268 privatesector P U B L I C P O L I C Y F O R T H E NUMBER NOTE 2004 SMEs, Growth, and Poverty FEBRUARY Thorsten Beck and Do Pro-SME Policies Work? Asli Demirgüç-Kunt This Note explores the relationship between the size of the small and Thorsten Beck medium-size enterprise (SME) sector and economic growth and (tbeck@worldbank.org) is senior financial economist, poverty reduction. A new study finds no support for the widely held and Asli Demirgüç-Kunt belief that SMEs promote higher growth and lower poverty. But it (ademirguckunt@ worldbank.org) is research does provide some support for the view that the quality of the PRESIDENCY manager, in the Development business environment facing all firms, large and small, influences Research Group of the World economic growth. VICE Bank. To accelerate growth and reduce poverty, inter- ments, direct government support of SMEs can national aid agencies provide assistance tar- boost economic growth and development. geted to small and medium-size enterprises Finally, some argue that the growth of SMEs (SMEs) in developing economies. The World boosts employment more than the growth of DEVELOPMENT Bank Group, for example, approved more than large firms because SMEs are more labor US$10 billion for programs supporting SMEs in intensive. So subsidizing SMEs may help reduce SECTOR the past five years. Does this pro-SME policy poverty. work? Even as international donors channel a large and growing amount of aid to subsidizing SMEs, PRIVATE The arguments for and against skeptics put forth four views questioning the Advocates of the pro-SME policy make three efficacy of this policy. First, some emphasize the core arguments for its effectiveness. First, they advantages of large firms. Large enterprises argue that SMEs enhance competition and may exploit economies of scale and more easily GROUP entrepreneurship and thus have economywide undertake the fixed costs associated with benefits in efficiency, innovation, and produc- research and development, boosting produc- BANK tivity growth. So direct government support of tivity. And empirical evidence from both indus- SMEs can help countries reap social benefits. trial and developing countries shows that large Second, proponents often claim that SMEs are firms offer more stable employment, higher generally more productive than large firms but wages, and more nonwage benefits than small WORLD are impeded in their development by failures of firms, even after differences in workers' financial markets and other institutions. Thus, education, experience, and industry are con- THE pending financial and institutional improve- trolled for. S M E S , G R O W T H , A N D P O V E R T Y D O P R O - S M E P O L I C I E S W O R K ? Second, skeptics challenge the assumptions been hampered by the lack of comparable cross- underlying pro-SME arguments. Some research country data. A new study provides the first cross- finds that SMEs are neither more labor intensive country evidence on the links between SMEs and nor better at creating jobs than large firms. economic growth and poverty reduction. Moreover, recent work finds that underdevel- oped financial and legal institutions do not hurt The provocative findings only SMEs. Indeed, research finds that such Comparable cross-country data are notoriously institutions constrain all firms from growing to hard to come by. For starters, countries define their efficient size. a small or medium-size enterprise in many dif- 2 Third, skeptics question the validity of con- ferent ways. Moreover, by definition, the data sidering firm size to be an exogenous determi- exclude the informal sector, where many such nant of economic growth. According to this firms operate in developing countries. So the "institutional" view, natural resource endow- data used in the analysis are the best available, ments, technology, policies, and institutions not the ideal. help determine the industrial composition and The study carries out an econometric analy- optimal firm size in a country. For example, sis using two newly compiled indicators of the some countries may have endowments provid- SME sector based on its share of manufacturing ing a comparative advantage in goods produced employment (box 1).1 The study also builds an most efficiently in large firms, while others have a comparative advantage in goods produced Box Measuring the SME sector most efficiently in small firms. Similarly, in 1 countries open to international trade, the opti- The study compiled two SME indicators for the mal firm size may be larger than in countries econometric analysis. One is the share of the SME sector that are less open. And some argue that firm size in the total official labor force in manufacturing, with reflects the margin between intrafirm transac- 250 employees taken as the cutoff for defining an tion costs and market transaction costs, such enterprise as small or medium-size. This indicator that as market transaction costs fall relative to provides a consistent measure of firm size distribution intrafirm transaction costs, the optimal firm size across