The rate of women’s entrepreneurship is high in Africa—higher than in any other region. However, this is not necessarily a sign of economic empowerment. Indeed, among entrepreneurs, the share of those who are selfemployed compared with those who are employers is highest in Africa, particularly in low-income sub-Saharan Africa. While women account for 40 percent of the non-agricultural labor force, they make up 50 percent of the self-employed but only 25 percent of employers.
Beyond the question of rates of entrepreneurship, there is also a question of whether there are performance gaps between men’s and women’s enterprises. Among employers, we find that—after accounting for differences in size, sector, and industry—any gender gap in performance becomes statistically insignificant.Among the self-employed, there is more variation and some evidence of gender gaps (particularly where women work part-time and/or in rural areas). Rather, where gender patterns are most striking is in firm size and sector and industry type: women are disproportionately found in smaller firms, in the informal sector, and in lower-value-added industries.
Thus the agenda for expanding women’s economic opportunities is one of enabling women to move into higher-value-added activities, both in terms of taking the step from selfemployment to being an employer, and in broadening the types of activities in which they engage. This chapter begins by looking at gender disaggregated patterns of entrepreneurship across regions, and then by income groups within Africa. It compares the performance of women’s and men’s enterprises, focusing on the performance of employers, as the enterprises they run have the greatest productivity and growth potential.
It examines the distribution by gender across types of entrepreneurial activities being pursued. It shows the importance of controlling for key characteristics of enterprises (sector, size, industry) and entrepreneur (particularly education) in accounting for most gender gaps in firm performance.
In understanding the differences in gender sorting across types of enterprises and entrepreneurial activities, the chapter examines gender differences in human capital and access to finance and assets. However, additional constraints in the investment climate could also be important—with women entrepreneurs well positioned to identify them and to propose solutions. Thus, the chapter concludes with a discussion of how to increase women’s participation in the policy dialogue addressing issues of relevance to entrepreneurs.
Where do women work? Using national household and labor force surveys from 137 countries, Figures 1a and b look at where women and men are economically active. Economic participation is subdivided into five employment categories, with a sixth category reflecting non-participation in the labor force. Employers (dark blue bars) are clearly a small share of the overall population for both women and men. Self-employment (pale gray bars) represents a much larger share. The shares that are in paid employment are represented by the black bars and unpaid workers by white bars. The share in agriculture (whether as self-employed, as an employer, or as a paid or unpaid employee) is represented by the light blue bars. There are a number of patterns that can be seen across regions. First, women are less likely than men to be in the labor force in every region.
Men’s labor force participation is both higher than women’s and exhibits less variation across regions. Women’s participation rates are highest in Africa (equivalently, the rate of those who do not participate in the labor force is lowest in Africa), and the gender gap in participation is lowest in Africa. Second, agriculture represents the most common form of employment within three regions.
It is highest in Africa, with little difference in gender shares. But the share of women participating in the non-agriculturallabor force in Africa falls, on average, to 25 percent. This is higher than it is in the Middle East and South Asia (less than 20 percent, but lower than the 28 percent in East Asia Pacific, 35 percent in Eastern Europe and Central Asia, and 40 percent in Latin America and the Caribbean).
Third, Africa and the Middle East and South Asia are the two regions where women’s share in selfemployment is higher than in wage employment. For men, in every region, wage earners outnumber the self-employed by at least two to one.
Eastern Europe and Central Asia is the region where wage employment is particularly high and self-employment relatively low. Fourth, rates of being an employer are low in all regions for both women and men. However, in aggregate, their activities account for a much higher share of overall employment and output, as their businesses employ those who report themselves as paid workers and unpaid workers. Fifth, gender gaps in wage employment are greater in Africa than in the other regions. The overall availability of wage work is lowest in Africa—and is disproportionately filled by men. One of the principal explanations for these different patterns is differences in income levels. Figure 2 looks within Africa, dividing countries by income levels. It is clear that there is significant heterogeneity within the continent, with the middle-income countries reporting patterns more similar to those of Latin America and the Caribbean or Eastern Europe and Central Asia than to low-income countries in Africa. Thus, high rates of agricultural activities and lower rates of being out of the labor force characterize the lowincome countries. In Africa’s middle-income countries, agricultural employment drops significantly. The share of those in wage work rises with country income and the share in self-employment falls. The share of employers, however, does not appear to vary significantly.
