American Economic Review: Papers & Proceedings 2015, 105(5): 396–401 http://dx.doi.org/10.1257/aer.p20151049

Behavioral Economics and Public Policy 102: Beyond Nudging†

By Saurabh Bhargava and George Loewenstein*

The United States, despite its affluence, faces Behavioral Economics BE as an alternative a number of daunting policy challenges. We rank framework. While ­traditional( ) economics offers last among 16 wealthy peers in life expectancy, regulatory and price-based solutions to deal infant mortality, and obesity, and have taken with market failures, such as those stemming very little action on climate change, despite from the presence of externalities e.g., pollu- the extremity of the threat it poses not only to tion or education , BE prescribes strategies,( and future generations but even those alive today.1 rationales, for the) use of policy to address the Complicating matters is that the causes of these “internalities” or “within-person externalities” policy problems are multi-determined and can that stem from( the failure of individuals to suc)- be engaged at different levels of policy analysis. cessfully pursue their own interests.2 For example, while obesity is proximally due In this essay, aimed broadly at students of to the persistent intake of excessive calories, ­policy and their instructors, we comment criti- it can (also be attributed) to more intermediate cally on the past, present, and future role of BE causes such as innovations in food production in public policy. We first describe the notable and delivery which have lowered the cost of successes of early applications of BE which consuming largely unhealthy foods, to policy have typically involved proximal interventions decisions such( as the subsidization) of corn and designed to improve behavior. We then argue hence corn syrup , or even to more deeply struc( - that BE can and should now aspire to influence tural factors, such) as the sociocultural forces that the design of policies aimed at the deeper causes have promoted the availability of “­super-sized” of policy problems. Through a set of guiding portions and physical inactivity. To combat principles, and case-studies, we seek to pro- obesity, a policymaker must decide whether to voke students of public policy to think about intervene proximally e.g., mandating caloric how to leverage the teachings of BE more fully displays , or more distally( e.g., reducing access to deliver policy solutions whose scope is com- to low-nutrient) foods . ( mensurable with the magnitude of contempo- Economics offers ) a useful framework for rary challenges. thinking about policy. However, the traditional approach in economics, by assuming fully ratio- I. The Birth and Early Successes of Nudge nal and perfectly informed individuals, assumes away many potentially problematic behaviors. The intellectual basis for applying BE to pol- Recognizing human limitations and their con- icy was formally articulated in two papers pub- sequences, many policymakers have embraced lished in 2003. Titled “Libertarian Paternalism” Thaler and Sunstein 2003 and “Regulation for ( ) * Bhargava: Social and Decision Sciences, Carnegie Conservatives: Behavioral Economics and the Case for ‘Asymmetric Paternalism’” Camerer Mellon University, 5000 Forbes Avenue, Pittsburgh, PA ( 15213 e-mail: [email protected] ; Loewenstein: et al. 2003 , both papers advocated an approach ( ) Social and Decision Sciences, Carnegie Mellon University, to policymaking) intended to benefit individuals 5000 Forbes Avenue, Pittsburgh, PA 15213 e-mail: gl20@ not acting in their own self-interest, but which andrew.cmu.edu . We thank Lynn Conell-Price,( Ania Jaroszewicz, Brigitte) Madrian, Nachum Sicherman, Cass imposed minimal burdens on those already Sunstein, and Richard Thaler for comments and feedback. acting rationally. This approach, as the titles † Go to http://dx.doi.org/10.1257/aer.p20151049 to visit the article page for additional materials and author disclo- sure statement s . 1 Relative rankings( ) on health outcomes are taken from a 2 Coined by Herrnstein et al. 1993 , an “internality” is 2013 report published by the Institute of Medicine and The produced when an individual fails( to consider) the full impact National Research Council. of a behavior on her present future utility. / 396 VOL. 105 NO. 5 BEHAVIORAL ECONOMICS AND PUBLIC POLICY 102: BEYOND NUDGING 397 of the papers suggest, was meant to appeal to not limit itself to proposing nano-size interven- both conservatives and progressives by prom- tions that may not significantly address the more ising to improve welfare while preserving basic causes, or the magnitude, of contemporary freedom of choice. The specific type of policy policy problems. Returning to the earlier exam- espoused by this framework was popularized