33663 Investment Climate Assessment: Improving Enterprise Performance and Growth in Tanzania November 2004 Table of Contents AT-A-GLANCE TABLE.................................................................................................................4 ACKNOWLEDGMENTS ...............................................................................................................5 EXECUTIVE SUMMARY .............................................................................................................6 CHAPTER 1: MACROECONOMIC PERFORMANCE AND BACKGROUND.......................17 I. BACKGROUND 17 II. GROWTH AND INVESTMENT 18 III. ECONOMIC MANAGEMENT 19 IV. TRADE 21 CHAPTER 2: ENTERPRISE PERFORMANCE AND PRODUCTIVITY..................................24 I. DATA 24 II. PRODUCTIVITY OF MANUFACTURING ENTERPRISES 25 III. HUMAN CAPITAL 33 IV. EXPORTING BY MANUFACTURING ENTERPRISES. 37 CHAPTER 3: CONSTRAINTS ON ENTERPRISE OPERATIONS AND GROWTH...............42 I. CONSTRAINTS ON ENTERPRISE OPERATIONS AND GROWTH 47 II. CONFRONTING INFORMALITY 70 III. OTHER CONSTRAINTS ON ENTERPRISES OPERATIONS AND GROWTH 76 CHAPTER 4: THE INVESTMENT CLIMATE AT THE SUBNATIONAL LEVEL.................81 I. TAX RATES AND ADMINISTRATION 81 II. INFRASTRUCTURE 81 III. CORRUPTION 83 III. BUSINESS REGULATIONS AND BUREAUCRATIC BURDEN 83 CHAPTER 5: POLICY IMPLICATIONS.....................................................................................85 I. REDUCINGTHE BURDEN OF TAXATION ON FORMAL ENTERPRISES 85 II. IMPROVINGTHE PERFORMANCE OF THE POWER SECTOR 87 III. ENHANCING THE EFFECTIVENESS OF FINANCIAL SERVICES AND ACCESS TO CREDIT 88 IV. REDUCING CORRUPTION 89 V. MAINTAINING MACROECONOMIC STABILITY 90 VI. IMPROVING TIES TO THE INTERNATIONAL ECONOMY 91 VII. INCREASING LABOR AND TOTAL FACTOR PRODUCTIVITY 93 POLICY MATRIX FOR IMPROVING THE INVESTMENT CLIMATE IN TANZANIA...................................................................................................................................94 ANNEX 1: THE TANZANIAN INVESTMENT CLIMATE SURVEY DATA..........................99 ANNEX 2: TECHNICAL APPENDIX FOR PRODUCTIVITY CALCULATIONS................102 ANNEX 3: RANKING OF OBSTACLES BY FIRM TYPE......................................................107 I. RANKINGS BY REGION 107 II. RANKINGS BY SECTOR 109 III. RANKINGS BY FIRM SIZE 111 IV. RANKINGS BY EXPORT STATUS 112 ANNEX 4: REGIONAL DISTRIBUTION OF FOREIGN DIRECT INVESTMENT...............113 ANNEX 5: STANDARD TABLES.............................................................................................115 REFERENCES ............................................................................................................................125 AT-A-GLANCE TABLE Investment Climate at a glance China, Tanzania and Kenya Macro environment China Tanzania Kenya GNI per cap, PPP $ 1995 2000/1 1995 2000/1 1995 2000/1 3400 GNI per capita (US$, PPP) 2650 3920 440 500 1000 1010 Population, mid year (millions) 1205 1262 30 34 26.7 30.1 2400 GDP growth (1991-95 and 1996-2000, avg %) 12.1 8.2 1.8 4.1 1.6 1.78 Openness (Imports+Exports/GDP) 45.7 49.1 46.0 48.6 71.4 62.1 1400 FDI inflows (net, % GDP) 5.1 3.6 5.1 3.8 0.4 1.1 400 Governance 1995 1996 1997 1998 1999 2000 Control of corruption2 -0.3 -1.00 -1.05 China Tanzania Kenya Rule of law2 -0.19 -0.49 -1.04 Political Stability2 0.39 -0.25 -0.86 No. of visits by gvt officials, avg per year 36 33 .. % of senior manager time with gvnt officials 7 16 .. PC per 1,000 people 20 Infrastructure Share of firms with own generator, % 27 .. 55 .. 70 15 Days to clear imports, longest in last year 12 .. 33 .. .. 10 Telephone lines in largest city (per 1000 people) 294 22.7 20 78.4 78.4 5 Personal computers (per 1000 people) 16 .. 1 0.6 4.9 0 Paved roads, % of total 22 4.2 .. 13.8 12.1 1995 1996 1997 1998 1999 2000 China Tanzania Kenya Finance Credit to Priv. Sector (% gdp) 150 Cost of capital (lending interest rate, %) 5.85 .. .. 28.8 22.3 Share of credit from financial institutions, % .. 8 .. 5 100 Credit to private sector (stock, % of GDP) 125 13 29 34 30 50 0 1995 1996 1997 1998 1999 2000 China Tanzania Kenya Source: WDI, Governance Research Indicators, ICU firm surveys 1/ or most recent available year 2 Scale of -2.5 to 2.5. Higher values correspond to better outcomes ACKNOWLEDGMENTS This Investment Climate Assessment (ICA) is based on an analysis of survey data collected by the Economic and Social Research Foundation (ESRF) in Dar es Salaam, Tanzania, and the Regional Program on Enterprise Development at the World Bank, in collaboration with the National Bureau of Statistics (NBS). The survey was conducted between April and July 2003. ESRF staff heavily involved in the project include Professor H.K. Amani, Josaphat Kweka, Oswald Mashindano, and John Kajiba. World Bank staff contributing to this report include Rita Almeida, George Clarke, Chad Leechor, Korotoumou Ouattara, Vijaya Ramachandran, and Tilahun Temesgen. John Paton played an important role in survey implementation. Ivan Rossignol, Andrew Stone, and Robert Utz acted as peer reviewers. Judy O'Connor, Linda Cotton, Odd-Helge Fjeldstad, Christiane Kraus, Melanie Marlett, Khalid Misra, Karen Rasmussen, Ben Tarimo, Simon Thomas, William Steel, and Michael Wong provided helpful comments. EXECUTIVE SUMMARY This Investment Climate Assessment is based on an analysis of data collected in the Investment Climate Survey of manufacturing firms in Tanzania. The survey was conducted between April and July 2003 by the World Bank and the Economic and Social Research Foundation (ESRF) in Dar es Salaam, in collaboration with the National Bureau of Statistics (NBS). In addition to data from this survey, the assessment draws on similar surveys of firms in the tourism and construction sectors, also conducted between April and July 2003; a survey of informal and micro enterprises conducted for World Development Report 2005; the Doing Business database; the World Bank's World Development Indicators; Investment Climate Surveys of firms in China, India, Kenya, and Uganda; and governance indicators from Kaufmann et al. (2003). The assessment compares measures of firm performance and the investment climate in Tanzania with similar measures in Kenya, Uganda, India, and China. The regional comparators, Kenya and Uganda, are chosen because of their geographic proximity to Tanzania, their joint membership of the East African Community, and the fact that similar Investment Climate Surveys were completed for these countries at the same time. India and China are chosen for different reasons. First, they have grown very rapidly for over two decades. These countries therefore provide a useful benchmark for the long-term progress that Tanzania might be able to achieve by improving its investment climate. Second, because of the large size of these countries and their growing importance in international markets, Tanzanian manufacturing firms will have to compete with enterprises from China and India as they expand into regional and international markets. Productivity Labor productivity. Value added per worker was about $2,028 for the median enterprise in Tanzania in 2003. This was higher than in Uganda ($960 per employee), but considerably lower than in India ($3,214 per employee in 1999), Kenya ($3,551 per employee), or China ($4,397 per employee in 2002). Workers were more productive in some sectors than in others. For example, in most sectors, value added per worker was lower in Tanzania than in Kenya, but the reverse was true in chemicals and paints, construction materials, and plastic products. Value added per worker was generally higher in larger firms. In very large firms (firms with 250 or more employees), the median enterprise produced $5,598 of value added per worker. In small firms (firms with between 10 and 19 workers), the median level was only $894 per worker. Value added per worker was particularly low for micro enterprises (firms with fewer than 10 workers)--only $474. Capital intensity. Labor productivity may vary across firms due to differences in capital intensity (i.e., the amount of capital per worker). The capital intensity of the median firm was lower in Tanzania ($6,853 per worker) than in Kenya ($9,731 per worker), but considerably higher than in Uganda ($1,421 per worker). Capacity utilization was similarly low in the three countries--between 57 and 61 percent. Larger enterprises in Tanzania were more capital-intensive. The median very large enterprise had about $12,944 of capital per worker. In comparison, the median small enterprise had only $4,655 of capital per worker. Total factor productivity. Differences in labor productivity can sometimes be explained by differences in the use of other factors of production (e.g., capital). To control for such differences, we used regression analysis to calculate total factor productivity (TFP) for firms in Tanzania, Kenya, and Uganda. Differences in TFP are differences in productivity that cannot be explained by differences in the use of labor, capital, or intermediate goods. Firms with higher TFP produce more goods with fewer inputs. By this measure, Tanzanian firms are less efficient than firms in the comparator countries. Average TFP is about 5.6 percent higher in Kenya than in Tanzania. Average TFP is also higher in Uganda, even though value added and sales per employee are almost twice as high in Tanzania as in Uganda. This discrepancy can be explained by differences in capital intensity. Firms in Tanzania produce more value added per worker than firms in Uganda because they have more capital per worker, not because they are more efficient. Total factor productivity, like labor productivity, is higher for large firms. Thus, the differences in labor productivity between large and small firms are not simply due to the fact that large firms employ more capital per worker. Firms that export also tend to have higher TFP than non-exporting firms--something that is true in many countries. Human Capital Education of the workforce. The quality of the workforce is a serious constraint on productivity in many developing countries. It appears to be an especially serious problem in Tanzania, even by regional standards. Workers in Tanzania tend to have considerably less formal education than workers in either Kenya or Uganda. Some 43 percent of workers in Tanzania have only a primary education, compared to 20 percent in Kenya and Uganda. The main cause of this gap appears to be Tanzania's low levels of secondary and vocational education. There is little difference between Tanzania and the comparator countries with respect to tertiary education. The education of the top manager also has an important influence on firm performance. On average, total factor productivity is 24 percent higher in enterprises where the manager has a university degree. In this respect, Tanzanian enterprises do well by regional standards. More managers have a university degree in Tanzania (68 percent) than in Uganda (40 percent) or Kenya (60 percent). However, top managers in China were far more likely to have a university degree (84 percent). Training. Enterprises in Tanzania were less likely to have formal training programs than enterprises in either Kenya or China (but not Uganda). Tanzanian firms with formal training programs also provided less training than their counterparts in the comparator countries. Enterprises without formal training programs generally reported that they were unable to afford a formal program or that informal training was sufficient. But contrary to the latter belief, formal training appears to pay off. Total factor productivity was 11 percent higher in Tanzanian enterprises with formal training programs. Technology. Tanzanian enterprises tend to use technology less intensively than enterprises in Kenya and China, but more intensively than enterprises in Uganda. About 68 percent of enterprises in Tanzania reported that at least some of their employees used a computer on the job, compared to 85 percent in Kenya and 96 percent in China. Similarly, about 58 percent of enterprises in Tanzania used email to communicate with clients and suppliers, compared to 71 percent in China and 79 percent in Kenya. Tanzanian enterprises that used technology more intensively were more productive and had faster sales and employment growth. The impact of HIV/AIDS. In addition to being a human catastrophe in terms of lost lives, HIV/AIDS undermines economic development. Firms in Tanzania generally appear to be responding less aggressively to HIV/AIDS than enterprises in Kenya. About 31 percent of Tanzanian enterprises had some type of prevention program by mid-2003, compared to 32 percent of Ugandan enterprises and 44 percent of Kenyan enterprises. The most common program involved displaying HIV prevention messages--most enterprises with a program provided this service. Close to half of enterprises with a program also reported that they provided counseling. Programs involving condom distribution, anonymous HIV testing, and financial support for dependents were far less common Exports Manufacturing enterprises in Tanzania are less likely to export and export less of their output than manufacturing enterprises in Kenya and China. The difference between Tanzania and Kenya is not entirely explained by differences in firm characteristics, such as enterprise size and sector of operations. When we controlled for these differences, enterprises in Tanzania were 18 percent less likely to export and exported 4.7 percent less of their output than similar Kenyan firms. This difference reflects the fact that trade and customs regulations are more burdensome in Tanzania than in Kenya. For example, it takes about seven days on average for exports to clear customs in Tanzania, compared to four days in Kenya. Barriers to Enterprise Operations and Growth Enterprises in Tanzania were most likely to rate tax rates, electricity, cost of financing, tax administration, corruption, access to finance, and macroeconomic instability as major or very severe obstacles. In addition, large enterprises and exporters rated customs and trade regulations as a serious problem. Tax rates. Tax rates were more likely to be considered a serious problem than any other constraint, with 73 percent of enterprises rating them as a major or very severe obstacle. In fact, more enterprises rated tax rates as a serious problem in Tanzania than in any of the other countries where Investment Climate Assessments had been completed by the end of 2003 (except Ethiopia and Brazil). Tax rates are not a new concern for Tanzanian enterprises. They also ranked as the leading problem facing enterprises in a survey conducted in 1999. Complaints about tax rates were common among most enterprises, with the exception of informal micro enterprises. This probably reflects the extremely low level of tax compliance within this group. Unregistered micro enterprises estimated that they reported about 19 percent of sales to tax authorities, whereas formal micro enterprises estimated that they reported about 28 percent of sales. In comparison, very large enterprises estimated that they reported about 83 percent of sales to the authorities. Despite complaints about tax rates, corporate income tax rates are similar to rates in other developing countries. But value-added tax (VAT) rates are somewhat higher than in the comparator countries. In addition, enterprises face many other national and local taxes. Nonetheless, tax revenues in Tanzania are relatively low. Total tax revenues were about 10 percent of GDP in 2000, compared to 23 percent in Kenya and 11 percent in Uganda. This gap appears to reflect problems associated with informality and evasion, as well as differences in economic structure. Formal enterprises in Uganda and Kenya estimated that they reported more of their sales to tax authorities (77 and 86 percent respectively) than enterprises in Tanzania (69 percent). The complaints about tax rates suggest that the government might want to consider reducing the tax burden on formal enterprises by cutting rates. But such moves must be combined with efforts to improve administration, reduce exemptions, and broaden the tax base. Unless the government improves revenue mobilization, rate reductions could harm Tanzania's public finances, undermine macroeconomic stability, and increase inflation. Tax administration. Some 58 percent of enterprises rated tax administration as a serious problem, making it the fourth-greatest problem facing Tanzanian firms. Reforms in this area should reduce the burden that tax administration imposes on enterprises, as well as improve compliance. Enterprises in Tanzania reported spending an average of seven days per year dealing with tax officials, compared to two to three days in Kenya, Uganda, and China. Despite recent reforms, managers did not report any improvement in this regard between 2001 and 2002. In addition, 21 percent of enterprises reported that tax inspectors requested gifts or informal payments during required meetings. Power. Despite recent reforms, about 59 percent of enterprises rated the power sector as a serious problem, more than in Kenya, Uganda, or China. Concern about power was particularly widespread among larger enterprises. Whereas over 60 percent of medium, large, and very large enterprises said that power was a serious problem, only 39 percent of informal micro enterprises, 52 percent of formal micro enterprises, and 44 percent of small enterprises said the same. The cost of power does not seem to be excessively high in Tanzania. The average price per kilowatt hour was no higher in Tanzania ($0.08) than in most of the comparator countries. However, reliability does appear to be a serious problem. The median enterprise in Tanzania reported losing 5 percent of production due to outages and surges. This was considerably higher than the median estimates for China (0 percent), Uganda (0 percent), and Kenya (3 percent). Problems with maintenance and the poor commercial performance of the Tanzania Electricity Supply Company (TANESCO) partially explain the sector's problems, but they are not the whole story. Per capita generating capacity is also lower in Tanzania than in the comparator countries. A severe drought, which has limited hydroelectric power generation, might exacerbate these problems in the near term. Finance. Tanzania has made considerable strides in developing its banking sector in recent years--privatizing several state-owned banks and allowing foreign banks to enter the Tanzanian market. But enterprises continue to report that access to finance and high interest rates are serious problems. About 58 percent of enterprises reported that cost of financing was a major or very severe constraint on operations and growth, and 48 percent reported the same for access to finance. Interest rates in Tanzania are in line with rates in the regional comparators, Kenya and Uganda. The median nominal interest rate was 13 percent in Tanzania, slightly lower than in Kenya (15 percent) and significantly lower than in Uganda (18 percent). Inflation was also slightly lower in Tanzania than in Kenya and Uganda, suggesting that real interest rates in the three countries were fairly close. But the median interest rate in China was significantly lower than in the three African countries (5.9 percent with inflation of about 1.2 percent). Despite the similarity in interest rates, access to credit appears to be worse in Tanzania than in Kenya. Only 20 percent of enterprises in Tanzania reported having loans from a financial institution, compared to 40 percent in Kenya (and 57 percent in China). Similarly, only 16 percent of investment is financed through bank lending in Tanzania, compared to 32 percent in Kenya. Smaller enterprises were far more likely to rate access to finance as a serious problem. They also financed far less investment through the formal financial sector--0 percent for informal micro enterprises on average, 2 percent for formal micro enterprises, and 5 percent for small enterprises. In comparison, very large enterprises financed 31 percent of investment through the banking sector. Corruption. Compared to other low-income countries, Tanzania performs well on most measures of governance (e.g., political stability, rule of law, and regulatory quality). But despite some recent progress, it trails many countries when it comes to controlling corruption. On one common corruption measure, only Kenya, among the comparator countries, performed worse. Corruption ranked as the fifth-greatest problem facing enterprises in Tanzania. About 33 percent of enterprises that did business with the government said that unofficial payments were typically needed to secure a government contract. The median reported payment was about 10 percent of the value of the contract. Furthermore, 35 percent of managers said that informal payments were typically needed to "get things done" in such areas as customs, taxes, licenses, and other government services. The median reported payment in this case was about 0.3 percent of sales. Bribes were also common when obtaining utility connections, applying for import licenses, undergoing inspections (particularly tax inspections), and participating in mandatory meetings with government officials. Micro enterprises appeared to avoid some of the burden of corruption by avoiding contact with bribe-taking institutions. Macroeconomic instability. Despite Tanzania's improved macroeconomic performance, enterprise managers continue to see macroeconomic instability as a major problem. About 43 percent rated macroeconomic instability as a major or very severe obstacle to enterprise operations and growth. Consequently, continued efforts to control inflation and maintain economic growth are important from an investment climate perspective. As part of these efforts, the government must continue to pursue fiscal stability. Cutting taxes on enterprises without steps that improve compliance, boost tax collection in other ways, or reduce spending would be risky. Customs and trade regulations. Although most enterprises did not rate customs and trade regulations as a serious obstacle, they were rated a serious constraint by more than 52 percent of very large enterprises (i.e., the enterprises most likely to engage in foreign trade). Customs delays illustrate the scale of the problem. The median delay for imports was 14 days in Tanzania, while the median delay for exports was 7 days--longer in both cases than in any of the comparator countries. In China, for example, the median delay for imports was 5 days, and the median delay for exports was 3 days. Confronting Informality Recent estimates suggest that the informal sector in Tanzania is equivalent to about 58 percent of gross national income (GNI)--more than in any of the comparator countries. By reducing the level of informality, the government could reduce the tax burden on formal enterprises without compromising macroeconomic stability. Reforms should focus on reducing the costs and increasing the benefits associated with operating in the formal sector. Reducing the regulatory burden on formal enterprises is one way to achieve this goal. Barriers to entry. Recent work has shown that informality is higher in countries where it is costly to register a business. Although most enterprises in Tanzania did not see business licensing as a serious obstacle to enterprise operations and growth, this result is likely to underestimate the impact of entry-deterring regulations. Businesses that are already operating--especially those that have been operating for a long time--may not see entry restrictions as an important barrier to their future operations and growth. For this reason, evidence from the Investment Climate Survey is supplemented with evidence from the World Bank's Doing Business database. According to this data, it takes about 35 days to fulfill all legal requirements for starting a business in Tanzania. Registering a business takes even longer in most of the comparator countries: 41 days in China, 89 days in India, 47 days in Kenya, and 36 days in Uganda. But the monetary cost of this process is considerably higher in Tanzania-- 204 percent of per capita GNI--than in any of the comparator countries. The most expensive procedures are getting the certificate of incorporation, getting a business license, and getting a company seal. Registration requirements are not the only barriers to entry. The process of obtaining infrastructure services--particularly power--also acts as a barrier. The median wait for an electricity connection in Tanzania was 30 days--twice as long as in any of the comparator countries. Micro and small enterprises faced far greater delays than larger firms. The median wait for very large enterprises was only 3 days. Labor regulations. Labor laws can also act as an entry barrier. According to the Doing Business database, labor laws are considerably more restrictive in Tanzania than in most of the comparator countries. But very few enterprises rated labor regulation as a significant obstacle--possibly because labor laws are poorly and unevenly enforced. Other evidence supports this explanation. Very large enterprises, which are more likely to face close scrutiny, were far more likely than smaller firms to rate labor regulations as a serious problem. Business regulations and inspections. The high cost of complying with regulations may also encourage enterprises to remain in the informal sector. The median enterprise in Tanzania reported having 15 inspections or meetings with government agencies per year--significantly more than in Uganda (5) and India (6). Micro enterprises appear to avoid a significant part of this burden by remaining informal. Other Constraints on Enterprise Operations and Growth One concern about perceptions data, such as data collected by the Investment Climate Surveys, is that enterprises that have adapted to certain constraints might not see them as problematic. For example, microenterprises without power did not consider electricity to be a serious problem even though access would probably improve their performance. For this reason, we also examine several potential problems that enterprises by and large did not identify as major constraints. Legal System. Only 20 percent of enterprises reported that the legal system was a major obstacle to their operations and growth. This is consistent with other evidence that suggests that the "rule of law" is relatively well established in Tanzania. Consistent with this, the court system appears to be relatively efficient. For example, according to the Doing Business database, it takes less time to get a court to enforce a standardized debt contract in Tanzania (127 days) than in Kenya (255 days) or China (180 days). However, there is still room for improvement. Enterprises in Tanzania were less likely to report that they could rely on the courts to enforce contracts and uphold property rights (45 percent of enterprises) than in Kenya (49 percent) or China (92 percent). Access to Land. Although larger enterprises were not very concerned about access to land, both formal and informal micro enterprises reported that it was a serious concern. Only access to financing and the cost of financing were greater problems for these groups. It appears to be an especially significant problem in rural areas. Telecommunications. Enterprises in Tanzania rated telecommunications as a less serious constraint than other infrastructure. Although the number of fixed-line phones in Tanzania is low even by regional standards (4 lines per 1000 people), the sector appears to perform well on other dimensions. Only 51 of 214 firms that used fixed line phones to communicate with clients reported losing service at some point in 2002. In comparison, 211 of 257 firms in Kenya reported losing service for at least one day over the same period. Transportation. About 23 percent of manufacturing enterprises regarded transportation as a major obstacle to enterprise operations and growth. However, transportation appears to be a greater concern for enterprises in other sectors--38 percent of enterprises in tourism reported the same. Access to sealed roads and railways appear to the greatest problems within this sector. Regional Differences Perceptions of different constraints varied by region. For example, establishments in Iringa/Mbeya, Kilimanjaro, and Dar es Salaam were most likely to report that tax rates were a major or very severe problem, while tax administration was a particularly serious concern in Iringa/Mbeya, Arusha, and Dar es Salaam. Firms in Arusha, Dar es Salaam, and Iringa/Mbeya were most likely to rate power as a major or very severe obstacle. The reliability of the power supply was a leading concern in Arusha and Dar es Salaam, while the wait to get a power connection was particularly long in Iringa/Mbeya. Enterprises in Morogoro, Dar es Salaam, and Arusha were most likely to report that import competition was a major or very severe obstacle, while enterprises in Tanga and Iringa/Mbeya were most concerned about corruption. Firms in Morogoro and Tanga reported that senior managers spent the most time dealing with government regulations, while the number of annual visits and inspections was highest in Tanga, Mwanza/Mara, and Iringa/Mbeya. Policy Recommendations Reducing the burden of taxation. Any proposal to reduce the burden of taxation must consider the likely impact on tax revenues--and subsequently, on macroeconomic stability and the country's ability to achieve broader social goals. The fact that many enterprises expressed concern about macroeconomic stability underlines the importance of safeguarding macroeconomic performance while reducing taxes. Anecdotal evidence suggests that complaints about tax rates largely reflect concerns about the multiplicity of national and local taxes and fees. A follow-up study on the marginal effective tax rates facing firms of different types (e.g., along the lines of a FIAS incentives review) could help provide specific recommendations on streamlining the tax system. This review should pay particular attention to opportunities for abolishing burdensome local fees and levies. The government should also improve revenue mobilization. Exemptions and evasion have eroded the tax base and created a highly distortionary system. The government should continue to reduce exemptions and expand the tax base by encouraging firms to enter the formal sector (by, for example, reducing the cost of business registration and the burden of regulation). Finally, the government should improve tax administration, both to increase compliance and to reduce the burden imposed on formal enterprises. Reforms should include steps to reduce corruption at the Tanzania Revenue Authority (TRA). Because corruption in individual agencies is affected by the general level of corruption in the country, progress will depend upon making progress in the general fight (see below). Allowing independent internal and external audits, protecting whistleblowers, and giving citizens a way of complaining about harassment can all help in this regard. In addition, two principles have proven effective in reducing corruption specifically in revenue agencies. (See Bird (2003) or Das-Gupta et al. (1999) for more detailed discussion.) The first is minimizing direct contact between tax officials and taxpayers--by automating and computerizing procedures, increasing the use of third-party data for assessments, and relying on tax withholding. A second useful principle is to organize tax agencies along functional lines (such as auditing, taxpayer assistance, and processing tax returns) rather than by tax type. This makes it harder for officials to develop relationships with taxpayers. Improving the performance of the power sector. Although TANESCO's commercial performance has improved since