Cost Benefit Analysis in U.S. Environmental Policymaking

[00:00:00] Brian Murray: This is Duke University. [00:00:05] Billy Pizer: Well, greetings. On the occasion of the 2016 Conference for the Chinese Society for Environmental Economics at Duke Kunshan University, a number of us here at Duke University were excited about the potential to make a contribution. This is particularly true given the topic of cost-benefit analysis and environmental policy meeting that's the theme of the conference. But unfortunately, none of us could travel to China. So instead, we decided to provide our contribution through this taped discussion of the topic. So before we get into a substantive conversation, I wanted to start with each of us giving a brief introduction for ourselves. I'm Billy Pizer. I'm an economist and professor at the Sanford School of Public Policy at Duke University. I've spent 20 years research and writing about environmental policy and four years in the federal government actually working on it. And my main connection with cost-benefit analysis has been work estimating the cost of environmental regulation and climate change in particular. But I've also spent quite a bit of time measuring the social cost of carbon used to compute the benefits of climate change mitigation. So, Jonathan, you want to introduce yourself? [00:01:12] Jonathan Wiener: Sure. Thanks, Billy. Greetings. I'm Jonathan Wiener, professor here at Duke University. It's a pleasure to be with you. I have been a professor here at Duke for more than 20 years, teaching courses on environmental law and regulation of risks. Before I came to Duke, I spent seven years working in the U.S. government, both in the Department of Justice and at the White House Council of Economic Advisors. And there I worked on what the guidelines should be for the use of cost-benefit analysis in environmental policy. Brian? [00:01:47] Brian Murray: I'm Brian Murray. I'm the interim director of the Duke University Energy Initiative, and I'm the director of the Environmental Economics Program at the Nicholas Institute for Environmental Policy Solutions and a research professor at the Nicholas School. Before coming -- and I've been at Duke for about a little over 10 years. And before I came to Duke, I was with an independent research institute that did a lot of advisory work for the U.S. EPA, including the development of methods for benefit-cost analysis. So I'm going to draw some from my experiences back then. [00:02:18] Billy Pizer: Excellent. Well, I didn't say anything about the format. What I'm going to do is I'm going to have each of us speak for a few minutes, generally about the use of cost-benefit analysis in environmental policymaking, particularly in the U.S., but potentially elsewhere. And then after each of us give some opening remarks, I will ask questions or potentially other people will ask questions of each other, and we'll do all of that for about 30 minutes. So, Jonathan, do you want to start? [00:02:44] Jonathan Wiener: Sure. Let me give a quick overview of the use of cost-benefit analysis in environmental policy in the United States. Of course, the idea of cost-benefit analysis goes back a long way, even back into the 18th century when people like Benjamin Franklin and Jeremy Bentham and the French engineer economists were developing the idea of comparing the pros and cons of a decision more accurately and quantitatively through the use of cost-benefit analysis. But it really became part of U.S. environmental law in the 20th century. And we could talk even about the Flood Control Act of 1936, which required cost-benefit analysis. But the main focus of a lot of attention is the environmental laws enacted in the 1970s. So I want to make three quick main points about that. First is that there's been a growing consensus in the United States in favor of the use of cost-benefit analysis in some version to evaluate regulation, including environmental regulation, but other types of regulation as well. And some of the laws enacted in the 1970s required cost-benefit analysis, some restricted it, and some were silent. And over time, the consensus has grown both through actions of the United States president. So every president of the United States since the 1970s, whether of the Democratic Party or Republican Party, has required the use of cost-benefit analysis to evaluate regulation. And through a White House office called the Office of Information and Regulatory Affairs, OIRA, the presidents have overseen that regulatory impact assessment process and the use of cost-benefit analysis. And in addition, the courts have become more open to or favorable to the use of cost-benefit analysis, including in some recent decisions in which the US Supreme Court has held that laws that don't even specifically require cost-benefit analysis are open to federal agencies using cost-benefit analysis. The second point is that this is spreading around the world. So more and more countries are adopting a regulatory oversight system using economic evaluation of regulation. So after the United States, starting in the 1970s, Australia, the European Union with its Impact Assessment Board now renamed last year the Regulatory Scrutiny Board, Mexico, South Korea, all around the world countries are increasingly adopting this form of oversight. And the third key point is that this method of analysis is not just to check or to hold back regulation or environmental protection. It's actually better understood as a guide to make good choices to protect the environment and public health in