71818 Cape Verde Investment Climate Assessment Regional Program on Enterprise Development (RPED) Africa Private Sector Group (AFTPS) ABBREVIATIONS ......................................................................................................................... 5 EXECUTIVE SUMMARY ............................................................................................................. 6 CHAPTER 1: INTRODUCTION ................................................................................................. 16 I. MACROECONOMIC BACKGROUND ................................................................................ 16 I.1 Economic Growth ................................................................................................. 16 I.2 Economic Policy ................................................................................................... 17 I.3 Macroeconomic Stability ...................................................................................... 18 I.4 Size of Government and Fiscal Deficit ................................................................. 18 II. EXTERNAL SECTOR ...................................................................................................... 19 II.1 Remittances ........................................................................................................... 19 II.2 Trade Policy .......................................................................................................... 20 CHAPTER 2: AN ANALYSIS OF FIRM PERFORMANCE...................................................... 25 I. THE INVESTMENT CLIMATE SURVEY ............................................................................ 25 I.1 Comparator Countries. .......................................................................................... 25 I.2 Firm Performance .................................................................................................. 27 II. LABOR PRODUCTIVITY ................................................................................................. 28 III. LABOR COSTS ............................................................................................................... 31 IV. CAPITAL PRODUCTIVITY ............................................................................................... 33 V. COMPETITION ............................................................................................................... 35 VI. IMPORTS ....................................................................................................................... 36 VII. TOTAL FACTOR PRODUCTIVITY (TFP) ......................................................................... 37 VIII. EXPORTS ...................................................................................................................... 43 CHAPTER 3: PERCEPTIONS ABOUT THE INVESTMENT CLIMATE IN CAPE VERDE . 51 I. MAIN PERCEIVED CONSTRAINTS ................................................................................... 52 2 II. DIFFERENCES IN PERCEPTIONS ACROSS DIFFERENT FIRMS ............................................ 54 III. REGIONAL DIFFERENCES IN PERCEPTIONS.................................................................... 55 CHAPTER 4: THE LABOR MARKET IN CAPE VERDE......................................................... 58 I. THE WAGE STRUCTURE IN CAPE VERDE ...................................................................... 58 II. ECONOMETRIC ANALYSIS OF DETERMINANTS OF HOURLY WAGES ................................ 64 III. MANUFACTURING WAGES IN CAPE VERDE COMPARED TO OTHER COUNTRIES ............. 70 IV. TRAINING ..................................................................................................................... 71 CHAPTER 5: ACCESS TO FINANCE ........................................................................................ 75 I. THE FINANCIAL SECTOR IN CAPE VERDE ..................................................................... 75 II. ACCESS TO FINANCE IN AN INTERNATIONAL PERSPECTIVE .......................................... 78 III. DIFFERENCES IN ACCESS BY FIRM TYPE ...................................................................... 83 IV. ACCESS TO FINANCE FOR MICROENTERPRISES .............................................................. 85 CHAPTER 6: OTHER ASPECTS OF THE INVESMENT CLIMATE ...................................... 89 I. ELECTRICITY ................................................................................................................ 89 I.1 Electricity and Water Sector Reform .................................................................... 89 I.2 Evidence from the Investment Climate Survey. .................................................... 90 I.3 Regional differences in the quality of electricity .................................................. 94 II. TAXATION .................................................................................................................... 95 II.1 Tax Policy ............................................................................................................. 96 III. INFORMALITY ............................................................................................................. 100 III.1 Informal Sector in Cape Verde ............................................................................ 100 III.2 Barriers to Becoming Formal .............................................................................. 101 IV. REGULATION .............................................................................................................. 102 IV.1 Regulatory reform in Cape Verde ....................................................................... 102 IV.2 The Burden of Regulation ................................................................................... 104 3 IV.3 Regulatory Uncertainty. ...................................................................................... 107 V. OTHER AREAS OF THE INVESTMENT CLIMATE ........................................................... 108 V.1 Corruption and other measures of governance .................................................... 108 V.2 Macroeconomic stability ..................................................................................... 112 V.3 Telecommunications ........................................................................................... 113 CHAPTER 7: CONCLUSIONS AND RECOMMENDATIONS .............................................. 119 I. IMPROVING FIRM COMPETITIVENESS........................................................................... 120 II. REDUCING BARRIERS TO EXPORTING. ......................................................................... 123 APPENDIX 1: INTERNATIONAL COMPARISONS OF PERCEPTIONS OF THE INVESTMENT CLIMATE......................................................................................................... 124 APPENDIX 2: STATISTICAL APPENDIX ON ACCESS TO FINANCE .............................. 126 APPENDIX 3: SAMPLE CHARACTERISTICS ....................................................................... 131 I. THE INVESTMENT CLIMATE SURVEY (ICS) ................................................................ 131 II. THE MICROENTERPRISE INVESTMENT CLIMATE SURVEY (MICS) .............................. 132 REFERENCES ............................................................................................................................ 134 4 ABBREVIATIONS ARE Economic Regulatory Agency AGOA African Growth and Opportunity Act ASCUDA Automated System for Customs Data BCA Banco Comercial do Atlantico BCV Banco de Cabo Verde BVC Bolsa da Valores de Cabo Verde CAS Country Assistance Strategy CVT Cabo Verde Telecom ECOWAS Economic Community of West African States EDP Electricitade de Portugal EPZ Export processing zone EU European Union FDI Foreign direct investment FIAS Foreign Investment Advisory Service GDP Gross Domestic Product GoCV Government of Cape Verde ICA Investment Climate Assessment ICS Investment Climate Survey IMF International Monetary Fund INE National Institute of Statistics IPP Independent power producer ISEG Instituto Superior de Economia e Gestão ISP Internet Service Provider LAD Least Absolute Deviations M2 Money and Quasi-Money MGF Matching Grant Fund MICS Microenterprise Investment Climate Survey MSE Micro and Small Enterprise NOSI Núcleo Operacional para a Sociedade de Informação NPL Non-performing loan OLS Ordinary Least Squares PPP Purchasing power parity PRGF Poverty Reduction and Growth Facility PRSP Poverty Reduction Strategy Paper PTI Portugal Telecom Internacional RPED Regional Program for Enterprise Development SITC Standard International Trade Classification SISP Sociedade Interbancária e Sistemas de Pagementos SME Small and Medium-Sized Enterprise UNCTAD United Nations Conference on Trade and Development US United States VAT Value-Added Tax WTO World Trade Organization 5 ACKNOWLEDGEMENTS The Cape Verde Investment Climate Survey was entirely enumerated by EEC Canada, a survey firm based in Montreal, Canada, and headed by Fares Khoury. The sampling frame and sample was constructed with considerable help from the National Institute of Statistics (INE), especially Francisco Tavares (President of INE). The report incorporates contributions from a team of World Bank staff and consultants including Philippe Alby, Rita Almeida, Sherri Archondo, Jessica Bloomgarden, George Clarke (team leader), Linda Cotton, Manju Shah, Ginger Turner, and Yutaka Yoshino. Giuseppe Iarossi, Jean Michel Marchat and Melanie Mbuyi managed the survey during the set-up and enumeration period. Julio Fortes also provided substantial help in setting up the survey. The peer reviewers were Mary Hallward-Driemeier and Alvaro Gonzales. We are also grateful to Iradj Alikhani, Joelle Dehasse, Manuela Francisco, Francoise Perrot, and Vijaya Ramachandran for advice and comments. 6 EXECUTIVE SUMMARY Cape Verde is a small country with a population of about 472,000 people spread over nine islands in a ten-island archipelago. It is more developed that most other countries in Sub-Saharan Africa. Its per capita Gross Domestic Product (GDP) per capita is US$5,715 in purchasing power parity (PPP) adjusted terms in 2004. The economy is dominated by the service sector, which accounted for about 75.5 percent of GDP in 2005. Cape Verde has grown steadily in recent years, with GDP growth averaging about 5.7 percent between 2000 and 2005 and reaching 6.3 percent in 2005. The fastest growing sectors are hotels (15 percent annual growth between 2000 and 2003), commerce (10 percent), construction (9 percent), and transportation and communication (7 percent). Agriculture (2 percent) and industry (0 percent) have been growing more slowly. As a small island economy, Cape Verde is heavily dependent upon external economies. Remittances and foreign aid are important sources of capital and the economy is heavily dependent upon imports for much of its consumption and investment. Recent growth in the tourism sector has further increased Cape Verde‘s integration into the world economy. Cape Verde has an extremely large trade deficit. Exports of goods and services were equal to about 17 percent of GDP. In comparison, imports were equal to about 53 percent of GDP. Cape Verde‘s exports are dominated by services, with only a very small amount of merchandise exports. Apparel, shoes and fishery products account for most manufacturing exports. Between 80 and 100 percent of each of those products, disaggregated to Standard International Trade Classification (SITC) 4-digit level, are exported to single countries (either Portugal or the United States). This high concentration of exports suggests that Cape Verde suffers from thin market channels and buyer-supplier networks. Cape Verde‘s economy is therefore vulnerable to external shocks, especially given its narrow export base. Prudent macroeconomic policy—particularly low inflation and a credible exchange rate peg—reduce Cape Verde‘s vulnerability. Expanding and diversifying the export base, however, would be a useful way of further reducing vulnerability and would allow Cape Verde to sustain and continue to improve its standard of living (World Bank, 2004a). The Investment Climate Survey The main sources of information for the Investment Climate Assessment (ICA) are two surveys carried out in Cape Verde in March and April 2006. Both surveys were conducted in two locations, Praia and Mindelo. The first survey, the Investment Climate Survey (ICS), covered formal enterprises with over five employees in manufacturing, retail trade, construction, and other services. Firms were randomly selected from lists provided by the National Institute of Statistics (INE). The second survey, the Microenterprise Investment Climate Survey (MICS), covered microenterprises in the same sectors. Firms in this sample were selected randomly in prescribed areas of the cities. This approach means that the survey will cover both registered and unregistered microenterprises. Because the two surveys were sampled using different methodologies—and because there is no way to weight the firms in the two samples—they are not pooled in the analysis. 7 One advantage that Investment Climate Surveys (ICSs) have over other firm-level analyses is that similar surveys have been conducted in a large number of other countries in Sub-Saharan Africa and throughout the world. This means that it is possible to benchmark firm performance and measures of the investment climate in Cape Verde against other countries. Because Cape Verde is more developed than most other countries in Sub-Saharan Africa, the comparator countries include some of the more developed countries in the region where investment climate surveys have been completed (Senegal and South Africa), other middle-income island economies (Mauritius and the Maldives), middle-income countries in the Caribbean (the Dominican Republic and Guyana) and lower middle-income countries in East Asia (Indonesia and the Philippines). Although none of these countries is ideal in all respects, most share some characteristics with Cape Verde. Because most investment climate surveys either cover only the manufacturing sector or do not contain weights that would make it possible to produce results for the entire economy, comparisons are presented only for manufacturing. Firm performance in the manufacturing sector. Labor productivity is relatively high in Cape Verde (see Figure 1). Manufacturing enterprises in Cape Verde are more productive than similar enterprises in the other lower middle income comparator countries. The median firm in Cape Verde produces about $6,100 of output per worker, substantially more than in Indonesia, the Philippines and the Dominican Republic (about $2,200, $2,400 and $2,300 per worker respectively). The median firm also produces slightly more than medina firms in Guyana and Senegal (about $5,500 and $5,600 per worker). Although manufacturing firms in Mauritius, the Maldives, and South Africa produce more value-added per worker ($7,000, $6,800, and $14,000 respectively), they have considerably higher per capita income. As a result, it is not surprising that labor productivity is also higher. Figure 1: Manufacturing firms are more productive in Cape Verde than in other middle income countries — although they lag behind firms in the most productive upper middle income countries. $20,000 $15,000 US Dollars $10,000 $5,000 $0 a s es l a ia de s ic ga iu ne an ric l es ub iv r ri t ne Ve Af pi uy d n au ep al do ilip Se G h e M M R ut In ap Ph n So C a ic in om D Value-Added per Worker Value-Added per worker in Garments Source: Investment Climate Surveys Note: All values are medians for enterprises with available data. Value added is calculated by subtracting intermediate inputs and energy costs from sales from manufacturing. Workers include both permanent and temporary workers. Values are converted to US$ using average exchange rates from World Development Indicators. Data were collected between 2002 and 2005 depending on survey period for each country. 8 A more detailed analysis of total factor productivity—a measure of firm performance that takes into account enterprises‘ use of capital—leads to a similar conclusion. Manufacturing enterprises in Cape Verde are more productive than in the lower middle income countries, but less productive than upper middle income countries such as South Africa or Mauritius. Despite being relatively productive, only 6 percent of manufacturing firms in Cape Verde export—fewer than in any of the comparator countries including other small island economies such as the Maldives and Mauritius. Moreover, very few firms both export and sell in domestic markets—about 2 percent of the sample. The low propensity to export is consistent with the macroeconomic evidence—merchandise exports are modest. The empirical analysis suggests several reasons for this:  Because there are large fixed costs associated with exporting, small firms are less likely to export than large firms. The small domestic market means that firms in Cape Verde are smaller than in many of the comparator countries, making it harder for them to export than the larger firms in the comparator countries.  Although productivity is high, so are labor costs. In fact, unit labor costs —a measure of labor costs that take productivity into account—are higher in Cape Verde than in any of the comparator countries. They are even higher than in South Africa and Mauritius, two countries where labor costs are also very high by international comparison. This makes it difficult for firms to compete on international markets.  Cape Verde‘s physical isolation and distance from major shipping routes makes exporting more challenging for firms in Cape Verde. Although this is likely to explain some of the difference between Cape Verde and its competitors, it is important not to overemphasize this. Firms from other countries manage to overcome this obstacle to compete with domestic firms within Cape Verde—most consumer and investment goods are imported. Moreover, firms in Cape Verde are less likely to export than firms in the small island comparator countries (e.g., the Maldives and Mauritius).  Firms report that customs and port procedures are relatively inefficient in Cape Verde. It takes longer for goods to complete all customs and port procedures in Cape Verde than in most of the middle income comparator countries. Results from the ICS do not distinguish between delays due to problems on the side of the importer (e.g., delaying paying duties), port delays, and customs delays. Given the recent improvements in customs administration in Cape Verde, it seems plausible that the delays might be due either to port delays or to delays on the side of the importers. In any case, delays might discourage firms from exporting. Perceptions about the Investment Climate. In addition to collecting information on firm productivity, the ICS also collects information on the quality of the investment climate—including on topics such as infrastructure, access to finance, crime and security, regulation, corruption, and taxation. Two types of information are collected: (i) perception-based measures that ask managers what they see as the major obstacles that their firm faces; and (ii) objective indicators such as production lost due to power outages, whether 9 the firm has a loan or overdraft facility, and amount of time managers spend dealing with government regulations. The report uses both types of indicator—and supplementary information from other sources—to assess constraints to enterprise operations and growth in Cape Verde and compare these with constraints in the comparator countries Perception-based measures provide a good starting place for an analysis of the investment climate. Although perception-based measures suffer from several problems, an enterprise manager probably has a better grasp of the immediate problems facing his or her business than government officials, academic researchers, or other outside experts. When asked about the problems that they faced, managers were most likely to say that the electricity was a problems—close to two-thirds of enterprise managers said that it was a major or very severe obstacle (see Figure 2). Managers were also likely to say that other problems were serious obstacles. Nearly 50 percent of managers said that tax rates were a major or very severe obstacle; over 40 percent of managers were concerned about the cost of financing; and over one- third of enterprise managers rated anti-competitive or informal practices and access to financing as serious problems. Firms were considerably less concerned about other issues—less than one fifth of firms said that other constraints such as legal system, transportation, labor regulation, business licensing, corruption, access to land, and macroeconomic policy were serious problems. Figure 2: Firms in Cape Verde are most concerned about electricity, tax rates, cost of financing, informal sector competition and access to financing % of firms saying issue is serious problem 80% 60% 40% 20% 0% bo ice n sp tion es d d on El tes ke ax gal tion c o E d io n g nd m an r ty Tr gu g ef eg ns to ty g n rim ra unic ion Fi rde tio Ta cin e in om cin f F itio ci ili ki dm ste fo Acc an lati R tio La a t r R ns at ce tab tri ss rup a n la a Te and stra o an R t o pet a Le ort Sy m uc ec u is x Ac Ins in n r i si Co in ss L ic m Se to C e, de om ne or rS A s t os La ct on lls th T C le Bu T ec ro al rm ac C M or In W Source: Investment Climate Surveys. For the most part, firms in different sectors of the economy (manufacturing firms, retail trade and service firms and hotels) had the same concerns. Electricity and tax rates are rated among the top five problems for all types of firms. Other areas, such as corruption, courts, and 10 macroeconomic policy are seen as problems by few firms in any sector. However, there were some differences. Manufacturing firms were more likely to be concerned about access to financing. Hotels were more concerned about the quality of the labor force in terms of education and skills and labor regulation. And firms in retail trade and other services appear to be more concerned about crime and theft. There were also some differences in perceptions between the two cities. First, although enterprises on both islands saw electricity as a serious problem, firms in Praia were far more concerned—nearly 80 percent of enterprises said that electricity was a serious problem on Praia compared to only about 40 percent in Mindelo. Second, firms on Praia were far more concerned about worker skills than firms in Mindelo. Finally, firms on Mindelo were far more concerned about financing and telecommunications than firms in Praia. Labor Markets There are significant inter-industry and intra-industry wage premiums. After controlling for the worker‘s education and skills, average wages in non-manufacturing firms as well as in older, larger and in formal firms tend to be higher than elsewhere. This is consistent with theories of compensating differentials or with efficiency wage theories. Unfortunately, with the available evidence it is hard to disentangle the different theories. There is robust evidence of high returns to schooling, to training and to specific occupational skills. An individual deciding to stay an additional year in school receives on average a return of 11 percent on this investment. Wages of individuals that received some training in the past are also 16 percent higher. This suggests that individuals have a strong incentive to invest in the acquisition of these skills since market returns are high. The results also suggest that women receive lower than men after controlling for observable differences. On average they earn 13 percent lower wages than men in the same occupation. Compared with other countries, wages in Cape Verde are high. Median wages in manufacturing for an unskilled worker is approximately US$2,000 a year. This is well above wages in Senegal or the Philippines and is approximately 50 percent of South Africa, one of the middle- income countries with the highest labor costs. However, we also find evidence that managers in Cape Verde tend to invest in on-the-job training which is taken to be a sign of the flexibility and adaptability of the workforce in the country. Access to Finance About 45 percent of manufacturing firms in Cape Verde have loans—higher than in any of the comparator countries. Consistent with this, firms in Cape Verde finance more new investment through bank loans than firms in the comparator countries. Moreover, this probably underestimates the relative performance of the financial sector on this dimension. The median firm in Cape Verde is smaller than the median firm in most of the other countries. Since small firms are typically less likely to have access to finance than larger firms, this might affect the relative share of firms with loans. Despite this, firms in Cape Verde are heavily reliant upon retained earnings for much of their investment. The average firm financed about 68 percent of new investment through internal 11 funds, higher than in 5 of the 8 comparator countries. Since they finance more with bank loans than in most of the comparator countries, this reflects the relatively modest levels of investment that firms in Cape Verde finance with money from informal sources (mostly family and friends) and non-bank formal sources. Access to credit is a greater problem for microenterprises than other enterprises in Cape Verde. Microenterprises are less likely to have loans and overdraft facilities, finance less of their short-term assets with bank financing and are more likely to say that they are credit constrained (i.e., that they would like a loan at current interest rates but are unable to get one). This, however, is the case in almost all developing countries. When compared to microenterprises in other countries, microenterprises in Cape Verde are not especially disadvantaged with respect to access to finance. Other Areas of the Investment Climate Electricity. When compared to low income countries in Sub-Saharan Africa, the power sector does not perform especially poorly. The median number of power outages that firms in Cape Verde reported was 8—in comparison, the median number of outages was 48 in Tanzania in 2003, 21 in Kenya in 2003 and 15 in Senegal in 2004 (Regional Program on Enterprise Development, 2004a; Regional Program on Enterprise Development, 2004b; Regional Program on Enterprise Development, 2005b). This is also lower than in Guyana, where the median firm reported 15 outages. However, it is significantly higher than in the other middle income comparator countries. In South Africa, Mauritius, Philippines, Indonesia and Maldives, the median manufacturing firm reported between 0 and 4 outages per year. Losses due to outages are higher in Cape Verde than in any of the comparator countries except the Dominican Republic. For example, although the median firm in Cape Verde reported fewer outages than median firms in Senegal or Guyana, losses due to outages were greater. Although this might seem peculiar, there are several plausible explanations. First, firms are affected differently by outages. For example, losses will be greater when outages result in equipment damage or when the enterprise is unable to make up for lost production by running extra shifts. Second, firms with generators are more able to cope with outages than other firms. Very few firms in Cape Verde have generators, potentially increasing the cost of outages. Tax Rates. About 55 percent of small and medium-sized enterprises in the services sector, 47 percent of in the manufacturing sector and 29 percent of hotels report that tax rates are a major or very severe constraint on their operations and growth. Despite the high levels of concern, marginal tax rates do not appear to be out-of-line with tax rates in other low and middle income countries. The VAT rate (15 percent) is lower in Cape Verde than in Senegal (20 percent) and the Dominican Republic (16 percent), is comparable to Mauritius (15 percent), and is higher than in Indonesia (10 percent), the Philippines (10 percent), and South Africa (14 percent). The top corporate tax rate is also not out of line with the comparator countries. The top tax rate (30 percent) is slightly lower than in Senegal (33 percent), the Philippines (32 percent) and Mauritius (35 percent) but is the same as in South Africa and Indonesia (30 percent). Marginal tax rates, however, might not provide a very accurate picture of the overall burden of taxation in the comparator countries. Many other things—including depreciation rates, other taxes that firms face, and a wide range of fiscal incentives that firms can be eligible for —affect the 12 total burden of taxation. The World Bank‘s Doing Business report (World Bank, 2005b) provides a more detailed estimate of the overall tax burden computed for a standardized firm. Unfortunately, Cape Verde was not included in the Doing Business Report in 2005 and data for the 2006 report were not yet available. Once data from the 2006 report is available, it will provide a better international benchmark than the simple marginal rates do. Competition from the informal sector. Another area of concern was unfair competition from informal firms. In practice, however, evidence from the investment climate survey and other sources does not generally suggest that informality is particularly high in Cape Verde. So why are firms so concerned about informality? One reason is that formal firms in Cape Verde tend to be small—and smaller firms are more likely to find themselves competing with microenterprises. Consistent with this microenterprises—and especially registered microenterprises—were especially concerned about informality. About 59 percent of registered microenterprises said that informality was a major or very obstacle, compared to 40 percent of very small formal enterprises, 36 percent of small formal enterprises and 21 percent of medium-sized formal enterprises. The concern about informality, therefore, might primarily reflect that most formal enterprises in Cape Verde are small. Regulation. The burden of regulation appears relatively high in Cape Verde—although not completely out-of-line with the comparator countries. On average, manufacturing enterprise managers spend a little more time dealing with government regulations in Cape Verde than in any of the comparator countries except for Senegal (see Figure 51). The average manufacturing enterprise manager reports spending about 12 percent of his time dealing with regulations and inspections, compared to less than 10 percent in South Africa, Philippines, Guyana, Maldives or Indonesia. A different pattern can be observed with respect to inspections. On the narrower—although more easily measured—indicator, the burden of regulation seems more modest in Cape Verde. On average, managers of manufacturing enterprise managers report fewer inspections in Cape Verde than managers in any of the comparator countries except for Indonesia and Maldives. Whereas the average manager reports only 2 inspections in Cape Verde, the average manager in South Africa report 14 and the average manager in the Dominical Republic reports over 30 inspections. Other areas of the Investment Climate. Firms were far less concerned about most other areas of the investment climate. For example, few firms saw corruption, macroeconomic instability, or telecommunications as serious obstacles. The objective evidence from the investment climate survey—and other evidence from other sources—seem consistent with this. For example, few firms reported that they needed to pay bribes to get things done, inflation and exchange rate instability is modest, and the telecommunications sector performs fairly well—although few firms reported using the Internet to communicate with suppliers and customers. Recommendations As noted above, manufacturing enterprises in Cape Verde are more productive than similar enterprises in the other lower middle income comparator countries. A more detailed analysis of total factor productivity—which also takes into account enterprises‘ use of capital—leads to a similar conclusion. Despite being relatively productive, few manufacturing firms in Cape Verde 13 export (only 6 percent). This is lower than in any of the comparator countries including the other small island economies such as the Maldives and Mauritius. So what can the Government of Cape Verde do to expand and diversify the country‘s export base? Some reasons why few firms export are structural and cannot easily be addressed through policy interventions. For example, the large fixed costs associated with exporting mean that small firms are less likely to export than large firms. Because firms in Cape Verde operate in a small domestic market, they are likely to be small and therefore will find it harder to export than the larger firms in the comparator countries. Similarly, Cape Verde‘s physical isolation and distance from major shipping routes makes exporting more challenging for firms in Cape Verde. These factors will make it difficult for firms to compete in labor intensive areas of manufacturing and emphasize the importance of developing other industries (e.g., tourism, financial services, or IT services). Although these constraints will generally make it harder for firms from Cape Verde to operate in international markets than firms in some of the comparator countries, this does not mean that the GoCV can not do anything. Policies aimed at improving competitiveness will both promote exports and improve the performance of domestic firms. Reducing barriers to trade might also promote exports. Improving Competitiveness Although the investment climate in Cape Verde is quite favorable, some problems remain. The biggest problem is the poor performance of the power sector. This, in turn, reflects the low level of investment in the sector that is the result of poor financial performance of the private operator, Electra. Two recent World Bank reports, a review of the infrastructure sector and the recent Public Expenditure Review, present various recommendations on how to improve sector performance (World Bank, 2006a; World Bank, 2006b). Perhaps the most important recommendation is to adjust tariffs to a level that would allow Electra to self-finance its investment requirements. In addition to taking steps to improve the financial viability of Electra, the infrastructure report includes several additional recommendations that might improve sector performance in ways that would improve services quality. These include resolving outstanding contractual issues between the GoCV and Electra; allowing independent power producers; and allowing auto- producers to sell power to the grid. Another important issue is reducing the high cost of labor. A first option is to increase the flexibility of labor regulations. Although firms did not indicate that labor regulations are a particular concern, previous studies have emphasized that labor regulations are inflexible. The inclusion of Cape Verde in the 2007 edition of the Doing Business report will provide a useful benchmark in this respect. A second option is to move into higher value-added sectors. Given the high premiums paid to educated and skilled workers, the Government will also have to take steps to improve the human capital of the workforce. In the medium-term, steps to improve secondary education will be important. In the short-term, the GoCV might look into other options such as supporting training programs. Although firms were very concerned about the burden of taxation, marginal tax rates do not appear to be out-of-line with other middle-income countries. This suggests the need for more 14 detailed information on what aspects of the tax systems might be improved without compromising revenues. One way of collecting this information would be to look at producing a more detailed report on the effective tax burden, such as those produced by the Foreign Investment Advisory Service (FIAS) Reducing Barriers to Exporting Another way of improving export performance would be to reduce remaining barriers to trade. A first option would be to reduce remaining tariffs and non-tariff barriers. Although tariffs were reduced in early 2004, they remain above international standards. Similarly, steps to improve the performance of the ports might also be useful. The recent World Bank report on infrastructure suggested several steps including improving regulatory administration and modernizing port administration as a complement to privatization. 15 CHAPTER 1: INTRODUCTION I. MACROECONOMIC BACKGROUND Cape Verde is a small country, with a population of 472,000 people spread over nine islands in the ten-island archipelago. It is more developed than most other countries in Sub-Saharan Africa. The country is classified as lower-middle income, with GDP per capita equal to US$5,715 (PPP) in 2004. I.1 Economic Growth Cape Verde has grown steadily since independence. Between 2000 and 2005, GDP grew at an average rate of 5.7 percent (see Figure 1). GDP growth continues to be relatively strong, with current estimates suggesting that the economy grew at about 6.3 percent in 2005. This rate of GDP growth is within the Government of Cape Verde‘s target of 5 to 7 percent per year. Figure 3: GDP growth has been strong over the past five years, averaging 6.3 percent 8 7 Annual percentage change 6 5 4 3 2 1 0 2000 2001 2002 2003 2004 2005 Source: World Bank (2006a) Cape Verde‘s economy is dominated by the services sector; tourism, transport, commerce and government spending accounted for 75.5 percent of GDP in 2005. The largest sectors in recent years have been commerce (19.6 percent of GDP in 2003), and transport and communications (18.6 percent of GDP in 2003). These two sectors have also been among the fastest growing, along with construction and hotels (see Table 1). Further growth is possible in both areas. In particular, Cape Verde has looked at the possibility of developing a transshipment hub at Porto Grande. Although 16 this might be challenging, progress has been made.1 In 2006, Maersk, a major shipping company, signed an agreement with the GoCV to have 72 ships stop in Porto Grande in Mindelo each year. Table 1: GDP Growth by Sector - Annual Percentage Change 2000 2001 2002 2003 Avg. Hotels 20.2 15.2 -6.6 32.9 15.4 Commerce 4.9 4.6 12.8 15.7 9.5 Construction 2.6 4.1 14.5 14 8.8 Transport & Communications 9.2 24.6 -6.7 -0.4 6.7 House renting 2.7 10.6 9 3.7 6.5 Public Services 5.4 0.5 5.4 11.6 5.7 Other Services 7.8 4.7 0.5 8 5.3 Banks & Insurance -16.5 20.1 6.8 6.4 4.2 Agriculture, Forestry & Livestock -1.1 -2.1 -2.4 13.7 2.1 Industry and energy -7.6 -10.2 14 4.8 0.3 Fishing 1.4 -17.3 -20.6 28.9 -1.9 Source: International Monetary Fund (2005a) Agriculture remains an important source of employment, accounting for about 23 percent of employment (compared to over 40 percent for services). Growth, however, has been modest, averaging only about 2 percent per year between 2000 and 2003. The results in the agricultural sector are mainly due to persistent drought. Despite the steady growth over the past decade, unemployment remains a pressing problem. According to the 2000 census, 22 percent of the labor force is unemployed (Ministry of Finance and Planning, 2004). Moreover, income distribution is uneven, with 37 percent of the population considered poor in 2002 and the majority of them in rural areas (67 percent). I.2 Economic Policy Since 1991, the Government of Cape Verde has been implementing a program of political and economic reform called Mudança which promotes the liberalization and diversification of the economy. The International Monetary Fund (IMF) sponsored Poverty Reduction and Growth Facility (PRGF) is consistent with the role of the government envisioned by Mudança and has achieved macroeconomic stability and progress on structural reforms, partly due to improvements in policy-making capacity since 2000. Macroeconomic stability has helped to create an environment conducive to increasing private investment, inflows of remittances and accumulation of emigrant deposits. The promotion of tourism is the cornerstone of the economic strategy. Public investment is increasing with a new international airport at Praia and on the island of Boa Vista. New private investment is coming through hotels on Sal and Boa Vista. Diversification is being pursued simultaneously however, because it is acknowledged that tourism can be vulnerable to the global security situation, the European economic situation and world oil prices. 1 For example PEP-Africa (2005) discusses some of the challenges including location, low returns from transshipment, and lack of support from detailed analyses. 