Measuring and transferring retail productivity: Evidence from a structural model and a mentorship intervention in Zambia

The retail sector is the biggest source of jobs across the developing world, accounting for a larger share of urban employment in most developing countries than manufacturing. Despite that, effective tools to measure retail productivity are lacking. Standard methods to measure productivity have been designed for the manufacturing sector, making them deficient for the study of retail. There is a lack of academic work that specifically focuses on retailers although they represent a disproportionate share of firm level interventions in developing countries.

To fill this evidence gap, this project implements a novel methodology to estimate productivity in the retail sector of developing countries. In collaboration with Prospero Zambia, the researchers will collect high-frequency data to construct a panel of 2,500 small and growing retailers selling food and conveniences. Using a structural model, they will provide new estimates of productivity across various relevant margins for retail shops including customer acquisition, efficient use of inputs, and inventory management. These measures will be used to guide a mentorship intervention that will reveal which margins of productivity are most easily transferrable across entrepreneurs through mentoring.

Other than potentially improving matches, this study will inform policymakers about which aspects of productivity can be taught and should be targeted by interventions.