Back to Buchanan? Explorations of welfare and subjectivism in behavioral economics

Abstract

In light of behavioral findings regarding inconsistent individual decision-making, economists have begun to re-conceptualize the notion of welfare. One prominent account is the preference purification approach (PP), which attempts to reconstruct preferences from choice data based on a normative understanding of neoclassical rationality. Using Buchanan’s notion of creative choice, this paper criticizes PP’s epistemic, ontological, and psychological assumptions. It identifies PP as a static position that assumes the satisfaction of given ‘true preferences’ as the normative standard for welfare. However, following Buchanan, choice should be understood dynamically as a process whereby preferences constantly regenerate. Accordingly, the meaning of welfare emerges from an ongoing quest for individual self-constitution. If this holds true, then rationality axioms cannot serve as a priori normative standards. Instead, creative imagination and learning processes must remain central to any understanding of welfare in economics.

Keywords:

• Behavioral welfare economics
• creative choice
• James M. Buchanan
• rationality, methodology
• subjectivism

Acknowledgments

I express my gratitude to Ratna Behal, Matías Petersen, Christian Schubert, the participants of the Adam Smith summer colloquium in Fairfax, VA, and two anonymous reviewers for helpful comments on a draft of this paper.

Notes

1. Behavioral welfare economics is the attempt to modify rational choice-based welfare economics so that it can be reconciled with insights from behavioral economics. The central question is: How can economists make coherent statements about welfare when the choices to which they look are inconsistent? For a more advanced overview, see Bernheim (2016).

2. ‘Choice consistency’ means that an individual acts on well-defined preferences that fulfill the rationality axioms of completeness and transitivity. Completeness, means that between any two alternatives A and B, an individual either strictly prefers A to B, strictly prefers B to A, or is indifferent between A and B. Transitivity means that for any three alternatives A, B, and C, if A is preferred to B and B is preferred to C, then A must be preferred to C.

3. In their ‘nudging approach’, Thaler and Sunstein follow this concept of the individual and give the rational agent normative authority over the psychological self. They (2008, p. 5) argue that ‘individuals make pretty bad decisions – decisions that they would not have made if they had paid full attention and possessed complete information, unlimited cognitive abilities, and complete self-control.’

4. In the parlance of dual-process theory, this is similar to privileging an individual’s planning ‘System 2’ self, as opposed to the short-term preferences of the spontaneous and affect-driven ‘System 1’ self. Although there is some affinity between dual-self models and the model of the inner rational agent that underlies the preference purification approach, they should not be confounded. Proponents of the dual process theory admit that System 2 can lead to decision errors (Kahneman, 2011, p. 45).

5. This is in line with a tradition of quasi-psychological interpretation of utility in welfare economics derived from the work of John Hicks and R.D.G. Allen. In the Hicks-Allen story, utility is given a mentalistic interpretation, insofar as it is assumed that individuals hold preference orderings over the available options in their mind. This contradicts Samuelson’s construction of utility functions. In the Samuelsonian world, a utility function is a summary of observed choice data. For Samuelson, the utility function does not causally explain individual choices; it only describes the preferences of the agent in a mathematically tractable way (Rizzo, 2014). PP supports the Hicks-Allen material interpretation of utility where welfare does not equate choice and mental preferences exist prior to the moment of choice.

6. A context-dependent choice would be the following: an individual picks A instead of B in C, but B instead of A in D, where C and D are normatively equivalent choice-frames. A classic example of the violation of this axiom is ‘preference reversal’ in cases of lotteries with low probabilities but high stakes in which the framing of the question (method of elicitation) controls the answers of the players.

7. Consider decisions involving ordered lists of options (such as items on a menu in a restaurant): if all potential orderings distort the expression of an individual’s preferences, then any pick of a true preference would be random.

8. In his most recent paper, Bernheim (2016) deviates from this strict assumption. The welfare-relevant domain now includes all choices apart from ‘characterization failures’. These are choices where an individual fails to correctly connect options with outcomes. Bernheim then constructs a welfare criterion based on the properties of choices within that domain. While this ambitious project now addresses ‘choice inconsistencies’, Buchanan’s subjectivist critique (presented in the following sections of this paper) still applies.

