MULTIPLE OUTPUT MEASURES OF RETURNS TO SCALE

Abstract

In a multiple output, multiple input setting, five measures of returns to scale are currently available in the literature. These measures have been useful empirical tools, but there remain some limitations regarding data availability and the choice of behavioral assumptions. To cope with these limitations, we will introduce three new measures of returns to scale and thus will close a gap in the literature. Then, we show that our three measures are theoretically equivalent to the existing five measures. Thus, the eight measures, with their corresponding data requirements and behavioral assumptions, are also reported here for the empirical purpose of gaining greater flexibility in the choice of behavioral goals and data requirements. In this way, we unify the results in the literature and our present study and, in addition, summarize the relationship among the eight measures diagrammatically.