1. The i-frame and the s-frame
The behavioral and brain sciences primarily focus on what we call the i-frame: On individuals, and their thoughts and behaviors. Public policy, by contrast, typically focuses on the s-frame: The system of rules, norms, and institutions usually studied by economists, sociologists, legal scholars, and political scientists.
Historically, i-frame insights engage with public policy through evidence about which s-frame policies will work. Thus, research on neural and cognitive mechanisms of imitation has been linked to the impacts of media violence (Bandura, Ross, & Ross, Reference Bandura, Ross and Ross1963; Hurley, Reference Hurley2004). The neuroscience and psychology of addiction has informed the regulation of recreational drugs, cigarettes, alcohol, and gambling (Robinson & Berridge, Reference Robinson and Berridge2000; Verdejo-Garcia et al., Reference Verdejo-Garcia, Lorenzetti, Manning, Piercy, Bruno, Hester and Ekhtiari2019; Volkow & Boyle, Reference Volkow and Boyle2018). Health psychologists, epidemiologists, and public health doctors have studied the physiological and psychological mechanisms that convert s-frame factors (e.g., status, inequality, isolation, food environments) into health outcomes (see, e.g., Harris, Bargh, & Brownell, Reference Harris, Bargh and Brownell2009; Leigh-Hunt et al., Reference Leigh-Hunt, Bagguley, Bash, Turner, Turnbull, Valtorta and Caan2017; Marmot, Reference Marmot2004; Marteau, Hollands, & Fletcher, Reference Marteau, Hollands and Fletcher2012; Pickett & Wilkinson, Reference Pickett and Wilkinson2015). Insights about individual psychology thus inform regulation, taxation, social support, and institutional reform. We advocate deepening and extending this work.
Recently, there has been increasing enthusiasm for a more direct approach: Using i-frame insights to create i-frame policies (Camerer, Issacharoff, Loewenstein, O'Donoghue, & Rabin, Reference Camerer, Issacharoff, Loewenstein, O'Donoghue and Rabin2003; Sunstein & Thaler, Reference Sunstein and Thaler2003; Thaler & Sunstein, Reference Thaler and Sunstein2003). Two founding papers identified individual limitations (e.g., excessive self-interest, present bias, confirmation bias), not systemic issues, as the source of social problems. Sunstein and Thaler (Reference Sunstein and Thaler2003, p. 1162) wrote “Drawing on some well-established findings in behavioral economics and cognitive psychology, we emphasize the possibility that in some cases individuals make inferior decisions in terms of their own welfare – decisions that they would change if they had complete information, unlimited cognitive abilities and no lack of self-control.” Camerer et al. (Reference Camerer, Issacharoff, Loewenstein, O'Donoghue and Rabin2003) likewise note “To the extent that the errors identified by behavioral research lead people not to behave in their own best interests, paternalism may prove useful.” The first three chapters of Nudge (Thaler & Sunstein, Reference Thaler and Sunstein2008), including the updated “final edition” (Thaler & Sunstein, Reference Thaler and Sunstein2021), contrast the biases and self-destructive behaviors of “humans,” with the rational actors of economic theory. Unlike traditional policies, i-frame interventions don't fundamentally change the rules of the game, but make subtle adjustments to help fallible individuals play the game better.Footnote 1
These approaches are not mutually exclusive. For example, the battle against cigarette smoking includes individual and systemic measures (e.g., gruesome labels on cigarette packages and tobacco taxes and smoking bans). Similarly, in pensions, the i-frame change of auto-enrollment is often part of wider changes (e.g., requiring or incentivizing executives to offer pensions to workers). Moreover, the boundary between i- and s-frame policies is not always clear-cut. For example, if individuals aren't sufficiently aware of a default setting, then changing that default could be a mandate by subterfuge: The veneer of free choice is maintained without the substance. But i-frame interventions that slide into s-frame mandates are against the spirit of the new approach, which is to encourage “good” choices while respecting individual liberty.
Freedom aside, shifting the focus to i-frame interventions is also pragmatically appealing. Traditional public policy measures often get snared in legislative thickets (especially when politics is polarized) and can be dauntingly costly. The hope is that “small changes can make a big difference.”Footnote 2 As the labels “libertarian paternalism” and “regulation for conservatives” hint, clever interventions to help people help themselves are intended to be politically uncontroversial.
The goal is not merely to create a smoother “interface” between government and citizens (by analogy, say, with mobile phone design), which we see as entirely appropriate. It is much more ambitious: To provide an alternative to traditional s-frame policies. For example, in a technology, entertainment and design (TED) talk the year before he became the British Prime Minister David Cameron, who established the first “nudge unit,” said “The best way to get someone to cut their electricity bill is to show them their own spending, to show them what their neighbors are spending, and then show what an energy conscious neighbor is spending…Behavioral economics can transform people's behavior in a way that all the bullying and all the information and all the badgering from a government cannot possibly achieve.” Presumably, the “bullying” and “badgering” is traditional regulation: Taxes and energy efficiency standards. Cameron hopes that i-frame solutions make old-fashioned s-frame approaches redundant.
We shared such hopes, and most of our own policy-oriented research has focused on i-frame interventions. But we now worry that in advancing i-frame solutions to problems, we have inadvertently assisted corporations that oppose s-frame reforms. These corporations consistently cast societal problems as issues of individual weakness and responsibility, the solutions to which involve “fixing” individual behavior.
In the remainder of this section, we outline our overall argument. Next, we illustrate our concerns in a series of case studies. Finally, we outline the crucial positive role that we believe the behavioral and brain sciences can and should play in informing s-frame policy.
Let us begin with an analogy: That seeing individual cognitive limitations as the source of society's problems is like seeing human physiological limitations as the key to the problems of malnutrition or lack of shelter. Humans are vulnerable to cold, malnutrition, disease, predation, and violence. An i-frame perspective would focus on tips to help individuals survive in a hostile world.Footnote 3 But human progress has arisen through s-frame changes – the invention and propagation of technologies, economic institutions, and legal and political systems has led to spectacular improvements in the material dimensions of life. Human physiology varies little over time. But the systems of rules and institutions we live by have changed immeasurably. Successful s-frame changes have been transformative in overcoming our physiological frailties.
Our suspicion is that the same is true of our cognitive frailties. Just as mechanisms for governing common resources help counteract self-interest (Cramton, MacKay, Ockenfels, & Stoft, Reference Cramton, MacKay, Ockenfels and Stoft2017; Ostrom, Reference Ostrom1990), many institutions help overcome psychological frailties (Heath, Larrick, & Klayman, Reference Heath, Larrick and Klayman1998; Laibson, Reference Laibson2018). For example, competition in science or the adversarial nature of legal disputes is a partial antidote to confirmation bias (Kunda, Reference Kunda1990) and motivated reasoning (Nickerson, Reference Nickerson1998). Likewise, the impersonal framing of the law counteracts favoritism (Greenwald & Pettigrew, Reference Greenwald and Pettigrew2014); limited liability may help overcome risk and loss aversion (de Meza & Webb, Reference de Meza and Webb2007), which might otherwise stifle entrepreneurial activity; workplace and state pensions help deal with the bias for present gratification (Laibson, Reference Laibson1997); social taboos and legal restrictions counteract visceral impulses (Loewenstein, Reference Loewenstein1996); and arbitrary markers for distinct cultural groups may help people coordinate their behavior (Efferson, Lalive, & Fehr, Reference Efferson, Lalive and Fehr2008). In short, history shows that the solution to individual frailty is to change the system, not to enhance the individual.
I-frame interventions alone are likely to be insufficient to deal with the myriad problems facing humanity. Indeed, disappointingly often they yield small or null results. DellaVigna and Linos (Reference DellaVigna and Linos2022) analyze all the trials run by two large US nudge units: 126 randomized controlled trials (RCTs) covering 23 million people. Although the average impact of nudges reported in academic journals is large – at 8.7% – their analysis yielded a mean impact of just 1.4%. Why the difference? They conclude that selective publication in academic journals explains about 70% of the discrepancy.Footnote 4 DellaVigna and Linos also surveyed nudge practitioners and academics, to predict the effect sizes their evaluation would uncover. Practitioners were far more pessimistic, and realistic, than academics, presumably because of their direct experience with nudge interventions.
Even when i-frame interventions are highly effective, their impact may be modest. For example, consider a recent large-scale field trial which showed that over 85% of Swiss individuals and 75% of businesses who were defaulted into a more expensive green energy tariff stuck with this tariff over many years (Liebe, Gewinner, & Diekmann, Reference Liebe, Gewinner and Diekmann2021). The authors estimate this mechanism could yield very large carbon savings. In an optimistically titled commentary on this work (“Green defaults can combat climate change”), Sunstein (Reference Sunstein2021, p. 548) begins by contrasting i- and s-frame approaches:
It has long been thought that to reduce environmental harm, the best approach is an economic incentive, perhaps a corrective tax. In recent years, however, increasing attention has been given to non-monetary interventions including “nudges,” such as information disclosure, warnings, uses of social norms, and default rules. A potentially promising intervention would automatically enroll people in green energy, subject to opt-out.
But the ultimate impact is likely to be slight. The energy system does not respond by instantaneously producing more green energy for newly defaulted consumers. Rather, existing green energy is reallocated, with no direct impact on the energy mix. Moreover, the policy could not be applied universally because there would be insufficient green energy to “go round.” Admittedly, if it could be rolled out almost universally, the policy might generate a sufficient price “premium” for green energy to boost investment – but that very premium would push people away from the green tariff, and likely generate a media and political backlash. Worse, investment costs would be inequitable and divisive, allowing free-riders to avoid investing in the public good of green energy. This type of i-frame intervention is not an alternative to the s-frame measures that have successfully decarbonized the power system in many countries.
There is a deeper concern: i-frame interventions may draw attention and support from crucial s-frame changes. Five increasingly specific lines of evidence suggest that this is a serious problem:
(1) The brain represents stimuli of all kinds in only one way at a time. Thus, once a representational “frame” is adopted, other frames are difficult to access. This is evidenced throughout the cognitive and neurosciences, from perceptual rivalry (e.g., Kornmeier & Bach, Reference Kornmeier and Bach2012), functional fixedness in problem-solving (Duncker, Reference Duncker1945), or the mind's apparent limitation to a single “mental model” in reasoning (e.g., Johnson-Laird, Reference Johnson-Laird1983). Irrespective of cognitive domain, different frames compete; where several are available, a focus on one tends to crowd out others.
(2) Work on causal attribution indicates that people see responsibility as divisible – the causal responsibility associated with one factor or agent varies inversely with that of any other (Chockler & Halpern, Reference Chockler and Halpern2004; Lagnado, Gerstenberg, & Zultan, Reference Lagnado, Gerstenberg and Zultan2013). This implies that voters and policy makers alike will judge s-frame causes as less important, when focusing on i-frame factors.
(3) A possible mechanism for such displacement effects is “single-action bias” (Weber, Reference Weber, Bazerman, Messick, Tensbrunsel and Wade-Benzoni1997). Weber found that US farmers who had adapted their agricultural practices to climate change were less supportive of government climate policies, and Hansen (Reference Hansen2004) found parallel results in Argentina. Weber (Reference Weber2006) hypothesized that action to cope with a problem reduces fear, and hence the perceived importance of other risk reduction strategies. This effect occurs even if the action taken is not the most effective option, or where multiple actions are needed.
