Measuring and Estimating Retail Productivity

About this event

We develop and apply a new method to estimate the productivity of small and medium retailers in developing countries. The method assumes a successful retail shop must 1) attract customers, 2) source and stock inventory, and 3) choose the right mix of products from the right suppliers. Our method estimates the productivity of a shop across all three dimensions. We apply our method to a novel retail-specific dataset collected through high frequency surveys of a sample of independent retailers in Lusaka, Zambia. Our results suggest the dimensions of productivity are correlated but also contain substantial independent variation. An entrepreneur can be highly productive in one aspect of running a shop but unproductive in others. Under these conditions, our model predicts that standard estimators would yield deeply misleading results. Policy interventions based on those estimators could actually lower social welfare.