Do rural development policy measures really affect farmers' behaviour and performance? A synthetic difference in differences estimation

This paper deals with the response of farmers to targeted policy measures. The research question consists of whether and to what extent this response, upon voluntary adoption and eventually driven by private motivations, also generates the outcome of societal interest. Challenges posed by these policies are multiple since they may admit a staggered adoption with often very few, at least initially, treated units. In order to deal with these challenges, the paper adopts a Synthetic Difference-in- Differences approach. Within this causal inference logic, an appropriate theoretical framework is elaborated to model farmers’ behavioural response to the policy distinguishing between the private and societal outcomes. This approach is applied to a balanced sample of Italian farms and to some selected measures of the second pillar of the EU Common Agricultural Policy over the period 2014-2022. Results point to the identification and estimation issues emerging when the entry into the treatment is staggered and treated units are few and heterogeneous. For some policy measures the estimated treatment effect is significant for the private outcome while it seems weaker and more volatile for the societal outcome.