In many theoretical models, it is assumed that there are costs associated with adjusting prices. (In this paper, the existence, type, and magnitude of such costs are investigated empirically. A discrete-choice dynamic-programming model that nests fixed and variable costs is developed, and econometric estimates of the adjustment-cost function are obtained. The data used to implement the model consist of weekly retail prices and sales of three brands of saltine crackers sold by four chains of grocery stores in a small U.S. town. The study has several distinguishing characteristics. First, the data are highly disaggregate (at the level of brand and store). Second, variables are sampled at weekly intervals. Third, the model nests fixed- and variable-adjustment costs and can therefore determine their relative importance. Finally, the discrete-choice model is both structural and dynamic.

This paper uses a unique data set to determine which dynamic model of tacit collusion best describes behaviour in a particular industry. The area investigated is a region of the Vancouver, British Columbia retail-gasoline market. Players are service-station managers who compete daily. Firms choose price in each period using strategies that depend on prices chosen in the previous period. Periodically, unanticipated demand shocks precipitate price wars. When shocks occur, the firms in the market must determine the new demand conditions and adjust their strategies. From an econometric point of view, slopes of intertemporal reaction functions are latent variables. The resulting system of equations with time-varying parameters is estimated via the Kalman filter. Different repeated-game oligopoly models correspond to different transition matrices for the latent variables. The models can thus be assessed in terms of their power to explain firm behaviour in this market.

We investigate the nature of price competition among firms that produce differentiated products and compete in markets that are limited in extent. We propose an instrumental variables series estimator for the matrix of cross price response coefficients, demonstrate that our estimator is consistent, and derive its asymptotic distribution. Our semiparametric approach allows us to discriminate among models of global competition, in which all products compete with all others, and local competition, in which products compete only with their neighbors. We apply our semiparametric estimator to data from U.S. wholesale gasoline markets and find that, in this market, competition is highly localized.