manage are also often of doubtful viability. In view of these disappointing results, the Government recently decided to reduce its involvement and intends to rely more on the private sector to promote the development of industry. Recent Economic Developments 1.12 The growth of the Burundi economy over the last decade was very uneven. From 1970 to 1975, real GDP increased at a low rate of 2.2% per year, about the same as population growth, mainly as a result of the slow growth of the agricultural sector. Gross fixed investment represented only 10% of GDP, while gross national savings averaged less than 2%, one of the lowest rate in the world. During this period, the economy also sustained the first oil shock and high international inflation. Despite some increases in export prices, the terms of trade deteriorated by more than 40%, resulting in income losses equivalent to 2.6% of the average 1970-75 GDP. In 1976-77, owing to good weather conditions, GDP expanded at a real rate of 7.6% per year, while the sharp increases in world coffee prices (more than 400%) permitted the country to accumulate substantial foreign exchange reserves (US$86 million at the end of 1977, equivalent to 10 months of imports).  - 6 - 1.13 The downturn which began in 1978 with a 55% decline in the country's terms of trade has not been reversed, despite Government's efforts to stimulate growth and contain the budgetary deficit. GDP at factor cost stagnated in 1979, recovered in 1980-81 to grow at an average rate of 7% per year but fell in 1982 by an estimated 1%. This erratic performance reflected to a large extent the effects of weather conditions on agricultural production, and the weight of that sector in overall GDP. The rest of the economy (secondary and tertiary sectors) grew steadily at an average annual rate of 7.3% between 1978 and 1982. Transport and communications, and energy and mining registered the highest growth, followed by industry which increased its share in the GDP from 6% to 8%. Table 2: Structure and Growth of GDP, 1978-82 (1970 prices) 1978 1982 Average Annual FBu million % FBu million % Growth rate 1978-82 Agriculture, Livestock 15,939 64 16,505 58 0.9 and Forestry Industry 1,386 6 2,226 8 12.6 Artisanat 683 3 748 3 2.3 Energy and Mining 87 - 151 1 14.8 Construction 1,255 5 1,561 5 5.8 Transport and Communications 489 2 912 3 16.9 Commerce 1,724 7 2,391 8 8.5 Administration 1,087 4 1,261 4 3.8 Other Services 2,367 9 2,797 10 4.3 GDP at Factor Cost 25,017 100 28,552 100 3.4 Source: Ministry of Planning, Burundi. N.B.: Data subject to changes as the Planning Ministry is currently revising GDP estimates for the period 1978-82. 1.14 Fiscal developments are more disquieting. As a result of the Government's efforts to raise the country's traditionally low investment rate, the Treasury incurred deficits since 1978 as budgetary savings could not keep pace with the fast increase in development outlays and had to be supplemented by advances from the central bank. Money supply expanded rapidly and, together with sharp increases in import prices, fueled inflation, which averaged 14.3% per year between 1978 and 1982. Public investments started to slow down in 1980, and the budgetary situation  - 7 - improved significantly during that year. In 1981, however, the Treasury deficit widened again to 2.8% of GDP as no tax was levied on coffee exports because of depressed market conditions, while receipts from import duties fell by 20%, reflecting the slowdown of economic activity. In 1982, the budgetary situation further deteriorated with the treasury deficit reaching 4% of GDP. 1.15 Another area of concern is the balance of payments, which has also been constantly in deficit since 1978. Over the last five years, Burundi lost almost FBu 8 billion (US$89 million) of foreign exchange, and at the end of 1982, for the first time since independence, its net reserve position turned negative. Poor export performance due to unfavorable world market conditions for coffee and high imports were the main factors behind this drastic deterioration of Burundi's external accounts. High inflows of foreign assistance during this period alleviated to a certain extent the burden for the economy. Investment Climate 1.16 Despite this difficult foreign exchange situation, Burundi still maintains a relatively liberal exchange system. Of particular interest to foreign investors is the possibility of repatriating 50% of distributed profits every year. The remaining may be repatriated after having been invested in savings bonds in Burundi, with half transferable after 2 years and the remainder after 5 years. Manufacturers can apply for foreign exchange at the central bank for payment to foreign contractors, provided the services performed are not available in Burundi. The advance import deposit scheme introduced in 1965 was abolished in March 1978. Thus, the investment climate remains good and recently some foreign investors of Asian origin from neighboring countries have settled in Burundi, bringing with them the needed capital and know how, and a good knowledge of working conditions in Africa. Many more are prospecting for investment opportunities as Burundi appears stable and safe compared to the chaotic situation prevailing in most of the region. The Government encourages manufacturers in many ways, including the guarantee of the domestic market and welcomes foreign investors to whom it grants generous fiscal advantages under its Investment Code. II. PERFORMANCE AND ISSUES IN THE MANUFACTURING SECTOR 2.01 Burundi started to collect industrial statistics in a systematic way only in the mid-1970's when more attention was given to industry and the need for information to better plan its development was felt. Since that time, surveys of manufacturing enterprises are periodically conducted by the Planning Ministry and cover about 40 to 50 firms, most of them with net assets exceeding FBu 25 million (US$210,000) and employing more than 30 people. Small enterprises are thus left out, but the surveys cover the modern sector reasonably well.  -8- 2.02 Because of Burundi's limited experience in collecting and processing industrial statistics, the data are often inconsistent and difficult to interpret. The reliability of the information is also affected by two other factors: (i) The sector is small and, except for clothing, wood furniture and a few other products, industrial branches often consist of fewer than five firms, or are dominated by a large one. Special events at the firm level, or the lack of information on a major enterprise, can distort the whole picture and undercut the value of year-to-year comparisons; and (ii) The sector is still very closely linked with commerce. Many industrial enterprises continue to have substantial commercial operations and often report them in the surveys together with their industrial activity. Efforts have been made by the Planning Ministry to separate these two activities, but it was not always possible because of the unsophisticated accounting systems used by most enterprises. The use of the OCAM accounting system in Burundi should improve the quality of the data in the future. 2.03 The latest year for which relatively comprehensive data are available is 1980. Consequently, the description of the structure of Burundi's manufacturing sector will be based on these data, complemented by information obtained by the mission. Table 3 presents the main characteristics of the sector. To minimize distortions, an average for the years 1979-80 is taken, whenever possible. Table 4 provides additional information on a number of selected firms, and is based on a questionnaire sent out by the mission. Industrial Structure 2.04 The picture which emerges from these tables is that of a sector geared toward meeting the domestic demand for basic goods. Exports are limited to processed coffee, tea, cotton and a few other products such as beer, soft drinks and cigarettes. As in many countries at the same stage of industrial development, food processing (excluding coffee and tea) dominates the sector, contributing 59% of total manufacturing value added and 17% of manufacturing employment. This branch is in turn dominated by one company, the brewery (BRARUDI), which accounts for more than two-thirds of the value added in food processing and 26% of its employment. The brewery is in fact the most important modern activity in Burundi after coffee, generating many commercial activities and financing directly about 35% of the Government ordinary (current) and extraordinary (mainly investment) budgets. Its impact on productive sectors is still very limited as most of its inputs, from malt to beer cases, are imported. The company is, however, a potential source of development for industry and agriculture. Recently, a number of projects designed to serve the needs of the brewery have been implemented (bottle caps and beer bottles), or are under execution (sugar) and if successful, they should help reduce the company's dependence on imports and increase its contribution to the economy.  Table 3: Main Characteristics of The Manufacturing Sector (1979-80 average, unless otherwise specified) As % of Total Sales As % of Total Production Cost Number of Total Value Wages Materials, Taxes Value Enterprises Sales Added Value Profit and Supplies Interest and Other Employment Added per Surveyed (FBu million) (FBu million) Added Before tax Salaries and Services Depreciation Payments Duties Costs (1980) Worker (000') (~1980) Coffee and Tea Processing 6 4,196 -530 13 4 24 56 8 4 4 4 2,237 237 Other Food Processing 11 3,882 2,269 58 46 15 75 4 1 4 1 1,154 1,966 Textiles (incl. clothing) 4 751 309 41 14 16 72 6 1 4 1 1,348 229 Wood and Wood Products 5 189 83 44 1 25 60 8 1 1 5 282 294 Mechanical Industries 8 503 190 38 11 29 61 3 2 2 3 643 295 Chemicals aind Co-struction 13 724 247 34 13 18 74 3 1 2 2 559 442 Materials Metal products (1981) 5 649 236 36 3 22 67 4 2 3 2 478 493 TOTAL MANUFACTURING 52 10,894 3,864 35 20 19 69 5 2 2 3 6,701 577 Source: EnquSte des Entreprises de Bujumbura. 1979-80. Minist6re du Plan, Service National des Etudes Statistiques, Burundi, and World Bank questionnaire sent to main manufacturing firms, February 1983.  Table 4: Economic and Financial Characteristics of Selected Mansfacrering Enterpriose, 1981 Year of An l cf total Operation Ovnership Capacity -PTdcFtion cent An Z of total nalen Import Duty or Maei Privete Utilieation Imported Labor V_ei n of cc-parable In-et-ent Major Creation Actiitry Foreign Gooernenr B-rundi (x) Inpets Cent Value Adddd Eports prodact Z Code Prhblens BRARUDI 1955 Seer and soft 59 41 _ 88 36 19 35 4 FSe SO/liter 10 years doty drinks relief aod 7 years tax holiday Rweg-rs 1973 Tea processing 93 7 - 64 19 22 20 97 - - Le wrid ma hoft prices and appreiatiee ef FEe against p0... sterling. Tore 1975 Tea processing 90 10 - n.a. 23 8 n.a. 95 - lSame as R.egera. Mlsote-ie 1981 Fleer - 100 - 15 55 6 n.a. - 50 5Yearne rta lity H-g-iihy M nrdt; redacd stility High pr-dactien cost. charges I...ffcis-t local chest. RAFINA 1952 Oil. 100 - - 30-40 4 44 46 - 100 - Sales affected by dintrihatios of oil given by U.S.A. COTEbU 1979/80 Te-tile- - 100 - 30 20 26 36 3 Ban en impo-rt of 5 yearn e-port tax Management; cottr fabrics relief; 2 years High labor ceots; tan hniiday Untl 1982 c-perStlon ice cheap inperts. RUZIZI 1962 Cotton proceaning - 100 n.a. 2 10 8 73 - - Old equipoent. maagmet. LOVINCO 1952 BIanker- 100 - - 55 45 25 40 - 57 Ipportt allowed only to the oxt-nt LOVINCO enrot satisfy the macbet ETC 1978 Cigarette- _ _ 100 30 86 4 12 25 150 50% duty relief for Competition frow 3Iyears smagglieg. 2yyeats tan holiday INAHU 1978 Pri,ting - 100 - 0.0. 30 34 35 - 10 iECARUDI 1963 Metal and WSed 100 - - n.a. 12 30 44 - - lrregularity of *aport F-rnit-re supply. HETALUIA 1952 Metal Predctes 100 - - 45 45 33 49 - 50 ALTECO 1967 Metal Sheets 100 - - n.a. 85 10 18 - 35 - Uteme-T-avhydro-E-rnndi 1963 Metal nod Plaatic 100 - - 40 70 6 29 1 - 4 years 1980 (itd. Tubes tan holiday ac-tiity) SAVONOR 1972 Soap 75 - 25 60 59 7 37 - 70 Duty relief 0 FADI 1976 Insecticides - 51 49 n.a. 64 10 32 - Inporto ailosed 3 yearn only if cspany duty relief caet nat7r,nfy lecal m arke t Haydry Industrien 1980 Matches 75 - 25 30 50 20 12 - Eon en isperrt 3re-f yearn Corpetition toting; 2 year iron osoggling. tan holiday Foot pred-ct qaliry. Fabriplaetic 1977 Plantic products - - 100 50 60 16 40 - 15 3 yearn doty Working capitol. relief; 5 years ta holiday CHANIC 1980 Oxygen and 99 - I onygen; 35 17 69 44 - -cetylene acetylene: 55 55 27 11 Source: World Bask Q.estionn-Mre.  - 11 -- 2.05 Compared to food processing (excluding coffee and tea processing), the other branches are relatively small with value added ranging from 2% to 14% of the sector's total. However, two of them (coffee and tea processing, and textiles) are very important employers, accounting for 33% and 20% of total labor force in manufacturing, respectively. This, again, is due to the presence of a large enterprise in each of these two branches: OCIBU (Office des Cultures Industrielles de Burundi) in the coffee and tea processing branch, and COTEBU (Complexe Textile du Burundi) in textiles. 