What Every Environmental Regulator Needs To Know About Cost-Benefit Analysis

If you are the person who signs off on a rule to clean up drinking water, to retire a coal fleet or to keep the last spawning run of salmon alive, listen closely.

The analytical scaffolding that is supposed to make your job neutral — cost-benefit analysis — is swaying under its own weight.

Regulators around the world were handed a simple instruction in the 1980s. Show, in dollars and cents, that the benefits of a regulation justify its costs. Executive orders demanded it. Courts sanitized it. For forty years, the cost-benefit apparatus has felt like solid ground to stand on. Sound like a bureaucratic backwater? It isn’t.

The ugly secret of environmental cost-benefit analysis is not that the math is hard. The deepest problem is that the math was designed for a simpler planet. Certain ecosystems were treated as static warehouses of stuff. Future generations were permitted to be discounted into irrelevance. Uncertainty was massaged into a single, confident probability distribution. And the whole exercise was considered objective when it was quietly, persistently, political.

Let’s unpack that.

The Calculation That Ate the World

Cost-benefit analysis, or CBA, is a family of techniques for comparing the total expected costs of a regulation with its total expected benefits. Expressed in monetary units, subtracted, divided, scrutinized. If benefits exceed costs, the rule is judged efficient.

This logic has been embraced, then attacked, then embraced again by every presidential administration from Reagan to Biden. In the United States, a centralized review process is conducted by the Office of Information and Regulatory Affairs — the so-called “gatekeeper” of regulation. The gatekeeper concept was designed to make sure agencies did not regulate on a whim.

The operating principle is deeply intuitive. It mirrors how ordinary people make decisions. You weigh the cost of a new roof against the financial pain of a leaky one. So what’s the harm in asking the Environmental Protection Agency to do the same?

Here’s the kicker. The ordinary act of weighing costs and benefits becomes perverse when the “benefits” belong to people who have not been born yet, the “costs” include the complete extinction of a species, and the “risks” involve feedback loops that our statistics have never observed.

A market failure occurs when prices do not tell the truth about social costs. Pollution is the textbook case. Standard CBA corrects for one market failure but ignores another: the failure of our measuring instruments.

The Ghost of Microeconomics

Every economics student learns the basic catechism. Preferences are revealed by choices. Markets make trade-offs. A dollar is a dollar is a dollar. And so, when economists value a wetland, they tend to look at what people would be willing to pay to preserve it. They construct hypothetical markets for clean air the way you might construct a hypothetical market for your grandmother’s smile.

Grandmotherly smiles, by the way, have been valued in the literature. Read that again. Researchers have attempted to price the non-use value of watching a sunset, the bequest value of leaving a healthy planet to your grandchildren, and even the intrinsic value of endangered fish.

There is a long tradition of respectable economists finding these mental gymnastics absurd. One of the most famous is the 1974 paper by Arrow and Fisher, which demonstrated that irreversible losses deserve a special premium in the ledger. And yet, forty years later, the premium is still treated as an afterthought in most regulatory documents.

Consider the snail darter. Back in the 1970s, a tiny fish blocked the construction of the Tellico Dam. The Supreme Court ultimately halted the project. Congress, furious, exempted the dam from the Endangered Species Act. Tellico was completed. The snail darter survived, though it was downlisted to threatened. The lesson? Cost-benefit analysis could not resolve the conflict because the two sides used different units. The Tennessee Valley Authority used kilowatt hours. The fish had no accountant.

The ghost of that conflict haunts every modern battle over northern spotted owls, delta smelt, and the manatee.

The Environmental Economist’s Toolbox

You might expect a different kind of economist to have stormed in with better tools. And, to be fair, there has been progress. Let’s stroll through the modern toolkit.

First, there is non-market valuation. Travel-cost methods estimate the willingness of visitors to drive to a national park, gasoline money and time as the implicit ticket price. Hedonic pricing studies compare houses near green spaces to identical homes a few blocks away. Contingent valuation asks citizens directly: what would you pay to avoid another summer of wildfire smoke? Several methods have been refined by careful scholarship and blessed by peer review.

Second, ecosystem services accounting was embraced by major international efforts such as the Millennium Ecosystem Assessment and the more recent Dasgupta Review on the Economics of Biodiversity. Nature, in this accounting framework, is treated as an asset, not a free input. The value of insect pollination has been estimated at more than one hundred billion dollars per year. The carbon absorption performed by forests is, similarly, measured as a service that would be staggeringly expensive to replace with technology.

Third, discounting has come under closer scrutiny. For a generation, regulators were told to discount future benefits using the observed rate of return on private investment. This handed enormous advantages to present-day polluters. At a seven percent discount rate, one dollar of avoided climate damage two hundred years from now is worth less than two cents today. At a typical private rate of return, catastrophic losses in 2300 are barely a rounding error in today’s spreadsheet.

