71513 v2 Zambia: Second Investment Climate Assessment Business Environment Issues in Diversifying Growth (In Two Volumes) Volume 2: Full Report Final December 2009 1 Contents Acknowledgment 3 Preface 4 Executive summary 5 Chapter 1 Introduction 11 Chapter 2 Productivity and Export Diversification 21 Chapter 3 Productivity and Domestic Market Distortions: Key Business Environment Issues 46 Chapter 4 Access to Finance 82 Chapter 5 Labor Markets and Human Capital 98 Bibliography 119 2 Acknowledgement The Second Investment Climate Assessment of Zambia was written by a team comprising Taye Mengistae (AFTFE, team leader), Sina Joy Grasmann (DECVP), James Habiyarimana ( Georgetown University), Inessa Love (DECRG), Manju Shah (AFTFE), and Colin Xu (DECRG). Marie Sheppard (AFTFE), Alvaro Gonzalez (AFTFP) and Gerardo Corrochano (AFTFE) provided overall guidance and support throughout the duration of the work. Magdi Amin (CEADR), Pablo Fajnzylber (LCSPR), and Jos Verbeek (AFTP2), kindly served as peer reviewers. The 2008 Enterprise Survey on which the report is based was carried out by the survey firm EEC Canada. 3 Preface The aim of the Second Investment Climate Assessment of Zambia is to provide firm level analyses of some of the key business environment factors influencing manufacturing employment and exports in the country. The report builds on an earlier investment climate assessment of Zambia that the World Bank issued in 2004. That report was based on the Productivity and Investment Climate Survey of 2003. The main data sources of the second assessment are the Zambia Enterprise Survey of 2008 and the 2003 survey itself. The report has two self-contained volumes, of which this is the second. The first volume lays out all the essential findings of the assessment without going into detail on underlying arguments and methodologies. This volume is the full report and presents, in addition to the findings of the first volume, details of the arguments and the evidence behind them. 4 Executive Summary This is the second World Bank assessment of Zambia’s investment climate. Its objective is to highlight some of the impediments to growth and export diversification in the current business environment in the country. It is also a sequel to an earlier assessment (World Bank, 2004), which also addressed this same issue. Like the first assessment, it is based on an analysis of enterprise survey data specifically collected for the purpose, namely, the World Bank’s Zambia Enterprise Survey of 2008, which the report analyzes along with data from the Zambia Investment Climate Survey of 2003 and similar data on a group of comparators selected from the World Bank Enterprise Survey (WBES) cross-country database. The comparators are drawn from the region and from high performing middle income economies in other parts of the world. Organization of the report A key component of the business environment in which firms operate is the state of the macro economy in terms of the levels and stability of prices, exchange rates and the cost of borrowing. Zambia’s economic performance in general and the reason why Zambia has not been able to diversify sufficiently into manufacturing and other tradable sectors has a lot to do with trends in these key variables. The report therefore opens with an overview of those trends in the introductory chapter in order to set the stage for the analysis of micro economic aspects of business environment in the following chapters. There are four other chapters in the report. Chapter 2 analyzes manufacturing productivity in an international perspective as a proximate determinant of manufactured exports. Chapter 3 discusses key business environment variables as underlying factors in manufacturing employment and productivity, and draws the main policy implications of the assessment. Chapter 4 is a more in-depth analysis of disparity in access to finance across firms and sectors as a major source of market distortions and allocative inefficiency. Chapter 5 discusses labor market issues with a focus on labor regulation, wage formation and on-the-job training. Productivity and export diversification Zambia needs to diversify its exports into manufactures and services. At present it is not as much of an exporter of labor intensive manufactures as its low wages should have made it to be. This is partly because the effect of low wages on unit labor costs is more than offset by Zambia’s labor productivity shortfalls relative to larger exporters of the products. The labor productivity shortfall itself is in part explained by underinvestment in equipment. However, even if Zambian manufacturing were as capital intensive as any of the comparators used in the report, Zambian workers would still not be as productive as their counterparts in those countries. A large part of the 5 labor productivity shortfall is in fact a total factor productivity (TFP) gap. Some of that gap reflects Zambia’s comparatively low ‘within-firm’ TFP, meaning that, relative to those in the comparators, Zambian manufactures operate, on average, further below the global technological frontier of their respective industries. The rest of the TFP gap is due to the fact that there is greater allocative inefficiency in Zambia’s industries. As a result, the correlation between within-firm TFP and market shares is comparatively weak in Zambia so that low-productivity firms tend to have higher market shares in Zambia than they would have in the comparators. Zambia’s business environment in an international perspective A key message of the report is that Zambia’s business environment has improved dramatically since the first investment climate assessment in 2004. However, there are also major business environment problems that need to be addressed as they are a large part of the reason why manufacturing productivity is as low as it is. The problems have reduced productivity and employment in manufacturing and other tradable sectors in two ways. One of these is that they have raised the cost of doing business in Zambia relative to that in other countries. This aspect of the effect of business environment problems on productivity and employment is analogous to that of an implicit tax imposed on all activities of all producers in Zambia at a flat rate thereby making them that much less competitive in world markets. In order to provide a sense of how much the country could be losing in jobs and productivity when business environment problems raise the cost of doing business in this way, the report compares averages of key business environment indicators between Zambia and other countries. Business Environment and Domestic Market Distortions The second way in which business environment problems are costing Zambia in terms of manufacturing jobs and output is by distorting domestic markets. In practice a business environment problem does not affect all firms to the same degree. It is more likely to affect some firms more than others for a variety reasons, and is therefore unlikely to add to the cost of doing business by the same amount for everyone. The effect of the problem from this point of view resembles not that of a flat tax applied uniformly to everyone, but that of a system of discriminatory tax rates that vary across sectors, locations, and, indeed, firms. Just as taxes of this kind would, differences between firms in the cost of doing business generate losses in employment and productivity by preventing factor productivity from equalizing at the margin across activities and producers, that is, by generating allocative inefficiency. The size of the loss involved here is determined, not by how grave the business environment problem in question is on average in Zambia, but by the variation in the incidence and intensity of the problem across domestic firms, the rule being that the greater is the dispersion across firms, the larger the loss in aggregate employment, productivity and exports. The basic point is that even if Zambia had the best set of average indicators of business environment on all dimensions by international standards, it could still in theory be less productive and poorer than comparators if there is too much disparity across firms within Zambia in terms of those indicators relative to the disparity in the other countries. In order to provide some sense of the extent of the loss in manufacturing employment and productivity due to the allocative inefficiency 6 arising from the disparity of business environment within Zambia, the report describes the extent to which key business environment indicators vary across business size and age groups, sectors, and regions. Zambia’s overall business environment has improved drastically According to the 2008 Enterprise Survey Zambia’s business environment has improved drastically on every major indicator since the 2003 survey. Business managers were asked in the 2003 survey to rate 17 potential business environment problems as obstacles to the growth of their businesses on a scale ranging from being ‘no obstacle’ to being a ‘severe obstacle’. The outcome was that ten of those potential problems were rated as major or severe obstacles by at least a third of survey respondents. Many problems were in fact rated in that way by 50 to 80 percent of respondents. These included, in that order of importance, macro-economic instability, high taxes, access to finance, power supply, crime, corruption, competition from informal firms, and skills shortage. By contrast, none of those ten potential obstacles was rated as major or severe by a third or more of respondents in the 2008 survey. Indeed, no other potential business environment problem was rated as such by more than a third of respondents. While much of the change in the average ratings has to do with changes in the composition of the sample between the two surveys in terms of size and sector distribution, a sizeable portion of it holds in most cases when comparisons are made controlling for shifts in sample composition. Moreover, the changes in ratings are matched by consistent improvements in hard indicators. Despite the improvement overall, serious business environment issues were also identified by respondents to the 2008 survey. These included (in more or less in that order of importance): (a) access to finance, (b) taxation, (c) product market competition and (d) the provision of physical infrastructure, particularly of power supply and transport. Competition and trade policies That low productivity firms tend to have larger market shares in Zambia than they would have in more advanced and better performing economies suggests that there should be some scope for increasing productivity through competition policy reforms. Indeed, the government is already considering some specific proposals for such reforms. The proposals include expanding the mandate of the ZCC beyond the regulation of mergers to the detection, prosecution and prevention of cartels and the abuse of dominant market power. These are commendable measures that could lead to significant allocative efficiency gains by reducing entry barriers to domestic industries. However, they are all measures for influencing the behavior of incumbent large players, which behavior is only one among several influences on entry and exit rates and on factor mobility between incumbent in domestic industry. Key among the other influences is the international integration of the economy in terms of the export orientation of domestic industries and the level of import penetration of domestic markets. At least as important are also a range of other business environment factors affecting the ability of potential entrants to respond to new investment opportunities. Some of 7 these relate to the direct regulation of entry by government. Some pose indirect barriers to entry, of which group the most notable are problems of access to finance and taxation and power shortages. Issues in the direct regulation of entry In Zambia, as in other countries, business licensing and the requirement of construction permits constitute the most common forms of direct regulation of entry by government. Since the first assessment, the time and pecuniary costs associated with meeting these requirements have fallen substantially. At the same time FIAS’ latest assessment is that Zambia can increase firm formation and entry rates by cutting the number of days needed to set up a business from the 18 days where it stands today to 8, and by reducing the time needed for obtaining construction permits. Access to Finance According to the Enterprise Surveys, inadequate access to finance is the most widespread indirect barrier to entry and to factor mobility in the current business environment in Zambia. More businesses had access to formal external finance in 2008 than did at the time of the 2003 survey, and therefore far fewer of them were complaining of lack of access in 2008. However, nearly a third of respondents in 2008 survey considered inadequate access to be a serious growth bottleneck-an assessment that is backed by several hard indicators of access. Moreover, as is the case in almost every other country, smaller and start up business have poorer access than larger and longer established ones. This distorts the distribution of factors and market shares between small and large firms, not only directly by impeding the mobility of resources among incumbent operators, but also indirectly by reducing entry and firm formation rates. The loss in employment and productivity resulting from the distortion adds to the loss due to Zambian businesses not having as good access to finance as their counterparts in many of the comparators used in this report. In Zambia the problem of inadequate access to finance is inseparable from that of macroeconomic instability. One of the most important steps the government can take to improve businesses’ access to finance is therefore to achieve lasting price and exchange stability, and to bring down government borrowing under control over the long term. In addition, there are major institutional gaps in the current financial system that need to be filled. To address one of the gaps, Zambia needs to promote the development of microfinance institutions at least to the point that these meet as much of the credit needs of microenterprises as similar organizations are doing in many other countries in the region. A second institutional gap is the lack of a workable credit information system, in which Zambia is being advised to invest in order to encourage commercial banks to participate in the SME credit market. SME’s demand for external finance for investment would continue to go largely unmet if commercial banks continue to stay out of the SME market because of lack of the minimum infrastructure needed for their participation. A reliable credit information system is a key part of that infrastructure. 8 Business taxation Another indirect barrier to entry and to factor mobility relates to small business taxes. Zambia is not a high tax economy by international standards, and the Enterprise Surveys suggest that the average business tax burden has fallen significantly since the 2004 assessment. But one in four business managers still think that the growth of their businesses is being seriously hampered by taxes that are too high. The report argues that the contrast between the relatively high complaint rate against high taxes and Zambia’s relatively low average tax rates is explained by the fact that the complaints have nothing to do with average tax rates, and are driven primarily by the marginal effective tax rate being way higher for SMEs than for large enterprise. Just like the disadvantage that SMEs have in terms of access to finance this is also a source of distortion that protects the market shares of larger firms among incumbents while reducing entry and firm formation rates. Like the disparity in access to finance between the two groups of firms, higher tax burdens on SMEs ultimately means less aggregate employment and lower aggregate productivity. Power shortages Despite unmistakable improvement in the situation in recent years, Zambian industry continues to suffer from serious power shortages. The shortages have meant frequent outages and long queues to get connected to the public grid for start ups and expanding businesses. The outages and connection delays have added significantly to the cost of doing business in Zambia relative to other countries. They are also a major source of allocative inefficiency, not only because they affect smaller and younger firms more than others among businesses already in operation, but also because they reduce firm formation and entry rates. To give a sense of the extent of the shortages, businesses reported to have lost 3.6 percent of yearly revenue to outages on average in 2008 while a business needed 80 days on average to get a new connection to the public grid that year. An indication of the magnitude of the allocative inefficiency associated with the shortages is that losses to outages were significantly higher in manufacturing businesses than in services. Within manufacturing, smaller businesses reported higher losses on average than larger one. The waiting period for getting connected to the public grid was also several times longer for small businesses than for large ones. Ultimately the shortages will be tackled only with the help of large investments in new generating and transmission capacity. The government is trying to facilitate these investments using institutional reforms designed to encourage private sector participation in the power sector. However, for needed investments to materialize in the long term, more rational electricity tariffs need to be instituted in the short term. The tariffs that are currently in force fail to cover the full cost of supply. This has undermined the financial viability of ZESCO. It is also a disincentive for private sector participation 9 in future investment programs. Other measures needed for improving ZESCO’s finances include reducing transmission and distribution losses and resolving the long standing issue of build-up of payment arrears by ZESCO’s public sector customers. Trade facilitation and transport costs Yet another potential source of productivity gains in Zambia relates to trade facilitation and transport costs. The trade liberalizing reforms that Zambia has carried out since 1991 are likely to have led to significant productivity growth in several ways. First, they are likely to have induced a reallocation of market shares from low productivity firms to high productivity ones within domestic industry. Secondly, they are likely to have increased the incentives for innovation by domestic firms. Third, they probably have provided opportunities for greater economies of scale in industries where exports have grown significantly. All indications are also that Zambia can realize more of these gains since there is significant room for opening up its economy even more through further trade facilitating measures and by reducing transport costs further beyond the remarkable achievement of the past decade in that regard. 10 Chapter 1 Introduction Issues Zambia’s GDP needs to grow at more than 7 percent a year if the country is to attain middle income status by 2030 as is envisaged in the Fifth National Development Plan (FNDP). The closest that the country ever got to that kind of growth was in the five years leading up to the current global recession, when output grew at 5-6 percent a year largely on the back of a copper price boom and a major debt relief. The sudden drop in copper prices that came with the global recession and the deceleration in growth it led to underscored a crucial risk factor in the structure of Zambia’ economy : the excessive dependence of growth on exports of a single commodity- copper. The consequences of the recession have been yet another reminder that diversification of exports and the attraction of FDI particularly into non-mineral sectors are necessary conditions for sustaining growth in what has been a low-income, low-savings economy since the 1980s. Ultimately diversification of exports can only come about as part of broader economic diversification, which is needed for faster and more equitable growth. As a low income economy, Zambia is characterized by extensive poverty and high unemployment, for which part of the blame lies in the concentration of exports, infrastructure and public services in the relatively capital intensive mining sector. Objectives and methodology The objective of the Second Investment Climate Assessment of Zambia is to help highlight some of the factors in the current business environment of the country that are impeding the diversification of growth and exports into manufacturing and services. The report builds on an earlier e assessment (World Bank, 2004), which also addressed this same issue using more or less the same methodology as that of this assessment. Like the first assessment, it is based on an analysis of enterprise survey data specifically collected for the purpose, namely, the World Bank’s Zambia Enterprise Survey of 11 2008, which the report analyzes along with the Zambia Investment Climate Survey of 2003 used in the first assessment and similar data a group of comparators selected from the broader World Bank Enterprise Survey (WBES) cross-country database. This database is maintained by the Enterprise Survey Unit of the World Bank and consists of returns to standardized surveys on more than 100 countries. The World Bank’s Enterprise Survey data contain simultaneous observations on policy variables and economic outcomes at the level of the individual business establishment. They are therefore useful for analyzing the correlations between employment, investment and exporting decisions and various elements of the business environment conditioning those decisions. As a way of highlighting the more important of those correlations the report benchmarks the performance of Zambian enterprises and their business environment against those of other economies in Sub-Saharan Africa and high performing middle income counties in other regions. Included in the first group are Angola, Botswana, Kenya, Mozambique, Nigeria, Tanzania, Uganda, and South Africa. Among the middle income comparators from other parts of the world are China, Malaysia, Thailand, Chile and Colombia. The Enterprise Surveys of 2003 and 2008 The World Bank Enterprise Survey of Zambia (2008) As the main source of the data for this report, the 2008 Enterprise Survey of Zambia was carried out by the survey firm, EEC Canada, and took place in the last quarter of 2007 and the first quarter of 2008. Full description of the sample design is provided in an annex to the full report. The total sample consisted of 603 enterprises selected from the cities of Lusaka (64 percent), Kitwe (13 percent), Ndola (13 percent), and Livingston (10 percent) and from about half a dozen two-digit manufacturing industries, retails and wholesales trade and other services (table 1.1). One hundred nineteen observations of the full sample were micro enterprises, which we define as establishments employing less than five workers. The remaining enterprises employed five workers or more. The remaining enterprises employed five workers or more. In addition to data on firms, the survey also collected labor market information on a sample of well over a 1000 workers selected from about one third the manufacturing establishments in the enterprise survey sample. 12 The Investment Climate Survey of Zambia (2003) Of the total sample of the 2008 Enterprise Survey, 87 businesses were revisits to a part of the sample of the Investment Climate Survey of 2003, on which the first assessment was based. We thus have repeat observations on a range of business environment and business performance variables in a four-year interval over a sizeable number of enterprises. The Investment Climate Survey of 2003 itself covered 207 manufacturing establishments. These too were drawn from the cities of Lusaka, Kitwe, Ndola and Livingston and from more or less the same industries as those covered by the 2008 survey (table 1.1). Survey Instruments The instrument of each of the 2003 and 2008 surveys was a written questionnaire that enumerators administered to enterprise managers through face to face to interviews, with significant variation in the instruments used for the survey of micro enterprises from those applied to larger manufacturers, which, in turn, differed from the survey instrument for retail businesses. Each variant of the instrument generated information on four broad areas with comparable observations across the two waves: managers’ ratings, on a common scale, of different aspects of their business environment; objective indicators of the various dimensions of same environment; financial, production, employment, assets, sales and technological information needed for the measurement of business productivity and growth; and key business characteristics such as business age, forms of business organization, and other entrepreneurial characteristics. Organization of the report A key component of the business environment in which firms operate is the state of the macro economy in terms of the levels and stability of prices, exchange rates and the cost of borrowing. Zambia’s economic performance in general and the reason why Zambia has not been able to diversify sufficiently into manufacturing and other tradable sectors has a lot to do with trends in these key variables. The report therefore opens with an overview of those trends in the introductory chapter in order to set the stage for the analysis of micro economic aspects of business environment in the following chapters. There are four other chapters in the report. Chapter 2 analyzes manufacturing productivity in an international perspective as a proximate determinant of manufactured exports. Chapter 3 discusses key business environment variables as underlying factors in manufacturing employment and productivity, and draws the main policy implications of the assessment. Chapter 4 is a more in-depth analysis of disparity in access to finance across firms and sectors as a major source of market 13 distortions and allocative inefficiency. Chapter 5 discusses labor market issues with a focus on labor regulation, wage formation and on-the-job training. Table 2.1 : Distribution of Enterprise Survey Sample by Business Chacteristics 2008 sample 2008 and 2003 pooled Number Percent Number Percent Distribution by industry groups: Food 117 19% 175 22% Textile and garments 80 13% 97 12% Machinery and metal products 40 7% 48 6% Chemicals and plastics 29 5% 41 5% Other manufacures 71 12% 87 11% Retail and wholesale trade 213 35% 214 26% Other services 44 7% 54 7% Other 9 1% 30 4% Total 603 100% 810 100% Distribution by location: Lusaka 383 63.52 487 60.2 Kitwe 81 13.43 106 13.1 Ndola 77 12.77 124 15.33 Livingston 62 10.28 93 11.37 Total 603 100 810 100 Distribution by employment size groups : Fewer than 5 workers ("micro") 119 20% 119 15% 5-49 workers ("Small") 366 61% 453 56% 50+ workers ("Large") 118 20% 238 29% Total 603 100% 810 100% Distribution by age groups : 1-9 years since established ("Young") 372 61.69 457 56.42 10 years or more since established ("Established") 231 38.31 353 43.58 Total 603 100 810 100 Distribution by exporting status: Non-exporter 533 88.39 656 80.99 Exporter 70 11.61 154 19.01 Total 603 100 810 100 Distribution by foreing equity participation: No foreign share in owneship 479 79.44 625 77.16 Some foreign ownership 124 20.56 185 22.84 Total 603 100 810 100 14 Macroeconomic background Economic Growth At independence, Zambia’s GDP per capital was several times larger than that of many of today’s successful economies in South East Asia. Yet Zambia today is one of the poorest countries in the world after decades of uninterrupted contraction in per capita incomes that came to an end only at the turn of the century, when a series of policy reforms that government undertook in the 1990s started bearing fruit to generate positive growth that continued up until the onset of the current global recession (figure 1.1). The Zambia Investment Climate Survey of 2003, on which the first assessment was based, captured the state of the manufacturing sector early in the upturn of 2001- 2008, but when the state of business confidence in the recovery was not very high. By the of the 2008 Enterprise Survey there had been several years of uninterrupted GDP growth which accelerated to a 6 percent annual rate, and manager’s perception of Zambia’s investment climate had consequently become quite positive as reflected in the new survey. Then came the global recession, in the light of which Zambia’s growth for 2009 is projected to fall to 4 percent a year before slightly recovering to 4.5 percent in 2010. While no one knows for sure what growth will look like over the next two to three years, it is clear that even the best of Zambia’s growth performance of the past five years falls short of the region’s average, and certainly is much less than those of many upper middle income countries (figure 1.1) and the annual average growth rate of 7 percent a year that would need to be realized over next two decades in order for Zambia to become a middle income economy by 2030 as it aspires to. Trade policy and export diversification Zambia has liberalized its trade policy a great deal since the early 1990s, having removed quantity restrictions and import licensing altogether and reduced and rationalized import tariffs. In 2003, it moved into a market determined exchange rate regime and stopped exchange controls. Largely as a result of those measures, Zambia’s exports are far more diversified today than they were in the mid 1990s, when copper and cobalt alone accounted for some 85 percent of merchandize exports. At the moment the share of non-traditional items including agricultural commodities, horticultural products and processed food constitute stands at more than 35 percent of merchandize exports. Combined with a series of other investor friendly policy changes including large scale privatization of mining 15 operations, trade and exchange rate policy reforms have also led to growth in FDI not only into mining but into agriculture and manufacturing as well. Figure 1.1 GDP per capita in 2000 US dollars, 1962-2008 However, copper exports and movements in the world price of copper still continue to dominate growth and employment outcomes. Just as the secular decline in the world price of copper was the single most important factor behind the persistent fall in per capita incomes since independence and up until turn of the century, a copper price boom was the main driver of the turnaround in growth observed during 2002 to 2008(figure 1.2). It was also the sharp dip in copper prices that the current global recession prompted which brought about the deceleration in growth being witnessed today, yet underscoring the fragility of any growth trends not based on sufficient diversification of exports away from copper. 