Posted: September 30, 2013
t is getting harder to run a constitution than to frame one," wrote a young Woodrow Wilson in 1887. Impatient with the American fixation on constitutional forms, Wilson famously called for the development of a homegrown science of bureaucratic administration. The science of curbing executive power, he urged, must now give way to the science of perfecting it—of "discovering the least irritating means of governing."
Today, Harvard Law School professor Cass Sunstein is the most influential exponent of the science of least irritating means. And over the last four years he tried his hand as practitioner, serving as President Obama's regulatory czar. His latest book, Simpler, offers a lucid account of this meeting of theory and practice, advancing an unapologetic defense of the Obama Administration's first-term regulatory agenda.
Sunstein promises the story of "the large-scale transformation in American government" that took place from 2009 to 2012. But it's not the story you might expect. Initiatives such as the Patient Protection and Affordable Care Act and Dodd-Frank Act serve as passing examples in a broader tale of how new regulatory techniques are making the federal government "much simpler" and the lives of Americans "longer, healthier, and easier."
The Office of Information and Regulatory Affairs (OIRA) was a testing ground for the scholarship of Cass Sunstein long before he took charge. Its lawyers and economists function as regulatory traffic cops. Since the Reagan Administration, OIRA has overseen the development of major new rules by federal agencies, mostly by scrutinizing the costs and benefits of big-ticket regulations with an eye toward reducing unnecessary burdens. The presidential directives that guide OIRA's work encourage regulators to maximize cost-effectiveness, use empirical evidence, study risk trade-offs, and consider flexible approaches instead of hard commands.
The development and defense of these methods owe much to Sunstein, who has long been a defender but not a flatterer of the administrative state. In a small library of books and articles, he has advanced a pragmatic defense of government intervention in private markets and the New Deal constitutionalism at its foundation. But he has also taken an unsquinting view of bureaucratic folly. In Risk and Reason (2002), Sunstein brilliantly dissects regulatory programs gone awry, particularly those of a 1970s vintage. The failures that he targets are primarily defects of technique: excessively costly tools, poorly drafted statutes, undisciplined analysis of risks, and priority-setting guided by interest group politics rather than reason. There's plenty wrong with the regulatory state, Sunstein teaches, but nothing that public-spirited experts can't fix.
* * *
In Nudge (2008), Sunstein and economist Richard Thaler argue that the insights of behavioral economics point to a "true, third way" between laissez faire and statist, command-and-control regulation. People err in predictable ways, the authors show. Regulators should intervene to correct those cognitive failures by nudging individuals toward optimal choices while still preserving the freedom to choose. Take, for example, the 2006 pension law that gave businesses an incentive to automatically enroll their employees in a retirement plan (with an easy opt-out). The law boosted participation rates, even though this nudge should have been irrelevant to pure rational actors with complete information.
Such "libertarian paternalism" aspires to prod people toward outcomes that they would select if they knew what the experts know and possessed the self-control to act on it. In Simpler, Sunstein seeks to show how he and his colleagues drew on this learning to launch a "major reorientation in national regulatory policy."
The conservative constitutionalist critique of the administrative state, dormant since the Supreme Court blessed the New Deal in 1937, once more walks the land. The individual mandate in the Affordable Care Act renewed interest in whether national economic regulation knows any constitutional limits. The new Consumer Financial Protection Bureau, with a budget drawn from the Federal Reserve Board rather than congressional appropriations, has reignited a debate over the constitutional pedigree of the "headless fourth branch." And new statutes giving regulators vast discretion have prompted some to dust off the Lockean principle that Congress is supposed to make laws, not deputize new legislators.
Readers looking for a treatment of these larger flashpoints will come away disappointed; Sunstein swears off such questions at the outset. "This book is not meant to address the question whether we should return to the days of Herbert Hoover," he explains, kicking a favorite straw man. The immediate project is to move beyond "sterile, tired" questions about the size and scope of government and to focus instead on "what really works."
* * *
Ever the pragmatist, Sunstein takes seriously the free-market conservative's lament that regulation is too often clumsy, costly, and counterproductive. He allows that there is some truth to this "general critique of regulation." One way to answer this critique would be to limit the sphere of government action. Devotees of Friedrich Hayek, for example, would say that regulation should be limited to correcting market failures such as negative externalities (e.g., pollution) or informational asymmetries (e.g., arcane home mortgages) that cannot be addressed through voluntary bargaining.