the government signed a two-year management contract with a private firm in May 2002, serious problems remain. Due to deferred maintenance and earlier underinvestment, the company needs significant resources to improve access and reliability. Current market conditions will make privatization difficult in the near term. Consequently, the government's first goal should be to extend the current management contract for three to five years. A new contract should be based on revised performance indictors aimed at reducing losses, increasing connections, and investing in system rehabilitation. In the medium term, the government should promote private participation in the power sector. In preparation for this step, the government should: (i) ensure that TANESCO can finance its own operations; (ii) restructure sector debt; (iii) unbundle generation, transmission, and distribution; and (vi) establish an independent regulator. Enhancing the effectiveness of financial services and access to credit. Access to finance remains problematic, especially for small and medium enterprises. The government should improve access by privatizing the National Microfinance Bank and other remaining state-owned banks, clarifying regulations for smaller microfinance institutions (MFIs), developing and strengthening umbrella organizations for MFIs, and reducing their reliance on external donors. At the same time, the government should avoid disrupting the development of self-sustaining savings and credit cooperatives (SACCOs) or NGO-based MFIs. Plans to open special SME windows at financial institutions or promote development banks should, therefore, be abandoned. Reforms should also focus on improving contract enforcement and increasing the efficiency of the judicial system. Useful steps would include: (i) upgrading the payments and securities settlement infrastructure; (ii) finalizing the credit registry/bureau project; and (iii) revising some banking regulations to relax unduly constraining barriers and tighten loopholes. Reducing corruption. The government is currently working to repeal the Anti- Corruption Act of 1971 and replace it with a new anti-corruption law. The revision will widen the definition of corruption, improve protection for whistleblowers, and place the onus on those being investigated to prove that assets were not amassed through corrupt or unethical behavior. These efforts should target the areas where corruption has been identified as a particularly serious problem. These include the Tanzania Revenue Authority, officials associated with labor regulations and social security, and the municipal police. Maintaining macroeconomic stability. Despite Tanzania's improved macroeconomic performance, enterprise managers remain concerned about macroeconomic instability. Therefore, the government must maintain fiscal discipline. This effort must include steps to improve revenue mobilization and tax administration, as discussed above. In addition, the government must ensure that Tanzania's public and publicly guaranteed debt remains sustainable. Provisional results of the annual debt sustainability analysis for the period ending in June 2003 indicate that this is currently the case, following the granting of irrevocable debt relief under the enhanced HIPC Initiative. However, the continued sustainability of Tanzania's external debt depends on the implementation of macroeconomic and structural reforms, as well as prudent management of external debt. Furthermore, if GDP or export growth is less than projected, or agreement on debt relief from non-Paris Club creditors is not reached, Tanzania's debt might not be sustainable. Improving ties to the international economy. Although exports have been growing--due primarily to increased sales of gold--few manufacturing enterprises export, and very few export much of their production. To address this problem, the government should improve the performance of the customs administration and eliminate overly burdensome trade regulations. Work on a customs reform strategy is currently proceeding with technical assistance from DFID. It is important that these efforts do not only focus on increasing revenue--it is just as important that they reduce the burden that excessive regulation puts on exporters and decreasing processing times. Increasing productivity. Labor productivity and total factor productivity remain low, especially when Tanzania is compared to the fast-growing Asian economies. The government could increase productivity by encouraging enterprises to invest in worker training, which is less common in Tanzania than in Kenya or China. Rather than become directly involved in worker training, the government should facilitate training and apprenticeship programs that are designed and implemented by the private sector-- perhaps by offering tax credits to firms that provide training. The government could also stimulate productivity growth by improving access to business education. This recommendation is based on the finding that enterprises with university-educated managers are more productive than other enterprises. Improving access to business education may also reduce inequality, offsetting the advantages of inherited ownership and family based business knowledge. It may be worthwhile for both governments and donors to revisit their priorities in this area at least with respect to business studies. CHAPTER 1: MACROECONOMIC PERFORMANCE AND BACKGROUND Tanzania is a country of 35 million people. In 2003, its GDP was $10 billion, and per capita income was about $285. (See Box 1.1.) The country's economic performance has improved significantly in recent years. Effective expenditure control and increased flexibility have strengthened fiscal policy and enhanced the efficacy of public services. The government has reduced inflation sharply, allowing foreign exchange markets to stabilize. The international reserves position has remained strong, despite modest export expansion and a major decline in export prices. And since 1995, the Tanzanian economy's growth rate has exceeded the average for sub-Saharan Africa, as well as for developing countries overall. But achieving the major policy goals set out in the Poverty Reduction Strategy Paper (PRSP) remains a substantial challenge.1 Box 1.1 Human development and poverty in Tanzania Tanzania's per capita income was about $285 in 2003. Life expectancy at birth fell from 50 years in 1990 to only 44 years in 2001--due primarily to HIV/AIDS. Infant mortality remains relatively high, with 104 deaths per 1,000 births in 2001 (up from 102 deaths per 1,000 births in 1990). Nevertheless, Tanzania has made significant progress in terms of human development in recent years. · Total adult literacy rose from 63 percent in 1990 to 76 percent in 2001. · Total youth literacy increased from 83 percent in 1990 to 91 percent in 2001. · The percentage of the population with improved access to water increased from 38 percent in 1990 to 68 percent in 2000. · Per capita GDP has grown at an average 1.6 percent per year since 1995. · The results of a household budget survey released in 2002 indicate that basic needs poverty has declined over the past decade from 39 percent to 35 percent (with significant differences, however, between urban and rural areas). Source: World Bank (2004c); National Bureau of Statistics (2002) I. BACKGROUND Tanzania became a new nation on April 26, 1964, emerging from the union of two newly independent states, Tanganyika and Zanzibar.2 Today, Tanzania is a constitutional republic with a multiparty political system and a directly elected president. Since 1984, the constitution has limited each president to two five-year terms. President Benjamin Mkapa is currently serving his second term. The dominant political party is Chama Cha Mapinduzi (CCM), which holds 206 out of 231 seats in the legislature. The next parliamentary (National Assembly) and presidential elections are scheduled for October 2005. Mr. Julius Nyerere served as president from independence until 1985. For the most part, he ran unopposed in elections. His economic and social policies were motivated largely by the Chinese communist model. In particular, his administration set up a series of collectivist farm ventures that gave rise to the ubiquitous "ujamaa" (community) villages. During this period, the economy stagnated. Political reforms began after Mr. Nyerere stepped down in 1985. The multiparty system was introduced in 1991, and the first parliamentary and presidential elections under this new system were held in 1995. II. GROWTH AND INVESTMENT Growth. Growth was modest in the early 1990s but accelerated dramatically in the second half of the decade. From 1990 to 1995, the economy grew at an annual rate of 1.8 percent. But real per capita income declined over this period because the population grew faster than the economy. Since 1995, growth has picked up, with real GDP increasing at an annual rate of 4.4 percent. In 2002, the most recent year for which data are available, growth was an impressive 6.2 percent. Since 1988, Tanzania has grown faster than Kenya (Figure 1.1). In the past few years, it has also grown faster than Uganda.3 Figure 1.1 Since 1988, GDP has grown faster in Tanzania than in Kenya, but more slowly than in Uganda (index: 1988 = 100) 300.0 200.0 Tanzania 100.0 Uganda Kenya 0.0 1988 1990 1992 1994 1996 1998 2000 2002 Source: World Bank (2004c). The mining sector has performed particularly well. With annual growth well above 10 percent over the past five years, it has come to account for about 2.5 percent of GDP. Tourism, along with hotel and retail businesses, has also been a leading sector, growing at well above 6 percent per year in the second half of the 1990s. In addition, policy reform and privatization in the transport and telecommunications sectors have improved performance and boosted the economy.4 Investment. Investment, as measured by gross domestic capital formation, has been relatively high in Tanzania. During the 1990s, total domestic investment, which includes both private and public investment, averaged 22.7 percent of GDP. This is considerably higher than the average for sub-Saharan Africa (15 percent) and about the same as the average for all developing countries (23 percent). The private sector accounts for the bulk of domestic investment (about 17 percent of GDP, or 73 percent of total investment). Public sector investment has been highly variable, ranging from more than 10 percent of GDP in 1990 to 3.4 percent in 1995. Total investment, however, appears to be tapering off, falling from about 25 percent of GDP in the early 1990s to below 20 percent in recent years (Figure 1.2). Figure 1.2 Public and private investment as a percentage of GDP 20.0 Private investment Public investment 15.0 10.0 5.0 0.0 1988 1990 1992 1994 1996 1998 2000 2002 Source: World Bank (2004c). The decline in investment does not appear to have slowed economic growth. GDP growth increased in the mid-1990s, even as capital formation slowed. Changes in the incremental capital-output ratio (ICOR) suggest that growing efficiency in the use of capital made this possible. In the early 1990s, the aggregate ICOR was relatively high at 12.2-- meaning that the capital stock needed to expand by 12.2 percent to produce a 1 percent increase in GDP. Since 2000, however, the ICOR has fallen to 3.3--a sharp improvement. III. ECONOMIC MANAGEMENT Inflation and exchange rates. Tanzania has made impressive progress in reducing macroeconomic instability. Inflation, as measured by change in the consumer price index, averaged a rapid 27 percent per year in the early 1990s. Since 1995, the rate has fallen to an average of 10 percent. In 2002, the most recent year for which data are available, the annual rate of inflation was 4.6 percent and falling. During much of this period, the Tanzanian shilling, which is under a managed float, depreciated against the major currencies. This depreciation did not fully compensate for the country's relatively high rates of inflation. As a result, Tanzania's currency strengthened in real terms against those of its trading partners. But since 2000, the real effective exchange rate has begun to decline, improving Tanzania's competitive position. Fiscal policy. The tax burden in Tanzania, which has averaged between 10 and 12 percent of GDP in recent years, is relatively low by regional standards. But external grants and foreign aid--amounting to 10 to 12 percent of GDP--have enabled the government to spend considerably more than it collects in taxes.5 The cash-budgeting system has helped keep spending under control, with fiscal deficits (after grants) generally remaining below 2 percent of GDP (Figure 1.3). Figure 1.3 Fiscal balance: revenues and expenditures (TSh millions) 2,000,000 1,500,000 Fiscal Balance Revenue & Grants 1,000,000 Expenditure 500,000 0 1990 1992 1994 1996 1998 2000 2002 -500,000 Source: World Bank (2004c). The adoption of the 2000 PRSP changed the government's expenditure profile. Due to increased spending on priority programs designated for poverty reduction, public expenditure is expected to rise from 18 percent of GDP to 22 percent of GDP in the years ahead. In fiscal year 2002, expenditure was roughly 20 percent of GDP, well on its way towards the target. To prevent these increased spending commitments from contributing to large deficits, the government has instituted a new system of expenditure monitoring that gives special attention to poverty-related spending. It will not be easy to achieve the PRSP goals without loosening fiscal policy or increasing the country's dependence on aid. The government plans to meet most of its new spending targets by increasing tax revenue from 12 percent to 15 percent of GDP. To reach this goal, it is adopting a two-pronged strategy. First, it is modernizing tax policy and administration. This involves moving toward a system of self-assessment and reorganizing the Tanzania Revenue Authority (TRA). Second, the government is undertaking reforms to enhance economic growth and promote business activity. These measures should expand the tax base. IV. TRADE Tariffs and non-tariff barriers. During the 1990s and 2000s, many countries abandoned protectionist policies, reducing both tariffs and non-tariff barriers. This trend did not pass Tanzania by. Imports did not increase as a share of GDP in the 1990s-- imports of goods and services accounted for 27 percent of GDP in 1990, 39 percent in 1995, but only 24 percent in 2002 (World Bank, 2003b). But tariffs have declined significantly since the late 1980s. The simple average tariff fell from 29.8 percent in 1988 to 21.6 percent in 1997 and then to 14.3 percent in 2001. Over the same period, the weighted average tariff declined from 22.8 percent to 9.5 percent--lower than in Kenya (17 percent), but higher than in Uganda (9 percent). Non-tariff barriers are similar in Tanzania and Kenya, but higher in Tanzania than in Uganda. The IMF gave Tanzania and Kenya a "2" on its 5-point scale for non-tariff barriers--higher values mean greater restrictions--while Uganda received a "1." Figure 1.4 Sectoral distribution of FDI in mainland Tanzania, 1999 (stock) Other Construction 7% 6% Retail and Mining and catering quarrying 13% 38% Finance and real estate 7% Agriculture, hunting and forestry 7% Manufacturing 22% Source: Tanzania Investment Center (2001). Foreign direct investment. After a period of capital flight in the 1980s, private capital inflows began in the early 1990s. Foreign direct investment amounted to only $21 million in 1993. Since then, however, it has consistently increased, reaching about $240 million in 2002. This has made Tanzania one of sub-Saharan Africa's leading investment destinations. But Tanzania's potential for attracting foreign investment is far from exhausted. Improvements to the investment climate could stimulate increased inflows. Although Tanzania has been successful in attracting foreign direct investment, investment has been concentrated in a few sectors and activities (Figure 1.4). FDI has been particularly significant in mining and quarrying. Its share of the total stock of FDI in mainland Tanzania increased from 31 percent in 1998 to close to 40 percent in 1999. In comparison, manufacturing accounted for around 20 percent of the stock of FDI. Similarly, mining and quarrying accounted for 67 percent of the flow of FDI in 1999, compared with 13 percent for manufacturing and 6 percent each for trade/catering, accommodation, and agriculture/hunting/forestry services. Balance of trade and payments. Like most countries in the region, Tanzania has seen its terms of trade deteriorate significantly over the past decade. Export prices have plummeted (especially for such key exports as coffee and cotton), while import prices have soared. Despite these adverse developments, the country has kept trade deficits from growing out of control. As noted above, Tanzania has progressively reduced trade barriers while improving the incentives for export production. It has been an uphill struggle. The modest and uneven growth of exports of goods reflects a combination of reductions in traditional exports and a sharp increase in a new export-- gold (Figure 1.5). Figure 1.5 Imports and exports of goods ($ millions) 2000 Exports Imports 1500 1000 500 0 1990 1992 1994 1996 1998 2000 2002 Source: IMF, International Financial Statistics. The trade deficit for goods has been quite large for most of the past decade-- typically above $500 million a year and more than 5 percent of GDP (Figure 1.6). (An exception was in 1997, when export earnings peaked due to a temporary change in commodity prices.) This deficit has been partially offset by a surplus in invisible transactions. Tanzania's substantial revenue from tourism is offset by equally large payments for services, leaving the service account more or less in balance. The income account is generally in deficit, although the shortfall has become smaller due to HIPC debt relief. The balance of current transfers, however, is positive and substantial. Thanks mainly to official grants, the current account deficit, while large and continuing, has remained below 5 percent of GDP. Figure 1.6 Balance of payments and trade 0 1990 1992 1994 1996 1998 2000 2002 -200 -400 -600 -800 -1000 -1200 Current account balance Trade balance Source: IMF, International Financial Statistics. Deficits in the current account are financed by inflows of foreign direct and portfolio investment. As noted earlier, foreign direct investment (equity) was $240 million in 2002. In addition, significant inflows of debt capital have allowed Tanzania's gross external reserves to grow steadily--from $600 million in 1998 to $2,000 million in 2003. CHAPTER 2: ENTERPRISE PERFORMANCE AND PRODUCTIVITY This chapter examines enterprise performance in Tanzania in terms of labor productivity, total factor productivity, human capital, and export behavior. It also compares the performance of Tanzanian firms to the performance of firms in Kenya, Uganda, China, and India. This allows us to benchmark Tanzanian firms against both regional and international competitors. Kenya and Uganda are natural comparators because of their geographic proximity to Tanzania and their joint membership of the East African Community. Furthermore, similar Investment Climate Surveys were conducted in the three countries at about the same time. In contrast, China and India may seem to have less in common with Tanzania. But it is useful to compare firm performance in these countries because as the world becomes increasingly integrated Tanzanian enterprises will increasingly find themselves competing with enterprises from outside East Africa, both in Tanzania and in export markets. China and India are especially important in this respect. Because of their large size and recent rapid growth, it is hard to think of any sector (for tradable goods) where Tanzanian enterprises will not find themselves competing with enterprises from these two countries. Comparing Tanzania with China and India is also interesting because both Asian economies have enjoyed strong and sustained growth in recent decades. Although Tanzania's recent economic performance has been impressive (Chapter 1), it is considerably harder to achieve sustained growth over several decades than to achieve shorter bursts of growth.6 Between 1992 and 2002, per capita GDP grew at an average annual rate of 8.7 percent in China and 4.0 percent in India. Between 1982 and 1992, the corresponding growth rates were 8.6 percent and 3.6 percent respectively. In comparison, growth of per capita GDP in Tanzania averaged only 0.8 percent between 1992 and 2002 (World Bank, 2003b).7 I. DATA The data for Tanzania used in this report come from two main sources: an Investment Climate Survey conducted between April and July 2003 and an informal firm survey conducted in the fall of 2003 for the 2005 World Development Report. Additional data are presented from surveys in the tourism and construction sectors that were conducted at the same time as the Investment Climate Survey. The data for Kenya, Uganda, China, and India come from Investment Climate Surveys conducted by the World Bank for these countries. The surveys for Kenya and Uganda were conducted in 2003, the survey for China was conducted in 2002, and the survey for India was conducted in 1999. The Investment Climate Survey for Tanzania covered 276 firms from the manufacturing sector. These firms were randomly selected from a sampling frame that was stratified by firm size, sector of operations, and location. The sampling frame was constructed using lists from various government sources. The most important list was provided by the National Bureau of Statistics. The data were collected in 8 industrial sectors and in 10 regions of mainland Tanzania, plus Zanzibar. The sectors and regions covered in the survey were selected based on the relatively high concentration of manufacturing firms in these areas. Annex 1 describes the sample in detail. Firms in the Investment Climate Survey are divided into several groups based upon their number of employees. Firms with 10 to 19 employees are classified as "small," enterprises with 20 to 49 employees are classified as "medium," enterprises with 50 to 249 employees are classified as "large," and enterprises with 250 or more employees are classified as "very large."8 The lists used to construct the sampling frame provide a fairly representative sample of firms that are registered with government agencies. However, micro enterprises and informal firms are underrepresented. Therefore, we supplement the Investment Climate Survey data with data from a survey of informal and micro enterprises conducted for World Development Report 2005. Survey enumerators visited areas where such enterprises were known to operate and randomly sampled firms with fewer than 10 employees.9 This sample included manufacturing and service enterprises, but to ensure comparability with the results from the Investment Climate Surveys, we only present results for manufacturing firms. The informal survey asked fewer and less detailed questions than the Investment Climate Survey. Therefore, comparisons are not always possible. Firms from the informal survey are divided into two groups. Firms that are registered with a government agency are designated as formal micro firms, and firms that are not registered with any agency are categorized as informal micro firms. Some of the firms in the informal survey are located in rural areas. When there are important differences between urban and rural micro enterprises, we present results separately for each group. Most of the formal micro firms were registered with local government agencies. Of the 382 firms in the survey, 163 were registered with local government agencies, 4 were registered with industry boards, and 33 were registered with central government agencies. Almost all of the enterprises that were registered with central government agencies and industry boards were also registered with local government agencies. Formal/registered firms appear to be distinct from informal/unregistered firms in several ways, but this distinction is not absolute. Some of the registered firms are informal to some degree--e.g., some registered firms appear to evade taxation. Similarly, some of the unregistered firms are formal to some degree--e.g., some are inspected by government officials. II. PRODUCTIVITYOF MANUFACTURING ENTERPRISES Labor productivity. One way to compare productivity across countries is to compare value added per employee, an indicator of labor productivity. Calculating value added per employee requires less information than calculating total factor productivity. Therefore, it is less affected by non-response patterns. (Many firms in the Investment Climate Survey did not provide enough information on capital to calculate total factor productivity.) In addition, this is the only approach that allows us to compare micro enterprises with larger enterprises, since the informal survey did not collect information on capital. Labor productivity in the manufacturing sector is significantly higher in Tanzania than in Uganda. But it is far lower in Tanzania than in China, India, and even Kenya. Value added per employee is about $2,028 in manufacturing enterprises in Tanzania. This figure is 54 percent lower than in China, 43 percent lower than in Kenya, and 37 percent lower than in India (Figure 2.1). Figure 2.1 Labor productivity is lower in Tanzania than in China, India, or Kenya $5,000 $4,397 $ $4,000 US $3,551 in $3,214 $3,000 $2,028 $2,000 productivity $960 $1,000 Labor $0 Uganda (2002) Tanzania Kenya (2002) India (1999) China (2002) (2002) Source: Investment Climate Surveys These averages hide substantial differences across sectors. For example, labor productivity is higher in Tanzania than in Kenya in chemicals and paints, construction materials, and plastic products (Table 2.1). These sectoral comparisons, however, should be treated with caution since they often rely upon only a few observations. Labor productivity also varies by firm size, with large firms producing more per worker than smaller firms (Figure 2.2). In 2002, the median value added per worker was $474 for micro enterprises, $894 for small enterprises, $1,717 for medium enterprises, $3,427 for large enterprises, and $5,598 for very large enterprises. Among micro enterprises, value added per worker was similar for urban and rural enterprises and somewhat higher for informal enterprises than formal enterprises ($560 and $313 respectively). Table 2.1 In some industries, labor productivity is higher in Tanzania than in Kenya Tanzania Uganda Kenya Agribusiness 3,074 942 4,270 (observations) (81) (122) (81) Chemicals and paints 9,220 2,738 4,272 (observations) (27) (18) (25) Construction materials 8,267 1,082 4,368 (observations) (11) (40) (17) Metals 1,303 1,383 3,360 (observations) (29) (21) (48) Furniture and wood 804 554 3,116 (observations) (65) (54) (20) Paper, printing, and publishing 2,722 3,537 6,045 (observations) (25) (23) (18) Plastic products 15,409 6,328 3,655 (observations) (7) (7) (19) Textiles, garments, and leather 950 1,617 2,013 (observations) (31) (15) (47) Source: Investment Climate Surveys Figure 2.2 Medium and large enterprises produce more value added per worker than small and micro enterprises $6,000 $5,598 US$) (in $5,000 $4,000 $3,427 worker $3,000 per $2,000 $1,717 $894 $1,000 $474 $313 $560 $476 $428 added $0 Value Micro formal rural urban Small Large large informal Medium Very Micro Micro Micro Micro Source: Investment Climate Surveys. Capital productivity. Capital intensity (capital per worker) is $6,853 for the median enterprise in Tanzania. This is lower than in Kenya ($9,731), but higher than in Uganda ($1,421) (Table 2.2). The difference between Tanzania and Uganda is driven primarily by differences in the capital stock. The average workforce is roughly the same size in both countries--130 employees in Tanzania and 125 employees in Uganda. Capacity utilization is roughly the same in the three countries--between 57 and 61 percent.10 Capital productivity, measured by the ratio of value added to capital, is also approximately equal in Tanzania and Kenya. Every dollar of capital in Tanzania generates around 30 cents of value added. But capital is much more productive in Uganda, where each dollar generates approximately 60 cents in value added. Table 2.2 Manufacturing firms in Tanzania have less capital per worker than manufacturing firms in Kenya, but more than firms in Uganda Tanzania Uganda Kenya Capital-labor ratio 6,853 1,421 9,731 Value added-capital ratio 0.32 0.66 0.30 Capacity utilization 57.3 58.4 60.3 Source: Investment Climate Surveys. Within Tanzania, capital intensity varies by type of firm. Capital intensity is higher in firms that export and foreign-owned firms: 4.1 times higher for exporters and 2.5 times higher for foreign-owned firms. Larger enterprises also use capital more intensively than smaller enterprises (Figure 2.3). Unfortunately, the informal survey did not include measures of capital comparable to those used in the Investment Climate Survey. Therefore, it is not possible to compare capital per worker for micro enterprises with similar measures for small, medium, or large enterprises. Figure 2.3 Larger enterprises have more capital per worker $18,000 $16,346 $16,000 US$) $14,000 $12,944 (in $12,000 $10,000 worker $7,682 $8,000 per $6,000 $4,655 $4,000 Capital $2,000 $0 Small Medium Large Very large Note:: Data on capital not available for micro enterprises. Source: Investment Climate Surveys. Total factor productivity. Cross-country comparisons of value added per employee give an indication of the relative efficiency of labor across countries. But differences in the use of other factors of production--notably capital--may also help explain differences in labor productivity. This seems particularly likely because both labor productivity and capital intensity are higher in Tanzania than in Uganda and lower than in Kenya. To control for these factor differences, we compute total factor productivity (TFP) using firm-level data. This procedure allows us to identify differences in output that cannot be explained by differences in the use of labor, capital, and intermediate goods. Table 2.3 reports the results of estimating a log linear production function using data for Tanzania, Uganda, and Kenya.11 The two models use manufacturing sales as the dependent variable and control for intermediate inputs used in production, total employment, and the use of capital (measured by the book value of capital multiplied by the rate of capacity utilization). Both models also include a dummy variable indicating whether the manager has a university degree to control for differences in managerial ability. Finally, both models include country dummies to pick up differences in total factor productivity that cannot be explained by differences in factor use or managerial ability. Model II includes sector fixed effects.12 As expected, the coefficients on the labor, capital, and inputs used in production are positive and significant. Enterprises with managers with university degrees are also more productive, although the coefficient is not statistically significant at conventional significance levels. However, after we control more completely for sector differences, this difference become statistically significant. (See Annex 2 for details.) After controlling for differences in the use of factors of production, we find that Tanzanian enterprises are less productive than their counterparts in Kenya and Uganda. Average TFP is approximately 5.6 percent higher in Kenya than in Tanzania. Average TFP is also 1.3 percent higher in Uganda than in Tanzania, even though sales and value added per employee are almost twice as large in