Adam Smith, in his Theory of Moral Sentiments, considered how an individual's moral sentiments affect their acts in the world in ways that are contrary to self-interest. Smith's first sentence sets up his thinking,

How selfish soever man may be supposed, there are evidently some principles in his nature, which interest him in the fortune of others, and render their happiness necessary to him, though he derives nothing from it except the pleasure of seeing it.
(Smith, 1790: I.I.1)

Let me be up front and acknowledge that I am only now reading The Theory of Moral Sentiments (I am currently a graduate student). Nevertheless, this first quote is intended to give a frame for the paper by Sam Bowles, who, in the title of his paper, 'Policies designed for Self-Interested Citizens may Undermine the "Moral Sentiments" makes an obvious allusion to this text. I cover Bowles's main theses and offer some additional points.

Sam Bowles's recent paper (gated) in Science focuses on exactly these "moral" imperatives in humans, the likes of which, he argues, have remained neglected in mainstream economics during the past half century. Bowles reviews economic experiments carried out all over the world, from Bogotà, Colombia to Zurich, Switzerland. His unifying theme is that individual "moral sentiments" matter and the self-interest of neoclassical economics is not as all-encompassing, or welfare promoting, as one might be led to believe. He argues instead for the acknowledgment of other-regarding preferences that operate in the economic realm and affect how individuals interact. At the core of his argument is the intuition that incentives that target individuals who are assumed to be solely self-interested will often fail because of this misunderstanding of human nature.

Bowles opens the paper with the example of the, now infamous, Haifa daycare centres in which the centres implemented a fine for late arrival of parents. Following this the incidence of lateness increased and did not return to the pre-fine levels after the fine policy was rescinded /Gneezy and Rustichini, 2000).

Bowles defines several ideas, using them as a basis to motivate a departure from self-interested behaviour. They are:

1. Framing
2. Endogenous preferences
3. Self-determination
   
4. Information conveyed by incentives
   

He argues that "incentives are part of a decision situation and may signal appropriate behaviour" (1606). In the Haifa daycare centre example, individuals did not engage in market interactions originally they would have experience 'shame' for arriving late. However, it became a market interaction in which 'lateness' was a consumer product that parents could pay for. Thus, the centres had 'framed' the policy such that it became a market interaction. Daniel Kahneman and Amos Tversky championed the understanding of framing, Kahneman was awarded the Nobel Memorial Prize in Economics for this and other work. Additionally, Bowles refers to work in Colombia by Juan-Camilo Cardenas which replicates a framing problem in public goods games. He contrasts 'communication' and 'government regulation' frames. It turned out that individuals, when 'regulated', were more likely to behave self-interestedly than in either the control or the 'communication' treatment. Cooperation broke down as a consequence. Hence, incentives, and the mechanisms that we use to implement them, need to be structured such that we are aware of the frames that they present, else behaviour could go in unpredictable directions.

Endogenous preferences are preferences that have been altered as a consequence of the experience of some state. The intuition here is that the experience of some specific incentive structure could permanently change the preferences of the individual. Public goods experiments revealed that individuals, who previously acted cooperatively in public goods games, begin to act self-interestedly after the application of a treatment and continue to behave self-interestedly after the treatment has ended. Hence, Bowles proposes that, "economies structures by differing incentives are likely to produce people with differing preferences" (1607).

Self-determination, or "control aversion", was one reason I found to be one of the most interesting reponses to incentive structures. In principal-agent experiments (similar to employer-employee relationships) individual agents selected a level of production that was non-zero even though their payoff was highest when they 'produced' nothing. The principal could choose to leave the agent free to choose an effort level, or impose a lower bound on effort. By standard economic theory the first treatment should produce zero production, the second treatment - production at the lower bound. Production was higher in the 'free' control experiment than it was in the 'bounded' treatment. It seems as though humans value autonomy highly and are willing to act against individuals who retract autonomy. Note however that it does not seem (by Bowles account) as though the experimenters attempted the counter-example, a situation when there would have been a fixed lower bound, which was then rescinded and allowed the principal to choose to impose a lower bound or not. The results of this could provide an interesting contrast to those above.