2.06 The share of value added in total sales is on average about 35%, with the exception of coffee and tea processing which has a share of only 13%, their value added being mainly generate,d in agriculture. This is relatively low compared with other countries. A surprisingly high share of value added is obtained by the food processing industry, which uses simple processes and relies mainly on imported inpuits. This is due to the extraordinarily high gross profit margin of the brewery, of which more than 80% go to the Government in the form of taxes. For the remaining branches, the variations in the share of value added seem to conform to the pattern observed in other countries, although on the low side. 2.07 With the exception of the wood products and metal processing groups, which have low profits before tax of 1% and 3% of sales, respectively, manufacturing appears quite profitable. (Table 3). This is another distinguishing feature of the sector and seems at first glance rather surprising, considering the relatively low capacity utilization (on average less than 50%), the rather poor quality of most products, and the high cost of doing business in Burundi. This is explained by the highly protective environment in which industry operates, the generous tax incentives granted by the Government, and the monopolistic situation enjoyed by most firms. With a few exceptions, manufacturers do not appear to have liquidity problems, despite the need to maintain high stocks of imports due to transport difficulties. Their main problems are the poor quality of the labor force, the difficulty of repairing broken equipment, and the frequent disruptions of supplies of crucial imports. 2.08 The general education level of manufacturing labor is very low, as can be seen in Table 5 below:  - 12 - Table 5: Level of Formal Education in Manufacturing, 1980 University Graduates 1.5 Secondary Education + less than 4 years Post High School Training 1.0 High School Graduates 0.9 Some Secondary Education 10.5 Primary Education 22.7 Some Primary Education 29.6 Illiterates 33.8 Total 100.0 of which: Technical Training (all levels) (3.3) Economics and Administration (all levels) (1.0) Other (95.7) Source: La Situation de l'Emploi en 1980, Revue de Statistique du Travail, Ministare des Affaires Sociales et du Travail, Burundi. December 1981. This situation reflects to a certain extent the acute shortage of skilled personnel in Burundi; but what is more worrisome is the apparent inability of manufacturing to attract the skilled manpower available in the country. Only 8% of Burundi's skilled labor work in manufacturing, as compared to 22% in construction, 19% in commerce, 13% in agriculture and banking. Indeed, manufacturing ranks only above mining and transport services, although its GDP share is higher than these two sectors combined. 2.09 Except for coffee and tea processing, manufacturing activity is mainly concentrated in Bujumbura (the capital and former business center of the Burundi-Rwanda Union before independence), where infrastructure is more developed. Ownership is largely in private hands and includes a relatively important foreign participation, particularly in large enterprises established in the 1950s and early 1960s. More recently created firms often have important private Burundian interests. Government control is mostly found in companies processing agricultural products for export or in large undertakings. However, the Government has become more active in the sector in the past few years, intervening either directly by promoting and financing important projects or through parastatal enterprises, many of which are now in a difficult financial situation.  - 13 - 2.10 A last distinguishing characteristic of Burundi's manufacturing is its close links with commerce. In fact, many industrial activities have groTwn out of commercial operations. 3/. Some traders entered manufacturing to diversify their activities, others because their import business became less profitable as a result of protection given to substitute products manufactured locally. The Government's favorable attitude toward industry is an important factor in this passage from commerce to manufacturing. Recent Sector Developments 2.11 Over the period 1977-82, manufacturing expanded at an average rate of 13% per year in real terms. Part of this growth was due to the start of production of new enterprises, such as plastic bags, textiles, cigarettes and matches. Among firms in operation before 1977, the highest growth was recorded for soap and soft drinks, the latter more than doubling its production over the last five years. Modern beer production stagnated in 1979, but recovered in 1980-82 when it. increased at an average annual rate of 15%. With a few exceptions, other manufactures also expanded rapidly. 2.12 Despite this growth in output, capacity utilization has remained low, at less than 50%. The reasons for this poor performance include the difficulty of finding equipment adapted t:o the small size of the domestic market, the inexperience of some manufacturers in dealing with foreign suppliers, and the high tariff barriers in neighboring countries, which make official exports almost impossible. Many firms have the capacity to supply Rwanda (some were even designed to serve this market) and other neighboring countries, but only produce for the small domestic market because of the difficulties involved in trying to export and because the present incentives system does not encourage them to look for foreign markets. While import substitution activities are encouraged in many ways, including quantitative restrictions, there is no positive assistance of any kind to exporters. On the contrary, they are penalized through the imposition of an export tax and by the absence of clear provisions for drawback of duties paid on imported inputs. In addition, the cost-plus price control system which does not specify at what level of capacity utilization the prices should be calculated actually allows firms to pass on all the costs and ensures them an adequate return, even when they are operating at very low capacity. 2.13 Between 1977 and 1982, an estimated 2,360 jobs were created in manufacturing at a cost of US$41,100 per job. Two projects, COTEBU and the extension of BRARUDI, accounted for 38% and 14% of the total new employment, respectively. The other 31 projects implemented during the period created only an average of 37 jobs each. In 1982, the share of manufacturing employment in total modern employment was at about the same 3/ Out of 30 enterprises which replied to the mission's questionnaire, 10 have substantial commercial activities, ranging from second-hand clothes (friperie) to car imports.  - 14 level as in 1978 (7%), given the expansion of employment in the other sectors, particularly public administration. Manufacturing growth during the period also yielded little structural changes and the relative importance of the different industrial branches has remained unchanged. Perhaps the greatest change since 1978 occurred in the ownership of industry. In 1982, out of a sample of 62 industrial enterprises, 20 were owned by the Government, 3 were mixed enterprises (with Government's participation ranging from 41% to more than 60%), and 39 were private, while before 1978, there were only 12 public industrial enterprises and two mixed enterprises. 2.14 So far, development of manufacturing has not been a cause of concern for the environment. With a few exceptions, the types of industries in Burundi are not generally pollutant and the sector is still small. It is however growing fast and rightly the Government insists that the environmental impact of industrial projects be carefully reviewed before an industrial license is granted. Health and safety of industrial workers are also important aspects that have recently attracted the Government's attention as the working conditions in some firms are very difficult. Execution of the Third Development Plan 2.15 With FBu 8,729 million (US$97 million) invested in industry between 1978 and 1982, the sector accounted for 14% of total investment under the Third Plan and achieved one of the highest implementation rates (54% of Plan investment target). However, nearly 75% of the industrial projects completed during the period were not initially included in the Plan. They were, for the most part, identified and executed by private entrepreneurs as can be seen in the Table 5 below. Of the 30 Plan projects, only nine were completed; nine are still under execution; nine have not yet started (most of them are now included in the Fourth Plan); and three were dropped.  - 15 - Table 6: Execution of Industrial Projects under the Third Plan Total Under Not yet Investment Completed Execution Started Dropped (FBu million) Plan Projects 9 9 9 3 6,198 of which: Textile Plant (1) (2,447) Projects Identified during 1978-82 24 6 - - 2,531 of which: Brewery Extension (1) (843) Total 33 15 9 3 8,729 Source: Draft Fourth Development Plan. Burundi, November 1982. 2.16 Projects selected by the Plan had indeed a very low rate of implementation. In addition, they were of much larger size than projects identified and implemented directly by private promoters. Excluding the two largest projects in each group (COTEBU, the textile plant, in the "Plan projects group" and the extension of BRARUDI in the "private promoters group"), the first group had an average investment of US$5.2 million per project and a cost per job created of US$37,700. Comparable figures for the second group are US$815,500 per project and US$25,000 per job created. 2.17 Distribution by ownership shows that, out of the 48 projects completed or under implementation, 31 belong to the private sector, 13 are fully Government-owned and four are joint venture between the Government and private interests. Government projects represented an investment of FBu 4,550 million, or an average size per project of FBu 349 million (US$3.9 million), whereas private projects amounted to FBu 1,909 million, i.e., an average project size of FBu 61.6 million (US$684,000). 2.18 Measured by the number of projects executed, their size (i.e., use of more appropriate technology), employment creation, and speed of execution, the private sector appeared to have performed better than the public sector. However, private sponsored projects are typically of the early import substitution stage, with many of them relying entirely on imported raw materials and having short pay-bacik periods. On the other hand, projects financed by the Government. are based mainly on local raw materials, are more complex and have longer gestation periods. They are also relatively large and require sophistlicated management.  - 16 - 2.19 COTEBU is an example. The project cost FBu 2,447 million (US$27.2 million) and was financed by a soft loan from the People's Republic of China. It has now 900 workers, but at full production of nine million meters of fabric could provide employment for 1,300 people. The factory is fully integrated vertically and includes spinning, weaving, dyeing, printing, finishing and all the ancillary workshops. It uses locally grown cotton, which is of good quality. Production started in early 1980, but sales only reached 2.7 million meters of fabric in 1982. COTEBU's financial situation is very weak. In 1981, the latest year for which financial statements are available, it had a loss of FBu 169 million and a negative cash flow of FBu 20 million. Working capital requirements amounted to FBu 364 million and have been covered by a Government subsidy. COTEBU faces a number of problems, including: (a) Competition from imported synthetic fabrics which are cheaper and of more varied design; (b) Competition from import of second-hand clothing which currently amounts to 2,500-3,500 tons per year; (c) High production costs resulting mainly from a too large and inefficient staff (wages and salaries accounted for 35% of production costs in 1981) and high energy consumption (fuel and electricity cost more to COTEBU tha,n cotton fiber); and (d) Poor management. Stocks are too large and affect working capital requirements. Budgetary control is inadequate and there is no financial planning or forecasting, which is surprising for a company of COTEBU's size whose activity is subject to wide fluctuations. 2.20 COTEBU's situation and prospects were analyzed in a recent study 4/ which recommended: (i) improvements in product quality; (ii) diversification of the production; (iii) strict control of production costs at all levels; (iv) establishment of a financial forecasting system and strengthening of the cash management system; and (v) recruitment of short-term technical assistance to help implement the above recommendations. 2.21 To protect its nascent textile industry, the Government is also considering raising tariffs on finished fabrics or prohibiting imports altogether. This would, however, be a step in the wrong direction as it would not provide COTEBU with the necessary incentive to improve its production and designs. Imports of second-hand clothing present a more difficult dilemma. On the one hand, this trade is affecting the survival 4/ Joel Malkin et al: Les Entreprises Publiques au Burundi, 1982. For a detailed analysis of COTEBU's problems and measures to strengthen the company's situation, the reader may refer to this study.  - 17 - possibilities of COTEBU, and is a critical iactor in any attempt to set up clothing industries or small garment shops. Second-hand clothing represents an unfair competition to a small textile industry as it is collected free in developed countries for resale at low prices in developing countries, but with substantial profits for the traders. This business has blocked the development of textile industry in many countries. On the other hand, second hand clothing is only what the impoverished masses could afford to purchase. Welfare considerations dictate that this trade should not be cut. It should however be closely controlled and reexamined with an eye on its long-term effects. 2.22 The average implementation period for all the industrial projects completed during 1978-82 was 25 months whiich is not excessively long but this was due to the large number of medium-size projects (less than FBu 90 million, or US$1 million). Indeed, most of the larger projects (tea factory, COTEBU, fish factory, cigarettes, extension of BRARUDI, etc.) were implemented over a period ranging from 36 to 48 months. Delivery delays of imported equipment due to transport difficulties, but particularly inadequate preparatory work (many projects were undertaken without proper feasibility studies) were the main reasons behind these implementation difficulties. 