9. The chosen paper are not necessarily representative for Buchanan’s oeuvre in general. Much of his work in public choice is explicitly built on a rational choice model of the individual. However, Buchanan himself had a ‘restless mind’ and he always focused on the ‘problems, blind spots and errors’ in a given area (Rizzo, 2014, p. 136). I will not deal with the inner tension between Buchanan ‘the subjectivist’ and Buchanan ‘the public choice theorist’, but focus on his unorthodox contributions to the conceptualization of the individual in economics. See Vaughn (2014) for an integrative discussion of both aspects of Buchanan’s work.

10. Buchanan himself does not use the term ‘agency’. However, I believe that it perfectly matches his notion of ‘an artifact, who constructs himself through his own choices’ (1979, p. 258). Agency is based on the notion of preference formation where preferences are endogenous to individual development and institutional environments. For an in-depth discussion about the difference between the concepts of agency and autonomy, see Davis (2011). Korsgaard (2009) presents a philosophical theory that resembles Buchanan’s conceptualization of the individual. She develops an account of agency in which individuals face the task of ‘making something’ of themselves, i.e. of constituting their identity through active choices and reason.

11. In addition, Buchanan thinks that his framework has more explanatory power than the Stigler-Becker model of rational choice. Buchanan points out (1979, p. 249) that stable preferences cannot reasonably account for the act of shifts in tastes. To him, the explanation of an individual’s shift in music consumption (let’s say from punkrock to jazz) makes only sense if we allow for an individual’s capability to imagine herself to be other than she is and take respective actions to achieve the imagined state. Buchanan stresses the importance of habituation in this process. In the Buchanan framework, an individual develops into the being she wants to be through intentional, repeated choices: ‘Insofar as man is wise or good, his “character” is acquired chiefly by posing as better than he is, until a part of his pretense becomes a habit’ (Knight quoted in Buchanan, 1979, p. 254). In the Stigler-Becker, there is no place for a forward-looking process of habituation and self-development. An individual is assumed to act upon given preferences; she cannot choose to develop a different set of preferences in the future and act accordingly in the present.

12. According to Buchanan (1991, p. 282), the usual epistemic argument reads as follows: ‘the individual is privileged as a choice maker because he or she knows better than anyone else what is “best” in terms of a given utility function’.

13. One could also think of a technological ‘super nudge’ where a computer algorithm perfectly steers individual consumption to its welfare-optimal point by means of visual or emotional choice framing. Buchanan (1991, p. 284) replies: ‘Arguments that involve reliance on experts in certain areas of choice must be addressed to individuals, as sovereigns, and it is individuals’ choice in deferring to expert-agents that legitimize the potential role of the latter, not some external assessment of epistemic competence, as such’ (288).

14. Buchanan (1991, p. 288) states: ‘Individuals are to be allowed to choose among potentially available alternatives simple because they are the ultimate sovereigns. And this conclusion holds independently of the state of knowledge possessed about either means or ends.’ Buchanan would probably agree with the positive assertion that nobody but the ‘man on the spot’ has the (often tacit) knowledge of time and place privileging him to correctly connect choice options with their likely outcomes. However, Buchanan thinks that this line of reasoning should not play any role in the normative justification of free choice.

15. One might wonder how fundamental Buchanan’s critique is. Does it imply that individual decision-making never follows ‘maximization subject to constraints’, even in simple choice situations when the available goods, their qualities and prices, as well as the preferred option are clearly known to the decision-maker (e.g. when buying cookies in the grocery store around the corner)? If one construes Buchanan’s subjectivist ontology strictly, people do not maximize anything in the neoclassical sense since there is no maximand in the utility function. If one interprets him loosely, then people are different maximizers at different points in time, i.e. their choices can be modeled with instantaneous utility functions. However, due to the fact that choice situations and the decision-maker change constantly, we should still abstain from modeling choice patterns by means of invariant utility functions.

16. Bernheim (2016, p. 43) admits:

… one can of course imagine environments in which choice mechanisms always distort preferences. However, we cannot accept such formulations without implicitly licensing all manner of mischief … Absent any setting that is free from an alleged distortion, we ought to question whether the associated conception of preference is merely a contrivance.

