Economics Letters 11 (1983) 25-31 North-Holland Publishing Company 25 ON THE ESTlMATION OF RETURNS TO SCALE USING VARIABLE COST FUNCTIONS Ronald R. BRAEUTIGAM * Northwestern University, Evanston, IL 60201, USA Andrew F. DAUGHETY * Universrty of Iowa, Iowa Crty, IA 52242, USA Received 14 June 1982 This paper compares two methods of estimating returns to scale based on variable cost functions. It shows specifically how the methods differ, and indicates why it is important to select an appropriate method when drawing inferences about policy. zyxwvutsrqponmlkjihgfedcbaZYXW 1. Introduction The importance of economies of scale as a determinant of industry structure, and in some cases as a justification for regulation, has long been recognized in the economic literature. In recent years there has been much progress in the measurement of scale economies, because of advances both in production and econometric theory. One of the most often employed methods of measuring economies of scale takes advantage of the availability of cost data and the duality between cost and production representations of technology. Several studies have proceeded by estimating a long-run cost function, and then using duality to infer the extent of scale economies. ’ * The authors would like to thank John Panzar and Fumio Hayashi for their helpful comments. ’ See, for example, Caves, Christensen and Swanson (1980a), and Spady and Friedlaender (1978), to name only two studies that have examined economies of scale using long-run cost functions. 0165-1765/83/0000-0000/$03.00 0 1983 North-Holland