2.23 On average, industrial projects were financed for 55% by local resources and for 45% by foreign loans arid equity as Table 7 below shows: Table 7: Financing of Industrial Investments, 1978-82 Number of Investment Financing Projects FBu-i1OToTn7-ff Local (%) Foreign (%) Public Sector 13 4,550 52 38 62 Mixed Enterprises 1/ 4 2,270 26 65 35 Private Sector 31 1,909 22 83 17 Total 48 8,729 100 55 45 Source: Draft Fourth Development Plan, Burundi, November 1982. 1/ Mixed enterprises are defined in Burundi as enterprises in which the Government has an equity participation of more than 25% but less than 100%.  - 18 - While the public sector used mainly foreign resources to finance its projects, the private sector succeeded in mobilizing locally most of its financing requirements. Initially, the P'lan had envisaged to finance most of the public industrial projects with local savings but had to rely more on foreign financing, because of the depression of the coffee market which prevailed all through the Third Plan period. This had not proved to be a problem as most of the large projects sponsored by the Government, such as the glass bottle project, the Mosso sugar complex or the coffee processing plant, were completed or are under execution. Issues in the Manufacturing Sector 2.24 The review of past performance shows that the manufacturing sector is confronted with a number of important issues which need to be addressed to increase its efficiency and contribution to the country's development efforts. The most important are summarized below. (a) Unused capacity 2.25 This is perhaps one of the most difficult problem facing the sector. On an average, existing firms operate at less than 50% capacity, with some as low as 15%-20%. In addition, according to the Planning Ministry, production capacity was overestimated in 90% of the projects implemented between 1977 and 1982; those already in operation use only 35% of their capacity. 5/ This poor performance is reflected in an exceptionally high fCOR, estimated by the Plan at 8.5, as compared to about 2.0 to 3.5 in most other countries. Burundi cannot afford to let this capital lie idle. Thus, rather than investing in new enterprises, every effort should be made to increase the sector's use of existing capacity. 6/ This is clearly the highest priority for the Government, particularly considering the impact that a higher capacity utilization could have on the country's employment situation. Given the small size of the domestic market, exports which are now insignificant should be encouraged in every possible way. There is also an urgent need to improve the country's limited project preparation capacity, which appears to be the main reason for the technical and economic design flaws of the past, and assist promoters in their negotiations with foreign equipment suppliers. (b) Dependence on Imports 2.26 At present, with the exception of coffee, cotton and tea processing, all the other manufacturing industries are based on imported 5/ Capacity here does not refer to full theoretical capacity. From discussion with manufacturers and officials of the Planning Ministry, it is the capacity that the manufacturer believes he can achieve, given availability of inputs. 6/ Obviously, this requires an assessment of existing firm's growth potential as the idea is not to increase capacity utilization of all existing firms but only those which are economically viable in the long-term.  - 19 -- raw materials. The weight of manufactured imports (raw materials, spare parts and equipment) is important on the balance of payments, accounting for an estimated 35%-37% of total merchandise imports. Available data do not permit an estimation of the net foreign exchange savings to the country resulting from local manufacturing production, but these are probably small, and are produced at a high cost of domestic resources. 2.27 Clearly this indiscriminate import substitution strategy cannot be pursued in the future as Burundi's already delicate financial situation could further deteriorate and require the establishment of mechanisms to allocate scarce foreign exchange. Burundi needs to be more selective in the choice of import substitution projects and make a special effort to promote exports. (c) Role of Government in Manufacturing 2.28 Since the mid-1970s, the Government has emerged as a dominant force in the development process because of its desire to accelerate the pace of development in Burundi and the conviction that the public sector was in a better position to mobilize the necessary resources and inject some dynamism into a sluggish private sector. In manufacturing, the forceful role of the Government manifested itself through participation in a number of joint ventures with foreign interests and in the establishment of public enterprises to manage large industri'al operations that the private sector has avoided so far because of their complex management and financial requirements. The performance of these public enterprises was analyzed in a recent IDA-financed study which pointed out that many of the problems were due to a lack of adequate management. The study outlined the serious financial implications for the Government of continuing to rely on public enterprises without action to improve their performance. It coincided with a drastic deterioration of the budgetary situation and has helped bring about a reassessment of the Government's role in industry. The draft Fourth Development Plan (1983-87) envisages a greater involvement of the private sector in manufacturing. It further recommends to make existing state-run manufacturing operations more efficient, to sell off enterprises to private investors where possible, and to phase out those that are unlikely to become viable in the medium-term. 2.29 A look at the proposed investment program, however, shows that the government's presence in manufacturing remains strong. Only 23% of the projected total investment are private, while 38% will be either fully or majority owned by the Government. The remaining 39% have a public participation ranging from 25% to 50%. Equally disturbing is the fact that six projects have a total investment cost of FBu 13.6 billion (US$151 million, or 60% of the proposed investment program). They all have an important Government participation, although past experience has shown that Burundi does not yet have the management capabilities to operate large industrial concerns. While inconsistencies are often unavoidable, to be credible and attract foreign donors interested in the development of Burundi, the Plan should present a consistent picture and propose an investment program which reflects its objectives.  - 20 - (d) Sector Effectiveness 2.30 Except for a few firms, the manufacturing sector is inefficient. Despite the natural protection and high tariffs, many firms need prohibition of competing products to operate. Their production costs are extremely high, in many cases exceeding the cif price of comparable imports, but the quality of the products is generally well below international standards. 2.31 It is true that the cost of doing business in Burundi is high. Firms have to hold large stocks due to frequent disruptions of supply of crucial inputs. The cost of petroleum products is prohibitive: about US$100 per barrel, including freight but exclusive of taxes, or twice what many other countries have to pay. However, many firms use electricity 7/, the tariff of which is reasonable (FBu 5-6/kwh for low voltage and FBu 7.6/kwh for high voltage), with the exception of BRARUDI and COTEBU which are the main industrial consumers of petroleum products. 2.32 Labor productivity appears low, although it is difficult to measure due to the scarcity of information. In the absence of better data, the mission attempted to compare the value added per worker in the main industrial branches in Burundi with a number of selected countries. The results are presented in Table 8 below and should be taken with caution because of distortions caused by factors such as overvaluation of currencies, data consistency and the use of different technologies. They nevertheless suggest that, with the exception of the foods, beverages and tobacco branch in which BRARUDI dominates, the value added per worker in Burundi is generally lower than in other countries. Burundian textiles and clothing has the lowest labor productivity, less than 40% that of Botswana and Zimbabwe and half that of Mauritius. The average for the sector shows that Burundi is below Botswana, a country also in its early stage of industrial development, but above Mauritius. This is due to the weight of export enterprises which are important and very labor intensive in Mauritius, the high tax component of the brewery in the value added of the foods, beverages and tobacco branch, and the fact that Burundian industries are highly protected and develop local value added at the expense of the consumer. 2.33 Burundian manufacturers face many constraints and work in a difficult environment. However, the monopoly situation that most of them enjoy is also not conducive to changes and efficiency improvements, particularly as they are assured of an adequate profit. In fact, many of the issues discussed above have their roots in the Government's policies and measures regarding import restrictions, import tariffs, exchange rate policy, investment incentives, etc. All these affect the performance and competitiveness of the manufacturing sector but have so far received little 7/ Private industry consumes about 35% of electricity in Burundi.  - 21 - Table 8: Value Added per Worker in Manufacturing: Selected Countries (US Dollars 1/) Burundi Botswana Mauritius Zimbabwe (1980) (1980/81) (1981) (1980) Foods, Beverages and Tobacco 21,800 2/ 11,300 4,500 3/ 9,500 Textiles and Clothing 2,500 6,500 5,000 6,900 Woods and Wood Products 3,300 2,300 2,800 4,700 Chemicals 4,900 6,600 7,700 15,300 Metal Products 5,500 2,600 6,000 10,400 Total Manufacturing 6,410 8,600 3,200 10,000 1/ Converted into US dollars using official exchange rates. 2/ Excluding coffee and tea processing. The brewery dominates and its tax aspects introduce a very particular bias. 3/ Excluding sugar. Source: Ministry of Planning, Burundi and varLous Industrial Sector reports, World Bank. attention from the Burundian decision-makers. Whereas the Plan is strong in setting out the Government's objectives, it does not clearly outline how they can be achieved. In its discussion of industrial policies, the Plan focuses mainly on the investment code, ignoring the effect of some of the other measures. To ensure that manufacturing industry fulfills the expanded role assigned to it by the Government, action over a much wider front may be needed. An evaluation of the most important measures affecting manufacturing is provided in Chapter III, together with an assessment of availability of finance for industry and the Government's efforts to promote the sector. Chapter IV discusses the regional arrangements concluded with neighboring countries and the need for Burundi to regain access to neighboring markets. Finally, Chapter V attempts to assess the medium term prospects of the sector and outlines the mission's major recommendations.  - 22 - III. POLICY FRAMEWORK AND INDUSTRIAL FINANCE A. Industrial Policies 1. Description and Implementation 3.01 In Burundi, the responsibility for the formulation and implementation of industrial policies is divided between four institutions. The Planning Ministry is responsible for overall industrial policy formulation and plays an essential role through the National Investment Commission (NIC) in the granting of benefits under the investment code. The central bank (Banque de la Rgpublique du Burundi - BRB) controls import licensing and all foreign exchange transactions, sets interest rates charged by financial institutions, and manages the country's exchange rate. The Finance Ministry influences industrial development through taxes and import tariffs. Finally, the Ministry of Commerce and Industry has the authority for delivering industrial licenses and controls industrial prices. Other institutions also indirectly affect industry, such as the Ministry of Social Affairs and Labor, which fixes minimum wages and salaries for different levels of skill. Recently, the Government created the Center for Industrial Promotion (CPI) and entrusted it with the task of preparing industrial projects and assisting enterprises in difficulty. Import Licensing 3.02 All imports require a license from BRB. Import licenses for goods that compete with local manufactures are only issued to the extent that the supply from domestic sources is insufficient. To estimate domestic demand in order to arrive at import requirements, BRB relies mainly on past sales data, to which a rough annual growth rate is applied. All domestic enterprises have to supply BRB with monthly production data. 3.03 The operation of the licensing system is efficiently managed by BRB and has been a powerful instrument for protecting existing manufacturing operations and, through its guarantee of the market, for the creation of new ones. However, this type of support for industry, which has no relation to efficiency, has a high cost. The removal of all foreign competition, coupled with limited domestic competition 8/ has led to the establishment of high-cost industry, while the absence of pressure to reduce costs and improve quality is likely to make them even less competitive. 8/ A license is required to establish an industrial firm and is generally not granted if an existing enterprise is able to supply the domestic market.  - 23 - Import Tariffs 3.04 Imports are subject to three different duties: an entry duty (varying from 0% to 10%), a fiscal duty (5% to 150%) and a statistical tax (a flat 3%). All are based on the CIF value, which includes a substantial transport cost element due to Burundi's landlocked position. 