Tanzania. The last result appears to be explained by the fact that average capital intensity is higher in Tanzania than in Uganda.13 Table 2.3 Determinants of firm-level productivity for manufacturing enterprises Model I Model II Observations 438 432 Intermediate inputs (natural log) 0.6 0.6 (0.03)*** (0.03)*** Capital (natural log) 0.12 0.12 (0.02)*** (0.02)*** Employment (natural log) 0.25 0.24 (0.05)*** (0.05)*** Manager degree 0.16 0.15 (0.12) (0.12) Tanzania (country dummy) 2.58 3.02 (0.27)*** (0.41)*** Uganda (country dummy) 2.69 3.06 (0.24)*** (0.39)*** Kenya (country dummy) 2.7 3.19 (0.30)*** (0.42)*** Agribusiness (sector dummy) 0.02 (0.26) Chemicals and paints (sector dummy) 0.24 (0.30) Construction materials (sector dummy) 0.1 (0.29) Metals (sector dummy) -0.25 (0.28) Furniture and wood (sector dummy) -0.34 (0.28) Paper, printing, and publishing (sector dummy) -0.19 (0.30) Textiles, garments, and leather (sector dummy) -0.3 (0.28) Adjusted R-squared 0.99 0.99 *Significant at 20% level; ** significant at 10% level; *** significant at 5% level. Note: Standard errors in parentheses. Dependent variable: logarithm of sales. Capital is the book value multiplied by the rate of capacity utilization. Model I does not include sector fixed effects. Regressions do not include a constant (so that three country dummies can be included). Omitted sector is plastics. Source: Authors' computations based on Investment Climate Surveys. Again, these aggregate differences hide substantial sector variation. For example, average TFP in the chemicals and paints sector is approximately 12 percent higher in Tanzania than it is in Kenya (Table 2.4). (See Annex 2 for a detailed analysis.) Table 2.4 In some manufacturing sectors, TFP is higher in Tanzania than in Kenya and Uganda. Tanzania Uganda Kenya Average TFP Agribusiness 3.2 3.3 3.6 (0.48)*** (0.43)*** (0.53)*** Chemicals and paints 5.3 4.7 4.7 (1.5)*** (1.4)*** (1.5)*** Construction materials 3.6 3.7 3.6 (1.3)*** (1.0)*** (1.3)** Metals 2.1 2.7 2.5 (1.2)** (1.0)*** (1.2)*** Furniture and wood 2.1 2 2.4 (0.5)*** (0.5)*** (0.6)*** Paper, printing, and publishing 2.6 3.6 3.1 (2.2) (2.2)* (2.6) Textiles, garments, and leather 2.3 2.3 2.5 (0.52)*** (0.51)*** (0.53)*** *Significant at 20% level; ** significant at 10% level; *** significant at 5% level. Source: Authors' computations based on Investment Climate Surveys. In order to understand which firms tend to perform better than others, Table 2.5 reports average differences in TFP by size and other firm characteristics. On average, very large firms (those with 250 or more employees) tend to have higher TFP than smaller firms. TFP is 16 percent higher in large firms than in small firms and 30 percent higher in very large firms. In other words, larger firms do not have higher labor productivity simply because they employ more capital per worker. They are also more efficient. Firms that export also have higher TFP than non-exporters, and firms with training programs have higher TFP than firms without such programs. Table 2.5 Average TFP is higher in large firms, firms that export, and firms with training programs Average TFP Firm size Small 2.8 Medium 2.8 Large 3.2 Very large 3.6 Education of the manager University degree 3.1 Non-university degree 2.5 Export capacity: Exporters 3.4 Non-exporters 2.7 Training effort Training 3.0 Non-training 2.7 Source: Investment Climate Survey. Annex 2 describes the data and econometric methodology in greater detail.14 The annex also presents results from additional analysis that includes several investment climate indicators. These results are discussed elsewhere in this chapter and in Chapter 3. III. HUMAN CAPITAL Education. A supply of workers with sufficient education and technological know-how is vital for productivity. In general, enterprises in Tanzania did not rate the availability of skilled workers as a major obstacle to operations and growth. Only 25 percent of enterprises rated the skills and education of workers as a major or very severe obstacle. This was slightly lower than in Uganda and Kenya, where 31 percent and 29 percent of enterprises respectively rated these factors as a major or very severe constraint. Nonetheless, workers in manufacturing enterprises in Tanzania tend to have less formal education than workers in Kenya or Uganda (Figure 2.4). About 51 percent of workers in Tanzania had more than a primary education, compared to 79 percent in Kenya and 76 percent in Uganda. This difference is due to low levels of secondary education and vocational training in Tanzania, not differences in tertiary education. Compared to employees in Kenya and Uganda, few workers in Tanzania had secondary or vocational education. But levels of tertiary education achievement were roughly similar across the three countries. Figure 2.4 Highest educational achievement of employees 100% 13% 12% 15% 80% 12% 25% 30% 60% 26% 40% 43% 32% 43% 20% 20% 20% 0% 6% 1% 4% Tanzania Kenya Uganda None Primary Secondary Technical/vocational University Note:: Totals may not equal 100 percent due to rounding. Source: Investment Climate Surveys. Another important factor in shaping productivity is the education of a firm's top manager. Labor productivity is higher in firms where the manager has a university degree. Total factor productivity is also 24 percent higher, on average, in such firms (Table 2.5).15 About 68 percent of enterprises in Tanzania had a general manager with a university education, compared to 60 percent of enterprises in Kenya and 40 percent of enterprises in Uganda. But general managers in China were far more likely to have university educations, with more than 84 percent of managers meeting this condition. Training. About 44 percent of enterprises in Tanzania reported that they offered formal training (i.e., beyond on-the-job training) to permanent employees. This figure was higher than in India or Uganda (28 and 30 percent of enterprises respectively), but lower than in Kenya (48 percent) and significantly lower than in China (71 percent). When asked why they did not invest more in training, enterprises that did not offer formal training programs were most likely to say either that training was not affordable or that in-house informal training was sufficient. Even when formal training was available, workers in Tanzania generally received less training than workers in most of the comparator countries. Enterprises with formal training programs in Tanzania reported that skilled workers received an average of about 3 weeks of training per year. Enterprises in Uganda also provided 3 weeks of training per year, while firms in Kenya provided 7 weeks. But companies in Asia reported much higher training levels: 15 weeks per year in India and 17 weeks per year in China. These differences matter because total factor productivity is significantly higher--11 percent, on average--in firms that offer a formal training program (Annex 2, Table A2.1). Use of technology. One of the most dramatic changes over the past few decades has been the rapid spread of information and communications technologies throughout the developed world. Although many of the wilder claims about the "new economy" have been tempered in recent years, there is a growing consensus that information technology is important for productivity. Similarly, several recent studies have concluded that by lowering communications and search costs, the Internet may have contributed to the recent increase in trade--particularly from developing to developed countries.16 Enterprises in Tanzania that use technology more intensively tend to perform better than other enterprises. (See Annex 2.) Enterprises where a greater percentage of the workforce uses computers have higher sales and employment growth, as well as higher total factor productivity. Similarly, firms that use email and Web sites to communicate with their clients or suppliers have higher average sales and employment growth, labor productivity, and total factor productivity. About 68 percent of enterprises in Tanzania reported that at least some of their workers used a computer on the job--more than in Uganda (47 percent), but less than in Kenya (85 percent) and China (96 percent). A second measure--the percentage of workers using a computer in the median enterprise--gives similar results. In the median enterprise in Tanzania, 5 percent of workers used computers on the job. This puts Tanzania ahead of Uganda (0 percent of workers), but considerably behind China (20 percent). Enterprises in Tanzania were also more likely than enterprises in Uganda to use email or the World Wide Web to communicate with their clients (58 percent, compared to 39 percent). But they were less likely to do so than enterprises in China (71 percent) or Kenya (79 percent). Still, Internet use has been growing quickly in recent years. In 1997, only 6 percent of Tanzanian enterprises had an Internet connection, but 30 percent of enterprises had Internet access by 2000, and 50 percent of enterprises had access by 2002.17 The impact of HIV/AIDS. HIV/AIDS is not just a tragedy in terms of lost lives. It can also undermine economic development and reduce the supply of skilled workers. According to UNAIDS, sub-Saharan Africa is the region of the world that is worst affected by HIV/AIDS. Within the region, some 26.6 million people are living with HIV, and an estimated 2.3 million people died from AIDS in 2003. Although it is difficult to estimate infection rates, UNAIDS estimated that in Tanzania, about 7.8 percent of people between the ages of 15 and 49 were living with HIV/AIDS in 2001.18 The incidence of HIV/AIDS in Tanzania was slightly higher than in Uganda (5 percent), but lower than in Kenya (15 percent).19 Awareness of HIV/AIDS testing in Tanzania is generally high. Although unskilled production workers were slightly less likely to know where to go for testing, over 80 percent of workers in all groups--and over 90 percent of management and professional workers--knew where to go. Across all levels, workers in Tanzania were at least as informed about testing as workers in Uganda. But enterprises in Tanzania appear to be responding less aggressively to HIV/AIDS than firms in other countries. Just under one-third of enterprises (31 percent) reported having some type of HIV/AIDS prevention program in place (Figure 2.5). This figure was significantly lower than in Kenya (44 percent of enterprises) and very slightly lower than in Uganda (32 percent). In all three countries, the most common type of program was the display of HIV prevention messages. Nearly all of the enterprises with a program (72 of 82 firms) took this approach. About 16 percent of enterprises offered counseling for HIV/AIDS, and 13 percent offered free condom distribution. Only a few enterprises offered anonymous HIV testing or financial support for dependents of workers with HIV/AIDS. Figure 2.5 Firms with HIV/AIDS prevention programs, by type, 2003 60% 44% 40% 38% 31%32% 27%24% 29% 23% 20% 16%14% 13% 9% 4%7% 4% 2%3% 3%2%4% 3% 0% Any programs HIV prevention Counseling for Free condom Anonymous Financial Other messages HIV/AIDS distribution HIV testing support for dependents Tanzania Uganda Kenya Note:: Enterprises can have multiple programs. Source: Investment Climate Surveys. The absence of a stronger response may reflect the fact that most enterprise managers believed that HIV/AIDS was not having a significant impact on their operations (Figure 2.6). Only 9 percent of enterprise managers reported that absenteeism among HIV-infected workers was affecting their workforce, and 8 percent cited absenteeism among workers with sick family members or friends as a problem. Even fewer managers reported high turnover due to sickness or death from HIV/AIDS (4 percent) or indicated that AIDS made it difficult to recruit essential workers (3 percent). These percentages were generally lower in Tanzania than in Uganda and Kenya. This suggests that awareness of HIV/AIDS is lower in Tanzania than in some of its neighbors. Figure 2.6 Percentage of firms saying that AIDS affects their workforce, 2003 20% 17%18% 18% 16% 15% 10% 9% 8% 8% 8% 5% 4% 3% 0% Absenteeism among Absenteeism among High turnover Recruiting essential infected workers caregivers workers* Tanzania Uganda Kenya *No recruitment data for Kenya or Uganda Source: Investment Climate Surveys. IV. EXPORTINGBY MANUFACTURING ENTERPRIS ES. The ability to compete in international export markets is a strong indicator of efficiency. Many recent studies have shown that enterprises that export are more efficient than enterprises that do not. One explanation is that exporting encourages enterprises to become more efficient (the "learning-by-exporting" hypothesis). Another is that enterprises are already more efficient are more likely to start exporting (the "self- selectivity" hypothesis). In practice, the evidence suggests that both hypotheses are at least partially true.20 As in other countries, exporters in Tanzania are more productive than non- exporting enterprises. Labor productivity is higher and sales growth is faster for exporters than for non-exporters (Annex 2, Table A2.1). Total factor productivity is also about 26 percent higher for exporters than non-exporters, although the difference is not statistically significant at conventional significance levels (Table 2.5). Although the data from the Investment Climate Surveys does not allow us to conclude that encouraging exports will result in improved productivity (i.e., to distinguish between the learning-by- exporting and self-selectivity hypotheses), it does suggests that this might be the case.21 Large enterprises are more likely to export than smaller firms and typically export more of their production. About 42 percent of large enterprises and 77 percent of very large enterprises export at least some of their sales, compared to only 4 percent of small enterprises (Figure 2.7). In addition, large enterprises export 19 percent of their output, while very large enterprises export 37 percent. In comparison, small enterprises export only 3 percent of output. Figure 2.7 Large enterprises are more likely to export and export more (as a percentage of sales) than smaller enterprises 100% 50% that 80% 77%37% 40% output of 60% 30% enterprises of export 42%19% 40% 20% exported percentage 20% 14% 5% 10% 4% 3% Percentage Average 0% 0% Small Medium Large Very large Percentage of enterprises that export Average percentage exported Note:: Data on exports not available for micro enterprises Source: Investment Climate Survey. These results are not surprising. First, the high fixed costs of exporting make it difficult for small enterprises to enter export markets. Second, typically only the most efficient firms can compete in international markets. Since labor productivity and total factor productivity are higher for large enterprises, it is not surprising that they export more. Finally, there may also be some degree of reverse causality. Exporting increases the size of the market that the firm is producing for, which may give exporters more growth opportunities. Exporting is more common in Tanzania than in Uganda, but less common than in Kenya or China. In the manufacturing sector, 26 percent of firms in Tanzania export some of their output, compared to 19 percent in Uganda, 57 percent in Kenya, and 38 percent in China. Firms in Tanzania also generally export a smaller share of production than their counterparts in Kenya or China (Figure 2.8). The differences between Tanzania and Uganda are due largely to differences in exporting behavior within Africa. In Kenya, 43 percent of firms exported to other African countries, compared to 13 percent in Tanzania and 11 percent in Uganda The difference between Kenya and Tanzania is far smaller for exports to industrialized economies. Only 7 percent of Kenyan firms exported to Western Europe or North America, compared to 9 percent of Tanzanian firms. Figure 2.8 Manufacturing firms in Tanzania are less likely to export than manufacturing firms in Kenya or China, 2003 60% 57% 50% 38% 40% 30% 26% 19% 19% 20% 17% 12% 10% 10% 0% Tanzania Uganda Kenya China Percent exporting Exports as percentage of sales Source: Investment Climate Assessments. The relatively modest number of enterprises exporting from Tanzania to Western Europe and North America may reflect the low level of ISO or other international certification. Only 12 percent of Tanzanian enterprises reported having ISO certification for either quality management (e.g., 9000/1/2) or environmental management (14000). Only 16 percent reported that any of their products were certified to international standards (e.g., ISO, BSI, UL). In comparison, 36 percent of enterprises in China reported that they had received ISO certification for quality management (e.g., 9000/1/2), and 45 percent reported that some of their products were certified to international standards. The cross-country differences in exporting may be due to observable differences between firms in these countries. For example, firms in high-exporting countries may be larger on average, or they may tend to operate in sectors where exporting is easier. To check whether this is the case, Table 2.6 presents a comparison of exporting behavior after controlling for other differences that might affect whether firms export or not.22 In all three countries--Tanzania, Kenya, and Uganda--large firms are more likely to export and typically export more (as a percentage of sales) than smaller firms--even after we control for other characteristics that might affect export behavior. Foreign- owned firms are also more likely to export and export more than domestically owned firms. In addition, exporting is more common in certain sectors. For example, firms that produce construction materials, furniture and wood, and metals tend to export less than other firms. Table 2.6 Probability of exporting and exports as a percentage of sales after controlling for other factors Probit Tobit Firm is an exporter Exports (dummy variable) (as a percentage of sales) Observations 699 557 699 557 Country-Tanzania -0.1833*** -0.1317** -18.2506** -5.2839 (dummy) (4.21) (2.45) (2.51) (0.70) Country-Uganda -0.2635*** -0.2033*** -29.9970*** -15.2556* (dummy) (5.54) (3.47) (3.83) (1.84) Age of establishment 0.0146 0.0399 1.4663 5.5854* (natural log) (0.68) (1.57) (0.46) (1.66) Workers 0.1363*** 0.1020*** 21.2863*** 15.9060*** (natural log) (8.67) (5.29) (9.03) (6.40) Majority foreign-owned 0.2268*** 0.1454** 30.4874*** 20.5129*** (dummy) (4.20) (2.40) (4.34) (2.82) Company uses Internet 0.3041*** 39.5554*** (dummy) (5.65) (4.73) Trade regulation -0.0631** -10.4834** (higher values mean greater obstacle) (2.02) (2.55) Sector­agribusiness -0.0364 -0.1236 20.7149* 10.7299 (dummy) (0.43) (1.26) (1.73) (0.88) Sector-chemicals 0.0100 -0.1095 5.4538 -10.7730 (dummy) (0.10) (1.06) (0.40) (0.76) Sector-construction materials -0.1549* -0.1891* -24.7034 -33.9002* (dummy) (1.81) (1.83) (1.61) (1.95) Sector-furniture and wood -0.1794** -0.2310** -27.0297* -33.9917** (dummy) (2.14) (2.44) (1.78) (2.09) Sector-metals -0.1855** -0.2620*** -22.4058 -38.9597** (dummy) (2.40) (3.22) (1.56) (2.57) Sector-paper, printing, and publishing -0.1134 -0.1749* -15.3482 -17.2551 (dummy) (1.27) (1.74) (1.02) (1.07) Sector-textiles, garments, and leather -0.0608 -0.1039 12.9009 10.1555 (dummy) (0.68) (1.01) (0.98) (1.02) Pseudo R-Squared 0.33 0.35 0.10 0.10 F test: Tanzania=Uganda 2.70 1.49 2.55 1.57 Prob>F 0.10 0.22 0.11 0.21 *Significant at 10% level; **significant at 5% level; *** significant at 1% level. Note: Z-statistics in parentheses. Source: Authors' computations based on Investment Climate Survey. But these firm characteristics do not fully explain cross-country differences in exporting behavior. Even after controlling for firm size, age, and sector, we find that Tanzanian firms are 18 percent less likely to export than Kenyan firms. In addition, they export about 4.7 percent less as a percentage of sales. Differences in technology use and government policy help explain this pattern. After controlling for differences in Internet use and trade and customs regulations, we find that Tanzanian firms are only about 13 percent less likely to export and export 2.9 percent less as a percentage of sales than similar Kenyan firms.23 Furthermore, the difference in exports as percentage of sales becomes statistically insignificant after controlling for these additional factors (We discuss the performance of the customs administration and other barriers to trade below.) CHAPTER 3: CONSTRAINTS ON ENTERPRISE OPERATIONS AND GROWTH One of the goals of an Investment Climate Assessment is to identify the leading factors that constrain enterprise productivity. As a first step, the Investment Climate Surveys ask enterprises whether they rate various problems as no obstacle, a minor obstacle, a moderate obstacle, a major obstacle, or a very severe obstacle to operations and growth. Although these rankings are not a definitive priority-setting tool, they can be a very useful starting point. Additional quantitative data--the number of power outages, for example, or the amounts enterprises report paying in bribes--can add weight to the survey results and help in comparing constraints across countries. In Tanzania, enterprises were most concerned about taxation: 73 percent of enterprises rated tax rates and 56 percent rated tax administration as a major or very severe obstacle (Figure 3.1). Finance was another serious problem. About 58 percent of enterprises identified the cost of finance and 48 percent identified access to finance as a major or very severe obstacle. The other three areas that firms frequently cited as important constraints were electricity (59 percent), corruption (51 percent), and macroeconomic instability (43 percent). Figure 3.1Percentage of enterprises rating problems as major or very severe constraints on enterprise operations and growth in Tanzania, 2003 80 73 59 58 60 56 51 48 43 40 31 31 27 25 25 25 24 23 20 20 12 12 0 rates Tax Electricityoffinance Corruptiontofinanceinstabilityregulations toland system licensingdisorder ofworkers practices regulations Cost administration uncertainty Legal and Access Transportation Tax Access Skills Labor Customs Businesstheft, Regulatory Telecommunications Macroeconomic Crime, Anti-competitive Source: Investment Climate Survey. Far fewer enterprises--generally less than 30 percent--reported that other potential obstacles were major or very severe. For example, only 12 percent of enterprises reported that labor regulations and telecommunications were major or very severe obstacles to operations and growth. Comparison with 1999. To examine the evolution of firm perceptions over time, we compared the Investment Climate Survey data with the results of the World Bank's 1999 World Business Environment Survey (WBES) (Figure 3.2). There are some differences between the two surveys. The WBES included service enterprises; it used a four-point scale (no problem, minor problem, moderate problem, and major problem); and it used some different categories. Still, some useful comparisons can be made. Respondents in both surveys identified the same leading areas of concern. In both 1999 and 2003, high tax rates were most likely to be seen as a major problem, and high interest rates, tax administration, and corruption were rated as major problems. Many enterprises also saw infrastructure as a major problem in both years. (The 1999 survey did not ask specifically about the power sector.) However, enterprises were considerably more likely to rate crime as a serious problem in 2003 than in 1999. Figure 3.2 Percentage of enterprises rating obstacles as major constraints on enterprise operations and growth, 1999 80% 72% 60% 52% 41% 40% 35% 27% 21% 21% 19% 20% 15% 12% 11% 8% 8% 3% 2% 0% rates rate Crime tax rates Inflation Corruption licensing High interest Infrastructure instabilityregulations administration regulations Policy regulationsregulationsExchangeregulations High Tax Labor Customs Business safety and exchange Fire Environmental Foreign Source: World Business Environment Survey. Differences by size. Perceptions of constraints varied appreciably by firm size (Table 3.1). First, small firms were far more likely to report that access to finance was a major problem. Some 69 percent of informal micro enterprises and 72 percent of formal micro enterprises rated access to finance as a major or very severe constraint, compared to 34 percent of large enterprises and 28 percent of very large enterprises. Access to finance was the problem most often identified by micro enterprises as a serious obstacle. In contrast, access to finance ranked sixth and eighth among the serious obstacles identified by large and very large enterprises. Very large enterprises generally regarded customs and trade regulations and workers' skills and education as more serious obstacles than access to finance. Second, not surprisingly, small informal enterprises were far less likely to complain about tax rates and administration--probably because they avoid most taxes by remaining informal. Third, smaller enterprises--especially informal micro enterprises-- were less likely to complain about the performance of the power sector. Fourth, micro enterprises were less likely to complain about macroeconomic instability. Fifth, whereas larger enterprises were mostly unconcerned about access to land, many micro enterprises rated it as a serious problem. Finally, very large enterprises were more likely to identify labor regulations and customs and trade regulations as serious problems. Table 3.1 Percentage of firms rating obstacles as major or very severe, by firm size All Informal Formal Very nonmicro micro micro Small MediumLarge Large Tax rates 73 23 53 76 75 72 80 Electricity 59 39 52 44 66 67 64 Cost of finance 58 54 57 60 61 50 63 Tax administration 56 17 40 55 63 53 64 Corruption 51 18 30 58 51 56 48 Access to finance 48 69 72 58 51 34 28 Macroeconomic instability 43 20 19 35 49 48 44 Customs and trade regulations 31 10 18 22 44 28 52 Economic policy uncertainty 31 18 25 33 36 30 24 Access to land 25 43 57 27 29 21 16 Crime 25 19 23 22 32 26 24 Skills and education of workers 25 18 32 24 25 27 32 Anti-competitive practices 24 17 22 24 24 22 44 Transportation 23 26 31 13 31 29 20 Legal system 20 7 17 15 24 21 28 Labor regulations 12 4 6 9 14 10 24 Telecommunications 12 18 24 15 8 11 20 Note: Bold indicates that the constraint ranked among the top five for enterprises of that size. Source: Investment Climate Survey. Differences by sector. In contrast, perceptions of obstacles did not vary greatly by subsector or between manufacturing and other sectors. Firms in manufacturing, construction, and tourism all ranked tax rates, tax administration, and cost of finance among the top five problems they faced. More than 50 percent of enterprises rated each area as a major or very severe constraint. Electricity also ranked among the top five obstacles for both manufacturing and tourism, and access to finance ranked among the top five for both construction and manufacturing. But very few enterprises in any of these sectors identified telecommunications or workers' skills as a major or very severe obstacle. However, there were some differences by sector. Enterprises in construction were far more likely to rate access to finance as a major or very severe obstacle (80 percent of firms) than enterprises in tourism (67 percent) and manufacturing (48 percent). On the other hand, only 46 percent of construction firms perceived electricity as a major or very severe problem, compared to 66 percent of firms in tourism and 59 percent of firms in manufacturing. Furthermore, although the relative rankings were similar across sectors, enterprises in tourism were generally more concerned than manufacturing enterprises about most obstacles (Table 3.2). Table 3.2 Percentage of firms reporting obstacles as major or very severe, by sector Manufacturing Constructio n Tourism Tax rates 73 66 72 Electricity 59 46 66 Cost of finance 58 69 70 Tax administration 56 73 62 Corruption 51 55 57 Access to finance 48 80 67 Macroeconomic instability 43 69 59 Regulatory policy uncertainty 31 43 54 Customs/trade regulations 31 46 43 Business licensing 27 44 47 Skills and education of workers 25 13 40 Transportation 23 13 38 Telecommunications 12 17 18 Note: Bold indicates that the constraint ranked among the top five for enterprises in that sector. Source: Investment Climate Survey. Differences by ownership. Foreign- and domestically owned enterprises had similar views about many obstacles to operations and growth (Table 3.3). Tax rates were the problem cited most frequently as a serious obstacle by both types of firms. In addition, both groups of enterprises perceived electricity, corruption, and tax administration as serious obstacles. There were, however, some differences. Foreign-owned enterprises were generally less likely to report that access to finance and cost of finance were serious problems. Some 61 percent of domestically owned enterprises reported that cost of finance was a major or very severe obstacle, compared to 46 percent of foreign-owned enterprises. Similarly, 52 percent of domestically owned enterprises reported that access to finance was a serious problem, compared to only 36 percent of foreign-owned enterprises. This may be because foreign-owned enterprises are more likely to have access to financing either fromparent companies or from banks in their home countries. Differences by export status. Exporters and non-exporters had similar perceptions in most areas (Table 3.3). Not surprisingly, however, exporters were more concerned about customs regulations and transportation. They were also far less likely to report that access to financing was a major problem. This may be because banks are more likely to lend to exporters, which are generally more productive than non-exporters. (See Chapter 2.) Table 3.3 Percentage of firms rating obstacles as major or very severe, by ownership and export status All Foreign Domestic Exporters Nonexporters Tax rates 73 75 73 75 73 Electricity 59 66 56 63 58 Cost of finance 58 46 61 53 59 Tax administration 56 55 55 64 54 Corruption 51 55 50 58 51 Access to finance 48 36 52 34 51 Macroeconomic instability 43 45 42 42 43 Regulatory policy uncertainty 31 33 30 24 34 Customs regulations 31 25 33 47 27 Business licensing 27 23 28 25 27 Crime, theft, and disorder 25 23 25 27 26 Skills and education of workers 25 21 27 27 24 Access to land 25 18 27 18 27 Anti-competitive practices 24 25 24 25 21 Transportation 23 29 21 33 19 Legal system 20 20 19 27 18 Labor regulations 12 18 10 17 9 Telecommunications 12 9 12 15 10 Note: Bold indicates that the constraint ranked among the top five for enterprises of that type. Source: Investment Climate Survey. In the rest of this chapter, we focus on individual constraints in greater detail. We also compare Tanzania with two regional comparators (Kenya and Uganda) and two likely competitors where firms have improved their productivity in recent years (China and India). First, we focus on those areas that are widely recognized as significant constraints on enterprise operations and growth in Tanzania (taxation, power, finance, corruption, and macroeconomic stability). We also examine trade and customs regulations, which are a problem for exporters and large enterprises. Second, we discuss regulatory barriers, such as barriers to entry, that encourage firms to remain informal. Barriers to entry may not be a major concern to enterprises that are already operating (i.e., the enterprises in the survey), but they can constrain productivity by restricting competition, deterring the creation of new enterprises, and encouraging informality. Finally, we present data on potential constraints that the surveys may have overlooked and compare