I Choose for Myself, Therefore I Am : The Contours of Existentialist Behavioral Economics

: Behavioral economics and existentialism both present informative perspectives on human choice. We argue in this article that the dialogue between the two approaches can enrich the current debate about the normative implications of behavioral economics. While behavioral economics suggests that our capacity to choose is constrained by cognitive biases and environmental influences, existentialism emphasizes that we can (and should) treat ourselves as free and ‘becoming’ beings in spite of the many constraints we face. Acknowledging these two perspectives in the form of a theoretical synthesis — which we propose to call existen-tialist behavioral economics — provides us with reasons why we should protect our choices ‘as our own’ and how doing so may be more difficult than we anticipate. It also provides a framework to analyze the threat of identity-shaping social and technological developments, such as prefer-ence-altering nudges and artificially intelligent prediction algorithms.


I. INTRODUCTION
Behavioral economics has been steadily on the rise in recent decades (Geiger 2017).It has been quite successful in exposing many of neoclassical economics' limitations in explaining and predicting human choice (Bernheim, DellaVigna, and Laibson 2018).More specifically, behavioral economics has helped to refine models of individual choice based on empirical evidence that challenges the axioms of neoclassical rationality (Sugden 2018, 7-14).Since at least the publication of the bestseller Nudge, by Thaler and Sunstein in 2008, there has been a lively debate about the normative implications of behavioral economics.At its core lies the question of how economists can reconcile the insights about the 'messiness' of people's revealed preferences with the logic of traditional welfare economics that assumes welfare to be the satisfaction of rational, well-ordered preferences.While the literature in behavioral normative economics is rich and diverse (for an overview, see Bernheim 2016, andSunstein 2020), there are two main approaches.
The first approach, behavioral welfare economics, is most prominently advocated by Thaler and Sunstein (2008), as well as by Bernheim andRangel (2007, 2009).It aims to define welfare as the satisfaction of preferences that are context-independent and time-consistent.While observed choice is sometimes not a proper indicator of individual welfare, and can be difficult to parse, behavioral welfarists assume that deep within people hold 'true' preferences; that is, preferences they would reveal if they "paid full attention and possessed complete information, unlimited cognitive abilities, and complete self-control" (Thaler and Sunstein 2008, 5).Such 'true' preferences are assumed to be well-ordered in that they fulfill the standard axioms of neoclassical rationality (Infante, Lecouteux, and Sugden 2016).It is the task of economists to identify choice frames in which individuals hold those 'true' preferences and take them as normative input for policymaking.
The second approach, which has been forcefully defended by Robert Sugden (2004Sugden ( , 2018)), aims to respect individuals' choices without referring to the underlying structure of preferences.It is grounded in the conviction that the behavioral welfarists lack a psychologically plausible theory that could explain the origin of rational preferences (Infante, Lecouteux, and Sugden 2016).Sugden argues for an alternative normative standard which he calls the opportunity criterion.This criterion says that, when comparing social states, economists should not be concerned with the rationality of individual preferences but should instead seek institutional arrangements that foster people's opportunities, that is, the size of their choice sets.
Much has been written about the advantages and pitfalls of these two approaches (see, for example, McQuillin and Sugden 2012;Dold and Schubert 2018;Lecouteux and Mitrouchev 2021).In this article, we do not want to dissect the intricacies of those approaches.However, we want to point out that both approaches suffer from a blind spot in that they do not systematically address the issue of agency freedom-individuals' capability to make choices that they perceive to be authentic and meaningful. 1 In doing so, both approaches underestimate the importance of the nature or quality of the choice process and, for instance, sidestep the important issue of whether a person has the capacity to reflect upon and possibly revise the values and beliefs that motivate her actions (Dold and Rizzo 2020).Also, the degree to which the social environment shapes and constrains individuals' processes of 'becoming' is not sufficiently addressed in either approach (Dold, van Emmerick, and Fabian 2020). 2 We think that this blind spot can be illuminated by the existentialist perspective we develop in this article.Such a perspective shifts the informational basis of normative economics away from welfare and opportunity toward authenticity and meaning.In doing so, we do not take a stance on the relative merits of behavioral welfare economics and the opportunity approach.Instead, our approach runs parallel to these as another ethical approach to behavioral economics.