17 On a broader level, the country‘s Poverty Reduction Strategy Paper (PRSP) was approved in September 2004 and is based on the National Development Plan for 2002-2005.2 The priorities identified are maintaining macroeconomic balances, modernizing the financial sector and increasing private sector competitiveness for growth. The World Bank‘s Country Assistance Strategy (CAS), approved in February 2005, is aligned with the PRSP and is based on three pillars: ensuring macroeconomic stability, supporting private sector led growth and alleviating poverty. I.3 Macroeconomic Stability The Government of Cape Verde (GoCV) has strengthened the exchange rate peg to the Euro (CVEsc110.265:€1) and increased its credibility by backing it with substantial international reserves. The peg was transferred to the Euro from the Portuguese escudo in 1998. The government‘s target is to have 2.5-3 months of import of goods and services by end-2007. The Government has been successful in this respect—reserve coverage increased from 1.0 to 3.4 months of current year imports between 2000 and 2005. The appreciation of the Euro (and hence the Cape Verdean escudo) against the dollar does not seem to be hurting the country‘s competitiveness due the strong links to Europe in terms of trade, tourism and remittance flows. The real effective exchange rate has been depreciating mainly due to low domestic inflation and fell 2.9 percent in 2004. The exchange rate peg helps to build anti-inflationary credibility in the currency and to control for inflation. The GoCV has set a target for inflation at around 2 percent. Annual average inflation has fallen steadily and averaged 0.5 percent, below target, between 2000 and 2005. I.4 Size of Government and Fiscal Deficit Although the reform program included elements that limited the role of the government, government spending increased in the early and mid-1990s. Government expenditure, which was only 25 percent of GDP in 1990, had increased to over 50 percent in 1994-53. Since this time, the Government has successfully reduced the size of government. By 2005, expenditures stood at 33 percent of GDP—higher than in the early 1990s, but lower than in the middle of the decade. The deficit was large during the mid 1990s and again during the lead up to 2000 elections when the deficit reached nearly 22 percent of GDP excluding grants and 16 percent including grants (see . The budget deficit probably had macroeconomic implications in terms of inflation and crowding out of the private sector, since it was mostly financed through the domestic banking system. Crowding out may have happened through a reduction in the volume of credits available to the private sector, rather than an increase in interest rates. However, the fact that government spending was significant in public investment in infrastructure could have offset this effect (Bourdet, 2001). In 2001, the deficit was much lower—about 12 percent of GDP excluding grants and 6 percent including grants. A large increase in foreign aid, 90 percent of which is usually in grant form, helped to consolidate the country‘s fiscal position along with the tax policy reforms. In 2 See Ministry of Finance and Development Planning (2004) 3 Data from World Bank (2006c). Data for 1994-1995 is roughly consistent with data from International Monetary Fund (1999, p. 4) 18 2005, the deficit was 8.5 percent of GDP excluding grants and 2.6 percent including grants. External debt and domestic public debt have also declined as a share of GDP. Figure 4: The government deficit has declined since 2000 due to both increased revenues and lower expenditures (as % of GDP) 5 0 2000 2001 2002 2003 2004 2005 -5 % of GDP -10 -15 -20 -25 Overall balance (excluding grants) Overall balance (including grants) Source: World Bank (2006a), Table 1.5 II. EXTERNAL SECTOR As a small island economy, Cape Verde is heavily dependent upon external economies. Remittances, and foreign aid, are important sources of capital, while the economy is heavily dependent upon imports for much of its consumption and investment. The recent growth of tourism (see Table 1) has increased Cape Verde‘s integration into the world economy. II.1 Remittances In 2000, twice as many Cape Verdeans lived abroad as lived in the country. Around two- thirds of families receive money from abroad, mostly from Western Europe and the United States. Emigrant remittances have been recorded between 10 and 15 percent of GDP during the early 1990s and 2000s (Table 2) and are mainly used by residents to fund construction and education. These figures are undoubtedly underestimated given that they do not account for certain sums such as those brought in during emigrant visits. Table 2: Private transfers have offset current account deficits, (2000-2005) 2000 2001 2002 2003 2004 2005 Emigrant remittances (percent of GDP) 15.9 14.2 13.6 11.3 11.3 Current account balance (percent of GDP) -15.3 -13.9 -17.1 -15.5 -12.4 -9.8 Overall trade balance -36.4 -37.1 -42.8 -37.5 -35.7 -36.3 Source: World Bank (2006a)and International Monetary Fund (2005a) Remittances are an important source foreign exchange and have helped maintain the fixed exchange policy through helping to increase international reserves. At the same time, an over- 19 dependence on remittances can lead to overspending on imports and can reduce the urgency to institute structural reforms. Inflows of emigrant remittances (and foreign aid) are offsetting current account deficits, which are due to high levels of imports relative to exports. Negative swings in remittance amounts can also be caused by events in host countries, such as war or changes in migration or labor policies. However, a recent IMF study on emigrant deposits stated that they did not seem to be affected by the introduction of the Euro or the events of September 11, 2001. Many island economies depend heavily upon remittances. For example, the Dominican Republic, Jamaica and Grenada have high remittance flows—and several of these have high current account deficits. Cape Verde, however, has exceptionally high inflows even relative to these economies. Although remittances are often used to finance consumption rather than investment, it is important to note that a large portion of remittances is put into special deposit accounts in Cape Verde (Bourdet and Falck, forthcoming). In 2004, gross remittance flows accounted for 20 percent of GDP and the net accumulation of emigrant deposits was almost 4 percent of GDP (International Monetary Fund, 2005b). This, in turn, has increased financial sector development, with a beneficial effect on investment, especially in real estate. Figure 5: Many island economies that rely on remittances have high current account deficits 40% Annual percentage change 20% 0% -20% -40% Maldives Mauritius Guyana Verde South Grenada Senegal Dominican Philippines Africa Cape Republic Source: World Bank (2006a; 2006c) II.2 Trade Policy Cape Verde has an extremely large trade deficit. Exports were equal to about 17 percent of GDP in 2005. In comparison, imports were equal to about 53 percent of GDP. The overall balance of payments, however, is in surplus due to inflows of emigrant remittances, which provide foreign exchange and help to maintain the fixed exchange policy. 20 Figure 6: Exports of goods and services are far lower than imports —leading to a large trade deficit 70 60 50 % of GDP 40 30 20 10 0 2000 2001 2002 2003 2004 2005 Exports of goods and services Imports of goods and services Source: World Bank (2006a) Cape Verde‘s export profile is influenced by geographical factors. As discussed earlier, Cape Verde is a service-oriented economy, with 75.5 percent of GDP coming from the services sector. In the most recent years for which comparable cross-country data are available (2003), merchandise exports accounted for only for 5 percent of total exports of goods and services, substantially less than any of the comparator countries (see Figure 7). Despite sizeble increases in 2005, they remain relatively low as a share of total trade. 4 Merchandise exports are much smaller than merchandise imports, less than 1/20th of the value of merchandise imports to Cape Verde in 2003. 4 Merchandise exports increased by 55% in 2005. This was mostly due to an increase in the exports of fresh fish, which increased by 200 percent as a result of the European Union (EU) decision to end its block on exports of fish from Cape Verde (World Bank, 2006a). 21 Figure 7: Merchandise Exports (2003) 100.0% 95.1% 2 90.0% 83.5% 1.8 79.5% 80.0% 74.6% 72.2% 1.6 70.0% 63.3% 1.4 60.9% 60.0% 1.2 50.0% 1 40.0% 35.7% 0.8 30.0% 26.0% 0.6 20.3% 20.0% 0.4 10.0% 5.1% 0.2 0.0% 0 Maldives Indonesia Jamaica South Africa Guyana Grenada Cape Verde Dominican Senegal Mauritius Philippines Share of Merchandise Exports in Total Exports of Goods & Services (%) Republic Export/Import Ratio of Merchandise Trade Source: World Bank (2006c) Cape Verde‘s exports are dominated by services. The thinness of the merchandise export base is apparent looking at the major exports and their destinations shown in Figure 8. Most manufacturing exports are apparel and shoes. Fishery products are also important—and increased dramatically in 2005.5 Portugal and the United States are the main export destinations for textiles and apparel. In fact, between 80 and 100 percent of each of those products (disaggregated to SITC 4-digit level) are exported to a single country (either Portugal or the United States). This high concentration of exports suggests that Cape Verde suffers from thin market channels and buyer- supplier networks of these exports. This is sharply different from Cape Verde‘s imports, where the set of source countries are more diversified. 5 Cape Verde's exclusive economic zone covers approximately 734,265 square kilometers of the Atlantic. The resource potential is estimated at 43,000 to 50,000 tons/year. However, the lack of adequate technology for deep-water fishing and the need to modernize the fleet leave approximately two thirds of the resource unexploited. Fish processing takes place in industrial units that use only part of their production capacity. Private investments, both national and foreign, have important roles to play in maximizing Cape Verde's industrial fisheries. The government's main priorities for the sector include modernization of the fleet, promotion of a strong and dynamic business community, development of infrastructure in fishing communities, and promotion of aquaculture projects and the processing industry. As noted earlier, the EU decision to end the block on fish exports from Cape Verde led to a large increase in fish exports in 2005 (World Bank, 2006a). 22 Figure 8: Major Products and Destinations TOTAL EXPORTS (2001-2003 Average) Major Destinations Total Portugal United States Mozambique Germany Ireland Mauritania Senegal Taiwan, Ch Spain $10,943,730 76.2% 17.8% 2.6% 1.3% 1.0% 0.9% 0.8% 0.6% 0.5% 100.0% Rank TOP 15 Export Products Major Destinations 1 6123: Parts of footwear Portugal 35.0% 100.0% 2 8432: Under garments,knitted of cotton Portugal 15.1% 100.0% ` 3 8432: Suits & costumes,women's,of textile fabrics U.S. Portugal 11.2% 98.3% 1.7% 4 8441: Shirts,men's,of textile fabrics Portugal U.S. Germany Japan U.K. 7.9% 59.7% 24.8% 13.3% 1.5% 0.7% 5 8423: Trousers,breeches etc.of textile fabrics Portugal Ireland 7.7% 91.9% 8.0% 6 2926: Bulbs,tubers & rhizomes of flowering or of foliage Portugal 6.7% 100.0% 7 8439: Other outer garments of textile fabrics Portugal U.S. Ireland Angola 4.9% 80.7% 11.5% 6.1% 1.6% 8 0360: Crustaceans and molluscs,fresh,chilled,frozen etc. Portugal Mauritania Senegal Taiwan, Ch 2.2% 30.5% 29.1% 21.3% 14.1% 9 8431: Coats and jackets of textile fabrics U.S. Portugal 2.0% 95.3% 4.5% 10 6114: Leather of other bovine cattle and equine leather Portugal 1.5% 100.0% 11 5417: Medicaments(including veterinary medicaments) Mozambique Portugal Angola 1.5% 62.6% 31.5% 5.7% 12 7712: Other electric power machinery,parts of 771-- Portugal 1.4% 100.0% 13 8510: Footwear Portugal 1.4% 100.0% 14 0371: Fish,prepared or preserved,n.e.s. including caviar U.S. Mauritania Netherlands Guinea 1.1% 51.5% 33.0% 7.4% 7.3% 15 8452: Dresses,skirts,suits etc,knitted or crocheted U.S. Portugal 1.0% 99.1% 0.7% * The percentage figures under the products represent the shares in the total Cape Verdean export value. (Only above 0.5% reported) * The percentage figures under the country names represent the share in the total export values of respective products. (Only above 0.5% reported) Source: UN COMTRADE/ World Bank WITS This structure is not unique to Cape Verde; it is found in many small island countries, where tourism and international transportation often dominate economic activity. Commercial service exports account for a very high share of Cape Verde‘s exports, with tourism a leading, and growing, source of foreign exchanges. Tourist arrivals have increased in recent years, with an average annual growth rate of 25 percent in the tourist market over the past four years. The low level of export diversification makes Cape Verde vulnerable to shocks due to a low level of export diversification. To reduce this vulnerability, recent trade policy reform has focused on improving the quality of exported products in order to support export diversification to developed countries. To do this, the Government has been attempting to take advantage of EU trade initiatives and the US African Growth and Opportunity Act (AGOA) to improve export quality. Cape Verde was declared eligible for AGOA on October 2, 2000 and eligible for apparel provision on August 28, 2002. Under the special rule for apparel, Cape Verde is eligible to use fabric and yarn from outside of the United States and Sub-Saharan Africa in apparel wholly assembled in Cape Verde until this provision expires in September 2007. In 2006, 37 countries were eligible for AGOA benefits, with 21 eligible under the special rule for apparel. Cape Verde's request for accession to the World Trade Organization (WTO) was circulated on 11 November 1999 and the General Council established a Working Party on 17 July 2000. A Memorandum on the Foreign 23 Trade Regime was submitted in July 2003 and the Working Party held its first meeting in March 2004.6 6 See http://www.wto.org/english/theWTO_e/acc_e/a1_capvert_e.htm. 24 CHAPTER 2: AN ANALYSIS OF FIRM PERFORMANCE I. THE INVESTMENT CLIMATE SURVEY This report is based upon the results of two surveys carried out in Cape Verde in March and April of 2006. Both surveys were conducted in two locations within Cape Verde, Praia and Mindelo. The first survey, the Investment Climate Survey (ICS), surveyed formal enterprises in manufacturing, retail trade, and other services. Firms with over five employees were randomly selected from lists provided by the National Institute of Statistics (INE) located in these areas and in the relevant sectors of the economy. The second survey, the Microenterprise Investment Climate Survey (MICS), surveyed microenterprises in retail trade, other services, light manufacturing, and construction. Rather than sampling firms based upon lists provided by government agencies, enterprises were selected using a judgmental route approach. Enumerators visited areas of the main cities identified by local experts as important locations for small and microenterprises. Following prescribed routes, they would count the number of enterprises. They would then sample every nth enterprise along the route to get a random sample over the entire area. In this way, it is possible to collect a random sample of microenterprises, including informal and unregistered enterprises. Because the two surveys use different surveying methodologies, we do not attempt to pool the two samples in the analysis. One important reason for not pooling is that we have no way to compute weights that would allow us to replicate the entire economy. I.1 Comparator Countries. One of the advantages that ICSs have over other firm-level surveys is that because similar surveys have been conducted in a wide range of other countries, it is possible to benchmark Cape Verde against other countries with respect to both firm productivity and measures of the investment climate. Cape Verde provides a relatively challenging country with respect to benchmarking. In particular, it is difficult to find relevant comparators within Africa—most African countries where Investment Climate Surveys have been completed are both considerably poorer than Cape Verde and considerably larger. For this reason, we benchmark Cape Verde against lower middle income countries in other regions, including several in the Caribbean. Although none of these are ideal comparators, they allow us to compare Cape Verde‘s investment climate against a range of economies that typically share some, but not all, of Cape Verde‘s unique characteristics. Pote ntial comparator countries include: 1. Africa: Senegal and South Africa 2. Indian Ocean: Mauritius and the Maldives 3. Caribbean: Dominican Republic and Guyana 4. Lower Middle-income Asia: Indonesia and the Philippines 25 These countries have been tentatively selected for the following reasons. Senegal is selected because of its close geographic proximity to Cape Verde. Senegal, however, is considerably poorer and larger than Cape Verde (see Table 3). South Africa is selected because it is the only other middle-income country in Africa where an Investment Climate Assessment has been completed. Like Senegal, South Africa is considerably larger than Cape Verde. In contrast to Senegal, it is considerably richer than Cape Verde, with per capita income almost twice the size in international dollars In addition to the two African countries, we also use two island economies located in the Indian Ocean. Like Cape Verde, the Maldives and Mauritius are relatively small—Mauritius is about twice the size of Cape Verde while the Maldives are about half the size. Both are considerably richer than Cape Verde, however, with per capita GDP about twice as large. 7 We also use two middle income countries in the Caribbean as a comparator country—one island and one additional country on the edge of the Caribbean. These are the only countries in the Caribbean region where data from Investment Climate Surveys are available. These economies share several notable characteristics with Cape Verde. One important common characteristic is that the Caribbean islands are also heavily reliant upon remittances and migration. Remittances are equal to 11 percent of GDP in the Dominican Republic and 8 percent of GDP in Guyana. Guyana, although not an island, is located on the edge of the Caribbean. It is close in size to Cape Verde and also close with respect to per capita GDP. The Dominican Republic is significantly larger—and slightly richer—than Cape Verde. We also select two lower middle-income countries in East Asia, Indonesia and the Philippines as the final comparator countries. These countries are primarily selected because they have similar per capita income to Cape Verde. Although both are island chains, they are considerably larger than Cape Verde and, consequently, do not have the same issues related to limited size (in terms of either population or geographic size) or physical isolation. 7 Although per capita GDP is not available in PPP adjusted terms for the Maldives, in current US$, per capita GDP is about twice as large ($2693 compared to $1328). 26 Table 3: Potential Comparator Countries Population Per capita GDP Remittances (millions) (PPP, int‘l $) (% of GDP) Cape Verde 0.5 $5,715 10.2% Senegal 10.5 $1,745 6.8% South Africa 45.6 $11,190 --- Mauritius 1.2 $11,944 --- Maldives 0.3 --- --- Dominican Republic 8.9 $7,326 11.0% Guyana 0.8 $4,294 8.0% Indonesia 217.6 $3,584 0.6% Philippines 83 $4,558 0.3% I.2 Firm Performance This section examines enterprise performance in Cape Verde. We first look at labor productivity, capital stock and capital productivity in the manufacturing sector and compare labor productivity with labor costs, to obtain the unit labor costs. These measures are compared across different firms within Cape Verde, and also to others internationally. This is followed by an analysis of total factor productivity. To ensure that the results are comparable across countries, and because the standard methodology is only appropriate for the manufacturing sector—results in this section only cover firms in that sector. 27 II. LABOR PRODUCTIVITY Labor productivity is measured by calculating the ratio of value added—intermediate inputs and energy costs subtracted from sales—to the number of workers. Manufacturing enterprises in Cape Verde are more productive than manufacturing enterprises in the other lower middle income comparator countries (see Figure 1). The median firm in Cape Verde produces about $6,100 of output per worker. This is substantially higher than for firms in Indonesia, the Philippines and the Dominican Republic (about $2,200, $2,400 and $2,300 per worker respectively) and is slightly higher than in Guyana and Senegal (about $5,500 and $5,600 per worker). In contrast, firms in Mauritius, the Maldives, and South Africa produce more output per worker ($7,000, $6,800, and $14,000 respectively) than firms in Cape Verde. The three countries with higher amounts of value- added per worker, however, have considerably higher per capita income. As a result, it is not surprising that labor productivity is also higher. Figure 9: Manufacturing firms are more productive in Cape Verde than in other middle income countries — although they lag behind firms in the most productive upper middle income countries. $20,000 $15,000 US Dollars $10,000 $5,000 $0 e l s a na s a lic s ga rd ve ne si iu ric ub a ne ne rit Ve di pi Af uy ep au al do ilip Se G e h M R M ap ut In Ph an So C ic in om D Value-Added per Worker Value-Added per worker in Garments Source: Investment Climate Surveys Note: All values are medians for enterprises with available data. Value added is calculated by subtracting intermediate inputs and energy costs from sales from manufacturing. Workers include both permanent and temporary workers. Values are converted to US$ using average exchange rates from World Development Indicators. Data were collected between 2002 and 2005 depending on survey period for each country. One problem with labor productivity is that labor productivity will generally be lower in sectors than are more labor intensive (i.e., sectors that use less capital per worker). If industry in Cape Verde was concentrated in sectors that were particularly labor intensive then labor productivity will be artificially low. One way of dealing with this, which we do in later sections, is to calculate total factor productivity—a measure of productivity that takes into account the firms‘ use of capital. Another approach is to focus more intensively on a single sector. In general, capital intensity will vary less within a single sector than it will between sectors. Controlling for this, therefore reduces concern about labor intensity. In this section, we focus on the garments sector — 28 an internationally traded good with a relatively well established production technology. When we focus on this sector alone, garment firms in Cape Verde remain more productive than firms in Indonesia, the Philippines and the Dominican Republic. Firms in this single sector also appear to be more productive in Cape Verde than in either Senegal or Guyana—and the difference between Cape Verde and these countries widens. Finally, although labor productivity is higher in Mauritius than in Cape Verde overall, it is lower in the garments sector. Garment firms in South Africa, however, are more productive than firms in Cape Verde. An important proviso with respect to these results is that productivity measures might appear artificially higher in countries with concentrated markets for firms that sell in domestic markets. If firms are able to exercise market power to raise prices, the measured value of their output will be higher than their actual value. As a result, labor productivity will appear higher than it would under perfect competition. Since, as discussed below, very few firms in Cape Verde compete on international markets, these calculations might overstate productivity in Cape Verde. There are significant differences between firms of different types within Cape Verde (see Table 4).8 Firms in the furniture and wood sector produce less output per worker than in other sectors. There are several plausible reasons for this. One is that firms in this sector appear to be less capital intensive than in other sectors and also appears to have lower capacity utilization (i.e., they believe that they could produce relatively more given their current workforce and capital equipment than they actually do). There is also a strong correlation between firm size and labor productivity—smaller firms produce considerably less than large firms. The median very small firm produces only about $3,000 of output per worker, compared to $27,000 for the median medium-sized firm. It is important to note, however, that there are relatively few firms with more than 50 workers, resulting in productivity being measured less precisely. There is also a strong correlation between enterprise size and capital intensity (capital per worker). Medium-sized firms appear to use more capital per worker than small or very small firms. As in many countries, foreign-owned enterprises are more productive than domestic firms. Whereas the median foreign-owned firm produces about $12,000 of value-added per worker, the median domestic firm produces only about $3,500 per worker. Potentially explaining this, foreign- owned firms appear to use more capital per worker than domestic firms. Finally, there is also a strong correlation between measures of human capital and labor productivity. Firms with formal training programs for their workers are close to three times more productive than firms without training programs and firms with university educated managers are about four times more productive than firms with managers with only secondary education. Although this suggests that these firms might be more productive due to their higher levels of human capital, they are also more capital intensive. The differences in productivity might therefore be the result of differences in human capital, but could also be due to differences in physical capital. 8 In most Investment Climate Assessments, we also include additional breakdowns. For example, we often include breakdowns for large and very large firms, and exporters and non-exporters. In Cape Verde, there were too few firms in these categories to provide breakdowns for them. 29 In the later section on total factor productivity, we are able to control for differences in physical capital intensity, making it possible to distinguish between the two hypotheses. Table 4: Median productivity by industry, size, and ownership in 2002 Value- Labor cost Unit Capital per Capital Capacity added per per Labor worker productivity utilization worker worker costs All $6,137 $2,797 0.49 $3,822 1.62 62 Sector Garments $6,564 $2,115 0.43 $3,524 1.81 63 Food and beverage $7,056 $4,274 0.36 $5,641 2.83 71 Furniture and wood $2,743 $1,535 0.49 $1,692 1.28 47 Size Very Small (5-9) $2,817 $1,624 0.49 $2,932 1.33 67 Small (10-49) $4,924 $2,288 0.60 $3,524 0.80 57 Medium (Over 50) $27,351 $6,611 0.35 $5,641 1.81 83 Ownership Domestic $3,587 $2,115 0.51 $2,932 1.33 62 Foreign $12,148 $4,338 0.36 $5,577 1.81 58 Training Formal training $9,325 $3,325 0.36 $4,338 1.33 60 No formal training $3,462 $2,035 0.61 $1,829 1.63 61 Manager Education University $13,839 $6,174 0.37 $5,609 2.43 64 Vocational $3,462 $2,065 0.67 $2,964 1.29 74 Secondary or less $2,807 $1,762 0.49 $2,030 1.33 52 Microenterprises All $1,663 $1,195 0.30 $1,128 1.19 Registered $2,175 $1,331 0.24 $2,707 1.20 Unregistered $1,556 $831 0.37 $395 1.09 Source: Investment Climate Survey. Note: See figures for detailed notes. Workers are permanent and temporary workers. Capital is the value of machinery and equipment if bought in current condition. All numbers are medians, except for capacity utilization which is a mean. All values are for manufacturing firms only. In addition to calculating labor productivity for the larger firms in the main survey, it is also possible to calculate productivity for the smaller firms in the microenterprise survey. On average, the microenterprises are less productive than the larger enterprises in the main survey. The median microenterprise in the light manufacturing sector produces about $1,663 of value-added per worker—lower than even the median very small firm (i.e., with between 5 and 10 workers). Among microenterprises, formal registered firms are considerably more productive than informal, unregistered firms ($2175 per worker compared to $1556 for unregistered firms). Although this could be because registration helps firms to improve their productivity in the medium-term—for example, if it helps them gain access to government support services or finance—causation might run in the opposite direction. That is, microenterprises that are more productive might either be more visible to the authorities or might be more likely to be in formal supply chains (e.g., to the service sector or government agencies). Thus, this does not immediately imply that registration improves productivity. Although microenterprise surveys have not been completed in many of the comparator countries—only Indonesia, South Africa and Senegal, it is possible to compare productivity with other middle income countries and other countries in Africa (see Figure 10). Overall, 30 microenterprises appear more productive in relative terms in Cape Verde than in other countries where microenterprise surveys have been completed. For example, although the average large firm in South Africa was considerably more productive than the average large firm in Cape Verde, microenterprises in these countries appear similarly productive. Also, although large firms in Senegal are only about 50 percent less productive than large firms in Cape Verde, microenterprises in Senegal are only one-third as productive. Figure 10 Labor productivity is higher for microenterprises in the manufacturing sector in Cape Verde than for similar firms elsewhere in Africa and other middle-income economies. value-added per worker (US$) $2,000 $1,500 $1,000 $500 $0 sh a e a l a a a il ga az si rd ny ni al ric de ne ne a m Ve Br Ke Af nz la te do Se h ng e Ta ua ut In ap Ba G So C Source: Microenterprise Investment Climate Surveys Note: All values are medians for manufacturing enterprises with available data. Value added is calculated by subtracting intermediate inputs and energy costs from sales from manufacturing. Workers include both permanent and temporary workers. Values are converted to US$ using average exchange rates for the relevant years from World Development Indicators. Data is for 2002, except for Cape Verde (2005) and South Africa (2004). III. LABOR COSTS Although labor productivity is relatively high in Cape Verde, labor costs are also quite high. Labor costs are equal to about $2,800 per worker for the median firm in Cape Verde (see Figure 11. This is considerably higher than for the median firms in the Dominican Republic ($1,400), Guyana ($2,000, Indonesia ($900), the Philippines ($1,200), and Senegal ($1800). As with labor productivity, labor costs were lower for the median firm in Cape Verde than they were for the median firm in any of the upper middle income economies. 31 Figure 11: Labor costs are high in Cape Verde —both in dollar terms and as percent of value added. Maldives Dominican Republic Guyana Guyana Dominican Republic Maldives Philippines Philippines Indonesia Indonesia South Africa South Africa Mauritius Mauritius Senegal Senegal Cape Verde Cape Verde $0 $2,500 $5,000 $7,500 0% 20% 40% 60% labor cost per worker (US$) labor cost per worker (% of value added) Note: All values are medians for enterprises with available data. Labor cost is the total cost of wages, salaries, allowances, bonuses and other benefits for both production and non-production workers. Value added is calculated by subtracting intermediate inputs and energy costs from sales from manufacturing. Workers include both permanent and temporary workers. Values are converted to US$ using average exchange rates from World Development Indicators. Data were collected between 2002 and 2005 depending on survey period for each country. Firms with high labor costs can remain competitive if their labor productivity is high enough. Although both labor costs and labor productivity are high, labor costs are relatively higher than labor productivity. Unit labor costs (the ratio of wages and salaries) were equal to 49 percent of value added in Cape Verde. This is higher than any of the comparator countries. For example, unit labor costs in South Africa—a country well-known for its high labor costs—were equal to only 45 percent of value-added. In the country with the lowest relative costs, the Maldives, unit labor costs were equal to only 21 percent of value-added. This suggests that manufacturing firms in Cape Verde might find it difficult to compete in international markets even though they are relative productive. Labor costs follow a similar pattern to value-added per worker. Larger firms, foreign-owned firms, and firms in the food and beverage sector have higher labor costs per worker than other firms in Cape Verde (see Table 4). In general, differences in labor costs per worker are smaller than differences in value-added per worker. For example, although labor costs are about twice as high in foreign-owned firms as domestic firms, labor productivity is about 3 times greater. As a result, unit labor costs are generally lower in firms with higher labor productivity (and labor costs). Unit labor costs are lower for medium sized firms, foreign-owned firms, firms with better formal training programs and better educated managers. 32 IV. CAPITAL PRODUCTIVITY Enterprises in Cape Verde use a relatively large amount of capital per worker compared to other lower middle income countries. When capital is valued at book value, the median enterprise in Cape Verde has about $2,500 of capital for every worker.9 This is the third highest among the comparator countries. The median firm in South Africa has about $3,500 of capital per worker and the median firm in Guyana has over $5,000 of capital per worker. Median firms in the Dominican Republic, the Maldives and Mauritius have slightly less capital per worker than in Cape Verde, while firms in the Philippines and Indonesia have considerably less capital. Figure 12: Enterprises in Cape Verde are relatively capital intensive—and capital productivity is consequently relatively low. Guyana Maldives Dominican Republic Dominican Republic Maldives Guyana Philippines Philippines Indonesia Indonesia South Africa South Africa Mauritius Mauritius Cape Verde Cape Verde $0 $2,500 $5,000 $7,500 0% 200% 400% 600% capital per worker (US$) capital productivity (capital over value added) Note: All values are medians for enterprises with available data. Capital is the net book value of machinery and equipment. Value added is calculated by subtracting intermediate inputs and energy costs from sales from manufacturing. Workers include both permanent and temporary workers. Values are converted to US$ using average exchange rates from World Development Indicators. Data were collected between 2002 and 2005 depending on survey period for each country. The productivity measure that is analogous to (the inverse of) unit labor costs is capital productivity—the ratio of value added to the net book value of machinery and equipment. Firms that produce a lot of output with only a small amount of capital will have higher capital productivity than firms that produce less output with more capital. In general, labor intensive firms (i.e., that rely relatively heavily on labor to produce their output) will have higher levels of capital productivity. 9 For the cross-country comparisons, the book value of capital is used since the preferred measure —the replacement cost—is not available for most of the comparator countries. The replacement cost is preferred for within-country comparisons as the data series is more complete and reliable for this measure and comes closer to the idea of the value of capital than the accounting measure, book value. 33 Given the high capital intensity of the median firm in Cape Verde, it is not surprising that capital productivity is relatively low for the median firm. Although capital productivity is lower for the median firms in Guyana—which is an outlier in terms of capital per worker—and the Dominican Republic than it is for the median firm in Cape Verde, firms in Cape Verde lag behind the other countries in this respect. For example, firms in Mauritius and South Africa produce over 50 percent more and over twice as much as firms in Cape Verde do. In addition to asking firms about labor and capital, firms are also asked about capacity utilization—how much they did produce last year relative to the maximum amount they could have produced if they had been using their current workforce and capital to the fullest extent possible. If capacity utilization is 50 percent, this suggests that the manager believes that the firms could have produced twice as much with its current amount of labor and capital. Firms in Cape Verde report low levels of capacity utilization. The average firm reported that capacity utilization was about 62 percent—lower than in any of the comparator countries, where average utilization rates varied between 68 percent (the Philippines) and 78 percent (South Africa). Figure 13. Capacity utilization is low in Cape Verde 100 Capacity utilization 80 60 40 20 0 a s s l s e a a lic ga ve ne iu an rd si ric ub rit ne ne Ve di pi Af uy ep au al ilip do Se G h e M R M ut ap In Ph an So C ic in om D Source: Investment Climate Surveys. Note: All values are means for enterprises with available data. Capacity utilization is directly reported by enterprise managers and is defined as the amount of output that is actually produced relative to the maximum amount that could be produced given current capital stock and employment. Data varies between 2002-2005, depending on survey period for each country. High levels of capital utilization suggest that the firm is using its resources more efficiently. This suggests that productivity could be even higher in Cape Verde if they were to use their resources more efficiently. Why did Cape Verde have relatively low levels of capacity utilization in 2006. One possibility is that firms in Cape Verde are using their resources inefficiently on a consistent basis. Another might be that there was some temporary shock that reduced capacity utilization for a relatively short period (e.g., a recession). Given that the economy was growing relatively rapidly in 2005, this seems less likely in this case. Temporary shocks, such as the expiration of the multi-fiber agreement, might explain low capacity utilization in the garments 34 sector, but does not appear to explain the very low productivity in the furniture and wood sector. A final alternative explanation, which is discussed below, is that firms might use excess capacity to deter entry. If this were the case, we would expect to see a negative correlation between market share and capacity utilization in Cape Verde. V. COMPETITION Cape Verde‘s small population spread over nine inhabited islands—although with over half the population on one island—limits the potential size of the market for manufacturing firms in the country. One result of this, noted above, is that most firms are relatively small. The largest manufacturing firm in the sample had less than one hundred employees. In addition to restricting firms, the small domestic market is likely to limit the number of firms in the country and, consequently, restrict competition. Consistent with this, firms in Cape Verde reported having fewer competitors and greater market share than firms in most of the comparator countries (see Figure 14). Even in Guyana, a country relatively similar to Cape Verde in terms of population income and per capita income, firms tended to report that they had more competitors than firms in Cape Verde. Firms in the Maldives, with a population that is about half the size of Cape Verde were slightly less likely to say they had one or zero competitors than firms in Cape Verde (18 percent compared to 20 percent), but were also less likely to say they had more than five competitors (46 percent compared to 57 percent) Figure 14: Firms in Cape Verde face less competition than similar firms in the comparator countries Senegal Maldives South Africa Guyana Indonesia Cape Verde Cape Verde 0% 20% 40% 60% % of firms 0 20 40 60 Five or less competitors Share of Local market One or less competitors Source: Investment Climate Surveys Note: Data varies between 2002-2005, depending on survey period for each country. Market share is the average reported market share in local markets. High transportation costs might also exacerbate the low level of competition. Most of the firms in the sample report that they primarily compete on a sub-national level. Only about one-third 35 of firms said that they operated primarily in national, as opposed to sub-national, markets. And only one firm said that they were primarily competing on international markets. The low level of competition might partially explain the low level of capacity utilization in the country. Firms with relatively high market shares often have lower levels of capacity utilization than firms in more competitive industries because they can use excess capacity to discourage entry by potential competitors. The correlation between market share and capacity utilization appears to hold in Cape Verde—the correlation between market share and capacity utilization is -0.11. VI. IMPORTS Manufacturing firms in Cape Verde rely heavily upon imported inputs in their production processes. About 62 percent of material inputs and supplies are either directly or indirectly imported. This is higher than most of the comparator countries—the average South African firm imports only about 23 percent of its inputs and the average firm in the Philippines imports only about 35 percent of its imports. Figure 15. Firms in Cape Verde are heavily dependent upon imported inputs 80 imports as % of sales 60 40 20 0 a s s l s e ca lic ga ve ne iu an rd ub ri rit ne Ve di pi Af uy ep au al ilip Se G h e M R M ut ap Ph an So C ic in om D Source: Investment Climate Surveys. Note: All values are means for enterprises with available data. Data varies between 2002-2005, depending on survey period for each country. The heavy reliance on imported inputs is a result of the small, service oriented domestic market—many manufactured inputs are not available locally. Two of the other small island economies—Mauritius and the Maldives—also import most of their supplies and inputs from abroad. Interestingly, this is not the case for Guyana or the Dominican Republic. In both countries, the average firm imports less than 30 percent of its inputs. Although the Dominican Republic is considerably larger than Cape Verde, Guyana is similarly sized. Given the poor relative performance of Guyana‘s ports (see Chapter 6) and its physical isolation from other countries, a heavy reliance on imported inputs is likely to contribute to high production costs for firms in Cape Verde. This, in turn, will make it more difficult for firms to get 36 involved in export markets and might lead us to underestimate value added per worker if inputs are more expensive in Cape Verde than they are in other countries. VII. TOTAL FACTOR PRODUCTIVITY (TFP) Although the measures of firm productivity discussed above provide useful information on firm performance, they can be misleading when considered in isolation. For example, although labor productivity is relatively high in Cape Verde, capital productivity is relatively low. To get an overall assessment of productivity, it is necessary to take both capital and labor use into account. The net impact of the different factors of production on firm-level productivity can be assessed by calculating total factor productivity. Differences in total factor productivity are those differences in output that cannot be explained by differences in the use of labor, capital and other inputs. Firms with higher total factor productivity are more efficient than other firms because they produce more with fewer inputs. Table 6 presents results from estimating a Cobb-Douglas production function using data for enterprises from twelve different manufacturing sub-sectors. To allow us to compare total factor productivity between Cape Verde and the middle-income comparator countries (Guyana, the Dominican Republic, Indonesia, Mauritius, the Maldives, South Africa, and the Philippines), we pool the observations for all of the countries into a single regression.10 In addition to allowing us to compare total factor productivity in Cape Verde with productivity in the comparator countries, this also gives us a sufficiently large sample of enterprises to compare total factor productivity across firms of different types. The production function is estimated using a stochastic frontier approach.11 As a robustness check, we also estimate the production function using a Least Absolute Deviations (LAD) estimator. This is used rather than Ordinary Least Squares (OLS), because it is more robust to outliers.12 The dependent variable is the natural log of sales and all regressions control for the enterprises‘ use of capital, intermediate inputs and raw materials, and workers. All models include country dummies to pick up differences in total factor productivity between the different countries and also to reduce problems associated with exchange rates. Sector Specific Production Technologies. In addition to the sector dummies, different sectors are also allowed to use different production technologies—that is, labor and capital intensities are not assumed to the same across sectors. Mechanically, this is done by multiplying 10 The low-income comparator countries, Kenya and Senegal, are omitted due to missing data and concerns about pooling data for firms from low and middle-income countries that might use different technologies. 