3.05 The tariff structure shows a familiar pattern: low duties (5%-15%) on essential foodstuffs (wheat flour, sugar) and on many inputs for agriculture and industry, including machinery and equipment; high duties on luxury consumer goods (generally 100% or more); and medium to high rates on imports competing with domestic products (i.e., those which are allowed in Burundi because the local manufacturer cannot meet the domestic demand) and on various consumer goods. The disparity in duty rates is further increased by the policy, set down in the Investment Code, of granting exemption of duty on many industrial inputs. 3.06 The average rate of duty collected on imports is 13%, and the revenue from this source accounts for about 20% of total Government revenue. During the period 1978-82, the cost of the exemptions was about FBu 1.1 billion (US$12 million), or about 9% of total revenue from import duties, which is not substantial, considering that duties exemptions under the Investment Code represent the main incentive for industrial investment. Unlike other countries, Burundi does not promote industry through establishment of industrial zones, service centers etc. No exemption of the statistical tax was granted during 1978-82. 3.07 As a means of protecting domestic industry, tariffs are of limited relevance as long as imports of competing products are prohibited. However, low duties on imported inputs do tend to make import-dependent processes relatively more attractive. Also some preference in favor of using capital rather than labor could in principle result from the low, often exempted, duties on machinery and equipment. Price Control 3.08 All imported and locally manufactured goods require price approval before they can be sold. Responsibility for price control lies with the Ministry of Commerce and Industry, and is implemented by the Department for Internal Trade. Prices are set on a 'cost-plus' basis, with manufactures receiving a negotiable net profit margin of 10% to 20%. Gross wholesale and retail mark-ups are also set for imported products, and vary between 15% and 30%. A staff of only ten persons is charged with reviewing all price submissions by traders and manufacturers, as well as with carrying out inspections. 3.09 Price control is mainly aimed at preventing producers and traders from making excessive profits in a monopolistic market. The measure has probably had some success in reducing profit margins in spite of inadequate inspection due to lack of staff. Its main drawback is that it discourages efforts towards greater efficiency. A fixed profit margin removes the incentive for manufacturers to reduce costs and become more efficient.  - 24 - While the price control officials can in principle disallow costs they consider avoidable, and thus exert some pressure on firms to cut out waste, the lack of adequate staff limits controls. Some investment opportunities might also have been lost as a result of price control because a fixed profit margin does not encourage firms to enter new and more risky fields. Because of its disincentive effect, price control is not in tune with the Government's concern about the growth of the productive sectors. 3.10 Investment Code. Burundi gives prominence to the Investment Code as an instrument of industrial policy. Established in 1967, and revised in 1979, the Code aims at encouraging investment in priority areas by offering certain privileges and guarantees. 9/ To qualify for these benefits, investments must (i) be in a priority sector as defined by the Plan; (ii) be of a minimum size, i.e., FBu 15 million (US$166,700) for new projects and FBu 10 million (US$110,000) in the case of extensions; and (iii) be considered satisfactory from a technical point of view, as well as in terms of job creation and value added, the provision of training, the impact on the balance of payments, etc. 3.11 The benefits offered include exemption of import duty on materials and equipment at the time of installation, exemption of duty on imported inputs for up to five years, and a tax holiday for up to five years. Of the various other kinds of assistance offered, the most important is the protection against competing imports which, however, the manufacturer could also obtain from the Ministry of Commerce and Industry and the central bank. In addition, firms may benefit from Government procurement, exemption of export duty, and facilities provided by industrial zones. 3.12 Projects of particular importance for the development of the country, and those established outside Bujumbura, may receive additional assistance. The former must either create a minimum number of jobs (100 for industrial enterprises, 150 for projects in agriculture or agro-industry), or consist of an investment of at least FBu 500 million (US$5.6 million) (agriculture and agro-industry) or FBu I billion (US$11 million) (other industries). Such investment may, in addition to the benefits already mentioned, enjoy a reduction in tax on profits for a further ten years, while an extension of the tax holiday to seven years is available for those located away from Bujumbura. 3.13 To obtain any of these advantages, the proposed investment is evaluated by two committees and must receive Cabinet approval. The Ministry of Planning plays a key role in the whole process. It chairs both the National Investment Committee (NIC), and the Technical Investment Committee (TIC), and is responsible for the Secretariat of the former. The Secretariat of the TIC comes under the Ministry of Commerce and Industry. The Ministry of Planning receives the original application and, once the project is approved, supervises its implementation. The TIC has some project appraisal expertise and considers the operation only from a 9/ An elaborate "Guide for Investors" provides potential entrepreneurs with a wide range of information on the Burundi economy and the institutional and administrative setting, including an outline of the Investment Code.  - 25 - technical and financial point of view, while the NIC places it in the context of the overall investment strategy, and recommends to the Cabinet the benefits to be offered. 3.14 The whole administrative review process is elaborate and time-consuming. To avoid delays in the execution of projects approved by the NIC, the application requires the promoters to submit a great amount of detail on the technical, financial, economic and legal aspects of the project. If all information is provided as requested, it takes about four months from the time of submission before the application is approved. Usually the process takes longer. 3.15 The benefits actually granted vary considerably from case to case. As Table 9 below indicates, only 26% of the applications passed by the TIC received all available benefits (though not for the maximum period). Exemption of the import duty on equipment is normally granted. The average tax holiday was only 2.5 years, compared to the five years allowed in principle. Table 9: Benefits Granted under the Investment Code 1980-82 1/ Benefits granted/refused Number of cases Exemption of duty on imported equipment only 9 Exemption of duty on imported equipment and tax holiday 11 Exemption of duty on imported equipment, tax holiday, and duty exemption for future imported inputs 9 Extension of benefits previously granted 5 Applications refused 8 Total applications 42 Source: Ministry of Planning. 1/ Applications received by National Investment Committee after approval by Technical Investment Committee. 3.16 The impact of the Code on investment decisions is probably overrated. Some of the main attractions of Burundi for investors are the country's political stability in a continent known for rapid changes, the liberal regulations with regard to the transfer of foreign exchange earnings, and the protection against imports of competing products. These are available outside the Code. The financial incentives offered by the latter are not inconsiderable, but there is some doubt as to how real they are. Given the monopolistic market structure and the 'cost-plus' system of price control, it is likely that firms paying import duties would be able  - 26 - to recover at least part of them through higher prices. Equally, the benefit of a tax holiday, may be more apparent than real in view of the relatively short period for which it is actually granted. Many firms are not very profitable in the initial years. When these limited advantages are set against the cost of going through the lengthy application process, the Code is expected to have limited success in bringing about investment that would not otherwise have occurred in the present economic environment. Its real attraction to investors is the protection that it provides against any change in the Government's present favorable policy toward industry. Although the code can also be changed, the advantages already granted are not withheld. 3.17 Foreign Exchange Regulations. The business community is allowed considerable freedom in their foreign currency transactions. The regulations specifically permit: (i) the repatriation of the capital originally invested once the project has ceased or been sold, as well as 50% of profits after tax every year. The remaining 50% can be transferred after having been invested in savings bonds in Burundi, with half transferable after two years and the remainder after five years. (ii) the remittance of debt service payments (interest and principal) on foreign loans; (iii) the transfer abroad of 60% of expatriates' earnings; and (iv) the payment for foreign goods and services required as inputs in the production process except when they are obtainable locally. 3.18 Company Tax. The tax rate on company profits is 45%, with a minimum of 1% of turnover. Standard depreciation provisions allow buildings to be written off over 20-33 years, and machinery, and equipment over 5-10 years. The same provisions apply to domestic and foreign enterprises. Wage and Salary Policies 3.19 The Government fixes minimum wagres and salaries for all levels of skill in the public and private sectors. The present range of minimum salaries dates from May 1982, when they were raised from the levels set in June 1977 (Table 10).  - 27 - Table 10: Minimum Salary by Level of Skill as of May 1, 1982 1/ Minimum Salary Housing per day per Month Allowance FBu US$ 2/ FBu US$ 2/ FBu US$ 2/ Unskilled - normal 140 1.2 3,,500 29.6 ) heavy 154 1.3 3,850 32.5 ) 600 5.1 specialized 170 1.4 4,250 35.9 ) Semi-skilled - normal 240 2.0 6,000 50.7 ) heavy 276 2.3 6,900 58.3 ) 1,000 8.4 specialized 294 2.5 7,350 62.1 ) Skilled - normal 351 3.0 8,775 74.1 ) heavy 403 3.4 10,075 85.1 ) 1,350 11.4 Highly skilled - - 15,625 132.0 1,875 15.8 Semi-Professionals - - 21,600 182.4 7,200 60.8 (Agents de Mattrise) Professionals - - 30,000 253.4 12,000 101.4 Source: Ordonnance Ministerielle of May 5, 1982. 1/ In addition, a monthly family allowance is payable of FBu 300 (US$2.5) for a wife/husband and FBu 150 (US$1.3) for each child. 2/ US$1 = FBu 118.4 While unskilled workers generally receive the minimum wage throughout the public and private sectors, qualified labor are generally paid in excess of the legal minimum, particularly in the private sector. All employees are entitled to a housing allowance, and annual salary increments of at least 2% are normally granted to reward length of service. 3.20 Skilled labor salaries in indust'ry compare unfavorably with those paid in the other sectors. With the exception of food and mechanical industries, which remunerate relatively well their higher staff, the other industrial branches mostly rank at the bottom of the scale in terms of employees' compensation (Table 11). The irelatively low salaries paid may make it more difficult for industrial firms to recruit high caliber people.  - 28 - Table 11: Monthly Salary of Skilled 'Labor by Sector, 1980 (In Burundi Francs) High Level Skilled Labor Professionals Professionals Services 24,999 Maximum 25,000-45,999 46,000 Minimum Commerce 24,999 Maximum 25,000-45,999 46,000 Minimum Food Industries 29,999 Maximum 30,000-44,999 45,000 Minimum Transport and Communications 23,999 Maximum 24,000-44,999 45,000 Minimum Mechanical Industries 24,999 Maximum :25,000-43,999 44,000 Minimum Banking and Insurance 21,999 Maximum 22,000-42,999 43,000 Minimum Energy 19,999 Maximum 20,000-39,999 40,000 Minimum Construction and Public Works 19,999 Maximum 20,000-39,999 40,000 Minimum Garages 24,999 Maximum 25,000-36,999 37,000 Minimum Chemical Industries 19,999 Maximum 20,000-34,999 35,000 Minimum Industrial Agriculture 24,999 Maximum 25,000-34,999 35,000 Minimum Textiles, Clothing and Leather 19,999 Maximum 20,000-33,999 34,000 Minimum Mining 16,999 Maximum L7,000-32,999 33,000 Minimum Construction Materials 15,999 Maximum 16,000-31,999 32,000 Minimum Non-Profit Organizations 14,999 Maximum 15,000-29,999 30,000 Minimum Administration 11,999 Maximum [2,000-24,999 25,000 Minimum Source: Ministere des Affaires Sociales et du Travail, Burundi. 3.21 In real terms the minimum wage has declined, notably in the late 1970s when the nominal wage remained unchanged wlhile inflation was high. Table 12: Changes in MiniLmum Wage, 1977-82 1977 1980 1982 Minimum Wage (FBu/day) 80 80 140 Minimum Wage Index in Current Prices (1977 100) 100 100 175 Consumer Price Index (1977 = 100) 100 177 205 Minimum Wage Index in Constant Prices (1977 100) 100 56 85 Source: Ministare des Affaires Sociales et du Travail, Burundi.  - 29 - 3.22 Employers pay a tax on the wage bill. This payroll tax is progressive, starting at 5% on a wage bil:L up to FBu 4 million (US$33,780) and rising to 30% for wage and salary payments in excess of FBu 7 million (US$59,120). Wages of employees earning Less than FBu 3,000 (US$33) a month are exempted, which, until the 1982 rise in the minimum wage, excluded wages of unskilled workers. 