ABSTRACT
Economists employ counterfactual causal claims of the form ‘if C did not happen, then E would not happen’ to establish singular (token-level) causal claims. It is useful to divide counterfactual claims into the manipulationist and Galilean counterfactuals. While the former type of counterfactuals can be used to address directly the question of what the effects of interventions are, the latter utilizes the possible-worlds semantics to describe causal dependencies among events. In such cases, causes are defined as necessary conditions of their effects. In order to establish counterfactual causal claims, economists either analyze counterfactually previously estimated econometric models or conduct case-study analysis. I argue that academic economists put forward Galilean counterfactuals. Employing the ‘causes as necessary conditions’ definition has serious policy implications. Given that effects usually result from a few (more than one) necessary conditions, policymakers willing to employ such evidence need to establish that all relevant factors present in the research setting also operate in the policy setting. Even though the manipulationist counterfactuals directly support the policies they describe, the evidence required to establish them is challenging to gather, what I exemplify with the case of the discussion of the effectiveness of Troubled Assets Relief Program.