16 Investment and Savings Exports have been of far greater importance in driving growth in Zambia than they would be in otherwise comparable economies because the investment rate has been relatively low in reflection of Zambia’s low domestic savings ratio which, while quite high compared to the average for Sub Saharan Africa as a whole, has also been always less than half the average for East Asia, and significantly less than the average for other resource-rich economies in the region (figure 1.3). In recent years much of the savings gap has been financed by the comparatively high volume of FDI that post 1991 reforms have helped attract (figure 1.4). Figure 1.2: Source: World Bank Zambia Competitiveness Study Mission, July 2009. 17 Figure 1.3 Gross fixed capital formation as percent of GDP Source : World Development Indicators (2009) Figure 1.4 FDI net average inflows 2002-2008 5.0 4.5 4.3 Foreign Direct Investment, 2002-2005 % average 4.0 4.0 3.9 3.5 3.1 3.1 3.0 2.9 2.7 2.5 2.0 1.7 1.5 1.1 1.0 0.7 0.5 0.3 0.1 0.0 d a a i a a na ya a s a c aw in ic n tiu ni bi ic si if i ila en a fr za h fr am ay ac al ri sw C A A ha au K M an al P Z n th ot T M M a T d ou B ar an ah S a -S si A ub t as S E 18 Interest rates and inflation High inflation and the high cost of borrowing were part of the forces that kept the investment rate low up until 2001. As late as 2003 the rate of consumer price inflation was more than 20 percent while the nominal lending rate averaged more than 40 percent, before the inflation rate eventually dropped to single digit (figure 1.5) and the lending rate fell almost by half as the copper price boom and the debt relief program of 2005 helped curtail the fiscal deficit from 13.4 percent in 2003 to 7.6 percent in 2006. Figure 1.5 Zambia: annual CPI inflation rate (percent), 1986-2008 19 Figure 1.6 Zambia: Real effective exchange rate (2000=100), 1980-2008 Real exchange rates The exchange rate of the Kwacha has been market determined since Zambia liberalized trade in its national currency in 2003 and stopped exchange controls. While the resulting flexible exchange rate provides an automatic balance of payments adjustment mechanism to swings in copper prices and other external shocks, this comes at the price of often extreme exchange rate volatility, which tends to work against fixed investment. After a very brief period of relative stability against the United States Dollar, the Kwacha appreciated sharply from November 2005 on the back of the copper price boom all the way up to the third quarter of 2008 when it depreciated sharply as copper prices tumbled with the onset of the global recession (figure 1.6). Between the third and fourth quarters of 2008, copper prices fell by about 60 percent while the Kwacha fell by 38 percent against the US Dollar between the fourth quarter of 2008 and the first quarter of 2009. The Kwacha has been appreciating steeply since the second quarter of 2009 as copper prices started to recover. 20 Chapter 2 Productivity and Export Diversification Introduction This chapter discusses proximate determinants of exports and productivity in Zambia’s manufacturing industries based on the 2008 and 2003 Enterprise Survey data. The aim is to lay the ground for an analysis in the next chapter of more deep-seated barriers to export diversification in the current business environment. It is clear that Zambia’s exports of manufactures in the five years leading up to the global recession would have been higher than they turned out to be if it were not for the significant appreciation of the Kwacha (figure 1.6) associated with the copper price boom of the time (figure 1.3). The 2008 Enterprise Survey suggests that other forces also helped bar Zambian manufacturers from export markets at the time. Zambia’s manufactured exports should have been higher than they actually turned out to be at the going exchange rates given Zambia’s low wages. The most immediate reason why they were not was that labor productivity in Zambia was even lower compared to that in more successfully exporting comparators. Zambia’s labor productivity shortfall itself was partly explained by the fact that Zambian workers were not as well equipped with fixed assets as their counterparts in those countries. It was partly due to the fact that industry aggregate TFP in Zambia was lower than in comparators. The lower industry aggregate TFP in Zambia itself partly reflected that the typical Zambian manufacturer had lower TFP than the typical manufacturer in the comparators. It partly reflected the greater allocative inefficiency of Zambian industry in the sense that low productivity firms tended to have larger market shares than they would have in more successfully exporting comparators. One of the indicators of allocative inefficiency in Zambian industry is the persistent gap in the rate of return to capital that the Enterprise Survey data show to exist between smaller firms and larger businesses. The rate of return to capital and the marginal productivity of capital are both persistently higher in smaller businesses employing less than 50 workers than they are in larger businesses of the same age group. This naturally raises the question of why smaller businesses are not investing more to take advantage of the higher expected rate of return. Part of the answer is 21 that small business investments are riskier than investments by larger firms. It is also often the case that smaller businesses are more risk averse. A part of the rate- of- return advantage of smaller firms is thus likely offset by larger risk premium that lenders and business owners themselves attach to small business investment. But there is also a part that cannot be explained by differences in risk exposure or in attitudes to risk- a part that should have led to higher investment by smaller businesses than what is actually observed in the data were it not for the presence of problems of business environment that deter the investment. An important example of such deterrence is that smaller businesses have inadequate access to credit to finance their investment projects because lenders discriminate against them for reasons unrelated to their assessment of the risk of default. Other instances of deterrents to small business investment are that smaller businesses have poorer access to other publicly provided goods such as reliable power supply and that business taxes in Zambia are regressive in that they penalize small business investments in favor of investments by larger firms. Measures of International Integration Zambia’s economy is fairly open to trade A measure of the openness of Zambia’s economy is that exports and imports have always added up to well over 50% of GDP. Zambia has also attracted in recent years more FDI relative to its GDP than most other countries in Sub-Saharan Africa, thanks, in part, to the wide range of policy reforms that has carried out since the early 1990s –including a major privatization program, and trade and exchange rate policy liberalization-. Thus, over the period 2002-2005, Zambia’s inward FDI averaged 4 percent of the country’s GDP, which rate was more than 50 percent higher than the average for Sub-Saharan Africa, and is quite high by the standards of middle income countries in the region and beyond including resource-rich ones (figure 1.4). 22 Table 2.1 : Distribution of Enterprise Survey Sample by Business Chacteristics 2008 sample 2008 and 2003 pooled Number Percent Number Percent Distribution by industry groups: Food 117 19% 175 22% Textile and garments 80 13% 97 12% Machinery and metal products 40 7% 48 6% Chemicals and plastics 29 5% 41 5% Other manufacures 71 12% 87 11% Retail and wholesale trade 213 35% 214 26% Other services 44 7% 54 7% Other 9 1% 30 4% Total 603 100% 810 100% Distribution by location: Lusaka 383 63.52 487 60.2 Kitwe 81 13.43 106 13.1 Ndola 77 12.77 124 15.33 Livingston 62 10.28 93 11.37 Total 603 100 810 100 Distribution by employment size groups : Fewer than 5 workers ("micro") 119 20% 119 15% 5-49 workers ("Small") 366 61% 453 56% 50+ workers ("Large") 118 20% 238 29% Total 603 100% 810 100% Distribution by age groups : 1-9 years since established ("Young") 372 61.69 457 56.42 10 years or more since established ("Established") 231 38.31 353 43.58 Total 603 100 810 100 Distribution by exporting status: Non-exporter 533 88.39 656 80.99 Exporter 70 11.61 154 19.01 Total 603 100 810 100 Distribution by foreing equity participation: No foreign share in owneship 479 79.44 625 77.16 Some foreign ownership 124 20.56 185 22.84 Total 603 100 810 100 Table 2.1: Distribution of Exporters and Foreign Invested Enterprises by Sector 2008 sample only Exporters Foreign invested enterprises Number Percent Number Percent Food 24 19.7 14 20.0 Textile and garments 8 6.6 13 18.6 Machinery and metal products 11 9.0 10 14.3 Chemicals and plastics 11 9.0 13 18.6 Other manufacures 17 13.9 9 12.9 Retail and wholesale trade 43 35.3 9 12.9 Other services 8 6.6 1 1.4 Other 1 1.4 Total 122 100 70 100 Some of this FDI went beyond mining to secondary and tertiary sectors including manufacturing and retail trade. Although the Enterprise Survey did not collect data on FDI, it gathered information on equity holdings of foreign investors in enterprises, which should capture the cumulative outcome of FDI inflows over the years leading up to the survey. Thus, that Zambia has attracted more FDI into its manufacturing and retail sectors than most other countries in the region in recent years shows up in the percentage of foreign invested enterprises being higher in the Zambia sample than in samples from the comparators shown in figure 2.1. Indeed the only fast growing middle income comparators for which foreign equity holdings are higher than those of Zambia in the chart are China and Thailand. Table 2.1 suggests that, within Zambia, FDI has been 23 reasonably evenly spread across industries with slightly greater concentration in food processing, textiles and chemical and plastics. Figure 2.1: Foreign invested firms-% of sampl e Botsw ana Sw aziland China Gambia Thailand Zambia08 Namibia Morocco Mozambique DRC Burundi Uganda06 Chile Rw anda South Koea A ngola A rgentina South A f rica08 Mauritania Tanzania06 Kenya07 Spain Poland Germany Guinea Guinea-Bissau Mexico Egypt Senegal Mali Brazil Colombia Nigeria 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 Export orientation and reliance on imported inputs Zambian industry is also highly dependent on imported inputs. According to the Enterprise Survey of 2008 about 35 percent of Zambian manufacturers import their supplies. This is about average for Sub-Saharan Africa, being a little lower than the average for resource-rich middle income countries in the region and a little higher than that for low income countries, but still makes Zambian industry far more reliant on imported inputs than nearly all the middle income comparators shown in figure 2.2. On the other hand, Zambian manufacturers are far less export oriented than their counterparts in middle income comparators from Asia and Latin America, although their export market participation rate is also average for Sub-Saharan Africa, being significantly lower than those of countries like Kenya, Namibia and Swaziland, but significantly higher than those of counties like Angola, Mozambique, Nigeria, and Tanzania (figure 2.3). 24 Figure 2.2: Share of imported inputs *% ) Morocco Rw anda Namibia Botsw ana Mauritania Gambia Guinea-Bissau Sw aziland Guinea Burundi Angola Zambia08 Malaysia Kenya07 DRC Senegal Argentina Mali Tanzania06 Egypt Colombia Poland Uganda06 South Koea Chile Mozambique South Af rica08 Spain Thailand Germany China Mexico Brazil Nigeria 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 The sharp contrast between the low export orientation of the Zambian Enterprise Survey sample and their comparatively high reliance on imported inputs illustrates the distance that Zambia has yet to cover in terms of its agenda of diversifying exports into manufacturing. This is because the coverage of survey basically coincided with the very labor intensive lines identified by the DTIS as the most promising in terms of exporting potential including food processing, textiles and garments, metal products (Zambia DTIS, 2005). The urgency of increasing the export orientation of labor intensive manufacturing industries such as these is underscored by strong indications in the survey data that the scope for increasing the rate of inward FDI as an alternative to export diversification as means of financing growth in imports has been diminishing since before the onset of the global recession. 25 Figure 2.3. Exporters-% of sample Thailand04 Morocco04 Malaysia A rgentina Chile China Egypt Spain Poland Brazil SouthKorea Kenya07 Colombia SouthA f rica08 Germany Namibia Guinea Sw aziland Senegal Mexico Zambia08 Uganda Mali Botsw ana Tanzania Mauritania Rw anda Gambia Guinea-Bissau Mozambique Drc Burundi A ngola Nigeria 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 Figure 2.4A: Value added per unit of fixed assets- MIC comparators Argentina South Africa03 South Africa08 Colombia Chile Mexico Morocco China Thailand Zambia08 Zambia03 Malaysia 0.00 2.00 4.00 6.00 8.00 10.00 26 Figure 2.4B:Value added per unit of fixed assets, SSA comparators Guinea Swaziland Gambia South Africa08 Burundi Angola Nigeria Tanzania06 Guinea-Bissau Botswana Rwanda Zambia08 Uganda06 DRC Mauritania Namibia Kenya07 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 One of the indications is that the rate of return on investment in Zambian industry is low compared to other African countries and to middle income comparators from other regions. Specifically, the marginal revenue productivity of capital is lower in Zambian manufacturing than in the manufacturing sectors of those comparators (figures 2.4A and 2.4B). There is also a similar pattern is the average gross profitability of fixed assets, which is lower for Zambian enterprises (figures 2.5A and 2.5B). We see in figure 2.5A that, although the expected gross rate of return on fixed investment is higher in Zambia than in countries like China and Malaysia, it is also significantly lower in economies that are far wealthier than Zambia, such as South Africa, Brazil and Argentina. And, although the same rate of return is higher in Zambia than in most comparators from Sub-Saharan Africa, it is also lower than many in its immediate neighborhood including Tanzania and Swaziland (figure 2.5B). Figure 2.5A: Gross profits per unit fixed assests Brazil South Africa03 Colombia Argentina South Africa08 Chile Zambia08 Zambia03 Mexico Thailand China Malaysia 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 27 Figure 2.5B: Gross profits pe r unit fix e d a sse sts, SSA compa ra tors Guinea Swaziland Burundi Tanzania06 South Africa08 Nigeria Zambia08 Botswana Namibia DRC Angola Gambia Kenya07 Rwanda Uganda06 Mauritania Guinea-Bissau 0.00 0.50 1.00 1.50 2.00 2.50 3.00 Not only is the export market participation rate of Zambian firms relatively low, but the rate has also declined in recent years as indicated by the fact that it was smaller for the 2008 sample than it was for the 2003 sample (table 2.3). Moreover, table 2.5 suggests that the prospects of the rate increasing are not good and had begun to deteriorate prior to the onset of the global recession. The main predictor of the deterioration in those prospects in the enterprise survey data is the declining marginal revenue productivity of capital in exporting businesses (columns 7 and 8). 28 Table 2_3: Probit model of export market participation Marginal Effects Dependent variable: dummy=1 for exporter 2008 sample 2003 sample Pooled Pooled 2008 sample 2003 sample Pooled Pooled Simple Simple Simple Radom effects Simple Simple simple Random effects probit probit probit probit probit probit probit probit Foreign invested 0.118 0.192 0.151 1.475 0.101 0.239 0.145 1.514 (4.29)** (4.43)** (6.08)** (6.41)** (3.45)** (4.08)** (4.95)** (5.15)** 2008-survey -0.279 -1.018 (8.46)** (4.41)** Garment and textiles 0.044 0.308 0.096 0.727 (1.35) (4.19)** (2.64)** (1.51) Metal products 0.093 -0.403 0.001 0.023 (1.98)* (6.14)** (0.03) (0.04) Chemicals and plastics 0.253 -0.042 0.156 1.857 (3.68)** (0.45) (3.07)** (2.82)** Other manufacures -0.002 -0.060 -0.020 -0.643 (0.08) (0.74) (0.75) (1.18) Retail and wholesale trade -0.085 -0.135 -2.403 (4.91)** (6.49)** (4.85)** Other services -0.082 -0.197 -0.113 -2.049 (4.73)** (2.16)* (4.55)** (2.79)** Other 0.096 0.214 1.518 (0.69) (1.43) (0.97) Constant -2.431 -1.493 (9.80)** (3.87)** Year dummies Yes Yes Yes Yes Yes Yes Yes Yes Enterprises 723 696 Observations 1206 621 1827 1827 1198 363 1567 1567 Sigma_u 2.66 2.74 Rho 0.86 0.88 Log likelihood -409.6 -421.3 -830.9 -542.7 -590.5 -225.7 -624.9 -388.5 Absolute value of z-statistics in parentheses * significant at 5% level; ** significant at 1% level Figure 2.6A: Ave ra ge a nnua l w a ge s pe r w orke r ('000 USD)- Middle income compa ra tors Argentina Chile South Africa08 South Africa03 Brazil Malaysia Poland Mexico Colombia China Morocco Zambia08 Thailand Zambia03 0.0 5.0 10.0 15.0 20.0 29 Figure 2.6 A2: Annual Sales and wages per worker ('000 USD)-middle income comprators Argentina Chile South Africa08 South Africa03 Brazil Malaysia Poland Mexico Colombia China Morocco Zambia08 Thailand Zambia03 0.0 20.0 40.0 60.0 80.0 100.0 120.0 Sales Wages Table 2_4: Probit model of being foreign invested Marginal Effects Dependent variable: dummy=1 for foreign ownership share Simple Probit Random effects (1) (2) (3) Probit Garment and textiles -0.110 -0.025 -0.087 -1.174 (3.37)** (0.34) (2.80)** (2.22)* Metal products 0.068 0.121 0.087 0.698 (1.16) (1.12) (1.68) (1.14) Chemicals and plastics 0.170 0.244 0.198 1.631 (2.37)* (2.68)** (3.49)** (2.40)* Other manufacures 0.033 0.056 0.044 0.015 (0.74) (0.73) (1.15) (0.03) Retail and wholesale trade -0.003 0.007 -0.200 (0.10) (0.23) (0.49) Other services -0.023 0.044 -0.003 -0.179 (0.48) (0.47) (0.06) (0.29) Other -9.734 (0.08) 2008 survey -0.078 -0.087 (1.92) (0.28) Constant -2.369 (6.13)** Observations 1188 363 1554 1567 Year dummies? Yes Yes Yes Yes Number of Enterpises 696 Sigma_u 2.88 Rho 0.89 Log likelhihood -590.5 -216.3 -808.9 -508 Absolute value of z-statistics in parentheses * significant at 5% level; ** significant at 1% level 30 Table 2_5: Exporting and the Average/Marginal Revenue Product of Capital OLS and Random GLS estimates Dependent variable= Log value added per unit book value of fixed assets (1) (2) (3) (4) (5) (6) (7) (8) 2008 Survey 2003 Surve Pooled-OLS Pooled GLS 2008 Survey 2003 Survey Pooled OLS Pooled GLS Exporter 0.035 -0.788 -0.553 -0.552 0.001 -1.053 -0.655 -0.493 (0.14) (5.95)** (4.64)** (4.08)** (0.00) (5.87)** (4.50)** (2.91)** 2008 Survey -16.183 -0.044 -16.388 -15.845 (10.91)** (0.33) (10.73)** (14.04)** Garment and textiles 0.298 0.199 0.238 0.284 (1.17) (0.82) (1.34) (1.36) Metal products 0.024 -0.070 0.169 0.179 (0.07) (0.21) (0.72) (0.65) Chemicals and plastics 0.213 1.216 0.796 0.631 (0.61) (3.91)** (3.39)** (2.18)* Other manufacures -0.306 0.440 -0.033 -0.187 (1.20) (1.51) (0.17) (0.83) Retail and wholesale trade -0.312 -0.171 -0.104 (0.32) (0.16) (0.07) Other services 0.078 0.052 -0.101 (0.25) (0.15) (0.21) Constant 0.432 0.613 0.521 0.549 0.417 0.617 0.488 0.431 (4.15)** (4.68)** (3.82)** (4.35)** (2.70)** (3.14)** (2.53)* (2.28)* Year dummies? Yes Yes Yes Yes Yes yes Yes Yes Enterprises 402 377 Observations 288 456 744 744 288 269 557 557 R-squared 0.26 0.07 0.16 0.28 0.17 0.21 Sigma_u 1.13 1.28 Rho 0.56 0.69 Absolute value of t-statistics in parentheses * significant at 5% level; ** significant at 1% level Productivity and Exports The effect of low wages on exports is being offset by that of low labor productivity A major contributing factor for declining export market participation by Zambian manufacturers in recent years undoubtedly was the steep appreciation of the Kwacha between 2003 and 2008 (figure 1.6). However, even controlling for exchange rate movements, export market participation rates could have been higher than shown in figure 2.3, given that Zambian wages have always been low compared with those of middle income comparators, and did not rise significantly over the period in question (figures 2.6A and 2.6B). Zambia’ exports of manufactures are as low as they are because the country’s advantage over middle income countries in terms of lower wages is more than offset by its disadvantage in terms of labor productivity (figures 2.6A2). Indeed Zambia’s unit labor costs are higher than those in countries like China, Brazil and Colombia, and have been so at least since the 2003 survey (figure 2.7A). Within 31 Africa, Zambia’s unit labor costs are significantly higher than those of Kenya, Swaziland, Botswana and Namibia (figure 2.7B). These countries and nearly all of the middle income comparators in fact have far higher average wages than Zambia. They therefore owe their lower unit labor cost shown in figure 7A to their average labor productivity being even higher than Zambia’s (figures 2.6B2). Figure 2.6B: Average annual wages per worker ('000 USD) South Africa08 Namibia Botswana Kenya07 Mozambique Swaziland Zambia08 Angola Tanzania06 Guinea-Bissau Gambia Mauritania Rwanda Uganda06 Senegal Nigeria Guinea Burundi DRC Mali 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 32 Figure 2.6B2: Annual sales and wages per woker ('000 USD)-SSA comparators South Africa08 Namibia Botswana Kenya07 Zambia08 Tanzania06 Mozambique Mauritania Swaziland Rwanda Burundi Uganda06 Guinea Angola Gambia Senegal Nigeria DRC Mali Guinea-Bissau 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 Sales Wages Proximate causes of low labor productivity Part of the reason that labor productivity is so much lower in Zambian manufacturers relative to their counterparts in East Asia and other comparators is that Zambian workers are not as well equipped with fixed assets. Thus the value of fixed assets per worker in Zambia is lower than that in all middle income comparators shown in figure 2.8A except those of Malaysia and Thailand. Within Africa, production is more capital intensive than Zambia in countries where manufacturing labor productivity is also higher than in Zambia. This group includes Kenya, Botswana and Namibia (figure 2.8B). 33 Figure 2.7B: Unit la bor cost, SSA compa ra tors South Africa08 Mozambique Uganda06 Senegal Angola Guinea Mali Guinea-Bissau Nigeria Rwanda Tanzania06 Zambia08 Gambia Kenya07 DRC Swaziland Mauritania Burundi Namibia Botswana 0.00 0.10 0.20 0.30 0.40 0.50 It also turns out that Zambia’s labor productivity shortfalls with respect to any of those comparators would not vanish even if Zambia’s capital to labor ratios were as high as the comparators’. In other words, part of the reason for the shortfall is that the average total factor productivity (TFP) is lower in Zambia (figure 2.9). Figure 2.8A: Book value of fixed assets pe r w orker ('000 USD)-middle income compa rators Argentina Chile Poland Morocco Brazil South Africa08 South Africa03 Colombia Mexico Zambia08 Zambia03 Thailand Malaysia 0.0 5.0 10.0 15.0 20.0 25.0 34 Productivity and allocative efficiency in domestic industry The aggregate TFP index shown in figure 2.9 is a weighted average of the TFP of the individual enterprises constituting the industry sample, with enterprise market shares as weights.1 It is therefore calculated as the sum of the (unweighted) average of enterprise level TFP -also known as within-firm TFP-and the sample covariance between enterprise TFP and enterprise market share. A positive covariance term implies that more productive firms have higher market shares. Considering changes over time this means that it is not necessary that average within-firm TFP increases for aggregate industry productivity to grow. The significance of this kind of productivity decomposition is that it highlights the fact that aggregate (or industry level) TFP often increases or falls even in the absence of significant changes in the average within-firm TFP as a result of the reallocation of market share between low productivity firms and high productivity ones. Figure 2.8B: Book value fixed assets per w orker ('000 USD)- Sub Saharan Africa Namibia Rwanda Kenya07 South Africa08 Tanzania06 Botswana Mauritania Zambia08 Burundi Gambia Uganda06 Angola Swaziland Nigeria DRC Guinea-Bissau Guinea 0.0 5.0 10.0 15.0 20.0 1 Let at be the weighted average of (log) TFP of a given industry in year t and let ait , be the log TFP of enterprises constituting the industry with respective market shares, sit , where i indexes enterprises. Then at can be written as Nt at  at   (sit  st )(ait  at ) , where letters with upper bars represent unweighted industry means of variables. i 1 35 Aggregate TFP is lower in Zambian manufacturing than in middle income comparators in part because average within-firm TFP is lower in Zambia (figure 2.10). This means that the typical Zambian enterprise operates further away from the world technological frontier than its counterparts in middle income countries because of poorer know-how or poorer work organization and management at the factory floor and beyond. A second explanation is that there is greater allocative inefficiency in Zambian industry compared to industries in middle income comparators as a result of which low productivity firms tend to have a larger market shares in Zambia than they would have in the comparators (figure 2. 11). 36 Figure 2.9: Aggregate TFP, all industry Brazil Thailand Malaysia Chile South Africa03 Morocco China Argentina Mexico Colombia South Africa08 Zambia08 Kenya07 Zambia03 Nigeria 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 Productivity shortfalls due to allocative inefficiency tend to be higher in industries exposed to relatively little competition because of barriers to domestic entry or because of insufficient import penetration. The shortfalls originate in distortions in domestic factor markets arising from problems of business environment and lead to the misallocation of capital, land and labor between sectors and groups of firms.2 The next chapter discusses some of the factors in the current business environment of Zambia that have reduced employment and productivity through the allocative efficiency losses they generate. Here we limit ourselves to describing patterns in the productivity data that are indicative of the kind of misallocation that produces the losses. 2 The distortions lead to a misallocation in the sense that employment and productivity would both be higher in the absence of the distortions. In the case where the misallocation is between smaller and larger firms because smaller firms suffer from capital shortages, society would increase employment and productivity by removing the distortions so as to facilitate the reallocation of capital in favor of small firms to the point where the rate of return on capital would be the same for both small and large firms. 