Navigation – Plan du site

AccueilNuméros15-4Revue des livresComptes rendusDavid M. Kreps, Ar­guing About Ta...

Revue des livres
Comptes rendus

David M. Kreps, Ar­guing About Tastes

Till Grüne-Yanoff
p. 679-684
Référence(s) :

David M. Kreps, Ar­guing About Tastes, New York: Columbia University Press, 2023, 202 pages, 978-023120991-5

Texte intégral

Afficher l’image
Crédits : Columbia University Press

1Everybody, even economists, concedes that people’s motivations change at least occasionally. Opinions sharply diverge, however, on whether this matters for economics as a social science. One group of naysayers, let’s call them “pragmatists”, argues that economics focuses on topics or decision horizons where motivational change—unless driven by changing information—has no relevance. Another group, the “reductionists”, claims that all such changes can be reduced to a stable internal motivational core, plus a production function that produces behavior-relevant preferences from this core and various external stimuli.

2In Ar­guing About Tastes, Kreps challenges both of these perspectives (though he does not explicitly distinguish between them). He contests the pragmatists by identifying economic phenomena that require accounting for motivational change, and he challenges the reductionists by arguing that all preferences as shaped by context, experience, intrinsic motivation, identity, self-perception, or the social environment. Kreps’s goal is not to present new material, but to persuade economists to model changing motivations, and thus break with the long-held orthodoxy of de gustibus non est disputandum: “economists should worry about tastes” (9) Yet his arguments also become entangled in some conceptual and methodological strictures, leaving it unclear whether the path he suggests is indeed open to economists.

3This slim book is based on the 13th Arrow Lecture held at Columbia University in 2021. As Kreps acknowledges, it offers only a small fraction of the possible ways to discuss people’s changing motivations and the determinants of these changes. The discussion is largely limited to the interaction between intrinsic motivation and extrinsic incentives, both in static and dynamic situations. Kreps makes two core assumptions in his discussion. First, that one can meaningfully distinguish between intrinsic and extrinsic motivation. Intrinsic motivation according to Kreps consists either in doing things that are inherently fun or interesting, internalizing the welfare of others or being wired to conform to social norms. Extrinsic motivation, by contrast, is driven by expected external rewards, including reciprocation or social approval for following a norm. Economists of course model extrinsic motivation changes all the time, most prominently perhaps due to changes in prices for goods and services. This is where Kreps’s second assumption comes in: he argues that also intrinsic motivation is subject to contextual influences in manners relevant for economic explanation and policy design.

4This second assumption draws on both economic and non-economic theories. Within economics, Kreps primarily references Akerlof and Kranton’s work on identity economics and Bowles’ research on how economic policies can erode ethical motivations. A person’s sense of identity—who she believes she is and how she fits into society—affects her utility function, as it reflects how well her choices align with the norms and ideals of their social identities. For example, a person seeing herself as an employee may view overtime work as an expression of loyalty to her employer, while as a parent may see it as disloyalty to her family. Crucially for Kreps, people internalize the welfare of different individuals based on these different identities, leading to distinct intrinsic motivations. Contextual factors often trigger switches between identities, thereby changing intrinsic motivations.

5These identity-mediated changes in intrinsic preferences can also be self-referential: the contextual features influencing a person’s identity may stem from their own behavior. For instance, the employee who frequently works overtime may interpret her own behavior as an expression of loyalty to her employer, thus stabilizing her identity and reinforcing her intrinsic motivation. Kreps draws here on Bem’s self-perception theory, integrating concepts from social psychology.

6A related issue is how changes in monetary or regulatory incentives can erode a person’s perception of an ethical interaction, transforming it into a market exchange. This echoes Bowles’ well-known point about “crowding out,” where intrinsic motivation is replaced by extrinsic incentives. The consequences can be drastic: a small fine might completely eliminate the ethical obligation to behave in a socially responsible manner, even if it isn’t high enough provide a monetary incentive to continue behaving in the same way.

