How Much Did Cash-for-Clunkers Boost Auto Sales?

The busy folks at the Council of Economic Advisers (CEA) released a quartet of studies today, covering the economic impacts of:

I suspect that other bloggers (not to mention the regular media) will have lots to say on the stimulus analyses, so I started my reading with the clunkers piece, which I found quite interesting.

News accounts often describe the program as a success because almost 700,000 people participated in it in just a few weeks. But, as CEA emphasizes in their new study, the fact that someone participated in the program does not necessarily mean that they bought a car because of it. Indeed, CEA estimates that the 690,000 auto sales under the program boosted 2009 auto sales by only 330,000:


What about the other 360,000?

  • Some would have happened anyway. People trade in clunkers and buy new cars all the time. CEA estimates that about 200,000 such new car sales would have happened in July and August without the program.
  • Some people held off on purchases in order to participate in the program. If you were going to trade in a clunker in June, the creation of the Cash-for-Clunkers program gave you a big incentive to delay your purchase until July or August. CEA estimates that 50,000 sales were delayed for that reason.
  • Some people accelerated purchases they were going to make later in the year. If you were planning to buy a new car in the fall, for example, you might have decided to go to the showroom a few months early in order to get in on this deal. CEA estimates that 90,000 sales were shifted forward from September, and another 20,000 were shifted from October through December (so the total forward shift in 2009 was 110,000).

Putting those factors together, CEA estimates the net effect of the program was to lift 2009 auto sales by 330,000, less than half of total participation in the program.

The Cash-for-Clunkers program cost about $2.88 billion, so one way to slice the numbers is to say that each additional auto sale cost taxpayers about $8,700 (=$2.88 billion divided by 330,000).

Was that money well spent? That’s a hard question I leave for another day. A careful answer would require analyzing a host of factors including the short-term boost to GDP from the program (which CEA analyzes), the longer-term reduction in GDP from the program (due to lower auto sales over the next few years), the cost of paying off the $2.88 billion borrowed to finance the program, the environmental benefits of getting newer cars on the road, and the wealth loss from euthanizing the clunkers rather than moving them into the used car market.

Note: CEA is careful to note that there are lots of uncertainties around each of its baseline estimates. It also reports a more pessimistic estimate (only 210,000 net new sales in 2009) and a more optimistic one (560,000).

10 thoughts on “How Much Did Cash-for-Clunkers Boost Auto Sales?”

  1. Very interesting post. I think it is pretty hard to see how $8,700 per car is a good deal for the taxpayers when the average MSRP for the top 10 models was around $16,000 and less than 40% were domestic nameplates.

    1. Yup, it’s going to be a heavy lift. If I have time, I want to see whether there’s any reason to believe that the hoped-for environmental gains might materially affect the cost-benefit analysis. Suggested links from readers are most welcome.

  2. I think this is a great example of opportunity cost. What else could that $2.88 billion be spent toward? Was that marginal cost worth the marginal benefit of 330,000 cars?

    I’ll let the economists decide those questions…

    1. Thanks Ann, that has some useful thoughts about some of the micro effects (value of the clunkers, distortion to the market, and the potential enviro benefits). They assume that all 700,000 cars are incremental. As you note on your related FB post, the fact that some cars weren’t incremental probably makes the program look worse in their framework.

      To be fair, though, I think they are too quick to dismiss the potential macro benefits of the program (at least in the short-run). The theory, at least, is that the program puts some idle resources to use, which is beneficial. It will be hard, though, for that benefit to overcome the other costs.

  3. the part that bothers me is that it’s not clear the program reduces the number of clunkers on the road. the average age of cars traded in and thus destroyed was around 10 years old. As Ann knows very well, cars often last 30 years. So by destroying 10 year old cars instead of reselling them, you raise the price of 10 year old cars making it less likely for someone with (far dirtier) 20 year old car from trading their car in for a 10 year old one.

    this one time editorial annoyed me talking about how the CARS program helps working class Americans, because it in fact could hurt all those who normally buy used cars, by making used cars more expensive. in favor of the relatively better off americans who buy new cars. (rough googling shows that only 20% of car purchases each year are new cars, most people buying a car, buy used)

    1. Hi Ben — If nothing else, this program provides a fascinating opportunity to do incidence analysis. I may have to find some way to use it in my teaching. As you say, one of the interesting channels is through the used car market, where some potential buyers will discover, I suspect, that prices are higher.

Comments are closed.

%d bloggers like this: