P1-1: Scale problems in productivity analysis in agriculture over space and time

PhD student: Daniel Mauricio Castro Medina
Thesis Committee: Prof. Bernhard Brümmer, Prof. Stephan v. Cramon-Taubadel, Prof. Thomas Kneib
Graduation Date: 2/2015

Thesis: online publication

At the scale of the individual farm, persistent technical inefficiency is usually viewed as indicative of knowledge problems of the farmer, while lack of technical progress at the sectoral scale is usually seen as a defect in the agricultural research and development process at the national or supra-national level. Consistent aggregation of productivity growth from farm scale to sector scale is only possible under restrictive assumptions. Among the components of overall productivity growth at the farm level, technical change (TC, i.e., an expansion or reduction of production possibilities) and technical efficiency change (TEC, i.e., differences in the extent to which production possibilities are fully utilized) are of particular importance.

The generic aggregation problem is exacerbated by the temporal and spatial dimension of agricultural production (e.g., time lag between input decisions and harvest, or factor rigidities). Farmers’ decision to adopt a new technology is conditional on the information available and their capacity to evaluate this information, which in turn heavily depends on local networks. Both the decision to adopt (impact on TC) and the success in fully utilizing new technologies (impact on TEC) are likely spatially dependent. Furthermore, the general role of learning-by-doing in agriculture has long been recognized. The aggregation from the individual to the sectoral scale should take these temporal and spatial dissemination processes into account.

A different scale problem arises when the perspective is reversed by looking at the impact of international trade on productivity growth. The consequences of trade liberalization on the livelihood of the rural poor, most of which are farmers, involve disaggregating the developments at the sectoral scale back to the farm level, where structural change in particular plays a crucial role.

Specific topics:

  • Aggregation of efficiency and productivity measures over space, time, and farms: The role of learning-by-doing and agglomeration effects
  • Interdependencies between agricultural trade and the development of sectoral productivity: The impact on rural livelihoods
  • Spatial spill-overs of investments between farms and down-stream processors and their impact on productive efficiency: Farm-farm versus farm-processor linkages


Here, high-quality farm level data are a precondition to tackle the discussed scale problems in productivity analysis. Whilst in the first two subjects, emphasis will be on regression-base decomposed error models (stochastic frontier analysis, and its spatio-temporal extension), the third one uses marked point processes.

back to overview of all projects