Jeff Jacoby, a conservative columnist for The Boston Globe, has a pretty thorough, well-cited, and convincing assessment of last summer's program, and the news isn't good.
First, it looks like the program didn't increase sales of new cars all that much, if at all. It just concentrated sales in July and August that would have otherwise happened a little earlier or a little later. "Of the 700,000 cars purchased during the clunkers frenzy, the estimated net increase in sales was only 125,000. Each incremental sale thus ended up costing the taxpayers a profligate $24,000."
Another problem, which wasn't discussed very much during the frenzy, has to do with the program's effect on the used car market. Because the cars people turned in had to be in drivable condition—and had to be destroyed—Cash for Clunkers ended up removing a lot of decent used cars from the market. That means that used cars are about 10 percent more expensive this summer than they were last summer. And that's no good for people who need to buy a car and can't afford to buy a new one.
But didn't the program at least make our fleet more energy efficient? Sort of, but it was absurdly expensive. "Researchers at the University of California-Davis calculated that the reduction of carbon dioxide attributable to the program cost no less than $237 per ton. In contrast, carbon emissions credits cost about $20 per ton in international markets."
And that's to say nothing of the environmental cost of producing the new cars.
This isn't necessarily an indictment of a Cash for Clunkers program in general. If we had required the new cars people were buying to be much cleaner, and required the old cars people were turning in to be much older and dirtier, the whole program might have lasted longer, burned slower, and had a better effect on emissions. As executed, however, it certainly looks like the program wasn't worth it.