Towards a Liberal Behavioural Political Economy: The Constitutional Approach and the Role of Capable Agency

At the core of the political economy advanced by Shaun Hargreaves Heap (hereafter, SHH) are three fundamental insights (SHH 2017, p. 263). First, experimental evidence from behavioural economics and psychology suggests that people’s preferences are not always clearly defined, raising doubts about the idea of equating a person’s welfare with the satisfaction of a core of stable preferences. Second, liberal political philosophy need not rely on this equivalence and need not prioritize preference satisfaction as the primary focus of policymaking. SHH (2023) highlights in particular how in On Liberty John Stuart Mill presents an alternative justification for valuing individuals. Mill argues that individuals possess a unique capacity for self-direction and individuality that they develop through thoughtful reflection, discourse, and experiential learning, all facilitated by an institutional framework defending individual liberty (Mill, 1859/2003). And third, when considering public policy, SHH argues that it is crucial to consider not only the degree of negative freedom granted by the institutional framework but also how, by affording people resources and positive freedoms conducive to preference learning and experiments in living, the rules governing people’s (inter)actions foster their capacity to live an autonomous, self-endorsed life.

We begin by outlining the key features of SHH’s contribution to behavioural political economy, focusing in particular on the implications he draws from behavioural economics and psychology for institutional analysis and design. We summarize his arguments in favour of autonomy as a core normative criterion for policy-making and briefly explain what he means by a ‘constitutional perspective’ on behavioural public policy. We then seek to extend SHH’s arguments in two ways. On the one hand, we aim to flesh out a conception of agency that we think is compatible with, and to an extent is already implicit in, SHH’s writings. On the other hand, we aim to develop the institutional implications of an agentic perspective on behavioural public policy by discussing economic and political conditions that are conducive to people having a sense of agency and to developing their agentic capabilities. We aim to connect SHH’s work with recent contributions to behavioural normative economics and psychology that take autonomy seriously, in particular one of the most influential theories of agency in psychology, namely selfdetermination theory (SDT). The vantage point provided by SDT suggests that SHH has an under-developed notion of agency, because his account focuses on autonomy but has much less to say about what SDT indicates are two other important aspects of agency, namely competence and relatedness. Further, we connect SHH’s account to the Bloomington School of political economy and recent empirical insights from the literature on public deliberation that have explored the reflexive relationship between institutional structure and agentic capabilities. The liberal constitutionalism supported by SHH requires people to possess certain capabilities not just in their role as private self-determining individuals but also in their capacity as citizens who actively take part in public discourse and can, if necessary, use the provisions of the constitution to defend their liberties.¹ In developing SHH’s account in these ways, we hope to engage with his political economy framework in a coherent and constructive manner.

Both in his theoretical analysis and policy prescriptions, SHH makes extensive use of insights from behavioural economics. A key finding, which is central to his analysis of normative issues, is that people often do not have stable, well-defined preferences. Rather, the context in which they make decisions shapes their preferences in important ways. For example, experiments on loss aversion and reference dependence indicate that people’s revealed preferences are sensitive to how choice options are framed, as in risk preference experiments when people facing mathematically identical situations are more inclined to take risks when exposed to loss framings (e.g., number of lives lost) than to gain framings (e.g., number of lives saved) (SHH, 2013, pp. 985–88, 2017, pp. 253–54, 2023, pp. 3–4). Experimental evidence also suggests that the extent to which people display so-called pro-social preferences – that is, ones encouraging them to take the interests of others into account when deciding how to behave – is similarly context-dependent (SHH, 2013, pp. 988–89, 2022, pp. 58–59, 2023, p. 5): ‘pro-social’ preferences can be eroded or crowded out when decisions are moved from a non-market to a market setting and when explicit financial incentives are introduced. In a similar vein, institutions which signal that citizens are trusted can strengthen people’s other-regarding concerns, while ones indicating distrust can have the opposite effect (SHH, 2013, p. 988, 2020a, p. 113, 2022, p. 60, 2023, p. 5).²

While the standard models of rational choice can be amended so as to explain some of these anomalies, no single extension can encompass all of them. Behavioural economic findings therefore challenge the generality of the standard rational choice model (SHH, 2013, p. 989). This does not, however, imply that people’s behaviour is erratic or unsystematic. This is because people have a general capacity for reason-directed action; they “latch on to ‘reasons’ for acting that can make their actions intelligible and predictable (even if not in a sense of rational choice or subjective expected utility maximizing)” (SHH, 2013, p. 985). This involves them constructing a ‘balance sheet’ of reasons for each available action and, when the number of reasons in favour of one action reaches some threshold, it is chosen (SHH, 2013, pp. 985–86).³ On this view, rather than decision-making being guided by pre-existing preferences, given prior to the process of choice, people often form their preferences in the course of making a decision, and individuals’ decisions are best understood by considering the cognitive mechanisms through which they gather, process, and act upon information and social cues in their environment, as well as by reference to the social and institutional context in which they find themselves (SHH, 2013, pp. 985–86, 996, 2023, pp. 4–5; also see Lichtenstein and Slovic, 2006; Grüne-Yanoff and Hansson, 2009; Stewart, 2009; McKenzie et al., 2018).⁴

