Self-selection as a problem in evaluating agrienvironmental programs

Since participation in agri-environmental programmes is voluntary, associated with opportunity costs (most of these programmes aim at reducing output), and financial compensation for participation is typically a fixed amount per hectare, a tendency is likely to arise for some sort of self selection: farms in favourable regions which face high opportunity costs from foregone output are less likely to participate than other, more disadvantaged farms. When trying to estimate the effects of such programmes on agricultural output, this self-selection bias can pose severe problems, as a direct comparison of participants with non-participants is likely to lead to erroneous results. This paper tries to highlight the effects of this self-selection process in the context of the Austrian OEPUL programme and its effects on grain yields. After identifying various sources of this problem, some possible remedies are presented and discussed. The empirical analysis is based on farm-level accounting data linked with the official agricultural support data (INVEKOS) for a sample of 1327 Austrian grain farms. Two years of data before (1993 and 1994) and one year (1997) after EU accession are utilized to estimate programme effects on grain yields.