In low-income countries, women make up approximately 42 percent of the non-agricultural labor force. However, they comprise half of the self-employed and unpaid workers, but only a quarter of the employers. In lower-middle-income countries, the share of women in employment categories is less skewed. In upper-middleincome countries, the share of self-employed women is not much higher than the overall rate of women in the non-agricultural labor force. The share of women among unpaid workers is higher, but from Figure 3 we also know this is only a small share of the labor force. What is true is that the share of women in selfemployment falls as income rises. However, the share of employers that are women remains relatively constant, at 25 percent. Explanations that account for women’s involvement as employers need to go beyond simple links to development, and are explored below after laying out the patterns of the different types of enterprises run by women and by men.
Types of enterprises run by women and men One challenge in comparing “women’s” and “men’s” enterprises is definitional. What criteria should be used in making this distinction? For some enterprises, this is not a meaningful distinction. Behind this question is the assumption that women and men may face different constraints or be able to draw on different resources in starting or running a business. For some types of firms this should not be relevant. For example, for firms that are state owned, are publicly traded, or are incorporated so that the enterprise is an independent legal entity, the gender of an individual owner is not likely to matter. However, for smaller firms, the characteristics of the entrepreneur could matter more.
For example, there might be gender gaps in property rights, in the ability to apply for credit, or in the likelihood of harassment from officials. For the vast majority of small firms, the same person is the owner, manager, and key decision maker within the business. Knowing the gender of that person is sufficient. However, for firms with multiple owners, or for firms where the owner is not the person running the firm, multiple definitions are possible. Ownership and decision-making control are two possibilities, with a further question of whether it is necessary to look only at the principal owner or decision maker, or whether the presence of female participation is sufficient. It is not that one is correct, but these two possible criteria imply varying degrees of inclusion in “women’s” enterprises that may affect the comparisons with “men’s.” The World Bank’s Enterprise Surveys provide a means of examining the importance of the different definitions—and the potential differences in the opportunities and constraints women and men may face in operating and growing their businesses.
The Enterprise Surveys provide detailed information on investment climate conditions and firms performance based on large, random samples of entrepreneurs.3 Now covering over 100,000 entrepreneurs in 100 countries, this database provides an important tool for looking at female and male entrepreneurs around the world. The Enterprise Surveys collect information on “female participation in ownership.” A follow-on survey in six African countries also collected information on the principal decision maker. In as many as half the firms with some female ownership, the woman is not the main decision maker.
Figure 5 illustrates that the distinction between having “female participation in ownership” and a woman as the primary decision maker running the business are not the same thing. Of establishments with multiple owners of whom at least one is female, half do not have a woman as a main decision maker and 35 percent (including 55 percent of partnerships) do not have a woman even participating in a decision-making role. This was not a random distribution of firms. It was the larger, more productive multiple-owner businesses that tended to include female members among the owners but not as decision makers. Beyond distinguishing between “female participation in ownership” and “women as prime decision maker,” we also look at sole proprietors where the owner and decision maker are almost always the same person. This makes distinctions along gender lines much clearer, but the firms in the sample often have fewer employees and lower levels of sales. For the larger Enterprise Survey sample, the share of enterprises with “female participation in ownership” and the share of sole proprietors who are women show that the former includes a higher share of “women enterprises.” While “female participation in ownership” averages over 25 percent across the region, there is considerable variation across countries, with Niger reporting 10 percent and Ghana just under 50 percent.
When restricted to sole proprietors, the shares of female firms are substantially lower (for example, in Swaziland and Botswana), but there are some exceptions (e.g.Ghana, Kenya, Rwanda, and Zambia,). Beyond looking at rates of ownership, the next section examines whether there are consistent differences by gender in the types of enterprises women and men run. As has been found in the literature,4 women are more likely than men to work in smaller firms, in the informal sector, and in lower-value-added sectors. This has been documented based on household survey data or on samples of microenterprises.5 The results here also show how the pattern changes when looking at the set of employers that largely operate in the formal sector. Size of the enterprise Using the “female participation in ownership” criterion, there is little difference in gender composition by size—until reaching fairly large firms in Africa. However, looking only at sole proprietorships, the share of women declines with firm size, even starting at firms with 10 or more employees. Sub-Saharan Africa has relatively lower female participation for all sizes of firms, and more so for larger firms (see in Figure 7 that female participation is roughly 35 percent in all size categories outside of sub-Saharan Africa but in that region it is roughly 28 percent for small- and medium-sized enterprises and 15 percent for large firms).