Alessandro Varacca (Università Cattolica del Sacro Cuore): “Causal Machine Learning in Agricultural Economics: Concepts and Applications."

Abstract: Causal machine learning (ML) is gaining attention as a useful extension to the general causal inference framework, especially when evaluating highly complex and potentially heterogeneous responses to a treatment or policy. Although ML methods can be very useful for working with large datasets characterizing non-linear data generating processes with many interacting variables, their application to program evaluation problems requires appropriate identifying assumptions and targeted adjustments. This means that off-the-shelf ML algorithms, which are designed for predictive purposes, may only be one of several components in the causal ML toolbox. This seminar will cover the nuts and bolts of causal ML, including all necessary identification requirements. The goal is to provide a general understanding of how these methods can be used in empirical analyses. Additionally, recent contributions to the evaluation of EU agri-environmental policies will be discussed, along with potential future directions.