the best way as possible. And there are several important environmental regulations in the United States that were promoted by cost-benefit analysis, including the phase-out of lead in gasoline, the phase-down of CFCs to protect the stratospheric ozone layer, the restrictions on trans fats in foods, and the restrictions on fine particulate matter, or PM2.5, in air pollution, all strongly favored by cost-benefit analysis. [00:06:17] Billy Pizer: Excellent. Brian, do you want to make your opening remarks? Sure. [00:06:22] Brian Murray: Yeah, I was going to speak to a couple of sort of broader conceptual and technical issues associated with performing benefit cost analysis. Some of this is drawn from the guidance that has been written up by the Office of Management and Budget to guide agencies on the use of benefit cost analysis for agency decisions. So the one thing is that it really -- it's very important for the agencies to identify the alternative courses of actions and the benefits and costs -- identify the benefits and costs of those actions and the timing. And I'm going to get to the timing in just a second. But the key issue about identifying these benefits and costs are that some of those benefits and costs can be monetized. And that's sort of the ideal in terms of applying an economic method like benefit cost analysis. Some can be -- can't be monetized, but they can be quantified. So we can figure out the physical flows of the benefits and costs as a result of some action. And some can only be identified. They can't be modified or quantified. And the agency -- I'm sorry, the Office of Management and Budget encourages the agencies still to -- to identify the benefits even when they're not monetized or quantifiable. So timing is really a critical element of benefit cost analysis, which means that if timing matters, then so does the notion of using discounting methods. And discounting methods means that a benefit or a cost that's incurred in the future is worth less in today's dollars than if it was incurred today. And so Jonathan alluded to sometimes the notion that maybe benefit -- some perceived benefit cost analysis as being a hurdle to environmental regulations because environmental regulations often require costs up front, but the benefits are incurred in the future. But that -- but it doesn't necessarily have to be that way. But I think that's where the genesis of that comes from. So why would -- why would you want to discount? I mean, you discount basically because the general notion is that decisions made on behalf of the government should reflect the preferences of the governed. And so if you figure that -- first of all, funds have alternative uses. If they're not going to be used to clean up the environment, they can be used for other things, whether it's other -- to produce other environmental outcomes or produce other goods and services that are of value to society. Another reason to discount is that people in everyday life reveal a strong preference for -- for discounting. I mean, we make decisions about how to invest. We incur insurance costs. And we -- we basically are risk averse in many ways, in part because we discount the future. And -- and -- and to the -- to the -- you know, in favor of actions that we take today. So -- and then another reason for doing it in large-scale public policy is the notion that future generations are -- are likely, based on history, likely to be wealthier than they are today. So the notion of reserving resources today for future wealthier generations raises some questions about intergenerational equity that is both controversial but also an important part of the discussions about benefit-cost analysis. There are often arguments leveled against discounting, basically the notion that it's biased against future generations, or -- or -- or -- or policies that generate benefits far into the future like climate change. So with -- there's several key issues with costs, though. One is, you know, in the -- in the sense of benefits, it's fairly straightforward, right? You know, we enact regulations to -- to produce cleaner air or safer food or the protection of -- bio-diverse ecosystems, but -- and so we -- we try to identify, quantify, monetize those. On the cost side, though, the real question is, what is society giving up to achieve those benefits? And there's -- there's two sort of main concepts and costs to think about. And one is direct sort of opportunity costs. What could the resources that are spent on the equipment and labor to clean up the environment -- what could those benefits -- what -- what could those resources produce elsewhere in society? What are the opportunity costs of tying them up to clean the air or clean the water? And so we typically value those by the market prices of the resources that are used -- the labor, capital, energy, materials. But there are also ancillary benefits and costs as well. So, for instance, there are greenhouse gas policies in the United States, and -- and there being many -- and I think Billy's going to talk about this in a bit, too -- there's, you know, sometimes they're evaluated based on the climate benefits, but they're also can be evaluated based on the other clean air benefits associated with taking the actions to reduce greenhouse gases. But on the cost side, too, there can be ancillary costs. So a safety cost might be associated with fuel efficiency standards. So more fuel-efficient cars might have higher rates of -- of driver mortality, and that's something that you would need to take into consideration. Regulation