In spite of his more nuanced framework, Bernheim (2016) still subscribes to the notion of ‘latent rationality’. He accepts preexisting modes of rational reasoning, in some form, as the normative standard, despite an initial qualification of the consistency criterion.

17. Besides subjectivist concerns, there is a purely economic reason against consistency as a normative criterion: the process of forming preferences is costly, it involves considerable (cognitive and non-cognitive) information costs. Therefore, a reasonable, i.e. cost-minimizing, person will not reveal a complete and transitive preference order in a world with positive transaction costs. On this point, see Whitman and Rizzo (2015).

18. In PP, the possibility that individuals might actually not want to hold a well-defined preference set is misleadingly assumed away. A person who imagines what person she wants to become might come to the conclusion that she does not want to act rationally. Her choices then look like they contradict each other when judged from the outside, however, the person herself may be totally fine with it.

19. Robert Sugden’s ‘opportunity criterion’ can be understood as a refined application and formalization of this Buchananite framework. Through a series of papers (2004, 2006, 2008, 2010), Sugden puts forward the argument that welfare ought to be measured in terms of an individual’s opportunity set in which she should be free to express her idiosyncratic volition, regardless of the incoherence or eccentricity of the underlying preferences. While individual preferences may very well be ‘unconsidered’, individuals’ ‘valuing the opportunity to satisfy them is considered’ (Sugden, 2006, p. 217). The normative role model is the ‘responsible agent’ who treats her past, present, and future preferences as fully her own. Sugden’s criterion implies that institutional arrangements (e.g. markets) are good to the extent that they help individuals maximize their opportunities. For other recent approaches to behavioral normative economics that emphasize Buchanan’s ‘procedural perspective’, see Hargreaves Heap (2013, 2017) and Schubert (2015a, 2015b).

20. In Natural and Artifactual Man, Buchanan does not address these material means for agency, but he is primarily concerned with the necessary ‘inner prerequisites’ that motivate individuals’ ongoing quest for self-constitution. In doing so, he follows a mentalistic tradition in welfare economics. Current findings in psychology support Buchanan’s focus on mental capacities as prerequisites for welfare. Folk notions of personal identity, for example, are largely informed by mental faculties with a particularly keen focus on moral traits (Strohminger & Nichols, 2014). Naturally, people do not only need the right mindset paired with the corresponding institutional setting, but also the material security for thinking and reflecting about self-constitution. See Hargreaves Heap (2017) for a discussion of some material prerequisites for individuality from a behavioral economic perspective. See also Sen (2009) for a more inclusive discussion over what material resources are required for agency capability.

21. In doing so, Buchanan is compatible with current literature on bounded willpower where self-binding contracts are seen as an alternative to paternalistic nudging to circumvent problems of ‘internalities’ (i.e. an individual’s current actions that impose costs on her future selves). On this point, see Schnellenbach (2016).

22. Buchanan mentions J. B. Bury’s classic book The Idea of Progress as an example and adds (1979, p. 245): ‘The hopes for man, individually and collectively, held out by the post-Enlightenment social philosophers may have been naive, especially when viewed from our age. But the lesson to be drawn is surely and emphatically not one of resignation to man’s fate as a natural animal’.

23. Buchanan does not clarify what exactly he means by ‘betterness’. Instead, Buchanan simply points out, on an abstract level, that shared ideas about progress can substantially contribute to growth and development. Buchanan’s point is complementary to Deirdre McCloskey’s view presented in Bourgeois Equality (2016) that liberal ideas of equal liberty and dignity, not the accumulation of capital or institutions are the main drivers for prosperous social developments.

24. Bentham (1983, p. 251) defines the educators role as ‘neither more nor less than that of a scout; a man who having put himself upon the hunt for consequences, for such consequences of a particular kind as have been found apt to result from a particular species of course, collects them as he can, and for the use of those who feel themselves disposed to accept of his services, spreads them out in their view.’

