I have striven not to laugh at human actions, not to weep at them, not to hate them, but to understand them.

Baruch Spinoza, Tractatus Politicus, 1677, ch. 1, sect. 4 (A3011)

1 Introduction

To live is to choose. Along with the big decisions in life, our profession, whom to marry, and how many kids we should have, we engage in more mundane decisions: what to eat, what to wear, and what movie to watch. We are what we choose. Sometimes we are happy with what we picked; other times we regret it and wish we had done it differently. This does not happen with ECONSFootnote 1 (or homo economicus), the theoretical agent that populates neoclassical economics. He is all-knowledgeable, has God-like calculation capacities, and never succumbs to temptation. He does not misbehave (Thaler, 2015a). But you and I are not like ECONS, are we? Nor are behavioral economists, who are like us, and fully reject this description of human behavior. We value the present too much, we get lost in our hopes, fears, and dreams. We procrastinate and we sometimes eat more than we should, just to promise not to do the same anymore. We are HUMANS. Behavioral economics findings were a welcome achievement as they improved our understanding of how real, fallible human beings operate in real life. But something went wrong along the way.

While behavioral economics began as an attempt to understand real human behavior, it took a normative turn, most notably with the “libertarian paternalism” approach popularized by Sunstein and Thaler’s book, Nudge (2008). Contrary to older notions of paternalism, which claimed special wisdom and tried to steer people away from immoral impulses and towards what they saw as “the good life” regardless of what individual preferences were, “libertarian paternalism” claims to make people better off, in terms of their own welfare. Moreover, libertarian paternalism aims to make people choose what they would have chosen “if they had complete information, unlimited cognitive abilities, and no lack of willpower” (Sunstein & Thaler, 2003, p.175). In other words, you’re a HUMAN, but you should be an ECON. And it is against this trend, and not the overall behavioral economics research program, that Mario Rizzo and Glen Whitman’s book, Escaping Paternalism: Rationality, Behavioral Economics, and Public Policy (2019), argue against. For them, behavioral economics performed an incomplete revolution. For the authors, for this revolution to be complete, behavioral economics should also depart from neoclassical economics on their normative view.

The argumentative strategy of the book starts first by questioning the concept of rationality used as a benchmark (homo economicus), what Rizzo and Whitman coined as “Puppet Rationality”,Footnote 2 just to conclude that it is far from evident that this should be the criteria for evaluating our actions and calling for a broader, more context-dependent concept of rationality: “inclusive rationality” (Rizzo & Whitman, 2019, p. 17). Rizzo and Whitman discuss behavioral economics’s findings on cognitive biases and question how these biases might divert us from our true desires. They argue that the evidence for bias is not as clear as portrayed by behavioral paternalists. Moreover, even if bias and mistakes can occur, it is different to recognize their existence in the abstract than to specify a specific instance. In addition, evidence for cognitive biases downplay the overall utility derived from “biased” human choices (assuming they are biased in the first place) in our general behavior. Rizzo and Whitman then adopt an “if-then” approach to address the following questions: If we can identify when individuals act contrary to their true preferences, how can policies be crafted to guide them closer to their genuine desires? How do we implement policies that account for self-debiasing? How do we bear in mind the interactions of bias? Given individual heterogeneity, how do we define an overarching policy capable of accommodating these differences and being just enough to correct the behavior of a particular individual? Such questions have not only philosophical and methodological implications, but also policy-relevant implications. Lastly, Rizzo and Whitman point out to the public choice problems of implementing these measures. Who can guarantee that “the man on the spot” is going to use the insights of behavioral economics for the common good and not for his own sake? How do we make sure incentives are not distorted? This review essay will follow the argumentative exposition of the book.