of the electric power sector might make the power supply somewhat less reliable if it's more dependent on intermittent renewable fuels, intermittent renewable energy. And then there are also concerns about innovation and loss of competitiveness. I think without closing -- because I know Billy wanted me to talk a bit about, you know, focus on costs. Let me just say there's two important distinctions to make in costs. And one is the real resource costs associated with any actions. And think of that as the goods and services required to comply with the regulation, the changes in practices or the changes in inputs. And then also the reductions in consumer and producer well-being due to regulation-induced changes in price or quantity. So in other words, if you produce less of a certain output, society consumes -- as a result of responding to the regulation, society consumes less and they lose some benefits there. And then the last thing is to understand the difference between those real resource costs and transfers, which is payments moving from one part of society to another. So one example of that would be with a pollutant tax, like a carbon tax. The polluting parties would be paying the government through tax transfers. And those are no net cost to society. That's just moves from one element of society to the other. But the polluting firms are also going to incur more expensive practices to reduce their pollution, and that is a net cost to society of the actions that they're taking. [00:12:44] Billy Pizer: What I wanted to do was briefly take a couple minutes to talk about the U.S. government's social cost of carbon. This is a dollar value that the U.S. government has established for valuing the benefits of climate change mitigation, particularly the reductions of carbon dioxide emissions. For example, the recent Clean Power Plan that regulates power plants. So this is a number that gets used explicitly in cost-benefit analyses on the benefits side. The central value is about $40 right now for emissions that are reduced this year, say. And that number rises over the next 20, 30 years to about $70 actually by 2050. And these are the dollar numbers that get multiplied by the emissions in order to figure out the benefits for reducing those emissions. So this idea of using a number on the benefits side for something like reducing carbon dioxide emissions actually highlights just two points I wanted to briefly mention. The first is that this valuation of benefits from reducing carbon dioxide emissions, reduced climate change benefits or the absence of climate change damages, it's an enormously complex undertaking. It's an enormously complex undertaking. You can look at the documentation that's on the EPA's website about how they did it. I've also been part of a two-year process under the National Academy of Science to review that. That report will be coming out in January if you're interested. But it's enormously complex. But nonetheless, that complexity has not been a reason that the government has shied away from doing it. It was actually a court decision that Jonathan could probably say more about that ruled that even though there was enormous uncertainty about the benefits from reducing carbon dioxide, reducing carbon dioxide, that the value was certainly not zero, and therefore it wasn't appropriate for the government to simply not value on those benefits. This kind of speaks to the point that Brian was saying, that in cost-benefit analysis or policy analysis generally, you see a variety of things that happen where the benefits are difficult to measure. Sometimes they just list the categories. Sometimes they try to quantify impacts without putting dollar values. And then in the cases where they're actually doing a dollar cost-benefit analysis, they monetize those impacts. So here the government was, or the courts were literally saying that you needed to use a value that was not zero. There's complexity, yet that's not an excuse for trying to put a number on it. The second point, I guess, which will lead into a couple of questions maybe, is that I would argue that cost-benefit analysis alone should not be the only element of policy-making and decision-making. Even if you're focused on it, there are obviously categories that are not valued. Even within the social cost of carbon, there are categories of benefits. For example, impacts from catastrophic damages or the risk of catastrophic damages that are poorly valued, if valued at all. There are other things like the impacts on ecosystems that are not valued. So it would be kind of unwise to assume that you can do everything with a mechanical calculation, even if it's informative. Just another thing that's very hard to get at in these analyses is distributional impacts. For example, on the social cost of carbon, on the one hand, we assume that dollar value impacts anywhere in the world are equally valuable as dollar impacts in the United States. But that doesn't get at the possibility that dollar impacts on a very poor household could be much more detrimental than dollar impacts on a very rich household. So those things can be difficult to get at. So with that kind of opening salvo from each of us, what I thought I would do is turn to a couple of questions. And I guess I'll start off with one that kind of came up from something that Brian said, and that was this issue of co-benefits or co-costs. And I guess one question, you know, for maybe for Jonathan, but Brian can also jump in, is, you know, how do you draw the boundaries over a cost benefit analysis? I