Psychologists and behavioral economists have identified many sources of human errors, including self-control problems, “present bias,” unrealistic optimism, and limited attention. Building on these underlying findings, a great deal of work has explored the possibility of enlisting libertarian paternalism, or nudges, to make people’s lives go better. Nudges preserve freedom of choice and thus allow people to go their own way. But in light of behavioral findings, there has also been increasing interest in asking whether mandates and bans have a fresh justification.1 The motivation for that question is clear: If we know that people’s choices lead them in the wrong direction, why should we insist on, or adopt a precommitment to, approaches that preserve freedom of choice? Some skeptics, notably Professors Ryan Bubb and Richard Pildes, object that behavioral economists have “trimmed their sails” by adopting an unjustified presumption in favor of choice-preserving approaches.2
It should be agreed that if a mandate would increase social welfare, suitably defined, there is a strong argument on its behalf. No one believes that nudges are a sufficient approach to violent crime. In the face of a standard market failure, coercion has a standard justification; consider the problem of air pollution. We know that there are “behavioral market failures” as well. If people suffer from unrealistic optimism, limited attention, or a problem of self-control, and if the result is a serious welfare loss, there is an argument for some kind of public response. We could certainly imagine cases in which the best approach is a mandate or a ban, because that response is preferable, from the standpoint of social welfare, to any alternative, including nudges.
Nonetheless, there are many reasons to think that if improving social welfare is the goal, nudges have significant advantages and are often the best approach. They may well have high benefits without high costs, and in any case their net benefits may be higher than those of alternative approaches. Five points are especially important.
First, choice-preserving approaches make sense in the face of heterogeneity. By allowing people to go their own way, they reduce the high costs potentially associated with one-size-fits-all solutions, which mandates often impose. Second, those who favor nudges are alert to the important fact that public officials have limited information and may themselves err. If nudges are based on mistakes, the damage is likely to be less severe than in the case of mandates, because nudges can be ignored or dismissed. Third, nudges respond to the fact that public officials may be improperly affected by the influence of well-organized private groups (the public choice problem). If so, the fact that people can go their own way provides an important safeguard, at least when compared with mandates. Fourth, nudges have the advantage of avoiding the welfare loss that people experience when they are deprived of the ability to choose. In some cases, that loss might be severe. Fifth, nudges recognize that freedom of choice can be seen, and often is seen, as an intrinsic good, which government should respect if it is to treat people with dignity.
In light of these points, it makes sense for regulators to give careful consideration to choice-preserving approaches, and generally to adopt a (rebuttable) presumption in their favor, at least where no standard market failure is involved.3 To adopt such a presumption, there is no need for any kind of sail-trimming, or for indulging a strong or naïve precommitment to freedom of choice, or for focusing on political feasibility.4 To be sure, the various points will have different degrees of force in different contexts. In some settings, for example, the interest in freedom of choice has overwhelming importance; in others, people do not much care about it, and its intrinsic value is modest. In the face of behavioral market failures, the presumption in favor of choice-preserving approaches is a pragmatic one, and an analysis of the context may suggest that a more aggressive intervention is justified. But from the standpoint of both liberty and welfare, the presumption matters.
Consider three illustrative problems.5
1. Behavioral economists have devoted a great deal of attention to automatic enrollment in retirement plans. Because of inertia and “present bias,” and perhaps because of optimistic bias, many employees fail to sign up. Automatic enrollment increases participation rates, and thus people’s savings, while also preserving freedom of choice. The problem is that if the default contribution rate is lower than what employees would otherwise choose (say, three percent, as it has been under many automatic enrollment plans), then the result of automatic enrollment might be to decrease aggregate savings, if and because the default rate turns out to be sticky.6 As Bubb and Pildes emphasize, this is an ironic result for those who want to use nudges to increase people’s welfare during retirement.7
The natural response, however, is not to abandon libertarian paternalism or choice-preserving approaches, but to choose a better default. One possibility is “automatic escalation,” which increases savings rates each year until the employee hits a predetermined maximum.8 Another possibility is to select a higher default contribution. No one denies that nudges can go wrong. If they do, the challenge is to get them right. The objection to unduly low default contribution rates operates within the family of libertarian paternalism.
