Calculating the costs of pollution

Last week, a divided court of appeals upheld what may well be the most important environmental rule in the nation’s history: the Environmental Protection Agency’s mercury standards. The regulation is expected to prevent up to 11,000 premature deaths, 4,700 heart attacks and 130,000 asthma attacks a year.

Critics of the mercury rule have focused on its expense. The EPA estimates it will cost $9.6 billion a year, with most of the burden falling on electric utilities. Indeed, the issue of cost is what split the court.

The Clean Air Act allows the EPA to regulate electric utilities under its hazardous air pollutants program only if it finds that such regulation “is appropriate and necessary.” Focusing solely on mercury’s effect on public health, the EPA made that finding.

That troubled Judge Brett M. Kavanaugh. In his dissenting opinion, he asked, quite reasonably, how the EPA could possibly conclude that regulation is “appropriate” without considering costs. He argued that it’s “just common sense and sound government practice” to take account of costs as well as benefits.

But the court’s majority had an answer. It noted the EPA had considered costs — not in its initial decision on whether to regulate mercury emissions but in its second and crucial decision about how stringent the regulation should be. The court’s majority also emphasized the benefits of the rule, which have been estimated to be worth from $37 billion to $90 billion, outweighing the costs by a factor of between 3 to 1 and 9 to 1.

For the future, the mercury controversy offers two lessons. The first is that no approach to environmental protection can afford to be indifferent to costs. In situations where Congress allows the EPA to take account of costs, the agency should do so, at least in considering the appropriate level of stringency.

The second lesson is more subtle. Calculations of the costs and benefits of a regulation should use the best available science. In the case of mercury, for example, the substance itself is a neurotoxicant, potentially affecting memory, language, attention and cognition. But it is not easy to quantify mercury’s adverse effects, and the EPA did not try.

The massive benefits identified by the EPA are expected to come not from mercury reductions but from “co-benefits” — that is, from reductions in other air pollutants that will result from efforts to reduce mercury emissions.

The vast majority of the quantified benefits of the mercury rule are a product of incidental reductions in emissions of just one other pollutant: particulate matter. That’s not all that surprising, given that reductions in particulate matter, which can cause serious health problems, accounted for about one-third to one-half of the total monetized benefits of all significant federal regulations from 2003 through 2012.

In coming years, the benefits of further reductions in particulate matter will be among the most contested issues in environmental regulation. No one doubts that particulate matter is harmful to human health, but we need answers to important questions about which particles are the most dangerous and how much damage they cause at low concentrations.

Increasingly, that will involve relying on so-called natural experiments to learn about the effects of pollution on health. Such experiments ask what happens to health when some practice or event (a regulatory action, for example) causes a significant, and sometimes abrupt, change in pollution for one group of people but leaves pollution unchanged for another, similar group of people.

One such study examined the health of two similar populations in China. One consisted of people who lived north of the Huai River, who received free, government-provided coal to heat their homes. The other was made up of people who lived south of the river and did not get free coal. Those heating their homes with the free coal were found to have life expectancies 5.5 years shorter than those who did not, due to the coal-generated particulate matter they inhaled.

A greater reliance on natural experiments would be an important improvement in regulatory policy. Currently, we rely on observational studies that, while informative, don’t necessarily tell us what we need to know to determine a regulation’s benefits. These studies compare health in places with high and low levels of pollution. But if, say, Los Angeles has higher levels of asthma than, say, Boise, Idaho, differences in the populations — such as people’s socioeconomic status or health habits — rather than differences in air pollution may be responsible. Researchers try to control for those differences with rigorous statistical methods, but it is not always easy to do so.

The court’s decision to uphold the EPA’s mercury rule means that the American people will be able to enjoy significant health benefits. As we celebrate that decision, we should ensure that the best science is brought to bear on decisions that affect present and future occupants of the planet.

Francesca Dominici is a professor in the department of biostatistics at the Harvard School of Public Health. Michael Greenstone is a professor of environmental economics at the Massachusetts Institute of Technology. Cass R. Sunstein is a professor at Harvard Law School; he worked on the mercury rule while serving in the Obama administration from 2009 to 2012.