33691 Pilot Investment Climate Assessment Obstacles to the Expansion of Eritrea's Manufacturing Sector Final Report December 2002 Linda Cotton, Tesfay Haile, Jean Michel Marchat, Peter Miovic, John Paton, Vijaya Ramachandran, Manju Kedia Shah Regional Program on Enterprise Development Africa Private Sector Group Contents ii List of Tables and Figures 3 Executive Summary 3 Acknowledgments 8 Abbreviations and Acronyms 9 1. Economic Growth and the Private Sector 10 2. Investment climate in an International Perspective 19 3. Conclusions and Recommendations 43 References 47 Annexes 48 List of Tables and Figures iii Table 1.1: GDP by industrial origin, 1993-2001 3 Figure 2.14: Use of formal and informal Figure 1.1: Real GDP growth borrowing and lending 24 (annual percentage) 5 Figure 2.15: Details on collateral required Figure 1.2: Manufacturing sector 5 for the most recent loan 25 Table 1.2: Key economic indicators, 1993-2001 6 Figure 2.16: Types of collateral utilized 25 Table 2.1: Distribution of capital Figure 2.17: Average interest rates and to labor ratio ($US) 11 duration of the most recent Lloan (by size) 26 Table 2.2: Distribution of value-added Figure 2.18: Average interest rates and duration per worker ($US) 12 of the most recent loan (by ownership) 26 Table 2.3: Distribution of value-added Table 2.5: Investment characteristics: to capital ratio 13 Eritrea Vs. other SSA countries 27 Figure 2.1: Ratio of wages to value-added Table 2.6: Distribution of investment by in Africa 14 firm characteristics 27 Figure 2.2: Ratio of wages to value-added Figure 2.19: Access to land 28 in Asia 15 Figure 2.20: Business environment 29 Figure 2.3: Average capacity utilization Figure 2.21: Business services: by size and ownership 16 perceptions of quality 32 Figure 2.4: Average capacity utilization Figure 3.1: Eritrea investment climate by ownership 16 policy matrix 37 Figure 2.5: Ranking of obstacles Table A.1.1: Sample frame to capacity utilization 17 (by region and sector) 41 Figure 2.6. Structure of permanent employment Table A.1.2: Sample frame (by sector, 2001) 18 (by size and sector) 42 Table 2.4. Mobilization by job position in 2001 Table A.1.3: Sample frame and (percentage) 19 theoretical sample 44 Figure 2.7: Monthly sash earnings 20 Table A.1.4: Sample structure Figure 2.8: Changes in the wage rate (by sector and region) 46 (by position) 20 Table A.1.5: Sample structure Figure 2.9: Changes in the wage rate (by sector and size) 47 (by sector) 21 Figure A.1.1: Employment distribution Figure 2.10: Changes in the wage rate (by size) 48 (by size) 22 Figure A.1.2: Employment distribution Figure 2.11: Overdraft facilities by size 23 (by sector) 48 Figure 2.12: Overdraft facilities by ownership 23 Figure A.1.3: Employment distribution Figure 2.13: Loans 24 (by region) 48 List of Tables and Figures iv Table A.2.1: Highest educational achievement Table A.5.7: Financial sector ­ auditing, of employees (%) 49 transaction costs, and property rights 74 Table A.2.2: Education of workers (percentage) 50 Table A.5.8: Labor and training in international Table A.2.3: Union membership 51 comparison and by firm characteristic 75 Table A.2.4: The determinants of wages in Table A.5.9: Regulatory burden and the Eritrean manufacturing sector 52 administrative delays by country 76 Table A.3.1: Percentage of white-collar Table A.5.10: Governance ­ workers (by sector and size, 2001) 57 uncertainty and corruption 77 Table A.3.2: Average number of workers per firm 58 Table A.3.3: Changes in number of workers per firm (1999 to 2001) 59 Table A.3.4.: Monthly cash earnings (by sector, size, location, and ownership) 60 Table A.3.5. Number of permanent full-time employees by sectors and sizeclass In 2001 62 Table A.3.6: Recent change in employment by sectors, 1999 to 2001 63 Table A.3.7: Average structure of employment in 2001 (c.pct) 64 Table A.3.8: Average age of employees by sizeclass and gender 64 Table A.3.9: Monthly cash earnings of workers in Nakfa by job position in April/May 2002 65 Table A.3.10. Monthly cash earnings of unskilled production workers 66 Table A.4.1. Details on the financial sector 68 Table A.5.1: ICA survey sample structure (Eritrea) 69 Table A.5.2: Competitors, suppliers and customers 69 Table A.5.3: Respondents' evaluation of general constraints 71 Table A.5.4: Infrastructure indicators 72 Table A.5.5: Sources of finance 72 Table A.5.6: Credits, loans and liabilities 73 Executive Summary v The Eritrea Investment Climate Assessment (ICA) is A perennial issue in the debate on whether based primarily on a survey conducted by the Africa can be globally competitive is that of the Regional Program on Enterprise Development (Africa competitiveness of African labor. Unit labor cost Private Sector Unit) in Spring 2002. The team covered measures the total cost per unit of output in a common manufacturing firms representing about 65 percent of currency, which enables international comparisons of total employment in manufacturing. The survey was competitiveness of labor. Because it is very difficult to designed to examine the cost of doing business and use physical measures of output across different to generate cross-country comparisons of countries, we use an approximate measure of unit competitiveness and the investment climate. labor cost which is the ratio of wages to value added The recent efforts of the Government of Eritrea to at the firm level, averaged across the sample of firms. promote private sector­led growth are grounded in a It is clear that while Eritrea's unit labor cost (0.38) is tradition of entrepreneurial spirit that has survived periods within the range of sub-Saharan African countries, it is of conflict and other shocks. Eritrea's small economy has much higher than that of several Asian countries a GDP estimated in 2002 at $US 644 million. In 2001, during periods of rapid economic growth. industry as a whole accounted for 22.3 percent (only 7.8 percent of which was manufacturing). The government The Dynamics of Factor Markets has recognized economic diversification, especially The Labor Market. Manufacturing employment has export diversification, as a key requirement for achieving been declining and wages are rising sharply, sustainable, broad-based growth in the new century. The presumably reflecting the impact of military loss of the Ethiopian market and rising debt means that mobilization In Eritrea, daily wages, overtime, bonus increasing exports is very important. Considerable and other benefits for skilled production labor are attention has been given to increasing comparative about $2.63 per day while unskilled workers receive advantage in labor-intensive manufacturing for export about $1.86 per day. While these numbers are not because of the potential for dramatic growth based on astronomically high, they are rising. experience in other developing countries. The survey found that about 15 percent of the white collar workers went on military duty in 2001. Cross-country Comparisons of This figure comes at the end of the mobilization Competitiveness process that has been taking place over the last Firms in Eritrea are more capital intensive than many several years. The point here, is not to discuss the other Sub-Saharan African countries. The median military rationale of the removal of such a portion of amount of capital per worker in Eritrea is just over the workforce but only to underline its detrimental $21,000; a very high amount given the size of the impact on manufacturing. The lack of demobilization country and its private sector. has deprived the private sector of a large segment of The ratio of value added to capital (value added its skilled workforce and negatively affects the quality per unit of capital), is used as an approximate measure of overall management and performance. of the productivity of capital. Given the extremely high The scarcity of labor is a major issue for the capital intensity in Eritrea, we would expect low returns Eritrean manufacturing sector. Our data strongly to capital, and hence very low capital productivity. confirm that labor is in fact scarce and that this Capital productivity in Eritrea is significantly lower than scarcity is driving up the cost of production. Between that of other countries in sub-Saharan Africa and the end of 2001 and May 2002, earnings increased beyond; its almost to the point of zero return. sharply over the sample by 17 percent. The increase Executive Summary vi in earnings was particularly significant for unskilled rates, there is real rigidity in the financial sector. The production workers (close to 22 percent) and also main concern with these rates is the lack of variance. very high for skilled production workers (close to 18 Interest-rates do not appear to take into account the percent). These are very significant wages changes, varying degrees of risk represented by the firms in our given that the time period considered is only one year. sample. This reflects the broader problems of the It highlights the enormous problem of scarcity of labor financial sector--by latching onto a 12 percent in an already difficult operating environment interest rate "ceiling," and paying depositors 5-6 percent, the financial sector has settled into lending at Financial Issues. Although there are some positive a risk-invariant rate of 9 percent while imposing elements in the country's financial sector--there substantial collateral requirements. seems to be adequate liquidity and an overall lack of corruption--it is unclear whether the availability of Access to Land. Firms were asked if "access to land" credit within the banking sector translates into was a significant obstacle to doing business in adequate access to low-cost credit and satisfaction of Eritrea. A close examination of the answers indicates demand at the firm level. in fact a bi-modal pattern, with companies usually Our results indicate that a large proportion of firms answering either "not a problem" or a "severe" in the private sector do have access to the formal problem. This division of answers could be due a financial sector, primarily through the use of deterrence effect. The results have also been overdrafts. However, most of the finance is clearly influenced by the fact that a majority of the 35 being used to meet working capital needs. privatized firms in Eritrea were included in the sample Investment is very low in Eritrea, even when compared study. These companies had unique privileges to the rest of Africa. Government policies to promote regarding access to land and other resources prior to investment and laws encouraging credit flows to the their transfer from government ownership. For those private sector are urgently needed to put the country firms that do need land and attempt to obtain it, land on a higher growth trajectory. is extremely difficult to get. On average, firms that About 40 percent of domestic private firms do not were able to obtain land within the last five years had have a loan; a majority of all other types of firms to wait 316 days and had to pay about 127,185 N indicated that they do not currently have a loan. All (roughly $US 9,080) for processing. firms in this category (except domestic privately owned firms) indicated that they did not apply for a loan. Of the firms that did not have a loan and had not Regulatory Constraints and Infrastructure applied for a loan, 75 percent indicated that they did During the RPED survey, owners and managers of not need a loan, and 17 percent argued that interest manufacturing firms were asked both qualitative rates are too high. The fact that such a large number "ranking" questions and quantitative "cost" quest- of firms did not apply for a loan may indicate powerful ions regarding infrastructure and government deterrence effects, such as the problem of accessing regulations. General business regulations and land to submit as collateral. The ratio of the value of infrastructure are covered, in order of importance for collateral to loan value is very high for domestic private the manufacturing firms in our sample. Some of the firms (172 percent) and high for other firms as well. more important issues outlined by the private sector With three state-owned banks that do not truly were macroeconomic stability and inadequate compete with each other and essentially fixed interest supplies of electricity. Executive Summary vii Macroeconomic Stability. The worst rating was given and so on) the average answer was consistently that to "Macroeconomic Stability" (defined as the stability they were not obstacles. Although some firms of inflation and exchange rates). The evidence complained about the burden of providing six months shows that macroeconomic instability is perceived of pay when a worker is fired (18 percent of the sample by the private sector as the largest factor dragging said that "layoff procedures and cost of retrenchment" down performance. These findings roughly were a problem), most firms agreed that the labor correspond to questions on the impact of lack of regulations were not the main problem and repeated foreign exchange on capacity utilization. The that the major problem was the lack of workers. majority of firms classified lack of exchange for imports as a severe problem (63 percent). Business Licensing. There is evidence that efforts to streamline the licensing of businesses have had some Electricity. While 27 percent of the sample answered success as well. Generally, the private sector sees the that electricity was not a problem, 18 percent license as a comprehensive instrument, which is considered it a "severe" problem and a further 11 reducing the amount of time required previously to and 15 percent, "major" and "moderate," obtain various licenses. For those firms established respectively. In addition, the average firm owner had within the last five years, the average number of to wait 99 days to obtain an electrical connection, licenses required was indeed one. The average although the distribution of answers was skewed by answer for the time it took to get a license was two one answer and a more meaningful number is days (for those firms that were established during the probably the median, 30 days. According to sample last two years). One should bear in mind, however, firms, there were an average of nine electricity that the RPED approach only covers existing interruptions (median equals five days) a month businesses that were by definition successful in during the year 2001. It is not surprising then that 43 overcoming these hurdles. Other assessments have percent of the sample firms own a generator from pointed out that potential investors are still having which they obtain an average of 22 percent of their business license applications turned down for electricity. Thus, the uncertain supply of electricity arbitrary reasons (World Bank 2002a). However, imposes higher business costs, both to acquire a overall, the average time spent by senior generator and to operate it when the public grid is management on government regulations was 5 not supplying electricity. percent, much lower than RPED results in other sub- Saharan African countries. Corruption. The majority of respondents (84 percent) in a ranking exercise answered that corruption is not a Main Conclusions significant obstacle to doing business in Eritrea. This The RPED survey of Eritrea yielded several results of perception was corroborated by answers to other value to the government in terms of designing policies related questions throughout the survey. regarding the private sector. The Eritrean manu- facturing sector is very capital intensive, capital Labor Regulations. A majority (85 percent) of productivity is quite low, and there is a shortage of respondents stated that labor regulations were not a skilled labor, causing wages to rise. Firms do have problem. When asked about specific types of labor access to the formal financial sector to some extent, regulations (hiring local or foreign workers, procedures primarily through the use of overdrafts. However, and cost of retrenchment, limits on temporary hiring, most of the finance is clearly being used to meet Executive Summary viii working capital needs and investment is very low. rates. While stabilization in the current situation in Interest-rates do not appear to take into account Eritrea will probably result in a major devaluation varying degrees of risk. Access to land is also of the official exchange rate, it will likely bring a presenting a severe problem for some firms. With number of benefits. Net earners or receivers of regard to the general business environment, access foreign exchange will be better off. Foreign aid will to foreign exchange, macroeconomic instability, and put greater resources in terms of Nakfa at the intermittent provision of electricity are among the disposal of government. Receivers of remittances largest obstacles. There exist a host of "second will be better off. Flow of remittances and foreign generation" issues that are not apparent at this time private investment is likely to rise as confidence because the labor shortage dominates above all other returns. Supply of foreign exchange will no longer constraints. It may well be the case that once labor is be much of a constraint on private businesses. freed up and some key problems are addressed, Costs of imported inputs will increase but will now there will be a surge in productivity and exports. At be correctly reflected on the firms' balance that point, the "second generation issues" will become sheets. To the extent that domestic production the binding constraints. competes with imports, the higher cost of foreign exchange will give the enterprises additional Recommendations protection. New export earners will emerge, A number of general and specific recommendations further increasing the available foreign exchange arise from the analysis and conclusions of this study. and thus help stabilize both nominal and real The two key obstacles to resuming growth are the lack exchange rates in the future. of foreign exchange and the severe shortage of labor. As these obstacles are removed, another set of Supply of Labor. The results of the survey constraints will likely become binding. The survey analysis clearly indicate that the current suggests that the top four obstacles in this "second overarching constraint is labor. The labor generation" are 1) infrastructure constraints such as: shortage affects firms' ability to grow and unreliable electricity supply, poor telecommunications, become competitive, either in the regional or and high costs for sea and air transport; 2) limited international marketplace. Due to the severe access to finance at a reasonable cost; 3) low levels of shortage of labor, firms are relatively less education and skills in labor force; and 4) limited productive in Eritrea. The ratio of capital to labor access to land. is much higher than the optimal ratio, and wages have been rising. Unit labor costs, a rough Foreign Exchange. The lack of availability of indicator of competitiveness, show that Eritrean foreign exchange is mostly a macroeconomic labor is expensive relative to labor in other parts issue. Above all, business people want a of the world. When coupled with other constraints predictable environment, which means slow and highlighted in the report, the labor shortage predictable changes in exchange rates, interest creates a severe drag on competitiveness in the rates, and the rates of inflation. This can only private sector. It may well be the case that once come about from macroeconomic stabilization. labor is freed up and other key problems Until that is achieved, inflation rates will continue mentioned in this report are addressed, there will to be high (double digit) and variable. That in turn be a surge in productivity and exports of existing will play havoc with exchange rates and interest firms and a high demand for new investment. Executive Summary ix Immediate and effective demobilization is Educated and Skilled Labor Force. The study necessary if the private sector in Eritrea is to found both direct and indirect evidence of the function normally. Demobilization will also make it negative impact of low levels of education and easier to achieve macroeconomic stability. skills of the Eritrean workers. At the direct level, Infrastructure. Power, telecommunications, and employers complained that even when workers transport are not currently the most important are available, they were not easily trainable binding constraints but may well become more because of a generally low level of education and problematic once labor is freed up and the private experience in jobs that require skill. Indirect sector starts growing. The uncertain supply of evidence comes from the fact that the unit labor power already imposes some extra costs on firms costs (ULCs) are high when compared to and may quickly emerge as a larger constraint successful exporters. A strong rise in productivity once the labor situation improves. would reduce ULCs, because ULCs in domestic Telecommunication reforms need to be pushed currency are just a ratio of nominal wages to labor vigorously to set the framework for a much productivity. But greater productivity is a function expanded landline and wireless access to the of better education and greater skills. world. Costs of sea and air transport are high and need to be brought in line with those of the Access to Land. Our survey shows that some firms neighboring countries. have serious problems with access to land; others report that they do not have a constraint. It is very Access to and Cost of Finance. At a general clear that the administrative process for accessing level, macroeconomic stabilization, by reducing land is problematic. Aggressive follow-up by the the crowding out of private sector by the public relevant authorities on existing applications for land sector, should increase the availability of funds for is urgently needed; the government needs to sort lending and decrease the average level of out exactly what is holding things up from the point interest rates. At a specific level, financial of view of administrative barriers. Using market allocation needs to take into account varying mechanisms to price the leasing of land may make degrees of risk. Low-risk investment will then sense as well; one possibility is to auction off leases benefit from lower interest rates, and high-risk for land. In most market economies, there is a "rent investment will at least get the opportunity to gradient"; that is, center-city land is much more succeed (or fail). Banks need to restructure to expensive than land on the outskirts of a city. achieve this as well as increase profitability. This Asmara would probably be no different once should also lead to lower collateral requirements, bureaucratic obstacles to pricing of land are at least in some cases. In addition, more training removed. Ease of exit from leases is just as in credit assessment would facilitate this shift in important. If sub-optimal decisions are made operating procedure. If these adjustments are not regarding leasing of land, entrepreneurs should be made, finance could very well be an even larger able to sell their leases just as easily as they buy obstacle for existing and potential firms once the them. An active and efficient market for land is labor supply is increased, especially where long- crucial to the growth of the private sector. Valuable term investment is concerned. capital resources should not be tied up in land indefinitely due to lack of a functioning marketplace. Acknowledgments x The