Tanzania's performance in these areas with that of regional and international comparators. I. CONSTRAINTS ON ENTERPRISE OPERATIONS AND GROWTH There appears to be broad consensus that certain areas of the investment climate pose particular obstacles to operations and growth. These areas include taxation, electricity, financing, corruption, and macroeconomic instability. In addition, trade and customs regulation are an additional area of concern for large enterprises and exporters. I.1 Taxation Almost 73 percent of enterprises in Tanzania rated tax rates as a major or very severe constraint on enterprise performance and growth--considerably more than rated any other obstacle as a major constraint and more than in any of the comparator countries (Figure 3.3).24 Among the countries where Investment Climate Assessments had been completed by the end of 2003, in only two cases (Brazil and Ethiopia) were enterprises more likely to rate tax rates as a serious obstacle. Fewer enterprises in Tanzania (58 percent) rated tax administration as a serious problem. However, tax administration ranked high among enterprises' serious concerns, falling just below tax rates, electricity, and cost of financing (see Figure 3.1). Figure 3.3 Enterprises in Tanzania rated tax rates and administration as greater problems than enterprises in the comparator countries 80% 73% 68% 70% 60% 56% 48% 51% 50% 40% 34% 36% 30% 24% 20% 10% 0% Tanzania China Uganda Kenya Tax rates Tax administration Source: Investment Climate Assessments. Problems with tax rates and administration were not unique to any one group of enterprises (Table 3.1). In the manufacturing sector, small, medium, large, and very large enterprises all rated tax rates as the greatest obstacle to enterprise operations and growth. (That is, tax rates were the obstacle rated as major or very severe by the largest number of enterprises in each group.) Enterprises in the tourism and construction sectors also rated tax rates as a serious problem, as did foreign- and domestically owned firms and exporters and non-exporters. In addition, each group ranked tax administration among the top obstacles it faced. However, micro enterprises were far less likely to rate tax rates and administration as major problems (Table 3.1). Although 53 percent of formal micro enterprises rated tax rates as a major problem, making it the third-greatest constraint for these firms after access to financing and cost of financing, only 23 percent of informal micro enterprises gave it the same rating. Similarly, whereas 40 percent of formal micro enterprises rated tax administration as a major or very severe obstacle (sixth overall), only 17 percent of informal micro enterprises rated it as a major or very severe obstacle (thirteenth overall). This probably reflects the low level of compliance among micro enterprises, especially informal ones. Table 3.4 Formal micro enterprises are more likely than informal micro enterprises to report that taxes are a major or very severe obstacle Firms reporting taxes are major or Formal Informal Urban very severe obstacle All micro micro micro Rural micro Tax rates 35% 53% 23% 40% 25% Tax administration 27% 40% 17% 28% 23% Source: Investment Climate Survey. Tax rates and revenues. Despite widespread complaints, tax rates in Tanzania do not appear to be significantly higher than rates in other developing countries. Corporate tax rates are 30 percent in Tanzania, China, Kenya, and Uganda and 40 percent in India. In contrast, the base rate for the VAT is slightly higher in Tanzania than in the comparator countries--20 percent in Tanzania, compared to between 14 and 18 percent in China, Kenya, and Uganda. One explanation for the difference between perceptions about tax rates and rates on these headline taxes is that enterprise managers might be concerned about other types of taxes. Consistent with this, informal discussions with several private sector representatives revealed that the multiplicity of national and local taxes is a major source of dissatisfaction. While tax rates in Tanzania are broadly similar to rates elsewhere in the region, tax revenues are relatively low. Total tax revenues were only 10.1 percent of GDP in Tanzania in 1999/2000, compared to 22.5 percent of GDP in Kenya and 10.8 percent in Uganda.25 In addition to reflecting differences in economic structure, the revenue gap probably reflects problems associated with evasion, informality, tax exemptions, and poor tax administration and enforcement. Informality and evasion. It is difficult to draw strong conclusions about tax evasion from Investment Climate Surveys--enterprise managers being unlikely to be especially forthcoming in this respect. However, the evidence from the survey is consistent with the macroeconomic evidence But evasion appears to be a greater problem in Tanzania than in the comparator countries. Managers in Tanzania estimated that the typical firm in their area of activity reported 69 percent of its sales for tax purposes.26 By contrast, the average estimates in Uganda and Kenya were 77 percent and 86 percent respectively. The high level of evasion is especially striking give that enterprises in Tanzania face a greater number of tax inspections and required meetings than enterprises in most of the comparator countries (see section below). Estimates for small, medium, large, and very large enterprises ranged between 68 and 83 percent of sales (Figure 3.4). Reporting rates were highest among very large enterprises and lowest among micro enterprises (about 25 percent). Consistent with the observations, that informal micro enterprises were far less concerned about tax rates and administration than other enterprises, informal micro enterprises reported the smallest share of sales to tax authorities--19 percent, compared to 28 percent for formal micro enterprises. The relatively small gap between informal and formal micro enterprises may reflect the fact that informality is a continuum, with many enterprises being informal to some degree.27 Figure 3.4 Micro enterprises report smaller shares of their sales to tax authorities than large enterprises 100 83 80 75 authorities 68 70 tax 60 to 40 25 28 27 19 20 (average) 20 reported 0 sales of Rural micro micro Small Large large % Allmicro micro Medium Very Formal Urban Informal Source: Investment Climate Surveys. One approach that the government could take to combat evasion would be to reduce the tax burden on formal enterprises. However, since the government's fiscal discipline has underpinned Tanzania's impressive macroeconomic performance, rate cuts would need to be combined with steps to improve revenue mobilization. One way for the government to do this would be to reduce costly tax exemptions and incentives.28 Another would be to encourage informal firms to enter the formal sector--and in so doing broaden the tax base--by reducing the regulatory burden on formal enterprises. As a first step in improving revenue mobilization, the government must improve tax administration, which remains problematic.29 For example, the VAT efficiency ratio (the ratio of VAT revenues to GDP divided by the VAT rate) in Tanzania (0.20) lags the average for sub-Saharan Africa (0.27). The government has already taken some steps to strengthen administration, including the establishment of the Large Taxpayer Unit in 2001, the introduction of taxpayer identification numbers in 2000, and the recently enacted five-year corporate plan for the Tanzania Revenue Authority. Tax Administration. Improving collection should not be the only goal of tax administration reform. The government should also reduce the burden that tax administration imposes upon enterprises. Enterprise managers in Tanzania reported that in 2002/03, they spent about seven days dealing with inspections or required meetings with tax officials. In comparison, enterprise managers in Kenya, Uganda, and China reported spending two to three days on such tasks. Despite recent reforms, managers generally did not report any reduction in this administrative burden. Although 36 managers reported fewer meetings with tax officials in 2002 than in 2001, 59 managers reported more meetings, and 149 reported the same number. The burden of tax administration is particularly high on more productive firms, which tend to be larger and are more likely to be foreign-owned and exporters (see Annex II). This acts as a tax on efficiency. In addition to discouraging firms from becoming more efficient, it also will discourage them from taking actions that might signal their efficiency, such as entering export markets. In addition, the government should combat corruption in tax administration. Despite recent reform efforts, 21 percent of enterprises that had required meetings with tax inspectors reported that gifts or informal payments to inspectors were expected or requested.30 The corresponding figures for the comparator countries were 7 percent in Uganda, 21 percent in China, and 38 percent in Kenya. The median value of the gift/informal payment was TSh 400,000 (about $400 in mid-2003). I.2 Power Although recent reforms may pay off in the future, the state of Tanzania's infrastructure has been a serious problem in recent years. A study by the International Finance Corporation's operations evaluation group reported that "Tanzania is uncompetitive even compared to other sub-Saharan African countries, because of unreliable and high-cost electricity, poor roads, high transportation costs, poor communications, and a largely agrarian labor force lacking the skills, training, or work experience required by modern, private enterprises." Within the infrastructure sector, the power sector generates the greatest concern. In the manufacturing sector, 59 percent of managers reported that electricity was a major or very severe obstacle to enterprise operations and growth (Figure 3.5). (The corresponding numbers for the construction and tourism sectors were 46 and 66 percent.) By contrast, only 23 percent and 12 percent of managers in manufacturing rated transportation and telecommunications similarly. This pattern was repeated in most of the comparator countries, with the exception of Kenya, where managers ranked telecommunications as a more serious problem than transportation. Figure 3.5 Power is the infrastructure area that managers in Tanzania were most concerned about 70% 59% 60% 49% 50% 44% 44% 40% 37% 28% 30% 23% 23% 19% 20% 17% 12% 10% 5% 0% Tanzania Uganda Kenya China Telecommunications Transportation Power Source: Investment Climate Surveys. While the rankings of the three areas were similar across countries, there were differences in the perceived severity of each constraint. Managers in Tanzania were more likely than managers in the comparator countries to report that electricity was a major or very severe problem (59 percent in Tanzania, compared to 28 percent in China, for example). But they were less likely than managers in Kenya and China to report that telecommunications was a major or very severe obstacle. They were also less likely than managers in Kenya to report that transportation was a major or very severe problem. There were also some differences within Tanzania by firm size. Micro and small enterprises were less likely than larger enterprises to find access to electricity to be a major or very severe constraint. Whereas over 60 percent of medium, large, and very large enterprises found electricity to be a major constraint, only 39 percent of informal micro enterprises, 52 percent of formal micro enterprises, and 44 percent of small enterprises found electricity a major constraint (Table 3.1). Access. Most enterprises in the Investment Climate Survey reported that they had access to electricity. Of 276 enterprises, only 4 (2 small and 2 medium) reported that they did not. But many micro enterprises--more than half--were not connected to the power grid (Figure 3.6). Enterprises without connections were far less likely to find electricity to be a major or very severe constraint (24 percent) than enterprises with connections (about 70 percent for both formal and informal micro enterprises). This suggests that most complaints about the sector focus on reliability or price, not availability of service. Figure 3.6 Less than half of micro enterprises--and even fewer informal micro enterprises-- are connected to the power grid 60% power 48% to 41% 40% 36% connected 20% enterprises of % 0% All micro Formal micro Informal micro Source: Investment Climate Surveys. Technical performance. The technical performance of the power sector is poor. In 2002, the median enterprise in Tanzania suffered power outages on 48 days, compared to 21 days in Kenya and 20 in Uganda. Outages and other problems impose real costs upon manufacturing firms. The median estimate for losses in production due to power outages and surges was 5 percent of production--considerably higher than in the comparator countries (Figure 3.7).31 In addition, 55 percent of managers in Tanzania reported that power fluctuations had damaged or destroyed some of their enterprise's equipment. The median loss was TSh 2 million, or 0.6 percent of the replacement value of plant, machinery, and equipment for the average enterprise in this group. Enterprises can reduce their vulnerability to outages by operating their own generators. But generators are expensive to buy and use. Among enterprises with generators, the median firm paid TSh 14 million for the generator. This represented 2.5 percent of the replacement value of plant, machinery, and equipment for the average enterprise within this group. Enterprises in Tanzania were more likely to have their own generators (55 percent of enterprises) than enterprises in China (27 percent) and Uganda (35 percent), but less likely to have generators than enterprises in Kenya and India (70 and 71 percent respectively) (Figure 3.21). Within Tanzania, the likelihood of owning a generator varied by sector. Firms in tourism were far more likely to have generators (74 percent of firms) than enterprises in manufacturing (55 percent) and construction (51 percent). This may be because power interruptions are more costly for firms in tourism. Unlike firms in manufacturing and construction, tourism enterprises cannot make up lost production by working additional shifts after power is restored. Figure 3.7 Performance of the power sector 80% 6.0% 70% 71% 70% 5% 5.0% 60% 55% outages) 4.0% 50% 3% power to generators) 40% 35% 3.0% due with 30% 27% 2.0% (% 20% losses 10% 1.0% 0% 0% (median 0% 0.0% Tanzania China Uganda Kenya India* * No data from India on lost sales due to power outages Source: Investment Climate Surveys. Not surprisingly, very few micro enterprises--less than 1 percent--owned or shared a generator (Table 3.5). (The figure was the same for micro enterprises with and without connections to the power grid.) But 29 percent of small enterprises and 88 percent of very large enterprises owned or shared a generator. Table 3.5 Performance of power sector, by enterprise size Very All micro Small Medium Large large Percentage of enterprises with generators 0.8 29.1 57.6 81.5 88.0 Number of days to get power connection 30 30 30 21 3 Number of days with power outages (median) 36 38 60 48 35 Percentage of sales lost due to power outages (median) 6.0 2.7 10.0 7.0 3.5 Source: Investment Climate Surveys. In other respects, there were no consistent differences by enterprise size. Micro enterprises with power connections experienced outages on about 36 days, compared to 38 for small enterprises, 60 for medium enterprises, 48 for large enterprises and 35 for very large enterprises. Consistent with this data, small and very large enterprises reported the smallest outage-related losses, and medium and large enterprises reported the largest losses. Prices. Although the cost of power is a concern for enterprises in Tanzania, the average cost per kilowatt hour (KWH) that enterprise managers reported was similar to price levels in most of the comparator countries. In mid-2003, the median cost per KWH was about $0.086 in Tanzania, $0.085 in China, $0.084 in Uganda, and $0.101 in Kenya. Capacity. Inadequate maintenance and the poor commercial performance of the Tanzania Electricity Supply Company (TANESCO) in the 1990s explain many of the electricity sector's problems. But at 0.02 kilowatts per capita, generating capacity is also lower in Tanzania than in most of the comparator countries. Generating capacity is 0.01 kilowatts per capita in Uganda, 0.03 kilowatts per capita in Kenya, 0.11 kilowatts per capita in India, and 0.25 kilowatts per capita in China. Tanzania's generating capacity increased from about 0.408 million KW in 1980 to 0.862 million KW in 2001--a 111 percent increase.32 But over the same period, the country's population grew from 18.6 million to 34.5 million--an 85 percent increase. In April 2002, Net Group Solutions, a South African management consultancy, assumed management of TANESCO. The transfer has already yielded some results in terms of improved commercial performance. However, further improvements in technical performance remain necessary. I.3 Finance Countries with well-developed financial systems (banks, stock markets, and bond markets) tend to grow faster than countries with less well-developed systems.33 Similarly, access to financial institutions appears to promote firm growth. Enterprises in Tanzania that had overdraft facilities expanded their sales and employment more rapidly than other firms (Annex 2).34 Tanzania has recently made considerable strides in developing its banking sector. The government has privatized several state-owned banks and allowed foreign banks to enter the Tanzanian market.35 But enterprises continue to report that access to finance and high interest rates are serious problems. About 58 percent of enterprises reported that cost of financing was a major or very severe constraint on enterprise operations and growth, and 48 percent reported the same for access to finance. Cost of financing ranked third among the constraints most frequently identified as major or very severe problems, and access to finance ranked sixth. Financial sector development. Despite recent reforms, financial markets are less developed in Tanzania than in most of the comparator countries. Claims on the private sector equal only 5 percent of GDP in Tanzania.36 In comparison, claims on the private sector are 6 percent of GDP in Uganda, 23 percent in Kenya, 29 percent in India, and 127 percent in China. Similarly, money and quasi-money, another common measure of financial sector development, equals only 20 percent of GDP in Tanzania, compared to 18 percent of GDP in Uganda, over 39 percent in Kenya, 58 percent in India, and 168 percent in China.37 In addition, Tanzania's stock market is embryonic. In 2002, only five companies were listed on the exchange, total market capitalization was equal to about 0.2 percent of GDP, and turnover was equal to only about 7.6 percent of GDP. In comparison, market capitalization was 0.6 percent of GDP in Kenya and 66 percent in India, and turnover was 18 percent of GDP in Kenya and 105 percent in India. Under these circumstances, it is not surprising that very few companies financed any investment by issuing equity. Of the 255 enterprises in Tanzania that reported how they financed investment and working capital, only 12 had any equity financing. Cost of finance. The cost of financing appears to be in line with rates elsewhere in the region. Based on reporting by enterprises with bank loans, the median interest rate was 13 percent in Tanzania--slightly lower than in Kenya (15 percent) and significantly lower than in Uganda (18 percent). The median loan period was three years in all three countries. Since inflation was slightly lower in Tanzania than in Uganda and Kenya, real interest rates were fairly close in the three countries in 2003. However, the median interest rate reported by enterprises in China was significantly lower than in the three African countries--5.9 percent with inflation of about 1.2 percent.38 Figure 3.8 Median interest rates are similar for micro enterprises and other firms, but median loan periods are shorter for micro enterprises 18 17 3.5 16 3 3 3 16 14 3 13 years rate 14 2.5 in 12 2 10 8 2 interest period 8 1.5 6 1 loan 1 Median 4 2 0.5 Median 0 0 Micro Small Medium Large Very large Interest rate Period Source: Investment Climate Surveys. Within Tanzania, perceptions of the cost of the financing did not vary systematically by firm size. Between 54 percent and 63 percent of enterprises rated cost of financing as a major or very severe problem (Table 3.1), and the median enterprise in most size groups reported paying between 13 and 17 percent on its most recent loan (Figure 3.8).39 (The median very large enterprise reported paying 8 percent.) Micro enterprises did not face higher rates than other firms, but they tended to have shorter-term loans. The median micro enterprise reported a loan period of one year, compared to two to three years for larger firms. Access to finance. Only 20 percent of enterprises reported that they had borrowed from a financial institution (Figure 3.9). This was similar to the rate in Uganda, but significantly lower than in Kenya (40 percent of enterprises) and China (57 percent). About 31 percent of enterprises in Tanzania reported having overdraft facilities--slightly higher than in Uganda (23 percent) and China (27 percent), but lower than in Kenya (73 percent). There was considerable overlap between the enterprises with overdraft facilities and enterprises with loans. Roughly 75 percent of enterprises with loans had overdraft facilities, compared to only 20 percent of enterprises without loans. About 63 percent of enterprises had neither loans nor overdraft facilities. Figure 3.9 Fewer firms have loans in Tanzania than in most of the comparator countries 80% 73% 70% 60% 57% 50% 40% 40% 31% 30% 27% 20% 23% 20% 20% 10% 0% Tanzania Uganda Kenya China Overdraft Loan Source: Investment Climate Surveys. Small and micro enterprises were more likely to rate access to finance as a serious problem (Table 3.1).40 Only 4 percent of small enterprises had loans and only 13.7 percent had overdraft facilities (Table 3.6). (The corresponding figures for very large enterprises were 48 and 64 percent.) In addition, small enterprises were more than twice as likely as very large firms to report that they were credit-constrained (i.e., that they would borrow more at current interest rates if they could). Table 3.6 Large and very large firms are more likely to have overdraft facilities and loans and less likely to report being credit-constrained Percent of firms Overall Small Medium Large Very large Has overdraft facility 31.4 13.7 32.8 44.6 64.0 Has loan 19.9 4.0 22.4 26.8 48.0 Is credit-constrained 44.8 64.3 48.1 34.2 40.0 Note: An enterprise is considered credit-constrained if it reported that it would borrow more at current interest rates if it could. Source: Investment Climate Surveys. Responses to questions about access to financing also varied by sector (Figure 3.10). In the construction and tourism sectors, 80 percent and 73 percent of enterprises rated access as a serious problem, compared to only 48 percent in the manufacturing sector. Despite this gap, manufacturing enterprises were less likely to have loans (20 percent) than firms in construction (50 percent) and tourism (58 percent). However, manufacturing firms were more likely to have overdraft facilities (31 percent) than firms in the other two sectors (17 percent in construction and 18 percent in tourism). Figure 3.10Construction firms are more likely than manufacturing firms to have loans 70% 60% 58% overdraft 50% 50% or 40% loans 31% 30% with 20% 20% 17% 18% firms 10% of & 0% Construction Tourism Manufacturing Loans Overdraft Source: Investment Climate Surveys. Sources of finance. Firms in Tanzania finance very little investment through the formal financial sector. Only 16 percent of investment is financed through bank lending, compared to 14 percent in Uganda and 32 percent in Kenya (Figure 3.11). Another 8 percent of investment is financed through other (non-bank) external financing, including non-bank financial institutions (investment funds, special development financing, government financing, NGO financing, and donor financing), trade credit, equity financing and leasing arrangements. But this average obscures the fact that very few enterprises have access to financing from most of these sources. For example, only two enterprises received money from non-bank financial institutions, but both used this financing to fund more than 70 percent of their investment. As a result, non-bank financial institutions accounted for close to 1 percent of investment for the average firm. Similar patterns can be observed for equity financing and other sources of external financing: a very small number of enterprises financed a lot of investment through these mechanisms, while most received no financing from them at all. Under these circumstances, it is not surprising that most enterprises finance investment primarily through retained earnings. Retained earnings account for 68 percent of new investment in Tanzania--slightly less than in Uganda (71 percent), but considerably more than in Kenya (54 percent). More than half of all enterprises (56 percent) with any new investment reported relying solely on internal funds. Another 11 percent of enterprises relied on a mix of internal funds and informal sources. In other words, more than two-thirds of the enterprises that invested did not obtain financing through the formal financial sector. Figure 3.11 Tanzanian firms rely more than Kenyan firms on retained earnings for investment 100% 8% 8% 9% 14% 80% 16% 32% 8% 7% 60% 5% 40% 68% 71% 54% 20% 0% Tanzania Uganda Kenya Internal funds Other external financing Banks Informal Note: Other external financing include leasing arrangements, investment funds, NGO/government/donor sources, trade credit, credit cards, and equity. Informal is family, friends, informal sources, and "other." Totals may not equal 100% due to rounding. Source: Investment Climate Surveys. Firms of all sizes used retained earnings and other internal funds to finance more than 50 percent of investment (Figure 3.12). But there were some significant differences by enterprise size. Large and extra large enterprises used bank lending to finance 20 and 31 percent of investment respectively. In contrast, bank loans accounted for 5 percent of investment at small enterprises, 2 percent at formal micro enterprises, and 0 percent at informal micro enterprises. Figure 3.12 Only large and very large firms finance large amounts of investment through the formal banking sector 0 2 100% 5 6 10 different 4 20 31 80% 40 40 9 6 1 from 12 2 60% 4 2 15 40% 81 80 financed sources 66 56 57 51 20% 0% investment Informal Formal Small (10- Medium (20- Large (50- Very large of micro (0-9) micro (0-9) 19) 49) 249) (250 and % up) Internal funds Other external financing Informal Banks Note: Other financial institutions include leasing arrangements, investment funds, NGO/government/donor sources, trade credit, credit cards, and equity. Informal is family, friends, informal sources, and "other." Totals may not equal 100% due to rounding. Source: Investment Climate Assessments. Despite recent initiatives to expand access to credit for micro enterprises through NGOs and other financial institutions, very few micro enterprises appear to use such loans. On average, informal and formal micro enterprises financed only 2 and 4 percent of investment with other external financing--considerably less than the 15 percent of investment that very large enterprises financed in this way. Within the category of other external financing, financial NGOs were the most important source of financing for informal micro enterprises (accounting for 3 percent of investment), and trade credit and "other" were the most important sources for formal micro enterprises (each contributing around 1 percent of investment). For large and very large enterprises, stock sales and trade credit were the most important sources of financing within the category of "other external financing." Only one medium enterprise reported any financing from a broad group of sources that included NGOs.41 Micro enterprises and the financial sector. Nearly half of the surveyed micro enterprises reported that they had deposits in banks, financial NGOs, or other informal financial institutions (Table 3.7). But only 19 percent held deposits with banks, and 2 percent had deposits with NGOs. A far greater percentage--about 30 percent--had deposits in other/informal institutions, such as savings and credit cooperatives (SACCOs) and money lenders. Micro enterprises in rural areas were less likely than urban ones to have bank deposits, but more likely to have other types of deposits. Similarly, informal micro enterprises were less likely than formal ones to have bank deposits, but equally likely to have deposits with NGOs or other/informal institutions. Table 3.7 Access to bank finance is a serious problem for micro enterprises, especially in rural areas Percentage of firms All micro Formal Informal Rural Urban Rate obstacle as major or very severe Access to finance 71% 72% 69% 75% 63% Cost of finance 55% 57% 54% 61% 43% Any deposits 48% 51% 46% 45% 53% Bank 19% 26% 15% 13% 31% NGO 2% 1% 2% 2% 1% Other/informal 30% 31% 30% 33% 24% Any loan 7.8% 7.9% 7.7% 8.6% 6.3% Bank 1.5% 1.3% 1.5% 2.2% 0.0% NGO 5.8% 5.2% 6.2% 5.7% 6.3% Other/informal 0.5% 1.3% 0.0% 0.7% 0.0% Note: Totals may not equal 100% due to rounding and because firms may have deposits in more than one type of institution. Source: Investment Climate Survey. Very few micro enterprises--0 to 2 percent--reported having bank loans. There was little difference in this regard between rural and urban enterprises or formal and informal ones. The proportion of enterprises that reported having loans from NGOs was slightly greater--5 to 6 percent--but still small. I.4 Corruption Compared to other low-income countries, Tanzania performs well on many measures of governance. In a recent study that rated nearly 200 countries, Tanzania ranked above the average for low-income countries on five of the six measures (Box 3.1, Figure 3.13). It ranked in the 39th percentile overall for rule of law, the 38th for voice and accountability, the 36th for political stability and government effectiveness, and the 34th for regulatory quality. However, on one dimension--corruption--Tanzania performed poorly, falling below the average for low-income countries. Corruption was worse in Tanzania (16th percentile) than in Uganda (19th percentile), China (42nd percentile), or India (50th percentile). In fact, of the comparator countries, only Kenya had a lower ranking (11th percentile).42 Box 3.1: Different aspects of governance Many researchers and practitioners have tried to produce aggregate statistics that make it possible to compare the quality of governance across countries and over time. Few of these studies cover the entire world or all topics. Furthermore, the questions used to elicit responses are usually not comparable across surveys. In order to increase country coverage, Kaufmann, Kraay, and Mastruzzi (2003) 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 government's commitment to announced policies. Regulatory quality: the quality of government policies. This measure focuses on the prevalence of market-unfriendly policies, such as price controls or inadequate bank supervision, and on perceptions about the regulatory burden facing businesses. 