Notice & Comment

Reviving Rationality: Three Cheers, Four Caveats, by Timothy J. Brennan

*This is the first post in a series on Michael Livermore and Richard Revesz’s new book, Reviving Rationality: Saving Cost-Benefit Analysis for the Sake of the Environment and Our Health. For other posts in the series, click here.


This recent book by Livermore and Revesz (LR) follows their 2011 book, Retaking Rationality: How Cost-Benefit Analysis Can Better Protect the Environment and Our Health. The important and, sadly, necessary lesson from their earlier volume is that cost-benefit analysis (CBA) can and often does work to the benefit of those favoring environmental and health regulations. It is not inherently a tool those against regulations can and do employ to serve private interests at the expense of the broader public.

Just when LR’s important contribution toward making cost-benefit analysis safe across the political spectrum looked secure, along came Donald Trump. Skeptics of cost-benefit analysis and its underpinnings in economics (that is, the intended audience for LR’s first book) saw this as evidence that CBA really was nothing more than an excuse to deregulate and pad the wallets of corporate interests.

This book is LR’s response. They persuasively argue that the core of the Trump regulatory program violated CBA principles. Trump’s order to eliminate regulations that imposed the same level of cost as any new regulation—the equivalent of a “regulatory budget”—ignored whether the deleted regulations had benefits that outweighed that cost. His proposed “two for one” rule, that no regulation could be adopted unless two others were dropped, similarly ignored CBA principles.

In fairness, a “regulatory budget” can be a response to a finding that regulators systematically overstate benefits or understate costs. A “two for one” rule can be a way to force agencies to weed out regulations that no longer pass a CBA. However, the response is not to reject regulations that pass a CBA test, but for regulators to do CBA better.

In what follows, I’ll make some occasionally critical observations on (and explain) four topics: incorporating private benefits into a CBA, ethical concerns beyond distribution, the importance of recognizing the losers, and co-benefits. Any criticism, however, pales beyond the overall value of LR’s contribution. Their Reviving Rationality satisfies a CBA, without question.


Behavioral economists have drawn attention to the possibility that people fail to act in their best interests due to cognitive limitations or psychological biases. These choice failures have led many to include what some call “private benefits” in CBA calculations, and LR largely join their ranks. “Private benefits” are those generated by a regulation that the recipients could have obtained on their own, were their actions revealing their true rather than mistaken preferences.

An example may help. Suppose that the government determines that consumers place too little value on high gas mileage when buying cars, and for that reason imposes a minimum average miles per gallon that a car company’s vehicles must achieve. This regulation increases the price of heavy sport utility vehicles relative to compact cars, leading many to purchase the latter instead of the former. “Private benefits” would include the gasoline cost savings when those who would have purchased SUVs get compacts instead. Under conventional economics, those purchasers would have said that they value the SUVs more than the compacts despite those gasoline cost savings, and thus they are worse off. Some other benefit, such as from reduced pollution, would be needed under CBA to outweigh this cost.

LR incorrectly treat behavioral economics and private benefits as only a slight adjustment to the overall CBA framework. This reflects a fundamental difference in worldview. Within conventional economics, the goal of CBA is to identify just those outcomes that would result if there were no market failures that impede the ability of people to purchase cleaner air, safer highways, additional fire engines, you name it. The appeal of CBA (and market outcomes it is designed to replicate) is that it depends not on the views of some independent authority but, rather, how much the beneficiaries would pay compared to the costs those on the other end would bear. Instead, the behavioral economics view is that the goal of CBA is to independently ascertain benefits and costs and choose those outcomes where the former exceeds the latter.

Undoubtedly people make mistakes, but taking that into account puts us in a far different empirical, political, and philosophical context than that which supports CBA.


The most familiar ethical consideration in CBA arises because its measure of benefit and cost, willingness to pay, depends on wealth or income. LR recognize this, but do not take on as explicitly ethical the issue of standing—whose benefits and costs count, and by how much. Climate policy brings these to the fore. As LR note, the on-the-record difference between Obama and Trump administration policies to regulate carbon emissions from electricity generation hinged not on climate science disputes, but on standing of non-US and future beneficiaries of US climate policies. These two differences reduced the “social cost of carbon” incorporated in proposed electricity sector regulations from $43 per ton under Obama to $1 under Trump.

