Proudhon famously said, “to be ruled is at every operation, transaction, and movement, to be noted, registered, counted, priced, admonished, prevented, reformed, redressed, corrected.” If Proudhon were living today, he certainly would have added “nudged.”
A nudge is a new term for something citizens of modern bureaucratized governments are intimately familiar. In their 2008 book Nudge, economists Richard Thaler and Cass Sunstein define the term as a restriction of “choice architecture” to ensure better outcomes. The idea is that whenever government acts, it inevitably creates certain policy defaults, and that, all things being equal, it’s better for those defaults to encourage good behavior over bad.
The canonical example of a nudge is that of organ donor registration. Most countries have an opt-in system, where citizens must sign a form to indicate that they consent to becoming an organ donor. Germany is one such country. The organ donor participation rate there is a paltry 12%. Here’s where the nudging comes in: imagine if instead of the system being opt-in, it were opt-out. Austria employs this strategy and with surprising success—over 99% of the population are registered as organ donors. The net effect is thousands of lives saved, as fewer of those on the organ transplant waitlist are turned away. All it takes is a conscious decision on the part of government to harness human inertia. Such is the power of the nudge.
As the winner of this year’s Nobel prize in economics, Thaler has seen his concept of nudging met with the official imprimatur of the economics establishment. It’s not difficult to see why. Thaler describes his philosophy as libertarian paternalism: Libertarian as in “the straightforward insistence that, in general, people should be free to do what they like-and to opt out of undesirable arrangements if they want to do so” and paternalist as in “the claim that it is legitimate for choice architects to try to influence people’s behavior in order to make their lives longer, healthier, and better.” Taken at face value, Thaler’s approach would seem to eliminate most of the hard trade-offs that governments face—a sort of regulatory fulfillment of the old Taoist desire for action-without-action. With economists as the Taoist sages, of course.
Naturally, the audaciousness of Thaler and Sunstein’s approach has attracted a considerable amount of criticism—with most of it focusing on the libertarian half of the libertarian paternalist formula. Critics, such as those at the Mises Institute, contend that Thaler’s program casts its net too wide: if state intervention is justified in cases of imperfect individual rationality, then there is no limiting principle on state action: Real people will always fall short of the textbook standard of rationality implied by homo economics, especially if that standard is interpreted by a group of bureaucrats eager to expand their influence. Ceding an inch to paternalism, in the Mises Institute’s view, creates a slippery slope that erodes individual liberty.
These criticisms no doubt contain some degree of truth. Indeed, they prove too much. Save for the most committed ideologues, no one really objects to switching to an opt-out system of organ donor registration. Moreover, the libertarian case against nudging offers no insight into precisely how nudging fails. From a principled libertarian perspective, there is little difference between the government urging more organ donation and the government urging citizens to consume 6 to 11 servings of carbohydrates a day other than the level of coercion it employs. Both actions employ some amount of coercion. Yet, the former, by most accounts, saves thousands of lives a year with minimal drawbacks, and the latter has contributed to the growing obesity epidemic and pushed dietary science behind by at least a decade. Something is obviously missing here.
Perhaps the problem lies less with the libertarian half of the equation and more with the paternalist half…
If you hold two mirrors toward each other facing inward, you will see a series of reflections-in-reflections tapering off to infinity. If those mirrors are crooked or distorted like those in a fun house, the resulting image will be a visual hodgepodge of weird tendrils and bulges. A child at a carnival standing between them would not recognize himself. If, on the other hand, the mirrors are smooth and flat, a man standing between will see an orderly line of little selves, each smaller than the last. An infinity mirror of this type does not cause distortion.
The same is true in political philosophy. Every ideology or political system is a mirror of sorts. If we hold the mirror of libertarian paternalism up to the world and gaze up at its reflection, what do we see? We see a world populated by irrational individuals, ensnared by all manner of cognitive biases—anchoring, the availability heuristic, status quo bias, herd mentality, etc. But, there’s hope! In this reflection—for it is a good reflection, perhaps slightly rose-tinted—we also see policy-makers benignly accounting for these biases and nudging citizens toward better ends—away from Big Gulps and penny stocks, toward mineral water and ETFs.
