Modelling Structural State Dependency in Agri-Environmental Schemes

Using discrete time duration models we examine the determinants of farmers entering and exiting the Rural Environmental Protection Scheme. We also extend our models to incorporate farm level heterogeneity (or “frailty”). We introduce this dynamic element into the models by using the random effects logit model estimator, with a set of time (year or REPS period) dummies included as explanatory variables. In the analysis of dynamic behavioral processes it is crucial to identify farm behavioral relations (including the corresponding random error terms), versus properties of the distribution of preferences and choice constraints across the population of farms in our data set. This identification allows us to establish a modeling framework that can be applied for analysing structural state dependence effects.