mean, how do you decide what to include and what not to include? And what do you do about these co-benefits or co-costs? [00:17:19] Jonathan Wiener: Yeah, that's a good question. And sometimes those aspects are named with different terminology. For instance, the U.S. government's guideline to federal agencies on how to address these kinds of effects refers to them as ancillary impacts. I'm thinking there of a document called Circular A4, which you may find interesting, which sets out the U.S. government's guidelines on how federal agencies should conduct their cost benefit analyses for regulatory decision making. So it talks about ancillary impacts, both benefits and harms may be ancillary impacts, but they're the unintended or not the target focus of the regulation. And in some earlier work that I did with John Graham, we talked about risk-risk tradeoffs, that any intervention to reduce one risk may also affect other risks, perhaps beneficially or adversely. So I would say, first, there's a strong argument in favor of asking, calling on regulatory officials to consider these ancillary impacts or risk-risk tradeoffs or co-benefits and countervailing harms. Because otherwise policymakers can get very focused on just the one target they're looking at and they neglect the side effects. And that would be like prescribing medication without considering the side effects. And if you go back to the 18th century, it's really what Benjamin Franklin was saying was the key to cost-benefit analysis is getting all the important consequences in front of the decision maker at the same time. So on the other hand, Billy raises a good point. How far should you go? How many ripples in the pond should one investigate? And actually, the concept of cost-benefit has a possible answer to that, which is the decision maker would want to be informed about the ancillary impact that could influence the decision. So you would want to do an analysis of how much the policy could be improved by considering these ancillary impacts, but also the cost of the additional analysis, including the delay. So you could think of a meta cost-benefit analysis of the degree to which the cost-benefit analysis should include these ancillary impacts. At least intuitively, one would want to consider the major effects that could change your decision from, say, one regulatory alternative to another. [00:20:04] Billy Pizer: Brian, did you want to add anything? [00:20:06] Brian Murray: Yeah. I mean, I think it seems fairly clear that the agencies, at least in terms of looking at the analyses that have been done, that they feel like they have the authority to go beyond the direct benefits, and the Circular A-4 is fairly clear about that. I want to use one example, and this is the regulatory impact analysis from the Clean Power Plan that was issued. And so Billy and I have both alluded to the Clean Power Plan, but this is a regulation of power plants in the United States. It's currently tied up right now with the U.S. Supreme Court. That's a separate 30-minute conversation we could have. But in the analysis that the EPA did, there were ancillary benefits and costs that they considered. And the ancillary benefits were in cleaning up greenhouse gases that are going to reduce other pollutants, like particulate matter, which have very high damage costs associated with that. And so when they estimated the benefits of the Clean Power Plan, a substantial portion of the benefits actually came from these so-called co-benefits, from the other air pollutants rather than greenhouse gases. So that, you know, sort of helped in the benefit-cost analysis in terms of, you know, favoring action. On the cost side, though, that the agency also felt like they had to examine the reliability concerns that were being raised. So it isn't just a matter of how much more expensive it would be or wouldn't be with producing lower carbon-compliant forms of energy, but also would it create reliability concerns for the electric grid in the United States. And so I think they worked sort of in partnership with another federal agency that looks at reliability concerns in the United States and determined that there should not be any particularly strong reliability issues associated with the rule. But both of those would be issues of ancillary benefits and costs that the agency has taken into consideration. [00:21:56] Billy Pizer: That's interesting. You can kind of imagine that issues of air pollution benefits, carbon benefits, and concerns about reliability might be the same sorts of things that the Chinese might be thinking about in terms of regulating their power sector. [00:22:08] Brian Murray: Yeah. If I can add just one thing. The one thing, though, that the ancillary benefits can sometimes raise issues about, and it's one of the challenges actually with a clean power plan, was whether or not you're double counting benefits. In other words, so if these other pollutants are regulated already, are you really going beyond, are you really reducing pollution even further, or is it just going to basically stay at the pollution levels that are already allowed by regulations? And if it's the latter, then, you know, you could say that that's a double counting exercise. In other words, it'll just free up other people to pollute more. However, if you can convincingly make the case that it will reduce those pollutants, I think the case could be made that it's a legitimate co-benefit that you should put in there and that you're not double counting. [00:22:54] Billy Pizer: Interesting. I mean, it's almost