Wednesday, 11 July, 2012 | 12:00 | Defense - PhD

Viliam Druska: “Essays on Efficiency Measurements”

Dissertation Committee:
Michal Kejak (chair)
Jan Kmenta
Daniek Münich
William C. Horrace



The unifying theme of this dissertation is the measurement of production efficiency, covering both parametric and non-parametric approaches to efficiency assessment. The first chapter considers the estimation of a fixed-effect panel data model with disturbances that are spatially correlated, based on a geographic or economic proximity measure. For when the time dimension is small (the usual panel data case), the study develops a generalized moments estimation approach based on a cross-sectional model from Kelejian and Prucha (1999). This approach is then applied in a stochastic frontier framework to a panel of Indonesian rice farms. Within this framework, spatial correlations are based on geographic proximity, and represent productivity shock spillovers across the production units. Using a Moran I test statistic, the first chapter empirically demonstrates that productivity shock spillovers may exist in this (and perhaps other) data sets, and that these spillovers have profound effects on technical efficiency estimation. The second chapter represents a logical extension of the first as it theoretically develops a random effect panel data model that accounts for spatial correlation across disturbance terms. The model is then applied within the framework of production frontier to the same data set of Indonesian rice farms. The study empirically confirms the impact of spatial correlation on the estimates of technical efficiency and compares the results with the outcomes from the first chapter. The empirical results also indicate that the technique developed here provides a viable alternative to the incorporation of time-invariant regressors in the equation specification. The third chapter addresses the drawbacks of the routine use of ratio analysis in the assessment of retailing performance. Applying multiple input–multiple output data envelopment analysis (DEA), the study assesses the technical and scale efficiency of the retail chain operation of a European mobile operator and identifies input excesses and means of reducing them. It also provides a review of parametric methodologies (COLS and SFA) and their use in testing the hypothesis of the constant returns to scale of the employed technology. The study concludes with policy recommendations for improvements in the productive efficiency of retail chain operations.

Full Text: “Essays on Efficiency Measurements” by Viliam Druska