(4) The “competition” between i-frame and s-frame explanations of behavior may be tilted toward the i-frame. Indeed, the tendency to underestimate situational factors and overestimate individual factors is viewed by many as the key finding of social psychology, known as the “fundamental attribution error” (Ross, Reference Ross and Berkowitz1977) or “correspondence bias” (Gilbert & Malone, Reference Gilbert and Malone1995). Thus, business interests advancing i-frame solutions may have benefit from a tailwind of human psychology.Footnote 5
(5) Finally, direct experimental evidence shows that the i-frame can “crowd out” s-frame considerations in policy-relevant contexts (Thøgersen & Crompton, Reference Thøgersen and Crompton2009). Hagmann, Ho, and Loewenstein (Reference Hagmann, Ho and Loewenstein2019) show that merely alerting people (including policy makers in one study) to the possibility of an i-frame intervention (a green energy nudge) reduces support for more substantive policies (a carbon tax). Moreover, they find that a green energy nudge appears to crowd out support for a carbon tax by providing the false hope that climate change can be addressed without costlier (but immeasurably more effective) policies. When individuals are informed about the limited impact of the green energy nudge, however, their support for a carbon tax increases.
Werfel (Reference Werfel2017) finds these effects in the field. Households in Japan who were randomly assigned to report actions they took to save energy were less supportive of a carbon tax, and those who reported more actions were especially unsupportive. Werfel concludes that the effect is “driven by an increase in the perceived importance of individual actions relative to government regulation” (Werfel, Reference Werfel2017, p. 512). Truelove, Carrico, Weber, Raimi, and Vandenbergh (Reference Truelove, Carrico, Weber, Raimi and Vandenbergh2014) find mixed results when it comes to support for recycling policies, but greater overall support for negative than positive spillovers.
Finally, Maki et al. (Reference Maki, Carrico, Raimi, Truelove, Araujo and Yeung2019) conclude in their meta-analysis of proenvironmental behavior (PEB) that spillovers from PEBs to intentions are positive; but spillovers from PEBs to actual behavior and, crucially, policy support, are negative (though small).Footnote 6 Collectively, these studies highlight a general propensity for i-frame solutions to undermine support for available s-frame policies.
Beyond crowd-out effects in public support, there are three further ways in which i-frame interventions can undermine s-frame policies. First, policies require human and financial resources: Pursuing one policy can interfere with pursuing others. Coronavirus disease (COVID) provides a recent illustration: ChinaFootnote 7 and New ZealandFootnote 8 relied on isolation, and did not initially push hard on vaccination. Furman (Reference Furman2016, p. 3) notes that “policymakers have a finite amount of time and attention, so every policy action taken has a cost in terms of other actions that they are unable to undertake as a result… Thus, even a high benefit-to-cost ratio may not be sufficient justification for pursuing a policy if it crowds out the time and attention that might have gone into other policies with higher absolute net benefits.”
Second, there will also be unavoidable crowd out of research resources. Social and behavioral scientists face constraints on time, effort, and funding, so that a focus on nudges almost inevitably reduces effort elsewhere. We are by no means calling for the suppression of specific types of research; but, as we argue below, a reprioritization could help both science and society.
Third, a focus on i-frame interventions can shift the standards of what counts as quality evidence for public policy. For many i-frame policies, RCTs are seen as the gold standard for evaluating and incrementally improving policy, and as the crucial contribution of behavioral insights research (Luca & Bazerman, Reference Luca and Bazerman2021).Footnote 9 But this gold standard itself pushes toward i-frame interventions (where different individuals may be randomly assigned to distinct interventions) and away from s-frame interventions where it is rarely possible to change the “system” for some subset of the population.Footnote 10, Footnote 11 As Hansen (Reference Hansen2018, p. 193) relates about his interactions with policy makers, “It is my repeated experience that we can quite easily run a letter-tweaking experiment involving thousands of taxpayers, but only provoke strenuous smiles when we say, ‘We could also try to rethink the policy assumptions.’”
S-frame policies are not inherently superior to i-frame policies. Many do not have their intended effect,Footnote 12 or even backfire.Footnote 13 But to evaluate the likely efficacy of s-frame policies, the natural approach is rarely experimental. Instead, the s-frame encourages us to ask where, when, and why a problem arose, and to explore differences between and within countries. Such analyses can provide clues about problems' origins, as well as ideas about how they could be addressed, perhaps by reversing the historical changes or adopting s-frame approaches that have proven successful elsewhere.
The idea that support for i-frame interventions crowds out support for more substantive and effective s-frame ones receives indirect support from another observation, which is the central focus of this review: The powerful and consistent support that i-frame interventions have received from interests that are opposed to s-frame reform. Picking up our previous analogy, slum landlords (by analogy with corporations opposing s-frame reform) will see illness as arising from poor hand-washing or unhygienic food and drink preparation. And well-intentioned behavioral scientists may suggest i-frame interventions to increase the use of soap and boiled water, probably to a little effect. But the i-frame perspective may itself weaken the impetus for tried-and-tested s-frame reform: Regulations to enforce quality housing, with heating, sanitation, and safe drinking water.
Over many decades we show that public relations specialists representing corporate interests have effectively deflected pressure for systemic change by reframing social problems in i-frame terms. They have learned to back i-frame interventions that pose little threat to the status quo while simultaneously lobbying heavily against proven s-frame changes that threaten their interests. The billions of dollars spent promoting i-frame interventions should make behavioral scientists uneasy. With the best of intentions, proponents of i-frame policy, including ourselves, may have inadvertently weakened support for crucial systemic changes.Footnote 14 As we review below, there is every reason to believe that this has happened.
These considerations do not imply that i-frame research should be abandoned. Indeed, many influential advocates of i-frame policies have long seen them as complementing, rather than replacing, s-frame policy (e.g., Sunstein, Reference Sunstein2022a; Thaler & Sunstein, Reference Thaler and Sunstein2021). But it does imply that behavioral scientists need to be aware of, and actively counter, any tendency to view i-frame interventions as alternatives to system change. Moreover, the relative impacts of i- and s-frame interventions strongly suggest that behavioral scientists should prioritize applying behavioral insights to s-frame reform.
1.1. Climate change and the i-frame
In the early 2000s, the world's second largest non-state-owned oil company, BP, began an enormous media campaign with the tag-line Beyond Petroleum to improve its environmental image. Mann (Reference Mann2021) documents how BP and its fossil-fuel allies had long challenged the reality of climate change by supporting climate-skeptical academics and discrediting legitimate climate scientists. As this approach became increasingly indefensible, they shifted gears. Rather than opposing climate science directly, they worked to reframe the problem of carbon reduction in i-frame, not s-frame, terms, beginning what Mann calls the “new climate wars”: promoting the idea that opposing climate change demands individual responsibility, not systemic reform.
A key strategy was to promote the personal “carbon footprint” (Safire, Reference Safire2008), in part through BP's carbon footprint calculator, which was completed by nearly 300,000 people in 2004 (Solman, Reference Solman2008). The campaign succeeded spectacularly: Individuals, campaigners, media organizations, and government agencies all created their own carbon calculators to help people reduce their impact on the planet.Footnote 15
BP's campaign promotes the i-frame by helping individuals reduce their own personal carbon footprint, and behavioral scientists have jumped aboard by testing, and advancing the implementation of, a variety of “green energy nudges.” The most prominent, based on ideas pioneered by Cialdini (Reference Cialdini1984) and highlighted by Cameron in his TED talk, involves showing people graphs comparing their own home-energy use with that of their neighbors (Schultz, Nolan, Cialdini, Goldstein, & Griskevicius, Reference Schultz, Nolan, Cialdini, Goldstein and Griskevicius2007).
BP's campaign may seem constructive, or at worst innocuous. But Mann (Reference Mann2021) suggests that it is in fact a clever exercise in framing: Describing a problem in a particular way to shape the solutions that come to mind (Chong & Druckman, Reference Chong and Druckman2007; Lakoff, Reference Lakoff2014). BP's campaign, which included personal appeals such as “It's time to go on a low carbon diet” (Learmonth, Reference Learmonth2020), frames the challenge of combating climate change as a matter of individual responsibility.Footnote 16
Carbon footprints have certainly attracted, and perhaps distracted, behavioral scientists including ourselves. In consulting and advisory work, we have thought a lot about what interventions can help individuals reduce their use of heating, insulate their homes, and shift to low-carbon transport and more plant-based diets (e.g., Chater, Reference Chater, Shorthouse and Hall2020a; see also Allcott & Rogers, Reference Allcott and Rogers2014). But we now doubt that carbon emissions can be substantially reduced by i-level interventions such as providing small incentives, better (or more transparent) information, more feedback, more awareness of social norms, or greener “defaults.”Footnote 17 Having a real impact will require systemic transformation on a huge scale: Changing how we heat our homes, travel, ship goods, and produce and consume food; rethinking manufacturing; and vastly expanding the production, storage, and transmission of green electricity. Such transformations would likely include a substantial carbon tax alongside extensive regulation (e.g., Cramton et al., Reference Cramton, MacKay, Ockenfels and Stoft2017; Energy Transition Commission, 2021; Markard, Reference Markard2018), as well as redistributive transfers to deal with issues of unequal impacts.
The case of carbon footprints is part of a wider pattern (which we illustrate in the next section):
(1) Corporations with an interest in maintaining the status quo promote public relations (PR) messages that the problem at hand is one of individual responsibility, and that people need help to exercise that responsibility more effectively. That is, the challenge is cast in the i-frame.
(2) Behavioral scientists enthusiastically engage with the i-frame, and focus on frailties of thought and behavior as the source of problems. It thus seems natural that behavioral scientists are well-positioned to solve them, by helping individuals overcome their limitations.
(3) There are hopes that i-frame interventions (nudges, providing individual-level incentives, information, and education) provide cheap and effective alternatives to conventional s-frame policies, such as regulation and taxation. This hope distracts from the s-frame. It also promises that “heavy-handed” s-frame interventions can be avoided in favor of cheap, incremental, and often politically palatable light-touch measures.
(4) The i-frame interventions yield disappointing results, and divert attention and effort from much needed s-frame reforms, bolstering the status quo.
(5) Corporations relentlessly target the s-frame, where they know the real leverage lies. They spend substantial resources on media campaigns, lobbying, think-tanks, and academic research sponsorship to ensure that the “rules of the game” reinforce the status quo.
Our focus in this paper is the unwitting (by academics) alignment of interests between corporations and behavioral scientists focusing on the i-frame. We leave aside the many direct attempts by business to coopt academia, dating back at least to the cigarette industry's mobilization of academics skeptical of a link between smoking and cancer (Brandt, Reference Brandt2012). Across the topics outlined below, there are direct initiatives by businesses to back academics who support specific industry-friendly positions. The same motivation likely underlies (more indirect) corporate campaigns to advance i-frame interpretations of societal problems.
Although some corporate tactics (e.g., regarding tobacco and climate disinformation) have been challenged as both legally and ethically deeply questionable, the broader tendency of companies to invest in PR and lobbying to promote their interests is almost inevitable, as predicted by economic theory (e.g., Grossman & Helpman, Reference Grossman and Helpman1994), and described in the empirical literature (e.g., Bombardini & Trebbi, Reference Bombardini and Trebbi2019). Here, too, an s-frame perspective is appropriate, rather than attributing the problem to the “greed” or other moral failings of individual executives. They too are working within the incentives and rules of a system which virtually requires that they promote their companies' interests, irrespective of their personal views. We take it to be uncontroversial that companies lobby to oppose s-frame reform. What is less obvious is that such PR and lobbying often operate by an indirect, but very powerful, mechanism: The promotion of i-frame solutions.
We now illustrate our argument in detail for four more policy domains (obesity, retirement savings, plastic waste, and rising healthcare costs), then more briefly for six others. Lastly, we outline a positive vision for an s-frame-oriented behavioral public policy.
2. Case studies: How i-frame behavioral public policy went wrong
Table 1 reviews prototypical i-frame and s-frame interventions for the four examples we discuss in detail, as well for climate change. Each case conforms (with variations) to the five steps outlined above.
