Our “Viewpoint” feature in the magazine gives economists and climate researchers the opportunity to provide a new perspective on an important topic. In this issue, two RFF scholars discuss how the Trump administration’s approach to benefit-cost analysis could have far-reaching impacts on federal environmental policy.
Thoughts on this? Send a response to the editor by letter at attn: Managing Editor, 1616 P St NW, Suite 600, Washington, DC 20036 or email at [email protected] for possible inclusion in the next issue of Resources or on the Common Resources blog.
The Trump administration has reshaped how environmental regulations are implemented and how benefit-cost analyses are conducted. RFF’s Art Fraas and Richard D. Morgenstern debate how enduring these changes might be, and what the Biden administration can do to reorient the regulatory process around sound science.
Resources: In light of our recent history of regulatory rollbacks, you both have expressed concern over a general erosion of trust in science and economic analysis. Looking now toward the prospect of applying science and economics to ambitious federal climate policy, do you think the general public can recover from what’s been interpreted as a recent debasing of economic analysis? Will people make space for scientists and economists again?
Art Fraas: Let’s take benefit-cost analysis as a major example: the whole idea came out of the Reagan administration. The Left has always been skeptical of benefit-cost analysis, even before the Trump administration. In about the last 10 years, as the US Environmental Protection Agency (EPA) has been able to generate large net benefits associated with its rules, the Right has become more and more skeptical of benefit-cost analysis as well.
Richard D. Morgenstern: The previous administration undermined the reputation of benefit-cost analysis by cutting away and attacking the benefit side in particular, but also by overemphasizing the cost side. Benefit-cost analysis by its very nature is a balancing of the two. If you attack one side and focus only on the other, then you've radically undermined the integrity of the discipline.
Having said that, I guess the question remains: Is benefit-cost analysis doomed forever, or is there a way forward?
Before we articulate a way forward, I would make the obvious point that EPA statutes are not by any means set up in a perfect way for the use of benefit-cost analysis. Each statute is a little different; there are some opportunities to use benefit-cost analysis in some sections of the Clean Air Act and some parts of the Clean Water Act and so on, but these opportunities are limited. Some court cases in the four or five years preceding the Trump administration actually were favorable to benefit-cost analysis, such as Michigan v. EPA. But an explicit incorporation of benefit-cost analysis is still not the way decisions are routinely made. The statutes don't require these analyses and don't even invite the analyses in large part.
At the same time, executive orders that have been issued going back to the Reagan era, and reaffirmed by several Democratic presidents since, always take note of the statutory limits but nonetheless establish that benefit-cost analysis can help us more fully understand the nature of the decisions we're making.
So, there's always been this tension. But even though we haven’t and won’t thrust benefit-cost analysis into the center stage of EPA decisionmaking anytime soon, there’s still great value in having a high-integrity set of analyses that can guide our thinking about the economic and environmental impacts of regulation.
The way forward could be for the new administration to make certain pronouncements, starting out, about benefit-cost analysis. It's a nuanced issue, but pronouncements could establish that benefit-cost analysis has an important role and should return, but with the level of integrity and quality to the analysis that was present before the Trump administration. There are several ways to do that.
One way is to speak publicly about the issues; a complementary way is through hiring. The new administration will be hiring a lot of new presidential appointees, some of whom undoubtedly will be asked at their confirmation hearings what they think of benefit-cost analysis. That’s an opportunity for the administration to go on record and make the case at the very outset, in the course of the confirmation process, for the value of benefit-cost analysis—not to embrace it in a way that departs from history, but to turn the clock back to the pre-Trump era.
Issues like co-benefits also clearly need to be revisited by the new administration.
Art Fraas: The Trump administration argued that the co-benefits—that is, other benefits that are ancillary to the objective of the regulation—should be downgraded and not considered in benefit-cost analysis. But the consensus among economists is that co-benefits are a critical element of benefit-cost analysis. That goes as well for costs—perhaps secondary cost effects also should be included in benefit-cost analysis.
Are these kinds of ancillary costs incorporated now—or, have they been in the past?
Art Fraas: Yes, although not fully. There's an argument, for example, that environmental rules may result in job loss. So, although unemployment benefits and re-training costs typically would be included, maybe substance abuse and suicide would not be.
Richard D. Morgenstern: There's always nuance in how far you go in quantifying benefits and costs. Anytime EPA does an analysis, they typically start out with some kind of list of the most important benefits, direct and indirect (i.e., co-benefits). They also start out with a list of the most important costs—also direct and indirect, such as job loss. It's always a challenge because doing these studies has constraints in terms of resources and time.