ple of obesity, while interventions intended as a “nudge” by a later best-selling book of that to nudge individuals into healthier eating may title. Nudges, such as the strategic placement modestly improve diet e.g., the caloric displays of fruits in the school lunch line, were defined ­mandated by the Affordable( Care Act ACA , as “any aspect of the choice architecture that such policies are not likely to address the( struc))- alters people’s behavior in a predictable way tural supply-side dynamics or ­socio-cultural fac- without forbidding any options or significantly tors that have contributed to the recent rise in changing their economic incentives” Thaler obesity. While nudges should remain an import- and Sunstein 2008, p. 6 . These interventions,( ant part of the policy toolkit, insights from BE according to Nobel Laureate) Daniel Kahneman, have the potential to expand this toolkit and yield “­medium-sized gains by nano-sized more aggressively address the underlying causes investments.”3 of problems. Because these fundamental causes The intellectual foundation laid by these writ- generally involve market failures due to exter- ings sparked a remarkable rise in the influence nalities, imperfect information and competition, of BE on public policy, the most celebrated and and widely recognized limits to consumer deci- widely adopted of which was the use of auto- sion making, the policies emerging from such matic defaults to sharply increase employee insights need not be more controversial than the savings for retirement. In addition, a series of policies espoused a decade ago by the initial set “informational nudges” applied simplification, of papers. social comparison, informational salience, and the reduction of “hassle” costs to improve II. Expanding the BE Policy Toolkit medical adherence, parental school choice, ( ) efficiency of home energy use, and take-up of We outline three principles to help guide social benefits see Madrian 2014 . Regulators ­policymakers in expanding the role of BE in further applied (BE to simplify disclosures) asso- public policy. The complexities of contemporary ciated with, for example, credit cards, insurance financial and health decisions, and the inability choice, and fuel economy, and even replaced the of individuals to navigate them, motivates a first infamous food pyramid with a simpler “food principle: policymakers ought to move beyond plate.” Beyond the specific policies it inspired, the simplification of choice environments toward the influence of BE is manifest in the estab- the simplification of the products and incentives lishment, growth, and global influence of the that underlie such choices. United Kingdom’s Behavioural Insights Team In recent years there has been an explosion BIT , which was charged with channeling evi- in the complexity of decisions confronting con- dence-based( ) insights from BE to improve the sumers. One study which examined the com- design and implementation of policies. The suc- plexity of 55,000 retail financial products e.g., cesses of the unit led other nations, including the for savings and investment offered to European( United States, to install similar institutions.4 consumers from 2002 to 2010) found that such While applauding these successes, and their products became increasingly complicated, and role in cultivating interest in evidence-based and that this complexity was correlated with higher BE-inspired policy, we believe that BE should firm profitability, lower product yield, and was often concentrated in markets populated with less sophisticated consumers Celerier and ( 3 Vallee 2014 . Consumers also face an expanded “Daniel Kahneman’s Gripe with Behavioral ) Economics.” The Daily Beast. http://www.thedailybeast. set of insurance options and credit instru- com/articles/2013/04/26/daniel-kahneman-s-gripe-with- ments, including payday loans and complex behavioral-economics.html. ­mortgages—a large fraction of which include 4 As a point of disclosure, Loewenstein has had numerous interactions with the BIT, and is on its advisory board, and teaser rates. Doing little to allay the problem, Bhargava has had a number of interactions with the Social and, in fact, likely exacerbating it, has been a and Behavioral Sciences Team in the United States. parallel surge of lengthy, and for a majority of 398 AEA PAPERS AND PROCEEDINGS MAY 2015 consumers, incomprehensible, informational example, the importance of fairness perceptions, disclosures from firms. emotion, and limited attention, implies that the In the midst of this growth in complexity, behavioral response to a new tax may depend surveys of financial literacy indicate that most not only on the underlying demand elasticity, Americans do not understand financial concepts but on the salience, complexity, nominal