countries. The second indicator is the share of falls. For institutional and technological rea- the SME sector in the total official labor force in manufacturing based on the official country definition of sons, this margin varies across industries and a small or medium-size enterprise, with the cutoff countries. So in this view pro-SME policies could ranging from 100 to 500 employees. This indicator takes distort firm size and potentially hurt economic into account that a country's economic and institutional efficiency. characteristics might determine whether a firm is defined Fourth, some skeptics, taking a "business as small, medium-size, or large. environment" view, doubt the crucial role of These indicators have shortcomings. For example, SMEs and instead emphasize the importance of information on SME employment outside manufacturing the business environment facing all firms, big (such as in agriculture and services) would be useful to and small. Low entry and exit barriers, well- have, but no cross-country data are available for the share defined property rights, effective contract of SMEs in other sectors. Another potential problem is that enforcement, and access to finance--all factors the definition of SMEs is restricted to formal enterprises, conducive to competition and private commer- yet informal enterprises may account for an important share of output in some economies. The analysis controls cial transactions--may encourage SMEs. But for the importance of the informal economy, however, by these skeptics focus not on SMEs, but on the incorporating estimates of the size of the informal sector environment facing all businesses. So, like other relative to the formal sector in each economy. Here the skeptics, they question the policy of subsidizing analysis relies on earlier work estimating the market value the development of SMEs. of output produced by the informal sector as a share of Much research has evaluated pro-SME argu- measured GDP. It uses measures of informal activity rather ments at the level of the firm, industry, or coun- than the informal labor force because few countries have try. But cross-country studies of the relationship data on the size of informal sector employment. between SMEs and economic development have indicator of the quality of the business envi- But the results do not take into account the ronment facing firms of all sizes--an aggregate possibility of reverse causation--that high measure of entry and exit costs, contract growth might lead to the emergence of many enforcement, and property rights protection. small enterprises. Moreover, the results could be Countries are considered to have a more effec- the outcome of a third factor, not included in the tive business environment if property rights are estimations, that might drive both the high share well protected, the costs of contract enforce- of SMEs and the high growth of the economy. ment and business registration are low, and the Regressions that explicitly control for reverse insolvency process is cheap, fast, and efficient. causation and the possibility of such a third fac- 3 The data show that the importance of SMEs in tor erode the significance of the relationship manufacturing varies greatly across countries between SMEs and economic growth. This find- and that those with a smaller share of SMEs also ing is consistent with the view that a large SME have a business environment that is hardly sector is a characteristic of fast-growing econo- conducive to doing business (table 1). The mies but not a cause of their rapid growth. cross-country econometric analysis yields some When the analysis focuses on income growth interesting results. among the lowest income quintile rather than the overall population, it again finds no evi- SMEs and growth: correlation or causality? dence for the importance of SMEs. Nor does it Regressions of GDP per capita growth, averaged find any statistically significant relationship over the 1990s, on either of the two SME meas- between the importance of SMEs and the depth ures and an array of other country characteris- and breadth of poverty across countries. tics that can account for differences in growth across countries show a strong relationship Business environment and growth between the importance of SMEs and economic While cross-country comparisons provide no growth. This relationship still holds when the evidence that SMEs cause economic growth analysis controls for the importance of the in- or reduce poverty, cross-country regressions formal economy and excludes transition provide qualified evidence that an effective economies and Sub-Saharan African countries business environment does cause growth. whose growth path might vary from those in Cross-country regressions of GDP per capita other regions. growth on the business environment indicator Table Size of SME sector and quality of business environment, selected countries, 1990s 1 Average GDP Average annual per capita, GDP per Percentage of 1990­99 capita growth, manufacturing Business (constant 1990­99 labor force environment Country 1995 