37 Figure 2.10:Average w ithin firm TFP, all industry Argentina South Africa03 Chile South Africa08 Brazil Thailand Malays ia Colom bia Morocco China Mexico Zam bia03 Zam bia08 Kenya07 Nigeria 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 Figure 2.11: Allocative efficiency index, all industry Brazil Malaysia China Thailand Morocco Chile Colombia Mexico South Africa03 South Africa08 Nigeria Kenya07 Zambia08 Argentina Zambia03 0.00 0.20 0.40 0.60 0.80 1.00 Productivity and patterns of misallocation of capital and labor The picture that he 2003 and 2008 Enterprise Survey data provide is that there has not been persistent distortion in the allocation of capital or manpower between sectors or regions in Zambia in recent years. While the 2003 Survey data do indicate that there probably was some misallocation by sector as well as by region, this seems to have been corrected by the time of the 2008 survey, to which correction some of the more recent policy reforms and improvement in Zamia’s business environment might have contributed. On the other hand, there is some evidence of persistent 38 misallocation of capital across the size distribution of business establishments, which is consistent with smaller firms having poorer access to finance or other business environment problems being costlier to them than to larger firms. Inter-industry allocation of capital The conclusion that there has not been persistent inter-industry misallocation of capital within Zambian manufacturing is based on patterns in the set of ordinary least squares regressions reported in table 2.6. The regressions relate to log units of the annual value added per unit of fixed assets of an enterprise to enterprise’s sector of activity and region of location. In this relationship the ratio of value added to fixed assets measures the average revenue productivity of capital. When production technology is Cobb-Douglas, the ratio is also strictly proportional to the marginal revenue productivity of capital, which should not vary between producers when there is no misallocation of capital assuming that neither the riskiness of projects nor the evaluation of the risk varies between producers. On this interpretation, the marginal revenue productivity of capital was significantly higher in some sectors than in others in 2003, according to column 2 of table 2.6. In the table, the coefficients of metal products and chemicals and plastics are positive and statistically significant at the 10% level or less, meaning that the marginal revenue productivity of capital in 2003 was higher in those industries than in the others represented in the table. Since we have no reason to assume that either the level of investment risk or business people’s attitude to it should differ by sector or region, this suggests that there was underinvestment in those industries at the time of the 2003 survey as a result of which aggregate output and productivity must have been smaller than they could have been at the time. This inference is also largely supported by the inter-industry pattern in the average gross profitability of fixed assets shown in column 6 of table 2.9. However, the inter-industry misallocation of capital that these two results imply for 2003 data appears to have been transient since we do not see the rate of return differences in the 2008 data (column 1 of table 2.6).3 The change could be the outcome of new investment having taken place since 2003 in the sectors of higher rates of return as automatic correction to the misallocation detected in the 2003 survey. 3 In the first column of table 2.6, none of the coefficients of the industry groups is statistically significant, which suggests that there is no statistically significant difference in the marginal revenue productivity of capital between industries in the 2008 survey data. There is therefore no overinvestment or underinvestment in any of the industries represented in the table, which is consistent with the result in table 2.9 (column 5) that the average gross profitability of fixed assets does not vary much between sectors. 39 Table 2_6: Average/Marginal Revenue Productivity of Fixed Assets OLS and Random Effects GLS estimates Dependent variable= Log value added per unit book value of fixed assets (1) (2) (3) (4) 2008-sample 2003-sample Pooled Pooled OLS OLS OLS GLS-Random effects Industry group (Base=Food) : Garment and textiles 0.304 0.365 0.220 0.243 (1.23) (1.34) (1.20) (1.16) Metal products 0.363 0.656 0.332 0.325 (1.05) (1.83) (1.30) (1.10) Chemicals and plastics 0.199 1.551 0.785 0.564 (0.60) (4.64)** (3.24)** (1.94) Other manufacures -0.400 -0.136 -0.073 -0.239 (1.63) (0.37) (0.37) (1.05) Retail and wholesale trade 0.231 -0.100 -0.192 (0.23) (0.09) (0.14) Other services 0.355 0.047 -0.158 (1.06) (0.14) (0.32) Location (Base=Lusaka): Kitwe -0.930 1.244 -0.327 -0.596 (3.22)** (2.67)** (1.40) (2.23)* Ndola -0.419 -0.836 -0.678 -0.589 (1.60) (3.50)** (3.76)** (2.77)** Livingston -1.381 0.183 -0.358 -0.617 (4.04)** (0.82) (1.89) (2.58)** 2008-survey -16.334 -15.841 (10.63)** (13.97)** Constant 0.714 0.067 0.452 0.531 (4.45)** (0.30) (2.20)* (2.71)** Year dummies? Yes Yes Yes Yes Observations 288 269 557 557 Enterprises 377 R-squared 0.33 0.14 0.20 Sigma_u 1.26 Rho Absolute value of t-statistics in parentheses * significant at 5% level; ** significant at 1% level Table 2_7: Average Revenue Productivity of Fixed Assets by Sector and Location OLS and Random Effects GLS Estimates Dependent variable= Log value added per unit book value of fixed assets 2008 Survey 2003 Survey Pooled Pooled OLS OLS OLS GLS (1) (2) (3) (4) Business size and age group (Base=young small) Established small business -0.318 0.691 0.137 0.045 (1.32) (2.20)* (0.72) (0.24) Young larger business -0.715 -0.753 -0.758 -0.797 (2.28)* (2.36)* (3.44)** (3.53)** Established larger business -0.215 -0.379 -0.239 -0.006 (0.84) (1.36) (1.35) (0.03) Limitited liability companies 0.276 -0.154 0.125 0.179 (1.29) (0.59) (0.76) (0.96) 2008 Survey -16.394 -15.839 (10.72)** (14.30)** Constant 0.733 0.458 0.622 0.614 (3.16)** (1.21) (2.29)* (2.29)* Year dummies? Yes Yes Yes Yes Industry controls ? Yes Yes Yes Yes Region controls? Yes Yes Yes Yes Observations 288 269 557 557 Enterprises 377 R-squared 0.35 0.23 0.23 Sigma_u 1.28 Rho 0.7 Absolute value of t-statistics in parentheses * significant at 5% level; ** significant at 1% level 40 Table 2_8: Average/Marginal Revenue Productivity of Fixed Assets by Sector and Location OLS and Random Effects GLS Estimates Dependent variable= Log value added per unit book value of fixed assets 2008 Survey 2003 Survey Pooled Pooled OLS OLS OLS GLS (1) (2) (3) (4) Log (Fixed assets/Number of workers) -0.813 -0.653 -0.756 -0.776 (23.63)** (14.70)** (28.47)** (28.37)** Log (Number of workers) 0.062 -0.082 -0.042 -0.045 (0.76) (0.94) (0.71) (0.74) Business size and age group (Base=young small) Established small business 0.066 0.030 0.067 0.102 (0.47) (0.13) (0.55) (0.83) Young larger business -0.362 -0.259 -0.120 -0.180 (1.46) (0.94) (0.66) (0.97) Established larger business -0.056 -0.200 0.008 0.107 (0.23) (0.74) (0.05) (0.62) Limitited liability companies 0.286 0.428 0.444 0.408 (2.27)* (2.22)* (4.23)** (3.56)** 2008 Survey -5.120 -4.168 (4.91)** (4.78)** Constant -3.686 0.804 0.683 0.753 (3.55)** (2.03)* (2.89)** (3.19)** Region controls? Yes Yes Yes Yes Industry controls? Yes Yes Yes Yes Year dummies? Yes Yes Yes Yes Observations 288 269 557 557 Enterprises 377 R-squared 0.79 0.59 0.70 Sigma_u 0.72 Rho 0.58 Absolute value of t-statistics in parentheses * significant at 5% level; ** significant at 1% level Inter-regional allocation of capital The 2008 survey data indicate underinvestment in Lusaka relative to the other three cities in the industries and sectors covered by the survey.4 This can be seen in column 1 of table 2.6, where the coefficients of those other cities are all negative, meaning manufacturing employment and productivity would have been higher if some of the investment in those cities had been made in Lusaka instead. It can also be seen in column 5 of table 2.9, where the average gross rate of return on fixed assets was significantly higher in Lusaka. The situation was quite different at the time of the 2003 survey, when there was evidence of underinvestment in Kitwe and not in Lusaka, and some of the capital that had been invested in Ndola could have been more productively utilized in any of the other cities. Regional misallocation of capital thus appears to have characterized Zambia in recent years according to the Enterprise Survey data. However, the shifting nature of the areas of underinvestment-now in Lusaka but in Kitwe earlier- suggests that, like the rate of return and productivity gap between sectors, the misallocation here is a transient phenomenon reflecting temporary friction than persistent spatial imbalances in manufacturing investment activities. 4 This assumes that there are no regional differences in the riskiness of investment projects, which may not be true. 41 Misallocation of capital between the SMEs and large firms While the Enterprise Survey data provide no evidence of persistent misallocation of capital between industries or regions within Zambia, they do show that, as a rule, smaller and younger firms have always been subject to more capital shortages than larger firms. By young enterprises we mean those that have been in business for less than 10 years. And by small enterprises we meant those employing less than 50 workers. Although total factor productivity does not vary by business size in the 2008 Enterprise Survey data, both the average rate of return to capital and the marginal revenue productivity of capital do. In particular both indicators are significantly higher for young small enterprises, than they are in larger enterprises of the same age group, which suggests that small enterprises operate subject to greater capital shortages than larger ones especially when they are starting out. Significantly, the pattern in the marginal productivity of capital holds up in the 2003 survey data as well, which is what suggests that shortages of capital among smaller firms has been a long term phenomenon. We see in the first column (for 2008) and the second column (for 2003) of table 2.7 that the marginal revenue productivity of capital in young small enterprises is significantly higher than it is in larger enterprises of the same age group. We also see in table 2.9 (column 5) that the average gross rate of return to fixed assets is higher in young small enterprises. Read in conjunction with table 2.7, the first and second columns of table 2.8 show that the higher marginal productivity of capital in smaller firms is despite the fact that there are no significant TFP differences between smaller and larger enterprises. Instead the gap in the marginal productivity of capital reflects the lower capital intensity of production in smaller firms. The persistently higher marginal revenue productivity of capital in smaller firms means that aggregate output and productivity would have been higher if some of the investment in the larger firms had gone to smaller businesses instead. In that sense the fact that smaller firms operate subject to capital shortages is costly to society. It is true that at least a part of the gap in expected rate of return between small enterprises and larger businesses reflects differences in investment risk or in attitude to such risk between the two groups of firms, in which case it does not necessarily represent misallocation of capital. However, a key premise of the next chapter of the report is that part of the difference in investment risk by business size groups does reflect differences in the business environment rather than in inherent capabilities of the two groups of firms. Thus at least part of the marginal productivity gap between the two groups of firms seems to be explained by lenders systematically overestimating the risk involved investment projects of smaller businesses more than their counterparts in developed countries would do because of the lack of a credit information system in Zambia. The next chapter will also argue that part of the gap in the rate of return to capital between small and large enterprises reflects other business environment problems that, either make small business projects riskier, or add to the cost of doing business of small businesses by more than they do that of larger businesses. 42 Table 2_9: Average Profitability of Fixed Assets by Sector and Location OLS and Random Effects GLS Estimates Dependent variable= Log gross profits per unit book value of fixed assets 2003 Survey 2008 Survey Pooled Pooled 2003 Survey 2008 Survey Pooled Pooled OLS OLS OLS GLS OLS OLS OLS GLS (1) (2) (3) (4) (5) (6) (7) (8) Business size and age group (Base=young and small) Established small business -0.115 0.069 -0.013 -0.021 -0.107 0.221 0.035 -0.025 (1.27) (0.71) (0.20) (0.31) (1.16) (1.62) (0.46) (0.32) Young larger business -0.343 -0.220 -0.262 -0.224 -0.370 -0.177 -0.276 -0.293 (2.81)** (2.22)* (3.51)** (2.84)** (3.05)** (1.28) (3.14)** (3.14)** Established larger business -0.106 -0.132 -0.131 -0.098 -0.137 -0.188 -0.138 -0.029 (1.11) (1.60) (2.21)* (1.59) (1.39) (1.55) (1.96) (0.39) Limitited liability companies 0.123 -0.074 0.037 0.101 0.145 -0.119 0.028 0.068 (1.51) (0.82) (0.62) (1.52) (1.75) (1.06) (0.43) (0.93) Location (Base=Lusaka): Kitwe -0.293 -0.117 -0.263 -0.278 (2.70)** (0.51) (2.76)** (2.62)** Ndola -0.138 -0.233 -0.202 -0.183 (1.34) (2.40)* (2.88)** (2.18)* Livingston -0.457 0.084 -0.103 -0.174 (3.60)** (0.84) (1.37) (1.89) Industry group (Base=Food) : Garment and textiles 0.065 -0.014 -0.037 -0.008 (0.66) (0.12) (0.51) (0.09) Metal products 0.081 0.106 0.071 0.071 (0.60) (0.68) (0.70) (0.60) Chemicals and plastics 0.076 0.511 0.242 0.202 (0.59) (3.43)** (2.45)* (1.74) Other manufacures -0.095 -0.196 -0.140 -0.136 (0.99) (1.31) (1.77) (1.49) Retail and wholesale trade 0.000 -0.563 -0.691 -0.663 (1.37) (1.69) (1.26) Other services 0.018 -0.085 -0.034 (0.12) (0.59) (0.17) 2008 Survey 0.751 0.658 -0.037 0.687 (1.29) (1.31) (0.44) (1.52) Constant 0.000 0.835 0.755 0.693 -0.065 0.896 0.877 0.808 (0.00) (6.84)** (8.82)** (8.36)** (0.11) (5.44)** (8.13)** (7.44)** Year dummies? Yes Yes Yes Yes Yes Yes Yes Yes Observations 260 424 684 684 260 249 509 509 Enterpises 375 348 R-squared 0.04 0.03 0.03 0.11 0.16 0.08 Sigma_u 0.41 0.45 Rho 0.51 0.62 Absolute value of t-statistics in parentheses * significant at 5% level; ** significant at 1% level Table 2_10: Average Productivity of Labor by Sector and Location OLS and Random Effects GLS Estimates Dependent variable= Log value added per employee 2008 Survey 2003 Survey Pooled Pooled OLS OLS OLS GLS (1) (2) (3) (4) Business size and age group (Base=young and small) Established small business 0.180 -0.433 0.084 0.194 (1.35) (1.80) (0.75) (1.68) Young larger business -0.026 -0.287 0.021 -0.219 (0.11) (1.14) (0.12) (1.34) Established larger business 0.160 -0.438 -0.007 0.074 (0.86) (2.02)* (0.05) (0.58) Limitited liability companies 0.495 0.646 0.559 0.526 (4.13)** (3.37)** (5.67)** (4.88)** Location (Base=Lusaka): Kitwe 0.177 0.685 0.320 0.251 (1.03) (1.98)* (2.18)* (1.53) Ndola 0.366 0.101 0.234 0.257 (2.11)* (0.55) (1.83) (1.71) Livingston -0.242 0.005 -0.171 -0.229 (1.30) (0.03) (1.33) (1.48) Industry group (Base=Food) : Garment and textiles -0.594 -0.385 -0.568 -0.627 (3.07)** (1.76) (3.98)** (3.76)** Metal products -0.098 -0.130 -0.158 -0.178 (0.39) (0.44) (0.83) (0.80) Chemicals and plastics -0.043 0.090 0.046 -0.058 (0.15) (0.34) (0.23) (0.24) Other manufacures -0.600 0.121 -0.317 -0.540 (2.98)** (0.47) (2.06)* (2.92)** Retail and wholesale trade 0.277 -0.288 0.299 0.255 (1.74) (0.36) (2.21)* (1.77) Other services -0.233 -0.843 -0.434 -0.406 (0.99) (3.20)** (2.46)* (2.01)* Other 0.377 0.452 0.378 (0.62) (0.79) (0.64) 2008 Survey 0.897 1.624 (5.61)** (2.74)** Constant -1.410 0.715 0.516 0.622 (1.50) (2.54)* (2.68)** (3.48)** Year dummies? Yes Yes Yes Yes Observations 593 275 868 868 Enterprises 680 R-squared 0.14 0.18 0.17 Sigma_u 1.12 Rho 0.76 Absolute value of t-statistics in parentheses * significant at 5% level; ** significant at 1% level 43 Allocation of labor Unlike the case with capital, there is no evidence of persistent misallocation of manpower between SMEs and larger businesses. Although the 2003 survey indicated that large firms were typically overstaffed at the time as shown in column 2 of table 2.10, there were no statistically significant differences in the marginal revenue productivity of labor by business size groups in the 2008 survey data (column 1 of table 2.10). The overstaffing observed in larger firms in 2003 thus seems to have disappeared by 2008, either because the phenomenon was one of friction and not a symptom of lasting distortions in the labor market, or because there had been improvement in the market’s functioning between the two surveys. Looking at the allocation of labor across regions and sectors, both the 2003 and 2008 surveys indicate labor shortages in the Copperbelt, on the one hand, and underemployment of labor in the garments and textiles industries, on the other (table 2.10). The underemployment observed in the garments industries could be due to sector specific problems in the functioning of the labor market, but could also reflect sector specific problems of capital shortages that the analysis of the returns to capital in tables 2.7 and 2.8 fails to pick up.56 Summary and conclusion Zambia’s exports are far more diversified today than they were in the mid 1990s, thanks in large part to the many investor friendly policy reforms it has made since then particularly in terms of liberalizing its trade and exchange rate policies. The starting point of this chapter has nonetheless been that the country needs to diversify its export even more. The reason that Zambia’s low wages have not translated into more exports of labor intensive manufactures is that manufacturing labor productivity has been too low compared to that in more successful exporters of those goods within the region and beyond. Part of the reason for this labor productivity shortfall is underinvestment in equipment relative to current manning levels in the manufacturing sector. A second reason is that Zambia’ manufacturing total factor productivity is low relative to the comparators, in part because within-firm TFP is 5 The evidence for this is that, in table 2.10, the marginal productivity of labor is persistently lower in those industries. Although there is substantial dispersion in the marginal productivity of labor between other sectors as well, this seems to be largely frictional to the extent that a sector where the marginal productivity of labor is low in the 2008 survey is not necessarily one for which the same is also true in the 2003 survey. 6 We should note also that there is evidence of labor shortages in Ndola in the 2008 data in the sense that the marginal productivity of labor was significantly higher in that city than it was in the other three at the time of the survey. However, this is a reversal of the pattern read in the 2003 data in which there is similar evidence of shortage in Kitwe. Jointly the two results suggest that, each instance of shortage is temporary. 44 comparatively low, and in part because the allocative efficiency of Zambian manufacturing industries is also relatively low. Productivity shortfalls due to allocative inefficiency tend to be higher in industries exposed to relatively little competition because of barriers to domestic entry or because of too little import penetration. The shortfalls originate in distortions in domestic factor markets including those in labor, credit and land. The most obvious instance of misallocation in the manufacturing and service industries covered by the 2003 and 2008 surveys is that small businesses are subject to significant capital shortages. The survey data also provide some evidence of underemployment of labor in the garments and textiles industries and of labor shortages in the Copperbelt. Chapter 3 describes some of the business environment problems contributing to these instances of misallocation by impeding the flow labor and capital from low return units and activities to high return units and activities. Why are not smaller businesses in Zambia investing more in fixed assets than they currently do since they would seem to put capital to more productive use than alternatives at this point? And what is it that forces overstaffing in labor intensive industries and labor shortages in the Copperbelt? The survey data suggest part of the answer lies in major business-environment differences within the Zambian economy between sectors and regions, but even more importantly, between small players and larger firms, between established businesses and young and prospective ones. 45 Chapter 3 Productivity and Domestic Market Distortions: Key Business Environment Issues Introduction Chapter 2 has argued that one of the main reasons that Zambia’s comparatively low wages have not translated to more exports of labor intensive manufactures is that their effect on unit labor costs has largely been offset by labor productivity shortfalls relative to countries that are more successful than Zambia in exporting those products. The labor productivity gap itself reflects underinvestment in equipment and comparatively low total factor productivity. In this chapter we link these investment and productivity shortfalls to problems in the country’s current business environment. Where the quality of data permits, we quantify the association between specific business environment variables and firm level employment and productivity.7 International comparisons Business environment problems can reduce employment and productivity in one or both of two ways. One of these is by raising the cost of doing business in Zambia relative to other countries. This aspect of the effect of business environment problems is best thought of as an implicit flat tax imposed on activities or transactions of all producers in Zambia that would make them that much 7 This is done with the strong caveat that there are major limitations to the Enterprise Survey data including that the sample is relatively small and has a comparatively small and short panel of just two time periods, which makes them ill suited for the identification of causal effects. 46 less competitive in world markets. Every producer in Zambia would be represented by a single implicit tax rate applying uniformly to everyone as far as this aspect of the effect of business environment on economic activity is concerned. The chapter seeks to provide a sense of how much Zambia could be losing in manufacturing employment and productivity when business environment problems raise the cost of doing business in the country relative to other countries by comparing averages of key business environment between Zambia and the international comparators used in chapter 2. Comparisons within Zambia The second way in which business environment problems cost Zambia in terms of employment and productivity is by distorting domestic markets. In practice a business environment problem does not affect all firms to the same degree. It is more likely that it impacts on some firms more than others for a variety reason. It is not therefore likely to add to the cost of doing business by the same amount on everyone. The proper tax analogy from this point of view is therefore, not one with a flat implicit tax rate, but one with that of a system of discriminatory tax rates that vary across firms, sectors and locations. Just as taxes of this kind would, differences between firms in the cost of doing business generate losses in employment in productivity by preventing factor productivity from equalizing at the margin across activities and producers, that is, by generating allocative inefficiency. In this case the magnitude of the loss involved is determined not by how much the business environment problem in question adds to the cost of doing business on average in Zambia, but by the dispersion in the added cost across firms and sectors within Zambia, the rule being that the greater is the dispersion, the larger the loss in aggregate employment and productivity. The basic point here is that even if Zambia had the best averages of all business environment indicators, it could still in theory be less productive and poorer than comparators if there is too much heterogeneity of business environment across firms within Zambia that leads to greater allocative inefficiency than in the comparators. In order to provide a sense of how far problems in the current business environment in Zambia could be working against the growth of manufacturing employment and exports through the allocative inefficiency they may generate, the chapter analyzes the variation in key business environment indicators on a number of dimensions, including business age, business size, sector of industry, region of location, exporting status and whether or not a business is foreign invested. The gravity of many business environment problems often varies significantly between business age groups in part because entry cohorts often differ in terms of technical know and capability. Established businesses could also respond differently from younger ones to the same problems because they have already incurred some (sunk) costs. Likewise, there could be scale economies in dealing with some business environment problems, which could put larger businesses at an 47 advantage over smaller ones. On the other hand, larger firms may be more exposed to predatory behavior by corrupt officials. Inter-industry technological differences could also translate to differences in how firms are susceptible to or cope with certain business environment problems. Moreover, such key aspects of business environment as physical infrastructure and governance often show significant regional variation. Lastly, it is common for some firms to find themselves in a better business environment than others as the outcome of a deliberate government policy-as beneficiaries of export promotion schemes or of special investment incentive, for example. Zambia’s overall business environment has improved drastically In trying to identify the main business environment influences on employment and productivity, the 2003 and 2008 Enterprise Surveys asked business managers to rate items on a list of 17 potential obstacles to business expansion. The ratings were on a scale of 0 to 4, 0 being no obstacle and 4 being severe obstacle, with 2 for moderate obstacle, and 3 for major obstacle in between. In addition, both surveys collected data from each responding enterprise on hard indicators of the state of most of the 17 potential obstacles. The purpose of collecting data on these ‘hard’ or ‘objective’ indicators was to get some concrete sense of the problems that managers would be “complaining� about in providing the ratings. The hard indicators have the added advantage of being objectively measurable and can be useful in monitoring the outcomes of specific business environment reforms. Figure 3.1: Respondents rating factor as a major/severe obstacle to growth (%)-Enterprise Surveys Access to f inance Tax rates Co mp etitio n f ro m… Electricity Co rrup tio n Crime and thef t Access to land Macro eco no mic instab ility Transp o rtatio n Custo ms and trad e … Skills & ed ucatio n o f … Tax ad ministratio n Teleco mmunicatio ns Busines licensing & … Lab o r reg ulatio ns 0 20 40 60 80 2008 2003 In figure 3.1 we compare the proportion of manages who rated problems in the indicated aspects of the business environment as major or severe obstacles as reported in both the 2008 and the 2003 surveys. The chart shows that the private sector’s view of Zambia’s business environment has become far mover favorable than it was in 2003. Of the fifteen items displayed in the chart, 10 were rated as major or severe obstacles by at least a third of respondents of the 2003 survey. These 48 included problems relating to, in that order of importance, macro-economic instability, high taxes, access to finance, power supply, crime, corruption, competition from informal firms, and skills shortage. It is thus a measure of the success Zambia has had in improving its business environment since then that none of those 10 ‘obstacle’ items were rated as major or severe by at a third or more of respondents during the 2008 survey. Indeed, no other potential business environment problem was rated as such by more than a third of respondents. While much of this discrepancy between ratings from the two surveys has to do with changes in the composition of the sample in terms of size and sector distribution, a sizeable portion of the change is a true time effect that survives comparisons controlling for shifts in sample composition. In addition the change in perceptions is matched by improvements in hard indicators in almost every case. Outstanding issues Figure 3.1 also points to at least four major business environment areas that should continue to be of serious concern to policy makers. These relate, more or less in that order of importance, to: a) problems of access to finance b) issues of taxation c) barriers to entry and to product market competition more generally, and d) problems of provision of physical infrastructure, particularly of power and transport The rest of the chapter is devoted to discussing these and other business environment problems from the points of view of i) the gravity of problems as reflected in managers ratings and also in objective indicators; ii) the magnitude of the adverse influence that the problems are likely to have on employment and productivity; and iii) policy reform measures that need to be taken or are being taken to address the problems In discussing the problems it is important to recognize that they impact on allocative efficiency not only in as far as they influence the mobility of resources among incumbent operators in a given set of industries, but also because they affect the rate of firm formation and entry rates into individual industries. It is also useful to make a distinction between problems that affect firm formation and entry rates directly, and those that do so only indirectly. Lastly, it is worth stressing that the distribution of resources and market shares among domestic produces depends on the degree international integration of domestic industry as much as it does on business environment problems not necessarily related to trade directly. We have therefore grouped business environment problems into three major categories in the discussion of the rest of the chapter: problems relating to the direct regulation of entry and competition, those posing indirect barriers to entry and factor 49 mobility, and factors impeding trade integration. We include in the first of these categories issues relating to competition policy and business regulation via licensing and permit requirements. Problems of access to finance, access to reliable power supply, and taxation form the second category. Table A3_1: Probit Model of Rating a Factor as Major Obstacle to Growth (Maximum Likelhood estimates) Acces to finance, maco instablitly, taxes, power shortage and informal competition Macroecon Tax Informal Access to omical Administra firm's Finance Instability Electricity Tax Rate tion competition Constant -0.61* -0.01 -0.59* -0.20 -0.57 -1.17*** (0.322) (0.358) (0.331) (0.318) (0.362) (0.328) Log (age) 0.08 0.26*** 0.04 0.05 0.07 0.21*** (0.062) (0.072) (0.065) (0.062) (0.072) (0.064) Small 0.37** 0.22 -0.14 0.07 0.22 