Table 1. Potential i-frame and s-frame interventions to address public policy problems
Although we primarily focus on behaviorally inspired i-frame interventions, i-frame thinking also includes legal disclaimers, conflict of interest disclosures, information provision (e.g., calorie or carbon labeling), which are aimed at helping people make better decisions. Behavioral scientists tend doubt of the effectiveness of such strategies – and with good reason (Golman, Hagmann, & Loewenstein, Reference Golman, Hagmann and Loewenstein2017; Loewenstein, Sunstein, & Golman, Reference Loewenstein, Sunstein and Golman2014).
2.1. Obesity
The problem of obesity is one of the major public health challenges facing the developed world: We are, collectively, eating too much and exercising too little (Hill, Wyatt, & Peters, Reference Hill, Wyatt and Peters2012). But why? Drawing on a vast literature on intertemporal choice from neuroscience (McClure, Laibson, Loewenstein, & Cohen, Reference McClure, Laibson, Loewenstein and Cohen2004), experimental psychology (Rachlin & Green, Reference Rachlin and Green1972), behavioral economics (O'Donoghue & Rabin, Reference O'Donoghue and Rabin1999), and philosophy (Parfit, Reference Parfit1984), researchers have often seen obesity as exemplifying the i-frame weakness of present-bias, by which lure of cake now overwhelms the long-term benefits of better health (see Chs. 13, 17, and 18 of Loewenstein, Read, & Baumeister, Reference Loewenstein, Read and Baumeister2003; but see also Ch. 16).
Yet variations in obesity over time and across counties reveal the limitations of such a perspective. There is no evidence that present bias has changed over time or place in ways that explain variations in obesity. Indeed, we know of no empirical evidence causally connecting obesity to present-bias. There is, however, strong evidence that people who migrate often take on the obesity characteristics of their new locality (Schulz et al., Reference Schulz, Bennett, Ravussin, Kidd, Kidd, Esparza and Valencia2006). Obesity is not the product of individual fallibility, but systemic factors.
The food industry encourages academics to focus on i-frame solutions to obesity, including attempts to deflect concern away from food as the source of the problemFootnote 18 and discredit academics with opposing views.Footnote 19 Brownell and Warner (Reference Brownell and Warner2009) identify the central plank of the industry's strategy: “Focus on personal responsibility as the cause of the nation's unhealthy diet,” taking the food system as a given.
Behavioral scientists have unwittingly jumped on this i-frame bandwagon, proposing and testing a wide variety of i-frame interventions.Footnote 20 Large bodies of research, including many papers by one of us, have explored (a) proximate interventions in people's interactions with food, including innovative calorie labeling and advanced ordering of meals (see Downs & Loewenstein, Reference Downs, Loewenstein and Cawley2011); (b) specially crafted incentives to motivate weight loss (e.g., Volpp et al., Reference Volpp, John, Troxel, Norton, Fassbender and Loewenstein2008); and (c) promoting exercise, typically by paying people to go to the gym (Charness & Gneezy, Reference Charness and Gneezy2009).
These i-frame interventions are often viewed as alternatives to legislation and taxation. In advocating for his HealthierUS initiative, for example, which promoted exercise and healthier food choices, former President George W. Bush expounded that:
We have a problem when people don't exercise and eat bad food. Obesity can cause serious health problems, like heart disease and diabetes…We must reverse the trend, and we know how to do it. It's exercise and good dieting. Good foods and regular exercise will reverse the trend and save our country a lot of money but, more importantly, save lives. (Bush, Reference Bush2003)
But i-frame interventions have proven disappointing: (a) Proximate interventions on food ordering produce small, though statistically detectable effects (e.g., VanEpps, Downs, & Loewenstein, Reference VanEpps, Downs and Loewenstein2016), (b) people regain weight once incentives are removed (John, Loewenstein, & Volpp, Reference John, Loewenstein and Volpp2012); and (c) although incentives increase gym visits, a measurable impact on actual weight has not been demonstrated (Charness & Gneezy, Reference Charness and Gneezy2009).
While publicly promoting an “i-frame” perspective on obesity, the food industry and agribusiness relentlessly lobbies at the s-frame: To maintain and modify laws and regulations to their advantage.Footnote 21 Individual consumers are no match for concentrated firms united by industry associations and armed with lobbyists, in line with Olson's (Reference Olson1965) classic analysis in The Logic of Collective Action. Individuals often care desperately about their waistlines and health, devoting huge amounts of time and money in (usually unsuccessful) attempts to get, and stay, thin (e.g., Polivy & Herman, Reference Polivy and Herman2002). It is not in any individual's interest to expend time or money to exert an infinitesimal influence on the overall food system.Footnote 22 But it is in the interests of a concentrated and highly organized food industry to spend vast sums to do so.
This argument in no way denies the fallibility of human behavior, or that such fallibility matters for s-frame public policy. Indeed, as we discuss in section 3, an i-frame understanding of human weakness is crucial to explaining how it can be exploited so effectively – for example, by producing and marketing products that cater to our evolved weakness for sugar and fat. As public health researchers Nestle and Jacobson (Reference Nestle and Jacobson2000) note, “Changes in the food environment help explain why it requires more and more willpower for Americans to maintain an appropriate intake of energy.” Moss (Reference Moss2013) documents a meeting of top food industry executives at which some acknowledged their leading role in the obesity epidemic, but could not agree on initiatives to curb it. Moss explains how the food industry uses the physiology and psychology of food consumption to reach consumers' “bliss points,” maximizing craving and in some cases suppressing cues for satiation.Footnote 23
Focusing on individual-level causes and remedies for obesity risks displacing researcher time, financial resources, and journal pages from deeper thinking and in-depth research about what caused the obesity epidemic, about s-frame interventions to reverse it, and about how to marshal behavioral science to help make such interventions successful. Misattributing problems to individual weakness rather than systemic factors also implicitly blames individuals – and encourages them to blame themselves – for their inability to swim against powerful currents they have little hope of resisting.
2.2. Inadequate provision for retirement
The citizens of the United States and many high-income countries are financially unprepared for retirement. Half of US families have no retirement savings whatsoever (Morrissey, Reference Morrissey2019), and over half of workers between 60 and 65 believe their savings are not “on track” for retirement (Board of Governors of the Federal Reserve System, 2021). Indeed, the median retirement account savings balance for people approaching retirement is just $21,000 (Morrissey, Reference Morrissey2019). Although the average Social Security payment is only $1,543/month (Social Security Administration, 2021), over two-thirds of US retirees identify Social Security as their primary income (Transamerica Center for Retirement Studies, 2020).
Inadequate saving vies with obesity as behavioral scientists' favorite illustration of “present-bias” (e.g., Laibson, Reference Laibson1997): The long-term benefits of saving are presumed to be overwhelmed by the immediate pleasures of spending. But, as with obesity, this story is implausible in a historical and cross-national context.
Preparing for retirement was, until quite recently, achievable for Americans. Until the 1980s, most companies offered pensions paying predictable amounts on retirement (typically pegged to their final salary).Footnote 24 Problems began with the emergence of “defined-contribution” retirement plans (a euphemism for “save for your own retirement”). Originally envisioned to supplement pensions,Footnote 25 in the 1980s, companies found they could drastically reduce the cost, administrative burden, and risks of pension schemes by offloading funding, and investment decisions, to their employees. And, once some employers ceased to fund pensions, competitive pressures forced their competitors to follow suit. Defined-benefit plans are now largely confined to public sector workers, and, even there, there is a trend toward defined-contribution schemes. Some employers do match employee contributions; but many offer no retirement plans at all.
Cross-national comparisons reveal that inadequate savings is not a universal problem. Australia is a particularly telling example.Footnote 26 Not long ago, it was one of the few countries with lower retirement savings levels than the United States, but it is now a frontrunner in retirement preparedness following far-reaching systemic reforms: Universalizing retirement saving, mandating substantial employer and employee contributions, and prohibiting withdrawals for almost any reason. The United States has taken the opposite route: Not mandating contributions by employers or employees, permitting withdrawals for a range of reasons, and even permitting borrowing against retirement funds.Footnote 27
A rapid shift back to conventional pensions would be financially onerous for most US companies, but especially for the financial services industry, built on the management of defined-contribution retirement plans. Not surprisingly, therefore, the industry adopts the i-frame: Taking the defined-benefit system as given, and focusing on helping people cope with it more adequately. Ads contrast the happy futures of those who put sufficient resources aside, with the struggles of those who do not. Likewise, the TIAA Institute, the research arm of the huge financial services company that manages many academics' retirement accounts, funds and posts on its website numerous research studies testing interventions to improve retirement preparedness, almost all of which aim to increase the level of saving within traditional defined-contribution plans.Footnote 28 From this perspective, the fate of struggling savers lies in their own hands. Such an i-frame perspective is understandable: Financial services firms could not be expected to propose policies that might put them out of business.
If defined-contribution retirement plans have such disastrous consequences, asks Atlantic author Frank Pasquale, then “why are policy makers so enamored of it?” Pasquale suggests that one reason is the hope that, with the right behavioral interventions (“nudges”), citizens can solve the problem on their own. But he suggests this hope is illusory:
Because the nudge is really a fudge – a way of avoiding the thornier issues at stake in retirement security. The most worrisome unexpected costs of old age, including medicine and personal care, should be addressed by politicians via programs such as Medicare and Medicaid. But by focusing on individuals' decisions to save up for retirement, they can shift responsibility.
This focus on the individual, rather than the wider social context, is not surprising, given that nudging comes out of microeconomics and psychology, two disciplines that tend to break the world into dyadic transactions between isolated individuals and firms. A sociological or political perspective, on the other hand, points to the real roots of retirement insecurity: a great shifting of risk from corporations to individuals. Workers can be urged to take all manner of “personal responsibility” for saving – but if their wages are stagnant while other costs are rising, it is hard to imagine that strategy really working.
Behavioral scientists have embraced the i-frame enthusiastically, proposing and testing different mechanisms to help people make the right pensions choices. As with green energy, one lever has been the power of defaults: Across the United States, the United Kingdom, and many other nations (OECD, 2020), people have been auto-enrolled into pension schemes, albeit with the possibility of opting out.Footnote 29 Present-bias has been tackled via “auto-escalation” – that is, allowing people to make low initial pension contributions, and ramp up contributions as their income grows (Thaler & Benartzi, Reference Thaler and Benartzi2004). In parallel, conventional i-frame interventions, such as improving customer understanding of pensions, and increasing “financial literacy” (Lusardi & Mitchell, Reference Lusardi and Mitchell2014) are proposed and tested (e.g., Mandell & Klein, Reference Mandell and Klein2009).Footnote 30
An interesting counterpoint to the i-frame interventions explored by behavioral scientists is provided by Willis's (Reference Willis2008) provocatively titled “Against Financial-Literacy Education.” Specifically critiquing i-frame interventions involving disclosure, Willis writes:
The dominant model of regulation in the United States for consumer credit, insurance, and investment products is disclosure and unfettered choice. As these products have become more complex, consumers' inability to understand them has become increasingly apparent, and the consequences of this inability more dire. In response, policymakers have embraced financial-literacy education as a necessary corollary to the disclosure model of regulation. (p. 197)
Willis questions whether such education really helps, and concludes:
Harboring this belief [that financial literary is the solution] may be innocent, but it is not harmless; the pursuit of financial literacy poses costs that almost certainly swamp any benefits… When consumers find themselves in dire financial straits, the regulation through education model blames them for their plight, shaming them and deflecting calls for effective market regulation… The search for effective financial literacy education should be replaced by a search for policies more conducive to good consumer financial outcomes. (p. 198)
i-Frame interventions provide a tempting alternative to urgent-needed s-frame reform: Radical reform of current defined-contribution plans.
