Switching Costs and Market Competitiveness: De-constructing the Relationship

The conventional wisdom is that switching costs raise prices and make markets less competitive. Dube, Hitsch and Rossi (2009, hereafter DHR) demonstrate a U shaped relationship between switching costs and equilibrium average prices; i.e., prices fall at low levels of switching costs and then rise as switching costs become very high. DHR show this result using a numerical solution of an infinite period game modeling stochastic consumer preferences with a logit demand model. The authors replicate the U shaped relationship between switching costs and equilibrium average prices in a two period Hotelling model with changing preferences. This analytically tractable model complements the empirical realism of the DHR model by enabling one to understand which of the features of the DHR model are essential for the U-shaped relationship between switching cost and market competitiveness. The authors conclude that the presence of preference heterogeneity (product differentiation) and changing preferences over time suggested by DHR are critical. Future research on switching cost should incorporate these features to obtain richer insights. Dube, Hitsch and Rossi (2009 hereafter DHR) question the conventional wisdom that switching costs make markets less competitive. They numerically solve for a Markov-Perfect stationary equilibrium of an infinite period game in which consumer demand is modeled using the empirically realistic logit model. They show that the average equilibrium prices have a U-shaped relationship with switching cost, with prices falling initially as switching costs rise and then rising beyond a threshold level of switching cost. We believe this is an important and novel finding because it provides a more complete description of the relationship between switching cost and prices; challenging the widely prevalent conventional wisdom that switching cost will raise prices and make markets less competitive. In arriving at this result, DHR expand on the original analytical model used by Klemperer (1987). First, they model demand using a logit model. The logit demand model allows for product differentiation as well as changing preferences over time. Second, they model an infinite period game, which is closer in spirit to Beggs and Klemperer (1992), but relax the assumption of infinite switching cost, that causes perfect lock-in of consumers to firms that they previously purchased from. The infinite horizon model allows us to characterize how firms balance the trade-off between “investing” in customer acquisition and “harvesting” the existing customer base in every period of a steady state equilibrium. In contrast, two period models artificially separate the investment motive for customer acquisition to be entirely in the first period and the harvesting motive to be entirely in the second period. Therefore, two period games are inadequate to capture steady state equilibrium behavior. Nevertheless, two period models are often used in analyzing the effect of switching cost on market prices and competition. Hence, it would be interesting to test whether the U shaped switching cost-market competition relationship identified in DHR can be replicated within the two period frame works by incorporating key features of the DHR model. By doing so, we hope to