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 of contracts. Control of corruption: the extent of corruption (i.e., the illegal use of public power for private gain). Figure 3.13 Tanzania does well compared to low-income countries on most governance measures--except corruption, 2002 Note: Thick, even-shaped hexagon depicts the average scores for low-income countries. Thick, uneven-shaped hexagon depicts scores for Tanzania. Dotted lines are 90 percent confidence intervals for estimates for Tanzania. Source: Kaufmann, Kraay, and Mastruzzi (2003). The government appears to have made some progress in reducing corruption since 1996, when a presidential commission led by Judge Wairoba produced the Wairoba Report.43 It has strengthened the institutional framework--notably through the Finance Act of 2001 and the Public Procurement Act of 2002--and adopted a clear zero-tolerance position on corruption.44 Perhaps as a result, between 1998 and 2002, Tanzania's ranking improved from the 9th percentile to the 16th percentile (Kaufmann et al., 2003).45 But given the imprecision of these estimates, this improvement may not be statistically significant. Moreover, corruption clearly continues to have an important impact on businesses. Grand corruption. Enterprise managers in Tanzania see grand corruption-- payments made to policymakers or senior bureaucrats in order to win government contracts and influence lawmaking--as a serious problem (Figure 3.14). Nearly 45 percent of managers said that payments to government officials that affected the content of government decrees had at least a moderate impact on their business, while 47 percent said the same of payments to members of parliament that affected their votes. These figures were higher than in Kenya and Uganda. Figure 3.14 Enterprises in Tanzania are more likely than firms in comparator countries to report that bribes to government officials and parliamentarians affect their business a 60% have 47% 45% bribes impact 40% 40% 40% 33% 30% greater 26% reporting or 20% 20% enterprises moderate of % 0% Tanzania Uganda Pakistan Kenya Source: Investment Climate Assessments. Informal payments to secure government contracts also appear common in Tanzania. Of enterprises that did business with the government, about 33 percent reported that unofficial payments were needed to secure government contracts. Within the subset of firms reporting that payments were needed, the median firm reported that about 10 percent of the contract value was needed to secure government work. This was similar to the median amounts reported in Uganda and Kenya (both 10 percent), but significantly higher than the median amount reported in China (2 percent). The amount paid in bribes is strongly correlated with firm performance--firms with faster sales growth and higher productivity pay larger bribes (Annex 2).46 Because corrupt officials often target the most productive and profitable enterprises, corruption can act as a tax on efficient firms. Petty corruption. Enterprises are also affected by petty corruption--payments made to lower-level government officials to "get things done" in connection with customs, taxes, licenses, and other services. About 35 percent of enterprise managers said that informal payments were typically needed for firms like theirs. Of the enterprises that reported that informal payments were needed, the median payment was about 0.3 percent of sales. Bribes were fairly common for many transactions, such as getting utility connections and applying for import licenses (Table 3.8). Table 3.8 Micro enterprises were less likely to report that bribes were needed to get things done Micro Small Medium Large Very large Percentage reporting bribes needed to "get things done" 23 30 44 41 36 Percentage saying gift/payment required for Mainline telephone connection -- 26 27 14 0 Electrical connection 27 32 35 27 14 Water connection 33 22 24 13 0 Import license -- 0 19 12 9 Operating license 34 15 30 14 6 Source: Investment Climate Surveys. Informal payments and gifts were also common during inspections and mandatory meetings with government officials. About 21 percent of enterprises that had inspections or meetings with officials from the Tanzania Revenue Authority (TRA) reported that bribes were requested, with the median amount being TSh 375,000 (Figure 3.15). Similarly, 25 percent of enterprises dealing with the municipal police reported requests for bribes. The median payment in this case was TSh 80,000. About 18 percent of enterprises reported requests from labor or social security officials, 13 percent from health and sanitation inspectors, 8 percent from environmental inspectors, and 7 percent from fire and building safety inspectors. Payments to these other inspectors were generally lower than payments to tax officials. Figure 3.15 Requests for bribes are especially common during tax inspections 30% 375 25% 400 21% 300 20% 18% 13% 200 enterprises) 8%115 thousands) of 10% 110 7% 75 50 80 100 (% (TSh 0% 0 Tax security Sanitation police Fire/building &social Municipal Environmental Labor Percentage reporting bribe requests Median amount Note: "Percentage reporting bribe requests" only includes enterprises that were inspected; "median account" only includes enterprises that reported positive amounts Source: Investment Climate Assessments. Micro enterprises are less likely than larger enterprises to find corruption a major or very severe constraint. Only 18 percent of informal micro enterprises and 30 percent of formal micro enterprises found corruption to be a serious problem, compared to between 48 and 58 percent of small, medium, and large enterprises. Micro enterprises were particularly unlikely to face requests for bribes. Only 15 percent of informal micro enterprises reported that bribes were needed to "get things done," compared to 38 percent of formal micro enterprises. Although micro enterprises were less likely to report that bribes were needed to "get things done," they were no less likely to report that bribes were needed for typical transactions. For example, 34 percent of micro enterprises reported that bribes were needed to get an operating license, compared to 15 percent of small, 30 percent of medium, 14 percent of large, and 6 percent of very large enterprises. The most plausible explanation is that micro enterprises are far less likely to interact with bribe-taking institutions, but at least as likely to encounter corruption when they do. For example, only 5 percent of micro enterprises reported getting an electricity connection in the two years prior to the survey, compared to 30 percent of large enterprises (Figure 3.16). Micro enterprises were also less likely to get operating licenses, water connections, and fixed-line telephone connections. Figure 3.16 Micro enterprises are far less likely to interact with institutions that demand bribes 100% 80% connection 67% 60% obtained 40% 33% 30% 25% 25% license that or 20% 14% 1% 5% 2% 0% Mainline Electrical Water Import license Operating enterprises telephone connection connection license of connection % Small, medium, and large enterprises Micro enterprises Source: Investment Climate Surveys. I.5 Macroeconomic Instability Despite recent improvements in macroeconomic performance, enterprise managers in Tanzania remain concerned about macroeconomic stability. About 43 percent rated macroeconomic stability as a major or very severe obstacle to enterprise operations and growth. In comparison, 51 percent rated it a major or very severe obstacle in Kenya, 45 percent in Uganda, and 26 percent in China. These perceptions are rooted in concern about such factors as inflation. In 2003, inflation in Tanzania was lower than in Kenya and Uganda but higher than in China (Figure 3.17). (Inflation in Tanzania was higher than the average rate in Kenya and Uganda from 2000 to 2002.) Although inflation has fallen in Tanzania over the past decade, the possibility that it might return remains a serious concern. The high interest rates observed in Tanzania-- and which are a serious concern to enterprises of all types--reflect this. The median interest rate that enterprises in Tanzania reported paying on their most recent loans was similar in real terms to the corresponding figures in Kenya and Uganda, but considerably higher than the rates observed in China. Until enterprises and bankers in Tanzania are convinced that the recent reductions are permanent, interest rates will remain high. Figure 3.17 In 2003, inflation was lower in Tanzania than in Kenya and Uganda, but higher than in China or India 18.0 16.5 16.0 14.0 14.0 12.0 9.8 10.0 9.3 7.9 7.7 8.0 7.2 5.9 6.0 5.2 4.4 4.0 3.8 4.0 1.5 2.0 1.2 0.4 0.0 Tanzania China Uganda India Kenya 1992-2002 2000-02 2003 Source: World Bank (2003b). Continued efforts to keep inflation under control and maintain economic growth are thus important from an investment climate perspective. Because of this, fiscal stability must remain a top priority. For example, reducing tax rates without taking steps to maintain fiscal stability might lead to medium-term macroeconomic instability, which would be costly for enterprises. I.6 Customs and Trade Regulations There is broad consensus that the barriers discussed above constrain operations and growth for most enterprises in Tanzania. Other barriers, however, impose significant burdens only on certain classes of enterprises. One prominent example is customs and trade regulations. Over 52 percent of very large enterprises rated customs and trade regulations as a major or very severe constraint--making it their fifth-greatest concern-- compared to 10 percent of informal micro enterprises, 18 percent of formal micro enterprises, and 22 percent of small enterprises. This is not surprising, given that larger firms are much more likely to engage in foreign trade. About 77 percent of very large enterprises exported some portion of their output, compared to only 3 percent of small enterprises. Similarly, 83 percent of very large enterprises directly imported some raw materials, compared to 18 percent of small enterprises. Exporting and importing in Tanzania are both hampered by poor customs administration. Among enterprises that engaged in foreign trade, the median firm reported that it took 14 days on average for imports and 7 days for exports to clear customs, once they had reached the point of entry or exit (Figure 3.18).47 The median wait was considerably shorter than the average typical wait because several enterprises reported exceptionally long delays.48 Managers also reported that clearance times were unpredictable, forcing them to hold additional inventory in anticipation of worst-case scenarios. For imports, the median enterprise reported that the longest delay it had faced was 21 days--considerable longer than the 14-day typical wait. For exports, the median enterprise reported that the longest delay it had faced was 12 days. Figure 3.18 Customs clearance times in Tanzania are long and unpredictable 35 33 30 25 21 22 20 18 Days15 14 12 12 10 7 5 0 Imports - typical Imports - longest Exports - typical Exports - longest wait wait wait wait Mean Median Source: Investment Climate Survey. Port and customs delays were considerably longer in Tanzania than in any of the comparator countries (Figure 3.19). The median delays for imports and exports in China were five and three days respectively--less than half the delays facing firms in Tanzania. Similarly, reported delays for imports and exports were seven and four days respectively in Kenya and seven and three days respectively in India. These delays have real costs. As discussed in the previous chapter, poor customs administration and overly restrictive trade and customs regulations discourage enterprises from exporting. Indeed, as noted previously, trade and customs regulation partially explain why enterprises in Tanzania export less than similar enterprises in Kenya. Since exporting has been linked to improved productivity--as well as an improved balance of trade--these delays and restrictions can have a real impact on enterprise performance. Figure 3.19 Customs clearance times are longer in Tanzania than in the comparator countries 16 14 12 8 7 7 7 Days 5 4 4 4 3 3 2 0 Tanzania Kenya Uganda India China Imports Exports Source: Investment Climate Surveys. II. CONFRONTING INFORMALITY As discussed previously, tax evasion and informality appear to be higher in Tanzania than in the comparator countries. Despite similar tax rates, revenues are lower in Tanzania than elsewhere in sub-Saharan Africa, and enterprises in Tanzania generally report a smaller share of revenues to tax authorities than enterprises in the comparator countries. Other estimates of the size of the informal economy, which take a very different approach towards estimating informality than the estimates above, also suggest that it is relatively large--measuring about 58 percent of gross national income (GNI) in Tanzania, compared to 43 percent in Uganda, 34 percent in Kenya, and much less in China and India (Figure 3.20).49 Governments can reduce informality by reducing the costs and increasing the benefits associated with becoming formal. By reducing corruption, for example, they can increase enterprises' willingness to deal with public institutions. In addition, governments can encourage firms to become formal by reducing the burden imposed upon formal enterprises by such factors as barriers to entry, business regulations and inspections, and labor regulations. Figure 3.20 Informality is higher in Tanzania than in the comparator countries 80 GNI) 58 of 60 % (as 43 40 34 economy 23 20 13 Informal 0 Tanzania China India Kenya Uganda Source: Doing Business database (World Bank, 2003a). II.1 Barriers to Entry Although some entry restrictions are necessary--requiring enterprises to register with tax authorities, for example--others do more harm than good. Several recent studies have shown that entry restrictions encourage firms to remain informal by making it difficult and expensive to enter the formal sector.50 Entry restrictions can also lead to corruption and reduce competition, thereby hurting economic performance in other ways.51 Within the OECD, for example, multifactor productivity and investment are significantly lower in sectors with greater restrictions on entry.52 Most enterprises in Tanzania did not rate business licensing as a serious obstacle to enterprise operations and growth. Only about 27 percent of enterprises said that it was a major or very severe problem (Figure 3.1). But these results probably understate the impact of such regulations. Businesses that are already operating are much less likely to view entry restrictions as an obstacle. Indeed, they may even welcome regulations that limit competition. For this reason, in discussing entry restrictions, we supplement the evidence from the Investment Climate Survey with evidence from the World Bank's Doing Business database. According to this data, in January 2004, it took about 35 days to fulfill all legal requirements for starting a business in Tanzania. Although this is considerably longer than the 5 days the same process takes in Australia, Canada, New Zealand, the United States, and Denmark (data for other OECD countries were not available), it is not exceptionally burdensome when compared to other developing countries.53 For example, business licensing takes about 41 days in China, 89 days in India, 47 days in Kenya, and 36 days in Uganda. But the monetary cost of this process as a percentage of per capita gross national income is far higher in Tanzania (204 percent) than in any of the other comparator countries. In contrast, business licensing costs 15 percent of per capita GNI in China, 50 percent in India, 54 percent in Kenya, and 132 percent in Uganda. The most expensive procedures in Tanzania are getting the certificate of incorporation ($249), getting a business license ($208), and getting a company seal ($100) (Table 3.9). Table 3.9 Entry procedures in Tanzania are time-consuming and very costly, January 2003 Duration Nature of procedure (2003) Procedures (days) Cost (US$) Apply for clearance of the proposed 1 3 0 company name Get a certificate of incorporation 2 7 249.36 Get a company seal 3 1 100 Get a business license 4 9 207.8 Get signature of town inspector 5 0* 0 Get signature of health inspector 6 0* 0 Apply for TIN 7 1 0 Register for VAT 8 1 0 Register for PAYE 9 1 0 Inspection by income tax officials 10 1 0 VAT/stamp duty inspection 11 3 0 Workmen's compensation insurance 12 1 0 National provident fund 13 7 0 Total: 13 35 $557.16 Source: World Bank (2003a). Legal restrictions on entry are not the only barrier to entry that new firms face. Access to infrastructure--especially electricity--is another important obstacle (Figure 3.21). The median wait for an electricity connection was 30 days in Tanzania, compared to 3 days in China, 14 days in Uganda, and 15 days in Kenya. Micro, small, and medium enterprises faced much longer median delays (30 days) than large enterprises (21 days) and very large enterprises (3 days). Obtaining a water connection also took longer in Tanzania than in any of the comparator countries, but the median delay--7 days--was far shorter than the wait for electricity. Getting a telephone connection was somewhat more time-consuming, but the median wait of 14 days was considerably shorter than in Kenya (60 days). Figure 3.21 Waits for electricity, water, and fixed-line telephone connections also create entry barriers 70 60 60 50 40 30 30 20 14 14 15 10 7 3 7 7 0 2 0 Tanzania China Uganda Kenya Days wait for electicity connection (past two years) Days wait for water connection (past two years) Days wait for phone lines (past two years) Note:: Only firms that received a fixed-line connection in the two years prior to the survey responded to the question about telephone service. Source: Investment Climate Surveys. II.2 Labor Regulations Hiring and firing restrictions can also discourage enterprises from entering the formal sector. According to data collected by the World Bank for the Doing Business database, the laws in these areas are stricter in Tanzania than in most of the comparator countries. Tanzania's hiring and firing restrictions were rated 57 and 49 respectively, based on an index that links tougher regulations to higher numbers. The corresponding scores for the comparator countries were 17 and 57 for China, 33 and 45 for India, 33 and 16 for Kenya, and 33 and 50 for Uganda. Despite Tanzania's strict labor laws, very few enterprises (only 12 percent) rated labor regulations as a serious obstacle to enterprise operations and growth. In part, this may be because labor laws are poorly and unevenly enforced. Other evidence supports this explanation. Very large enterprises, which are most likely to face strict monitoring, were far more likely to rate labor laws as a major or very severe obstacle (24 percent) than small enterprises (9 percent) or informal micro enterprises (4 percent) (Table 3.1). Consequently, in addition to making labor laws more flexible, the government should make enforcement more even and effective. II.3 Business Regulations and Inspections Business regulations and inspections are a third set of factors that might encourage enterprises in Tanzania to operate informally. On average, enterprises in Tanzania reported that senior managers spent about 15 percent of their time dealing with government inspections, regulations, and paperwork. This was slightly higher than in Kenya (13 percent) and China (12 percent) and significantly higher than in Uganda (4 percent). Similarly, the median enterprise in Tanzania reported 15 inspections or meetings with government agencies (including tax officials) per year, significantly more than in Uganda and India (5 and 6 inspections respectively). However, enterprises in Kenya and China both reported a slightly greater number of inspections (16). Micro enterprises appear to avoid a significant part of this burden by remaining informal. Managers in informal micro enterprises spent only 3 percent of their time dealing with government regulations, compared to 13 percent in formal micro enterprises (Figure 3.22). In general, the larger the enterprise, the more time managers spent dealing with government regulations.54 Figure 3.22 Informal micro enterprises avoid much--but not all--of the burden of regulation 20 19 19 with 16 15 dealing 13 13 time 10 regulation 5 3 management of % 0 Micro Micro formal Small Medium Large Very large informal Source: Investment Climate Surveys. Tax administration was the area of regulation that enterprises were most likely to identify as a serious problem. More than half of all enterprises rated it as a major or very severe obstacle, and 92 percent reported that they had required meetings with tax inspectors, with the median firm within this subset reporting seven meetings per year (Figure 3.23). Fewer enterprises reported having required meetings or inspections from other agencies, and the number of required meetings was lower for those that did. Figure 3.23 Percentage of firms with inspections and number of inspections per year, by government agency 100% 92% 9 7 80% 66% 6 60% 50% 52% 4.5 4 48% 40% 31% 2 2 2 3 20% 0% 0 Tax Labor/social Fire/building Santitation Municipal Environmental inspectorate security safety police Firms reporting inspections/required meetings Median number of meetings Note:: Number of visits is the median number for firms with inspections. Source: Investment Climate Surveys. Overall, the impact of business regulations does not appear to vary greatly by sector. The vast majority of establishments are inspected by government agencies. For example, more than 90 percent of manufacturing firms, 85 percent of tourism establishments, and 71 percent of construction firms reported tax inspections (Table 3.10). But firms in the construction sector were less likely to be inspected than enterprises in other sectors, while firms in the manufacturing sector were most likely to be inspected. For firms that were inspected, the average number of inspections was similar across sectors. For example, the median firm in all sectors received seven inspections by the tax inspectorate in 2002. Table 3.10 Percentage of firms with inspections and median number of inspections, by sector Social Tax Building Security Safety Police Inspectorate Sanitation Municipal Labor/ Fire/ Environmental Percentage of firms with inspections Construction 71 30 19 19 21 19 Tourism 85 58 23 48 18 32 Manufacturing 92 66 50 52 31 48 Median number of inspections for firms with inspections Construction 7 2 2 3 5 2 Tourism 7 3 1 3 5 2 Manufacturing 7 4 2 2 5 2 Source: Investment Climate Survey. Finally, the bureaucratic burden associated with regulation is much higher for exporters than those who do not export. The number of inspections and days spent dealing with regulatory issues are higher for exporters (47 and 20) than non-exporters (28 and 14). This may further discourage exporting. III. OTHER CONSTRAINTS ON ENTERPRISES OPERATIONS AND GROWTH One concern about perception data, such as the data collected by the Investment Climate Surveys, is that enterprises that have adapted to certain constraints may not see them as problematic. For example, many micro enterprises without access to power do not consider electricity to be a serious obstacle to their operations and growth--even though access to power would probably improve their performance. Similarly, non- exporting firms are less likely to find transportation a significant constraint. Consequently, average perceptions of transportation may be more positive in a country where low-quality transportation prevents most enterprises from exporting than in a country where better transportation allows more firms to engage in external trade. Such examples suggest that it is worthwhile to examine potential problem areas that enterprises by and large did not identify as major constraints. These include the legal system, access to land, telecommunications, and transportation. III.1 Legal System Few enterprises in Tanzania--only 20 percent--viewed the legal system as a major or very severe obstacle to enterprise operations and growth (Figure 3.15). Labor regulations and telecommunications were the only areas that ranked lower among enterprises' concerns. This finding is consistent with other studies, which have generally found that the "rule of law" is relatively well established in Tanzania--at least when compared to other low-income countries (Figure 3.13). On this measure, Tanzania scored in the 39th percentile among developing and developed economies, outperforming Kenya (14th percentile) and Uganda (21st percentile). But Tanzania did less well than the faster-growing comparator countries, China (52nd percentile) and India (57th percentile). Consistent with this evidence, the court system in Tanzania appears to be relatively efficient. Among enterprises that had been involved in lawsuits regarding payment disputes, the median firm reported that it took about 12 weeks to resolve a case, compared to 4 weeks in China and 52 weeks in Kenya (Figure 3.24). These results, however, might be misleading if firms refrain from bringing complicated cases to court in countries with weak legal systems. Consistent with this, a slightly different pattern emerged from a World Bank survey that sought to control for the complexity of cases. This study asked lawyers and legal experts in over 130 countries how long it would take in their country to get a court to recover a standardized debt contract (that is, a contract that was standardized with respect to size, type of contract, nature of the parties being sued, etc.).55 In this survey, Tanzania (127 days) outperformed Kenya (255 days) and China (180 days), as well as South Africa (207 days) and India (365 days), but not Uganda (99 days). Figure 3.24 Court cases in Tanzania are resolved relatively quickly 60 case 52 50 resolve to 40 weeks 30 of 20 12 12 number 10 4 Median 0 Tanzania China Uganda Kenya Source: Investment Climate Assessments. But other evidence suggests that there is significant room for improvement. Only 45 percent of enterprises in Tanzania agreed that they could rely on the courts to enforce contracts and uphold their property rights in a business dispute--fewer than in Kenya (49 percent), Uganda (69 percent), or China (92 percent). In addition, the prevalence of corruption brings into question the overall effectiveness of the legal system. III.2 Access to Land Few of the formal enterprises in the Investment Climate Survey were concerned about access to land. Only about 25 percent rated it as a major or very severe problem, making it the 10th-most serious problem overall (Figure 3.1). Larger enterprises were less likely to say that access to land was a serious concern: only 21 percent of large enterprises and 16 percent of very large enterprises reported that it was a major or very severe problem. Small and medium enterprises were more likely to find access to land to be a problem, but they generally did not regard it as a serious constraint (Table 3.1). In contrast, both formal and informal micro enterprises found access to land to be a serious concern. About 57 percent of formal micro enterprises and 43 percent of informal micro enterprises said that it was a major or very severe obstacle (Figure 3.25). Only access to financing and cost of financing were greater problems for these groups. Access to land was an especially serious problem for rural micro enterprises. About 52 percent of rural micro enterprises reported that access to land was a major or very severe obstacle, compared to only 42 percent of urban micro enterprises. Figure 3.25 Rural enterprises--and enterprises that have been forced to move in the past 12 months--are more likely to say access to land is a major or very severe obstacle a 100% land to 80% 76% problem 57% access 60% 48% 52% 43% 42% 45% find severe 40% that very or 20% 0% enterprises major All micro Formal Informal Rural Urban Forced Not of % micro micro micro micro to move forced to move Source: Investment Climate Surveys. III.3 Telecommunications As in most of the comparator countries other than Kenya, enterprises in Tanzania generally rated the telecommunications sector as a less serious constraint on operations and growth than other infrastructure sectors (Figure 3.15). Despite this, the number of fixed-line phones in Tanzania is relatively low, even by regional standards. At the end of 2001, there were about 4 fixed lines per 1,000 people in Tanzania--only slightly more than in 1994. In contrast, in China there were over 100 fixed lines per 1,000 people, and in India, there were 38 fixed lines per 1,000 people (Figure 3.26). Figure 3.26 Tanzania's telephone infrastructure is relatively underdeveloped, 2001 300 250 2001 110 in 200 users 150 of 100 6 138 Number 50 12 14 19 38 0 4 3 11 Tanzania Uganda Kenya India China Fixed lines Mobile Note:: Data for Tanzania were not available after 2001. Source: International Telecommunication Union (2003). Although fixed-line coverage stagnated in the 1990s, cellular telephony expanded rapidly. The first cellular license in Tanzania was issued to Mobitel in 1994. Following the licensing of a second firm in 1995, cellular service expanded quickly.56 At the end of 2000, there were five registered mobile companies in Tanzania, although one of these, Tritel, withdrew from Tanzania at the beginning of 2001. In 2001, there were 12 cellular phones per 1,000 people, up from 1 per 1,000 people in 1997. Although data were not available for 2002 or 2003, this expansion appears to have continued in recent years. By other measures, the performance of the telecommunications sector also seems reasonable, especially when compared to other utilities. For example, of the 214 firms that reported using fixed-line phones to communicate with clients, only 51 firms reported losing service at some point in 2002--while 211 firms reported power outages and 91 reported insufficient water supply. In comparison, 211 of 257 firms in Kenya reported losing telephone service for at least one day over the same period. III.4 Transportation Roughly 23 percent of manufacturing enterprises regarded transportation as a major or very severe obstacle to operations and growth, compared to 13 percent of enterprises in construction and 38 percent of enterprises in tourism. Among specific transportation-related services, sealed roads and railways drew the most criticism from manufacturing firms: 18 percent of managers with access to sealed roads and 19 percent of managers with access to railways rated these services as severe problems. In contrast, only 1 percent rated airfreight as a severe problem, and 0.3 percent gave the same rating to postal services. In addition, 13 percent of enterprises reported that they lacked access to railways, 10 percent reported that they lacked access to airfreight services, and 6 percent reported that they lacked access to sealed roads. Only 2 percent said that they did not have access to postal services, and only 1 percent said that they did not have access to trucking services. Despite these gaps in service, Tanzania compares favorably with Uganda and Kenya, especially with respect to railways. Some 43 percent of enterprises in Uganda and 33 percent of enterprises in Kenya said that railway services were not available in their region--far more than the 13 percent in Tanzania. Enterprises in Uganda were also more likely to say that trucking services (16 percent), sealed roads (8 percent), airfreight (33 percent), and postal services (5 percent) were unavailable. Enterprises in Kenya were more likely to lack access to airfreight (11 percent) and trucking services (3 percent), but less likely to lack access to postal services (1 percent) and sealed roads (3 percent). CHAPTER 4: THE INVESTMENT CLIMATE AT THE SUBNATIONAL LEVEL Whereas the previous chapter examined the investment climate at the national level, this chapter examines differences in investment climate indicators among regions in Tanzania. I. TAX RATES ANDADMINISTRATION Tax rates and