Nonparametric Estimation of Demand with Switching Costs: the Case of Habitual Brand Loyalty

We study habitual brand loyalty, one of the earliest empirically-studied forms of switching costs and a classic source of structural state-dependence in consumer demand. Auxiliary instruments and economically-motivated restrictions can tighten nonparametric bounds on the extent of brand loyalty in choice panel data. We also prove that the canonical dynamic discrete-choice model, nested in our nonparametric framework, has “built-in” exclusion restrictions that semiparametrically identify the discount factor, in general, and point identify it for standard parameterizations of switching costs. Case studies of several large consumer goods categories show that brand loyalty accounts for at least 10.8% but no more than 72.2% of the observed choices across categories studied. In some categories, it accounts for over 90% of observed repeat-purchase behavior. Consumers are found to be forward-looking, but more impatient than would be implied by the real rate of interest.