Subjective Causality in Choice

Abstract: Choices based on observational data depend on beliefs about which correlations reflect causality. An agent predicts the consequence of available actions using a dataset and her subjective beliefs about causality represented by a directed acyclic graph (DAG). We identify her DAG from her random choice rule. Her choices reveal the chains of causal reasoning that she undertakes and the confounding variables she adjusts for, and these pin down her model. When her choices generate the dataset used, her behavior affects her inferences, which in turn affect her choices. We provide necessary and sufficient conditions for testing whether her behavior is compatible with such a model.