Mentors or Teachers? Microenterprise Training in Kenya By Wyatt Brooks, Kevin Donovan, and Terence R. Johnson * We use a randomized controlled trial to demonstrate that inexpe- rienced female microenterprise owners in a Kenyan slum benefit from mentorship by an experienced entrepreneur in the same com- munity. Mentorship increases profits by 20 percent on average with initially large effects that fade as matches dissolve. We con- duct a formal business education intervention, which has no effect on profits despite changes in business practice. Our results demon- strate that missing information is a salient barrier to profitability, but the type of information matters: access to the localized, spe- cific knowledge of mentors increases profit while abstract, general information from the class does not. JEL: O12, L26, M53 In urban areas of developing countries, microenterprises are ubiquitous and a large fraction of the workforce is engaged in entrepreneurship. 1 Many of these very small firms generate low profit and have few or no employees. Understanding the reason that these firms have low profits, and why they generate so little * Brooks: University of Notre Dame, 3060 Jenkins Nanovic Halls, Notre Dame, IN 46556, wbrooks@nd.edu. Donovan: University of Notre Dame, 3060 Jenkins Nanovic Halls, Notre Dame, IN 46556, kdonova6@nd.edu. Johnson: University of Notre Dame, 3060 Jenkins Nanovic Halls, Notre Dame, IN 46556, tjohns20@nd.edu. Thanks to conference and seminar participants at Arizona State, Notre Dame, the World Bank, York University, the Chicago Fed Development Workshop, Development Day (Michigan), the Econometric Society NASM (Penn), IGC Growth Week at LSE, the NBER Summer Institute (Productivity, Entrepreneurship, and Development), NEUDC (MIT), the World Bank ABCDE for comments and insights, especially Nick Bloom, Paco Buera, Bill Evans, Xavi Gin´ e, Joe Kaboski, Molly Lipscomb, David McKenzie, Ezra Oberfield, and Chris Woodruff. This paper also benefitted sub- stantially from discussions by Paulo Bastos, Jing Cai, and Dan Keniston, along with comments and suggestions from two anonymous referees. We thank the Ford Family Program and the Helen Kellogg Institute for International Studies for financial support, including Bob Dowd and Dennis Haraszko for their help coordinating the project, and Lawrence Itela, Jackline Aridi, and Maurice Sikenyi for their ex- cellent work managing the project in Dandora. Brian Mukhaya provided outstanding research assistance. This research was conducted under University of Notre Dame IRB protocol 14-05-1801. Appendices are available online. 1 According to Gollin (2008), developing countries have substantially higher rates of entrepreneurship than rich countries. The proportion of workers that identify as entrepreneurs is 75.3 percent in Nigeria, 74.5 in Bangladesh, and 50.4 in Guatemala. This compares to 8.2 percent in the United States and 10.6 percent in the United Kingdom. 1