Throughout the article, we treat existentialism as a branch of ethics that treats agency freedom and the quest for authenticity as central to human existence (Olafson 1967;Webber 2018).The notion of authenticity helps identify situations in which humans are 'unfree' (Golomb [1995] 2012; Crowell 2020).At its core lies the question: do I succeed in being the author of my life, or am I mainly a product of the situational framing or the wider social environment I find myself in?Taking this existentialist perspective can enrich positive economic theorizing by shedding light on a wide range of behaviors that traditional choice theory has trouble explaining, such as "mountaineering" (for the latter, see Loewenstein 1999, 318).In this article, we illustrate ways in which the existentialist 1 The capability approach of Amartya Sen (1985Sen ( , 1999) ) is the closest approach in normative economics to the account we sketch in this article.Sen is famous for highlighting the important normative role of agency freedom and for arguing that a person's sense of agency is both intrinsically valuable and of central importance to her well-being.We deviate slightly from Sen in our definition of agency freedom.For Sen, "[a]gency freedom [is the] freedom to achieve whatever the person, as a responsible agent, decides he or she should achieve".It refers to "what the person is free to do and achieve in pursuit of whatever goals or values he or she regards as important" (Sen 1985, 203).It is remarkable that the Senian perspective has gotten little traction in the debate on the normative implications of behavioral economics in recent years.For exceptions, see Davis (2011, chap. 8) andDold, van Emmerick, andFabian (2020).For an excellent philosophical discussion of Sen's idea of agency freedom, see Binder (2019).
2 By 'becoming', we mean the process through which a person's preferences change over time.Depending on her level of self-consciousness, a person might or might not be aware of her own process of 'becoming'.
perspective can enrich the discussion about the normative implications of behavioral economics.
Our argument is based on insights from empirical psychology that highlight individuals' inherent desire for authentic, meaningful choices which, in turn, has a considerable impact on human behavior and economic activity (Loewenstein 1999;Karlsson, Loewenstein, and McCafferty 2004;Phelps 2013).Authentic and meaningful choices are those that we can feel are in line with our developing values and the narratives we are constructing of our lives.There is evidence that a person's perception of meaning and authenticity is essential for her physical and mental wellbeing (Ryan andDeci 2006, 1566).When people perceive that they fulfill meaningful roles in their families and communities that align with their overall values, they show a higher ability to cope with uncertainties, withstand unhealthy temptations, endure hardship, and work more steadily toward difficult goals (Routledge and Bitzan 2020, 2).
Some readers might wonder why we have chosen to bring together such disparate strands of literature.There are two primary reasons we believe that a dialogue between existentialist philosophy and behavioral economics is fruitful.First, we think existentialism and behavioral economics are effective complements.Both begin with the same object of investigation, choice under uncertainty.Behavioral economics uses the tools of decision theory and existentialism uses philosophical reasoning.In doing so, the starting point for both is the acknowledgment that human decision-making is bounded by unknown outcomes and ambiguous probabilities. 3 Moreover, where existentialism may sometimes assume an extreme conception of individual freedom (Webber 2018, 2), behavioral economics often leans toward emphasizing our cognitive flaws and the situational influences that trigger "systematic mistakes" (Bernheim and Rangel 2007, 16).Hence, the two have much to gain from each other.A combined perspective allows us to conceptualize individuals who are, at the same time, internally and situationally constrained, but who are also striving for freedom and self-realization.
Second, existentialism and behavioral economics occupy parallel positions within their respective disciplines.Just as behavioral economics emerged as a response to the formal rigidity of rational choice axioms in neoclassical economics, existentialism can be seen as an antagonist to analytic philosophy's focus on formal properties of language (as exemplified in Wittgenstein's Tractatus Logico-Philosophicus) and the idea that "logical analysis is the main business of philosophy" (Bertrand Russell, quoted in Jones 2009, 13).Existentialists like Heidegger were seeking "to free philosophy from logic" (Matthews 2003, 171). 4Both Heidegger and the French existentialists (de Beauvoir, Camus, Merleau-Ponty, Sartre) wanted to shift philosophy's focus away from 'narrow' problems of language and logic to 'wider' reflections of the actual lived experience of people in an uncertain world (Kaufman 1956, 12).Central to existentialism is the thought that people do "not have an inbuilt set of values that they are inherently structured to pursue.Rather, the values that shape a person's behavior result from the choices they have made" (Webber 2018, 4).This is very similar to the idea in behavioral economics that people don't 'come to market' with a set list of master-preferences but often construct them in the process of choosing (Lichtenstein and Slovic 2006).
The structure of this article is as follows.In section II, we discuss some of the core tenets of existentialist philosophy: the notions of freedom, agency, and authenticity.Section III sketches the contours of what we call existentialist behavioral economics.We explore both the work of earlier behavioral economists (in particular, Kahneman and Tversky) and a newer, more philosophically inclined branch of behavioral economics.Section IV presents the two core elements of existentialist behavioral economics: constrained freedom and the pervasive influence of the social environment on individual choice.Sections V and VI illustrate practical implications of existentialist behavioral economics.

