Citation Context

...treams are discounted sufficiently slowly, then a firm would reduce the discounted value of its returns by failing to collude. Therefore, for all firms to adopt the collusive strategy would be a noncooperative equilibrium. The trouble with these formulations, from an applied industrial organization viewpoint, is that incentives in these equilibria are so perfect that the deterrent mechanisms are never observed. Then it may be difficult to infer from econometric time-series evidence whether the observed market data is the outcome of a quasicompetitive or collusive equilibrium (cf. T. Bresnahan [2]). The substance of the present contribution is that this perfection is an artifact of the certainty world in which these models are formulated. When the considerations of imperfect information, which played a decisive role in Stigler's theory, are reintroduced, optimal incentive structures may involve episodic recourse to the kind of shortrun unprofitable conduct which would have been characterized as "price wars" or "punishment" previously. Our argument has three parts. First, we frame a precise definition of collusion in terms of industry conduct. Second, we show that collusive conduct may,...

Citation Context

...to the definition offered in this section, and that evidence that this particular Nash equilibrium occurs in a specific industry is not the only evidence relevant to forming an opinion about the extent of collusion in various sectors of the economy. However, even though the direct applicability of our model is severely limited, it would be valuable to examine an industry for which it would be appropriate. We believe that the American rail freight industry in the 1880's was one example of an industry which satisfies our structural conditions quite well. Studies of that industry by Paul MacAvoy [7] and Thomas Ulen [16, 17] have produced qualitative conclusions which are consistent with our model. Recent econometric work by Porter [10] (based on the extensive time series data collected by Ulen) strengthens these conclusions. 3. PRICE PROCESSES GENERATED BY COLLUSION The equilibrium discussed in the preceding section is noteworthy because it reverses the traditional interpretations of a certain kind of industry price pattern. According to these traditional interpretations, an episode in which price drops sharply, remains low for some time, and then sharply rises again without there being ...

Citation Context

...fered in this section, and that evidence that this particular Nash equilibrium occurs in a specific industry is not the only evidence relevant to forming an opinion about the extent of collusion in various sectors of the economy. However, even though the direct applicability of our model is severely limited, it would be valuable to examine an industry for which it would be appropriate. We believe that the American rail freight industry in the 1880's was one example of an industry which satisfies our structural conditions quite well. Studies of that industry by Paul MacAvoy [7] and Thomas Ulen [16, 17] have produced qualitative conclusions which are consistent with our model. Recent econometric work by Porter [10] (based on the extensive time series data collected by Ulen) strengthens these conclusions. 3. PRICE PROCESSES GENERATED BY COLLUSION The equilibrium discussed in the preceding section is noteworthy because it reverses the traditional interpretations of a certain kind of industry price pattern. According to these traditional interpretations, an episode in which price drops sharply, remains low for some time, and then sharply rises again without there being an apparent cost or deman...