Regional Program on Enterprise Development Subramanian, and Axel Peuker. Jacki Edlund-Braun (RPED) team is grateful to Sumana Dhar for extensive edited the final version of this report. comments and overall guidance. Andrew Singer also RPED is a member of the Africa Private Sector provided very useful and thorough comments. Group (AFTPS) of the World Bank and conducts The RPED team would also like to thank the private enterprise surveys in various countries in Sub- following people for facilitating this work and Saharan Africa. It is also a member of the Bank-wide providing comments and suggestions: Dr. Woldai Investment Climate Unit, which carries out private Futur, Economic Advisor to the President; Prof. enterprise surveys all over the world. Additional Abraham Kidane, Economic Advisor; Dr. Giorgis information can be found on its web site: Teklemikael, Minister of Trade and Industry; Amaha www.worldbank.org/rped Kidane, Director General, Department of Industry, Ministry of Trade and Industry; Akberom Tedla, Secretary General, Eritrean National Chamber of Commerce; Taddese Beraki of the Eritrean National Chamber of Commerce; and Kiflemariam Zerom of Business Consulting Services. The team is indebted to Country Director Makhtar Diop and to the Eritrea Resident Mission for their guidance--in particular, Emmanuel Ablo, Samuel Zerom, and Saba Tekle. In addition, RPED is grateful to the Integrated Framework Team and the Eritrea Country Team for their advice and for generous financial support. Linda Cotton participated in the preparatory mission and led the pilot survey for this exercise. Jean Michel Marchat led the survey in Eritrea during April- May 2002 and was assisted by a team composed of Linda Cotton, Dr. Tesfay Haile (General Manager and Consultant, FABI Consulting & Trading, PLC), and John Paton, as well as several local enumerators hired by FABI Consulting. Manju Kedia Shah provided econometric analysis of cross-country data. Hooman Dabidian provided analytical analysis. Also Vijaya Ramachandran task-managed the survey and the writing of report. Peter Miovic presented the report to the Government of Eritrea and made extensive revisions to the final document, based on consultations in Asmara. Several World Bank colleagues provided useful advice and suggestions; the team thanks Ibrahim Elbadawi, Ian Bannon, Iain Christie, Michael Wong, William Steel, Melanie Marlett, Demba Ba, Uma Abbreviations and Acronyms xi AFTPS Africa Private Sector Group AGOA Africa Growth and Opportunity Act DRP National Demobilization and Reintegration Process EU European Union FDI Foreign Direct Investment FIAS Foreign Investment Advisory Service GDP Gross Domestic Product GOE Government of Eritrea HAMSET HIV, Malaria, Sexually Transmitted Diseases, and Tuberculosis in Eritrea HIPC Highly Indebted Poor Country IBRD International Bank for Reconstruction and Development ICA Investment Climate Assessment IDP Internally Displaced Person IFC International Finance Corporation IMF International Monetary Fund IT Information technology MAP Multi-Country HIV/AIDS Program for Africa NICE National Insurance Corporation of Eritrea OLS Ordinary Least Squares PFDJ People's Front for Democracy and Justice RPED Regional Program on Enterprise Development SSA Sub-Saharan Africa ULC Unit labor cost UN United Nations USAID United States Agency for International Development WHO World Health Organization 1. Economic Growth and the Private Sector 1 Introduction 2 Private Sector Development and Growth: Looking Ahead 2 Macroeconomic Stability 4 Public and Private Investment 6 Conclusion 8 1. Economic Growth and the Private Sector 2 Introduction two countries due to trade disputes, the level of charges in Eritrean ports, as well as the negative The current Government of Eritrea (GOE) is committed Ethiopian reaction to Eritrea's introduction of a new to improving economic development and raising the currency, the Nakfa, in place of the birr. After a series living standards of the population. The private sector, of skirmishes over the poorly demarcated southern with emphasis on the export sector, is seen as the border with Ethiopia, war broke out. A peace engine of growth that will help jump start the small agreement signed in Algiers in June 2000 created a economy and eventually lead to long-term growth. 25-km-wide demilitarized zone along the 1000-km- The recent efforts of the GOE to promote private long border. Both sides have agreed to a border sector­led growth are grounded in a tradition of decision made by an independent Boundary entrepreneurial spirit that survived periods of relative Commission in The Hague in April 2002, and the inactivity. Industrial development in Eritrea began physical demarcation is still in process. during the Italian colonial period (1890 to 1941), when a large number of firms were established in light manufacturing (food and beverages) and con- Private Sector Development and Growth: struction materials (cement, brick, and tiles). Looking Ahead Ethiopian Emperor Haile Selassie annexed Eritrea in 1962, which started the war for independence and The RPED survey discussed in this paper is designed caused a decline in industrial production. Despite the to take a close look at the costs of doing business in nationalization of assets in the mid-1970s under the Eritrea and to make cross-country comparisons of socialist Derg regime (1974 to 1991), Eritrea still private sector competitiveness. This Investment accounted for 30 percent of Ethiopian industrial Climate Assessment (ICA) is structured in the production and was one of the most industrialized following manner: Chapter 2 begins the analysis with areas in East Africa. It is clear today that some of the a discussion of productivity in the framework of other commercial and industrial skills endured these African countries. It then moves on to cover the phases of relative idleness. various factor markets and ends with a broader look at The three-decade-long battle for independence the regulatory environment and constraints posed by left farmland and infrastructure devastated by the the existing infrastructure. The third and final chapter early 1990s. An estimated 80 percent of the country's draws conclusions from the survey and discusses infrastructure had been destroyed, and the loss of their policy implications.1 military personnel and civilians has been estimated at Eritrea's small economy has a gross domestic 150,000 (Bruchhaus and Mahreteab 2000). Eritrea product (GDP) estimated in 2002 at $US644 million. In began the difficult task of reconstruction, and its 2001, industry as a whole accounted for 22.3 percent provisional government decided in 1992 to (only 7.8 percent of which was manufacturing), demobilize 57,000 soldiers (60 percent of the total services for 59 percent (18 percent of which is "public 95,000, one-third of whom were women). After a administration"), and agriculture for 18.7 percent of referendum, the country proclaimed independence on Eritrea's GDP (Table 1.1). Out of a population May 24, 1993. estimated in 2002 to be 4.3 million, around 80 percent Just a few years into the new nation's life, lives in rural areas and is dependent on agriculture. however, war resumed with Ethiopia in May 1998. The data presented are very rough estimates and may Starting in 1997, tension began to mount between the not be reliable indicators of the true growth rates in the 1. Economic Growth and the Private Sector 3 Table 1.1. GDP by industrial origin, 1993-2001 (in millions of constant Nakfa, 1992=100) 1993 1994 1995 1996 1997 1998 1999 2000 2001 Agriculture 481.8 658.5 581.5 548.0 550.8 866.7 800.3 454.4 616.8 Crops and livestock 385.7 554.8 472.3 437.4 443.7 755.5 683.8 331.3 488.2 Staple crops 109.0 217.0 160.1 134.6 135.3 335.5 279.2 97.0 195.6 Cash crops 54.2 107.9 79.6 66.9 67.3 166.8 138.8 48.2 97.2 Livestock 222.5 229.9 232.6 235.9 241.1 253.2 265.8 186.1 195.4 Forestry and fishing 96.1 103.7 109.2 110.6 107.1 111.2 116.5 123.1 128.6 Forestry 92.6 95.3 98.0 100.8 103.7 106.7 109.8 113.0 116.3 Fishing 3.5 8.4 11.2 9.8 3.4 4.5 6.7 10.1 12.3 Industry 337.5 384.6 466.2 638.4 764.5 715.1 735.2 688.4 732.2 Mining and quarrying 4.0 4.0 4.0 4.1 8.8 6.8 6.1 3.0 3.0 Manufacturing 166.1 176.6 213.0 240.9 256.6 233.5 234.6 242.9 256.2 Handicrafts and small industry 28.1 32.7 38.1 50.2 93.0 85.0 88.4 91.5 96.5 Electricity and water 24.0 26.2 27.8 29.5 32.2 32.9 35.0 35.5 35.8 Building and construction 115.3 145.1 183.3 313.7 373.9 356.9 371.1 315.5 340.7 Services 1331.1 1659.3 1729.3 1845.7 1951.5 1813.1 1868.3 1856.1 1941.3 Distribution services 828.4 970.2 1074.4 1148.8 1222.7 1019.2 1029.4 967.6 1025.6 Trade, wholesale, and retail 567.8 641.4 717.8 763.9 808.9 648.9 655.4 589.8 625.2 Transport and communications 260.6 328.8 356.6 384.9 413.8 370.3 374.0 377.8 400.4 Other services 502.7 689.1 654.9 696.9 728.8 793.9 838.9 888.5 915.7 GDP at factor costs 2150.4 2702.4 2777.0 3032.1 3266.8 3394.9 3403.8 2998.9 3290.3 Indirect taxes less subsidies 144.2 130.0 125.2 137.4 155.8 133.0 146.7 123.6 135.6 GDP at market prices 2294.6 2832.4 2902.2 3169.5 3422.6 3527.9 3550.5 3122.5 3425.8 Source: International Monetary Fund. various sectors, because a proper system of the GOE approved 500 new companies, although it is compiling national accounts has not yet been not known how many of those actually materialized. instituted in Eritrea. The Business Licensing Office was designed to be a From the outset, the GOE viewed the private one-stop shop. The Chamber of Commerce and other sector as the key to economic growth and took steps vehicles were used to gather input from the business to promote its growth. The new government began to community. Industrial estates were established on the operate the state-owned industrial sector inherited at outskirts of Asmara, Dekamhare, and Dubarwa (with independence on a commercial basis and little infrastructure), and housing estates were started progressively began to privatize it. By the end of 2001, for investors including the diaspora. it sold off 35 of 41 large public enterprises. It The government has recognized economic reformulated the investment code and established the diversification, especially export diversification, as a key Eritrean Investment Center. Between 1992 and 1997, requirement for achieving sustainable, broad-based 1. Economic growth and the Private Sector 4 growth in the new century. The loss of the Ethiopian Macroeconomic Stability market and the rising debt mean that increasing exports is more important now than ever. Considerable attention Firms require a stable macroeconomic environment has been given to increasing comparative advantage in to lower risk and facilitate long-term financial, labor-intensive manufacturing for export because of the employment, and investment decisions. After potential for dramatic growth based on experience in independence, the GOE was quite successful in other developing countries. The government is stabilizing most major macroeconomic indicators. concentrating on industries in which there is already a However, many of the gains in macroeconomic significant level of skills in the country and which require stability achieved during the early years of relatively low levels of capital such as garments, knitwear, independence were reversed during the border and shoes. In addition, the government hopes to take war. Regaining this ground will take much longer advantage of the fact that the Cotonou Agreement gives than did the initial stabilization, partly due to a Eritrea preferential access to the European Market and continued perceived military threat from Ethiopia that the U.S. Africa Growth and Opportunity Act (AGOA) and partly because Eritrea's room for maneuver is gives Eritrea duty-free access for garments and quota- now much smaller than during the post- free access for the next eight years. independence stabilization. Significant progress was made after independence regarding the liberalization of trade GDP Growth policy. The 1994 legal notice 18/1994 reduced the GDP growth averaged 7.4 percent from number of import tariffs to twelve. Capital goods, raw independence in 1993 through 1997 (Figure 1.1). The materials, and semi-processed goods have only a 2 border war disrupted this growth, which dropped to percent tariff. Basic goods duties range from 3 to 20 0.3 percent in 1999 and became negative 11.9 percent, non-basic goods range from 25 to 40 percent in 2000. Growth resumed in 2001, led by percent, and luxury import duties go as high as 200 recovery of agricultural production. percent. In addition, customs procedures were The manufacturing sector expanded by about 50 simplified. In the mid-1990s, the government began percent from 1993 to 1997, according to rough major investments in infrastructure, roads, electricity, estimates compiled by the International Monetary dams, and port operations to support the further Fund (Figure 1.2). Then the manufacturing sector development of exports. Nonetheless, after a peak in was largely stagnant during the war years 1995, exports decreased and then collapsed in 1997 (1998­2000), and most likely a large percentage of and 1998. In 1997, Ethiopia reacted negatively to the that production was related to war efforts. The introduction of the Nakfa in place of the birr, and when private sector has tended to concentrate on the war began in 1998, trading ceased altogether. services, namely imports and distribution. Exports to Ethiopia were 63 percent of total exports in Manufacturing sector growth has begun to pick up goods in 1997. Neither was there much diversification slightly in 2000 and 2001. Although the in the type of goods exported. Exports of both goods manufacturing sector constitutes a small percentage and services saw a continual decline from 1997 to of GDP, the GOE emphasized developing it because 1999, with recovery beginning in 2000 and 2001. Total it is central to the strategy of growth through merchandise exports increased from US$20.1 million exporting labor-intensive light manufacturing. in 1999 to US$36.7 million in 2000 but then declined to US$19.8 million in 2001. 1. Economic growth and the Private Sector 5 Figure 1.1. Real GDP growth Inflation (annual percentage) Inflation, which was traditionally low in Eritrea (and Percent previously in Ethiopia), was brought firmly into the 15 single-digit range by 1996­1997 (Table 1.2). During the war years (1998­2000), inflation (end-year to end- 10 year) averaged 15.5 percent, peaking at 26.8 percent in 2000. A large part of the increase during this period 5 was due to food price inflation spurred on by the disruption of agricultural production, due to the 0 Ethiopian invasion of the western lowlands in May 2000. The inflation rate then started to come down, ­5 falling to an annual rate of 7.7 percent by the end of 2001. While there are no firm numbers yet available for ­10 inflation during 2002, it would be surprising if it has not moved back into double-digit range, given the ­15 continued high defense expenditures, slow progress 1993 1995 1997 1999 2001 on demobilization, problems with financing with some 1994 1996 1998 2000 of the donors, and problems on the border with Source: World Bank, World Development Indicators. Sudan. Exchange Rate Figure 1.2. Manufacturing sector In 1997, the dual exchange rate was unified, and the Bank of Eritrea was established. This was also the Constant prices (millions of Nakfa, 1992=100) year that the new currency, the Nakfa, was introduced, Percent of GDP (at factor costs) on par with the birr. By early 1998, the exchange rate Constant prices Percent GDP was stable, backed by reserves that reached nearly 5 300 9 months' of imports of goods and services, and 15 foreign exchange bureaus were trading foreign exchange. 200 6 During the war, the government kept the nominal exchange rate from depreciating in line with inflation and the real exchange rate appreciated. The parallel market for foreign currency began expanding. By 100 3 2000, foreign exchange was being rationed, pushing the black market rate to as much as twice the official rate. While this gap was significantly reduced for a while during 2001, the foreign exchange premium 0 0 1993 1995 1997 1999 2001 grew again during 2002 and was in the 60­70 percent 1994 1996 1998 2000 range by October 2002 (although in an admittedly thin and volatile market). It is likely to be a harbinger of Source: International Monetary Fund. renewed inflationary pressures. 1. Economic growth and the Private Sector 6 Table 1.2. Key economic indicators, 1993-2001 1993 1994 1995 1996 1997 1998 1999 2000 2001 Real GDP growth (%) -2.5 9.8 2.8 9.2 7.7 3.9 0.3 -11.9 9.7 Inflation (%) change, end of year 9.6 6.8 11 3.4 7.7 9 10.6 26.8 7.7 Fiscal balance -7.0 -13.3 -25.1 -20.6 -6.0 -39.9 -56.3 -47.8 -32.3 (incl. official grants) as % of GDP Exports of goods & services 93.3 143.3 171.4 200.2 203.3 110.4 65.7 97.5 147.3 (US$ million) Imports of goods & services 283.7 409.3 545.5 571.5 589 596.1 597.4 498.6 522.8 (US$ million) Current account balance 7.6 3.8 -10.2 -21.2 -5.7 -33.0 -40.1 -33.0 -29.7 (excl. transfers) as % of GDP Current account balance 26.4 20.8 3.5 -8.2 2.2 -24.8 -29.1 -16.7 -12.6 (incl. transfers) as % of GDP Reserves in months of imports 3.9 5.6 4.6 4.1 4.9 1.4 1.1 0.9 1.2 of goods & services (end-year) Gross national savings/GDP (%) 81.9 29.2 11.6 14.9 40.7 12.1 13.3 19.2 22.9 * GDP growth rates for 1993 and 1994 have been reported in various documents over the past few years. These figures, still unofficial, report GDP growth to have been 9.9 percent in 1993 and 25 percent in 1994. Resolution of the difference awaits the authorities' publication of the official estimates of GDP since independence. The current account balance estimates (excluding and including transfers) for the period 1993-96 are taken from the IMF staff report for the 1998 Article IV consultation. Source: World Bank, "Revitalizing Eritrea's Development Strategy," Washington, D.C., 2002, processed. The current dual exchange rate regime implies foreign aid, and domestic and foreign investment that those firms with access to foreign currency at the (including that of the Eritrean diaspora) will be official rate are in effect subsidized. Those without needed to foster the vigorous development of the access are forced, if they can afford it, to obtain private sector. foreign currency on the parallel market to make new capital investment or purchase raw materials, spare Government Budget parts, and so on. At times even this black market While the new GOE was largely able to institute fiscal currency is unavailable in sufficient quantities. discipline after the end of the thirty-year war, the budget was hit hard by the 1998­2000 border war. Between 1993 and 1997, the government was by Public and Private Investment and large able to pay for the demobilization and reintegration of soldiers, the resettlement of In addition to macroeconomic stability, the refugees, the support to families of killed and government strategy for growth requires a great deal disabled soldiers, and the increase in the salaries of of public and private resources. Public expenditures, underpaid civil servants. Their ability to achieve this 1. Economic growth and the Private Sector 7 without plunging into debt was due in part to reversed. Private transfers declined to 27 percent, improving tax collection and increasing non-tax and donor support increased to 34 percent of GDP. revenues. Non-tax revenues rose to nearly 23 Debt in net present value terms in 2002 as a percent of GDP in 1997, almost half of which came percent of exports of goods and services reached 311 from charges for Massawa and Assab port services. percent (preliminary estimate), more than the Highly On top of that, the GOE was able to reduce Indebted Poor Country (HIPC) eligibility criteria. corporate income taxes, eliminate export taxes, and However, the external debt was contracted on lower and reduce the dispersion of import tariffs. concessional terms, so the debt service to exports When the border war began with Ethiopia in 1998, ratio is more reasonable, estimated at 23.5 percent however, the port revenues dried up, public and projected to fall thereafter. Increasing exports to enterprises were no longer bringing in sufficient levels sufficient to repay these debts will be extremely revenue, military expenditures expanded, and an difficult to achieve. income surtax was introduced to help cover these lost revenues. Public expenditure averaged 90 Balance of Payments percent of GDP between 1998 and 2001, meaning Before the war the current account deficit (including that the state became the dominant actor in the official transfers) was more or less in balance, and economy. The budget deficit increased from 6 domestic and foreign debt remained low. During the percent of GDP in early 1998 to 48 percent in 2000, war, the current account deficit (including transfers) mostly due to military expenditures, reconstruction, reached a high of 29.1 percent of GDP in 1999, which and humanitarian needs. The deficit was financed was subsequently brought down to 12.6 percent in primarily with domestic banking resources. Domestic 2001. Gross international reserves that covered five debt reached 128 percent of GDP by 2000. While the months of imports at the end of 1997 dwindled to less numbers for 2001 and 2002 are not yet known, than one month of imports by 1999 and have renewed double-digit inflation and a large premium remained in that range since. on foreign exchange in the parallel market in mid- 2002 suggest that serious macroeconomic Foreign Direct Investment imbalances continue. Foreign direct investment (FDI) has not taken off, despite a promising start in 1996. Between 1996 and Debt 2000 FDI inflows have remained largely stagnant, The country was debt free at independence and staying between US$35 million and US$39 million. established a debt management unit within the According to a recent World Bank document, several finance ministry in 1997. The external debt rose to foreign companies had signed joint ventures in areas US$321 million by the end of 2000, surpassing 50 such as fisheries and agribusiness but backed out percent of GDP in that year. for various reasons (World Bank 2002a). The Dependence on support from the diaspora has government's strategy includes trying specifically to been shifting toward dependence on aid, as diaspora attract the diaspora, but also FDI in general. A inflows declined somewhat over the past few years survey of foreign and diaspora investors is currently and aid receipts increased strongly. Private transfers being carried out by Foreign Investment Advisory averaged 45 percent of GDP from 1993 to 1997. Net Service (FIAS) to identify the perceptions of various donor support was 16 percent of GDP during the groups of potential investors about constraints to same period. By 2001, these ratios were essentially investment in Eritrea. 