Advocates of behavioral interventions acknowledge the difficulties with defined-contribution plans, but argue that i-frame interventions should be part of the solution. For example, Thaler (Reference Thaler2009) writes “Everyone's lost a lot of money on their 401(k) plans. I've heard some people calling them 201(k) plans. So it's even more important to get people to be saving more for retirement. Behavioral economics has helped us learn a lot about how to do that. One simple way is… automatic enrolment.” Former President Barack Obama called automatic enrollment a “common-sense, practical solution” to retirement savings (Jacobson, Reference Jacobson2012).
Auto-enrollment and auto-escalation are among the most ingenious and elegant i-frame interventions in any domain. Yet their impact has been disappointing, despite often being seen as the major success story for behavioral public policy.Footnote 31 David Laibson, once a leading advocate of i-frame solutions, concluded in a 2020 keynoteFootnote 32 that neither auto-enrollment nor auto-escalation have moved the needle on retirement saving.
First, even without auto-enrollment, many employees end up enrolling in the firms' retirement plans; auto-enrollment only slightly accelerates the process (Choi, Laibson, Madrian, & Metrick, Reference Choi, Laibson, Madrian, Metrick and Wise2004). Second, and more significant, is the problem of “leakage” (Argento, Bryant, & Sabelhaus, Reference Argento, Bryant and Sabelhaus2015): Employees often remove funds from retirement savings accounts, for example, when changing jobs, or borrow at low interest rates using their retirement balances as loan collateral. Third, auto-enrollment cannot help the many workers at companies which don't offer matches or provide no plan at all.Footnote 33 Incentives for companies to implement auto-enrollment were built into the Pension Protection Act of 2006, and pension enrollment did rise substantially at those firms that offered their workers defined-contribution pension plans (Engelhardt, Reference Engelhardt2011). But even when defaults apply, workers are often defaulted to low rates of contribution (Butrica & Karamcheva, Reference Butrica and Karamcheva2013). A decade after these reforms, US retirement saving remains stagnant (Morrissey, Reference Morrissey2019).
In the United Kingdom, auto-enrollment has been particularly undermined by default contributions often being set very low. Thus, many people wrongly believe they have “ticked the pensions box” while remaining woefully under-prepared for retirement (Decision Technology, 2017). Indeed, the tendency of low defaults to actually reduce contribution rates for workers who otherwise would have saved more was documented in the very first paper on the impact of defaults (Madrian & Shea, Reference Madrian and Shea2001).
Responding to an earlier paper (Loewenstein & Chater, Reference Loewenstein and Chater2017), in the concluding pages of Nudge: The Final Edition, Thaler and Sunstein (Reference Thaler and Sunstein2021) rightly stress that pension reforms in the United States and United Kingdom involve both i- and s-frame changes. For example, the NEST pension scheme in the United Kingdom (which helps employers of all sizes provide workplace pensions), is almost entirely an s-frame reform. Crucially, employers are required to provide such pensions. The i-frame “nudge” element – that they are defaulted in with an opt-out – is a relatively minor detail. It is typically the s-frame issues that really matter: Whether, as in the United States, employees can withdraw money from, or take loans against, their pension; or in the United Kingdom, the default pension contribution level.
For pensions, unlike most of the topics we discuss, there has been relatively little active lobbying against reinvigorating defined-benefit schemes or their equivalent. We suspect this is because there are few firms that would benefit from a shift back to such schemes, and many that would be harmed. If strong public support were to emerge for such reform, we would anticipate a reaction from the financial services industry paralleling that in health care (see below).
2.3. Plastic waste
The production and disposal of plastics, cans, bottles, bags, and containers provides a further illustrative example. Plastic bags clog sewage systems, kill about 100,000 marine mammals every year, and degenerate into toxic microplastics that pollute oceans and landfills. Worldwide, shoppers use around 500 billion single-use plastic bags annually.Footnote 34
Readers who see reducing waste as a matter of individual responsibility may be surprised, as we were, to discover that this i-frame perspective can be traced to the influence of industry. Consider, for example, the famous ‘Crying Indian’ ad (Mann, Reference Mann2021, pp. 52–60). In the ad, an actor in Native American dress paddles a birch bark canoe on water that becomes increasingly polluted, pulls his boat ashore, and walks toward a bustling freeway where a passenger hurls a paper bag out a car window. The ad concludes with an encapsulation of the i-frame perspective “People start pollution. People can stop it.” The wider “Keep America Beautiful” campaign (ubiquitous from the 1950s until today), of which the ad was a part, was conceived and funded by beverage and packaging corporations including the American Can Co., Owens-Illinois Glass Co., and later Coca-Cola and Dixie Cup.Footnote 35
Behavioral scientists have generated many potential interventions, particularly focusing on reducing littering (e.g., Keep Britain Tidy, 2015). For example, pictures of “watching eyes” are widely deployed in the United Kingdom, in the light of studies indicating that these prime prosocial behavior (Bateson, Nettle, & Roberts, Reference Bateson, Nettle and Roberts2006; Haley & Fessler, Reference Haley and Fessler2005), and especially litter reduction (Bateson et al., Reference Bateson, Robinson, Abayomi-Cole, Greenlees, O'Connor and Nettle2015). A highly cited intervention tested in Copenhagen in 2011 involves painted footprints leading to brightly colored trash bins, was found to reduce littering by a 46%.Footnote 36 Unfortunately, despite its apparent success, this approach does not seem to have been tested further, and appears to have been implemented in one other locality: Stirling, Scotland.Footnote 37 This highlights a broader problem: Even where interventions do work, they are difficult to sustain or scale-up. To our knowledge, there are currently no proven anti-littering initiatives operating at scale with a strong evidence base.Footnote 38 Putting “watching eyes” on packaging (which reduced littering of a leaflet in one field study) may be scalable (Bateson et al., Reference Bateson, Robinson, Abayomi-Cole, Greenlees, O'Connor and Nettle2015). However, considerations of cost, displacement of other packaging information, and potential diminution in impact if “watching eyes” become almost ubiquitous, all argue for caution.
For over 50 years the oil and plastics industries have further resisted efforts to curb plastic packaging by promoting the myth that large-scale plastic recycling is technically and economically feasible, in order to allay concerns about new plastic.Footnote 39 Yet, according to the Environmental Protection Agency (EPA), less than 10% of plastic has been recycled in the last 40 years. An National Public Radio (NPR) investigation titled “Plastic Wars: Industry Spent Millions Selling Recycling – to Sell More Plastic” found internal documents from the 1970s confirming that the industry always knew that recycling at scale would never be economically viable. Moreover, the plastics and oil industries have created and funded nonprofit organizations with names that belie their true purpose. The promisingly proenvironment sounding “Earth911” (https://earth911.com/about-earth911-mission-and-history/), for example, with industry partners including ExxonMobil, which promotes the recycling myth and focuses on the i-frame, states that “Thousands of individual small changes create a large, positive impact.”
While promoting the i-frame publicly, the food, beverage, and packaging industries have correctly identified that s-frame change is far more important. Indeed, s-frame interventions banning or taxing plastic use have proven highly effective. For example, in San Jose, CA, a plastic bag ban led to 89% fewer plastic bags in storm drains (60% in rivers and residential areas), and the average number of bags used per person decreased from 3 to 0.3.Footnote 40 As standard political economy considerations would predict, industry has therefore lobbied heavily against such s-frame interventions, with great success. Although in the United States only two states (CA and HI) have banned plastic bags, 10 (AZ, FL, IA, ID, IN, MI, MN, MO, MS, and WI) have legislated statewide preemptive bans on banning plastic bags, preventing municipalities from imposing bans or fees. These bans aren't spontaneous expressions of public hostility to an obscure policy; they arise from concerted lobbying.Footnote 41
Corporate interests have also actively opposed “Extended Producer Responsibility” measures for packaging, cigarettes, bottles, and other waste, an s-frame approach that aims to make producers bear the full social and environmental cost of their waste, thereby incentivizing product redesign to reduce that waste (Walls, Reference Walls2006). Where implemented, such schemes can be highly effective (Hanisch, Reference Hanisch2000; Walls, Reference Walls2006), and the approach is gaining momentum in the European Union (EU) and United Kingdom.Footnote 42
2.4. The high cost of US health care
Health care is increasingly expensive, especially in the United States, both for individuals and the economy at large; and results are often disappointing. As usual, time trends and international comparisons reveal this. The United States has not always been an outlier. In the 1980s, the US population was primarily insured in managed care plans with incentives to insurers and providers to keep down costs (Draper, Hurley, Lesser, & Strunk, Reference Draper, Hurley, Lesser and Strunk2002). Providers were mainly paid salaries, which were high but not lavish. In the 1980s, however, the United States began a crucial systemic shift: To a fee-for-service system and highly fragmented private insurance market, with high administrative costs and incentives to over-provide expensive, low benefit tests and services, as well as overpriced medications (Lesser, Ginsburg, & Devers, Reference Lesser, Ginsburg and Devers2003). US healthcare costs soon departed dramatically from those in comparable countries, and at this point US health costs are roughly twice the Organization for Economic Co-operation and Development (OECD) median, with no better than median results on almost all measures of health and health care. Higher US health costs do not arise because Americans are, individually, less healthy than people in other countries. For example, smoking rates in the United States are far lower (14% in 2019) than many countries with much lower health costs (e.g., France, Germany, and Spain; all with smoking rates substantially higher than 25% in 2021).Footnote 43
The US healthcare industry (e.g., insurers and providers) promotes an i-frame perspective: That high healthcare costs stem from poor health, which itself depends on individual fallibility. The message is conveyed through actions such as the provision of rewards for exercise (or subsidization of fitness clubs), despite little evidence that such incentives impact health (Redmond, Solomon, & Lin, Reference Redmond, Solomon and Lin2007).
Behavioral economists have often taken a similar line. In a typical passage from a large literature, Loewenstein, Brennan, and Volpp (Reference Loewenstein, Brennan and Volpp2007) wrote:
Individual behavior plays a central role in the disease burden faced by society. Many major health problems in the United States and other developed nations, such as lung cancer, hypertension, and diabetes, are exacerbated by unhealthy behaviors. Modifiable behaviors such as tobacco use, overeating, and alcohol abuse account for nearly one-third of all deaths in the United States. (p. 2415)
A huge range of i-frame interventions have been proposed to improve health.Footnote 44 But incentives, reminders, and apps have shown little success, either in changing behavior or improving outcomes (e.g., Volpp et al., Reference Volpp, Troxel, Mehta, Norton, Zhu, Lim and Asch2017). In parallel, more traditional i-frame solutions, such as providing information (e.g., alcohol labeling), injunctions on products (e.g., “please drink responsibly”), and industry-funded self-help programs (e.g., https://www.drinkaware.co.uk/) have typically yielded disappointing results, as with obesity.
Behavioral researchers have also proposed i-frame inventions to help people reduce their own healthcare costs by optimizing their choice of insurance plan, for example, using calculation aids and defaults (Johnson, Hassin, Baker, Bajger, & Treuer, Reference Johnson, Hassin, Baker, Bajger and Treuer2013). The researchers extrapolated from their promising results that the approach could save Americans $9 billion/year (although scaling up i-frame interventions is often difficult and disappointing; see Kalkstein et al., Reference Kalkstein, De Lima, Brady, Rozek, Johnson and Walton2022). s-Frame differences between the US system and that of comparable countries account cost over $1 trillion/year (Centers for Medicare and Medicaid Services, 2021). Even the best i-frame intervention is no substitute for s-frame reform.