A second reason why behavioural insights suggest that decisions are often context-dependent is the observation that people follow rules (SHH, 2022, p. 56). When people face complicated or new decisions where they are uncertain about what they want and what the consequences of their actions will be, they tend to rely on behavioural rules of thumb (heuristics) or established habits. Since subjects typically experience laboratory settings as novel and abstract, experiments are typically not identifying preferences directly but rather the rules and habits people use on a daily basis (SHH, 2022, p. 56). People internalize those rules and habits within specific social and institutional contexts and they are activated by contextual cues that are salient to the decision-maker at the moment of choosing (SHH, 2020, p. 113). This way of understanding choice can help explain the comparatively higher rate of cooperation observed in the laboratory in social dilemma situations that are labelled ‘Community Game’ compared to social dilemma situations that are labelled ‘Wall Street Game’ (Liberman et al., 2004). Rule-following behaviour can be socially beneficial, as when people productively overcome free rider problems, for instance by encouraging tit-for-tat behaviour. But it can also be socially problematic when people internalize rules and habits that produce rigid belief systems that prevent collective action, as in the case of people believing in the righteousness of the caste system or denying the existence of global warming (Dold, 2023a).

We end this sub-section on a slightly more speculative note, by making a suggestion that does not feature prominently in SHH’s recent work on behavioural economics but which seems to us to be consistent with his emphasis on the importance of rule-following (and also with his earlier work, undertaken prior to his ‘behavioural turn’, on the importance of rule-following as a way for people to express who they are [SHH 1989, pp. 3–4]). Our suggestion is that the perspective developed by SHH is one whereby rule-following is arguably important for understanding how people are able to develop the preferences they choose to cultivate. Here we follow the work of James Buchanan, who in his paper on “Natural and Artifactual Man”, portrays people as artifactual beings who, within certain limits, possess the capacity to construct their individual characters and identities (Buchanan, [1979] 1999, p. 255). Buchanan uses the term ‘artifactual’ to denote man’s capacity for “becoming different from what he is”, to acknowledge that, within limits, we as human beings enjoy the capacity to construct ourselves as individuals” (Buchanan, [1979] 1999, pp. 247, 252). The way people do so, on Buchanan’s account, is by imposing rules upon themselves, conformity with which helps them to develop aspects of their character and identity. Buchanan considers the example of a repentant smoker, who imposes upon himself a rule prohibiting him from smoking. If he adheres to the rule, then over time, “he will find that he does become different from the person he was. His preferences shift; he becomes the non-smoker that he had imagined himself capable of becoming” ([1979] 1999, p. 253). In this way, Buchanan contends, people are able to cultivate new preferences and thereby forge a new identity. Buchanan’s emphasis on the endogenous creation of rules, whereby people devise a personal ‘constitution’ to constrain their actions and thereby over time shift their preferences and identity, seems to us to be compatible with, but not identical to, SHH’s emphasis – in works such as his book Rationality in Economics – on how acting in accordance with existing, publicly given rules can help people to express a sense of their identity (SHH 1989, pp. 3–6, 148–77; also see SHH 2001).⁵

For SHH, behavioural economics’ findings have significant implications for normative and institutional analysis (2013, pp. 986, 991–96; 2017, 2023). Standard welfare economics conceptualizes a person’s well-being in terms of preference satisfaction. But if people’s preferences are context-dependent, and may not even exist prior to decision-making, then that normative standard is called into question. The absence of stable preferences, given prior to the process of choice, makes it hard for welfare economists and policy-makers to identify a set of ‘true’, underlying preferences that can serve as a reliable benchmark for making normative judgments; “there is perhaps”, SHH (2013, p. 995) contends, “no ‘correct’ preference to have that policy should support”.⁶ The evaluative standard upon which neoclassical welfare economics relies is thereby undermined – as is the standard of behavioural welfare economics with its prominent policy agenda of libertarian paternalism (2017, p. 253).⁷

SHH argues that behavioural normative economics should focus instead on the assessment of “the conditions under which people’s preferences form and not simply with how best to satisfy them” (SHH, 2013, p. 998; also see p. 986). What matters in particular is the extent to which people are able to form their preferences in an autonomous way:

People have a sense of autonomy, according to SHH, if they feel that they are the authors of their own lives. This requires not only freedom from external coercion – that is, negative freedom – but also that people enjoy a measure of positive liberty, as SHH makes clear in quoting a passage from Isaiah Berlin:

