Should you be worried about mercury emitted from power plants?

Sure, but only if you are a pregnant woman, who during gestation consumes about 220 pounds of fish caught from exclusively the top ten percent most polluted fresh waters of the United States, despite all the signs along these rivers and lakes warning “DO NOT EAT THE FISH!”

Don’t take my word for it. I’m simply relaying EPA science. And not the ‘bad” kind produced by the Trump administration; rather, I’m talking about virtuous EPA science as practiced by the Obama administration.

A little background: mercury emissions aren’t a direct threat to humans, but instead settle onto water bodies, and then make their way up the aquatic food chain. Because mercury is a neurotoxin, the fear is that pregnant women can engender developmental disorders in their offspring by eating fish that have bio‐​accumulated the toxin.

In the course of promulgating the Obama‐​era Mercury and Air Toxics Standards for power plants, the EPA stated that it considers “IQ loss estimates of 1–2 points as being clearly of public health significance,” even though this low a number rests comfortably within the error of measurement inherent to an IQ test. According to the EPA’s analysis, the Mercury Rule was necessary to prevent an IQ loss of 1.1 points supposedly suffered by children born to a putative population of pregnant women from substance families, who during their pregnancies eat 220 pounds of self‐​caught fish reeled in from the most polluted bodies of fresh water. Notably, the EPA failed to identify a single member of this supposed population. Instead, these women were modeled to exist.

Even under EPA’s ultra‐​accommodating analysis of its rules’ benefits, the agency pegged the benefits of the Mercury Rule at a mere $6 million. In stark contrast, the agency estimated that the rule would cost about $10 billion annually, making it one of the most expensive regulations ever.

On its face, such an imbalanced cost‐​benefit ratio is plainly unreasonable. But the EPA pointed to the rule’s “co‐​benefits,” which were estimated to be dwarf the rule’s costs.

So, what are these “co‐​benefits”?

Retrofitted air pollution controls employ either a filter or a chemical reaction to capture pollutants from the power plant exhaust flue. Though these controls are optimized for the specific pollutants they are designed to reduce, other pollutants also are captured. With the Mercury Rule, the EPA claimed that “co‐​benefits” attendant to the required mercury controls would amount to $37 billion annually.

On the one hand, I agree with Prof. Cass Sunstein, who has argued that it would be foolish to ignore readily evident costs and benefits, regardless whether they are direct or indirect. On the other hand, there’s an element of duplicity behind the EPA’s co‐​benefits, to which I object.

Co‐​benefit pollutants are known as “criteria pollutants,” and they are regulated by the Clean Air Act. Indeed, criteria pollutants are regulated at a level that is “requisite to protect the public health” with “an adequate margin of safety.” That is, these pollutants are regulated by an entire statutory program at standards that go beyond what is necessary to protect public health.

The EPA, moreover, is not allowed to consider costs when it regulates “criteria” pollutants. It is somewhat ironic that the agency is attributing billions of dollars worth of benefits to the reduction of a pollutant beyond stringent public health limits set by EPA without considering costs.

The other problem with EPA’s “co‐​benefits” is the likelihood of double‐​counting. Since the George W. Bush administration, the EPA has justified most of its major rules by relying on “co‐​benefits.” Yet the agency isn’t keeping a running tab of its claimed “co‐​benefits.” As a result, it’s almost certain that the agency has counted the same benefits twice or more. 

Recently, EPA Administrator Andrew Wheeler initiated reforms regarding the agency’s use of of costs and benefits in the rule‐​making process. While there’s no reason for the agency to categorically ban the use of co‐​benefits, the EPA should render their use reasonable. To this end, the EPA should refine its analysis to account for doubly‐​counted co‐​benefits. The agency also must inform the public how much of the co‐​benefits valuation can be ascribed to pollution reductions below a level that the agency already had determined to be “requisite to protect the public health” with “an adequate margin of safety.”