A Bayesian approach to dynamic efficiency and productivity measurement
by Ioannis Skevas
Date of Examination:2017-02-06
Date of issue:2017-02-17
Advisor:Prof. Dr. Bernhard Brümmer
Referee:Prof. Dr. Thomas Kneib
Referee:Dr. Grigorios Emvalomatis
Files in this item
Name:PhD_dissertation.pdf
Size:1.25Mb
Format:PDF
Abstract
English
The vast majority of the efficiency and productivity measurement literature has been based on the static viewpoint of the firm. Few studies have developed the dynamic analog of static efficiency measurement, introducing the notions of long-run efficiency and inefficiency persistence. However, these few existing dynamic efficiency studies have not provided any empirical evidence on the driving forces of firms' long-run efficiency and inefficiency persistence. Furthermore, calculation of Total Factor Productivity growth has predominantly been based on static efficiency specifications. This dissertation departs from previous dynamic efficiency studies by shedding light on the drivers of long-run efficiency and inefficiency persistence. Additionally, a dynamic efficiency specification is used to calculate and decompose Total Factor Productivity growth. Three models for dynamic efficiency measurement are developed and applied to the case of German dairy farms. Estimation proceeds using Bayesian techniques. The first model is based on the argument that the efficiency levels achieved by farms in the long-run may vary due to their different characteristics and the varying degrees of their inefficiency persistence. The conventional dynamic stochastic frontier model is extended to allow for such heterogeneity in the long-run efficiency of farms. The results suggest that farms achieve different long-run efficiency levels mainly due to their different characteristics. The second model aims to provide empirical evidence on the driving forces of farms' inefficiency persistence. By assuming that inefficiency persistence arises due to the existence of pecuniary and non-pecuniary adjustment costs, the model tests whether financial aid and managers' experience have an impact on inefficiency persistence. The empirical findings reveal a high degree of inefficiency persistence through time, which increases with the amount of subsidies received. Older farmers exhibit higher inefficiency persistence as opposed to younger ones, presumably due to their lack of motivation to adopt state-of-the-art technologies. The third model calculates and decomposes total factor productivity growth of German dairy farms using a dynamic specification for the efficiency component. The results report a high total factor productivity growth rate that is mostly attributed to technical progress. The model is also able to capture steep efficiency and total factor productivity growth changes that may have been induced by the high milk price volatility which occurred in the German dairy sector. The dynamic efficiency model is favored by the data when tested against a model that imposes a very restrictive time structure on efficiency, and a model that imposes no time structure on efficiency scores.
Keywords: Bayesian techniques; long-run efficiency; inefficiency persistence; Total Factor Productivity growth; Stochastic Frontier Analysis; German dairy farms