II. EXISTENTIALISM: A BRIEF OVERVIEW
Existentialism evades precise definition, particularly because so many of its leaders refused to accept the label (Kaufman 1956, 37).Existentialism's very DNA contains within it a resistance to labels; the existentialists celebrate individuality and self-definition, to lump them into a single mass is to violate their very credo (Irwin 2015, 11).They also frequently disagreed with each other and even with their past selves.For example, both Sartre and Heidegger experienced a "turn" later in life, by which their philosophies transformed into something very different from their original iterations (Irwin 2015, 14).While Sartre's Being and Nothingness clearly takes inspiration from Heidegger's Being and Time, Heidegger is famously quoted as exclaiming of Being and Nothingness, "How can I even begin to read this rubbish!"(Bakewell 2016, 201).The two also held radically different political views, as did many of the philosophers within the tradition.
Despite all of this, we will use the term 'existentialism' as a blunt instrument.When we reference what 'the existentialists think', we mean the recurring patterns of existentialist thought, at all times asking our readers to keep in mind that existentialism is a colorful, diverse set of ideas.The reason we have chosen to do it this way is because, despite their disagreements, the existentialists share fundamental, core ideas and it is this core to which we wish to attach ourselves.5

II.I. The Existentialist Core
The existentialists are united in rejecting a perspective in which philosophers are seen as neutral observers who seek a general, formal account of what it means to be human (Crowell 2020).The existentialists "are opposed to any system which abstracts from the hopes and fears of the individual or which attempts to fit man into some sort of impersonal schematism" (Roberts 1952, 469).Existentialism is grounded in lived experience and in the acknowledgment of the individual holding an essentially subjective, not a bird's eye view.In existentialist thought, "genuine philosophizing must well up from a man's individual existence and address itself to other individuals to help them to achieve true existence" (Kaufman 1956, 23).Thus, it is a philosophy designed for one's life as one leads it, diverging significantly and intentionally from the rigid theorems and tight structures of logical-formal inquiry.
To this end, one hallmark of existentialist writing is that much of it is done through story.Nietzsche and Kierkegaard rely on metaphor and allegory, Sartre wrote many plays, and Camus is perhaps most famous for his novel The Stranger ([1942] 1970).People understand their life through narrative, through the "drama of decision-making" (Harari 2018, 55), and existentialists employ literature to explore this human drama, to spotlight subjective experience and to give voice to those parts of our lives which we do not fully understand.The existentialists remind us vigorously that "man is not primarily a thinker![…] He is a volitional actor, a being who makes choices and lives by them" (Killinger 1961, 306).
Existentialists explore possible responses to the hardest challenge of what it means to be human: to live with the recognition of death, the absurdity of life, and the "benign indifference of the universe" (Camus [1942(Camus [ ] 1970, 120), 120), which does not reveal any objective meaning to us.This thought is powerfully articulated by Albert Camus in his famous article, The Myth of Sisyphus ([1942] 1955).Camus compares the human condition to Sisyphus' fate: each of us is consigned to push our 'burden' up a hill only to lose control of it once again, a task void of meaning.Further, if Sisyphus turns inwards, he does not discover any immutable moral law that could serve as a substitute for life's absurdity (Webber 2018, 14).Paradoxically, this situation creates a radical freedom for him: if neither life's meaning nor his own nature is fixed, he has the freedom to make himself, to become different from what he is through his own choices.Sisyphus cannot change the task assigned to him by the gods, but he is still free to choose what those circumstances mean to him.The world will not become meaningful on its own nor will it listen to his laments.But he can become resolute, accept 'the courage to be' and march up the hill 'in spite of' the awareness of his finite existence and the seemingly absurd situation he is thrown into (Tillich 1952).Consequently, rather than giving in to dark cynicism or acquiescence in light of the universe's indifference, existentialism can be seen as a philosophy of exploration, creativity, and energetic resistance.Existentialists contrast this emphasis on self-exploration with potential forces of unfreedom, showing that in order to be free, we must understand the ways in which we are not (Roberts 1952, 471).The notion of authenticity plays an important role in clarifying situations of 'unfreedom'.

II.II. Unfreedom and Authenticity
The French existentialists are famous for placing freedom and responsibility at the core of their philosophies.But it was German philosopher Martin Heidegger who particularly emphasized the notion of authenticity. 6n his 1927 magnum opus, Being and Time, Heidegger describes what the individual is and what she must resist in order to become her authentic self.Heidegger uses his own terminology, unsatisfied with the limits imposed by a language laden with preconceptions.Thus, he labels his conception of individuals' existence as "Dasein", loosely translated as "beingthere" (Bakewell 2016, 60).When Heidegger references 'being-there', he refers to a type of Being far and above our typical use of the word.Lowercase being generally refers to any living thing; but Being means something much more expansive.For the purposes of this article, Being can be thought of as 'consciousness'.
By 'being-there', Heidegger refers to an existentialist concept, that of being "situated" or "embodied" (Crowell 2020).Various existentialist thinkers have their own interpretations of this concept, but the general thrust is that individuals are, in Heidegger's terms, "always already": we are always already within a situation or circumstance and cannot detach our experience, our identity, or our perceptions from this situated position (Wrathall 2002, 220).We belong to a world that precedes us, and we are always enmeshed within it.There is no "view from nowhere" (Nagel 1986), no detached objectivity available to us (Bakewell 2016, 65).
Heidegger argues that constraint is a necessary part of what it means to be human and, consequently, understanding constraints is necessary to live authentically.Part of those constraints is "das Man", translated as "the they" (Bakewell 2016, 78) or "the one" (Dreyfus 1995, xi).'The one' is a term that describes situations in which a person follows common practice-she does what 'one' does; for instance, the person could be 'eating as one eats' (for example, with fork and knife) or 'driving as one drives' (for example, on the right side of the road).As both examples indicate, conforming to 'the one' is not necessarily negative.7It can save time and energy if a person does not have to determine how to go about routine tasks.Moreover, conforming to 'the one' fulfills a positive coordinative function as it establishes reliable expectations of what other people do.It can, however, become problematic when a person does not just conform to common practices in certain parts of her life, but when she uncritically internalizes the norms and values of her social surroundings.Then, conformity becomes conformism; her actions are succumbing to the roles and patterns laid down by 'the one'.
For Heidegger, 'the one' is not merely external pressure (for example, social conventions) but also something that lives within us (for example, emotions such as self-doubt or shame) that we must struggle against (Bakewell 2016, 78).In order to achieve authenticity, we must distinguish the line between the functional aspects of 'the one' and the conformist influence that lives both inside and outside us.Authentic Being is not a refusal of 'the one', it does not demand that we extract ourselves from our emotions and our socio-cultural embeddedness.Instead, it demands that a person develop an attitude toward and consciously commit to certain elements of herself and her socio-cultural surrounding that are meaningful.
Consider the following example: a person who simply does action  because this means following the prevailing norm in her peer group-'it is what one does'-is living inauthentically.However, if this very same person does  because it reflects values she identifies with, then-following Heidegger's logic-we can say that she acts authentically.In the latter case, the person will perceive doing  as a meaningful act, as being in line with who she is or wants to be.If a person has the capacity of committing to authentic choices on a continuous basis, she is depicting agency freedom.In this sense, agency freedom requires an acknowledgment of the socio-cultural constraints, but demands also that we rebel against these limitations when they threaten our capacity of making authentic choices.For existentialists, the rebellion is particularly important when the constraints are the result of ourselves or those around us, rather than the result of an indifferent universe we cannot control.8

III. TWO STRANDS OF BEHAVIORAL ECONOMICS
While behavioral economics has been instrumental in exposing many of neoclassical economics' limitations in explaining and predicting human behavior, we argue in this section that the mainstream view within behavioral economics, which we will call strand one, is still limited by a mechanistic view of individuals.However, a new subfield of behavioral economics, which we will refer to as strand two, has begun to adopt a more nuanced vision of individuals qua economic agents.While still limited by the logic of economic models, this more recent strand of behavioral economics privileges a dynamic view of choice and identity, which has productive commonalities with existentialist thinking.

