Entry model of multiple agents : empirical application to domestic air transport within the European Union

This paper is composed of two parts. The first part of the paper deals with an example of two firms theory game of entry. After specifying, the set of firms, the firms space of pure strategies and the profit functions, we consider different types of games with respect to the information structure and the sequence of move of the two firms. We present the impact of different assumptions on the probability distribution of the outcomes of the game. With respect to the information structure, there exist two types of games: the complete information games where information about firms entry cost is perfectly known by both firms and asymmetric information games in which firms entry cost is a private information. So every firm knows its own cost of entry and have a partial information about its opponent entry cost (opponent?s cumulative distribution function). Firms can participate either in a simultaneous move game or in a sequential move game. The combination of the nature of the structure of the information with the rule of move, gives arise to four games whose outcomes are different. We show that in some cases multiplicity of equilibria exists. The second part of the paper begins by showing how previous literatures have treated the problem of multiplicity of equilibria. We conclude that to come over this problem and to guarantee the uniqueness, like in einav(2003), one should use a sequential move asymmetric information game. Within the framework of domestic air transport within the European union, the sequential move doesn?t appear to be a realistic assumption and an appropriate empirical model of oligopoly market structure must estimate simultaneously the decision of all companies. Due to the fact that the simultaneous move asymmetric information entry game becomes computationally intractable for more than 2 firms, we decide to opt for the simultaneous move complete information entry game. To deal with the problem of multiplicity of equilibria we adopt the same approach as berry (1992). In our empirical model the firms profits are specified as a function of the demand characteristics of the market, as well as the equilibrium number of companies present in the market. Like in berry (1992) the model specification supports both types of heterogeneity: observable and unobservable. Four models estimations are performed: three of them stem from specific constraints on parameters and the last one is the most general one. A comparison between models is built to find the best fit data model.