Returns to Scale in Retail Trade*

This paper attempts to estimate genuine scale effects in retail trade from a cross section of retail stores in Israel. This is done by estimating a simple production function for several retail branches and employing the faithful old direct Cobb‐Douglas structure with value added as output and labor and capital inputs. And indeed despite the well‐known peculiarities of the retail industry, a cross section estimation produces “normal” production‐function estimates with reasonable input elasticities. The estimates also identify marked increasing returns‐to‐scale parameters, higher in food and lower in branches less affected by consumer participation and geographical dispersion. These increasing returns may explain a good part of the increase in sales per unit of inputs observed in time series.