III.I. Neoclassical Economics and Its Discontents
Neoclassical economics has been the dominant economic paradigm for decades and can be characterized by the combined assumptions of maximizing utility, market equilibrium, and stable preferences, which are "used relentlessly and unflinchingly, [and] form the heart of the economic approach" (Becker 1976, 5).Individuals are assumed to be rational; that is, they act according to a set of "integrated preferences" (Sugden 2018, 7) that are stable and internally consistent.Integrated preferences are assumed to be context-independent, meaning they are unchanged by situational influences, such as informational framing or fleeting emotions (Sugden 2018, 5).This understanding of the individual is obviously simplistic, yet the simplicity is believed to be a large part of its appeal.Neoclassical "parsimony" (Tirole 2002, 636) makes modelling easier in that it abstracts from random deviations in individual rationality, which is a crucial virtue of economic models that seek to give a general account of average human behavior in market settings.Behavioral economists argue, however, that neoclassical economics' explanatory and predictive power is handicapped by its misunderstanding of individuals' systematic deviations from rationality.In his magnum opus, Thinking, Fast and Slow, Daniel Kahneman (2011) details the core results of his research with Amos Tversky on systematic biases and heuristics in decision-making, instructing the reader to imagine that our minds operate under two (metaphorical) mental systems, System 1 and System 2. System 1 thinks fast.It jumps to conclusions and, as a result, makes biased judgments.Yet, it allows us to react quickly and to conserve cognitive energy in situations that cannot be solved with hedged bets and rough estimates.System 2 thinks slow.It demands more of our cognitive resources but enables us to solve complicated problems, which require care and deliberation.
One example Kahneman uses to illustrate System 1's hasty thinking is the anchoring effect.It describes the tendency of individuals to base their estimates on an available cue (the 'anchor'), for example, an initial piece of information or a random number.Kahneman describes an experiment in which participants were asked to guess whether the value of an item X was higher or lower than a number provided by the experimenter.The item X included the height of Mt.Everest or the number of member states in the United Nations.The median guess of participants who were shown a high initial anchor was much larger than that of participants shown a lower anchor (Jacowitz andKahneman 1995, 1163).Such an effect occurs even when the anchor is completely unrelated to the estimation task.Critcher and Gilovich (2008) found that estimations of athletes' performance can be anchored by their jersey's number and that sales forecasting is influenced by a product's model number.Despite the fact that the anchor is unrelated to the task in both cases, System 1 still grabs onto it as an unconscious shortcut and systematically biases participants' guesses.
The anchoring effect is merely one of many ways that individuals do not act like the optimal problem-solvers neoclassical economics assumes them to be.Other systematic deviations from rational choice include people's assessment of gains and losses as to reference points, their overweighing of small and under-weighing of large probabilities, and their choices being affected by how options are framed (Kahneman 2011).In general, behavioral economics in the tradition of Kahneman and Tversky has been a powerful rebuttal to the neoclassical assertion that perfectly rational, utility-maximizing agents are good proxies for average human decision-making.