By contrast, Sunstein's principal answer is not to limit the occasions or grounds for government regulation, but rather to shore up the "empirical foundation" for the methods and means of regulation. The critical task is to apply expert analysis to determine whether a particular intervention will produce an outcome that "maximizes net benefits." The trick is to make bureaucracy more bureaucratic in the old sense of the word—clinical, rational, detached from interest-group politics. Central to that project is evaluating rules based on their projected costs and benefits, a process Sunstein sees as a "natural corrective" to regulation based on biases or mere politics. ("Get your mind out of the gutter," he admonished aides who cited the opinions of this or that lobby group or business.)
Many on the left were anxious about Sunstein's commitment to cost-benefit analysis, but the administration's record suggests those concerns were overwrought. No doubt Sunstein was an advocate within the Executive Branch for a more balanced approach to regulation, including the high-profile return of the $90-billion Ozone Rule in 2011. He also launched a government-wide effort to prune unnecessary regulations already on the books. But on the real measure of regulatory output-big-ticket rules with effects exceeding $100 million—President Obama's first term was the most aggressive of any administration since pertinent records were kept. In fact, according to the administration's own estimates, the cost of rules issued in 2012 alone exceeded the cost of all rules in the entire first terms of Presidents George W. Bush and Bill Clinton, combined.
* * *
That record notwithstanding, conservatives can still find much to admire in Sunstein's defense of cost-benefit analysis. In Simpler, he is at his best showing how more disciplined thinking can help us to avoid common fallacies that skew regulation-fallacies such as exaggerated perception of risk. Consider a 2008 railroad safety law enacted by Congress in response to a California commuter train crash. (My example, not his.) The law requires installation of autopilot-like systems on passenger trains. The proposed rule's multibillion-dollar price tag exceeded its benefits by about 20 to 1, but Congress rushed to do something in response to an alarming (but relatively small) risk. Sunstein notes that cost-benefit analysis can have a "cooling effect" on such impulses because it prompts policymakers to act based on evidence rather than anecdote.
Analyzing costs and benefits is an excellent way to size up the economic effects of a proposed rule and to compare it to alternatives. But Sunstein tends to rely on this methodology as a substitute for a clear philosophy about the proper ends of regulation and limits of government. "If the net benefits are high," he explains, "we have good reason to go forward, whatever our abstract misgivings about regulation." Of course, another term for "abstract misgivings" is "principles." As Reagan Administration regulatory czar Christopher DeMuth has argued, a cost-benefit regulator asks only whether a proposed regulation would benefit one group more than it would hurt another. That is certainly useful information. But it does not tell us whether the benefits sought are proper ends of government, or how the costs imposed will affect individual rights that are themselves ends—including personal liberty, property rights, and other abstract misgivings.
* * *
These questions do not much trouble Sunstein. One explanation lies in his theory of state action. In his 1993 book, The Partial Constitution, he argues that government does not stay "neutral" when it declines to intervene in private affairs. Existing rights and distributions of goods are themselves "creations of law" because "people own things only because the law permits them to do so." On this view, there should be no "presumption" in favor of government non-intervention. Both enforcement of, and interference with, property or contract rights are equally state action. Redistributive policies, for example, are no more a product of state action than the laws that protect property ownership. In this sense, your claim to your home is no different from your neighbor's claim to government-subsidized housing. "The line between positive and negative rights is thus selected," Sunstein concludes; nothing permanent or pre-political marks that line. It is easy to see how this rejection of a baseline anchored in natural rights and the consent of the governed would persuade any regulator that the possibilities for government intervention are virtually limitless.
Accordingly, when Sunstein claims the purpose of regulation is to promote "social welfare, broadly understood" he meansbroadly: health, wealth, equity, fairness, and "easier" living. He subordinates the question of ends to the study of means, a common pitfall in the science of public administration. Traditional political science began with rational deliberation about ends, and therefore about the human good. Those ends limited and guided the choice of means. But as political scientist Herbert Storing famously argued, the modern science of public administration has little to say about ends because it presumes that "rationality does not extend to the sphere of ends." Sunstein seems to operate from the same premise. Indeed, he emphasizes that the "factual" aspects of regulation (dealing with means and their consequences) can often overcome disputes about "values" (dealing with the ends and limits of state intervention). The practical import of this narrow focus on means is that the regulator's domain has few, if any, ex ante limitations—because we first have to run the numbers.