11 See Kumbhakar and Lovell (2000) for a description of stochastic frontier models. The model estimated assumes that technical inefficiency component is distributed with an exponential distribution, while the white noise component is distributed with a normal distribution. Firms that are outliers for value added per worker and capita per workers (i.e., enterprises in the top and bottom 2.5 percent of firms) are dropped in this part of the analysis so that results will not be driven by outliers and to ensure convergence. 12 Due to concerns about outliers, LAD estimators are often used when estimating production functions. See, for example, Greene (2000, pp. 449-450). 37 the sector dummies by capital, intermediate inputs and labor.13 A joint test of the significance of the interaction terms rejects the null hypothesis that the coefficients are equal across sectors at a one percent level or higher for all regressions in Table 6. This suggests that it is inappropriate to pool enterprises from different sectors into a single model without controlling for sector differences in factor intensities between sectors. Differences between Countries. After controlling for differences in productivity due to differences in the use of different factors of production, and sector of operations, (i.e., using results from column 1 of Table 6), enterprises in Cape Verde appear to be more productive than similar enterprises in most of the lower middle-income comparator countries. On average, firms in the Philippines were about 20 percent less productive than firms in Cape Verde, firms in the Dominican Republic were about 15 percent less productive, and firms in Indonesia were about 6 percent less productive (see Figure 16). Firms in the other lower middle income country—Guyana—were about 10 percent less productive than firms in Cape Verde. The differences in productivity were not, however, statistically significant except for the Philippines. Figure 16 Total factor productivity is higher in Cape Verde than in the Philippines, Indonesia or the Dominican Republic, but lower than in the upper middle income countries in the sample. (as % of TFP in Cape Verde) 200% 150% 100% 50% 0% Guyana South Indonesia Mauritius Africa Dominican Maldives Philippines Republic Total Food Furniture Source: Investment Climate Surveys Note: TFP is calculated based upon the coefficient estimates for the country dummies from the regression in column 1 of Table 6 and a similar regression only including firms in the furniture/wood and food and beverage sectors. Formula for effect of a dummy variable is from Kennedy (2003, p. 123). See Halvorsen and Palmquist (1980) for the derivation. In contrast, firms in the upper middle income comparator countries were significantly more productive than firms in Cape Verde. Firms in South Africa were about 30 percent more productive, firms in Mauritius about 40 percent more productive, and firms in the Maldives about 58 percent more productive. These differences are all statistically significant. 13 Formally, the coefficients on capital, labor, and intermediate inputs are allowed to vary across sectors. This system is estimated by interacting sector dummies with these variables. The model is then estimated after adding the sector dummies and interaction terms to regression. These coefficients are not reported in the table due to space constraints. 38 Results are similar when we repeat the TFP regressions only including firms in the food and beverage and wood and furniture sectors. This makes fewer functional form assumptions than are required to pool firms in difference sectors of the economy. We chose these sectors because they account for most of the firms in the sample for Cape Verde. The main difference are that firms in the food and beverage sector are less productive in South Africa than they are in Cape Verde, while firms in the wood and furniture sector are more productive in Indonesia than they are in Cape Verde. It is important to keep an important caveat in mind when comparing total factor productivity across countries. Although including country dummies is useful because doing so means that the coefficients on other variables will not be affected by assumptions about exchange rates (i.e., the dummies will control for exchange rates in cross-country regressions where monetary variables are in logs), exchange rates can make the coefficients on the country dummies difficult to interpret. That is, the coefficients on the country dummies will depend upon the exchange rate as well as on productivity differences. If a country‘s exchange rate is overvalued relative to its long-run equilibrium value, the coefficient on that country‘s dummy will appear artificially large (as will value added per worker). Economies of Scale. If large enterprises were consistently more productive than small enterprises, the sum of the coefficients on labor and capital would be greater than one. If this were the case, total production would more than double if the number of workers and amount of capital were doubled. In practice, the sum of the two coefficients is generally close to one for most of the sectors. For the base sector, wood and furniture, the coefficients sum to 1.01 in both the stochastic frontier and least absolute deviation models (columns 1 and 3 respectively). Table 5 presents sums of the coefficients for each sector and the test of the null hypothesis that they sum to one (model from column 1). These results suggest in most sectors—with the possible exceptions of garments and chemicals where large enterprises appear to be more efficient and electronics—large enterprises are neither more nor less productive than small enterprises. At least for the firms in this study—formal manufacturing firms with over 5 employees—economies of scale do not appear to be very important Table 5: Test for Constant Returns to Scale Sum of P-value for test of Sector Coefficients CRS Furniture 1.01 0.76 Food and Beverages 1.03 0.17 Garments 1.08 0.00 Chemicals 1.17 0.00 Machinery 1.44 0.27 In the context of this study, this suggests that the small size of firms in Cape Verde is unlikely to explain the difference in productivity between firms in Cape Verde and the larger firms in most of the comparator countries. In the two main sectors that firms in Cape Verde operate in — furniture and food and beverages—the sum of the coefficients is very close to one suggesting only very modest economies of scale. 39 Other Enterprise Characteristics. In addition to capital, intermediate inputs and labor, a series of sector dummies, the interaction terms, and the country dummies, the regressions also include a series of variables to control for other enterprise characteristics. These variables include a dummy variable indicating that the enterprise exports, a dummy variable indicating that it is foreign owned, a dummy variable indicating that the firm uses e-mail to communicate with clients and suppliers, a dummy variables indicating that the manager has a university or technical education, and a dummy variable indicating that the enterprise has a formal training program. The average exporter is about 5 to 7 percent more productive than similar enterprises that do not export.14 The coefficient is statistically significant in LAD model, but not statistically significant in the stochastic frontier model. This result is consistent with other studies that have found similar results for different sets of countries.15 Economists have suggested two possible explanations for the correlation. One is that exporting might result in productivity improvements for the firms that are doing it (the ‗learning-by-exporting’ hypothesis). The discipline of competing in international markets might encourage enterprises to improve their productivity or might expose them to foreign technologies or modes of production. The other explanation is that since firms have to be efficient to compete on international market, only firms that are already efficient are able to export (the ‘self-selectivity’ hypothesis). Although inefficient firms might be protected from international competition in domestic markets by natural barriers (e.g., high transportation costs) and policy barriers to trade (e.g., government tariffs and quotas or inefficient ports or customs administration), they are unable to enter international markets. It is important to note that the two hypotheses are not mutually exclusive. Even if efficient firms are more likely to start exporting, this does not rule out the possibility that exporting will help them increase their productivity further.16 Due to the cross-sectional nature of the data included in this study, we are unable to distinguish between the two competing hypotheses in this analysis. The results also suggest that foreign-owned enterprises are more productive than domestically owned enterprises. The point estimates of the parameters in columns 2 and 4 suggest that fully foreign-owned enterprises are about 9 to 11 percent more productive than fully domestically-owned enterprises. The coefficients on the dummy variable are statistically significant at a 10 percent level or higher in both regressions. This is also consistent with previous results that have often found that foreign-owned enterprises are more productive than domestic enterprises in developing countries. (Saggi, 2002) 14 The median exporter in the sample included in the regression in column 1 exports about 25 percent of its output. 15 The large literature on this topic is summarized in Tybout (2003) and Keller (2003). 16 The evidence appears to support both hypothesis to some degree. For example, several econometric studies that have looked at whether enterprises improve their productivity before or after they start exporting have found that productivity improvements precede exporting, providing support for the self-selectivity hypothesis. See for example, Clerides et al. (1998), Bernard and Jensen (1999), Liu et al. (1999) and Aw et al. (2000). However, case studies often support the ‗learning by exporting‘ hypothesis. Studies of exporters in Korea and Taiwan found that export buyers were an important source for new technologies, which they provided in various forms including complete blueprints, information about manufacturing processes and quality control methods, technical advice and on-site plant inspections, and training for technical and production staff (Westphal, 2002) 40 Firms that use the internet to communicate with other enterprises appear more productive than other firms.17 Rather than reflecting actual internet use, it seems likely that this reflects difference in the intensity of technology use rather than internet use. Enterprises that use technology more intensively are between 15 and 23 percent more productive than similar firms that do not. Previous studies of firm performance in Africa have stressed the importance of formal training on firm performance.18 Similar results appear to hold in the sample of middle-income countries in this sample. Firms with formal training programs appear to be about 10 percent more productive than similar programs without training programs. In contrast to the previous results, this result does not appear to be highly robust Enterprises with better educated managers are more productive than other enterprises. An enterprise with a manager with a university education is about 1 to 5 percent more productive than a similar enterprise where the manager has only a primary or secondary education. In contrast to the previous results these results are not statistically significant. The coefficients are not significantly different from zero in either regression. 17 Quiang, Clarke, and Halewood (2006) find similar results for a larger sample of developing countries using Investment Climate Surveys 18 See, for example, Regional Program on Enterprise Development (2005a; 2005b) 41 Table 6: Determinants of Firm-Level Productivity for Manufacturing Enterprises Frontier Model Least Absolute Deviations Sales (natural log) Observations 1749 1579 1863 1676 Factors of production Intermediate Inputs 0.6189*** 0.6280*** 0.6823*** 0.6610*** (natural log) (13.92) (13.72) (27.20) (20.92) Workers 0.3195*** 0.2678*** 0.2389*** 0.2252*** (natural log) (5.33) (4.30) (7.23) (5.40) Capital - book value 0.0826** 0.0627 0.1063*** 0.0957*** (natural log) (2.12) (1.59) (4.46) (3.23) Countries Guyana 0.1189 0.1528 0.0244 0.0405 (dummy) (1.01) (1.32) (0.30) (0.40) Indonesia -0.0565 0.0497 -0.2043** -0.1795* (dummy) (0.48) (0.42) (2.56) (1.77) Maldives 0.5810*** 0.3497* 0.4536*** 0.2356 (dummy) (3.21) (1.91) (3.85) (1.55) Mauritius 0.4140*** 0.2525 0.2618*** 0.1827 (dummy) (3.21) (1.59) (3.01) (1.39) Philippines -0.2039* -0.1856 -0.1847** -0.1963** (dummy) (1.78) (1.62) (2.35) (1.97) South Africa 0.2988** 0.2013* 0.1931** 0.1474 (dummy) (2.51) (1.70) (2.39) (1.45) Dominican Republic -0.1570 -0.1478 -0.2892*** -0.2460** (dummy) (1.18) (1.13) (3.27) (2.23) Enterprise characteristics Foreign-owned 0.0959* 0.0868** (dummy) (1.85) (2.03) Has formal training program 0.0950** -0.0158 (dummy) (2.39) (0.47) Internet use 0.2322*** 0.1473*** (dummy) (5.02) (3.79) University educated manager 0.0568 0.0076 (dummy) (1.36) (0.22) Exporter 0.0488 0.0652* (dummy) (1.19) (1.89) *** Significant at 1 percent level ** Significant at 5 percent level * Significant at 10 percent level. t-statistics in parentheses. Dependent variable is log of value-added. Capital is book value of capital. a Coefficients are reported for firms in the furniture sector. In addition, sector dummies and sector specific production functions are included for 12 additional sectors (Textiles, Leather, Food, Beverages, Metals and Machinery, Electronics, Chemicals and Pharmaceuticals, Wood and Furniture, Non-metallic and Plastic Materials, Paper, Printing and Publishing, Other Manufacturing, Autos and Auto Parts). The dummies are interacted with capital and labor to allow sector-specific production technologies. Do these additional factors explain the difference in productivity between Cape Verde and the other countries in the study? A natural question is whether these differences explain any of the differences in productivity between countries in the regression. That is, after including these variables, do the coefficients on the country dummies become smaller or less statistically significant. In the stochastic frontier regressions, the coefficients become smaller in absolute terms for six of the seven comparator countries—all except Guyana. Moreover, statistically significant coefficients become statistically insignificant for two of the four countries (Mauritius and the Philippines). That is, after controlling for these differences, the differences in productivity between Mauritius and Cape Verde and the Philippines and Cape Verde become statistically insignificant. Results are similar for the LAD regressions. Coefficients fall in absolute value in four of seven 42 cases and become statistically insignificant in three (Mauritius, Maldives, and South Africa). This suggests that some of the productivity differences between Cape Verde and the other countries can be explained by differences in ownership, export behavior, Internet use, training, and manager education. Is Cape Verde Different? The previous results are from regressions that include data from all eight countries included in the empirical analysis. A natural question is whether the results appear to hold in Cape Verde as well as the other countries or whether Cape Verde is different from the other countries in the sample. Unfortunately, there are too few observations to estimate a separate regression for Cape Verde. The previous results strongly indicate that it is inappropriate to pool all firms in a single regression without including sector specific intercepts and allowing the coefficients on capital, labor and intermediate inputs to vary across sectors. Consequently, over 20 coefficients need to be estimated – with only about 41 enterprises from Cape Verde included in the final sample, it would not be feasible to estimate a separate model for Cape Verde. Therefore, we take a different approach, re-estimating the model allowing the coefficients on enterprise characteristics to take different values for enterprises from Mauritius and enterprises from the other countries. To do this, model 1 from Table 6 is re-estimated including interaction terms for the enterprise characteristics and the Cape Verde dummy. This allows foreign ownership, exporting, internet use, manager education, and training to affect firms in Cape Verde differently from firms in other countries. Although the point estimates of the coefficients on the variables were slightly different for Cape Verde than for the other countries, the differences were not large in most cases. The main exceptions were Internet use, which appears to have a lesser impact in Cape Verde than elsewhere, and manager education, which appears to have a significantly larger impact in Cape Verde (see Table 7). However, even for these two cases, the results might be due to the small sample size. In all cases, the null hypotheses that the coefficients are equal for Cape Verde and the other countries could not be rejected for any of the variables (either singly or jointly). Overall, these results suggest that enterprise characteristics that affect enterprise performance in the other countries have a similar effect in Cape Verde. Table 7: Test that coefficients are the same for Mauritius and other countries Coefficient for Coefficient on Coefficient for Significance all interaction Cape Verde level Foreign Ownership 0.10 -0.06 0.03 88% Training 0.09 0.00 0.10 99% Internet Use 0.24 -0.29 -0.05 22% Manager Education 0.05 0.27 0.32 33% Exporting 0.05 0.01 0.06 98% VIII. EXPORTS Very few of the manufacturing firms in Cape Verde are involved in exporting. Only 6 percent of manufacturing firms in Cape Verde export—fewer than in any of the comparator countries including the other island economies. Moreover, very few firms are involved in both exporting and domestic markets. Only one firm in the sample both exports and sells in domestic 43 markets. The low propensity to export is consistent with the macroeconomic evidence presented in Chapter 1—merchandise exports are relatively modest. Other countries, including the other island economies, export more. About 17 percent of firms in the Dominican Republic and 28 percent of firms in the Maldives export. About 8 percent of firms in the Maldives sell their output in both export and the domestic markets, while all of the exporters in the Dominican Republic sell some of their output in domestic markets. In Mauritius, a relatively small island economy that has managed to develop a large manufacturing sector, about two-thirds of manufacturing firms export with the average firms exporting close to 40 percent of its output. Although a large number of these are only involved in exporting (about 20 percent of firms), most sell some of their output domestically. Figure 17: Firms in Cape Verde are less likely to export than firms in the comparator countries Guyana Guyana Maldives Maldives Dominican Republic Dominican Republic Indonesia Indonesia Philippines Philippines Mauritius Mauritius South Africa Senegal Senegal South Africa Cape Verde Cape Verde 0% 20% 40% 60% 80% 0 10 20 30 40 % of firms exporting exports (% of sales) Source: Investment Climate Surveys Note: Data varies between 2002-2005, depending on survey period for each country. Why do so few firms export in Cape Verde export? The small size of the domestic market is likely to play a significant role in this. The large fixed costs associated with setting up an international distribution or service network will generally make exporting easier for large enterprises. Further, large enterprises generally have better access to finance than small enterprises—especially in developing countries—making it easier for them to finance these costs.19 Many previous studies have found that large firms are considerably more likely to export in developing countries than small firms are—something that remains true even after controlling for 19 See Schiantarelli (1996) for a review of the literature on firm size and financial constraints. 44 the potential for reverse causation (i.e., the fact that entering export markets allows firms to grow).20 Clerides and others (1998) find evidence consistent with this for Colombia, Mexico and Morocco. Similarly, Grenier and others (1999) found that large Tanzanian enterprises export more than smaller enterprises. Finally, using data from several countries in sub-Saharan Africa from the mid- 1990s and 2000s, Bigsten and others (2004), Söderbom and Teal (2003) and Clarke (2005) found similar results. Even when we restrict the sample to Micro and Small Enterprises (MSEs) with less than 50 workers, enterprises in Cape Verde appear less likely to export than other firms than similar enterprises in other countries. None of the MSEs in Cape Verde exported, while 7 percent exported in the Dominican Republic, 26 percent exported in Guyana, and 29 percent exported in Maldives. Moreover, about 40 percent of South African MSEs and close to 50 percent of Mauritian MSEs export. This suggests that differences in exporting are not simply the result of the small size of firms in Cape Verde. Another possibility might be that firms in Cape Verde mostly operate in manufacturing sectors that mostly are involved in domestic markets. It is more difficult to test whether this is the case than to see whether the low propensity to export is due to firm size. To do this, we estimate a regression model that estimates the probability that firms in Cape Verde export relative to firms in the comparator countries. The model is described in more detail in an appendix. Before controlling for any firm characteristics, including size and sector, firms in Cape Verde are considerably less likely to export than firms in any of the comparator countries (see Table 8, Column 1). The coefficients on the country dummies in the regression, which show how much more likely firms in that country are to export than firms in Cape Verde, are all positive and statistically significant. This suggests that before controlling for any firm characteristics, firms in Cape Verde appear to be between 27 percent (compared to firms in the Dominican Republic) and 63 percent (compared to firms in South Africa) less likely to export than firms in the comparator countries. To see if this is due to firms in Cape Verde operating in sectors that are generally unlikely to export, sector dummies are added to the regression (column 2). The coefficients on the country dummies can now be interpreted as the difference in probability that firms will export between Cape Verde and the comparator countries after controlling for sector of operations. This does not appear to have a large impact on the coefficients—they remain statistically significant and positive and appear similar in size to the coefficients before controlling for sector—four increase slightly and four decrease slightly. To see whether other firm characteristics explain why firms appear unlikely to export, additional control variables are added to the regression with sector dummies. These include number of workers (as a proxy for size), a dummy variable indicating that the company is foreign-owned, a dummy variable indicating whether the firm has a formal training program, a dummy variable indicating whether the firm has an Internet connection, and a dummy variable indicating that the manager of the enterprise has a university degree. 20 See Biggs (2003) for a summary of the literature on this topic. 45 Table 8: Probability that firm exports Firm exports (dummy) Observations 2680 2674 2335 2330 Sector Dummies No Yes Yes Yes Countries Guyana 0.4316*** 0.4617*** 0.5287*** 0.5059*** (dummy) (3.52) (3.90) (5.11) (4.61) Indonesia 0.5288*** 0.4866*** 0.2927** 0.2491** (dummy) (4.37) (3.92) (2.40) (1.97) Maldives 0.3810*** 0.4309*** 0.2935* 0.2928* (dummy) (2.77) (3.17) (1.95) (1.95) Mauritius 0.5811*** 0.5817*** 0.5304*** 0.5211*** (dummy) (6.05) (6.00) (4.71) (4.43) Philippines 0.5028*** 0.4408*** 0.3410*** 0.3597*** (dummy) (4.11) (3.48) (2.83) (2.95) South Africa 0.6307*** 0.6227*** 0.4257*** 0.4065*** (dummy) (5.74) (5.61) (3.64) (3.39) Dominican Republic 0.2724** 0.3126** 0.2744** 0.2627** (dummy) (2.01) (2.31) (2.13) (2.00) Senegal 0.4985*** 0.4775*** 0.4693*** 0.4161*** (dummy) (4.43) (4.10) (4.10) (3.26) Enterprise characteristics Workers 0.1510*** 0.1520*** (natural log) (13.74) (13.67) Foreign-owned 0.1861*** 0.1771*** (dummy) (5.08) (4.82) Has formal training program 0.0067 0.0090 (dummy) (0.23) (0.31) Uses Internet 0.1717*** 0.1733*** (dummy) (5.51) (5.54) University educated manager 0.0642** 0.0677** (dummy) (2.15) (2.28) Unit labor costs -0.3361** (sector-country averages) (2.14) The coefficient on number of workers is positive and statistically significant. This is consistent with previous studies that have found that large firms are more likely to export than small firms (see above). The coefficient on the dummy variable indicating that the company is foreign- owned is also statistically significant and positive—foreign owned firms are more likely to export than domestically owned firms. This is not surprising. Firms that are foreign-owned are more likely to have distribution networks setup, especially in their home countries. To the extent that their parent company has already borne the fixed costs associated with entering export markets, the firm should find it easier to use these previous efforts to reduce their own costs doing so. The regression also includes a dummy variable indicating whether the company has a formal training program and the education of the manager. These variables are included to test whether firms with greater human capital are more able to enter export markets. For example, if training improves quality and flexibility, it might allow domestic firms to more easily adjust their process to meet differing requirements. Similarly, firms with better educated managers might also find it easier to do the same. The coefficient on whether the firm has a formal training program is positive, but is small and is statistically insignificant. This suggests that there is only a modest impact of training on exporting. In contrast, the coefficient on the dummy variable indicating a university 46 educated manager is statistically significant and positive. Firms with university educated managers are 7 percent more likely to export than firms with less educated managers. Finally, the regression includes a variable indicating whether the firm uses the Internet to communicate with clients and suppliers. Previous studies using both firm-level data and aggregate trade level data have found a causal link between Internet use and export behavior (Clarke, 2001; Clarke and Wallsten, forthcoming; Freund and Weinhold, 2004) The evidence from this study is consistent with these earlier studies. Firms that use the Internet are 17 percent more likely to export than firms that do not. Together these factors appear to explain some of the difference between Cape Verde and the comparator countries. The coefficients on the dummy variables remain statistically significant, but are small in six of eight cases. For example, the additional variables explain about 44 percent of the difference in propensity to export between Indonesia and Cape Verde and 32 percent of the difference between the Philippines and Cape Verde. As noted earlier in this chapter, and discussed in greater detail in Chapter 3, although firms in Cape Verde are relatively productive, labor costs are even higher in relative terms. This might make it difficult for firms in Cape Verde to export. Although tariff and non-tariff barriers and the factors that make it difficult to export (e.g., physical isolation) allow firms to operate in domestic markets, high labor costs might make it difficult for them to export. To see if this is the case, we include unit labor costs in the regression. Because of concerns about endogeneity, the variable used in the analysis is the average unit labor cost for firms in that country and sector of the economy. The coefficient on unit labor costs is negative and statistically significant. This suggests that firms that have to cope with high unit labor costs are less likely to export than similar firms with lower costs. Even this, however, does not appear to explain the low propensity to export. Although it reduces the size of the coefficients on seven of the eight dummy variables, the coefficients remain statistically significant in all eight cases. Together this suggests that observable firms differences and differences in labor costs do not fully explain the low propensity to export for firms in Cape Verde. So what might explain the remaining difference between Cape Verde and the other comparator countries? Cape Verde‘s physical isolation might play a role. Since Cape Verde has no land borders with any other countries, this limits its potential exports. Previous studies have noted that manufacturing firms in Africa that export mostly do so to neighboring countries (Clarke, 2005), putting firms from Cape Verde at a relative disadvantage when compared to other countries. For example, although Cape Verde is a member of Economic Community of West African States (ECOWAS), its intra-regional trade with other ECOWAS members is significantly smaller than for other ECOWAS countries (Figure 4). Moreover, trade and market diversification are also hampered by the fact that there are few shipping channels that connect Cape Verde with the rest of the world. 47 Figure 18: Cape Verde has much less trade relations with other ECOWAS countries. 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% Benin Burkina Cote Guinea- Mali Niger Senegal Togo Cape Gambia Ghana Guinea Liberia Nigeria Sierra Faso d'Ivoire Bissau Verde Leone Export within ECOWAS Import within ECOWAS Source: IMF Direction of Trade Statistics Box 1: The Experience of Mauritius Mauritius was once dependent on limited number of agricultural exports. Today, Mauritius has a large manufacturing sector, especially in the textiles and clothing sectors. The successful manufacturing sector growth in Mauritius is often associated with its export-processing zone (EPZ) established in the early 1970s using Taiwan, China as a model. The fiscal incentives to investors, such as tax holidays or duty drawbacks, are common to other countries where EPZs were established. However, unlike other EPZs, over 50 percent of the investment came from local entrepreneurs. Instead of creating economic enclaves with plants set up by foreign multinational corporations with little local spillover effects, the EPZ in Mauritius has been the driver of local manufacturing sector growth with forward and backward linkages. The EPZ started showing positive results from the early 1980s. The number of EPZ firms increased form 9 in 1971 to over 500 in 2000. Likewise, employment in the EPZ has been impressive, increasing from 644 in 1971 to over 90,000 in 2000. First it gave sugar factory owners an alternative activity, which carried them through the agricultural crisis in the late 1970s. Second, it has built confidence and increased wages in other sectors of the economy. Third, it brought foreign investors, global business linkages, and new ideas into the Mauritian economy, allowing Mauritians to modernize and build investors, global business linkages, and new ideas, allowing Mauritians to modernize and build investor confidence in all sectors especially tourism. Although its location is a distinct disadvantage with respect to trade, this disadvantage should not be overemphasized. Although its location makes trade difficult, it does not make it impossible—Cape Verde manages to import a wide range of consumer goods from other countries. Moreover, other island economies, such as Mauritius, have managed to develop strong exports in labor-intensive manufacturing goods despite physical isolation from the continent (see Box). Firms from Cape Verde also have some advantages over firms in other African economies that still manage to export. In particular, they do not have to export their output through neighboring countries, as landlocked countries in Africa do, to reach international markets such as the United States as Europe. Finally, it is important to note that firms in Cape Verde have a lower propensity to export than firms in other small island economies such as the Maldives or Mauritius that suffer similar problems related to physical isolation. 48 Other factors, such as the performance of ports and customs, might also discourage trade. In the Investment Climate Survey, firms that directly export and import goods are asked about the time it takes to complete all port and customs procedures in the country. Because there are so few firms that export in Cape Verde, it is difficult to estimate the average number of days for exports to complete all procedures after arriving at the port or point of exit. For this reason, we focus on imports—more firms in Cape Verde were involved in importing that exporting. In practice, there is a very strong correlation between number of days for goods to clear imports and days to clear exports. Among the comparator countries, the correlation is 0.99—countries that take a long time to process imports also typically take a long time processing exports. It took 8 days for the median manufacturing firm‘s direct imports to complete all customs and port procedures in Cape Verde—higher than in any of the comparator countries except Guyana (see Figure 1). For example, it took only three days for the median firm‘s direct exports in the Maldives, 4 days in Mauritius, and 5 days in South Africa. This is consistent with the perceptions data—firms in Cape Verde were more likely to say that they were concerned about trade and customs regulations than firms in most of the other comparator countries (see Chapter 3). Previous studies have found that firms are far less likely to export and export less when trade and customs regulations are onerous and port and customs procedures are slow (Clarke, 2005). The survey does not distinguish between port delays, customs delays and delays on the side of the importer (e.g., due to unpaid custom duties or incorrect documents). Recent improvements in customs procedures suggest that the delay might be due to ports rather than customs. After the Foreign Investment Advisory Service (FIAS) study was completed, the Government adopted several initiatives to reduce delays due to custom procedures. In February 2003, the Government adopted Automated System for Customs Data (ASACUDA++), a computerized customs management system developed by United Nations Conference on Trade and Development (UNCTAD) covering most foreign trade procedures. According to customs, manifests are sent via ship, everything is computerized and only about 15 percent of containers are inspected. In 2005, customs introduced container seals which permit merchandise in domestic transit to no longer require customs‘ agents to physically accompany the containers to their final destination. This has also substantially reduced delivery times. Customs officials report that it takes only about 34 to 48 hours to clear customs procedures when documents are in order and duties are paid. 49 Figure 19: Customs and port procedures are slow in Cape Verde. Days for imports to clear customs 15 10 5 0 e l s a a s a s ga rd ve ne si iu an ric ne ne rit Ve di pi Af uy au al do ilip Se G e h M M ut ap In Ph So C Source: Investment Climate Surveys Note: All values are medians for enterprises with available data. A final issue is remaining barriers to trade. Although Cape Verde liberalized its trade regime in early 2004, with customs tariffs being simplified and lowered, tariffs are still above international standards. There are 7 tariff bands ranging between 0 and 50 percent. Taxes on imports have been streamlined and rationalized but the external current account continues to register extremely high deficits due to high imports of capital goods. To the extent that tariffs provide protection for inefficient domestic producers, they can result in import substitution and can discourage exports. 50 CHAPTER 3: PERCEPTIONS ABOUT THE INVESTMENT CLIMATE IN CAPE VERDE In addition to collecting information on firm productivity, the Investment Climate Survey also collects information on the quality of the investment climate—including on topics such as infrastructure, access to finance, crime and security, regulation, corruption, and taxation. Two types of information are collected: (i) perception-based measures that ask managers what they see as the major obstacles that their firm faces; and (ii) objective indicators such as production lost due to power outages, whether the firm has a loan or overdraft facility, and amount of time managers spend dealing with government regulations. The report uses both types of data—and supplementary information from other sources—to assess constraints to enterprise operations and growth in Cape Verde and compare these with constraints in the comparator countries. I. Perceptions about Constraints to Enterprise Operations and Growth Perception-based measures provide a good starting place for an analysis of the investment climate. Although perception-based measures suffer from several problems, an enterprise manager probably has a better grasp of the immediate problems facing his or her business than government officials, academic researchers, or other outside experts. However, they have several drawbacks. First, it is difficult to quantify and aggregate perception-based data across firms. Consequently, it is difficult to assess exactly what would need to be done to reduce the constraint. Second, although managers may be aware of a problem, they might not be aware of the underlying causes. For example, managers might know that it is difficult get bank loans to finance new investment, but might not know the underlying reasons for this (e.g., lack of competition in the banking sector, government debt issues crowding out private investment or problems with land registration that prevent firms from using land as collateral). Third, enterprise managers‘ interests might not always be consistent with society‘s interests. Most managers would like subsidized credit or to be charged less for power or water if they believed that the cost of providing these services would be borne by someone else. Similarly, most managers would be happy to face less competition even if the cost to society outweighed the benefits to their firm. Fourth, the perceptions of managers of existing enterprises might not reflect all obstacles to private sector growth. Managers of existing enterprises that have already completed registration procedures might not be concerned about entry costs even if they remain high. Similarly, they might rate some issues as lesser problems because they have structured their businesses in ways to minimize those costs. For example, if transportation costs are especially high in some areas, existing firms might only be located close to transportation facilities. Finally, because firms‘ experiences and expectations differ significantly between countries, it is difficult to compare perceptions across countries. Because of these concerns, although we use the perception-based data as a starting point, we supplement this information with objective measures of the investment climate taken from the ICS and other sources when appropriate. This chapter discusses the perception-based measures, while later chapters discuss specific issues in greater detail, supplementing this information with objective data. The additional objective data allow us to benchmark Cape Verde‘s investment climate against the investment climates of other countries. 51 Box 2: Perceptions about the investment climate—do they vary across countries? Many people claim that perceptions about the investment climate are uninformative—all firms in all countries would like lower taxes, better access to finance, lower interest rates, cheaper and more reliable infrastructure, less crime, and less burdensome regulation. If this were true, asking firms about these obstacles would provide little information—firms would always complain about certain things and would always like improvements along all dimensions. Given this criticism, a natural question is how much firms‘ perceptions reflect specific aspects of the investment climate in Cape Verde and how much they merely reflect complaints firms have throughout the world. This issue is discussed in greater detail in Appendix 1, which compares responses in Cape Verde to those of firms in other countries. Some complaints are more common than others. For example, tax rates rank among the top five complaints in five of nine comparator countries and the cost of financing and macroeconomic instability rate among the top five in six of nine. However, of the 17 areas of the investment climate, only trade and customs regulation and telecommunications fail to rank among the top 5 complaints in any country and none ranks in the top five for more than two-thirds of the countries—perceptions do vary between countries. I. MAIN PERCEIVED CONSTRAINTS Firm managers were asked to rate seventeen areas of the investment climate on a five-point scale that indicated how great an obstacle that area was to their business‘s operations and growt h. The issues that managers were most likely to say were a major or very severe obstacle (i.e., 4‘s and 5‘s on the five-point scale) to their firms‘ operations and growth are: (i) electricity; (ii) tax rates; (iii) cost of financing; and (iv) anti-competitive or informal practices (see Figure 20) Managers were most concerned about the electricity sector—close to two-thirds of enterprise managers said that this was a major or very severe obstacle (see Figure 2). Managers were also likely to say that other problems were serious obstacles. Nearly 50 percent of managers said that tax rates were a major or very severe obstacle; over 40 percent of managers were concerned about the cost of financing; and over one-third of enterprise managers rated anti- competitive or informal practices and access to financing as serious problems. Firms were considerably less concerned about other issues—less than one fifth of firms said that other constraints such as legal system, transportation, labor regulation, business licensing, corruption, access to land, and macroeconomic policy were serious problems. 