But there is a more fundamental objection, which questions freedom of choice altogether. Suppose that people opt out of 401(k) plans for reasons that are bad, in the sense that the decision to opt out makes their lives worse (by their own lights). Perhaps the relevant people have a general (and unjustified) distrust of the financial system or of their employer, and so they elect not to save at all. Perhaps they suffer from an acute form of present bias. Perhaps those who opt out are most likely to suffer (a lot) as a result of doing so. If so, the argument for a mandate gains force on welfare grounds. Indeed, critics might go further and argue for some kind of comprehensive welfare assessment, by public officials, of optimal savings rates, and ask those officials to build mandates on the basis of that assessment.9
This conclusion cannot be ruled out in principle, but there are reasons for considerable caution, rooted in the five points catalogued above. In assessing the rationality of those who opt out, public officials might be wrong. As compared to a nudge, a mandate might include people in the system who would benefit from such inclusion, but it might also get people into the system who would be harmed. It is important, and may be difficult, to know the size of the two groups or the magnitude of their losses. Those who opt out might do so not for bad reasons, or because they are ignoring their future selves, but because they need the money now; perhaps they are making a sensible tradeoff between their current and future welfare. To say the least, a comprehensive welfare assessment of optimal savings rates is exceedingly difficult on both normative and empirical grounds, especially in view of heterogeneity in the population and economic and demographic changes over time.
Moreover, any form of coercion will impose a welfare loss on choosers, some of whom want to exercise their autonomy and would undoubtedly be frustrated to find that they cannot. And if freedom of choice has intrinsic value or can promote learning, then there are additional reasons to avoid mandates. To be sure, the social security system does compel savings, and for a mixture of legitimate reasons; some people think that the program should be expanded to increase the level of mandatory savings. But even if so, private retirement plans have an important place for savers, and the question is whether the current voluntary system should become more coercive. If the evidence demonstrates that a more coercive approach would make people’s lives better, that approach should be on the table. But the fact of heterogeneity and the risk of government error argue strongly in the direction of nudges.
It is true that default rules tend to be “sticky.” But it is also true that people reject defaults when that harm is clearly apparent to them. That fact is an important safeguard.10
2. In the context of financial products, behavioral economists have emphasized the potential value of two nudges: disclosure and defaults. Both of these tools played a role in the Credit Card Accountability Responsibility and Disclosure Act11 (CARD), enacted in 2009. One of its provisions is a small nudge: every month, companies must disclose the interest savings from paying off the full balance within thirty-six months, instead of making only minimum payments every month.12 It is easy to be skeptical about disclosure requirements of this kind.13 But existing research finds that the consequence has been to reduce interest payments by $74 million a year — not a huge amount, but far from trivial.14
The CARD Act contains other seemingly modest provisions designed to limit credit card fees. For example, companies are forbidden to impose fees on cardholders who go over their credit limit unless cardholders agree to opt in to authorize that practice.15 In addition, banks must give cardholders a forty-five-day advance notice of rate increases, and they must inform cardholders of their right to cancel the account before such increases go into effect.16 Cardholders must also be provided with statements that inform them exactly how long it would take to pay the outstanding balance if they made only the minimum monthly payments.17
What are the effects of these provisions? The answer is that they have contributed to substantial decreases in both over-limit fees and late fees — with the overall package saving U.S. credit card users no less than $20.8 billion annually.18 Notably, cardholders with low credit scores appear to be the biggest beneficiaries.19 To be sure, and importantly, the package includes mandates as well as nudges, and it would be useful to quantify the effects of both — but there is no question that nudges have played a beneficial role.
In terms of default rules, a great deal of skeptical attention20 has been paid to the regulatory effort by the Federal Reserve Board to protect consumers from high bank overdraft fees.21 To provide that protection, the Board issued a regulation, in 2009, banning banks from automatically enrolling people in overdraft “protection” programs; instead, customers have to sign up.22 In principle, the regulation should have had a very large effect, but the available evidence, catalogued in an important article by Lauren Willis, suggests that the effect has not been nearly as large as might be expected.23 The reason is that people are opting into the program, and thus rejecting the nonenrollment default, in significant numbers.24
But we should be careful before drawing strong conclusions from this example, and it would be premature to conclude that the regulation has failed. Large numbers of people are no longer enrolled in the programs. The precise figures remain unclear, but the overall level of opt-in seems to be around fifteen percent; to that extent, the opt-in default has been at least somewhat sticky.25 Moreover, the largest proportion of people who opt in are those who actually go over their checking limits.26 For such people, it is plausible to think that opting in is a good idea. If they cannot borrow from their bank, they might have to borrow from someone else — which would mean a level of inconvenience (an important consideration in light of the scarcity of time) and potentially equivalent or higher interest rates. With these points in mind, we might consider the possibility that the Federal Reserve’s approach is, on balance, a tribute to the sense and value of choice-preserving approaches.
It is true that to evaluate the existing situation, we would need to know a lot more about the population of people who opt in. Perhaps they are insufficiently informed or attentive. Perhaps they are paying interest charges unnecessarily. If so, there is a good argument that further steps should be taken, perhaps in the form of textual reminders that they are about to incur charges. Nor can we exclude the possibility that in some settings (for example, those involving teaser rates), a mandate or ban might be justified to protect customers.27 But to reach that conclusion, we would need to have a clear sense that the benefits justify the costs, and to investigate the risk of unintended bad consequences.28
3. Motor vehicles emit pollution, and the use of gasoline increases American dependence on foreign oil. On standard economic grounds, there is a market failure, and the best response seems to be some kind of tax designed to ensure that drivers internalize the social costs of their activity. But behaviorally informed regulators would be inclined to suspect that when consumers purchase cars, many do not give sufficient attention to the costs of operation, and hence that they buy too many fuel-inefficient vehicles. Even if they try, consumers might not have a sufficient understanding of those costs, because it is not simple to translate differences in miles per gallon (MPG) into economic and environmental consequences. An obvious nudge would be disclosure, in the form of a fuel economy label that would correct the kind of behavioral market failure that comes from limited attention.29
But it would be possible to wonder whether the label would prove entirely effective; this is an empirical question. If it is not, people will purchase cars that are less fuel-efficient than would be socially optimal. True, a corrective tax might reduce that problem, but if, at the time of vehicle purchase, consumers neglect fuel costs, it might be best to combine the tax with some kind of subsidy for fuel-efficient cars. And if consumers (at the time of purchase) are genuinely inattentive to the costs of operating a vehicle, then it is not entirely impossible that fuel economy standards, which are strongly disfavored on standard economic grounds, might themselves turn out to be justified.
In exploring the case for mandates rather than nudges, Bubb and Pildes make precisely this argument.30 In support of that argument, they might have explored two kinds of consumer savings from fuel economy standards, not involving externalities at all: money and time. In fact, the vast majority of the quantified benefits from recent fuel economy standards come not from environmental improvements, but from money saved at the pump; turned into monetary equivalents, the time savings are also significant. For the most recent and ambitious of those standards, the Department of Transportation found consumer economic savings of about $529 billion; time savings of $15 billion; energy security benefits of $25 billion; carbon dioxide emissions reductions benefits of $49 billion; other air pollution benefits of about $14 billion; and less than $1 billion from reduced fatalities.31 The total projected benefits are $633 billion over fifteen years, of which a remarkable eighty-four percent come from savings at the pump, and eighty-six percent from those savings along with time savings.32
The problem is that on standard economic grounds, it is not at all clear that these consumer savings should be allowed to count in the analysis, because they are purely private savings and do not involve externalities in any way. In deciding which cars to buy, consumers can certainly take account of the private savings from fuel-efficient cars. Where is the market failure? If the problem lies in a lack of information, the standard economic prescription overlaps with the behaviorally informed one: provide that information so that consumers can easily understand it.
In this context, however, there is a serious risk that a nudge will be inadequate. Even with the best fuel economy label in the world, consumers might well be insufficiently attentive to those benefits at the time of purchase, not because they have made a rational judgment that they are outweighed by other factors, but simply because most of their focus is on other variables. (How many consumers think about time-savings when they are deciding whether to buy a fuel-efficient vehicle?) If so, a suitably designed fuel economy mandate — hard paternalism, and no mere nudge — might end up producing an outcome akin to what would be produced by consumers who are at once informed and attentive. If the benefits of the mandate greatly exceed the costs, and if there is no significant consumer welfare loss (in the form, for example, of reductions in safety, performance, or aesthetics), then there is reason to believe that the mandate does serve to correct a behavioral market failure.
Of course we should be cautious before accepting that conclusion. Behavioral biases have to be demonstrated, not simply asserted; perhaps most consumers do pay a lot of attention to the benefits of fuel-efficient vehicles. The government’s numbers, projecting costs and benefits, might be wrong. It is important to emphasize that consumers have diverse preferences with respect to vehicles, and regulation might end up decreasing consumers’ access to vehicles with attributes that many of them prefer. (Fortunately, the use of fleet-wide averages helps to ensure that a great deal of freedom of choice is maintained.) With these qualifications, the argument for fuel economy standards, made by reference to behavioral market failures, is at least plausible. In this context, nudges (in the form of an improved fuel economy label) and mandates (in the form of standards) might march hand-in-hand.