3.23 Some comparative information on labor rost in several countries is given in Table 13. The figures which are indicative only due to the problem of data comparability, show that ]3urundi, while ranking somewhere in the middle, relative to some other countries in Africa, has earnings per worker that are considerably higher than in some Asian countries. Table 13: Average Monthly Earnings per Employee in Manufacturing in Selected CountrLes, 1930 Monthly Earnings Index 1980 GNP/capita (US$) (Burundi = 100) (US$) Burundi 102.50 100 200 Malawi 62.50 61 230 MaurLtius 66.00 64 1,060 Kenya 145.00 1/ 141 1/ 420 Zambia 199.50 1/ 195 1/ 560 Bangladesh 27.00 26 120 Sri Lanka 36.50 36 270 India 59.00 2/ 58 2/ 240 Source: UN - Yearbook of Labour Statistics and World Bank Atlas. 1/ 1979. 2/ 1978. 3.24 With wages and salaries accounting for about 20% of production costs in industry, measures that raise the cost of labor may have some impact on employment. The minimum wage legislation is not thought to have had much effect on employment, given its decline in real terms, but the payroll tax, by making labor more expensive, may have discouraged its use and favored small firms as the tax rate rises with the size of the payroll.  - 30 - Exchange rate policy 3.25 Until recently, the Burundi franc (FBu) was pegged to the US dollar at the rate of FBu 90 = US$1.00. Thus, the exchange rate of the FBu vis-a-vis the currencies of Burundi's main trading partners reflected the fluctuations in the value of the US dollar. Between 1977 and 1980 the FBu depreciated against these currencies by 14.4% (22.2% against the main non-dollar currencies) as a result of the weakness of the dollar (Table 14). When the latter strengthened in 1981 and 1982, this not only reversed the earlier depreciation, but resulted in a net appreciation of the FBu of almost 9% over the period 1977-82 (13.2%if the dollar is excluded). The sharp appreciation of the FBu since 1980 has coincided with a rapid deterioration in the balance of payments and the external reserves position. 3.26 The exchange rate is a key determinant of the international competitiveness of manufacturing industry. The Government's past policy of linking the FBu to the dollar, and not to a basket of currencies reflecting Burundi's trading pattern, meant that price relationships between Burundi and non-dollar markets experienced considerable changes, and must have had an unsteadying effect on industry. In practice, the effective banning of competing imports had insulated Burundi's import substituting industries from much of the effects of exchange rate fluctuations, though at a high cost to the economy. Exporting industries had no such protection against exchange rate changes. The important devaluation of the Burundi Franc (30%) and the decision to link it to the SDR taken in November 1983 reduced the anti-export bias in Burundi's incentives system and should in the long-term help shift resources toward export activities.  - 31 - Table 14: Exchange Rate Movements of Burundi Franc vis-a-vis Currencies of Burundi's Major Trading Partners, 1977-82 1/ (percent) Share in 1977/80 trade Percentage change with major Weighted percentage change 1977-80 1980-82 1977-82 partners 1977-80 1980-82 1977-82 US dollar - - - 35 - - - Belgian Franc 22.3 -35.8 -21.5 20 4.5 -7.2 -4.3 Deutsche Mark 27.6 -25.3 -4.6 15 4.1 -3.8 -0.7 French Franc 16.3 -35.8 -25.4 11 1.8 -3.9 -2.8 Japanese Yen 18.4 -8.9 7.8 7 1.3 -0.6 0.5 Pound Sterling 33.3 -24.8 0.3 6 2.0 -1.5 ne.g Kenyan Shilling 11.6 -32.1 -24.2 6 0.7 -1.9 -1.5 Weighted average 14.4 -18.9 -8.8 Weighted average excluding US dollar 22.2 -29.1 -13.5 Source: Annex table 15. I/ A negative sign indicates an appreciation of the Burundi franc vis-&-vis the foreign currency. Export tax 3.27 In principle, all exports are subject to an export tax but exemptions can be granted by the Finance Ministry. For manufactured goods, it is a flat 3% on the f.o.b. value. When duties have been paid on imported inputs, exemption of the export tax can in principle be obtained. 2. Impact of Industrial Policies 3.28 Assessing the combined impact on. industry of a wide range of policies is a complex task. Given the long-term objective of creating an economically efficient manufacturing sector, questions arise as to how policies (i) encourage existing industries to improve their performance; and (ii) influence the establishment of new enterprises.  - 32 - 3.29 One way of approaching this problem is to try and measure the effect of Government intervention on the financial outcome of manufacturing operations, or in other words to measure the amount of assistance enterprises receive. Such assistance is provided in many different ways: by protecting enterprises against foreign competition (through tariffs or import restrictions) and, thus, allowing them to produce at higher costs, or by granting them fiscal advantages through direct subsidies, etc. On the other hand, certain policies may have a negative impact on a firm's results, such as an overvalued exchange rate, price controls, selective taxes etc. It is the net effect of all these influences and their relative impact on manufacturing activities that has to be measured. 3.30 An indication of the degree of assistance an activity receives is provided by a comparison of the value of its output (or better, its value added) at domestic costs and prices with that at world prices, i.e., cif prices). If this is done for a wide range of manufacturing operations, some insight is gained into the level and structure of assistance to enterprises. Such an exercise requires detailed and reliable information on the cost structure of firms as well as on comparable world prices. 3.31 In Burundi a first step toward the measurement of Government assistance to industry has been taken (Annex 1). If this exercise is to yield useful results, considerably more time will have to be spent on clarifying the data on production costs through visits to enterprises, as well as on obtaining relevant price comparisons. The latter presents a particular difficulty in Burundi because imports of goods that could compete with local products are effectively prohibited. However, in spite of these difficulties, a number of broad conclusions can be drawn with respect to the overall impact of Government intervention on the basis of the earlier review of individual policies: (i) A major drawback of the existing set of policies is that they provide little guidance toward the establishment of economically viable industries. By shutting out foreign competition from a market with strong monopolistic tendencies, the Government is creating an environment in which uneconomic enterprises could be financially viable. (ii) Once an enterprise is established it is given little incentive to improve its performance. In a situation where the scope for competition is limited by the small size of the market, and where there is already a considerable degree of natural protection because of the country's geographical position, the Government has removed the threat of foreign competition as well. Tariff protection is a better policy instrument as it introduces an element of competition and puts sone pressure on manufacturers to maintain financial discipline an.d pay attention to the quality of their products. At present such pressure comes only from smuggling. The open-ended protection provided to local industry is conducive to the emergence of inefficient, high-cost industries.  - 33 - (iii) Government intervention has a strong anti-export bias. Import controls provide a blanket protection for industries producing for the local market, but there is none for exporting industries. In fact, the latter are taxed, and, in the absence of an effective duty drawback scheme, are put at a disadvantage vis-a-vis their competitors. The sharp appreciation of the FBu over the past two years further made exporting an unattractive proposition. The recent devaluation, however, eliminated this disadvantage for exporters. (iv) Government policies are changing the relative prices of capital and labor in some ways. While the payroll tax is increasing the cost of all but unskilled labor, capital is made cheaper by exemptions of import duty on capital goods. Whether these policies induced changes in relative factor prices and had any effect on the use of labor or capital is hard to ascertain. However, they ignore the Government's concern regarding employment creation. (v) While this report focuses on the manufacturing sector, the question of how assistance to manufacturing enterprises compares to that to other activities is important. If one sector is put in a privileged position, it tends to draw resources away from others, at a potential loss to the economy. Considering sectors that have close links (agriculture, industry and commerce), it would appear that industry is generally favored relative to agriculture, but is possibly at a disadvantage vis-a-vis commerce. Though there are exceptions (e.g., wheat), many crops (notably coffee and green tea) are produced at costs that are internationally competitive, while many manufactured goods would not be able to compete without substantial protection. A comparison of commercial and industrial enterprises' earnings shows a much higher return to capital invested in the former (Table 15).  - 34 - Table 15: Some Results of a Survey of Enterprises, 1978 (in FBu million, unless otherwise specified) Commerce Industry Services Fisheries Construction Number of Enterprises (No.) 220 40 97 14 8 Equity 1,288 1,000 1,582 294 65 Sales 9,918 3,233 2,147 259 552 Total Labor Cost 401 513 519 133 237 Interest 80 48 65 5 33 Amortization 126 85 306 16 15 Gross Profit 1,166 489 650 88 29 Value Added 1,773 1,136 1,540 242 314 Number of Employees (No.) 2,368 3,179 2,079 1,626 2,192 Profit/Sales (M) 12 15 30 34 5 Profit/Equity (%) 91 49 41 30 45 Source: Ministry of Planning. B. Industrial Finance 1. Description of the Financial Sector 10/ 3.32 For a country at a still early stage of financial and economic development, Burundi has a relatively large number of financial institutions. Apart from the central bank (BRB), the financial system comprises: (i) 3 majority foreign-owned commercial banks (Banque Commerciale du Burundi - BanCoBu; Banque de Crgdit de Bujumbura - BCB - and Banque Belgo-Africaine - BBA); (ii) a development bank (Banque Nationale pour le Dgveloppement Economique - BNDE); (iii) a public sector resource mobilization and financing institution (Caisse de Mobilisation et de Financement - CAMOFI); (iv) a holding/investment company (Societe Holding Arabe Libyen Burundais); (v) a savings bank (Caisse d'Epargne de Burundi - CADEBU); (vi) the social security system (Institut National de Securit6 Sociale - INSS); (vii) an insurance company (Socifte d'assurances du 10/ For a detailed analysis of the financial sector, see Burundi Financial Sector Report, World Bank, July 1982.  - 35 - Burundi - SOCABU); and (viii) the Postal Office Checking system (Office des Cheques Postaux - CPP). CAMOFI, SOCABU and the Societe Holding were created in the late 1970's and CADEBU's previously modest role was significantly expanded in 1976/77 when it was charged with implementing the newly introduced obligatory savings scheme. Recently, the Government created two new institutions: a development financier (Societe Burundaise de Financement - SBF), and a fund for investing in and lending to financial institutions (Fonds de Promotion Economiqu.e - FPE). Total assets of the financial system amounted to FBu 34.5 billion (US$388 million) in 1980. With the exception of the commercial banks, all the other institutions are completely or, to a significant extent, Gavernment-owned. BNDE is the financial institution most involved in financing industrial projects, although CAMOFI, Societ6 Holding, and SBF can and do provide term loans and/or equity to industry. Since 1978, the three commercial banks have also become active in industrial lending with the introduction of the medium-term discountable credit ratio. 11/ 2. Interest Rates 3.33 Interest rates charged by commercial banks and BNDE, including those on loans made with foreign lines of credit, are regulated by BRB. CADEBU's interest rates are not subject to the general BRB regulation, but are set by CADEBU's board, which, however, is chaired by the Vice-Governor of the BRB. CAMOFI also is not subject to the general regulation; its rates are specified in its statutes or are determined by its Management Committee. Nevertheless, through its presenc:e on CAMOFI's board, BRB can influence the decision for those rates in principle freely determined. Finally, Societe Holding, while not specifically cited in the interest rate regulations, has been advised to follow BRB's guidelines. 3.34 Burundi's present interest rate structure is shown in Tables 7 to 10 of the Statistical Appendix. Lending rates on short-term discountable credits (less than two years) range from 6% for export credits, 7% for capital imports to 10% for imports of non essential goods. Equipment credits are charged 8.5%. Discountable medium-term (two to seven years) and long-term credits (over seven years) to industrial enterprises are 9% and 11%, respectively. Non-discountable credits are more expensive, ranging from 11% to 15%. Although the inflation rate is forecast to slow down in the future (about 10% per year as compared to more than 14% over the last five years), interest rates on discountable credits, which represent the bulk of credits distributed to the economy, are low and generally negative in real terms. They should be increased to reduce the pro-capital bias of present policies discussed in para. 3.31(iv) and encourage employment. 11/ Since 1978 commercial banks are required to allocate a minimum of 8% of their resources for medium-term discountable credit. The measure is designed to encourage the traclitionally conservative banks to participate in the country's development efforts.  - 36 - 3. Credit Regulation and Availability of Finance for Industry 3.35 Since May 1978, practically all credits extended to industrial enterprises are subject to control by the central bank as the regulation requires that credit must be submitted for prior review by BRB whenever it would cause the cumulative amount of all credits outstanding (excluding credit prefinancing exports of coffee and other export credits after shipment) to a given enterprise to exceed FBu 3 million (US$25,340). However, once this limit is reached, BRB can assign to individual enterprises global discount ceilings for short-term credit needs. These ceilings are reviewed annually and, if needed, may be readjusted. Credits granted within the global ceilings are not subject to BRB review but if banks grant credits, which would cause the ceilings to be exceeded, these credits would not be discountable. Medium- and long-term credits are also eligible for discount with prior BRB approval. In principle, long-term lending is the domain of specialized institutions such as BNDE, SBF, or the Societe Holding. However, commercial banks may be authorized on a case-by-case basis to lend long-term, but only that portion of the credit which does not exceed seven years would be eligible for discount. 3.36 The system is complex and cumbersome but does not appear to have been a constraint for manufacturing enterprises. Global ceilings assigned to individual enterprises are generally adequate and meet their short-term needs. In 1982, only a few enterprises had exceeded their ceilings and virtually no firm visited by the mission mentioned credit as their major problem. Term loans are also readily available, given the large number of financial intermediaries involved in financing industry. The evolution of credit to industry and other economic activities is given in Table 16. 3.37 The table shows that over the period 1979-82, credit to industry almost quadrupled, increasing its share in total credit outstanding to the economy from 6% in 1978 to 15% in 1982. Industry was only second to construction in the holding of term loans, accounting for 24% of all medium- and long-term credits distributed during that period. Availability of finance has not been a constraint to the development of the sector. On the contrary, during these past few years of high inflation and relatively low borrowing costs, enterprises might have sought more credits than they needed and have financed marginal projects as shown by the number of industrial operations now in difficulty. The situation may be different in the future with the difficult balance of payments situation facing Burundi and the continued high demand for credit on the part of the Government. The availability of foreign exchange at BNDE and the comfortable resource position of the newly created Great Lakes Countries Development Bank (Banque de Developpement des Etats des Grands Lacs - BDEGL - para. 4.06) should help mitigate the impact of a difficult local situation, although short-term credits may be more difficult to obtain from a financial system busy to serve the needs of the Government. At present, the liquidity situation of the manufacturing sector is satisfactory.  Table 16: Credit Outstanding by Branch of Activity, 1978-82 1/ (Year End, FBu million) 1979 1980 1981 1982 Short M. & Short M. & Short M. & Short M. & Term L.T. Total Term L.T. Total Term L.T. Total Term L.T. Total Agriculture 55.0 152.3 207.3 188.5 39.9 228.4 101.0 48.7 149.7 134.2 186.7 320.9 Industry 324.0 150.6 474.6 551.2 354.2 905.4 539.8 767.3 1,307.1 699.4 1,006.8 1,706.2 Civil Works, Construction and Public Works 493.0 466.9 959.9 215.1 816.7 1,031.8 227.0 1,499.7 1,726.7 437.8 2,129.4 2,567.2 Transports 35.6 271.9 307.5 99.7 202.2 301.9 53.7 124.3 178.0 78.7 41.3 120.0 Services 2/ 75.5 360.1 435.6 47.2 338.2 385.4 74.4 361.2 435.6 73.6 356.7 430.3 Commerce 4,898.3 155.4 5,053.7 5,165.1 254.7 5,419.8 6,901.5 228.3 7,129.8 5,211.6 218.5 5,430.1 of which: coffee (2,822.1) (-) (2,822.1) (2,776.3) (-) (2,776.3) (4,723.7) (-) (4,723.7) (3,124.9) - (3,124.9) Miscellaneous 210.5 19.9 230.4 376.2 36.4 412.6 149.3 85.2 234.5 746.1 202.6 948.7 Total 6,091.9 1,577.1 7.669.0 6,643.0 2,042.3 8,685.3 8,046.7 3,114.7 11,161.4 7,381.4 4,142.0 11.523.4 1/ Same breakdown not available for earlier years. 2/ Mainly Tourism. Source: BRB, Burundi.  - 38 - C. Industrial Promotion Efforts 1. Past Efforts and Present Situation 3.38 Until 1977-78, BNDE was the only institution in Burundi with some expertise in preparation and appraisal of small- and medium-sized industrial projects. Under the first credit to BNDE, IDA helped strengthen this institution's project appraisal unit by financing the services of an industrial expert who stayed with BNDE until the expiration of his contract in 1982. The unit is now staffed with two experts financed by the Federal Republic of Germany and Belgium, and two local economists, but its capacity to appraise and supervise projects still need strengthening. Recently, BNDE received funds from EIB to finance feasibility studies for large projects, particularly those of an industrial nature. The other financial intermediaries do not have any project preparation/appraisal capability and rely mainly on outside expertise to evaluate the viability of potential operations. The commercial banks only extend credits to their own clients whose credibility and financial situation they know. In general, financial institutions in Burundi work with collateral and emphasize secured lending. 3.39 In the late 1970s, Burundi received financial and technical assistance from UNDP and UNIDO to develop its project preparation capability. A UNIDO team posted in the Ministry of Trade and Industry helped prepare a number of industrial projects (plastic bags, flour mill, insecticides etc.) and assisted a few enterprises in their starting period (flour mill, tannery). The team also assisted in establishing the Industrial Promotion Center (CPI), which was formally created in May 1981 as an autonomous institution under the supervision (tutelle) of the Minister of Trade and Industry. CPI has a very broad mandate and can prepare industrial projects, assist enterprises as well as conduct industrial research, create training centers or establish and manage industrial estates. The Center is managed by a Board of Directors consisting of representatives of the Ministries of Trade and Industry (Chairman of the Board), Finance (vice-chairman), Planning, Agriculture, and Labor; of the Central Bank, the Chamber of Commerce, and a representative named by the financial institutions. In 1981 and 1982, the Center received subsidies from the Government totalling FBu 17.5 million. Its current budget is FBu 12 million, of wVich FBu 10 million comes from the Economic Promotion Fund to help undertake a few project studies. The center has established close relations with SBF which intends to rely on it to firm up its project pipeline. 3.40 The Center has recently started operation with technical assistance from UNIDO. It is managed by a Director-General and its Burundian staff consists of five economists, one engineer and two lawyers recruited in 1982 and early 1983.  - 39 - 2. Issues and Recommendations 3.41 The creation of CPI was a major step taken by the Government to strengthen the country's project preparation and appraisal capability; and its establishment as an autonomous institution should give it the necessary independence of action to become effective. The Government has rightly separated CPI from the Ministry of Commerce and Industry whose main functions are administration and control of industrial activities. It has thus avoided the risk too often encountered in other countries of having CPI burdened with administrative tasks and be regarded by private entrepreneurs as an arm of the Administration through which they have to go to obtain the necessary authorizations for their projects. 3.42 Admittedly, there is some overlapping of functions between CPI and the Industry Department (Ministry of Commerce and Industry), as the latter is also in charge of identifying andl preparing industrial projects. This has, reportedly, hampered the functionaing of CPI but is not, in the mission's opinion, a major issue as there is ample room for cooperation between the two institutions in this critical area. CPI's impact on project preparation depends on the quality of its staff and the ability of its management to demonstrate that the institution can carry out the tasks for which it was created. 3.43 A more important issue appears to be the broad mandate given to CPI which is supposed to cover everything from technical assistance to industrial research and management of industrial estates. Although all these functions seem well-conceived, there is a need for narrowing the scope of CPI's activities and set clear priorities to avoid dispersion of efforts. In the mission's opinion, CPI's activities should be essentially geared toward the private sector and its main task should be to assist private promoters in preparing their projects. It should, to the extent possible, avoid being involved in the preparation of important operations or assistance to large enterprises which would strain its limited capability. Specifically, CPI's tasks would cover the following areas: (i) Identification and preparation of small and medium industrial projects for private promoters in coordination with banks and Ministries involved in industrial develoDpment; (ii) Technical advice to entrepreneurs in thes selection and procurement of equipment and in the preparation of loan requests to local financial institutions; (iii) Assistance to entrepreneurs in o!btaining the necessary administrative clearings and authorizations, and during implementation of their projects; (iv) Assistance to enterprises in difficulty in collaboration with the Ministry of Commerce and Industry; (v) Exploration of new local and foreign market outlets; and  - 40 - (vi) Promotion, through fairs, visits, seminars, etc., of investment opportunities in Burundi. 3.44 CPI will also need to establish close working relations with the financial institutions involved in the financing of industry and make efforts to enhance its image with the business community as its success depends to a large extent on the perception of private promoters of the institution's ability to help them overcome administrative bottlenecks and implement their projects. IV. REGIONAL ARRANGEMENTS A. Existing Arrangements 4.01 For a small landlocked country like Burundi, good relations with neighboring countries are of utmost importance. Because of historical ties with Rwanda and Zaire, Burundi entered into a cooperation agreement with these countries during the second half of 1970s, but recently also joined other regional arrangements which include its Anglophone neighbors with whom Burundi has important economic relations. A description of existing arrangements is provided below and will be followed by an assessment of their effectiveness. 1. The Economic Community of the Great Lakes Countries 4.02 The Communaut6 Economique des Pays des Grands Lacs (CEPGL) was created in September 1976 and embraces Burundi, Rwanda and Zaire, a region of about 2.4 million km2 and 37.5 million inhabitants (15 million inhabitants, if only the Kivu region of Zaire is included). Its objectives are, first and above all, to ensure the security of the three member states and their populations and, second, to promote trade and economic integration of the region. The institutional framework established for achieving these objectives includes: an Executive Secretariat, a Regional Development Bank (BDEGL), and two specialized agencies: the Institute of Agronomic and Zootechnical Research (IRAZ) and the Organization for Energy (EGL). 4.03 The Executive Secretariat located in Gisenyi, Rwanda, is run by an Executive Secretary and two Deputy Executive Secretaries from the two other countries. It has a staff of about 40 professionals transferred from the Government services of their respective countries, and two Departments dealing with (i) political and social matters, security and immigration; and (ii) economic affairs, transport and communications. The Secretariat's budget comes from member countries' contributions and amounted to about  - 41 - FBu 260 million in 1983. Most decisions, including the definition of the Community's general policy, adoption of the budget and selection of the location for a community project, are made by the Conference of the Heads of State, the supreme body of the community. The Executive Secretariat has only an execution role and prepares meetings and studies for decision by the Heads of State. 4.04 The Secretariat receives support from a number of countries including the United States, Belgium and France which contribute to its budget and provide experts and funds for carrying out feasibility studies. It has established close links with the Economic Commission for Africa which set up an agency in Gisenyi in 1977 (the Gisenyi MULPOC - Multinational Programming and Operational Center) to assist the secretariat's staff in identifying and implementing projects and programs. Recently, UNDP approved a technical assistance project aimed at strengthening the Secretariat and the Regional Development Bank and providing funds for carrying out feasibility studies of regional projects. 4.05 Although CEPGL is entrusted with promoting economic cooperation in all productive sectors, its main preoccupation has been industry. There, its role has been to finance feasibility studies requested by the member countries and to recommend the location of regional projects which, in principle, are guaranteed the market of the three countries. Among industrial projects at various stages of study or implementation are the glass bottles project in which IFC has a participation, bottle caps, pharmaceuticals, agricultural implements, ceiments, plastic products and sugar. 