52 Figure 20: Firms in Cape Verde are most concerned about electricity, tax rates, cost of financing, informal sector competition and access to financing % of firms saying issue is serious problem 80% 60% 40% 20% 0% in ion to ty or nan r sp tion un tion co Ed ion rm ces d d ion ef eg ns m n n g os om ng ec s nd ty eg ng e e tio io Ta cin ili Se to F rd m ste ci R tio it at La i r R si ce tab t t ss upt c an ula Te and stra Ac t a ula a tri t Le orta iso an t o pe bo icen R a m uc r S Ad Sy c Ac Ins rr x i i s i Co in El i l ss L fF ga ic n C m om e, de Tr ne s rim ra ct on La ke ax ls th T C le kil Bu ec T ro al ac C M fo or In W Source: Investment Climate Surveys. In addition to being asked about how great an obstacle an issue was on a five-point scale, firms were also asked what they saw as the biggest obstacle. While the first measure gives an idea about the breadth of concern—that is it allows us to see how many firms saw that area of the investment climate as a serious problem—the second gives some idea about the depth of concern. Since firms can only rank one issue as their greatest concern, a relatively small number of firms could give a single issue prominence on this measure. In the first measure, since firms can rate multiple problems as 4s or 5s, these small group issues would be unlikely to standout. For example, if exporters only make up a relatively small part of the population and are particularly concerned about trade regulation or transportation, these areas might not look like they are major problem on the first measure (i.e., only a small number of firms rated it as a major problem). Even a small group, however, can make an issue appear important using the second measure. Although the two questions do not always give identical answers, the responses to the two questions were generally consistent in Cape Verde. In particular, electricity appears to be the most serious problem using either measure. About 40 percent of firms said that electricity was the biggest problem—far more than any other issue. Other rankings were also similar—about 14 percent said competition with the informal sector was the biggest problem and 13 percent said tax rates were. About 6 percent said access to financing was their biggest concern, while 8 percent said the cost of financing was. No other problem was rated as the greatest concern by more than 3 percent of firm managers. 53 II. DIFFERENCES IN PERCEPTIONS ACROSS DIFFERENT FIRMS For the most part, there is no significant divergence of opinion across firm type (see Table 9). Most notably, managers of all types of firm rated electricity as a serious problem. Between 56 and 100 percent of firms of all types rated electricity as a serious or severe obstacle, putting it among the top five obstacles using this rating for all types of firms. There was similar consistency in the measure based upon the greatest problem. Over 40 percent of manufacturing firms, 36 percent of microenterprises, 35 percent of retail and other service enterprises and 63 percent of hotels said that electricity was the biggest problem. Hotels were most likely to say that electricity was a serious obstacle. This probably reflects the geographical distribution of the hotels in the sample—most were in Praia where firms were more likely to see electricity as a serious problem (see below). Manufacturing firms were more concerned about electricity than retail and service firms. In contrast, this probably reflects the greater vulnerability these firms have to poor sector performance rather than geographical distribution—in contrast to hotels, most manufacturing firms are located in Mindelo where electricity is a lesser problem than in Praia. Table 9: Concerns across sectors Sector—SMEs Micro Ownership Retail and Foreign- Domestic Manufacturing other Hotels All Owned Owned services Electricity 63.2% 60.3% 91.1% 55.8% 100.0% 61.9% Access to Financing 49.3% 30.3% 8.9% 28.8% 0.0% 37.3% Tax Rates 47.3% 54.6% 29.0% 37.5% 30.1% 52.1% Cost of Financing 45.6% 43.4% 17.7% 25.0% 21.0% 44.0% Informal Sector Competition 36.6% 36.9% 26.6% 51.0% 21.0% 37.8% Trade Regulation 25.7% 24.6% 18.9% 10.6% 28.7% 23.7% Transportation 22.3% 18.0% 18.9% 9.6% 40.2% 17.4% Tax Administration 21.8% 23.8% 8.9% 16.3% 7.7% 22.9% Crime, theft and disorder 20.5% 31.5% 18.9% 36.5% 46.1% 25.5% Business Licensing 16.7% 16.9% 18.9% 15.4% 13.2% 17.7% Labor Regulation 16.7% 15.6% 37.9% 0.0% 11.9% 18.9% Legal System 16.0% 23.7% 8.9% 9.6% 10.5% 21.2% Corruption 15.2% 17.7% 0.0% 11.5% 7.7% 16.0% Macroeconomic Instability 11.5% 9.4% 0.0% 14.4% 0.0% 10.0% Telecommunications 9.0% 31.9% 18.9% 3.8% 11.9% 25.7% Access to Land 7.2% 15.3% 8.9% 11.5% 10.5% 12.7% Worker Skills and Education 4.5% 25.6% 63.3% 5.8% 34.2% 23.2% In addition to electricity, competition from the informal sector and tax rates were fairly consistent concerns, rating in the top five problems for almost all types of firms. The only exception was that foreign-owned enterprises were less concerned about informal sector competition and microenterprises were less concerned. In addition to these issues that firms consistently rate as serious problems, some additional issues are consistently rated as minor problems—few firms of any type rated tax administration, the legal system, macroeconomic policy, corruption, or access to land as serious obstacles. Transportation was also not seen as a major concern by most types of firms. Nevertheless, on other issues, there is a noticeable divergence of opinion. First, manufacturing firms are more likely to rate access to financing as a serious obstacle than other types 54 of firms. Domestically owned firms are also more concerned about financing that foreign owned firms are. Whereas 37 percent of domestically owned firms said that access to financing was a problem, no foreign-owned firms did. The same basic pattern is observed when looking at the cost of financing. About 21 percent of foreign firms rated cost as a serious obstacle—compared to 44 percent for domestic owned firms. This could suggest that the mostly foreign-owned banking sector favors foreign-owned firms, could be because foreign-owned firms are different from domestic firms in other ways (e.g., they could be bigger or in different sectors of the economy), or could be because foreign-owned firms have access to finance in their home country. This issue is discussed in more detail, using objective data, in Chapter 5. Second, hotels are more likely to rate workers‘ skills and education and labor regulation as serious obstacles than other types of firms. Other enterprises do not generally see these areas of the investment climate as major issues. For example, about 63 percent of hotels rated workers‘ skills as a serious obstacle, compared with less than 5 percent of manufacturing firms. This might reflect the distribution of firms in the sample. Most hotels in the sample were in Praia—and as discussed below, firms in Praia were generally more concerned about this issue. Third, firms in retail and other services and informal sector are more likely to rate crime, theft and disorder as a significant problem. This is fairly common in investment climate surveys that cover this sector—retail enterprises in particular are particularly vulnerably to theft. 21 They have stock that can readily be converted to cash and, in contrast to manufacturing firms who can keep potential thieves out more easily, they need to keep their premises open to the public. In summary, different types of firms have roughly the same concerns in Cape Verde. Electricity and tax rates are rated among the top five problems for all types of firms. The fact that there is convergence on these perceptions and the large proportion of firm rating these issues as significant problems underline the relative importance of these two obstacles. Other areas, such as corruption, courts, and macroeconomic policy are seen as problems by few firms in any sector. However, some significant differences appear across firm type as manufacturing firms seem to experience problems with access to financing. The hotel sector appears to be the only one where the quality of the labor force in terms of education and skills and labor regulation is considered as a serious obstacle to enterprises operations and growth. And firms in retail trade and other services appear to be more concerned about crime and theft. III. REGIONAL DIFFERENCES IN PERCEPTIONS In addition to varying by sector, size and ownership, perceptions about the investment climate often differ by region. This is especially likely in an archipelago such as Cape Verde where there are significant differences between the islands. As noted earlier, firms were sampled in the two largest cities in Cape Verde—Praia, the capital city, and Mindelo. Praia is located on Santiago, while Mindelo is on São Vicente. 21 Results are similar, for example, in South Africa. See Regional Program on Enterprise Development (2005c). 55 There were several similarities between the concerns of firms on the two islands. Firms on both islands complained about tax rates and the unfair competition from the informal sector (see Figure 2). Few firms rated other areas, such as corruption, courts, business regulations, labor regulation, and macroeconomic instability, as major or very severe obstacles on either of the islands. Figure 21: Although enterprises on Mindelo and Praia generally had similar concerns, there were some differences in priorities and some differences in the intensity of problems % of firms saying issue is serious problem 100% 80% Mindelo Praia 60% 40% 20% 0% y n to ons n ns on reg on n r s co inis n Ac reg nce Ta ity ce nd Tr t of me rts ills ilit to io tio rm ate tio io ss atio ti c an c la at ab ou sk dm pt ri i tra ro abo ula tri ta a se at r om ul C ru fin fin to C st ec or x ul ne nic rs g al or in sp ss ke re El u C ic ce an m or r ss os fo e m W Ac Tr a ad ce In C x L si Ta le ec Bu Te ac M Source: Investment Climate Surveys. However, there were some differences. First, although enterprises on both islands saw electricity as a serious problem, firms in Praia were far more concerned—nearly 80 percent of enterprises said that electricity was a serious problem on Praia compared to only about 40 percent in Mindelo. Second, firms in Praia were far more concerned about worker skills and crime than firms in Mindelo. Finally, firms in Mindelo were far more concerned about access to finance, the cost of finance, and telecommunications than firms in Praia. For the most part, it makes intuitive sense that regional differences would be most visible in these areas. In general, perceptions about areas of the investment climate where there would be unlikely to be significant regional differences—such as tax rates, many areas of regulation and macroeconomic stability—should be similar across regions. Other areas, such as electricity or financing, would more naturally vary between regions depending on the quality of infrastructure in each location. If labor markets between islands were highly integrated, differences in perceptions about the availability of skilled labor should be small. This does not appear to be the case—as noted earlier, firms in Praia were more concerned about the availability of skilled workers. 56 Although differences between the islands with respect to perceptions about investment climate constraints could be due to differences in the extent of the problem, it might simply reflect other differences between the firms on the island. For example, retail trade and services firms are typically more concerned about crime than manufacturing firms and smaller firms are more concerned about access to finance. If there are differences with respect to the population of firms on the two islands, differences in perceptions could be the result of differences in the demographic makeup of firms in each city rather than differences in the extent of the problem. To test whether differences are due to differences with respect to the types of firms operating on each island, we ran multivariate regressions controlling for sector, size, and ownership. The differences in perceptions about finance, electricity, telecommunications and access to skilled labor remain statistically significant. The only major difference is that the difference in perceptions about crime is not statistically significant after controlling for differences between firms in the two cities. 57 CHAPTER 4: THE LABOR MARKET IN CAPE VERDE This chapter focuses on the labor markets in Cape Verde. Understanding a country‘s labor market is important for a number of reasons. The ability to work is the only asset that many people have and is vital for a country‘s economic development and poverty reduction strategy. Work provides individuals with income, reduces social exclusion and produces a sense of dignity and self- esteem. By creating opportunities for such work, efficient labor markets can reduce poverty. The cost of labor, however, is also one of the important determinants of firm productivity. Having an educated and flexible workforce is vital in an increasingly complex and global economy. Sound labor market policies and programs help workers manage risks associated with unemployment, income variability, and poor working conditions. Furthermore, in allocating labor to its most efficient use in the economy and encouraging employment and investment in human capital, a well- functioning labor market contributes to growth and development. Although economic growth in Cape Verde has been relatively strong over the past decade (average growth of 5.2 percent in 2000-2004), the number of formal sector jobs has grown slowly. As a result, many workers have become self-employed – either in agriculture and non-agriculture activities – or have entered the informal sector (self employment in Cape Verde accounted for about 25 percent of the labor force in early 2000). It also translated into a significant increase in the number of micro and small firms in these sectors. As a consequence the strong growth of the 90s and the more recent moderate growth have failed to improve the labor markets in the country and hence, did not significant contribute to a reduction in poverty. The analysis of the labor market in this chapter will cover mostly urban areas. The chapter is divided into two parts.22 First, we analyze the main differences in average wages across different types of firms in Cape Verde. Here we study also the main determinants of labor productivity, which will be approximated by average wages. Second, we compare labor productivity in Cape Verde vis-à-vis with other comparator countries focusing on manufacturing firms.23 I. THE WAGE STRUCTURE IN CAPE VERDE Economic growth affects labor not only through job creation and job destruction but also through changes in the wage structure. In the short run, the supply of workers with given skills is very inelastic. Hence, the effect of growth on the labor earnings is mostly driven by the effect of 22 The main issues in urban labor markets differ significantly from the issues in rural labor markets. Given the data constraints, in this section we only focus on the former. When reading the chapter one should keep in mind that urban labor markets are characterized by having: (1) more heterogeneous production process, more heterogeneous labor and higher wage variance (2) higher returns to education and skills and a higher concentration of these types of workers (3) lower dependence on seasonality (4) higher geographical concentration of production activities and (5) more intense policy regulation and union activity. 23 We do not devote special attention to labor market regulations in Cape Verde because the managers who answer the survey do not consider this topic a priority for the private sector development in Cape Verde. Managers where asked to rate the importance of labor regulation as an obstacle to the growth of their firms. Only 14% of the managers reported that labor regulations were a major obstacle and none of the firms reported it as being severe obstacle. Moreover, none of the firms considered labor regulations as being the major constraint affecting their operation and only two firms reported it has being the second most important constraint. This can also be a reflection of the recent changes in the labor law that significantly simplified the contracting and licensing procedures. 58 growth on the wage structure. In those sectors having the largest increase in demand there is an increase for sector specific skills, which in turn should translate into increased wages.24 Figure 1 shows the annual average wages and average years of schooling in manufacturing and services. Average wages in services are 24 percent higher than in manufacturing. The difference partly explained by differences in the quality of the workforce. The average workers in the service sector has 9.1 years of schooling, one year more than the average years in manufacturing. Figure 22: Average wages per employee are one and half times higher in services than in manufacturing –although workers are also better educated in this sector. Services Services Manufacturing Manufacturing 500 1500 2500 3500 4500 0 2 4 6 8 10 Wages per employee (USD) Years schooling Source: Investment Climate Survey. Figure 2 shows that there are also significant differences in wages within the manufacturing sector. The figure reports the average hourly wages for different manufacturing sectors—Food and Beverage, Garments and Textiles, Wood Products and other subsectors (including Chemicals and Paints, Plastics, Machinery and equipment, Electrical equipment and Construction). Averages wages are lower in the Wood sector than in other sub-sectors of manufacturing. Average wages in food and beverage production and in other sub-sectors are more than twice as high as wages in the Wood sector. 24 In the medium run, the relatively higher wages of the more scarce workers should create an incentive to the investment in skills and education so that wage skill premium is reduced. However, the impact of growth on labor earnings takes time and depends on the country initial conditions. 59 Figure 23: There is a significant heterogeneity in hourly wages within manufacturing. 4000 Av. Wages per employee (USD) 3000 2000 1000 0 Food and beverage Garments and Wood Products Other Textiles Source: Investment Climate Survey. Figure 3 documents that differences in average wages are also striking when comparing microenterprises from the microenterprise survey with the small and medium sized firms from the formal sector survey. Microenterprises are those with less than 5 employees, small firms have between 5 and 50 employees, while medium-sized firms have more than 50 employees. Small and medium-sized firms together pay wages that are on average two and half times higher than in micro firms. Part of these wage differences are explained by differences in human capital. The average worker is a small or medium sized firm has over 3 years more of schooling that the average worker in a microenterprise. There are also significant differences in investment in on-the-job training. While approximately 40 percent of the small and medium-sized firms offer training programs, only 14 percent of the microenterprises do so. This disadvantage of the micro firms has been established for other developing countries (see, for example, Regional Program on Enterprise Development, 2006). 60 Figure 24: Average wages per employee are two and a half time greater for Small and Medium sized firms than for Microenterprises. Small & Small & Medium Medium Micro Micro 500 1500 2500 3500 4500 0 2 4 6 8 10 Wages per employee (USD) Years schooling Source: Investment Climate Survey. Note: Results include both service and manufacturing firms. The finding that average wages are higher for small and medium sized firms than in microenterprises is sometimes interpreted as evidence of urban labor market segmentation. However one should be cautious since workers in the two sectors might differ in several characteristics (observable or unobservable) and this could affect their selection into each type of firm. Recent evidence from several middle income countries in Latin America actually suggests that there is significant mobility across formal and informal labor markets (Maloney, 1999).25 25 In this analysis we use the size of the firm as a proxy for formality. In table 2 we report more formal tests of the wage differences across firms of different sizes using worker level data. The advantage of the comparisons using the worker level data is that, unlike the differences in average wages reported in Figures 3 or 4, we actually control for several observable worker characteristics that proxy for the worker‘s human capital (e.g., education, experience, gender). However, to the extent that workers differ in unobservable characteristics (ability, preferences) that affect their selection into different sectors, these estimates would still be biased. 61 Within the group of small and medium sized firms we also find significant differences in average wages per employee between young and old firms as well as between small and medium sized firms.26 Figure 4 reports that average wages per employee are over twice as high in medium- sized firms and older firms than in smaller and younger firms. Again, part of these differences is explained by the higher human capital in these firms (either measured with average years of schooling or training of the workforce).27 Figure 25: Average wages per employee are higher in older and larger firms. Annual wages per employee (USD) 4500 4000 3500 3000 2500 2000 Young Firms (<5 Old Firms (> 5 Small Firms (5-50 Medium Firms (>50 years) years) workers) workers) Source: Investment Climate Surveys Note: Includes all firms in service and manufacturing sectors. To investigate further the differences across firms sizes, Table 1 below reports differences in different human capital characteristics between very small (5-10 employees), small (10-50 employees) and medium sized firms (>50 employees), both for manufacturing and services.28 There is robust evidence that in larger firms, the workforce is more educated and skilled, and also tends to be older and more experienced than in smaller firms. We expect these differences to translate into higher worker productivity and, hence, into higher wages. 26 We define medium sized firms as those with more than 50 employees and old firms as those with more than 5 years. 27 Large and small firms differ in the average wages per employee but also in other performance indicators. For example, employment in larger firms (more than 50 employees) grew 11.3% versus 5.5% in smaller firms in the period 2003-2005. 28 Statistics reported in Table 1 refer only for firms with more than 5 employees (excludes micro firms) since they are based on the information collected in the worker‘s module of the investment climate survey in Cape Verde. Worker level information was no t collected for microenterprises 62 Table 10: Medium sized firms have a workforce with a better human capital. Percent Skilled Av. Years Av. Av. Potential Percent Firms Av. Age of the workers in the schooling in Experience of experience of with on-the-job workforce workforce the workforce the workforce the workforce training Services Very Small (5-10 workers) 0.51 7.05 33.29 7.10 20.23 .276 Small (10-50 workers) 0.71 7.75 35.06 8.29 21.31 .435 Manufacturing Very Small (5-10 workers) 0.15 5.47 27.28 3.84 15.81 .127 Small (10-50 workers) 0.40 6.82 31.28 7.01 18.46 .263 Medium(>50 workers) 0.65 8.58 38.29 9.79 23.89 .575 Source: Investment Climate Survey. The size of the differences in average wages per employee across firm size categories (reported in Figure 25) are too large to be explained only by the differences reported in the previous table. Figure 26 reports that larger firms pay higher wages both for manufacturing and services separately. In particular, medium sized firms in manufacturing pay 75 percent higher hourly wages than small firms, and almost 200 percent more than very small firms. Figure 26: Hourly wages are higher for larger firms in both manufacturing and services. Medium Small Small Very Small Very Small 0.0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.5 Hourly wages (USD) Hourly wages (USD) Services Manufacturing Source: Investment Climate Survey. The size-wage premiums are also prevalent for the sample of skilled and unskilled workers. The size-wage premium is larger for the more skilled workers than for the unskilled workers. In manufacturing, medium sized firms pay 95 percent more than small firms for skilled workers and 35 percent more for unskilled workers. However, Figure 27 shows that even for the unskilled workers the size wage premium is still very significant. The wage premium for unskilled workers between very small firms (5-10 workers) and small firms (10-50 workers) is 55 percent for services and 5 percent in manufacturing, while it is 49 percent between small firms (10-50 workers) and medium firms (>50 workers) in manufacturing. 63 Figure 27: Hourly wages for unskilled workers are higher for larger firms in both manufacturing and services. Medium Small Small Very Small Very Small -0.5 0.5 1.5 2.5 Hourly wages (USD) Hourly wages (USD) Services Manufacturing Source: Investment Climate Survey. In sum, these differences show that, for a given skill group, larger firms tend to pay higher wages in both the manufacturing and non-manufacturing sectors. This evidence is consistent with efficiency wage theories in the urban labor market. Assuming that average wages in the firm are related to the firm‘s profitability, as it is the case in these theories, one could reconcile these findings since labor productivity is higher in larger firms (see Chapter 2). II. ECONOMETRIC ANALYSIS OF DETERMINANTS OF HOURLY WAGES The evidence on this topic will be more formally analyzed next when we investigate the determinants of wages. For this we will use information on hourly wages reported in the employer- employee data collected for each firm.29 The availability of worker and firm characteristics allows us to more formally test differences in hourly wages across the different types of firms. Since worker level data was only collected for firms in the large firm surveys in manufacturing and services (excluding hotels), microenterprise employees are omitted from this analysis. 30 The wage structure describes the market prices for various labor market skills (measured and unmeasured) and rents received for employment in particular sectors of the economy. This analysis 29 For each firm outside of the microenterprise sector, detailed information was collected at the worker level for up to 10 employees. The sample of workers covers the different occupational categories within the firm. 30 In this analysis we exclude those workers at the top and bottom 1% of the distribution of hourly wages. The industries included are Services (excluding hotels), Agro-industry, Garments, Wood products, Chemicals and Paints, Plastics, Machinery and equipment, Electrical equipment and Construction materials. The survey collects information on the current wage and on the number of hours of work in a week. These two variables are converted into monthly units and then divided to obtain hourly wages. 64 follows the standard approach of estimating a Mincer-type equation. The dependent variable is the logarithm of hourly wages and the independent variables are measures of human capital and firm characteristics. Mincer (1974) proposes an approximation for the financial private returns to education which can be readily estimated using cross-section data at the individual level. Studies using data from all regions of the world have examined these issues using this specification of wage structure for both male and female workers. The results of the least square estimation are reported in column (1) to (4) of Table 11. Column (1) reports the results including dummy variables for each industry (10 in total), years of schooling, years of on-the-job experience, squared years of on-the-job squared, a dummy variable for female employees and a dummy variable for union membership. Column (2) adds a dummy variable for whether the worker received any on-the-job training in the firm. Column (3) adds a dummy variable for firms with 10-50 employees and for firms with more than 50 employees, and column (4) adds the logarithm of the total number of employees.31 Table 11: The determinants of hourly wages in Cape Verde- Worker level regressions (1) (2) (3) (4) Observations 447 447 447 447 R-squared 0.4 0.41 0.42 0.42 Sector Dummies included (9) Yes Yes Yes Yes Schooling 0.1212 0.1138 0.1103 0.1106 [0.0095]*** [0.0098]*** [0.0099]*** [0.0099]*** Experience 0.0769 0.0725 0.0697 0.0715 [0.0120]*** [0.0120]*** [0.0121]*** [0.0120]*** Experience squared -0.0018 -0.0017 -0.0016 -0.0017 [0.0005]*** [0.0005]*** [0.0005]*** [0.0005]*** Female -0.1461 -0.136 -0.1372 -0.1351 [0.0574]** [0.0570]** [0.0569]** [0.0569]** Member of Union -0.0213 -0.0321 -0.0357 -0.053 [0.0627] [0.0623] [0.0630] [0.0631] Received Training - 0.1817 0.168 0.1576 [0.0636]*** [0.0641]*** [0.0646]** Firms>=10 & <51 employees - - 0.0755 - [0.0699] Firms with more 50 employees - - 0.2726 - [0.1403]* Log (Employees Firm) - - - 0.0927 [0.0477]* Constant -0.8753 -0.8596 -0.8804 -1.0783 [0.0888]*** [0.0882]*** [0.0946]*** [0.1428]*** Source: Investment Climate Survey. Note: Standard errors in brackets. * Significant at 10 percent; ** significant at 5 percent; *** significant at 1 percent The private return to an additional year of schooling in Cape Verde is between 11 percent and 12 percent.32 These numbers are similar to or greater than the private returns found in other developing countries. For example, Duflo (2001), using census data for Indonesia, finds a 7.8 percent return using a sample of males born between 1950 and 1972. Using a similar approach to 31 Omitted category in the regression is formal firms with 5-10 employees. 32 Alternatively, we have also estimated the same model as in column (1) but defining three mutually exclusive schooling categories: primary, secondary and tertiary education. The results (not reported) show that there are 39% returns to primary education, 77% to secondary education and 146% to a university degree in Cape Verde. 65 the one used in this chapter, the Investment Climate Assessment in South Africa estimates a similar average return—between 11 and 12 percent a year (Regional Program on Enterprise Development, 2005c). Returns in South Africa, however, seem to be especially high. Schultz (2004) shows that wages are higher for more educated workers in Ghana, Côte d‘Ivoire, Kenya and South Africa, Nigeria and Burkina Faso.33 He finds that the wage gains associated with each year of education completed range from about 5 percent-20 percent, with South Africa having the highest returns. Overall, these numbers suggest that individuals in Cape Verde have a strong financial incentive to acquire more education. These estimates are large when compared to the returns to investment in physical capital. Moreover, they are significantly greater than the OLS estimates of the returns to schooling for developed countries, which are approximately 5 percent and 10 percent. Angrist and Krueger (1991), Card (1995) and Lemieux and Card (2001) find returns that are close to 7 percent using the United States (US) or Canadian data.34 Wages also depends upon the worker‘s seniority within the firm. If seniority has a large effect on wages, then workers have a greater incentive to finance firm-specific capital and have a lower incentive to leave the firm.35 The size of the return to seniority is important in assessing the costs of dislocation from work, which has been subject of much research, especially in developed countries.36 The wage returns to an additional year of experience of on-the-job experience are lower than the returns for an additional year of schooling: approximately 70 percent for 10 years of seniority. Hourly wages rise with years of experience for most of the years in the sample. When we include potential experience in the regression, the coefficient on experience falls from 0.07 to 0.05.37 Again, this number is high and very close to the one reported for South Africa in that country‘s Investment Climate Assessment. It is also above the estimates found for developed countries. For example, Altonji and Shakotko (1987) and Topel (1991) report that 10 years of seniority raises the log wage by 0.26 and 0.3 respectively. It should be noted that OLS estimation might overestimate the effect of seniority on wages. Both experience and tenure are likely to be correlated with unobserved individual ability and job match heterogeneity, potentially biasing OLS 33 Schultz (2004) uses household surveys collected from 1985 to 1999 in six African countries. 34 The principal challenge in any effort to estimate the effects of education on wages is identification. Individual education and average schooling levels are correlated with wages for a variety of reasons, so the observed association between schooling variables and wages is not necessarily causal. In this cense one should be very cautions when interpreting the OLS returns presented in table 2. One type of bias relates to the omitted ability variables. If ability has an independent positive effect on earnings and if schooling and ability are correlated, the OLS returns will be an upward biased. The first attempts to address this problem included additional control variables, such as aptitude scores and parental background variables in the Mincer equation. Measurement error could also bias the OLS estimates. In the case of classic measurement error (uncorrelated with schooling), this leads to a downward bias in the estimates, possibly offsetting the ―ability‖ bias. More recently a literature focuses on the need to consider the issue of heterogeneity in schooling choices more explicitly (Card, 1999) 35 This is also related with models of wages that emphasize the use of deferred compensation as an incentive, insurance, or sorting mechanism (see, for example, Farber, 1999; Gibbons and Waldman, 1999). 36 Studies that examine the wage losses from layoffs include Addison and Portugal (1989), Altonji and Williams (2006) or Farber (1999) survey the literature on dislocated workers. 37 This specification is not reported. The magnitude of the squared term remains very similar. 66 estimates upwards. For example, experience on the job (or tenure) will be positively correlated with unobserved individual ability in the likely event that individuals with low productivity have high quit and layoff propensities. Wage structure might affect the relative earnings of groups like women, who tend to be employed in lower paying sectors of the economy. As in most data sets, the results in table 2 show that, all else constant (i.e., after controlling for industry, experience, and education), females tend to receive lower wages than similar males. The female wage gap is quite substantial —around 13 percent to 14 percent. It is worth emphasizing that this relates to differences across average hourly wages between males and females keeping all else constant (human capital and firm characteristics, including size and industry). Finally, there is no evidence that union membership increases wages— the difference in wages between union members and non-members is not statistically significant after controlling for other factors that affect wages. Work-related training is important for providing the workforce with the necessary skills for maintaining and enhancing the competitiveness of the firms and the economy. In the previous section, it was shown that firms with training programs tend to be more productive than other firms. The data from the individual level data is consistent with this—using wages as a proxy the worker productivity, trained workers earn more than untrained workers. Column (2) shows that all else constant (e.g., sex, education and experience), individuals who have received on-the-job training earn 16 to 18 percent more than those that do not receive training. This number is similar to the 20 percent return that studies have been found for developed countries (e.g., see Groot (1995) for evidence for the Netherlands) although smaller than the one found for South Africa (25 percent-30 percent). The magnitude of the effect in Cape Verde is substantial when compared with other investments, such as investment in schooling. This has been found for other countries using different data sets. Although this suggests that training is an important determinant of productivity and wages, the OLS estimates of the effect of training reported are likely to be biased upwards. If the ―best‖ workers self select into training or into firms that offer training, the estimated coefficient on training might be larger than the actual effect. Controlling for selectivity is difficult in the training context because it is hard to find variables that affect training decisions but do not affect earnings. 38 The firm-size effect on wages is also analyzed in table 2. The findings in column (3) and (4) show that after controlling for human capital and other worker characteristics still leaves a considerable wage premium. Column (3) shows that, keeping all else constant, workers in firms with more than 50 employees receive 27 percent higher hourly wages than workers in smaller firms. Alternatively, in column (4), we see additional evidence that workers in larger firms also tend to receive higher wages: an increase in total size of the firm by 1 percent is associated with 9 percent wage premium. This finding suggests that hiring better quality workers might not be the only reason 38 Recent literature has addressed the problem of the nonrandom selection into training by having an instrument for training or a difference-in-differences estimator (Heckman and others, 1997). Recent studies that use panel data and control for selection using individual fixed effect tend to find that training has a significantly smaller effect on wages. An example for this approach are and Blundell et al (1999). They find effects that are less than 1% per year. Barron, Berger, and Black (1999) also find small effects of training on wages based on fixed effect estimation. 67 for average wages in larger firms to be higher. This observation runs counter to most standard theories of the labor market. Brown and Medoff (1989) proposed several potential explanations for this observation, and several papers for developed and developing countries went to considerable effort to sort them out.39 The firm-size effect, however, seems to survive even those analyzes that control for endogeneity of firm size and individual heterogeneity (Idson and Oi, 1999). It is worth noting that in all the specifications we also find significant evidence of inter- industry wage differentials. Hourly wages in manufacturing are all else constant 21 percent lower than average wages in services. Also, wages in Chemicals and Paints are 23 percent higher than wages in agro-industry. In other words, not all the differences in average wages reported in figure 1 or in figure 2 are explained by the differences in worker characteristics. Evidence for other countries have also shown that even after accounting for worker and job characteristics, wage differences across sectors are rather stable.40 However, it is difficult with the available data to disentangle the importance of the different theories. One possibility is that these differences are explained by compensating differentials (either based on worker‘s quality or job characteristics). Another possible explanation for the inter industry wage differences is that they are based on ―true‖ wage differences across firms, even for exactly identical workers. Such differences could arise, in models of rent sharing and efficiency wages. Another interesting topic is whether there are significant returns to occupational skills. For that we compare the hourly wages of individuals with different occupations within the firm, after controlling for differences across individuals in education, gender, experience and training intensity. One reason for some occupations to have higher wages is because there could be more top managers and professional employees in firms that are better. More generally, firms might have different wage setting mechanisms that could be a function, for instance, of the firm‘s productivity and we expect wages to be higher for all occupations in the ―better‖ firms. Because of the well - known endogeneity problems of including a performance measure as explanatory variable in wage equations, we explore the panel dimension of the survey data to account for this. In the empirical work, we account for the unobservable firm specific characteristics that are likely to affect the skill premiums within each firm by including a firm time invariant effect in the model.41 The results for the five occupation categories - managers, top professionals, skilled production, unskilled production, non-qualified workers and ―other‖ workers - are reported in Figure 28. 39 Brown and Medoff (1989) discuss six explanations for the firm size effect. One is individual heterogeneity, the second is a compensating wage differential and the third is a union threat argument. They also consider a rent sharing argument, following from a premise that large firms have more market power, and the argument of worker monitoring being more difficult in larger firms. Finally, they consider the idea that large establishment have to pay more to recruit better employees. 40 Moreover because these differences tend to be similar across countries they are likely not to be explained by particular government interventions in the labor market or particular collective bargaining arrangements. 41 Computationally this is possible because we observe wages for 10 different employees within each firm, so that we can account for the fact that some firms have on average higher wages than other firms. 