| This thesis examines the efficacy of the simulation model of Werden and Froeb (1994) for predicting the price effects of horizontal mergers. The approach incorporates a logit demand specification within a multi-product price-setting game. Based on the observed market shares and prices, costs are estimated and then used to predict the post-merger equilibrium in prices. I apply this approach to the case of the 1986 merger of Northwest Airlines and Republic Airlines. The simulations predict that fares would rise by 0.80% on unconcentrated routes and by 9.01% on concentrated routes. To compare these predictions with what actually transpired, the actual fare changes are estimated using the method of Kim and Singal (1993). This method entails comparing the change in fares on routes affected by the merger (that is, routes involving Republic and/or Northwest) with comparable routes not impacted by the merger. Using this method, actual fares went down by 2.38% on unconcentrated routes and went up by 5.08% on concentrated routes. Predicted fare changes were then between 3% and 4% higher than actual fare changes.; A parameter in the logit demand specification, denoted beta, influences the elasticity of firm demand and is central to the simulations approach. The problem is that existing approaches arbitrarily specify a value for beta. In the second essay, a method is developed and implemented for estimating beta using pre- and post-merger data on unconcentrated routes. Assuming the estimated cost pass-through rate of Breuckner and Spiller (1994), the estimated value of beta is quite close to what has been used for the airline industry and what I use in the first essay. Using the estimated value would at most result in a slight decrease in predicted fares on routes with more than two pre-merger carriers and a slight increase in predicted fares on routes with two pre-merger airline carriers. |