A funny thing happens, though, if you hold up two of these mirrors. The benevolent bureaucrats so beloved from the reflection of the solitary mirror now appear as if — gasp! — they too are subject to the same set of cognitive biases afflicting ordinary citizens. And it gets worse: in this double-refracted world, the bureaucrats are not so benevolent — in fact they are rather self-interested. And rather than nudging people toward The Good, they nudge in ways which flatter their favorite politicians; they nudge according to the dictates of their corporate backers; they nudge with an eye to what they think will look good in a research paper; they nudge, in short, in a way that is human.
Ironically, what dooms Theater’s philosophy is a lack of attention to his original premise. The starting place, and impetus for nudging is a recognition that man is not homo economicus. What Thaler forgets is that economists aren’t homo economicus either. For there to be nudging, there must be not only be a nudgee but a nudger as well. And for the system as a whole to function, the interests of both parties must be aligned.
The conceit of this thought experiment would seem to be born out by reality. Academic bureaucrats have many biases which prevent them from nudging correctly. The first problem is time-horizon. The efficacy of a nudge—especially in a domain like public health—is measured in years or decades. But nudgers’ careers are not. The academic-bureaucrat’s career is based on publishing interesting papers and advising high-profile government commissions (as well as appearing in Hollywood blockbusters). But, if those papers and interventions fail or fail to replicate a decade after the fact, the negative consequences are minimal—the prestige of publication has already been earned, and the retraction, if it comes at all, is unlikely to be noticed.
Perhaps it is true that “printing tax bills on the pink paper typically used for debt collection [leads] to an improvement in the prompt payment rate of between three to five percentage points,” as the Thaler-advised Behavioural Insights Team (BIT) found in Singapore in 2015. But will this still be true in 2020? Or in the UK instead of Singapore? Or once pink paper tax bills become ubiquitous? Or perhaps the entire effect is a statistical fluke? If one is looking through a pile of data to find nudges, one is sure to find them solely due to chance.
According to a groundbreaking study by the Open Science Collaboration, one-third to one-half of the findings published in top-tier psychology journals fail to replicate. And the problem is compounded for nudges, which by their very nature, depend upon the correctness of a lengthy chain of assumptions. For a nudge to be effective, one must establish: Firstly, that there exists a common cognitive bias. Second, that bias will persist even outside the confines of a controlled experiment. (It is often the case that effects isolated in an N=25 study of college sophomores seeking beer-money will fail to replicate in the real world). Third, that an isolated bias is addressable by a particular intervention. And lastly, that the intervention will be stable and not create a new social equilibrium in which the intervention is no longer effective. People are, after all, highly adaptable. As such, we should expect not only the measurements to change, but the underlying statistical distributions to change as well. Effective nudging is thus a tall order. And given psychology’s abysmal track record, there is good reason to be skeptical.
The story gets worse when we consider, even after excluding all the complications of normal nudging, that bureaucrats often aren’t acting in good faith. The link between smoking and lung-cancer was first established by Nazi scientists in 1939. It took decades for Americans to re-discover this truth, in no small part due to the influence of the tobacco lobby. Tobacco lobbyists not only funded pro-smoking studies but also waged a sophisticated disinformation campaign on the public. (They also created several successful nudges. For instance, did you know that more doctors smoke Camels than any other cigarette?) But Big Tobacco is just one tale in the greater saga of the regulatory capture of bureaucratized science. The dairy industry nudged generations of schoolchildren into drinking milk every day based on bogus studies. The Fannie Mae Foundation published numerous studies defending the Community Reinvestment Act and its attempt to nudge banks into providing mortgages to people with low (or no) income—setting the stage for the 2008 Housing Crisis. Purdue Pharmaceutical spent millions of dollars funding studies “proving” OxyContin was safe and non-addictive, and millions more on nudging doctors to prescribe the drug to patients—helping create the Opioid Crisis. There is no shortage of these stories. Nearly every pathology in American society has at least some scientific support behind it, and inevitably some of that science is translated into policy. What’s a nudger to do?