2 I might be irrational, but I am not a fool

Because the path of our lives is determined by our choices, human beings always wanted to understand how we make choices. It comes as no surprise that economics, for some the science of choice, is particularly interested in understanding human behavior and rationality. As economics became more formal and its reliance on mathematics increased, neoclassical economics brought a figure of human beings to prominence: homo economicus. Richard Thaler and Cass Sunstein, prominent behavioral economists, have initiated a substantial challenge, driven by their skepticism toward this theoretical figure. As they put it, “[i]f you look at economics textbooks, you will learn that homo economicus can think like Albert Einstein, store as much memory as IBM’s Big Blue, and exercise the willpower of Mahatma Gandhi. Really. But the folks that we know are not like that. Real people have trouble with long division if they don’t have a calculator, sometimes forget their spouse’s birthday, and have a hangover on New Year’s Day. They are not homo economicus; they are homo sapiens” (Sunstein and Thaler, 2008, p.6). The “folks we know” are HUMANS. While it may seem evident, the reminder that we are not all like Einstein and the effort to genuinely understand human behavior was a refreshing theoretical innovation brought forth by behavioral economics. However, with the insights from behavioral economics, a normative branch known as “Libertarian Paternalism” emerged. This branch aims to correct individuals and prevent them from making mistakes to which they are predisposed. And it is this rash move, that is, this urge to correct alleged decision failures, that Rizzo and Whitman strongly oppose. The reason? It is not that clear that people are being irrational, after all. As they put it “despite having rejected rationality as a model of how people do behave, the behavioral paternalists still accept rationality as a model of how people ought to behave” (Rizzo & Whitman, 2019, p. 16). But Rizzo and Whitman pose the question: Why should we compare individuals to a pristine intellect like Einstein and label everything that doesn’t meet this standard as irrational? After all, what truly defines rationality? And why should puppet rationality be the benchmark against which we evaluate our actions? For the authors the concept of rationality should be broader, what they called “inclusive rationality”. This way of envisaging rationality acknowledges that human decision-making entails deliberate actions directed toward a given goal, while simultaneously acknowledging the practical constraints of our understanding and cognitive capacities. Rizzo and Whitman posit that when confronted with intricate decisions, the utilization of heuristics can prove cognitively efficient and advantageous.

The book offers a compelling argument against using neoclassical rationality as the benchmark for evaluating our actions. Rather than focusing in on when decisions go wrong, it’s crucial to understand that individuals make choices based on their unique contexts. In short, we need to seek to understand economic choices as if the chooser were humans (Boettke & Candela, 2017). In essence, decision-making should be viewed as a dynamic, ever-evolving journey, not just a single, static moment. In a sense the concept of inclusive rationality follows from a long tradition in economics that goes “from Adam Smith to Vernon Smith” (Boettke & Candela, 2017, p.5). This is a tradition that “that is grounded in the decision calculus of individuals, but requires neither the heroic assumptions of omniscience, nor that the individuals are interacting with others in frictionless environments” (Boettke & Candela, 2017, p.6).

Hence, the concept of inclusive rationality, while rejecting the standard neoclassical homo economicus, is still consistent with the rational choice model. Elinor Ostrom captured quite well how this way of envisaging rationality distinguish itself from the neoclassical rationality. As Ostrom posits, “consistent with all models of rational choice is a general theory a theory of human behavior that views all humans as complex, fallible learners who seek to do as well as they can given the constraints that they face and who are able to learn heuristics, norms, rules, and how to craft rules to improve achieved outcomes” (Ostrom, 1998, p. 9). Learning, self-discovery, and even mistakes are essential parts of this process. Consequently, Rizzo and Whitman acknowledge human beings make mistakes and errors, but they argue that would be very difficult for an external observer to pinpoint when one is happening. As they put it, “[w]e do not, however, claim that everyone is always fully rational. We are happy to concede that they are not. But it is one thing to say people make mistakes; it is another to identify which actions clearly and definitively are, in fact, mistakes. In the rush to characterize certain “anomalies of choice” as violations of rationality, behavioral paternalists have been insufficiently subjectivist.” (Rizzo & Whitman, 2019, p. 17). While they rightly criticize libertarian paternalism for not considering subjectivism enough, one might fear that the inclusive rationality concept swings the pendulum too far, toward a radical subjectivism approach. In fact, Dekker and Remic (2024) argue this lowers the threshold for rationality so far that nearly all actual choices are subjectively rational, what they call the “subjectivist retreat.” This, for Rizzo and Whitman, should not, however, be seen as a shortcoming of the book. Rather, this might indeed be what Rizzo and Whitman intended in the first place: shift the focus of discussion to broader definitions of rationality, taking subjectivism seriously, and disregarding the quixotic quest of identifying irrationality in our actions.

