Critical Review and State-of-the-Art of PMP Models: An Application to Greek Arable Agriculture

Mathematical programming models have been extensively used for assessing the impacts of agricultural policy especially in case of radical shifts. Although mathematical programming (MP) was first introduced as a farm management tool, regional bottom-up models have been implemented as an instrument of policy analysis. Regional models include classical normative mathematical programming models (with linear or non-linear objective function) and Positive Mathematical Programming (PMP) ones. PMP was first introduced in 1995 to calibrate LP models and relies upon the assumption that the economic agent’s observed behavior is in fact optimal. The aim of this paper is to compare two different PMP methods on their ability to predict crop mix changes due to the decoupling and crosscompliance measures. For this purpose, a sample of 70 farms from Thessaly is used by means of a survey. Both models are validated against base year observations (2005) and predictions of farmers’ reactions for the following year (new CAP implementation in Greece) are compared with the actual observations (2006), revealing the strengths and the weaknesses of the examined methodologies.