I blogged a few days ago about the Edmunds.com analysis of the Cash for Clunkers program, which concluded that the program was an expensive bust, costing the government roughly $24,000 for each extra car sold.
Now a group called NADA is weighing in on the subject. In a press release issued by NADA, they quote their chief economist Paul Taylor as saying the following:
It’s really not that hard to determine a credible cost estimate for the Clunkers program. You subtract projected sales from actual sales for July and August when the Clunkers program was operating, and divide the program’s $3 billion by that number.
When you do that calculation, you come up with a cost to the government per extra car sold of $4,587.
If you are in the mood for a little test of your economic acumen, re-read that quote above and see if you can figure out why it is completely and utterly wrong.
That quote makes no sense economically, of course, because people have control over the timing of when they make their purchases. If you tell people the price of anything is going to go way up in the future (because the Clunkers program is ending), they will make their purchase earlier. This is especially true with automobiles, which are durable goods. It would be less true, of course, with something perishable like a meal at a fast-food restaurant.
With or without a Ph.D. in economics, it should be obvious that the wrong way to judge the success of the Clunkers program is without factoring in shifting of the timing of purchases.
When I see a mistake as egregious as this, I usually suspect it is more likely the result of someone trying to intentionally deceive the public rather than an error of logic. So the first thing I do is try to figure out the incentives of the group that is making the statement.
In this case, I was not surprised to find out that the initials NADA stand for National Automobile Dealers Association. One of their main purposes: to represent auto dealers on Capitol Hill. Their incentive: to say Cash for Clunkers worked so that the program is renewed and more government funds are funneled to auto dealers.
I know it is Paul Taylor’s job to figure out ways to make it seem like Cash for Clunkers was a success, but it is bad for the field of economics when people calling themselves economists make ridiculous, erroneous statements like this one.
If nothing else, getting an economics Ph.D. should teach someone how to complicate and obfuscate the issue so that it isn’t so obvious to outsiders that the argument makes no sense.
(Hat tip: David Cushman)
and 
. You can also watch it on 
I have a very naive question. Accepting the Edmunds figure of $24,000 per car, and assuming that the dealer got $3,500 or $4,500 per car…where did they other $20,500 or $19,500 go?
I’m sure there’s a simple answer, but my poor brain isn’t coming up with it.
Does Mr. Levitt agree with Edmunds? The article poses good questions but without an independent analysis which would include an estimate of the important stimulative effects of this program during the depths of an unprecedented recession, there is little to learn from this article.
Given the 4,500 dollars maximum government contribution per clunker turned in, the NADA estimate looks kind of straight-forward to me.
Jim
neither measure is accurate – one is too optimistic the other too pessimistic – and each has the same flaw – fails to account for changed behavior
- in one case they fail to account for the cheap price simply shifting the time of purchase
- in the other the they simply fail to account for the possibility the the demand would have been substantially lower than in other years
“If nothing else, getting an economics Ph.D. should teach someone how to complicate and obfuscate the issue so that it isn’t so obvious to outsiders that the argument makes no sense.”
I thought that’s what any Ph.D. program teaches…
I love the last sentence of this piece.
I also love what people do with statistics(not). One can really do anything with them. For example, there is a statistic going around that says medical malpractice reform will only very slightly help medical costs. Of course, the statistic doesn’t take into account the cost of defensive medicine.
It should be illegal to make such a mess of mathematics.
@Prof. Levitt,
I wish you would study the entitlements giving rise to the single mothers. EX. in my county**, 50% of all babies are born to single mothers. Women with an income of less than $35,000 are eligible for Medicaid benefits including maternity benefits. Our county ave. household income is $37,000. What women would chose marriage?
The plagues of entitlements are reaching up into the middle class.
**Winnebago County, IL
ps.. my daughter enjoyed your class
And it’s a lot more complicated than you make it as well.
While there may certainly be some consumers who, fueled by the fact that the clunkers program was ending, bought cars earlier. But that number could be very (VERY) small.
Another fact — and one that is far more important in buying decisions — is consumer confidence and the willingness to spend. As we all know, one huge factor in the failure of the economy to right itself is the reticence of consumers to spend. And spening on big ticket items like cars is a huge leap of faith.
People made this leap because the program was attractive to them. Is that bad? Hardly. It gave the economy a temporary but very important boost. People who can only finf fault with programs such as these never have their own fix for the economy; they are just narcissists who want us to think they know better.
Dr. Levitt: What is the right number and how would you calculate it?