Unpacking the Determinants of Entrepreneurship Development and Economic Empowerment for Women in Kenya

Working Paper
Published on 1 April 2019


The International Labour Organization’s (ILO)’s Gender and Entrepreneurship Together training programme (GET Ahead) seeks to enhance women’s opportunities in entrepreneurship through knowledge and skills development in business and management. The programme brings a unique gender perspective and interactive approach in order to expose women, primarily from low-income settings, to business and working environments, development of business ideas, product design and management practices. The programme began in Thailand in 2001 and has now been carried out in at least 21 countries, serving approximately 400,000 women. This study assesses an intervention implemented in Kenya with the core objective of facilitating business growth. It offered women with small businesses in rural markets a five-day training course. Its delivery followed ILO’s implementation manuals and relied on local training providers. A year after training, some women who had received GET Ahead were offered complementary group and individual support services over a five-month period.

This impact evaluation focuses on understanding the impact of GET Ahead – and subsequent mentoring – on profitability, sales and business survivorship for treated firms. The second research question explores changes in secondary outcomes among treated firms. Next, the study focuses on the intervention’s impact on primary outcomes of non-treated firms; lastly, it seeks to understand the causal chain of process and mechanisms. It departs from the hypothesis that training and mentoring should lead to an increase in business knowledge, greater deployment of business practices and a potential increase in profitability or sales.

We worked with a sample of 3,537 women in 157 markets in four counties of Kenya: Kakamega, Kisii, Embu and Kitui. These regions are largely rural, with most of the population below the poverty line. These markets are typically small and remote, largely consisting of women operating a limited variety of businesses.

We used a two-stage randomised experiment to allow for estimation of the causal impact of training participation. First, we allocated markets to treatment or control. Then within the treatment market we randomly selected 1,172 women to invite to training. A comparison of women invited to training in the treatment markets and women in the control markets enables us to estimate the impact of training. Comparing the women in treatment markets who were not invited to training with women in the control markets enables a measurement of spillover impacts of operating in a market where others are trained. We conducted four rounds of follow-up surveys – two after one year, and two after three years – to measure impacts.

After three years, treated women are earning higher profits than the pure control group, with no spillover impact on other women in their markets. They also are more likely to have surviving businesses and higher weekly sales. This increase in business income is accompanied by improvements in mental health and subjective wellbeing. Examining impacts at market level shows that the treated markets have more customers each week, as well as higher sales. There is no change in the rate of new business entry into these markets. Treated firms show an increase in the proportion of good business practices used. A key mechanism for this market growth appears to be the training they received: business owners began keeping more reliable opening hours and diversifying their range of products. Both factors make the market more attractive for customers and allow overall sales to grow, rather than reallocating sales from one business to another.

Treatment impacts are stronger after three years than after one year. Few firms receive finance, and training does not increase the use of credit. As a result, it seems that firm owners have had to slowly build their inventories by reinvesting profits over time. Mentoring intervention was costlier per woman and has not had a significantly different impact than training alone.

Many evaluations of business training programmes have struggled to find significant impacts. A key reason for this has been the use of relatively small samples with heterogeneous firms. This lack of statistical significance has been interpreted by some as evidence that training seldom works, rather than the correct interpretation that there is a lack of evidence as to whether it works or not. The treatment impacts here are not larger than those found in prior studies, but they are much more precise and show that training can have a positive impact. The impact of US$2.60 per week is not transformative, but it represents an important increase in income for poor women. Moreover, and importantly, this benefit to trained women does not come at the expense of other women operating in their same markets. It appears that training passes a cost-benefit test, since gains would need to last 1.5 years to offset the costs.


David McKenzie

World Bank

Frank Odhiambo

Innovations for Poverty Action

Susana Puerto

International labor Organization