Intertemporal Preferences and the Adoption Decision for Bluetooth Speakers

The adoption decision for durable goods is intertemporal by definition. However, estimating utility and discount functions from revealed preference data using dynamic discrete choice models is difficult because of an inherent identification problem. To overcome this issue, we use stated preference data. Specifically, we employ the experimental design of Dubé, Hitsch, and Jindal (2014), where future prices are known and that elicits intertemporal adoption decisions for Bluetooth speakers in a discrete choice framework. We estimate several models of discounting (e.g., static, myopic, geometric, and quasi-hyperbolic) and find considerably lower discount factors than typical market interest rates would suggest. The values are also smaller compared to respondents' matching-based discount factors, even though the correlation is positive and significant. Furthermore, there are substantial differences in discounting across respondents (i.e., heterogeneity in time-preferences) and lastly, there is no strong empirical evidence for quasi-hyperbolic discounting. Thus, the standard economic model seems to be appropriate for the data at hand.