Historically, EPA has been criticized by the business community for going light on the costs and heavy on the benefits. Unsurprisingly, the agency also has been criticized by the environmental community, which often says that analyses look at the subtle costs but don’t focus on the benefits, too. That's a long-standing debate.
In terms of the climate-related regulations that the previous administration has rolled back, and the additional environmental regulations that could have been implemented by a different administration, EPA has lost more than four years.Richard D. Morgenstern
The Trump administration announced as a matter of principle that co-benefits were going to be downplayed or diminished in their analysis—whereas in the past, the constraints on evaluating either benefits or costs have been driven more by data and analytic capacity that the agency can access. I would say the Trump approach was a huge step in the wrong direction. And I believe that’s also the consensus of the economics community: it's been a step in the wrong direction.
The Trump administration actually used co-benefits in at least one regulatory rollback. Can you talk about that example and speak to how this type of inconsistency plays into the public perception of science and economics?
Art Fraas: The National Highway Traffic Safety Administration and EPA issued a joint regulation that rolled back an Obama-era rule, saying that substantial safety benefits would be associated with the rollback of the rule. Their argument was that if we adhered to the more stringent standards, fewer new cars would be purchased, and we would miss the safety benefits associated with relatively new vehicles. So, by citing co-benefits here and not elsewhere, for example, the Trump administration was inconsistent in its treatment of co-benefits.
The Trump administration was inconsistent in other ways, as well; it adopted inconsistent baselines, for example. That's another part of the declining integrity in the application of economic analyses that we’ve seen lately.
I think, as a broader matter, part of the erosion of confidence in benefit-cost analysis is that it seems like one could make reasonable assumptions and come out either way on most of these rules. You could show that a rule has net benefits, or you could look at the same rule with an analysis that differs in only one or two assumptions and show that the rule has net costs associated with it.
Richard D. Morgenstern: The inconsistency applied by the administration cuts across a number of areas. Typically, you’d think of the US government as having a position on benefit-cost analysis. And in fact, the Office of Management and Budget (OMB) has issued multiple guidance documents that apply to all regulatory agencies—not just EPA. What we've seen in the previous administration is EPA taking on some of the issues by themselves (with co-benefits as one example), and not OMB taking on those issues government-wide.
The transportation rule technically was a joint US Department of Transportation (DOT) and EPA rule, but it's my understanding that that piece of the analysis was brought to the table by DOT. So, what you have is two agencies of the US government, both of which have legal obligations to issue rules, but both operating under different criteria. One of them is generally trying to back away from co-benefits—that is, the Trump EPA—and the other one, DOT, is adhering to the traditional role of co-benefits. In addition, they have been criticized for the particular studies they used, which showed the reduction Art mentioned in terms of health and safety from lower sales of new cars; those studies have not been widely vetted in the academic community and have not been given the legitimacy that you would expect for studies that would quantify an effect like that. So, a number of different inconsistencies come across here, which I think the new administration has the opportunity to fix.
OMB could play a constructive, positive role by issuing, reinforcing, or updating guidelines. Discount rates and the social cost of carbon are other areas in which the Trump administration departed from mainstream thinking, and a new administration has the opportunity to fix it. The new administration could start through the appointments process, and then it can find specific rules which demonstrate the importance of these factors. In doing so, the administration would be signaling to the scholarly community and the broader regulatory community that it’s going to adhere to serious and mainstream economic principles when conducting these analyses. That doesn't obviate the fact that new rules may still be rejected because of certain legal constraints—but analysis should be free of the manipulation and bias that’s been introduced lately.
How can we incorporate benefit-cost analysis objectively, impartially, and optimally?
Richard D. Morgenstern: Economic science, like any other science, is complex and best evaluated by experts in the field. The peer review process tends to sort out the stronger from the weaker arguments, and I'd say peer review is essential. The Trump administration abolished EPA’s Environmental Economics Advisory Committee, which I think is very unfortunate, both in substance and for the broader anti-science message that such a move sends. The message was that we can just manipulate the regulations and the data, and we don't really need peer review. To their credit, EPA set up an ad hoc subcommittee (of which Art Fraas is a member) to review the agency's guidelines. And that seems like a step in the right direction. But the ad hoc subcommittee was established only after EPA had abolished the Environmental Economics Advisory Committee.