0.36*** (0.166) (0.181) (0.165) (0.159) (0.189) (0.167) Medium 0.17 0.43** -0.03 0.18 0.27 0.42*** (0.191) (0.205) (0.190) (0.182) (0.213) (0.189) Export -0.19 -0.001 0.21 -0.11 0.11 0.01 (0.164) (0.170) (0.160) (0.156) (0.178) (0.159) Foreign Owned -0.29** 0.10 -0.07 0.07 0.12 0.09 (0.147) (0.160) (0.151) (0.140) (0.163) (0.144) Lusaka 0.81*** -0.19 0.07 0.40** -0.09 0.12 (0.200) (0.222) (0.197) (0.201) (0.222) (0.197) Copperbelt 0.62*** -0.13 -0.11 0.42** -0.12 0.15 (0.219) (0.241) (0.215) (0.216) (0.239) (0.214) Sole propreitorship 0.07 -0.35 -0.20 -0.60*** -0.47* 0.16 (0.185) (0.260) (0.209) (0.213) (0.285) (0.190) 2008 -0.98*** -1.77*** -0.59*** -0.81*** -0.86*** -0.47*** (0.138) (0.149) (0.138) (0.132) (0.155) (0.134) Food -0.2 -0.08 0.43*** -0.15 -0.18 -0.01 (0.165) (0.191) (0.178) (0.162) (0.188) (0.164) Textile and Garments 0.24 -0.22 -0.30 -0.14 -0.38 -0.22 (0.192) (0.240) (0.212) (0.200) (0.248) (0.197) Metal -0.25 -0.09 0.14 -0.05 -0.50** -0.24 (0.213) (0.235) (0.226) (0.206) (0.247) (0.210) Chemical -0.31 -0.15 -0.20 -0.12 -0.43 -0.18 (0.279) (0.312) (0.312) (0.275) (0.314) (0.284) Loglikehood -301.409111 -216.4747 -272.842 -304.7081 -207.8872 -295.3273745 No. of Observations 509 511 511 511 511 511 Table A3_2: Probit Model of Rating a Factor as Major Obstacle to Growth (Maximum Likelhood estimates) Acces to land, corruption, crime, skills shortage and customs and trade regulation Shortage of Customs Access to Skilled and Trade Land Corruption Crime Workers Regulations Constant -1.01*** -0.68** -0.04 -0.41 -0.68* (0.411) (0.337) (0.343) (0.355) (0.366) Log (age) -0.02 0.144** 0.11 0.01 -0.03 (0.076) (0.067) (0.069) (0.069) (0.071) Small -0.17 0.47*** -0.06 -0.13 -0.3 (0.200) (0.177) (0.174) (0.178) (0.183) Medium -0.07 0.42** -0.14 -0.11 0.27 (0.228) (0.197) (0.202) (0.203) (0.198) Export -0.35* 0.28* -0.05 -0.05 0.14 (0.212) (0.160) (0.170) (0.174) (0.172) Foreign Owned 0.20 0.13 -0.01 -0.15 0.24 (0.174) (0.151) (0.159) (0.163) (0.157) Lusaka 0.65*** -0.44** 0.51*** 0.05 0.42* (0.264) (0.199) (0.200) (0.215) (0.239) Copperbelt 0.10 -0.33 -0.83*** 0.10 0.3 (0.298) (0.216) (0.224) (0.232) (0.254) Sole propreitorship 0.31 -0.55** -0.29 -0.46 -0.67** (0.217) (0.264) (0.245) (0.276) (0.341) 2008 -0.47*** -0.84*** 1.05*** -0.95*** -0.8*** (0.167) (0.139) (0.145) (0.150) (0.151) Food -0.24 0.07 0.62*** 0.16 -0.06 (0.196) (0.185) (0.190) (0.197) (0.196) Textile and Garments 0.22 -0.05 -0.02 0.11 0.15 (0.218) (0.228) (0.247) (0.243) (0.237) Metal -0.35 0.42** 0.35 0.16 -0.13 (0.275) (0.220) (0.238) (0.240) (0.245) Chemical -0.47 -0.03 -0.09 -0.09 0.52 (0.357) (0.294) (0.319) (0.316) (0.316) Loglikehood -191.6583 -250.0270536 -233.8367 -220.5425 -211.99156 No. of Observations 509 510 511 511 511 50 Issues in direct regulation of entry and competition Competition and trade policies The fact that low productivity firms tend to have larger market shares in Zambia than they would have in a typical middle income economy suggests that there could be some scope for increasing productivity through competition policy reforms. Zambia has had a competition law since 1994, when it enacted the Competition and Fair Trading Act, primarily as a safeguard against anti- competitive behavior by the large players in domestic markets that emerged from the privatization and FDI deregulation programs of the early 1990s. The Competition and Fair Trading Act mandates the Zambian Competition Commission (ZCC) to regulate mergers and acquisitions. However, it is not clear how far this power has in fact been used by the ZCC to regulate the evolution of the structure of Zambian industry since its establishment, and whether it has the right tools and resources needed for exercising that kind of influence. There are now proposals in the MCTI for expanding the ZCC’s mandate into the pursuit of a more activist competition policy beyond the regulation of mergers (GOZ, 2009). Specifically, it is being proposed that the ZCC expand its scope to the detection, prosecution, and prevention of abuse of dominant market power and cartels, that it be provided with the authority and resources commensurate with the expansion of the mandate, and that a Competition and Consumer Protection Tribunal be established to provide for due process in the implementation of the new policy. These are measures which, if successfully implemented, could generate significant allocative efficiency gains in as far as they could lower entry barriers to domestic industries. However, it is also important to recognize that competition policy is only one of several complementary tools for promoting competition in domestic industry. Competition policy is essentially about influencing the behavior of larger players in specific industries and in the economy as a whole. Such behavior is often a critical determinant of entry and exit rates and of the distribution of market share among incumbents, especially in a country such as Zambia, where, as the MCTI latest policy statement on competition and consumer protection points out, industries tend to be concentrated to a greater degree than in an advanced economy because of the small size of national markets (GOZ, 2009). But the behavior of potential entrants and the constraints conditioning it is also as critical. There are also important influences other than competition policy on the behavior of potential entrants just as there are on the market power and behavior of incumbents. Of these influence by far the most important is foreign trade. Indeed, it is the trade liberalizing measures that Zambia has carried out since 1991 that have influenced the most the competitive pressure under which domestic firms and industries operate. It is clear that the economy can be opened up even more to trade through a range of trade facilitation measures that could significantly reduce trade costs and non tariff barriers 51 to trade. At least as important determinants of entry and exit rates, and hence ultimately of the market structure and productivity of domestic industry, as openness to trade and competition policy, are also a range of determinants of the ability of potential entrants-big and small- to respond to new investment opportunities. We include among these determinants the direct regulation of entry as well as indirect barriers to entry including lack of access to finance and to other basic services such as power supply. Direct regulation of entry Business licensing and the requirement of construction permits are probably the most ubiquitous forms of direct regulation of entry by government in most countries. In Zambia anyone setting up a new business needs to have an investment certificate from the Zambia Investment Center (ZIC). They also need to obtain operating licenses from the local and central government, and have the business registered with the Patent and Company Registration Office (PCRO). Because the costs associated with getting license and legal status are incurred prior to the start of operations, established businesses rightly treat them as sunk costs. They are therefore unlikely to put those costs high on their lists of “business obstacles�, as responses to the 2003 and 2008 enterprise survey showed. Less than one in 10 of the managers of non-micro enterprises rated the requirements as a major or serious obstacle to business operations in the 2008 survey (table 2). This was not much different from the proportion of respondents who thought likewise in the 2003 survey. It was also true of all business size groups and industry groups. Yet, a 2004 World Bank study (World Bank 2004b) showed that the time and pecuniary costs associated with licensing and permit requirements were formidable enough to have substantially reduced rates of new entry and firm formation at the time. The study noted that the basic business licenses alone took five to six weeks to get for Zambians and nine to 16 weeks for foreigners, and made a series of recommendations aimed at shortening the time needed for both groups. Chief among the recommended measures were that the discretion that the ZIC exercised in issuing investment certificates be eliminated, the number of licenses required be reduced, the frequency of renewals be cut, and that the registration process be regionally decentralized by opening regional and local offices of the PCRO. Thanks largely to the implementation of some of these recommendations, Zambia today counts as one of the countries in Sub-Saharan Africa where it is easiest to set up a business. The total time needed to set up the standardized Doing-Business company was estimated at 40 days at about the time the FIAS recommendations were made in 2004. This was the time needed to complete the 6 procedures that the standardized company was expected to go through. According to the 2009 Doing Business report, the number of required procedures is still the same as it was six years ago, but the number of days needed to complete them has dropped to 18. Most of the saving in time occurred in 2008 and 2009, when the time needed for registration was cut from 7 days to 3 and the number of days needed for VAT registration fell from 21 to 10 (World Bank, 2008). 52 These are large improvements, especially given the shortness of the time interval in which they were made. However, they are not good enough compared to more than 60 better performing countries outside of the regions. More importantly, FIAS’s assessment is that Zambia still can and should cut the number of days needed to set up a business from 18 days to 8 by reducing the number of licensing and registration procedures from six to five, by cutting the number of days needed for VAT registration even further and by setting up a one-stop service center for business registration (World Bank 2009b). The need for greater improvement is clearly evident in two related areas that are probably as important determinants of entry rates as the cost of business registration and licensing. These are the time needed to secure permits for business construction projects and the ease of access to land for business premises. Unlike business registration and business licensing, construction permits have been an area in which things have worsened in recent years. The number of days needed to obtain the Doing Business standardized construction permit was first measured in 2006, when it was estimated to be 165. The number of procedures involved then was 16. The number of procedures increased to 17 in 2008 and has remained unchanged since then. The number of days needed to complete all procedures increased to 196 in 2007 and then to 254 in 2009 and has not changed yet since. FIAS is recommending that the number of procedures be cut to 14 in order to bring about significant reduction in the total number days needed to secure permits. The 2004 FIAS report found that obtaining land for business premises was a major problem for any investor in Zambia. The report traced the problem to the fact that very little of available was titled and registered, which made the process of identifying suitable plots exceedingly difficult and time consuming. It also found the land registration and acquisition process to be over centralized but poorly coordinated, and called on the government to improve the situation by establishing, codifying and publicizing procedures and service delivery standards for land administration agencies. The 2004 FIAS report also made specific recommendations for facilitating FDI including some that required amendments to the Immigration and Deportation Act or to the Investment Act with a view to making it easier for foreigners to make obtain work permits and business licenses and make it easier for their businesses to hire and fire workers. Entry regulation, competition and informality Being licensed by and registered with authorities imposes costly norms of technology and transactions on firms in return for benefits in the form of cheaper access to key business services and to publicly provided goods. The decision whether or not to get a license is based on how a particular entrepreneur evaluates the balance between those benefits and the explicit and implicit costs of being licensed and registered. Other things being equal, higher costs of registration, licenses and routine compliance with rules would increase the number of firms operating informally by avoiding both the registration and the licenses. By making it easier and less costly for businesses to 53 register, get a license and keep one, government helps reduce the size of the informal sector thereby helping boost overall economic growth. The government can also make registration and licenses more attractive by increasing its effectiveness in excluding the unregistered and the unlicensed from the services that licenses entitle their holders to. Moreover, the more diverse are those services and the higher they are in quality, the less attractive is it for a firm to operate unregistered and unlicensed. A reasonable indicator of how high the cost of registration and licensing are relative to the benefits they entitle firms to is the proportion of survey respondents who think that being registered and licensed puts them at a competitive disadvantage relative to the unregistered and the unlicensed. Both the 2003 and 2008 surveys provide information on that proportion based on a question which asked respondents to rate the competition that they faced from unregulated businesses as an obstacle to their growth on a scale ranging from 0=no obstacle to 4=severe obstacle. The evidence based on the ratings provided is that the costs of registration and licensing have come down significantly since 2003, that operating informally is a less attractive proposition of Zambian firms today than it was then. For, in the 2008 survey, the proportion of respondents that rated competition from informal firms as a major obstacle to growth was about 26 percent as compared to the 39 percent of respondents who felt likewise in 2003. Although some of this change reflects a shift in sample composition in terms of distribution by size and industry, we can see from column 3 of table A3_1 that a significant share of the drop in the complaint rate remains after accounting for those changes. Despite the drop in the complaint rate between the surveys, the fact that more than one in four of business covered by the 2008 survey felt that they were held back by competition from the informal sector means that there is some room for raising business formation and business formalization rates by reducing further the cost of registration and licensing. As is to be expected, high licensing and registration costs deter the formalization of smaller enterprises more than they do that of larger ones. For, as the second panel of table 3.3 suggests, the cost of being formal is higher for smaller businesses than for larger ones for any given business age group, and higher for older established business more than relatively new entrants younger ones for any size group. For example, in the 2008 survey, the complaint rate about competition from informal firms was 33 percent for established small firms as compared to 28 percent for large businesses of the same age group, 19 percent for young small businesses, and o 12 percent for young large businesses. 54 Indirect barriers to entry and factor mobility Problems of access to finance A major source of distortion that Zambia’s economy shares with others is that smaller and start up business do not have as much access to external finance as larger and longer established ones. One aspect of the distortion pertains to the allocation of resources and market shares between operating small firms and existing larger businesses. A second aspect is that that business entry rates are lower than what they would be if smaller and younger businesses had as good access to finance as larger and longer established businesses. An indication of the relative importance of this distortion is that inadequate access to finance was the factor that the largest percentage of businesses in the 2008 Zambia Enterprise Survey rated as a major obstacle to their growth from the list shown in figure 3.1. We use the term “inadequate access to finance� to designate situations in which businesses find interest rates to be too high as well as those in which lenders would not extend credit under any terms. Understood in this broad sense, access to finance has always been a major business environment issue in Zambia, and is one of the consequences of a long history of high inflation and currency fluctuation. As the 2004 assessment argued, the key access problem at the time of 2003 was indeed that these factors of macroeconomic instability pushed real lending rates to prohibitive levels, the average Kwacha loan rate standing as high as 50 per cent at one point. Perhaps as importantly, the uncertainty that large and growing budget deficits and the external debt problem created made bankers very reluctant to extend long term loans to anyone. 55 Table 3.1: Repondents rating factor as major or sever business obstacle (%)- Enterprise Survey 2008, Full Sample Full sample SMEs and Large Micro Manufacturing Retail Other industry/ Enterprises Services Number=603 Number=304 N=122 Number=58 Number =119 Acccess to finance 27.7 27.2 15.5 30.3 41.2 Competion from informal firms 26.4 21.8 24.7 26.5 37.8 Tax Rates 25.5 25.9 23.2 25.3 27.7 Court System 17.8 18.8 12.5 26.3 13.0 Access to Land 15.8 12.4 15.5 14.1 24.4 Corruption 14.6 15.2 9.2 7.1 26.1 Electricity 13.8 20.2 5.6 11.1 12.6 Crime 12.9 12.8 12.0 6.1 20.2 Macro Instability 12.3 11.9 10.6 11.1 16.0 Transport 10.3 10.7 9.2 10.1 10.9 Tax Administration 10.1 7.0 9.9 8.1 18.5 Customs and Trade Reg. 9.6 9.5 10.6 8.1 10.1 Skills shortage 6.6 9.1 5.6 7.1 2.5 Licensing and permis 6.5 4.1 5.6 7.1 11.8 Telecommunications 3.7 4.9 0.7 4.0 4.2 Labor Regulation 3.7 3.3 4.9 5.1 1.7 Political Instability 1.5 0.8 2.1 2.0 1.7 Table 3.2: Repondents rating factor as major or sever business obstacle (%)- Enterprise Survey 2008 Manufacturing Enterprises only Foreign Not-foreign Small Large Exporter Non-exporter Invested invested Access to finance 31.8% 23.3% 16.7% 30.9% 18.2% 31.9% Access to land 12.4% 12.6% 9.5% 13.0% 10.6% 13.0% Competition from informal firms 23.4% 24.3% 21.4% 24.0% 27.3% 22.7% Licensing & permits 7.0% 1.0% 2.4% 5.3% 4.5% 5.0% Corruption 12.4% 17.5% 31.0% 11.5% 15.2% 13.9% Crime 11.4% 12.6% 19.0% 10.7% 10.6% 12.2% Customs and trade reg 6.5% 15.5% 19.0% 8.0% 19.7% 6.7% Electricity 16.9% 19.4% 21.4% 17.2% 15.2% 18.5% Labor regulations 2.5% 6.8% 9.5% 3.1% 1.5% 4.6% Macroeconomic instability 8.0% 19.4% 26.2% 9.5% 16.7% 10.5% Skills shortage 7.0% 11.7% 14.3% 7.6% 9.1% 8.4% Tax admin 8.5% 7.8% 4.8% 8.8% 12.1% 7.1% Tax rates 23.9% 31.1% 28.6% 26.0% 37.9% 23.1% Telecom 5.5% 3.9% 4.8% 5.0% 0.0% 6.3% Transport 12.4% 7.8% 14.3% 10.3% 9.1% 11.3% 56 Table 3.3: Percent of respondents rating problems as a major obstacle to growth Size group: Access to finance and to infrastracture: Unreliable power Problems of Poor access Skills shortage supply transport to finance 2003 2008 2003 2008 2003 2008 2003 2008 Young and small 30% 14% 20% 10% 62% 26% 38% 7% Established and small 34% 11% 43% 12% 69% 25% 40% 6% Young and large 39% 24% 39% 0% 37% 14% 25% 7% Established and large 46% 14% 25% 12% 51% 22% 38% 12% Total 39% 14% 30% 10% 54% 24% 36% 8% Taxes, macro economic stability, and regulation: High taxes Macro economic Competition from Problems of tax instability the informal sector admin 2003 2008 2003 2008 2003 2008 2003 2008 Young and small 63% 25% 78% 6% 28% 19% 28% 9% Established and small 66% 22% 89% 13% 17% 33% 29% 9% Young and large 61% 29% 61% 12% 33% 12% 31% 7% Established and large 50% 29% 75% 22% 25% 28% 27% 5% Total 57% 25% 75% 11% 26% 24% 28% 8% Labor regulation, corruption and crime: Labor regulation Corruption Crime 2003 2008 2003 2008 2003 2008 Young and small 8% 3% 53% 9% 53% 13% Established and small 20% 5% 54% 15% 46% 10% Young and large 14% 5% 33% 17% 42% 10% Established and large 21% 5% 45% 13% 55% 9% Total 17% 4% 46% 12% 50% 11% 57 Table 3.4 Selected business evironment indicators by industry and size groups Power supply, crime, bribes, all enterprises, micro included Output loss to outages Cost of crime or Bribes paid as % (% of sales) of security of sales (% of sales) 2003 2008 2003 2008 2003 2008 Size groups: Young and small 5.1 5.6 9.0 5.9 2.5 1.6 Established smal 6.2 3.6 7.8 5.4 2.5 0.9 Young and large 3.3 6.1 0.8 4.5 1.1 0.5 Established and large 4.1 3.1 1.7 4.9 0.6 0.4 Total 4.5 4.6 5.1 5.4 1.5 1.1 Industry groups: Food 4.7 6.2 7.3 0.9 0.9 Textile and garm 2.7 4.8 . 3.6 2.8 1.8 Machinery and me 0.9 4.0 . 0.7 2.0 1.0 Chemicals and pl 6.4 3.0 . 3.0 2.0 0.4 Other manufacure 3.4 5.7 2.0 3.8 1.2 2.5 Retail and whole 0.0 3.8 3.7 6.5 0.0 0.9 Other services 4.4 1.6 . 1.5 0.9 0.2 Other 5.3 5.4 . 10.0 1.9 0.3 Table 3.5 Selected business evironment indicators by industry and size groups Regulation and Transport , all enterprises, micro included Management's time Unionized workforce Inventory left of intermediate for dealing with officials (%) inputs on new delivery (%) (number of days) 2003 2008 2003 2008 2003 2008 Size groups: Young and small 14.6 5.5 22.5 5.8 41.3 16.1 Established small 13.7 5.3 29.7 17.8 52.4 23.6 Young and large 12.3 8.3 41.3 26.0 36.9 17.1 Established and large 14.0 7.0 57.1 39.5 46.0 31.0 Total 13.8 5.9 42.6 16.2 44.5 20.5 Industry groups Food 11.6 5.7 46.8 10.0 38.3 18.3 Textile and garm 21.5 4.3 56.3 17.6 49.6 12.9 Machinery and me 15.6 10.4 38.0 33.1 31.9 34.2 Chemicals and pl 14.1 6.3 25.7 27.4 19.2 20.6 Other manufacure 12.9 7.0 37.1 23.8 99.2 22.0 Retail and whole 15.0 5.1 0.0 8.3 90.0 22.6 Other services 21.8 7.0 18.9 21.7 27.8 9.2 Other 12.6 3.6 43.0 37.4 42.7 . 58 Improvement in small business access to finance One of the main finding of the 2008 Enterprise Survey was that firms’ access to credit had improved a great deal compared to what it was in 2003. Some 27pecent of respondents to the 2008 Enterprise Survey rated inadequate access to finance as a major obstacle to business growth. This compares with the more than 60 percent of respondents to the 2003 survey who gave the same rating to the same problem. While some of the drop in the complaint against inadequate access to finance reflected changes in the composition of the sample in terms of business size, age, industry, and location, it is also clear from the first column of table A3.1 that much of it holds up when we control for those attributes. The change is also backed up by improvement in hard indicators of access between the two surveys. Many of these indicators improved sharply between the surveys as key indicators of macroeconomic stability improved in the wake of a debt relief program and a copper price boom. Thus as the inflation rate fell from 21 percent in 2003 to under 10 percent in 2008, the average nominal interest rate dropped from 28 percent to 19.3 percent while the average value of collateral to loan ratio fell from 324 percent to 141 percent. As a consequence, the percentage of small businesses that had active bank loans rose from 19.5 percent to 28 percent. Large as these improvements are, the values of indicators for 2008 themselves do not compare well with those in other comparatively successful African economies. In particular, nominal and real interest rates are quite high in Zambia by the region’s standards while the share of bank financing in firm level fixed investment and working capital in Zambia is one of the lowest in the region. The value of the collateral to loan value ratio is higher in Zambia in many other countries in the regions while the proportion of small businesses that have active bank loans and the share of SMEs that have active bank credit lines are both lower in Zambia. The problems of access to finance that these comparisons point to generate losses of employment and productivity at two levels. At one level, the fact that Zambian firms do not have as good access to finance as their counterparts, say in Kenya or Thailand, because they face higher finance charges and higher collateral requirements makes them invest less and employ less than their counterparts in those countries. In turn, the lower investment and hiring rates often make Zambian firms less productive than those counterparts in as far they mean that Zambian firms use inferior technology or fail to exploit economies of scale that their counterpart do. 59 Access to finance, employment and productivity Tables 3.8, 3.9 and 3.10 describe the correlation between access to finance, on the one hand, and enterprise level employment and productivity, on the other, while controlling for differences in other business environment variables and key business characteristics. The aim is to outline the mechanics by which differences in access between firms can translate to employment and productivity outcomes. Collectively the tables show that firms that have poorer access to finance have higher marginal revenue productivity of capital among those firms in one or more of three ways. There are two possible reasons for this. One of these is that inadequate access to finance forces firms to be less capital intensive than they would otherwise be. A second is that inadequate access to finance forces firms forego economies of scale they would have exploited with better access, that is, forces them to reduce investment and hiring rates at a given ratio of fixed assets to employment. A reading of tables 3.8, 3.9 and 3.10 along these lines suggests that both of these firm level outcomes of inadequate access to finance have been at work in Zambia. Specifically the tables show that businesses that have better access to trade credit and to bank credit tend to have higher labor productivity, not necessarily because they are inherently more productive, but often because they have access to better technology as a consequence of their better access to finance. This can be seen from the fact that the marginal revenue productivity of capital and the average rate of return on fixed assets are both consistently lower in enterprises that have better access to finance only because the value of equipment per worker is higher for such enterprises (tables 3.10 columns 2 and 3). Access to finance and allocative efficiency The effects on employment and productivity of the average Zambian firm having poorer access to finance than international comparators are reinforced by the allocative efficiency losses arising from the fact that the ease of access to finance also varies between firms within Zambia, notably between SMEs and large firms, and between start-ups and established businesses. Those allocative efficiency losses add to Zambia’s international productivity and employment shortfalls if the variation in access within Zambia is greater than that between firms within international comparators. For example, a higher cost of borrowing for the average Zambian firm than for the average firm in another country, would reduces employment and productivity in Zambia relative to that other country, Zambia’s employment and productivity shortfalls vis-à-vis the other country being even larger if, in addition, 60 the dispersion in the cost of borrowing also happens to be larger in Zambia.8 Later in the report, we will compare a wider range of hard indicators of access to finance between Zambia and other countries from the point of view of the average firm. Unfortunately we are unable to compare the dispersion of access indicators in Zambia with those of comparators. We are therefore unable to say whether and how much Zambia’s international productivity shortfalls could be attributed to the variability of access being high in Zambian by international standards. Still we can get some sense of the potential importance of allocative efficiency losses as a source of Zambia’s overall productivity shortfalls by looking at the variation in access indicators across firms within Zambia alone. We can also get a sense of how far these losses are likely to have changed between the two survey years by looking at changes in the dispersion of indicators between the surveys. Looking at the 2008 survey data, we see in tables 3.1 and 3.2 that, on average, manufacturers were more likely to report access to finance as a major or severe obstacle (27 percent) than those in retail trade (16 percent). There were also similar gaps, first, between exporters (17 percent) and non- exporters (31 percent) and, secondly, between enterprises that were foreign invested (18 percent) and those that were not (32 percent). However, these gaps are probably less significant differences of access between small enterprises and larger ones. As pointed out earlier, although there was no significant TFP difference between these two groups of enterprises, it was also clear that smaller enterprises underinvested in equipment relative to larger ones, the indication of this being that the marginal revenue productivity of capital was significantly higher in smaller enterprises than larger ones, as indeed was the average profitability of fixed assets. Table 3.1 indicates that at least one explanation for the underinvestment by small enterprises is their poorer access to finance relative to large firms. Thus that the proportion of those rating problems of access to finance as a major or severe obstacle to business growth in the 2008 survey data was significantly higher for smaller firms (32 percent) than larger one (23 percent) –table 3.2-and for micro enterprises (41 percent) than larger businesses-table 3.1. Complaint rates were also significantly higher for younger businesses (table 3.3). Table 3.3 also suggests that the difference between large firms and small firms in terms of the rate of complaint about access to finance has not changed much between the two surveys, although the difference between established large firms and young large firms has come down drastically. 