Unobserved heterogeneity in productivity analysis of panel data: applications to meat chain firms and global growth in agriculture
von Jonathan Holtkamp
Datum der mündl. Prüfung:2015-02-12
Betreuer:Prof. Dr. Bernhard Brümmer
Gutachter:Prof. Dr. Bernhard Brümmer
Gutachter:Prof. Dr. Meike Wollni
EnglischProduction analysis matters in agricultural economics because technology shapes developments in agriculture as well as in related sectors, such as the food industry. Pro- ductivity and technical efficiency (TE) are key aspects in this context. In empirical analysis, inefficiency can be confounded with unmeasured attributes. As panel data are used to analyse productivity trends, it is possible to account for unobserved hetero- geneity. However, in the context of efficiency analysis, estimation of the corresponding stochastic frontier (SF) model is a challenge. Several econometric methods have been proposed. This thesis comprises applications that are of interest in agricultural economics and applies a novel fixed effects SF approach. This estimation technique allows to distinguish between inefficiency and unobserved heterogeneity. The applications address agricultural production at the firm-level as well as at the country-level. As meat production is a major activity in the European food chain, data on slaughtering and meat processing firms as well as on livestock farms are used. On a global scale, agriculture is supposed to meet a variety of challenges where technological development is one important component in improving agricultural performance. The ultimate objective is to identify trends in total factor productivity (TFP), and thereby, to better understand the developments in these sectors. Differentiation between technical inefficiency and unobserved heterogeneity is supposed to result in more reliable predictions of potential improvements. Progress of the European meat industry proves to be divergent when different sub- sectors are considered. The basic productivity of a firm is derived from the estimated effects. In case of meat processing firms, a higher basic productivity seems to be an advantage with respect to productivity growth. Productivity spillovers may exist within the European meat chain. The analysis relates TFP growth of livestock farms to TFP growth of meat firms. The sources of pro- ductivity change are of different relevance between the two sectors. Results for regional spillover effects are ambiguous. The last application provides an update of productivity analysis using FAO data on global agricultural production. Earlier studies typically neglect the panel structure of the data. Results show that actual TE scores are higher and less dispersed when accounting for unobserved heterogeneity. Consideration of the input variable feed leads to lower rates of productivity change.
Keywords: production analysis; total factor productivity; stochastic frontier analysis; technical efficiency; unobserved heterogeneity; aggregation; agriculture; livestock production; European meat industry; meat supply chain