Doubting the Health Harms of Air Pollution — Not a “Mistake,” but Not “Just Science” Either

Tony Cox (directly behind EPA Administrator Scott Pruitt)
  • Something is amiss that not one word in the 120 slides mentioned the fact that agencies’ estimates of regulatory cost are also plagued with assumptions that could invalidate key relationships. After all, where do we get the ideas that prices will rise or that jobs will be destroyed if we require pollution controls? From economic modeling, of course — where claims of future causality at best come from observations of past associations. So encouraging judges to ignore benefits but to swallow whole accusations of cost is a half-right lecture, which is to say a wholly incomplete piece of advocacy;
  • Cox identifies legitimate problems such as confounding (a spurious association between exposure and disease, actually caused by an unobserved third factor linked to both), but failed to tell the judges that epidemiologists have cared about these problems for many decades and have refined statistical methods to reveal and correct for them;
  • He, of course, counsels for better information, for waiting to decide until we can resolve these important unknowns. This, of course, is both respectful of the science and (conveniently?) a recipe for inaction over action. It feeds into a legal/regulatory system in which the status quo allows exposures to continue unabated until a final decision is made that the free market is (oops!) creating health and environmental costs in excess of its economic benefits;
  • Tony is only interested in how government might regulate harms as causal when they are not, and so he fails utterly to mention the mirror-image problem of ignoring harms as non-causal when they are. In fact, I can easily argue that risk analysis currently sets a higher bar to guard against these false negatives than it needs to be strengthened to prevent false positives. Consider the perennial rule that an association found in an epidemiology study cannot be considered “real” unless it is statistically significant at the p<0.05 level. That means that an environmental exposure that appears to double the risk of disease cannot be regulated if there is a 5 percent chance that the exposure does not increase risk at all. I support this kind of gatepost, but it seems ironic (these days, a euphemism for “deliberate”) that Tony wants us to make regulations less stringent when we aren’t 100% sure an effect is causal, but isn’t interested in changing our practice of doing nothing when we are “only” 90% sure an association is real.
  • Most importantly, Tony has waded waist-deep into the Big Muddy of decision-making under uncertain consequences, a terrain that simply cannot be navigated without imposing value judgments upon the choice. He is part of a long and frustrating tradition of above-the-fray scholars who believe (or claim to believe) that they are merely providing “objective” advice, yet nothing could be further from the truth. A doctor who says that a patient should get an amniocentesis (because she is above the age where the odds of having a child with Down syndrome exceed the odds of the test itself causing a miscarriage) is forcing her to regard the consequences of the two bad outcomes as equally dire, when any other weighing of the two outcomes could be her true and valid preference. A consultant who says that you should find the “best estimate” of how long it will take to drive to the airport to catch your flight is telling you — without telling you! — that you regard being 4 minutes late to the gate as better than being 5 minutes early (the former mistake is “closer to the right answer” than the latter). So when Tony tells judges that they should regard a “possibly causal” exposure as less worthy of intervention than a “known causal” one, he is steering them to the view that false positives (or over-regulation) are to be avoided — and by counseling them to “exclude” harms not proven causal, to the view that false positives are far more dire than the opposite. This is a perfectly defensible value judgment, but it is indefensible to hide it by failing to mention that any other weighing of the two is also valid. Maybe ozone and fine particles aren’t as dangerous as the face-value epidemiologic studies would indicate: but if we regulate them less stringently (or not at all) because of real or manufactured doubt, we would be balancing the costs of error in strict and profound ways, none of which Tony wants to talk about. In my long experience, when experts don’t want to admit their values it’s because they prefer to impose their values covertly, and prefer to belittle those who advocate for different values and who do so transparently.
Causality is important — and controversial!



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Adam M. Finkel

Adam M. Finkel

Risk assessment expert, former federal government regulator (OSHA), choral singer and conductor