Small and medium sized enterprises (SMEs) are being targeted as never before for their potential to stimulate growth and create jobs. Many of the development agencies (World Bank, EBRD, Islamic Development Bank, among others) have plans to expand programs for financing and supporting SMEs. Most recently at the World Economic Forum in Jordan, government officials from countries in the Middle East and North Africa, as well as from the US and EU, talked up their potential for creating jobs in the region. This would be good news if it was true, but a growing body of evidence suggests that SMEs are not the silver bullet that we had hoped for.
While small firms do create the most jobs, they also destroy the most jobs, and ultimately what matters for the labor force is net job creation. Studies using industrial surveys, which are the most rigorous, tend to find that net job growth comes from large firms and startups. The most comprehensive work has been done on the United States, where researchers examine firms as opposed to establishments. Enterprise data may lead to size misclassifications; for example, think of McDonald’s where each restaurant is small but the firm is huge. Census data also includes information on entry and exit over time, which is not observable in subsamples of the universe of firms. Recent work on the United States shows that once firm age is controlled for there is no systematic relationship between firm size and job creation; employment creation is all about new and young firms. Using census data for Moroccan and Tunisian firms, we find similar results: startups and large firms account for the bulk of net job creation.
As shown in cross-country studies of Africa, Latin America, and the rest of the world, one reason SMEs are not the main net job creators is that they tend to be less productive than large enterprises and account for a smaller share of productivity growth. A sizable share of the productivity gap between developed and developing countries can thus be explained by the estimated 30 percent share of SMEs in economic activity in developed countries as compared with a 60 percent share for developing countries. Of special concern is that SMEs in developing countries tend to be exceptionally stagnant as compared with their developed country peers. A recent study estimates that the failure of small firms to grow into large firms lowers productivity growth in manufacturing by 25 percent in Mexico and India as compared with the US.
It is not surprising startups and large firms are more productive and generate both jobs and growth. This is precisely what the theoretical literature on heterogeneous firms tells us. High-productivity firms will attract more resources and grow faster, and in equilibrium be relatively large. A few high-productivity startups will also grow fast and absorb labor.
Promote Entry and Startups not SMEs
Economic development is about moving resources to their most productive uses. Promoting entry and startups is therefore likely to make more sense than broad SME promotion schemes. As noted, young firms do account for a sizable share of net job creation and productivity growth. In addition, the density of startups is a major source of growth. Regions with more startups relative to the working age population in Mexico, India and the US create more jobs and grow faster. Ensuring ease of entry to start a business and access to finance for entrepreneurs is therefore likely to prove more valuable than providing access to finance for mature SMEs.
Do SME Interventions Really Work?
Even if SMEs were found to promote job creation or growth, there would still be the issue of whether programs designed to promote SMEs are effective. Do SME financing, training, or logistics interventions facilitate the creation of jobs and faster growth? And if so, which programs are most effective?
Unfortunately, there has been no randomized experiment that I am aware of that evaluates whether efforts to promote SMEs lead to improved employment or productivity outcomes. There have been a number of ex-post attempts to measure whether SME programs work, especially in Latin America, but selection bias haunts these studies. In particular, more productive firms are likely to apply for and receive financing or training, so better performance of the so-called treatment group as compared with a similar control group could be a result of selection. Somewhat surprisingly given this bias towards a positive impact, these studies do not offer an unambiguously strong role for SME promotion, highlighting a need for caution going forward. For example, a study of Chile that examines different types of programs finds no evidence that credit programs alone are successful at promoting employment, productivity, or wages. Only interventions that encourage upgrading or better logistics help firms. A study of Mexican SMEs finds that the implementing agency matters-programs under the Ministry of Science and Technology have positive effects, but programs under the Labor Ministry lead to lower output, sales and exports. This perverse outcome suggests the possibility of capture in poorly designed programs. Thus, choosing interventions and designing programs carefully is critical to ensure against negative results.
If the trend toward SME targeting is to continue, we need a better understanding of the effectiveness of various programs in their stated goals. Assuming SMEs are especially dynamic, leading to the private sector to finance them, SME interventions may be unnecessary. There would need to be a distortion or an externality that the intervention addresses for it to be warranted. In this case, identifying the distortion is key to designing the appropriate program. A carefully designed randomized experiment on SME interventions and entrepreneurship initiatives should be implemented to ensure that the millions of dollars flowing into this channel are well spent and high expectations not dashed.
Join the Conversation