III.II. Strand Two Behavioral Economics
In recent years, a second type of behavioral economics has emerged that puts more emphasis on preference change and social-not just situational-aspects of decision-making: "strand two behavioral economics" (Hoff and Stiglitz 2016, 28). 9 It explicitly models the individual as an enculturated actor who is shaped by her own choices and socio-cultural surrounding.If the main goal of Kahneman and Tversky's program ('strand one behavioral economics') was to contest neoclassical hyperrationality, then strand two's goal is to contest neoclassical stasis.
In neoclassical economics, "neither individuals' utility functions nor their preferences are changed by choices they make" (Davis 2003, 49).The self is what it is, and it stays that way: "In the standard analysis of choice, 9 Hoff and Stiglitz state that strand two behavioral economics "is an attempt to broaden economic discourse by importing insights into human behavior not just from psychology, but also from sociology and anthropology" (2016,25).We want to point out that they particularly stress the idea from sociology that the social structure influences individual identity.However, they do not investigate to the same degree the question of how individual action transforms social structure.For a discussion of a comprehensive sociological approach to identity, see Davis (2011, 79-81, section 4.3.1).
individuals can never choose anything other than their most preferred option as determined by their preferences and feasible set.Preferences, then, determine choice" (Davis 2003, 49).In this sense, behavioral change is understood as the outcome of the mechanistic interplay between one's stable preferences and changes in prices or income. 10There is no room to account for a person's freedom to hold different preferences in the future. 11 While 'strand one behavioral economics' of Kahneman and Tversky allows for distorting factors in the choice process (for example, anchoring or framing affects), it still subscribes to a static view of human nature: individuals are ultimately constituted by their core of 'true' preferences.The benchmark for 'good' choice is the satisfaction of those true preferences that are assumed to be stable and integrated (Rizzo and Whitman 2020, 80).It neither explains the psychological foundation of those 'true' preferences, nor does it allow for the possibility that a person's 'true' preferences are endogenous to her social environment or that they may change over time (Infante, Lecouteux, and Sugden 2016).This means that traditional behavioral economics portrays a simple, dualistic view of what constitutes human choice: there is a stable inner ECON that is struggling to come to the surface, but situational influences (such as arbitrary anchors or informational framing) are throwing HUMANS off course.Hoff and Stiglitz (2016) contrast this static understanding with strand two thinking which seeks to explore how social identity impacts and is impacted by economic behavior.They begin with an idea familiar to traditional behavioral economics: priming.They discuss an experiment in which the participant's prime is their own identity.A group of bankers were told they would be playing a nation-wide coin-flipping game and were instructed to flip a coin and report the results, with each "winning toss" worth $20 (Hoff and Stiglitz 2016, 34).However, some of the bankers were primed beforehand to recall their role as banker while others were not.Individuals who were primed to think of their banker identity acted less ethically than those who were not primed.The researchers explain, "the context of the moment of decision [the identity prime] influences choices even when the context should be transparently irrelevant to the decision" (Hoff and Stiglitz 2016, 26).Thus, in this case, the way that individuals thought of themselves, that is, how close-at-hand their identity as banker was, altered their choice behavior.Put existentially: how subjects believe that they 'should' act is shaped by their understanding of how 'one' acts in a certain role (here: the role of a banker).
While this experiment applies notions of identity to economics, it still does not represent a significant departure from traditional behavioral economics.Yet Hoff and Stiglitz go further, suggesting that identity is not only relevant as an external, temporary influence on choice but as an internal, long-lasting one as well, manifesting through "cultural mental models" (2016,25).This position is a radical departure from the typical economic perspective.To evidence their point, Hoff and Stiglitz cite an experiment performed in India where both low-caste individuals and high-caste individuals were asked to play a two-player assurance game in which each anonymous player chooses to 'hunt hare' or to 'hunt stag' over a series of rounds.Hunting hare offers a low reward consistently, regardless of the other player's action, whereas hunting stag offers a high reward only if the other player also chooses hunting stag.Low-caste individuals were more likely to choose stag even after being 'burned' in the first round by the other player choosing hare.Alternatively, high-caste individuals were more likely to switch to hunting hare after an unsuccessful first round, exhibiting a "retaliatory" response (Hoff and Stiglitz 2016, 39).The authors posit that this discrepancy can be explained by highcaste individuals choosing to protect their ego and sense of status, noting that "a survey provided supporting evidence: high-caste men were much more likely than low-caste men to think it appropriate to beat up an individual who, because of adverse or ambiguous circumstances, had caused one a loss" (Hoff and Stiglitz 2016, 39).This research implies that the social (not just the situational) context, in which identity is expressed, affects the choices people make. 12Hoff and Stiglitz elaborate on this idea by proposing a model of the self as a set of potential selves, each of which will be differentially triggered according to context.While a person may become more like one of these identities over time, she will not solidify into a self that is more 'real' or 'true' than the other selves she could have become.Her identities and preferences are not subject to some ultimate destination, though they are moderated by previous choices and the circumstances in which the person finds themselves.
If we left the discussion here, strand two behavioral economics would seem as deterministic as strand one; it only ascribes a greater role to a person's inculcated identity in shaping her choice patterns.However, Hoff and Stiglitz argue that the reverse is also true: an individual's choices will influence her identity.They theorize that individuals engage in a process of "learning by consuming", in which current consumption influences downstream choice (2016,31).By this process, preferences are not static; rather, old preferences inform the development of new preferences, and the choices a person makes in the past will necessarily affect the ones they make in the future.For instance, a person who starts consuming jazz music will end up with a preference for live jazz concerts that she did not have at the outset.This means that strand two behavioral economics acknowledges that preferences are not only endogenous to the social environment somebody is enculturated in, but that they are also changing as a function of one's past choices.

IV. ELEMENTS OF EXISTENTIALIST BEHAVIORAL ECONOMICS
Strand two behavioral economics paints a more dynamic picture of individual choice and identity that comes closer to the existentialist understanding we sketched in section II of this article.However, it is important to note that it is not yet describing the freedom and agency existentialist philosophers had in mind.While Hoff and Stiglitz make clear that in their account "[man] is a social being.He is social not just in what he does, but also in the lenses through which he perceives himself and the world", they also argue that "[man] does not choose the lenses that he uses.They are socially acquired and socially activated" (2016,39).Their account underscores how impactful the social environment is for the development of individual identities.Yet, it does not allow for individuals to step back from their environment and their past choices and rank the various selves or social identities they are exposed to.This is a result of the fact that strand two's conceptualization of the individual still conceives of social identities as parts of the individual's utility functions.It does not leave room for intentional choices about what 'type of utility functions' a person wants to hold in the future.So, there is a tension between existentialism's emphasis on active self-creation and strand two's emphasis on the dominant influence of the mental model ('the lens') that comes with one's social environment.In what follows, we want to sketch the contours of existentialist behavioral economics that build on elements of Hoff and Stiglitz and combine them with the economic approach of a pioneer of strand two thinking, James M. Buchanan.

IV.I. Constrained Freedom
According to Buchanan ([1979Buchanan ([ ] 1999)), it is essential that economists acknowledge that real-world people learn, grow, and seek to change themselves.Buchanan ([1979Buchanan ([ ] 1999, 247) , 247) criticizes traditional economics for not acknowledging change as the individual's defining feature, arguing that people are crucially motivated by 'becoming': individuals have the ability to shape themselves and make choices in pursuit of the identity they wish to achieve.They may not always have the same preferences, or even understand fully what these preferences are, so they will value the freedom to pursue these preferences as they arise and shift.Part of this venture involves for Buchanan a recognition that individuals cannot always be best understood through their utility functions or given preferences.By this view, individuals are defined, at their very core, by an ironic lack of definition.They are uncertain and opaque, even to themselves, and always in flux. 13ike in existentialist philosophy, the individual in Buchanan's framework has the freedom to self-create but they are not unconstrained, free to become anything.Instead, Buchanan points out, humans are "natural and artifactual" ([1979] 1999, 247), both determined and free.As natural beings, we are limited by the confines of our body, our family, our culture, and our circumstances.Yet, we are also artifactual; we can construct ourselves, pursue different ends, change our mind, resist, and engage in processes of becoming different.Because individuals are not entirely determined, there exists "a large set of possible persons that one might imagine himself to be, or might imagine himself capable of becoming" (Buchanan [1979(Buchanan [ ] 1999, 250), 250).People are not subject to any teleology but are instead the agents of their own experiences and identities; their intentional choices direct them into one or another version of themselves.14This point adds an important nuance to strand two behavioral economics in that it argues that an individual's self-identity is the result of their purposeful choices that impact and are impacted by their environment.

