A note on estimating expenditure elasticities in metropolitan areas

Several attempts have been made to explain the causes of variations in retail sales between geographic areas. The factors manipulated in such studies have included the explanatory variables, the size of the geographic unit under analysis, and the type of measure employed. The present study presents a method that reduces the likelihood of biased elasticity estimates and therefore facilitates the explanation of the causes of retial sales variation. The authors wish to acknowledge the suggestions and helpful comments provided by Dr. Herbert L. Lyon throughout the preparation of this project.