tax administration are leading concerns throughout the country. These issues were more likely than any other obstacle to appear among the top three obstacles in each region covered by the survey (Annex 3). But concern over tax rates was particularly high in Iringa/Mbeya, Kilimanjaro, and Dar es Salaam. In these regions, 90 percent, 79 percent, and 76 percent respectively of respondents rated tax rates as a major or severe obstacle (Figure 4.1). Concern over tax administration, in comparison, was highest in Iringa/Mbeya (82 percent), Arusha (65 percent), and Dar es Salaam (62 percent). Figure 4.1 Enterprises in Iringa/Mbeya, Kilimanjaro, and Dar es Salaam were most likely to report that tax rates were a major or very severe problem very 100 or 80 major a 60 as 40 obstacle 20 problem severe 0 the Arusha Tanga rating Zanzibar Kagera Morogoro % Iringa/Mbeya Kilamnajaro esSalaam Dar Mwanza/Mara Tax rates Tax administration Source: Investment Climate Survey. II. INFRASTRUCTURE Regional variation was more significant when it came to the performance of the power and telecommunications sectors. About 76 percent of firms in Arusha, 72 percent of firms in Iringa/Mbeya, and 71 percent of firms in Dar es Salaam reported that power was a major or very severe problem, compared to 20 percent of firms in Kagera (Table 4.1). Consistent with this evidence, enterprises in Arusha and Dar es Salaam reported more power outages than firms in other regions. (The median number of outages was 94 in Arusha and 60 in Dar es Salaam.) Enterprises in Iringa/Mbeya reported significantly fewer outages--the median number was 30--but the median waiting time for a new connection was longer in Iringa/Mbeya (60 days) than in any other region. Overall, firm perceptions are highly correlated with objective measures of sector performance. The correlation between the percentage of enterprises reporting that power is a major or very severe problem and the median waiting time for a new connection is 0.53 across regions; the correlation between the same perceptions and the median number of power outages is 0.63. Table 4.1 The reliability of the power sector varies across regions Percentage of enterprises rating it as Median waiting time Median reported major obstacle for connection outages Power Telecom Power Telecom Power Telecom Arusha 76 34 15 14 94 18 Dar es Salaam 71 6 30 7 60 20 Iringa/Mbeya 72 45 60 -- 30 66 Kagera 20 7 2 14 12 2 Kilimanjaro 46 13 19 20 30 24 Morogoro 50 10 38 20 75 -- Mwanza/Mara 43 7 30 18 36 94 Tanga 48 17 30 30 28 14 Zanzibar 47 50 -- 38 30 265 Source: Investment Climate Survey. Enterprises generally regarded telecommunications as a less serious problem than power, except in Zanzibar. In Zanzibar, about half of all firms reported that telecommunications was a major or very severe problem. This is not surprising, given that the median wait for a fixed line was 38 days in Zanzibar--longer than anywhere else--and the median number of outages was 265--more than double the number in any other region. Across regions, the correlation between the percentage of enterprises rating telecommunications as a major or very severe problem and the median waiting time for a fixed line is 0.656; the correlation between the same perceptions and the median number of service outages is 0.659. One of the regions where telecom services appear to be most unreliable (Iringa/Mbeya) also has the highest percentage of firms that use email to communicate with clients and suppliers (82 percent). This may be a strategy adopted by businesses to counter the problem of frequent phone interruptions. On the other hand, the other region where reliability is especially poor (Zanzibar) has the lowest proportion of firms (5 percent) that use email. Overall, these results suggest that the telecommunications infrastructure is worse in Zanzibar than elsewhere in Tanzania. III. CORRUPTION Corruption appears to be a serious obstacle to enterprise operations and growth across Tanzania. But perceptions of its severity vary significantly by region (Figure 4.2). The regions where the largest numbers of firms rated corruption as a serious problem were Tanga (74 percent), Iringa/Mbeya (64 percent), and Dar es Salaam and Arusha (55 percent each). Kagera had the smallest proportion of enterprises rating corruption as a major or very severe obstacle (27 percent). Figure 4.2 Enterprises in Tanga, Iringa/Mbeya, Dar es Salaam, and Arusha were most likely to report that corruption was a major or very severe obstacle 100 severe 80 74 very 64 or 60 55 55 53 43 38 38 major 40 a 27 as obstacle 20 0 corruption Tanga Arusha Kagera rating Iringa/Mbeya esSalaam Morogoro Zanzibar Kilimanjaro % Dar Mwanza/Mara Source: Investment Climate Survey. III. BUSINESS REGULATIONS AND BUREAUCRATIC BURDEN The bureaucratic burden varies significantly across regions (Figure 4.3). The burden is greatest in Morogoro, where the average firm reports that senior managers spend 32 percent of their time dealing with government regulations and inspections. Tanga (24 percent), Iringa/Mbeya (21 percent), and Dar es Salaam (18 percent) are next on the list. Senior managers spend the least time dealing with regulations and inspections in Zanzibar (6 percent), Kagera (9.6 percent), and Kilimanjaro (9.7 percent). The regions where enterprises report the highest median number of inspections and required meetings are Iringa/Mbeya (30) and Mwanza/Mara (35). But such meetings appear to require less management time in these regions than in other areas, such as Morogoro and Tanga. Figure 4.3 Management time dealing with regulations and numbers of inspections and required meetings, by region 40% 35 35 40 (%) 32% 30 visits 30% of time 24% 21% 22 20% 17% 15% 17 16 20 11 11% 10% 10% 9 number 10% 6 6% Management 0% 0 Median Tanga Arusha Kagera Morogoro Zanzibar Iringa/MbeyaesSalaam Kilimanjaro Dar Mwanza/Mara Management time (%) Median number of visits Source: Investment Climate Survey. CHAPTER 5: POLICY IMPLICATIONS This Investment Climate Assessment emphasizes the need for reform in several areas. These include reducing the burden that taxes and tax administration impose on enterprises, improving access to credit (especially for micro and small enterprises), reducing corruption, maintaining macroeconomic stability, increasing ties to the international economy, improving the performance of the power sector, and improving enterprise productivity. Although the government has recently made progress in many areas--and such progress continues--continued reform remains necessary. I. REDUCING THE BURDENOF TAXATION ON FORMAL ENTERPRISES Issues. High tax rates were consistently rated among the greatest obstacles in Tanzania by formal enterprises of all types. More enterprises rated taxes as a major or very severe problem in Tanzania (73 percent) than in almost all of the other countries for which Investment Climate Surveys had been completed by the end of 2003. The exceptions were Ethiopia (74 percent) and Brazil (84 percent). Tax administration was not as serious a problem as tax rates. However, over 50 percent of enterprises in the manufacturing sector found tax administration to be a major or very severe problem, making it the fourth-greatest obstacle overall. As noted in Chapter One, in order to increase spending to meet the PRSP goals without losing fiscal discipline or increasing aid dependence, the government must increase domestic revenue mobilization. Accordingly, the government has increased its revenue target from about 12 percent of GDP to about 15 percent of GDP. Achieving this goal will be difficult since the private sector already considers the tax burden to be excessive, and tax administration is seen as corrupt and ineffective. These challenges emphasize the importance of reducing the size of the informal sector. Tax evasion and informality are high in Tanzania, both in absolute terms and relative to the comparator countries. By reducing evasion and informality, the government could lessen the burden on formal enterprises without losing fiscal discipline. Recommendations. In some areas--e.g., reducing corruption or reducing the time and cost of business licensing--steps to improve the investment climate involve relatively modest tradeoffs. For example, speeding up processing might increase administrative costs. In other cases, such as setting business tax rates, some investment climate goals must be balanced against others and against broader social goals. For example, although reducing the tax burden is an important goal, it must be done in a way that does not compromise macroeconomic stability.57 Despite recent improvements in this area, over 40 percent of enterprises rated macroeconomic instability as a major or very severe obstacle to operations and growth. Despite complaints about taxes, corporate tax rates are no higher in Tanzania than in the comparator countries, and VAT rates are only slightly higher (by 2 and 6 percentage points). But based on anecdotal evidence, the multiplicity of national and local taxes appears to be a major cause of dissatisfaction. A follow-up study (e.g., along the lines of a FIAS incentives review) could be helpful in generating detailed recommendations for streamlining the tax system. In addition, attention should focus on improving taxcollection, both to reduce the burden on law-abiding enterprises and to reduce economic distortions. As noted in Chapter 3, tax evasion and exemptions have eroded Tanzania's tax base. Consequently, despite comparable tax rates, tax revenues are far lower than elsewhere in the region. The government has recently taken several steps to combat this problem.58 Effective October 2003, the authorities revoked all government notices granting tax exemptions that were issued before 1997. The government has also curtailed the number of new exemptions by agreeing that no new companies will be added to the list of strategic investors maintained by the Tanzania Investment Center. In July 2003, the government abolished some local government taxes and levies that were thought to be particularly burdensome for businesses. And finally, it has been considering a revised income tax act, which was designed with technical assistance from the International Monetary Fund. The government is also taking steps to improve tax administration. In June 2003, it approved a new five-year corporate plan for the Tanzania Revenue Authority (TRA). Now it must focus on implementing the new plan and monitoring its impact on the TRA's performance. In assessing such reforms, the government should consider their impact both on revenue and on the costs of compliance for enterprises. Increasing the VAT threshold, for example, would reduce revenues but would also reduce the burden that VAT imposes on small formal enterprises. Other efforts to improve tax mobilization should focus on encouraging businesses to operate in the formal sector--for example, by reducing the regulatory burden that enterprises face. At present, senior managers spend more time dealing with government regulations in Tanzania than in any of the comparator countries. It is also more costly to register a business in Tanzania than in any of the comparator countries. Reducing the cost of registering a business--and more generally reducing the burden of regulation-- would encourage informal micro enterprises to expand their operations and enter the formal sector. The government should also focus on reducing corruption in tax administration. Das-Gupta et al. (1999) and Bird (2003) discuss specific actions that can be useful in reducing corruption in revenue agencies. One general principle is to minimize direct contact between tax officials and taxpayers--by automating and computerizing procedures, increasing the use of third-party data for assessments, and relying on tax withholding. A second useful step is to reorganize the tax agency along functional lines (such as auditing, taxpayer assistance, and processing tax returns), rather than by tax type. This makes it harder for officials to develop relationships with taxpayers. Broader strategies for addressing corruption in civil service organizations can also help, such as allowing independent internal and external audits, protecting whistleblowers, and giving citizens a channel for complaining about harassment. II. IMPROVING THE PERFORMANCE OF THE POWER SECTOR Issues. The poor performance of the power sector is a serious problem in Tanzania. Electricity was second only to tax rates in the percentage of manufacturing enterprises that rated it as a major or very severe problem. Enterprises in the tourism sector were even more likely to be concerned about power, with close to two-thirds rating it as a serious problem. And although large enterprises tended to rate it as a greater obstacle than small or micro enterprises, many micro enterprises did not have access to power at all. Firms' negative perceptions are consistent with quantitative measures of sector performance. The median enterprise in Tanzania reported that it suffered power outages on 48 days in 2002--more than in any of the comparator countries. Median production losses due to power outages were also higher in Tanzania than in the comparator countries. The government has already taken some steps to improve the power sector's performance. In May 2002, it signed a two-year management contract with Net Group Solutions, a South African management consultancy, to assume management of TANESCO. Since then, the commercial performance of TANESCO has improved: for example, the collection rate increased from 76 percent to 92 percent over this period. In addition, reliability has improved to some degree. But serious problems remain.59 In particular, TANESCO's technical performance remains problematic. Due to deferred maintenance and underinvestment in earlier years, the company requires significant resources to improve access and reliability. In the near term, TANESCO's problems may actually increase.60 A severe drought has limited hydroelectric power generation, which could lead to serious problems with load shedding in the near future. Furthermore, high oil prices have increased the cost of thermal capacity, forcing the government to provide additional budgetary support in 2003, despite TANESCO's improved commercial performance.61 Recommendations. In the medium term, the government needs to ensure that TANESCO can finance its operations without government assistance. TANESCO's management approved a 4.3 percent tariff increase effective May 1, 2004. In addition, a proposed comprehensive tariff study will update estimates of long-run power costs. This study will also examine such issues as indexing of tariffs, time-of-day tariffs, quota mechanisms for application in the event of shortfalls, and geographically differentiated tariffs. The government should decide in the medium term on a schedule for privatizing TANESCO. This decision should be based upon international market conditions and the completion of necessary reforms. In preparation for attracting private investment, the government could take various actions, including: (i) unbundling generation, transmission, and distribution; (ii) restructuring power sector debt; and (iii) establishing an independent regulatory body and revising supporting legislation to establish a successful track record with respect to regulatory decisionmaking. Under current market conditions, investor interest in TANESCO is likely to be relatively low, making privatization difficult in the near term. As a result, the government should extend the current management contract for three to five years. A new contract should be based on revised performance indicators aimed at reducing losses, increasing connections, and investing in system rehabilitation. III. ENHANCING THE EFFECTIVENESS OF FINANCIAL SERVICES AND ACCESS TO CREDIT Issues. Enterprises in Tanzania consistently find access to financing and, to a lesser extent, the cost of financing to be serious obstacles to enterprise operations and growth. The problem is least severe for very large and foreign-owned enterprises and most severe for micro enterprises, especially those in rural areas. Micro enterprises financed very little investment with loans from banks, instead relying on retained earnings and loans from family and friends. Despite the expansion of financial NGOs in recent years, very few micro enterprises--only about 6 percent--reported having loans from such institutions. In an effort to address these problems--at least for larger firms--the government launched a series of financial sector reforms in 1991. The Banking and Financial Institutions Act of 1991 liberalized interest rates, eliminated administrative credit allocation, strengthened the Bank of Tanzania's role in regulating and supervising financial institutions, restructured state-owned financial institutions (including the privatization of the two largest banks), and significantly liberalized domestic financial intermediation, allowing local and private banks to enter the market. In addition, the Cooperative Societies Act of 1991 provided the basis for the development of savings and credit cooperatives (SACCOs) as privately owned and organized equity-based institutions.62 The government also created extensive legal and regulatory structures--some of which were initially flawed. The leading example is the Land Act of 1999, which was meant to facilitate the availability of mortgage-based finance. Recent amendments to the Land Act, passed in early 2004, should facilitate mortgage-based financing in the future. Despite this progress, there has been little growth in financial services for the poor.63 To address this problem, the government has focused on strengthening the microfinance sector. Approved in February 2001, the National Microfinance Policy articulates a clear vision and strategy for the development of a sustainable microfinance industry. Recommendations. The goals of further reforms should be to strengthen the financial system in ways that improve access for micro and small enterprises. In addition to these efforts, broader efforts focusing on the institutional environment, such as increasing access to credit information and improving contract enforcement and the efficiency of the judicial system, will be helpful. To improve access for micro and small enterprises: · The government should complete the privatization of state-owned banking institutions--notably, the National Microfinance Bank (NMB), Tanzania Investment Bank (TIB), and Tanzania Postal Bank (TPB). NMB has considerable potential to become an important source of credit for small businesses, as it currently provides payment and other facilities to 750,000 depositors. · The government should clarify its regulatory intentions regarding smaller microfinance institutions (MFIs) and adopt a light handed and delegated approach to small providers of basic financial services. · The government should encourage the development and strengthening of umbrella organizations, as well as greater reliance by smaller MFIs on local banks, rather than external donors. · The government should consider creating an agency for cooperative financial institutions that focuses on building capacity in these institutions and strengthening financial infrastructure. · The government should abandon plans to open special SME windows at financial institutions or create government-promoted development banks. Although in theory these institutions are attractive, government directed credit has often ended up failing to benefit the groups that it is supposed to benefit and loans are often not repaid. For example, a recent study of 18 industrial development banks found that over half their loans were in arrears.64 Under current conditions, such initiatives could disrupt the development of self-sustaining SACCOs or NGO- based MFIs. In addition to these initiatives targeting micro and small enterprises, several broader approaches will improve access for all enterprises. · The payments and securities settlement infrastructure should be upgraded. · The credit registry/bureau project should be finalized and available for reporting by all licensed lenders. · Regulations concerning the treatment of leasing companies should be clarified and amended. · Some banking regulations should be revised to relax unduly constraining barriers (collateral requirements for large loans, aggregate large loan limits, fixed asset ceilings for banks) and to tighten loopholes (exposure to individual offshore banks). IV. REDUCING CORRUPTION Issues. In most areas related to governance--voice and accountability, political stability, government effectiveness, the rule of law, and regulatory quality--Tanzania performs better than other low-income countries. But not in corruption. Despite government efforts, progress in combating corruption in the past five years has been slow. In fact, Tanzania's score and ranking deteriorated in the most recent survey by Transparency International.65 Recommendations. The government is working to repeal the Anti-Corruption Act of 1971 and replace it with a new Anti-Corruption Law. The revision will widen the definition of corruption, increase protection for whistleblowers, and place the onus on those being investigated to prove that assets were not amassed through corrupt or unethical behavior. The government has also made some progress in improving the institutional framework related to corruption (e.g., the Public Procurement Act of 2001 and the National Anti-Corruption Strategy and Action Plan, NASCAP) and raising public awareness of the issue.66 But additional action along several dimensions would be helpful. First, Anti- Corruption Action Plans should be rolled out to local government authorities. Second, the government should increase its capacity to monitor implementation of the NASCAP. Third, publication of the new code of conduct for public service is currently underway. Once the code has been published, the government needs to publicize it and encourage support from senior management. In addition to changing the incentives of public employees, reforms should also reduce the incentives that firms have to offer bribes. Streamlining regulation to reduce the burden on enterprises would promote this goal, as would reducing the discretion that officials have in interpreting regulations. Such measures are particularly important in agencies, such as the Tanzania Revenue Authority, where corruption has been identified as a particular problem. Other areas where action should be a priority include agencies involved with labor regulation and the municipal police. V. MAINTAINING MACROECONOMIC STABILITY Issues. Since the mid-1990s, Tanzania's macroeconomic performance has been quite impressive. Real GDP has increased at an annual rate of 4.4 percent, and the annual growth rate reached an impressive 6.2 percent in 2002. Tanzania has also made impressive progress in reducing macroeconomic instability. Although consumer price inflation has averaged 10 percent per year since 1995, it has been lower in recent years. Despite these gains, enterprise mangers remain concerned about macroeconomic instability. About 43 percent of enterprise managers in Tanzania said that macroeconomic instability was a major or very severe problem, with larger enterprises tending to be more concerned than small and micro enterprises. This continued concern emphasizes that it takes a relatively long time to dispel concerns about macroeconomic instability. As a result, the government must be very careful about taking any actions that will undermine stability in the medium term. Recommendations. To maintain macroeconomic stability, the government must maintain fiscal discipline by increasing revenue mobilization in line with spending. Key measures include improving tax administration and reducing informality by reducing the burden of regulation, as well as entry costs. The government must also ensure that Tanzania's public and publicly guaranteed debt remains sustainable. Provisional results of the annual debt sustainability analysis for the period ending in June 2003 indicate that this is currently the case, following the granting of irrevocable debt relief under the enhanced HIPC Initiative. However, future sustainability depends on the implementation of macroeconomic and structural reforms, as well as prudent management of external debt. Furthermore, Tanzania's debt may not be sustainable if GDP or export growth is less than projected, or if the government fails to reach agreement with non-Paris Club creditors on debt relief. VI. IMPROVING TIES TO THE INTERNATIONAL ECONOMY Issues. Although exports have been growing in recent years--primarily due to increased exports of gold--manufacturing exports remain low. In 2002, only about 26 percent of firms exported any part of their production, and exports accounted on average for 12 percent of an exporter's sales. This was considerably lower than in Kenya, even after controlling for enterprise size and sector. The poor export performance of Tanzanian firms reflects several factors. Productivity tends to be low in Tanzania, making it difficult for firms to compete on international markets.67 In addition, relatively few firms have obtained international certification for any of their products. Finally, the poor performance of customs authorities and burdensome customs and trade regulations restrict export opportunities. Customs delays for both imports and exports were longer in Tanzania than in any of the comparator countries. As part of its efforts to improve export performance, the government passed the Export Processing Zone (EPZ) act in May 2003. By November 2003, two zones had been established in Dar es Salaam, and three companies, all of which were expected to export 100 percent of their output, had been licensed. Recommendations. As a first step, the government should thoroughly review existing practices in customs administration and trade regulation. Work on a customs reform strategy is currently proceeding with technical assistance from DFID. In addition to boosting revenue, these reforms should focus on shortening processing times and reducing the burden created by excessive regulation. Recent reforms in Ghana might be a useful model in this respect (Box 5.1). Box 5.1 Reducing customs delays through computerization in Singapore and Ghana In recent years, governments have used computerization to dramatically reduce processing times for imports and exports. Rather than submit forms to multiple agencies, a trader can now electronically submit a single document that contains all the information required by different agencies. This information is submitted to all relevant agencies, and individual agencies respond with the necessary permits or request additional information. This process reduces transaction costs and minimizes direct contact between traders and public officials, potentially reducing opportunities for side-payments. After Singapore adopted these methods in 1989, average processing times fell from between two and four days to a few minutes, and the number of required documents fell from between 3 and 35 to 1. Freight forwarders estimate that the program has reduced the cost of handling trade documentation by 20 to 35 percent. Singapore's success, and that of a similar program in Mauritius, inspired the government of Ghana to adopt a program called TradeNet. Before the program, importers estimated that the shortest clearance time at sea ports was four days, while the average clearance time was several weeks. After the program was implemented, about 14 percent of clearances at Tema port took less than a day, and only 11 percent took more than five days. At the airport, average clearance times fell from three days to four hours, with 18 percent of clearances taking less than two hours. Although computerization can reduce delays, it will not succeed unless procedures are modified to fully exploit its benefits. Before implementing TradeNet, the Ghanaian customs administration was already using a standard software package to help process imports. But procedures were not designed to take advantage of the package. For example, rather than being submitted electronically, customs declarations still had to be manually entered into the database, a process that took up to 24 hours. Consequently, the technology was underused. Source: De Wulf (2004); World Bank (1998). The government should also ensure that the newly created Export Processing Zones do not undermine other investment climate-related goals. First, the government should continue to license only firms that export all of their output and should enforce strict regulations that prevent output from leaking into the domestic economy. Second, given the problems associated with revenue collection, the government should consider the revenue implications of granting any additional exemptions. Finally, the government should not allow EPZs to distract it from the broader goals of reducing import and export restrictions and improving customs administration. Exporting by enterprises operating in domestic markets generates far more important spillovers than exporting by firms operating only in EPZs. In addition to these actions, the government could expand market access through regional integration by negotiating customs unions with neighboring countries and establishing a common external tariff. It should also continue the process of financial and economic liberalization, maintain a flexible exchange rate, and continue to reduce and simplify tariffs. VII. INCREASING LABOR ANDTOTAL FACTOR PRODUCTIVITY Issues. Labor productivity in Tanzania is higher than in Uganda, but considerably lower than in Kenya and the fast-growing Asian economies, China and India. Enterprises in Tanzania tend to be less capital-intensive (i.e., to have more capital per worker) than enterprises in Kenya. But this factor alone does not explain differences in productivity. After controlling for differences in the use of capital, labor, and intermediate inputs, we find that total factor productivity is about 5.6 percent higher in Kenya than in Tanzania. TFP is also higher on average in Uganda than in Tanzania, even though value added and sales per employee are almost twice as high in Tanzania as in Uganda. The discrepancy is explained by differences in capital intensity in the two countries. Firms in Tanzania produce more value added per worker than firms in Uganda because they have more capital per worker, not because they are more efficient. Recommendations. Enterprises in Tanzania were less likely to have formal training programs than similar enterprises in Kenya or China, but those that did were more productive. Encouraging enterprises to invest in worker training might therefore be a useful way of improving firm-level productivity. Rather than become directly involved in worker training, the government should facilitate training and apprenticeship programs that are designed and implemented by the private sector. For example, it could offer tax credits to firms that provide worker training. Enterprises with university-educated managers were also more productive than other firms.68 Consequently, the government should consider exploring ways to improve access to business education. In addition, improving access to secondary and vocational training could help boost productivity. While most workers in Tanzania had at least a primary education, and the share of workers with some tertiary education was similar to the corresponding figures in Kenya and Uganda, workers in Tanzania were far less likely to have secondary or vocational levels of education than workers in the other two countries. POLICY MATRIX FOR IMPROVING THE INVESTMENT CLIMATE IN TANZANIA Major Policy Specific Policy Issues Observations/Comments Policy Suggestions/Ongoing Reforms Area · Individual taxes in Tanzania are