CCP Estimation of Dynamic Discrete Choice Demand Models with Segment Level Data and Continuous Unobserved Heterogeneity: Rethinking EV Subsidies vs. Infrastructure

Abstract

When multiple groups of consumers reside in the same market, we determine that we can write each group’s conditional choice probabilities (CCPs) as a function of unobserved consumer heterogeneity. Moreover, we can specify choice probabilities of one group as a function of another by shifting the unobserved component. Armed with our novel CCP estimator, we develop an approach to identify and estimate a dynamic discrete demand model for durable goods with nonrandom attrition of consumers and continuous unobserved consumer heterogeneity but without the usual need for value function approximation or reducing the dimension of state space by ad hoc behavioral assumptions. We illustrate the empirical value of our method by estimating consumer demand for electric vehicles (EVs) in the state of Washington during the period of 2016–2019. We also determine the impact of a different federal tax credit based on the electric range of a car rather than the size of the battery, which was the existing policy during the data period, and we evaluate how best to seed a nascent market that presents indirect network effects to drive faster adoption. Should the government incentivize adoption through consumer tax credits or through EV infrastructure?.