If s-frame changes caused the problem, then reversing those changes is surely the most natural solution. We know from history, however, that such s-frame reform meets concerted opposition. Following Olson's logic of collective action, the concentrated interests of the healthcare sector trump the diffuse benefits that system reform could give individuals. As President Obama (Reference Obama2020) wrote on the challenges of even modest reform:
Unlike the insurance companies or Big Pharma, whose shareholders expected them to be on guard against any change that might cost them a dime, most of the potential beneficiaries of reform – the waitress, the family farmer, the independent contractor, the cancer survivor – didn't have gaggles of well-paid and experienced lobbyists roaming the halls of Congress.
Health care is poor value-for-money in the United States because there has not been the political consensus to drive through s-frame reforms. Without that consensus, the power of special interests to dilute and derail change is considerable: The United States' major recent attempt at reform, the Affordable Care Act, left most problems facing US healthcare intact, as President Obama implicitly acknowledges above.Footnote 45
The world provides a number of successful healthcare systems with better services and far lower costs than in the United States. The key to lowering healthcare costs is to move decisively to a system proven to work elsewhere.Footnote 46 Insights from the behavioral sciences may thus be best focused primarily on understanding how damaging s-frame policies become embedded, and how to build consensus for s-frame reform, rather than “patching” the problem with new i-frame interventions.
2.5. The broader picture
The pattern we have identified applies more widely. Here, we briefly consider six further areas: educational inequalities, discrimination, privacy, misinformation, addiction to prescription drugs, and gun violence.
2.5.1. Educational inequalities
Across most of the world, although elites obtain high-quality education for their children, educational opportunities for the disadvantaged are often poor (UNESCO, 2020). It is uncontroversial that educational inequality is a systemic phenomenon (Merrow, Reference Merrow2017). As affluent parents send their children to private schools, their interest in maintaining the quality of publicly funded schools declines, hurting the quality of public schools (Scott & Holme, Reference Scott and Holme2016). In consequence, people at decreasing levels of affluence find it worthwhile to make the financial sacrifice to send their children to private schools, creating a pernicious “tipping” effect (Darling-Hammond, Reference Darling-Hammond2017). Even within the publicly funded school system, similar feedback loops occur between school catchment areas and property prices, which can rapidly divide localities into affluent communities with “good” schools and less affluent communities whose children are consigned to “bad” schools. In the United States, the divide is exacerbated because education is funded by local property taxes (EdBuild, 2019). Inequalities in education can substantially be reduced with the right systems in place: Most Scandinavian countries, for example, have well-funded universal education with no significant private educational sector (Abrams, Reference Abrams2016).Footnote 47
How have behavioral scientists contributed to the debate? Much of our work has focused, not on changing the system, but on helping individual students: Shifting students' attributions for outcomes from a “fixed mindset” to a “growth mindset” (Dweck, Reference Dweck2008; Hochanadel & Finamore, Reference Hochanadel and Finamore2015), instilling “grit” (Duckworth, Reference Duckworth2016), and reducing “stereotype threat” (Steele, Reference Steele1998). Much of the research along these lines has hinted, or even explicitly proposed, that these interventions can counteract the impact of low-quality education.Footnote 48 Here, as in the many other cases we discuss, there is the real danger that well-intentioned research providing a false hope of radical change from i-level interventions will undermine public pressure for fundamental systemic change.
2.5.2. Discrimination
Poor and unequal education is, of course, closely linked with race- and class-based discrimination, not just in education, but in housing, nutrition, criminal justice, economic opportunities, and beyond. These are highly entrenched systemic problems that warrant far-reaching s-frame reforms. With its embrace of diversity, equity, and inclusion as top goals for institutions, academia, perhaps more than any other profession, have taken the problem to heart. Yet the interventions that are proposed and embraced – mainly dealing with individual-level solutions such as measuring, acknowledging, and combatting “implicit bias” (Banaji & Greenwald, Reference Banaji and Greenwald2016), are, we suspect, likely to make only a small dent in the problem (Dobbin & Kalev, Reference Dobbin and Kalev2018). We believe it is crucial that these policies reinforce, rather than distracting from, the case for deep systemic changes, including a massive reallocation of resources and opportunities.
2.5.3. Privacy
The rapid transition to the digital age has seen rules for maintaining privacy lag far behind technological and commercial innovations that undermine privacy. Currently, even with the protections put in place by the EU's GDPR,Footnote 49 privacy is unachievable for anyone who owns a smart phone, shops at supermarkets, drives a car, or browses the web. We each leave a digital trail that is all too easy for companies, governments, or malign individuals to track and exploit.
Technology companies promote i-frame solutions while opposing tighter s-frame regulation. The movement toward “notice and consent,” whereby people click a consent button allowing their data to be used, is a paradigm example. Here, a behavioral perspective provides a useful corrective, pointing out that few people read, let alone understand, the lengthy and legalistic policies attached to products, apps, and services (Loewenstein et al., Reference Loewenstein, Sunstein and Golman2014); and in any case, they have little choice but to consent, or be denied access.
As elsewhere, behaviorally inspired i-frame interventions have been proposed (e.g., Acquisti et al., Reference Acquisti, Adjerid, Balebako, Brandimarte, Cranor, Komanduri and Wang2017). A particular puzzle is the “privacy paradox” (Acquisti, Brandimarte, & Loewenstein, Reference Acquisti, Brandimarte and Loewenstein2015; Barnes, Reference Barnes2006): People claim to care about privacy, yet readily reveal private information when on-line. Merely identifying the puzzle seems implicitly to blame individuals for their carelessness. But achieving digital privacy is not within the power of individuals, however motivated they might be. As elsewhere, s-frame regulation, rather than individual-level prompts, is crucial (Acquisti et al., Reference Acquisti, Brandimarte and Loewenstein2015).
2.5.4. Misinformation
In today's politically polarized atmosphere, the problem of misinformation is especially pressing. Rational public debate requires agreement on the facts. But in many countries, and especially the United States, there are powerful interests actively promoting conspiracy theories and “alternative facts,” sowing confusion and uncertainty among the general public. Again, regulation lags far behind technological and social change.
Behavioral science provides a powerful i-frame analysis of why people are so vulnerable to misinformation – and should be taken to imply that protecting against these vulnerabilities requires s-level interventions. People are excessively credulous (Gilbert, Tafarodi, & Malone, Reference Gilbert, Tafarodi and Malone1993), strongly underestimate the power of conflicts of interests (Dana & Loewenstein, Reference Dana and Loewenstein2003), and are influenced by the many nonepistemic benefits of new information: Reducing cognitive dissonance, shoring up personal beliefs systems, creating or cementing identification with “like-minded” others, providing ammunition in hypothetical or real debates, and many more (Chater & Loewenstein, Reference Chater and Loewenstein2016; Wojtowicz, Chater, & Loewenstein, Reference Wojtowicz, Chater, Loewenstein, Cogliati-Dezza, Schulz and Wu2022).
There have been wide-ranging academic discussions on how to tackle misinformation (Zucker, Reference Zucker2020), but a major focus of behavioral science has been on i-frame interventions, such as training individuals to detect fake news (van der Linden, Roozenbeek, & Compton, Reference van der Linden, Roozenbeek and Compton2020). One representative study on misinformation about climate change (van der Linden, Leiserowitz, Rosenthal, & Maibach, Reference van der Linden, Leiserowitz, Rosenthal and Maibach2017), for example, forewarned participants that some political actors try to mislead people on the issue, and gave facts and arguments to refute the misinformation before they encountered it. This “inoculation” had some of its intended effect, although one might wonder about the scalability of such an approach given the huge quantity and diversity of misinformation. Likewise, Pennycook, McPhetres, Zhang, Lu, and Rand (Reference Pennycook, McPhetres, Zhang, Lu and Rand2020) showed a powerful impact on truth-discernment and information forwarding of either asking research participants to judge the accuracy of a piece of information or reminding them that information might be inaccurate. Disappointingly, this finding barely replicated (Roozenbeek, Freeman, & van der Linden, Reference Roozenbeek, Freeman and van der Linden2021) and quickly disappeared, and again seems difficult to scale. Yet another approach involved having individuals play a game – Bad News Footnote 50 – in which they seek to distinguish between real and fake news (Basol, Roozenbeek, & van der Linden, Reference Basol, Roozenbeek and van der Linden2020). With a very large sample, rates of correct detection increased slightly, although even this small effect is difficult to evaluate, as the study lacked a control group.
The problem of misinformation is urgent. If behavioral scientists could find an effective i-level remedy in advance of systemic reforms, this would be hugely ideal. We worry, however, that the “promise” of i-level solutions (which, we suspect, will continue to disappoint) will reduce the perceived need for s-level change, which would surely entail the dramatic tightening of regulation of social media. The negative consequences of “information pollution” are, after all, potentially even more damaging to society than chemical pollution, by destabilizing the common base of facts that must underpin any well-functioning democracy.
2.5.5. Addiction to prescription drugs
There has been a wide coverage of the corporate malpractice and government complicity that created a wave of addiction and overdoses that currently kills more than 100,000 Americans each year. Purdue, the company most notorious in fueling the disaster, heavily funded academic studies promoting the idea that pain was under-treated and that opioids provide the best treatment. Purdue-funded academics baselessly claimed that only 1% of patients put on opioids become addicted, and even promoted the bizarre concept of “pseudo-addiction,” according to which people who appeared to be suffering from withdrawal were actually suffering from under-treatment (Greene & Chambers, Reference Greene and Chambers2015).
Crucially for the present argument, Purdue consistently promoted an i-frame perspective on the problem it had helped create, portray its addict victims as weak-willed, irresponsible, individuals. Purdue's Richard Sackler, for example, wrote in an email detailing his company's proposed legal and PR defense, “We have to hammer on the abusers in every way possible. They are the culprits and the problem. They are reckless criminals” (emphasis added).Footnote 51 Highlighting the i-frame puts the focus of federal and state policy makers on law enforcement, targeting the illegal use of opiates, but not restricting medical prescriptions – the s-frame intervention that could have had a decisive impact. Moreover, framing addiction as a crime not a disease led addicts to hide their addiction from doctors and others who could potentially help, and compounded the misery of the addicts by adding self-blame to the other devastating consequences of addiction.Footnote 52
While advancing the i-frame perspective to the media and government, Purdue relentlessly lobbied against s-frame regulation to limit opioid prescribing. Just how powerful s-frame actions could have been is indicated by international comparisons. For example, Germany, the country second to the United States in opioid prescriptions (and hence a conservative point of comparison), resisted efforts by Purdue to foist opioids on patients, and managed to largely avoid the addiction epidemic and rash of overdoses experienced in the United States.Footnote 53
2.5.6. Gun violence in the United States
Why are there so many more mass shootings, and gun-related murders and suicides in the United States than in other developed nations? The consensus in criminology is that systemic factors are decisive: The availability of cheap and powerful firearms is a distinctive feature of the United States. Many nations have imposed strict s-level regulations on weapons, rules on gun ownership, and on locking guns safely. Such s-level interventions have generally proven remarkably successful. For example, increasing restrictions on firearms in the United Kingdom has led to steady declines in gun-deaths by homicide, to around 30 per year in England and Wales in 2020.Footnote 54 By contrast, in the United States, figure is over 50 per day. There have been only two mass shootings in Great Britain in more than 20 years, whereas mass shootings in the United States (in which at least four people are killed) occur almost daily.Footnote 55 The National Rifle Association (NRA) has fought every attempt at regulation, adopting the ubiquitous catch-phrase “guns don't kill people; people kill people,” succinctly encapsulating the i-frame perspective.