The standing issue for non-US residents comes down to a calculation like this: Suppose a US policy would impose a net cost of $5 on US citizens but produce $10 in benefits to non-US residents. Perhaps that policy should be adopted, reflecting utilitarian-like arguments or Kantian duties. However justified, it is not a matter of CBA economics, at least not in the absence of a mutual international agreement.

The future generations issue is not usually treated as a standing issue. It is typically not either/or, but how much a dollar in benefit to those generations should count in a CBA against costs borne in the present to produce those benefits, measured by the discount rate. One can think of this as “partial standing,” where the lower the discount rate, the greater the standing of future generations. Economics has much to contribute to the discount rate discussion, in particular, illuminating the opportunity cost of climate policies over other investments that would benefit future generations. However, it cannot tell us our obligations to future generations.

LR discuss standing, but treat it as part of the science of CBA rather than an ethical choice to be made before doing CBA. I think that many fear that acknowledging that an issue is ethical makes it somehow arbitrary or “squishy.” That is unfortunate, if for no other reason that claiming that value judgments are scientific forces those with other values to become “science deniers.”


A second ethical consideration meriting more attention than LR provide is whether to compensate the losers. This is not the same as whether to give greater weight to the willingness to pay of poor people compared to wealthy people, although it has consequences for that issue. If the winners compensated the losers, as is generally possible in principle for policies that pass a CBA, everyone would gain, eliminating the need to change the balance between winners and losers because one side may be richer or poorer. But whether to compensate is different from whether to weigh benefits and costs differently, although failure to compensate can necessitate such a weighting.

There may be another benefit to compensating. It is hardly a headline to point out that many in this country and around the world feel alienated from their governments. One reason could be a sense among many that the government treats them as only a number in a policy calculation. The aversion to CBA that has motivated LR may reflect this.

Agencies that implement regulations supported by CBA are rarely if ever in a position to compensate those who lose. The Environmental Protection Agency cannot compensate workers in a polluting sector who lost their jobs. A modest proposal to show that regulators using CBA are not treating those on the losing end as mere numbers would be that any regulatory impact assessment include suggestions as to how Congress, which controls the budget, might compensate the losers directly or with programs to mitigate the downsides.


LR devote a chapter to “indirect benefits” or “co-benefits”, the (net) benefits that a regulation may create beyond the specific problem it was proposed to solve. For example, an EPA rule proposed to control mercury emissions passed a CBA largely because of reductions in other air pollutants that mercury controls would bring about.

LR devote time to this because of two Trump administration errors they identify. The first is inconsistency. LR make a persuasive case that the Trump administration chose when it wanted to count outside costs but not outside benefits, e.g., in looking at alleged auto safety effects when deciding against stronger fuel economy standards that would reduce the size of cars.

The other error LR and many others identify is the simple failure to count co-benefits. At one level, this seems a no-brainer. A CBA should count all benefits and all costs, as best it can. The argument seems so obvious that one perhaps should ask if something else is going on.

Finding what that might be begins with an unstated question: Why is a certain class of benefits called a “co-“benefit? The answer is, “Because they weren’t included in the name of the regulation.” The co-benefit question could be recast as asking whether the name of the regulation matters.

Arguably it does, not for economic reasons, but to facilitate the regulatory process. If the agency wants to count effects on Y as well as X in the regulation, it can propose to regulate X and Y and not just X. There would be no “co-benefits”, only “benefits.” Notice and comment mandated by the Administrative Procedure Act might turn up better ways of addressing both X and Y. LR’s analysis of the legal standing of the justification for including co-benefits is persuasive, but I cannot help but think that there must be some reason why the issue might be more subtle.


I’ve highlighted some differences in approach and emphasis with those in LR’s book. However, I want and need to reiterate that those differences pale in context with the importance of LR’s important overall mission—persuading skeptics that cost-benefit analysis is not just reasonable but is on their side when they are right and can help support their cases. LR are largely successful in achieving it. The more this book is circulated, the better, especially in public policy classrooms, where it might have the most influence and do the most good for the longest time.

Timothy Brennan is a Professor of Public Policy and Economics at the University of Maryland, Baltimore County and a Senior Fellow at Resources for the Future.

Print Friendly, PDF & Email