UCLA

In this dissertation, I empirically examine the effectiveness of financial incentives on consumer adoption of new products. Specifically, I study the impact of two subsidies, which are federal tax credits and rebates, on electric vehicle sales in New York. The first chapter studies heterogeneity in the subsidy effect on electric vehicle adoption across two types of incentives. The second chapter develops a dynamic discrete choice model for demand of electric vehicles and compares the average impact of tax credits and rebates per spending through counterfactual simulations. The third chapter discusses the policy implication of revisions in federal tax credits and studies the impact of the protectionist policy on domestic and foreign-produced EV sales.

The first chapter investigates how each type of subsidy affects consumer adoption of EVs differently and how consumers dynamically optimize their consumption decisions in this market. There are two noticeable differences between rebates and tax credits. First, while consumers receive rebates directly at the point of sale by applying the amount to reduce the purchasing price of EVs, consumers have to wait to enjoy tax credits till the end of each tax year. Also, tax credits are non-refundable, so consumers who do not have any tax liability cannot benefit from these credits. Due to these disparities, it is important to understand the effectiveness of each subsidy respectively to promote the adoption of EVs with a more efficient subsidy scheme in the future. This chapter provides descriptive evidence on the relative effectiveness of the two subsidies. The findings indicate that rebates are more effective in promoting EV sales than tax credits : for every thousand dollars, rebates lead to an 3.9 ~ 4.1% increase in new EV purchases, while tax credits increase EV sales by only 1.3%. Moreover, I find that current EV sales are associated with an expected subsidies change in the future, implying that households dynamically optimize their consumption decisions in this EV market.

The second chapter takes a structural approach to conduct a more rigorous analysis on the impact of rebates and tax credits. Based on the empirical observations in the first chapter, I specify a dynamic demand model in which consumers are assumed to discount the financial benefits of tax credits. Along with coefficients of vehicle characteristics, I estimate the discount factor as well. In this context, the variation in EV sales in response to anticipated changes in subsidies and the difference between the variation in market shares with respect to tax credits and rebates provide sources for identifying and estimating the discount factor. Using the estimated coefficients and discount factor, I run counterfactual simulations to compare the effectiveness of each subsidy. Direct rebates result in promoting 918 additional EV purchases per $10 million, while tax credits only increase EV sales by 582 per $10 million. The results suggest that direct rebates are a more effective means to promote a wider adoption of EVs. Furthermore, at alternative levels of government spending, the average impact of rebates is still significantly higher than that of tax credits. This implies that with rebates only, the government could have saved 23% of the current subsidy costs to accomplish the same level of EV adoption.

In the third chapter, I study the recent major changes in federal tax credits for electric vehicles and further examine the impact of one controversial revision on EV sales. Under the Inflation Reduction Act (IRA), there were four revisions on tax credits : the removal of the phase-out structure, the availability of tax credits at the point of sale, the MSRP requirement, and the final assembly plant requirement. I focus on the requirement of EVs to have undergone the final assembly in North America to qualify for tax credits. Using the estimated dynamic demand model in the second chapter, I quantify the impact of this requirement on domestic-produced and foreign-produced EV sales through counterfactual simulations. If EVs manufactured outside North America no longer qualify for tax credits, their sales decrease by 42.6%, while sales of EVs with final assembly in North American increase by only 1.2%. This implies that removing tax credits for foreign-produced EV models does not directly translate into consumers' adoption of domestic-produced EVs. Also, this requirement undermines the overall effectiveness of subsidies on EV adoption by 8.6%.