That is why the British government’s Stern Review on the Economics of Climate Change was such a bombshell. Stern, in 2006, argued that standard discount rates are ethically indefensible when we are talking about the survival of entire regions and cultures. He used a near-zero pure rate of time preference. Economists erupted. But the core insight stuck. Discounting is not a law of nature. It is a moral choice, usually made by men in windowless offices.

How to Redesign Cost-Benefit Analysis for Turbulent Times

So where does this leave the actual regulator? Your inbox is filled with comments from industry lawyers demanding rigorous analysis. Your case docket is filling with citizen suits. Your professional credibility rests on your ability to justify decisions with evidence.

Here is the secret of cost-benefit regulation design, and it has been hiding in plain sight for decades. The answer is not to abandon quantitative analysis. The answer is to make the analysis wrong in the right places. Catastrophic risks must be treated differently from nuisance risks. Irreversible actions demand a precautionary margin. Distributional harms cannot be cancelled out by aggregate net gains.

The following table outlines the contrast between the old textbook paradigm and a climate-informed regulation design.

That’s a substantial shift in analytic culture. Shifts of this kind are not accomplished by a Twitter thread. They are accomplished by redesigning the rules of the game — by a series of concrete steps that an agency can take without waiting for a new statute.

Step 1: Take the Discount Rate Seriously, Because It Does the Heavy Lifting

If you change one parameter in an environmental cost-benefit model, make it the discount rate.

The pure rate of time preference is the rate at which society trades present consumption against future consumption. High rates communicate impatience. Low rates communicate stewardship. Every regulator is, whether they admit it or not, picking a side in a intergenerational equity debate.

For environmental regulation design, declining discount rates are the most defensible approach. Why? Because the far future is clouded by uncertainty about economic growth, technological change, and catastrophic feedback. A constant discount rate implies that the future is exactly as risky as the present. Shakespeare’s observation about the readiness is all — no, the discount rate is all.

Several European governments, including France and the United Kingdom, have already adopted schedules in which the discount rate falls as the time horizon lengthens. The U.S. Office of Management and Budget still mulls a circular requiring a 7% rate, which operates as a heavy thumb on the scale against the future.

Step 2: Put a Price on the Unpriced, but Keep a Watchful Eye

The second step in modern CBA design is the rigorous measurement of non-market benefits. You cannot manage what you do not measure, the saying goes, but you also cannot measure everything.

Quantitative ecologists and environmental economists have spent years building high-quality estimates for things like wetland storm protection, mangrove carbon storage, and urban tree shade. Those values should be the starting baseline in every analysis. Regulators have a tendency to assign a zero to anything that is difficult to measure. Zero is rarely the right number. When the measurement error is enormous, the correct treatment is a range, not a zero.

Natural capital accounting is now being institutionalized. The United Nations adopted the System of Environmental-Economic Accounting; dozens of countries are compiling accounts for their forests, fisheries, minerals, and soils. Useful data is emerging that can feed directly into regulatory design. If a ministry of planning can look at the national balance sheet, then so can a local utility board.

The political trick is to make these estimates transparent. Non-market valuation will always be attacked by someone with a calculator, which is why the models and the survey instruments should be published in full. One should let the light shine in, as Justice Brandeis supposedly said. Actually, sunlight is said to be the best of disinfectants — a phrase often attributed to him.

Step 3: Confront Irreversibility Head-On

Extinction is forever, and that fact is too often buried in chapter nine of an environmental impact statement.

Arrow and Fisher’s insight from 1974 was simple. If a decision is irreversible and the future will bring new information, then there is a value to waiting. Waiting has an option value. Every investor understands this. You do not bulldoze a building that might contain a lost da Vinci painting simply because the land is worth more as a parking lot.

The regulation of deep-sea mining offers a live example. The ocean floor contains polymetallic nodules that could yield cobalt for electric vehicle batteries. But the same sediment hosts ecosystems that scientists have barely cataloged. In this case, the option value of not mining is enormous. The potential technological substitutes for cobalt, meanwhile, are advancing rapidly. Wait and wait and wait.

Environmental regulation should include a rule that says catastrophic consequences receive a weight that is not derived from smooth probability curves. This is sometimes called a “safe minimum standard.” When the outcome is irreversible and potentially catastrophic, cost-benefit analysis must yield to a precautionary constraint.

Step 4: Use Distributional Weights

Another uncomfortable secret is that a simple sum of benefits and costs can justify almost any form of environmental injustice.

Here is the basic problem. A dollar of benefit received by a rich person is summed identically with a dollar received by a poor family in a flood zone. But the marginal utility of money is higher for the poor. The flood zone family suffers more, in welfare terms, than the same loss experienced by a billionaire who owns a beach house in Malibu.

Contemporary research on inequality and regulation suggests that the same aggregate net benefits are judged very differently when you display the distributional consequences. A regulation that reduces asthma rates among children of color has a higher social value than one that reduces allergies among hedge fund managers.

To design good rules, an agency must compute the net benefits for at least three groups: the bottom income quintile, the middle, and the top. This transformation is cheap to run. It hardly changes the complexity of a model. Yet it exposes the hidden choices that are being made by the modeler.