1. Economic growth and the Private Sector 8 To a large extent these perceptions include the lack Conclusion of confidence in a lasting peace, which will take some time to develop. In addition, despite the progress made The border war has been a major setback and has in reforming the economy, business operations are still imposed tremendous costs in terms of lives, food hampered by various obstacles. Anecdotal evidence security, and economic development. In the long run, suggests that decisions made on land, access to however, there is hope that Eritrea can regain the finance, and access to foreign exchange are based on momentum that was interrupted. The attitudes of self- opaque procedures influenced more by the reliance and determination that sustained the country government than by market forces. during the war for independence are still present As discussed further in chapter two, business throughout Eritrean society. But expectations for regulations have seen some improvement but more is dramatic increases in productivity and growth may needed, especially to facilitate FDI. This will become need to be tempered with realistic and practical clearer as more foreign investors attempt to enter the consideration of all of the challenges ahead, some market. The absence of foreign investors means not resulting from conflict and others not related to war only forgone capital investment, exports, and jobs, but equally difficult to overcome. but also forgone knowledge and skills. It is in this context that the GOE requested RPED to conduct a survey of the manufacturing sector as a Foreign Aid whole. This work was done as part of a series of The World Bank is the major source of external funds projects regarding trade and investment issues. for economic recovery, demobilization, and Included in these efforts are the development of a reintegration programs. The World Bank and the "Crash Program on Exports" guided by Andrew European Union (EU) are also the principal Singer, an "Assessment of Policy and Performance in supporters of the government in its efforts to improve Eritrea's Services Sector" prepared by Carsten Fink, power generation/distribution and rural road and an investment perceptions survey currently rehabilitation. The rehabilitation of the ports of being carried out by the Foreign Investment Advisory Massawa and Assab is also being undertaken with Service of the International Finance Corporation and the backing of the World Bank. In addition, the World International Bank for Reconstruction and Bank is financing the government's Economic Development (IFC/IBRD). Recovery Program. Other significant economic assistance tends to focus on rural development and agriculture. Danish and Italian assistance is centered Notes on agricultural development and education. The U.S. Agency for International Development (USAID) 1. The sampling design is one of a stratified random programs focus on the growth of rural enterprises to sample. The Ministry of Industry and Trade develop agricultural and related support enterprises, provided RPED with a list of officially registered with emphasis on high value exports. The UN Food companies with more than 10 employees in 2000, and Agriculture Organization assists the Ministry of which accounts for about 65 percent of total Agriculture with weather forecasting, crop reporting, employment in manufacturing. Clusters were then and food security impact assessments. Funds from defined on the basis of location, size, and sector, the International Fund for Agricultural Development and firms were sampled randomly from each. The support irrigation systems. sample actually surveyed has 79 firms, giving a 1. Economic growth and the Private Sector 9 sampling rate of 35.2 percent. The sample frame and sample exhibit two prominent characteristics: the weight of the Asmara region is high, and firms with more than 100 workers in total employment are dominant. (For a more detailed description of the methodology, see Annex I.) 2. Investment Climate in an International Perspective 10 A. Firm-Level Productivity and Cross-Country Comparisons 11 B. Factor Markets 16 C. Regulatory Constraints and Infrastructure 27 Conclusion 33 2. Investment Climate in an International Perspective 11 A. Firm-Level Productivity and Cross- Cameroon. The median amount of capital per worker Country Comparisons in Eritrea is just over $21,000; a very high amount given the size of the country and its private sector. At the core of any discussion of private sector viability Ratio by size measures capital per worker in each is the issue of firm-level productivity. Using value- size category, relative to the median for each country. added per worker as an approximate measure for Therefore, a ratio of 0.79 for microenterprises in Eritrea productivity, we compare Eritrean firms with firms from indicates that capital per worker is US$16,816 (or 0.79 other parts of Sub-Saharan Africa. We begin our x US$21,288). Large firms in Eritrea have a capital- analysis by describing the capital-labor ratio for labor ratio of 2.14, more than twice the amount of several African countries, including Eritrea, by size capital per worker as the median Eritrean firm. In all and by sector (Table 2.1). countries, the capital intensity increases with firm size, Table 2.1 clearly shows that firms are Eritrea are with large firms generally having more than double the more capital intensive than all other Sub-Saharan amount of capital per worker as compared to medium African countries, with capital accumulation in Eritrea and small firms. Across sectors, we see that the textile almost double that of the next highest country, and furniture sectors tend to have below average Table 2.1. Distribution of capital to labor ratio ($US) Cameroon Côte Ghana Kenya Tanzania Zambia Zimbabwe Eritrea d'Ivoire Median 11,496 5,469 946 4,511 4,065 4,417 4,382 21,288 Ratio by size Micro 8,771 2,319 279 1,669 1,852 2,309 1,183 16,818 Small 9,289 4,884 731 4,055 3,309 4,064 3,050 20,011 Medium 13,197 9,511 1,279 3,848 3,480 4,201 3,729 34,061 Large 24,785 8,964 1,401 5,427 3,671 4,554 4,001 45,556 Very large 15,956 8,340 1,893 4,782 7,431 4,987 5,241 ­ Ratio by sector Food 13,301 9,565 1,907 4,534 4,191 4,174 4,846 26,610 Metal 13,450 5,114 1,650 2,643 3,032 3,905 4,965 25,971 Textile 8,220 2,237 805 2,314 2,638 3,419 3,138 16,179 Furniture 9,760 4,578 1,115 4,128 4,411 2,460 2,550 3,832 Note: In all instances, the mean was greater than the median, and due to the skewed distribution, the median is a more appropriate measure of central tendency. Source: World Bank, RPED Eritrea, 2002. 2. Investment Climate in an International Perspective 12 capital intensity, while the food and metal sectors tend Europe. Firms that ordered capital equipment in to have above average capital-labor ratios. 1996­1997 might also have been forced to let some The extraordinarily high degree of capital of this equipment sit idle in the years since, at least in intensity found in this sample of Eritrean enterprises part due to the war. One firm in our sample ordered is backed up by anecdotal evidence. One firm owner equipment about four years ago; this equipment has said that he had to rent additional space to store been stored in a warehouse simply because the firm capital equipment that had previously been used to owner has not been able to secure access to the produce for the Ethiopian market. In addition to the needed additional land. problem of reduced demand, access to land had Compounding these problems is the military prevented him from building additional space. Some mobilization of labor--about 12.5 percent of firms reported that they could easily add another shift professionals in the sampled firms were recruited if faced by higher demand, indicating that at present, into the army in 2001. Many managers confirmed there was plenty of idle capacity. In the leather that they did not have enough employees, tanning industry, some firm owners complained of particularly skilled ones. Hence, a variety of reasons lack of labor and lack of quality leather to export to for the lack of expansion in the manufacturing sector Table 2.2. Distribution of value-added per worker ($US) Cameroon Côte Ghana Kenya Tanzania Zambia Zimbabwe Eritrea d'Ivoire Median 9,656 1,122 1,304 3,337 1,862 2,962 3,999 1,786 Ratio by size Micro 4,838 272 767 1,595 1,460 1,780 2,175 1,857 Small 7,059 1,076 1,198 3,337 1,411 3,110 3,087 1,340 Medium 19,196 1,404 1,132 3,374 2,272 3,012 3,879 2,518 Large 19,418 2,087 2,476 4,655 2,080 4,123 3,999 1,804 Very large 17,226 2,025 3,463 2,830 3,754 4,668 4,919 na Ratio by sector Food 14,088 1,652 2,628 5,349 3,441 3,415 7,182 2,697 Metal 8,922 1,361 1,438 1,929 1,940 4,404 3,815 3,233 Textile 6,363 376 1,008 1,832 1,030 1,973 3,119 1,072 Furniture 7,705 997 917 2,676 890 1,582 2,559 2,840 Note: In all instances, the mean was greater than the median, and due to the skewed distribution, the median is a more appropriate measure of central tendency. For Eritrea, other sectors were also included in the survey. Source: World Bank, RPED Eritrea, 2002. 2. Investment Climate in an International Perspective 13 emerge from the evidence--lack of demand, low difference between Eritrea and other countries in our availability of labor, lack of access to land, and lack sample. Micro-enterprises are much closer to the of quality inputs. The last three of these constraints median in Eritrea than micro-enterprises in other could present serious obstacles to the expansion of countries. The productivity difference between small supply should the firms be able to find markets for and large firms is much more significant in other their products. countries than in Eritrea, where there is virtually no Value-added per worker in Eritrea, as described in difference between micro and large firms. By sector, Table 2.2, is comparable to that of other Sub-Saharan we see that sectors with high capital-labor ratios tend African countries. It is almost equal to that of workers to have higher labor productivity. in Tanzania and higher than that of workers in Ghana Table 2.3 describes value-added per unit of and Côte d'Ivoire. Across size classes, labor capital, which is an approximate measure of the productivity is lowest for the small size class (10­49 productivity of capital. Given the extremely high employees) and highest for the medium size class capital intensity in Eritrea, we would expect low (50­99 employees). It is interesting to note the returns to capital and hence very low capital Table 2.3. Distribution of value-added to capital ratio Cameroon Côte Ghana Kenya Tanzania Zambia Zimbabwe Eritrea d'Ivoire Mean 0.89 2.45 1.41 0.81 0.66 0.64 0.88 0.14 Standard deviation 3.11 3.90 5.17 3.52 4.99 3.59 3.08 0.26 Ratio by size Micro 0.65 1.77 2.42 1.07 1.60 1.03 1.49 0.11 Small 1.11 2.36 1.28 0.70 0.31 0.54 0.84 0.12 Medium 0.87 2.32 0.95 0.84 0.40 0.47 0.70 0.29 Large 0.80 2.65 0.65 0.73 0.37 0.59 0.93 0.13 Very large 0.88 2.97 0.77 0.48 0.19 0.78 0.75 n.a. Ratio by sector Food 0.85 2.01 1.47 0.91 0.60 0.66 0.91 0.19 Metal 0.99 3.09 1.11 0.83 0.67 0.68 0.62 0.09 Textile 0.78 1.67 1.99 0.86 0.53 0.47 1.01 0.59 Wood 0.85 2.76 1.01 0.75 0.72 0.70 0.98 0.85 Note: All values are in U.S. dollars. For Eritrea, other sectors were also included in the survey. The wood and furniture sector included only two furniture makers, with much higher value-added to capital ratios compared to others in the survey. The averages for the comparator countries are across three waves of data gathered in the 1990s. Due to small sample size and lack of observations on capital, figures for Côte d'Ivoire may be biased. 2. Investment Climate in an International Perspective 14 productivity, and this is exactly what we find. Capital When converted to a common currency, it enables productivity in Eritrea is by an order of magnitude international comparisons of competitiveness of labor. lower than capital productivity in other countries. Unit labor cost in U.S. dollars is defined as: Across countries, capital productivity tends to fall as size of the enterprise increases. There seems to be no ULC = (w.L/Q).(1/e) strong pattern in capital productivity when data are where w is the manufacturing wage disaggregated by sector. L is the amount of labor employed The main conclusion emerging from the data is that Q is a physical measure of output the extremely high capital-labor ratios and the resulting e is the exchange rate defined as domestic near-zero returns to capital are likely to be due to the currency per U.S. dollar severe shortage of labor in Eritrea. It remains to be seen whether other constraints on the supply or demand ULC can also be reformulated as a ratio of nominal side will emerge once demobilization of the army wage (w) to labor productivity (Q/L). For a country to relieves the labor constraint. have a low (competitive) ULC, it has to do one of three things (or a combination thereof): (i) keep nominal Unit Labor Costs in Eritrea wages low, (ii) keep its exchange rate competitive, Finally, let us consider the unit cost of labor in Eritrea and/or (iii) increase its labor productivity. Because it is versus other countries in Sub-Saharan Africa. Unit very difficult to obtain comparable physical measures labor cost measures the total cost per unit of output. of output across different countries, we use an Figure 2.1. Ratio of wages to value-added in Africa (approximating for unit labor costs) 0.7 0.59 0.6 0.5 0.42 0.40 0.4 0.38 0.39 0.39 0.35 0.35 0.36 0.36 0.33 0.3 0.26 0.2 0.1 0 Botswana Cameroon Côte Ghana Kenya Madagascar Malawi Nigeria Tanzania Zimbabwe Zambia Eritrea 1990 1990s d'Ivoire 1990s 1990s 1984 1983 2001 1990s 1990s 1990s 2002 1990s Source: World Bank, RPED Eritrea, 2002. 2. Investment Climate in an International Perspective 15 Figure 2.2. Ratio of wages to value-added, show that Africa has higher ratios of wage to labor Asia (Approximating for unit labor costs) productivity relative to Asia at roughly equivalent Ratio stages of development. When data from Africa for the 0.3 1980s are compared with Asian data from the 1960s 0.27 0.26 and 1970s, it is clear that earnings in Africa are about 0.24 two-thirds higher than was the case historically in Asia, 0.21 and African productivity is about one-fourth lower. 0.2 Disaggregation of the Eritrean data shows that the 0.16 unit cost of labor is very high in certain size categories--very small firms have the highest unit 0.1 labor cost (0.69), small and medium-sized firms have unit labor costs in the general range for Sub-Saharan Africa, and large firms have very high unit labor costs 0 (0.55). When broken down by sector, textiles and Indonesia South Malaysia Taiwan Thailand garments and furniture have reasonable unit labor 1981 Korea 1970 1961 1970 cost ratios (0.20 and 0.26, respectively), while other 1963 sectors have ratios between 0.60 and 0.77. Interestingly, government-owned firms have very high Source: Lindauer and Velenchik, 1994; and RPED Surveys, 1990s. unit labor cost ratios of 0.60,whereas party-owned firms have a ratio of only 0.22, although the sample size of the People's Front for Democracy and Justice approximate measure of ULC, which is the ratio of (PFDJ)-owned firms is small and hence may not be wages to value-added at the firm level, averaged statistically representative. Foreign firms have a lower across the sample of firms (w.L/p.Q), where p is the unit labor cost than domestic firms. deflator for physical value-added. Several explanations have been offered for high Figures 2.1 and 2.2 describe the ratio of wages to unit labor costs in Africa, including the effect of non- value-added for several countries in Asia and Sub- market forces such as unionization and labor Saharan Africa, including Eritrea in 2002. It is clear regulations that have resulted in high wages in the that while Eritrea's unit labor cost (0.38) is within the formal sector and the low labor-land ratio. Abundant range of Sub-Saharan Africa countries, it is much supplies of cheap labor have tempered wage higher than that of several Asian countries during the increases in Asia. In the Eritrean case, the labor period of their rapid economic growth. Daily wages, scarcity that is supposed to be prevalent on the overtime, bonus, and other benefits for skilled African continent is compounded by the extremely production labor are about $2.63 per day, whereas high numbers of people who have been mobilized into unskilled workers receive about US$1.86 per day. the army relative to the size of the Eritrean population. Chapter three will go into further detail on the dynamics of the labor market. Capacity Utilization By definition, unit labor costs are high in countries The average capacity utilization by firms changed that have high wages and low labor productivity. Apart only marginally between 1999 and 2001, averaging 53 from overvalued exchange rates that have hampered percent for the overall sample. Examining across firm Africa's competitiveness, the data on unit labor costs size and ownership categories, we find that capacity 2. Investment Climate in an International Perspective 16 Figure 2.3. Average capacity utilization utilization was lowest for firms in the small size by size (mean) category (10­49 employees) and not much different Percent across other size classes (Figures 2.3 and 2.4). 75 Across ownership categories, we see sharp differences in capacity utilization rates. Government- 63.4 61.9 60.0 owned firms have the highest capacity utilization, while foreign firms have the lowest. The survey did not 50 provide sufficient detail to ascertain the reasons for 46.4 these differences (type of goods produced, access to capital, inputs, and so on). Examining the obstacles to higher capacity utilization, we see that the most significant factor 25 inhibiting firms is the lack of foreign exchange to purchase imports (Figure 2.5). The second highest- ranking obstacle was the lack of skilled labor, followed by equipment breakdowns and shortage of raw 0 materials. Surprisingly, lack of demand had the lowest Very small Small Medium Large ranking overall, suggesting that obstacles to boosting Source: World Bank, RPED Eritrea, 2002. capacity utilization came mainly from the supply side. B. Factor Markets Figure 2.4. Average capacity utilization by ownership (mean) Labor Market Percent The total Eritrean labor force consisted of an 80 estimated 2.2 million persons in 2002, or about 50 71.4 percent of the total population. The manufacturing labor market is a tiny fraction of the overall labor 60 58.5 market. According to the last estimate available, the 50.1 manufacturing sector accounted for about 1.3 percent of the labor force in 1999 (World Bank 2001a). 40 38.0 The objective of this section is to provide some insight into the structure and characteristics of the Eritrean labor market, based on data gathered during 20 the 2002 RPED survey. A large portion of this survey was devoted to the collection of labor data. The 79 surveyed enterprises provided information from 0 Private Government PFDJ Foreign internal records and from worker interviews (a sample domestic of up to 10 employees was taken from each firm). In the first part of this chapter, we describe the salient Source: World Bank, RPED Eritrea, 2002. features of employment in Eritrean manufacturing. In 2. Investment Climate in an International Perspective 17 Figure 2.5. Ranking of obstacles labor-intensive activities. These sectors account for 45 to capacity utilization and 24 percent of the manufacturing employment in 2001, respectively (Figure 2.6). Very small firms account for only 0.8 percent of Lack of demand 1.84 employment in the sample while large enterprises Poor infrastructure 1.97 employ almost 74 percent of the workers (Annex III, Table A.3.5). This result is consistent with what was Shortage of local raw 2.44 found in RPED surveys in other African countries. materials/inputs Finally, the importance of Asmara in terms of Lack of working 2.44 employment is obvious. About 83 percent of the firms capital were located in that area, and these firms account for Shortage of imported 2.57 about 87 percent of the employment in 2001. raw materials/inputs Another feature of the distribution of employment Equipment 2.78 breakdowns is of special interest: the share of management and professionals (white-collar workers) in the average Shortage of 3.52 skilled labor firm's workforce. It is often argued that a possible explanation for the comparatively high cost of labor in Shortage of 4.38 forex for inputs Africa is attributable to an excess of white-collar workers, which was measured at 20 to 30 percent in 0 1 2 3 4 5 previous RPED surveys in Africa. Eritrea is quite No obstacle Severe obstacle different in this respect. In our sample, only about 11 Source: World Bank, RPED Eritrea, 2002. percent of the workforce is composed of white-collar workers, about half the number found in other countries (Table A.3.1). the second part, we provide information on the This is probably due to the scarcity of labor in characteristics of the workforce. Finally, in the third Eritrea. A common complaint of firms was that many of part, we examine various aspects of earnings, their skilled employees had been enlisted in the army changes in wages, and wage determination. because of the war effort. The survey data we gathered confirm this complaint. Structure of the Manufacturing Labor Market Overall, 2.58 percent of the employees of the Using detailed firm-level survey data we examine the surveyed firms were mobilized in 2001. While in patterns of employment within manufacturing and the absolute terms this number is small, a closer look at changes in employment across various stratification the mobilization rates by job position reveals variables for the 1999­2001 period. The data suggest important negative impacts on firm performance that the distribution of employment is quite uneven (Table 2.4). Mobilization rate of professionals (12.44 across regions and sectors and that employment in percent) is much larger than for any other position. manufacturing has been declining recently. When this is considered in conjunction with The sample does not differ from many developing mobilization of managers (2.56 percent), it means that countries in that permanent employment in about 15 percent of the available white-collar workers manufacturing is concentrated in the textile, leather, were on military duty in 2001. In addition, mobilization and garments and food and beverage industries, two seems to have followed a very rational pattern in 2. Investment Climate in an International Perspective 18 Figure 2.6. Structure of permanent employment unusually large share in Eritrea can be explained by (by sector, 2001) the fact that in times of conflict, women tend to Chemicals/paints Construction materials replace men sent to the front line. Food and beverage Furniture The point here is not to discuss the military Metal Paper/printing/ rationale of the removal of such a portion of the publishing workforce but only to underline its detrimental impact Plastics Textile, leather and on manufacturing. In effect, it deprives the sector of a garments large segment of its skilled labor and negatively affects the quality of overall management and performance. 