Lilia Qian

Under current standards for evaluating scientific evidence, economic models of antitrust are frequently excluded from legal consideration, but not always for reasons that make them genuinely unreliable. This paper clarifies why antitrust models face difficulties when subjected to methodological scrutiny: the employment of model-based counterfactual arguments under an epistemically defective ‘but-for’ structure of causation. A more flexible causal framework, the weak-necessity structure of causation, is suggested as a means of developing and evaluating model-based counterfactuals. The framework is demonstrated in application to causal mechanisms derived from economic theories as well as those implicit to many regression models.