inci- such as interest, inflation, expense ratios or dence, and framing of the tax. Likewise, if wage diversification, and are unaware of the terms that expectations are subject to loss aversion, tiered govern their borrowing or the incentives facing ­unemployment benefits may be more effective those who provide them with products and ser- than standard unemployment insurance in short- vices. Not surprisingly, these deficits predict a ening unemployment spells DellaVigna et al. range of suboptimal financial behaviors with 2014 . ( respect to investments, savings, credit card debt, Beyond) improving the design of policies, choice of insurance, and mortgage delinquency insights from BE could be used to help mobi- Hastings, Madrian, and Skimmyhorn 2013 . lize citizens, through education and persuasion, Of( particular policy concern is that the burdens) rather than simple nudging, in support of poli- of complexity may be greatest for the econom- cies aimed at ambitious issues such as economic ically disadvantaged Mullainathan and Shafir inequality or climate change. 2013 . Simpler and more( standardized products and incentives) hold the promise of easing such III. Three Case Examples burdens and reducing inequities. A second principle is that policymakers should Examples of policies that are poorly designed aggressively protect consumers from behavioral for real humans, or that fail to address the deep exploitation by firms. The traditional economic causes of unwanted outcomes, abound. We dis- rationale for intervening in competitive markets cuss three, intentionally disparate, examples of to protect consumers is the existence of informa- policy domains in which nudges alone provide tional asymmetries. However, recent theoretical inadequate remedies but BE, broadly conceived, work e.g., Heidhues and Kőszegi 2014 sug- could help to deliver more meaningful solutions. gests that( the presence of unsophisticated) con- sumers may lead to a steady-state­ in which firms A. Health Insurance exploit consumer biases and inattention through, for example, the strategic use of product com- While BE is already an active and construc- plexity or marketing. tive participant in the domain of health, the We contend that the possible exploitation of insurance side of the market is a setting in which consumer biases should broaden the conven- poor decision-making may require a bolder tional rationale for consumer protection. While approach. In recent years, there has been a trend economists generally favor disclosures as an toward expanding the number of health plan efficient means of instituting such protections, choices available to consumers, often through given the practical limits of even well designed public or employee-sponsored exchanges. The disclosures c.f., Loewenstein, Sunstein, and rationale for greater choice is to better enable Golman 2014( , more aggressive measures may consumers to choose appropriate plans and to be required—for) example, taxation, or the improve prices and quality via provider com- explicit regulation of choice. Policies of this petition. Neither goal is likely to be fulfilled if type should not be particularly controversial consumers make uninformed plan choices. when, in their absence, firms would be moti- In our own recent research, we analyzed the vated to exploit consumers’ weaknesses. insurance choices of over 50,000 employees of A third principle is that BE should be lever- a large US firm which offered its employees an aged to improve the design and implementa- expansive, standardized, health plan menu in tion of policies based on traditional economics. which nearly all low-deductible plan options Given that an individual’s behavior is deter- were financially dominated by otherwise mined not by actual, but perceived, incentives, equivalent high-deductible plans Bhargava, BE points to a range of tactics that could enhance Loewenstein, and Sydnor 2015 . A majority( of the efficacy of conventional economic incentives employees, and especially low-income) employ- Congdon, Kling, and Mullainathan 2011 . For ees, selected dominated plans, which resulted in ( ) VOL. 105 NO. 5 BEHAVIORAL ECONOMICS AND PUBLIC POLICY 102: BEYOND NUDGING 399 unnecessary spending equivalent, on average, protections Acquisti, Brandimarte, and to 42 percent of annual premiums. Experiments Loewenstein 2015( . reported in the paper suggest that these adverse To the extent that) privacy disclosures are choices were associated with a lack of basic mandated, the “informational nudge” approach health plan literacy, that targeted education had would express them in a simple, vivid, and only moderate success in improving choice standardized fashion to heighten attention and quality. In recognition of the difficulty of choos- understanding. However, given that even