4% 11% Remuneration The data presented below come from the main survey questionnaire as well as worker interviews. In total, 45% 25% 521 individual interviews of workers were conducted. At each firm, up to ten workers were interviewed with at least one worker from each major job position. The 2% 2% 6% 5% analysis focuses on the level and structure of remuneration and takes into account the components Source: World Bank, RPED Eritrea, 2002. of cash remuneration including wages plus overtime pay, bonuses, and other benefits. which white-collar workers in sectors of potential Wage Levels interest for the military were recruited. Mobilization for According to firm accounting data for 2001, the white-collar workers was significant in chemicals and structure of remuneration is such that wages account paints, construction materials, food and beverage, for about 84.5 percent of total cash earnings, overtime and textile, leather, and garments--all sectors pay accounts for about 8 percent, and the various requiring skills that can be applied to military bonuses or benefits account for about 7.5 percent. purposes with little additional training. Unlike what is found in many other African countries, An indirect confirmation of the impact of the overtime pay, bonuses, and other benefits are a small mobilization on the structure of employment is part of the total cash earnings for workers (about 15.5 provided by the exceptional importance of women in percent). For example, in Nigeria, on the basis of 2000 the Eritrean manufacturing sector. The average share wage data, these elements amounted to 35 percent of of women in manufacturing in Sub-Saharan Africa was cash earnings. about 38.5 percent in the late 1990s (World Bank Figure 2.7 shows the breakdown of average 2001a). In Eritrean manufacturing, women amount to monthly cash earnings by job function, expressed in 55.8 percent of the permanent employees, larger than dollar terms (values in current Nakfa are reported in the average, as well as larger than the rate found in Annex III, Table A.3.5). Monthly cash earnings in other RPED surveys in Africa. For example, in Côte April/May 2002 averaged $71.50.2 On average male d'Ivoire in the mid-1990s, women account for 7.8 workers earn about 67 percent more than female percent of the employed workers, and in Nigeria in employees. This difference decreases at higher levels 2001 they accounted for 9.5 percent. Such an of skill and education, and at the management level 2. Investment Climate in an International Perspective 19 Table 2.4. Mobilization by job position in 2001 (percent) Job position Chemicals Construction Food Furniture Metal Paper, Plastics Textile, Total and paints Materials and printing, leather beverage publishing and garments Management 0.00 17.39 2.22 0.00 0.00 0.00 0.00 0.00 2.56 Professionals 7.14 8.57 23.40 0.00 0.00 0.00 0.00 4.88 12.44 Skilled Production 6.67 13.19 8.18 16.22 0.00 0.00 0.00 0.00 1.97 Workers Unskilled 0.00 8.75 3.63 0.00 0.00 0.00 13.79 0.34 2.72 Production Workers Non Production 2.94 5.83 1.66 0.00 0.00 0.00 0.00 0.00 1.50 Workers Source: World Bank, RPED Eritrea, 2002. the women are actually paid more than their male production workers (close to 22 percent) and also counterparts (however, this finding should be treated very high for skilled production workers (close to 18 with caution given the limited number of female percent). These are very significant wage changes, workers in this category). given that the period of change is only five months. Labor regulations have been made simpler since There is no easily available index of prices, so it is independence in Eritrea, there is no minimum wage hard to know whether the wages also rose in real in the private sector, and wage scales have been terms. Anecdotal evidence, however, suggests that decompressed (World Bank 2002a). Eritrea is while prices have been rising, they are unlikely to have somewhat exceptional in this regard. Our survey risen as much as wages in a five-month period. results confirm this, showing that for 80 to 93 In addition, while labor shortages are widespread, percent of the interviewed firms, hiring procedures it seems that they are especially acute in metal and and layoff procedures are not a problem. In addition, paper, printing, and publishing, where the increase in almost 90 percent of the firms report having no earnings is the highest of all the sectors. Interestingly, problems with regulations dealing with the hiring of it seems that the smaller the firm, the larger the temporary employees. increase in earnings. The increase in earnings is also smaller for foreign-owned firms than for local ones. Recent Changes in Wages Among regions, Massawa area workers fared worst; Our data on wages also suggest that labor is scarce their wages actually registered a decline. and that this scarcity is driving up the cost of production. Between December 2001 and May 2002, Worker Surveys earnings increased sharply over the sample by 17 Additional results from worker surveys support the percent (see Figures 2.8­2.10). The increase in following conclusions (results are detailed in earnings was particularly great for unskilled Annex II): 2. Investment Climate in an International Perspective 20 Figure 2.7. Monthly cash earnings Education. Despite higher gross enrollment rates Male Total Female since the war, the data we collected suggest that US $ by position this strategy has not yet provided the 250 manufacturing sector with the properly educated workforce it needs. There is a huge gap in 200 training, because it was virtually nonexistent 150 during the Derg regime (1974­1991). 100 HIV/AIDS. The government, donors, and local groups are taking a great deal of initiative at an 50 early stage in the spread of this disease, which is 0 extremely encouraging. Nevertheless, there is still Profess- Unskilled Appren- a stigma strongly associated with the disease, ional Skilled prod. Non tice which is seriously inhibiting open discussion and Manage- prod. worker prod. ment worker worker behavioral change. Given the obvious implications for public health and the potential numerous Note: Computed on the basis of the wages provided economic impacts, this is quite disconcerting. by workers in April/May 2002 and converted into U.S. Wage determinants. Regressions confirm that dollars using the May 2002 official exchange rate of US$1 = 13.5 Nakfa. wages are in part determined by factors other Source: World Bank, RPED Eritrea, 2002. than worker characteristics, which suggests that the labor market is not operating on an entirely competitive basis. Figure 2.8. Changes in the wage rate (Percentage change, end by position 2001 Conclusion to May 2002) This section outlined the main characteristics of the manufacturing labor market in Eritrea, which is currently Percent in a difficult situation. Manufacturing employment has 25 been declining, while wages are probably rising in real terms, presumably reflecting the impact of military 20 mobilization. The distribution of employment and wage 15 changes remains uneven across regions and sectors. The manufacturing industry is dominated by the 10 Asmara area, and the textile, leather, and garments and food and beverage sectors. 5 It seems that recent events have strongly affected the structure of this market. The mobilization process 0 has substantially reduced the number of available Overall Profess- Unskilled white-collar workers and thus deprived firms of a sample ional Skilled prod. Non Manage- prod. worker prod. valuable asset at a difficult time. It has also led to an ment worker worker unusually large proportion of women in manufacturing, which is typical of a wartime economy. The most Source: World Bank, RPED Eritrea, 2002. critical issue at present regarding the labor market is 2. Investment Climate in an International Perspective 21 Figure 2.9. Changes in the wage rate (by sector) Percent 25 21.65 20.03 20 18.85 19.12 17.16 17.47 15 13.71 12.53 10 5 0 Overall Construction Food and Furniture Metal Paper/printing/ Plastics Textile, leather sample materials beverage publishing and garments Source: World Bank, RPED Eritrea, 2002 the labor shortage. This negatively impacts the current in the formal private sector. The reintroduction of performance of firms. Rising real wages are a factor in conscription for civilian purpose has, according to the high cost and low competitiveness of Eritrean anecdotal accounts, encouraged some young men to firms. They are also likely to reduce external go into hiding or even leave the country. Taken competitiveness when they are not offset by currency together, resolving the labor issue may be one of the depreciation or higher labor productivity. Thus, it government's greatest challenges, if it is to succeed in seems that a step-up in the pace of the current its strategy of growth and poverty reduction that relies demobilization program and the parallel establishment on the use of supposedly abundant yet productive of appropriate re-training programs for former soldiers low-cost labor to compete in the global markets. should remain at the top of the policy agenda. A word of caution should be added here. While Finance and Investment demobilization will undoubtedly increase the supply of The performance of firms in any country is labor, it may not completely solve the problem of labor inextricably linked to the functioning of financial shortage. Eritrea seems to have been continually markets. A properly functioning financial sector is a depopulated during the thirty-year war for prerequisite for the growth of enterprises and for independence as well as the recent border war with overall economic development. Eritrea is well Ethiopia. Skills of older workers have eroded, and positioned in many ways compared to other younger workers have never had the opportunity to be countries in Sub-Saharan Africa. Until 2000, there trained. In addition, a large number of families receive was adequate liquidity in the financial sector and remittances, easing the pressure to find employment an overall lack of corruption in terms of financial 2. Investment Climate in an International Perspective 22 Figure 2.10. Changes in the wage rate these costs are much lower compared to those faced (by size) (Percentage change, by enterprises in other African countries. However, end 2001 to May 2002) most of the finance is clearly being used to meet Percent working capital needs; investment is very low in 25 Eritrea, even when compared to the rest of Africa. This is likely to have a negative impact on future economic 20.83 20 19.33 19.73 growth. Government policies to promote investment 17.16 and laws encouraging longer-term credit flow into the private sector are needed to put Eritrea on a higher 15 growth trajectory. 11.38 Demand for Finance. Demand for finance from 10 firms consists of two components: finance for meeting working capital requirements and finance for fixed 5 investment in equipment, buildings, or land. For both these needs, and given the fungibility of capital, firms 0 can either use internal cash flow or borrow from Overall Very Small Medium Large capital markets by using overdrafts or bank loans. In sample small a perfectly functioning credit market, internal and Source: World Bank, RPED Eritrea, 2002 external capital would act as perfect substitutes. In reality however, this rarely occurs due to market imperfections. The extent of market imperfection is allocation (World Bank 2001b). However, due to the determined by the differential access to external transfer to the government of some old liabilities credit and cost of credit faced by firms. that the Central Bank of Eritrea had toward the In Eritrea, we see that firms' access to external Ethiopian banking system and the taking on of credit--bank loans and overdrafts in particular--has further government debt during 2001­2002, it is no declined in recent years (World Bank 2002a). Figure longer clear how liquid the system still is and 2.11 below examines the use of overdrafts by firms in whether the supposed availability of credit within our sample. We see that fewer than half of very small the banking sector translates into adequate access firms have access to overdraft facilities, and 45 to low-cost credit and satisfaction of demand at the percent of large firms use this type of credit. When firm level. broken down by sector, there are two sectors Our survey results indicate that a large proportion (chemicals/paints and textiles and leather) in which of firms in the private sector do have access to the over half the firms have access to overdraft facilities. formal financial sector, primarily through the use of Sharp differences exist across ownership types (Figure overdrafts. A significant number of small and medium 2.12). We see that the largest users of overdraft firms have current loans. Compared to other Sub- facilities are domestic private firms. All other types of Saharan African countries, private Eritrean firms are ownership, especially government owned and foreign much more likely to use bank borrowing privileges. firms, have access to internal financial resources not While firms complain about the stringent demand for available to others, therefore having a limited demand collateral, their inability to collateralize land, and the for overdraft facilities. PFDJ firms have overdraft high cost of interest payments, it is worth noting that facilities but are least likely to use them. 2. Investment Climate in an International Perspective 23 Access to loans also differs significantly across Figure 2.11. Overdraft facilities (by size) firm types. Figure 2.13 indicates whether or not firms % of firms with overdraft facility/line of credit currently have a loan, disaggregated by firm Average % unused over the last 6 months ownership. About 40 percent of domestic private firms Percent do not have a loan; a majority of all other types of firms 75 indicated that they do not currently have a loan. As to the reason, 100 percent of all firms (except domestic 61.54 55.00 55.56 privately owned firms) indicated that they did not 50 apply for a loan. Only 10 percent of domestic private 46.84 47.37 45.00 firms indicated that they had applied for a loan and were turned down. Of the firms that did not have a 29.53 loan and had not applied for a loan, 75 percent 25.00 25.25 25 indicated that they did not need a loan, and 17 percent argued that interest rates are too high. The fact that 92 percent of government-owned firms and 3.50 67 percent of partly government-owned firms do not 0 Sample Very Small Medium Large have loans is quite remarkable, leading one to small question how these firms are financed. That such a large number of firms overall did not Source: World Bank, RPED Eritrea, 2002. apply for a loan may indicate powerful deterrence effects, such as the problem of accessing land to count as collateral. With two state- and one PFDJ- Figure 2.12. Overdraft facilities (by ownership) owned bank that do not truly compete with each other % of firms with overdraft facility/line of credit (interest rates are essentially fixed on the deposit side Average % unused over the last 6 months and capped at 12 percent on the lending side), there Percent is real rigidity in the financial sector (World Bank 100 2002a). In addition to a need for further training of credit analysts, there are anecdotal reports of 83.33 significant government interference on how final 75 decisions on credit applications are made. We further examine the characteristics of loans 57.69 50.00 and overdrafts, focusing on the private domestic firms 50 only. This group is most likely to be credit constrained 33.33 and require formal sector borrowing. Figure 2.14 27.08 25.00 25 below presents the proportion of firms currently using 20.00 16.67 bank loans and overdrafts and compares them to 5.00 other Sub-Saharan African countries. 0.00 0 Given the tedious procedure required for Domestic Gov. Mixed Priv. Mixed PFDJ obtaining a loan compared to the relatively easy Private Local/ Priv. Foriegn Gov. applications for overdraft facilities, one would expect Source: World Bank, RPED Eritrea, 2002. firms to maximize their use of overdrafts before 2. Investment Climate in an International Perspective 24 applying for loans. Overdrafts are likely substitutes Figure 2.13. Loans for loans in this regard. However, if there is an overall Percent of firms currently having a loan credit constraint, then we would expect loans to 75 complement overdrafts. The data show that the likelihood of having a loan increases with firm size 60.78 and then falls for the largest firms. It also indicates 50.00 that overdrafts and loans are substitutes for each 50 44.87 other. A firm that has overdraft facilities is less likely to use loans as well. 33.33 The RPED survey also shows that almost three- 25 quarters of new investment is financed by retained 16.67 earnings or internal funds; the remaining amount is financed by local banks. There is virtually no access 8.33 to foreign banks (no firms in our survey report having 0.00 0 access to such loans). Collateral is required from Sample Foreign Mixed Mixed PFDJ almost all firms in the sample (Figure 2.15). The ratio private priv. priv. Domestic Govern- local/ govern- of collateral is very high for domestic private firms private ment foreign ment (172 percent) and high for other firms as well. The most common collateral provided is buildings, Source: World Bank, RPED Eritrea, 2002. followed by a combination of buildings and machinery, (Figure 2.16). Figures 2.17 and 2.18 show average interest rates Figure 2.14. Use of formal and informal and average duration of loans. The main concern with borrowing and lending these rates is the lack of variance; interest-rates do (private domestic firms only) not appear to take into account the varying degrees of Overdraft Bank loan risk represented by the firms in our sample. 0.8 Rather, the rates are remarkably similar across 0.69 sectors and firm types (which is the case for short- 0.61 0.62 term liabilities as well). This reflects the broader 0.6 0.58 0.52 problems of the financial sector--by using a 12 0.07 0.45 percent interest rate "ceiling" and paying depositors 0.4 0.30 0.21 5-6 percent, the financial sector has settled into 0.06 0.33 0.06 0.07 lending at a risk-invariant rate of 9 percent while 0.25 0.23 0.22 imposing substantial collateral requirements. In 0.2 other words, collateral, rather than interest rates, is being used to ration finance. It may well be the case 0 that the optimal scenario would include issuing T- Cameroon Ghana Tanzania Zimbabwe bills at about 4-5 percent, paying depositors a lower Eritrea Côte Kenya Zambia rate, and lending to firms according to market d'Ivoire criteria including risk. This may in fact lead to a Sources: RPED Surveys; Mazumdar and Mazaheri, 2002. substantial number of firms having access to money 2. Investment Climate in an International Perspective 25 Figure 2.15. Details on collateral required at a rate lower than 9 percent. However, this for the most recent loan approach denies loans to high-risk, high-return Collateral needed opportunities that almost surely exist and, if taken, Collateral ratio could be beneficial to the development of the Percent Eritrean economy. 200 Other results from the RPED survey are interesting as well. Checks and domestic currency 172 wires take an average of 1.4 days to clear, while foreign wire transfers take about 6.3 days (Table 120 A.4.1). External accounting experts audit about 90 100 100 100 101 percent of firms. Close to 90 percent of firms reported 90 that audited statements are necessary to obtain bank credit. Finally, domestic firms reported that about a third of debt was due within 90 days, another third was due within three to twelve months, and a final 0 third was due in over a year's time. Foreign firms Domestic Government Mixed private/ reported that about 50 percent of debt was due in 90 private Government days; the rest was long-term debt (which is generally Source: World Bank, RPED Eritrea, 2002. not accessible to domestic firms). Investment Characteristics in Figure 2.16. Types of collateral utilized, International Comparison Collateral provided for the most recent loan Table 2.5 compares the fundamental investment characteristics of firms in Eritrea to other countries in Building Guarantee from another bank Sub-Saharan Africa. We see that the percentage of Machinery Machinery and building firms in Eritrea that are currently investing is lower than in most countries in the comparator sample. Only Tanzania and Cameroon have a lower percentage of firms investing. The differences in the mean investment 23% to capital ratio are more striking. We see that Eritrea has the lowest average investment ratio (0.02). Even for firms who do invest, the average investment rate is only 6 percent. In other countries such as Tanzania, 10% 64% while only 19 percent of firms in the sample invested, those firms made large investments, with a mean 3% investment rate of 165 percent. Comparing differences across size classes, we see that investment increases with firm size up to the medium size category and then drops for the largest size class. Of those investing, firms in the small size Source: World Bank, RPED Eritrea, 2002. category have the highest investment rate (10 2. Investment Climate in an International Perspective 26 Figure 2.17. Average interest rates and percent), compared to less than 3 percent for other duration of the most recent loan (by size) size classes. Further investigation revealed that many Average interest rate of the firms that were not making investments in the Loan average duration (months) larger size classes were those that were recently 75 privatized and were heavily indebted. 60.0 Across ownership categories, we see that while 59.13 government and PFDJ firms are most likely to invest, their investment rates are much lower than those in 50 46.53 the private sector. Private firms that are investing are 41.78 making proportionately bigger investments than other firms. This may indicate a gradual movement of 25 resources away from the public sector. 12.0 9.81 Conclusion 9.75 9.71 9.94 10.0 There seems to be adequate liquidity in the financial 0 sector and an overall absence of corruption in terms Sample Very Small Medium Large of financial allocation. Interest rates are remarkably small similar across sector and firm types and therefore do Source: World Bank, RPED Eritrea, 2002. not appear to take into account varying degrees of risk. This situation impedes the possibility of a substantial number of firms having access to finance Figure 2.18. Average interest rates and at a rate lower than 9 percent. The flip side of this duration of the most recent loan (by ownership) situation is that a number of high-risk, high-reward opportunities are most likely being lost. A related Average interest rate development is Eritrea's low investment rate Loan average duration (months) compared to other African countries. 150 120.0 Access to Land According to the Land Proclamation of 1994, the 96.0 100 government owns all land in Eritrea. Eritreans are granted usufructuary rights. Domestic and international investors can obtain land leases, but the 60.0 government retains final dispossession rights. 46.53 50 41.83 In Figure 2.20 (next section), numerous business environment factors were rated on a scale of 1 to 5, 9.81 9.79 with one equaling "no problem" and 5 equaling 9.0 12.0 9.0 "severe problem" for operations. Access to land 0 Sample Domestic Foreign Govern- Mixed ranked on average between a "minor" and a private private ment Priv./Gvt. "moderate" problem. A closer examination of the answers indicates a bi-modal pattern, usually either Source: World Bank, RPED Eritrea, 2002. "not a problem" or a "severe" problem (Figure 2.19). 