On a model of environmental performance and technology gaps

In this paper we consider a stochastic directional technology distance function to re-examine the results of recent research in which the authors estimate a generalized directional distance function using programming methods, derive technology gaps and, in a second stage, they fit a Markov process to the technology gaps. One problem is that in the second stage efficiencies and gaps are themselves estimated. Moreover, the authors consider two groups (Annex I and non-Annex I countries according to the Kyoto protocol). We allow for endogeneity of good and bad outputs and inputs, endogenously determined groups of countries, endogenous directions for each country and group, and a distribution of technological gaps (with respect to the meta-technology) which is based on a Markov process. We use a semi-parametric directional technology distance function and we propose stochastic envelopment of different frontiers allowing for its own “meta-inefficiency”. All quantities of interest are estimated jointly using numerical Bayesian techniques.