IV.II. The Influence of 'The One' and the Power of System 2 Thinking
Buchanan acknowledges that individuals' capacities to reflect upon and direct their own processes of becoming are shaped by their socialization and upbringing.He ascribes an important role to education.It can "provide persons with both an array of imagined prospects and some means of valuation" (Buchanan [1979(Buchanan [ ] 1999, 254, 254) that lie at the heart of the capacity to form one's own identity.However, Buchanan believes that adults, by and large, possess the necessary motivational and cognitive means to intentionally choose different paths of becoming (Dold 2018).In doing so, he might have underestimated the powerful role of cultural lenses and cognitive models individuals often adopt unconsciously and uncritically-a feature that both strand two behavioral economics and existentialist philosophy highlight.
If it is true, as Hoff and Stiglitz (2016, 51) argue, that our environment has the power to become a dysfunctional or dominating influence on our processes of identity formation, then existentialists are right to warn us from das Man. 15 Hoff and Stiglitz show that individuals become, in part, the identity that their context brings forth from them.If one wishes to self-create and avoid passively assuming the "they-self" (Bakewell 2016, 78), then individuals must cultivate their own self-understanding, uphold the capacity to creatively imagine a different future for themselves, and critically assess the ways in which their environment presses upon them.As Hoff and Stiglitz argue, the experience we have in the world is the result of the lens through which it is filtered.But as Buchanan argues, we also have the freedom and the burden to build ourselves the lens that will help us best cope with our world and allow us to choose the life that we find meaningful.
How can we shape this lens?We think that behavioral economic research gives us a partial answer, in particular Kahneman's distinction between 'thinking, fast and slow'.The existentialist individual, though free and self-created, will not be immune to impulsive reactions or to expressions of self which may not be entirely her own.To resist the world all the time would be exhausting and lead to diminishing returns, especially in cases where das Man fulfills socially productive, coordinative functions.Thus, social influence cannot and should not be banished entirely, and it will likely emerge most frequently through the operation of the fast and unconscious System 1.In order to cultivate a more secure, self-directed identity, the existentialist individual can look to her slower System 2. Here, the individual can learn to prepare for and moderate the outbursts of System 1 such that slowly, with time, she will learn to identify and reconsider das Man, channelling it to the extent that it is conducive to her own agency freedom.Some might argue the radical freedom proposed by existentialists is nothing but the philosophical embodiment of neoclassical rationality.However, we think that such a position ignores that System 2 thinking does not need to coincide with integrated preferences; it only describes the mode of critical-creative reasoning, not its outcome in terms of formal properties of an individual's choice patterns.Authentic choices require becoming aware of oneself and the world around us, it is the ability to step back from one's 'given' preferences by means of System 2 thinking.Recalling Hoff and Stiglitz's mental models, identity colors the world we see as well as our participation in it; if we want to make choices that are meaningful and authentic, we must be aware of the kind of lens we are behavioral economics can help us identify situations where das Man over-determines individual choice and suppresses an individual's awareness of what it means to choose creatively and freely.
working with and aim to transform that lens in cases where we sense a lack of agency freedom.

V. TWO APPLICATIONS: NUDGING AND ARTIFICIAL INTELLIGENCE
We believe that the dialogue between strand two behavioral economics and existentialist thinking we sketched in the previous section leads to a more realistic understanding of individual choice, which stresses both the normative significance of and practical obstacles to agency freedom.As we argue in this section, that dialogue can also help us understand some of the dangers in recent policy innovations like nudges and artificial intelligence.