For SHH, then, autonomy understood as self-determination requires that people are able to step back from their momentary preferences or habits and become aware of the ways in which their preferences are shaped by their situational context and wider socio-cultural environment. In SHH’s words: “To own our preferences is to have an identity and to own them requires that we know why we have them” (SHH, 2017, p. 255). In this sense, autonomy demands that people know the reasons why they have the preferences and habits they possess; they must be “able to reflect as sovereign agents on what matters to them and so have reasons to value some substantive ends over others” (2022, p. 61; also see 1989, pp. 4–5, 148–49, 2017, p. 255 and 2023, p. 2). Such self-direction – a genuine sense of personal authorship – is intrinsically valuable (as indicated by the passage from Berlin quoted above). It is also instrumentally valuable, as such individuals are more likely to choose doings and beings that are in their intrinsic and informed interests (SHH, 2017, p. 254; also see Delmotte and Dold, 2022, pp. 89–90).⁸

The move from preference satisfaction to autonomy as the benchmark for normative evaluation is accompanied by a corresponding shift in the focus of policy, away from the outcomes generated by people’s actions and towards the procedures or rules that shape their ability to reflect and act upon their evolving preferences:

SHH terms this “the constitutional approach to public policy” (2017, p. 252). It is a procedural approach, whereby – as he writes – policy “is not directed to achieve a specific outcome … it is directed at getting the rules within which action takes place right” (2017, pp. 255–56). Policies understood as changes in the ‘rules of the game’ are evaluated based on their contribution to people’s autonomy rather than on whether they further specific behavioural outcomes, such as a particular increase in pension contributions or a specific loss in weight. As SHH puts it, the constitutional approach seeks to “avoid taking a view on people’s true interests while still being active … [in being] directed at the conditions under which people acquire the sense of interest on which they act” (2013, p. 995). This alleviates to some extent the epistemic challenge facing (behavioural) welfare economists and policy-makers, because it does not require them to identify people’s ‘true’ preferences (as required by policies that target specific behavioural outcomes). Instead, it requires them only – we use the term advisedly – to identify ways of assessing and improving the process through which people make decisions (Dold, 2023b; also see SHH, 2013, p. 996). In short, the constitutional approach to behavioural public policy aims to devise rule-based interventions that are prior to preference and welfare considerations and that enable individuals to think critically about their goals and the means to achieve them.

More specifically, SHH argues that this perspective has at least two significant policy implications. The first concerns the requisite constitutional framework. Drawing on Mill’s On Liberty, SHH argues that people should be free to act as they see fit, so long as their actions do not harm others.⁹ Under this general and abstract rule – the harm principle – people are afforded the freedom to explore a wide variety of different lifestyles. The experience of engaging in such experiments in living will help them to make more informed decisions about what preferences – and, ultimately, what character and identity – to acquire, thereby contributing to their sense of autonomy. As SHH puts it, liberty, which in his Millian conception is equally available to all under the rule of law, “makes available the practices of thinking and acting through which we develop the ability to discern and desire what is best” (SHH, 2017, p. 255; also see 2023, p. 6).

However, while liberty in this sense is necessary for individual autonomy, it is insufficient (SHH, 2023, p. 6). If people are to be able to engage in experiments in living in such a way that they can explain why they decided to cultivate certain preferences, as autonomy requires, then they will “need not only a constitution of liberty, but also resources for thinking and reflecting” (SHH, 2017, p. 256).¹⁰ Although SHH does not quite put it this way, those resources are both agentic and social-structural (cf. Lewis and Dold, 2020, pp. 1161–62, 1172–75). So far as the agentic side of things is concerned, people need to possess certain “cognitive capacities” (SHH, 2023, p. 6) if they are to be able to think critically about their preferences. There is a role for policy in helping to ensure that people have these capabilities, for example through the education system (SHH, 2023, p. 6).

People’s capabilities are also enhanced by the presence of social-structural or institutional “conditions (e.g. … the media, the family, vibrancy of the arts world) that support reflection on what preferences to hold” (SHH, 2013, p. 996).¹¹ Such institutions facilitate and constrain people’s efforts to cultivate their preferences and identities, as for example when a vibrant arts scene and a thriving cultural life afford people myriad examples of paths and styles of life that can serve as provocation to or role models for one’s way of living – “material for critical reflection and discussion” – to inform their decisions about what preferences to acquire (SHH, 2023, p. 6). SHH also argues in favour of the provision of a basic income, on the grounds that people will not have the time to reflect on and take responsibility for their choices about what kind of person to become if they lack security and material resources (SHH, 2017, pp. 256–58, 2023, p. 6, 9–10). The basic income is constitutional in SHH’s sense because, like public education and health care, it is built into the rules governing society (2017, p. 257). For SHH, those rules are the backdrop against which people develop their individuality. On the public financing side, SHH argues for a constitutionally constrained redistributive tax system (2017, pp. 258–60). The rules of taxation should be updated sporadically in light of the experience of how they seem to have contributed to citizens being able to engage in the development of their individuality (2017, pp. 258–59). But adjustments to these rules should only be made from time to time to preserve their constitutional status and to insulate them against the “unfettered process of democratic politics” (2017, p. 259).

