Deep Dive Episode 43 – Re-Considering Co-Benefits in Environmental Regulation
This Deep Dive episode brings you the recording of a teleforum co-sponsored with the Federalist Society’s Environmental Law practice group.
In this episode, Adam Gustafson and Daniel Farber discuss various approaches to considering co-benefits in the cost-benefit analyses of new air pollution regulations, and whether the standing approach is the most efficient and cost-effective.
Operator: Welcome to Free Lunch, the podcast of The Federalist Society’s Regulatory Transparency Project. All expressions of opinion are those of the speakers. On March 29th, RTP co-sponsored a teleforum discussion with the Environmental Law Practice Group entitled, “Reconsidering Co-Benefits in Environmental Regulation.” This call discussed how the EPA’s reconsideration of its co-benefit accounting methodology in the context of its mercury standard could have significant implications for environmental regulation. It featured Professor Daniel Farber of the University of California, Berkeley and Adam Gustafson, who is a partner at Boyden Gray & Associates. We hope you find it interesting and enjoyable.
Adam Gustafson: If you’ve ever read Astrid Lindgren’s Pippi Longstocking, you know the story of Pippi and the tin can. The eccentric, red-headed nine-year-old finds a large, rusty old tin can in the grass and decides that it will come in handy to put over your head and pretend that it is midnight. While her friends watch her, speechless, Pippi puts the can on her head and starts walk around the block. Before you know it, she stumbles over a fence and falls flat on her stomach with a crash. “Now, see that,” says Pippi, “If I hadn’t had this can on me, I’d have fallen flat on my face and hurt myself terribly.”
Pippi Longstocking had received her training in cost-benefit analysis from certain environmental regulators who seem to be able to justify the costs of any action, no matter how quixotic, and Pippi shows us how. The trick is to identify a substantial benefit from the action to weight against its cost, no matter how disconnected that benefit may be from the problem at hand and no matter how easily the same benefit might have been realized in another less costly way. If the comparison seems unfair, consider the Obama administration’s Mercury and Air Toxic Standard, which was intended to reduce mercury emissions from power plants. The rule was projected to generate relatively small, quantifiable mercury-related benefits, only $4 to $6 million per year, much less than the estimated annual compliance costs of $9.6 billion, billion with a B.
But EPA’s regulatory impact analysis also included $37 to $90 billion in ancillary benefits, also known as co-benefits. Specifically, these were reductions of particulate matter that EPA said would occur as a result of the mercury control technology that power plants would have to install to comply. EPA counted the particulate matter co-benefits, even though particulate matter is a criteria pollutant that EPA already regulates under a different provision of the Clean Air Act. EPA’s cost-benefit accounting appeared in the Mercury rule’s regulatory impact analysis, as required by executive order.
But the agency disclaimed any reliance on that analysis in promulgating the rule. It said that the statutory command to determine whether regulation is appropriate and necessary did not require the agency to consider cost. As you might expect, the Mercury rule was challenged by the power industry, which called it, quote, “the most costly rule ever issued by EPA.” The Supreme Court held that EPA had misinterpreted the statute when it refused to consider cost. “No regulation is appropriate,” wrote Justice Scalia for the Court, “if it does significantly more harm than good.” The Court expressly reserved the question whether, quote, “the agency could have considered ancillary benefits when deciding whether regulation is appropriate and necessary,” end quote.
On remand, the Obama administration doubled down. EPA concluded that the Mercury rule was cost-justified because its co-benefit particulate matter reductions exceeded the costs of the rule. The Trump administration changed course in December. Under Administrator Wheeler, EPA proposed a revised supplemental cost finding that, if finalized, would rescind the prior finding that the Mercury rule is appropriate and necessary. Specifically, EPA now proposes to exclude the particulate matter co-benefits from its analysis and instead to compare the direct costs of compliance to the direct benefits of controlling mercury emissions. EPA is not proposing to rescind the Mercury rule itself, which power generators have already complied with. My position is that EPA’s decision to exclude particulate matter co-benefits is appropriate in this instance. Let me briefly outline five reasons why.