like a general equilibrium analysis that you need to figure out is whether or not reducing the emissions of, say, sulfur and nitrous oxides at the power plants might be. It might be then used as an excuse to weaken emission controls at other sources. [00:23:10] Brian Murray: Yeah, or whether or not the pollutant itself is quantitatively limited or is limited in terms of its emissions rate, right? So if it's quantitatively limited, like some-- Like a capped. Yeah, if they're capped. Yeah, if they're capped, then you might not actually be getting any further benefit. But if they're admitted by rate, you may be. [00:23:29] Jonathan Wiener: So I think what this is illustrating is that we recognize increasingly that the world is very interconnected and that we don't deal with just one risk at a time, even if our policymaking is targeted at one risk. Really, there's a web of multiple risks that we're addressing. And it can be important but complicated to figure out for this policy decision, what will the set of multiple changes that it actually generates be in society? And as Brian is saying, we don't want to double count, that is, changes that would have occurred anyway or that are occurring because of some other policy. On the other hand, we don't want to neglect the multiple changes, the second order and third order changes, if they're important to the decision that come from that main policy. So it's valuable to keep that in mind, that is, what is the overall purpose of this analytic exercise in informing our decisions? And that suggests two other quick points. One is, as Billy was saying earlier, at least in the United States, use of cost-benefit analysis is a tool to inform decision-making. It is not the decision itself. So the decision itself still has to be taken by publicly accountable officials considering a range of factors, some of which are included in the cost-benefit analysis, but some of which may be difficult or not included in the cost-benefit analysis. And the second point is, we can actually learn about the ability of these analyses to encompass all of these multiple effects by looking back at some of these analyses. So a lot of the cost-benefit analyses in the United States have been done prospectively to try to decide whether to adopt a new policy or not. And all of these presidential executive orders since the 1970s have emphasized that prospective or ex-ante use of cost-benefit analysis. There is a lot of interest and demand for retrospective review to look back at past analyses and existing regulations and say, how well are these regulations actually working? And how accurate were those ex-ante cost-benefit analyses? And I think that's an area that's ripe for further research. We have a project at Duke University looking into how governments are doing retrospective review of regulation. And the possibilities for learning about improved analytic methods in the future. [00:26:12] Brian Murray: If I could just -- no, why don't you go ahead, Billy. [00:26:16] Billy Pizer: Well, I was going to say two things, but then I'm happy to let you jump in. This is, you know, not a scripted activity here. We probably have about, I would guess, five more minutes at least, then we should wrap up. The two things I just wanted to quickly mention, you know, one is, Jonathan and Brian, you both mentioned Circular A4 as being an interesting reference point that people might want to look at. The other thing I was going to mention, if people are interested in cost-benefit analysis in the U.S., as a reference document are the EPA guidelines for cost-benefit analysis, which are a really long document that the EPA has issued to guide cost-benefit analysis at that agency. The other thing is people can also look at the regulatory impact analyses, which I believe are required for all major environmental regulations and are published in the Federal Register, although a lot of them are sitting on the EPA website as well. So if people are interested in seeing what cost-benefit analyses actually look like in practice, they can do that. It's all in the public record. Brian, what did you want to say? [00:27:17] Brian Murray: I was just going to say, you know, it's interesting when Jonathan talked about retrospective versus prospective. One of the things I'm not sure gets fully appreciated, I mean, I think people get the notion, I'm going to look at the benefits, I'm going to look at the costs, if the benefits exceed the costs, it's a good thing. But where I think benefit costs can be even more effective is looking at ranges of alternatives. And so we could look at, if we just were saying thumbs up or thumbs down, you could have a benefit cost analysis that says thumbs up, that should be good. But it doesn't mean that the regulation itself couldn't be designed better to raise the net benefits. So there may be some other option where you could increase the benefits more than you would increase the costs, and that's something that you'd want to do. So that's one thing. The other thing is this point about really the decision, as you were saying, the decision really is with the decision makers. This is just a tool. And it's worth pointing out, and you alluded to this some in your opening comments, Jonathan, is that different statutes that the EPA abides by give them more and less leeway about how they can use benefit cost analysis. There are some rules that essentially say you really can't take costs into consideration, and there are others, like the major rules that basically say a regular thing. [00:28:27] Billy Pizer: The