As we have seen, SHH argues that behavioural normative economics should pay less attention to preference satisfaction and more to the conditions under which people’s preferences form. SHH is in our view correct in doing so and in arguing for the importance of autonomy for behavioural normative economics. However, beyond brief references to “mental capacities”, “cognitive capabilities”, and “critical capabilities” (SHH, 2017, p. 256, 2023, p. 6), he offers relatively little detail either on the types of capacity people need to act autonomously or on the concrete policies that might promote them. In this section of the paper, we build on SHH’s insights, both by elaborating on those concepts and also by connecting SHH’s work with other recent contributions to behavioural normative economics and psychology that take autonomy seriously, locating it as part of a broader emphasis on the role of agency (Banerjee et al., 2023; Dold and Lewis, 2023; Dold et al., 2023; Hertwig and Grüne-Yanoff 2016, 2017; Sher et al., 2022; McKenzie et al., 2018).

The approaches taken by SHH and by those emphasising the importance of agency appear to be broadly compatible (with some terminological differences concealing much substantive agreement). For SHH, ‘autonomy’ requires that people be “able to reflect as sovereign agents on what matters to them”, in particular about “what preferences to hold and how to act on them”, so that people become the “authors of who they are” (2022, p. 61, 2013, p. 995, 2023, p. 2). While SHH uses the term ‘agency’ relatively infrequently in his recent work on behavioural economics, he does appear to be referring to what other recent contributions to behavioural normative economics mean by that term, namely “people’s ability to form intentions and act freely on them” and thereby to “own […] the process of behavioural change” (Banerjee et al., 2023, p. 1) so that they are “the authors of their own lives” (Dold and Lewis, 2023, p. 2).¹² Like SHH, these agency-oriented contributions criticise the nudging literature that advocates exploiting citizens’ cognitive biases (e.g., status quo bias or anchoring) in order to directly target behavioural outcomes (e.g., eat less sugar, work out more, etc.). In contrast, agency-oriented approaches aim to enhance people’s reasoning capacities and thereby improve the quality of the cognitive process that precedes choice. Examples of policies to which the agency-centric approach gives rise are (i) assistive cues that improve the quality of the information for people’s decisions and (ii) boosts that target people’s competence by enhancing the cognitive processes through which people process information (McKenzie et al., 2018, Figure 1).

Assistive cuing might involve providing an evaluator charged with the task of assessing an applicant’s score on an unfamiliar standardized test with the mean test score achieved by all candidates. The cue provided by the mean score enables the evaluator to obtain a better picture of the applicant’s abilities by offering a clearer indication of his/her ranking relative to the average (McKenzie et al., 2018). An example of boosting would be to train people in temptation bundling, a simple heuristic that can help people overcome self-control problems. By bundling instantly gratifying ‘want’ experiences (e.g., listening to music or podcasts) with ‘should’ activities that provide delayed rewards (e.g., going to the gym or doing the dishes), it can lead to a higher uptake of ‘should’ activities (Milkman et al. 2013).

Both assistive cues and boosts focus on epistemic factors in process facilitation and understand agency mainly in terms of decision-making competence, that is, as a capacity for making correct means-end calculations (Dold, 2023b). In both cases, people’s reasoning capacities are enhanced by increasing their domain-specific knowledge of what kind of behaviour would lead to a given desired outcome. In doing so, however, these process-oriented interventions only cover one aspect of agentic capacities, namely competence. Insights from positive psychology, in particular SDT, suggest that competence is a crucial ingredient for people to feel that they are the authors of their own lives. SDT defines ‘competence’ as “the experience of effectiveness and mastery … [enjoyed] as one capably engages in activities” and involves people feeling able to act with efficacy in pursuit of their given goals (Vansteenkiste et al., 2020, p. 3). However, SDT also indicates that competence is insufficient for doing justice to the sense of agency SHH has in mind, in particular the idea that people have a comprehensive sense of leading a self-directed life. For the latter to obtain, SDT highlights two other basic psychological requirements besides competence, viz. autonomy and relatedness (Deci and Ryan, 2000; Ryan and Deci, 2006, 2017, 2020). Both of these can also be understood as factors that facilitate the decision-making process, but are not directly concerned with decision outcomes.¹³