The behavioral argument on behalf of fuel economy mandates should not be read for more than it is worth. Even if it is convincing, it certainly does not establish that in the face of human error, mandates are generally preferable to choice-preserving alternatives. As we have seen, the advantage of such alternatives is that they often have higher net benefits, because they reduce the costs of imposing solutions on heterogeneous populations, reduce the risks associated with government error, and avoid the welfare (and other) costs associated with eliminating freedom of choice. It is true that in the end, mandates might ultimately turn out to be justified. But here as elsewhere, it is important to give careful attention to less intrusive, choice-preserving alternatives, and generally to adopt a (rebuttable) presumption in their favor, at least when a standard market failure is not involved.
* Robert Walmsley University Professor, Harvard University. The author is grateful to Tyler Cowen, Lucia Reisch, and Richard Thaler for valuable suggestions and to Matthew Lipka for superb research assistance.
1. Professor Sarah Conly, for example, argues on behalf of coercion when the analysis of costs and benefits justifies it. See generally Sarah Conly, Against Autonomy (2012). In the same general spirit, Professors Ryan Bubb and Richard Pildes argue, in How Behavioral Economics Trims Its Sails and Why, against any kind of presumption of precommitment in favor of choice-preserving approaches. 127 Harv. L. Rev. 1593 (2014).
2. See Bubb & Pildes, supra note 1.
3. To be sure, choice-preserving approaches might make sense in the face of such failures as well (such as an absence of information).
4. This is my principal disagreement with Bubb and Pildes. The interest in choice-preserving approaches is based on an appreciation of the five considerations outlined here; it should hardly be seen to reflect a capitulation to political constraints or as an artificial effort to truncate the set of potential policy responses.
5. All of these are explored in detail in the Bubb and Pildes article. See Bubb & Pildes, supra note 1, at 1607-1677.
6. See Bubb & Pildes, supra note 1, at 1618–19.
8. See Shlomo Benartzi & Richard H. Thaler, Behavioral Economics and the Retirement Savings Crisis, 339 Science 1152, 1152 (2013).
9. See id.
10. See John Beshears et al., The Limitations of Defaults 8 (Sept. 15, 2010) (unpublished manuscript), available at http://www.nber.org/aging/rrc/papers/onb10-02.pdf, archived at http://perma.cc/9P94-TQZM.
11. Pub. L. No. 111-24, 123 Stat. 1734 (2009) (codified in scattered sections of 5 U.S.C., 11 U.S.C, 15 U.S.C, 20 U.S.C., and 31 U.S.C.).
12. Id. § 201, 123 Stat. at 1743–45.
13. Bubb & Pildes, supra note 1, at 1598.
14. Sumit Agarwal et al., Regulating Consumer Financial Products: Evidence from Credit Cards 4 (Nat’l Bureau of Econ. Research, Working Paper No. 19484, 2013), available at http://www.nber.org/papers/w19484, archived at http://perma.cc/XW3V-GDUD.
15. § 102, 123 Stat. at 1738–40.
16. Id. § 101, 123 Stat. at 1735–36.
17. Id. § 201, 123 Stat. at 1743–45.
18. See Agarwal et al., supra note 14, at 3.
20. See Lauren E. Willis, When Nudges Fail: Slippery Defaults, 80 U. Chi. L. Rev. 1155, 1174–75 (2013) (explaining the regulation); Bubb & Pildes, supra note 1, at 1655–56 (drawing on and endorsing Willis’s critique).
21. Requirements for Overdraft Services, 12 C.F.R. § 205.17 (2010).
22. See id. § 205.17(b).
23. See Willis, supra note 20.
24. Id. at 1184.
26. Todd J. Zywicki, The Economics and Regulation of Bank Overdraft Protection 17–19 (George Mason Univ. Law & Econ. Research Paper No. 11-43, 2013), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1946387, archived at http://perma.cc/3TBK-KNCA.
27. Bubb & Pildes, supra note 1, at 1597–98.
28. Zywicki, supra note 26, at 25.
29. Indeed, such a nudge is in place. See Cass M. Sunstein, Simpler (2013).
30. Bubb & Pildes, supra note 1, at 1675–77.
31. Nat’l High. Traf. Safety Admin., Final Regulatory Impact Analysis: Corporate Average Fuel Economy for MY 2017–MY 2025 Passenger Cars and Light Trucks 49 tbl. 13 (2012).