* * *
Picking up in Simpler where he left off in Nudge, Sunstein describes behavioral economics' lessons for public administration. The key premise is that man is not the homo economicus that rational choice theory posits. Our choices are often guided by habit or gut instinct, rather than reflection and calculation. We procrastinate, go with the flow, underestimate risks, sacrifice the long-term for the short-term, and exercise poor self-control, among other mistakes. In short, we do everything our parents warned against.
That's where Sunstein's "choice architects" come in. From cafeteria designers to federal energy regulators, choice architects have an opportunity to enlighten our irrational preferences and nudge us toward eating less junk food and driving more energy-efficient cars, among other goods.
On the question of methods, Sunstein's application of behavioral economics is interesting and nuanced. The old-school central planner's method was rigid mandates and bans. But informed by behavioral science, choice architects can devise simpler interventions that are more effective and "freedom-preserving," such as well-designed disclosure and default rules. Sunstein makes a persuasive case that government policy can be at once less irritating and more effective by addressing people as they are.
But when it comes to showing how the Obama Administration applied these insights to soften and simplify regulatory tools—Simpler's point—Sunstein comes up short. He reports, for example, that the Agriculture Department devised a sleek new "food plate" to replace the outdated "food pyramid," in which a stick figure appeared to be running up stairs away from grains and vegetables. And the Transportation Department updated its fuel efficiency labels to inform car buyers how much they will save (or lose) at the pump. These sorts of innovations are sensible as far as they go. But a "revolutionary" change in regulatory policy they are not.
Quite the contrary, on the most consequential issues, the old command-and-control approach won out during Sunstein's tenure. In addition to those new fuel efficiency labels (a nudge), federal regulators issued a slew of energy-use mandates for everything from refrigerators to pickup trucks. In addition to requiring businesses to auto-enroll employees in health insurance plans with the ability to opt out (another nudge), the administration imposed the individual mandate. And in addition to a new Labor Department disclosure rule designed to ensure that union leaders look out for their members' interests, the National Labor Relations Board brought an unprecedented enforcement action against Boeing for opening a non-union plant in South Carolina. In these and other major areas, nudges were more often used to reinforce, not soften, regulatory shoves.
* * *
More striking than Sunstein's treatment of regulatory methods is his explanation of how "behavioral market failures" can supply "new grounds for government action." A behavioral market failure occurs when a common error prevents a lot of people from making welfare-enhancing choices that they would make—if they knew what experts know. That's where government steps in "to protect choosers from themselves."
Sunstein sees cognitive errors infecting all sorts of decisions. People save less, eat worse, and consume more energy than they should due to myopic thinking and poor self-control. He is especially concerned that millions of Americans do not donate organs because they are just lazy about registering. When the government does not believe that people's choices are promoting their own welfare, it is often defensible to "influence or alter people's choices for their own good."
On the old rational choice theory, such paternalism would seldom pass cost-benefit analysis. Economists would assume people drive Ford F-150s and drink Big Gulps because they value them more than Chevy Volts and carrot juice. But what Sunstein terms "behaviorally informed cost-benefit analysis" allows planners to substitute their own values to correct for what consumers would choose if they were not so childlike, impulsive, and thirsty. These assumptions are powerful. For example, when regulators accounted for perceived consumer irrationality, the estimated economic impact of one very costly EPA rule swung from $150 billion in net losses to $1.6 trillion in net gains. But there's the ancient problem: whocorrects whom? Political scientist Herbert Simon laid the foundation for behavioral economics in his groundbreaking exploration of "bounded rationality." Shopaholics, chain smokers, and big gulpers were not his subjects. Rather, Simon discovered bounded rationality through his observation of bureaucrats working in a government agency—the Milwaukee Parks Department. As it turns out, bureaucrats are people, too. But more than that, they are people whose decisions are plagued by public choice pathologies—rent-seeking, tunnel vision, action bias—that Sunstein himself has chronicled.
Sunstein does not explain why interventions designed by government officials will be less prone to error or bias than the free choice of individuals. Nor does he grapple seriously with the vast potential for abuse. The regulatory ambition to make people's lives "longer, healthier, and easier" by correcting for their own failings admits of no principled limitation. Yet the only answer he musters is that we can oppose paternalism driven by "impermissible motivations." With no principle to distinguish the impermissible from the permissible, we are left with the bare assurance that this will be worked out case-by-case.