SDT understands autonomy as “a sense of initiative and ownership in one’s actions” which occurs “when one’s actions, thoughts, and feelings are self-endorsed and authentic” (Ryan and Deci, 2020, p. 1; Vansteenkiste et al., 2020, p. 3). Autonomous action is “really proceeding from its reputed author” (Ryan and Deci, 2006, p. 1561). This means that people are not just choosing the ‘right’ means to given preferences, but are aware and in control of their preference formation process and are therefore able to reflect on and pursue preferences that are authentic and non-alienating (Christman, 2009). As Ryan and Deci (2006, p. 1561) put it, “for an act to be autonomous it must be endorsed by the self, fully identified with and ‘owned’.” SDT’s notion of autonomy is similar to the one defended by SHH. For instance, as noted above, SHH (2013, p. 995–96) also emphasises the centrality of individuals’ feeling of being in charge of the preference formation process and their ability to reflect on what preferences to hold as the hallmark of autonomy. However, SDT helps to elaborate on and advance the notion of agency implicit in SHH’s work by distinguishing competence and relatedness as additional aspects of agency, in addition to autonomy.

According to SDT, relatedness refers to people’s intrinsic “need to feel belongingness and connectedness with others” (Ryan and Deci, 2000). On this view, leading a self-directed life requires individuals to be socially embedded in a respectful peer group or community (Kachanoff, 2023). One reason is that an essential part of relatedness is “experiencing oneself as giving or contributing to others” (Ryan and Deci, 2017, p. 11); feeling authorship over one’s life often involves being part of peer groups in which one can develop one’s identity and express one’s pro-social preferences. Second, experiencing social bases of self-respect and nonhumiliation is instrumentally valuable for the development of competence and autonomy; respectful communities nurture one’s self-image as a dignified being of equal worth to others (Ryan and Deci, 2003).

According to SDT, a person feels a sense of agency – of authorship and self-determination – when their psychological needs for competence, autonomy, and relatedness are met: they feel a sense of mastery in what they do (competence); they believe they are in control of the process of preference formation (autonomy); and they feel connected to others in their pursuits (relatedness).¹⁴ For the development and satisfaction of all three psychological needs, the availability of opportunities is, of course, crucial. But if people are to have a comprehensive sense of agency, then they arguably need more than just a wide range of opportunities from which to choose (Dold and Rizzo, 2021). They also need certain agentic capabilities, the possession of which enables them to exploit those opportunities and to form preferences in a competent, autonomous, and socially embedded way (Dold and Lewis, 2023, pp. 2, 7–8).¹⁵ Those capabilities seem likely to include ones such as the following:

• the cognitive capacity to envisage oneself as, at least in part, the product of one’s own making and to reflect critically on one’s current preferences and character;

• the imagination needed to envisage new preferences and a new character or identity;

• the ability to critically assess situational and social influences on individual preference formation processes;

• the ability to evaluate how well certain new preferences will contribute to the development of one’s broader goals and values (e.g., one’s overall character and identity);

• the commitment needed to cultivate those preferences until they have been acquired to the requisite degree;

• the capacity to discuss successfully with other people whether one has done enough to count as possessing a particular kind of preference (such as a ‘refined’ taste for classical music);

• the ability to connect with and care for others, to form bonds, and express one’s identity in nurturing social settings.

From the vantage point provided by SDT, the possession of such capabilities is central to people enjoying feelings of autonomy, competence, and relatedness, and therefore of agency. In short, possessing agentic capabilities will help people to be more able to make the most of whatever opportunities to develop their preferences they enjoy, enhancing their sense of agency and (therefore) their well-being (Ryan and DeHaan, 2023; Kasser and Ryan, 1996).

The term ‘capabilities’ is being used here in the sense in which it is developed by Amartya Sen, namely to refer to the set of ‘functionings’ – the types of activities and goals, including kinds of life – people can achieve (Sen 1985, 1999; also see Dold and Lewis, 2023, pp. 5–7). There are two places in SHH’s writings where he refers to capabilities in ways that suggest he would find our use of that concept to elaborate on the notion of autonomy congenial. First, in Rationality in Economics, SHH briefly suggests that the Senian notion of capabilities can be used to elaborate on what it means for people to be autonomous, writing that “[t]he capability of choosing is what underpins a reflective capacity; without it, the project of autonomy cannot get off the ground.” He also suggests that “Sen’s concept of ‘capability’ corresponds closely with the Berlin (1958) positive sense of freedom” (SHH 1989, pp. 195–96). When taken in conjunction with SHH’s statement, noted in section 2.2 above, that positive freedom – now closely associated with Senian capabilities – is an essential requirement for people to enjoy a sense of autonomy, these remarks suggest that our use of the concept of capabilities to elaborate on the notion of autonomy is consistent with SHH’s views.

