Yesterday I tipped The Audit’s cap to a nifty bit of analysis from Calculated Risk pointing out that the real cost of the Cash for Clunkers and first-time-homebuyers subsidy programs were far higher than what’s been reported.

But Michael S. Cullen in comments and reader Doug Smith in an email both point out that CR’s analysis is only of the cost side not the benefit side, which is true, though it in no way takes away from CR, who explicitly said he was only tallying costs, not doing a cost-benefit analysis.

So what about that upside?

Daniel Gross at Slate took a stab at that last week:

If we use Taylor’s estimate, about 250,000 extra cars were purchased (40 percent of 625,000). And if each cost $29,000, those sales generated about $7.3 billion in revenue in the space of a few weeks. That’s a pretty good return on $2.6 billion in government spending. Let’s be more conservative. Say only 20 percent of the clunker traders were extra demand, and the cars they bought cost $25,000 each. That’s still an extra $3.125 billion in sales for dealers. What’s more, the sales represent only a portion of the economic impact. Ford, for example, announced that it is increasing production of some models.

But CR estimates 320,000 additional sales in August alone— so let’s do our own math with those numbers.

If the average sales price for those cars was $26,300 (that’s what Comerica Bank said the average sales price for a new car was in the second quarter), that gives you $8.42 billion in additional sales in August alone. That’s almost triple the $2.88 billion the government spent on the whole program.

But as Smith points out, there’s a multiplier effect on economic activity. Those sales filter through the economy in the form of increased bonuses for salesmen, saved jobs at auto plants, more hauls by truckers, etc. When they spend money, it in turn saves jobs in other industries, who in turn spend money and save jobs in other industries, and so on. I’ve got no idea what a multiplier effect might be for this (have at it in comments), but manufacturing jobs have the highest multiplier of any industry.

That’s not to mention the boost in taxes to state and local coffers, who will take anything they can get these days.

The true extent of the economic benefits of Cash for Clunkers won’t be known until sales numbers come in for September and October. There’s no doubt that many of the sales were just pulled forward a couple or three months by the subsidy, the question is how many. And to prove there are always unintended consequences with anything, the program seems likely to make it more expensive for poor folks to buy cars. Also, there’s the opportunity cost: What else would have or could have been done with that $2.9 billion? And some of the manufacturing benefits went overseas. It’s unclear how much since many foreign cars are now made in the U.S.

But the calculation also doesn’t account for the non-economic benefits of the program, as Justin Hyde of The Detroit Free Press pointed out to me on Twitter.

The new cars bought with the subsidies get 9.2 mpg more—a 58 percent jump—than the cars traded in. Those drivers will not only be spending less on gas (perhaps $600 a year), they’ll be polluting less, though that benefit is pretty microscopic.

The benefits of less pollution aren’t anywhere near enough in and of themselves to justify the spending—the dollar-to-carbon-saved ratio is high.

Consider them an added bonus. Oh, and getting those heavy cars off the roads makes them safer for the rest of us.

Ryan Chittum , a former Wall Street Journal reporter, is deputy editor of The Audit, CJR's business section. If you see notable business journalism, give him a heads-up at rc2538@columbia.edu.