68 Figure 28: The wage skill premiums are quantitatively important even after controlling for the firm’s time invariant characteristics but are lower than in South Africa or Senegal. Cape Verde South Africa Senegal Production Production Production Unskilled Unskilled Unskilled Production Production Production Skilled Skilled Skilled Top Top Top professionals professionals professionals Managers Managers Managers 0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5 -0.5 0.0 0.5 1.0 1.5 % Gain by occupation Group % Gain by occupation group % Gain by occupation group Source: Investment Climate Survey The results in the figure show significant returns to skilled occupations within each firm in Cape Verde. The omitted category in the regression analysis is ―other workers‖, which includes, among others, health workers, accountants, security workers. Managers or firm owners earn on average 52 percent higher wages. This skill wage premium is still quantitatively important for the professionals or skilled production workers, which earn on average 56 percent and 44 percent more than the ―other workers‖ group, respectively. The table also reports the returns to skills for South Africa and Senegal. The returns to skilled workers are significantly lower in Cape Verde than in South Africa. As noted in the Investment Climate Assessment for South Africa, the exceptionally high wage premiums observed in South Africa probably reflect the significant shortage of skilled workers in South Africa. In fact, managers and top professionals in South Africa receive premiums that are more than twice as large as the returns in Cape Verde. In contrast, the relative returns in Cape Verde are not much different than the relative returns in Senegal. It is worth emphasizing that the magnitude of the occupational wage premiums in Cape Verde is significant, especially for managers and top professionals. As a consequence, individuals who invest in the acquisition of these skills should get high returns for their investment. These returns could reflect the shortage of this type of workers in Cape Verde which is, however, much smaller than the one in South Africa. 69 III. MANUFACTURING WAGES IN CAPE VERDE COMPARED TO OTHER COUNTRIES To compare the attractiveness of Cape Verde relatively to other countries in the establishment of labor intensive industries we compare average wages for different countries. Whereas the evidence in the productivity chapter looks at the average cost of labor using data from the firm‘s income statement, this compares wages for different categories of production workers for a typical worker in that group. Figure 9 compares annual wages per employee in manufacturing in Cape Verde versus other comparator countries. Median annual wages in Cape Verde are nearly twice as high as in the Philippines. Wages in Cape Verde are slightly higher than in Guyana and are slightly lower than in Mauritius. They are far lower than wages in South Africa, which are over four times greater in South Africa than in Cape Verde. This pattern is very similar when we analyze wages by skilled and unskilled production workers. Wages in Cape Verde are similar to or greater than the comparator countries, except for South Africa. Figure 29: Wages in Cape Verde are far lower than in South Africa, but higher than in Senegal or the Philippines All Workers Skilled Production Unskilled Production Workers Workers South Africa South South Africa Africa Mauritius Mauritius Mauritius Guyana Guyana Guyana Philippines Philippines Philippines Cape Cape Cape Verde Verde Verde 0 5000 10000 0 5000 10000 0 5000 10000 Median Wages per employee Median wages per employee Median wages per employee Source: Investment Climate Survey Differences in wages for unskilled workers are especially important when looking at the competitiveness in labor intensive industries (e.g. textiles). The results in Figure 29 suggest that Cape Verde would find it difficult to remain competitive in labor intensive export industries, even without considering high transportation costs. Costly labor could be one of the reasons explaining the low levels of foreign direct investment (FDI) in Cape Verde. 70 One possibility is that the samples of firms across the countries analyzed are very different, potentially biasing the results. To test the robustness of the results we have compared the wages per employee across countries only for firms with more than 50 employees. The results again show that the mean average wages in Cape Verde for unskilled workers are well above than those in other countries. The only exception is again South Africa but the size of the gap between the two countries is smaller. Wages for unskilled workers are twice as high and wages for skilled workers are nearly one and a third times as high in South Africa when compared with Cape Verde when the analysis is restricted to this class of firm. IV. TRAINING One explanation for these wage differences is that the workforce in Cape Verde is actually more productive than the workforce in other countries because they have a better human capital. One proxy to measure the quality of the human capital is the share of firms that offers on-the-job training - internal or external to the firm – to its workforce.42 Figure 30 shows the share of firms offering on-the-job training in Cape Verde vis-à-vis in other countries. Figure 30:Firms in Cape Verde tend to offer more training than firms in Guyana, Indonesia, Philippines or Senegal. Share of firms with on-the-job training (internal or 1.0 0.8 0.6 external) 0.4 0.2 0.0 Cape Verde South Africa Mauritius Guyana Senegal Philippines Indonesia About 42 percent of the firms in Cape Verde offered on-the-job training. This is higher than in Guyana, Indonesia, the Philippines and Senegal, but it is considerably lower than in Mauritius or South Africa (60 percent and 64 percent of the firms train, respectively). It is also well below the 42 Firms in each of the comparator countries where asked whether they offered any on-the-job training to their employees and if so, what fraction of skilled and unskilled workers was offered this training. However, one should keep in mind that the quality of the training offered might be quite different across countries. Unfortunately, there is no way we can account for this. 71 share of firms training in China (84 percent) or Brazil (67 percent), but is close to the share of firms training in Poland 48 percent. The incidence of on-the-job training depends very strongly on the specific technology used, and thus on the firm‘s sector of activity. Moreover, large firms tend to be more likely to train their workers than smaller firms. To the extent that firms are credit constrained or if there is a fixed cost in the supply of on-the-job training to the workforce (e.g., buying training equipment), bigger firms are better positioned to offer formal training programs than smaller firms are. This could happen because larger firms have an easier access to finance or because they have more potential trainees to spread the fixed costs of the training. Hence, at least part of the pattern in figure 8 could be driven by the sample‘s sector and composition. To address this we estimate a simple binary model of the incidence of training on different firm characteristics. Column (1) includes only country dummy variables (omitted country dummy is Indonesia), column (2) adds dummy variables for small firms (10-50 employees) and medium firms (more than 50 employees), column (3) adds the manager‘s education and column (4) controls for industry dummies (13 groups). The results of the maximum likelihood estimates are reported in the Table 12 below. Table 12: Firms in Cape Verde are more likely to train than firms in Senegal or in the Philippines after controlling for size, sector and the manager’s education. (1) (2) (4) (3) Industry Dummies? No No Yes Yes Observations 2534 2534 2203 2371 Cape Verde 0.659 1.2995 1.54691.4172 [0.1918]*** [0.2017]*** [0.2194]*** [0.2107]*** Mauritius 1.0945 1.3029 1.45431.3309 [0.1081]*** [0.1134]*** [0.1965]*** [0.1197]*** Philippines 0.0141 0.1241 -0.02710.0217 [0.0760] [0.0796] [0.0944] [0.0908] Senegal 0.3713 0.8349 1.0327 0.948 [0.1213]*** [0.1326]*** [0.1560]*** [0.1482]*** South Africa 1.2201 1.2724 1.28461.3364 [0.0760]*** [0.0792]*** [0.1046]*** [0.1002]*** Small Firms (10-50 employees) - 0.222 0.14620.2567 [0.1359] [0.1742] [0.1636] Medium Firms (>50 employees) - 1.0936 0.93021.1335 [0.1376]*** [0.1758]*** [0.1642]*** Manager‘s Education (years schooling) - - 0.1666 - [0.0252]*** Constant -0.8468 -1.7194 -1.8673 -2.3949 [0.0543]*** [0.1461]*** [0.1857]*** [0.2174]*** Source: Investment Climate Survey. Note: Standard errors in brackets. * Significant at 10 percent; ** significant at 5 percent; *** significant at 1 percent. Dependent variable is a dummy variable equal to 1 for firms reporting that offer some formal training program to its workforce and zero otherwise. Even after controlling by the sector and size composition of firms, it is still true that firms in Cape Verde tend to offer more on-the-job training than firms in Indonesia, Philippines or Senegal. Taken together, we interpret this as evidence that wages are relatively high in Cape Verde but that the workforce is flexible enough so that almost half of the sampled firms offer some form of on-the- job training to their employees. In fact, in the survey managers where asked whether the education and efficiency of the workforce was a major obstacle to the operational activity of the firm. Only 5 72 percent of the managers reported that this was a major or severe problem and none of the firms reported that skills were the most important obstacle facing the firm‘s operations. One of the reasons why managers in Cape Verde might be more likely to offer on-the-job training programs relates with their educational level. More educated managers are more likely to perceive the benefits of on-the-job training so that the incidence of training would be higher in firms with more educated managers. However, this does not seem to be the case in Cape Verde. Managers here, as well as in Senegal, have at most some vocational training, while managers in Indonesia or the Philippines have at most a graduate degree. The results in table 3 reinforce this finding. In column (4), controlling for industry, size and for the education of the manager, firms in Cape Verde are more prone to offer training than firms in Indonesia, Philippines or Senegal. In addition to looking at the share of firms offering on-the-job training programs (intensive margin), it is also important to know the coverage of the training, measured by the number of workers involved in the training (extensive margin). Figure 11 reports the share of workers that receive on-the-job training, by skill groups, in Cape Verde vis-à-vis other countries. Figure 31:Firms in Cape Verde train a significant number of skilled and unskilled workers. Poland Poland Brazil Brazil China China South Africa South Africa Philippines Philippines Mauritius Mauritius Guyana Guyana Cape Verde Cape Verde 0 20 40 60 80 100 0 20 40 60 80 100 Sh. Skilled workers Sh. Unskilled workers Source: Investment Climate Survey. The coverage of training in Cape Verde is almost half of the workforce: firms that train in Cape Verde do it on average for 40.1 percent of the skilled workforce and 47.3 percent of the unskilled workforce. This number is quite above the coverage in the Philippines or Mauritius and it is quite close to the coverage in South Africa, where firms train 45 percent and 46.3 percent of the skilled and unskilled workforce, respectively. However, the coverage is quite below in some other middle income countries. For example, in Brazil 77 percent, in Poland 79.9 percent and in China 69.1 percent of the skilled workers receives some training. The same advantage holds for unskilled workers in these countries. 73 We have contrasted the coverage of training implied by the firm level data with the coverage of training implied by the worker‘s responses to the survey. 43 Approximately 46.5 percent and 17 percent of the skilled and unskilled workers respectively, reported having received some on the job training in the past. The discrepancy of the coverage for the unskilled workers might result from the fact that in the worker‘s survey the training question refers to past training and, therefore could refer to a previous employer, or simply because managers do not answer correctly. Unfortunately, with the available information it is impossible to disentangle the exact reasons for this difference. 43 Workers where asked whether they have received on-the-job training in the past, how long it took and how the training was financed. However, it should be noted that this is past information and that it might not have taken place within the surveyed firm. 74 CHAPTER 5: ACCESS TO FINANCE I. THE FINANCIAL SECTOR IN CAPE VERDE Cape Verde only developed a modern banking sector a little over a decade ago (World Bank, 2004a). In 1995, Cape Verde‘s monobank was split into a central bank, Banco de Cabo Verde (BCV), which is responsible for regulation, and a commercial bank, Banco Comercial do Atlantico (BCA). Caixa Economica de Cabo Verde (CECV), which had started as a branch with the Post Office in 1928 and was transformed into a commercial institution in 1985, was already operating.44 Despite the recent transformation, Cape Verde‘s financial sector is better developed than the financial sectors of most low-income countries in Sub-Saharan Africa (World Bank, 2005a). A recent IMF study that compared financial sector development in Cape Verde to 12 low- income in Sub-Saharan Africa found that credit to the private sector was higher in Cape Verde (34 percent of GDP) than in any of the other countries (average of 12.4 percent) (International Monetary Fund, 2005b). Cape Verde also performs well relative to other lower middle-income countries. Money and quasi-money (M2) and private sector credit are greater as a percent of GDP than in the Dominican Republic, Indonesia, the Philippines, or the Maldives (see Figure 32). For the most part, however, the differences are relatively modest. Although private sector credit was lower in Cape Verde than in the two upper-middle comparator countries in Africa, Mauritius or South Africa, this might not be surprising given that these countries have considerably higher per capita GDP. Figure 32: The financial sector is Cape Verde is better developed than it is in low-income countries in Africa and is comparable to middle-income countries in other parts of the world 100 75 % of GDP 50 25 0 us s a a de al ia s ic ve ne an ric g l es ub iti r ne Ve di Af pi uy r n au ep al do ilip Se G h e M M R ut ap In Ph n So C a ic in om D Private sector credit (% of GDP) M2 (as % of GDP) Source: World Bank (2006c) 44 The history of CECV is presented on its website (http://www.caixaeconomica.cv/eng/company/aboutus.asp) 75 The banking sector dominates the financial system in Cape Verde, with commercial banks accounting for 87 percent of financial system assets. There are four commercial banks with 44 agencies, seven non-bank financial institutions (two insurance companies, one venture capital, two exchange houses, leasing and the Sociedade Interbancária e Sistemas de Pagementos (SISP), and four offshore banks. Five additional off-shore banks have requested licenses. The banking sector generally performs well on most measures of financial soundness. Although the level of non-performing loans (NPLs) is higher than in most developed economies, it is lower than elsewhere in Sub-Saharan Africa and has been declining. In 2002, the combined, bank-wide non-performing loan (NPL) portfolio (as a percentage of the total) was over 9 percent. As a result of improved supervision efforts by the BCV, accelerated collection efforts and increased write-offs of bad loans by the commercial banks, the credit portfolio of the commercial banks has been steadily improving. In 2004, it was 7.23 percent and in 2005 it was 6.31 percent. It is expected that the results for 2006 will be even better—by May 2006, for instance, the largest commercial bank had reduced its NPL portfolio to 6.9 percent (down from about 9 percent). Moreover, banks appear to be relatively profitable—return on equity is higher than global benchmarks and than in other countries in Sub-Saharan Africa (International Monetary Fund, 2005b). Although this suggests that the banking sector is performing relatively well, some concerns remain. One is that the banking sector in Cape Verde is highly concentrated. There are only four major banks, with two of these accounting for close to 90 percent of assets. Moreover two of the four banks—the first and third largest—are subsidiaries of the same Portuguese state-owned bank. The second largest bank, accounting for about 25 percent of total assets at the end of 2005, is owned by two Portuguese banks. The final bank, which was acquired by a group of private domestic investors in 2004, accounts for only 3 percent of total assets. Although in a small country such as Cape Verde, economies of scale make a certain level of concentration inevitable, the highly concentrated banking sector could potentially lead to low levels of competition. Intermediation costs were relatively high in the mid-2000s (World Bank, 2005a), with nominal interest rates remaining high despite falling inflation (International Monetary Fund, 2005b). Reserve requirements were reduced from 19 to 17 percent in 2005 and to 15 percent in 2006. This helped to reduce commercial bank consumer and commercial lending rates to between 8 and 11 percent buy mid-2006. Although the banking sector is relatively well developed, other sources of funds are more limited. Although insurance companies could potentially be a source of long-term funds, the insurance market is very small. Although total premiums increased by 54 percent between 2004 and early 2005, most of the growth in premiums was due to increases in premiums from mandatory car insurance, which accounts for 55 percent of total insurance premiums and provides little or no long-term capital to the market. Life insurance, which could be an important source of long-term capital, remains underdeveloped, accounting for only 1.3 percent of insurance premiums. In many developed countries, stock markets are an important source of funds. The stock market in Cape Verde is small with only a limited number of listed companies—although activity has been increasing in recent years. In 2002, the Government decided the cost of maintaining the country‘s stock market, the Bolsa da Valores de Cabo Verde (BVC) was too high and put the 76 initiative on hold. Following a review of the financing sector, completed as part of a development policy review, the Government decided to re-launch the BVC. It was inaugurated on October 15, 2005. As the first issue, the Government placed 26 percent of the shares in the formerly state- owned tobacco company on the exchange. These shares were almost three times oversubscribed. Since this time, shares of several additional companies (BCA, Garantia and CECV) have also been listed. The Government anticipates that shares in Improfac, Emprofac and Enacol (pharmaceutical and oil companies) will be the next to be listed. The Government also placed 44 treasury bonds dating back to 1993 and the new Treasury bill issues on the BVC. The BVC operates according to international standards. It has developed an extensive training program, which includes a masters course in finance, taught by professors from Instituto Superior de Economia e Gestão (ISEG) in Portugal. To date, over 1,000 persons have been trained in various courses. These courses are provided on a cost recovery basis and the BCV has been operating without Government support. Two important initiatives have recently been launched under the leadership of the Chambers of Commerce, and with support from the World Bank, to improve access to credit. A Task Force has been created to: (a) assist the institutionalization of the recommendations resulting from the dialogue between Government and the Private Sector; and (b) ensure that Government and the Private Sector are able to operationalize, monitor and evaluate the implementation of agreed recommendations. This Task Force includes representatives from the Chambers of Commerce, various Commercial, and Financial Institutions and Heads of Departments within the Ministries of Government that deal most directly with the Private Sector, namely Commerce, Tourism, and Cape Verde Investments. The first initiative was an agreement between the commercial banks and the Chambers of Commerce in setting up a Commercial Credit Information System. This will be comprised of a database of commercial firms‘ clients (individuals and other entities) that purchase goods and services on credit. The database will be available and accessible to participating firms either through on-line, voice or fax connections. It will provide credit history and other financial information about the prospective client and thus allow for better informed sale and credit decisions. Participating institutions will have access to this essential credit information. This project, estimated to cost some US$35,000 equivalent, will be jointly financed by the commercial banks (which will provide the required software and servers), the Chambers of Commerce (which will provide the relevant computer, voice and fax equipment) and a World Bank project (which will finance related training and marketing). This initiative will encourage debtors to be more responsible in paying their bills and encourage the development of additional financial products, such as ―factoring‖. The Chambers of Commerce have been working with interested financial institutions on the issues of information privacy and constitutionality of the system. The Chambers of Commerce have commissioned a legal opinion which allays those concerns and details how the system can comply with existing legislation. Modeling of the software and the database that will comprise the system is being done by a Brazilian firm and is already functional. A practical demonstration was held during the May 2006 Public-Private Forum and a testing phase will now ensue. The Chambers intend to have the system up and running by the third quarter of 2006. The second important initiative will help improve access to credit for those firms that have taken steps (through the matching grant fund, MGF) to strengthen their capacity. Although the 77 commercial banks have consistently indicated their willingness to expand credit, they each see a limited number of bankable projects. The banks argue that although project ideas submitted may often be good, the studies accompanying them are rarely up to required standards. On the other side, firms believe that although their project ideas are good and proven viable by studies financed with the MGF resources, their business plans often do not receive the attention they believe is warranted from the banks and are too infrequently funded. Agreement was reached that a training program for consultants that assist the MGF beneficiaries will be developed and implemented under the auspices of the Chambers. This should lead to a pool of certified consultants capable of producing bankable projects that would be approved by the Chambers of Commerce and would not have to undergo the detailed scrutiny that would otherwise be the norm. The program will also include information activities regarding lines of credit conditions, and its consultants will work with staff of banks‘ commercial departments so that better rapport is created between the private sector and the banks. The task force created under the auspices of the protocol has been meeting regularly and an action plan for 2006 should be submitted for validation in the upcoming Forum. II. ACCESS TO FINANCE IN AN INTERNATIONAL PERSPECTIVE Although firms—especially domestic firms and those in the manufacturing sector—rated access to finance and cost of financing as serious obstacles to their operations and growth, the objective evidence from the Investment Climate Survey is mostly consistent with the idea that the financial system is not an especially great constraint on enterprise operations and growth. Although not exceptional when compared to the comparator countries, firms in Cape Verde do not appear to be especially disadvantaged either. About 45 percent of manufacturing firms in Cape Verde have loans (see Figure 33).45 This is higher than in any of the comparator countries. In comparison, about 40 percent of manufacturing firms from South Africa and Senegal and about 20 percent of manufacturing firms from Guyana and the Maldives have loans. Consistent with the previous results, firms in Cape Verde generally finance a greater share of new investment through bank loans than firms in the comparator countries (see Figure 33). The average firm in Cape Verde financed about 22 percent of new investment with bank loans. This was higher than in any of the comparators except Guyana (25 percent) and Mauritius (35 percent). If anything, this probably underestimates the relative performance of the sector on this dimension. The median firm in Cape Verde was smaller than the median firm in most of the other countries. Whereas the median firm in Cape Verde had only 15 employees, the median firm in Guyana had 18, the median firm in Senegal had 25, the median firm in the Maldives had 33, and the median firm in South Africa had 95. Since small firms are typically less likely to have access to finance than larger firms, this might affect the relative share of firms with loans. 45 As noted earlier, for reasons of comparability, we compare responses for manufacturing firms in the cross country comparisons. 78 Figure 33: Firms in Cape Verde have good access to long-term sources of finance even relative to other middle income countries. Mauritius Dominican Republic Guyana Maldives Senegal Senegal South Indonesia Africa South Africa Guyana Philippines Dominican Republic Maldives Cape Verde 0% 25% 50% 75% 100% Cape Verde % of new investment 0% 20% 40% 60% Banks Internal Funds % of firms with loan Informal Sources Non-Bank Formal Source: Investment Climate Surveys Note: Cross-country comparisons are only for manufacturing enterprises Despite this, firms are heavily reliant upon retained earnings for much of their investment. The average firm financed about 68 percent of new investment through internal funds. This was higher than 5 of the 8 comparator countries. Since they finance more with bank loans than in most of the comparator countries, this reflects the relatively modest levels of investment that they finance with money from informal sources (mostly family and friends) and non-bank formal sources. None of the manufacturing firms reported using loans or gifts from family or friends or moneylenders and only 4 of the 27 firms that reported new investment said that they used financing from other formal sources. Of these firms, 3 reported financing from formal financial institutions other than banks and 1 firm—a private domestic firm—reported receiving government financing. Although loans appear to be a lesser problem in Cape Verde than in most of the comparator countries, short-term financing appears to be a greater constraint. Firms in Cape Verde were less likely to have overdraft facilities than firms in the comparator countries (see Figure 34). Only about 17 percent of manufacturing firms in Cape Verde had overdraft facilities. In comparison, about 20 percent of manufacturing firms in Indonesia, 36 percent of firms in Guyana, and 88 percent of firms in Mauritius had overdraft facilities. Consistent with this, firms in Cape Verde finance only a small portion of their short-term capital needs through financing from the banking sector (11 percent of total). The only country where manufacturing firms financed less of their short-term needs with bank financing was the Philippines (8 percent). 79 Figure 34: Firms in Cape Verde use banks to finance less of their short-term working capital needs than firms in other countries. Mauritius Mauritius Guyana South Africa Indonesia Senegal South Africa Dominican Republic Maldives Guyana Senegal Dominican Republic Philippines Philippines Maldives Indonesia Cape Verde 0% 25% 50% 75% 100% Cape Verde % of short-term assets 0% 25% 50% 75% 100% Banks Internal Funds % of firms with overdraft Informal Sources Non-Bank Formal Source: Investment Climate Surveys Note: Cross-country comparisons are only for manufacturing enterprises Moreover, as with new investment, firms also rely less heavily on informal sources (family and friends and money lenders) and non-bank formal sources than in any of the comparator countries. None of the firms in the sample reported using financing provided by family and friends or money lenders and only 3 of 40 firms reported using financing provided by other formal sources. In all the cases where firms reported using other sources of formal financing to finance short-term assets, they reported using credit from suppliers or clients. As a result, firms were more reliant upon retained earnings to finance short-term assets than firms in any of the comparator countries. On average, manufacturing firms in Cape Verde financed about 82 percent of short-term assets with internal funds. In comparison, manufacturing firms in Indonesia financed about 48 percent, firms in South Africa financed about 66 percent and firms in the Maldives financed about 74 percent. Firms that had not applied for a loan were asked why they did not have one. For the manufacturing firms in Cape Verde, only 2 of 11 firms that had applied for a loan had been rejected. The other 30 firms had never applied. The most common response (about 50 percent of firms) was that they did not need one. The next most common response was that interest rates were too high (about 23 percent of firms). The remaining firms said that collateral requirements were too strict (6 firms) or replied ‗other‘. 80 Based upon these responses, just over 18 percent of firms appear to be credit constrained (i.e., that they would like a loan at current interest rates but don‘t have one). 46 This was relatively high compared to other countries (see Figure 16). In comparison, only about 2 percent of firms in South Africa, 9 percent of firms in the Dominican Republic, and 17 percent of firms in Guyana, and just under 18 percent in Senegal. Figure 35: Only a relatively modest number of firms reported that they would like to have a loan but could not get one. (% of firms that are credit constrained) 20% 15% 10% 5% 0% Cape South Dominican Guyana Senegal Verde Africa Republic Source: Investment Climate Surveys Note: Firms are considered credit constrained if they responded to a question about why they did not have a loan with a response indicating that they could not get one. For example, if they responded that they did not think they would get one, they did not have enough collateral, they had applied but had been rejected, or that bank procedures were too complicated. Cross-country comparisons are only for manufacturing enterprises Why do so many firms report being credit constrained despite the relatively large number with bank loans? One plausible explanation is that other types of formal financing are less well developed in Cape Verde than in the other middle income comparator countries. That is, although the relatively heavy dependence upon retained earnings to finance both short-term assets and new investment could be because firms have no need for external financing—for the most part firms prefer to use internal sources when available—the evidence on credit constraints suggests that this might not be the only reason. That is, firms might rely upon internal funds due to problems with getting external funds. Consistent with this, firms that report being credit constrained rely upon internal funds more heavily for financing than firms that report not being credit constrained (94 percent of short-term assets and 100 percent of new investment for credit constrained firms compared to 80 percent and 60 percent for non-constrained firms). Although the banking sector appears relatively well developed, it seems harder to raise money from other sources (both formal sources and from family and friends). In addition to providing information on access to credit, for firms with loans the Investment Climate Survey also provides information on the terms of the enterprise‘s most recent loan 46 Note that this is after weighting responses – the unweighted mean is higher. 81 including the period of the loan (from the time that the loan was awarded) and the interest rate of the loan. The availability of long-term loans might explain how firms use bank loans to finance long-term investment rather than just short-term assets more than in other countries. About 60 percent of manufacturing firms in Cape Verde reported that they had loans with durations of over three years. This compares relatively favorably with many of the other middle-income countries. Less than a quarter of firms in Indonesia and the Philippines, less than half of firms in the Dominican Republic, and less than 60 percent of firms in the Maldives and Senegal reported that their most recent loan had a duration of more than three years. Loan periods were longer in South Africa, Mauritius and Guyana, however. Figure 36 A large number of firms have loans with durations greater than 3 years. (% of firms with loans of over three years) 100% 75% 50% 25% 0% Guyana Maldives South Verde Senegal Indonesia Philippines Mauritius Africa Dominican Cape Republic Source: Investment Climate Surveys Note: Cross-country comparisons are only for manufacturing enterprises Although periods are relatively long, real interest rates also appear quite high in Cape Verde. The median firm with a loan reported an interest rate of 12 percent—slightly higher than in the Maldives, South Africa, the Philippines, Senegal, and Mauritius and considerably lower than in Indonesia, Guyana and the Dominican Republic. Inflation, however, has been relatively modest in Cape Verde. As a result, real interest rates appear quite high in relative terms. As noted previously, real lending rates have risen considerably due to a sharp drop in inflation (International Monetary Fund, 2005b). 82 Figure 37 Nominal interest rates are relatively low in Cape Verde—but so is inflation. 30 25 20 rate (%) 15 10 5 0 Verde Senegal Indonesia Guyana Maldives South Mauritius Africa Dominican Philippines Cape Republic -5 Median Interest Rate Inflation Source: Investment Climate Surveys and World Bank (2006c) Note: Inflation rates are averages between 2002 and 2004. Note: Cross-country comparisons are only for manufacturing enterprises III. DIFFERENCES IN ACCESS BY FIRM TYPE The previous section provides some evidence that access to finance is relatively good in Cape Verde, even when compared to other middle income countries. This sub-section looks at differences in access across firms of different types within Cape Verde. The results presented in this section are based upon an econometric analysis that is described in detail in Appendix 2. The advantage of regression analysis is that it allows us to simultaneously control for multiple differences at the same time (e.g., size, ownership and sector of operations), making it easier to see where the differences are. This section describes the main results from the econometric analysis. Firm Size. In most other countries where ICSs have been conducted, large firms have better access to credit than smaller firms do. This is also true in Cape Verde. After controlling for other things that might affect access to credit (e.g., size, ownership, and sector of operations), large firms appear to have better access to finance. Increasing firm size by 10 workers increases the likelihood that the average firm will have a loan by 8 percentage points, increase the likelihood that it will have an overdraft facility by 2 percentage points, reduces the interest rate by 0.6 percentage points, and reduces the likelihood that the firm reports being credit constrained by 10 percentage points. 47 Although small firms appear to be less likely to have access to bank loans or overdraft facilities, this does not mean that they rely heavily on informal sources such as money lenders or even loans from family and friends. Microenterprises and small firms financed less that 2 percent of the new investment in 2005 using loans from informal sources. The average large firm does not 47 See Appendix 2 for issues related to statistical significance and robustness of results. The results for loans and being credit constrained were statistically significant at a 1 percent level, while the results for overdraft facilities and interest rates were only significant at a 20 percent level. 83 use informal sources of funds at all. The main difference, therefore, appears to be with respect to internal funds—microenterprises finance about 85 percent of short-term assets with internal funds compared to 78 percent for small firms and 69 percent for medium-sized firms. In addition to relying more heavily on banks for financing new investment and short-term assets, medium-sized firms are also more likely to have access to other sources of formal financing. The main source of formal financing was trade and commercial credit (e.g., from suppliers). This accounts for all other sources of formal financing reported by medium-sized firms. Although this pattern is not uncommon, it seems plausible that this might reflect characteristics of the financial sector in Cape Verde. Recent research suggests that banking sector concentration is especially problematic for Small and Medium Sized Enterprises (SMEs) (Beck and others, 2004). Given that there is concern about access to credit for SMEs in Cape Verde, this might be a serious concern (World Bank, 2005a). The banking sector is also heavily dominated by foreign banks. Although there have been questions about foreign banks‘ willingness to provide credit to SMEs, it is important to note that the empirical evidence on whether this is the case is mixed. Although some studies in developing countries have found that foreign-owned banks are less likely to lend to SMEs than similar domestic banks, other studies have found that bank ownership has little impact on lending to SMEs.48 Ownership. The results from the econometric analysis do not support the idea that foreign- owned firms have better access to finance than domestic firms—they were no more likely to have loans or overdraft facilities, were not less likely to be credit constrained and those with loans did not appear to have better terms (e.g., lower interest rates or longer loan durations).49 Education of Manager. Better educated managers might find it easier to get loans than less well educated managers for a number of reasons. They might have better connections (e.g., with bank managers) than other firms. Or they might find it easier to fulfill the bureaucratic requirements needed to get a loan, such as completing a business plan or keeping accurate financial records. Consistent with this, firms with university educated managers were more likely have audited accounts than firms with managers with only a secondary education. Whereas only 25 percent of firms with a manager with a secondary education or less had audited accounts, about 44 percent of firms with a manager with a vocational degree had audited accounts and 49 percent of firms with a manager with a university degree did. The econometric results suggest that firms with better educated managers have easier access to credit. After controlling for other differences between firms, firms with university educated managers were 11 percentage points more likely to have a bank loan, 14 percentage points more likely to have an overdraft facility, the average interest rate was 3 percentage points lower and the average loan duration was 33 months longer than firms with a manager with a secondary education or less. Firms with managers with vocational educations also appear to have better access to credit than firms with managers with a secondary education or less. They were 21 percentage points more 48 See Clarke and others (2003) for a summary of the literature on this topic 49 The coefficients on these variables were not statistically significant at even a 20 percent level. 84 likely to have loans, 14 percentage points more likely to have overdraft facilities, the average interest rate was 2 percentage points lower and the average loan duration was 20 months longer. Sector of operations. After controlling for size, location, and other firm characteristics, there was little evidence that access to credit varied by sector. The only exception was in the that firms in the retail trade and services sectors were 18 and 22 percentage points less likely to say that they were credit constrained than firms in the manufacturing and construction sectors. Since these firms were no more likely to say that they had loans or overdraft facilities than other firms, this suggests that this is because they are less likely to demand credit. Consistent with this finding, firms in the manufacturing sector also report that they are more concerned about access to credit (see Chapter 3), with close to 50 percent saying that they are concerned about access to finance compared to only about 30 percent of firms in retail and other services. Region. Despite the fact that managers of firms in Mindelo appear to be more concerned about access to finance than firms in Praia (see Chapter 3), the objective data do not suggest that firms in Praia have better access to credit than firms in Mindelo. IV. ACCESS TO FINANCE FOR MICROENTERPRISES As in many countries, the econometric analysis suggests that access to finance is more difficult for smaller firms. Larger firms were more likely to have loans, more likely to have overdraft facilities, pay lower interest rates and are less likely to say that they are credit constrained. There are many reasons why this might be the case. For example, small firms are less likely to keep detailed, audited accounts. In addition, they might be less likely to have managers with the skills needed to write detailed business plans and be less likely to be able to afford the fixed costs of hiring consultants to do so for them. Managers of small firms might also be less likely to have connections with bank officials. Finally, poor credit information and the fixed costs associated with making loans makes lending to small businesses both relatively expensive and relatively risky for banks discouraging them from entering this market. Since access to credit is difficult for microenterprises in many countries, it is important to assess whether small firms appear to be particularly disadvantaged in this respect in Cape Verde. To see whether this appears to be the case, Figure 38 shows the percent of short-term assets that microenterprises and larger firms financed with bank financing in several countries where both formal and informal surveys have been completed.50 This measure is used for the comparisons because comparable data is not available for many of the other measures of access to finance (e.g., whether firms have loans from banks). 50 Informal surveys have not been completed in any of our standard comparator countries except South Africa, Senegal and Indonesia. 