8 Greater dispersion in the cost of borrowing would add to the loss in employment and productivity in Zambia by making the dispersion in the marginal productivity of capital and labor in Zambia to be higher than it would be in the other country. 61 What needs to be done Access to finance and macroeconomic stability In Zambia the problems of access to finance is inseparable from that of macroeconomic stability. For the extremely high cost of borrowing and tight credit that Zambian business faced at the time of the 2004 assessment was an immediate consequence of the high rates of inflation and currency volatility of the time. Similarly, the fall in real interest rates and the expansions in business lending that was captured in the 2008 survey came about thanks to a sharp drop in the rate of inflation to a single digit and the stabilization of the Kwacha that a large debt relief program and a copper price boom brought about by helping drastically reduce government borrowing. Things had improved so much by the time of the 2008 survey that less than 15 percent of survey respondents thought of macroeconomic instability to be a major obstacle to business growth as compared to the 80 percent of respondents to the 2003 survey who thought likewise. However, as the consequences of the global recession showed, and as indeed should be apparent from the role that arguably fortuitous factors played in the comparative stability of the 2005 to 2008 period, Zambia is far from having mastered the forces of macroeconomic instability inherent in its economic structure. Until it does the country would continue to see-saw between the threat of Dutch disease in good times and of near-credit crunch scenarios of bad ones.9 Achieving lasting price and exchange stability and bringing government borrowing under control over the long term is therefore probably the most important step Zambia can take to improve businesses’ access to finance. Access to finance and credit information 9 Even at the time of the 2008 survey macroeconomic instability was of serious concern to larger businesses, some 20% of which rated it as a major or severe obstacle to their growth, the preoccupation this time being more with the appreciation of the Kwacha than anything else (tables 3.2 and 3.3). Apart from its influence on the cost of borrowing macroeconomic instability can harm employment and growth in exporting industries when it generates exchange rate volatility or appreciation, which impacts investment decisions directly to the extent these increase the uncertainly of returns to investment. Some evidence of this type of influence of macroeconomic instability can be seen in tables 3.8, and 3.9 and 3.10. In the tables, we capture that influence by including a dummy variable for businesses which have rated macroeconomic instability as a major or severe obstacle. A joint reading of the coefficients of the three variables as described suggests that macro economic instability reduces total factor productivity, not only through its influence on access to finance, but also by generating excess capacity more directly. The tables also suggest that the same excess capacity may in turn have reduced manufacturing employment, not only by discouraging further investment in fixed assets, but also in as far as it means lower manning levels of equipment. 62 SMEs and microenterprises complain so much of lack of access to finance because most commercial banks do not extend loans to such businesses as a matter of policy. Micro finance institutions that meet a significant part of the credit needs of microenterprises in other African countries and low income economies more generally are nowhere near as common in Zambia, and currently serve at most 50,000 clients (Martinez, 2006). And although there are non-bank financial institutions designed to lend to SMEs, experts believe that the loan products that they offer are not attractive enough to such businesses, either being of too short a cycle to provide investment finance or carrying prohibitively high interest charges. There is thus an urgent need in Zambia for the development of micro finance institutions, on the one hand, and, on the other, for providing the commercial banking sector with the infrastructure it needs to participate in the SME credit market. A missing key component of this infrastructure is a workable credit information system. For one of the reasons why interest charges are often too high at present when Banks do lend to SMEs occasionally is the extremely high risk premiums that they attach to the return on lending to small borrowers. A large part of this premium reflects Banks’ lack of reliable information on this particular group of borrowers, which is a problem that well developed credit information systems have helped mitigate in advanced economies. Zambia’s credit information system is rudimentary at present, the first credit reference bureau, Credit Reference Bureau Africa Limited (CRBAL), having opened only in 2007. A recent FIAS assessment ( World Bank 2009) stresses that Zambia has a long way to go before it can build on this a reliable credit system that will have measureable influence on bank lending to SMEs. At the moment the CRB covers a tiny fraction of potential borrowers, and more importantly, collects data only from Banks and other financial institutions, and has yet to expand its sources to retailers, trade creditors and utility companies in order to capture a larger share of the population of potential borrows, and generate more reliable information on those already in its database. Taxation Yet another source of distortion in Zambian industry is that the marginal effective tax rate is higher for SMEs than it is for larger businesses. Just as the disadvantage of SMEs in terms of access to finance does, this distortion protects the market shares of larger firms among incumbents while reducing rates of firm formation and entry. As we say this, we should stress that Zambia is not a particularly high tax economy by international standards. For example, the Doing Business database shows that the ‘standard’ Zambian firm pays a much smaller percent of its profits in total taxes than its counterparts, not only in middle income countries, but also in low income comparators such as Kenya, Ghana, Malawi and Tanzania (figure 3.2). Similarly, although labor contributions are comparatively high in Zambia –higher than South Africa, Kenya, and Thailand, they are also less than those in Tanzania, Ghana, and most middle income comparators shown in figure 3.3. 63 Indeed, the average business tax burden may have fallen significantly since the 2004 assessment, of which one indication is that a far smaller proportion of business managers complained of high taxes in the 2008 survey than did in the 2003 survey. Specifically, about 26 percent of respondents to the 2008 survey thought high taxes were a major obstacle to the growth their businesses as compared to more than 60 percent of respondents who thought likewise in the 2003 (figure 3.1). As can be seen from table A3.2 (column 3), the change in the complaint rate applies to all groups of firms. Nonetheless, the proportion of business which felt held back by high taxes in 2008 was quite high in absolute terms and nearly as high as the proportion of those who complained inadequate access to external finance as a growth obstacle. Marginal effective tax rates are higher for the non-corporate sectors The reason that so many businesses are complaining of high taxes despite Zambia’s standard tax rates not being high relative to other countries is that survey respondents’ ratings are not based on standard tax rates. The ratings are more likely to reflect mangers’ assessment of the effective marginal tax rate, that is, the amount by which the sum total of direct taxes and other fiscal obligations net of subsidies and other incentives increases for every dollar of fixed investment made by the business. Stern and Barbour (2005) calculate that the effective marginal tax rate in Zambia falls in the 20-25 range for small businesses. This is very high compared to the 5 to 10 percent range they calculate for other sectors of the economy. Zambia’s effective marginal tax rate so calculated would be significantly lower than South Africa’s (about 32 percent) and of Rwanda’s (30-50 percent), but most Zambian small businesses would probably compare their tax burden with larger businesses within Zambia in responding to the survey question, and are quite unlikely to benchmark rates internationally.10 Just like the differences in tax burdens between the corporate and non-corporate sectors, differences in tax burdens that may exist among various groups of firms reduce productivity and employment due to the allocative inefficiency they generate. In tables 3.8 to 3.9 we describe the direct correlation between business tax burden and firm level job creation rates and productivity by using as a proxy for tax burden a dummy for whether or not a business rated high taxes as a major or severe obstacle to growth. The conclusion we draw from a joint reading of the three tables is that those complaining of high taxes are employing fewer workers than others either because they invest less or employ more capital intensive technology.11 10 We should stress here that the small business category used in Stern and Barbour (2005) includes many businesses that would count as “large� according to the classification used in tables 3.1 through to 3.4. This is important when we try to interpret the differences between business groups in terms of ratings of high taxes as an obstacle to growth. 11 Although there were no significant differences between sectors of activity in complaint rates against high taxes in the 2008 survey (table 3.1), larger enterprises (by our classification) were more likely to rate high taxes as a major or severe obstacle to growth than smaller ones within the manufacturing sector (table 3.2). Foreign invested manufacturers were 64 Figure 3.2 Figure 3.3 Power shortages Although the situation has improved drastically in recent years, Zambian industry has suffered from chronic power shortages for quite a while now. The shortages have meant frequent outages and long queues to get connected to the public grid for start ups and expanding businesses. Like many other business environment problems, the shortages have added significantly to the cost of doing business in Zambia relative to other countries. They have also been a source of significant allocative also more likely to rate high taxes likewise than others (table 3.2), although the difference between exporters and non- exporters was not very large in this regard. 65 inefficiency in the economy. Again, the inefficiency occur not only because the shortages affects smaller and younger firms than the larger and the more established among businesses already in operation, but also because they are likely to reduce firm formation and entry rates. One indication that the shortages have lessened recently is that the average time taken to get connected to the public grid dropped from 120 days in the 2003 survey to about 80 in the 2008 survey (figure 3.5). Although the frequency of outages did not change much, average revenue losses due to outages also fell from 4.5 percent a year to 3.6 percent (figure 3.4). As a result the share of managers who considered power shortages to be a serious obstacle to business growth dropped from 40 percent to 20 percent (figure 3.1). Moreover, the complaint rate fell for all business size groups, industries and locations (tableA3_1). Still, an 80-day waiting time to get an electrical connection indicates a serious supply shortfall as does an average revenue loss of 3.6 percent due to outages. Figure 3.4 The analysis reported in tables 3.8 to 3.10 suggests that the shortages reduce employment by reducing the rate of investment overall and by making production more capital intensive within each industry. The second of these effects occurs when firms are forced to switch funds from their wage bill to capital expenditure on backup generators and other equipment related to mitigating the effects of the shortages. Over the long term power shortages can reduce the rate of fixed investment across the board because of the high degree of complementarity between capital and energy as inputs. 66 Outages and long queues to getting connected There were nearly 36 outages in the year leading up to the 2008 survey, which works at an outage every 10 days. Outages could lead to loss of sales by forcing downtime. They could also cause wastage of material in-process at the point of the power interruption and damage equipment thereby adding to routine maintenance costs. These costs should not be confused with the investment costs that frequent outages may also lead to in the form of outlays on backup generators, and surge protection devices, or with the additional investment costs of substituting existing equipment with alternatives less susceptible to damage due to power fluctuations. The sharp drop in managers’ concern with power shortages between the two surveys seems to have more to do with steep declines in such investment costs and in the waiting time to getting connected to the public grid, than it had with the relatively small drop in recurrent losses to outages. This conclusion is supported by the fact that the proportion of businesses running backup generators dropped from 38 percent in 2003 to 11percent in 2008 (table 3.3). Figure 3.5: Power Power shortages as a source of allocative inefficiency To the extent that outages are more frequent and waiting queues to getting connected are longer in Zambia than those in comparators (figure 3.4), they make doing business in Zambia more costly than in the other countries, which in turn reduces Zambia’s aggregate employment and productivity. 67 To these losses of employment and productivity should be added those stemming from the allocative inefficiency caused by the fact that outages are more frequent in some sectors than in others and for some firms than others, while longer queues for getting connected are likely to mean lower firm formation and entry rates. It is important that the pattern of variation in these shortage indicators did not change between the two surveys. In both surveys, the most striking contrast in revenue losses due to outages was that between manufacturing businesses and those in retail and other services. As should be expected those losses were significantly higher in manufacturing businesses than in services. Within manufacturing, smaller businesses reported higher losses on average than larger ones for each business age group. Similarly, the waiting period for getting connected to the public grid was several times longer for small businesses than for large ones(table 3.6). Table 3.6: Indicators of provision of infrastructure Infrastructure Indicators 2008 Other Full sample small Large Lusaka cities Frequency of power outages (times last yr) 36.28 37.01 34.83 42.77 25.76 % of production lost due to power outages 3.46 3.61 3.16 4.57 1.66 Have own generator (%) 11% 8% 17% 12% 9% % of production lost in shipment 1.93 1.50 3.70 2.64 1.61 No. of days to obtain a telephone connection 19.86 19.62 20.00 20.55 17.00 No. of days to obtain a electricity connection 79.56 164.17 52.84 98.88 45.22 Infrastructure Indicators 2003 Other Full sample small Large Lusaka cities Frequency of power outages (times last yr) 37.2 29.8 44.9 37.8 36.7 % of production lost due to power outages 4.5 5.6 4.3 5.2 3.9 Have own generator (%) 38.2 27.3 60.6 33.3 42.3 % of production lost in shipment 2.68 4.7 2.3 – – No. of days to obtain a telephone connection 132.5 135 21.7 183.6 13.3 No. of days to obtain a electricity connection 120.7 47 162.8 90 123.8 These patterns in hard indicators are also reflected in those in rates of complaint about the shortages (table 3.1). Thus in the 2008 survey only about 6 percent of retail businesses rated power shortages as a serious obstacle to growth (as opposed to 20 percent in manufacturing). Although the rate did not differ much between small enterprises (17percent) and larger ones (19 percent) within manufacturing (table 3.2), complaint rates become higher for larger businesses when we control for 68 the age of businesses (table 3.3). In particular, relatively young and large enterprises –meaning those that have been in business for less than 10 years but employing 50 workers or more-were more likely to rate power shortages as a major or severe obstacle to business in the 2008 (24percent) than small enterprises of the same age group (14 percent). A similar pattern is also seen in the 2003 data, in which the complaint rate among relatively young large businesses was about 30 percent as compared to about 39 percent for large businesses of the same age group. What needs to be done The root cause of power shortages in Zambia is that there has not been any major addition to generating capacity since 1970, while demand for electricity is estimated to have increased by 4 percent a year since then with notable acceleration since the 1990s.12 The ultimate solution to the shortages will therefore require large investments in generating and transmission capacity. Indeed, the government has a five year plan for such investments. It has also undertaken a series of institutional reforms since the mid 1990s in order to make the plan viable. The reforms include the legalization of private investment in the power sector by the Electricity Act of 1995, the establishment of the Energy Regulation Board in 1995, and the 1999 government policy framework for encouraging private sector participation in power generation and transmission development. In the short-term, the government should increase the prospects of success for its long term investment plans even more by helping institute a more rational pricing system for the sector. At the moment the electricity tariff in does not cover the full cost of supply. For example, a recent ERB commissioned study indicates that the tariff is on average 45 percent below the cost of service. Underpricing has direct implications to the prospects of long term investments in additional capacity, not only because it is undermining the financial viability of ZESCO, but also because no other new or potential player would be willing to invest in new generation capacity at the current tariff rates. Here also there are already important initiatives with the government taking steps to revise tariffs. However, low tariffs are not the only factor behind ZESCO’s weak financial position. Other perhaps equally important issues that need to be addressed in order to improve ZESCOs performance include excessively high transmission and distribution losses, the build-up of payment arrears by ZESCO’s public sector customers and financial losses arising from operating inefficiencies one would not expect in a commercial enterprise. 12 As a result the country has been experiencing load shedding since the early 2000. For example, in 2005, the total installed capacity was 1,732 MW, mostly hydro generation, and the peak demand was 1,330 MW, while total generation was 8884 GWh. 69 Trade logistics and trade facilitation To date the most important competition enhancing policy developments in Zambia have been the trade liberalization measures the country has implemented since 1991. While there is no formal study of the impact of those measures on the structure of domestic industry in Zambia, recent studies based on data from developing and OECD economies show that the kind of trade policy reforms that took place in Zambia help lower domestic prices and mark up rates not least of all by influencing the behavior of large players in the domestic economy. There is also solid evidence that increased openness to trade leads to large productivity gains in two complementary ways. On the one hand, it generates allocative efficiency gains by inducing a reallocation of market shares from low productivity firms to high productivity ones. It also raises average within-firm total factor productivity by providing firms with greater incentives for innovation. To these forms of trade induced productivity growth should be added a third one, which is the economies of scale that export markets often help domestic firms realize. It is quite likely that Zambia has benefited from one or more of the three forms of productivity growth over the last decade, especially in light of the fact that data from the Enterprise Surveys suggest that aggregate manufacturing TFP went up between 2003 and 2008 as a result of allocative efficiency gains (figures 2.9 and 2.11). All indications are also that Zambia can realize similar gains in the future to the extent that there is room for opening up its economy even more by improving its customs administration in the short term and by reducing transport costs in the long term. Trade Facilitation The Doing Business standard cargo takes 53 days and 64 days respectively to export from and import to Zambia by ocean transportation at a cost of USD 2664 for exports and USD 3335 for imports. This puts Zambia’s 2009 global rank for the ease of trading across borders at 157 in a field of 183 countries, and at a clear disadvantage with most of the comparators used in this report (figure 3.6). Although adverse geography has a lot to do with why it is more difficult to trade with Zambia than with so many other countries, there is also scope for opening up Zambia’s economy further to trade through trade facilitating reforms given that more than two third of the time needed to export from Zambia and about half of what is needed to import to it is taken up by document preparation and document handling (World Bank 2009b). FIAS has therefore been recommending reducing the number of customs documents and streamlining their handling as major trade 70 facilitation measures (World Bank 2009a).13 One of the specific recommendations for streamlining document handling is to extend the Customs Accredited Client Program to SMEs. This is a scheme whereby businesses go through a comprehensive accreditation process the completion of which would allow them to bypass most inspections at entry and exit points. At the moment only the largest exporters and importers are eligible for the program. Other measures that FIAS has recommended for streamlining customs clearance include the use of inland clearance facilities, better clarification and codification of customs valuation rules, and improved staff training. Transport costs Although transport accounts for the largest share of trading costs in Zambia, it is also an area where it is not necessarily obvious that there is a great deal of room for cost saving. This is not only because Zambia’s geography can hardly be more adverse from the point of view international trade.14 It is also because, as Raballand et al. (2008) point out, Zambia has among the lowest transport costs for landlocked countries in Sub-Saharan Africa. Raballand et al. (2008) attribute Zambia’s comparatively low transport costs to two factors. One of these is the investment that Zambia made over the past decade in improving road conditions, which is of crucial importance given that more than two-third of the volume Zambia’s trade relies on road transport and is with neighboring countries. The second is that Zambia has succeeded in cutting freight tariffs by making its transport sector highly competitive through deregulation measures that opened it to up to foreign competition and to foreign direct investment. As a result of those measures the share of Zambian operators in the domestic market has dropped to no more 40 percent, which share, however, those operators are maintaining at competitive tariff rates matching those of much larger foreign competitors without any form of direct or indirect support from the government. At the same time, Raballand et al. (2008) point out that Zambia can reduce freight transport costs significantly further by lowering fuel costs and by reducing delays involved in “border-post operations�. Costs would be even lower if it had not been for adverse spillovers of South Africa’s banning of imports of second hand trucks to protect its motor industry against foreign competition. 13 World Bank (2004) made 15 recommendations for streamline importing and exporting procedures some of which required amendment to the Customs and Excise Act. 14 Current ports serving Zambia are Dar es Salaam (1970 Km away from Lusaka), Durban (3000km from Lusaka and Beira (1400 km from Lusaka). Apart from long transit time of up to 10 days by road and 25 days by rail, the use of either of these ports is said to be expensive ranging between 50 USD per ton and 160 USD per ton (Raballand et al. 2008). 71 The assessment that there has been rapid improvement in road transport in Zambia is borne out by the differences between the 2003 and 2008 surveys in how managers thought of transport problems as business environment issue (figure 3.1). In 2008 barely 10 percent of managers rated poor transport as a major drag on business growth as compared to 35 percent who thought that way in the 2003 survey. Much of the complaint about transport in 2003 was against the poor condition of roads on account of which business estimated to have lost 2.7 percent of annual sales at the time. The lower complaint rate in the 2008 survey was matched by businesses reducing estimated losses of shipment in transit to less than 2 percent (table 3.6). Yet another indicator of the improvement in the availability and quality of transport services between the surveys is the change in the level of precautionary stocks of materials that firms maintained. The stocks fell significantly from a 40 day inventory in 2003 to 20 days in 2008 (table 3.6). Telecom services An important element of trade logistics that has improved a great deal since the early 2000s but one in which there is also significant room for improvement is telecom services. Less than one in ten of respondents in the 2008 survey saw the availability and quality of those services as a major constraint to business expansion. By contrast, about a third of respondents to the 2003 survey thought that way (figure 3.1, table 3.1 and table 3.2). This change in perception of telecommunications services as business environment issue reflected the improved availability of land line phone connections to businesses as well as the reduced cost of mobile phones. In 2003 the average length of the waiting period for getting a connection was 132 days (table 3.6). 15 This came down to 80 days in the 2008 survey. Related indicators of improvement in telecom services between the 2003 and 2008 surveys include that more than 50 percent of businesses covered in 2008 reported that they used the internet regularly, and that internet usage among 87 enterprises covered by both surveys had gone up by 22 percent between the surveys. Important as these changes are, a waiting period of nearly 3-months for fixed telephone lines is simply too long by any standard. 15 The first assessment linked the long queue for fixed lines and the high cost of mobile phone to problems to the organization, governance and ownership of the Zambia Telecommunication (ZAMTEL) and competition policy issues relating to the dual role of ZAMTEL as a regulator and competitor of private operators in mobile phone services. 