Breaking the metal ceiling, one enterprise at a time

Inclusive Economy28 Jan 2015Francisco Capos

Female participation in entrepreneurial activities is higher in Sub-Saharan Africa than in any other region. Given that female entrepreneurs tend to hire other women – even after taking into account the sector, person’s age, and type of firm – expanding female-owned businesses can spur employment and the participation of women in the labour market.

On average, businesses owned by women in Africa do not perform as well as those owned by men in terms of profitability, survival rate, average size and growth trajectory (see for example here, here, here, and here). Assuming the equal distribution of entrepreneurial skills between men and women initially, gender issues that lower opportunities for women may hinder the country’s competitiveness and development.

Current research by the World Bank’s Africa Region Gender Innovation Lab finds that gender differences in entrepreneurship outcomes can be explained by the systematically different strategic choices that men and women make, which are, in turn, driven by gender-specific underlying constraints. Among these strategic choices, a consistent and major determinant of gender-observed differences in performance and growth is the ‘sector’ in which the firm operates. Differences in sectoral choice among men and women are significant, with the vast majority of female entrepreneurs clustering in low value-added industries.

In a recent study (a preliminary draft of which is provided here), we examine the role of factors that may hinder or help women entrepreneurs wishing to enter male-dominated sectors. Our area of study was Uganda, where only 6% of women operate in male-dominated sectors (defined as sectors where over 75% of enterprises are male-owned), while 34% of men have businesses in these sectors. We used mixed methods to examine potential differences when businesses start up between ‘crossover’ entrepreneurs (i.e., women who own businesses in male-dominated industries) and ‘non-crossover’ entrepreneurs (i.e., women entrepreneurs in traditionally ‘female’ sectors).The study found that the average profit of a crossover firm is about three times higher than that of a non-crossover firm. It also found that women who cross over into male-dominated sectors earn as much as men in those sectors. Women who break into male-dominated industries are clearly different than their peers, but no theory of entrepreneurship exists to explain this difference.

We examined three potential explanations. First, the women who cross over could be ‘superwomen’ with superior innate skills and abilities. Second, these women could be entrepreneurs of average ability, but with psychosocial characteristics that have helped them to circumvent or overcome the norms that identify certain professions as ‘male’ (note: by psychosocial characteristics, we mean the interrelation between social and individual factors expressed in someone’s behaviours and mind). Third, it could be that these are the few women who have somehow managed to overcome structural inequalities in education or human capital, generally, and access to finance, enabling them to move into these higher return sectors.


Before looking at the groups of determining factors, there is one potential market factor that can explain why some women do not cross over: access to information. Women who remain in female-dominated sectors simply do not know that they are making less money than those who cross over. About 75% of the non-crossovers that make less than the crossovers believe that they actually make the same or more as female entrepreneurs in male-dominated sectors. While information is clearly one factor preventing some women from crossing over, there are other factors in play, which we examined through regression analysis.

Skills and abilities

Our results show that, while the women who cross over may be different in some dimensions of their personality, they are not more likely to consistently score higher in the skills that would aid them in becoming successful entrepreneurs. We used two cognitive tests measuring working memory and fluid intelligence – indicators of their ability to understand complex problems and problem-solve – and concluded that these are not important in explaining women’s entry into male-dominated sectors.

We also considered a variety of non-cognitive tests to gauge a respondent’s innate entrepreneurial spirit. This class of measure may be useful in predicting the success of entrepreneurs over time (see here and here). As with the cognitive measures, crossovers did not score consistently better than non-crossovers.

Psychosocial factors

Individual characteristics interact with women’s environment in a range of ways that might impact on their choice to cross over into a male-dominated profession. The employment choices of the entrepreneur’s father and mother or the composition of one’s siblings were found to have no significant effect on the probability of crossing over. Our analysis also indicated that being a widow or not having founded the business are not central to becoming a cross over.

In seeking to understand the importance of different psychosocial factors, we analyzed the role models that women reported as important to them in their youth. We found that women with a male role model were 20–28% more likely to be a crossover, suggesting that men’s involvement may be critical in influencing the transition of women to a higher profit sector. The importance of certain kinds of male role models was, again, evident when we asked who first exposed the entrepreneur to the male-dominated sector. Crossovers are more likely to have been introduced to the sector by their father, other male family members, or male friends and community members.

On the other hand, non-crossovers are more likely to have been introduced to their sectors by their mothers and teachers. Indeed, teachers are one of the main sources of exposure to non-crossover sectors. This is especially true for tailors, who learn to sew at school or through churches and NGOs. This suggests that Uganda’s current education system is one of the factors reinforcing the gender segregation of labour.

In terms of what started the process of crossing over, we found that external influences are far more important than the entrepreneur coming up with the idea herself. A suggestion by someone else, observing others, and being offered a job in the sector by a friend or family member are all significant and positive in predicting the probability of crossing over. These results seem to confirm the importance of active engagement by others in enabling women to move into male-dominated sectors.


Most forms of human and financial capital do not appear to play a role in crossing over, even though capital and skills acquired in the market are factors typically associated with gender gaps in entrepreneurship. Education, the most common measure of human capital, does not significantly affect the probability that a woman is a crossover. This also holds true for age and years of experience in a previous job, which captures all accumulated experience. In addition, none of the entrepreneurs in our sample – whether crossover or non-crossover – had selected their sector based on its capital requirements.

Conclusion and recommendations

Based on this analysis, we provide a set of policy recommendations. Given the lack of tested interventions in this area and the limitations of descriptive analysis, experimentation and the rigorous impact evaluation of these policies are critical.

The first ingredient in a policy to support women who wish to crossover is information. Findings from an evaluation of Kenya’s national vocational training programme suggest that information may be important in changing attitudes. However, information alone is not usually enough. Second, supportive engagement with individuals who can help guide female entrepreneurs as they seek to enter and operate businesses in male-dominated sectors is another important element. Ideally, mentors would be drawn from the entrepreneur’s existing network of family and friends. In our qualitative work, both crossovers and non-crossovers agreed that they would allow their daughters to cross over if someone reliable introduced them to the sector.

Third, while we did not find skills to be a binding constraint on women who crossover, our analysis suggests that it can be important to be exposed to the sector early on. One way to access early exposure is through work or apprenticeships, ideally in businesses owned by people within a trusted network.

In addition to these key elements, a number of other considerations may help women to cross over into male-dominated sectors. For instance, our analysis highlights the importance of targeting younger women entering the labour market; engaging key players in the local power structure in the face of potential disruption to long-standing norms; and support for crossovers beyond just support to enter the sector.