P3-8: Partial and Total Factor Productivity at Different Scales: Measurement, Decomposition, and Implications

PhD Student: Yashree Mehta
Supervisors: Prof. Dr. Bernhard Brümmer
Department: Agricultural Economics and Rural Development

Project Description:
Scaling problems in efficiency and productivity analysis are present in many research questions of interest in agricultural economics: Aggregation problems, complex, often hierarchical data structures, and increasingly complex value chains for agricultural and food products. The last two aspects are exemplary fields of research for the position to be filled.

Many farm level datasets exhibit a hierarchical structure: The highest spatial resolution is attained at the plot scale, often combined with very detailed information on specific labor tasks, and the use of fertilizer, seeds, etc.,. This detailed information is often not available for all plots of the whole farm so that at the scale of the individual farm, simple aggregation methods would lead to some information loss. The farm scale is usually the level of decision-making so that efficiency and productivity studies should usually take this scale as the observational unit of interest. Higher spatial scales are often of additional interest, e.g., if sector-wide results for policy analysis and recommendations are relevant. Analyses of structural efficiency are important in this context.

An increasingly important characteristic of modern economies is the role of interconnected production activities along many stages of a value
chain (each of which, in turn, might exhibit multiple linkages, so that the term value web might be more appropriate). Food products are not exempt from this development, which is visible both in international trade (global value chains) and in domestic food processing (local value chains). E.g., at the global scale, multinational corporations respond to this in some cases by vertical integration, in other cases by reliance on certification schemes. From the perspective of the individual farmer (at the local scale), the overall efficiency and competitiveness of local downstream processors is of particular importance because ultimately, his/her profitability will not only depend on the productivity in farming (which is under his/her direct control) but also on the productivity in downstream firms (which is clearly not under control of the farmer). Thus, profitability at the sectoral level will crucially depend on the productivity in each stage of the value chain.

For both these issues, methods based on Bayesian estimation of hierarchical models are promising candidates, and could be further explored in this project.