T’was the season of gift-giving and on December 9, outgoing EPA Administrator Andrew Wheeler delivered a parting gift for his successor in the form of a new regulation: Increasing Consistency and Transparency in Considering Benefits and Costs in the Clean Air Act Rulemaking Process.

The new Rule is offered as a simple housekeeping measure designed “to ensure consistent, high-quality analyses [and to] codif[y] best practices for benefit-cost analysis in rulemaking.” Some observers find it relatively harmless; but others are not so sanguine. We view it as a sort of Trojan Horse—seemingly innocuous on its face, but harboring content that will hamper, and may undermine, EPA’s efforts to confront the climate crisis and protect the safety of the air we breathe. Here are a few reasons we came to that view.

The problems begin with the word “codify.” For many years, EPA’s Cost-Benefit Analyses (CBAs) have been governed by non-binding guidelines. Readers of this blog will appreciate that guidelines can be adjusted to fit the circumstances of individual rules, and adapted over time as scientific understanding of best practice evolves.

No longer. The new Rule “codifies,” by our count, at least forty-two separate analytical steps the agency “must” undertake in virtually every CBA. Each one of these procedures is, according to EPA itself, meant to be “enforceable against the Agency” in a court of law. True, the rule offers a few qualifiers, such as “to the extent feasible” or “to the extent supported by scientific literature.” But those phrases themselves are left undefined, inviting more litigation and judicial second-guessing over the scope of the exceptions they provide.

Moreover, after this Rule, parties will be able to litigate, and invite courts to second-guess, not only agency decisions about which analyses to undertake, but also how to perform each one. To take just one of many possible examples: The rule requires the agency to select “human health endpoints” for analysis, based on “scientific evidence that indicates there is a clear or likely causal relationship between pollutant exposure and effect” and “sufficient data and understanding to allow the agency to reasonably model a [dose-response curve].” Who will decide when, or whether, each of these vague conditions is met? And what is the agency to do when, as often happens, the latest science indicates a potential health hazard but it doesn’t definitively prove a clear causal connection or support a dose-response curve? What will happen to the precautionary principle then? Henceforth, the shadow of litigation will hover over every agency judgment call made at every step of the analysis, and the steps will now be many.

The foreseeable result of all this “codification” will be an explosion of procedural litigation and a return to the very sort of judicial improvisation with procedure that the Supreme Court rejected in Vermont Yankee—as EPA is forced to adopt the slowest and most burdensome analytic procedures (and the slowest and most elaborate approach to implementing each procedure) for fear of unpredictable second-guessing by a later court. This time, however, judicial second-guessing will be given legal cover by the Rule, without need of the APA.

Media coverage of the rule has largely overlooked this ossification problem, focusing instead on the controversy surrounding the Rule’s treatment of co-benefits, defined as benefits other than those directly targeted by the rule in question. Some are encouraged by the fact that the new Rule does not ban the reporting of co-benefits. It merely requires, in the name of “transparency,” that EPA report co-benefits separately from primary benefits. What’s wrong with that? Again, this Trojan Horse rule appears innocuous at first glance, but further examination reveals a deeper mischief.

Understanding that mischief requires context, well-provided in a 2019 article by Professor Revesz and Kimberly Castle published in the Minnesota Law Review. In a nutshell, EPA and other regulatory agencies have consistently followed the practice of counting in their CBAs all benefits and all costs—direct or indirect, primary or ancillary. This was not politics, just textbook economics as practiced by Democratic and Republican Administrations and routinely accepted by courts for decades.

The controversy began when regulatory critics started noticing that many EPA rules derive huge “net benefit” numbers not from the monetized benefit of regulating the emissions they directly target but from the ancillary benefit (or co-benefit) of reducing particulate matter (or “PM”) emissions. For example, EPA calculated the net benefits of its 2012 Mercury Air Toxic Standard at $24 at $80 billion. But 99 percent of those benefits were attributable not to the massive reductions emissions in air toxics that were this rule’s raisin d’etre, but to the salutary fact that the measures power plant operators adopt to reduce mercury also reduce PM emissions. In fact, the mercury standard would have shown negative net benefits of over $9 billion but for the counting of PM co-benefits. This is typical of EPA’s air toxics rules, for which PM co-benefits often represent 100% of the monetized benefits.