Behavioral scientists have at times pursued i-frame policies to combat gun violence. In New York City, the behavioral insights firm ideas42 (founded by Harvard behavioral economists) was asked by the city to conduct a campaign to “discourage would-be shooters from carrying guns” (Gardiner, Reference Gardiner2017). Researchers at the University of Chicago Crime Lab point to field experiments showing that interventions to promote cognitive behavioral therapy techniques among male youths reduce violent crime arrests (Heller et al., Reference Heller, Shah, Guryan, Ludwig, Mullainathan and Pollack2017). Although these types of interventions might prove useful in the unlikely event that they could be rolled out to a broader population, there is a risk that the promise of such approach could nudge policy makers away from the s-frame actions so urgently required to address the structural roots of gun violence.Footnote 56
2.6. A success story: Tobacco
Perhaps the best evidence that corporate interests can be overcome and problems (largely) solved via s-level reforms comes from the long but ultimately successful battle against cigarettes in many countries. In the United States, government interventions played a key role in decreasing the smoking rate, from around 50% in the mid-1900s to below 15% today.Footnote 57 A range of different factors turned the tide of public opinion, including Surgeon General Luther Terry's 1964 report definitively linking cigarettes and cancer, and, later, the movement opposing second-hand cigarette smoke, ultimately resulting in legislation and regulation, against tobacco.
Despite concerted and well-funded opposition from the tobacco industry, s-frame reforms, starting shortly after the 1964 report, collectively contributed to the decline in cigarette sales and smoking (Cole & Fiore, Reference Cole and Fiore2014). These include large cigarette excise tax increases, clean indoor air laws, efforts to prevent adolescents from purchasing tobacco, more dramatic labeling of cigarette packing, and the pressure and consequences of litigation against the tobacco industry by private individuals, the states, and the US Department of Justice (DOJ). The success of these efforts shows both that individual initiatives can, under the right conditions, overcome corporate resistance, and that s-frame policies can address entrenched problems. Although some policies (e.g., labeling) have more of an i-frame flavor, others (taxes and clean indoor air laws) are squarely s-frame; and the far-reaching nature of the policies taken as a whole is unambiguously s-frame in character.Footnote 58
3. Toward an s-frame behavioral public policy
We have argued that i-frame interventions won't provide cheap and effective solutions to pressing social problems. In retrospect, perhaps this should have been obvious, as the message has been conveyed, repeatedly, by colleagues in political science, law, and sociology.
Our faith in i-frame interventions came from attributing diverse societal problems to frailties in individual behavior. But the history of culture, technology, law, science, technology, and politics is not merely one of human potential continually undermined by human folly (though there is plenty of folly). It is also a story of how humans can flourish despite our physical and cognitive weaknesses, by reshaping the rules and systems by which we live. The invention of language, writing, diagrams, maps, and notations of all kinds allows us to share and store our ideas, overcoming the limitations of our memories. Religious, moral, and judicial systems keep selfishness in check. The division of labor helps overcome individual limitations in knowledge and skill acquisition, and radically increases efficiency. Legal and political institutions help us coordinate our actions, determine the allocation of power and property, and save us from Hobbes's “war of all against all.” The adversarial processes of the courts, political debates, and scientific exchange mitigate confirmation bias and related effects. And these institutions are entwined with the invention of money, joint stock companies, taxes, governments, the market economy, international organizations and agreements, and the logistical, informational, and financial architecture underpinning modern economies – allowing us, collectively, to achieve, understand, and produce far more than we could operating as lone individuals. In short, the history of humanity is one of astonishing s-frame innovation (Hayek, Reference Hayek1945; Ostrom, Reference Ostrom1990; Polanyi, Reference Polanyi1941; Sugden, Reference Sugden1989). This innovation has occurred despite our cognitive failings, and, in fact, in remediation of them.
Given that human society and its decision makers have historically demonstrated an extraordinary ability to create rules, systems, and institutions to solve social problems, why do the urgent challenges discussed here remain unaddressed? The answer is not, we believe, that these problems are particularly intractable. For most of the problems discussed here, tried-and-tested s-frame solutions are available, many of which are currently successfully implemented in some parts of the world. Nor is the problem any lack of will, attention, long-term focus, or deficiency in moral fiber. Rather, these problems remain unresolved primarily because of the active and coordinated efforts to block s-frame reform by concentrated commercial interests who benefit from the status quo (see Mayer, Reference Mayer2017), and who seek to maintain it in part by promoting the perspective that these problems are solvable by, and the responsibility of, individuals.
This pattern of opposition to change is another constant of human history. s-Frame, and indeed technological, innovations have been continually and actively opposed by powerful interests that benefit from the status quo, and whether such opposition succeeds has dramatic consequences for mass prosperity and well-being (Acemoglu & Robinson, Reference Acemoglu and Robinson2012). It has been argued that the same pattern arises regarding corruption, dictatorships, and even civil wars (Collier & Hoeffler, Reference Collier and Hoeffler2004). Deep and persistent problems arise not because individual humans are not sufficiently ingenious, far-sighted, or unselfish enough to solve them; but because powerful groups benefit from, and defend, the status quo, whatever the consequences for the population at large.
Looking back, we realize that we, and many of our colleagues, had excessive faith that a specific and quite narrow subfield of research on individual judgment and decision making could substantially help address some of society's most pressing problems. By understanding present-bias, loss-aversion, and judgment biases such as over-confidence, we thought it might be possible to redesign the decision-making environment – the “choice architecture” – perhaps in quite subtle ways, that would help nudge the individual “players” in society to make better choices for themselves and society at large. But the real problem lies not in human fallibility, but in institutions, laws, and regulations that render such fallibility largely irrelevant. In short, we had mistaken deep systemic problems of political economy and conflicts of interest, for problems of individual human folly and responsibility.
But individual-level research remains crucial to informing s-frame policy. Systems operate through their impact on individuals, and their design, operation, and impact depend crucially on human psychology. There are long traditions of applied work in health and educational psychology, clinical psychology, political psychology, and criminology, as well as basic findings in the behavioral and brain sciences, that are directly relevant to the design, implementation, and testing of s-frame public policy. Here we illustrate this relevance by considering three key issues: Seeing the problem, increasing public support for effective s-frame policies, and policy design.
3.1. Seeing the problem
The role of human psychology is important, first, for understanding when and why people perceive the existence of a problem that warrants attention. If people are unaware of (or doubt the reality of) climate change, the rising epidemic of obesity, or the crisis in retirement savings, they are unlikely to seek out or support policy solutions (Weber, Reference Weber2006).
Unfortunately, our minds and brains are not well-adapted for identifying and reacting to long-term systemic problems, however severe. Natural selection operates primarily at the level of individual, and most human evolution occurred in radically simpler times, when most of what mattered for survival and reproduction was in our local environment and occurring in the immediate present. This is especially true of our evolutionarily older emotion system, which evolved to help us deal with immediate threats, such as falling from heights, attacks from predators (Gray, Reference Gray1987), and problematic social interactions involving norm violations or uncooperativeness (Frank, Reference Frank1988). Our emotion system is ill-adapted to responding to slowly evolving, complex, large-scale social problems.
Our emotion system is adaptive. If an adverse situation persists over time, or worsens gradually, our emotional reactions diminish (Frederick & Loewenstein, Reference Frederick, Loewenstein, Kahneman, Diener and Schwarz1999). Emotions evolved to motivate action. But when we fail to act, or action brings no immediate result, it is taken as a sign that maintaining emotions serves no function. Our emotion system is, therefore, not well designed to motivate action against most of the problems discussed in this paper, such as climate change, obesity, and gun violence, that have gradually climbed to levels which, if we experienced them abruptly, would horrify us. As Dubos (Reference Dubos1965) wrote prophetically in Man Adapting, “This very adaptability enables [us] to become adjusted to conditions and habits which will eventually destroy the values most characteristic of human life.”
Our emotion system is also largely oriented to the present, which is a major cause of present-bias (McClure et al., Reference McClure, Laibson, Loewenstein and Cohen2004). In part because our emotion system is so much more responsive to immediate than delayed outcomes, we fail to clamor for solutions to problems that threaten us in the future. Governments may be in an even worse position than individuals, trapped in a short-term election cycle or concerned about imminent unrest.
Finally, the most effective way to diminish negative emotional reactions to perceived threats is often not to tackle the threats themselves, but to ignore them or persuade ourselves that they don't exist – a major theme in the literature on “fear appeals” (e.g., Leventhal, Reference Leventhal1970; Witte & Allen, Reference Witte and Allen2000). As Marshall (Reference Marshall2015, p. 228) writes in Don't Even Think About It: Why Our Brains Are Wired to Ignore Climate Change, “The bottom line is that we do not accept climate change because we wish to avoid the anxiety it generates and the deep changes it requires.”
Emotions are also oriented to the vivid and the tangible, and to narratives, rather than to facts and statistics. Constantino and Weber (Reference Constantino and Weber2021) insightfully argue that narratives
play a vital role in shaping environmental publics, policy and politics. They can be strategically crafted and disseminated, or they can emerge, be reinforced or revised through social relations. To the extent that those with vested interests in the existing system also have power over information flows, uncertainty may create the conditions for the intentional manufacturing of narratives that reproduce existing power relations and serve those interests, including discourses of denial, uncertainty and delay. (p. 152)
Constantino and Weber review evidence that narratives have played a key role in forestalling action on climate change (Bushell, Buisson, Workman, & Colley, Reference Bushell, Buisson, Workman and Colley2017; Lamb et al., Reference Lamb, Mattioli, Levi, Roberts, Capstick, Creutzig and Steinberger2020), and also have the potential to motivate successful reform (Hinkel, Mangalagiu, Bisaro, & Tàbara, Reference Hinkel, Mangalagiu, Bisaro and Tàbara2020).
Our emotional reactions are often remarkably disconnected from factors that are most important for survival and well-being. We cry in movies about fictional characters, but not when we read about calamities in the newspaper. We are outraged by someone jumping line at a restaurant, but unperturbed by extreme wealth inequality. We are swayed more by stories than statistics (Johnson, Bilovich, & Tuckett, Reference Johnson, Bilovich and Tuckett2023). Again, this lack of proportionality makes us vulnerable to manipulation. Powerful interests are often perfectly aware of these features of human emotions, and actively exploit them. We can be manipulated into risking our life in war, or into committing atrocities, by primal appeals to identity, including nationalism; and we can be distracted from crucial policy challenges by the emotional appeal of “hot-button” issues (Lobel & Loewenstein, Reference Lobel and Loewenstein2005).
3.2. Increasing public support for effective s-frame policies
Human psychology critically determines which policies people support – and in a democratic system (or an authoritarian one in which rulers need to maintain popularity) public support can have a powerful influence on policy. Applying behavioral science to this issue is therefore an important development (e.g., Goldberg, Gustafson, Ballew, Rosenthal, & Leiserowitz, Reference Goldberg, Gustafson, Ballew, Rosenthal and Leiserowitz2021; Rinscheid, Pianta, & Weber, Reference Rinscheid, Pianta and Weber2021; Sherman, Shteyn, Han, & Van Boven, Reference Sherman, Shteyn, Han and Van Boven2021).
Consider emotional adaptation, discussed above, which crucially shapes reactions to beneficial s-frame policies, both before and after implementation. People systematically underestimate how much they will adapt (Mazar, Tomaino, Carmon, & Wood, Reference Mazar, Tomaino, Carmon and Wood2021; Riis et al., Reference Riis, Loewenstein, Baron, Jepson, Fagerlin and Ubel2005; Ubel, Loewenstein, & Jepson, Reference Ubel, Loewenstein and Jepson2005). This provides a powerful brake on the public appetite for systemic change, and a tendency to want to maintain the status quo (Samuelson & Zeckhauser, Reference Samuelson and Zeckhauser1988; loss aversion exacerbates this problem, Tversky & Kahneman, Reference Tversky and Kahneman1991). Those opposing the s-frame reforms needed to shift world economy to net zero carbon emissions, or to reform pensions, health care, or the redistribution of wealth, have found that threats to the status quo (e.g., to the “American way of life”) are highly effective tools in resisting reform.
