Reviewed by Michael Klenk
Material incentives, such as taxes, fines, or subsidies, are, in economic orthodoxy, considered to be the proper way to deal with market failures and incomplete contracts. For instance, a firm cannot specify in an enforceable contract that an employee will work as hard as he can because there is always room for defection. However, the firm can offer a material incentive, such as a bonus for good performance, to tap into the employee’s self-interest to achieve the desired outcome. Similar assumptions about the determinants of behaviour and the right way to influence it, namely by relying on the material self-interest of individuals, dominate practically all areas of social policy design. The complexity of the task of achieving large-scale coordination, as is required in firms or societies, perhaps forced this myopic view on the motivations and preferences of human beings. It would be profoundly complex to tailor institutions to the actual preferences of subjects, and dangerously naive to count on their civic-mindedness. Despite obvious problems with this approach, one should remember the global financial crisis, in which the pursuit of self-interest broke the system; there is no clear alternative to-date. Those concerned with the design of economic incentives keep trying to figure out the right mechanism, and the proper setup, that will make the self-interest of individuals entice them to act for the greater good.
In The Moral Economy, Samuel Bowles offers a brilliant criticism of the self-interest-centred approach to policy design, and proposes, as an alternative, a synthetic approach, that takes into account both self-interest, and the evidently strong “social preferences” of most human beings (26). His main thesis is that that assumption of the separability of incentive and social preferences is false: behavioural experiments show that economic incentives influence what we value and, in particular, they affect our social preferences, often adversely.
A simplified example helps to clarify Bowles’ thesis. Assume that the people of Modernville value punctuality as a social or moral duty. We can think of the value citizens place on punctuality as a determinant of their behaviour. The mayor of Modernville, however, is dissatisfied; too many citizens are late for their appointments at the town hall, and he imposes a hefty fine for latecomers. In doing so, he effectively appeals to the citizens’ self-interest, which we can see as another determinant of their behaviour. We got two determinants of behaviour now: values or social preferences (’It is my duty to be punctual!’) and self-interest (’I don’t want to pay the fine!’). In line with economic orthodoxy, the mayor thinks that both determinants are separable: appealing to the citizen’s self-interest will not affect their social preferences. Hence he can both impose the fine on lateness and live amongst people who still value punctuality for its own sake. Not so, says Bowles. What people value is affected, often adversely, by incentives that appeal to their self-interest and this, he argues, calls for a new paradigm in policy design. Material incentives are certainly required for a well-governed society, but they must be used to enhance values, rather than in abstraction of them.
The first two chapters contain, respectively, a brief summary of the book’s main argument and an illuminating account of the intellectual history of social policy design insofar as it is relevant to Bowles’ argument. Early views on ‘good government’, such as that of Aristotle, considered “inculcating good habits in citizens” as central to the task (12). Later developments, however, increasingly focused on providing proper laws and institutions, in abstraction of the actual preferences of subjects (11-13). Of course, as Bowles readily admits, this shift of focus is sensible to coordinate the behaviour of a multitude of individuals with differing goals and motivations (65). Relying (solely) on the goodwill of individuals would be unmanageable and unlikely to yield efficient outcomes. However, he criticises that belief in the coordinating power of material self-interest went too far (22). Not only are most contracts incomplete (33), as many economists would admit, but material incentives cannot fix the problem because the crucial assumption of separability is false (23).
Bowles elaborates his argument for nonseparability in chapter three, where he provides a conceptual description, as well as empirical evidence for its occurrence in laboratory behavioural experiments featuring economic games. Two determinants of behaviour are not separable if “social preferences” (i.e. behavioural motives that are not consistent with maximising one’s own material self-interest) are affected by the incentive that appeals to self-interest (47). For example, the effect of the fine introduced in Modernville is not separable from the social preferences of citizens (’it is my duty to be punctual!’) if the behavioural effects of the fine cannot simply be added to the behavioural effects of social preferences (50). Bowles distinguishes three forms of non-separability. The mere presence of an incentive may affect social preferences (categorical crowding out), an incentive may affect social preferences depending on its extent (marginal crowding out), and, finally, values might offset the effects of incentives and reduce net contributions to a public good (strong crowding out, 51).
A typical experiment used by Bowles tests the ‘social preferences’ of subjects by comparing actual contributions in economic games with predicted contributions of (imagined) agents that are purely self-interested. If the actual contributions are higher than what a purely self-interested agent would contribute, Bowles infers that subjects exhibit ‘social preferences’ (42). A second stage of the experiment introduces a material incentive (to one group of subjects) and compares actual contributions either to contributions predicted of a purely self-interested subject or to non-incentivized contributions of subjects in a control group. In typical crowding-out cases, Bowles finds that actual contributions of incentivized subjects are less than those of non-incentivized subjects or what would be expected of a self-interested subject.
