Back to basics: The Comanor–Wilson MES index revisited

Abstract

The present article attempts to investigate the validity of the Comanor–Wilson Minimum Efficient Size (MES) measure. The basic assumption is that firms that have exhausted scale economies are in non-increasing returns to scale. The same firms are also assumed to have a size greater than MES estimated on sales (total turnover), employment or fixed assets. Data Envelopment Analysis (DEA) is used, on a sample of firms in three Greek manufacturing industries, to classify firms in operation according to increasing or non-increasing returns to scale. On the basis of the results of the DEA input oriented model, the MES measure correctly predicts over 85% of the cases. A probit model is applied to those cases that are not identically predicted by MES concerning returns to scale. Results indicate that technical efficiency, size and age are the factors that compel MES to yield the same prediction as the DEA approach.