Distributional weights have been proposed in economics for more than a century. They were always rejected as “too political.” The irony is that no weights at all — which is to say, weights of one for every dollar regardless of whose hand it falls into — is also a political choice. And an indefensible one, at that.

Step 5: Design the Institution, Not Just the Rule

The final lesson comes from the field of institutional economics. The quality of a regulation is determined by the incentives surrounding the regulator.

If analysts know their jobs will be ended by political revenge for an unwanted finding, you will see fudged assumptions. If the agency’s career staff are allowed to update their models as new data arrives, you will see trust. Every environmental regulatory agency on Earth is a monument to this simple trade-off.

Regulators should be required to publish a retrospective review of their major rules after five or ten years. Did the predicted benefits appear? Were the costs overestimated? Retrospective review used to be a popular talking point in Washington, DC, but it was rarely funded. We are spending enormous sums on prospective analysis and almost nothing on the empirical confirmation of whether we got it right.

The ideal institutional design is something like a learning loop. The rule is set. Outcomes are measured. Models are updated. Then, if necessary, the rule is revised. Such adaptive management is standard practice in fisheries, forestry, and water resources. It ought to be standard practice in the regulation of greenhouse gases, too.

There’s a reason this pattern recurs. Ecology is a turbulent, continuously surprising science. A one-time forecast that presumes stationarity is guaranteed to be wrong in ways that are expensive or irreversible.

The Question No Model Can Answer

Let’s return to the moment when you, the regulator, are staring at the final benefit-cost table. The numbers say the dam should be built, or the mine should be opened, or the oil lease should be auctioned. But your instinct says no.

Is the instinct just a failure of nerve? Or is it a warning from a deeper cognitive system that noticed something the spreadsheet did not?

Here’s the thing. Cost-benefit analysis was created to make decision-making more rational. But in environmental economics, rationality must expand to include the complex systems on which our economies depend. There is no economic value to be calculated if the biosphere ceases to support human civilization. Dr. Seuss’s Lorax, who spoke for the trees, was trying to explain this concept to the Once-ler. The Once-ler kept counting Thneeds.

Modern environmental economics is a discipline at war with its own legacy. It is brilliant at valuing clean air after an EPA regulation is analyzed, but it was slow to value the atmosphere before it was polluted. It is rigorous about interest rates but squeamish about extinction risks. That is changing.

A new generation of economists has been recognized with major academic prizes for demonstrating that climate damages are nonlinear, that economic growth itself is threatened by warming, and that catastrophe is a measurable event inside the model. The old assumption of smooth, incremental damage has been replaced by convex functions. This is not an arcane statistical point. It changes policy dramatically. It means the optimal carbon price is much higher than conventional estimates predicted.

What would Pigou say? Arthur Pigou, the father of welfare economics, believed that the goal of government is to maximize social welfare. He had no affection for false precision. He would, one suspects, be horrified by the ritualized pseudoscience that often masquerades as rigor.

Decision-Making Under Deep Uncertainty

One of the latest frameworks to influence regulation design is called Robust Decision Making, or RDM. Rather than asking “what is the probability of each outcome?”, RDM asks a different question: “What actions will produce acceptable outcomes across a wide range of plausible futures?”

The regulator no longer pretends to know the probability distribution of the climate in 2080. Instead, dozens of futures are simulated, and the virtues of competing rule designs are compared. This approach was developed at RAND Corporation and has been applied to water management in Southern California, and it has influenced climate adaptation planning in many states.

RDM forces no single answer onto the policy. It gives managers the confidence to choose infrastructure that is flexible, reversible, and adjustable.

Analogous logic should permeate the clean air act, the clean water act, and every statute where uncertainty is the rule rather than the exception. If a rule is hard to reverse, you bear the burden of proof. If you can adjust next year, you can take a step forward and measure.

Not every regulation can be temporary, obviously. Bans on hazardous chemicals are not reversible in any meaningful sense. But regulatory design should always try to learn.

Let’s briefly circle back to the title of this article. What does every environmental regulator really need to know about cost-benefit analysis? The answer might feel anticlimactic. Know who holds the pen. Know what assumptions are sitting inside the spreadsheet, silently shaping the outcome. Know that the discount rate is a moral statement.

And know that the beauty of a well-designed analysis is not that it creates certainty. It is that it maps the uncertainty honestly, so that politicians, not anonymous modelers, are forced to make the hard choices.

None of this means that cost-benefit analysis should be thrown overboard. Good regulation needs tools of comparison. Markets still matter. Benefits still deserve to be measured.

We just need to abandon the fiction that a ratio is a verdict. It is evidence. A regulator who treats evidence with humility, who is transparent about the limits of every estimate, is worth more than a computer model run by a thousand contractors.

The future will not be summed up in one number. It will be navigated, year after year, through careful attention to feedback, distribution, and the long view. That is the secret that good environmental economists have always understood. It is time the rest of the world caught up.