The second place where SHH reflects on the notion of capabilities in ways suggestive of the account of agentic capabilities advanced here is in his discussion of expressive rationality. For SHH, expressively rational conduct is action “primarily taken for its symbolic properties: that is, the action says some, usually, evaluative thing about the person taking it. Thus the person appears ‘honourable’, ‘caring’, ‘moral’, ‘just’, ‘fair’, etc. through such expressive acts” (SHH 1999, pp. 142–43; also see n. 8 and n. 12 above). But as SHH has often observed, people cannot simply declare that their actions are worthy in the relevant way; for if the criteria for making such judgements were simply personal, evaluations would become purely “self-serving” and would therefore “prove a weak foundation on which to base on evaluation of worth” (SHH 2021, p. 45).¹⁶ It follows, SHH argues, that if such judgments of worth are to enable people to cultivate and express a sense of their identity, then they must be made by reference to intersubjectively shared social rules about what counts as certain kinds of conduct. Those rules serve as (relatively) independent standards of evaluation by reference to which the merits of people’s actions can be assessed (SHH 2000, p. 158, 170 n. 3. 2002, pp. 258–59). However, while such rules do impose restrictions on what constitutes behaviour of certain kinds, and so can provide the (relatively) independent standard needed for people to make judgements of the worth of their actions, they do not apply themselves. On the contrary, they must be interpreted and applied by people to particular cases, leaving scope for negotiation and debate between the relevant people about whether a particular action satisfies some rule (and so counts as worthy in some way). It is in this context that, as SHH briefly observes, people’s capabilities – in a Senian sense – become important, because the capabilities possessed by the relevant parties may be an important influence on the outcomes of such discussions and therefore upon people’s ability to develop their identities. As SHH puts it, “Although ‘capabilities’ are often associated with a variety of resources, there is an aspect of capabilities that has not received much attention: the resources for discussing and debating ideas about what is worthy in society” (SHH 2008, p. 90; also see SHH 2009b, p. 72). This, for SHH, indicates a possible item on the agenda for future research on the “political economy of identity formation” (SHH 2008, p. 90) and it is one to which our discussions, both above and in what follows, seek to contribute.¹⁷

In this section, we elaborate on the conditions that are conducive to people acquiring agentic capabilities and thereby support individual agency. Recently, SDT research has begun to focus on the underlying macro-level institutions and wider societal structures that shape people’s agentic capabilities and, therefore, their feelings of agency (Ryan and DeHaan, 2023; also see Ryan et al., 2019).¹⁸ The focus of attention then becomes the economic and political conditions that equip (to a degree) people with the capabilities needed to support autonomous, competent reflection on their preferences (Lewis and Dold, 2020, pp. 1174–75; Dold and Lewis, 2022, pp. 113–14). This approach is compatible with SHH’s view that a constitutional perspective on behavioural public policy would be centrally concerned with the character of the institutions that constrain and enable people’s actions, including their efforts to reflect on their preferences (SHH, 2013, pp. 995–96, 2017, p. 253, 2023, p. 6, 8–10).¹⁹

SDT supplies a critical and comparative perspective on economic and political institutions, evaluating them by analysing their influence on people’s feelings of competence, autonomy and relatedness. According to SDT, psychological needs are the mediators of the processes through which economic, social and political arrangements translate into a sense of agency. SDT highlights that both economic conditions (access to resources) and political conditions (freedoms and rights) influence people’s sense of agency via their influence on the extent to which those basic psychological needs are satisfied (Ryan and DeHaan, 2023). While SDT does not refer much to capabilities, it is arguably through their impact on people’s capabilities, and therefore on the capacity to satisfy their basic needs, that these conditions affect people’s sense of agency.²⁰

Economic conditions (factors such as income, wealth, and occupational type) can facilitate or thwart the satisfaction of people’s psychological needs and thereby contribute to their sense of agency. People in occupations that are perceived to be of a lower socio-economic status (SES) report less job satisfaction, more emotional exhaustion, and lower vitality, all of which indicate a low sense of agency (Gonzalez et al. 2014). People in higher SES occupations have more resources to invest in their knowledge and skills (competence), are more likely to be exposed to a wider set of social circumstances that stimulate critical thinking (autonomy), and have the means to connect with and care for others (relatedness). Economic resources allow people to travel and meet people, and thus develop their imagination; they foster a sense of control over what preferences to cultivate and give people the freedom to pursue preferences they have reason to value (Sen, 1999). Yet SDT highlights that it not only matters how much resources people have access to, but also what the prevailing economic conditions are, in particular how wealth is distributed around them. Di Domenico and Fournier (2014) find that socioeconomic inequality in subjects’ geographic vicinity is a negative predictor of feelings of competence and autonomy. This is a point acknowledged by SHH (2017, p. 256) who highlights the potentially detrimental effect of inequality on the scope for ‘experiments in living’ and the development of people’s individuality. As he writes, “the character of the constitution that matters for public policy purposes is not just the liberty it gives, but also its redistributive bias. This is because an individual’s acquisition of individuality depends on both their freedom and their resources” (2017, p. 263). Furthermore, relatedness is typically higher in jurisdictions with effective redistributive policies since “inequality affects trust, community life, and violence, and … it predisposes people to be more or less affiliative, empathic or aggressive” (Wilkinson and Pickett, 2020, p. 236). This is compatible with SHH’s viewpoint, according to which a basic income financed by a progressive tax is an effective redistributive system that enables more of the populace to engage in experiments in living compared to the currently dominant means-tested system of welfare benefits (SHH, 2017, p. 257). A positive effect of basic income for relatedness might stem from its egalitarian treatment of all citizens, which signals equal worth of all community members and can thus help nurture one’s sense of belonging and embeddedness.