V.I. Existentialist Behavioral Economics and Nudging
The most prominent example of behavioral welfarism is the policy paradigm of libertarian paternalism which aims to honor freedom of choice while helping individuals satisfy their 'true' preferences (Thaler and Sunstein 2008).It does so through the use of 'nudges', that is, alterations in decision frames that influence behavior by making certain choices easier or more salient.For example, the manager of a college cafeteria might decide to put apples on an eye-level shelf at the cashier and candy bars on a top shelf in the back of the room in order to nudge students into picking the apple.While nudges can be an effective tool to achieve given policy goals, such as achieving a higher degree of sustainable energy consumption (Kaiser et al. 2020) or tax compliance (Holz et al. 2020), from an existentialist perspective, there is reason to question the underlying narrative that tries to justify interventions on the grounds of 'true' preferences.
One such reason is that, as we have already discussed, individuals may not even have given preferences, instead using their choices to experiment and determine what to value (Hargreaves Heap 2017).Thus, "the challenge is to find economic institutions that can be recommended to individuals who do not know what they prefer", and whose so-called 'true' preferences are frequently opaque (Sugden 2018, 15).One solution behavioral welfare economists have used to remedy this lack of information and assess whether or not an enacted policy (like a nudge) actually satisfied an individual's 'true' preference is the "as judged by themselves" (AJBT) criterion (Paul and Sunstein 2019).AJBT states that an action taken to improve someone's welfare can be judged as successful or not successful only from the point of view of the targeted individual.Since the nudge might alter a person's revealed preferences, Paul and Sunstein (2019) advocate that nudged individuals must assent to the nudge post hoc; that is, after being nudged, they approve of the outcome achieved.For example, if an individual is automatically enrolled in a 401(k) retirement program and then, after the fact, reports she is satisfied with this outcome, then the AJBT criterion deems default-enroll justified.
Yet this solution is imperfect, particularly because people become different in the process of choosing.Choices are 'transformative' (some more than others), and this quality threatens AJBT's applicability.If the individual post-choice is much changed from the individual pre-choice, then post-choice self-evaluation will not capture the opinions of the prechoice self.Imagine someone trying to decide whether or not to have a child.If they become a parent, they will be fundamentally changed and now occupy a different identity than if they remained childless.In one sense, the difference between the pre-knowledge and post-knowledge selves may not matter; if the pre-choice self no longer exists, perhaps the only experience that matters now is that of the post-choice self.However, this does not address the issue in its entirety because even if the postknowledge self is satisfied, they cannot know the experience of the counterfactual childless scenario and therefore do not know if their choice was the optimal one.This is of course an inherent feature of individual decision-making in an uncertain world, and its philosophical intricacies may not be of much significance for personal well-being.However, when economists make policy recommendations for a large population (for example, in terms of incentivizing having offspring through child allowances), those philosophical intricacies need to be addressed since the standard for decision-making has to be higher in politics, and preference change has severe consequences for the application-and policy prescriptionsof stated preference approaches in cost-benefit analysis. 16.II.Sisyphus Confronts AI Artificial intelligence (AI) is a broad term, referring generally to machines that can perform problem-solving tasks at or above the level of a human (Bringsjord and Govindarajulu 2020).In this section, we want to focus on narrow AI.This is the AI used to predict our choices and recommend things to us, whether in the form of products for purchase, video streaming, or partners on dating sites.Prediction algorithms may seem a small issue relative to hot-button issues in artificial intelligence like the future of work, the ethics of self-driving cars, or the defense against a superhuman consciousness.However, from an existentialist behavioral economic point of view, these technological questions should not obscure AI's immediate, everyday impact on our agency freedom.In fact, we believe that prediction algorithms' apparent low risk and widespread appeal makes recognizing their potential pitfalls all the more pressing.
The primary appeal of artificially intelligent prediction algorithms is that they can track a myriad of past individual choices, making connections and recognizing patterns that individuals themselves cannot remember.Moreover, prediction algorithms can make use of 'datification', which compounds observations about one individual with the observations of millions of others.17In this way, AI can identify the beats per minute a person exercises to most effectively, locate exactly which subsub-genre of comedy makes her laugh the hardest, or expose destructive health habits she never recognized herself.AI also saves transaction costs associated with choice, like the time and effort required to gather information and compare options.In this sense, AI frees up cognitive resources by outsourcing elements of a laborious decision-making process.Consequently, a large part of AI's allure is that it unburdens us from many of the responsibilities of choice, making our lives simpler, easier, and more comfortable.
However, as pointed out by existentialism, choice is how individuals navigate the world and construct their identities; if they are not making their own active choices, do they relinquish the very means they have for creative personal growth, undermining their own agency?Considering the logic of existentialist behavioral economics, we think that there are (at least) two reasons for why the answer might be 'yes'.
The first reason is that artificially intelligent algorithms do not just help us satisfy our given preferences in a more effective way.Instead, especially in those situations where individuals do not have well-defined preferences yet, there is a risk that we substitute our own preference formation process with that of an AI algorithm.This substitution may be benign or even beneficial so long as it is accompanied by individuals' System 2 thinking and active choosing.18However, we can identify Heidegger's logic of das Man in modern algorithms.AI is powerful in part because it uses the will of the crowd to determine and inform our particular preferences.Based on data from millions of users, the algorithm tells us what 'one' should choose in a certain area of life, for example, what 'one' should do to become fit or healthy. 19Strand two behavioral economics shows us that this problem is a real one: we are very much susceptible to 'the herd'.Our preferences and identities frequently reflect the environment we find ourselves in and we adopt a self that accords with the other selves we see.This is by no means a new phenomenon, but the addition of AI expedites and expands the traditional process.Where before preferences spread locally, they now spread globally; where before they spread in months or years, they now spread in days or weeks.There will be less space for differentiated local pockets of taste and opinion to emerge and challenge each other; instead, preference adoption can become much more universal and, therefore, much more threatening to authentic individual choices.
The second reason is that the delegation of active choice to AI algorithms might lead to an erosion of our ability to experiment with different identities and reflect on our preference formation process.AI's personalized recommendations may stunt individual growth by biasing us toward preferences we revealed in the past.Challenging our own prevailing tastes and opinions can be difficult and uncomfortable, often requiring conscious effort on the part of the individual.If AI is designed to give us what we craved in the past, then it will likely reinforce and amplify specific elements of our identities.In addition, the fewer choices we make for ourselves, the less trained we are in making intentional choices and engaging in System 2 thinking.We can think of an individual's capacity to make active choices as a muscle that atrophies in the long run if not used frequently enough (Rizzo and Whitman 2020, 252-253;Schubert 2021).
While an individual might actively choose to restrict her choice set knowing that choice overload could impede processes of self-experimentation, a systematic outsourcing of active choice via AI might make it more likely in an event that requires serious effort and perspective-taking that individuals will not be equipped with the skills of creative imagination and evaluation that Buchanan ([1979] 1999, 254) identified at the heart of an individual's authentic process of becoming.As AI gets better at knowing what individuals will likely consume, individuals will have fewer incentives to explore beyond their comfort zone and try something unfamiliar or untested, yet this is precisely how humans learn and improve.There is neurophysiological evidence that we learn predominantly from making mistakes, not from confirmed expectations (Schultz 2016).And in line with the existentialist credo that lived experience lies at the heart of personal growth, individuals need "to feel responsible for [their] actions if those actions are to be a source of learning" (Hargreaves Heap 2017, 257).If we allow artificial intelligence to do our learning for us, we're not really learning, nor will we be prepared for the problems AI cannot solve for us.20