2. Investment Climate in an International Perspective 27 Table 2.5. Investment characteristics: Eritrea vs. other SSA countries Cameroon Côte Ghana Kenya Tanzania Zambia Zimbabwe Eritrea d'Ivoire Percentage of firms 27% 42% 46% 39% 19% 33% 72% 32% investing Investment /capital 0.06 0.08 0.07 0.04 0.03 0.03 0.07 0.02 Investment /capital 0.20 0.19 0.15 0.10 0.16 0.10 0.11 0.06 (Investing firms only) Note: Data for Eritrea are for 2000. Côte d'Ivoire data are for 1995. All other countries rates are for 1993. Source: World Bank, RPED. Table 2.6. Distribution of investment processing. Nevertheless, these numbers should be by firm characteristics interpreted with caution, since this question only applied to about 20 percent of the sample, and the answers were highly skewed by one or two firms (the By firm size Micro Small Med. Large median answer for time was 180 days and 500 N in % Investing 12.5% 35% 38% 29% costs). Anecdotal evidence shows that some I/K 0.02 0.10 0.01 0.02 enterprise owners and managers were frustrated with the lack of information on how their applications By ownership Priv. Govt. PFDJ Foreign for land allocation were processed. Those who had categories Dom. applied had no idea how long it would take, what % Investing 28% 38% 50% 20% stage they were currently in, or even what all of the I/K 0.10 0.02 0.01 0.01 stages were. This may lead to another reason for some firms reporting that they had "no problem" Note: I/K is the investment to capital ratio. obtaining land. Having seen from others how difficult Source: World Bank, RPED Eritrea, 2002. it is to get land, they simply decided to make do with the land they already had. To such firms the What could explain this bipolarity of responses? application process would seem opaque and carried One possibility is that the results have been out in a discretionary manner. influenced by the fact that a majority of the 35 privatized firms in Eritrea were included in the sample study. These companies had unique C. Regulatory Constraints and privileges regarding access to land and other Infrastructure resources prior to their transfer from government ownership. Many other firms had a much harder time A number of improvements have been made in the getting access to land. On average, firms that were business climate by the GOE since independence. able to obtain land within the last five years had to These achievements give credence to their expressed wait 316 days and had to pay about 127,185 N for commitment to encouraging the growth of the private 2. Investment Climate in an International Perspective 28 Figure 2.19. Access to land Macroeconomic Instability --severity of obstacle The term macroeconomic stability was defined to Number of firms interviewees as the stability of the inflation and 30 exchange rate. The evidence shows that instability is 28 perceived by the private sector as the largest factor dragging down performance. These findings roughly correspond to the reasons firm owners and operators 20 gave for low capacity utilization (discussed in Section A). Respondents were asked specifically about the 15 impact of the lack of foreign exchange on capacity utilization, and the majority (63 percent) classified it as 10 "severe." Firms seemed more optimistic about the future, however. Specific questions were asked regarding the 3 2 2 firm owner/manager's expectations of macroeconomic 0 indicators during the coming years. Those firms that No Minor Moderate Major Severe answered seemed to be quite hopeful, but many problem hinged their answers on the understanding that Source: World Bank, RPED Eritrea, 2002 nothing would disturb the current peace process between Ethiopia and Eritrea. Inflation is expected to begin to come down to an sector. Among other issues, progress has been made average of 10.8 percent beginning in 2003 (these on privatization, reforming the investment code, and figures reflect the inflation rate in Asmara, rather than business license procedures. Nevertheless other the lower national average). On average, these firms factors are holding back efficient operation and thought that the official exchange rate would continue growth of enterprises. During the RPED survey, rising, but slowly to 15.2 N/$ by the end of the year owners and managers of manufacturing firms were 2002 and to about 16 N/$ in 2003. These figures belie asked both qualitative "ranking" questions and the fact, however, that large numbers of firms (on quantitative "cost" questions regarding infrastructure average 24 percent of the respondents) were so and government regulations.3 This section covers unsure of the future that they could not even estimate general business regulations and infrastructure, in these figures. This type of uncertainty is extremely order of importance for the manufacturing firms in our detrimental to the ability of firms to make decisions sample. Next, the supply and quality of business regarding production levels, financing, and investment. services are examined. Finally, the firms' experiences with regulations specific to trade are presented. Infrastructure The results of the qualitative ranking questions The war for independence had a devastating impact on business regulation and infrastructure are on infrastructure. Since independence the country illustrated below and reveal the private sector's had made considerable improvements in roads, perception of areas for improvement as well as power, water, and telecommunications. As a result, some of the positive aspects of the investment infrastructure as a whole ranked as only a "minor" climate (Figure 2.20). constraint to capacity utilization in Eritrea (see 2. Investment Climate in an International Perspective 29 Figure 2.20. Business environment Figure 2.5 earlier in this chapter). This is at variance with typical SSA countries where infrastructure Business usually ranks among the top three obstacles to licensing expanding production--the other two being Crime, theft, corruption and tax regulation/administration disorder (Brunetti, Kisunko, and Weder 1997). Nevertheless, Corruption the survey does point to a few areas that could use some improvement, especially the reliability of Labor regulation electricity supply and telecommunications. Anti-competitive practices Electricity Cost and access to electricity is particularly important Transport for the manufacturing sector. Although the government has made good progress on power supply, with the Tax admin opening of the new Hirghigo Power Plant in Massawa, Customs/ there are anecdotal reports that outages due to testing regulations of the plant and network have interfered with production. Some firm owners also informed the RPED team that Telecom whereas in the past, the national power company would Economic policy notify them of upcoming tests, they rarely do so now. For uncertainty enterprises producing certain continuous-process items Tax rates such as plastic and soap, the unexpected cessation of power can lead to weeks without production while Access to land machines are cleaned. While 27 percent of the sample answered that electricity was not a problem, 18 percent Access domestic credit considered it a "severe" problem and a further 11 and Worker skills/ 15 percent, "major" and "moderate," respectively. education In addition, the average firm owner had to wait 99 days to obtain an electrical connection, although the Electricity distribution of answers was skewed by one answer and a more meaningful number is probably the Cost of financing median, 30 days. From conversations with the firm Macroeconomic owners and managers, obtaining electricity seems instability particularly time consuming and bureaucratically 0 1 2 3 4 difficult for higher capacity industrial connections. Degree of problem 0 = No problem According to sample firms, there was an average 1 = Minor of nine electricity interruptions (median equals five 2 = Moderate 3 = Major days) a month during the year 2001. It is not 4 = Severe surprising then that 43 percent of the sample firms own a generator from which they obtain an average of Source: World Bank, RPED Eritrea, 2002. 22 percent of their electricity. In towns outside of 2. Investment Climate in an International Perspective 30 Asmara, a slightly higher percentage of firms (46 its performance. In addition, and as a result, percent) have generators. As expected, the average retrenchment during the privatization process was less. estimated cost of supplying electricity with generators The perceptions of the manufacturing firms in our (2.00 Nakfa /KwH) is double than the average sample attest to these improvements as well. A reported cost of electricity from the public grid (1.00 majority (85 percent) of respondents stated that labor N/KwH, about 0.07 $US/KwH).4 Thus, the uncertain regulations were not a problem. When asked about supply of electricity imposes higher business costs, specific types of labor regulations (hiring local or both to acquire a generator and to operate it when the foreign workers, procedures and cost of public grid is not supplying electricity. retrenchment, limits on temporary hiring, and so on), the average answer was consistently that they were Water not obstacles. Although some firms complained about On average most firms answered that water was not a the burden of providing six months of pay when a problem at all. However, 48 percent of firms reported worker is fired (18 percent of the sample said that using a private well to obtain water for production, and "layoff procedures and cost of retrenchment" were a firms reported an average of seven interruptions to the problem), most firms agreed that the labor regulations public water supply each month during 2001. were not the problem and repeated that the major problem was the lack of workers. Telecommunications Some firm managers/owners told RPED The message obtained through the survey on enumerators that they understood that businesses telecommunications is mixed. Although in Figure 2.3 could apply in writing to retrieve essential skilled the majority of respondents reported no problems with workers from the front, but that even large companies telecommunications, the average (median) firm had to had not been successful in their attempts to do so. wait 65 days to obtain a mainline telephone Even the Crash Program for Exports was not connection. In addition, there was a reported average successful in getting back the key workers in the of seven interruptions to telephone service in these factories participating in the program. Consequently, sample firms per month during 2001. smaller companies were discouraged from even trying. Some firms expressed the belief that Labor Regulation and Business Licensing exceptions to the rules were made for government- Labor regulation and business licensing have been owned firms and ministries and more skilled workers areas of activity during recent years for the were allowed to stay. Demobilization will obviously do government, and the progress clearly shows. As the most to quickly increase the pool of labor, and discussed in greater detail in Section A on productivity skilled labor in particular. At that point, the labor and Section B on the labor market, the Eritrean labor regulations in place will be tested to a greater extent market has been significantly liberalized since than they have been so far. independence. At independence, wages were not There is evidence that efforts to streamline the market-determined, firing was more difficult, and the licensing of businesses have had some success as civil service was inefficient, partly due to too many well. Generally, the private sector sees the license as a employees. Hiring practices and regulations on comprehensive instrument, which is reducing the retrenchment in particular have improved. These amount of time required previously to obtain various adjustments, along with reorganization and refocusing licenses. For those firms established within the last five of the civil service, helped the government to improve years, the average number of licenses required was 2. Investment Climate in an International Perspective 31 indeed one. The average answer for the time it took to complaints that party-owned firms, being closely get a license was two days (for those firms that were linked to the government, have occasionally used established during the last two years). One should bear their "pull" to get things done. in mind, however, that the RPED approach only covers With future economic growth, greater existing businesses that were by definition successful opportunities will likely emerge involving larger sums in overcoming these hurdles. Other assessments have of money and the relatively more recognizable types pointed out that potential investors are still having of corruption problems (bribery) could increase. At business license applications turned down for arbitrary the moment, however, the absence of bribery is a reasons (World Bank 2002a). However, overall, the bright spot in the business outlook for Eritrea and average time spent by senior management on gives it a comparative advantage vis-ŕ-vis its potential government regulations was 5 percent, much lower competitors in much of Africa. than RPED results in other Sub-Saharan countries. Business Services Corruption and Crime While most firms thought domestic business services Not surprisingly, for anyone familiar with Eritrea, were available and affordable, the perceptions of corruption and crime/theft/disorder are considered their quality varied (Figure 2.21), with accounting virtually nonexistent. The majority of respondents (84 and engineering/design services rated at the good percent) in this ranking exercise answered that end of the scale and information technology (IT) at corruption is not a problem. This was corroborated by the poor end. other questions throughout the survey. One asked Firm owners/managers describe IT consulting specifically about the "total value of gifts or bribes services as inadequate because when follow-up required" during inspection visits by government support after hardware installation is available (which officials from the Health Ministry, Fire Inspector, or is not always the case), it often does not fix the Labor and Social Security Ministry. The universal problems encountered by the firm. answer was zero. When asked in general what A number of firms expressed the need for more percentage of revenue is used for informal payments to competition in the provision of insurance to the state- "get things done," the answers averaged 0.2 percent. owned National Insurance Corporation of Eritrea (NICE). By comparison, in a global private sector survey There are no regulatory barriers to the entry of conducted in 1997 by the World Bank, results for Sub- independent insurance agencies; indeed, one domestic Saharan Africa showed that on average 50 percent of company is reportedly preparing to begin operations in entrepreneurs were frequently asked for these types of insurance, Augaro Financial Services. The war, however, payments (Brunetti, Kisunko, and Weder 1997). has undoubtedly made Eritrea a less-than-desirable The difficulty in assessing the presence of market at present for foreign investors in this sector, as corruption lies in the fact that a more complete has the restriction on advertising for foreign companies. definition of corruption includes more than simply the payment of bribes. Preferential treatment given Trade Regulations through discretionary power in the awarding of As discussed in Chapter 1, the Government of Eritrea contracts or allocation of land or foreign exchange is has taken various measures to try to increase trade also considered corruption and can have quite since independence. At first the value of merchandise distortionary effects on the growth of the private exports increased from US$36.1 million in 1993 to sector. There are hearsay, but frequently heard, US$95.3 million in 1996. But during this period Eritrea 2. Investment Climate in an International Perspective 32 was not successful in diversifying manufacturing Figure 2.21. Business services: export destinations away from Ethiopia. With the onset perceptions of quality of war with Ethiopia, Eritrea's exports collapsed, bottoming out at only $20 million in 2001. IT services 2.6 Only 14 percent of the firms surveyed reported Management any exports. None of these are government- or PFDJ- 2.9 and marketing owned firms. These firms are of all size classes and Insurance 2.9 exported on average 77 percent of their sales. Not surprisingly, these firms are mostly in the food and Legal services 3.0 beverage sector and the textile/leather/garments Engineering sector, two of the largest sectors within the sample 3.1 and design (see discussion of the sample in Annex I). Accounting 3.3 Only a very small number of firms used the available trade incentives. Only three firms reported 1 2 3 4 using the duty drawback scheme and the same 1 = Very poor number used the duty exemption on imported inputs. 2 = Poor 3 = Good Only two establishments used the duty exemption for 4 = Very good imported capital goods. It is not clear what the reason is behind the underutilization of these provisions. Source: World Bank, RPED Eritrea, 2002. Perhaps the firms are not fully aware of what incentives are available to them or perhaps they feel the incentives are not worth the time and effort Some firms remarked that there has been a required to take advantage of them given the relatively significant improvement over the last few years. Yet low import tariffs. Duty drawback systems have not others said that so much of the dealings with the worked well in most developing countries. Duty customs agency depended on "whom you know" and exemption was used on a case-by-case basis for the whether your company is "well known" by customs firms participating in the Crash Program on Exports. officials. On the basis of the data, however, the There is now a duty exemption regulation drafted, but conclusion would have to be that import customs are it has not yet been passed into law. not a binding constraint for most firms. But the cost of A majority of the sample firms, 71 percent of the imports can be high because of the unavailability of sample, imports inputs directly, and 59 percent of the foreign exchange at the official rate. Some companies firms used a clearing agent for imports. Firms ranked pointed out that even when they can afford to buy "customs procedures" between "no problem" and a foreign exchange at a higher rate on the parallel "minor" problem (Figure 2.20). Data on the length of market, they lose money because the import invoice time taken for imports to clear the customs suggest lists the official foreign exchange rate. This that there is some room for improvement. The average underestimates the cost of inputs, artificially raising (mean) wait for imports last year (from the moment they profits and therefore taxes. One firm pointed out that arrive at the port to when they have cleared all they even had to pay a 2 percent bank fee on foreign customs procedures) is 12 days and the average exchange transactions to get goods out of customs, longest wait was 23 days. For exports the figures are 4 even though they had bought the currency on the days (average wait) and 4 days (average longest wait). black market. 2. Investment Climate in an International Perspective 33 Expectations of Firm Growth Notes Manufacturing company managers/owners are even more optimistic about their own growth potential than 2. These numbers are computed using the official the economy's ability to recover. Generally, the exchange rate that prevailed in April/May 2002. percentage of sales growth expected for 2002 is 25 The exchange rate on the parallel market may percent and for next year it is 44 percent. Firms plan have been substantially different at the time. to increase investment by 27 percent during both 3. Despite the limitations of qualitative ranking years. Imports for the same time periods are expected questions, especially concerning the ability to to increase by 30 percent and 34 percent, compare data across countries, the RPED survey respectively. Not surprisingly, the most optimism was considers them to be a useful gauge to contribute in the textile/leather/garments and food/beverage to policy priority discussions at a national level. sectors. However, these estimates hinge on the 4. The cost for publicly supplied electricity is expectation that peace with Ethiopia will continue, obtained by taking all charges (fixed costs and trade will resume, and demobilization will be those dependent on consumption) on a typical completed soon. monthly bill and dividing by KwH consumed. The privately supplied electricity cost is generally estimated by asking the value of the equipment, Conclusion the number of operatives and their cost, the cost of servicing in terms of parts and labor, and the cost What should be kept in mind is that the obstacles of fuel and the amount used per week. outlined in this and previous sections are not static. Simply because certain infrastructure services or regulations are currently viewed as sufficient does not mean this will remain the case. Once current obstacles are overcome to a reasonable extent, and the private sector begins to grow, there will be pressure on the other services and resources, which may then become the main obstacles to subsequent growth. For example, as noted in a World Bank study of the services sector in Eritrea, competition-promoting policies are still in their infancy but may well become a priority as markets develop, (World Bank 2002b). As the economy grows, there may also be greater opportunities for corruption and crime to emerge. As the demobilization program proceeds and more skilled labor is freed up, there could very well be more demand for investment and credit, which may or may not be accessible in adequate amounts. It is for these reasons that RPED surveys are generally repeated with an interval of three to five years, to track the shift in the main obstacles to the private sector. 3. Conclusions and recommendations 34 3. Conclusions and Recommendations 35 A number of general and specific recommendations additional protection. New export earners will emerge, arise from the analysis and conclusions of this study, further increasing the available foreign exchange and (Figure 3.1). The two key obstacles to resuming thus help stabilize both nominal and real exchange growth seem to be the lack of foreign exchange and rates in the future. the severe shortage of labor (Figure 2.5). As these obstacles are removed, another set of constraints will Supply of Labor likely become binding. The survey suggests that the The results of the survey analysis clearly indicate that top four obstacles in this "second generation" are: the current overarching constraint is labor. The labor shortage affects firms' ability to grow and become Infrastructure constraints, such as unreliable competitive, either in the regional or international electricity supply, poor telecommunications, and marketplace. Due to the severe shortage of labor, high costs for sea and air transport; firms are relatively less productive in Eritrea. The ratio Limited access to finance at a reasonable cost; of capital to labor is much higher than the optimal Low levels of education and skills in labor force; ratio, and wages have been rising. Unit labor costs, a and rough indicator of competitiveness, show that Eritrean Limited access to land. labor is expensive relative to labor in other parts of the world. When coupled with other constraints Foreign Exchange highlighted in the report, the labor shortage creates a In the first instance the lack of foreign exchange is severe drag on competitiveness in the private sector. mostly a macroeconomic issue. Above all, business It may well be the case that once labor is freed up and people want a predictable environment, which means other key problems mentioned in this report are slow and predictable changes in exchange rates, addressed, there will be a surge in productivity and interest rates, and rates of inflation. This can only come exports of existing firms and a high demand for new about from macroeconomic stabilization. Until that is investment. Immediate and effective demobilization is achieved, inflation rates will continue to be high (double necessary if the private sector in Eritrea is to function digit) and variable. That in turn will play havoc with normally. Demobilization will also make it easier to exchange rates and interest rates. While stabilization in achieve macroeconomic stability. the current situation in Eritrea will probably result in a major devaluation of the official exchange rate, it will Infrastructure likely bring a number of benefits. Net earners or Power, telecommunications, and transport are not receivers of foreign exchange will be better off. Foreign currently the most important binding constraints but aid will put greater resources in terms of Nakfa at the may well become more problematic once labor is disposal of government. Receivers of remittances will freed up and the private sector starts growing. The be better off. Flow of remittances and foreign private uncertain supply of power already imposes some investment is likely to rise as confidence returns. extra costs on firms and may quickly emerge as a Supply of foreign exchange will no longer be much of a larger constraint once the labor situation improves. constraint on private businesses. Costs of imported Telecommunication reforms need to be pushed inputs will increase but will now be correctly reflected vigorously to set the framework for a much expanded on the firms' balance sheets. To the extent that landline and wireless access to the world. Costs of domestic production competes with imports, the higher sea and air transport are high and need to be brought cost of foreign exchange will give the enterprises in line with those of the neighboring countries. 