Political conditions – the rights and freedoms formally afforded to people under the prevailing constitutional and legal regime – also matter for satisfying people’s basic psychological needs and, therefore, for shaping their sense of agency. A political regime that affords citizens core liberties – freedom of action, association, and expression – has a direct, positive impact on their sense of autonomy, because it allows people to engage in the experiments in living that are crucial for their sense of being the authors of their own lives (Bradshaw et al., 2023). This is, of course, why SHH emphasizes the importance for people’s sense of autonomy of a liberal constitutional order that protects them from external coercion and affords them the opportunity to act as they see fit (subject to the constraints imposed by the harm principle).

But it is also noteworthy that, as the political economists Elinor and Vincent Ostrom have argued, formal constitutional rules and procedures are necessary but insufficient for a functioning liberal political order of the kind advocated by SHH. The reason is that the rules and procedures comprising a constitution of liberty do not automatically apply and enforce themselves.²¹ Rather, they must be invoked and applied by citizens to defend their liberties, not least when their freedom to engage in some novel course of action is challenged. That in turn requires at least some citizens to be skilled at using the provisions of the constitution and legal system to challenge and check the use of the state’s coercive powers – whether by elected politicians, unelected bureaucrats, or other groups of citizens – so as to sustain their freedom to engage in experiments in living.²² However, if people lack those “self-governing capabilities”, as Vincent Ostrom (1997, p. 4, 31) terms them, then they will be unable to make effective use of the relevant constitutional and legal procedures and their freedoms may be eroded (also see E. Ostrom, 2006, p. 9; Ostrom and Ostrom, [2004] 2014, p. 86). Just as we argued above that people’s efforts to develop their personal identities are facilitated by the possession of certain kinds of agentic capability, so we contend here that when acting in their capacity as citizens people also need certain ‘public’ agentic capabilities if they are to sustain the constitution of liberty that underpins their freedom to explore a variety of different lifestyles.

The process through which policies and laws are established matters for the development of these public agentic capabilities. This point is acknowledged by SHH, who – like the Ostroms – draws on Tocqueville to argue that people’s civic mindedness is developed “from the experience of an extraordinary range of civic association”, that “the origin of these practices of civic association for [Tocqueville] is political participation”, and that political life “teaches the means of combination” (SHH 2020a, p. 210, quoting Democracy in America).²³ Recent insights from the empirical literature on public deliberation support SHH’s conjectures and suggest that political participation fosters public agentic capabilities and thus contributes to people’s sense of agency in a number of ways. First, jurisdictions that actively include citizens in the democratic decisionmaking process (e.g., by means of deliberation in public citizen forums)²⁴ can enhance citizens’ competence, for instance by increasing their understanding of political topics and correcting preference distortions caused by either active manipulation or passive overemphasis on symbolically potent issues (Niemeyer, 2011). SHH (2020a, p. 115) emphasizes that this competence building is particularly important for “sniffing out” positive-sum games in politics (that is, those opportunities or projects where collective action serves all or most people’s interests such as public health initiatives like clean water and clean air). The idea that public deliberation and political participation can help build competences follows from the insight that “humans are … poor monadic reasoners but not poor group reasoners” (Chambers, 2018, p. 37).²⁵ Group reasoning displays better results than individual reasoning on many issues, including bias detection, information search, and depolarization (Mercier and Sperber, 2011).²⁶ By exposing them to facts about policies and different evaluative standards in the public discourse, people can become aware of some of the unquestioned situational and socio-cultural context effects (Niemeyer, 2011).²⁷

Second, the experience of being an active member of a community can inculcate a sense of autonomy, understood as ownership of and identification with the political process, even if one disagrees with the policies ultimately chosen. Taking part in public deliberation and local rule-making enhances people’s sense of autonomy because they “feel that there has been fairness in representation and that they have had their due voice in governmental processes” (Ryan and DeHaan, 2023, p. 1162). Being actively engaged and vocal within a community provides individuals with the confidence to openly share their viewpoints, recognizing the potential of their voices to influence results.