2.1 Do I know what I want?

The whole endeavor of libertarian paternalism boils down to helping people make the decisions they would do if they were not affected by bias and impulses. In short, the goal of libertarian paternalism is to make people act according to their own, true preferences. In more detail, neoclassical models of choice rely on the assumptions that the (1) agent’s preferences are well-defined, meaning they adhere to the principles of completeness and transitivity, (2) the agent’s preferences remain consistent and unchanged over time, and (3) the agent consistently acts in accordance with her established preferences. This notion of rational choice implies that the agent’s decisions and actions align with her preferences consistently and without deviation (Rizzo & Whitman, 2019, p.79). Grocery purchases are easy for ECONS: they know exactly what to buy, not only today but for all eternity. They know, even if they never tried, that they prefer a Kitkat over a Snickers (and all other existent chocolates). But are real preferences like this? Rizzo and Whitman answered in the negative. First, preferences need to be discovered. Do you know a friend of yours that says, “How do you know you don’t like this food if you never tried”? He is onto something. In other words, it is through experimentation that one learns what he likes or dislikes. Hence, it should not be considered irrational, as in the case of neoclassical rationality, to not have “their minds fully made up on the first day about the relative desirability of all alternatives in all states of the world at all points in time” (Rizzo & Whitman, 2019, p.57). Second, and following Buchanan (1979), they argue that preferences do not exist independent of the act of choice. Buchanan (1979, p.111) posited that “[i]ndividuals do not act so as to maximize utilities described in independently existing functions. They confront genuine choices”, highlighting the dynamic and, most importantly, emergent nature of preferences in the moment of choice itself. Put in other words, preferences are formed in the moment of choice and not merely discovered. He further argues that “[t]he sequence of decisions taken may be conceptualized, ex post, in terms of ‘as if’ functions that are maximized.” These ‘as if’ functions, however, “are, themselves, generated in the choosing process, not separately from such process.” Taking this insight to its conclusion, Buchanan challenges the notion of an omniscient designer being able to mimic what true preferences would be: “There is no means by which even the most idealized omniscient designer could duplicate the results of voluntary interchange.” Both aspects rob the “Archimedean point” that behavioral economists use to judge their outcomes. After all, if preferences do not exist independent of our choices, and arise only when interact with other agents, how can we judge a given behavior? If the individual does not know what their own preferences are, how can you nudge him to make him choose according to something that is not even there? Rizzo and Whitman go further and argue that even if “deep down” real preference did exist, we should not expect human beings to always act according to them. Why? We have better things to do with our time than to “consider (possibly hypothetical) pairs of options, settling on preferences over them, and making sure that all such preferences are mutually consistent” (Rizzo & Whitman, 2019, p.59). If you see a Snickers good enough on the shelf of the supermarket, why bother if this is the best possible chocolate you can pick and if last week you picked up a KitKat instead? How terrible would life be if the only chocolate you tried was a Kitkat? And, indeed, how irrational that would be?

2.2 What do you mean I am biased?

The discoveries made by behavioral economists draw attention to the existence of some alleged biases in our behavior: the tendency to overvalue the present moment (“present bias”) that leads us to spend more money than we should, eat more than we need and procrastinate; the tendency to devalue statistics and overestimate our chances of success (“optimism bias”); a greater preference for something just because we already have it (“endowment effect”); or even the “availability bias”, that is, our tendency to overestimate the frequency of a given event just because it is fresher in our memory (for example, greater reluctance to fly after hearing news of a crash). All of this, upon reflection, is quite appealing and something we all experience. Is this irrational? And if so, by what criterion? Consider the following example: an individual who spends most of his income during an early phase of his life and reaches retirement age with very little money. To an external observer, this person overvalued the present moment and therefore did not save enough (suffered from “present bias”). Next, the analyst would say that mechanisms should be created to make this person save more. However, it is quite possible to imagine that this person has a greater preference for enjoying his money when he is younger than in old age (greater utility). How do we decide which of the two is correct? Is it a preference or a bias? Aren’t we being normative? This example demonstrates that behavioral economics, while challenging the neoclassical standard of rationality by highlighting systematic deviations from the model, is still adhering to it on normative grounds. In short, behavioral economics faces a critical tension: it acknowledges that individuals frequently depart from the description of the homo economicus in predictable ways, yet it often prescribes corrective measures that implicitly endorse that standard of rationality. It is this latter move that Rizzo and Whitman consider unwarranted.