First, particulate matter is one of six criteria pollutants that EPA already regulates through its National Ambient Air Quality Standards. The statutes that govern that program are complicated. To this day, I rely on Professor Farber’s Environmental Law in a Nutshell to get it straight, but the important point is that the statutory scheme for particulate matter and other criteria pollutants is distinct from EPA’s regulation of mercury and other hazardous air pollutants under Section 112. When the vast majority of a rule’s benefits are derived from pollutants that EPA is supposed to control under a different provision, it’s hard to avoid the conclusion that EPA is evading the scheme Congress established. Chief Justice Roberts made this point in oral argument in the Michigan case. “It’s a good thing if your regulation also benefits in other ways,” he said, “but when it’s such a disproportion, you begin to wonder whether it’s an illegitimate way of avoiding the quite different limitations on EPA that apply in the criteria program.”
Second, it’s not just that EPA ought to regulate particulate matter under a different statute – it already has. EPA set a particulate matter standard of 12 micrograms because it deemed that the level requisite to protect public health with an adequate margin of safety, in the words of the statute. When it imposed that standard on the States, EPA quantified the benefits of reducing particulate matter emission to that level. To the extent the Mercury rule counted the benefits of reducing particulate matter above the National Ambient Air Quality Standard, it was double counting.
Third, and relatedly, the Obama administration’s approach to particulate matter co-benefits treated all reductions equivalently, even if they occur below the level EPA has already deemed requisite to protect the public health. But EPA selected that 12-microgram standard because it was, quote, “somewhat below the lowest concentration shown by key epidemiological studies to cause adverse health effects,” end quote. If particulate matter reductions in low concentration areas don’t actually benefit human health, then it makes no sense to assign them the same value as reductions in high concentration areas. The Trump EPA alludes to this problem in a memorandum to the docket for the new proposal. It sites, quote, “uncertainties regarding the relationship between PM”—that is particulate matter exposure—“and the risks of premature death at low PM concentrations,” end quote. It goes on to say that, quote, “these uncertainties are particularly important because air quality has improved over time, reducing the fraction of the U.S. population experiencing elevated PM exposures,” end quote.
The fourth reason that I think it’s appropriate for EPA to exclude particulate matter co-benefits is that relying on these co-benefits to justify an otherwise unjustifiable rule is not efficient regulation. Even if there were no double-counting or overcounting in EPA’s co-benefit assessment, relying on co-benefits encourages roundabout ways of controlling harmful air pollution. Are mercury controls really the most efficient way of controlling particulate matter? That seems unlikely, and EPA itself has determined that an annual standard is, quote, “the most effective and efficient way to reduce total population risk associated with both long- and short-term particulate matter exposures,” end quote.
Finally, relying on particulate matter co-benefits undermines the cooperative federalism that Congress built into the Clean Air Act. When EPA sets a National Ambient Air Quality Standard for particulate matter of any other criteria pollutant, it’s the states who get to decide how to implement that standard. Using particulate matter co-benefits to justify the Mercury rule avoids that legislative scheme by empowering EPA to regulate powerplants directly.
In conclusion, EPA’s decision to exclude ancillary reductions of criteria pollutants from its cost-benefit analysis makes sense in this circumstance. It promotes rational and efficient regulation, and it is consistent with the rule of law. And I look forward to hearing from Professor Farber.
Micah Wallen: Thank you, Adam, for those opening remarks. We will now switch over to Professor Farber.
Prof. Dan Farber: Well, I wish I had thought of a Pippi Longstocking story to start with. It’s always great to have the opportunity to participate in Federalist Society events, and I’m happy for the invitation to speak today. I’d like to start with the general question of co-benefits and then talk a little bit about some of the specifics. I’m not going to try to respond to each of the five arguments because I think that would take a little bit too long, but I’m happy to talk about them during the Q&A.
So I’d like to start with just the general issue of co-benefits and cost-benefit analysis. And I think that it’s important to go back even to President Reagan’s initial executive order, which lays out the argument for using cost-benefit analysis. And that’s, generally, that we want to do things that improve society’s welfare, and, therefore, we want to do things where the benefits of a regulation are greater than the cost. And that has been always interpreted by the White House and by pretty much every economist to mean all the costs and all the benefits.
Obviously, you might have some reason for wanting to assume co-benefits or various kinds of costs, but if you do that, you’re not really doing cost-benefit analysis anymore. And you’re not really focusing on the question of will a regulation make society better off. So to the extent there’s a broader attack on co-benefits, it seems to me that that, essentially, is an attack on the idea of doing cost-benefit analysis as the basis for regulation.