statutes. The statutes are what? [00:28:29] Brian Murray: I'm sorry. Yes. The statutes prohibit the use of benefit cost analysis explicitly in the decision making. It's another real issue about whether implicitly they're still making decisions along those lines. But that's, so I think that's really good. You know, I did one benefit cost analysis at the state level. And the state basically said do the benefit cost analysis, and then, but the rules that they use to make a decision is if it raised the cost at all, they weren't going to let it happen. [00:28:56] Billy Pizer: You mean cost of electricity or something like that? When you say cost. [00:28:59] Brian Murray: This was going to be cost of basically managing hog waste in the state of North Carolina. If it made it any more expensive, it was sort of irregardless of the benefits. So they had the leeway to do that. And at the federal government, there's a little bit more prescription going on about what they can and cannot do. A lot more prescription. I'll let Jonathan go. Do you want to pick up on that? [00:29:19] Jonathan Wiener: Yes. Well, just to agree with a couple of points that Brian made. First, on comparing multiple alternatives and trying to pick not just one that has net benefits, but the best option. So the U.S. presidential executive orders say that. So they say, for example, that not only should the benefits justify the cost, but they also say that the agency should try to choose the alternative that maximizes net benefits. And then the second point about that some of the federal statutes in the United States restrict cost benefit analysis. And that's interesting because in the 1970s, there are some laws that required cost benefit analysis and some laws that seem to restrict it. But it's very rare to find a statute that actually says to the agency, do not conduct or do not consider the costs and benefits. Typically, they used other language or they were silent. And in the 70s, 80s, 90s, courts often held that if the statute was silent about cost benefit analysis, then the agency could not use it. But in the last 10 years, that's changed. And the U.S. Supreme Court has held in at least two major decisions, one in 2009 about the Clean Water Act and one in 2015 about a special provision of the Clean Air Act that where the statute is silent or uses general language like appropriate, then the agency is either allowed or required to use cost benefit analysis. So I think that's moving more towards this consensus that I was describing both among presidents and the courts in favor of at least some version of cost benefit analysis. [00:31:03] Billy Pizer: So the question, we were talking about alternatives and how, you know, the idea is not just to find something as positive net benefits, but to find something ideally that has the most net benefits of some feasible choices. How important is the baseline? I mean, I keep thinking that, you know, if you were thinking about a carbon policy and your baseline was nothing at all, that would be one scenario. If your baseline was very stringent enforcement of tighter air pollution regulations in the future, that would be a different baseline. So it seems like, I'm just wondering how you think about this, but it seems like in addition to the questions of like, what's the policy choice that's in front of you and like, what's the best choice? It may, does it not also depend on kind of what you're assuming in the baseline as well? And is that at all controversial? Like, I'm also thinking like the people in China might be familiar with this, but the Energy Information Administration, for example, you know, has a policy or a internal approach to modeling where they only take the baseline as things that are enacted currently, and they don't really think about perspective regulations that are on the horizon. And that seems like that would affect some of their analysis, but I'm wondering how you guys think about that. [00:32:16] Brian Murray: You know, my quick thought about that is simply that the more aggressive the baseline, then the magnitude of the benefits should be smaller by the action that you're taking because you're getting a lot for free, at least in the benefit cost analysis parlance, because it's part of the baseline. You typically think of benefits and costs relative to a baseline. So that's one thing. The other is, you know, it's also you're moving farther up the cost curve and farther down the demand curve. So I mean, I guess the net benefits would generally be smaller. [00:32:45] Jonathan Wiener: I would just add that this is an issue in retrospective review as well. We may think, oh, if we look back, we can see the actual cost of benefits. But in order to do a good retrospective or ex-post review, we have to construct a counterfactual estimate of what the world would have been like without the regulation. So that requires constructing that baseline and trying to envision if this policy had not been adopted, how would the world have evolved? So that's a challenge. Right. Both ways in time. That's a challenge. [00:33:22] Billy Pizer: Okay. Well, look, I want to just end by thanking our virtual audience, wherever you might be, and also thanking Jonathan and Brian for their time this afternoon. Thank you, Billy. Thanks, Bill. And we should do this again. This was fun. Thank you, Junji. [00:33:39] Jonathan Wiener: Thank you, Junji. We hope to come visit you in person in China soon. Yeah, and feel free to contact us if you have any follow-up questions. [00:33:45] Brian Murray: Produced by Duke University.