Yet once an effective s-frame policy is implemented, people often adapt surprisingly quickly. Janusch, Kroll, Goemans, Cherry, and Kallbekken (Reference Janusch, Kroll, Goemans, Cherry and Kallbekken2021), for example, examined individuals' acceptance of a “congestion charge” before and after its implementation in a six-player-two-route congestion game. Although the charge curbed congestion effectively, people often vote against it initially. But when the positive effects of the charge were experienced, many embrace an s-level reform they had previously resisted.
Policy makers, too, may significantly underestimate how rapidly people can adapt to new circumstances and how quickly social norms can change (e.g., initial resistance to masks rapidly reversed in many countries during the COVID-19 pandemic; Denworth, Reference Denworth2020). Indeed, people consistently underestimate how much of their behavior is driven by habits (Mazar & Wood, Reference Mazar and Wood2022) and social norms (Cialdini, Reference Cialdini2005) rather than preferences – and hence overestimate how much they will dislike a shift to new patterns of behavior.
Moreover, adaptation can lead us into futile “arms races,” in which competition expends resources to no-one's overall benefit (Frank, Reference Frank1985, Reference Frank2005; Hirsch, Reference Hirsch1976). Frank argues that goods can be divided into those that increase human welfare directly (e.g., freedom from pain) and “positional” goods that are valued partly because we have them, and others do not (obtaining a place in a prestigious college, winning a race, or holding high political office).
Frank argues that a larger than optimal fraction of consumer spending is devoted to what are primarily positional goods (e.g., large houses, fast cars, and “luxuries” of all kinds), creating an arms race that funnels human activity and economic resources to activities that leave people, in aggregate, no better off. Frank's argument is bolstered by neural and behavioral evidence that the physiology and psychophysics of the senses are inherently comparative (Laming, Reference Laming1997), with only crude judgments of absolute level magnitudes such as loudness or brightness (e.g., Stewart, Brown, & Chater, Reference Stewart, Brown and Chater2005). Similarly, reward value is coded relatively in at least some neural systems (e.g., Tremblay & Schultz, Reference Tremblay and Schultz1999), and behavioral experiments (e.g., Ariely, Loewenstein, & Prelec, Reference Ariely, Loewenstein and Prelec2003; Vlaev, Seymour, Dolan, & Chater, Reference Vlaev, Seymour, Dolan and Chater2009) as well as research on happiness (Boyce, Brown, & Moore, Reference Boyce, Brown and Moore2010; Clark, Frijters, & Shields, Reference Clark, Frijters and Shields2008) tells a similar story. The accumulation of money (rather than leisure, time with family, and so on) may itself generate an arms race leading to a large loss of human welfare. The challenge of diffusing such arms races (e.g., by s-frame measures such as taxation and redistribution; Frank, Reference Frank2005), is therefore crucial, though not straightforward.
Psychological insights can also provide direct guidance for designing policies that will garner popular support. For example, banning single-use plastic bags might be perceived as intruding on individual rights. But charging consumers a token amount for using single-use plastic bags is remarkably effective in reducing their use (Homonoff, Reference Homonoff2018).
These same “implementational” questions arise when considering how to implement a carbon tax. Psychological insights, and research using psychologically informed research methods, can contribute tremendously to design decisions regarding whether a carbon tax should be imposed upstream (e.g., on miners, drillers, manufacturers, or retailers) or downstream (on consumers), if such a tax should be integrated with the price of the product or segregated (Chetty, Looney, & Kroft, Reference Chetty, Looney and Kroft2009), and, crucially, how tax revenues should be returned to the public. Moreover, some of the same psychological forces that undermine calls for immediate climate action can also help make interventions more palatable (see Loewenstein & Schwartz, Reference Loewenstein and Schwartz2010; Schwartz & Loewenstein, Reference Schwartz and Loewenstein2017). If people discount the future and ignore small changes, then it may be appropriate to use capital markets to generate the dividend from future carbon tax revenues in an up-front lump sum. Indeed, a “people's payout” model, in which carbon tax revenues are largely or entirely redistributed, rather than supporting government spending, has gathered enough support to be implemented in many provinces in Canada (Nuccitelli, Reference Nuccitelli2018). Behavioral research on these questions will be crucial in making carbon taxes publicly acceptable (Carattini, Kallbekken, & Orlov, Reference Carattini, Kallbekken and Orlov2019; Kallbekken, Kroll, & Cherry, Reference Kallbekken, Kroll and Cherry2011).
3.3. Improving policy design
The behavioral and brain sciences can also provide i-frame insights that inform better s-frame policies. “Behavioral insights” have been at the heart of the i-frame interventions defining the nudge movement. But individual-level psychology is equally important in designing effective s-frame interventions. Table 2 illustrates the many ways in which the behavioral and brain sciences can inform public policy. We briefly discuss a selection of these issues here.
Table 2. Many roles of the behavioral and brain sciences in policy design and implementation
Uncontroversially, s-frame policies should be as “ergonomic” as possible, and they frequently fail badly in this regard. For example, claiming tax credits or benefits often involves navigating a baffling bureaucratic process, excluding many of the people they are intended to benefit (Goldin, Reference Goldin2018); financial, medical, environmental, or nutritional information is often uninterpretable to consumers and does not improve their choices (Loewenstein et al., Reference Loewenstein, Sunstein and Golman2014); processes by which the public express their preferences (e.g., regarding preferred school options for their children) can be mystifying (Johnson, Reference Johnson2022); and information disclosure (e.g., about restaurant hygiene standards) is often optional, while consumers often fail to appreciate the significance of omitted information (Gurney & Loewenstein, Reference Gurney and Loewenstein2020; Sah & Read, Reference Sah and Read2020).
A valuable lesson from the behavioral insights movement has been that ergonomics matters just as much for government policies as for the personal computer (PC) or smart phone (see, e.g., Norman, Reference Norman1988; Thaler & Sunstein, Reference Thaler and Sunstein2008, Reference Thaler and Sunstein2021). Designing policy around the consumer can frequently make the difference between success and failure, and policy design should be guided primarily by behavioral insights. Policy, like any complex good or service, is best designed by multidisciplinary teams, with subject experts, designers, user-experience specialists, ethnographers, anthropologists, and psychologists, alongside behavioral insights specialists.
3.3.1. Improving the policy-making process
Another crucial target for the behavioral sciences is improving how policy is made. To optimize the process of policy development (including influences of lobbying and even corruption), scrutiny and consultation (with external bodies and other parts of government), and legal and political “sign-off” (see, e.g., Sunstein, Reference Sunstein2022b), systemic factors will, again, be crucial. But individual psychology remains important: Do policy makers effectively prioritize the most impactful policies (Toma & Bell, Reference Toma and Bell2022)? Are they overconfident, both individually and in potentially self-reinforcing group discussions? Is there suspicion of ideas borrowed from other countries or contexts that are “not invented here” (e.g., Katz & Allen, Reference Katz and Allen1982), which may impede policy development?
Here the interplay between s- and i-frame analyses is particularly intricate (e.g., Mercier & Landemore, Reference Mercier and Landemore2012). The process by which diverse opinions and interests are combined provides checks and balances against psychological quirks. Open public scrutiny, or the consultation with a range of interests, may reduce the tendency to “lock in” to particular viewpoints (Chater, Reference Chater2020b), by uncovering counterarguments and evidence (e.g., Callon, Lascousmes, & Barthe, Reference Callon, Lascousmes and Barthe2009). Conversely, policy-making environments with “like-minded” people, or where there is pressure to be on the “winning side” of any debate (if debate occurs at all), may amplify individual biases, by squashing counterarguments and evidence (Sunstein, Reference Sunstein1999), leading to group polarization (Bray & Noble, Reference Bray and Noble1978), pluralistic ignorance (Miller & McFarland, Reference Miller, McFarland, Suls and Wills1991), and group think (Janis, Reference Janis1972). How to make group interaction improve, rather than impede, policy design is a key topic for further investigation.
3.3.2. Understanding and reversing industry exploitation of human psychology
Industry often exploits consumer psychology for its own ends.Footnote 59 We have already discussed the food industry's search for “bliss points” for ultraprocessed foods (Moss, Reference Moss2013). Just as understanding the psychology and physiology of appetite and eating helps industry identify such products (typically nutritionally poor and energy-dense), so that same understanding can shape s-frame regulation to protect consumers. Slot machines (and other gambling products and services) are deliberately designed to maximize the tendency to keep gambling and the desire to return (Schüll, Reference Schüll2012) – capitalizing on human desires for immediate “hits,” loss-chasing, present-bias, and so on. Arguably entire industries, including alcohol, cigarettes, gambling, and pay-day lending, are partially dependent on “hooking” consumers. Similarly, “click-bait,” “fake news,” and the propagating of extreme opinions by social media algorithms are all designed to keep our collective eyeballs on our screens; day-trading platforms encourage unsophisticated investors to “burn” their money by overtrading, and so on. Here, too, effective regulation requires psychological insight into when and how people can be exploited to their detriment.
3.3.3. s-Frame changes that improve i-frame decision making: Helping individuals make better choices
Improving individual decision making has been the focus of i-frame behavioral insights. But often the most powerful way to help people make better decision is not merely to modify their “choice architecture,” but to fundamentally change the “rules of the game.” Thus, eliminating conflicts of interest between professionals and their clients (e.g., in medicine or finance) is likely to be more effective than requiring disclosure (Cain, Loewenstein, & Moore, Reference Cain, Loewenstein and Moore2005, Reference Cain, Loewenstein and Moore2011; Kanter & Loewenstein, Reference Kanter and Loewenstein2019; Larkin et al., Reference Larkin, Ang, Steinhart, Chao, Patterson, Sah and Loewenstein2017), or educating consumers to detect potential conflicts. Similarly, removing conflicts between operational and safety considerations (e.g., by separate chains of command, and being bound by agreed protocols) is typically the priority in safety critical contexts (e.g., airlines, medicine), rather than helping individuals manage these conflicts in the moment.
s- And i-frame approaches can still often be mutually reinforcing. For example, i-frame measures, such as health warnings on cigarette packets or antismoking public information campaigns, may increase public support for s-frame measures including advertising bans, and outlawing smoking in public places (Sunstein, Reference Sunstein2022a). Similarly, standardized procedures, such as checklists in aviation and medicine (e.g., Gawande, Reference Gawande2009), may enhance s-frame processes for scrutinizing performance (e.g., adherence to procedures is more easily monitored).
Finally, in a democracy, key individual decisions citizens make is through voting – and a crucial systemic challenge is to maximize turnout. Recent work (Mazar, Tomaino, Carmon, & Wood, Reference Mazar, Tomaino, Carmon and Wood2022) has revealed that people dramatically underestimate the impact of “frictional” factors (e.g., long-distances to travel) on voter turnout. Citizens who are particularly prone to this bias are especially supportive of measures that would increase frictional effects. An electoral system based on good understanding of the determinants of individual behavior may be crucial for maintaining a healthy democracy.
3.3.4. Avoiding psychologically naïve policy prescriptions
A central topic is behavioral economics is the impact of incentives on behavior (Gneezy, Meier, & Rey-Biel, Reference Gneezy, Meier and Rey-Biel2011). Everyday psychological intuitions, rational choice models, and reinforcement learning theories in psychology (Skinner, Reference Skinner1938), neuroscience (Schultz, Dayan, & Montague, Reference Schultz, Dayan and Montague1997), and machine learning (Sutton & Barto, Reference Sutton and Barto2018) emphasize the power of incentives. But, although carrots and sticks matter, an overly simple view of human psychology as maximizing utility may lead to incomplete policy prescriptions.
