A note on competing merger simulation models in antitrust cases: can the best be identified?

Advanced economic instruments like simulation models are enjoying an increased popularity in practical antitrust. There is hope that they being quantitative predictive economic evidence can substitute for qualitative structural analysis and lead to unambiguous results. This paper demonstrates that it can be theoretically impossible to identify the most appropriate simulation model for any given merger proposal. Due to the inevitable necessity to reduce real-world complexity and multi-parameter character of merger cases, the comparative fit of proposed merger simulation models with mutually incompatible predictions can be the same. This is valid even if an ideal antitrust procedure is assumed. This insight is important regarding two aspects. First, the scope for partisan economic evidence cannot be completely eroded in merger control. Second, simulation cannot eliminate or substitute for qualitative reasoning and economically informed common sense.