Valid Inference in Single-Firm, Single-Event Studies