The learning effects of subsidies to bundled goods: a semiparametric approach

Can temporary subsidies to bundles induce long-run changes in demand due to learning about the relative quality of one of its constituent goods? This paper provides theoretical and experimental evidence on the role of this mechanism. Theoretically, we introduce a model where an agent learns about the quality of an innovation on an essential good through consumption. Our results show that the contemporaneous effect of a one-off subsidy to a bundle that contains the innovation may be decomposed into a direct price effect, and an indirect learning motive, whereby an agent leverages the discount to increase the informational bequest left to her future selves. We then assess the predictions of our theory in a randomised experiment in a ridesharing platform. The experiment provided two-week discounts for car trips integrating with a train or metro station (a bundle). Given the heavy-tailed nature of our data, we follow \cite{Athey2023} and, motivated by our theory, propose a semiparametric model for treatment effects that enables the construction of more efficient estimators. We introduce a statistically efficient estimator for our model by relying on L-moments, a robust alternative to standard moments. Our estimator immediately yields a specification test for the semiparametric model; moreover, in our adopted parametrisation, it can be easily computed through generalized least squares. Our empirical results indicate that a two-week 50\% discount on car trips integrating with train/metro leads to a contemporaneous increase in the demand for integrated rides, and, consistent with our learning model, persistent changes in the mean and dispersion of nonintegrated rides. These effects persist for over four months after the discount. A simple calibration of our model shows that around 40\% to 50\% of the estimated contemporaneous increase in integrated rides may be attributed to a learning motive.