4.06 BDEGL, the other regional institution concerned with industrial development, was established in 1977 and began operation in Goma, Zaire in March 1980 with an authorized capital of SDR 25 million, which was increased to SDR 50 million in 1982. As of April 1983, SDR 33.4 million have been subscribed by the three member Governments and a number of national and international institutions, and about SDR 10 million paid-in. 12/ The bank is still a young institution and will need considerable technical assistance to improve its procedures and systems in project appraisal and supervision, as well as in financial management and loan administration. Officially created to finance regional projects requiring substantial financing, BDEGL can also intervene in smaller operations, which are normally the sphere of the three national development banks (BNDE in Burundi, Banque Rwandaise de D6veloppement - BRD - in Rwanda, and Societ6 Financiere de Ddveloppement - SOFIDE - in Zaire). 12/ BDEGL's capital is expressed in Units of Account (which are equivalent to the SDR) but should be paid-in by the three Governments and national institutions, half in foreign exchange and half in local currencies, and by foreign institutions and countries (ADB and Belgium) entirely in foreign exchange.  - 42 - So far, BDEGL has made one loan of about US$2 million for the Ruzizi II regional hydroelectric power project. It is negotiating the purchase of 10% of the Burundi's Government's equity participation in the glass bottle project and is considering co-financing with other institutions the rehabilitation of the Kiliba sugar complex and the Katana cement plant. Besides these regional operations, BDEGL's pipeline includes projects which are already in the three national development banks's pipeline. Aware of its limited appraisal capacity, BDEGL emphasizes co-financing and relies on the national development banks's experience to prepare viable projects. The technical assistance provided by Belgium and UNDP (for which the World Bank is the executing agency) should help develop that needed capacity, but the role that BDEGL can play in the development of the region needs to be defined more clearly. At present, the bank's main asset is its foreign exchange resources. Thus, it can help finance a few large projects, which are beyond BNDE and BRD's capacity, although such projects are a rarety as the main constraint to industrial development in Burundi and Rwanda is not the availability of finance but the lack of viable projects. Such is, however, not the case in Zaire, and this is where BDEGL may concentrate its interventions. 4.07 In 1978, the three countries also entered into a Trade and Customs Cooperation Agreement, whereby certain imports originating in the member countries do not require import licenses but only import notices. This agreement was reinforced by a Monetary Arrangement concluded the same year which stipulated that payments for these imports would be made through special clearing accounts to be settled at the end of each quarter in the convertible currency of the creditor country's choice. The effectiveness of these arrangements are assessed in para. 4.12 below. 2. Other Regional Arrangements 4.08 Burundi is a member of the Kagera Basin Organization (KBO), a regional institution common to Rwanda, Uganda and Tanzania as well, created in 1977 with UNDP assistance to accelerate the development of the Kagera river basin (60,000 km2 and 6.7 million inhabitants in 1980). The organization consists of the Commission for the Management and Development of the Kagera River Basin and the Secretariat which has three Departments: Research and Statistics; Projects, Planning and Execution; and Administration and Management. The commission is the decision body of the organization and is composed of one representative from each of the four member countries. The headquarters of KBO is in Kigali, Rwanda. 4.09 At present, KBO's highest priority is agriculture and its efforts have focussed on designing with UNDP assistance an action program centered on food production with the objective of reaching self-sufficiency as soon as possible. The program also includes provision for infrastructure to ensure transport of goods in the region and production of hydroelectric power to reduce the member countries' dependence on imported energy. Although considered a key factor in the development of the Basin, industry is not given a high priority at this stage because of the need to develop a  - 43 - common agricultural base. The importance of KBO for the Burundian industry should therefore be seen with a long-term perspective and will depend to a large extent on the successful implementation of the agricultural program and the development of economic relations and trade among member countries. 4.10 In early 1983, Burundi also joined the Preferential Trade Area (PTA), an arrangement concluded by 18 Eastern and Southern African States in late 1981 to facilitate intraregional tradle. The PTA arrangements aim not only at tariff reduction (10%-70% on selected commodities traded in the region), but also includes wide-ranging cooperation agreements, such as clearing and payment facilities, and industrial development. However, these arrangements are not yet fully operational. B. Effectiveness of Existing Arrangements 4.11 The efforts deployed to conclude or join regional arrangements show the determination of the Government to cooperate with its neighbors and to enlarge its economic horizon. The effectiveness of these arrangements is assessed below. The discussion will focus on CEPGL, which has a more structured framework, and because the other arrangements are either only marginally concerned with industry, or are of too recent creation to be evaluated. 4.12 Although the political commitment of the member Governments to regional cooperation is indisputable, CEPGL has so far not been effective in promoting regional trade and establishing a common industrial strategy for the three member countries. The Trade and Monetary Arrangements (para. 4.07) did not work well because of the large differential between the official and the parallel rates of the ZaiLre currency vis-a-vis the Rwanda Franc and the Burundi Franc. The system was mainly used by Zairian businessmen until it was suspended in 1982 because Zaire did not have the foreign exchange to settle its account with Burundi and Rwanda. 4.13 CEPGL's attempt to coordinate and promote industrial development also proved to be an impossible task, given national ambitions and interests. Sugar is an example. Burundi is going ahead with its Mosso Sugar Project for which financing from a number of Arab funds and ADB has reportedly been obtained, although the Kiliba sugar plant in Zaire, the rehabilitation of which is planned, would be sufficient to supply the whole region. Rwanda is also considering building a sugar complex of 15,000 tons capacity because the present one is too small (2,000 tons/year) for its needs. 4.14 Tariff harmonization is another example. There is no plan in the near future to reinforce cooperation in this area. Yet, such cooperation is essential to the viability of the three regional projects which are now in operation or under construction: the glass bottles project (Burundi), bottle caps (Burundi), and agricultural implements (Rwanda).  - 44 - According to recent information, without a preferential tariff it is doubtful that Burundi would be able to export its glass bottles or caps to Rwanda as the ex-factory cost of the products is already higher than imports, partly because of the appreciation of the FBu. The glass bottles project has received all the advantages reserved in the Burundi's investment code and the Government will probably take the necessary measures to secure the domestic market. But without exports, the project will not be able to operate profitably as the present demand in Burundi is well below the break-even point currently estimated at 10 million bottles. The other regional projects (pharmaceuticals, methane gas, metal boxes, plastic products, cement, etc.) are still for the most part at the pre-feasibility or market study stage. CEPGL expects that some of them would be prepared by UNIDO and financed by UNDP under its technical assistance project to the Great Lakes States. 4.15 Official trade among CEPGL countries is negligible (Table 17). Over the period 1978-82, Burundi's imports from Rwanda and Zaire represented less than 1.5% of its total imports and consisted of a few products, such as cement, sugar and vegetables (mainly from Zaire). On the export side, the amounts involved were even smaller in absolute terms, consisting mostly of rice and wood panels bought by Rwanda. Burundi exported little to Zaire, except in 1981 when sales to that country increased more than sixfold to reach FBu 225 million (US$2.5 million). Manufactured goods (mainly beer and cigarettes) accounted for 38% of the total and agricultural products (meat and beans) for the remaining 62%. The CEPGL clearing arrangement in place at that time appears to have been the main factor behind this surge of exports to Zaire since, valued at the official exchange rate, Burundian products were very competitive compared to other sources of supply. 13/ 4.16 During the same period, Burundi's main trading partners in Africa were two countries outside the CEPGL zone with a relatively developed industrial sector: Kenya and Tanzania. Trade with these countries was very unbalanced as Burundi imported manufactured goods from them, but could not offer much in exchange, except for some tea to Kenya. 4.17 The low level of official trade and the grim picture depicted above do not mean that all manufacturing firms in Burundi are restricted to their small domestic market. Those producing essential or easily transportable consumer goods such as beer, plastic products, soap, etc., can and do take advantage of the unrecorded trade between countries in the region which, although not documented, is thought to be important. 13/ Zairians importing from outside the CEPGL region should purchase foreign exchange to settle their transactions at the parallel market, the rate of which was 3 times higher than the official rate in 1981. In September 1983, Zaire devalued massively and set its currency at about US$1 = Z 30, while the official pre-devaluation rate was US$1 = Z 5.7.  - 45 - Table 17: Share of Selected African Countries in Burundi's External Trade, 1978-82 (Value in FBu million; share in percentage of total imports or exports) Jan. - Oct. 1978 1979 1980 1981 1982 Imports Exports Imports Exports Imports Exports Imports Exports Imports Exports Zaire Value 47.7 4.7 117.4 3.9 190.2 34.4 144.5 224.5 139.6 n.a. Share 0.5 0.1 0.9 0.1 1.3 0.6 1.0 3.5 0.9 n.a. Rwanda Value 20.2 71.9 22.5 50.6 20.6 81.0 29.6 79.0 17.6 n.a. Share 0.2 1.2 0.2 0.5 0.1 1.4 0.2 1.2 0.1 n.a. Total CEPGL Value 67.9 76.6 139.9 54.5 210.8 115.4 174.1 303.5 157.2 n.a. Share 0.7 1.3 1.1 0.6 1.4 2.0 1.2 4.6 1.0 n.a. Tanzania Value 362.9 - 511.0 - 423.2 - 576.2 1.5 500.9 n.a. Share 4.1 - 3.7 - 2.8 - 4.0 - 3.2 n.a. Uganda Value 0.6 - 0.2 - 2.2 - 0.4 - n.a. n.a. Share - - - - - - - - - n.a. Total KBO 1/ Value 383.7 71.9 533.7 50.6 446.0 81.0 606.2 80.5 518.5 n.a. Share 4.3 1.2 3.9 0.5 2.9 1.4 4.2 1.2 3.3 n.a. Kenya Value 586.4 - 954.5 77.1 692.2 25.9 984.8 53.2 601.1 n.a. Share 6.6 - 7.0 0.8 4.6 0.4 6.8 0.9 3.8 n.a. Total Africa Value 1,101.8 112.9 1,792.4 147.5 1,572.0 145.5 2,485.3 431.9 1,937.4 n.a. Share 12.5 1.8 13.1 1.6 26.7 2.5 17.1 6.7 12.3 n.a. I/ Including Rwanda. Source: BRB, Burundi.  - 46 - Traditionally, this trade mainly involved the CEPGL countries, with Eastern Zaire supplying Rwanda and Burundi with coffee and other agricultural products in exchange for manufactured goods, but now includes Tanzania and Uganda as well. The reasons most often advanced for the expansion of this trade include disparities in the availability of goods due to a tightening of exchange control practices in some countries, and exchange rates that do not reflect market forces. Burundian beer and soft drinks are reportedly being sold in Tanzania or exchanged for salt, groundnut oil or mattresses. Zaire exports coffee, beans, milk, palm oil, etc. to Burundi to buy fabrics, beer, pans, and second-hand clothes. However, Rwanda seems to benefit the most from this market because of lower prices due to its conservative monetary and fiscal policies and comparatively lower import costs. Moreover, Rwandese products are starting to penetrate the Burundian market. 4.18 It is generally recognized that controlling this trade is costly and difficult, if not impossible, given the existence of cultural and other links between Burundi and its neighbors, and the largely unpatrolled frontiers. Goods or funds move across borders in response to the demand or price differentials and could not be effectively controlled in the absence of direct efforts to correct the causes of the imbalances that lead to such flows. This would argue for some coordination of policies between Burundi and its immediate neighbor, Rwanda, because as long as Burundi pursues a more liberal monetary policy, the pattern of unrecorded trade which has developed between the two countries is likely to continue. Such coordination already exists in an important area of mutual interest: the setting of the producer's price for coffee which has been kept unchanged over the last five years and was identical in the two countries from 1978 to November 1983, when expressed in US dollars. It would be to the mutual benefit of both countries if this coordination could be extended to monetary and fiscal matters as pursuing divergent policies in these areas would be self defeating in the long run, given the similarities of their economies and the permeability of their borders. A revitalized CEPGL could be a forum where discussion of these matters could take place. 