Nested stochastic possibility frontiers with heterogeneous capital inputs

This paper studies the productivity impact of heterogeneous capital inputs of selected EU-15 member countries and of the U.S. at the macroeconomic level. The stochastic possibility frontiers approach of Battese and Coelli (1992) applied here is used to identify neutralities or non-neutralities between different heterogeneous capital and labor inputs. Owing to the introduction and estimation of two-stage nested translog possibility production frontiers, the otherwise huge parameter space for the seven input factors included in the model is reduced significantly. This gives more robust estimates of the remaining parameters. Due to the detailed data, specific types of biased technological change in heterogeneous capital inputs can be tested. Furthermore, time-varying inefficiency trajectories for each country are obtainable. Annual data from 1980 to 2004, calculated and published by the Groningen Growth and Development Centre, are used in the empirical analysis. The results obtained shed new light on how fast technological progress in a global economy can shift comparative advantages between countries. In particular the different factor specific impacts of ICT and non-ICT capital stocks give a more detailed picture of the structural dynamics between factor inputs than do most other empirical studies using more aggregate factor input data.