The answer is: Just about everyone who was not in the tank for the Obama Administration predicted this (from my Princeton classmate Henry Payne):
When Congress gave away $3 billion for buyers to trade in their "clunkers" and buy new cars in August, lawmakers thrilled as buyers swamped showrooms to take advantage of the big discounts. "Cash for clunkers has captured the public's attention . . . (it) has the possibility to truly jumpstart our economy," said Rep. Candice Miller (R., Mich.). Other, more sober analysts, warned that the clunkers program was only stealing from future sales.
September sales are in, and sobriety can take a bow.
Edmunds.com reports that "September's light-vehicle sales rate will fall to 8.8 million units . . . the lowest rate in nearly 28 years, tying the worst demand on record. After the cash-for-clunkers program boosted August sales to their first year-over-year increase since October 2007, demand has plunged. In at least the last 33 years, the U.S. seasonally adjusted annual rate has only dropped as low as 8.8 million units once -- in December 1981 -- with records stretching back to January 1976."
The real popularity of the program was always due to the fact that the government was throwing money away and people rushed to pick it up. Edwards.com estimated the Feds purchased vehicles with average blue book values of just under $1500 for $3500 to $4500. That means that the government purchased cars that blue booked at just over a billion dollars for three billion. I you suddenly offered to buy all of your neighbors' cars for three times what they were worth, you'd be popular too. It was a $2 billion giveaway, and people rushed to pick the cash up like one of those money drops in the outfield of a minor league baseball game. In doing so, the government made a trivial change in the overall fleet fuel economy, in the process overpaying for Co2 reduction by a factor of 20.
Update: The study linked above shows the government paying over $400 per ton of Co2 reduced in the Clunkers program. The 20x factor cited was based on an estimated clearing price of a tone of Co2 in a future cap and trade system. This is hypothetical, as currently a ton of Co2 offsets trades right now in the US at 20 cents. At this price, the program overpaid by a factor of 2000. To be fair, this reflects both estimated pricing as well as a discount for the likelihood of a cap and trade bill passing.