72 Table 3_8: Profitability and business environment-Hausman Taylor estimates Dependent variable= Log (gross profits per unit book value of fixed assets) (1) (2) (3) ENDOGENOUS REGRESSOS## Has credit line (bank plus trade) -0.024 (0.11) Bribes as % of sales -0.052 (1.00) Mgmt time dealing with officials (%) -0.020 (1.93) Unionized work force (%) -0.003 (0.89) Output lost to power outage (% of sales) 0.063 (1.62) Days of inventory of inputs at time of new delivery 0.001 (1.26) Cost of crime/security (% of sales) -0.074 (0.32) Reported major obstacle: Macro instability 0.511 (1.80) High taxes 0.444 (1.58) Tax admin -0.295 (0.84) Skills shortage 0.831 (2.13)* Business size and age group (Base=young and small) Established small business -0.086 -0.183 -0.143 (0.92) (2.13)* (1.61) Young larger business -0.245 -0.145 -0.193 (1.76) (1.15) (1.39) Established larger business -0.085 -0.155 -0.122 (0.71) (1.40) (0.99) Log (business age) 0.082 0.094 (2.56)* (2.70)** EXOGENOUS REGRESSORS# Log (employees at start up) -0.090 -0.102 (3.89)** (3.76)** Business owner is female -0.012 0.003 (0.25) (0.03) Years of schooling of business owner -0.007 0.025 (0.24) (0.44) Constant 0.105 0.275 -0.250 (0.96) (1.33) (0.20) Observations 2014 2014 1969 Enterprises 691 691 676 Sigma_u 0.32 0.4 1.01 Rho 0.15 0.23 0.65 # Other exogenous regressos include majority share holder's ethnicity. Industry, region and year dummies Absolute value of z-statistics in parentheses; * significant at the 5% level. Significant at the 1% level Full set of results in table A3_1 73 Table 3_9: Average/Marginal Revenue Productivity of Fixed Assets- Hausman Taylor Estimates Dependent variable= Log value added per unit book value of fixed assets (1) (2) (3) ENDOGENOUS REGRESSORS Log (Number of workers) -0.077 (0.81) Log (Fixed assets/Number of workers) -0.007 -0.007 (1.79) (1.64) Business environment variables : Has credit line (bank plus trade) -0.006 -0.141 0.015 (0.03) (0.63) (0.07) Bribes as % of sales -0.037 -0.008 -0.035 (0.84) (0.18) (0.79) Mgmt time dealing with officials (%) -0.032 -0.033 -0.032 (3.19)** (3.26)** (3.21)** Unionized work force (%) -0.006 -0.007 -0.006 (2.03)* (2.21)* (1.94) Output lost to power outage (% of sales) 0.111 0.139 0.112 (3.02)** (3.72)** (3.09)** Days of inventory of inputs at time of new delivery 0.002 0.001 0.002 (1.53) (1.46) (1.56) Cost of crime/security (% of sales) 0.150 0.207 0.150 (0.71) (0.73) (0.70) Reported major obstacle: Macro instability -0.636 -0.741 -0.629 (2.47)* (2.78)** (2.46)* High taxes 0.485 0.536 0.483 (1.85) (2.00)* (1.86) Tax admin -0.779 -0.694 -0.749 (2.32)* (2.00)* (2.25)* Skills shortage 0.272 0.220 0.276 (0.71) (0.57) (0.73) EXOGENOUS REGRESSORS Log (employees at start up) -0.067 -0.086 -0.075 (2.67)** (3.40)** (2.99)** Business owner is female 0.001 0.025 0.003 (0.01) (0.26) (0.03) Years of schooling of business owner 0.036 0.041 0.038 (0.71) (0.74) (0.73) Constant 0.234 -0.453 0.224 (0.20) (0.31) (0.19) Observations 1983 1981 1983 Number of enterprises 677 677 677 Sigma_u 0.92 1.10 0.97 Rho 0.67 0.74 0.69 # Other exogenous regressos include majority share holders ethnicity. Industry, region and year dummies Other endogenous regessos include age-size groups of table 3.8. Absolute value of z-statistics in parentheses; * Significant at 5% level, ** significant at 1% level. Full set of results reported in Table A3_2 Table 3_10: Determinants of Profitability and Average/Marginal Revenue Productivity of Capital Fixed Effects (Within) IV Estimator Dependent variable : Log(value added/fixed assets) Log(gross profits/ (1) (2) fixed assets) Log (Fixed assets per worker) -0.746 (2.65)** Business environment variables : Output lost to power outage (% of sales) -0.181 0.042 -0.227 (1.78) (0.27) (1.81) Unionized work force (%) -0.002 0.012 -0.008 (0.39) (0.61) (1.10) Cost of crime/security (% of sales) -1.520 2.188 -2.388 (0.73) (0.22) (1.15) Bribes as % of sales -0.120 -0.054 -0.058 (0.97) (0.27) (0.44) Mgmt time dealing with officials (%) 0.021 0.035 0.014 (0.78) (0.73) (0.49) Days of inventory at time of delivery -0.003 -0.009 0.000 (0.88) (0.72) (0.05) Has credit line (bank plus trade) -0.971 0.531 -1.356 (1.57) (0.40) (2.04)* Reported major obstacle: Macro instability 1.498 -2.858 1.877 (2.16)* (1.65) (2.57)* High taxes -2.829 0.883 -2.580 (2.43)* (0.22) (2.00)* Tax admin 1.059 0.881 0.686 (1.53) (0.31) (0.94) Skills shortage 0.224 0.877 0.555 (0.42) (0.72) (0.90) Year dummies? Yes Yes Yes Observations 1269 1269 1269 Enterprises 386 386 386 Joint Significance Test 1.83 1.71 4.69 Sargan statistic 2.83 2.79 1.36 p-value of Sargan statistic 0.41 0.25 0.71 Absolute value of z-statistics in parentheses * significant at 5% level; ** significant at 1% level ****Excluded instruments: start up employement size, gender of owner, ethnicity of owner, location dummies, sector dummies 74 Figure 3.6 Figure 3.7 Other issues: corruption, crime, skills shortage and labor regulation In this section we discuss, very briefly, four business environment issues that were seen to be major growth obstacles by large proportions of respondents to the 2003 survey, and were treated at some 75 length in the 2004 assessment, but were no longer as high on the priority lists of respondents to the 2008 Enterprise Survey. The issues are corruption, crime, skills shortage and labor regulation. The last two issues will be taken up also later in chapter 5. Corruption The 2008 Enterprise Survey indicated that Zambian firms faced far less corruption than they did five years earlier. About 15 percent of respondents to the survey saw corruption as a major growth obstacle, which, while quite high a complaint rate in absolute terms, was also large improvement over the 46 percent of respondents to the 2003 survey who thought of corruption that way. The complaint rate was even higher among small enterprises in 2003 at about 55 percent. This change in subjective ratings applied to all business size groups and to all sectors of activity (table A3.2). It is also consistent with changes observed in bribe payments: the average amount of bribes that respondents of the 2008 paid was significantly smaller than the average reported in 2003 (table 3.4). It was also consistent with the drop between the surveys in the proportion of business managers who thought they were expected to pay bribes to government officials (figure 3.7). Regardless of how much of improvement it represents over the rate for 2003, that 15 percent of respondents to the 2008 survey complained against should make corruption an important business environment problem that merits policy makers’ attention. It is also significant in this context that complaint rate is much higher at 30% among exporters and particularly high at both ends of the size distribution of firms (tables 3.1 and 3.2) . Crime Crime is another area in terms of which Zambia’ business environment improved significantly between the 2003 and the 2008 surveys. In the 2003 survey, 49 percent of respondents thought that crime was a major or severe obstacle to the growth of their businesses, the proportion being even higher for large firms at 53 percent. In the 2008 survey the proportion of businesses who complained of the problem to the same degree was about 13 percent. This is obviously drastic improvement. But the complaint rate in the 2008 survey is high enough to make crime still an important area of concern about the current business environment. The complaint rate for 2008 is even higher than the overall average for exporters (19 percent) and microenterprises (20 percent), although these rates themselves represent improvement over what 76 was observed for each of these groups in the 2003 survey (table 3.4) . As importantly, the explicit cost that Zambian businesses incur as victims of crime as expense on crime prevention is quite high (figure 3.8). In 2008 these costs were 5.1 percent of the annual sales revenue on average, which is a relatively small drop from the average for 2003. Figure 3.8: Skills shortages and labor regulation Ten percent of respondents to the 2008 survey rated skills shortage as a major constraint to their expansion. Again this is a large enough complaint rate to make skills development a significant policy issue, especially given that the rate is even higher among exporters (14 percent) and large firms more generally (12 percent). At the same time the rate was also a welcome contrast to responses to the 2003 survey in which some 36 percent rated skills shortage the same way. Moreover, the change in ratings applied to all business size groups and sectors, and seems to have little to do with the change in the composition of the sample between the two surveys. Labor regulation was never near the top of the list of constraints respondents identified in the 2008 survey or in the 2003. Nonetheless there was a large drop in the number of businesses who saw labor regulation as major growth obstacle between the two surveys, namely, from 17 percent in 2003 to less than 4 percent 2008. Again the change applied to all size groups and sectors, the drop being 77 most pronounced in the case of exporters (22 to percent to 10 percent) and large firms more generally (21 percent to 7 percent) . The changes in the perception of labor regulation as a growth obstacle coincided with the decline in unionization between the two surveys as (table 3.5), although unionization rates remain high among large enterprises of all age groups, ranging from 26% among young businesses to just below 40% among longer established ones. The decline occurred across the board including all business size groups and all sectors of industry. Summary and conclusion One of the main messages of this chapter has been that Zambia’s business environment has improved drastically in practically all of the areas in which the 2004 assessment found it to be seriously lacking. At the same time, the chapter has highlighted key business environment problems that are still working against the growth of manufacturing employment and productivity in Zambia. These include problems relating to access to finance, business taxation, product market regulation, and the provision of physical infrastructure. The problems reduce employment and productivity at two levels. On one hand, they raise the cost of doing business in Zambia relative to that in other countries. On the other hand, they distort domestic markets in as far as they affect different sectors of the economy and different firms within each sector to different degrees. The effect of the first of these on employment and productivity is best thought of as that of an implicit flat tax imposed on activities or transactions of all producers in Zambia that would make them that much less competitive in world markets. The effect of the second is similar to employment and output losses arising from the allocative inefficiency that idiosyncratic tax rates generate. Business environment problems reduce employment and productivity by generating allocative efficiency losses in as far as they influence the mobility of resources among incumbent operators in domestic industries, or reduce the firm formation and entry rates, or impede the international trade integration of the industries. That low productivity firms tend to have larger market shares in Zambia than they would have in more advanced and better performing economies suggests that there could be some scope for increasing productivity through competition policy reforms, for which some specific proposals are being considered already by the government. The proposals include expanding the mandate of the ZCC beyond the regulation of mergers to the detection, prosecution and prevention of cartels and the abuse of dominant market power. These are commendable measures that could lead to significant allocative efficiency gains by reducing entry barriers to domestic industries. However, they are inevitably all about influencing the behavior of incumbent larger players, which is only one 78 among several influences on entry and exit rates, and on factor mobility among incumbent in domestic industry. At the top of the list of the other influences is the international integration of domestic industry (in terms of export orientation and openness to competition from imports). At least as important determinants of entry and exit rates and factor mobility, and hence ultimately of the level of competition and productivity in domestic industry as openness to trade and competition policy, are also a range of factors affecting the ability of potential entrants to respond to new investment opportunities. These include the direct regulation of entry by government and indirect barriers to entry stemming from problems of access to finance and taxation and from power shortages. In Zambia, as in other countries, business licensing and the requirement of construction permits constitute the most common forms of direct regulation of entry by government. One aspect of the improvement in Zambia’s overall business environment since the first assessment is that the time and pecuniary costs associated with meeting these requirements have been reduced substantially. At the same time, FIAS’ latest assessment is that Zambia can increase firm formation and entry rates by cutting the number of days needed to set up a business from 18 days where they stand today to 8 and by making similar reductions in the time needed for obtaining construction permits. Turning to indirect barriers to entry, the most important of such barriers in the current business environment in Zambia is inadequate access to finance according to the Enterprise Surveys. As is the case in almost every other country, smaller and start up business do not have as much access to external finance as larger and longer established businesses. While the distortion that this creates probably relates mainly to the distribution of factors and market shares across the size distribution of incumbent operator, inadequate access to finance also reduces firm formation and entry rates, which adds to the protection of market shares of incumbents. The problem of access to finance in Zambia is inseparable from that of macroeconomic stability, and achieving lasting price and exchange stability and bringing government borrowing under control over the long term is the most important step Zambia can take to improve businesses’ access to finance. Then there are major institutional gaps in the current financial system that need to be addressed in order to improve small business access to finance. To meet one of these gaps, Zambia needs to promote the development of microfinance institutions at least to the point that these meet as much of the credit needs of microenterprises as they are doing in many other countries in the region. A second institutional gap relates to the credit needs of SMEs, which Zambia is being advised to fill by investing in the development of a credit information system. Another indirect barrier to entry relates to small business taxes. Although Zambia is not a high tax economy by international standards, and the average business tax burden has fallen significantly in recent years, a major source of distortion in is still that the marginal effective tax rate is higher for SMEs than it is for larger businesses. Just like the disadvantage of SMEs in terms of access to finance, this distortion protects the market shares of larger firms among incumbents while reducing rates of firm formation and entry. 79 The third major indirect barrier to entry is that, despite the unmistakable improvement in the situation in recent years, Zambian industry continues to suffer from serious power shortages that have meant frequent outages and long queues to get connected to the public grid for start ups and expanding businesses. The outages and connection delays have added significantly to the cost of doing business in Zambia relative to other countries. They are also a major source of allocative inefficiency not only because they affect smaller and younger firms more than others among businesses already in operation, but also because they reduce firm formation and entry rates. Ultimately the shortages will be solved only with large investments in new generating and transmission capacity, which the government has sought to facilitate through a series of institutional reforms designed to encourage private investment in the sector. However, for needed investments to materialize in the long term, the government needs to help institute a more rational pricing system for electricity in the short term. The electricity tariffs that are in force now fail to cover the full cost of supply. This has not only undermined the financial viability of ZESCO, but is also a clear disincentive for private sector participation in future investment programs. Other measures needed for improving the finances of ZESCO include reducing transmission and distribution losses and resolving the build-up of payment arrears by ZESCO’s public sector customers. Yet another potential source of productivity growth in Zambia relates to trade facilitation and transport costs. The trade liberalizing reforms that Zambia has carried out since 1991 are likely to have led to significant productivity growth in as far as they are likely to have induced a reallocation of market shares from low productivity firms to high productivity ones within domestic industry, increased the incentives for innovation by domestic firms, and provided opportunities for greater economies of scale. All indications are also that Zambia can realize more of these gains to the extent that there is room for opening up its economy even more to trade through more trade facilitating measures and by reducing transport costs. 80 Chapter 4 Access to Finance This chapter provides an overview of differences between groups of firms in terms of hard indicators of use and cost of credit. To the extent that there are real differences between firms in terms of access to external finance, it is very likely that the differences result in misallocation of capital across the size distribution of firms in the economy, which ultimately means fewer jobs and lower productivity than would be possible if the allocation of capital were optimal. Getting reliable measurers of gaps in access to credit is a necessary step towards having a sense of how much Zambia could be losing in employment and productivity due to the misallocation of productive assets among firms and between sectors of the economy. Trends in the cost of borrowing One of the key indicators of financial sector development is the ratio of domestic credit to private sector to GDP. This ratio is very low in Zambia, only 9.6 percent (figure 4.1). Nominal interest rates are also high in Zambia, averaging 17 percent in 2006, according to the Enterprise Survey data (Figure 4.2). Only Ghana and Malawi have higher rates than this. Fortunately, the inflation rate has come down steeply in the recent years, to single digit levels. This has brought real interest rate down to 8 percent, which is reasonable and falls in the middle of the distribution (close to South Africa and Namibia). The high nominal interest rates could be due to high overhead costs, and especially high staff costs, and lack of competition. Another factor is the high proportion of bank non performing loans, which is quite high in Zambia – about 11 percent of total gross loan portfolio in 2005 (Figure 4.3). Only Ghana and Thailand have slightly higher non-performing loans than Zambia, with most countries on the lower side. On a positive side, the non-performing loans have improved since 2001, when the non-performing loans ratio was at nearly 25 percent, although the ratio slightly increased in 2004 and 2005. 81 Figure 4.1: Bank Credit to the Private Sector International Comparison: Domestic Credit to Private Sector, 2006 Bank Credit to the Private Sector, 2001-2006 Credit provided by banking sector (% of GDP) Zambia 9.6 Domestic credit to private sector (% of GDP) Tanzania 12.2 60 Malawi 12.3 Ghana 17.5 50 Botswana 20.1 Kenya 27.7 40 Percent of GDP Namibia 63.6 Mauritius 78.0 30 South Africa 78.4 20 Thailand 88.0 Malaysia 113.2 10 China 114.4 0 20 40 60 80 100 120 140 0 Percent of GDP 2001 2002 2003 2004 2005 2006 Source: WDI Database Figure 4.2: Cross-Country Comparison of Interest Rates Interest Rates Real Interest Rates Malawi 27 Malawi 13 Ghana 21 Ghana 10 Zambia 17 Zambia 8 Kenya 14 Kenya -0.3 Botswana 13 Botswana 1 South Africa 13 South Africa 8 Tanzania 12 Tanzania 6 Namibia 12 Namibia 7 Mauritius 11 Mauritius 2 China 6 China 4 0 5 10 15 20 25 30 -2 0 2 4 6 8 10 12 14 Percent of firms Percent of firms Source: World Bank Enterprise Surveys Figure 4.3: Bank Nonperforming loans International Comparison: Bank nonperfoming loans to total gross loans, 2005 Bank nonperfoming loans to total gross loans (%) South Africa 1.5 25 Namibia 2 Kenya 5.2 20 Malaysia 9.9 15 Percent China 10.5 Zambia 10.8 10 Thailand 11.1 5 Ghana 13.9 0 2 4 6 8 10 12 14 16 0 Percent 2001 2002 2003 2004 2005 Source: WDI Database. 82 Indicators of access to credit Only 43% of businesses covered in the 2008 Enterprise Survey have made use of any credit product, such as line of credit, loan or overdraft. This puts Zambia below average among its comparator countries – better than Namibia, Botswana, China, Tanzania and Ghana, but worse than Malawi, Malaysia, Kenya and others. It is not therefore surprising that nearly one in five of surveyed firms rate inadequate access to finance as a major or severe obstacle to their growth (figure 4.4). External finance is used by firms to finance their working capital purchases (inventories, accounts receivables) and investment in productive assets (property, plant and equipment). Two commonly used measures of the quality of access to finance are percent of bank finance used by firms to pay for their working capital and investment. Zambia scores poorly on both of these measures (figure 4.5). On average firms in Zambia finance only 4 percent of working capital and 5 percent of investment using bank funds. This is below all the other comparator countries. Figure 4.4: Access to Finance Obstacle and Credit Products Use Percent reporting major/severe obstacle Percent with overdraft or line of credit Ghana 68 Thailand 86 Kenya 45 Mauritius 84 Botswana 43 Kenya 73 Malaysia 72 Malawi 39 Malawi 68 Namibia 38 South Africa 64 Tanzania 38 Zambia 43 Mauritius 31 Namibia 40 Zambia 19 Botswana 36 South Africa 14 China 31 Malaysia 10 Tanzania 28 Thailand 8 Ghana 22 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 90 Percent of firms Percent of firms Source: World Bank Enterprise Surveys Figure 4.5: Sources of Working capital and Investment Financing Total Banks Working Capital 80 35 34 30 26 26 60 57 25 23 46 20 20 19 Percent Percent 18 40 34 31 32 31 15 29 28 30 12 27 26 11 22 20 10 20 15 7 6 11 11 5 5 8 10 9 5 4 7 7 3 6 4 5 1 2 2 1 0 1 1 1 0 0 0 Botswana China Ghana Kenya Malawi Malaysia Mauritius Namibia South Tanzania Thailand Zambia Botswana China Ghana Kenya Malawi Malaysia Mauritius Namibia South Tanzania Thailand Zambia Africa Africa Working Capital Investment Trade credit Family, friends or informal sources Source: World Bank Enterprise Surveys 83 When finance from formal institutions such as banks is not available, firms might be able to substitute other sources of funds – such as trade credit or funds from family, friends and other informal sources. In Zambia, trade credit is relatively well-developed and is used to finance about 26 percent of working capital finance on average, which is on the high end of the distribution among its comparators. Smaller firms have less access Access to credit is significantly more difficult for microenterprises than for small enterprises and significantly more difficult for small enterprises than for large enterprises (figure 4.6). Microenterprises are more likely to report that access to finance is one of the top 3 obstacles, are less likely to have a bank account and significantly less likely to have access to any of the credit products (loans, overdrafts or line of credit). Among micro enterprises only 19 percent have any of the credit products, compared with 28 percent of small, 57 percent of medium and 72 percent of large firms. These differences between small, medium and large firms are comparable to those observed in the 2003 survey. Both micro and small enterprises also have very high rejection rates when they apply for bank loans – 47 percent for micro and 59 percent for small enterprises, compared with 23 percent for medium and 16 percent for large enterprises. On the demand side, micro enterprises are also less likely to state “no need for a loan� as a reason for lack of loan application. When they do get loans, micro enterprises are likely to be charged higher interest rates – 23 percent, compared to 19 percent for small and medium enterprises and 13 percent for large firms. Among the non-micro enterprises, small firms have less access than medium or large firms. They are less likely to have any credit products, less likely to apply for loans, but more likely to be rejected for a loan. However, the difference between size groups in terms of subjective ratings of access to finance as a growth constraint are not as pronounced as what these hard indicators of access might suggest. Thus, 34 percent of small firms name access as one of the top 3 obstacles, relative to 40 percent of medium and 27 percent of large firms. Still, microenterprises report to be most severely constrained s – 51 percent of them claim access as one of the top 3 obstacles. 84 Figure 4.6: Access Indicators by Firm Size 120 9998 100 92 80 72 64 Percent 57 59 57 60 51 47 4646 40 40 34 27 28 28 24 23 23 23 19 1919 20 16 1312 13 0 Percent Percent with Percent with Applied for a Rejected No need for Interest Rate reporting a bank credit loan application a loan finance as account products one of the top 3 obstacles Micro Enterprise Small Medium Large Source: World Bank Enterprise Surveys As in many countries around the world most enterprises in Zambia rely primarily on retained earnings to finance working capital and investment (figure 4.7). However, the same pattern described above is visible here as well - micro and small enterprises have significantly less access to bank finance. Micro and small enterprises use about 2 percent of bank finance for working capital and even less for investment, compared with 6-7 percent for medium and large enterprises. Micro enterprises also use less finance from suppliers and customers – they finance 17 percent of working capital, vs. 25-30 percent for small, medium and large. This is because micro enterprises rely more heavily on finance from family and friends, financing up to 8 percent of working capital from this source, relative to 1-2 percent for other size groups. These patterns are common around the world and are signs of systematic difficulties with access that are more pronounces for smaller size firms. 85 Figure 4.7: Sources of Finance for Working Capital and Investment by Firm Size Working Capital Investment 1 2 100% 2 2 1 100% 1 1 5 4 8 2 7 7 17 25 28 30 80% 80% 2 2 6 7 60% 60% Percent Percent 97 93 85 86 40% 40% 72 71 64 62 20% 20% 0% 0% Micro Enterprise Small Medium Large Micro Enterprise Small Medium Large Retained earning Total Bank Retained earning Total Bank Suppliers and customers Family, friends or informal sources Suppliers and customers Family, friends or informal sources Source: World Bank Enterprise Survey International Differences in SME Access Relative to Large Firms That SMEs have less access to credit than large firms is a worldwide phenomenon and is not limited to Zambia. Figure 4.8 compares access in Zambia separately for SMEs and for large firms with the set of comparator countries. Left panel reports credit product usage by SMEs and the right panel reports credit usage by large firms. Figure 4.8: Access Indicators-SMEs vs. Large Firms Percent with overdraft or line of credit (SME) Percent with overdraft or line of credit (Large) Mauritius 83 Thailand 91 Thailand 77 Mauritius 89 Kenya 72 Tanzania 86 Malawi 68 South Africa 82 Malaysia 67 Malaysia 80 South Africa 59 Namibia 78 Zambia 39 Kenya 76 Namibia 34 Zambia 72 Botswana 33 Malawi 68 China 20 Ghana 66 Ghana 20 China 40 Tanzania 14 Botswana 38 0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90 100 Percent of firms Percent of firms Source: World Bank Enterprise Surveys The gap between SME access and large firms’ access is fairly large – the difference between credit product use for large firms and SMEs is 33 percent in Zambia, i.e. 72 percent for large minus 39 86 percent for SMEs. This is better than the gap in Namibia (44 percent), Ghana (45 percent) or Tanzania (71 percent), but worse than in other comparator countries. In other words, while SMEs have less access than large firms in every comparator country, the difference is larger in Zambia than in 8 out of 11 comparator countries. Characteristics of Loan products In the micro enterprise sample only 10 firms have either loan or line of credit and in the SML sample 74 firms do (table 4.1). Most loans are issued by private commercial banks, although state- owned banks have a non-trivial impact with over 10 percent of loans in the sample. Non-bank financial institutions (likely NGOs and other special-purpose lending institutions) are important in the micro enterprise sample and provide 60 percent of all micro-enterprise loans; but lenders of this type provide only 9 percent of loans to SML sample. Table 4.1: Loan Providers Micro Enterprise SML Enterprise No of No of Type of financial institution Obs. Percent Obs. Percent Private commercial banks 2 20% 55 74% State-owned banks and/or government agency 1 10% 10 14% Non-bank financial institution 6 60% 7 9% Other 1 10% 2 3% Total 10 100% 74 100% Source: World Bank Enterprise Survey About half of loans were obtained in 2006 or after, with the earliest loan in 1996. The majorities of the loans are short-term and have median maturity of one year, while three-quarters of the loans have less than 3 years maturity. This suggests that long-term finance is scarcely available in Zambia. The size of loans varies from 30 thousand ZMK to 160 billion ZMK, with the median of about 250 million. As a fraction of the estimated current value of the firm’s fixed assets the average size of a loan is about 30 percent, while the median is 17 percent, which indicates a relatively low leverage. The average and median interest rates are about 19 percent, which is relatively high, even given inflation rate of about 9 percent in 2006. 87 Table 4.2: Loan Characteristics 25th 75th Standard Variable N Min Median Max Mean Percentile Percentile Deviation Year of Approval 86 1996 2006 2006 2007 2007 2005.8 1.51 Amount at the time of approval (millions) 86 0.30 45 250 800 160,000 2,860 17,300 Amount scaled by assets 63 0.00 0.03 0.17 0.33 1.78 0.29 0.38 Average annual interest rate (%) 85 2 11 19 25 60 19.37 10.32 Total duration in months 86 2 12 12 30 84 21.53 17.00 Collateral as a percentage of loan amount 76 0 100 120 200 600 153.71 98.54 Source: World Bank Enterprise Survey Most firms have to post collateral to secure loans – over 90 percent of the firms with loans use collateral. The most popular type of collateral is land and buildings, with 73 percent of all firms use this type of collateral. Next is machinery and equipment, used by 35 percent of all firms and personal assets, used by 28 percent of all firms. However, personal assets are predominantly used by micro enterprises (70 percent of micro enterprises use personal assets as collateral, vs. only 22 percent of non-micro enterprises) and firms with unlimited liability (57 percent of unlimited liability firms use personal assets, vs. only 17 percent of LLC firms.) Accounts receivables and inventories are used as collateral less frequently – overall only about 14 percent of firms use it. However, only medium and large firms that have access to loans secured by movable assets like receivables and inventories. The ability to use movable assets and receivables as collateral is an important indicator of financial market development and sophistication, but is not very common in Zambia. Increasing access to this type of collateral would be especially beneficial for smaller firms. The collateral to loan values measure percent of asset value that can be financed by banks. A “loan to value� ratio of 0.8 (which is considered prudent by lending standards) translates into collateral to loan ratio of about 125 percent, which could serve as a useful benchmark. In Zambia median collateral to loan ratio is about 120 percent, not far from the benchmark ratio of 125 percent. 