Furthermore, some economists argue that some biases can have a positive effect on our behavior. A classic example is the aforementioned “optimism bias”, where there is some evidence that individuals more prone to this bias end up effectively having greater professional success. Moreover, it is possible to argue that people who suffer more from this bias are happier. Isn’t it good to be optimistic? The essential point is not to deny the existence of biases (although some are debatable), but to highlight that it is not possible to understand how these affect our behaviors when analyzed holistically and thus demonstrate that these biases should always be “corrected.” It is also important to consider that most of the biases studied by behavioral economics are done in isolation. However, several biases may be “acting” together, possibly reinforcing or contradicting each other. Imagine that John has to study for an exam. On the one hand, he is very impatient and overvalues the present, which makes him want to have fun with friends and not study (“present bias”); on the other hand, he has great confidence in the potential benefits that a good result in this exam may have (“optimism bias”), which makes him study more. Which of the two biases is stronger? Do they cancel each other out? Should we try to change the way John acts and, if so, how? The answer is anything but obvious. Lastly, let’s consider, for instance, the famous endowment effect and the so-called loss aversion bias. Every bookworm understands that the value of a book goes beyond its market price. A book can transport its reader to a specific time, evoking memories and emotions. In essence, there’s a profound emotional bond between a reader and a book. So, regardless of the book’s market price, many might find it impossible to put a monetary value on such a cherished item. While most people naturally exhibit this behavior, behavioral economics sees it as a sign of irrationality. The individual, in this context, is succumbing to the “endowment effect,” which suggests they value something more because they already possess it. This effect is closely tied to loss aversion, a psychological phenomenon where people feel more pain from losing an item than the joy they experience from acquiring it. Once we own something—like the cherished book—our fear of the loss that comes with parting from it heightens its perceived value. This explains why readers might assign an “irrationally” high value to a book they own, even if its market value is significantly lower. In the neoclassical framework, for any given object or service, the willingness to accept (WTA) and the willingness to pay (WTP) should be identical. Using our example, if you were willing to buy a particular book for 20 dollars, you should also be willing to sell that same book for 20 dollars. However, as we’ve observed, this is often not the case. An individual might demand “more to give up a benefit they already possess than they would pay to acquire it: WTA is higher than WTP” (Rizzo & Whitman, 2019, p.106). But as Rizzo and Whitman note, “it’s hard to see, however, what is irrational… there are two valuations, but they are not, strictly speaking, inconsistent because the circumstances under which each valuation is expressed differ.” (Rizzo & Whitman, 2019, p.106). According to the authors, behavioral economists overlook this context. With neoclassical theory in mind, they perceive this as suboptimal behavior. Furthermore, Rizzo and Whitman concede that ignoring loss aversion might be justifiable in situations of intense market pressure. For instance, if someone were starving, it’s plausible that even their most treasured book would be exchanged for food. But in the absence of such conditions, they argue that it’s unclear why valuing things we already own should be considered irrational. In sum, Rizzo and Whitman contend that the evidence of bias isn’t as definitive as behavioral paternalists suggest and that even if biases exist, identifying them in theory is different from pinpointing a specific occurrence of it. Crucially, they argue that the potential benefits or usefulness of such biases in our overall behavior are often understated.

3 Slippery slope

Rizzo and Whitman argue that not only might behavioral paternalist policies restrict freedom when they are enacted, but they also have the potential to gradually shrink it over time, leading to more interventions in the future. As they put it, “accepting behavioral paternalist policies creates a risk of accepting, in the long run, greater restrictions on individual autonomy” (Rizzo & Whitman, 2019, p. 350). In short, behavioral paternalism is prone to slippery slope dynamics. In one of the best chapters of the book, “Slippery Slopes in Paternalist Policymaking,” Rizzo and Whitman excel in demonstrating why this is the case.

The first factor to consider when evaluating paternalistic policies is the inevitability of change over time, both in the policymakers who design and implement these policies and in the target groups affected by them. This raises a crucial question: How can we ensure that a policy, designed at a specific moment with certain groups in mind, remains suitable as time changes? What mechanisms for feedback are in place to evaluate and adapt the policy’s effectiveness? Additionally, how can we prevent that what begins as a minor regulation, established by one set of policymakers and readily accepted by the target group, is deemed too lenient by subsequent policymakers?