Having seen the pitfalls of a naive reliance on bureaucratic science, we are now a position to explain the challenge of paternalism. What unites all the examples of regulatory failure is not their deviance from scientific consensus—quite the contrary. Nor does the failure stem from restricting individual liberty. (This is perhaps a by-product of nudging, but it does not explain the underlying problem.) No. What dooms all these would-be paternalists is a lack of interest in the long-term success of their interventions and a lack of any natural bond of sympathy or understanding with those whom they are nudging. We can summarize both of these points by using the now-classic Talebbian formulation: The problem is that paternalists have no skin in the game. That “skin” may be reputational (perhaps the esteem of one’s fellow nudgers) or it may be material (perhaps a payment tied to a particular outcome). But, it must be a real incentive. And that incentive must be aligned with the interest of society if the nudge is to succeed.
The mistake at the heart of the nudging mindset is the view that abstract knowledge is sufficient to govern, and that pure, detached rational deliberation is capable of producing such knowledge. This idea has a long and distinguished philosophical pedigree—and to do it full justice would require a lengthy exegesis of centuries of intellectual history of the sort I will not provide here—but ultimately, it’s false. Governments are made of people. People have their own biases and motivations. Without the proper incentives, people will behave in ways that are not aligned with their supposed goal. The question is not what a bureaucracy knows, but what the consequences will be if an intervention succeeds or fails. If the consequence of failure is nothing, there will be more failures. If the consequence of success is acclaim or remuneration, there will be more successes. It is true that no set of incentives can guarantee success, but at the scale of government, misaligned incentives mean almost certain failure.
When spelled out in simple terms, the logic of skin in the game seems unimpeachable. Yet, even economists — experts who, above all others, should be focused on incentives — find themselves unable to think clearly when it comes to institutional incentives. The reason is not too difficult to discern: there’s no incentive for them to do so! As a class, economists have no interest in analyzing the incentives that shape their field if doing so will lead to less demand or less prestige for their services. Every priest caste maintains that it’s blasphemy to question their position in the hierarchy. Why should the economist caste be any different?
Thaler himself is unusually exemplary in that he has accumulated a rigorous track record of the application of (a subset) of his ideas in the real world. That track record is the performance of his mutual funds, which attempt to capitalize on behavioral anomalies in the market. Unfortunately, it appears both his “Fuller & Thaler Behavioral Small-Cap Equity Fund” and “Undiscovered Managers Behavioral Value Fund” are struggling to beat the market. (The latter fund has returned 162.40% versus 173.87% against the iShares Russell 2000 ETF since its 2004 inception; the former 54.21% versus 110.07% since 2011.) The financial world is different from the policy world and arguably a much more difficult domain to add value to. But still, if the great new behavioral insights which led to the concept of nudging in the first place fail to beat the market, it’s not exactly an inspiring image.
Paternalists would do well to cultivate the virtue of humility. If even Nobel prize-winners find it difficult to apply their ideas in the real world, then run-of-the-mill bureaucrats should be extra cautious. Good nudging, like good government, is not easy. There will always be the temptation to reduce government to a simple formula—such as implementing the findings of the most recent peer-reviewed paper. But yielding to this temptation, as promising as it may seem, is a mistake. The problem of government is genuinely difficult. There are no shortcuts. Without considering the broader institutional incentives that regulatory bodies operate within, any arbitrary intervention is likely to end in failure. Nudging isn’t easy.
There is no doubt the paternalists of the world will continue to note, register, count, price, admonish, prevent, reform, redress, and correct. But they should do so in a spirit of humility with their incentives aligned. Let this be their nudge.
Nudge, p. 5
As this is Jacobite, it should be noted that the credit arguably goes to James I, who wrote A Counterblaste to Tobacco in 1604, which correctly describes tobacco as “dangerous to the Lungs”, and more controversially, as “resembling the horrible Stigian smoke of the pit that is bottomelesse.”