Now, the specific argument here is that we should truncate the consideration of benefits in determining whether regulating toxic emissions from coal-powered plants is appropriate. I do think it’s important, even though there is a special provision for powerplants, to keep in mind that this is part of Section 112, which is a very rigorous scheme for regulating toxics. Congress, apparently, put a lot of weight on the idea that toxic chemicals are a bad thing to put in the environment and that we should try to do something about that.
Now, it seems to me the question here is whether the term “appropriate” excludes consideration of co-benefits or, indeed, whether it excluded consideration of cost as EPA originally thought. The Supreme Court concluded that costs do have to be considered, and the opinion by Justice Scalia defines costs in very broad terms, to include, for example, environmental harms that might be caused by a regulation, not just monetary costs. And it seems to me there’s a strong argument for parallelism that, if we’re going to include all the costs, we really ought to be including all of the benefits.
Now, again, we may have reasons for not doing that, but I think it’s important to keep in mind that, if we do decide that it’s not appropriate to consider co-benefits, we’re saying that we’re willing to pass up an opportunity to do something which will have major benefits for society above the costs and that, in fact, will, according to the, at least, initial regulatory analysis EPA did, will save hundreds of lives in the name of statutory interpretation. It seems to me that the right answer here is that the full range of costs should be considered because I think, otherwise, we have to embrace the view that Congress preferred, or that the word appropriate mandates, that we take actions that will result in very harmful pollution that will be very detrimental to society when regulating would cost much less than the benefits of the regulation.
On the more specific points that Adam raised, I think the question — I want to just talk about the issue of Ambient Air Quality Standards, and I guess I wanted to point out a couple of things here. In terms of emission levels that are — or air quality levels that are above the Air Quality Standards, that is that are not in compliance with the Clean Air Act, the reality is that that includes a fair amount of the country. There are a lot of areas that are not in compliance, and, in those areas, anything that reduces the level of particulates will, according to the analysis underlying the Air Quality Standards themselves, will have a direct effect on human health. In terms of areas that are below the Air Quality Standard, I think it’s important to keep in mind that the Air Quality Standards are not set on the basis that there’s zero risk below the standard.
In fact, for, I think, pretty much all the major air pollutants, we don’t have any reason to think that there’s a threshold at which the risk of harm drops to zero. So in setting the Air Quality Standards, EPA is trying to decide, for the country as a whole, at what level should we set standards that will prevent harm to broad portions of the public based on available evidence. But that doesn’t mean that there’s no risk below that level. And to the extent that there is risk, potentially, below that level, it’s reasonable for EPA to take further action. So I don’t really see the argument about Ambient Air Quality Standards as being as decisive as Adam does.
And in terms of whether EPA is pulling a fast one, Congress did require EPA to make a determination about whether it’s appropriate to regulate. It’s not something EPA just made up as sort of a dodge to avoid using other parts of the statute. They’re carrying out the function that Congress gave them. So I think the right answer is different than the one EPA is proposing. I think it’s a closer question, whether the answer proposed by EPA is sufficiently reasonable to withstand Chevron — to qualify for Chevron deference. And at that point, I think it becomes hard to tell because what is reasonable is somewhat in the eye of the beholder. From my point of view, I guess I’m not sure that it’s reasonable to say that we’re willing to knowingly sacrifice, you know, a significant number of lives by adopting a restrictive interpretation of the statute. But I can imagine that some judges would disagree with me and would conclude that it’s reasonable.
Of course, there would be some irony in EPA relying on Chevron, given that the current move among conservatives, I think, including members of the Trump administration, has been to jettison Chevron.
The final thing I’d like to say is just that this is an issue that only a lawyer could love. We’re dealing with a threshold determination about whether it’s okay to issue a regulation, but we’re making that decision at a time when the regulation has already gone into effect. The industry has complied, and the industry does not seem to have any interest in overturning the regulation. And in fact, according to the EPA, even overturning the threshold would not have an actual effect on the requirements on the industry. So we’re dealing with kind of a little bit of angels dancing on the head of pin issue here.
As I was saying, I think you have to go to — you know, I think it takes three years of law school to convince you this is the kind of thing that we really ought to be spending our time worrying about. And with that, let me wait for the Q&A and for any comments that Adam might have.