3. Conclusions and Recommendations 36 Access to and Cost of Finance authorities on existing applications for land is urgently At a general level, macroeconomic stabilization, by needed; the government needs to sort out exactly reducing the crowding out of private sector by the what is holding things up from the point of view of public sector, should increase the availability of administrative barriers. Using market mechanisms to loanable funds and decrease the average level of price the leasing of land may make sense as well; one interest rates. At a specific level, financial allocation possibility is to auction off leases for land. In most needs to take into account varying degrees of risk. market economies, there is a "rent gradient"; that is, Low-risk investment will then benefit from lower center-city land is much more expensive than land on interest rates, and high-risk investment will at least get the outskirts of a city. Asmara would probably be no the opportunity to succeed (or fail). Banks need to different once bureaucratic obstacles to pricing of restructure to achieve this as well as increase land are removed. Ease of exit from leases is just as profitability. This should also lead to lower collateral important. If sub-optimal decisions are made requirements, at least in some cases. In addition, regarding leasing of land, entrepreneurs should be more training in credit assessment would facilitate this able to sell their leases just as easily as they buy them. shift in operating procedure. If these adjustments are An active and efficient market for land is crucial to the not made, finance could very well be an even larger growth of the private sector. Valuable capital obstacle for existing and potential firms once the labor resources should not be tied up in land indefinitely supply is increased, especially where long-term due to lack of a functioning marketplace. investment is concerned. Educated and Skilled Labor Force The study found both direct and indirect evidence of the negative impact of low levels of education and skills of the Eritrean workers. At the direct level, employers complained that even when workers are available, they were not easily trainable because of a generally low level of education and experience in jobs that require skill. Indirect evidence comes from the fact that the unit labor costs (ULCs) are high when compared to successful exporters. A strong rise in productivity would reduce ULCs, because ULCs in domestic currency are just a ratio of nominal wages to labor productivity. But greater productivity is a function of better education and greater skills. Access to Land Our survey shows that some firms have serious problems with access to land; others report that they do not have a constraint. It is very clear that the administrative process for accessing land is problematic. Aggressive follow-up by the relevant 3. Conclusions and Recommendations 37 Table 3.1. Eritrea investment climate policy matrix Constraints Proposed solutions Planned actions Constraint Specific Degree of Recom- Political Technical Period of Active WB Planned Other Category Constraint Importance mended Difficulty Difficulty Implemen- Actions WB Action Donor Action Solutions tation Macro- Foreign Very High Macro- High Low Medium CEM Continue Dialogue economic Exchange economic Term Dialogue Dialogue Instability Rate stabilization Continue to Instability implement ERP Labor Shortage Very High Demobiliz- Very High High Medium to Economic Full imple- Multi-donor Market of Labor ation Long Term Recovery mentation of demobiliz- Program demobiliz- ation (ERP) ation program program Labor Education High · Training for Low Low Medium Demobiliz- Launch Market and Skills demobilized Term ation education of Workers solders program project · Increased general level of education Infra- Consistency Very High Reform of Low Low Short Term ESW on Launch of Financing of structure of Supply the power education; power power of Electricity sector planning/ project generation preparation for education project Finance Access Very High · Introduce High High Short Term Dialogue to Finance private with banks government · Restructure re: power Financial sector project · Training in credit assessment Land Access High · Identify Very High High Long Term to Land administrative barriers · Establish market-based mechanism for leasing land References 38 Bloom, David, Ajay Mahal, Jaypee Sevilla, and River UNAIDS/WHO (UN AIDS Program/World Health Path Associates. 2001. "AIDS and Economics." Organization). 2000. Epidemiological Fact Sheet, Paper prepared for Working Group One of the 2000 Update. Geneva. WHO Commission on Macroeconomics and ----. 2002. Epidemiological Fact Sheet, 2002 Health, November. Update. Geneva. Brunetti, Kisunko, and Weder. 1997. "Institutional World Bank. 1994. Eritrea: Options and Strategies for Obstacles for Doing Business: Data Description Growth. Vol. 1. Washington, D.C. and Methodology of a Worldwide Private Sector ----. 1996. Eritrea, Poverty Assessment. Survey", The World Bank. Washington, D.C. Bruchhaus, Eva Maria, and Ammanuel Mahreteab. ----. 2001a. World Development Indicators. 2000. "`Leaving the Warm House': The Impact of Washington, D.C. Demobilization in Eritrea," in Demobilization in ----. 2001b. Financial Sector Report. Washington, D.C. Sub-Saharan Africa: The Development and ----. 2002a. Revitalizing Eritrea's Development Security Impacts, ed. Kees Kingma. New York: Strategy. Washington, D.C. Processed. St. Martin's Press. ----. 2002b. Assessment of Policy and Performance EIU (Economist Intelligence Unit. 2001. Eritrea: in Eritrea's Service Sector. Washington, D.C. Country Profile 2001. London. Processed. ----. 2002. Eritrea: Country Report 2002. London. Eritrea, State of. 2001. Transitional Economic Growth and Poverty Reduction Strategy, 2001­2001. Asmara. IMF (International Monetary Fund). 2000. Eritrea: Statistical Appendix. Staff Country Report 00/55. Washington, D.C. ----. 2001. The Gender Gap in Education in Eritrea in 1991-1998: A Missed Opportunity? WP/01/94. Washington, D.C. Levy, P. S., and S. Lemeshow. 1991. Sampling of Populations, Methods and Applications. New York: John Wiley and Sons. Lindauer, David, and Ann Velenchik. 1994. "Can African Labor Compete?" in Asia and Africa: Legacies and Opportunities in Development, eds. David Lindauer and Michael Roemer. San Francisco, Calif.: ICS Press. [[Mazumdar and Mazaheri. 2002. Please give full cite. Cited in figure 2.14 in short form.]] Ministry of Industry and Trade. 2001. Listing of Formal Establishments. Asmara, Eritrea. Annexes 39 Annex 1. Sampling Methodology 40 Annex 2. Results from the Worker Survey 49 Annex 3. Supplemental Labor Tables 56 Annex 4. Supplemental Finance Tables 68 Annex 5. Standard Investment Climate Tables 69 Annex 1. Sampling Methodology 40 In the early 1950s, when Eritrea was federated with (50­99 employees), and "Large" (100 and more Ethiopia, it had a relatively sophisticated urban and employees). industrial infrastructure. However, this structure was Table A.1.1 and Table A.1.2 show the breakdown downgraded by the initial reorientation of industrial of the sample frame among regions, sizes, and development toward Ethiopia and the consequences sectors. They also show employment figures. The of the various wars. importance of the Asmara region is immediately At present, the government is striving to achieve apparent. About 82.6 percent of the firms and 86.0 an economy led by the private sector. However, to percent of the employment in the manufacturing date, the private sector has tended to concentrate on sector are concentrated in this region. All sectors imports and distribution, receiving little investment have firms in the Asmara region. and generating few exports. The manufacturing Not surprisingly, the Eritrean manufacturing sector sector accounted for about 7.8 percent of GDP has an uneven size distribution (Table A.1.2). While in1999, a level similar to the average share of most enterprises are "small" (about 57 percent of the manufacturing in GDP for low-income Sub-Saharan firms have between 10 and 49 employees), they only Africa countries (9.8 percent over the 1995­1999 account for 18.3 percent of the employment. However, period). Total employment in manufacturing is the 33 firms with more than 100 employees provide estimated at 25,000 workers in 2000 (EIU 2001). the bulk of the employment. As a whole, large firms account for 68 percent of total employment. Description of the Sample Frame In short, the sample frame exhibits two prominent The sampling design strategy was--as usual in RPED characteristics: the weight of the Asmara region and surveys--one of a stratified random sample. The the dominance of firms with more than 100 workers in Ministry of Industry and Trade provided RPED with a total employment. This sample frame accounts for list of 222 officially registered companies with more about 65 percent of total employment in than 10 employees for 2000.5 Fourteen firms were manufacturing. added to this list, on the basis of the information available in other registries (mainly the directory of the Sample Derivation Chamber of Commerce and the listing of officially Clusters were defined on the basis of location, registered businesses). In addition, 12 firms were size, and sector. This three-level stratification removed, leaving a final sample frame of 224 formal implied 48 clusters, each of Nh elementary units . manufacturing firms.6 Firms were randomly selected inside most of the The sample frame was then stratified by location, clusters (nh firms for each cluster). For a small sector, and size. Four regions were defined: number of very small clusters, all firms were "Asmara," "Massawa," "Keren," and "Other." Similarly, selected (Table A.1.3). eight broad sectors were defined: "Chemicals, Paints Due to the small size of the sample frame and the and Pharmaceuticals," "Food and Beverage," "Metal," three levels of stratification, there was a risk of (i) "Paper/Printing and Publishing," "Plastics," "Textile, missing key companies due to the small size of many Leather, and Garments," "Furniture," and clusters, and (ii) having a biased distribution of "Construction." Finally, to compare Eritrea to other employment. With these concerns in mind, firms RPED datasets for various African countries, four located in the "Other" regions were not sampled due standard size classes were used: "Very Small" (0­10 to their marginal impact, both in terms of number of employees), "Small" (11­49 employees), "Medium" firms and employment. In addition, for the regions to Annex 1. Sampling Methodology 41 Table A.1.1. Sample frame (by region and sector) Asmara Keren Massawa Other Total Area Area Area Areas Chemicals/Paints/ Number of firms 12 1 13 Pharmaceuticals Total employment 434 11 445 Average nber of employees/firm 36.2 11.0 34.2 Standard deviation nber of empl. (30.0) na (29.6) Construction Material Number of firms 35 4 4 1 44 Total employment 2,106 67 449 12 2,634 Average nber of employees/firm 60.2 16.8 112.3 12.0 59.9 Standard deviation nber of empl. (101.8) (5.0) (67.9) na (94.9) Food and Beverage Number of firms 52 6 6 12 76 Total employment 3,666 673 342 421 5,102 Average nber of employees/firm 70.5 112.2 57.0 35.1 67.1 Standard deviation nber of empl. (166.8) (247.8) (63.6) (77.3) (156.6) Furniture Number of firms 20 2 22 Total employment 589 223 812 Average nber of employees/firm 29.5 111.5 36.9 Standard deviation nber of empl. (20.1) (98.3) (37.5) Metal Number of firms 16 16 Total employment 829 829 Average nber of employees/firm 51.8 51.8 Standard deviation nber of empl. (51.3) (51.3) Paper/Printing/ Number of firms 10 1 11 Publishing Total employment 483 15 498 Average nber of employees/firm 48.3 15.0 45.3 Standard deviation nber of empl. (72.1) na (69.1) Plastics Number of firms 7 7 Total employment 155 155 Average nber of employees/firm 22.1 22.1 Standard deviation nber of empl. (13.8) (13.8) Textile, Leather and Number of firms 33 1 1 35 Garments Total employment 5,705 35 25 5,765 Average nber of employees/firm 172.9 35.0 25.0 164.7 Standard deviation nber of empl. (344.4) na na (335.8) Total Number of firms 185 11 13 15 224 Total employment 13,967 751 1,049 473 16,240 Average nber of employees/firm 75.5 68.3 80.7 31.5 72.5 Standard deviation nber of empl. (182.0) (182.4) (67.7) (69.0) (171.7) Note: Based on an MoIT list of the 2001 industry survey which reports 2000 data. Source: World Bank, RPED Eritrea, 2002. Annex 1. Sampling Methodology 42 Table A.1.2. Sample frame (by size and sector) Very Small Medium Large Total Small (10-49) (50-99) (100 (0-10) and +) Chemicals/Paints/ Number of firms 10 2 1 13 Pharmaceuticals Total employment 201 141 103 445 Average nber of employees/firm 20.1 70.5 103.0 34.2 Standard deviation nber of empl. (9.8) (16.3) na (29.6) Construction Material Number of firms 5 26 7 6 44 Total employment 50 568 513 1,503 2,634 Average nber of employees/firm 10.0 21.8 73.3 250.5 59.9 Standard deviation nber of empl. (0.0) (12.3) (17.8) (149.6) (94.9) Food and Beverage Number of firms 22 39 5 10 76 Total employment 220 879 335 3,668 5,102 Average nber of employees/firm 10.0 22.5 67.0 366.8 67.1 Standard deviation nber of empl. (0.0) (12.5) (13.5) (295.2) (156.6) Furniture Number of firms 3 14 4 1 22 Total employment 30 354 247 181 812 Average nber of employees/firm 10.0 25.3 61.8 181.0 36.9 Standard deviation nber of empl. (0.0) (11.8) (12.8) na (37.5) Metal Number of firms 1 10 2 3 16 Total employment 10 253 151 415 829 Average nber of employees/firm 10.0 25.3 75.5 138.3 51.8 Standard deviation nber of empl. na (8.0) (29.0) (53.4) (51.3) Paper/Printing/ Number of firms 2 7 1 1 11 Publishing Total employment 20 146 90 242 498 Average nber of employees/firm 10.0 20.9 90.0 242.0 45.3 Standard deviation nber of empl. (0.0) (7.9) na na (69.1) Plastics Number of firms 2 5 7 Total employment 20 135 155 Average nber of employees/firm 10.0 27.0 22.1 Standard deviation nber of empl. (0.0) (13.5) (13.8) Textile, Leather and Number of firms 3 17 4 11 35 Garments Total employment 30 432 253 5,050 5,765 Average nber of employees/firm 10.0 25.4 63.3 459.1 164.7 Standard deviation nber of empl. (0.0) (9.2) (4.6) (493.6) (335.8) Total Number of firms 38 128 25 33 224 Total employment 380 2,968 1,730 11,162 16,240 Average nber of employees/firm 10.0 23.2 69.2 338.2 72.5 Standard deviation nber of empl. (0.0) (11.2) (14.9) (342.9) (171.7) Note: Based on an MoIT list of the 2001 industry survey which reports 2000 data. Source: World Bank, RPED Eritrea, 2002. Annex 1. Sampling Methodology 43 be surveyed, all of the "large" firms and most of the The sample and sample frame distributions differ "medium"-sized firms were kept in the sample, as significantly for a few sectors (Figure A.1.2.). Two were the firms in the "Plastics" sector. The rest of the extreme cases are the "Food and Beverage" sector, firms were sampled randomly from each of the whose share in employment is overestimated in the remaining clusters. A theoretical sample of 98 firms sample while the importance of the "Construction" was thus drawn with a sampling rate of 43.7 percent sector is probably larger in reality. (Table A.1.3). In addition, as shown by Figure A.1.3, the sample The sample effectively surveyed differs slightly reproduces the distribution of employment by region from this theoretical sample but retains its major reasonably well. characteristics, as well as those of the sample frame: The total sample includes 9,057 workers, which the importance of firms with more than 100 workers account for about 56 percent of the employment in the and the prominence of the Asmara region. The sample frame. Available statistics suggest that the sample actually surveyed is 79 firms, giving a employment represented in the sample is also about sampling rate of 35.2 percent. In the sample 36 percent of Eritrea's total manufacturing surveyed, firms with more than 100 workers account employment in 2000. In short, the distribution of for 72 percent of the employment. Also, 83.5 percent employment in the surveyed sample is probably quite of sample enterprises are located in the Asmara correct with respect to the size of firms and their region. In addition, the Asmara region accounts for location; however, the "fit" is imperfect for the sectoral 83.6 percent of the employment in the sample (Table distribution. The sample surveyed takes into account A.1.4 and Table A.1.5). the dominance of "large" firms and the importance of The difference between the theoretical sample and the Asmara region. Hence, this sample can be the sample that was actually surveyed came about considered reasonably representative of the main because some firms refused to be interviewed and characteristics of the manufacturing sector in Eritrea others had either closed or changed sectors. In total, given available statistics. about one-fifth of the firms listed in the theoretical sample were no longer relevant for the survey exercise. Some of these "missing" firms were replaced Notes with "new" firms with the same characteristics.7 To confirm the validity of the surveyed sample, 5. This listing provided the name, address, activity, the different employment distributions of the sample and number of employees of formal frame and the surveyed sample were compared. The manufacturing firms existing in 2000 that had comparison of the employment distribution by size of been interviewed during a survey in 2001. firm, sector, and location is illustrated in the next 6. Firms with activity not directly related to three figures. manufacturing or that had disappeared in the Figure A.1.1 shows the distribution of employment meantime were eliminated. among the different size classes. Clearly, the 7. Replacements had to belong to the same sector of distribution in the sample does not differ greatly from activity, to be of a similar size, and to operate in the distribution in the sample frame. "Large" and the same location. "medium" firms do have, however, a slightly larger share in the sample, while the opposite is true for "small" and "very small" enterprises. Annex 1. Sampling Methodology 44 Table A.1.3. Sample frame and theoretical sample Cluster h Employment Sample Weight nh Nh/N Sample Probability Size Frame of Selection Location Class Sector Nh nh/Nh 1 Asmara Region Very Small Construction Material 50 5 0 0.0000 0.0000 2 Asmara Region Very Small Food and Beverage 130 13 1 0.0769 0.0045 3 Asmara Region Very Small Furniture 30 3 0 0.0000 0.0000 4 Asmara Region Very Small Metal 10 1 0 0.0000 0.0000 5 Asmara Region Very Small Paper/Printing/Publishing 20 2 2 1.0000 0.0089 6 Asmara Region Very Small Plastics 20 2 2 1.0000 0.0089 7 Asmara Region Very Small Textile, Leather and Garments 30 3 0 0.0000 0.0000 8 Asmara Region Small Chemicals/Paints/Pharmaceuticals 190 9 8 0.8889 0.0357 9 Asmara Region Small Construction Material 467 20 3 0.1500 0.0134 10 Asmara Region Small Food and Beverage 664 28 7 0.2500 0.0313 11 Asmara Region Small Furniture 312 13 5 0.3846 0.0223 12 Asmara Region Small Metal 253 10 1 0.1000 0.0045 13 Asmara Region Small Paper/Printing/Publishing 131 6 4 0.6667 0.0179 14 Asmara Region Small Plastics 135 5 5 1.0000 0.0223 15 Asmara Region Small Textile, Leather and Garments 372 15 8 0.5333 0.0357 16 Asmara Region Medium Chemicals/Paints/Pharmaceuticals 141 2 2 1.0000 0.0089 17 Asmara Region Medium Construction Material 415 6 3 0.5000 0.0134 18 Asmara Region Medium Food and Beverage 285 4 2 0.5000 0.0089 19 Asmara Region Medium Furniture 247 4 2 0.5000 0.0089 20 Asmara Region Medium Metal 151 2 1 0.5000 0.0045 21 Asmara Region Medium Paper/Printing/Publishing 90 1 0 0.0000 0.0000 22 Asmara Region Medium Textile, Leather and Garments 253 4 2 0.5000 0.0089 23 Asmara Region Large Chemicals/Paints/Pharmaceuticals 103 1 1 1.0000 0.0045 24 Asmara Region Large Construction Material 1174 4 4 1.0000 0.0179 25 Asmara Region Large Food and Beverage 2587 7 7 1.0000 0.0313 26 Asmara Region Large Metal 415 3 3 1.0000 0.0134 27 Asmara Region Large Paper/Printing/Publishing 242 1 1 1.0000 0.0045 28 Asmara Region Large Textile, Leather and Garments 5050 11 10 0.9091 0.0446 29 Keren Region Very Small Food and Beverage 30 3 2 0.6667 0.0089 30 Keren Region Small Chemicals/Paints/Pharmaceuticals 11 1 1 1.0000 0.0045 31 Keren Region Small Construction Material 67 4 2 0.5000 0.0089 32 Keren Region Small Food and Beverage 25 2 1 0.5000 0.0045 33 Keren Region Large Food and Beverage 618 1 1 1.0000 0.0045 continued... Annex 1. Sampling Methodology 45 Table A.1.3. Sample rrame and theoretical sample (continued) Cluster h Employment Sample Weight nh Nh/N Sample Probability Size Frame of Selection Location Class Sector Nh nh/Nh 34 Massawa Region Small Construction Material 22 1 0 0.0000 0.0000 35 Massawa Region Small Food and Beverage 109 4 1 0.2500 0.0045 36 Massawa Region Small Furniture 42 1 0 0.0000 0.0000 37 Massawa Region Small Textile, Leather and Garments 35 1 1 1.0000 0.0045 38 Massawa Region Medium Construction Material 98 1 1 1.0000 0.0045 39 Massawa Region Medium Food and Beverage 50 1 0 0.0000 0.0000 40 Massawa Region Large Construction Material 329 2 2 1.0000 0.0089 41 Massawa Region Large Food and Beverage 183 1 1 1.0000 0.0045 42 Massawa Region Large Furniture 181 1 1 1.0000 0.0045 43 Other Regions Very Small Food and Beverage 60 6 0 0.0000 0.0000 44 Other Regions Small Construction Material 12 1 0 0.0000 0.0000 45 Other Regions Small Food and Beverage 81 5 0 0.0000 0.0000 46 Other Regions Small Paper/Printing/Publishing 15 1 0 0.0000 0.0000 47 Other Regions Small Textile, Leather and Garments 25 1 0 0.0000 0.0000 48 Other Regions Large Food and Beverage 280 1 0 0.0000 0.0000 16, 240 224 98 0.4375 Source: World Bank, RPED Eritrea, 2002. Annex 1. Sampling Methodology 46 Table A.1.4. Sample Structure (by sector and region) Asmara Keren Massawa Total Chemicals, Paints and Number of Firms 9 1 10 Pharmaceuticals Total Employment 251 8 259 Construction Materials Number of Firms 6 2 1 9 Total Employment 632 22 158 812 Food and Beverage Number of Firms 11 3 5 19 Total Employment 2,004 650 625 3,279 Furniture Number of