88 However, some firms are required to post much more collateral to loan value, with over quarter firms posting 200 percent or more of loan value. The average is about 150 percent, which is somewhat on a high end both relative to the benchmark of 125 percent and relative to other countries. The good news is that collateral requirements have come down as a share of loan value since the 2003 survey, when firms reported a average of 300 percent collateral to loan value. However, the collateral requirements remain high relative to other countries (figure 4. 9). Zambia scores in the upper half of the cross-country distribution of average collateral value relative to loan requirement. Figure 4.9: Collateral Requirement Value of Collateral as Percentage of Loan Value Kenya 182 Namibia 147 Tanzania 146 Zambia 141 Botswana 124 Ghana 117 Mauritius 112 Malawi 109 South Africa 103 Thailand 88 China 83 Malaysia 77 0 20 40 60 80 100 120 140 160 180 200 Percent of firms Source: World Bank Enterprise Surveys Small firms and micro enterprises frequently cite inability to meet collateral requirements as reasons for application rejections (table 4.3). Microenterprises cite this reason 57 percent of the time, small 33 percent and medium 22 percent. No large firms mention this reason for loan rejection (but the sample of large firm loan rejections is small – only 2 large firms were rejected for loans). Among other reasons for application rejections are insufficient profitability (more frequent in smaller firms) and incompleteness of loan application, which is surprisingly more frequent in small and medium firms but not in micro enterprises. Some firms choose not to apply for loans because simply because they do not need any. However, microenterprises are significantly less likely to cite this reason (only 26 percent of them do not apply because of lack of demand) than small, medium and large enterprises. Clearly the larger the enterprises – the less prominent this reason is. This suggests that smaller firms have more of the unmet demand for loans than larger firms. 89 Among other reasons, the next most important one for micro enterprises is “do not think it would be approved� – 21 percent of them do not apply for this reason, while only 5 percent of small and 2 percent of medium enterprises cite this reason. Micro enterprises also more likely to find application procedures to be more complicated. Larger firms do not apply for reasons of unfavorable interest rates. This puts them in a different category – they do not apply because of high cost of finance, while smaller firms clearly lack access to finance, which means they might be unable to get credit even at a high cost. Table 4.3: Reasons for Applications Rejections and Lack of Loan Applications Panel A: Reasons for Application Rejections Panel B: Reasons for Lack of Loan Application SML Enterprise SML Enterprise Micro Micro Reason Reason Enterprise Enterprise Small Medium Large Small Medium Large Collateral or cosigners 57% 33% 22% 0% No need for a loan 26% 47% 53% 64% unacceptable Application Insufficient 29% 22% 11% 0% procedures are 13% 11% 6% 5% profitability complicated Problems with credit Interest rates are not 0% 17% 21% 29% 16% history/report favorable Incompleteness of Collateral requirement 14% 28% 33% 0% 18% 9% 3% 2% loan application are unattainable Concerns about level Size of loan and of debt already 0% 0% 11% 0% maturity are 1% 1% 2% 0% incurred insufficient Did not think it would Other objections 0% 17% 22% 100% 21% 5% 2% 0% be approved Total 100% 100% 100% 100% Other 3% 6% 6% 14% Total 100% 100% 100% 100% Sample Size 7 18 9 2 Sample Size 104 234 104 44 90 Access to finance by degree of formality In this section we consider another important aspect of access to credit – the degree of firm’s formality measured by a registration status for micro enterprises and legal status for small, medium and large firms. We separate limited liability companies, LLC, from unlimited liability firms, including sole proprietorships and partnerships. In addition we make distinguish between registered and unregistered microenterprises. Enterprises are classified as “registered� if they have at least one of the following:  registered name with the Office of the Registrar or other government institution responsible for approving company names,  registered with the Office of the Registrar, the local courts, or other government institutions responsible for commercial registration,  an operating or trade license or otherwise registered for a general business license with any municipal agency,  obtained a tax identification number from the tax administration or other agency responsible for tax registration. In the micro enterprise sample, about 80 percent of microenterprises are registered with at least one of these agencies. The difference between registered and unregistered is very significant with respect to access to credit. Among unregistered microenterprise, only 16 percent have a bank account, while 77 percent of registered enterprises do (figure 4.10). The difference in usage of credit products is similarly striking – only 4 percent of unregistered enterprises and 23 percent of registered use any credit products, that is nearly 6 times difference. 91 Figure 4.10: Access Indicators for Formality and Age Breakdown 120 120 96 98 96 100 100 91 90 80 77 80 Percent Percent 60 56 54 55 60 57 56 50 51 49 49 47 40 43 39 40 34 40 36 37 35 31 33 28 30 29 24 25 27 23 22 23 21 16 17 18 16 16 17 19 18 20 14 20 14 12 8 4 0 0 Percent Percent with Percent with Applied for a Rejected No need for Interest Rate Percent Percent with Percent with Applied for a Rejected No need for Interest Rate reporting a bank credit loan application a loan reporting a bank credit loan application a loan finance as account products finance as account products one of the one of the top 3 Registered Micro Enterprise No Registered Micro Enterprise Yes top 3 obstacles Legal Status Unlimited Legal Status LLC obstacles 1-5 yrs 5-10 yrs 10+ yrs Source: World Bank Enterprise Survey Usage of credit products tends to increase with degree of formality (figure 4.10). Unregistered micro enterprises have the least usage are more likely to rate problems of access to finance as the biggest growth obstacle. Registered microenterprises are more likely than unregistered micros to use credit products and less likely to rate problems of access to finance as the biggest obstacle to growth. In turn, registered microenterprises are less likely to use credit products than SMEs and more likely to rate problems of access to finance as the biggest growth obstacle. Among registered SME’s LLCs are more likely to use credit products than others and less likely to rate issues of access to finance as the biggest obstacle to their growth. There is a similar pattern in the use of financing sources (figure 4.11). More formal firms finance higher percent of working capital and investment using bank finance. Formal firms also rely more on supplier credit than less formal firms. For example, unregistered microenterprises use 7 percent, registered 19 percent, unlimited liability 22 percent, and LLCs use 28 percent of supplier credit to finance working capital. Therefore, supplier credit cannot be seen as a substitute to bank financing and follows the same pattern – firms that are better able to obtain bank finance appear to also be better able to obtain supplier credit.16 Most surprisingly, registered micro enterprises use more funds from family and friends than unregistered micro enterprises (i.e. 9 percent vs. 3 percent), which may suggest that family and friends might be more willing to lend to more formal enterprises. Registration could be seen by informal creditors as a signal of higher quality or more serious intentions for the business.17 16 It’s important to keep in mind that the demand for bank credit or supplier credit is unobservable, so lack of demand might be a reason for lower usage of both – supplier credit and bank credit in less formal enterprises. However, the fact that they are less likely to say “no need for a loan� as a reason for lack of loan application suggest that lower demand cannot explain lower access in less formal firms. 17 Same caveat about potential lack of demand apply (see previous footnote). 92 Figure 4.11: Sources of Finance by Degree of Formality and Size Breakdown Working Capital Investment 100% 3 2 2 1 1 100% 1 1 2 2 1 4 1 1 3 1 9 6 5 7 4 5 5 4 22 30 26 28 24 80% 19 80% 3 3 4 3 5 3 60% 60% Percent Percent 100 97 90 89 89 91 87 89 40% 40% 67 72 69 66 66 67 20% 20% 0% 0% No Yes Unlimited LLC 1-5 yrs 5-10 yrs 10+ yrs No Yes Unlimited LLC 1-5 yrs 5-10 yrs 10+ yrs Registered Micro Legal Status Firm Age Registered Micro Legal Status Firm Age Enterprise Enterprise Retained earning Total Bank Retained earning Total Bank Suppliers and customers Family, friends or informal sources Suppliers and customers Family, friends or informal sources Source: World Bank Enterprise Survey Differences by age It has been widely documented that younger firms without a proven track record experience more severe financing constraints. These firms are more opaque because less information is available about them to the banks and often they are more risky (i.e., more likely to fail). To test whether age of the firm affects access to credit, the sample is divided into several groups of firms based on their age: SMEs and larger firms that are 5 years old or younger; firms between 6 and 10 years old and firms that are more than 10 years old. There is indeed some indication that younger firms have less access to credit products (Figure 4.10): only 27 percent of firms aged less than 5 years have any credit products, compared to 35 percent of firms in 6-10 years range and 57 percent for firms over 10 years. However, the subjective indicators of access obstacle are fairly similar across age group, and demand for loans (or more precisely, the “no need for loan� as a reason for lack of loan application) is also fairly similar, with younger firms have slightly more demand for loans. Younger firms are much more likely to be rejected for loans – 56 percent of their loan applications get rejected as compared to about 30 percent for firms older than 5 years. It appears to be slightly easier for younger firms to finance investment rather than working capital (figure 4.11). For example, firms under 5 years old finance 6 percent of investment with bank loans, relative to 4-5 percent of older firms, but the difference in working capital finance is reversed – i.e. older firms have higher proportion of working capital financed by banks, though the differences in magnitude are small. 93 Differences by Industry, Region and Foreign Ownership Access to finance also varies by industry (table 4.4). Specifically, manufacturing firms are more likely to use credit products (54 percent of them use at least one credit product, vs. 39 percent of retail and 33 percent of other industries). The difference in usage of bank finance is more pronounced in investment – manufacturing firms use 10 percent of bank finance for investment, vs. only 5 percent for working capital; while the rest of the firms only use 0-3 percent of bank finance either for investment or for working capital. This could be because manufacturing firms are able to provide better collateral, especially for machinery and equipment finance. However, manufacturing firms are also more likely to claim access as one of the three top obstacles. Exporters have better access to credit than non-exporters – they are more likely to use credit products (74 percent of them do, while only 38 percent for non-exporters), less likely to cite finance as one of the top 3 obstacles (26 percent of exporters and 37 percent of non-exporters) and they pay lower interest rates (13 percent vs. 19 percent). Among firms with different ownership characteristics, foreign owners and firms with white owners have easier time with obtaining credit: 60 percent of foreign owned firms and 67 percent of firms with white owners use any credit products, relative to only 36 percent of domestic firms and 37 percent of firms with African owners. Foreign and white-owned firms also pay lower interest rate, claim lower subjective access obstacles and are less likely to be rejected for loans. There are also some regional differences in access. Firms in Lusaka are less likely to use any credit products (34 percent relative to 50-70 percent in other regions) even though they report similar demand for loans (i.e. the “no need for loan� is similar across the regions). However, firms in Lusaka pay lower interest rates (16 percent vs. 17 percent in Kitwe, 19 percent in Ndola and 21 percent in Livingston). Kitwe has the highest loan rejection rates – 52 percent, while Ndola has the lowest – 7 percent (however, the sample of loan applications is very small, so the rejection rates are not estimated with much precision). 94 Table 4.4: Access and Firm Characteristics: Industry, Region, Exporters and Ownership Industry Ownership Owner Owner ethnicity Exporter Region Gender Manufacturing Livingston Domestic Foreign African Female Lusaka Ndola Kitwe Other Other White Retail Male Yes No Sample Size 304 115 65 366 118 283 201 298 51 135 425 59 301 66 64 53 Percent reporting finance access as one of the top 3 42% 29% 33% 37% 31% 37% 33% 40% 23% 31% 37% 26% 35% 33% 39% 27% obstacles Percent with a bank account 93% 96% 97% 95% 95% 96% 95% 94% 97% 97% 94% 100% 95% 93% 99% 88% Percent with overdraft, line 54% 39% 33% 36% 60% 42% 45% 37% 69% 46% 38% 74% 34% 67% 71% 54% of credit or loan Percent -- Applied for a loan 28% 9% 15% 18% 19% 21% 14% 17% 31% 16% 16% 29% 16% 25% 25% 22% Percent -- Rejected 34% 22% 42% 39% 23% 40% 19% 44% 43% 10% 32% 40% 35% 52% 7% 19% application Percent - No need for a loan 37% 54% 56% 45% 54% 47% 49% 45% 42% 54% 49% 42% 47% 45% 50% 53% Interest Rate 16.8 19.1 16.9 18.0 16.1 17.2 17.6 19.0 18.3 15.3 18.7 13.3 16.3 17.1 19.6 21.1 Percent that use bank 99% 95% 100% 98% 98% 98% 97% 98% 98% 99% 98% 100% 99% 96% 97% 94% finance for working capital Average amount of bank 68% 65% 71% 70% 64% 68% 67% 71% 61% 63% 68% 65% 66% 76% 66% 83% finance for working capital Percent that use bank 93% 100% 95% 96% 96% 97% 95% 96% 97% 96% 97% 94% 99% 100% 80% 59% finance for investment Average amount of bank 81% 96% 93% 89% 88% 88% 90% 89% 80% 91% 89% 87% 94% 77% 74% 0% finance for investment Source: World Bank Enterprise Survey Conclusions Zambia has a low level of financial development and provides insufficient access to finance especially to small and micro enterprises. Interest rates are relatively high, again particularly for smaller firms. The high interest rates are partially due to high overhead costs and lack of competition 95 in the banking sector. Term finance is largely unavailable for all firms, and the average loan maturity for the Enterprise Survey sample is just one year. About 43% of the firms in the Enterprise Survey sample use at least one credit product – including overdraft, line of credit or a loan. This is about average, relative to other countries. However, access to loans (as opposed to overdraft facilities) is lower than reported in an earlier survey. As a result, bank finance plays an unimportant role in financing of working capital and investment in Zambia – providing only 4-5 percent of total capital. This is lower than those for all of the comparators used in this report. Even though supplier credit is relatively well developed in Zambia, it follows the bank credit: firms that are better able to obtain bank finance appear to also be better able to obtain supplier credit, and therefore it is not a perfect substitute. As in many other countries, smaller firms have less access to external finance than larger firms. There are differences in access among different sectors: foreign firms, exporters and LLC’s use more credit products and appear to have more access. As in many other countries, firms that have audited financial statement use more credit products and report lower obstacles. One indication of difficulty of access is that collateral requirements are quite high, with many firms required to post over 100 percent of collateral relative to value of the loans, and the average of about 150 percent. However, the collateral requirements show a noticeable improvement since the earlier survey in 2004, which reported average collateral values of over 300 percent. The ability to use movable assets and receivables as collateral is an important indicator of financial market development and sophistication, and it is lagging in Zambia. Increasing access to this type of collateral would be especially beneficial for smaller firms. Overall, despite some improvements (particularly in interest rates and collateral requirements) the overall usage of credit products and the contribution of bank finance to working capital and investment have declined since the 2003 survey. The results show that financial sector is quite underdeveloped in Zambia and further efforts to deepen the financial sector reforms and lower intermediation costs, are crucial to reach the Millennium Development Goals (MDGs). One potential area of reform is indicated by the fact that state-owned banks provide a noticeable share of loans in the Enterprise Survey sample, which suggesting the need for further expansion of private sector banking t further diminishing role of the state in the banking sector (via privatization of NSCB) is recommended. High nonperforming loans affect the soundness of the banking sector and banks’ ability to finance productive investment. Stricter risk management practices and prudential regulation and supervision 96 of the banking sector, as well as better access to credit information would improve banks’ ability to offer finance to small and medium enterprises. Measures aimed at increasing competition in the banking sector would allow more financial providers to go downstream and offer more credit to smaller firms. Other contributing factors that affect access to finance are investor protection, property registration and bankruptcy procedures, all of which are not well developed in Zambia. These are important features of the business environment that are necessary for improving access to finance and thus reforms of the legal framework and judicial system should be a priority for policymakers. 97 Chapter 5 Labor Markets and Human Capital Introduction A well-functioning labor market will be vital to Zambia fully succeeding in its economic diversification drive. This chapter uses data provided by the personnel managers of surveyed firms together with individual-level employee data from the 2003 and 2008 Enterprise Surveys to describe the labor market in the manufacturing, retail and services sectors. The chapter starts with a broad description of firm perceptions on a variety of labor market constraints and firm responses to these constraints including training. The chapter then examines wage-setting behavior using firm- and worker-level data. The data used for this chapter comes from a survey of 484 firms from the manufacturing (304), retail (142) and services sector (38). Because of data limitations, wage-setting behavior is examined only for manufacturing firms. The individual-level data comes from about 1000 workers matched to the sampled firms in the manufacturing sector. The average enterprise workforce is 89 workers in the manufacturing sector, 17 in the retail sector and 129 in the services sector. In the manufacturing sector, firms in the plastics and textiles sub- sectors have a larger labor force than other firms with average employment of more than 80 workers. In the services sector, construction firms employ an average of 70 workers, while transport services firms employ over 100 workers. The workforce of the typical firm in the manufacturing sector is 20 percent female, 6 percent part- time, 9 percent seasonal. 17 percent of the workforce in the chemicals sector is temporary compared to 9 percent in the food sector. 93.5 percent of firms report that their typical worker has more than 6 years of schooling. Part-time workers in the retail and services sectors constitute about 3 percent of workforce. The average worker in the employee sample is 31.3 years old, has 6.8 years of working experience, has been with the current firm for 4.1 years and has completed 11.2 years of schooling. The gender composition of the employee sample is slightly higher than the average estimates derived from the firm sample: 22.3 percent of the sampled workers are female. Skills Shortages as a Growth Constraint Labor market constraints are very low on firms’ lists of the impediments to growth. Two constraints are pertinent to this chapter: the extent to which an inadequately educated workforce and labor 98 regulations constrain the growth and operations of enterprises. For both constraints, an overwhelming majority of firms in Zambia do not perceive either to be a major or very severe impediment to growth. Less than 10 percent of all manufacturing firms report either constraint to be a major or very severe impediment. The same holds true for the retail and services sector: less than 6 percent of retail firms and 8 percent of other services firms report being inhibited by either constraint. For firms in Zambia, the lack of a well educated workforce is much more important constraint than labor regulations. About 9 percent of manufacturing firms report the shortage of skills as a major or severe impediment to growth. Figure 5.1 shows the proportion of firms that report being constrained by a poorly educated workforce in Zambia and a set of comparator countries. It is striking that the sample of firms surveyed in Zambia in 2006 have the lowest proportion reporting major constraints. Five times the proportion of firms report inadequate skills in neighboring Malawi, more than three times the proportion in Namibia and over twice the proportion in Angola. With the exception of South Africa, middle income countries register a higher proportion of firms unhappy with the quality of the workforce. We include the un-weighted proportion of manufacturing firms surveyed in 2003 as a guide to trends in firm perceptions of formal education over the years leading up to the 2008 survey. While it is important to bear in mind how differences in the two samples potentially confound the comparison, there has been a very large downward trend in concerns about skills in Zambia. Across the two samples, the proportion of firms concerned has dropped from just over 35.8 percent to 8.6 percent: a decline of nearly 80 percent in the proportion of firms concerned about formal worker training. To rule out the possibility that differences in the samples are driving this result, we restrict our analysis to those firms that were surveyed in both years. A total of 87 firms were surveyed across the two years. We first point out these panel firms are considerably more likely to report skills to be a problem in 2007 than non-panel firms (10.3 percent vs. 7.1 percent). However, what is even more striking is that 41.4 percent of these firms reported that skills were a major or severe problem in 2003. The proportion of panel firms reporting constraints due to the education/skills of workers declined by 31 percentage points (41.4 percent in 2003 vs. 10.3 percent in 2007). 99 Figure5.1: Manufacturing Firms in Zambia are at the bottom of the pack with regards to perceptions of skills in the labor force % of firms saying that skills are a serious problem 60 49.7 50 42.9 40 35.8 32.6 30 30.7 30 25 22 20 16.9 8.6 8.9 9 10 0 Kenya Malaysia Tanzania Ghana Angola China Malawi Thailand Mauritius South Africa Zambia 2007 Zambia 2003 Namibia Source: World Bank Enterprise Surveys Note: Cross-country comparisons are only for manufacturing enterprises. The figure shows percentage of firms that report that skills shortage is a major/severe constraint to firm operation in all the countries shown Figure 5.2 shows the proportion of firms reporting worker skills as a function of a firm’s total workforce. As firm size increases, the likelihood of reporting worker skills as an impediment increases. However, the gradient is not very steep, and peaks at only about 10 percent. We advance three tentative explanations of why firms report diminished concerns with the formal education of the workforce. Firstly, it is possible that firms have made the necessary input-mix adjustments that are compatible with a low skills abundant workforce. Secondly, it is possible that the quality/quantity of formal training has risen sufficiently to match firm needs. Finally, it is possible that there has been a broad based rise in the supply of workers with adequate formal education. 100 Figure 5.2: Concerns about Worker Skills Concentrated amongst large firms Worker skills/education .5 .4 .3 .2 .1 0 -.5 -.4 -.3 -.2 -.1 0 1 2 3 4 5 6 7 8 Log employment Source: World Bank Enterprise Surveys Using the data on the 87 firms that are surveyed in both rounds we test for which of the first two explanations has the greatest traction. For the first explanation we examine whether firms that switch from reporting major/severe impediments to minor or less are also firms for which the skill composition of their workforce is reducing.18 A simple fixed effects regression of changes in reports 18 It is also possible that firms that succeed in hiring either local or foreign skilled workers experience an increase in skill composition and are likely to report diminished concerns about skilled labor availability. In this case, we would expect the opposite effect. 101 on skill composition changes is inconclusive on the direction of the association. While there is a small increase in skill composition across the two surveys, data limitations preclude a resolution of this explanation. Similarly the explanation that improvements in training explain this large secular decline in concerns about formal education of the workforce is not supported by the data. We do have to point out that our measure of training is rather crude and does not discriminate between high quality and low quality training programs. It is quite possible that the likelihood of providing training does not change across survey years but the quality of training offered does. We test an additional implication of this potential explanation: namely that if the optimal firm response to the formal education of workers crisis was training, we should expect to observe a higher proportion of unskilled workers receiving training. The data used for this test comes from only about 30 of the firms surveyed in both rounds. However, it does suggest that the proportion of unskilled workers receiving firm financed training rises by 19 percentage points. However this rise is uniform across all panel firms and is not concentrated amongst firms reporting minor or less concerns about the formal education of workers. Finally, it is possible that the supply of sufficiently skilled workers has increased between rounds. An examination of the proportion of urban dwellers with a secondary school certificate across three years for which the Demographic Health Surveys are publicly available for Zambia does not show a sharp rise in schooling. In the ten years between 1992 and 2002, there is only a 5 (2) percentage point increase in the proportion of female (male) urban dwellers who have completed secondary school. While it is possible that the decline in concerns is also driven by the major policy shift by the government to provide free primary education starting in 2002, the data are unable to support or refute this possibility. An examination of the number of years of schooling of a typical worker in the manufacturing sector in international perspective provides some insights that are useful in resolving this question.19 It is important to point out that simple comparisons of years of schooling completed could under- or over-estimate differences in schooling achievement given cross-country differences in the quality of a year of education. As table 5.2 shows, the typical worker in the modal firm in Zambia has between 7-12 years of schooling. More than three quarters of firms report that their typical worker has between 7-12 years of schooling. This is only just a shade lower than the corresponding estimate for South Africa. A higher proportion of firms report typical education levels of more than 12 years: 18 percent of Zambian firms report average education levels of more than 12 years of schooling—higher than in any of the comparators. 