4.19 CEPGL should also play a more active role in promoting industrial cooperation. The importance of this cooperation is recognized by the three member countries and the need for it is evidenced by the existing duplication of production capacity within the region, particularly between Burundi and Rwanda where the industrial sector has essentially the same structure and produces the same goods. Because of high tariff barriers and the difficulties to export, some firms in Burundi or Rwanda even had to establish themselves in the other country to capture the market. So far, CEPGL has tackled this coordination task by focussing on the promotion of new projects and has devoted most of its efforts to help, with limited success, carry out feasibility studies. This action is needed and should be pursued because the region is richly endowed with natural resources which should be exploited in common. However, CEPGL may also want to address the constraints limiting regional trade and initiate discussion with a view to reducing tariff barriers among its member countries.  - 47 - 4.20 This question is of utmost importance to Burundi for which regaining access to neighboring countries' markets is essential, given the large unused production capacity of its industrial firms. At present, Burundi's exports of manufactured goods are relatively modest accounting for less than 3% of total exports in 1981 and are for the most part bought by the other CEPGL countries. Their composition is given in Table 18 below: Table 18: Composition of Burundi's Recorded Manufactured Exports, 1977-82 (FBu million) Jan-Oct 1977 1978 1979 1980 1981 1982 Fibrocement Products 32.3 37.8 35.8 66.4 51.2 39.1 Soft Drinks 0.3 0.6 2.2 2.0 0.1 0.4 Beer - - -- 4.3 9.6 62.2 Oxygen 2.2 5.5 4.3 7.7 8.3 7.0 Metal Products 3.1 3.9 2.1 1.4 - 7.2 Cigarettes - - - - 72.0 42.3 Other 10.2 33.9 10.4 19.9 29.0 29.5 Total 48.1 81.7 54.8 102.1 170.2 187.7 Source: BRB, Burundi. 4.21 Without the substantial increase in exports to Zaire in 1981 and 1982 due to the CEPGL clearing arrangements, Burundi's exports of manufactured goods would have registered a decline compared to 1980. These clearing arrangements were suspended in late 1982 (para. 4.12) and, as a result, Burundi's manufactured exports to Zaire would probably not increase much in the future. Rwanda has traditionally been Burundi's main client. Exports to that country have, however, stagnated over the last few years at about FBu 80 million (US$900,000) a year and would certainly benefit from an agreement to reduce tariffs with Rwanda. The tax revenue forgone by the Government would be marginal, given the low level of official trade, but Burundi would have to open its small market to Rwandese products. This may impinge on the Government's present import substitution policy and create additional problems for some industrial firms, but in the long run, the benefits that Burundi could derive would by far exceed the costs of opening up its market. At present, both countries maintain relatively high tariff and non-tariff barriers for the protection of local industry, with no regional preference. This protection, which was designed to allow time for  - 48 -- new firms to develop, has tended to perpetuate, resulting in relatively high-cost domestic production and discouraging trade. Table 19: Tariff Rates on Selected Manufactured Goods in Burundi and Rwanda (in percent of cif value) Burundi Rwanda Fiscal Duty Imnport Duty Fiscal Duty Import Duty Cigarettes 1/ 150 Exempted 130 20 Soap 50 3 50 20 Blankets 2/ 50 7 40 10 Plastic Utensils 30-50 5 15 exempted Mattresses 55 10 40 20 Wood Chairs 100 10 80 20 Cushions, Pillows 55 10 40 20 Brushes, Brooms 45 2 30 10 Metal Sheets 15 5 10 5 1/ Imports prohibited in Burundi since 1982. 2/ Imports in both countries allowed only to the extent the local firm cannot satisfy the market. Source: Tarif des Douanes. Burundi and Rwanda. 4.22 Another area where CEPGL countries would need to devote some attention to in the future concerns the trade and industrial policies to adopt vis-a-vis non-member countries. CEPGL countries trade more with Kenya and Tanzania than among themselves. Kenya emphasizes the promotion of exports as a matter of policy, particularly since the emergence of a manufacturing surplus capacity after the breakup of the East African Community (EAC). Tanzania presently faces a difficult economic and financial situation, but when growth resumes could become an important competitor for export markets. Should CEPGL adopt a protection policy against these non-member countries or should they be invited to join the community? The issue is further complicated by the fact that Rwanda and Burundi are members of KBO which includes Tanzania but not Zaire and Kenya. Both Rwanda and Burundi are also parties in the PTA arrangements, but not Zaire. To a large extent, the future of CEPGL would depend on the resolution of this issue.  - 49 - V. PROSPECTS AND STRATEGY FOR INDUSTRIAL DEVELOPMENT A. The Fourth Development Plan (1983-87) 5.01 Since the sharp decline in coffee prices in 1978 and in spite of a bumper coffee harvest in 1981, economic events in Burundi have taken a turn for the worse. At present, coffee :Ls no longer a source of revenue for the Government, whereas it previously accounted for approximately 25% of total budgetary revenue. On the balance of payments side, lower foreign exchange receipts and reduced foreign aid inflows in the face of continued increase of imports have created a serious current account imbalance which will be difficult to resolve in the absence of balance of payments assistance. 5.02 The Fourth Plan, drawn up when all indications are pointing to a difficult period ahead for the Burundi economy, gives a prominent role for the directly productive sectors: agriculiture and industry. While for the previous five years industry had been allocated 12% of total planned investment, its share has been increased to 20% in this new plan. The main objectives in pushing for a rapid growth of industrial production, apart from its effects on income, are the creation of jobs and a strengthening of the balance of payments. Decentralization of industry, which could spread its benefits more widely, is also to be encouraged. Also links between the agricultural and industrial sectors are seen as the key to the achievement of these aims, with industry being required to rely on local raw materials to a much greater extent than in the past. 5.03 The Plan outlines very ambitious targets for industry. It projects an average annual growth rate of 17%. New projects are expected to create 25% of all industrial value added during the plan period. This is to be achieved with the help of an investment of FBu 18 billion (US$200 million), representing 18% of the country's GDP in 1981 and twice the amount invested in industry during tlle Third Plan Period (1978-82). The new projects are expected to rely more on local inputs for their needs and would create 3,800 new jobs at an average cost of US$35,500 per job; another 2,500 jobs will be generated by existing enterprises through better capacity utilization.  - 50 - Table 20: Investment and Employment Creation in Manufacturing During The Third and Fourth Plans Third Plan Fourth Plan 1978-1982 1983-1987 (Actual) (Objectives) Investment (1981 FBu million) 8,729 17,948 Number of Jobs 2,359 6,293 1/ Number of Projects 48 62 - Completed (33) (-) - Under execution (15) (-) 1/ Of which, 2,500 new jobs resulting from a better capacity utilization of existing enterprises. Source: Draft Fourth National Development Plan - Burundi. 5.04 On the policy and institutional front, the Plan is proposing to take measures in areas such as public enterprises, investment incentives and productivity increase in existing enterprises. Concerning public enterprises, and building up on the recommendations of the recent IDA-financed report on the subject, the Plan proposes to transfer to the private sector some of the viable public projects and to liquidate those which are unlikely to become profitable in the long run. With respect to investment incentives, the Plan is consi.dering increasing the advantages presently available under the investment: code and making them more accessible to small enterprises. Finally, regarding productivity increase, the Plan recommends granting financial and fiscal advantages to encourage existing enterprises to increase the use of their fixed assets. The Proposed Investment Program 5.05 The Plan's investment program has been prepared in great detail, identifying all projects and spelling out their funding requirements as well as their expected results, including the employment and value added to be created and the balance of payments impact. However, for many of these projects no pre-feasibility or feasibility study has been prepared yet, and the figures given are thus highly tentative. 5.06 The first striking feature of this proposed investment program is the large number of projects slated for implementation during the plan period (62 compared to 33 executed over the past five years). Even if all the projects were already well defined and prepared (which they are not),  - 51 - it is doubtful that existing institutions would be able to handle this large investment program with sophisticated technological and managerial requirements and complex financing plans. 5.07 The second striking feature is the relatively large size of the projects, with an average investment of FBu 365 million (US$4.1 million) per project, resulting in an investment cost per job of US$35,500. Six projects 14/ account for more than 60% of the proposed investment, adding vulnerability to the program and contradicting the implicit objective of emphasizing small and medium size projec:ts which use technologies better adapted to Burundi's market size. 5.08 All these six projects will eiLther be owned by the Government or have a large public participation and this is another striking feature of the plan which goes against the declared policy of assigning a more important role to the private sector in industry. Admittedly, industrial investment must be viewed against the dual objective of employment creation and balance of payments alleviation. The latter type of projects are often capital-intensive and more complex and, thus, could only be promoted by the Government as Burundi's private sector appears more motivated by small projects requiring simple technologies. Elsewhere in this report, the missionI suggested that Government's parl:icipation in the sector should not be excluded because of the complexity and long gestation periods of some important projects. It should, however,, be used with caution as their managerial requirements may exceed the country's present capacity. 5.09 Ten projects are considered decisive by the Plan for achieving the objectives assigned to the industrial sector. Excluding the agricultural component of the sugar facl:ory, they represent a total investment of FBu 16,152 million (US$179 million), of which FBu 12,249 million (US$136 million) during 1983-87, 14/ They are: a sugar complex in Mosso, a new brewery in Gitega, two coffee factories for OCIBU, a cement plant, a fertilizer plant and VERRUNDI, the Glass bottle factory.  - 52 - Table 21: Largest Industrial Projects of the Fourth Plan (FBu Million) Total Cost 1983-87 Agro-Industries Sugar Complex (Mosso) 1/ 2,750 2,228 Brewery (Gitega) 2,640 1,320 BRARUDI Extension 800 800 Coffee Factories 2,450 1,723 Slaughterhouse 703 703 Distillery (Mosso) 450 450 Construction Materials Cement Plant 2,500 2,250 Wood and Paper Particle Board 579 579 Chemicals Fertilizers (Phosphates) 1,200 840 Glass VERRUNDI 2,080 1,456 TOTAL 16,152 12,349 1/ Excluding the agricultural component,. Total cost of this project is currently estimated at FBu 6,180 miLlion (US$69 million). 5.10 Two of these projects : the BIARUDI extension (150,000 hl/year) and the slaughterhouse in Bujumbura are still at an early stage of preparation and estimates of investment costs indicated in the Plan may have to be revised subsequently. The execution of the Gitega brewery project has reportedly started with a loan from a French bank. However, market studies will have to be made to assess the possibility for export as two breweries with a total capacity of 1,150,000 hl/year will be too large for the local market. Present beer consumption is about 600-700,000 hl/year and is expected to increase at t:he same rate than in the past (8%-9% per year). 5.11 Feasibility or pre-feasibility studies are planned or are being carried out for three other projects:  - 53 - (i) the distillery which at full zapacity is estimated to produce about 4 million liters of alcohol from sugar molasses and other agricultural products as a partial substitution for imported gasoline. The project is e:xpected to generate 80 new jobs at a cost of US$62,500 per job. (ii) the fertilizer plant with a capacity of 100,000 tons which would use the phosphate deposit at Matongo. Two hundred new jobs would be created at a cost of US$66,700 per job; and (iii) the cement plant using the limestone deposits under the phosphate deposit in Matongo. Demand for cement is estimated at 100,000 tons per year. The plant would have a maximum production capacity of 120,000 tons per year and would create 250 new jobs at a cost of US$111,100 per job. 5.12 The sugar complex in Mosso is already at an advanced stage of execution as a joint venture between foreign investors and the Burundi Government. Total cost of this project!, including sugar plantation and infrastructure, is estimated at about US$69 million and would be financed by the Burundi Government, a small equit:y investment from the foreign investors and loans from a number of local banks and foreign institutions, including ADB. Present sugar consumption in Burundi is estimated at about 8,000-9,000 tons per year. The project should be in operation in 1985 and will produce 6,000 tons of sugar that year, gradually reaching its maximum production capacity of 15,000 tons in 1990. The feasibility study estimated the ex-factory price at US