by Jerry Cayford
The rigged rules that govern our economy are being rewritten right now. And the fight is fierce. “The most powerful agency you’ve never heard of” (as the media calls the Office of Information and Regulatory Affairs) is revising its main guidance telling federal agencies how to structure regulations. That is, OIRA is rewriting the rules that federal agencies must follow in writing their own rules that govern the industries they regulate.
What makes this rulemaking earthshaking is that the people doing it are trying to unrig decades of rigged rules, and getting pushback from powerful players. The magnitude of the stakes can be seen in the public comments on OIRA’s revision of its guidance, Circular A-4. It’s complicated, obviously, but there is one point on which everything else turns—OIRA’s most controversial and consequential proposal. I am going to explain that central point.
Here is a thumbnail. Agencies are required to use Cost–Benefit Analysis (CBA) to justify their regulations as increasing overall social welfare. A huge contributor to the rigged rules in our society is that this formally mandated Cost–Benefit Analysis has a logical fallacy at its core that systematically favors the wealthy: it defines social welfare as increased by more total wealth (productivity), regardless of who gets the money. This definition of welfare forces federal agencies to design their regulations to maximize wealth, which inevitably favors those who already have it, for many reasons that I throw together under the adage “It takes money to make money.” Think of wealth production as an industry with economies of scale and barriers to entry.
The new proposal changes the rules. It tweaks CBA to weight the dollars a policy generates according to who gets them (and who pays them), instead of just counting the total. It is not a new idea, but it is a radical one, and the hornets’ nest is buzzing.
The criterion that CBA uses to judge whether a policy increases social welfare is called the “Kaldor–Hicks potential compensation criterion,” named for two giants of twentieth century economics. In 1939, economics was somewhat in disgrace because of the Great Depression. Wanting to make their discipline more rigorous, economists confronted a fundamental challenge known as “the problem of interpersonal utility comparisons.” According to utilitarianism (the ethical philosophy guiding economics), the best policy is the one that produces the greatest good (happiness, utility) for the greatest number. The problem is that happiness is subjective, so there is no common measure by which you can compare it between people: no common measure, no comparison, no possibility to add and subtract utilities among a bunch of people to get a net total. This problem is accepted as insoluble by virtually all economists (and philosophers), then and now. Nicholas Kaldor, though, invented an ingenious way around the problem. (Linked papers are on JSTOR.)
Kaldor builds on an earlier argument by Vilfredo Pareto. Even if you cannot compare utility between people, individuals can still assess utility for themselves. Pareto showed that people’s rankings of their own subjective preferences can—under certain conditions—be sufficient to draw conclusions about social welfare. A change that makes some people better off without making anyone worse off (a “Pareto improvement”) raises overall social welfare, no measure of subjective utility or comparison between people necessary. Every simple economic transaction is an example of a Pareto improvement. Albert, who wants pie to go with his coffee, buys Janet’s piece of pie for a dollar. Albert prefers the pie to a dollar, and Janet prefers a dollar to the pie, so the exchange makes them both better off, and presumably no one else is affected at all. The world is better. This very simple and very important bit of reasoning is fundamentally why we think markets are good. They enable us to knowingly make changes for the better.
The big changes of public policy, though, get complicated. They help some people and hurt others. Unable to make interpersonal comparisons, economists couldn’t see how to assess such policy changes. Kaldor’s striking insight was that compensating those hurt by a change—by transferring some of the winners’ gains to the losers so that the losers, too, end up better off—could turn even big, complex changes into Pareto improvements. Using compensation, many actions can be shown to raise overall social welfare. With elaborations by John Hicks, Kaldor designed a rigorous method to create public policies that improve the world, even without being able to make interpersonal utility comparisons. If a policy change brings pie to many Alberts and causes many Janets to lose their pie, we can compensate all the Janets with dollars we take from all the Alberts, and the policy then makes everyone better off and the world a better place. Economics is saved from disgrace!
Then, very quickly, it all goes wrong. The logic of Kaldor and Hicks was solid, but their Achilles heel was politics. The vulnerability comes from “dividing ‘welfare economics’ into two parts: the first relating to production, and the second to distribution” (Kaldor 551). The economist’s part is to figure out which policies efficiently create enough wealth to potentially make everyone better off. The politician’s part is then to distribute that wealth so that everyone actually ends up better off. Only if both parts are completed, and compensation is given, can we know—without interpersonal comparisons—that a policy improves the world.
Hicks anticipates the difficulty their method will encounter in practice: “Yet when such reforms have been carried through in historical fact, the advance has usually been made amid the clash of opposing interests, so that compensation has not been given, and economic progress has accumulated a roll of victims, sufficient to give all sound policy a bad name” (711). He probably did not anticipate the supreme irony that “Kaldor–Hicks” would be the bad name given to this refusal to compensate policy’s victims. But that’s what happens.