US$) (percent) in SMEsa indexb Brazil 4,327 0.63 59.80 ­0.34 Cameroon 653 ­1.74 20.27 ­1.98 Germany 30,240 1.43 70.36 0.82 Indonesia 963 3.09 79.20 ­1.37 Korea, Rep. of 10,508 5.41 78.88 1.03 Turkey 2,865 2.49 61.05 ­0.12 Ukraine 1,190 ­7.17 5.38 ­0.56 United States 28,232 1.93 52.54 2.26 a. Small and medium-size enterprises are classified on the basis of the official country definition. b. The business environment index is a composite of four measures relating to the protection of property rights, the costs of contract enforcement and business registra- tion, and the cost and efficiency of the insolvency process. It has an average value of zero and a standard deviation of one. Higher values indicate a more effective busi- ness environment. Source: Meghana Ayyagari, Thorsten Beck, and Asli Demirgüç-Kunt, "Small and Medium Enterprises across the Globe: A New Database," Policy Research Working Paper 3127, World Bank, Development Research Group, Washington, D.C., 2003. S M E S , G R O W T H , A N D P O V E R T Y D O P R O - S M E P O L I C I E S W O R K ? and an array of other potential growth factors ating aggregate growth, the results do not show a strong relationship between an effective suggest that the business environment has an business environment and economic growth. effect on poverty beyond its effect on the over- Moreover, they show that this relationship still all economy. holds when the analysis takes into account the Together, these findings have important pol- viewpoint possibility that faster-growing countries might icy implications. They suggest that rather than adopt more effective business regulations or directly subsidizing SMEs and aiming for a large that a third factor might drive both an effective number of small enterprises, policymakers is an open forum to business environment and economic growth. should focus on creating a business environ- encourage dissemination of So, unlike for SMEs, there is evidence that an ment that allows easy entry and exit for firms public policy innovations for effective business environment is not just a char- and assures entrepreneurs and financiers that private sector­led and acteristic of successful economies but also plays property rights and contracts will be enforced. market-based solutions for an important part in their success. development. The views Cross-country comparisons show that the published are those of the positive relationship between an effective busi- authors and should not be ness environment and income growth holds as Note attributed to the World much for the lowest income quintile as for the 1. The raw data were collected from national and Bank or any other affiliated rest of society. But there are limits: the results do international sources, such as the United Nations, Inter- organizations. Nor do any of not show that a good business environment has American Development Bank, and Organisation for the conclusions represent an effect on poverty reduction beyond its posi- Economic Co-operation and Development. For details official policy of the World tive effect on GDP per capita growth. see Meghana Ayyagari, Thorsten Beck, and Asli Bank or of its Executive Demirgüç-Kunt, "Small and Medium Enterprises across Directors or the countries Policy implications the Globe: A New Database," Policy Research Working they represent. Cross-country comparisons suggest a strong pos- Paper 3127 (World Bank, Development Research Group, itive association between SME development and Washington, D.C., 2003). To order additional copies economic growth. But this relationship does not contact Suzanne Smith, hold up when the analysis controls for reverse managing editor, causation or for a third factor that might drive Room I9-009, both growth and the emergence of many SMEs. The World Bank, 1818 H Street, NW, Moreover, cross-country comparisons do not Washington, DC 20433. show that SMEs do much to boost the incomes of the poor or that they have a significant rela- Telephone: tionship with the depth and breadth of poverty. 001 202 458 7281 So while a thriving SME sector is a characteristic Fax: of flourishing economies, the results do not sup- 001 202 522 3480 port the contention that SMEs accelerate growth Email: and reduce poverty, calling into question the ssmith7@worldbank.org policy of directly subsidizing their development. But cross-country comparisons do point to Produced by the potential benefits of policies that strengthen Maryland Composition the business environment in ways that foster Company, Inc. competition and facilitate commercial trans- actions for all firms, large and small. The Printed on recycled paper results show that a measure of the business environment--an index incorporating informa- tion on entry and exit barriers, the protection of property rights, and the efficiency of contract enforcement--is associated with the growth rate of GDP per capita. But while a sound business environment tends to help the poor by acceler- T h i s N o t e i s a v a i l a b l e o n l i n e : h t t p : / / r r u . w o r l d b a n k . o r g / V i e w p o i n t / i n d e x . a s p