A Generalized Linear Programming Approach to International Production Efficiency Rankings

Abstract

Productivity studies generally compare the Solow residuals for the more developed countries such as the OECD countries.1 In this paper we follow up on the frontier estimation approach of Fare, Grosskopf, Norris and Zhang (1994) to 17 OECD countries, and develop a generalized efficiency index for a much larger set of 57 countries both developed and underdeveloped by employing measures of per capita gross national product and resource availability indicators compiled by the World Bank and other international institutions. Using a Data Envelopment Analysis (DEA) based linear programming approach, we maximize the components of per capita GDP subject to minimizing specific resource intensity measures. Because DEA provides a units invariant comparison of relative efficiency, this programming approach yields a ranking of the most robustly efficient to the most robustly inefficient countries. As expected the more developed countries tend to be more robustly efficient, but intriguing contrasts between developed and less developed nations exist. We believe that the analytical tools developed in this paper have significant implications for meaningful comparison of cross-national data both at the inter governmental and non- governmental (NGO) level. Used with appropriate precautions, the DEA based analytical procedure developed here can be used by trans-national organizations including MNCs in allocation of scarce resources in the public and private sectors.