Analyzing a Clunker

Megan McCardle has pointed out that the 'successful' cash for clunkers program has resulted in a rather nasty hangover for automakers. Her argument is that:

Cash for Clunkers moved a bunch of auto sales forward, causing people who thought they might replace their car in the next year or two to rush into the showrooms.

This is a true but incomplete explanation. The Cash for Clunkers sales likely came from several sources:

1. Once the program was announced many buyers likely postponed purchases until they could take advantage of the subsidy. So sales in the period immediately prior to the start of the program were artificially lowered.
2. As Megan points out, people who intended to buy a car in the next year or so likely brought forward their purchase in order to take advantage of the subsidy. So sales in the period after the end of the program were (and will continue to be) artificially low.
3. People who were weighing up possible purchases opted to buy a car, as opposed to say new appliances or a trip to Disneyland, in order to take advantage of the subsidy. So higher car sales were balanced in part by lower sales in other parts of the economy.
4. People who might otherwise have saved or paid down debt opted to purchase a car in order to take advantage of the subsidy.

Sales resulting from explanations 1 and 2 were simply a direct wealth transfer from future taxpayers to car buyers. There wasn't any real impact on the number of cars sold, just the timing, so the car companies aren't really better off. Any stimulative effect on economy would be the result of car buyers choosing to spend rather than save their subsidy windfall.

By contrast, sales resulting from explanation 3 were rather more insidious. In addition, to the wealth transfer from future taxpayers to car buyers these sales also include a wealth transfer from non-car companies to their automotive brethren. In other words, politically favored companies got some increased sales at the expense of those with less political pull.

Overall, these sales don't represent any increase in aggregate demand. Again, any stimulative effect on economy would be the result of car buyers choosing to spend rather than save their subsidy windfall.

The subset of sales driven by explanation 4 was the most economically useful. Additional cars sold without an offsetting loss of sales in other parts of the economy actually represent an increase in aggregate demand.

So the utility of the program as a Keynesian stimulus depends on the proportion of sales driven by explanation 4 and the extent to which the windfalls enjoyed by car buyers driven by explanations 1, 2 and 3 were spent rather than saved. It is possible, even likely, that the effective stimulus was substantially less than the $3 billion plus the program cost.

Taking a broader view, this program suffers from the same problem that dogs most stimulus initiatives. A small fraction of the economy will experience direct benefits. The remainder of the economy will simply see future costs in the form of higher taxes, interest rates and inflation, and a more politicized economy. For this reason, stimulus efforts that aren't linked to future productivity improvements are, at best, a questionable idea.