25. In Bentham’s view, the purpose of education is ‘[for] the use of each man to lay before his eyes a sketch of the probable future more correct and compleat than, without the benefit of such suggestion, … to assist them in making reflections and drawing comparisons – in taking a correct and compleat account of the past – and from thence in drawing inferences and forming eventual calculations and eventual conjectures in relation to the future’ (1983, p. 251).

26. Matusov, von Duyke, and Kayumova (2016) develop a pedagogical approach that builds explicitly upon Buchanan’s procedural idea of the individual. In their authorial agency account ‘becoming, transcending, and transformation’ are seen as the central aims of education. According to the authors, the purpose of teaching is to facilitate a learning process that leads individuals into situations in which they can test and investigate their ideas and preconceived notions, assuming new responsibilities, and developing new questions and concerns.

27. This idea is also reflected in Schubert (2015a) and Davis (2011, p. 158ff.). It also resonates with Buchanan and Vanberg (1991, 2002) who apply the same rationale to individuals in the market process. They see the market as particularly valuable from the perspective of individuals who develop their preferences since they provide the chance to try out different preferences over time.

28. Based on Buchanan’s framework, it is difficult to formally model individual learning processes at all since individual learning standards themselves are endogenous and are determined by the development of the individual. To illustrate, what is considered a ‘correct treatment’ of a piece of information can change since individual preferences are contingent to the idiosyncratic trial-and-error based project of self-constitution.

29. In practice, it might be challenging to preclude paternalistic tendencies of the educator. Even a ‘scout’-type educator in an open-ended curriculum might influence her students’ processes of preference formation by guiding their attention to certain means of imagination and valuation. This could result in ‘means paternalism’, i.e. a situation in which individuals are nudged to use certain heuristics for how to perceive their uncertain future. Admittedly, promoting individual agency might always be a contradictory endeavor that requires balancing complex problems of social reflexivity. On specific pedagogical challenges of the authorial agency account, see Rajala, Kumpulainen, Rainio, Hilppö, and Lipponen (2016); on reflexivity and personal development, see Hurrelmann (1988).

30. In order to fulfill the full potential of self-constitution, Buchanan points out that individuals need to develop a constitutional attitude or virtue: the ‘artifactual man’ must bear the responsibilities for the creation of his own social environment, as he is in the position to ‘make his own history’ (1979, p. 257). Buchanan (1989) states that individuals need to learn some sort of ‘ethical responsibility’ for constitutional decision-making. It remains an open question how institutions would look like that foster such an attitude. See Sunstein (2017) for a recent discussion about how to design incentives that could encourage citizens’ constitutional attitude.

31. Some economists might argue that the acceptance of agency is a value judgment that is not part of economics but belongs to moral philosophy. Rizzo (2014) gives a convincing answer to this critique: advocating agency is a value judgment just like the advocacy for preference satisfaction would be. For any construction of a normative theory of the individual, the decisive litmus test is whether individual value agency or the satisfaction of rational preferences.

32. In a later article with Buchanan and Vanberg (2002, p. 128) states: ‘If there is one central constitutional implication of radical subjectivism, it is the recognition that a constitutional framework which accounts for the creativity of the human mind has to be one that allows for, and provides, favorable conditions for learning and adaptation at all levels.’

33. According to Humboldt (1792), Bildung describes both the educational process and a state of mind arrived at through education. Bildung enables individuals to cultivate reason as a creative faculty for self-development. For Humboldt, the two prerequisites for self-development are (a) the freedom to follow one’s individual aims; and (b) the exposure to a variety of circumstances and social experiences. On the Humboldtian notion of the individual, see Spong (2011).

34. Buchanan’s distinction between the deterministic (natural) and creative (artifactual) parts of human decision-making resonates with the agency model in realist social theory in that it avoids the polar extremes of voluntarism (social structure portrayed as the mere creation of human agency) and determinism (human agency portrayed as totally determined by social structure).

35. Behavioral economic insights back Buchanan at this point. Individuals seem to value agency per se: they experience procedural utility when they are engaged in active, consenting choices (Deci & Ryan, 2000; Frey, Benz, & Stutzer, 2004) and exhibit preferences against having their preference-formation process influenced by third parties (Sugden, 2016).