Measuring the Impact of Promotions on Brand Switching Under Rational Consumer Behavior

Abstract

Logit choice models have been used extensively to study consumer choice behavior and promotion response. A common finding is that promotion has a strong effect on brand choice. This paper examines whether brand switching elasticities derived from these models may be over-estimated due to rational consumer adjustment of purchase timing to coincide with promotion schedules, and whether this bias can be addressed by a dynamic structural model. We develop a dynamic structural model of choice/incidence that traces the process by which consumers make optimal buying decisions. We then conduct three analyses. First is a simulation based on synthetic data. We show that if the structural model is correct, brand switching elasticities are over-estimated by stand-alone logit choice models. Adding a nested logit model of choice and incidence improves the estimates, but not completely. Second we estimate these models on real data. The results indicate that the structural model fits better and produces sensible coefficient estimates. We then observe the same pattern in choice elasticities as we did with the simulation. Third, we predict sales for a major change in promotion policy - a 50% increase in promotion frequency. The reduced form models predict much higher sales levels than the dynamic structural model. This is a prime illustration of the Lucas Critique. Our overall conclusion is that reduced form model estimates of brand switching elasticities can be overstated, and that a dynamic structural model is best for addressing the problem. Reduced form models that try to capture the same phenomena as the dynamic model, especially if they model incidence, can partially, although not completely, address the issue. We discuss the implications of these findings for researchers and managers.