Another problem arises as consequence of the way behavioral paternalists regard their policies. Sunstein and Thaler position their approach to paternalism as being more flexible and less coercive compared to non-libertarian paternalism. They maintain that “[t]he libertarian paternalist insists on preserving choice, whereas the non-libertarian paternalist is willing to foreclose choice. But in all cases, a real question is the cost of exercising choice, and here there is a continuum rather than a sharp dichotomy” (Sunstein & Thaler, 2003, p. 1185). Paradoxically, it’s this very ambiguity that makes their approach more vulnerable to slippery slope risks. Rizzo and Whitman explore how the inherent ambiguity is problematic. They discuss, for instance, the strategy of nudging people by using risk narratives, appealing to people’s emotions. The logic behind this is to counteract people’s tendency toward optimism bias, which can lead them to underestimate, for example, the risks of behaviors like smoking on their life expectancy. To fight this, risk narratives and strong and impactful images are employed to trigger an emotional response. You might ask what is problematic with this approach. After all, if the intended goal is achieved, the policy is being used to the better. Rizzo and Whitman clearly demonstrate the problem: “For risk narratives to be effective, they must be sufficiently frightening or visceral. And therein lies the problem: there is no objective line between ‘not frightening enough’ and ‘too frightening’” (Rizzo & Whitman, 2019, p. 364). Because of this “line-drawing problem,” it is not clear what is a clear rational action or decision in the face of some existent (or perceived) risk. While it’s widely accepted that not smoking is healthier than smoking, the rationality (or lack of) of smoking becomes less clear-cut when considering the risk involved in smoking a low number of cigarettes daily. The distinction between smoking, say, 2, 4, or 6 cigarettes a day becomes increasingly ambiguous, illustrating the complexity of determining harm and rational behavior in the context of such risks.

Another aspect to consider when analyzing the slippery slope dynamics of behavioral policies arises because of the unintended or unexpected changes in the behavior of target agents in response to the policies. This unpredictability not only arises from the interactions of biases discussed earlier but also from the possibility that such policies may reduce individuals’ inclination to self-regulate their own behavior. When biases interact and individuals become less inclined to self-regulate, it is probable that new problems will emerge, and existing problems may worsen. This scenario sets the stage for slippery slopes, as “policymakers will have incentives to engage in further interventions” (Rizzo & Whitman, 2019, p. 365). Consider the example of fat taxes, discussed in the book (Rizzo & Whitman, 2019, p. 365–366), which are based on the straightforward economic principle that increasing prices should lead to reduced consumption. However, the availability of credit complicates this story. If people can access credit, they may not reduce their consumption levels; instead, they might resort to borrowing money, thereby saving less. Consequently, not only does the policy fail to decrease fat consumption, but it also leads to reduced savings among the population. In turn, this outcome could prompt policymakers to develop new strategies to encourage people to save more, illustrating how one intervention can lead to another to address unforeseen consequences arising from the initial policy.

The first half of that chapter shows how slippery slopes are likely to follow even in the presence of rational (in the neoclassical sense) policymakers. In the second half, Rizzo and Whitman explore the effects if the rationality assumption is abandoned and it is assumed that the same biases affecting individuals also influence policymakers. In other words, what is the impact of these biases on the slippery slope dynamics? As Rizzo and Whitman show, they are likely to significantly increase it. They discuss some biases, but four biases are especially relevant for the slippery slope argument: action bias, overconfidence, confirmation bias, and present bias. The impact of action bias is almost self-explanatory: policymakers have a propensity to act, leading them not only to devise policies initially but also to further their interventions once they start. As for the overconfidence bias, it allows policymakers to be overly optimistic about their ability to gather all relevant information and in their capacity to craft adequate policies. Moreover, this confidence also prevents them from interpreting unintended results as failures. These tendencies are going to be reinforced by the confirmation bias: if a policy fails to achieve the intended goal, this is taken as further proof they were right. The takeaway? It is necessary to double down and devise more aggressive interventions. But the most important bias when it comes to enhancing slippery slope dynamics is the present bias. To recapitulate, present bias is the tendency to overvalue the present moment in comparison to the future. In the context of policymaking, this can drive policymakers to prioritize the short-term advantages of a policy while overlooking its potential long-term costs. Hence, the incentives are to reap the benefits of a policy while burdening future policymakers with its costs. However, the future always arrives, and with that, these new policymakers now must deal with the problem. This is particularly the case if we also consider that policymakers discount the future in a hyperbolic fashion. As Rizzo and Whitman stated, “[h]yperbolic discounting implies that when policymakers are faced with a policy proposal that is appealing in the present but which creates a danger of bad policies being adopted further down the line, they will be inclined to focus on the former at the expense of the latter” (Rizzo & Whitman, 2019, p. 383).