Micah Wallen: Tossing it back over to you, Adam. Is there anything you wanted to respond to there before we switch over to audience Q&A?
Adam Gustafson: Well, just very briefly, on the points that Professor Farber made about reductions of PM, particulate matter, above and below the National Ambient Air Quality Standard. So to the point that much of the country is still out of attainment with the particulate matter standard, I think that’s true, but the question is how should the country be brought into attainment. And under the legislative and regulatory scheme that Congress has put in place and that EPA has implemented, the way to do that is through state implementation plans, that states propose and EPA approves. And the problem that I see with trying to achieve those reductions above the NAAQS through this alternative mechanism is that it really throws cooperative federalism under the bus.
The whole premise of the NAAQS program is that states get the first crack at it, and what we have here is EPA taking that prerogative away and saying, “We know better. This is how we’re going to bring you down below the level of the NAAQS.”
For reductions below the standard, I don’t know enough to disagree with the claim that there’s no threshold below which the risk drops to zero. But that is different from saying that the risk is the same at every level, and EPA has decidedly not made that finding. Instead, when it established the 12-microgram standard for particulate matter, it said very clearly that that was the standard, below which — no standard below that would be warranted to provide requisite protection, that it’s neither more nor less than needed to provide an adequate margin of safety. So if there is a risk below the National Ambient Air Quality Standard, it is certainly a lower risk than the risk above the standard. That’s why we have that standard in the first place.
And EPA’s historic approach to particulate matter and ozone co-benefits is to treat all reductions the same, which doesn’t make any sense, especially when the cost avoidance that is being monetized here is the cost of avoided mortality. And it’s just not true as a matter of science or common sense that the effect of particulate matter on mortality is the same at any level. Professor Farber said EPA should take further action if there’s still a risk, and I wouldn’t disagree with that. But I would point out that the Clean Air Act provides the way forward. It requires the agency to reevaluate its National Ambient Air Quality Standards for PM and other criteria pollutants and then to set more stringent standards if those are requisite to protect human health. So I’ll leave it at that.
Micah Wallen: I’m not seeing any questions jump in right off the bat. I will ask both of you what you think of the other issues and arguments present in this case and whether or not they are relevant.
Adam Gustafson: I defer to Professor Farber.
Prof. Dan Farber: I was just going to defer to you. So I guess one of the issues that I would like to hear more about, and I think Adam may be somewhat more knowledgeable than I am, is the question of whether overturning the threshold finding that regulation is appropriate would actually leave the regulation intact. It seems kind of peculiar in some ways, but, on the other hand, it seems equally peculiar to say that, once the regulation is there, the question of whether it is — think about regulating remains very relevant. So I don’t know. Adam, if you can shed any light on that, that would be great.
Adam Gustafson: It’s an interesting question. It might seem strange at first blush that the agency is proposing to change its reasoning, to change its conclusion that a Mercury rule is appropriate and necessary but not to actually propose rescinding the rule. And I think Professor Farber alluded to the rationale that these are largely sunk costs because the regulated community has already come into compliance with the standard. And so I think the rationale might be that there’s really no cost savings to achieve by rescinding the rule. Rather, from my perspective, this action seems more important as a statement of agency policy and as a potential precedent for future rules under Section 112 and perhaps other statutes as well. And to that end, I think it’s worthwhile for the agency to take this opportunity to reconsider its reliance on co-benefits and to come up with some standards that it can apply consistently for treating co-benefits.
Micah Wallen: All right. And we do have a question in the queue right now. So without further ado, we will go to our first audience question.
Caller 1: Yes, thank you. My question is for Adam. Under what circumstances would you say consideration of co-benefits is appropriate?
Adam Gustafson: That’s a tricky question. Let me put it this way. I think the argument for considering co-benefits would be much stronger if they are not already regulated — if those pollutants are not already regulated by another provision of the Clean Air Act. A lot of the co-benefits that are relevant here are fair game under Section 112, which talks about not just mercury, but all hazardous air pollutants. EPA can expand that list as necessary when it can define new hazardous air pollutants. So I would say that maybe there’s really no need for co-benefits in this circumstance — co-benefit pollutants because EPA has, at its disposal, the ability to define new hazardous air pollutants and to regulate them together with mercury.