85 Microenterprises do not rely heavily upon the banking sector to finance their working capital needs anywhere in the world (see Figure 38). Although the average microenterprise in Cape Verde financed only about 8 percent of its short-term assets with bank financing—more than in any of the comparator countries, this was higher than in any of the comparator countries. The average microenterprise relied upon banks for less than 3 percent of their working capital needs in all of the middle income comparator countries (Brazil, Indonesia, and South Africa) and upon them for less than 6 percent in the three other countries in Sub-Saharan Africa. Figure 38: Microenterprises in Cape Verde are relatively less disadvantaged with respect to access to finance than in other countries Microenterprises Larger Firms Brazil Brazil Indonesia Indonesia Guatemala Guatemala Bangladesh Bangladesh Kenya Kenya Senegal Senegal Tanzania Tanzania South Africa South Africa Cape Verde Cape Verde 0 20 40 0 20 40 % of short-term assets financed through banks % of short-term assets financed through banks Source: Investment Climate Surveys Note: International comparisons are for manufacturing firms only This does not appear to be simply due to Cape Verde having a better developed financial market than in the other countries. Larger firms in Cape Verde finance less of their short-term assets with bank financing than in any of these countries. The average large manufacturing enterprise in Cape Verde financed more of its short-term assets than the average microenterprise in the manufacturing sector in Cape Verde (11 percent compared to 8 percent). However, this remained less than for large manufacturing firms in other countries. In all of the comparator countries, the average firm financed over twelve percent of their working capital needs through the banking sector and in three of the countries, Bangladesh, Brazil and Kenya, they financed more than one-quarter of their needs in this way. This suggests that although access to finance is more difficult for small firms in Cape Verde, it is not as difficult as in some other countries. Overall, this does not provide strong support for the hypothesis that small enterprises are especially constrained due to the characteristics of the financial sector (i.e., highly concentrated and heavily reliant upon foreign bank). 86 Figure 39 Microenterprises rely more heavily upon retained earnings and less heavily on formal financing. % of short-term assets 100% 75% 50% 25% 0% Micro Medium Small Internal Funds Banks Informal Sources Other Formal Sources Source: Investment Climate Surveys and World Bank (2006c) Note: Inflation rates are averages between 2002 and 2004. Although this suggests that microenterprises in Cape Verde are not especially disadvantaged when compared to microenterprises in other countries, access to credit does appear to be more difficult for the average microenterprise than for the average small or medium-sized enterprise in Cape Verde (see Figure 40).51 Microenterprises are less likely to have loans and overdraft facilities, finance less of their short-term assets with bank financing and are more likely to say that they are credit constrained (i.e., that they would like a loan at current interest rates but are unable to get one). Although access to credit is harder for registered microenterprises than it is for larger SMEs, access to credit is especially hard for unregistered microenterprises (see Figure 40). Whereas only 4 percent of unregistered microenterprises had loans or overdraft facilities, 28 percent of registered microenterprises had a loan and 5 percent had an overdraft facility. Although the difference between unregistered and registered firms might explain why microenterprises appear relatively disadvantaged, it is important to note that registration is a choice that firms make. This affects the interpretation of the results. In particular, it suggests that increasing registration might not increase access to credit as much as the raw numbers might suggest. Since access to credit is seen as one of the main benefits for registration—84 percent of microenterprise managers said that registration made access to finance better—those microenterprises that think they have the best chance of getting access to finance might be the most likely to register. If an enterprise owner knows that he is unlikely to get credit even if he does register—for example if he knows he has little collateral or an impaired credit record—then he might decide not to register. As a result, we are likely to overestimate the importance of registration 51 Note that Figure 40 is for ALL microenterprises while Figure 38 is only for manufacturing enterprises. 87 since firms that know that they will not qualify for financing might decide not to register for this reason. Figure 40: Registered microenterprises have better access to finance than unregistered microenterprises – but worse access than SMEs. 75% 50% % of firms 25% 0% % with bank loan % with overdraft % of short-term % credit constrained facility assets financed through banks Unregistered micro Registered micro Small and medium Source: Microenterprise Investment Climate Survey Note: Includes both service and manufacturing firms for all categories 88 CHAPTER 6: OTHER ASPECTS OF THE INVESMENT CLIMATE The previous two chapters discussed access to finance and the skills and availability of labor. This chapter discusses other issues related to the investment climate including infrastructure, regulation, taxation, governance and macroeconomic stability. Rather than focusing on just firm perceptions, this chapter looks at objective measures of the investment climate such as power outages, tax rates, and the time spent dealing with government regulations. The chapter first discusses several areas of the investment climate that firms identified as serious constraints on their operations and growth: the performance of the power sector, taxation and competition from the informal sector. Along with financing, which was discussed in the previous chapter, firms were more likely to say that these areas of the investment climate were serious problems and were most likely to identify them as the biggest investment climate related constraint. In addition to these areas of the investment climate, the chapter also looks at the burden of regulation—an issue identified as a serious constraint in other studies. After this, the chapter looks at several areas of the investment climate that firms did not generally identify as serious constraints, including corruption, macroeconomic instability and telecommunications. This provides a useful check on firms‘ perceptions. I. ELECTRICITY The quality of a country‘s infrastructure can have a large impact on enterprise performance. In addition to the direct costs of dealing with poor infrastructure—purchasing generators and expensive fuel to cope with frequent power outages—poor infrastructure can increase indirect costs and drive down productivity. Finally, if firms are especially vulnerable to the supply of power, water and telecommunications, they might adopt less advanced production methods to get around the variation in quality and supply. This in turn may also affect their productivity. Across all sectors and for all types of firms, firms in Cape Verde saw the poor performance of the power sector as the greatest constraint on their operations and growth. Over 60 percent of firms said that the power sector was a major or very severe obstacle and about 40 percent said it was the greatest investment climate-related obstacle they faced. Moreover, complaints about sector performance were not limited to only a small segment of the market. Electricity was the greatest concern for large firms in both services and manufacturing, for hotels, and for microenterprises. Similarly, it ranked as the top constraint for both foreign and domestic enterprises. I.1 Electricity and Water Sector Reform A single company, Empresa Pública de Electricidade e Agua (Electra), is the sole provider both electricity and water and sewerage in both Praia and Mindelo, the two cities covered in this survey.52 In the electricity sector, independent power producers are allowed to enter the generation market and build local distribution networks. In practice, however, no independent power 52 This description of the sector draw heavily on ―Infrastructure Reform, Regulation, and Competitiveness in Cape Verde‖ (World Bank, 2006b) 89 producers had entered the market as of mid-2006 and, as a result, Electra operates as a de facto monopoly in both generation and distribution. In the water and sewerage sector, although Electra has a monopoly over most of the market, municipalities are responsible for service provision outside of Praia, Mindelo and the islands of Sal and Boa vista. The state-owned electricity and water company, a majority share (51 percent) of Electra, was sold to a Portuguese consortium of Electricitade de Portugal (EDP) SA and IPE-Aguas de Portugal SGPS SA. All generation and distribution activities, including those that had been conducted by municipalities, were transferred to Electra under the concession contract. Relations between the Government of Cape Verde and Electra have not always been smooth. Although negotiations were started in 1999, they were suspended in 2000 before being completed in 2002. In 2001, the government froze electricity and water tariffs, creating a fuel-tariff deficit estimated at US$17.5 million, and causing power cuts in the capital. Electra‘s financial problems were further exacerbated by a growing amount of government and private sector arrears over this period. In 2003, the Government signed an agreement with Electra that was intended to improve sector performance. The memorandum allowed tariffs to increase automatically as fuel and other operating costs increased. Following this, the Government continued to negotiate with Electra to reach a detailed agreement on compensation for the tariff deficit that had accumulated before negotiations started. In June 2005, the World Bank mediated a deal in which the Government of Cape Verde paid Electra US$10 million in compensation. The concession contract between Electra and the Government was initially the main document that guided regulation of the private concessionaire. In late 2004, however, the Government set up the Economic Regulatory Agency (ARE) to regulate water and electricity tariffs. It was hoped that this would prevent future conflicts between the Government and Electra. In practice, however, relations between the two have remained tense (World Bank, 2006b). The change in ownership failed to alleviate supply problems. Although outages were reduced considerably in the early years of the contract—falling considerably between 1998 and 2002—as discussed below they remain problematic (World Bank, 2006b, p. 42). Losses in transmission and distribution were equal to about 18 percent in 2003—high compared to international standards (World Bank, 2006b) Cape Verde‘s geography has contributed to the relatively high cost of service in the archipelago. The electricity generation system is highly fragmented since it is split across the different islands. Accessibility to water is also limited—underground water resources are limited and the islands are in the semi-arid Sahel region. In Sub-Saharan Africa, only Djibouti has lower water resources per capita than Cape Verde. Increasingly, Cape Verde has had to rely upon desalination plants for water, which accounted for about 75 percent of production by 2003 (World Bank, 2006b) I.2 Evidence from the Investment Climate Survey. Despite the significant structural changes that have occurred in electricity sector in Cape Verde, firms in Cape Verde remain concerned about sector performance. As noted in Chapter 3, 90 over 60 percent of firm managers said that electricity was a major or very severe obstacle to their enterprises‘ operations and growth. Objective indicators of sector performance are consistent with the perception-based indicators. Despite recent improvements in terms of coverage and expansion of the electric network in Cape Verde, power outages remain common. The median enterprise in the manufacturing sector reported about 8 outages in 2004 (see Figure 2). The average number of outages was considerably higher (over 30) due to some firms reporting a very large number of outages. The average outage lasted six hours. When compared to low income countries in Sub-Saharan Africa, this is not especially high—for example, the median number of outages was 48 in Tanzania in 2003, 21 in Kenya in 2003 and 15 in Senegal in 2004 (Regional Program on Enterprise Development, 2004a; Regional Program on Enterprise Development, 2004b; Regional Program on Enterprise Development, 2005b). It is also lower than in Guyana, where the median firm reported 15 outages. However, it is significantly higher than in the other middle income comparator countries. In South Africa, Mauritius, Philippines, Indonesia and Maldives, the median manufacturing firm reported between 0 and 4 outages per year. Figure 41: Power outages in Cape Verde are more common and more costly than in the better performing middle-income economies Dominican Republic Guyana Philippines Senegal Senegal Mauritius Philippines Indonesia South Africa Mauritius Indonesia Guyana Maldives South Africa Cape Verde Cape Verde 0 10 20 30 40 50 0 5 10 15 Median number of power outages Median losses due to power outages Average number of power outages Average losses due to power outages Source: Investment Climate Surveys Note: Cross-country comparisons are only for manufacturing enterprises Losses due to outages are higher in Cape Verde than in any of the comparator countries except the Dominican Republic—which had a far greater number of outages than in any of the other comparator countries. For example, although the median firm in Cape Verde reported fewer outages than median firms in Senegal or Guyana, losses due to outages were greater. Although this might seem peculiar, there are several plausible explanations. First, firms are affected differently by 91 outages. For example, losses will be greater when outages result in equipment damage or when the enterprise is unable to make up for lost production by running extra shifts. Second, firms with generators are more able to cope with outages than other firms. As discussed below, few firms in Cape Verde have generators, potentially increasing the cost of outages to firms. The number of outages differs slightly by firm type—the median service firm reported slightly more outages than the median manufacturing firm (see Figure 42). The median service firm also reported greater losses due to outages. While the median manufacturing firm reported losses equal to 3 percent of sales, the median service firm reported losses of over 11 percent of sales. This occurs even though firms in the service sector are more likely to have generators than manufacturing firms—more than 45 percent of retail and other services firms in the sample have a generator, compared to only 35 percent of manufacturing firms. Figure 42: Firms in retail and other services sector face less outages, but lost more output, than manufacturing firms Microenterprises 5 Microenterprises 3 Retail and other Retail and other 8.8 11.1 services services Manufacturing 7.6 Manufacturing 3 0 5 10 0 5 10 15 Median number of power outages Median losses due to power outages Source: Investment Climate Surveys Considering the poor reliability of the power sector in Cape Verde, relatively few firms operate generators. Only 39 percent of manufacturing firms operate a generator in Cape Verde (see Figure 43). This is considerably lower than in the comparator countries where firms face a large number of outages. About 52 percent of firms in Senegal, 64 percent of firms in Guyana and 88 percent of firms in the Dominican Republic operate generators. In fact, the proportion of manufacturing firms in Cape Verde which operates a generator is comparable to Indonesia, Philippines, Mauritius and Maldives—countries that display a much more consistent electricity supply than Cape Verde. Not surprisingly, smaller firms are less likely to operate generators than larger firms are (see Figure 43). Medium-sized firms are more than twice as likely to operate generators as small firms are. Given the high fixed costs of purchasing and running a generator, this is not surprising. The 92 small size of firms in Cape Verde probably explains the relatively low use of generators in this country despite the unreliability of the power supply. Indeed, whereas 88 percent of medium-sized firms in Cape Verde have generators, only 39 percent of medium-sized firms in Indonesia, 32 percent in Mauritius, and 45 percent in Philippines operate generators. Domestically owned enterprises are less likely to have generators than foreign-own enterprises. One reason might be that foreign-owned firms‘ production process may rely more heavily on electricity, so they need generators more than domestic enterprises. However, foreign firms also tend to be slightly larger than domestic firms, potentially partially explaining the difference. Figure 43: Fewer firms have generators than in other selected countries Dominican Republic 88.0 Medium 87.9 Guyana 64.0 Small 41.6 Senegal 52.4 Indonesia 39.4 Foreign 71.6 Mauritius 39.3 Domestic 41.8 Maldives 39.2 Philippines 36.6 Hotels 81.1 Retail and South Africa 9.4 41.8 other services Manufacturing 39.1 Cape Verde 39.1 0 20 40 60 80 100 0 20 40 60 80 100 % of firms with generators % of firms with generators Source: Investment Climate Surveys Note: Cross-country comparisons are only for manufacturing enterprises In addition to being interested in the reliability of power, we are also interested in the cost of power. Firms in Cape Verde spend about the same on power as firms in most of the comparator countries. For example, median firms in Senegal, Indonesia and Mauritius all spend about 1.5 percent of sales on electricity. Electricity costs in Cape Verde are far below costs in Guyana and the Dominican Republic, where the median firms spend over five percent of sales on power. These countries also face the highest number of power outages and are more likely to own generators than firms in other countries. The median manufacturing firm spends more on power (1.4 percent of sales) than the median firm in the services sector (0.8 percent). This is far above what the median hotel spends— about 0.2 percent of total sales. Microenterprises appear to spend more on power than larger firms do—about 6 percent of sales compared to only 0.8 percent for small firms and 1.0 percent for medium-sized firms. 93 One interesting difference between Cape Verde and other countries where microenterprise surveys have been conducted is that microenterprises were more likely to say that they were connected to the power system than in other countries. All of the microenterprises covered in the microenterprise survey—all located in Mindelo and Praia—reported that they were connected to the power system. In comparison, only 80 percent of microenterprises in a similar survey in South Africa reported that they were connected to the power system (Regional Program on Enterprise Development, 2006). Figure 44: Firms in Cape Verde spend less on electricity than in the comparator countries Dominican Republic 8 Medium 1 Guyana 5.1 Small 0.8 Philippines 2.9 Microenterprises 6.6 Indonesia 1.6 Mauritius 1.4 Senegal 1.3 Hotels 0.2 South Africa 0.7 Retail and other 0.8 Maldives 0.4 services Manufacturing 1.4 Cape Verde 1.4 0 5 10 15 20 0 5 10 Median electricity costs as % of sales Median electricity costs as % of sales Source: Investment Climate Surveys Note: Cross-country comparisons are only for manufacturing enterprises I.3 Regional differences in the quality of electricity As noted in Chapter 3, firms in Praia were more likely to say that power was a major or very obstacle than firms in Mindelo. A natural question is whether the objective data suggest that the performance of the power sector is worse in Praia than in Mindelo. To see whether this is the case, this section presents several measures of sector performance for the two cities separately. The objective indicators are consistent with the measures of perceptions—power sector performance appears to be a greater concern for firms in Praia than for firms in Mindelo. Enterprises were more likely to say that they had to deal with power outages (100 percent in Praia compared to only 73 percent in Mindelo) and faced more outages per month (15 for the median firm in Praia compared to only 3 for the median firm in Mindelo). Firms in Praia were also more likely to have generators (67 percent compared to 18 percent)—consistent with the idea that they are more concerned about the performance of the power sector. Although the samples are relatively small when we compare the responses of manufacturing firms in Praia with manufacturing firms in the comparator countries, the results suggest that the 94 power sector in Praia performs at least as poorly as in the worst performing of the comparator countries—the median number of outages (20 for manufacturing firms in Praia) was higher than the median number in Guyana and Senegal (15).53 In contrast, the median number of outages reported by manufacturing firms in Mindelo (3) was comparable to the median number of outages in the Philippines (2), South Africa (4) or Mauritius (4). Figure 45: Firms in Praia are more likely to face power outages, face outages more often and are more likely to have generators. % with generator Ave # of outages % with outages Median # of % saying outages power major problem 0% 25% 50% 75% 100% 0 10 20 30 Mindelo Praia Mindelo Praia Source: Investment Climate Surveys II. TAXATION The performance of the power sector was the area of the investment climate that enterprises of all types were most concerned about. Overall, the area that enterprises were next most likely to say was a serious obstacle was tax rates. Close to 50 percent of the larger enterprises covered in the main sector said that tax rates were a major or very severe obstacle to enterprise operations and growth. Fewer microenterprises expressed concern—only about 40 percent of microenterprises said that tax rates were a severe obstacle. Not surprisingly, unregistered microenterprises were less likely to express concern than registered microenterprises (28 percent compared to 40 percent). Firms were generally far less likely to express concern about tax administration. Only about 22 percent of large firms and 16 percent of microenterprises rated tax administration as a major or very severe obstacle. This makes tax administration as the tenth largest constraint for large firms and the seventh largest constraint for microenterprises in Cape Verde. 53 Note that the results in Figure 45 include firms in the service sector, while the cross-country comparisons are only for manufacturing firms. 95 II.1 Tax Policy The Government of Cape Verde has improved revenue collection in recent years. The introduction of a single rate Value-Added Tax (VAT), set at 15 percent, in January 2004 has improved the revenue base. In June 2002, the GOCV simplified import duties; capital goods and raw materials now carry total customs duties of either 0 or 5 percent, down from a level of about 20 percent. The corporate income tax has been reduced from 35 percent to 30 percent of net profits. Tax administration and expenditure planning have also improved. Both the International Monetary Fund and the World Bank have supported reforms designed to improve the functioning of the tax system. In particular, the World Bank‘s 2003-2008 Growth and Competitiveness project includes a component to support tax reform (World Bank, 2003). Further improvements, however, are possible. Firms are experiencing delays getting VAT refunds. Administration and control systems are weak. Tax incentives are generous and complex and need to be rationalized— something to which the Government of Cape Verde is committed. II.1.1 Tax Rates About 55 percent of small and medium-sized enterprises in the services sector, 47 percent of in the manufacturing sector and 29 percent of hotels report that tax rates are a major or very severe constraint on their operations and growth. Microenterprise managers are also concerned—38 percent of microenterprises said the same. Despite the high levels of concern, marginal tax rates do not appear to be out-of-line with tax rates in other low and middle income countries. For example, the VAT rate (15 percent) is lower in Cape Verde than in Senegal (20 percent) and the Dominican Republic (16 percent), is comparable to Mauritius (15 percent), and is higher than in Indonesia (10 percent), the Philippines (10 percent), and South Africa (14 percent). Although VAT rates are not the lowest among the comparator countries, they are also not the highest. The top corporate tax rate is also not out of line with the comparator countries. The top tax rate (30 percent) appears to be slightly lower than in Senegal (33 percent), the Philippines (32 percent) and Mauritius (35 percent) but is the same as in South Africa and Indonesia (30 percent). 96 Figure 46: Tax rates are similar or lower in Cape Verde to the comparator countries. Dominican Republic Indonesia Indonesia Philippines Philippines Mauritius Guyana Dominican Mauritius Republic Maldives South Africa South Africa Senegal Senegal Cape Verde Cape Verde 0 10 20 30 40 50 0 5 10 15 20 VAT Rate Top Corporate Tax Rate Source: IMF-Letter of Intent of the government of Cape Verde; www.fita.org; www.ita.doc.gov; www.doingbusiness.org. Note: The corporate tax rate in Senegal will be reduced from 33 percent to 25 percent starting from January 2006. Given that rates do not appear to be seriously out-of-line with the comparator countries, what explains the high level of concern about tax rates in Cape Verde? One important point is that tax rates typically rank among enterprises‘ greatest concerns in investment climate assessments. According to the 2005 World Development Report, enterprise managers ranked tax rates among the top five obstacles in over 4 out 5 low income countries where Investment Climate Surveys had been completed (World Bank, 2004b). This is also true in the comparator countries—tax rates ranked in the top five concerns in four of the eight comparator countries (see Chapter 3). Moreover, although it did not rank in the top five constraints in the other four countries, about thirty percent of firms said it was a serious problem in two of the remaining four countries (the Maldives and Mauritius) and over 50 percent did in the Dominican Republic. Thus, although tax rates were not especially high compared to the comparator countries, firms in these countries also often saw tax rates as a major or very severe constraint. Another important point is that the marginal tax rates might not provide a very accurate picture of the overall burden of taxation in the comparator countries. In addition to the marginal tax rates, the burden of taxation is affected by many other things including depreciation rates, other taxes that firms face, and a wide range of fiscal incentives that firms can be eligible for. The World Bank‘s Doing Business 2006 report (World Bank, 2005b) provides a more detailed estimate of the overall tax burden computed for a standardized firm. Unfortunately, Cape Verde was not included in the Doing Business 2006 in 2005 and data for Doing Business 2007 were not yet available. Once data from Doing Business 2007 are available, they will provide a better international benchmark than the simple marginal rates do. Although the Investment Climate Survey might suggest that tax rates might be a problem in Cape Verde, it is important to keep in mind that although taxes represent a cost to firms and reduce their incentives to invest and create jobs, governments need revenue to cover the cost of providing 97 public services—including those that improve the functioning of the investment climate. All countries struggle with how to strike a balance between these priorities in an efficient and equitable way. The high level of concern does suggest that some additional analysis might be useful. In particular, FIAS produces reports on the effective tax burden that look at the overall burden of taxation in far more detail than would be available in a standard report such as Doing Business 2006. Given the level of concern, it might be worth looking at this question in greater detail. II.1.2 Tax Administration Only 22 percent of enterprise managers say that tax administration is a serious obstacle to enterprise operations and growth. As in the case of tax rates, domestic enterprises are more likely to say that tax administration is a serious problem than foreign-owned enterprises—about 23 percent of domestic firms reports that tax administration is a serious problem compared to only 8 percent of foreign enterprises. Hotels are less concerned about tax administration than other enterprises—9 percent of them rates it a serious obstacle, compared to 24 percent of firms in retail and other services and 22 percent of firms in manufacturing sector. Objective indicators support the perception that tax administration is relatively less burdensome in Cape Verde than in comparable countries. In addition to the perception-based measures, firms were also asked about the number of tax inspections that they faced during the previous year. The average manufacturing firm in Cape Verde reports less than one inspection. In comparison, the average manufacturing enterprise in Mauritius and Indonesia reports 2, the average manufacturing firm in Philippines reports 4, the average manufacturing firm in South Africa reports 4 and the average manufacturing firm in Senegal reports 7.7 inspections. Medium enterprises and hotels face a slightly higher number of inspections and required meetings than other enterprises (see Figure 47) but the burden is still not significant. For example, the average manufacturing enterprise, the average microenterprise or the average enterprise in retail and other services reports less than one inspection or required meeting while the average hotel has about 1.4 inspections in a year. Similar patterns can be observed in other countries, although the increase in number of inspections is usually much greater. For example, the average small enterprise in Senegal had about four inspections or required meetings, while the average large firm had more than 19. Concentrating administrative effort on large and profitable firms generally makes sense from a revenue-maximizing approach, but can end up imposing a large burden on a few enterprises. 98 Figure 47: Large firms, foreign-owned firms and hotels have more required meetings with tax inspectors than other firms. 2 1 0 el s ro g m n l ic al ce in ig t ot iu ic Sm es ur re vi H M ed om Fo ct er M fa rs D u he an ot M d lan ai et R Average number of tax inspections Source: Investment Climate Survey. Tax compliance generally appears good in Cape Verde. Firms report that they believe that ―firms like theirs‖ generally report most of their revenues to the tax authorities. On average, manufacturing enterprise managers say that firms like theirs report about 87 percent of revenues to the tax authorities. Although lower than in South Africa (about 90 percent) and Mauritius (87.8 percent), this is considerably higher than in Senegal (less than 20 percent) and above other selected countries. Figure 48: Firms report a greater share of revenues in Cape Verde than in most of the other countries. 100 80 60 40 20 0 a es e al s ia ca lic iu an rd eg es ub n ri rit Ve pi Af uy n n ep au ilip do Se G th e R M ap In Ph u n So C a ic in om D Percent of revenues reported to tax authorities Source: Investment Climate Survey. Note: Firms are manufacturing firms only. 99 III. INFORMALITY The area of the investment climate that firms were the fourth most likely to say was a major or very severe obstacle to their operations and growth is competition with the informal sector. When formal firms find themselves competing with informal enterprises, they can often find themselves at a competitive disadvantage. Informal firms avoid much of the burden of taxation and regulation, reducing their costs relative to formal firms. On the other hand, informal firms have some disadvantages. They will find it more difficult to access financing and will be less able to turn to the courts for support when they are in disputes with suppliers or customers. III.1 Informal Sector in Cape Verde Recent estimates suggest that the informal sector accounts for about 40 percent of employment in Cape Verde – mostly in small, family-owned firms (Ministry of Finance and Planning, 2004). Although this is lower than—or similar to—other countries in Sub-Saharan Africa, it is slightly higher than in lower middle-income countries in East Asia (see Figure 49). Figure 49: Although informality in Cape Verde is about the same as in other lower middle-income countries and is lower than in other countries in Sub-Saharan Africa, the informal sector remains an important source of employment. 60% % of Employment 40% 20% 0% e ya i re a n s an rd ne si oo oi n ne nz Ve Ke pi Iv er do Ta ilip d' am e ap e In Ph C ot C C % of employment Source: Ministry of Finance and Planning (2004); Ayyagari and others (2003) Note: Comparable data were not available for most of the comparator countries. Evidence from the survey of microenterprises also suggests that informality is not exceptionally high in Cape Verde. As part of the microenterprise survey, microenterprises were asked if they were registered with central or local government agencies. About three-quarters of the informal firms in the survey reported that they were registered with one or more government agencies. Although this might overestimate the extent of formality among microenterprises — unregistered firms might be reluctant to admit that they are even when they are confident that their responses will remain confidential—the firm conducting the investment climate survey reported that 100 most of the microenterprises, which were identified using area sampling, were on the lists that were obtained from the Bureau of Statistics.54 One other possible reason for the high level of concern about informality might be the small size of the firms in the sample. Smaller firms might be more likely to find themselves competing with microenterprises than larger firms. Consistent with this microenterprises—and especially registered microenterprises—were especially concerned about informality. About 59 percent of registered microenterprises said that informality was a major or very obstacle, compared to 40 percent of very small formal enterprises, 36 percent of small formal enterprises and 21 percent of medium-sized formal enterprises. The concern about informality, therefore, might primarily reflect that most formal enterprises in Cape Verde are relatively small. III.2 Barriers to Becoming Formal Although the concern about informality appears to be reflect the small size of many firms in Cape Verde, this does not lessen the importance of the issue. The small size of the domestic market, combined with the difficulty that the small firms in Cape Verde are likely to have entering export markets (e.g., due to the high cost of labor, Cape Verde‘s physical isolation, and other issues), means that many firms are likely to remain small in the near term. Reducing unfair competition from informal and unregistered firms is therefore an important goal. Figure 50: The financial cost of taxation and registering are the main reasons microenterprises don’t register—but problems related to getting information on registering are also important. 50% 25% 0% s n ts ts n n g er te tio tio io in en en st ra er at a tra gi m m ul rm st x re re re is eg Ta i eg fo in i ui to qu rR in m fr eq e re ad ng to m bo lr y Ti ita os ti La x or et Ta ap C at G ul C eg R All micro Registered Unregistered Source: Investment Climate Survey. In the microenterprise survey, registered and unregistered microenterprises were asked what they say as the main obstacle to becoming formal. Microenterprises were most likely to say that the 54 Note that this does not necessarily imply that the firms are fully formal (i.e., that they comply with all regulations or that they pay taxes). It just means that they were included in the industrial census conducted by the Bureau of Statistics. 101 financial costs of paying taxes, the financial costs of registration and the difficulty of getting information on registration requirements and procedures were the major or very severe obstacles to registration. Over one quarter of microenterprises said that capital requirements and tax administration were major or very severe obstacles. In contrast, microenterprises were considerably less likely to be concerned about the time it takes to register a firm, labor regulations, or other regulatory burdens (e.g., inspections). Registered microenterprises had a slightly different perspective from unregistered microenterprises. Registered microenterprises were more likely to stress the financial costs of registration (cost of registration and capital requirements) and the burden of taxation (both financial and administrative). They were far less likely to see difficulties related to getting information on how to register as a serious problem than unregistered enterprises. Increasing the ease of getting access to the information needed to register should therefore be an important step towards reducing information. A recent initiative to bring the informal sector into the formal sector, driven by the Chambers of Commerce, should be helpful in this respect. To help accomplish this, the initiative will help informal and unregistered microenterprises get more organized to better manage their resources and help them set up accounting systems adapted to their size. IV. REGULATION For the most part, the larger firms in the formal firm survey in Cape Verde were not particularly concerned about the burden of regulation. Of the 17 areas of the investment climate, the area of regulation that was of the greatest concern was trade regulation—although even this ranked only 7th among the 17 potential obstacles. Tax administration was next (10th), followed by courts (11th), labor regulation (13th) and business licensing (14th). Microenterprises were generally even less concerned. The area of greatest concern, tax administration, was ranked as a serious problem by only 16 percent of firms. Business licensing was ranked as a serious problem by 15 percent, while all other areas were ranked as a serious problem by 10 percent or fewer of the microenterprises. IV.1 Regulatory reform in Cape Verde The low level of concern appears broadly consistent with other studies. In a 2002 survey of 120 companies in Cape Verde, conducted by the International Finance Corporation‘s Foreign Investment Advisory Service (FIAS), the companies were less concerned about administrative barriers than companies in many African countries where FIAS has conducted the same survey. Less than 20 percent of the surveyed firms reported being unsatisfied with the quality of regulatory procedures or their relations with the administration, and only 3 percent claimed that they were very unsatisfied. Although the FIAS survey suggested that firms were not especially worried about regulatory procedures, the analysis suggested some areas where procedures could be streamlined. In particular, the analysis showed that it takes between 96 and 200 days for a foreign investor to be established in the industrial sector (see Table 13). 102 Table 13: Number of Days to Establish a Business, 2002, FIAS Number of days Procedure Minimum Maximum Eligibility to foreign investment status 30 60 Business creation and registration 8 15 Tax registration 2 5 Registration with labor authorities 5 10 Industrial Status 6 30 Access to land 15 30 Building permit 30 50 Total 96 200 Source: FIAS (2003). Based on the results of the FIAS survey, the study identified three high priority areas: 1. The special procedures required to gain access to sectoral incentives. 2. Land acquisition and registration, due to the rigidity of legislation, the absence of a land register and controversies surrounding past expropriations. 3. Operating formalities related to foreign trade, namely import licences. In addition to the previous concerns, one of the main conclusions of the FIAS report was that the officials enforcing regulations regarding the private sector did not have the right mindset regarding the importance of the private sector. Because of this, the report concluded that technical solutions alone would not fix the system and that problems would emerge if perceptions about the private sector remained unchanged. Since the time, the Government and other organizations have taken several steps to address these issues. The first high priority area identified above—fiscal incentives—was the focus of a recent study conducted by the International Monetary Fund‘s Fiscal Affairs Department. The initial ―Study on the Economic Impact of the Fiscal Incentives‖ and recommended that th e Government recruit a local consultant to compile legislation relating to the incentive structure. Once the Final Report is received, a workshop will be held to validate the findings and a legal consultant will be recruited to prepare a draft ―single fiscal incentives law.‖ With respect to the third priority, the World Bank‘s Growth and Competitiveness project has supported legislation to simplify the import process (see Chapter 2). In addition to these actions, the Government of Cape Verde has taken additional steps in its efforts to reduce the burden of regulation. On February 2, 2005 the Government created the ―Grupo de Trabalho para a Reduçao das Barreiras Administrativas ao Investimento (GTRBAI, the Working Group for the Reduction of Administrative Barriers to Investment). This group includes representatives from all line ministries, Banco de Cabo Verde, Cabo Verde Investimentos, ENAPOR, (Ports Authority) the Chambers of Commerce and the World Bank. The Prime Minister leads GTRBAI and meets on a weekly basis to ensure that the project advances. This group has been merged with a task force on judicial reform. The Government has also taken steps to reduce the time it takes to register a business. Núcleo Operacional para a Sociedade de Informação (NOSI or Operational Information Society Nucleus) was created in January 2004 to promote and implement the policy measures for the Information and 103 Communication Technology sector. This organization has been charged with implementing an ambitious E-government project, which will simplify the interaction between government and citizens. In addition to addressing the citizens‘ needs throughout their life (life cycle events), the project targets the needs of businesses throughout their functional life (idea, creation, registration, operation, termination). The end result will be much more efficient and user friendly services. The first phase of the program is almost ready and ―live data‖ is being used to test the system. It is expected that one will be able to license a company in 24 to 48 hours by the first quarter of 2006. In addition, the ex-post inspection process for licensing has been changed. Previously, one could not obtain a license until the company was inspected. Now, the license can be issued immediately but is conditional on the inspection. This ensures that any environmental and safety issues are addressed adequately. In the area of legal reform, the Ministry of Justice and the Chambers of Commerce have successful completed their initiative to bring arbitration to Cape Verde. A new arbitration law and accompanying regulations have been approved, three workshops have been carried out to discuss the organizational structures, rules and procedures and the physical premises are being equipped during the first quarter of 2006. IV.2 The Burden of Regulation In addition to presenting evidence on perceptions about the burden of regulation, firms are also asked some quantitative questions on the time they spend dealing with government regulations and the number of inspections that they have. The first measure is a broader measure of the burden of regulation, since it includes things such as the time it takes to fill out required forms (e.g., tax forms), the time it takes to comply with regulatory demands (e.g., ensuring that the production process meets required standards), as well as the time spent meeting with inspectors. Figure 51: The burden of regulation is costly in terms of time in Cape Verde compared to other countries Indonesia 1 Maldives 2 Maldives 2 Guyana 3 Philippines 7 Indonesia 6 Mauritius 8 Philippines 9 Guyana 11 Mauritius 11 Dominican Republic 35 Dominican Republic 12 South Africa 14 South Africa 10 Senegal 19 Senegal 15 Cape Verde 2 Cape Verde 12 0 10 20 30 40 0 10 20 30 40 % of management time dealing with government regulations Average number of inspections per year Source: Investment Climate Survey Note: International comparisons are for manufacturing enterprises only to ensure comparability across countries. 