VI. AGENCY-FRIENDLY INSTITUTIONS
Admittedly, to claim one's agency is difficult.One reason is, as strand two behavioral economics shows, we are constantly pressed upon by situational and environmental influences.Existentialist behavioral economics does not dispute that these constraints exist or are meaningful.It rather argues that we, both as individuals and policymakers, should think about strategies to reduce their dominance.Family, upbringing, culture-we cannot escape any of them.Yet this is exactly the existentialist point: a fundamental retreat into social isolation is neither possible nor desirable, but an existentialist economic lens can help us identify both enabling and oppressive forms of identity-shaping influences.Furthermore, it can help us envision institutional reforms-for instance, in education or work environments-that foster the idea of individuals as claimants of their own freedom.
Consider the example of 'boosting': boosts have been proposed as an alternative to nudging.Unlike nudges, which are goal-directed and exploit behavioral biases (for example, the status quo biases in the case of defaults), boosts "foster people's competence to make their own choicesthat is, to exercise their own agency" (Hertwig and Grüne-Yanoff 2017, 973). 21In general, boosts are interventions that focus on creating and promoting individuals' cognitive and motivational competencies.They either target competencies directly (for example, by teaching decision strategies or changing procedural routines) or indirectly by altering the choice environment (for example, by changing the informational representation).Furthermore, instead of externally redesigning the choice environment for individuals based on an agent's assumed 'true' preferences, boosts provide individuals with the ability to redesign their proximate choice environment (Samuli and Hertwig, forthcoming).For instance, agency freedom in cyberspace can be boosted when users are given the option to see the full history of a post or to customize how their news feed is designed and sorted (Lorenz-Spreen et al. 2020).Such boosts can increase agency freedom by helping individuals become aware of the power of situational influences not just in cyberspace, but in other areas of their lives as well.
If strand two behavioral economics is right in emphasizing that cultural and environmental influences are powerful identity-shaping forces in our lives, then social institutions play a large role in determining the kind of people we become.In acknowledging the reflexivity between social structure and individual preferences, the existentialist perspective shifts the institutional focus from outcomes to the rules and processes that performatively shape choices.At the same time, the existentialist perspective we have described in this article asks from economists a certain degree of epistemic humility.Hargreaves Heap rightly notes that economists cannot know in advance in any detail what agency freedom will consist of for a particular person; it is an individual capability, "a state that does not depend on some set of final outcomes (like the experience of friendship or love), it is a state that depends on the character of the route taken to it" (2020,117).Institutions play an important role in determining 'the character of the route'.They can foster individual selfreflection and active choice, or they can nudge and steer people.The existentialist perspective we developed in this article defends the former and challenges the latter.Agency-friendly institutions "would seem naturally to be concerned with the conditions (for example, the educational system, the media, the family, the vibrancy of the arts world) that support reflection on what preferences to hold" (Hargreaves Heap 2013, 996).Naturally, the repeated exchange of goods and services on markets can also contribute to self-reflection and individual processes of preference learning.By means of 'experiments of living', markets can be seen as arenas that foster agency freedom and complement the aforementioned sociocultural institutions (Delmotte and Dold, forthcoming).

VII. CONCLUSION
Artificial intelligence and nudging are but the newest in a long line of technological innovations that promise to help individuals make better choices.They offer to unburden us from a laborious and error-prone decision-making process.Existentialist behavioral economics can provide us with a conceptual approach that helps us assess what the impact of those external decision devices might be.This approach does not call for us to become luddites that shun social and technological innovations altogether.Instead, it invites us to weigh the costs and benefits of delegating active choices.Choice is difficult, and nudging and AI purport to assume that burden for us.But if we use them unreflectively and unflinchingly, we do so at a potential cost to our agency freedom.
We hope that philosophically inclined economists and philosophers of economics find some merit in the dialogue between existentialism and behavioral economics we sketched in this article.We believe that the core tenets of existentialism can add to contemporary behavioral economic thinking by offering a richer understanding of human choice that conceptualizes individuals as being in flux and motivated by their quest for meaning and authenticity.At the same time, we think that behavioral economics does provide the existentialist perspective with a much-needed empirical grounding that rightly highlights how pervasive the influence of situational factors and cultural mental models on human choice is.The integrated perspective of existentialist behavioral economics conceptualizes individuals as beings who are constrained but at the same time motivated to develop preferences that they can fully embrace as their own and whose satisfaction leads to meaningful choices.Such a perspective invites behavioral economists to move away from efficiency and preference satisfaction as the benchmarks of policymaking.Instead, it opens a much-needed discussion about the character of institutions that constrain or enable individuals' agency freedom.intersection of behavioral economics, philosophy of economics, and history of economic thought.Contact e-mail: <malte.dold@pomona.edu>Alexa Stanton graduated from Pomona College magna cum laude in 2020, with a major in Philosophy, Politics, and Economics (PPE), and a minor in Computer Science.In her thesis, Alexa examined the links between existentialism and behavioral economics, and how these links inform perspectives on artificially intelligent prediction algorithms.She now works as a consultant at Transom Consulting Group.Contact e-mail: <alexastanton@gmail.com>