📝 SummarXiv

Abstract

We bridge quasi-experimental and structural approaches to develop a robust method for retrospective merger evaluation. First, we show that the standard difference-in-differences estimation equation can be interpreted as a "reduced form" of our structural equations, formally $\textit{bridging}$ these methodologies. Demand and cost parameters enable the construction of a robust measure of the merger price effect that remains identified even when the difference-in-differences approach is not, and can be decomposed into efficiency gains versus conduct changes among the remaining firms. Second, we enhance the structural approach by weighting each market's contribution to the moment conditions using market- and time-specific synthetic weights when the number of treated markets is limited. Applying this methodology to three major airline mergers, we find that modest efficiency gains are entirely offset by increased coordination, particularly in later mergers. The $\textit{synthetic GMM}$ refinement sharpens these findings, uncovering anti-competitive effects that standard approaches miss.