not significantly · Steps to broaden the tax base would allow the higher than rates in other government to reduce the burden on formal developing countries (except enterprises without compromising macroeconomic Close to three-quarters of VAT which, at 20 percent is stability. Steps to reduce informality--including enterprises in Tanzania slightly higher than in the reducing entry costs and the burden of regulation-- comparator countries). are vital parts of this agenda Reducing the perceive tax rates to be a · Tax evasion is a serious · Although recent reforms in tax have been useful, burden of serious problem. Many enterprises also rate tax problem in Tanzania. For tax administration remains a serious concern. Further taxation administration as a major example, Tanzania's VAT reforms should reduce the burden that administration obstacle. efficiency ratio (0.20) is imposes upon enterprises, rather than simply increase lower than the average for tax revenues. sub-Saharan Africa (0.27). · Although the government has abolished some Tax exemptions and local taxes and levies that were particularly incentives in some sectors of burdensome, the large number of taxes remains a the economy have reduced tax concern. revenues. · More than half of firms in · The commercial performance of the power sector The poor quality of Tanzania rated power supply has improved since 2002, when the government Tanzania's infrastructure as a serious problem, with signed a two-year management contract with Net firms in Dar Es Salaam and Group Solutions. This contract should be extended Improving the is a serious impediment to enterprise performance. Arusha being most concerned. for three to five years. The new contract should be performance of · Many micro enterprises based on revised performance indicators that focus on the power Within this broad area, are not connected to the reducing losses, increasing connections, and investing sector enterprises were most concerned about the power grid. in system rehabilitation. performance of the power · The reliability of power · In the medium term, the government should set a sector. supply is a particular concern. timeframe for introducing greater private-sector Enterprises in Tanzania participation in the sector. In the near term, the reported greater losses due to government should start preparing TANESCO for Major Policy Specific Policy Issues Observations/Comments Policy Suggestions/Ongoing Reforms Area power outages than this step by (i) unbundling generation, transmission, enterprises in the comparator and distribution; (ii) restructuring power sector debt; countries. and (iii) establishing an independent regulator and revising supporting legislation. · The government should also ensure that TANESCO can finance its own operations without government assistance. · Efforts to improve the performance of the · Enterprises of all sizes banking system should continue. Important steps find the cost of financing to include: (i) completing the privatization of state- be a serious impediment. owned banks; (ii) upgrading the payments and Despite Tanzania's improved securities settlement infrastructure; and (iii) revising macroeconomic stability, real banking regulations to relax remaining barriers. Despite progress in interest rates remain high. · In particular, the government should complete the Access to credit is an Enhancing banking sector reform, · privatization of the National Microfinance Bank access to credit, enterprises in Tanzania especially great problem for (NMD), which has the potential to grow as a source micro and small enterprises. especially for continue to rank the cost of credit for SMEs. small and of financing and access to These enterprises rely · The credit registry/bureau project should be primarily upon retained finalized and available for reporting by all licensed micro finance as serious earnings and upon family and enterprises constraints on enterprise lenders. operations and growth. friends to fund investment and · Regulations regarding smaller microfinance working capital. institutions and leasing companies need to be · Despite the expansion of clarified and amended or adopted. financial NGOs in recent · Plans to develop special SME windows at years, very few micro financial institutions and government-promoted enterprises reported having development banks should be abandoned, as they loans from these institutions. could disrupt the development of self-sustaining SACCOs and NGO-based microfinance institutions. Improving Tanzania outperforms the · The Investment Climate · In recent years, the government has made accountability average low-income Survey confirms that progress in fighting corruption. The Finance Act and reducing country on most corruption is a significant (2001), the Public Procurement Act (2002), the 1996 Major Policy Specific Policy Issues Observations/Comments Policy Suggestions/Ongoing Reforms Area corruption governance indicators. problem in Tanzania. More reform of the Tanzanian Revenue Authority (TRA), But corruption remains a that 50 percent of enterprises and the government's clearly stated zero-tolerance problem. said that it was a major or position on corruption have been important steps. very severe obstacle to · But corruption remains a serious problem, operations and growth. especially in relation to tax administration and · Both grand corruption and government procurement. The government should petty corruption are seen as focus on these areas. serious problems by firm · More broadly, the government should continue to managers. involve the mass media in the fight against corruption, strengthen the "Bureau of Prevention of Corruption," and reduce the burden that regulation imposes on formal enterprises. · In recent years, · The government should maintain its efforts to Tanzania's macroeconomic control inflation and promote strong economic performance has been quite growth. impressive. Per capita GDP · The government should continue to support grew at an average rate of 3.2 reforms to strengthen private investment, increase percent per year between efficiency in the goods and financial markets, and Despite improvements in 2000 and 2002. Inflation maintain a reasonable level of foreign reserves. macroeconomic averaged a modest 5.2 percent · The government should maintain market-oriented Maintaining performance in recent between 2000 and 2002 and policies of monetary restraint. It should also promote macroeconomic years, enterprise was even lower in 2003. This and sustain positive real interest rates in order to keep stability managers remain is a testament to the inflation low and improve allocation of financial concerned about government's careful fiscal resources. macroeconomic stability. stance and its financial sector · The continued sustainability of Tanzania's reforms. external debt is contingent upon the sustained · Tanzania's public and implementation of macroeconomic and structural publicly guaranteed external reform policies, as well as prudent management of debt is currently considered to external debt. If GDP or export growth is less than be sustainable, following the projected, or agreement on debt relief from non-Paris granting of irrevocable debt Club creditors is not reached, Tanzania's debt might Major Policy Specific Policy Issues Observations/Comments Policy Suggestions/Ongoing Reforms Area relief under the HIPC not be sustainable. Initiative. · Proposals to reduce tax rates must take into account their impact on government revenues. · Tanzanian manufacturing enterprises are less likely to · The government should improve the functioning export than similar enterprises of the customs administration to shorten processing Improving the Despite recent reforms, in Kenya. times. performance of · Trade regulations and · The government should enlarge markets and the export Tanzania's exports have grown only modestly and customs administration enhance regional cooperation by negotiating customs sector and remain a serious problem for unions with neighboring countries and establishing a reducing unevenly, with gold large enterprises and common external tariff distortions in exports increasing sharply in recent years. exporters. · The government should continue financial and foreign trade · Clearing customs takes economic reforms and liberalization, maintain a almost twice as long in flexible exchange rate, and continue to reduce and Tanzania as in any of the simplify tariffs. comparator countries. Labor productivity is Workers have less education lower in Tanzania than in Kenya. Although labor and training in Tanzania than in the comparator countries · The government should encourage enterprises to Enhancing productivity is higher in for which data were available. invest in formal worker training by facilitating productivity Tanzania than in Uganda, With respect to education, the training and apprenticeship programs that are and this is because enterprises main difference between designed and implemented by the private sector. competitiveness in Tanzania are more capital-intensive. Total Tanzania and the comparator · The government should expand access to factor productivity is countries was that secondary secondary and vocational education. and vocational education were lower in Tanzania than in both Uganda and Kenya. less common in Tanzania. ANNEX 1: THE TANZANIAN INVESTMENT CLIMATE SURVEY DATA The 2003 Tanzanian survey covered about 480 establishments, including 276 firms from the manufacturing sector. These manufacturing firms operated in 8 industrial sectors and in 10 regions of mainland Tanzania, as well as Zanzibar. The sectors and regions covered in the survey were selected because they had relatively high concentrations of manufacturing firms. Regional distribution of sample firms. The regions covered in the survey are: four regions on the eastern coast (Arusha, Kilimanjaro, Tanga, and Dar es Salaam), three in the center (Morogoro, Iringa, and Mbeya), one island (Zanzibar), and three in the north (Kagera, Mwanza, and Mara). In the final sample, three regions had only a few manufacturing firms: Iringa (4), Mbeya (7), and Mara (5). Therefore, firms in Iringa were combined with firms in neighboring Mbeya, and firms in Mara were combined with those in Mwanza (Figure A1.1). Figure A1.1 Regional distribution of firms Zanzibar (21) 8% Iringa/Mbeya (11) 4% Morogoro (10) 4% Kagera (15) Dar es Salaam 5% (112) 40% Tanga (23) 8% Kilimanjaro (24) 9% Mwanza/Mara Arusha (29) (31) 11% 11% Source: Investment Climate Assessment. Sectoral distribution of sample firms. The survey covers eight subsectors in manufacturing that were chosen for their importance in Tanzania. The subsectors are: furniture and wood; textiles, garments, and leather; construction materials; paper, printing, and publishing; chemicals and paints; metals; plastics; and agribusiness. Since there were only seven firms in the plastics subsector, they were combined with firms in the chemical and paints subsector (Figure A1.2). Figure A1.2 Sectoral distribution of firms Paper, printing, & publishing (25) 9% Chemicals & Construction plastics (34) materials (11) 12% 4% Textiles, garments, Metals (29) & leather (31) 11% 11% Wood and furniture (65) Agribusiness (81) 24% 29% Source: Investment Climate Assessment. Size distribution of sampled firms. Most of the manufacturing firms in the survey were small, with the median firm having 31 employees. Larger firms were most common in Iringa/Mbeya and Tanga, where the median establishments had 234 and 68 workers respectively (Figure A1.3). Figure A1.3 Size distribution of firms 500 500 411 400 400 300 300 200 150 200 120 126 126 99 92 100 234 74 100 33 15 68 54 41 27 21 21 15 9 31 0 0 Tanga Arusha Total Kagera Morogoro Zanzibar Iringa/Mbeya esSalaam Kilimanjaro Dar Mwanza/Mara Median Average Source: Investment Climate Assessment. ANNEX 2: TECHNICAL APPENDIX FOR PRODUCTIVITY CALCULATIONS This annex presents results from an analysis of the effect of investment climate variables on four different measures of firm performance: sales growth, employment growth, labor productivity, and total factor productivity (TFP).69 The investment climate variables include measures of firm internationalization, infrastructure, technology, labor quality, governance, corruption, and finance. The general reduced-form equation estimated in this section is: Yi = Xi + ICi +i (A.1) Where Yi is the measure of firm performance (i.e., sales growth, employment growth, labor productivity, or total factor productivity), X is a set of control variables for size of the firm (proxied by employment) and sector of operations, and IC is a measure of the investment climate for the firm i. The coefficient of interest in this analysis is, therefore, . Unfortunately, many firms failed to answer some of the survey questions. Consequently, if multiple investment climate variables were included simultaneously, many observations would be lost. To avoid this problem, we estimate the effect of each measure individually. This also makes it easier to interpret the coefficient for each investment climate measure. The dependent variables are: the average annual growth rate of manufacturing sales between 2000 and 2002; the average annual growth rate of employment between 2000 and 2002; and two measures of productivity. Labor productivity is the value of manufacturing sales per employee in 2002. Total factor productivity (TFP) is calculated using the methodology described in the next section. Differences in TFP across firms at a given moment in time are differences in output that cannot be explained by differences in the use of labor, capital, and intermediate goods. Firms with higher TFP are more efficient--that is, they produce higher value with fewer inputs. Estimating total factor productivity. Firms that operate in different sectors of activity are likely to use technologies that differ in capital and labor intensity. Therefore, to estimate TFP, we specify a different production function for each sector, using sector- specific parameters and input shares. We assume that every firmoperating in a given sector has the following production function: ln Si = + ln Ii + ln Ei + ln Ki +Hi + µi (A.2) where S is the manufacturing sales of firm i at time t, I are the firm's intermediate inputs, E is the total number of employees, K is the stock of capital, and H is the firm's human capital.70 Heterogeneity across firms is captured by the unobserved firm productivity shock, µi . The time subscript is omitted because all the variables refer to the same period (2002). The estimated TFP for Tanzanian firms is the residual of this regression, µ^ , (i.e. the part of total sales that is not explained by sector affiliation, intermediate inputs, and the stock of physical and human capital). Intermediate goods include raw materials, electricity, fuel, and other sources of energy.71 The stock of capital is proxied by the book value of machinery and equipment multiplied by the firm's capacity utilization rate. Firms with higher rates of capacity utilization produce more, but this is entirely explained by their use of the technology. To measure a firm's stock of human capital, we use a dummy variable indicating whether its manager holds a university degree. Unfortunately only a limited number of firms provided enough information to compute TFP. Of the 276 firms in the sample, only 57 percent provided information on sales and intermediate inputs. Because equation (A.2) is estimated at the sector level, we are left with very few observations per sector. To increase the sample size when computing TFP, we include firms in the neighboring countries of Uganda and Kenya. (These firms are not included in the second-stage regressions of TFP on investment climate measures.) The surveys for all three countries were conducted at approximately the same time, using essentially the same variables and methodology. This procedure assumes that firms in Kenya and Uganda use the same technologies as Tanzanian firms operating in the same sector. To control partially for country-specific effects, we include dummy variables indicating country of origin in the regression model. These variables capture differences in the level of productivity among the three countries.72 Unfortunately, many firms in Kenya and Uganda also failed to answer some of the questions needed to calculate TFP. Only 62 percent of Kenyan firms and 69 percent of Ugandan provided data on all the variables. Since this non-response pattern may create selection bias (if the firms that answer all questions are not a random sample of the original sample of firms), the results from the second-stage regressions with TFP as a dependent variable should be treated with caution. With the addition of firms from Kenya and Uganda, the sample we use for estimating TFP contains 441 firms, with Tanzanian firms accounting for approximately 30 percent of the sample. The sector composition of the sample is as follows: agribusiness (33 percent); chemicals and paints (6.7 percent); construction materials (9.9 percent); metals (12 percent); furniture and wood (16 percent); paper, printing and publishing (7.1 percent); chemicals and plastics (3 percent); and textiles, garments, and leather (12 percent). For each sector, we estimate Equation (A.2.) by ordinary least squares. This method assumes that firm-level productivity shocks are contemporaneously independent of the explanatory variables in equation (A.2.). The estimates for the country fixed effects allow us to compare sector differences in TFP levels across counties. The results are reported in Table 2.3 in Chapter 2. Once we control for differences in the levels of intermediate inputs and other factors of production, average total factor productivity is lower in Tanzania than in Kenya and Uganda. However, the results of this analysis vary significantly across sectors. For example, while TFP in chemicals and paints is 11 percent lower in Kenya than in Tanzania, TFP in paper, printing, and publishing is 19 percent higher in Kenya than in Tanzania. Estimating the impact of the investment climate on firm performance. After calculating TFP and the other measures of firm performance, we regress these variables on the investment climate variables. The results of estimating equation (A.1) for sales growth, employment growth, labor productivity, and TFP in Tanzania are presented in Table A2.1. For each investment climate measure, we report the number of observations in the regression, the estimate of the coefficient , and its standard error (in parentheses). Because the sample size is small, we use a single star (*) to designate significance at a 20 percent level, two stars (**) to designate significance at a 10 percent level, and three stars (***) to designate significance at a 5 percent level. To test the robustness of the results, we performed some sensitivity checks. First, we ran the same regressions omitting size and sector controls. The results presented are robust in the sense that the sign of the significant coefficients did not change, even though the degree of significance of some variables changed. Second, we included regional controls in the regressions. Once again, the results are not sensitive to this change. Third, we tested the robustness of the labor productivity results by changing the definition of labor productivity to value added per employee. Table A2.1 Estimates of investment climate regressions for Tanzania Sales Growth Employment Growth Labor Productivity Total Factor Productivity Obs.Coefficient. Obs. Coefficient Obs.Coefficient Obs. Coefficient Age 179 -0.007*** 254 -0.003*** 154 -0.241 124 -0.006 (-0.004) (-0.001) (-0.225) (-0.011) Firm Characteristics Private individual ownership 159 0.001 226 0.064** 138 2.150 111 0.181 (0.084) (0.040) (5.875) (-0.265) Manager degree 179 0.121 252 0.044 155 10.963*** 125 0.607*** (0.100) (0.040) (5.884) (0.287) Foreign ownership 178 0.048 256 -0.010 153 1.835 125 -0.347 Internationalization (0.115) (-0.043) (6.811) (0.358) Exporter 177 0.174** 253 0.033 153 12.873** 125 0.372 (0.102) (0.043) (6.335) (0.316) Power outages 142 0.000 201 0.000 116 0.029* 94 -0.001 (0.001) (0.000) (0.049) (-0.002) Computer 178 0.007*** 252 0.002* 154 0.182 125 0.016* (0.003) (0.001) (0.170) (0.011) Infrastructure and 0.127* 250 0.061* 152 10.388** 122 0.508** Technology Email 175 (0.101) (0.038) (6.193) (0.296) Web 172 0.266*** 247 0.085*** 150 13.571** 121 0.594** (0.110) (0.041) (7.457) (0.378) Research and development 177 -0.067 253 0.100*** 154 0.674 125 0.138 (-0.108) (0.042) (6.546) (-0.351) Labor Share of non-permanent workers 126 0.247* 183 0.010 103 14.906 81 -0.322 (0.191) (0.075) (14.719) (-0.661) Training share 132 -0.099 190 0.039 114 8.284 110 0.333* (-0.114) (0.050) (8.196) (0.270) Bribe 179 0.319*** 255 0.008 154 7.819** 124 0.587*** (0.087) (0.036) (5.320) (0.266) Governance and 113 0.018*** 170 -0.001 96 -0.113 78 -0.009 Corruption Bribe/contract value (0.006) (-0.003) (-0.220) (-0.018) Tax inspection meetings with officials 180 0.001 258 -0.001 155 0.145** 125 0.000 (0.001) (-0.001) (0.077) (0.004) Overdraft facility 180 0.136* 258 0.061* 155 -2.193 125 0.184 (0.099) (0.039) (-6.072) (0.303) Finance Credit suppliers 180 0.001 258 0.000 155 0.214*** 125 0.011*** (0.001) (0.000) (0.073) (0.003) Significant at the 20 percent level; ** significant at the 10 percent level; *** significant at the 5 percent level. Source: Authors' computations based on Investment Climate Surveys. ANNEX 3: RANKING OF OBSTACLES BY FIRM TYPE I. RANKINGSBY REGION (Percentage of firms that ranked each obstacle among the top three) Dar es Salaam Arusha 80% 100% 86% 65% 80% 60% 60% 40% 24% 34% 22% 40% 21% 20% 17% 24% 20% 13% 20% 14% 14% 3% 2% 7% 3% 3% 0% 0% 0% shortage reguln toland shortage reguln toland admin/rates Corruption imports Corruption Skill admin/rates imports Tax infastructurefrom Lack ofdemandoffinance Access Tax infastructureSkill from Lack ofdemandoffinanceAccess Business/Labor Business/Labor InadequateCompetition Access/Cost Inadequate Access/Cost Competition Mwanza/Mara Kilimanjaro 80% 71% 60% 54% 60% 42% 40% 40% 23% 20% 20% 16% 13% 13% 10% 10% 8% 8% 8% 8% 6% 6% 3% 0% 0% 0% 0% shortage demand toland Corruption reguln shortage toland admin/rates imports offinance Skill of admin/rates imports reguln ofdemand Corruption finance of Skill Tax from infastructure Lack Access Tax infastructurefrom Lack Access Business/Labor Competition Inadequate Access/Cost Business/Labor InadequateCompetition Access/Cost Tanga Kagera 100% 80% 78% 80% 60% 60% 60% 40% 33% 40% 27% 27% 26% 17% 20% 13% 20% 13% 9% 7% 7% 7% 7% 4% 4% 0% 0% 0% 0% Corruption reguln imports demandoffinance shortage toland shortage toland of reguln Corruption Skill admin/rates importsofdemandoffinance Skill Taxadmin/ratesinfastructure from Lack Access Tax infastructurefrom Lack Access Inadequate Business/Labor Access/Cost Access/Cost Business/Labor Competition InadequateCompetition Morogoro Iringa/Mbeya 80% 70% 100% 91% 60% 80% 40% 60% 40% 20% 20% 20% 40% 27% 27% 20% 10% 18% 18% 20% 0% 0% 0% 0% 0% 0% 0% 0% 0% reguln toland shortage Corruption shortage toland admin/rates imports reguln imports Skill Corruptionoffinance admin/rates Skill ofdemandoffinance Tax infastructure from LackofdemandAccess Tax infastructurefrom Access Lack Inadequate Business/Labor Access/Cost Business/Labor Competition InadequateCompetition Access/Cost Zanzibar 80% 67% 60% 40% 33% 29% 19% 20% 14% 10% 5% 0% 0% 0% shortage toland admin/rates ofdemand Corruptionimports reguln Skill offinance Tax Lack infastructure from Access Business/Labor Inadequate Access/Cost Competition II. RANKINGSBY SECTOR (Percentage of firms that ranked each obstacle among the top three) Chemical/Paints Metal 80% 71% 80% 62% 60% 60% 40% 32% 40% 21% 18% 18% 21% 21% 21% 20% 12% 12% 20% 14% 10% 10% 3% 3% 3% 3% 0% 0% Corruption shortage toland imports reguln toland admin/rates shortage imports Corruptionreguln Skill ofdemand offinance admin/rates Skill ofdemand offinance Tax infastructure from Lack Access Tax from infastructure Lack Access Inadequate Business/Labor Access/Cost Competition CompetitionInadequate Business/Labor Access/Cost Agribusiness Woodwork/Furniture 100% 60% 57% 79% 80% 40% 60% 26% 25% 40% 32% 20% 15% 19% 12% 12% 20% 16% 12% 7% 8% 5% 6% 2% 1% 3% 0% 0% toland reguln shortage toland admin/rates imports reguln admin/ratesofdemand Corruptionimports Skill Tax infastructure from CorruptionshortageofdemandoffinanceAccess Skill Lack Tax Lack infastructure from offinanceAccess InadequateCompetition Business/Labor Access/Cost Inadequate Business/Labor Access/Cost Competition Apparel/Leather Construction Materials 80% 74% 80% 64% 60% 60% 40% 40% 29% 27% 23% 20% 16% 16% 18% 18% 10% 10% 20% 9% 9% 3% 0% 0% 0% 0% 0% 0% toland admin/ratesimports Corruption reguln toland shortage Corruption reguln admin/rates imports ofdemand Tax from infastructureSkillshortageofdemandoffinanceAccess Lack Tax infastructureSkill offinance Access from Lack Business/Labor Access/Cost Business/Labor Competition Inadequate Inadequate Access/Cost Competition Paper/Printing All Sample 80% 72% 80% 69% 60% 60% 40% 28% 40% 24% 24% 24% 27% 18% 17% 20% 12% 12% 13% 12% 12% 4% 4% 20% 5% 2% 0% 0% reguln toland Corruption reguln toland shortage admin/rates shortageimports Corruptionofdemandoffinance admin/rates imports Skill Tax infastructureSkill from Lack Access Tax infastructure from Lack ofdemandoffinanceAccess Inadequate Access/CostBusiness/Labor Business/Labor Access/Cost Competition Inadequate Competition III. RANKINGSBY FIRM SIZE (Percentage of firms that ranked each obstacle among the top three) Small (1-19 workers) Medium (20-99 workers) 80% 80% 75% 59% 60% 60% 40% 32% 40% 18% 18% 21% 21% 21% 21% 20% 15% 13% 12% 14% 6% 20% 10% 3% 4% 3% 0% 0% shortage reguln toland toland admin/rates imports Corruptionshortageimports reguln demandfinance Skill admin/rates Skill of of Tax infastructure from CorruptionofdemandoffinanceAccess Lack Tax from infastructure Lack Access Inadequate Business/Labor Access/Cost Competition CompetitionInadequate Business/Labor Access/Cost Large (100+) 80% 75% 60% 40% 32% 19% 20% 15% 7% 6% 5% 4% 1% 0% toland admin/rates imports Corruption shortage reguln Skill Tax infastructurefrom Lack ofdemandoffinanceAccess InadequateCompetition Business/Labor Access/Cost IV. RANKINGSBY EXPORT STATUS (Percentage of firms that ranked each obstacle among the top three) Exporters Non-Exporters 77% 80% 80% 66% 60% 60% 40% 31% 40% 25% 20% 17% 19% 15% 12% 20% 16% 16% 13% 13% 9% 5% 3% 1% 4% 2% 0% 0% Corruptionimports reguln toland shortage imports admin/rates from Skill offinance ofdemand toland admin/rates Corruptionofdemand shortage reguln Skill Tax infastructure Lack Access Tax infastructurefrom Lack offinanceAccess Inadequate Business/Labor Access/Cost Competition InadequateCompetition Business/Labor Access/Cost ANNEX 4: REGIONAL DISTRIBUTION OF FOREIGN DIRECT INVESTMENT Tanzania has succeeded in attracting a growing amount of foreign direct investment (FDI), both in absolute terms and as a share of total FDI inflows to Africa. But this investment is concentrated in a few regions. (Domestic investment is believed to be highly concentrated as well.) One explanation for this pattern may lie in differences in the investment climate across regions. A closer look at the regional distribution of FDI in mainland Tanzania shows that two regions (Dar es Salaam and Mwanza) accounted for close to 60 percent of the stock of FDI (in 1998 and 1999) and close to 57 percent of the flow of FDI (in 1999) (Table A4.1). Dar es Salaam had the greatest share of the stock of FDI (36 percent in 1999), followed by Mwanza (24 percent). But the flow of FDI was much higher to Mwanza (34 percent of the total flow to mainland Tanzania in 1999) than to Dar es Salaam (23 percent). Consequently, Dar es Salaam's share of the FDI stock fell from 40 percent in 1998 to 36 percent in 1999, while Mwanza's share rose from 20 percent in 1998 to 24 percent in 1999. Table A4.1 Stocks and flows of FDI by region, 1998 and 1999 Stocks of FDI Flows of FDI Region (% of total) (% of total) 1998 1999 1999 Dar es Salaam 39.7 35.6 34.0 Mwanza 20.0 23.6 16.5 Shinyanga 6.8 9.1 22.8 Arusha 9.7 8.0 15.6 Mara 4.1 6.8 4.1 Morogoro 7.3 5.9 2.7 Iringa 5.4 4.2 1.6 Tanga 3.1 2.6 1.3 Tabora 1.3 2.0 1.0 Kilimanjaro 1.5 1.5 0.4 Mbeya 0.7 0.4 0.3 Dodoma 0.3 0.3 0.3 Pwani 0.1 0.2 -0.5 Total 100.0 100.0 100.0 Source: Tanzania Investment Report (2001). A similar pattern is found in many of Tanzania's neighbors. In many African countries, the economic infrastructure tends to be particularly well developed in a few regions, particularly those including capital cities. As a result, investment is concentrated in those areas. For example, between 1992 and 2000, close to 40 percent of FDI in Ethiopia went to Addis Ababa.73 This explanation also helps explain the distribution of FDI in mainland Tanzania. Dar es Salaam is the capital city and the main commercial center of mainland Tanzania, and its economic infrastructure is more developed than in other regions. Mwanza is also a commercial center, with abundant mineral resources, including gold. Conditions in Mwanza, which includes the lake zone (Lake Victoria), are also conducive to agriculture and fishing. Other regions that have received relatively substantial amounts of FDI include Shiyanga, Arusha, and Mara. Shingaya and Mara are located in the lake region, where there are substantial mining activities. Arusha is a main commercial center in the northern zone with great potential in mining, tourism, and manufacturing. It also has relatively good rail and road networks and hosts the headquarters of the East African Community (EAC). As one might expect, the regional distribution of FDI is closely related to the relative economic strength of each region (Figure A4.1). The regions that accounted for the greatest shares of Tanzania's GDP between 1980 and 1994 (Dar es Salaam, Arusha, and Mwanza) are also three of the top four regions in terms of FDI inflows and stocks. These three regions together contribute more than 35 percent of the mainland's GDP. Figure A4.1 Average annual regional contributions to GDP, 1980-1994 Paper/Printing/ Publishing (25) 9% Chemicals & Construction Plastics (34) Materials (11) 12% 4% Textile/Garments & Metals (29) Leather (31) 11% 11% Woodwork, Furniture (65) Agro-Industry (81) 24% 29% Source: Planning Commission, based upon national accounts of Tanzania. ANNEX 5: STANDARD TABLES Table A5.1. Sample Firm Size (%) Firm Activity (%) Sample Sample Small (<100 employees) 74.28 Agribusiness 29.35 Large 25.72 Chemicals and paints 9.78 Construction materials 3.99 Market Orientation (%) Metals 10.51 Sample Furniture and wood 23.55 Exporter (>=10 % sales) 18.84 Paper, printing, publishing 9.06 Non-exporter 81.16 Plastics 2.54 Textiles, garments, leather 11.23 Firm Location (%) Firm Ownership (%) Sample Sample Dar es Salaam 40.58 Publicly listed 2.55 Arusha 10.51 Private held, limited 62.18 Mwanza/Mara 11.23 Publicly held, limited 2.18 Kilimanjaro 8.70 Cooperative 5.45 Tanga 8.33 Sole proprietorship 17.45 Kagera 5.43 Partnership 7.27 Morogoro 3.62 Government-owned 1.09 Iringa/Mbeya 3.99 Other 1.82 Zanzibar 7.61 Table A5.2. Competitors, suppliers, and customers er Capacity Capacity Tanzania Uganda Small Large Exporter- Export Low High Non Average NumberDomestic private firms 245.5330.65277.14144.81256.93 212.94261.32166.24 of Competitors State-owned firms 62.07 0.14 66.77 1.00 75.93 46.08 64.15 6.00 Foreign-owned firms 170.5 9.87122.51243.26199.04 106.29137.32254.95 Average NumberDomestic private 185.3615.89 183.8190.25189.47 173.46169.54250.49 of Suppliers State-owned firms 2.75 0.05 3.64 0.04 3.74 0.25 3.29 0.03 Foreign-owned firms 26.619 1.1311.16870.26325.377 29.96622.51147.389 Average NumberDomestic private firms 388.0887.37399.01354.62404.71 340.13398.99 343.6 of Customers State-owned