In line with his plea to cast a light into the black box of actual human preferences in economics, Bowles traces the “cognitive processes” (77) that account for the nonseparability of material interests and moral sentiments in chapter four. He proposes four different explanations of the phenomenon of nonseparability, all of which portray “incentives as information” that frame the decisions of individuals, and act as a “guide to their behaviour” (85). Showing acute sensitivity to different cases, Bowles distinguishes incentives that function like “bad news” for individuals, suggesting something about the intention of the designer of the incentive, or the nature of the task to be done (86). Jncentives may cause subjects to “switch their morality off”, because the incentive signals that the situation calls for materialistic rather than moralistic deliberation (96). Moreover, finally, Bowles finds that some incentives are best explained as giving away an unwelcome attempt to control the subject (103). A brief reflection on the relation between cognitive processing styles (i.e. affective vs. deliberative) and preferences types, which Bowles considers to have “no simple mapping” (110), concludes Bowles’ cognitive analysis of the problem of nonseparability.
Chapter five addresses a fundamental objection; why do subjects in capitalist societies, such as Denmark or the UK, who are confronted with abundant material incentives, still show strong social preferences in experimental settings and related case studies? After all, Marx’s prediction of “universal venality” does not accurately describe these societies, but if Bowles were correct, we should expect the crowding out of social preferences. Bowles’ answer, arrived at through intercultural comparison of experimental evidence, is that the rule of law, and other properties of liberal societies, rather than markets, positively influences civic preferences. The data he refers to shows that subjects from less liberal societies indeed take fewer chances in economic games – they behave less cooperatively. Bowles, however, resists the often heralded interpretation of the ‘civilising effect of the market’. The difference is better explained, he argues, by pointing to differences in institutional settings. Citizens that are protected from worst-case outcomes when engaging in social interactions, for instance, through laws that reliably protect them from loss of property or personal injury, have learnt to take chances at cooperating, even with complete strangers, and have thus grown accustomed to ‘take changes’ in cooperative settings (145). Hence, Bowles argues, the properties of liberal civic cultures often lead to crowding in and subjects act more cooperatively than what we should expect of purely self-interested agents (150). This is good news, argues Bowles, for it shows how material incentives and social preferences could work in tandem.
Bowles addresses the repercussions of his findings in chapter six. Material incentives are necessary, but not sufficient for a well-governed society, and, as Bowles emphasises in a discussion of the field of mechanism design, designing better incentive systems aimed at self-interested individuals without paying attention to the actual preferences of individuals is very unlikely to be successful on a broad scale (161-6). That leaves legislators with a dilemma; we need material incentives as most contracts are incomplete, yet using them unwittingly, without paying attention to the actual social preferences of subjects, threatens to diminish the basis of our liberal societies (186, 202). In other words, we often need the ‘carrot and the stick’, but using it the wrong way may turn even good-willing mules into the most stubborn, selfish beasts.
The final chapter is an emphatic “mandate” (187) for solving the legislator’s dilemma. Bowles revisits some of the experiments and case studies he discussed earlier in the book and finds that incentives and social preferences are an ungainly mix when incentives were not accompanied by a “moral message” (203). In line with his analysis of material incentives as information, he recommends that the possible ‘this-is-a-scenario-for-egoists’ message of incentives must be offset by a signal that shows emphasises the constitutive social and moral aspects of the relevant situation (220). The mayor of Modernville would thus be well advised to add to his fine on lateness a public campaign praising punctuality.
Bowles’ insights about the nonseparability of self-interest and values, and the implications that follow, are intuitively compelling; who is not in favour of treating others nicely, and who would not value a bit more camaraderie and trust in the business world? Moreover, who does not lament that this is often not possible, given the ‘forces’ of the market? Bowles’ admirable contribution is to provide the empirical evidence and theoretical analysis to show that such pleas to a more moral economic environment are quite right in an important sense and far from ill-advised romanticism.
Bowles’ argument has interesting implications for the normative debate about the responsibilities of businesses. The question is whether the normative and moral responsibilities of businesses extend beyond the pursuit of their material self-interest. Call an affirmative answer to this question an ‘inclusive’ view and a negative answer an ‘exclusive’ view. Note that this is a normative, rather than a descriptive question; we want to know what businesses ought to do, rather than what they, in fact, do. To answer the question, we might determine the likely outcomes of either alternative. Bowles argument suggests that outcomes are sub-optimal, on many plausible ways of measuring outcomes, if we opt for exclusive views. Alternatively, we might dig deeper and assess the normative principles that would support inclusive and exclusive views in the first place. Here, the interesting implication of Bowles’ argument is that it takes away much of the support for advocating outcome-based arguments in favour of exclusive views and it is hard to see how exclusive views could otherwise be supported. This leaves proponents of exclusive views in business ethics with a puzzle: opt for a position reminiscent of egoism, or accept an inclusive view about the social responsibilities of businesses.
‘The Moral Economy’ is a compelling, deep, and carefully crafted plea to pay attention to the social preferences of people when designing institutions. Although Bowles’ practical policy recommendations lack some of the thoroughness of his problem analysis, his book is required reading for anyone who wishes to understand the difficulty of orchestrating mutually beneficial outcomes in groups.
25 January 2017