Buried in Joseph White's interesting Wall Street Journal piece about the enormous power of incremental improvements in auto fuel economy (versus so-called moonshots) sits an interesting statistic:

"Mr. Fedewa estimates auto makers can deliver 10% or better fuel efficiency improvement for about $500 a vehicle." (That is Eric Fedewa, described as "a vice president with the automotive consulting firm CSM Worldwide.")

$500 for 10 percent improvement!

What if the government just paid the costs upfront as part of the stimulus package?  Could we invest in fuel economy standards as a sort of time-release piggy bank?

Last year Americans bought 7.87 million light trucks. Paying an extra $500 per vehicle would cost about $3.9 billion and at last years prices of over $3 a gallon and this years of over $2, the investment would paid off in two years. By year three, we would have saved more than a billion gallons of gas on that model year alone, and it would free up a billion dollars in gasoline money to be spent on other things that year (if gas were merely $2.65 a gallon.) In other words: we could pump the $500 into cars now, and the stimulus money would hit the auto industry now, but in three years the benefits of the savings on gasoline would be hitting the economy as a whole in increased disposable income.And every year after that, too.

Want to stimulate without adding to the deficit? The government could finance the $500 at 0 percent interest rate to the manufacturers, who then pass the cost along (again at zero percent) as part of the financing for auto consumers.