Firms 4 4 Total Employment 171 171 Metal Number of Firms 6 6 Total Employment 420 420 Paper/Printing/Publishing Number of Firms 7 7 Total Employment 418 418 Plastics Number of Firms 7 7 Total Employment 187 187 Textile, Leather and Number of Firms 16 1 17 Garments Total Employment 3,488 23 3,511 Total Number of Firms 66 6 7 79 Total Employment 7,571 680 806 9,057 Annex 1. Sampling Methodology 47 Table A.1.5. Sample Structure (by sector and size) Very Small Medium Large Total Small Chemicals, Paints Number of Firms 2 7 1 10 and Pharmaceuticals Total Employment 16 157 86 259 Construction Materials Number of Firms 1 3 3 2 9 Total Employment 8 71 326 407 812 Food and Beverage Number of Firms 2 7 3 7 19 Total Employment 16 304 266 2,693 3,279 Furniture Number of Firms 1 1 2 4 Total Employment 15 24 132 171 Metal Number of Firms 3 2 1 6 Total Employment 109 164 147 420 Paper/Printing/ Number of Firms 1 4 1 1 7 Publishing Total Employment 7 102 90 219 418 Plastics Number of Firms 7 7 Total Employment 187 187 Textile, Leather Number of Firms 1 9 2 5 17 and Garments Total Employment 8 255 170 3,078 3,511 Total Number of Firms 8 41 14 16 79 Total Employment 70 1,209 1,234 6,544 9,057 Annex 1. Sampling Methodology 48 Figure A.1.1. Employment Distribution Figure A.1.3. Employment Distribution (by size) (by region) Sample frame Surveyed sample Sample frame Surveyed sample Percent Percent 80 100 80 60 60 40 40 20 20 0 0 Very Small Medium Large Asmara Keren Massawa Other small Region Region Region Regions Figure A.1.2. Employment Distribution (by sector) Sample frame Surveyed sample Percent 40 30 20 10 0 Chemicals/Paints/ Construction Food/ Furniture Metal Paper/Printing/ Plastics Textile/Leather/ Pharmaceuticals Materials Beverage Publishing Garments Annex 2. Results from the Worker Survey 49 Worker Characteristics Table A.2.1. Highest educational achievement The data from this section come from the worker of employees (percent) interviews, a subsample of 521 employees from the manufacturing sector. Male Female Total University Degree 6.17 1.47 4.30 Age and Gender Tertiary Education 8.44 6.37 7.62 Workers employed in Eritrean manufacturing tend to Secondary 32.14 41.18 35.74 be middle-aged, on average 42 years old (Table A.3.4 Non Vocational in Annex III). Women are usually much younger than Secondary Vocational 3.57 4.41 3.91 their male colleagues (34 and 47 years old, Junior Secondary 16.88 9.31 13.87 respectively). In the Eritrean context, this gap may be Primary 24.35 20.10 22.66 due to the recent conflict and the mobilization of No Education 8.44 17.16 11.91 qualified young men, thereby widening the gender- age gap. The younger workers tend to concentrate in Source: World Bank, RPED Eritrea, 2002. very small firms, while older ones are found in large firms. In addition, most of the well-trained workers tend to be older because widespread worker training increase the stock of human capital and ultimately largely took place only during the Italian colonization favor growth. By the end of the war for and the rule of Emperor Haile Selassie. There is a independence, female and male illiteracy rates were huge gap in training from the period of the socialist about 80­90 percent, mainly due to destruction of Ethiopian Derg regime (1974­1991). infrastructure and disruption of social services The mean tenure of the sample's workers (the brought on by the war (World Bank 1996). Quite period they have been employed at the firm) was appropriately, education has been given a prominent about 10 years. Skilled production workers stayed the role in Eritrea's development strategy (IMF 2001; longest in their firms (almost 14 years). Experience in World Bank 2002a). Despite higher gross enrollment other firms is also significant. While employees have rates since the war, the data we collected suggest been unemployed for about 3 years on average in the that this strategy has not yet provided the past, they have either been self-employed or manufacturing sector with the properly educated employed by another company for about 6 years. workforce it needs. This is understandable given that However, almost half of the workers currently in the raising a population's education level takes private sector (48 percent) have previously been considerable time and the initial level of education employed by the government. The duration of this was quite low. involvement was significant, on average 18 years, and Conflict is the reason for interruptions in includes time spent in the army. education for 35.7 percent of the interviewed workers. Data indicate that within the subsample, Education about 12 percent of the workers received no The need for improvement in workforce qualifications education and about 22 percent only reached and education is especially acute in Eritrea. The links primary education. Most of workers went to between growth and education are well documented. secondary non-vocational school (35.7 percent). The The externalities generated by improved education dominance of this type of secondary education in the levels (better adaptability, easier learning-by-doing) workforce is very similar to what was found in many Annex 2. Results from the Worker Survey 50 Table A.2.2. Education of workers (percentage) University Tertiary Secondary Junior Primary No degree education secondary education Sector Chemicals/Paints 3.39 6.78 32.20 15.25 25.42 16.95 Construction 3.33 6.67 36.67 11.67 23.33 18.33 Materials Food and Beverage 10.08 9.24 29.41 15.97 20.17 15.13 Furniture 0.00 7.14 50.00 21.43 17.86 3.57 Metal 8.00 18.00 34.00 14.00 22.00 4.00 Paper/Printing/ 0.00 8.33 68.75 10.42 10.42 2.08 Publishing Plastics 0.00 0.00 30.77 5.13 41.03 23.08 Textile, Leather 1.83 4.59 46.79 14.68 23.85 8.26 and Garments Size Very Small 0.00 0.00 2.17 19.57 21.74 6.52 Small 2.12 5.51 39.41 13.98 24.15 14.83 Medium 4.55 13.64 42.05 15.91 18.18 5.68 Large 9.15 9.86 34.51 10.56 23.24 12.68 Ownership Foreign 7.27 5.45 32.73 14.55 23.64 16.36 Local 3.94 7.88 40.48 13.79 22.54 11.38 Source: World Bank, RPED Eritrea, 2002. RPED surveys in Africa. Of noticeable importance is the firm is locally owned (Table A.2.2). The largest the fact that only 3.9 percent of the workers received proportion of workers with a university degree is secondary vocational training, a type of education found in sectors with significant technological that could provide manufacturing with some of the requirements (that is, chemicals and paints, food skilled workers it needs. and beverages, and construction materials). The As confirmed by other studies (IMF 2001), there are overall level of education of the workforce is not still some significant differences in education levels very high, and it seems that very few workers have between men and women in Eritrea. Overall, it seems benefited from any training since they began that women have less access to higher education than working. Only 9.7 percent of the workers received men. Only 7.8 percent of the female workers interviewed any training from their former employers and only had a university degree or tertiary education, while this 9.2 percent of them benefited from training in their was the case for 14.6 percent of the male workers we current jobs. Again, gender disparities remain met. A similar conclusion can be drawn for female strong. While 12.8 percent of the male workers access to primary and junior secondary education. received some training from their current The level of education of workers varies widely employers, only 4.9 percent of the female workers according to sector, size of firm, and whether or not had such opportunities. Annex 2. Results from the Worker Survey 51 Table A.2.3. Union membership % of firms whose Unionization employees belong Rate (%) a union Sample 58.2 91.4 Sizeclass Very Small 12.5 98.0 Small 50.0 93.1 Medium 69.2 88.0 Large 85.0 90.8 Sector Chemicals/Paints 30.0 99.7 Construction Materials 66.7 96.8 Food and Beverage 57.9 92.4 Furniture 50.0 91.0 Metal 83.3 81.2 Paper/Printing/Publishing 42.9 89.7 Plastics 57.1 81.3 Textile, Leather and Garments 70.6 93.7 Source: World Bank, RPED Eritrea, 2002. Unions Estimates of Earning Functions At the beginning of 2002, according to the data, union Our data suggest that the issue of earnings differentials membership was fairly high in Eritrea. On average, in Eritrea deserve greater attention. It is quite plausible almost 58.2 percent of the firms report having workers that the traditional view of an integrated competitive that belong to a union (Table A.2.3). On average, labor market does not hold in this country. A way to when there is union membership in a firm, about 91.4 clarify the origin of these differentials is to estimate a percent of the workers belong to it. few wage equations. Some results are presented in The percentage of firms that report having Table A.2.4. In these equations the dependant variable employees in a union increases, not unexpectedly, with is the log of the individual earnings (which includes their size. While only 12.5 percent of the very small firms wage, over-time pay, bonuses, and other benefits).8 report to have employees belonging to a union, 85 The starting point is the estimation of a simple percent of large firms have unionized employees. For all wage equation related to the individual characteristics firms, the unionization rate ranges between 88 and 98 of the workers. Related results are reported in equation percent. In terms of sectors, the distribution is uneven. (1) in Table A.2.4. It is assumed that employers are At the upper level of the distribution is the metal sector, able to distinguish between workers' productivity, in which about 83 percent of the firms report to have which depends on their level of education, origin, and workers belonging to a union. At the other extreme, only experience, and set wages accordingly. Equation (2) 30 percent of the firms in the chemicals and paints differs from Equation (1) by the inclusion of sectoral sector report having employees in a union. dummies and firm-specific variables.9 Annex 2. Results from the Worker Survey 52 Table A.2.4. The determinants of wages in the Eritrean manufacturing sector Equation(1) Equation(2) Intercept 4.9091 4.7705 (0.283) (0.275) Workers Characteristics Education 0.0527 0.0515 (in years) (0.005) (0.005) War -0.0211 ns -0.0346 (dummy, equal to 1 if impact on education) (0.042) (0.041) Experience with firm 0.0217 0.0259 (in years) (0.007) (0.007) Experience (squared) -0.0002 ns -0.0003 (in years) (0.000) (0.000) Other professional experience 0.0117 0.0126 (in years) (0.002) (0.002) Gender 0.3151 0.2850 (dummy equal to 1 if male worker) (0.049) (0.050) Weekly hours worked 0.0182 0.0188 (0.006) (0.006) Trained 0.2551 0.1863 (dummy equal to 1 if any training after school) (0.061) (0.058) Sector dummies Yes Firms Characteristics Age of the Firm 0.0018 (in years) (0.001) Foreign Ownership 0.0651 ns (dummy equal to 1 if majority shareholder (0.079) is foreign) PFDJ Ownership 0.3937 (dummy equal to 1if majority shareholder is PFDJ) (0.084) Firm in Asmara -0.1176 (dummy equal to 1 if firm located in Asmara) (0.062) N 513 513 F-statistic 51.362 25.796 R-squared 0.449 0.499 Significance levels: 1 %; 5 %; 10 %; ns = non significant Dependent Variable: Log of Earnings (Earnings= sum of individual wage, overtime pay and bonus on a monthly basis) Annex 2. Results from the Worker Survey 53 In both equations, the variables that account for 70,000 are said to currently be living with HIV human capital--the number of years of education, the infection. According to UNAIDS, as of late 2001, the years of experience with the same firm, the work prevalence rate of HIV/AIDS is 2.8 percent for the experience outside the firm (in years)--have the adult population (ages 15-49 years), 4.8 percent for expected influence and are statistically significant. mobilized troops, and 25.5 percent for female bar For example, the higher the endowment of human workers. While the infection rate is one of the lowest capital in an individual worker, the higher his wage. on the continent, population dynamics can cause a The impact of training by firms is also positive and quick jump in this rate. significant, whether in equation (1) or (2). This is also Demobilization, flows of refugees, and movements the case for the duration of work variable. This of internally displaced persons can work to spread the confirms that part of the remuneration is adjusted on disease among the national population. The the basis of how long the worker has been employed demobilization of 200,000 soldiers in Eritrea has at the firm. The gender dummy is highly significant in begun and is scheduled to last for the next two years. both equations and confirms the wage disparities Some 5,000 soldiers (3,600 of which are female) were mentioned earlier. Interestingly, the result of the war discharged in June 2002, marking the end of the pilot dummy (a proxy based on whether or not a worker's phase of the National Demobilization and studies were interrupted by the various wars) is not Reintegration Process (DRP). Normalization of border significant in both equations. crossings with Ethiopia will also have an effect, given In equation (2), firm-level variables such as age of that Eritrea's most important neighbor has an the firm and PFDJ ownership are relevant. Working at HIV/AIDS adult prevalence rate of 6.4 percent. The a PFDJ firm entails statistically significant higher movement of over a million internally displaced earnings, consistent with our previous descriptive persons (IDPs), as well as the repatriation of refugees statistics. However, contrary to what is often found in from Sudan, could also increase the risk of spreading other surveys, working at a foreign firm does not the disease. By the end of 2001, only about 45,000 significantly affect the level of wages. Finally, IDPs were unable to return to their land because of geographical factors influence earnings; the Asmara landmines. An estimated 300,000 to 500,000 Eritrean dummy is also significant. These regressions suggest refugees fled to Sudan during the thirty-year war that wage levels are determined not just by worker (Bruchhaus and Mahreteab 2000). Sudan has an characteristics but by other variables as well. HIV/AIDS adult prevalence rate of an estimated 2.6 percent, although, like Eritrea, this is based on limited HIV/AIDS surveillance data. Although malaria is currently the biggest killer in In addition, changes in beliefs and behavior can Eritrea (between 70 to 90 percent of deaths among all influence the spread of the disease. Health sector ages in 1997 and 1998), HIV is initiating its spread in professionals have stated that the increasing openness the country. Unless a significant, well-planned, and of the country to rest of the world, in both physical and focused intervention is maintained, there is no reason psychological terms since the end of the Derg regime, why the disease will not follow the pattern found in has started a change in mores and attitudes, especially other countries. those of young people in urban areas. The first case of AIDS in Eritrea was reported in While the impact of AIDS on public health is clear, 1988. As of late 2001, there are 6, 873 people the disease's effect on economies is complex. On a diagnosed with the disease. An estimated 60,000 to macroeconomic level, the lack of good quality time- Annex 2. Results from the Worker Survey 54 series data on the spread of AIDS in Eritrea in particular The study's findings were similar to previous UNAIDS and in Africa in general hinders the ability to data, with an adult prevalence rate of 3.0 percent, 4.6 disaggregate it from other factors affecting economic percent in the army, and 22.8 percent among female growth. However, the potential impacts are numerous. bar workers. The high number of respondents who Unlike most other fatal diseases, HIV infection occurs reported not believing that they were at risk (72 most often in people of working age, meaning (1) a percent) is troubling. Even "at risk" populations, such reduction in income and savings, (2) additional costs as female bar workers and soldiers, felt they were not are incurred for training and recruiting, (3) an increase at risk (around 60 percent). This may be due to the in absenteeism for the people living with HIV/AIDS and high number of people who seem to take precautions. those caring for them, (4) an increase in health However, a number of more detailed investigations spending in the public budget, to the detriment of other revealed, for example, that a woman can believe she growth-enhancing areas such as infrastructure, (5) the is safe if she only has one sex partner, even if that resulting reduced productivity can lead to a reduction partner has multiple partners, revealing a profound in the economy's attractiveness to foreign investors, misunderstanding of how the disease is transmitted. and (6) a reduction in tax revenue, (Bloom et al. 2001). The Government of Eritrea is aware of the risks Questions related to HIV/AIDS were asked in the posed by the recent and imminent demographic shifts public health section of the worker surveys. Five to ten and is working to ensure that the HIV/AIDS rate is not workers of various types were surveyed at each firm increased. It is hoped that any increase can be for a total of 521 respondents. When asked "Do you avoided with the cooperation of the government and know anyone who you believe died of AIDS?" the multilateral and bilateral donors. The GOE has majority (62 percent) responded "No." Perhaps more launched a nationwide campaign against HIV/AIDS interestingly, 37 percent responded, "Yes," a with the goals improving awareness and prevention as particularly high percentage, given the low well as treatment for those 60,000­70,000 Eritreans prevalence rate of less than 3 percent in the country. already infected. However, 68 percent of workers had no idea (or were The GOE and the World Bank (as part of a $40 unwilling to divulge) what share of their co-workers million project) financed a program to control HIV, they believe is currently HIV positive; 32 percent Malaria, Sexually Transmitted Diseases, and responded that they thought none of their co-workers Tuberculosis in Eritrea (HAMSET) beginning in 2000. currently had the disease. The program's objectives are to coordinate and It is quite encouraging that when asked "Do you improve treatment of these communicable diseases take any particular precautions to avoid getting that are often related. In connection with this project, a AIDS?" most workers (86 percent) answered "Yes." A new national blood bank has been established in demographic health survey undertaken in 1995 found Asmara. At least 20 voluntary counseling and that the proportion of people citing at least two treatment centers have been established in hospitals acceptable ways to protect against HIV infection was around the country, and the first "free standing" HIV high, 79 percent (UNAIDS/WHO 2000). testing facility was recently opened in Asmara. Over In June 2002, the government released data from 100 counselors have been trained. A pilot project has the latest nationwide health survey conducted in been started to provide antiretroviral drugs to antenatal 2001, showing that general awareness is high: 99 clinics to prevent mother-to-child transmission of HIV percent of respondents had heard of HIV/AIDS (4,753 (antiretroviral drugs are unavailable in Eritrea; those people between 15 and 49 years old were surveyed). who can afford it buy them overseas). Annex 2. Results from the Worker Survey 55 In addition, the World Bank's Multi-Country HIV/AIDS Program for Africa (MAP) has offered a $40 million no-interest credit to Eritrea. The Bank approved MAP in September 2001, giving a dozen countries financial and technical support for the implementation of national HIV/AIDS strategies (Benin, Burkina Faso, Cameroon, Central African Republic, Eritrea, Ethiopia, Gambia, Ghana, Kenya Madagascar, Nigeria, Uganda). The support is directed to the government, communities, and civil society organizations. USAID, the Italian Cooperation, and UN agencies are also contributing to information dissemination activities in schools, via radio, television, and print media. The fact that a great deal of initiative is being taken by the government, donors, and local groups at an early stage is extremely encouraging. Nevertheless, there is still a stigma strongly associated with the disease that inhibits open discussion and behavioral change. Notes 8. Equations have been estimated using ordinary least squares (OLS). In addition, White's consistent t ratios are reported in the table in order to correct for heteroscedasticity in the data. 9. These sectoral dummies have not been reported in the table for ease of reading. Half of them are significant at the 10 percent level. Annex 3. Supplemental Labor Tables 56 Employment Decline intersectoral variation is high. Earnings are highest in We now examine the changes in manufacturing the food and beverage and metal industries and are employment from 1999 to 2001. All the firms in the the lowest in the plastics sector. On average, cash survey were operating during this period. The latest earnings for workers increase with size and differ year a firm in the sample was created was 1998. Since greatly depending on geographic location. Firms with there is no volatility (entry/exit) in the sample, an foreign equity provide higher wages than their examination of employment changes in existing firms domestic counterparts. It seems that PFDJ-owned provides a good picture of changes in overall firms tend to provide much higher cash earnings than manufacturing employment. Not surprisingly, over the firms with other types of ownership ($104.40 a month last three years, the average size of manufacturing versus a sample average of $71.50). firms has decreased (Table A.3.2). While on average 129.0 staff were employed by firms in 1999, this number was about 106.9 in 2001, representing a 17.1 percent decline. Domestic firms were strongly affected; their average size declined by 18.1 percent between 1999 and 2001. In terms of size, very small firms--which are mainly local enterprises--faced a decrease of 24.4 percent in their average size over the 1999­2001 period. Textile, leather, and garments and food and beverages were the two most strongly affected sectors. The average size of firms in those sectors declined by 28.5 and 17.3 percent, respectively. Note that these two sectors were among those that had the highest rates of white-collar mobilization. In addition, the loss of the Ethiopian export market probably has had a negative impact on the shrinking of manufacturing employment in Eritrea. The burden of the decline in employment has not been borne equally by permanent and temporary workers (see Table A.3.3). Although the decline in permanent employees between 1999 and 2001 is significant (11.4 percent), the decline in temporary workers is much larger (19.5 percent). The reduction in temporary workers was especially important for very small and small firms, where 44.5 and 54.3 percent of the temporary workers were laid-off, respectively. A crude way to analyze these earnings differentials is to examine the distribution of average cash earnings by sector, size, location, and ownership of the firms. According to Table A.3.4. Eritrea does not differ greatly from other African countries. In effect, Annex 3. Supplemental Labor Tables 57 Table A.3.1. Percentage of white-collar workers (by sector and size, 2001) Very Small Medium Large Total small Chemicals/Paints Average percentage 11.8 