Efficiency Measurement with Unbalanced Panel Data: Tunisian Evidence in the Textile Industry.

This paper is concerned with the estimation of stochastic frontier production functions with unbalanced panel data when unobservable firm efficiency levels are related to explanatory variables. We perform the "weighted-means" instrumental variables panel data model proposed by R. Gardner (1998) to emphasize technical efficiency in the Tunisian textile industry during the period 1983-1994. Further, we assume a Cobb-Douglas production frontier where input uses are expressed in "efficiency units", in order to assess the extent of embodied technical progress and the impact of workers' skills on productivity. The mean technical efficiency of the firms is found to be 57%.