Sunstein and Thaler (2008) contend that debates over slippery slopes often miss the central issue regarding the efficacy of behavioral interventions. They argue the focus should primarily be on the extent to which these interventions achieve their intended objectives, rather than on whether they might necessitate additional policies. However, this perspective does not directly address the slippery slope concerns raised by Rizzo and Whitman. They do not claim that slippery slopes are the sole issue with behavioral paternalism. Indeed, in their book, they present numerous arguments to address “the question of whether our proposals (Sunstein and Thaler) have merit in and of themselves” (Rizzo & Whitman, 2019, p. 393). Thus, the slippery slope argument is just one of many reasons to tread carefully within this research program. Furthermore, Sunstein and Thaler’s call to evaluate the merit of their proposals and to consider ways of avoiding slippery slope risks is untenable, as discussed in this section and as the authors elaborate in their book. As Rizzo and Whitman argue, this framing is merely a variation of the notion of “do the right thing now, and resist doing the wrong thing later” (Rizzo & Whitman, 2019, p. 394). However, “if the slippery slope argument is correct, there is a causal (albeit probabilistic) connection between initial interventions and subsequent ones” (Rizzo & Whitman, 2019, p. 394). The authors conclude brilliantly that “The slope risk must be counted among the costs of the initial intervention” (Rizzo & Whitman, 2019, p. 394). Rizzo and Whitman present in this chapter another reason to be wary of behavioral interventions, demonstrating the essence of the slippery slope in policymaking.

4 Do you really want to help me? Public choice problems of behavioral interventions

Take as given that people make mistakes and imagine for a moment that all methodological and epistemological issues raised by Rizzo and Whitman are overcome and that behavioral economists really know how to design policies that make people better off according to their own standards. How can we be sure that policies are going to be implemented with that end in mind? We cannot, Rizzo and Whitman argue.

First, fully rational policymakers might “face incentives to adopt undesirable paternalist policies” (Rizzo & Whitman, 2019, p. 310). Second, policymakers have to deal with the same bias that behavioral economics tries to highlight. In their view, we are committing once again what Harold Demsetz named “the Nirvana Fallacy”: comparing less-than-ideal individual decision makers to flawless and all-known public decision makers that can correct these failures (Demsetz, 1969). However, this need not be the case, and “however bad the status quo may be, the intervention could be worse” (Rizzo & Whitman, 2019, p. 310). This is especially the case if we take the public choice problems raised in the book seriously.

Public choice literature invites us to envisage policy design and implementation without romance (Buchanan, 2003), considering the incentives and epistemological issues at place. When analyzing the policies in a realistic way, often we find that policymakers usually fail to fulfill the general public interest, rather being influenced by interest groups or by their own private interests. All public policy is vulnerable to this problem but Rizzo and Whitman argue that behavioral paternalism policies are especially so. For one, consider the different consequences of mistakes: policymakers seldom face significant personal costs from flawed legislation affecting the targeted groups, whereas individuals directly suffer from their own poor decisions. This difference in the consequences of choices prompts individuals to gather information more carefully before acting than policymakers. Additionally, policymakers are unlikely to stay for long periods of time, as discussed in the previous sections. Hence, incentives exist to implement policies that have large benefits upfront, even if the long-run costs outweigh the present benefits. Once again, the incentives are not necessarily aligned with the public interest.