104 On this broader measure, the burden of regulation appears relatively high in Cape Verde— although not completely out-of-line with the comparator countries. On average, manufacturing enterprise managers spend a little more time dealing with government regulations in Cape Verde than in any of the comparator countries except for Senegal (see Figure 51). The average manufacturing enterprise manager reports spending about 12 percent of his time dealing with regulations and inspections, compared to less than 10 percent in South Africa, Philippines, Guyana, Maldives or Indonesia. However, a different pattern can be observed with respect to inspections. On the narrower—although more easily measured—indicator, the burden of regulation seems more modest in Cape Verde. On average, managers of manufacturing enterprise managers report fewer inspections in Cape Verde than managers in any of the comparator countries except for Indonesia and Maldives. Whereas the average manager reports only 2 inspections in Cape Verde, the average manager in South Africa report 14 and the average manager in the Dominical Republic reports over 30 inspections. The burden of regulation varies somewhat across different types of firms (see Figure 52). In most countries, the burden of regulation is greater for large firms than it is for small firms. This pattern is less evident in Cape Verde than in other countries. Managers of microenterprises report spending only about 5 percent of their time dealing with government regulations, less than larger enterprises. However, on average, managers of medium-sized firms spend about 13 percent of their time dealing with government regulations, slightly less than small firms (17 percent). In contrast to the indicator of time spent dealing with regulations, the evidence on inspections is more consistent with evidence from other countries. The average microenterprise has less than one required meeting or inspection per year, the average small enterprise has about 2 required meetings or inspections per year, while the average medium-sized firm faces about 8 required meetings and inspections in a year. Among microenterprises, the burden of regulation is greater for registered microenterprises. Whereas the average manager of registered microenterprise faces about 1.2 inspections per year and spends nearly 7 percent of their time dealing with government regulations, the average manager of an unregistered microenterprise has only 0.5 inspections per year and spends about 1 percent of their time dealing with government regulations. The ongoing burden of registration is thus not negligible for microenterprises in Cape Verde. The burden is also heavier on hotels than retail and other services or manufacturing firms. Managers of hotels spend more time dealing with government regulations and face more frequent required meeting and inspections than managers of other firms. Managers of domestically owned firms spend more time dealing with government regulations than managers employed in foreign- owned firms but face fewer inspections in a year than their colleagues in foreign-owned enterprises. This may be due to the fact that foreign-owned enterprise managers—like medium firm managers— are more trained to cope with the burden of regulation and are able to minimize the time spent dealing with it. 105 Figure 52: Hotels face a greater regulatory burden than other firms. Domestic Domestic Foreign Foreign Medium Medium Small Small Microenterprises Microenterprises Hotels Hotels Retail and other Retail and other services services Manufacturing Manufacturing 0 2 4 6 8 0 10 20 30 Time spent dealing with regulations Number of inspections per year Source: Investment Climate Survey Most required meetings and inspections are with tax inspectors (see Table 13). About 89 percent of hotels have at least one required meeting with tax inspectors. Moreover, when hotels are inspected they face a higher number of inspections than firms in other sectors. Firms in the retail and other services sector are less likely to have required meetings with tax inspectors than manufacturing firms and hotels. About 79 percent of hotels and 55 percent of manufacturing firms also meet with labor inspectors. This is less common in the retail and other services sector. Finally, about 70 percent of hotels face sanitation or epidemiology inspections. This is the most frequent type of inspection for hotels—about two in a year. This is a good signal for the tourist industry if it ensures hotel quality. Consistent with other evidence that corruption is low in Cape Verde, very few firms report that informal payments are expected or requested during labor or sanitation inspections. Although still relatively uncommon, bribe requests were more common during tax inspections than other types of inspection—almost 9 percent of firms reported that informal payments are expected or requested during tax inspections. 106 Table 14: Inspections and bribe requests by type of inspector. Percent of firms Average number of inspections for firms Percent of firms reporting inspections reporting bribes that were inspected requested Manufacturing Services Hotels Manufacturing Services Hotels Total Tax Inspectorate 67% 37% 89% 0,9 0,7 1,4 9% Labor and Social Security 55% 23% 79% 0,8 0,4 1,2 2% Sanitation/Epidemiology --- 32,7% 69,6% --- 0,5 1,9 5% IV.3 Regulatory Uncertainty. Only 36 percent of manufacturing firms‘ managers disagree with the statement that government officials‘ interpretations of regulations affecting their establishment are consistent and predictable (Figure 53). This is far below that observed in Dominican Republic (65 percent), Indonesia (56 percent), Guyana (53 percent), Philippines (49 percent) and Senegal (46 percent) and just above what is reported for Mauritius (33 percent). In fact, concerning managers‘ faith in regulation, Cape Verde looks very similar to South Africa (38 percent) and Maldives (37 percent). Figure 53: Firms in Cape Verde have more faith in regulation than firms in other countries Dominican Republic Domestic Indonesia Foreign Guyana Philippines Maldives Medium Mauritius Small Senegal Hotels South Africa Retail and other services Cape Verde Manufacturing 0 20 40 60 80 Microenterprises 0 20 40 60 % of firms believing that regulations are % of firms believing that regulations are inconsistently or unpredictably applied inconsistently or unpredictably applied Source: Investment Climate Survey The types of firms which face the lowest number of inspections—namely firms in retail and other services sector and small enterprises—are also the ones which believe the most that regulations are inconsistently or unpredictably applied. However, a significant fraction of foreign- owed firms seem to be affected by the inconsistency of regulations as they face a larger number of inspections than domestic firms. Hence, the number of inspections and the managers‘ faith concerning regulation do not appear to be strongly correlated. Managers‘ opinion about officials‘ interpretations of regulations seems to be much more linked with the quality of inspections rather than the quantity. 107 V. OTHER AREAS OF THE INVESTMENT CLIMATE Firms in Cape Verde were considerably less concerned about other areas of the investment climate, such as corruption, macroeconomic instability, the performance of the telecommunications sector. In this section of the report, we discuss several of these areas to see whether the objective indicators appear consistent with the measures of perception and to see if the objective data provides any other interesting points of comparison between firms in Cape Verde and the other comparator countries or between different groups of firms within Cape Verde. V.1 Corruption and other measures of governance Only about 15 percent of small and medium-sized firms and 12 percent of microenterprises in Cape Verde said that they saw corruption as a major or very severe obstacle to their operations and growth. This is not because firms are never concerned about corruption. In four of eight comparator countries, firms rated corruption among the top five problems. This section looks at other evidence on corruption—both results from previous studies and the objective indicators from the investment climate survey to check that the results are consistent with the perception-based indices. V.1.1 Evidence from Previous Studies Previous studies have also found that corruption is relatively low in Cape Verde. Aggregating data from a number of empirical studies, Kauffman, Kraay, and Mastruzzi (2005) ranked Cape Verde at about the 66th percentile for corruption (see Figure 54). This suggests that Cape Verde outperforms both other countries in Sub-Saharan Africa (regional average of 30th percentile) and lower middle income countries (average of the 39th percentile). Consistent with this, Cape Verde has also performed well in other studies that have looked at corruption. For example, it did well in the 2004 Afrobarometer survey.55 Corruption is not the only area related to governance where Cape Verde performs well. It also ranks relatively well in all of the other six governance indicators compiled by Kauffman, Kraay and Mastruzzi (2005). These measures, which are described in detail in Box 3.1, include voice and accountability, political stability, government effectiveness, regulatory quality and the rule of low. Even its lowest ranking, government effectiveness, was still in the 50th percentile and well above both the averages for the region (28th percentile) and its income category (40th percentile). 55 Cape Verde is not included in Transparency International‘s most recent Corruption Perceptions Index because too few surveys were available that included it (Transparency International, 2005). 108 Figure 54: Cape Verde performs well on most measures of governance including corruption Source: Kauffmann, Kraay and Mastruzzi (2005) Note: Inner dotted line is average for Sub-Saharan Africa. Outer line is values for Cape Verde. Percentile rank indicates the percentage of countries worldwide that rate below the selected country (subject to margin of error). Higher values imply better governance ratings. In addition to performing well relative to countries in the same region and income levels, Cape Verde also performs well relative to most of the comparator countries used in this study. Although it ranks slightly below the best performing comparator countries, such as South Africa and Mauritius, the difference is not statistically significant. It ranks above most of the middle income comparators in the Caribbean and East Asia (see Table 15). For example, the Dominican Republic ranked in the 41st percentile, Guyana ranked in the 45th percentile, and the Maldives ranked in the 60th percentile. Cape Verde has several laws and regulations in place to fight corruption. Giving or accepting a bribe is a criminal act for which conviction could result in eight years imprisonment. Cape Verde has also established the High Authority against Corruption that, along with the Judiciary Police, specifically targets and combats corruption.56 56 Strategis, Industry Canada. Cape Verde Country Commercial Guide FY 2003: Invest Climate. http://strategis.ic.gc.ca/epic/internet/inimr-ri.nsf/en/gr108886e.html. 109 Table 15: Corruption is low relative to most of the middle income comparator countries Percentile Rank Cape Verde 66.5 Senegal 43.3 Mauritius 67.0 South Africa 70.9 Dominican Republic 41.4 Guyana 44.8 Indonesia 17.7 Philippines 36.5 Maldives 60.6 Source: Kaufmann, Kraay, and Mastruzzi (2005) Box 3.1: Governance Indicators In recent years, many researchers and practitioners have tried to produce aggregate statistics that can be used to compare the quality of governance across countries and for single countries over time. Few of these studies cover the entire world or all topics. Furthermore, although the studies often cover similar topics, responses and questions are usually not comparable across surveys. In order to increase country coverage, Kaufmann and others (2002; 2003; 2005) combined information from as many as 60 mostly subjective indices from other sources to produce six measures that capture different aspects of regulation, corruption, and governance. The six measures are: Voice and Accountability: the extent to which citizens of the country are able to participate in the selection of the government. Political Stability: the likelihood that the government will be destabilized or overthrown by possibly unconstitutional and/or violent means, including terrorism. Government Effectiveness: the quality of public service provision and the government bureaucracy, the competence and independence of the civil service, and the credibility of the go vernment‘s commitment to adhering to announced policies. This measure primarily focuses on ―inputs‖ that governments need to implement good policies and deliver public goods. Regulatory Quality: the quality of government policies. This measure is ―output‖ rather than ―input‖ based, in that it focuses on the prevalence of market-unfriendly policies such as price controls or inadequate bank supervision, as well as perceptions about the burden imposed on businesses by regulation. Rule of Law: the extent to which individuals have confidence in and abide by the rules of society. This includes perceptions about the incidence of crime (both violent and non-violent), the effectiveness and predictability of the judiciary, and the enforceability or contracts. Control of Corruption: the extent of corruption (i.e. the illegal use of public power for private gain). 110 V.1.2 Objective indicators from the Investment Climate Survey The objective indicators on corruption from the Investment Climate Survey are consistent with both firm perceptions and with previous studies that have looked at corruption in Cape Verde—corruption does not appear to be a serious constraint on firm performance in Cape Verde. The most direct measure of corruption in the Investment Climate Survey is a question that asks whether ‗gifts or informal payments‖ are needed to ―get things done with regards to customs taxes, licenses, and regulations, services, etc.‖ for a typical firm in the industry. The question is asked indirectly—rather than directly—so that firms can answer without implicating themselves. Only about 2 percent of manufacturing firms said that firms like theirs typically had to pay bribes to get things done. This is lower than any of the comparator countries, although it is very close to the number for South Africa (see Figure 2). In the other comparator countries at least 20 percent of firms reported that bribes are needed to ―get things done‖. Figure 55: Corruption is quite low for formal firms in Cape Verde. Medium Maldives Philippines Small Indonesia Very small Dominican Republic Microenterprises Senegal Guyana Hotels South Africa Manufacturing Services Cape Verde 0% 20% 40% 60% 0% 20% 40% % of firms reporting bribes % of firms saying bribes are needed Source: Investment Climate Surveys Note: Cross-country comparisons are only for manufacturing enterprises Even though Cape Verde has the lowest percentage of bribes, it is important to note that when a bribe is required, bribes tended to be relatively high relative to sales. The median bribe for firms that reported bribes in Cape Verde was 3 percent of their sales. South Africa follows a similar pattern in that it has a low percentage of firms reporting bribes, but a relatively high median bribe for those firms that do report them. This could reflect the low number of observation resulting in unstable estimates. Another explanation for the negative correlation between percentage of firms reporting bribes and the median bribe of these firms who report being bribed is that bribes are less common in these countries because of stricter laws and better enforcement, so that when they do occur, corrupt officials have to compensate for the risk involved with larger bribes. 111 In addition to the cross country comparisons, it is possible to break down bribes for different types of firms. Bribes were relatively uncommon in all sectors of the economy for small and medium-sized firms—less than 10 percent of firms in the manufacturing, service and hotel sectors reported that bribes were needed to get things done. Microenterprises were far more likely to report that bribes were needed than small or medium-sized enterprises. Close to one-third of microenterprises reported that bribes were needed to ―get things done‖. One possible explanation for this might be that informal and unregistered firms have to pay bribes to avoid any legal problems and restrictions. If this were the case, however, we would expect unregistered microenterprises to be more likely to pay bribes than registered microenterprises. This does not appear to be the case—26 percent of unregistered but 35 percent of registered firms reported paying bribes. V.2 Macroeconomic stability Firms in Cape Verde were also relatively unconcerned about macroeconomic instability (i.e., inflation and exchange rate instability). Only 9 percent of small and medium-sized firms and 14 percent of microenterprises said that macroeconomic instability was a major or very severe problem. Once again, this is not typical. In the comparator countries, macroeconomic instability ranked among the top five problems in six of eight countries. The low level of concern about macroeconomic instability is consistent with other evidence. Inflation, in particular, has been relatively low in Cape Verde. Between 2002 and 2004, it averaged 2.3 percent. This was lower than in any of the comparator countries except Senegal and the Maldives. Figure 56: Inflation has been relatively low in Cape Verde. 30 inflation rate (%) 25 20 15 10 5 0 Maldives Mauritius Guyana South Verde Indonesia Senegal Dominican Philippines Africa Cape Republic -5 Source: World Bank (2006c) Note: Inflation rates are averages between 2002 and 2004. Cape Verde‘s success in this respect reflects the success of the country‘s fixed rate peg to the Euro. As the World Bank‘s recent Development Policy Review (World Bank, 2004a) argues that fixed rate peg has served Cape Verde well. Although the fixed rate peg was tested by fiscal slippages in 2000, it was strengthened by the new Central Bank law, which was enacted in 2002. 112 The peg to the Euro works relatively well due to the fact that the European Union is the main source of both imports and exports for Cape Verde—as well as accounting for two-thirds of remittances. V.3 Telecommunications Very few firms complained about telecommunications services in Cape Verde. Only about 24 percent of firms said that it was a major or very severe problem—making it rank 8th overall. Firms in the retail trade and service sectors were most concerned about telecommunications—close to one third of firms in these sectors said that it was a serious problem. In comparison, only 9 percent of manufacturing firms and 19 percent of hotels said the same. Microenterprises were far less concerned than larger enterprises, with only 4 percent of these firms seeing it as a serious issue. The low level of concern about telecommunications among manufacturing firms in Cape Verde is similar to the level in other middle income countries. Telecommunications did not rank among the top five constraints in any of the comparator countries, with less than 10 percent of firms seeing it as a serious obstacle in seven of the nine countries. In part this probably reflects that manufacturing firms rely less upon reliably telecommunications services than they do on reliable power, macroeconomic stability, and access to finance. However, it probably also reflects that telecommunications reform has been quite successful throughout most of the world. The introduction of competition through cellular services means that coverage has exploded throughout the world in recent years. V.3.1 Sector reform in Cape Verde As in many other countries, telephone coverage has increased rapidly over the past ten years. In 1994, there were about 5 mainline telephone subscribers for every 100 people in Cape Verde (see Figure 57). By 2004, there were 16 mainline connections and 14 cell phones—an increase of over 500 percent. Figure 57: Telephone coverage has been increasing rapidly in Cape Verde since 1998 —mainly due to extremely fast growth in cellular coverage. 35 30 25 20 15 10 5 0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 Cellular subscribers per 100 inhabitants Main lines per 100 inhabitants Source: ITU (2005). 113 In part this reflects the successful sector reform that has taken place in Cape Verde over the past decade.57 In 1994, the Government of Cape Verde enacted a new postal and telecommunications law, transferring responsibility for regulation to the Telecommunications Directorate and splitting the former postal and telecommunications company into two separate entities—one responsible for postal service and the second responsible for telecommunication services. In 1996, the Government of Cape Verde privatized Cabo Verde Telecom (CVT), selling a 40 percent share to Portugal Telecom Internacional (PTI) for US$20 million in a competitive bidding process. The new company was given a 25-year exclusivity period on fixed network services (telephones, telex and data transmission). At the end of 1997, a cellular license was provided to the recently privatized CVT. CVT is also the only Internet Service Provider (ISP) in the country. As noted in World Bank (2006b), in the absence of interconnection rates regulation and the exclusivity on international and data communications, the potential for competition remains limited. In an effort to further liberalize the sector, the government has started to renegotiate the concession contract with PTI—a process that will take some time and may result in a large settlement with fiscal implications. At the end of November 2005, the GOCV announced that it planned on ending Cabo Verde Telecom‘s monopoly on international calls on January 1, 2006 and its fixed line monopoly on January 1, 2007. The Government has also moved to end Cape Verde Telecom‘s (CVMOVEL) monopoly on mobile telephony, issuing a second GSM license. Following the failure of an international bidding process, the Government entered direct negotiations directly with a US-based company. V.3.2 Sector performance in an international perspective Telephone coverage in Cape Verde is significantly higher than the average for Sub-Saharan Africa (World Bank, 2005a). In 2005, there were about 8 fixed and mobile subscribers for every 100 inhabitants in Sub-Saharan Africa, compared to close to 30 in Cape Verde (see Figure 33). Cape Verde also performs relatively well when compared to other some of the other middle income comparator countries with respect to fixed-line coverage. Fixed line coverage is higher than in any of the comparators except Mauritius and Guyana. Cellular coverage is also higher than in most low-income countries in Sub-Saharan Africa. However, it is lower than in many of the other middle income comparator countries. There were fewer cellular customers per 100 inhabitants in Cape Verde than in any of the middle-income comparator countries. This might not be surprising—previous studies have emphasized the importance of competition in increasing cellular coverage (Li and Xu, 2004; Wallsten, 2001). The recent award of a second cellular license is likely to therefore provide a welcome boost in coverage. 57 These paragraphs draw heavily upon ―Infrastructure Reform, Regulation, and Competitiveness in Cape Verde‖ (World Bank, 2006b) 114 Figure 58: Although fixed-line coverage is high in Cape Verde relative to the comparator countries, Cape Verde lags behind the leaders in terms of cellular coverage and Internet use. Philippines Indonesia Indonesia Philippines Guyana Guyana Dominican Rep. Dominican Rep. Maldives Maldives Mauritius Mauritius South Africa South Africa Senegal Senegal Sub-Saharan Africa Sub Saharan Africa Cape Verde Cape Verde 0 25 50 75 100 0 10 20 30 Subscribers per 100 inhabitants Internet users as % of population Cellular subscribers per 100 inhabitants Main lines per 100 inhabitants Source: International Telecommunication Union (2005). Note: All subscriber data for the Maldives and Sub-Saharan Africa are from 2004. Internet users are from 2004 for Philippines and Mauritius are also for 2004. Other data are for 2005. Internet coverage is also low. According to data from the International Telecommunication Union (2005), although there were more Internet users per 100 people in Cape Verde than in Senegal and Sub-Saharan Africa in 2005, there were fewer than in any of the middle-income comparator countries (see Figure 58). Although differences in the number of Internet users at the individual level might not necessarily reflect differences among businesses, the evidence from the Investment Climate Survey also suggests that coverage is low (see Figure 59). Enterprises in Cape Verde were less likely to use e-mail to communicate with suppliers and customers than enterprises in any of the comparator countries. Moreover, this probably understates the actual gap to some degree. Coverage has been increasing rapidly in recent years and while, the Cape Verde Investment Climate Survey surveyed firms in 2006, most of the other surveys were conducted between 2003 and 2005. The gap in Internet use probably reflects several factors. One is that small firms might have less financial and human resources than large firms, making it less likely that they will use the Internet than larger firms. This appears to be the case in Cape Verde. Whereas over 90 percent of medium and large firms used e-mail to communicate with suppliers and customers, only 60 percent of small firms, 30 percent of very small firms and ten percent of microenterprises did the same. Given the small size of firms in Cape Verde, Internet use in Cape Verde might be lower than elsewhere for this reason. Although firms in Cape Verde were less likely to use the Internet to 115 communicate with clients and suppliers than firms in the Philippines or Indonesia, SMEs were more likely to do so. SMEs in Cape Verde, however, were less likely than firms in the leading countries (e.g., the Maldives, South Africa or Mauritius) to use e-mail to communicate with clients and suppliers. Other factors might also have played a role, however. Previous studies have found that Internet connectivity is lower in countries where regulatory barriers make it harder for new companies to provide Internet services (Clarke and Wallsten, forthcoming; Wallsten, 2005). Cape Verde Telecom‘s monopoly over Internet service might have retarded sector development. Figure 59: Firms in Cape Verde are less likely to use e-mail to communicate with clients and suppliers than in the comparator countries. Maldives Maldives Indonesia Philippines Philippines Indonesia Dominican Republic Dominican Republic Guyana Guyana South Africa South Africa Mauritius Mauritius Senegal Senegal Cape Verde CapeVerde 0% 25% 50% 75% 100% 0% 25% 50% 75% 100% % of firms with internet access % of SMEs with internet access Source: Investment Climate Surveys. Another reason why internet use is low in Cape Verde is that firms that operate entirely in local markets have a far lower incentive to use the Internet than firms competing in national or international markets. When firms operate primarily in local markets they can rely upon face-to- face interactions or can use standard telecommunications services cheaply. In contrast, the high cost of telecommunications services mean that firms operating in international markets have a greater incentive to use the Internet. As noted in Chapter 2, firms in Cape Verde are particularly likely to operate primarily in local markets. There were other differences in e-mail use between firms within Cape Verde. As in other middle income countries, foreign-owned firms are more likely to use e-mail than domestic firms (Clarke, 2003). This might reflect either higher levels of technical sophistication or the need for the firm to communicate with their parent companies overseas. Another is that firms with better 116 educated managers are more likely to use e-mail than other firms. Finally, the differences in regional coverage (i.e., between Praia and Mindelo) appear to be relatively modest. Figure 60: Large firms, foreign-owned firms and firms with better educated managers were more likely to use the Internet than other firms were. 100% % that use e-mail 75% 50% 25% 0% n lo e ca y o ity Se ng O fac il a s ni al n ll Fo c l al ta a Vo dar on ice ti ig rg Ve icr ai io de n rs Sm es Sm ri an Re re Pr tio ct La M tu v ve in n om ru r co M ry nd st D Se U u .a er C th ed M M Source: Investment Climate Survey. Despite the monopoly over fixed line services in Cape Verde, the quality of services does not appear to be especially poor. The median firm that got a connection in the past two years reported that it took about seven days to do so (see Figure 61). This is slightly longer than in most of the middle income countries where similar data were available, but was not excessively long. There was difference between Mindelo and Praia in this respect and differences were relatively modest with respect to sector. The shorter delays for service and retail firms might reflect the fact that these firms were more likely to be located in commercial districts that might be easier to connect. 117 Figure 61: Although longer than in the best performing countries, it does not take an excessively long time to get a connection in Cape Verde. Indonesia Mindelo Philippines Praia Mauritius Construction Senegal Manufacturing South Africa Other Services Retail Cape Verde 0 5 10 15 0 5 10 Days to get connection Days to get connection Source: Investment Climate Assessment V.3.3 Regional Differences in the quality of telecommunication services Firms in Mindelo were over twice as likely as firms in Praia to say that telecommunications services were a major or very severe obstacle to their current operations (35 percent and 15 percent of firms respectively). The survey provides some objective evidence that the quality of telecommunications services is worse in Mindelo than in Praia. In Praia, the average firm that had got a fixed line telephone connection in the past 2 years reported it took about 6 days. In Mindelo, it took almost twice as long. Although quality might be worse, there was less evidence that access is worse in Mindelo. In the sample of larger (non-micro) enterprises, about 52 percent of firms in Mindelo and 51 percent of firms in Praia used e-mail to communicate with clients and suppliers. Evidence from the microenterprise surveys, where some additional questions on access were asked, also provides little evidence of a gap between Mindelo and Praia with respect to access to services. In fact, firms in Mindelo were more likely than firms in Praia to have fixed line telephone connections (72 percent compared to 66 percent), faxes (17 percent compared to 13 percent), and cell phones (67 percent compared to 61 percent). About 10 percent of microenterprises in both cities used e-mail to communicate with clients and suppliers. 118 CHAPTER 7: CONCLUSIONS AND RECOMMENDATIONS Labor productivity is high in Cape Verde. Manufacturing enterprises in Cape Verde are more productive than similar enterprises in the other lower middle income comparator countries. Although manufacturing firms in Mauritius, the Maldives, and South Africa produce more output per worker than firms in Cape Verde, they have considerably higher per capita income and so it is not surprising that their labor productivity is also higher in these countries. A more detailed analysis of total factor productivity—which also takes into account enterprises‘ use of capital— leads to a similar conclusion. Despite being relatively productive, few manufacturing firms in Cape Verde export (only 6 percent). This is lower than in any of the comparator countries including the other small island economies such as the Maldives and Mauritius. Moreover, very few are involved in both exporting and domestic markets. The low propensity to export is consistent with the macroeconomic evidence—merchandise exports are relatively modest. This makes the economy vulnerable to external shocks. Although prudent macroeconomic policy—particularly low inflation and a credible exchange rate peg—has reduce this vulnerability, expanding and diversifying the export base would be a useful way of further reducing it and would allow Cape Verde to sustain and continue to improve its standard of living (World Bank, 2004a) So what can the Government of Cape Verde do to encourage diversification? Some reasons why few firms export are structural and cannot easily be addressed through policy interventions. For example, the large fixed costs associated with exporting mean that small firms are less likely to export than large firms. Because firms in Cape Verde operate in a small domestic market, they are likely to be small and therefore will find it harder to export than the larger firms in the comparator countries. Similarly, Cape Verde‘s physical isolation and distance from major shipping routes makes exporting more challenging for firms in Cape Verde. These factors will make it difficult for firms to compete in labor intensive areas of manufacturing and emphasize the importance of developing other industries (e.g., tourism, financial services, or IT services). Although this explains some of the difference between Cape Verde and its competitors, it is important not to overemphasize this issue. Firms from other countries manage to overcome Cape Verde‘s relative isolation to compete with domestic firms within Cape Verde—most consumer and investment goods are imported. Moreover, firms in Cape Verde are less likely to export than firms in the small island comparator countries (e.g., the Maldives and Mauritius). In summary, these constraints will generally make it harder for firms from Cape Verde to operate in international markets than firms in some of the comparator countries. However, this does not mean that the GoCV can not do anything. Policies aimed at improving competitiveness will both promote exports and improve the performance of domestic firms. Reducing barriers to trade might also promote exports. 119 I. IMPROVING FIRM COMPETITIVENESS. In many respects the investment climate in Cape Verde appears quite favorable. Corruption is low. The economy is stable on a macroeconomic basis. The telecommunications sector performs well. However, some problems remain. Reducing the burden that these areas of the investment climate impose upon firms will improve competitiveness in both domestic and international markets. Improving the performance of the power sector. The area of the investment climate that firms were most likely to say was a serious obstacle was electricity. The objective data is consistent with this—outages are more common in Cape Verde than in most of the middle income comparator countries. Losses due to outages are especially high because few firms have generators. The high fixed cost of buying and operating a generator makes it difficult for small firms to afford them. Although specific recommendations on how to improve power sector performance are beyond the scope of an ICA, two recent World Bank reports, the first on infrastructure and the second a Public Expenditure Review have highlighted the need for sector reform in Cape Verde.58 As discussed in detail in the Public Expenditure Review (World Bank, 2006b), the main problem facing the sector is the poor financial performance of Electra (see Box). Until tariffs are revised to bring them to a level that will permit adequate self-financing ratios to meet investment requirements, the viability of the energy sector will remain in doubt, with a consequent impact on the investment climate. In addition to taking steps to improve the financial viability of Electra, the infrastructure report recommended several additional policy changes that might also improve sector performance in a way that would improve service quality (World Bank, 2006b). These include:  Resolving outstanding contractual issues between the government and the concessionaire. Negotiations between the Government and the concessionaire to settle outstanding contractual obligations (e.g. with regards to tariff adjustments, arrears, investments) were conducted in 2005, but the relationship between the parties is still tense and may continue to adversely impact on the performance of the sector. Among these issues are the implementation of the automatic adjustment mechanism, arrears of the central government and municipalities and parastatals to Electra, the VAT refund, and the non participation of the municipalities in the recapitalization, which was due in 2005 (World Bank, 2006a).  Allowing independent power producers (IPPs). The law does not clearly define the procedures and requirements to license independent power producers to sell to Electra. There should be clear guidelines regarding the competitive process and the purchase agreement. Similarly, since Electra is in charge of the dispatching from IPPs; these guidelines should also include an implementation mechanism under which capacity from the 58 There reports are World Bank (2006a) and World Bank (2006b). 120 most efficient plants be called first. A policy to facilitate independent producers could usefully include a price that would be paid to anybody contributing power to the grid.  Allowing auto-producers to sell their power to the grid. Auto-producers are also allowed, by law, to sell their power to the grid. Guidelines should be defined in order to avoid that Electra is required to purchase higher cost KwH. Reducing the relative cost of labor. Although productivity is high, so are labor costs. In fact, unit labor costs—a measure of labor costs that take productivity into account—are higher in Cape Verde than in any of the comparator countries, including South Africa and Mauritius, where unit labor costs are also very high by international comparison. This makes it more difficult for firms to compete on international markets, especially in labor intensive sectors, such as garments. So what can government do to improve the situation? One option would be to increase the flexibility of labor regulation. Although firms in Cape Verde did not indicate that labor regulations are a particular concern, previous studies have emphasized that labor regulations are relatively inflexible. Moving into higher value-added sectors might also make sense. However, given the high premiums paid to educated and skilled workers, which indicate a shortage of skilled workers in some regions of the country, the Government should also take steps to improve the human capital of the workforce. Increasing the supply of skilled workers will reduce the relative cost of these workers and improve Cape Verde‘s competitiveness in these areas. In the medium -term, a policy to reform and improve primary and secondary education would seem to be natural. In the short-term, the Government might look into other options such as supporting training programs —although firms in Cape Verde already invest heavily in these. 121 Box: Brief summary of background on financial sustainability of Electra from Public Expenditure Review (World Bank, 2006b) In December 1999, the formerly state-owned electricity and water company, Electra SA, was privatized through the sale of 51 percent of the share value of the production assets of electricity, water and waste water treatment to a joint-venture of Electricidade de Portugal (EDP), and Aguas de Portugal (ADP), the electricity and water utilities in Portugal. The Government maintained 34 percent of the share value, and the municipalities retained 15 percent. The Privatization Agreement provided Electra the exclusive rights for electricity generation up to December 31, 2003, and up to December 31, 2005 for water production. In May 2002, the Government entered into a 36-year concession agreement with Electra for the management and operations of electricity and water distribution, and the collection and treatment of wastewater. The concession also provides exclusive rights for these services to Electra in its areas of concession throughout the concession period. In recent years, Electra‘s financial results have been unsatisfactory. The main problem is that fuel prices have skyrocketed and tariffs have not been adjusted accordingly. This has led to serious financial problems and the depletion of the company's capital. The failure to adjust tariffs partly reflects the institutional failure of the Multisectoral Regulatory Agency, which was supposed to adjust tariffs when fuel price increases were above the benchmark agreed in January 2000. Because the Multisectoral Regulatory Agency was not established, no adjustment mechanism was put in place until the Economic Regulatory Agency (ARE) started functioning in August 2004. This led to a ―tariff deficit‖ of US$13.2 million over the past five years. The continued of erosion of Electra‘s capital in 2003 and 2004 led to a recapitalization (CVE 1.4 billion). Electra‘s poor financial results have been exacerbated by low collection efficiency. Billing collection rates for public agencies and public companies are 65 percent and 57 percent. As of April 2006 overdue electricity and water consumption bills amounted to CVE 1.9 billion (US$21 million). Simulations using a financial model for Electra developed by ARE suggested that before the increase in fuel prices adopted on April 27, 2006, a 10-percent tariff increase would be needed to achieve financial viability. However, after factoring in the 60 percentage average increase in fuel costs that was implemented in April 2006, the ARE model suggested that a tariff increase on the order of 50 percent would be needed. On June 1, 2006, ARE announced a 25.4 percent increase in electricity tariffs. Although, as discussed in greater detail in the Public Expenditure Review, some mitigating factors might reduce the magnitude of needed additional increases, it does not appear that this increase will meet Electra‘s revenue requirement. As a result, the Public Expenditure Review concludes that further tariff increases will be required to bring tariffs to a level that will permit Electra to self-finance its investment requirements. Source: World Bank (2006b). Improving the tax system. Firms were also concerned about the burden of taxation. Despite a high level of concern, tax rates do not appear to be out of line with other middle-income countries. As discussed in the report, there are several possible reasons for this. One is that marginal tax rates might not provide an accurate picture of the overall burden of taxation in the comparator countries. Many other things—including depreciation rates, other taxes that firms face, and a wide range of fiscal incentives that firms can be eligible for—affect the total burden of taxation. 