7While these theories highlight the possibility of intrinsic preference change due to contextual (regulatory) influences, they don’t specify the strength of these various contextual or self-referential influences, nor do they predict which fine will be too small to maintain previously ethical behavior. To make this trade-off more precise, Kreps presents a tongue-in-cheek principal-agent model featuring Columbia’s Joseph Stiglitz attempting to incentivize David Kreps to deliver a lecture. The model starts with principal Stiglitz offering an honorarium and setting a no-show penalty, while agent Kreps decides whether to participate and how much effort to invest, based on the incentives and risks. In the model, Kreps accepts and exerts socially optimal effort if both the honorarium and the penalty are positive and contingent solely on Stiglitz’ evaluations. However, in the real world, no one imposes monetary penalties on honorary speakers. Why not? The proposed answer is that monetary punishment could reduce the speaker’s intrinsic motivation. If the reduction in intrinsic motivation is strong enough, no acceptable deal involving a penalty may exist. Alternatively, even if a deal is acceptable, Stiglitz might be better off not imposing any penalty, as the quality of Kreps’ talk would depend on his intrinsic motivation, which the penalty could undermine.

8These models are highly idealized, and Kreps almost apologetically calls them “caricature models,” wondering whether “the math is hiding what is going on.” The formal results are not particularly strong—they offer only sufficient, but not necessary, conditions for an acceptable deal in the first model, and the results from the ‘crowding out’ models follow obviously and directly from the additional assumptions. Instead, Kreps presents a series of numerical scenarios with different honoraria and penalties, showing “with reasonable parameter values … how powerful intrinsic motivation can be and how quickly extrinsic incentives … can become counterproductive” (61). I found this helpful for illustrating the models’ mechanics, specifying plausible numerical possibilities. However, they didn’t fully convince me that these are more than possibilities—“that ‘could be’ should be ‘can be’ and, very often, ‘is’”, as Kreps puts it. Instead they just remain how-possibly explanations amongst many others.

9Kreps stresses that such modeling exercises challenge the economic mainstream. The term “orthodox” appears 25 times in the book, and references to “mainstream economics” and “the traditional view” are frequent. Indeed, Kreps positions himself as an iconoclast: “A colleague suggested that … insofar as I succeed in convincing others to reject de gustibus, it would ruin economics. That’s a heavy charge. But it is one I am prepared to endure” (6). I found this claim somewhat exaggerated for two reasons.

10First, despite his declarations, it remains unclear what the “orthodox view” actually is and how Kreps’ ideas differ from it. In a footnote, Kreps clarifies that “my argument is with the broader principle that economists should leave the explanation of tastes to other social scientists and the (often implicit) further assumption that tastes or preferences are fixed, unchanging” (174). Yet even mainstream economists recognize that preferences can change—expected utility theory, for example, can explain preference changes over actions based on shifting beliefs, and price changes can explain why consumers come to prefer not to buy certain products. Obviously, that’s not what Kreps has in mind; he argues for intrinsic preference change after all.

11But there are many theoretical attempts to model the change of some intrinsic preferences, while insisting that all these changes can be reduced to interactions of some unchanging motivational core with changes in non-motivational factors. Lancaster’s household production function comes to mind here—preference for inputs (e.g. labor and materials) derive from household production capacities and preferences for outputs. Stigler and Becker’s approach is another example, deriving preferences for goods and activities from the interaction of an immutable core of motivational factors and changing external factors such as economic conditions, education, advertising, and peer effects. Even Akerlof and Kranton’s identity economics might fit this mold, as it explains preferences through the influence of identity on an otherwise stable set of motivations.

12Kreps’ relationship to these reductionist accounts remains unclear. On one hand, he critiques Becker’s approach for trying to preserve the assumption of unchanging preferences by modifying the arguments in the individual’s utility function (40). On the other hand, he praises Akerlof’s identity economics for rejecting the “unchanging preference” assumption, while offering his own modified version of it in form of a utility function uA(x;i)—the utility of person A in state x if she perceives her identity as i (35). How these two approaches differ he does not really explain. Ultimately, Kreps concedes that Becker too is in the business of “arguing about tastes” but rejects his model for relying too heavily on unobservable variables.

13This is of course a methodological, not a conceptual argument, so I was left wondering whether Kreps sees his own theoretical position as related to, or substantially different from these reductionist approaches. If he claims that all intrinsic preferences change, what are his reasons for doing so? And is the intrinsic—extrinsic distinction, besides its obvious methodological challenges, even conceptually clear? After all, Kreps puts a lot of weight on a specific causal feature: extrinsic preferences are driven by expected external rewards, while intrinsic preferences are not. Yet if intrinsic preferences are influenced by contextual factors, don’t these factors also act as external rewards? If a certain context shapes a person’s utility function because it reflects how well her choices align with the norms of the social identities she accepts for herself, doesn’t that alignment consideration constitute or function as an expected external reward? More needs to be said here to stabilize this intrinsic/extrinsic distinction for the purpose of modeling motivational change.

