Technical efficiency and total factor productivity of United States states, 1977--1986: Multi-output distance function approach

Explaining and understanding the existence of spatial productivity differentials and its determinants have been central concerns of regional economists and policy makers. As evidenced by a vast literature, many questions still remain unresolved among economists as well as regional scientists and policy makers. There exists substantial room for improving estimation methods.;This study estimated production efficiency and total factor productivity growth for 48 U.S. contiguous states for the period 1977--1986. This study employs a multi-output (private and public sector outputs) production specification via an output distance function. To consider technical inefficiency explicitly, three frontier function approaches were used: a Data Envelopment Analysis approach; a deterministic parametric frontier model; a stochastic parametric frontier model.;The estimation results showed that choice of estimation model matters in empirical productivity studies at the regional level. By using various estimation models, this study found the existence of significant productivity differentials across regions. Technical efficiency has increased during the study period at a decreasing rate. Public capital, education, and unionization have positive impacts on technical efficiency, while the ratio of vehicle miles traveled to road supply, crime rate, the share of urban employment to the state employment, and highly concentrated industrial structure have negative impacts on technical efficiency. However, the relative size of the public sector and the size of population are not likely to have significant effects on technical efficiency.;During the study period, U.S. states achieved productivity growth via technical efficiency improvements rather than technical change. The states with lower technical efficiency achieve higher productivity growth by catching up with the states on the frontier.;More efforts to accommodate more recent years would be a useful research direction. It would be also helpful to apply the approach used in this study to urban areas for investigating their agglomeration factors.