122 The Doing Business Report 2007, which will include Cape Verde for the first time, will provide better information on the total tax burden that firms in Cape Verde face. However, even this is unlikely to provide detailed information on what aspects might be improved without compromising revenues. One way of collecting more detailed information would be to look at producing more detailed reports on effective tax burdens. For example, the Foreign Investment Advisory Service (FIAS) produces reports on the effective tax burden that look at the overall burden of taxation in far more detail than would be available in a standard report such as Doing Business. Given the level of concern, it might be worth looking at this question in greater detail. II. REDUCING BARRIERS TO EXPORTING. Another way of improving export performance would be to reduce remaining barriers to trade. This could be done by: Reducing remaining barriers to trade. Cape Verde has a liberal trade regime, with customs tariffs being simplified and lowered in early 2004. However, tariffs are still above international standards, with 7 tariff bands ranging between 0 and 50 percent. Taxes on imports have been streamlined and rationalized but the external current account continues to register extremely high deficits due to high imports of capital goods. Improving port performance. As noted in the report, it takes longer to clear goods through customs and ports in Cape Verde than in most of the comparator countries. Given the work that has already been done on improving customs administration, it seems plausible that improving port performance might further reduce this burden. The recent World Bank report on infrastructure provides some recommendations on how this might be done. Recommendations include:  Improving regulatory administration. While the Government is in the process of designing and implementing a privatization process, attention needs to be dedicated to introducing appropriate institutional mechanisms to administer the required economic and regulatory functions. This will entail increasing the capacity of the newly created Port and Maritime Institute.  Modernizing port administration. As a complement to the privatization, a set of measures could be implemented to improve the overall policy of port development in Cape Verde. These would include, but not be limited to: (a) preparing a national port master plan and completion of the plan for each port; (b) establishing regulations of the areas of jurisdiction and expansion for each port; (c) developing a computerized port documentation system; (d) ensuring the development of the congested port infrastructures; and (e) acquiring the necessary equipment for the secondary ports. 123 APPENDIX 1: INTERNATIONAL COMPARISONS OF PERCEPTIONS OF THE INVESTMENT CLIMATE One common complaint about perception-based measures of the investment climate is that people often claim that they are uninformative—all firms in all countries would like lower taxes, better access to finance, lower interest rates, cheaper and more reliable infrastructure, less crime, and less burdensome regulation. As a result, it is claimed that perception-based rankings provide little information. Given this criticism, a natural question is how much firms‘ perceptions reflect specific aspects of the investment climate in Cape Verde and how much they merely reflect complaints firms have throughout the world. Table 16 compares responses in Cape Verde to those of firms in other countries. Some complaints are more common than others. For example, tax rates rank among the top five complaints in five of nine comparator countries and the cost of financing and macroeconomic instability rate among the top five in six of nine. However, of the 17 areas of the investment climate, only trade and customs regulation and telecommunications fail to rank among the top 5 complaints in any country and none ranks in the top five for more than two-thirds of the countries— perceptions do vary between countries. The main issue of concern to the manufacturing sector in Cape Verde—electricity—appears to be much more a significant obstacle than in other countries. One exception is the Dominican Republic, where 84 percent of firms said that it was a serious concern. Firms in Senegal and Guyana also cited electricity as a major problem—although far less often than firms in Cape Verde did. Only 35 percent of firms in Senegal and 41 percent in Guyana rates electricity as a major problem compared with more than 65 percent in Cape Verde. Firms in Cape Verde were less likely to complain about other issues, such as macroeconomic policy, corruption and workers‘ skills and education than firms in most other countries. Only 12 percent of firms in Cape Verde rates macroeconomic policy as a significant problem, compared with more than 50 percent of firms in Indonesia and over 60 percent in Dominican Republic. In terms of corruption, Cape Verde does marginally better than South Africa and Guyana, but significantly better than in the other countries. Manufacturing firms in Cape Verde also do not seem to face significant problems concerning workers‘ skills and education, compared with other selected countries. Only Philippines and Indonesia registered a scoring rate below 20 percent, compared with less than 5 percent of firms rating workers‘ skills and education as an obstacle in Cape Verde. 124 Table 16: An International Comparison of Constraints South Dominican Cape Verde Senegal Mauritius Philippines Indonesia Maldives Guyana Africa Republic (2005) (2003) (2005) (2003) (2003) (2005) (2004) (2003) (2005) Electricity 63.2% 34.7% 9.2% 12.6% 33.4% 22.4% 19.6% 41.0% 84.4% Access to Financing 49.3% 61.5% 12.8% 32.6% 13.5% 17.7% 72.3% 30.7% 35.6% Tax Rates 47.3% 52.1% 19.0% 28.4% 30.4% 29.4% 32.6% 16.7% 50.6% Cost of Financing 45.6% 73.8% 16.3% 48.1% 23.0% 28.8% 67.4% 55.8% 50.0% Informal Sector Competition 36.6% 51.0% 15.8% 38.5% 24.3% 17.3% 14.9% 12.5% 51.7% Trade Regulation 2.7% 39.3% 16.6% 23.2% 21.7% 16.0% 6.1% 18.4% 17.8% Transportation 22.3% 34.7% 10.4% 14.9% 18.3% 16.7% 17.6% 16.7% 21.7% Tax Administration 21.8% 51.1% 10.4% 21.7% 25.1% 29.4% 5.0% 8.7% 27.8% Crime, theft and disorder 20.5% 18.6% 29.3% 25.4% 26.5% 22.3% 35.4% 30.1% 60.0% Business Licensing 16.7% 8.2% 3.4% 46.4% 13.5% 20.7% 2.0% 6.1% 10.6% Labor Regulation 16.7% 14.0% 33.3% 26.9% 24.7% 26.0% 24.0% 10.6% 16.1% Legal System 16.0% 13.5% 8.7% 23.0% --- 24.9% 42.6% 8.1% 31.7% Corruption 15.2% 45.8% 16.3% 37.8% 35.2% 41.9% 40.0% 17.8% 75.0% Macroeconomic Instability 11.5% 31.2% 33.6% 38.5% 38.4% 50.2% 10.2% 44.2% 60.0% Telecommunications 9.0% 3.5% 3.6% 5.5% 11.3% 9.1% 3.9% 24.7% 2.8% Access to Land 7.2% 32.0% 3.3% 21.9% 14.8% 13.1% 64.0% 27.8% 10.6% Worker Skills and Education 4.5% 20.8% 35.4% 42.1% 11.9% 19.1% 25.0% 40.4% 35.6% Source: Investment Climate Surveys Finally, we compare the other four most rated issues in Cape Verde—access to financing, tax rates, cost of financing and anti-competitive or informal practices—with what was found in the other countries. In terms of access to financing, Cape Verde stands between Dominican Republic and Senegal, which displays the highest scoring rates concerning these issues. Concerning tax rates, manufacturing firms in Cape Verde look very similar, in terms of perception, to what can be found in Dominican Republic and Senegal—the two worst countries in terms of perception concerning this item. In terms of cost of financing and anti-competitive practices, manufacturing firms in Cape Verde can be compared with what can be observed in Mauritius. 125 APPENDIX 2: STATISTICAL APPENDIX ON ACCESS TO FINANCE I. ECONOMETRIC METHODOLOGY We examine the question of how different factors, including ownership, affect access to credit in Cape Verde, estimating different versions of the equation below: Finance i  �1  � 2 Firm Characteristicsi  � i The dependent variables are various measures of access to finance for firm i. The measures include whether the firm has a loan from a bank or an overdraft facility, loan duration of the most recent loan for firms with loans, interest rate for the most recent loan for firms with loans and whether the firm is credit constrained. We define a firm as being credit constrained if wants a loan at current interest rates but does not have one (e.g., if it was rejected for a loan in 2005 or reported that it did not apply for a loan because it did not think it would get one). If the firm has either got a loan in 2005 or reported that it did not apply because it did not want one at current interest rates it is classified as not being credit constrained. Several of the dependent variables are dummy variables. These models are estimated as simple Probit models. The other two variable are continuous and the models are therefore estimated as simple probit models. A large literature has shown that financing tends to be a greater problem for small and medium-sized enterprises. 59 The regressions therefore contain a variable to control for enterprise size—the number of workers in the enterprise. Firm‘s financing needs are also likely to depend upon its customers. Ideally, we would want to include a dummy variable representing whether the enterprises exports. A priori, it is difficult to assess whether the coefficient on this variable should be positive of negative. On the one hand, exporters are likely to be more efficient than other enterprises, suggesting that they might find it easier to get financing. On the other, if their greater efficiency translates into higher profits —not necessarily the case since international markets are likely to be more competitive that domestic markets, they might have less demand for bank loans. In practice, because there are so few exporters in the sample (only two), we omit this variable. The regressions include a dummy variable indicating that the firm is partially foreign- owned. Foreign-owned enterprises might be able to get financing either in their home country or from their parent company and, therefore, be less likely to depend upon local banks for financing. If better educated managers are more likely to have contacts that allow them to get loans or find it easier to deal with bureaucratic requirements and paperwork, then they might find access to finance a less serious constraint than other firms. Consistent with this, firms with better educated managers were more likely have audited accounts than firms with managers with only a secondary education. Whereas only 25 percent of firms with a manager with a secondary education or less had 59 See Schiantarelli (1996) for a review of the literature on firm size and financial constraints. 126 audited accounts, about 44 percent of firms with a manager with a vocational degree had audited accounts and 49 percent of firms with a manager with a university degree did. This remains true after controlling for additional variables (i.e., the ones included in the main regressions). In a simple probit regression of a dummy variable indicating that the firm has audited accounts on a dummy variable indicating that the manager has a university education and other control variables (e.g., size, ownership, location and sector), the coefficient on the dummy indicating education remains positive. In addition to these variables, the regressions also include sector and location dummies. II. ECONOMETRIC RESULTS Firm Size. As in most other countries where ICSs have been conducted, large firms have better access to credit than smaller firms do. The point estimates of the parameters indicated that larger firms are more likely to have loans, more likely to have overdraft facilities, pay lower interested rates and are less likely to say that they are credit constrained (i.e., to say that they would like to have a loan but are unable to get one). For two of these four variables, the coefficient on firm size (natural log of the number of workers) is statistically significant at a 10 percent level or higher. The coefficient in the final variable, duration of the loan, is negative—indicating that large firms have shorter-term loans—but is statistically insignificant. Table 17: Effect of firm characteristics on access to credit Probit Probit OLS OLS Probit Firm has loan Firm has overdraft Firm is credit Interest Rate Duration of loan (dummy) (dummy) constrained Observations 200 199 68 68 189 Firm Characteristics Number of Workers 0.1388*** 0.0279 -1.1384 -8.3520 -0.5040*** (natural log) (3.78) (1.41) (1.35) (0.78) (4.43) Foreign-owned -0.1075 0.0444 -0.8200 -30.0444* 0.0851 (dummy) (0.72) (0.58) (0.94) (1.98) (0.19) Manager Education University Educated 0.1130 0.1448** -2.9685** 33.0675* -0.3812 (dummy) (1.13) (2.15) (2.36) (1.97) (1.25) Vocational Educated 0.2075** 0.1383** -2.3519** 20.6739 0.0632 (dummy) (2.01) (2.04) (2.05) (0.98) (0.23) Sector of Operations Construction -0.0197 0.0909 2.8635* -36.3729 0.3951 (dummy) (0.09) (0.72) (1.71) (1.56) (0.43) Retail 0.1012 -0.0125 -1.0078 -12.1801 -0.5738** (dummy) (1.12) (0.26) (1.00) (0.87) (2.17) Other Services -0.0535 -0.0242 1.0162 -4.0050 -0.7702*** (dummy) (0.56) (0.52) (0.76) (0.27) (2.58) Region Praia -0.0327 0.0429 0.7794 2.7896 0.4726** (dummy) (0.46) (1.19) (0.60) (0.18) (2.22) R-squared --- --- 0.16 0.10 --- Source: Investment Climate Survey Note: t-statistics in parentheses. *** Significant at 1% level ** Significant at 5% level * Significant at 10% level. The effect appears to be relatively large. The point estimates of the coefficients suggest that increasing firm size for the average firm by 10 workers increases the likelihood that the firm will have a loan by 8 percentage points, increase the likelihood that it will have an overdraft facility by 2 percentage points, reduces the interest rate by 0.6 percentage points, and reduces the likelihood that the firm reports being credit constrained by 10 percentage points. 127 Although small firms appear to be less likely to have access to bank loans or overdraft facilities, this does not mean that they rely heavily on informal sources such as money lenders or even loans from family and friends. Microenterprises and small firms financed less that 2 percent of the new investment in 2005 using loans from informal sources. The average large firm does not use informal sources of funds at all. The main difference, therefore, appears to be with respect to internal funds—microenterprises finance about 85 percent of short-term assets with internal funds compared to 78 percent for small firms and 69 percent for medium-sized firms. In addition to relying more heavily on banks for financing new investment and short-term assets, medium-sized firms are also more likely to have access to other sources of formal financing. The main source of formal financing was trade and commercial credit (e.g., from suppliers). This accounts for all other sources of formal financing reported by medium-sized firms. Although this pattern is not uncommon, it seems plausible that this might reflect characteristics of the financial sector in Cape Verde. Recent research suggests that banking sector concentration is especially problematic for Small and Medium Sized Enterprises (SMEs) (Beck and others, 2004). Given that there is concern about access to credit for SMEs in Cape Verde, this might be a serious concern (World Bank, 2005a). The banking sector is also heavily dominated by foreign banks. Although there have been questions about foreign banks‘ willingness to provide credit to SMEs, it is important to note that the empirical evidence on whether this is the case is mixed. Although some studies in developing countries have found that foreign-owned banks are less likely to lend to SMEs than similar domestic banks, other studies have found that bank ownership has little impact on lending to SMEs.60 Ownership. The coefficients on the dummy variable indicating that the firm was foreign- owned were statistically insignificant in most cases and the signs do not appear to follow a consistent pattern. The one statistically significant coefficient—on loan duration—suggests that loans to foreign-owned enterprises are shorter term on average than loans to domestically owned enterprises. This might suggest that foreign-owned firms are not less concerned about access to financing and the cost of financing because access is easier for foreign owned firms per se, but that foreign owned firms are different from other firms (in terms of size and sector for example) in other ways. However, the weak results for this variable might simply reflect the small number of foreign- owned enterprises in the sample. Only 13 enterprises were foreign-owned making it difficult to find statistically significant results. Education of Manager. Better educated managers might find it easier to get loans than less well educated managers for a number of reasons. They might have better connections (e.g., with bank managers) than other firms. Or they might find it easier to fulfill the bureaucratic requirements needed to get a loan, such as completing a business plan or keeping accurate financial records. Consistent with this, firms with university educated managers were more likely have audited accounts than firms with managers with only a secondary education. Whereas only 25 percent of 60 See Clarke and others (2003) for a summary of the literature on this to pic 128 firms with a manager with a secondary education or less had audited accounts, about 44 percent of firms with a manager with a vocational degree had audited accounts and 49 percent of firms with a manager with a university degree did. The empirical results are consistent with the idea that firms with better educated managers have easier access to credit. After controlling for other differences between firms, firms with university educated managers were 11 percentage points more likely to have a bank loan, 14 percentage points more likely to have an overdraft facility, the average interest rate was 3 percentage points lower and the average loan duration was 33 months longer than firms with a manager with a secondary education or less. Firms with managers with vocational educations also appear to have better access to credit than firms with managers with a secondary education or less. They were 21 percentage points more likely to have loans, 14 percentage points more likely to have overdraft facilities, the average interest rate was 2 percentage points lower and the average loan duration was 20 months longer. In all but two cases, these differences were statistically significant. Sector of operations. After controlling for size, location, and other firm characteristics, there was little evidence that access to credit varied by sector. The coefficients on sector dummies were statistically insignificant both singly and jointly in most model specifications. The only exception was in the regression for whether the enterprise is credit constrained. Firms in the retail trade and services sectors were 18 and 22 percentage points less likely to say that they were credit constrained than firms in the manufacturing and construction sectors. These differences were statistically significant at a 5 percent significance level or higher. Since these firms were no more likely to say that they had loans or overdraft facilities than other firms, this suggests that this is because they are less likely to demand credit. Consistent with this finding, firms in the manufacturing sector also report that they are more concerned about access to credit (see Chapter 3), with close to 50 percent saying that they are concerned about access to finance compared to only about 30 percent of firms in retail and other services. Region. Despite the fact that managers of firms in Mindelo appear to be more concerned about access to finance than firms in Praia (see Chapter 3), the objective data do not suggest that firms in Praia have better access to credit than firms in Mindelo. The coefficient on a dummy variable indicating that the firm is located in Praia is statistically insignificant in most model specifications. In fact, the one exception is in the regression for whether the firm is credit constrained—firms in Praia were more likely to say that they were credit constrained than firms in Mindelo. Audited Accounts. One common complaint of banks is that although they would like to provide loans to firms, they are unable to because the firms do not keep audited accounts and are unable to provide detailed business plans outlining their investment needs. Keeping audited accounts will generally make it easier to get bank loans since it will be easier for these firms to document firm performance and will also make it easier for them to provide detailed business plans. To test whether this appears to have an impact on access to credit, we add an additional dummy variable indicating that the firm has audited accounts to the base regressions. Because of concerns about exogeneity (i.e., that firms might only have detailed accounts because they were required to in order to get a loan), we do not include this variable in the base regression. 129 For the most part, the point estimates of the coefficients are consistent with the hypothesis that having audited accounts is important in terms of getting access to credit. In particular, the positive coefficient in the regression for whether the firm has an overdraft facility suggests that firms with audited accounts are 5 percentage points more likely to do so, while the negative coefficient on this variable in the regression indicating that the firm is credit constrained suggests that firms with audited accounts are about 10 percentage points less likely to be credit constrained. There is less evidence that having audited accounts affects the terms of the loan—the coefficients are statistically significant at lower significance levels and do not have consistent signs. There is also little evidence that firms with audited accounts are more likely to have a bank loan—the coefficient is small and statistically insignificant. Table 18: Effect of firm characteristics on access to credit Firm has Firm has loan Duration of Firm is credit overdraft Interest Rate (dummy) loan constrained (dummy) Observations 199 199 68 68 189 Audited Accounts Firm has audited accounts -0.0013 0.0551 -1.0316 -11.4501 -0.1178 (dummy) (0.02) (1.27) (0.86) (1.06) (1.49) Firm Characteristics Number of Workers 0.1435*** 0.0214 -1.0402 -7.0592 -0.1553*** (natural log) (3.73) (1.09) (1.24) (0.68) (4.06) Foreign-owned -0.1224 0.0444 -0.6254 -27.0859* 0.0178 (dummy) (0.82) (0.59) (0.61) (1.90) (0.12) Manager Education University Educated 0.1253 0.1356** -3.0353** 31.9879* -0.1089 (dummy) (1.25) (2.06) (2.39) (1.90) (1.15) Vocational Educated 0.2064** 0.1350** -2.5602** 18.4534 0.0303 (dummy) (1.99) (2.02) (2.25) (0.88) (0.31) Sector of Operations Construction -0.0355 0.1360 2.2725 -43.0798* 0.0963 (dummy) (0.16) (0.99) (1.21) (1.88) (0.29) Retail 0.1058 -0.0098 -1.0195 -12.1390 -0.1714** (dummy) (1.16) (0.20) (0.97) (0.87) (2.03) Other Services -0.0426 -0.0333 1.0897 -3.0859 -0.1924** (dummy) (0.43) (0.73) (0.80) (0.20) (2.19) Region Praia -0.0272 0.0476 0.6094 0.5087 0.1567** (dummy) (0.38) (1.28) (0.44) (0.03) (2.22) Source: Investment Climate Survey Note: t-statistics in parentheses. *** Significant at 1% level ** Significant at 5% level * Significant at 10% level. 130 APPENDIX 3: SAMPLE CHARACTERISTICS This report is based upon the results of two surveys carried out in Cape Verde in March and April of 2006. Both surveys were conducted in two locations within Cape Verde, Praia and Mindelo. The first survey, the Investment Climate Survey (ICS), surveyed larger formal enterprises with 5 or more employees. The second survey, the Microenterprise Investment Climate Survey (MICS), surveyed microenterprises with fewer than 5 employees. Because, as discussed below, the two surveys use different surveying methodologies, we do not attempt to pool the two samples in the analysis. One important reason for not pooling is that we have no way to compute weights that would allow us to pool the two samples. Because, as discussed below, we did not conduct a rigorous census for microenterprises, we have no way to calculate weights that are comparable to the weights that we calculate for the sample or larger formal firms. I. THE INVESTMENT CLIMATE SURVEY (ICS) The Investment Climate Survey in Cape Verde targeted establishments in the manufacturing, retail trade, hotel, and other service sectors. Firms were into the following broad categories: (i) Manufacturing; (ii) retail trade; (iii) other sectors (construction; wholesale trade; hotels; transportation, storage and communications; and (iv) computer-related activities. The sampling frame was constructed from a list of firms with over five employees located in Praia and Mindelo. The list was provided by the National Institute of Statistics (INE). Firms on the list were randomly selected and contacted to take part in the survey. Since certain sectors were oversampled to give a sufficient sample for international comparisons (e.g., food and beverages, apparel and other manufacturing firms), weights were constructed to allow us to calculate representative samples for the entire economy of these urban areas of Cape Verde. The sample included 109 enterprises in these three sectors (see Table 19). Nearly 40 percent of the sample was made up of manufacturing firms—41 firms in total. Another 20 percent of the sample was in the retail trade sector (21 firms), with the remaining firms located in the residual category, other. The residual sector included 11 hotels. Table 19: Sample characteristics for the ICS (larger firm) sample (unweighted) Total Number of Firms 109 Sector Size Manufacturing 38% Very Small (5-9) 36% Retail 19% Small (10-49) 55% Other 43% Medium (50 and up) 9% of which: hotels 10% Ownership Location Any Foreign 12% Praia 55% Any State 2% Mindelo 45% Even in this sample, most of the firms were relatively small. Over one-third of the formal firm sample has between 5 and 9 employees (referred to as very small firms). An additional 55 percent of the sample had between 10 and 49 employees. Most of these were also small —close to half of the small firms had between 10 and 15 employees. Only about 20 percent of the small firms had over 30 employees. The remaining firms with more than 50 employees are referred to as 131 medium-sized. These are also relatively small—only 2 had more than 200 employees and none had over 300. In most Investment Climate Surveys, additional categories (‗large‘ and ‗very large‘) with between 100 and 499 employees and 500 employees and up are constructed. Because of the very small number of enterprises in these categories, these firms are pooled in with the ‗medium-sized‘ firms in this sample. Most of the firms were privately owned by Cape Verdeans. About 12 percent of the sample had some foreign ownership, with most of these being majority foreign-owned. Foreign ownership was most common in the hotel sector (about 30 percent of firms in this sector). Only 2 firms had any government ownership. Because of this, statistics are not broken out for state-owned firms. The sample was fairly evenly divided between Praia and Mindelo, with about 55 percent of firms located in Praia. Firms in the sample for Mindelo were more likely to be in the retail trade and service sectors (24 percent and 39 percent) and less likely to be in the manufacturing sector (31 percent) than in Mindelo (14 percent, 35 percent and 40 percent respectively). II. THE MICROENTERPRISE INVESTMENT CLIMATE SURVEY (MICS) In the Investment Climate Surveys (ICSs), firms are randomly sampled from lists of registered firms provided by the Government statistical offices. Because lists of unregistered firms are not available in most developing countries—almost by definition—we cannot follow this approach in the MICS. Within these areas, through discussions with INE and other persons with local knowledge, we identified those parts of the city where microenterprises are known to operate. Within these geographical areas, EEC Canada conducted a quick survey of the area following a prescribed route and counting businesses along the route. After conducting this quick survey, they would then interview every nth firm to get sample elements spread evenly across the geographical areas. In this way, it is possible to collect a random sample of microenterprises, including informal and unregistered enterprises. Although this allows us to construct a random sample in the covered areas, because we do not conduct a rigorous census of the entire country, or even of the entire city, the samples are not necessarily representative of microenterprises throughout the country. In addition, because the judgmental route survey does not cover the entire city, we are unable to construct weights that would allow us to pool the ICS and MICS. In most countries, most of the enterprises covered in these surveys are not included in the lists provided by the statistical agency. In contrast, EEC Canada reported that most of the microenterprises in the MICS sample for Cape Verde were also included in a list of enterprises with fewer than 5 employees provided by INE. The sample contained 104 microenterprises. Firms were divided into four broad categories: retail trade, light manufacturing (e.g., woodworking, welding and dressmaking), construction and services (e.g., hairdressers and automobile mechanics). Most of the enterprises were either in the light manufacturing sector (30 percent) or the retail trade sector (62 percent). About 9 percent were involved in providing other services, while none were in the construction sector. Firms were included in the survey if they had at least one full-time employee and no more than five full-time employees at the time of the survey. The median size was 2 employees, while the average size was 2.2 employees. Just over half of the firms in the sample were located in Praia, with the remaining firms located in Mindelo. 132 Table 20: Sample characteristics for the MICS (larger firm) sample Total Number of Firms 104 Sector Size Light Manufacturing 30% Micro (1-5) 100% Retail 62% Other Services 9% Location Construction 0% Praia 54% Registration Mindelo 46% Registered with any agency 76% Not registered 24% 133 REFERENCES Addison, John T., and Pedro Portugal. 1989. "Job Displacement, Relative Wage Changes, and Duration of Unemployment." Journal of Labor Economics 7(3):281–302. Altonji, Joseph G., and Robert A Skakotko. 1987. "Do Wages Rise with Job Seniority?" Review of Economic Studies 54(3):437–459. Altonji, Joseph G., and Nicolas Williams. 2006. "The Effects of Labor Market Experience, Job Seniority, and Job Mobility on Wage Growth." Research in Labor Economics 17(1998):233–276. Angrist, Joshua D., and Alan B. Krueger. 1991. "Does Compulsory School Attendance Affect Schooling and Earnings?" Quarterly Journal of Economics 106(4):979–1014. Aw, Bee Yan, Sukkyun Chung, and Mark J. Roberts. 2000. "Productivity and Turnover in the Export Market: Micro-level Evidence from the Republic of Korea and Taiwan (China)." World Bank Economic Review 14(1):65–90. Ayyagari, Meghana, Thorsten Beck, and Asli Demirgüç-Kunt. 2003. "Small and Medium Enterprises Across the Globe." Policy Research Working Paper 3127. World Bank, Washington DC. Barron, John M, Mark C. Berger, and Dan A. Black. 1999. "Do Workers Pay for On-the-Job Training?" Journal of Human Resources 34(2):235–252. Beck, Thorsten, Asli Demirgüç-Kunt, and Vojislav Maksimovic. 2004. "Bank Competition and Access to Finance: International Evidence." Journal of Money, Credit and Banking 36(3):627–648. Bernard, Andrew B., and J. Bradford Jensen. 1999. "Exceptional Exporter Performance: Cause, Effect or Both?" Journal of International Economics 47(1):1–25. Biggs, Tyler. 2003. "Is Small Beautiful and Worthy of Subsidy? Literature Review." World Bank: Washington, D.C. Processed. Bigsten, Arne, Paul Collier, Stefan Dercon, Marcel Fafchamps, Bernard Gauthier, Jan Willem Gunning, Abena Oduro, Remco Oostedorp, Catherine Pattillo, Mans Soderbom, Francis 134 Teal, and Albert Zeufack. 2004. "Do African Manufacturing Firms Learn from Exporting." Journal of Development Studies 40(3):115–141. Blundell, Richard W., L. Dearden, C Meghir, and B Sianesi. 1999. "Human Capital Investment: The Returns from Education and Training to the Individual, the Firm and the Economy." Fiscal Studies 20(1):1–23. Bourdet, Yve, and Han Falck. Forthcoming. "Emigrants' Remittances and Dutch Disease in Cape Verde." International Economic Journal. Bourdet, Yves. 2001. "Reforming the Cape Verdean Economy - The Economics of Mudanca." In M. Lundahl, eds., From Crisis to Growth in Africa? London, UK: Routledge. Brown, Charles, and James Medoff. 1989. "The Employer Size-Wage Effect." Journal of Political Economy 97(5):1027–1059. Card, David. 1995. "The Wage Curve: A Review." Journal of Economic Literature 33(2):785–799. --------. 1999. "The Causal Effect of Education on Earnings." In Orley Ashenfleter and David Card, eds., Handbook of labor economics. New York, NY: Elsevier Science, pp. 1801--1863. Card, David, and Thomas Lemieux. 2001. "Education, Earnings, and the 'Canadian G. I. Bill.'." Canadian Journal of Economics 34(2):313–344. Clarke, George R. G. 2001. "Does Internet Connectivity Affect Export Performance? Evidence from the Transition Economies." World Bank: Washington DC. Processed. --------. 2003. "The Effect of Enterprise Ownership and Foreign Competition on Internet Diffusion in the Transition Economies." Comparative Economic Studies 46(2):341–370. --------. 2005. "Beyond Tariffs and Quotas: Why Don't African Manufacturers Export More?" Policy Research Working Paper 3617. World Bank, Washington DC. Clarke, George R. G., Robert Cull, Maria Soledad Martinez Peria, and Susana Sánchez. 2003. "Foreign Bank Entry: Experience, Implications for Developing Countries, and Agenda for Further Research." World Bank Research Observer 18(1):25–40. 135 Clarke, George R. G., and Scott J. Wallsten. Forthcoming. "Has the Internet Increased Trade? Developed and Developing Country Evidence." Economic Inquiry. Clerides, Sofronis K., Saul Lach, and James R. Tybout. 1998. "Is Learning by Exporting Important? Micro-dynamic Evidence from Colombia, Mexico, and Morocco." Quarterly Journal of Economics 113(3):903–947. Duflo, Esther. 2001. "Schooling and Labor Market Consequences of School Construction in Indonesia: Evidence from an Unusual Policy Experiment." American Economic Review 91(4):795–813. Farber, Henry S. 1999. "Mobility and Stability: The Dynamics of Job Change in Labor Markets." In Orley Ashenfleter and David Card, eds., Handbook of labor economics. New York, NY: Elsevier Science. Foreign Investment Advisory Service. 2003. "Cap Vert: Parcours de l'Investisseur." World Bank: Washington DC. Freund, Caroline, and Diana Weinhold. 2004. "The Effect of the Internet on International Trade." Journal of International Economics 62(1):171–189. Gibbons, Robert, and Michael Waldman. 1999. "A Theory of Wage and Promotion Dynamics Inside Firms." Quarterly Journal of Economics 114(4):1321–1358. Greene, William. 2000. Econometric Analysis. 4th Edition. Upper Saddle River, NJ: Prentice-Hall. Grenier, Louise, Andrew McKay, and Oliver Morrissey. 1999. "Exporting, Ownership, and Confidence in Tanzanian Enterprises." World Economy 22(7):995–1011. Groot, Wim. 1995. "The Wage Effects of Investments in Enterprise-Related Training." Empirical Economics 20(1):133–147. Halvorsen, Robert, and Raymond Palmquist. 1980. "The Interpretation of Dummy Variables in Semilogarithmic Equations." American Economic Review 70(3):474–475. Heckman, James J., Hidehiko Ichimura, and Petra Todd. 1997. "Matching as an Econometric Evaluation Estimator: Evidence from Evaluating a Job Training Programme." Review of Economic Studies 64(4):605–654. 136 Idson, Todd L, and Walter Y Oi. 1999. "Workers Are More Productive in Large Firms." American Economic Review Papers and Proceedings 89(2):104–108. International Monetary Fund. 1999. "Cape Verde: Recent Economic Developments." International Monetary Fund: Washington DC. --------. 2005a. "Cape Verde: Fifth Review Under the Three-Year Arrangement Under the Poverty and Growth Facility--Staff Report." International Monetary Fund: Washington DC. --------. 2005b. "Cape Verde: Selected Issues and Statistical Appendix." International Monetary Fund: Washington DC. International Telecommunication Union. 2005. World Telecommunication Indicators, 2003/2004. Geneva, Switzerland: International Telecommunication Union. Kaufmann, Daniel, Aart Kraay, and Massimo Mastruzzi. 2003. "Governance Matters III: Governance Indicators for 1996-2002." World Bank Policy Research Report Series 3106. Washington, D.C. --------. 2005. "Governance Matters IV: Governance Indicators for 1996-2004." World Bank: Washington DC. Processed. Kaufmann, Daniel, Aart Kraay, and Pablo Zoido-Lobaton. 2002. "Governance Matters II: Updated Indicators for 2000/01." Policy Research Working Paper 2772. Washington DC. Keller, Wolfgang. 2003. "International Technology Diffusion." University of Texas Department of Economics: Austin, Tx. Processed. Kennedy, Peter. 2003. A Guide to Econometrics: 5th Edition. Oxford UK: Blackwell Publishing. Kumbhakar, Subal C., and C. A. Knox Lovell. 2000. Stochastic Frontier Analysis. Cambridge, UK: Cambridge University Press. Li, Wei, and Lixin Colin Xu. 2004. "The Impact of Privatization and Competition in the Telecommunications Sector around the World." Journal of Law and Economics 47(2):395– 430. 137 Liu, Jin-Tan, Meng-Wen Tsou, and James K. Hammitt. 1999. "Export Activity and Productivity: Evidence from the Taiwan Electronics Industry." Weltwirtschaftliches Archiv 135(4):675– 91. Maloney, William F. 1999. "Does Informality Imply Segmentation in Urban Labor Markets? Evidence from Sectoral Transitions in Mexico." World Bank Economic Review 13(2):275– 302. Mincer, Jacob. 1974. Schooling, Experience and Earnings. New York, NY: Columbia University Press. Ministry of Finance and Planning, Republic of Cape Verde. 2004. "Growth and Poverty Reduction Strategy Paper (GPRSP)." Ministry of Finance and Planning: Praia, Cape Verde. PEP-Africa (IFC Private Enterprise Partnership for Africa). 2005. "MCC Cape Verde Sector Growth and Investment Porgram: Phase 1 Report. Annex 4: Transport Services." International Finance Corporation: Washington DC. Qiang, Christine Zhe-Wei, George R. Clarke, and Naomi Halewood. 2006. "The Role of ICT in Doing Business." In World Bank, eds., Information and Communications for Development 2006: Global Trends and Policies. Washington DC: World Bank, pp. 57--86. Regional Program on Enterprise Development, Africa Private Sector Group. 2004a. "Enhancing the Competitiveness of Kenya's Manufacturing Sector: The Role of the Investment Climate." Washington DC: World Bank. --------. 2004b. "Investment Climate Assessment: Enterprise Performance and Growth In Tanzania." World Bank: Washington DC. --------. 2005a. "Mali: Une évaluation du Climat des Investissements." World Bank: Washington DC. --------. 2005b. "Senegal: Une évaluation du climat des investissements." Regional Program on Enterprise Development, Africa Private Sector Group: Washington DC. --------. 2005c. "South Africa: Investment Climate Assessment." World Bank: Washington DC. --------. 2006. "The Investment Climate For Microenterprises in South Africa." World Bank: Washington DC. 138 Saggi, Kamal. 2002. "Trade, Foreign Direct Investment, and International Technology Transfer: A Survey." World Bank Research Observer 17(2):191–235. Schiantarelli, Fabio. 1996. "Financial Constraints and Investment: Methodological Issues and International Evidence." Oxford Review of Economic Policy 12(2):70–89. Schultz, T. Paul. 2004. "Evidence of Returns to Schooling in Africa from Household Surveys: Monitoring and Restructuring the Market for Education." Journal of African Economies 13(2):95–148. Soderbom, Mans, and Francis Teal. 2003. "Are Manufacturing Exports the Key to Economic Success in Africa?" Journal of African Economics 12(1):1–29. Topel, Robert H. 1991. "Specific Capital, Mobility, and Wages: Wages Rise with Job Seniority." Journal of Political Economy 99(1):145–176. Transparency International. 2005. "Corruption Perceptions Index 2005." Transparency International: Berlin, Germany. Tybout, James R. 2003. "Plant- and Firm-Level Evidence on "New" Trade Theories." In E. Kwan Choi and James Harrigan, eds., Handbook of International Trade. Malden, MA: Blackwell Publishers, pp. 389--415. Wallsten, Scott J. 2001. "An Econometric Analysis of Telecom Competition, Privatization, and Regulation in Africa and Latin America." Journal of Industrial Economics 49(1):1–19. --------. 2005. "Regulation and Internet Use in Developing Countries." Economic Development and Cultural Change 53(2):501–523. Westphal, Larry E. 2002. "Technology Strategies for Economic Development in a Fast Changing Global Economy." Economics of Innovation and New Technology 11(4-5):275–320. World Bank. 2003. "Cape Verde-Growth and Competitiveness Project. Project Appraisal Document." World Bank: Washington DC. --------. 2004a. "Development Policy Review: Building on a Strong Record To Meet New Challenges." World Bank: Washington DC. 139 --------. 2004b. World Development Report 2005: A Better Investment Climate For Everyone. Washington DC: World Bank. --------. 2005a. "Country Assistance Strategy for the Republic of Cape Verde." World Bank: Washington DC. --------. 2005b. Doing Business 2006: Creating Jobs. Washington DC: World Bank. --------. 2006a. "Cape Verde: The Challenge of Increasing Fiscal Space to Meet Future Pressures." World Bank: Washington DC. --------. 2006b. "Infrastructure Reform, Regulation, and Competitiveness in Cape Verde." World Bank: Washington DC. --------. 2006c. World Development Indicators. Washington, D.C.: World Bank. 140