14Kreps is more persuasive in countering the pragmatic view that economics should focus on areas where motivational change is irrelevant, except when driven by changes in information. He offers numerous examples of economic phenomena that challenge such self-imposed limitations: Can stock options make employees internalize the welfare of their employers? Can intrinsic motivation in nurses improve both patient care and nurse satisfaction? What company cultures foster intrinsic motivation among employees, and are these conducive to economic success? Do monetary incentives crowd out socially responsible behavior in both public and for-profit organizations? Can intrinsic and extrinsic rewards align, and if so, how do they need to be designed to avoid crowding out? Kreps raises many such questions, and the solutions he gestures to always involve models of intrinsic preference change. Those who deny the relevance of these issues to economics will be hard pressed, I believe, by Kreps’ points.

15And yet I find that spotting the hole is easier than plugging it. Kreps thinks that by incorporating social psychology theory into economics, he can enhance economic modeling in the relevant directions. Yet mainstream economics has some pretty deep methodological commitments. One is the restriction to incentive-aligned choice data. Another related one is to conceptualize preferences as nothing but revealed by choice. As long as one sticks with these commitments, there is no chance to empirically support the various distinctions Kreps proposes.

16For one thing, if one accepts the standard belief-desire account of action as well as the changeability both of beliefs and desires, then for every observed choice, there is an infinite set of possible belief-desire combinations revealed by that choice. This makes changing beliefs and preferences empirically underdetermined within a revealed preference framework.

17Moreover, in terms of choice, there is no discernible difference between doing things for intrinsic or extrinsic reasons. Of course, one can seek or design situations that exclude reciprocity, public observation, or external rewards. But what matters is individuals’ beliefs (and even implicit beliefs) about such situations, not any objective feature. As long as I believe that my norm-abiding behavior will be reciprocated (however ludicrous that belief might be), I will be extrinsically motivated to abide by the norm. Revealed preference approaches can’t capture this distinction. Kreps hints at this in a few places, at one point admitting that “we must be comfortable thinking of her ‘expressing’ preferences over this richer space” (25). But how this is to be done, and how economists should overcome their often justified reservations about using expressed preferences as data for testing their models, remains unsaid.

18Instead, Kreps points to another, but equally mysterious source of empirical data: “casual empiricism”. In the way Kreps puts it, casual empiricism can either be unsystematic small sampling, or drawing on intuition and “common sense” (148), or both. Pointing to the role of intuition in economic theorizing of course has a long tradition, ranging from Smith through Keynes to Simon. But those authors referred to intuition as a contributor of theory building—Kreps in contrast seems to see it as a source of theory justification, placing it often in direct connection to more systematic empirical evidence gained through laboratory and field experiments. This is unfortunate. The models discussed in this book cannot be empirically justified under the standard methodological commitments of current economic science. Perhaps this should lead to a change of these commitments; perhaps it means that these models must remain unjustified. Yet “casual empiricism” is not an alternative empirical justification. It lacks the systematicity, transparency and reproducibility of genuine empirical evidence, and thus is just a placeholder for what we wish we had but don’t. Invoking it just muddles the methodological and conceptual problems that still beset modeling preference change.

19To conclude, Kreps’ Ar­guing About Tastes is an engaging and thought-provoking attempt to rouse his economist colleagues from their dogmatic non est disputandum slumber. While it highlights numerous economic phenomena that seem to require an explanation in terms of preference change, the book ultimately falls short of addressing the conceptual and methodological issues that have long plagued efforts to model such changes.

Haut de page

Pour citer cet article

Référence papier

Till Grüne-Yanoff, « David M. Kreps, Ar­guing About Tastes »Œconomia, 15-4 | 2025, 679-684.

Référence électronique

Till Grüne-Yanoff, « David M. Kreps, Ar­guing About Tastes »Œconomia [En ligne], 15-4 | 2025, mis en ligne le 01 décembre 2025, consulté le 17 janvier 2026. URL : http://journals.openedition.org/oeconomia/19623 ; DOI : https://doi.org/10.4000/15ead

Haut de page

Auteur

Till Grüne-Yanoff

Division of Philosophy, KTH Royal Institute of Technology. gryne@kth.se

Haut de page

Droits d’auteur

CC-BY-NC-ND-4.0

Le texte seul est utilisable sous licence CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont susceptibles d’être soumis à des autorisations d’usage spécifiques.

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search