And in fact, this rule was addressed not only at mercury but at other HAPs, other hazardous air pollutants. The problem is EPA just had a hard time quantifying the benefits of any of those reductions. So I haven’t really given a good answer to your question, but I think the more attenuated the co-benefits are from the target of the agency action, the less likely it is that they are appropriate subjects for consideration in cost-benefit analysis.
Micah Wallen: Did you have anything you wanted to add on to there?
Prof. Dan Farber: Two things. Well, the first is I think it’s a mistake to think that the Clean Air Act, or the environmental statutes generally, are so carefully designed that there’s never any overlap between sections. For that reason, I think, you know, it’s sort of a mistake to say, “Well, you can’t regulate here because you could have regulated another way.” I think that will often be true. For example, toxic substances could also be reduced by the effects of state implementation plans, and I don’t see why we should necessarily ignore those reductions in setting up state implementation plans. If I were a state regulator, I think I would want to regulate sources that would allow me to both get to the Ambient Air Quality Standards and reduce toxins at the same time.
In fact, the statute itself calls in several places for considering non-air quality environmental effects, even though those might well be regulated by different environmental statutes such as the Clean Water Act. So I just think that Congress was dealing with very complex problems and often provided alternate routes to deal with similar — where you would have effects on other issues.
In terms of whether co-benefits generally should be reduced, you know, again, I want to say that, to the extent you’re doing that, you’re really moving away from the idea that we want to decide on regulations based on whether they improve the welfare of society, which is the theory behind cost-benefit analysis. If you start saying, “Well, we’re not going to count all of the costs or all of the benefits,” you’re really saying, “Well, we’re no longer interested in that goal of maximizing social welfare. At least, we’re not interested in making that a basis for deciding on particular regulations.”
Now, there’s certainly an argument to be made for that. I have a lot of environmentalist friends who are opposed to cost-benefit analysis because they think it’s just a mistake to give equal weight to costs and regulatory benefits. And they have some arguments for that position. And I think there’s some arguments for the other position, maybe that we should be less worried about regulatory benefits than about costs. I think that’s kind of the instinct of the Trump administration, anyway. But whichever of those approaches we take, I think we can no longer say we’re engaged in cost-benefit analysis.
Micah Wallen: Adam, did you have something to add?
Adam Gustafson: Sure. If I could, I’ll just respond to two of those points that Professor Farber made. On the question of the statutory reference to non-air quality health and environmental impacts, I would just point out that, in context, it’s pretty clear that those considerations are relevant to the extent that they render some potential standards non-achievable. The statute says that the administrator shall require the maximum degree of reduction in emissions that the administrator determines is achievable, taking into consideration the costs of the reduction in any non-air quality health environmental impacts. So I don’t see that as an invitation to regulate more stringently. Rather, I think it’s a consideration for the agency’s determination as to whether any given standard is achievable.
So any effect on energy prices, for instance, could deter the agency from regulating because it would make that given standard not achievable. I don’t think it’s a license to regulate more stringently. On the general point that executive orders since Reagan have required regulations to do more good than harm, I wholeheartedly agree with that, and I think that that is a recipe for good policy. I would point out that what is going on in EPA’s recent proposal for a supplemental cost finding doesn’t really go to the 12866 regulatory impact analysis. In fact, EPA expressly carves out regulatory impact analysis. It says that reporting co-benefits in regulatory impact analyses is, quote, appropriate for informing the public about the potential effects of the regulatory action.
So I think this proposal will have a limited effect on how cost-benefit analysis is done on the whole. This proposal is really focused on what benefits EPA is allowed to consider to outweigh costs in this particular context of 112. So I don’t think that we’ll see EPA immediately stopping any consideration of co-benefits in regulatory impact analyses.
Micah Wallen: I’m not seeing any questions lined up. Adam and Professor, did you have any closing remarks?
Adam Gustafson: Nothing further from me, but it’s been a pleasure. And I really appreciate the chance to talk with you.
Prof. Dan Farber: Let me just echo all of that. I’ve really enjoyed the conversation. I think I’ve come away with a better appreciation of the arguments, the other side. And as always, I appreciate the willingness of the Federalist Society to provide forums for a diversity of viewpoints. So thank you.