One of the seminal insights from public choice literature is the phenomenon of rational ignorance among voters. This concept posits that voters may choose not to be informed on policy issues when the costs of doing so outweigh the perceived benefits. How does that relate to the public choice problems that beset behavioral policies? Simply because this ignorance from the citizens gives them more room to implement policies.

However, the issue of rational ignorance is compounded by another finding from public choice theory: not all groups are equally uninformed. Certain groups, particularly those that stand to gain or lose significantly from behavioral policies, become highly informed and well-organized. Their aim is to influence legislation in a way that favors their specific interests.

When we combine the rational ignorance from most people with influence exerted by these special groups, we have the dynamics of concentrated benefits and diffuse costs, highlighted by Mancur Olson. (Olson, 1965; Rizzo & Whitman, 2019, p. 310). Hence, yet again, we can see how likely it is that, while rhetorically aiming at increasing public welfare, these policies are likely to work in favor of a particular special group. The motivations behind pressure groups aren’t solely financial; some are propelled by ideological or moral reasons. This distinction is captured in Bruce Yandle’s concept of “Baptists and Bootleggers.” (Yandle, 1983) Here, “Baptists” represent those driven by ideology, while “Bootleggers” are motivated by financial gains. It is interesting to note that, despite the fact that this dynamic can occur in any particular regulation, Yandle made this distinction when discussing a paternalist one - the banning of selling alcohol on Sundays (Rizzo & Whitman, 2019, p. 319). But Rizzo and Whitman point out that a new group of Baptists, in contrast from the religiously or morally driven groups, exist: the ones advocating policies from behavioral insights. Hence, we can expect “coalitions of both old paternalists (coming from a religious or moral perspective) and new paternalists (coming from a behavioral perspective), as well as the usual financially motivated special interests” (Rizzo & Whitman, 2019, p. 320). Because the support for these policies is coming from different sources and pressure groups, they would likely be longer and prone to expansion.

5 The future of behavioral economics

The biggest achievement of the book, not always recognized by economists, at least explicitly, is to highlight the tension between behavioral economics and welfare economics. While behavioral economics is under some criticism in the book, it is not seriously challenged, and I dare to say is even welcomed by the authors. The main “fire” of the book is turned towards the “libertarian paternalism” project, making salient its flawed normative foundations and their shaky methodological assumptions. As Fred McChesney has written: “All economics is behavioral–it is a social science, about how people behave when acting and interacting” (2013, p. 44). But while studying human action is worthwhile and necessary, contrasting real actions against the benchmark of homo economicus and deeming everything that falls short from it irrational, might not be. This is especially the case when the criteria to assess rational action is far from clear.

As the authors show in the book, evidence from bias is not as definitive as otherwise thought. They argue that distinguishing between bias and preference in specific actions or decisions is challenging, both philosophically and methodologically. This complexity has significant implications for policymaking. Even if we assume the existence of “true preferences,” designing policies that accurately reflect individuals’ actual desires is problematic. Moreover, Rizzo and Whitman demonstrated how these so-called biases interact with each other, increasing our difficulty in understanding human action even further. Furthermore, some of those tendencies might have positive impacts on our overall well-being.

How do we account for this complexity? If people engage in “self-debiasing”, how do we consider that in designing policies? And, acknowledging the diversity among individuals, how can we craft policies that are sufficiently adaptable to correct specific behaviors of heterogeneous agents? Lastly, the authors also point out the public choice problems of this research program. Who can guarantee that the policymaker is going to use the insights of behavioral economics to do “the right thing”? Moreover, given the slippery slope dynamics, how can we ensure that this does not lead to a progressive loss of liberty?

As a result of all these tensions, the authors advocate for a broader conception of rationality, one that embraces subjectivism and recognizes the limits of our understanding and reasoning capabilities, an approach they term “inclusive rationality”. This perspective seeks to broaden our understanding of rational behavior by acknowledging and integrating the complexity of human decision-making processes and the diverse contexts in which they occur.

In recent years, and to a great extent as a result of the critique in this book, we have witnessed a burgeoning literature challenging the welfare implications of behavioral economics, and one of its main offshoots: libertarian paternalism. Despite the apparent lack of consensus, we should not despair and the lively and insightful discussion that we are witnessing is the main proof of how profound this book is and how you cannot pass without it if you are interested in behavioral economics, its normative aspect, and the policy implications that ensue from it.