Working Paper No. 12-10
Theory and Evidence of Switching Costs in the Market for College Textbooks
October 2012; revised May 2013
This paper develops and estimates a model of switching costs in the market for college textbooks. First, in a theoretical setting, this paper characterizes the professor's adoption decision which includes a trade-off between time and course quality. The professor faces a time cost when he switches textbooks. This switching cost leads to state dependence and adoptions of textbooks that are sub-optimal for students. In a two-period life cycle model, switching costs are shown to lead to higher prices and shorter revision cycles. Predictions of the theoretical model are tested empirically using a unique 30-semester history of professors' textbook adoptions. Using rookie professors as a counterfactual, switching costs are identified to be large and significant in a random utility model that allows for observed and unobserved professor preference heterogeneity. Results show both book-specific and edition-specific switching costs affect adoption decisions in directions and magnitudes consistent with the theoretical model. A final empirical analysis identifies heterogeneous switching costs cross textbooks and suggests switching costs tend to dampen competition, leading to increased textbook prices. In my sample, switching costs account for about 9% of textbook prices; however, there is no evidence of correlation between switching costs and revision cycles.
JEL classification: L14, D23, I21
Keywords: Switching costs, Mixed Logit, Random Utility Model, Textbooks, Higher Education