16.9 8.9 na 15.1 Standard Deviation (0.98) (7.91) na na (7.14) Construction Materials Average percentage 0.0 7.7 na 9.3 7.7 Standard Deviation na (6.82) na (10.16) (8.50) Food and Beverage Average percentage 0.0 14.5 17.6 9.6 11.6 Standard Deviation (0.00) (8.22) (13.20) (4.05) (8.81) Furniture Average percentage 0.0 13.6 3.8 na 5.3 Standard Deviation na na (1.54) na (5.91) Metal Average percentage na 17.3 8.4 7.4 11.2 Standard Deviation na (6.25) (4.13) na (6.09) Paper/Printing/ Average percentage (20.00) (19.20) 15.8 6.0 16.9 Publishing Standard Deviation na (12.66) na na (10.26) Plastics Average percentage na 11.6 na na 11.6 Standard Deviation na (4.24) na na (4.24) Textile, Leather and Average percentage 11.1 10.5 8.0 1.4 7.6 Garments Standard Deviation na (6.06) (7.49) (0.19) (6.28) Average percentage 6.8 13.5 11.1 7.2 10.8 Standard Deviation (7.83) (7.55) (8.95) (6.35) (7.96) Source: World Bank, RPED Eritrea, 2002. Annex 3. Supplemental Labor Tables 58 Table A.3.2. Average number of workers per firm (permanent and temporary workers, 1999-2001) 1999 2000 2001 Ownership Local 139.49 123.17 114.21 Std (254.21) (217.93) (215.17) Foreign 41.63 39.00 42.25 Std (36.30) (23.87) (30.83) Sizeclass Very Small 11.25 8.75 8.50 Std (5.39) (2.55) (2.20) Small 30.49 27.92 25.92 Std (21.38) (18.26) (9.89) Medium 84.67 79.77 71.85 Std (47.71) (28.87) (13.39) Large 375.30 344.45 323.00 Std (370.95) (317.02) (324.63) Sector Chemicals/Paints 25.50 25.90 27.20 Std (22.13) (23.20) (24.11) Construction Materials 103.00 90.22 92.78 Std (114.94) (92.40) (85.22) Food and Beverage 175.42 172.58 145.11 Std (204.14) (204.96) (173.40) Furniture 43.00 42.75 44.25 Std (31.49) (27.21) (31.28) Metal 65.83 70.00 64.00 Std (45.82) (46.30) (35.67) Paper/Printing/Publishing* 41.60 59.71 62.43 Std (40.96) (75.27) (81.29) Plastics 27.43 26.71 26.14 Std (11.90) (11.56) (12.23) Textile, Leather and Garments 279.80 206.53 200.12 Std (447.39) (362.55) (379.41) Sample 129.05 114.65 106.92 Std (242.25) (208.15) (205.21) * This average sectoral increase is highly biased by one firm that doubled its employment between 1999 and 2001. The sector should be considered an outlier rather than representative of a sectorwide improvement Source: World Bank, RPED Eritrea, 2002. Annex 3. Supplemental Labor Tables 59 Table A.3.3. Changes in number of workers per firm (1999 to 2001) 1999 2001 Change % change Number of Permanent Workers Very Small firms 72 58 -14 -19.44 Small 905 911 6 0.66 Medium 987 906 -81 -8.21 Large 6 117 5 285 -832 -13.60 Total 8 081 7 160 -921 -11.40 Number of Temporary workers Very Small firms 18 10 -8 -44.44 Small 162 74 -88 -54.32 Medium 29 28 -1 -3.45 Large 1 389 1 175 -214 -15.41 Total 1 598 1 287 -311 -19.46 Source: World Bank, RPED Eritrea, 2002. Annex 3. Supplemental Labor Tables 60 Table A.3.4. Monthly cash earnings (by sector, size, location, and ownership) US$ Nakfa Sector Chemicals/Paints Average Value 77.47 1,045.87 Standard Deviation (43.5) (586.7) Construction Materials Average Value 66.66 899.89 Standard Deviation (43.8) (591.0) Food and Beverage Average Value 81.95 1,106.31 Standard Deviation (57.4) (774.5) Furniture Average Value 72.62 980.36 Standard Deviation (28.6) (385.9) Metal Average Value 88.48 1,194.42 Standard Deviation (49.1) (662.7) Paper/Printing/Publishing Average Value 69.21 934.38 Standard Deviation (41.6) (561.3) Plastics Average Value 49.44 667.38 Standard Deviation (23.2) (313.4) Textile, Leather and Garments Average Value 59.90 808.71 Standard Deviation (38.3) (517.6) Size Very Small Average Value 60.21 812.89 Standard Deviation (41.5) (560.6) Small Average Value 61.92 835.96 Standard Deviation (41.3) (557.2) Medium Average Value 83.88 1,132.32 Standard Deviation (47.0) (635.0) Large Average Value 82.99 1,120.33 Standard Deviation (50.0) (674.7) Location Asmara Average Value 70.44 950.97 Standard Deviation (44.9) (606.7) Keren Average Value 60.50 816.79 Standard Deviation (49.8) (672.7) Massawa Average Value 89.24 1,204.78 Standard Deviation (50.4) (680.7) Ownership Foreign Average Value 77.47 1,045.87 Standard Deviation (58.5) (789.2) Local Average Value 70.84 956.41 Standard Deviation (44.4) (600.0) continued... Annex 3. Supplemental Labor Tables 61 Table A.3.4. Monthly cash earnings (by sector, size, location, and ownership) (continued) US$ Nakfa Detailed Domestic Private Average Value 63.55 857.88 Ownership Standard Deviation (40.2) (543.1) Foreign Private Average Value 97.06 1,310.28 Standard Deviation (77.3) (1 044.1) Government Average Value 82.8 1,118.2 Standard Deviation (44.8) (604.7) Mixed Priv. Local/Foreign Average Value 64.43 869.79 Standard Deviation (49.2) (664.2) Mixed Priv./Gvt Average Value 71.22 961.50 Standard Deviation (39.6) (534.6) PFDJ Average Value 104.37 1,409.05 Standard Deviation (54.9) (740.8) Note: Computed on the basis of the wages provided by workers in April/May 2002 and converted into U.S. dollars using the May 2002 official exchange rate of US$1 = 13.5 Nakfa. Source: World Bank, RPED Eritrea, 2002. Annex 3. Supplemental Labor Tables 62 Table A.3.5. Number of permanent full-time employees by sectors and sizeclass in 2001 Very C.Pct Small C.Pct Med. C.Pct Large C.Pct Total C.Pct small Chemicals/Paints 17 (29.31) 161 (17.67) 90 (9.93) - (0.00) 268 (3.74) R.Pct (6.34) (60.07) (33.58) (0.00) (100.00) Construction 8 (13.79) 67 (7.35) - (0.00) 681.00 (12.89) 756 (10.56) Materials R.Pct (1.06) (8.86) (0.00) (90.08) (100.00) Food and Beverage 16 (27.59) 90 (9.88) 232 (25.61) 1 436.00 (27.17) 1 774 (24.78) R.Pct (0.90) (5.07) (13.08) (80.95) (100.00) Furniture 3 (5.17) 22 (2.41) 137 (15.12) - (0.00) 162 (2.26) R.Pct (1.85) (13.58) (84.57) (0.00) (100.00) Metal - (0.00) 54 (5.93) 194 (21.41) 121.00 (2.29) 369 (5.15) R.Pct (0.00) (14.63) (52.57) (32.79) (100.00) Paper/Printing/ 5 (8.62) 86 (9.44) 95 (10.49) 217.00 (4.11) 403 (5.63) Publishing R.Pct (1.24) (21.34) (23.57) (53.85) (100.00) Plastics - (0.00) 164 (18.00) - (0.00) - (0.00) 164 (2.29) R.Pct (0.00) (100.00) (0.00) (0.00) (100.00) Textile, Leather 9 (15.52) 267 (29.31) 158 (17.44) 2 830.00 (53.55) 3264 (45.59) and Garments R.Pct (0.28) (8.18) (4.84) (86.70) (100.00) Total 58 (100.00) 911 (100.00) 906 (100.00) 5 285 (100.00) 7 160 (100.00) R.Pct (0.81) (12.72) (12.65) (73.81) (100.00) Annex 3. Supplemental Labor Tables 63 Table A.3.6. Recent change in employment by sectors, 1999 to 2001 Permanent Workers 1999 2001 Change % change n Chemicals/Paints 251 268 17 6.77 10 Construction Materials 854 756 -98 -11.48 9 Food and Beverage 2 024 1 774 -250 -12.35 19 Furniture 146 162 16 10.96 4 Metal 375 369 -6 -1.60 6 Paper/Printing/Publishing 188 403 215 114.36 7 Plastics 176 164 -12 -6.82 7 Textile, Leather and Garments 4 067 3 264 -803 -19.74 17 Total 8 081 7 160 -921 -11.40 Temporary workers Chemicals/Paints 4 4 0 0.00 10 Construction Materials 73 79 6 8.22 9 Food and Beverage 1 309 983 -326 -24.90 19 Furniture 26 15 -11 -42.31 4 Metal 20 15 -5 -25.00 6 Paper/Printing/Publishing 20 34 14 70.00 7 Plastics 16 19 3 18.75 7 Textile, Leather and Garments 130 138 8 6.15 17 Total 1 598 1 287 -311 -19.46 Annex 3. Supplemental Labor Tables 64 Table A.3.7. Average structure of employment in 2001 (C.pct) Very Small Medium Large Total small Chemicals/Paints Permanent workers 100.00 97.58 100.00 na 98.53 Temporary workers 0.00 2.42 0.00 na 1.47 Construction Materials Permanent workers 100.00 100.00 na 89.61 90.54 Temporary workers 0.00 0.00 na 10.39 9.46 Food and Beverage Permanent workers 100.00 84.11 96.67 59.98 64.35 Temporary workers 0.00 15.89 3.33 40.02 35.65 Furniture Permanent workers 23.08 100.00 96.48 na 91.53 Temporary workers 76.92 0.00 3.52 na 8.47 Metal Permanent workers na 100.00 92.82 100.00 96.09 Temporary workers na 0.00 7.18 0.00 3.91 Paper/Printing/ Permanent workers 100.00 84.31 100.00 92.34 92.22 Publishing Temporary workers 0.00 15.69 0.00 7.66 7.78 Plastics Permanent workers na 89.62 na na 89.62 Temporary workers na 10.38 na na 10.38 Textile, Leather and Permanent workers 100.00 93.68 100.00 95.93 95.94 Garments Temporary workers 0.00 6.32 0.00 4.07 4.06 Total Permanent workers 85.29 92.49 97.00 81.81 84.76 Temporary workers 14.71 7.51 3.00 18.19 15.24 Table A.3.8. Average age of employees by sizeclass and gender (number of years) Very Small Med. Large Total small Male 36.4 49.4 46.1 46.7 46.5 Female 29.0 34.1 32.6 36.3 34.2 Total 35.1 42.3 40.9 43.0 41.6 Annex 3. Supplemental Labor Tables 65 Table A.3.9. Monthly cash earnings of workers in Nakfa by job position in April/May 2002 Male Female Total Management Average Value 2 186.15 2 775.00 2 237.36 Standard Deviation (763.0) (1 025.3) (778.3) Frequency 21 2 23 Professional Average Value 1 636.96 1 496.91 1 596.42 Standard Deviation (781.8) (645.9) (739.2) Frequency 27 11 38 Skilled Production Worker Average Value 1 174.47 717.03 1 069.26 Standard Deviation (518.9) (365.8) (523.9) Frequency 154 46 200 Unskilled Production Worker Average Value 748.95 474.76 607.60 Standard Deviation (405.5) (184.3) (339.8) Frequency 78 83 161 Non Production Worker Average Value 941.97 734.78 808.77 Standard Deviation (418.9) (395.2) (413.9) Frequency 35 63 98 Apprentice Average Value 150.00 na 150.00 Standard Deviation na na na Frequency 1 na 1 Total Average Value 1 147.19 686.32 965.85 Standard Deviation (646.0) (462.2) (622.4) Frequency 316 205 521 Note: Computed on the basis of the wages provided by interviewed workers in April/May 2002. Annex 3. Supplemental Labor Tables 66 Table A.3.10. Monthly cash earnings of unskilled production workers (by sector, size, location and ownership in April/May 2002) US$ Nakfa Sector Chemicals/Paints Average Value 53.05 716.15 Standard Deviation (33.5) (452.0) Construction Materials Average Value 45.66 616.44 Standard Deviation (25.5) (344.1) Food and Beverage Average Value 41.81 564.41 Standard Deviation (20.0) (270.3) Furniture Average Value 49.43 667.27 Standard Deviation (15.8) (212.8) Metal Average Value 66.28 894.82 Standard Deviation (45.9) (619.3) Paper/Printing/Publishing Average Value 49.00 661.54 Standard Deviation (28.8) (388.5) Plastics Average Value 43.62 588.91 Standard Deviation (16.2) (219.1) Textile, Leather and Average Value 35.30 476.60 Garments Standard Deviation (14.8) (199.9) Size Very Small Average Value 29.25 394.88 Standard Deviation (10.9) (147.2) Small Average Value 42.99 580.33 Standard Deviation (22.1) (298.3) Medium Average Value 54.42 734.69 Standard Deviation (24.9) (335.9) Large Average Value 50.07 675.89 Standard Deviation (32.6) (439.6) Location Asmara Average Value 46.92 633.39 Standard Deviation (26.7) (360.1) Keren Average Value 28.25 381.33 Standard Deviation (8.8) (119.2) Massawa Average Value 48.95 660.86 Standard Deviation (15.7) (212.2) Ownership Foreign Average Value 47.35 639.16 Standard Deviation (24.0) (324.0) Local Average Value 44.77 604.36 Standard Deviation (25.4) (342.3) continued... Annex 3. Supplemental Labor Tables 67 Table A.3.10. Monthly cash earnings of unskilled production workers (by sector, size, location and ownership in April/May 2002) (continued) US$ Nakfa Detailed Ownership Domestic Private Average Value 42.96 579.91 Standard Deviation (22.7) (306.9) Foreign Private Average Value 63.85 862.00 Standard Deviation (30.9) (416.8) Government Average Value 43.4 585.3 Standard Deviation (26.4) (355.8) Mixed Priv. Local/Foreign Average Value 37.75 509.63 Standard Deviation (17.8) (240.2) Mixed Priv./Gvt Average Value 38.89 525.00 Standard Deviation (8.2) (110.2) PFDJ Average Value 77.89 1 051.50 Standard Deviation (41.4) (558.9) Note: Computed on the basis of the wages provided by interviewed workers in April/May 2002 and converted into U.S. dollars using the May 2002 official exchange rate of 1US $ = 13.5 Nakfa. Annex 4. Supplemental Finance Tables 68 Table A.4.1. Details on the financial sector A. Share (%) of Total Borrowing Denominated in Foreign Currency Sample 2.86 Domestic Private 0.00 Foreign Private 100.00 Government 0.00 Mixed Priv./Gvt 0.00 PFDJ 0.00 B. Clearing of Payments Duration (Days) Charge (% of Transaction) a. Check 1.44 0.31 b. Domestic Currency Wire 1.43 0.76 c. Foreign Currency Wire 6.33 1.64 C. Review of Accounts by External Auditors (Pct. C) Sample 89.74 Very Small 50.00 Small 91.89 Medium 100.00 Large 95.00 89.74 D. Audited Statements Necessary to Obtain Bank Credit (Pct.C) Sample 86.67 Very Small 71.43 Small 91.43 Medium 92.31 Large 80.00 E. Share of Outstanding Debt Due (Pct.C) Foreign Local Within 90 days 50.00 31.18 Within 3 to 12 Months 0.00 36.01 More than a Year 50.00 28.69 Annex 5. Standard Investment Climate Tables 69 Table A.5.1. Standard investment climate tables Firm Size (%) Sample Firm Activity (%) Sample Small (<100 emps) 74.68 Chemicals/Paints 11.39 Large (100+ emps) 25.32 Construction Materials 11.39 Food and Beverage 22.78 Market Orientation (%) Furniture 5.06 Exporter (>= 10% sales) 90.91 Garments 3.8 Non-Exporter 9.09 Metal 7.59 Non Metal 1.27 Firm Ownership (%) Paper/Printing/Publishing 8.86 Public Enterprise 12.66 Pharmaceuticals/Medical Products 1.27 Public Share Company 3.8 Plastics 8.86 Private Share Company 53.16 Textile and Leather 17.72 Sole Proprietorship or Partnership 18.99 PFDJ-owned 7.59 Firm Location (%) Others 3.8 Asmara 83.54 Keren 7.59 Massawa 8.86 Annex 5. Standard Investment Climate Tables 70 Table A.5.2. Competitors, suppliers and customers Eritrea Pakistan China India Morocco Small Large Low High capacity capacity Average Number Domestic 14.69 85.90 NA NA NA 9.49 29.50 8.69 30.71 of Competitors Private Firms State- 0.29 0.70 NA NA NA 0.31 0.25 0.28 0.29 Owned Firms Foreign- 0.14 2.90 NA NA NA 0.10 0.25 0.19 0.05 Owned Firms Average Number Domestic NA 93.20 NA NA NA NA NA NA NA of Suppliers Private Firms State- NA 1.30 NA NA NA NA NA NA NA Owned Firms Foreign- NA 7.90 NA NA NA NA NA NA NA Owned Firms Average Number Domestic NA 133.10 NA NA NA NA NA NA NA of Customers Private Firms State- NA 1.30 NA NA NA NA NA NA NA Owned Firms Foreign- NA 14.30 NA NA NA NA NA NA NA Owned Firms Annex 5. Standard Investment Climate Tables 71 Table A.5.3. Respondents' Evaluation of General Constraints to operation, % of firms evaluating constraint as "major" or "very severe" Eritrea Pakistan China India Morocco Small Large Low High capacity capacity A.Telecommunications 13.92 9.20 23.5 NA NA 15.25 10 11.11 23.81 B.Electricity 36.71 39.20 29.7 NA NA 32.2 50 31.48 47.62 C.Transportation 17.95 10.00 19.1 NA NA 18.64 15.79 18.87 19.05 D. Access to Land 21.52 20.40 14.7 NA NA 25.42 10 22.22 23.81 E. Tax rates 29.11 45.60 36.8 NA NA 33.9 15 27.78 33.33 F. Tax administration 15.19 46.10 26.7 NA NA 15.25 15 12.96 23.81 G. Customs and Trade 8.97 24.60 19.3 NA NA 10.34 5 9.26 10 Regulations H. Labor Regulations 5.06 15.00 20.7 NA NA 3.39 10 1.85 9.52 I. Skills and Education of 40.51 12.80 30.7 NA NA 38.98 45 46.3 23.81 Available Workers J. Business Licensing and 2.53 14.50 21.3 NA NA 3.39 0 3.7 0 Operating Permits K.Access to Financing NA 37.60 22.8 NA NA NA NA NA NA (e.g. collateral) L. Access to Domestic Credit 25.32 NA NA NA NA 28.81 15 27.78 19.05 M.Access to Foreign Credit 21.79 NA NA NA NA 22.03 21.05 18.52 25 N. Cost of Financing 36.71 42.60 21.8 NA NA 38.98 30 37.04 33.33 (e.g. interest rates) O.Regulatory Policy Uncertainty 29.11 40.10 32.9 NA NA 25.42 40 25.93 42.86 P.Macroeconomic Instability 79.75 34.40 30.2 NA NA 81.36 75 79.63 76.19 (inflation, exchange rate) Q. Corruption 2.53 40.40 27.3 NA NA 1.69 5 3.7 0 R.Crime, theft and disorder 1.27 21.50 20.0 NA NA 1.69 0 1.85 0 S. Anti-competitive or informal 7.59 21.30 23.7 NA NA 10.17 0 9.26 4.76 practices T. Legal system/conflict resolution NA NA 7.1 NA NA NA NA NA NA Annex 5. Standard Investment Climate Tables 72 Table A.5.4. Infrastructure indicators Eritrea Pakistan China India Morocco Small Large Low High capacity capacity Freq of power outages 105.44 14.50 NA NA 15.9 91.42 144.00 105.65 123.00 (times last yr) % of production lost due to 5.47 5.40 2.00 NA NA 4.98 6.83 5.55 6.14 power outages Have own generator (%) 43.04 41.80 16.20 68.94 16.73 38.98 55.00 46.30 38.10 Have own well (%) 48.10 43.80 15.60 50.77 29.14 47.46 50.00 46.30 52.38 % of production lost in shipment 0.36 na 1.20 NA NA 0.47 0.03 0.43 0.26 No. of days to obtain a 256.33 47.30 12.00 NA NA 303.36 58.80 292.48 46.67 telephone connection No. of days to obtain an 98.68 46.80 19.00 NA NA 125.13 19.30 92.26 135.00 electricity connection No. of days to obtain a NA NA NA NA NA NA NA NA NA water connection Share of Exports/ GDP NA 16.7 18.90 NA NA NA NA NA NA Table A.5.5. Sources of finance Eritrea Pakistan China India Morocco Small Large Low High capacity capacity Share of working capital from: retained earnings 73.96 65.40 51.5 30.37 61.95 72.76 77.45 72.53 79.76 banks, other financial institutions 23.45 7.40 20.6 36.10 19.55 23.78 22.50 25.19 16.38 trade credit 0.26 4.60 4.1 NA 7.58 0.34 0.00 0.38 0.00 equity 0.01 12.70 0.6 13.04 2.56 0.00 0.05 0.02 0.00 informal sources 1.28 1.30 8.6 NA NA 1.72 0.00 1.89 0.00 other 1.04 8.20 6.3 20.48 8.31 1.40 0.00 0.00 3.86 Financing of new investments from: retained earnings 63.08 55.60 NA NA NA 62.00 66.33 60.90 65.29 banks, other financial institutions 31.17 8.20 NA NA NA 37.22 13.00 31.79 31.18 trade credit 0.00 1.70 NA NA NA 0.00 0.00 0.00 0.00 equity 2.67 14.10 NA NA NA 0.00 10.67 2.56 3.53 informal sources 0.58 2.60 NA NA NA 0.78 0.00 0.90 0.00 other 2.50 11.00 NA NA NA 0.00 10.00 3.85 0.00 Annex 5. Standard Investment Climate Tables 73 Table A.5.6. Credits, loans and liabilities Eritrea Pakistan China India Morocco Small Large Low High capacity capacity Share with overdraft or line 47.44 22.80 21.80 NA 77.38 48.28 45 47.17 47.62 of credit Percent of credit that is currently 29.53 43.50 29.70 NA 25.46 21.16 55.56 35.40 19.60 unused Share with a loan from a bank or 44.87 19.50 45.80 NA NA 50.00 30.00 47.17 38.10 financial institution For the most recent loan or overdraft: Share that require collateral 85.71 80.30 82.80 NA NA 89.66 66.67 88.00 75.00 Average value of collateral 168.06 70.90 86.80 NA NA 174.11 130.25 179.60 149.88 required (as % of the loan) Average interest rate on loan 9.81 14.80 5.40 NA NA 9.78 10.00 9.80 9.79 Average duration of the loan (mns) 46.53 8.30 15.10 NA NA 44.61 60.00 45.17 60.86 Share of your total borrowing 2.86 0.50 7.70 9.49 3.89 3.45 0.00 0.00 12.50 denominated in foreign currency Share of long-term liabilities 29.50 7.10 9.40 28.69 8.96 33.81 17.50 33.19 18.11 (1 year or more) in total liabilities Share of short-term liabilities in 66.53 13.50 47.00 22.82 46.63 63.36 75.36 63.84 74.20 total liabilities Share of equity earnings 52.80 63.40 43.70 48.48 30.73 55.85 43.15 53.49 43.69 (or share capital) and retained in total liabilities Annex 5. Standard Investment Climate Tables 74 Table A.5.7. Financial sector -- auditing, transaction costs, and property rights Eritrea Pakistan China India Morocco Small Large Low High capacity capacity Share of firms whose financial 89.74 41.60 NA NA NA 87.93 95.00 88.68 90.48 statements are audited by outside auditors: Days to clear the following payments through your financial institution check (days) 1.44 1.90 4.5 NA NA 1.43 1.47 1.46 1.33 a domestic currency wire (days) 1.43 2.40 4.8 NA NA 1.45 1.40 1.94 0.80 a foreign currency wire (days) 6.33 3.20 3.0 NA NA 7.67 5.00 7.00 3.00 Share of land that is: owned (%) NA 87.90 44.00 NA NA NA NA NA NA leased or rented (%) NA 11.60 22.20 NA NA NA NA NA NA Average length of lease/rental NA 185 75.4 NA NA NA NA NA NA contract (mns) Share of buildings that are: owned (%) NA 90.50 56.90 NA NA NA NA NA NA leased or rented (%) NA 9.00 43.20 NA NA NA NA NA NA Average length of lease/rental NA 91.5 53.2 NA NA NA NA NA NA contract (mns) Annex 5. Standard Investment Climate Tables 75 Table A.5.8. Labor and training in international comparison and by firm characteristic Eritrea Pakistan China India Morocco Small Large Low High capacity capacity Labor composition 91.13 86.60 85.30 NA NA 93.49 84.19 93.78 88.10 Share of workers that are permanent Share of permanent workers 41.35 3.10 42.50 NA NA 39.27 47.50 41.37 44.92 that are female Share of temporary workers 29.46 NA 19.40 NA NA 27.11 32.78 32.76 33.03 that are female Share of permanent skilled workers 2.94 1.95 NA NA NA 3.38 1.74 2.66 1.79 that are foreign nationals Labor turnover New employees as share of total 14.92 8.30 NA NA NA 15.46 13.34 16.38 12.93 Employees that left as share 17.32 5.50 10.20 NA NA 17.52 20.06 12.45 of total Average time to fill for skilled NA 1.50 5.2 NA NA NA 16.74 NA NA technician vacancy Average time to fill production/ NA 1.30 13.9 NA NA NA NA NA NA service worker vacancy Excess workforce due to NA 3.60 16.70 NA NA NA NA NA NA regulatory restrictions Training and education Share of workforce with less than NA 59.30 1.10 NA NA NA NA NA NA 6 years schooling Share of workforce with more than NA 9.50 3.00 NA NA NA NA NA NA 12 years schooling Share of firms offering formal 9.59 11.10 71.70 NA NA 10.91 NA 8.16 14.29 training Share of skilled workers NA 36.00 47.70 NA NA NA 5.56 NA NA receiving training Labor Unrest Total days lost to labor disputes NA 1.3 0.3 NA NA NA NA NA NA or civil unrest Annex 5. Standard Investment Climate Tables 76 Table A.5.9. Regulatory burden and administrative delays by country Dimension Eritrea Pakistan China India Morocco Small Large Low High capacity capacity Interpretations of regulations NA 64.8 14.10 NA NA NA NA NA NA consistent, predictable (% disagreeing) % senior management's time 4.96 10.10 11.80 NA NA 3.45 9.66 7.23 4.43 spent dealing with regulations % revenues typically paid to 1.55 2.00 1.40 NA NA 2.05 0.11 5.26 0.29 officials to "get things done" % total firm revenues typically 84.20 NA 95.90 NA NA 81.46 91.33 88.62 85.16 reported for tax purposes Total wait in days for business NA NA 22.00 NA NA NA NA NA NA registration (days) Inspections a) Total days spent in inspections 6.32 41.10 29.10 10.7 NA 1.48 20.37 20.18 1.26 or required meetings with officials (days) b) % of meetings/inspections by NA NA NA 68.00 NA NA NA NA NA Local Authorities c) Total cost of fines or seized 0.02 0.23 0.60 NA NA 0.03 0.00 0.07 0.00 goods (% Sales) d) % of interactions in which NA NA 2.20 NA NA NA NA NA NA informal payment requested e) If yes, value? (% Sales) NA 0.15 NA NA NA NA NA NA NA Imports Avg. days to clear customs (days) 11.79 17.20 7.90 10.60 2.70 8.49 20.80 18.14 9.93 Longest delay to clear customs 22.62 30.20 12.50 21.20 5.40 14.15 45.20 34.79 18.75 (days) Exports Avg. days to clear customs (days) 3.20 9.60 5.40 5.00 1.70 3.46 2.75 1.00 3.43 Longest delay to clear customs 3.90 17.10 8.00 9.20 2.70 2.17 6.50 1.00 4.22 (days) Annex 5. Standard Investment Climate Tables 77 Table A.5.10. Governance -- uncertainty and corruption Eritrea Pakistan China India Morocco Small Large Low High capacity capacity Uncertainty How consistent/predictable are NA 14.3 30.0 NA NA NA NA NA NA government interpretations of regulations? Share of profits reinvested in NA NA NA NA NA NA NA NA NA the firm? Confidence in the judiciary NA 4.1 2.4 NA NA NA NA NA NA Percent of payment disputes NA 30.2 5.3 NA NA NA NA NA NA resolved in the courts Planning Horizon for NA NA 2.3 NA NA NA NA NA NA Investments (months) Corruption Percent of revenues that are 1.55 NA 1.4 NA NA 2.05 0.11 0.29 5.26 needed for informal payments? % Saying Gift/Payment Required for: a) A mainline telephone connection NA NA NA NA NA NA NA NA NA b) An electrical connection NA NA NA NA NA NA NA NA NA c) A construction permit NA NA NA NA NA NA NA NA NA d) An import license NA NA NA NA NA NA NA NA NA e) Operating license NA NA NA NA NA NA NA NA NA % of Revenue Reported by 84.20 NA 95.9 NA NA 81.46 91.33 85.16 88.62 Typical Establishment for Tax Purposes