Too Large or Too Small? Returns to Scale in a Retail Network

Performance in retailing is usually evaluated by routine use of ratio analysis, but due to the univariate nature of this simple management tool there are many drawbacks to the obtained results. Therefore, the aim of this study is to demonstrate successful employment of parametric and non–parametric methods for evaluating technical performance in retailing. We also show how to utilize DEA results, when parametric methods do not satisfactorily perform due to their strict distributional assumptions. Results of this study are used to optimize the retail chain of a European mobile telecommunication network operator by providing estimates of and recommendations for improvements in the productive efficiency of the chain operations. Estimates of store–level technical and scale efficiency indicate that a majority of stores are operating in the decreasing returns to scale region of the production possibility set. The employed methodology allows us to identify input excesses and to address a means of reducing them.