The villain of this history is economist Tibor Scitovsky, who wrote a muddled rebuttal to Kaldor and Hicks in 1941. His paper is famous for one point he gets right, but it’s his bad arguments that carry the day. Scitovsky advocates for economists to ignore the politicians’ distribution part and to push their productivity part alone. He scoffs at the demand for rigor that only both parts of the Kaldor–Hicks method together can supply: the economist “may also renounce his claim to purity and base his own recommendations” (80) on productive efficiency alone. That is, do the economist’s production part, and pretend you don’t need the distribution part. The weakness of two parts is the temptation to separate them.
Now, productive efficiency is a wonderfully convenient measure. It is an indisputably useful thing to know and has the hard-nosed numerical objectivity that economists want. By itself, it is innocently the economist’s part of the Kaldor–Hicks two-part program for a better world. The logical fallacy is only introduced when we claim to infer social welfare from productive efficiency, without the compensation part. That inference requires an interpersonal utility comparison: if we consider a policy’s monetary impacts as equivalent to welfare, we are presuming that dollars to winners and dollars from losers are equal in utility, which we cannot know. Yet over the next decades, economists build Cost–Benefit Analysis on Scitovsky’s unsound position. (I use capitals to emphasize that this is a formally defined procedure, not the generic weighing of pros and cons we all do all the time.)
In the new “science” of Cost–Benefit Analysis, productive efficiency measured in money becomes a proxy for utility. And since money is easy to compare between people, compensation becomes unnecessary for calculating a total: production gains that cover losses (plus some) are taken to imply a net gain in social welfare. A policy that creates enough wealth to potentially compensate those it hurts “is considered welfare enhancing according to the Kaldor–Hicks criterion” (to quote the Environmental Protection Agency’s Guidelines for Preparing Economic Analyses). This is pure fantasy. The potential to compensate creates only the potential to enhance welfare; it takes actual compensation to create actual enhancement.
Unwilling to throw away their claim to purity—phony though it now is—economists need a fig leaf to put over this fantasy criterion’s naked logical fallacy. So they name it after Kaldor and Hicks, as if those icons of rigor had originally invented and endorsed it. The fact remains, though, that CBA based on merely potential compensation cannot justify policy decisions.
EPA’s weaselly phrasing (“is considered welfare enhancing”) is hardly accidental. The welfare economics literature is rife with convoluted wording that tries to imply that Cost–Benefit Analysis is sound without quite claiming it is, because everyone knows it is not. But no matter, CBA is up and running. And it runs for forty years, deep into our bureaucratic processes, skewing the economy’s rules and regulations to increase wealth inequality. Until OIRA’s current rulemaking.
OIRA goes to the heart of the matter in this new draft of Circular A-4: the use of money as a proxy for utility. The interesting twist, though, is its strategy: it is more one-upmanship than corrective. Instead of attempting to return to rigor—instead of rejecting the interpersonal utility comparison implicit in CBA’s treating a dollar as having the same utility to everyone—OIRA is saying, basically, “Fine. If that’s the game, let’s really try to compare how the utility of money varies among people.” The way OIRA proposes to estimate money’s varying value is through another fundamental economic concept called “diminishing marginal utility.” It’s less complicated than it sounds.
“Diminishing marginal utility” is something everyone experiences and understands: when you need something, it is valuable, but its value decreases as you get more, then get enough, and then get too much. It is practically our most basic intuition about how utility works in real life. One way to say what is wrong with using money as a proxy for utility is that money, whether it goes where it’s needed or where it isn’t, is counted the same. OIRA’s solution is to make money a better proxy by giving it the diminishing character so essential to utility.
OIRA proposes to multiply the dollar value of policy impacts by a weighting factor that varies depending on the income of the person affected: costs or benefits to a person of median income are multiplied by a weight of one, but that weighting factor increases the poorer someone is and decreases the richer someone is. (The actual formula is hilariously technical, and it is also not mandatory for agencies. The details are important, but this will do for now.) In this way, the calculated value of a policy’s benefits (in dollars) is greater when they go where they are needed and diminishes the less they are needed. But do these weights capture utility?
In a prescient paragraph, Hicks considered weighting and “rejected as unsatisfactory” both sides in today’s debate: CBA’s “equal weights, 1, 1, 1, . . .” and also OIRA’s variable weights, with their “attention to variations of the marginal utility of money between rich and poor” (700). Of course, Hicks is right. Technically, this new method has the same logical fallacy as the old one. Since assigning weights to different people’s marginal utility of money is by definition an interpersonal utility comparison, new CBA is no more objective than old CBA. Presumably, OIRA and its critics both know this: weighting marginal value cannot be a technical fix of CBA’s flaw. The argument is really a moral and political one.