19 Education data was not collected in the retail/services sectors 102 Table 5.2: Percent of firms saying that the average worker in the firm has completed different levels of schooling 0-6 years 7-12 years >12 years Mauritius 97 0 4 Angola 49 49 2 Malawi - - - Zambia 6 76 18 Namibia 33 62 5 Tanzania 35 57 8 Malaysia 4 72 24 South Africa 10 78 12 Source: World Bank Enterprise Surveys Note: Cross-country comparisons are only for manufacturing enterprises. For most workers, skills development after formal schooling takes place through two distinct channels: learning-by-doing the tasks specific to ones job and on-the-job training. The range of skills acquired through both of these mechanisms includes both general skills that can be used outside the firm and sector- and firm-specific skills. The level of skills gained through learning by doing is heavily influenced by labor turnover. We measure labor turnover in Zambia by looking at the average tenure of workers surveyed as part of the labor module. The typical worker has been with the current firm for an average of 4 years (median of 3.3). Relative to the profile of firm vintage, the turnover rates implied by these estimates are moderate. In addition, there is no discernible relationship between firm size and worker tenure, although workers in the smallest firms report slightly lower average tenure than the rest of the sample. By this measure, learning by doing is fairly limited due to moderate to high turnover of workers. The second mechanism of human capital deepening is through firm-based training. The ability of firms to impart the requisite skills will depend on a variety of factors that include the extent of demand for skills development, the availability and cost of external training by specialized firms, and financial and space constraints at the firm level. In particular, the demand for skills development is affected by the concern that firms will not be able to recoup the productivity gains resulting from training since a worker can leave (see Acemoglu and Pischke, 1999) and utilize the skills gained 103 elsewhere.20 Below we examine the extent to which firms support skills development through on- the-job training and characterize the firms that provide on-the-job training. Just over 30 percent of manufacturing firms provide training to their workers. However this average conceals considerable variation in the likelihood of providing training. Figure 5.2 shows the likelihood of providing training as a function of log of firm size. Firm size is a useful proxy for firm quality. The training-size gradient is very clear. As a firm grows in size it is more likely to provide training to its workers. Above 200 employees, the likelihood of providing training is equal to or exceeds 40 percent. There are several candidate mechanisms for this relationship. Firm size could proxy for factors such as higher wages and other perks associated with working for a larger firm, that lead to lower turnover of workers. Secondly, firm size might also reflect higher liquidity and accompanying demand for training. This could also include having the space in which to conduct training sessions or the flexibility to cover for workers receiving training. 20 This difference between the social and firm-level returns to training is behind the basis for public intervention to support skills development at the firm level. 104 Figure 5.3: Likelihood of providing firm-based training as a function of firm size. Firm-based Training vs Firm Size 1 .8 .6 .4 .2 0 -.2 -.4 -.6 -.8 -1 0 1 2 3 4 5 6 7 8 Log employment Source: Investment Climate Survey Note: The figure shows the likelihood that firms provide firm-based training as a function of log firm size. The vertical line corresponds to a firm size of 100 employees. 105 To understand which of these relationships drives the firm size gradient shown in figure 5.2, we use regression techniques to identify the correlates and determinants of firm-based training. At the firm level, the regression results confirm the gradient shown in figure 5.2. Firm size matters and firms with more than 100 employees are about 25 percentage points more likely than firms with less than 20 employees to provide on-the-job training. Firms that have their accounts audited by an external auditor (a proxy for access to formal sources of external finance) are also more likely to provide training. Finally firms actively engaged in HIV prevention and treatment interventions are also more likely to provide training.21 These results are consistent both with the liquidity thesis and that firms that care (or are very sure) about the longevity of their workforce are more likely to provide firm- based training. We do not find any strong associations between unionization rates, export status, schooling or capacity utilization and training (see technical appendix). At the individual level, we examine the correlates of receiving any training that includes both self- financed and firm-provided training and the determinants of firm-financed training only. 21 percent of all workers report having undertaken either one or both types of training and 16 percent report having received firm-financed training. Across both categories of training but more pronounced for self-financed training, we find a strong association between individual schooling and receipt of any training. An extra year of school is associated with a 1-6 percentage point increase in the likelihood of receiving any training. Worker experience is also positively associated with both definitions of individual training suggesting that learning by doing and firm-based training are complements.22 This result has important implications for skills development of young workers at the beginning of their careers. Finally we find evidence for the fact that union members are nearly 30 percentage points more likely to receive training than non-union members. However, this result is driven primarily by individuals undertaking self-financed training. Surprisingly, we find that other factors constant, individuals working in the largest firms are less likely to receive firm-based training. This contradicts the strong firm size gradient documented in figure 5. 2. We conjecture that either differences in how training is defined by firms and workers or sampling biases account for this discrepancy. Compared to a set of comparator countries, a very low fraction of manufacturing firms in Zambia do not provide on-the-job training (table 5.1). In particular, relative to the more developed economies in the comparator set, skills development in Zambia is poor. Only one out of three firms provides training in Zambia, compared to over 70 percent of firms in Thailand and China and over 60 percent in South Africa and Mauritius. Even amongst the less developed comparators, skills development as measured by the proportion of firms providing training is very low in Zambia. More than 40 percent of firms in Kenya, Tanzania, Malawi and Namibia provide on-the-job training. Only Angola has a lower proportion of firms that provide training. 21 This result is consistent with Ramachandran et. al. (2006) who show that firms whose production technologies are skill intensive are more likely to invest in HIV prevention and treatment programs. 22 This result is also consistent with the fact that older workers might be less likely to leave the firm. 106 Conditional on provide training, firms in Zambia compare more favorably with respect to the proportion of the workforce that is trained. Interestingly, there is a moderate negative correlation between the percent of firms offering training programs and the proportion of workers trained (- 0.3). For example, China ranks first with respect to the percent of firms with training programs but ranks in the bottom half for percent of workers trained. It is important to point out, that the data used in this table cannot account for quality differences in training provided or the duration of training provided both within and across countries. Table 5.1a: Firm-based training: prevalence and percent of workers trained Country % Firms Offer Training % Production workers % Non-production trained workers Trained Angola 20 55 13 Malawi 52 - - Zambia 2003 34 - - Zambia 2007 31 62 21 Kenya 41 66 50 Malaysia 42 81 76 Tanzania 42 69 31 Namibia 44 69 46 Mauritius 62 34 18 South Africa 64 45 47 China 72 48 25 Thailand 76 --- --- Source: World Bank Enterprise Surveys Note: Cross-country comparisons are only for manufacturing enterprises. A comparison of the 2003 and 2007 samples suggests that the provision of firm-based training has changed little. Restricting our analysis to an examination of those firms that are surveyed in both rounds, we find that 31 percent of the panel firms provided training in 2003 compared to 37 percent in 2007. The difference, while moderately large is not statistically indistinguishable from zero. 107 Table 2b: Firm-based training-a comparison of the same firms across two years Variable 2003 2007 % Firms provide formal training 31.00 36.80 Number of observations 87 Labor Regulation Labor regulations govern the terms under which firms hire, utilize and fire workers. These terms include remuneration guidelines, leave and over-time policies and separation policies. We investigate the extent to which this regulatory regime is an impediment to firm operation in Zambia. As with concerns about the quality of the workforce, labor regulations do not impede firm operation/growth. Less than 5 percent of firms in the sample (manufacturing (3.4 percent), retail (4.9 percent) and services (2.6 percent)) find labor regulations to be a major or severe constraint to growth and operation. In international perspective, Zambia registers the lowest proportion of firms that report being constrained by labor regulations. Crucially, Zambia is well ahead of its immediate central and southern African neighbors on this matter: the next lowest proportion is nearly twice as high as in Zambia. The graph shows although low in absolute terms, labor regulations are considerably more constraining in Southern Africa than in Zambia. Figure 5.4 below also shows that the trend in firms’ concerns about labor regulations has been sharply downward. Nearly 17 percent of firms in 2003 reported serious concerns about labor regulations compared to 3.4 percent in 2007. Given that sampling variation is a possible source of gap between the two estimates, we restrict the data to the firms that are observed in both time periods. This analysis confirms the downward trend: 8.5 percent of the panel sample report concerns about labor regulations in 2002 compared to only 1.9 percent in 2006. 108 Figure 5.4: Hardly any manufacturing firms in Zambia complain about labor regulations. % of firms saying that labor regulations are a serious 30 27.9 20.7 20 16.3 16.6 16.9 14.5 15 12.7 11.4 problem 10 6.1 6.7 3.4 0 Kenya Malaysia Tanzania Ghana Angola China Malawi Thailand Mauritius South Africa Zambia 2007 Zambia 2003 Namibia Source: World Bank Enterprise Surveys Note: Cross-country comparisons are only for manufacturing enterprises. The figure shows percentage of firms that report that labor regulation is a major/severe constraint to firm operation. The findings above are not entirely consistent with other evidence. The Doing Business report (2008) collected detailed information on the legal implications of labor laws regarding hiring, firing, and the rights and obligations of employers and employees. Based upon the legal stipulations, the report calculates objective measures that assess how strict labor regulation is in a country. In the 2008 report, Zambia is ranked 121 out of 178 countries surveyed in 2008 (figure 5.5). This is at variance with the proportions reported above. The low ranking for Zambia is driven primarily by the imputed cost of firing. As a fraction of a worker’s wages, a firm must incur a cost equivalent to 178 weeks of the worker’s wage to accommodate severance and other arbitration fees. A plausible explanation for this variance is that as with many burdensome requirements, firms find ways to avoid spending the time and resources stipulated by the letter of the law. This might be efficient if both employers and employees would prefer to avoid the delays and costs involved in separation. However, if either one of the parties experiences considerably less protection than is provided by the law, then this evasive action has costs and improvements in the functioning of the courts or a revision of the laws governing labor need to be revised. The estimates in figure 5.4 above suggest that firms are not unhappy with the framework (both formal and informal) for dealing with employees. 109 Figure 5.5: Doing Business finds labor regulations to be particularly burdensome in Zambia. 200 172 175 Ranking out of 175 countries 151 150 138 121 125 100 86 90 91 75 61 66 49 50 43 33 25 0 Kenya Malaysia Tanzania Ghana China South Angola Malawi Thailand Mauritius Africa Zambia Namibia Source: World Bank Doing Business (2008) Wage Formation Assuming uniform worker productivity across countries, the level of wages paid to workers determines the competitiveness of the manufacturing sector. The level of wages (relative to productivity) and its growth trajectory is particularly important given that Zambia has actively courting foreign direct investment. Given the advantages described above, of a low-regulatory burden and a relatively well educated workforce, it is important that wage levels remain competitive to support an attractive low-cost production environment. Rising wages that are not commensurate with productivity gains are likely to result in the flight of FDI to more favorable destinations and greater competitive pressure from imports. This section compares median monthly wages for production workers across the comparator countries. Once again, it is important to point out that these comparisons do not account for differences in human capital or the sectoral composition of manufacturing in the comparator countries. Figure 5.6 below shows the median monthly wage in dollars paid to production workers. 110 Figure 5.6: Median monthly wages for production workers are a little higher in Zambia than they are in China, but lower than in most of the other African economies median monthly wage for production workers (in US$) $500 $412 $400 $365 $299 $300 $210 $200 $157 $116 $89 $100 $85 $100 $0 Malaysia Tanzania Mauritius Kenya Thailand Zambia Namibia China South Africa Source: World Bank Enterprise Surveys Note: Cross-country comparisons are only for manufacturing enterprises. The figure shows median monthly wages in constant 2005 US$. Deflators and exchange rates are from World Bank (2007). The median monthly wage for a full-time permanent production worker in Zambia is $100. Wages in Zambia are generally lower than in most of the Sub-Saharan Africa comparators. With the exception of Tanzania, the median wage in Zambia is 14 percent lower than the corresponding wage in Kenya, and one third or less of the wage in Namibia or South Africa. However and very crucial to competitiveness in any sector in manufacturing, median monthly wages for full-time permanent production workers is higher in Zambia than in China. Coupled with considerably higher worker productivity in China, the competitiveness implications of the higher remuneration are not encouraging. Comparisons across firms in Zambia Given the high relative wage costs in Zambia, it is important to identify the correlates of remuneration in Zambia. Firstly, more than half of manufacturing firms report that the modal labor contract in their firms is between the firm and the employee. Only 30 percent report that wages are set by unions and a further 18 percent state that legal minimum wage stipulations determine their 111 remuneration policies. The variation in average remuneration permits an investigation of wage setting mechanisms. The insights derived from this exercise potentially inform the design of policies to address unchecked wage increases. In this direction we examine the variation of wages across firm size, unionization rates and firm activity. Figure 5.7 shows average worker remuneration as a function of firm size. As has been shown in a variety of settings in both developed and developing countries, large firms pay higher wages. As the figure below shows, a production worker in a firm with 100 employees would earn more than 30 percent more than if he/she worked in a firm with only 10 employees. While the gradient shown in figure 7 seems to belie the operation of a competitive labor market in which a worker gets paid their marginal product, there are a number of explanations for this result. Figure 5.7: Larger firms pay higher wages to production workers Monthly Wage, Production Workers 300 250 200 150 Wages, 2005$ 100 50 0 -50 -100 0 1 2 3 4 5 6 7 8 Log employment 112 Among the predominant explanations include efficiency wages, the role of collective bargaining, search frictions and fairness norms. In order to identify which of these mechanisms best explains the wage pattern above, we employ regression techniques that control for differences in monitoring costs, collective bargaining arrangements, fairness norms and selective matching of high quality workers and better firms. The Appendix presents detailed econometric results that can effectively test some of these mechanisms in a regression framework in which competing wage-setting mechanisms are represented by one or more control variables. The findings are striking. Firstly, we find significant variation in the wages paid across sectors within the manufacturing sector. Secondly, we find that the relationship documented in figure 5. 7 remains strong even after controlling for a battery of firm characteristics. The point estimates suggest that a firm with more than 100 employees pays nearly 36 percent more than a firm with less than 20 employees. Similarly a firm with 20-100 employees pays nearly 30 percent more than a firm with less than 20 employees. These two results have usually been cited as evidence for the efficiency wage mechanism in which wages play a crucial role in inducing worker effort. Two other results suggest the role of another competing mechanism. Increasing the share of a firm’s workers that are unionized by 1 percent is associated with a 0.3 percent increase in the average wage of a production worker. While the effect of unions on remuneration has been based on the union’s control of the size of the workforce and a share of firm rents, it is not common to explicitly control for firm rents. We measure these by calculating gross profit (net revenue that excludes labor costs in the cost calculation). Increasing a firm’s gross profit by 1 billion kwacha increases the average wages of a production worker by only 0.5 percent. In addition, we find that firms that provide training pay higher wages. A one percentage point increase in the likelihood of providing training is associated with a 0.2 percent increase in the average wage. The direction of causation here is not clear. However this result lends support to the profit-sharing mechanism whereby productivity gains flowing from increased training are shared between employer and employee. The results suggest that efficiency wage considerations as well as profit sharing driven in part by the role of unions are paramount in wage setting in Zambia’s manufacturing sector. While the evidence for the profit sharing mechanism is strong in a statistical sense, the substantive significance is relatively small. This analysis suggests that the level of wages is not driven by organized labor but perhaps by considerations to induce greater effort from workers. While unions exert a small effect on average wages, the proportion of workers in Zambia that belongs to a union is moderate. Among the comparator countries, only Malawi, Mauritius and Thailand have lower unionization rates. Just under 20 percent of workers in Zambia’s manufacturing sector are members of a union. Unionization rates are even lower in the retail sector (8.9 percent) but a little higher in the services sector (25.5 percent). However, there is a strong firm size gradient. Unionization rates in firms with more than 100 employees are nearly 7 times higher than in firms 113 with less than 20 employees. The large difference in unionization rates observed across the two survey years is predominantly due to differences in the samples used. Restricting the comparison to firms that were surveyed in both years, shows no significant change in the rate of unionization. Figure 5.8: Unionization rates in Zambia are moderate 70 65.5 60 57.6 % of workers in a union 50 41.3 40 36.9 34.8 31.3 29.4 30 18.7 20 11.9 13.4 10 2.3 0 Kenya Malaysia Tanzania South China Malawi Thailand Mauritius Zambia Africa Zambia Namibia 2007 2003 Source: World Bank Enterprise Surveys Note: Cross-country comparisons are only for manufacturing enterprises. The figure shows the percentage of workers in the manufacturing sector that are unionized Having examined the wage setting motivations of firms, we turn to an examination of the worker characteristics that firms value the most. By measuring the returns to various worker attributes, this exercise is useful in delineating the scope for the labor market to support poverty eradication in Zambia. The regressions used here rely on the survey of 1000 workers that collected vital information on remuneration, worker experience and tenure and other demographic characteristics. An extra year of schooling increases earnings by about 8 to 10 percent—this is on the high end of the distribution of returns to schooling found in other developing countries. More experienced workers also appear to earn more, with an additional year of experience increasing wages by about 3 to 4 percent at the beginning of a worker’s career but falling thereafter. Unlike in many other settings, we do not find any evidence of a gender gap. Assuming similar attributes such as schooling and experience, the earnings of women are statistically indistinguishable from those of their male counterparts. The result on union membership is negative and significant in most of the 114 specifications estimated. This result is at odds with the firm-level results and suggests a potential problem with how representative the sample is. The point estimate on the gross profits variable suggests that a billion Kwacha increase in net revenues is associated with an increase in worker remuneration of 2 percent -- about four times higher than the effect measured at the firm level. In addition, we find that exporting firms pay nearly 30 percent more than non-exporter and foreign owned firms pay less than domestically owned firms. HIV Prevention, Impact of Illness and HIV on Absenteeism While there is tentative evidence that HIV prevalence rates in Zambia have reached a plateau or are declining, the current prevalence rates are high and have far reaching effects on the private sector. The last national prevalence survey found that one in six adults aged 15-49 were HIV positive with higher prevalence rates amongst women and men aged 35-39. A recent government publication on the quality and coverage of HIV/AIDS services finds that while there has been a rapid improvement in service provision, the level of access is confined to urban areas and coverage remains very low. The authors go on to call on all stakeholders, including the private sector, to support the provision of these services (Service Provision Assessment, CSO GRZ 2005). The decision for firms and the private sector in general to provide HIV/AIDS services is driven by two principle considerations. Firstly, firms are unlikely to engage in an area where public provision is efficacious. Secondly, even when public provision is poor, the return on providing these services must be positive. In the area of HIV prevention, an infection averted implies higher productivity of that worker – through fewer days of absence and greater effort available when at work. A firm weighs the cost of implementing HIV prevention strategies against the benefit of retaining a healthy worker (see Ramachandran et. al 2005 for results on what firms in East Africa are doing and Acemoglu and Pischke 1999 for a discussion on firm provision of productivity enhancing programs). Nearly one third of firms are involved in the provision of HIV prevention information. As figure 5.8 shows, there is a steep firm size gradient with larger firms considerably more likely to provide HIV- related information. While this might be driven by differential demand for reliable labor supply, it likely suggests that firms are either producing their own information campaigns or that these are costly to obtain. Since larger firms tend to be less liquidity constrained, they can afford to pay for the information materials required. A lower proportion of firms are engaged in the provision of condoms for their workers. On average only about 1 in 10 firms provides condoms for their workers. As with information, there is a discernible firm size gradient. Unlike information materials which do not require frequent or bulk purchases, condom provision can be a severe financial burden for small and medium sized firms. It 115 is also difficult to ensure that benefiting employees actually use the condoms and not sell them outside the firm. Only 6 percent of firms provide HIV testing services for its workers. Given the expense associated with testing, most of the firms that provide testing are very large firms. As the right panel shows, the likelihood of providing HIV testing is negligible for most of the firm size distribution. Figure 5.9: Firm Provision of HIV Prevention Technologies and Firm Size Likelihood of Engagement in HIV prevention vs Firm Size Information Provide Condoms HIV Testing 1 1 1 .8 .8 .8 .6 .6 .6 .4 .4 .4 Likelihood of providing Likelihood of providing .2 .2 .2 0 0 0 -.2 -.2 -.2 -.4 -.4 -.4 -.6 -.6 -.6 -.8 -.8 -.8 -1 -1 -1 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8 Log employment Log employment Log employment Notes: Source Investment Climate Survey. The figures are generated using a Fan-locally weighted regression with bootstrapped standard errors. Dashed lines represent the 95% confidence intervals. Firms were asked to report on the impact of illness in general and HIV in particular on the attendance patterns of workers. In particular firms were asked if illness or HIV had had a perceptible effect worker absenteeism in the past 24 months. 36 percent of firms report that illness was responsible for high absenteeism amongst workers. Furthermore, 19 percent of firms reported that illness in the family of workers was responsible for high absenteeism. When asked specifically about HIV/AIDS, only 7 percent of firms report a link between high absenteeism and HIV infection of workers and only 9 percent of firms report a link between HIV infection in the family and high absenteeism. 116 As figure 5.9 shows, high absenteeism appears to be driven primarily by traditional illnesses such as malaria and other infectious diseases. Across a large part of the firm size distribution, the likelihood of reporting a link between illness and high absenteeism is tightly bound in a narrow range between 30-35 percent. In contrast to illness in general, the impact of HIV on worker absence is constrained to the largest firms in the sample. For smaller firms, a negligible proportion report a link between HIV and worker absence. Figure 5.10: Impact of Illness/HIV on Worker Absence Proportion of Firms Reporting Absenteeism due to Disease Burden Worker Illness Workers HIV+ 1 1 .8 .8 .6 .6 .4 .4 .2 .2 Likelihood Likelihood 0 0 -.2 -.2 -.4 -.4 -.6 -.6 -.8 -.8 -1 -1 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8 Log employment Log employment Notes: Source Investment Climate Survey. The figures are generated using a Fan-locally weighted regression with bootstrapped standard errors. Dashed lines represent the 95% confidence intervals. Workers miss an average of just over 0.4 days per month due to own illness and a further 0.3 days per month due to illness in the family. Figure 5.11 shows the comparison in worker absenteeism across Namibia, Kenya, Tanzania and South Africa. Only South Africa has lower absenteeism levels than Zambia. Using the estimates for South Africa as a reasonable standard, a firm in Zambia loses 117 about 5 days a year due to worker absenteeism.23 This is equivalent to just under 2 percent of working time in a calendar year.24 Figure 5.11: Absenteeism is higher in Zambia than in South Africa but on par if not better than other comparators Average days absent in last 30 days 1.2 1 0.8 0.62 0.61 0.59 Days 0.6 0.53 0.46 0.44 0.38 0.4 0.31 0.3 0.2 0.05 0 Own Illness Family illness Own Illness Family illness Own Illness Family illness Own Illness Family illness Own Illness Family illness Zambia Namibia Kenya Tanzania South Africa Male Female Total Source: World Bank Enterprise Surveys Note: Cross-country comparisons are only for manufacturing enterprises. The figure shows the average number of days that a worker reports having been away from work due to own illness or illness in the family over the last 30 days. 23 Differences in the disease environment likely account for some of the differences observed in absence rates across countries. 24 This assumes a working calendar of just under 250 days. However, given that Zambia has a high number of holidays, lost time due to illness is likely higher than 5%. 118 References Africa Finance Private Sector Development Unit.2009. ‘Second Investment Climate Assessment of South Africa,’ World Bank: Washington DC Martinez, J. de Luna. 2006. ‘Access to Financial Services in Zambia,’ World Bank Policy Research Working No. 4061. Raballand, G., C. Kunaka, and B. Giersing. 2008. “The Impact of Regional Liberalization and Harmonization in Road Transport Services: A Focus on Zambia and Lessons for Landlocked Countries,� World Bank Policy Research Working Paper No. 4482. Washington DC. Republic of Zambia, Ministry of Commerce, Trade and Industry, 2009. ‘National Competition & Consumer Protection Policy-Draft No. 1,�mimeo, May. Stern, R. and P. Barbour. 2005. “Designing a Small Business Tax System that Enhances Growth: Lessons from Africa,’ mimeo, World Bank. World Bank. 2004a. An Assessment of the Investment Climate in Zambia. World Bank: Washington, DC. --------------. 2004b. Zambia: Sectoral Study of the Effective Tax Burden. World Bank: Washington, DC. -------------. 2005. Zambia: Diagnostic Trade Integration Study. World Bank: Washington, DC. -------------. 2004. Zambia: Analysis of the legal and regulatory framework underpinning investment Environment. Foreign Investment Advisory Service World Bank: Washington, DC. -------------. 2009a. Doing Business In Zambia: Reform Memorandum. Washington, DC. ------------. 2009b. Doing Business 2010. World Bank: Washington, DC. 119