Parametric and Non-Parametric Calculation of Technical Efficiency: Application on the Czech Local Action Groups

The aim of the paper is to calculate technical efficiency of the Czech Local Action Groups (LAGs) by parametric and non-parametric approaches and to compare the results. LAGs are actors concerned with rural development. They receive subsidies via Rural Development Program for the Czech Republic for years 2007-2013. Hence, the efficiency of the finances' usage should be examined. Their inputs and outputs are multiple. Firstly we used Stochastic Frontier Analysis (SFA) and estimated translog output distance function. The model was fitted by maximum likelihood. The true fixed effect assumed truncated-normal distribution of the inefficiency term. Heterogeneity among LAGs was explained in the mean of the inefficiency function. Secondly, the Data Envelopment Analysis (DEA) was used. Models of Charnes, Cooper and Rhodes assuming constant and of Banker, Charnes and Cooper for variable returns to scale were calculated. We used panel data of all 112 LAGs observed in 2012 and 2013. Average inefficiency was estimated by SFA at 21.28 % and efficiency at 89.29 %. The inefficiency decreased and efficiency increased overtime. On the other hand, CCR model predicted lower efficiency (56.04 %) as same as the BCC model (62.83 %). In both cases it decreased overtime.