firms 66.51 1.14 55.29 99.39 53.31 100.92 72.95 32.51 Foreign-owned firms 44.162 4.3725.096100.6843.244 46.53231.503100.05 Table A5.3. Respondents' evaluation of general constraints on operations (% of firms evaluating constraints as "major" or "very severe") - Low High Non Tanzania Uganda China Small Large Capacity Capacity Exporter Exporter Telecommunications 11.8 5.2 23.5 10.0 16.7 9.9 17.4 11.4 13.5 Electricity 58.9 44.5 29.7 58.3 60.6 56.266.7 57.8 63.5 Transportation 22.9 22.9 19.1 22.1 25.0 23.321.7 19.6 36.5 Access to land 24.6 17.4 14.7 19.5 38.9 25.123.2 25.9 19.2 Tax rates 73.4 48.3 36.8 73.4 73.6 70.980.9 73.6 72.5 Tax administration 55.7 36.1 26.7 53.3 62.5 55.257.4 54.5 60.8 Customs and trade regulations 31.5 27.4 19.3 26.3 45.8 27.742.6 28.3 45.1 Labor regulations 2.5 10.8 20.7 2.5 2.7 1.5 5.6 2.2 3.8 Business licensing and operating permits 27.4 10.1 21.3 26.8 29.2 25.433.3 28.0 25.0 Access to finance (e.g., collateral) 48.3 45.0 22.8 49.5 45.1 53.034.3 52.1 32.0 Cost of finance (e.g., interest rates) 57.8 60.3 21.8 56.3 62.0 58.456.1 60.3 46.9 Economic and regulatory policy uncertainty 31.5 27.6 32.9 31.8 30.6 32.229.4 33.8 21.6 Macroeconomic instability (inflation, exchange rate) 43.0 45.4 30.2 43.9 40.3 42.644.1 42.5 45.1 Corruption 51.1 38.2 27.3 50.0 54.2 52.247.8 50.5 53.8 Crime, theft, and disorder 25.5 26.9 20.0 26.6 22.2 25.724.6 25.1 26.9 Anti-competitive or informal practices 24.3 31.1 23.7 25.0 22.2 21.731.9 23.6 26.9 Table A5.4. Infrastructure indicators - Non Low High Tanzania Uganda China India Small Large Exporter Exporter Capacity apacityC Freq of power outages (times last year) 67.238.6 NA NA 69.063.5 63.1 67.3 64.5 75.3 % of production lost due to power outages10.8 6.3 2.0 NA 12.2 6.5 9.3 11.1 11.2 9.1 Have own generator (%) 55.435.316.2 68.945.185.9 86.0 48.6 52.0 65.7 Have own well (%) 34.713.015.6 50.825.460.9 52.9 30.6 34.0 35.4 % of production lost in shipment 1.8 NA 1.2 NA 1.9 1.8 2.2 1.8 2.2 0.9 Days to obtain telephone connection 23.133.212.0 NA 25.118.4 25.3 22.3 21.3 30.7 Days to obtain electricity connection 54.638.319.0 NA 54.457.8 61.3 52.0 60.6 40.1 Table A5.5. Sources of finance - Non Low High Tanzania Uganda China India Small Large Exporter Exporter Capacity Capacity Share of working capital financed from: Retained earnings 74.0 80.0 51.5 30.4 78.4 59.2 56.1 78.2 75.0 71.7 Banks, other financial institutions 13.8 5.7 20.6 36.1 10.1 26.0 27.7 10.5 14.1 12.8 Trade credit 5.6 5.3 4.1 NA 5.3 6.9 6.9 5.2 5.4 6.6 Equity 3.0 1.8 0.6 13.0 2.1 6.3 4.0 2.9 1.5 6.7 Informalsources 0.4 0.4 8.6 NA 0.2 0.9 0.0 0.5 0.5 0.1 Other 3.2 2.7 6.3 20.5 3.9 0.7 5.4 2.7 3.6 2.1 Share of new investment financed from: Retained earnings 68.0 71.1 NA NA 70.5 61.7 54.4 72.2 71.5 61.6 Banks, other financial institutions 17.3 11.6 NA NA 15.8 21.1 25.9 15.0 14.7 22.5 Trade credit 2.0 0.5 NA NA 0.9 4.6 0.2 2.5 2.3 1.4 Equity 4.8 2.0 NA NA 4.5 5.5 10.9 3.3 4.5 3.8 Informal sources 1.9 1.5 NA NA 1.6 2.7 1.5 2.1 1.5 3.2 Other 6.0 4.5 NA NA 6.8 4.4 7.1 4.8 5.5 7.4 Table A5.6. Credits, loans, and liabilities - Non Low High Tanzania Uganda China India Small Large Exporter Exporter Capacity Capacity Share with overdraft or credit line 31.7 22.7 21.8 NA 23.8 54.5 49.0 27.8 28.3 41.4 Percentage of credit that is currently unused 24.4 NA 29.7 NA 26.7 21.5 24.7 23.4 27.8 16.9 Share with a loan from a bank or financial institution 19.9 20.0 45.8 NA 12.2 43.9 36.0 15.4 16.1 29.4 For most recent loan or overdraft: Share that require collateral 91.2 92.0 82.8 NA 92.9 89.7 94.7 88.6 90.9 90.9 Average value of collateral 114.6 116.0 86.8 NA 109.5 119.5128.0 111.5 127.8 92.8 Average interest rate on loan 16.3 16.7 5.4 NA 15.0 17.4 10.8 17.0 13.2 22.1 Average duration of loan 39.2 43.8 15.1 NA 37.9 40.3 33.5 42.8 43.9 28.4 Share of total borrowing denominated in foreign currency 13.7 8.7 7.7 9.5 7.7 30.9 33.9 8.6 12.6 15.4 Share of long-term liabilities (1 year or more) in total liabilities NA NA 9.4 28.7 NA NA NA NA NA NA Share of short-term liabilities in total liabilities NA NA 47.022.8 NA NA NA NA NA NA Share of equity earnings and retained capital in total liabilities NA NA 43.748.5 NA NA NA NA NA NA Table A5.7 Financial sector--auditing, transaction costs, and property rights a - Non Low High Tanzani Uganda China India Small Large Exporter Exporter Capacity Capacity Share of firms whose financial statements are audited by outside auditors 79.7 59.0 NA NA 73.9 98.5 98.0 75.5 79.1 81.4 Days to clear following payments through your financial institution: Check 5.1 NA 4.5 NA 4.1 7.7 5.0 5.1 4.2 4.5 Domestic currency wire 4.4 NA 4.8 NA 5.0 3.3 4.1 4.5 4.2 4.8 Foreign currency wire 6.3 NA 3.0 NA 7.3 4.8 5.8 6.5 6.9 5.0 Share of land that is: Owned 67.6 62.7 44.0 NA 62.8 79.5 71.4 67.6 69.9 60.9 Leased or rented 32.4 34.3 22.2 NA 37.2 20.5 28.6 32.4 30.1 39.1 Average length of lease/rental contract (years) 1.7 24.3 6.3 NA 1.5 3.3 4.0 1.2 1.6 2.0 Share of buildings that are: Owned 74.8 73.1 56.9 NA 69.5 90.1 83.1 74.1 75.3 73.7 Leased or rented 25.2 26.5 43.2 NA 30.5 9.9 16.9 25.9 24.7 26.3 Average length of lease/rental contract (years) 1.3 2.3 4.4 NA 1.2 2.5 2.1 1.1 1.3 1.5 Table A5.8. Labor and training - China Non Low High Tanzania Uganda mallS Large Exporter Exporter Capacity Capacity Labor composition Share of workers that are permanent 51.00 85.30 56.09 53.00 47.00 51.60 52.00 51.20 51.00 Share of permanent workers that are female 16.62 42.50 13.29 13.10 26.00 22.40 15.40 15.45 19.80 Share of permanent skilled workers that are foreign nationals 2.83 NA 2.94 2.44 4.00 4.30 2.49 2.30 4.26 Labor turnover New employees as share of total 17.13 NA 4.40 16.50 18.16 16.69 15.80 18.00 14.80 Employees that left as share of total 11.1 10.2 1.60 14.70 4.36 5.10 13.00 8.80 9.20 Average time to fill skilled technician vacancy 6.08 5.20 NA 5.30 7.16 5.48 6.30 6.50 5.20 Average time to fill production/service worker vacancy 5.10 13.90 NA 3.60 8.05 5.06 5.10 5.80 3.20 Excess workforce due to regulatory restrictions 21.60 16.70 NA 22.10 20.50 9.75 26.20 26.10 8.36 Training and education Share of workforce with less than 6 years of schooling 48.0 1.1 12.76 49.1 43.5 45.2 48.0 48.8 47.8 Share of workforce with more than 12 years of schooling 13.7 3.0 8.48 12.9 16.0 12.5 19.2 13.9 12.4 Share of firms offering formal training 44.0 71.7 29.66 35.0 71.9 66.6 39.0 40.3 57.1 Share of skilled workers receiving training 7.3 47.7 3.40 4.1 16.6 8.3 6.8 5.8 10.8 Labor unrest Total days lost to labor disputes or civil unrest 0.3 0.3 0.52 0.4 0.02 0.05 0.38 0.29 0.37 Table A5.9. Regulatory burdenand administrative delays, by country China Capacity High Tanzania Uganda Small Large Exporter- Exporter Capacity Non Low % of senior management time spent dealing with 15.7011.80 4.00 14.6018.6018.5015.0816.40 14.05 regulations % of revenues typically paid 1.54 1.40 2.44 2.00 0.22 1.30 1.61 1.80 0.93 to officials to "get things done" % of total firm revenues typically reported for tax 69.1095.9076.7166.6076.5068.0069.6069.50 69.70 purposes Total wait in days for 17.2022.00 NA 13.5026.5028.2014.1016.21 20.82 business registration (days) Inspections a) Total days spent in inspections or required 26.8029.1013.4320.7046.2053.0020.9023.30 39.60 meetings with officials (days) b) % of meetings/inspections 35.4 NA 19.4 36.1 35.3 28 36.1 36.1 36.9 by local authorities c) Total cost of fines or seized goods (% of sales) 0.28 0.6 0.05 0.39 0.01 0.01 0.33 0.4 0.0049 d) % of interactions in which informal payments are 6.6 2.2 6.69 7.1 NA NA 7.6 8.3 NA requested e) If yes, value? (% of sales) 0.59 NA 0.32 0.77 NA NA 0.65 0.83 NA Imports Average time to clear customs (days) 18.3 7.9 5.79 17.9 19.4 16.5 19 18.9 15.4 Longest time to clear customs (days) 32.7 12.5 11.15 35.1 30.36 27.8 34.9534.97 28.5 Exports Average time to clear 11.5 5.4 3.53 16.9 7.8 11.9 11.3 13.35 8.1 customs (days) Longest time to clear 22.2 8.0 6.03 32 15.2823.78 19.2 25.6 16.04 customs (days) Table A5.10. Governance­uncertainty and corruption - Non Low High Tanzania Uganda China Small Large Exporter Exporter Capacity Capacity Uncertainty How consistent are government interpretations of regulations? 56.5048.63 30.00 55.6 61.10 55.50 56.20 56.70 57.10 % of profits reinvested in the firm 28.0024.42 NA 25.7 34.90 37.10 26.30 24.50 38.80 Confidence in the judiciary (% disagree) 52.8033.33 2.40 54.6 50.70 51.80 53.00 52.20 55.70 % of payment disputes resolved in the courts 27.7026.66 5.30 23.0 34.40 NA 30.70 30.00 20.60 Planning horizon for investments (months) 11.80 NA 2.30 11.3 13.70 9.70 12.60 11.50 11.90 Corruption % of revenues typically paid to officials to "get things done" 1.54 2.44 1.40 2.04 0.22 1.30 1.61 1.80 0.93 % saying gift/payment required: To get connected to and maintain public services 18.8030.87 NA 24.6 4.10 32.0 14.20 21.10 11.70 To obtain electrical connection 26.5050.65 NA 32.7 14.80 47.8 18.90 29.60 17.60 To obtain construction permit 15.9039.57 NA 21.7 4.30 38.8 8.10 20 0 To obtain import license 10.1047.71 NA 13.9 4.00 30.0 2.10 10.90 7.60 To obtain trading license 16.8038.56 NA 20.4 6.60 29.4 14.30 18.40 12.50 % of revenue reported by typical establishment for tax purposes 69.1087.31 95.90 60.60 76.50 68.6 69.60 69.50 69.70 Table A5.11. Technology indicators - China Small Large Non Low High Tanzania Exporter Exporter Capacity Capacity Firms with ISO certification (%) 12.3 50.4 4.0 38.5 24.0 9.6 11.8 13.4 Forms of technology innovation (% of firms) Developed new product line 32.7 44.7 30.8 38.5 39.2 31.4 29.7 41.2 Upgraded an existing product line 58.4 56.2 64.6 60.8 58.6 53.8 72.1 Introduced new technology that has changed how main product is produced 32.0 39.5 26.9 46.2 37.3 31.0 27.7 45.6 Discontinued at least one product line 18.2 14.9 29.2 25.5 17.1 17.9 19.1 Agreed a new joint venture with foreign partner 4.1 2.0 10.8 5.9 3.8 3.6 4.4 Agreed a new licensing agreement 8.6 23.7 7.0 13.8 7.8 9.0 7.7 11.8 Form of technology acquisition Embodied in new machinery or equipment 41.1 50.9 38.3 48.4 39.1 42.2 38.8 46.2 Hiring key personnel 32.4 18.9 25.5 51.6 50.0 28.1 29.0 38.5 Licensing or turnkey operations from international sources 4.0 23.7 2.1 9.7 2.2 4.5 4.4 3.1 Licensing or turnkey operations from domestic sources 1.2 1.1 1.6 0.0 1.5 1.6 0.0 Developed or adapted within the establishment 45.1 79.2 47.9 37.1 37.0 47.2 48.6 36.9 Transferred from parent company 7.5 19.4 3.7 19.4 10.9 7.0 6.6 9.2 All other 81.0 55.9 83.0 77.4 82.6 80.9 81.4 81.5 REFERENCES Alesina, Alberto, Silvia Ardagna, Giuseppe Nicoletti, and Fabio Schiantarelli. 2003. Regulation and Investment. NBER Working Paper 9560. Cambridge, Mass.: National Bureau of Economic Research. Beck, Thorsten, Ross Levine, and Norman Loayza. 2000. "Finance and the Sources of Growth." Journal of Financial Economics 58(1­2): 261­300. Bird, Richard M. 2003. "Administrative Dimensions of Tax Reform." University of Toronto, Canada. Broadman, Harry G., and Francesca Recanatini. 2001. "Seeds of Corruption--Do Market Institutions Matter?" MOCT-MOST: Economic Policy in Transitional Economies 11(4): 359­92. Clarke, George R. G, and Scott J. Wallsten. 2004. "Has the Internet Increased Trade? Evidence from Developed and Developing Countries." Trade and Industry Monitor 30: 6-17. Clarke, George R. G, and Lixin Colin Xu. 2004. "Privatization, Competition and Corruption: How Characteristics of Bribe Takers and Payers Affect Bribes to Utilities." Journal of Public Economics 88(9­10): 2067­97. Clarke, George R. G. 2001. "Does Internet Connectivity Affect Export Performance? Evidence from the Transition Economies." World Bank, Washington, D.C. Processed. Cotton, Linda, James Habyarimana, Chad Leechor, Jean Michel Marchat, John Paton, Vijaya Ramachandran, Manju Kedia Shar, Michael Wong, William Kalema, and Charles Ntale. 2004. "Competing in the Global Economy: An Investment Climate Assessment for Uganda." World Bank, Regional Program for Enterprise Development, Washington, D.C. Das-Gupta, Arindam, Michael Engelschalk, and William Mayville. 1999. "An Anti-Corruption Strategy for Revenue Administration." PREM Note 33. World Bank, Washington, D.C. De Wulf, Luc. 2004. "Tradenet in Ghana: Best Practice in the Use of Information Technology." Background paper for World Development Report 2004. World Bank, Washington, D.C. Processed. Djankov, Simeon, Rafael La Porta, Florencio Lopez-de-Silanes, and Andrei Shleifer. 2002. "The Regulation of Entry." Quarterly Journal of Economics 117(1): 1­37. Dollar, David, Mary Hallward-Driemeier, and Taye Mengistae. 2003. "Investment Climate and Firm Performance in Developing Countries." World Bank, Washington, D.C. Processed. Easterly, William, Michael Kremer, Lant Pritchett, and Lawrence H. Summer. 1993. "Good Policy or Good Luck? Country Growth Performance and Temporary Shocks." Journal of Monetary Economics 32(3): 459­83. ESRF (Economic and Social Research Foundation) and (FACEIT) Front Against Corrupt Elements in Tanzania. 2002. "The State of Corruption in Tanzania: Annual Report 2002." Economic and Social Research Foundation, Dar es Salaam. Fjeldstad, Odd-Helge. 2002. "Fighting Fiscal Corruption: The Case of the Tanzania Revenue Authority." Chr. Michelsen Institute, Bergen, Norway. Processed. Freund, Caroline, and Diana Weinhold. 2000. "On the Effect of the Internet on International Trade." International Finance Discussion Paper 693. Board of Governors of the Federal Research System, Washington, D.C. ------. 2002. "The Internet and International Trade in Services." American Economic Review 92(2): 236­40. Haggarty, Luke, Hadji Semboja, Frew Gebreab, Henry Mgombelo, and K. M. N. Kazungu. 2002. "Reform of Telecommunications in Tanzania." World Bank, Washington, D.C. Processed. International Monetary Fund. 2003. "Tanzania: Selected Issues and Statistical Appendix." Washington, D.C. ------. 2004. "First Review under the Three Year Arrangement Under the Poverty Reduction and Growth Facility." Washington, D.C. International Monetary Fund and World Bank. 2003. "Poverty Reduction Strategy Paper-- Progress Report Joint Staff Assessment." Washington, D.C. International Telecommunication Union. 2003. World Telecommunication Indicators, 2002/2003. Geneva. Kaufmann, Daniel, Aart Kraay, and Massimo Mastruzzi. 2003. "Governance Matters III: Governance Indicators for 1996­2002." World Bank, Washington, D.C. Processed. La Porta, Rafael, Florencio Lopez-de-Silanes, and Andrei Shleifer. 2002. "Government Ownership of Banks." Journal of Finance 57(1): 265­301. Levine, Ross, Norman Loayza, and Thorsten Beck. 2000. "Financial Intermediation and Growth: Causality and Causes." Journal of Monetary Economics 46(1): 31­77. Li, Wei, and L. Colin Xu. 2002. "The Impact of Privatization and Competition in the Telecommunications Sector around the World." World Bank, Washington, D.C. Processed. National Bureau of Statistics Tanzania. 2002. Household Budget Survey 2000/01. National Bureau of Statistics, Dar es Salaam, Tanzania. Nicoletti, Giuseppe, and Stefano Scarpetta. 2003. "Regulation, Productivity and Growth: OECD Evidence." Economic Policy 36: 9­51. Rodrik, Dani. 2003. "Growth Strategies." Harvard University, Cambridge, Mass. Processed. Ros, Agustin J. 1999. "Does Ownership or Competition Matter? The Effects of Telecommunications Reform on Network Expansion and Efficiency." Journal of Regulatory Economics 15(1): 65­92. Schneider, Friedrich. 2002. "Size and Measurement of the Informal Economy in 110 Countries around the World." University of Linz, Austria. Processed. Svensson, Jakob. 2003. "Who Must Pay Bribes and How Much? Evidence from a Cross Section of Firms." Quarterly Journal of Economics 118(1): 207­30. Tanzania. 2000. "Tanzania Poverty Reduction Strategy Paper (PRSP)." Dar es Salaam. Tanzania, Ministry of Finance. 2000. "National Microfinance Policy." Dar es Salaam. Tanzania Investment Center. 2001. "Tanzania Investment Report: Report on the Study of Foreign Private Capital Flows in Mainland Tanzania." Dar es Salaam. Transparency International. 2002. "Transparency International Corruption Perceptions Index 2002." Berlin. ------. 2003. "Transparency International Corruption Perceptions Index 2003." Berlin. 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, Mass.: Blackwell Publishers. United Nations Programme on HIV/AIDS. 2002. "Report on the Global HIV/AIDS Epidemic." Geneva. Wallsten, Scott J. 2001. "An Econometric Analysis of Telecom Competition, Privatization, and Regulation in Africa and Latin America." Journal of Industrial Economics 49(1): 1­19. World Bank. 1998. "Project Appraisal Document: Ghana Trade and Investment Gateway Project." Washington, D.C. ------. 2002. Globalization, Growth and Poverty. Washington, D.C. ------. 2003a. Doing Business in 2004. Washington, D.C. ------. 2003b. World Development Indicators. Washington, D.C. ------. 2004a. "Program Document for a Poverty Reduction Support Credit." Washington, D.C. ------. 2004b. "Technical Annex: Emergency Power Supply Project." Washington, D.C. ------. 2004c. World Development Indicators. Washington, D.C. ------. 2004d. World Development Report 2005: A Better Investment Climate for Everyone. Washington D.C. World Bank and International Monetary Fund. 2003. "Financial Sector Assessment Tanzania." Washington, D.C. 1 See Tanzania (2000). The PRSP is available on the World Bank's website at: http://poverty.worldbank.org/files/TanzaniaPRSP.pdf. 2 Tanganyika became fully independent in December 1961, and Zanzibar became independent in December 1963. The two formed the United Republic of Tanganyika and Zanzibar in April 1964. The country was renamed the United Republic of Tanzania on October 29, 1964. 3Between 2000 and 2002, per capita GDP grew at an average annual rate of 3.2 percent in Tanzania and 2.1 percent in Uganda (World Bank, 2004c). 4Haggarty et al. (2002) discuss telecommunications reform in Tanzania. 5Data are from World Bank (2004c) 6Rodrik (2003) notes that relatively modest reforms can result in impressive short-term growth spurts, but that it can be very difficult to sustain these in the medium to long term without more comprehensive reform. Rodrik (2003) lists many examples of short-term growth spurts. On a more general level, Easterly et al. (1993) show that growth rates within countries are only very weakly correlated across decades. 7 Although the difference in growth rates might seem relatively modest, the small differences add up over a decade. Over this period, per capita GDP increased more than 120 percent in real terms in China, 50 percent in India, but only 12 percent in Tanzania. 8Enterprises with fewer than 10 employees are omitted because some of these enterprises might have also been sampled in the micro survey. 9In practice, a small number of the firms included in the informal survey had 10 or more employees. These firms are not included in the averages presented in this section. 10In the productivity analysis below and in Annex 2, many observations are dropped due to missing data. In the sample we use to estimate total factor productivity, we do not find significant differences between capital utilization rates in Tanzania and Uganda (approximately 60 percent in both countries). Capacity utilization is slightly higher in Kenya (63 percent). 11We focus on comparisons with Kenya and Uganda because of concerns about pooling data from China and India with data from the three African countries. Differences in technologies, accounting practices, and sectors of operation make comparisons with the Asian countries difficult. 12As noted below, in Annex 2 we also present results allowing the coefficients on inputs to vary between sections. 13For the same sample that was used in the regression analysis reported in Table 2.3, we find that the median of the capital-labor ratio is 3.6 times larger in Tanzania than in Uganda. However, this sample is much smaller than the sample used to obtain the statistics in Table 2.3 because most of the firms report missing values for the book value of capital, as well as for the depreciation of the capital stock. Both values are used in the construction of the capital measure. 14For presentational ease, this table pools the data for firms across all sectors, allowing for sector-specific intercepts. In Annex 2, we also estimate a production function that allows the capital and labor intensities to differ across sectors. 15 The differences in total factor productivity become statistically significant after controlling for sectoral differences (see Annex 2). 16 See, for example, Freund and Weinhold (2000), Freund and Weinhold (2002), and Clarke and Wallsten (2004) 17 Almost all enterprises with Internet access used email or the World Wide Web to communicate with clients. Several enterprises without Internet access also reported that they used email or the World Wide Web to communicate with clients or suppliers. Managers at these firms may be using home connections or alternative sources of Internet access, such as Internet cafes. 18 Most recent country-specific estimates available from UNAIDS Web site as of end- 2003 taken from United Nations Program on HIV/AIDS (2002). 19Recent estimates for Kenya--including data from the Kenya Demographic and Health Survey (DHS)--suggest that the percentage of people living with HIV/AIDS might be slightly lower than earlier estimates suggested. A press release from UNAIDS, however, questioned the accuracy of results from the Kenya DHS ("UNAIDS response to Kenyan HIV prevalence survey"), noting the high number of people who refused to take the test. 20For a recent summary of literature on this topic, see World Bank (2002). 21The main problem is the data from the Investment Climate Surveys is cross-sectional. Without a panel, it is difficult to know whether improvements in productivity preceded or followed entering export markets. 22Table 2.6 presents results from regressions of a dummy variable representing whether the firm exports or not and a limited dependent variable indicating exports as a percentage of sales. The regression with the dummy variable as the dependent variable is estimated as a probit model. Results from the probit estimation are presented as marginal effects. Since exports are bounded between 0 and 100 percent of sales, the regression with exports as a percentage of sales as the dependent variable is estimated as a Tobit model. 23We measure trade and customs regulations by using a variable that assesses the extent to which trade and customs regulations are perceived as an obstacle to firm growth. To control for the potential for endogeneity, rather than use each firm's own perceptions of trade and customs regulations, we substitute the average perceptions for firms in the same sector and region. Other studies using investment climate measures have used similar approaches to deal with endogeneity. See, for example, Dollar et al. (2003). 24For example, enterprises in Tanzania were more likely to rate high tax rates as a major problem than firms in almost two-thirds of the countries in the 1999 World Business Environment Survey (WBES). The WBES asked firms in all surveyed regions about infrastructure, access to finance, policy instability/uncertainty, inflation, exchange rate, functioning of the judiciary, street crime, and organized crime. The survey also asked about regulatory constraints, including high tax rates and tax administration. 25See International Monetary Fund (2003) 26 Enterprise managers were asked about the `typical firm' in their industry rather than they own firm so that they could avoid implicating themselves. 27It may also reflect the way that the question was asked. If registered enterprises felt that unregistered enterprises in the same sector and of similar size were an "enterprise like yours," they may have included these enterprises in their response. 28International Monetary Fund (2003). 29International Monetary Fund (2003). 30Reforms included the establishment of the independent Tanzania Revenue Authority in 1996, measures designed to limit political interference in tax administration and to allow the authority to pay salaries that were higher than they could if the agency remained part of the civil service. See Fjeldstad (2002). 31 Median losses were 0 percent of production in Uganda and China and 3 percent in Kenya. Mean losses were positive for all countries, but still higher in Tanzania than elsewhere. 32Data is available on the U.S. Energy Information Agency web page (http://www.eia.doe.gov/emeu/international/electric.html#IntlCapacity). 33Evidence suggests that causation runs from financial sector development to growth, not the other way. See, for example, Beck et al. (2000) and Levine et al. (2000) 34Moreover, firms that purchase a higher percentage of inputs with supplier credit tend to have higher labor productivity. 35 La Porta et al. (2002) show that government ownership of banks reduces financial sector development. 36 Claims on the private sector include gross credit from the financial system to individuals, enterprises, non-financial public entities not included under net domestic credit, and financial institutions not included elsewhere. 37Money and quasi-money comprise the sum of currency outside banks; demand deposits other than those of the central government; and the time, savings, and foreign currency deposits of resident sectors other than the central government. 38 The median loan period was, however, shorter in China (one year). For loans with periods between two and four years, the median interest rate in China was about 6 percent. 39 Because relatively few enterprises reported having loans, we do not try to break down micro enterprises into small groups (e.g., urban/rural or informal/formal). 40 Comparable data were not available for micro enterprises. 41 In the formal firm survey, funds from NGOs were reported as part of a broader category than included investment funds, special development funds, state services, government, donors, and NGOs. No finer breakdown within this category was available. 42 Kaufmann, Kraay, and Mastruzzi (2003). Tanzania performed slightly better (tied as the 92nd-least corrupt country) than Uganda (113th) in the most recent survey by Transparency International (TI). However, the TI rankings use results from fewer studies than Kaufmann et al. (2003). 43 See Economic and Social Research Foundation (ESRF) and Front Against Corrupt Elements in Tanzania (FACEIT) (2002) for a general summary of the Wairoba Report and progress against corruption over the past decade. 44 Economic and Social Research Foundation (ESRF) and Front Against Corrupt Elements in Tanzania (FACEIT) (2002). 45 Consistent with the improvements observed in Kaufmann et al. (2003), Transparency International's rating for Tanzania has also improved modestly. 46 Annex 2 presents results from the Tanzania Investment Climate Survey that show a positive correlation between the incidence of bribes and average sales growth, labor productivity, and total factor productivity. Svensson (2003) and Clarke and Xu (2004) present similar results for Uganda and countries in Eastern Europe and Central Asia respectively. 47 The Investment Climate Survey asked enterprise managers how long it takes from the time that goods arrive at the point of entry or exit to the time that they clear customs. It did not ask how much time was attributable to customs processing and how much was attributable to port operations because managers typically do not have access to this information. However, for the most part, the problems in Tanzania appear to be related to customs, rather than port performance. For example, the Economist Intelligence Unit (May 2004) notes, "One of the more successful privatizations carried out by the government is that of Dar es Salaam container port. This is now being run by Tanzania International Container Terminal Services (TICTS), a company that is financially controlled by a Hong Kong-based company... While TICTS has benefited from the long-term project embarked on by the government in dredging the port's entry channel, its own investment in facilities has been a key factor in improving services and it now has the capacity to handle 250,000 containers annually." 48 For example, nine enterprises reported typical delays of over 50 days. 49 See Schneider (2002) for more detail. The estimates are based upon physical inputs (electricity), currency demand, and a model (DYMMIC) approach. 50 Djankov et al. (2002) find, using data from the Doing Business database for 85 developing and developed economies, that informality is greater in countries with greater barriers to entry. 51 Broadman and Recanatini (2001) find that corruption is higher in transition economies where entry barriers, which they measure using subjective data from the World Business Environment Survey, are higher. Djankov et al. (2002) find similar results using different data. 52 Using data from 17 high-income OECD economies for several manufacturing and business service industries between 1984 and 1998, Nicoletti and Scarpetta (2003) show that multi-factor productivity is lower in sectors where regulation is stricter. Alesina et al. (2003) present results for investment. Because the measures of regulation are sector-specific and are measured over time, both this study and Nicoletti and Scarpetta (2003) control for country- and sector-specific effects that might affect investment and productivity. 53 Data are from the Doing Business Database. See World Bank (2003a). 54 Comparable data on number of inspections were not available for micro enterprises. 55See World Bank (2003a) for a complete description of the methodology. 56Many studies have shown that telecommunications services expand more when there is competition. See, for example, Ros (1999), Li and Xu (2002), Wallsten (2001). 57 Observers often suggest that Tanzania's improved macroeconomic performance is at least partially due to improved fiscal performance. International Monetary Fund and World Bank (2003). 58International Monetary Fund (2004) discusses recent progress related to tax reform and tax administration reform in greater detail. 59World Bank (2004b). 60World Bank (2004b). 61International Monetary Fund (2004) 62 Ministry of Finance (2000) 63 World Bank and International Monetary Fund (2003) 64See Chapter 6 of World Bank (2004d) for a recent review of this literature 65The absolute scores should be treated with caution due to the relatively large standard errors associated with the estimates. The change in score between 2002 and 2003 was far smaller than the standard error in either year. See Transparency International (2002) and Transparency International (2003). 66World Bank (2004a) discusses progress with corruption and potential future actions in greater detail. 67 One explanation for the low productivity of non-exporting firms is that firms have to be relatively efficient in order to compete on export markets (i.e., the self-selectivity argument). See Tybout (2003) and World Bank (2002) for summaries of the literature on this topic. 68See Cotton et al. (2004). 69 In Tanzania, few firms reported the book value of current assets. This is the main reason why we do not report the results for the relationship between the investment rate (total investment/stock capital) and the investment climate variables. 70 We performed the same analysis including regional dummy variables, but the results for estimated TFP did not depend upon the inclusion of these variables. 71 Firms reporting missing values for electricity, fuel, and other intermediate factor costs were included in the sample. (We treated the missing values as zeros). Firms that reported missing values for raw materials were excluded from the sample. 72 Chapter 2 compares labor productivity in the three countries without accounting for differences in other inputs and finds that labor productivity is higher in Kenya than in Tanzania and Uganda. 73 Ethiopia: Investment Climate Assessment Report, 2003.