Consumer choice and market outcomes under ambiguity in product quality

摘要

When product quality is ambiguous, consumers do not compare subjective expected utilities of each product before making a purchase. Instead, they behave in a boundedly rational way and rely on their initial preference for a product to simplify the decision process. The goal of this study is to present a boundedly rational model of consumer choice that explicitly accounts for the interaction between predisposition and ambiguity. The consumer decision process follows the hypothesis testing framework.The interaction term (ambiguity ⅹ predisposition) is the threshold that a competing product must meet to overcome a consumer's initial favorable predisposition toward the original product. A multi-attribute utility model of consumer choice is presented in which a product's valuation depends on price, quality, consumer predisposition and ambiguity present in the product-market environment. The model is a boundedly rational alternative to the Bayesian updating model where price, quality, ambiguity and predisposition are viewed as attributes. It is shown that if quality is not ambiguous or if the consumer is not predisposed toward either product, then behavior is consistent with expected utility maximization. The article also investigates the equilibrium price, profit and market share that each firm would obtain in a duopolistic setting (two vertically differentiated products(firms). In such a competitive environment, the demand curve depends on the interaction between the degree of ambiguity and predisposition. When ambiguity is high or the customer base is more partisan, price competition between firms diminishes and higher prices result. It is also evident that even when the price-quality trade-off is identical for all consumers, the lower-quality firm in vertical competition will capture a segment of consumers in the presence of ambiguity and predisposition. The article also discusses how equilibrium prices in the presence of loyal customers are derived.