sim- ing between plans, the insurance exchanges of plified information may be ignored or misinter- the ACA adopted a standardized online interface preted, simplified disclosures are unlikely to be and organized plans into metal tiers that were sufficient. Rather, the optimal strategy to protect linked to actuarial value. However, in another, consumers in these settings with asymmetric ongoing project, we find that the metal labels, interests may entail regulations that explicitly rather than facilitating more efficient choices, restrict firm use of information to purposes slightly worsened them. judged to be in the consumer’s interest and con- A superior solution to attempts to simplify the sistent with reasonable expectations—a strategy choice interface, but not fitting the definition of recently adopted by the European Union. a nudge, would be to mandate the provision of dramatically simplified and standardized insur- C. Climate Change ance policies or to even restrict plan menus to include only reasonable options. A recent paper BE has in recent years played a positive role Loewenstein et al. 2013 demonstrated not only in addressing climate change—for example, by that( a simplified product) of this sort could be providing input in the design of new fuel effi- understood by consumers, but indicated its fea- ciency labels or the use of social comparisons sibility; the simplified product was actually mar- to reduce home energy use. This role can and keted by the insurance company that funded the should be enlarged. Despite the severity of the research. The company’s copay-only plan listed threat posed by climate change, international different prices for different services but elimi- and domestic efforts toward concerted action on nated the deductibles and coinsurance that are a climate change policy have largely failed. major source of confusion. Requiring the provi- Standard economic theory offers clues as sion of such a simplified product, and dictating to the causes of this paralysis. The notion of a its characteristics, would improve plan choice free-rider problem helps to explain the failure and increase the likelihood that firms compete of individuals and states to unilaterally change on price and product quality. their behavior. Economic theory also provides regulatory and market-based solutions for deal- B. Privacy and Information Disclosure ing with externalities, such as a carbon tax, or a system of cap-and-trade. The Internet age, despite its many dividends, Although the path forward is clear, the pau- poses a grave challenge to individual privacy city of actual progress may be due, in large part, and, with that, critical questions for policymak- to psychological factors, such as motivated ers. Activities in which many of us routinely disbelief, the ostrich effect, confirmation bias, engage—e.g., e-mailing,­ texting, social media, present-bias, adaptation, and intangibility. In Internet searches, cell phone and E-ZPass a brilliant book titled Don’t Even Think About use—leave digital traces which make us vul- It: Why our Brains are Wired to Ignore Climate nerable to commercial exploitation, discrim- Change, George Marshall 2014 draws heavily ination, and unwanted monitoring. While the on BE to explain why individuals( ) and nations disclosure requirements favored by traditional are failing to act on, and in some instances economics should, in theory, enable individu- even denying the existence of, climate change. als to safely navigate the increasingly complex Marshall then draws on the same research to privacy landscape, limited attention, motivated identify strategies to capture the attention of the reasoning, and biased assessments of probabil- population and mobilize individuals and nations ity lead most to simply ignore such disclosures to take action. or, more worrisomely, to infer that the presence Research on collective action shows how con- of privacy disclosures implies non-existent­ centrated interests, such as energy firms, may 400 AEA PAPERS AND PROCEEDINGS MAY 2015 yield a strong political advantage over atomis- most reliable source of financial security in later tic individuals. However, concentrated inter- life. ests, such as tobacco companies, have in prior instances, with sufficient public backing, been surmounted. In conjunction with nudges aimed REFERENCES at facilitating environment-related behaviors, insights from BE could help to build sustained Acquisti, Alessandro, Laura Brandimarte, and and informed public support for, and enhance George Loewenstein. 2015. “Privacy and the efficacy and palatability of, the traditional Human Behavior in the Age of Information.” economic policies that offer the best hope for Science 347 6221 : 509–14. long-term solutions. Bhargava, Saurabh,( George) Loewenstein, and Jus- tin Sydnor. 