The other point I'll make is that, when legitimate uncertainty exists among experts, we have ways of incorporating uncertainty into benefit-cost analyses—and this has been done for many years at EPA. Whether it's done rigorously and vigorously enough, one can debate. But one can subject the uncertainty analysis to peer review as well.
Art Fraas: To some extent, I agree that the statutes certainly did not think about benefit-cost analysis very often. Over the past 40 years, though, benefit-cost analysis has intruded into a lot more areas of the statutes than one would have expected in 1970. And benefit-cost analysis also—perhaps unfortunately—has taken on the role of defending the rule after the rule already has been promulgated. So, the agency can say, “We adopted this stringent regulation (or this rollback), and look: it has net benefits.”
When legitimate uncertainty exists among experts, we have ways of incorporating uncertainty into benefit-cost analyses... Whether it's done rigorously and vigorously enough, one can debate. But one can subject the uncertainty analysis to peer review as well.Art Fraas
A second point is, maybe benefit-cost analysis has taken on a reputation of being able to provide a very precise answer, when in fact substantial uncertainty often is associated with the choice of assumptions that underlie the benefit-cost analysis. So, it should not be too unsettling that by changing one or two assumptions, you can substantially swing where the benefit-cost analysis points.
Richard D. Morgenstern: Looking back over EPA’s 40-year history of doing these studies, benefit-cost analysis increasingly has been useful and has helped the agency think through complex issues. A growing number of people inside EPA would recognize that it's also true that Congress, in more recent legislation (I'm thinking of the Safe Drinking Water Act of 1996), explicitly introduced benefit-cost analysis as a criterion—whereas in the earliest legislation in the Clean Air Act and Clean Water Act, benefit-cost analysis was either omitted or was treated negatively. The implication was that other criteria should drive the decisions.
Having said that, I think the biggest limitation of benefit-cost analysis probably is not its treatment of discounting or co-benefits—although both of these are important problems—but rather the agency's failure to consistently consider alternatives at the very beginning of the regulatory process. There's a tendency—both a human and an institutional tendency—that once you point in a certain direction, you tend to build a case to support that direction. We all do that—it's human nature. And if I were going to advise a new Biden administration, I would probably put the most emphasis on instituting careful, up-front consideration of alternatives. Maybe not a full benefit-cost analysis done on each option, but at least some analysis and review, which could then play a role in further, more specific, and more detailed regulatory development.
Art Fraas: I think that's a great idea. Within EPA and other agencies, establishing a rule involves the development of the rule, with the economic analysis stovepiped separately—which means that program decisions about the rule don't get informed by the regulatory analysis. In some cases, the rule development and economic analysis come together only at the point when the assistant administrator is making a decision about whether to go forward with the proposal. It would be useful to include, at the beginning of the rule development process, some analysis of alternative approaches.
Another point is that, while I think it would be hard to set up, it would be worthwhile to have a third-party institutional review of agency benefit-cost analysis that’s separate from the administration. Economics panels currently don’t tend to comment on the Regulatory Impact Analyses (RIAs) of rules. Maybe a really important rule, like the Clean Power Plan, gets that kind of attention. But a lot of the major rules don't. Most rules get comments from the affected industry or environmental organizations. But there’s no institutional, third-party review of the quality of RIAs.
Richard D. Morgenstern: I hate to disagree with my esteemed colleague, but I think the regulatory development process is already quite complicated. When RIAs were first introduced back in the 1980s, their cost (recognizing inflation) was a couple hundred thousand dollars. Now they cost several, oftentimes multiple, millions of dollars. For big rules especially, these reviews can cost a lot. Since public comment periods happen with these rules, and the OMB conducts reviews, the addition of RIA reviews generally would not be worth the additional time and resources involved.
Other than the lost time involved, which is an irreversible loss—if you have the commitment and the willingness ... then almost anything is reversible.Richard D. Morgenstern
However, I do have another thought—one that I think Art will agree with—which is that EPA should embrace, as the government is starting to embrace, retrospective analysis of regulation. That is to say, let's look back at some of these rules and see how they’ve worked out compared to what we thought.
Art and I have been working together in this area for some time now. Our work so far has revealed that in some situations, the agency has gotten the costs wrong by understating them, along with quite a few situations in which EPA has gotten the benefits wrong by understating them. By looking back at rules, you can find many of these examples—and we have a new paper coming out on the subject that talks about lessons we can draw and ways of advancing this work in the future.