By shifting the debate to subjective terrain, OIRA is better positioned to unrig the rules. Embracing the lack of objectivity that both sides share draws attention to that lack, which makes their opponents’ pretense of it untenable. (The comments are still full of that pretension to objectivity, anyway.) And on subjective terrain, weighted and diminishing marginal utility is enormously more compelling than the frankly nonsensical constant marginal utility implicit in using productivity to proxy welfare. Furthermore, with weights directly embedded in the calculus of policy impacts, distributional effects are no longer relegated to after-the-fact measures that politicians can easily ignore. The new draft of Circular A-4 is a good opening to the debate on CBA’s problems.
That debate has started full force in the public comments (though it plainly will not end there). Everyone wants a say. There are multi-signatory comments from MIT economists, environmental groups, petroleum industry groups, progressive advocates, civil rights groups, international scholars, banking and finance, and economists from all over. Organizations focused on immigration (ILRC), poverty (PRRAC), consumers (CFA), LGBTQ+ (SAGE), animal welfare (HSLF), impact valuation (IFVI), and anti-monopoly (AELP). Associations for small business (NFIB), water districts (AMWA), and the gas industry (GAAS). Think tanks and academies of various stripes (CEI, AEI, RFF, EPI, CBPP, GWRSC), prominent experts (Nordhaus, Gollier, Adler), and many, many other individuals (including me). (The full list of comments submitted to OIRA is at https://www.regulations.gov/document/OMB-2022-0014-0001/comment, but put them in oldest to newest order, because after the first two hundred or so, 4000-plus copies of a letter of support swamped the list.) There’s one from Ganesh Sitaraman, the Elizabeth Warren ally who was Biden’s first choice to lead OIRA and this whole process—until his confirmation was blocked by Kyrsten Sinema, the plutocrat’s friend. Lastly, foreshadowing the next plot development, 26 Republican-controlled states submitted a joint comment signed by their attorneys general, which I think we can take as announcing that they are going to sue to keep the game rigged.
My own comment tells the history and logic of Cost–Benefit Analysis in more detail. I give the story a traditional morality tale structure (known to all): the protagonist (economics) gets a position of privilege (advisor) from a magical or higher being (reason), but with a prohibition attached (you’ll have to read it); unable to resist violating the prohibition, our protagonist is cast out to wander in the wilderness. In my comment, I advocate for Kaldor’s and Hick’s original method, which is both logically and morally superior to either old or new CBA. Politically, though, forty years of anti-regulatory and anti-redistributive propaganda—which includes the creation of OIRA and unsound CBA—have rendered radioactive the tax rates of the previous forty years (top income bracket over 90 percent in the 1940s and ’50s, 70 percent in the 1960s and ’70s). So, my position is not on the table, today. Still, studying this OIRA rulemaking has made me ponder where rigorous logic does not go. I will end, then, with some thoughts on the moral and political debate.
Compensation in Kaldor’s and Hicks’s argument is methodological: it is a means to get around the problem of interpersonal utility comparisons. But compensation also gets them past the issue of marginal utility. How does Hicks know, I wondered, that the marginal utility of money varies between rich and poor? He can’t. It requires an interpersonal comparison to know that. Rigorously, we know diminishing marginal utility only within personal preferences, not between people. Rigorously, we cannot prove even the most absurd maldistribution—give all the money to that person; the rest of you go hungry—is not welfare maximizing. And yet we all know a dollar generates more utility for a poor person than a rich one. It’s self-evident. But it’s a truth known only on the terrain of politics and morality. OIRA is right to shift the debate.
There are times when even the most abstract topics can suddenly take the most human face. “We hold these truths to be self-evident, that all men are created equal.” This famous statement has had a wide range of contentious interpretations, from an empty piety about legal fairness to an extremist demand for material equality. Somewhere in the middle, I glimpse the self-evident truth that varying marginal utility is trying to describe: that every human has the same capacity to experience utility (value, happiness). I do not need to peer into another’s heart, only to see their humanity. One person can easily have a million times as much money as another, but we find the idea weird and incomprehensible that one person could have a million times another’s capacity for joy.
The rulemaking OIRA is engaged in now is designing procedures that will in turn shape policies with huge impacts on our lives. The most dramatic change in Circular A-4 is the introduction of weights for marginal utility that vary by income. These weights, by their nature cannot be known to be right or wrong. We are on the terrain or politics, where we belong. I wish OIRA luck in making the case that their formula estimating the marginal utility of money to different people does justice to our intuition—cherished as a self-evident and founding truth—that every person’s capacity for life should weigh equally in the cost–benefit calculus of the rules and regulations that govern us.