Chater & Loewenstein (C&L) have produced a powerful manifesto (with a bit of mea culpa thrown in). They claim that behavioral scientists (themselves included) have mistakenly focused on improving individual decision making (the “i-frame”), and in so doing have unwittingly deflected attention from necessary changes in the rules of the game – societal institutions and policies – that shape individual decisions (the “s-frame”). They provide convincing evidence that:

C&L advocate “applying behavioral insights to s-frame reforms” and provide a series of important suggestions in this direction. I'm sorry that C&L did not address how economics – in light of these new behavioral insights – would have to change to contribute to this project (Loewenstein is an economist). This would require a new approach to mechanism design, the field of modern economics that – like Adam Smith, David Hume, and the other classical economists – unabashedly engages in social engineering, by deriving systems of property rights, incentives, prohibitions, and other rules of the game (“mechanisms”) that will support desired societal outcomes.

Over the past three decades, behavioral evidence has challenged two assumptions that were once standard among economists: That people are good at maximizing expected utility and that we are amoral and self-regarding. So (here comes my behavioral economics bumper sticker): People are both dumber and nicer than we thought. My response is a mini manifesto for a science of s-frame reforms without the brainy and self-interested Homo economicus, summarized in three points (I wonder if C&L agree).

First, we have to give up the idea that a single concept – utility – can be used both to predict behavior and to evaluate the results of a policy or set of institutions. Evidence that we are not very good choosers – as Daniel Kahneman showed – is very bad news to economists (Kahneman, Reference Kahneman1994). It means that in order to explain behavior we need to extend “preferences” beyond “tastes” (something worth satisfying) to encompass weakness of will, bad habits, even addictions.

But if we incorporate these self-harming and regret-inducing reasons for behavior into our utility function, utility is no longer something that a benevolent planner or a citizen would like to see maximized. As a result, our workhorse concepts such as Pareto efficiency lose their normative appeal. Does satisfaction of a nicotine-craving count as a plus when we tote up economic benefits and costs? Or a racist desire to interact only with members of one's own ancestral group? When evaluating a successful drug cessation program, do we really want to include as a cost the forgone drug induced high of a recovered formerly addicted person (Chaloupka et al., Reference Chaloupka, Warner, Acemoglu, Gruber, Laux, Max and Sindelar2014)?

Second, we have to give up the idea that whatever utility is, it is not comparable across people. Jeremy Bentham and the other classical economists conceived of what we now term utility as “pleasure and pain” experienced by people in differing and comparable degrees. If we can compare utilities across people, then given Gossen's “law of the saturation of wants” (“diminishing marginal utility” to economists) we can (and often do) conclude that transferring wealth from the very rich to those of limited means will do more good than harm.

The dictum that we cannot (and must not) compare utilities across people was introduced to economics – ostensibly on grounds of parsimony – almost a century ago by Lionel Robbins. It is still widely accepted and taught to students, who find it odd once they realize that it precludes not only the (above) utilitarian argument for redistribution of wealth, but also statements like: “I'll pick up the kids today, dear, it will be less trouble for me than for you.” Denying the coherence of such statements (the offending phrase is in italics) precludes the most elementary sentiments of regard for others that are commonly observed in experiments and essential to a well-ordered society.

Third, we have to give up trying to design institutions and policies that can be expected to work well no matter what people are like. This means abandoning the assumption that everyone is a self-interested “knave” (David Hume's dictum, echoing Machiavelli; Hume, Reference Hume, Green and Grose1898), recognizing that widely shared ethical and other-regarding values are essential to policy implementation, and taking on board the possibility that our favorite policy instruments – incentives and constraints designed for knaves – may sometimes be ineffective, even counterproductive.

Given its currency in legal, economic, and policy-making circles it may seem odd that nobody really believes the assumption that people are amoral and self-interested to be literally true. Instead, the assumption has been advanced on grounds of prudence, not realism. Even Hume, just a sentence after announcing his maxim about citizen-knaves, warned the reader that the maxim was false in fact.

But letting Homo economicus be the behavioral model of the citizen, the employee, the student, or the borrower when it comes to designing laws, policies, and business organizations is anything but prudent. Policies that follow from this paradigm sometimes make the assumption of universal amoral selfishness more nearly true than it might otherwise be. Consistent with the research half a century ago by Edward Deci, Mark Lepper, and their coauthors, recent experimental evidence suggests that incentives and constraints sometimes crowd out other-regarding and ethical preferences (Bowles & Polania-Reyes, Reference Bowles and Polania-Reyes2012; Deci, Reference Deci1971; Lepper, Greene, & Nisbett, Reference Lepper, Greene and Nisbett1973).

In a similar vein C&L warn of “public ‘reactance’ against incentives for policies which citizens see as ineffective, unfair, or infringing liberty” (target article, sect. 3.3.4, para. 3) referring to evidence from the COVID pandemic (on which, see also Schmelz, Reference Schmelz2021). No matter how cleverly designed to harness the “avarice” of knaves (as Hume put it), incentives cannot alone provide the foundations of good governance (Bowles, Reference Bowles2016).

C&L provide a rich starting point (including a gold mine of references) for the formulation of a behavioral mechanism design, whether along the lines proposed above, or in some other direction. Progress in this project may require going beyond the i-frame and s-frame dichotomy (a brilliant rhetorical device for the purposes of C&L's manifesto) to include intermediate arenas of sociality in communities and civil society, in other words a c-frame (Bowles & Carlin, Reference Bowles and Carlin2020).