The reason for this numerical imbalance, it is important to note, is not a lack of evidence of harm from mercury and other air toxics. They pose well-documented, serious risks to human health and the environment. The numbers problem arises from the fact that the benefits they yield are not necessarily of the type that are conducive to quantification and monetization using known methods and available data. Similarly, a study published by one of us last year showed that in 80% of the major rules issued between 2002 and 2015, EPA was unable to quantify and monetize whole categories of benefits that the agency itself described as “important,” “significant,” or “substantial.”

Particulate matter, on the other hand, also has generated an enormous body of scientific study over decades, but those studies have yielded a clear dose-response curve that makes it possible to estimate how many lives are likely to be saved (or illnesses avoided) from reduced PM exposure. Moreover, that curve exhibits no scientifically-supportable threshold of no-harm, meaning that every significant reduction in PM exposure saves lives. The quantification and monetization of those lives has yielded huge net benefit numbers for all sorts of rules that incidentally reduce particulate emissions while targeting other important air pollutants for which benefits can’t be so easily monetized—like air toxics and greenhouse gases.

Rather than welcome the reality that it is possible for certain rules to reduce both their targeted pollutants and particulates—killing two lethal birds with one stone—regulatory critics have chosen to attack the inclusion of PM co-benefits as somehow illegitimate and even illegal. They call it a devious way of “beautifying an otherwise economically indefensible rule” with inflated net benefit numbers, a backdoor device for regulating PM federally which otherwise would be left to states, and a source of double-counting of benefits as PM benefits are used to jack up benefit numbers in multiple rules.

Castle and Revesz patiently review and dispel each of these myths in their 90-page law journal article, but that has not stopped the campaign against the inclusion of co-benefits in agency CBAs. Their intervention may, however, have helped deter EPA from categorically excluding co-benefits in its latest rule: as we have seen, the Rule merely requires the segregation and separate reporting of primary and co-benefits in all rules going forward, without excluding co-benefits altogether.

What is wrong with separate reporting? Nothing, until you appreciate how regulatory critics plan to exploit that segregation, as candidly laid out in the Competitive Enterprise Institute’s (CEI) comment on the proposal. That comment suggests that CEI and its allies will use EPA’s reporting of co-benefit-free estimates as an agency affirmation of the validity of excluding co-benefits in impact analysis—at least as one credible mode of analysis. They will then proceed to report the non-co-benefit estimate as the “real” benefit of the rule. They will claim that the only “real” benefits are quantified and monetized benefits; that the only legally cognizable benefits are those directly targeted by the particular statutory provision that authorizes the rule; and that monetized co-benefits, now that they have been isolated, should be excluded. Separate reporting of benefits and co-benefits may not change the outcome of the analysis immediately, but it will fuel a campaign of rhetorical broadsides and legal challenges to each rule that includes co-benefits and that, one suspects, is the point.

There are other major problems with this new Rule. For example, it requires the reporting of net benefits to society at large in all cases, while failing to acknowledge that reporting the “net benefits” of a rule is affirmatively misleading if the supporting analysis omits or elides important un-monetized benefits or distributional impacts. Indeed, the new Rule is conspicuously silent on the subject of distributive impacts analysis or environmental justice. We stop with the deficiencies discussed above solely for lack of space. They are sufficient, we believe, to establish that this rule is not about transparency, it is not right, and it is not harmless. It should be repealed as soon as possible, through means we leave for other scholars and/or another day.

Richard W. Parker, professor of law at UConn School of Law, specializes in administrative law and environmental law and regulation.

Amy Sinden, professor of law at the Temple University Beasley School of Law in Philadelphia and a member scholar at the Center for Progressive Reform, specializes in environmental law and regulation.