Competitive pricing behavior in the auto market: A structural analysis

In a competitive marketplace, the effectiveness of any element of the marke ting mix is determined not only by its absolute value, but also by its rela tive value with respect to the competition. For example, the effectiveness of a price cut in increasing demand is critically related to competitors' r eaction to the price change. Managers therefore need to know the nature of competitive interactions among firms. In this paper, we take a theory-driven empirical approach to gain a deeper understanding of the competitive pricing behavior in the U.S. auto market. The ability-motivation paradigm posits that a firm needs both the ability a nd the motivation to succeed in implementing a strategy (Boulding and Stael in 1995). We use arguments from the game-theoretic literature to understand firm motivation and abilities in different segments of the auto market. We then combine these insights from the game-theoretic literature and the abi lity-motivation paradigm to develop hypotheses about competition in differe nt segments of the U.S. auto market. To test our hypotheses of competitive behavior, we estimate a structural model that disentangles the competition effect from the demand and cost effects on prices. The theory of repeated games predicts that firms with a long-run profitabil ity objective will try to sustain cooperative pricing behavior as a stable equilibrium when conditions permit. For example, markets with high concentr ation and stable market environments are favorable for sustaining cooperati ve behavior and therefore provide firms with the ability to cooperate. The theory of switching costs suggests that in markets in which a firm's curren t customers tend to be loyal, firms have a motivation to compete very aggre ssively for new customers, recognizing the positive benefits of loyalty fro m the customer base in the long run. As consumer loyalty in the market incr eases, the gains from increasing market share by means of aggressive compet itive behavior are more than offset by losses in profit margins. Firms ther efore have the motivation to price cooperatively. Empirically, we find aggressive behavior in the minicompact and subcompact segments, cooperative behavior in the compact and midsize segments, and Ber trand behavior in the full-size segment. These findings are consistent with our theory-based hypotheses about competition in different segments. In estimating a structural model of the auto market, we address several met hodological issues. A particular difficulty is the large number of car mode ls in the U.S. auto market. Existing studies have inferred competitive beha vior only in markets with two to four products. They also use relatively si mple functional forms of demand to facilitate easy estimation. Functional f orms of demand, however, impose structure on cross-elasticities between pro ducts. Such structure, when inappropriate, can bias the estimates of compet itive interaction. We therefore use the random coefficients logit demand mo del to allow flexibility in cross-elasticities. We also use recent advances in New Empirical Industrial Organization (NEIO) to extend structural estim ation of competitive behavior to markets with a large number of products. W e use the simulation-based estimation approach developed by Berry et al. (1 995) to estimate our model. A frequent criticism of the NEIO approach is that its focus on industry-spe cific studies limits the generalizability of its findings. In this study, w e retain the advantages of NEIO methods but partially address the issue of generalizability by analyzing competitive behavior in multiple segments wit hin the auto industry to see whether there is a consistent pattern that can be explained by theory. Theoretical modelers can use our results to judge the appropriateness of their models in predicting competitive outcomes for the markets that they analyze. A by-product of our analysis is that we also get estimates of demand and co st apart from competitive interactions for the market. Managers can use the se estimates to perform "what-if" analysis. They can answer questions about what prices to charge when a new product is introduced or when an existing product's characteristics are changed.