During July and August of 2009, Americans purchased nearly 700,000 new automobiles or light trucks under the government's "cash-for-clunkers" program--officially the Car Allowance Rebate System, or CARS. Under the terms of the purchase, qualified buyers traded in their old vehicles to receive a rebate on the price of a new vehicle; the amount of the rebate--either $3,500 or $4,500--depended on both the model that was traded in and the model that was acquired. The direct cost to taxpayers of subsidizing these purchases was roughly $2.8 billion.
The program had two primary goals: to reduce energy consumption and pollution by inducing consumers to replace their old gas-guzzlers with newer, fuel-efficient models, and to serve as a stimulus for the troubled automobile industry and the U.S. economy as a whole. In this article, we take a close look at the second objective, exploring how the CARS program affected sales and, more important, production in the automobile industry. (1) Drawing on our findings, we then propose an answer to the more difficult question of the program's impact on U.S. gross domestic product (GDP). (2)
In line with other recent studies, our analysis shows that the CARS program had only a transitory cumulative effect on sales. We estimate an initial impact of about 450,000 additional automobile sales, but these were essentially shifted from the periods before and (especially) after the program. We calculate that by January 2010, the cumulative effect of the CARS program on auto sales was essentially zero. (3)
More important for our purposes, however, is the program's effect on automobile production. While most studies of CARS have focused on sales, we argue that the program's success as a stimulus to the economy hinges on its impact on production. Production increases are more likely to translate into higher GDP and employment, GDP being fundamentally a measure of production. Sales clearly rose at the outset, but given that the industry can simply let inventory stocks absorb the increase in sales, higher sales alone do not imply higher GDP or employment. (4)
Overall, we find that the program had a very modest and short-lived effect on production. Production was about 200,000 units higher during the program (in comparison to the 450,000 increase in sales). Sales from September 2009 to January 2010 were correspondingly lower than they would have been in the absence of CARS, which allowed producers to have lower production while still replenishing their inventory stocks. Thus on a quarterly basis we estimate that the CARS program shifted production by around 100,000 units from 2009Q4 and 2010Q1 to 2009Q3.
These calculations suggest that the program had a negligible direct effect on GDP, shifting less than roughly $2 billion (or less than one-tenth of 1% of GDP) into 2009Q3 from the subsequent two quarters. This contrasts starkly with a study released by the Department of Transportation (DOT) in the immediate aftermath of the program, which concluded that CARS had given a substantial boost to both GDP and employment. In the final section of the article, we discuss why our conclusions differ from those of the DOT study.
Our findings raise broader questions, beyond just this particular program, about the efficacy of efforts to stimulate short-term spending. It is clearly a mistake to gauge their impact by just looking at the spending they induce, without looking at the supply side. It is reasonable to expect that producers will use other margins than production and employment to respond to temporary movements in sales, so the only question is quantitative: How large is the gap between the sales and production impact? We find it to be large in the case of "Cash-for-Clunkers."
In what follows, we will lay out the economics behind our analysis by examining how rebates and other transitory price changes affect consumer purchases and manufacturers' inventory decisions. …