Of course, regulation recast as cognitive-error correction can obscure the government's true motivation. Sunstein's favored nudges generally point to progressive ends, but a nudge less congenial to those on the left might illustrate the point. Years ago, the late James Q. Wilson proposed that states ought not ban abortion, but should instead adopt a particular kind of informed consent law: women seeking an abortion should be shown an ultrasound before making their decision. Suppose that such a law, much like other disclosure-based nudges, were grounded in behavioral evidence that people "prefer an outcome we value and can visualize to one that we value but cannot visualize" (as Wilson argued). Would progressives like Sunstein be satisfied that such an intervention is merely a "freedom-preserving nudge"? Or would they cry foul and argue that the true aim of such a law is to restrict abortion access? Surely an honest debate would turn not on whether a "behavioral market failure" exists, but rather on arguments about ends—about the nature of liberty and the morality of abortion. Behavioral science has much to say about how people will respond to a regulation, but it is a poor guide to identifying new grounds for regulation.
* * *
Many Americans instinctively bristle at paternalistic regulation. They don't like being told what to do. Sunstein sees such skepticism as "an exceedingly valuable part of a culture that prizes liberty," but he argues that the liberty objections to his own "soft paternalism" are surmountable.
He first insists that there is a crucial distinction between "means paternalists" (nudgers like himself) and "ends paternalists" (let's call them meddlers). The meddler steers you toward ends that he thinks most important, like "chastity or longevity," even if you disagree. The nudger is not so bossy. He is "reluctant to question people's ends." Like "the GPS in your car," he just helps get you where you want to go.
But, of course, people have many destinations. Even the most diffident nudger must ultimately select which of the "people's ends" to facilitate, often amidst competing ends. And how will he select? Sunstein already told us: the nudger is guided not by people's actual preferences, but rather by what they would choose if they knew what he knows. Would any self-respecting meddler claim any different? Ultimately this facile distinction masks the reality that the means paternalist, no less than the ends paternalist, steers people toward ends that he thinks important and good. The tools may differ but the project is the same.
Then there is Sunstein's assertion that liberty is not endangered so long as people can "opt out" of default choices that government has devised. He highlights, for example, the uptick in organ donation that would follow from a rule that presumes every adult is a donor, unless you opt out. If we shift for a moment from the fixation on utility of means to a consideration of ends, it should be immediately obvious that such a rule would be an attack on self-ownership—a pre-political right well understood by the founders and their contemporaries. Once again, though, Sunstein voices no such qualms because he rejects the notion that government action should be judged against a "status quo" baseline. The "default rule" that your organs belong to you is no less a product of state action than the default rule that your organs will be harvested. Ultimately, such a philosophy is antithetical to the idea of limited government, for it sees the hand of the state everywhere. The main question becomes whether that power could be deployed more expertly.
* * *
Finally, Sunstein argues that too many choices can make people "far worse off," like diners flustered by a long lunch menu. He sees it as an advantage, not a drawback, that soft paternalism relieves people of the burden of choosing for themselves in small things. His idea is explicitly that our "decision-making energy" is a scarce commodity, and the government should intervene to help conserve it. "Often the best approach is to make things automatic," Sunstein writes, "so that if people do nothing at all, they'll be just fine." Alexis de Tocqueville warned against precisely such administrative mother-henning. A government that substitutes its judgment for ours "where simple good sense can suffice" threatens the very stuff of self-government—free will and self-reliance. Tocqueville foresaw a nameless species of gentle regulation that "does not break wills, but softens them, bends them, and directs them," and thereby "renders the employment of free will less useful and more rare." Policies that cultivate a dependence on government in making small choices, he warned, could eventually degrade the ability of a free people to choose in large affairs—including "choosing well those who lead them."
Is it implausible that people whose choices in everything from diet to light bulbs are increasingly directed by the state might come to rely less on their own will and judgment? That they might become less self-reliant and more infantile? Sunstein's examples of regulatory innovations will strike many readers as so reasonable that there is little to fear. But as the author himself notes, it is ideas that exert lasting influence. "Indeed the world is ruled by little else," as John Maynard Keynes wrote and Sunstein reminds us. The real question is whether Cass Sunstein's administrative science is built on sound ideas—principles suited to the character of a republican people and consistent with limited government. The answer, based on this book, is no.