Third, public deliberation and political participation can also enhance the sense of relatedness when people realize in public forums how much they actually agree with others, in particular regarding the nature and the legitimacy of the relevant issues, even though they may not agree on concrete policy outcomes or the veracity of different claims. The metaconsenus on the rules of discourse and the political issues at stake builds empathy and the experience of a shared sense of purpose in decision-making (Niemeyer, 2011; Ryan and DeHaan, 2023). This means, as SHH (2020a, p. 116) highlights, that even if one ends up as the loser in a zero-sum political game (e.g., because one belongs to the richer part of the population and has to pay high taxes that exceed the personal benefits received), the feeling of being part of the process makes the acceptance of the outcomes of the rules enacted more likely. In other words, people “come to internalize the practice of rule-following through political participation” (2020a, p. 117).

Finally, taking part in public deliberation and local rule-making can lead to a mutually enforcing effect between autonomy and relatedness. If people view institutions as controlling, perhaps because they have been externally imposed, then their sense of autonomy suffers which can cause a crowding out of relatedness understood as the willingness to express intrinsic pro-social motivation (Ostrom 2005). Conversely, if the political process gives people their due voice, then “self-esteem is fostered, and the individuals feel that they are given more freedom to act, which enlarges self-determination”, thereby nurturing and crowding in relatedness and pro-social motivation (Ostrom 2005: 113; also see Ostrom 2000, pp. 5–6, 9–11, 13 and 2005, pp. 76–79. 94–95).²⁸ These insights support SHH’s conjecture that an inclusive, voice-based political decision-making process encourages people facing social dilemmas to take each others’ interests into account, boosting cooperation by turning perceived zero-sum interactions into positive some games and thereby depolarizing an emotionally charged political landscape (SHH, 2020a, pp. 115–11; 2020b, section 3).

Consistent with SHH’s approach, we have discussed the contours and implications of a political economy programme that takes behavioural insights seriously and emphasises agency rather than preference satisfaction as a normative benchmark for behavioural public policy. We suggested that insights from positive psychology in the form of SDT can help flesh out the notion of agency that is latent in SHH’s approach. We explored some of the concrete institutional implications of this extension of SHH’s framework by discussing empirical findings on economic and political conditions that are conducive to the development of people’s sense of agency. Following SHH’s work, we argued that an agency-centric perspective shifts the focus from the redesign of concrete choice architecture (e.g., in the form of nudges) to a more general analysis of the institutional conditions that nurture or stunt the development of people’s capability to reflect on, act on, and identify with their evolving preferences. We argued that institutional reforms need to go beyond boosts and assistive cueing that focus narrowly on competence and also consider autonomy and relatedness as additional dimensions of agency-centric policy-making and institutional design. We highlighted the importance of public deliberation and participation for the development of agentic capabilities. The empirical literature suggests a crucial role for deliberation within groups for people’s civic mindedness and the quality of individual reasoning in the context of preference formation processes.

We believe that SHH rightly emphasizes that a core motivation for a liberal behavioural political economy ought to be that individual preference-formation processes are not dominated by social and cognitive factors that diminish a person’s sense of agency. In contrast to libertarian paternalism, such an approach to policy would not take a strong view about precisely what people’s preferences should be. It would therefore, we believe, be consistent with SHH’s preferred constitutional approach to policy, seeking to “avoid taking a view on people’s true interests” and instead being “directed at the conditions under which people acquire the sense of interest on which they act” (SHH, 2013, p. 995). Admittedly, a political economy that emphasises how people’s agentic capabilities can contribute to their sense of agency poses new challenges. A core challenge might be the question of whether ordinary citizens actually see themselves reflected in the approach’s emphasis on self-reflection and individuality. People throughout history have shown a robust tendency to conform to social norms, fads, and fashions. SDT provides powerful empirical arguments for why people ought to be concerned with the development of a sense of agency – not only for its own sake but also for its instrumental value for psychological well-being. However, it is an open question whether people will find its promises appealing or, alternatively, too psychologically demanding.

While this paper has elaborated on the notion of agency inherent in SHH’s account, it has not fully fleshed out the notion of practical reason to which SHH subscribes. In emphasizing the critical role of reasoning in the selection of ends, as well as means, SHH acknowledges Kant’s account of practical reasoning as a viable alternative to the instrumental rationality commonly assumed in conventional economics, but without wholeheartedly embracing it. Consequently, future research could explore the question of the specific account of practical reason that best complements SHH’s account. Furthermore, it remains unclear where SHH stands on the issue of creative agency and determinism, in particular the question of whether acknowledging creative agency requires a departure from deterministic models of choice that portray people’s conduct as a unique, utility-maximizing response to their circumstances. We think it does, but SHH does not clearly say so. We alluded to Buchanan’s oeuvre in this paper, but future work might explore further the implications of creative agency for economic theory and policymaking by discussing connections between SHH’s approach and the ideas of thinkers, such as Buchanan and G.L.S. Shackle, whose work engages deeply with that topic.