One weakness of the rational, maximizing perspective is that it underplays the importance of perceived autonomy, fairness, wider ethical considerations (Rai & Fiske, Reference Rai and Fiske2011), and the “logic of appropriateness” (March & Olsen, Reference March, Olsen, Goodin, Moran and Rein2008) that guides so much human behavior (i.e., doing what we are believe we are supposed to do). Thus, a criminal justice system based on deterrence, reliance on share-options to incentivize management, or attempts to pay for prosocial behavior and fine antisocial behavior, may need to be reconsidered. As ever, direct evidence from comparison across nations, organizations, and real-world incentive systems will likely play a dominant role in evidencing any s-frame policy changes (e.g., Jeppson, Smith, & Stone, Reference Jeppson, Smith and Stone2009; Nagin & Pepper, Reference Nagin and Pepper2012).
A second complication with a purely incentive-based approach to policy is public “reactance” against incentives for policies which citizens see as ineffective, unfair, or infringing liberty (Taylor & Asmundson, Reference Taylor and Asmundson2021). Such reactance need not be grounded in justifiable concerns or firm evidence, but also in baseless conspiracy theories (e.g., COVID is a hoax, COVID vaccinations lead to sterility, etc., e.g., Imhoff & Lamberty, Reference Imhoff and Lamberty2020). In such circumstances, incentives may be counterproductive, by increasing suspicion of government motives.
3.4 Wider issues for the role of behavioral science
3.4.1. Implications for research methods
In a policy-making regime emphasizing s- over i-frame reforms, there will be an expanded role of the social and behavioral sciences in predicting the likely consequences of alternative s-frame reforms (see, e.g., Janusch et al., Reference Janusch, Kroll, Goemans, Cherry and Kallbekken2021). We discussed above how emphasis on the “gold standard” of field experimentation may reinforce to the focus on i-frame policies. Yet quasi-experimental studies can often substitute for experiments, when different countries, states, and other entities implement specific reforms at different times. Similarly, regression discontinuity designs are informative when a policy measure is applied based on an abrupt qualifying threshold (e.g., income, age, or test scores). Pioneered by psychologists in the 1960s (e.g., Campbell & Ross, Reference Campbell and Ross1968; Campbell & Stanley, Reference Campbell, Stanley and Gage1963), these and related approaches have been refined by economists. These developments provide valuable tools for rigorously evaluating s-frame policies.
3.4.2. Building the information environment for debate about s-frame reform
Meaningful debate over s-frame reform, in whatever domain, requires meeting key preconditions for constructive discussion, and opponents of s-frame reform will often work hard to undermine these preconditions. Three factors appear particularly crucial to stymying agreement: Lack of a sufficient common ground on the relevant “facts”; excessive polarization, such that any issue becomes a proxy for all others (and perhaps for social identity); and lack of trust in the good faith of the “other side.” Opponents of s-frame reform engage in disinformation (e.g., big tobacco on the dangers of smoking; the fossil-fuel industry's attempt to undermine climate science; Oreskes & Conway, Reference Oreskes and Conway2011), thus undermining a common ground of facts from which consensus might be reached. Another common tactic is to align policy problems with existing polarized debates (e.g., painting climate activists or healthcare campaigners as anti-capitalist or anti-freedom). Personal attacks (e.g., on climate scientistsFootnote 60 or public health experts), further undermine trust in the good faith of those with whom we differ; and acrimonious social media interactions are often sufficient to block reasoned debate.
Improving the information environment for public debate, and countering active attempts to corrupt it, is a key research topic. Without high-quality public debate based on a shared evidence base, gaining support for systemic change is likely to be very difficult. How can such a goal be furthered? We suspect that substantial s-frame changes are likely to be required, for which consensus will be difficult precisely because that basis has been progressively undermined. It is outside our scope and expertise to identify the most effective s-frame measures: But possibilities include dramatically reducing the concentration of media ownership; imposing rules of impartiality on news providers; treating social media providers as publishers (and hence subject to the laws and regulations that apply to them); associating social media profiles with traceable human identities (addressing both the prevalence of bots, malicious disinformation, and allowing legal redress to defamatory posts); and requiring social media companies to open their algorithms to public scrutiny. Which of these policies will be effective? Which will backfire? Although the specifics of many of these issues are new, some insights can be gleaned by experiences in other countries at different points in time. Insights from social psychology on belief and attitude formation, trust, in-groups and out-groups, social identities, and so on, will also be clearly relevant in predicting which interventions are likely to work.
3.4.3. Where to draw the line on “heavy-handed” paternalism
We advocate a more heavy-handed public policy than that inherent in the nudge approach. But where should the line be drawn on regulation? The behavioral and brain sciences won't answer this question. The public, through normal democratic processes, must balance freedom-to-choose and freedom-from-temptation (or addiction). But behavioral insights should inform this debate, for example, regarding the power of visceral impulses (hunger, thirst, sex, pain, etc.) which can overwhelm a person's attention and drive behaviors that may not align with their long-term interests (Critchley & Harrison, Reference Critchley and Harrison2013; Loewenstein, Reference Loewenstein2006). The physiology and psychology of addiction is particularly crucial (e.g., Elster & Skog, Reference Elster and Skog1999) to distinguish addiction from free consumer choice (Heather & Segal, Reference Heather and Segal2017). Similarly, understanding of individual differences, including psychiatric disorders, will help in clarifying whether some groups of people may be especially vulnerable, and how they can be protected.
Exploitation arises, too, from cognitive rather than motivational vulnerability. Products can be misleading and overly complex; advice can be distorted by conflicts of interest (see Table 2). Drawing the line between acceptable marketing (e.g., legitimately putting goods and services in a good light) and malpractice cannot, again, be resolved by scientific evidence – it is political choice to be made by the electorate and its representatives. Here too, insight from the behavioral and brain sciences should inform such deliberations. For example, if product complexity is too great for people to make stable choices (or assess which products are appropriate for which people or purposes), this “sludge” may bamboozle consumers into make choices against their own interests (Sunstein, Reference Sunstein2020; Thaler, Reference Thaler2018; Thaler & Sunstein, Reference Thaler and Sunstein2021). Similarly, the fact that people largely discount disclosed conflicts of interest (Loewenstein, Sah, & Cain, Reference Loewenstein, Sah and Cain2012) should raise alarm bells for regulators relying on mandatory disclosure (e.g., Loewenstein et al., Reference Loewenstein, Sunstein and Golman2014). If privacy disclosures are demonstrably incomprehensible, they clearly cannot usefully inform choice.
4. Conclusion
Our goal has been to provoke discussion of how behavioral science can best inform public policy. We have argued that our field has been excessively focusing on policy interventions targeting individual behavior, and that (1) many critical public policy challenges arise from problematic systemic policies, which are defended by the commercial interests they benefit; (2) those commercial interests promote the virtues of i-frame solutions, while lobbying against s-frame reform; (3) many behaviorally oriented academics, including ourselves, have inadvertently reinforced the ineffective i-frame perspective; and (4) i-frame interventions yield disappointing results, and more importantly, can reduce support for effective s-frame policies.
We have focused on how behavioral scientists have inadvertently assisted efforts by corporate interests to resist systemic changes, but the idea that corporate interests craft the rules to benefit themselves is hardly original (see, e.g., Acemoglu & Robinson, Reference Acemoglu and Robinson2012). Nor is the idea that commercial interests promote individualistic perspectives to avoid regulation. Giesler and Veresiu (Reference Giesler and Veresiu2014, p. 841) coin the term “responsibilization” to refer to processes “through which responsibility is shifted away from the state and corporations” and toward the “responsible consumer.” Giridharadas (Reference Giridharadas2019) notes that we seem to have collectively
lost faith in the engines of progress that got us where we are today – in the democratic efforts to outlaw slavery, end child labor, limit the workday, keep drugs safe, protect collective bargaining, create public schools, battle the Great Depression, electrify rural America, weave a nation together by road, pursue a Great Society free of poverty, extend civil and political rights to women and African Americans and other minorities, and give our fellow citizens health, security, and dignity in old age.
We see informing s-frame interventions as the future of behavioral public policy. Behavioral scientists' excessive enthusiasm for i-frame policy has reduced the impetus for systemic reform, just as corporations interested in blocking change intend. We have been unwitting accomplices to forces opposed to creating a better society.
We echo Furman's (Reference Furman2016, p. 8) call for “behavioral scientists to look further up in the branches toward higher-hanging and potentially better fruit. That entails starting from the big questions… and then determining what behavioral insights and research, often as complements to more traditional policy tools, are needed to help solve them.”
Acknowledgments
The authors gratefully acknowledge invaluable research and editorial help from Aden Halpern, help on the privacy and education sections from Alessandro Acquisti and Najeeb Shafiq, respectively, and valuable comments from Max Bazerman, Linda Dezso, Angela Duckworth, Simon Dedeo, Emily Ho, Cait Lamberton, and Jules Lobel, Cass Sunstein and Richard Thaler. This paper has benefitted enormously from comments from four anonymous reviewers and from discussions with researchers with a wide spectrum of viewpoints (some, but by no means all, overlapping with our own), including Peter Bayliss, Ed Gardiner, Craig Fox, David Hagmann, Dave Nussbaum, Adam Oliver, Magda Osman, David Weiss, and Ivo Vlaev. The authors also thank Laura Jekel and Rosa Stipanovic for their invaluable help with references and proofreading.
Financial support
Nick Chater was supported by the ESRC Network for Integrated Behavioural Science (grant number ES/K002201/1).
Competing interest
Both authors have served on the academic advisory board of the UK Behavioural Insights Team. N. C. is co-founder and director of Decision Technology (www.dectech.co.uk), a research consultancy that has worked on consumer behavior with retailers, banks, energy companies, the gambling industry, food delivery companies, café chains, telecoms and media companies, and charities. G. L. consults with health insurers Florida Blue, Highmark, and United Healthcare. None of the ideas expressed in this paper are supportive of the interests of these organizations, and in some cases could be viewed as conflicting with them.
Target article
The i-frame and the s-frame: How focusing on individual-level solutions has led behavioral public policy astray
Related commentaries (33)
An inconvenient truth: Difficult problems rarely have easy solutions
Behavioral market design
Behavioral mechanism design
Behavioral public policy in practice: Misconceptions and opportunities
Behavioral winter: Disillusionment with applied behavioral science and a path to spring forward
Community-engaged research is best positioned to catalyze systemic change
Conspiracy theory
Don't throw the individual perspective out while waiting for systemic change
Expectations, opportunities, and awareness: A case for combining i- and s-frame interventions
i-Frame interventions enhance s-frame interventions
Individual-level solutions may support system-level change â if they are internalized as part of one's social identity
It's always both: Changing individuals requires changing systems and changing systems requires changing individuals
Misdiagnosing the problem of why behavioural change interventions fail
Moral psychology biases toward individual, not systemic, representations
Moving from i-frame to s-frame focus in equity, diversity, and inclusion research, practice, and policy
Nudges, regulations, and behavioral public choice
Nudging is being framed
On Skinner's pendulum: A framework for assessing s-frame hope
Optimizing behavior change through integration of individual- and system-level intervention approaches
Real systemic solutions to humanity's problems require a radical reshaping of the global political system
Structural problems require structural solutions
The influence of private interests on research in behavioural public policy: A system-level problem
The psychology and policy of overcoming economic inequality
The real cause of our complicity: The preoccupation with human weakness
The social sciences are increasingly ill-equipped to design system-level reforms
The “hearts-and-minds frame”: Not all i-frame interventions are ineffective, but education-based interventions can be particularly bad
Unpacking the nudge muddle
Use behavioral research to improve the feasibility and effectiveness of system-level policy
Using effective psychological techniques to subvert a US sociopolitical context
When nudges have societal-level impact
Why a group-level analysis is essential for effective public policy: The case for a g-frame
Wise interventions consider the person and the situation together
“More effective” is not necessarily “better”: Some ethical considerations when influencing individual behaviour
Author response
Where next for behavioral public policy?