Lastly, in reference to information conveyed by incentives, Bowles describes a 'trust' investment game, in which pairs of individuals act as investors and trustees. The investors could choose to 1) 'trust' the individual for back payments, 2) propose a back payment level and choose to impose a fine if the back-transfers was below this level, or 3) propose a back payment level and drop the option to use the fine. The situation in which the fine was available, but dropped as an option by the principal resulted in the highest levels of back transfers, the straight fine treatment resulted in the lowest level of back transfers. Bowles segues into a brief discussion of social preferences, I couldn't find a definition in the paper, so I went to his textbook, Microeconomics, wherein he defines social preferences as follows, "in choosing to act, individuals commonly take account not only of the consequences of their actions for themselves but for others as well" (Bowles, 2006: 96).Typically social preferences include reciprocity motives and altruism. Hence, he establishes that, "Incentives imposed by peers who do not stand to benefit personally do not social preferences and are often synergistic with them" (1608). For this he brings support from experiments in public goods games in which punishment curtails self-interested behaviour and reinforces cooperation.

At this juncture I think it pertinent to introduce evidence from Herrmann et al's (2008) paper on antisocial punishment across societies. They introduce the 'dark side' of punishment, showing that individuals who contribute 'too much' are punished (hence the term 'antisocial punishment') in certain cultures (I blogged on this previously here). Bowles briefly refers to Herrman et al.'s paper, but I think that the social preferences involved here are definitely as exciting as those discussed by Bowles generally, and, in fact, provide further evidence for the dynamic nature of social norms. With the international evidence that they provide (experiments with fifteen societies across the globe) there are dramatically different norms when it comes to punishment. Spitefulness and angry retaliation are as much a part of the human emotional arsenal as the beneficial and prosocial preferences are. It is important therefore that these antisocial preferences are also taken into account when policymaking: when deciding on incentive structures.

Bowles summarizes a few points on problems of efficiency and self-interest. If a principal is able to introduce an incentive structure and increase their own payoffs, then they will do so, regardless of whether the entire surplus diminishes as a consequence. Contrary to commonly held views, individuals interacting together who are not self-interested can produce a higher total surplus (through, say cooperating in a public goods game). Thus, although agents respond particularly negatively to certain incentive structures, some self-interested profit-maximising principals will introduce these incentive structures to maximise their own payoffs, while the pie itself diminishes is size. This is not the kind of world many of us woud like to work in, most economists would promote the 'efficient solution'.

Bowles maintains then that moral sentiments and material interests are 'nonseparable'. The problem being that behaviour is not only 'acquisitive', but 'constitutive'. He argues furthermore that these interests can act synergistically, especially in instances where there is substantive rule of law and other strong institutional frameworks. Individuals supposedly wish to be cooperative, but also want to ensure that they are not the only cooperator in a world of defectors. Their cooperation is conditional. Contingent on strong rule of law, individuals' social preferences are maintained and they uphold cooperative behaviour.

Bowles's 'take home message' is that policy makers need to take consideration of the experimental evidence that is presented to them. They should be aware of the prosocial behaviours that policies to promote self-interest might undermine, else the total social surplus might diminish: unintended consequences have been the undoing of many a well-intended policy. The good news though for market-minded individuals is that good institutions can act synergistically with social preferences to produce good consequences from the market. It seems as though individuals who "interact in market-oriented economies are more likely to be fair-minded" (1607). Capitalism has not lost the day, it simply needs to rethink some of its economic models for policy-making.


References:
Smith, Adam. The Theory of Moral Sentiments. Library of Economics and Liberty. Retrieved July 12, 2008 from the World Wide Web:http://www.econlib.org/Library/Smith/smMS1.html

Bowles, S. (2006), Microeconomics: Behavior, Institutions and Evolution, Russell Sage Foundation and Princeton University Press, New York, US.



Bowles, S. (2008). Policies Designed for Self-Interested Citizens May Undermine "The Moral Sentiments": Evidence from Economic Experiments. Science, 320(5883), 1605-1609. DOI: 10.1126/science.1152110



Gneezy, U., Rustichini, A. (2000). A Fine Is a Price. The Journal of Legal Studies, 29(1), 1.

Herrmann, B., Thoni, C., Gachter, S. (2008). Antisocial Punishment Across Societies. Science, 319(5868), 1362-1367. DOI: 10.1126/science.1153808

______
Aside: Two comments on Bowles' article are offered by Gavin Kennedy here and here. Another in reason magazine here.