2015. “Do Employees Make Sensi- IV. Conclusion ble Health Plan Decisions? Evidence from a Menu with Dominated Options.” Unpublished. In this essay we have discussed how BE has, Camerer, Colin, Samuel Issacharoff, George and can continue to, influence public policy. Loewenstein, Ted O’Donoghue, and Mat- While not dismissing prior successes, which have thew Rabin. 2003. “Regulation for Conser- served as critical precursors, we argue that the vatives: Behavioral Economics and the Case structural causes underlying many policy prob- for ‘Asymmetric Paternalism’.” University of lems, such as the increased decision complexity Pennsylvania Law Review 151: 1211–54. faced by individuals, demand more aggressive Celerier, Claire, and Boris Vallee. 2014. “What applications of BE. Stronger policy interven- Drives Financial Complexity? A Look into tions should not be taken lightly, given concerns the Retail Market for Structured Products.” about overreach and unexpected outcomes, but Unpublished. students of behavioral public policy should be Congdon, William J., Jeffrey R. Kling, and Send- made aware of the full complement of policy hil Mullainathan. 2011. Policy and Choice: tools—including taxes, subsidies, regulations, Public Finance through the Lens of Behavioral nudges, and educational campaigns—at their Economics. Washington, DC: Brookings Insti- disposable. The principles in this essay can be tution Press. applied far beyond our few examples to a range DellaVigna, Stefano, Attila Lindner, Balázs Reizer, of policy domains in public finance Congdon, and Johannes F. Schmieder. 2014. “Refer- Kling, and Mullainathan 2011 , health( Volpp ence-Dependent Job Search: Evidence from et al. 2011 , and education Lavecchia,) Liu,( and Hungary.” Unpublished. Oreopoulos) 2014 . ( Hastings, Justine S., Brigitte Madrian, and Wil- There are, in )fact, examples of successful liam L. Skimmyhorn. 2013. “Financial Lit- mergers between traditional and appropriately eracy, Financial Education, and Economic ambitious BE approaches. One, which we have Outcomes.” Annual Review of Economics 5 already alluded to, is the set of policies which, 1 : 347–73. in recent decades, helped to sharply curb smok- Heidhues,( ) Paul, and Botond Ko˝szegi. 2014. “Using ing. Traditional levers such as taxes and bans, Information about Naivete to Price Discrimi- supplemented with more psychologically moti- nate.” Unpublished. vated measures—vivid warning labels, massive Herrnstein, Robert, George Loewenstein, Drazen public health campaigns, and the more recent Prelec, and William Vaughan Jr. 1993. “Utility development of medical patches and devices Maximization and Melioration: Internalities designed to provide less harmful substitutes for in Individual Choice.” Journal of Behavioral tobacco—produced one of the most success- Decision Making 6 3 :149–85. ful policy initiatives in US history. And, while Lavecchia, Adam M., Heidi( ) Liu, and Philip Oreo- nudges have commendably encouraged many poulos. 2014. “Behavioral Economics of Edu- employees to save, many others, particularly cation: Progress and Possibilities.” National among the impoverished, confront retirement Bureau of Economic Research Working Paper with little to no savings. For many of these cit- 20609. izens, Social Security, a heavy-handed policy Loewenstein, George, Joelle Y. Friedman, Bar- conceived decades ago, remains the main and bara McGill, Sarah Ahmad, Suzanne Linck, VOL. 105 NO. 5 BEHAVIORAL ECONOMICS AND PUBLIC POLICY 102: BEYOND NUDGING 401

Stacey Sinkula, John Beshears, et al. 2013. Mullainathan, Sendhil, and Eldar Shafir.2013. “Consumers’ Misunderstanding of Health Scarcity: Why Having Too Little Means So Insurance.” Journal of Health Economics 32 Much. New York: Times Books. 5 : 850–62. Thaler, Richard H., and Cass R. Sunstein. 2003. Loewenstein,( ) George, Cass R. Sunstein, and Rus- “Libertarian Paternalism.” American Eco- sell Golman. 2014. “Disclosure: Psychology nomic Review 93 2 : 175–79. Changes Everything.” Annual Review of Eco- Thaler, Richard H., (and) Cass R. Sunstein. 2008. nomics 6 1 : 391–419. Nudge: Improving Decisions About Health, Madrian, Brigitte( ) C. 2014. “Applying Insights Wealth, and Happiness. New Haven, CT: Yale from Behavioral Economics to Policy Design.” University Press. Annual Review of Economics 6 1 : 663–88. Volpp, Kevin G., David A. Asch, Robert Galvin, Marshall, George. 2014. Don’t( Even) Think and George Loewenstein. 2011. “Redesigning About It: Why Our Brains Are Wired to Employee Health Incentives––Lessons from Ignore Climate Change. New York: Blooms- Behavioral Economics.” New England Journal bury USA. of Medicine 365 5 : 388–90. ( ) This article has been cited by:

1. Oren Bar-Gill, Cass R. Sunstein. 2015. Regulation as Delegation. Journal of Legal Analysis 7, 1-36. [CrossRef]