The agency can help with the implementation of retrospective analysis by building some elements into the design of rules that will make retrospective analysis easier to do later on. And EPA can identify alternatives that could be evaluated in this way—for example, EPA could investigate the distributional impacts of major environmental problems, or they could initiate other data collection that would facilitate retrospective analyses. The value of this would be considerable.
Retrospective analyses certainly need not be done for all rules, but they could be done for major rules. It seems like this kind of practice could help increase the legitimacy of benefit-cost analysis and help improve it over time, which I think is an objective that everybody seeks.
Art Fraas: Just to defend my idea a bit: I think this kind of third-party review would be separate from the rule development process, and one way to do it could be a National Academy of Sciences effort to take, for example, 20 RIAs to evaluate and critique, suggesting ways to improve the RIA methodology. At one time, the Congress approved but never funded that kind of systematic review in the Government Accountability Office. It's a possible mechanism for keeping the agencies a little more honest.
And going back to the question of the extent to which recent rules and rollbacks are reversible or not: In terms of the climate-related regulations that the previous administration has rolled back, and the additional environmental regulations that could have been implemented by a different administration, EPA has lost more than four years. As the new administration comes in, it will take time to develop a new record to support climate-related regulation. And the new rules will have to go through formal proposal, take comments, respond to those comments, and then become a final rule. And there may even be a court review after that. So it may be a delay of at least six years—maybe even more—associated with the four-year interlude of the Trump administration.
Some have mentioned the “brain drain” from federal agencies, which might require a rebuilding of agencies such as EPA.
Richard D. Morgenstern: That's an important point; the brain drain has been significant in terms of the number of people who have retired early, quit, or been pushed out in some cases—along with the number of new people who have not been brought in at the junior level with the Trump administration. There's not as much fresh blood as there normally would be in the agency, which is a problem.
Now that we’ve been talking about the potential changes that the new administration can make: Are those changes going to be durable?
Richard D. Morgenstern: I think the problem with durability, as we've seen in recent history, is that there's very little that's truly durable—certainly not on the regulatory side of affairs. Changing rules is possible; established procedures exist for doing that. And if rules are not the operative mechanism, then it's even easier to change guidance documents. The whole effort to undermine the use of benefit-cost analysis is really one of the mechanisms by which the recent administration has tried to change regulations—by simply changing and deviating from practices that have been established in executive orders.
In fact, some of the greenhouse gas rules from the Obama administration aren't really rules—they’re executive orders. And those can be turned around readily. If it’s an actual regulation, then you have to go through a more formal process with the Administrative Procedure Act. But that's totally doable, as we've seen in the past.
By that logic, is everything reversible?
Richard D. Morgenstern: Other than the lost time involved, which is an irreversible loss—if you have the commitment and the willingness to go through the sometimes lengthy, time-consuming, and resource-consuming process, then almost anything is reversible.
Art Fraas: I agree for the most part with that, but where a rule is fully adopted and industry largely has absorbed the costs of compliance, the push to change the rule dissipates. In addition, where the courts get involved, agency flexibility can be constrained in the future.
Maybe this is reaching a little too far, but the Supreme Court changed substantially during the Trump administration. Some are concerned over the potential reversal of the flexibility that EPA has been able to exercise in interpreting statutes; the Supreme Court might in a case find that those actions are not allowed. One possibility (though I'm not sure how it would be teed up) would be if Massachusetts v. EPA went back to the Supreme Court: it's possible that the Supreme Court could come out with a different decision. There's also a long-standing court decision, the Chevron decision, which defers to the expertise of the agencies that administer the statute. The present makeup of the Supreme Court may challenge that decision, and the thrust of that change would require Congress to be very specific in its direction to the agencies on what sorts of regulations they can adopt. And that would mean, for example, that EPA would be quite constrained in developing new regulatory programs for climate change under the Clean Air Act.
Richard D. Morgenstern: There’s also an active debate in the legal community on the extent to which the court—even before Amy Coney Barrett’s confirmation—might take an opportunity to interpret Chevron in a way that constrains the agency. That's certainly a possibility, and I think it raises a question: How does EPA in the new administration go forward with climate regulation? And the answer, I think, is: very carefully, because of potential legal challenges.
You could imagine a situation in which the agency invests a lot of time and resources in developing a rule, and then it encounters legal challenges. Potentially, the rule even could be overturned by the Supreme Court. Thus, the agency needs to be cautious in how it proceeds.