At a high-level, off-the-record meeting concerning energy security that I attended earlier this week in Washington featuring New York Times columnist Tom Friedman, former CIA director James Woolsey, and energy consultant Daniel Yergin, a study came up in the course of discussion that has been bobbling around for a while now just below the radar screen regarding oil subsidies. The study, co-authored by major Republican C. Boyden Gray and published in a conservative law journal out of the University of Texas, alleges that the oil industry is subsidized to the tune of $250 billion a year, and that claim was marshaled to support the case for countervailing ethanol subsidies. If a careful guy like Boyden Gray — no enemy of business community he — has come to this conclusion, then there must be something to it, right? At least, that’s what many of the attendees were telling each other.

Now, this is a pretty remarkable claim given that the most aggressive yet credible oil subsidy estimates I’ve ever seen come from economist Douglas Koplow of Earth Track. He argued in a 1998 study for Greenpeace (not available electronically as far as I know) that total oil subsidies range from $18–40.6 billion if you count not just subsidies targeted at the oil industry but (1) those that help multiple industrial sectors as well, and (2) embrace some pretty ambitious claims about the chunk of defense spending that would disappear if the military’s oil mission were to disappear.

Look, I like Boyden personally. He’s been a generous contributor to the Cato Institute over the years and he’s gone out of his way to help promote many of our scholars here in Washington. But a close look at this paper of his speaks volumes about the poverty of the policy conversation in Washington with regards to energy.

Boyden’s argument boils down to this: chemical substances found naturally in gasoline such as benzene and other aromatic hydrocarbon compounds are imposing severe health costs on society. In a perfect world, the oil companies would have to compensate victims for those harms, but the federal government largely protects those companies from liability. This constitutes an implicit subsidy to the industry.

Boyden alleges that the direct harms from the various toxic emissions from gasoline total about $64 billion a year. But those aromatics also contribute to the formation of particulate matter (PM) in the atmosphere, and the harms from PM that can be traced back to aromatic gasoline emissions totals at least $200 billion a year. Boyden rounds the sum to $250 billion a year (which works out to about $1.78 a gallon in 2005) and argues that “leveling the playing field” would justify an equivalent subsidy to the ethanol industry. Ethanol subsides, he says, amounted to only $1.4 billion a year (the CBO estimate of the lost revenue associated with the federal fuels tax credit which, by the way, represents only a fraction of the total subsidies going to ethanol), so there’s a lot of room left to justify ethanol subsidies to the moon.

Boyden is right that the aromatics found in gasoline impose human health risks, and the regulatory history he tells about how Congress has dealt with this issue in the past is rather good. But his cost estimates relating to these emissions are drawn essentially from the ether.

His $64 billion estimate for the benefits associated with reducing aromatic emissions is simply the costs associated with reducing past industrial toxic air emissions. Huh? How did costs become benefits? Well, there are no independent estimates of the benefits. But the EPA asserts that the benefits from those previous industrial emission reductions exceed the costs so.… Even if the EPA’s claim were correct, there’s no reason to assume that the cost of reducing toxic air emissions from point-sources x years ago has anything to do with the cost of reducing toxic air emissions from automotive tailpipes today.

Boyden’s estimate for the costs associated with PM formation that can be traced back to gasoline likewise emerges from a problematic set of assumptions. He posits that 40 percent of all fine PM mass is carbon based (which seems fair enough) and then assumes that half of this mass (when adjusted for population exposures) can be attributed to gasoline emissions (which is not so fair enough; his own footnote suggests that only 4–33 percent of PM 2.5 can be traced back to tailpipe emissions). Using the benefit estimates associated with ambient PM concentration reductions from the recently established off-road diesel fuel regulations allows Boyden to come up with about $200 billion in benefits, although it’s unclear how he traces those costs to aromatic tailpipe emissions out of the total universe of motor vehicle tailpipe emissions.

I doubt whether anybody who’s citing Boyden’s study with gusto has ever gotten around to reading this particular sentence on p. 52; “We emphasize that these are, necessarily, speculative estimated, based on various heuristic assumptions that cannot easily be proven (or refuted, given basic uncertainties).” I’ll say. Normally, claims that cannot be proven or disproven are called “baseless opinions” (or, alternatively, “religious beliefs”). Let’s posit that we shouldn’t use either as the basis for public policy.

If Boyden was familiar with the literature on tailpipe emissions, he wouldn’t need to go through such analytic contortions. The man who probably knows more than any other person in the United States about the issues surrounding the environmental cost estimates associated with gasoline consumption is Mark Delucchi, a research scientist at the Institute for Transportation Studies at the University of California at Davis. His own economic calculations based on epidemiological work by others finds that environmental costs associated with toxic air emissions from motor vehicle tailpipes ranges from a lower-bound estimate of $87 million a year to an upper-bound estimate of $1.62 billion a year in 1991 dollars (which translates to $116 million-$2.16 billion in 2005 dollars) – a tiny fraction of the $64 billion estimate coming from Boyden.

Delucchi does not break down the PM emissions associated with gasoline aromatics, but he does report that the environmental costs associated with all the particulate emissions from motor vehicle tailpipes ranges between a lower-bound estimate of 16.7 billion and an upper-bound estimate of $266.4 billion. However, Delucchi reports that “we are uneasy with this result, even as an upper-bound,” because it’s heavily weighted by one study in the literature (Pope et al.) and that study is both anomalous and methodologically problematic. Regardless, keep in mind that Boyden is concentrating his fire not on all the particulate matter coming out of automotive tailpipes, but that subset of particulate matter formed as a consequence of the aromatic emissions. Given the small percentage by weight and volume that aromatics constitute within a gallon of gasoline, it’s clear that Boyden’s estimate is wildly off even if we use Pope et al.

By the way, it’s worth noting that the toxic air emissions associated with ethanol are even greater than the toxic air emissions associated with conventional gasoline, so even if Boyden’s estimates were correct, they do not justify countervailing subsidies for ethanol, the remedy for the problem suggested in Boyden’s paper.

One could spend a lifetime swatting down papers like this. That such weak arguments have no problem gaining currency in Washington demonstrates that policymakers simply cannot differentiate between analytic wheat and chaff. But such is the stuff that policy is made, particularly when the analytic “chaff” is politically convenient.