What price the moral high ground?

• Author: Frank, Robert H.

1. Introduction

   Are people fundamentally selfish? In textbook accounts of rational choice, economists generally take an agnostic position on this question: a person's tastes are her own business, and rationality requires merely that she act efficiently in pursuit of whatever preferences she happens to hold. This approach affords obvious flexibility. A person's anonymous gift to charity, for example, can be readily explained as the consequence of a taste for helping others.

   But this flexibility comes at the expense of what George Stigler might have called the "crank-case oil" problem. If we see a person drink the used crankcase oil from his car, and he then writhes in agony and dies, we can assert that he must have really liked crankcase oil. (Why else would he have drunk it?) Virtually any behavior, no matter how bizarre, can be "explained" after the fact by simply assuming a taste for it.

   With this difficulty in mind, many economists assume - implicitly or explicitly - that people are essentially self-interested. Thus, in Gordon Tullock's words, "the average human being is about 95 percent selfish in the narrow sense of the term" [21; 12, 12]. This approach has generated many powerful insights into human behavior. It explains, for example, why car pools form in the wake of increases in gasoline prices and why the members of "service" organizations are more likely to be real estate salespersons, dentists, chiropractors, insurance agents, and others with something to sell than to be postal employees or airline pilots.

   But the assumption of selfishness is not without drawbacks of its own. The most apparent is that, for every behavior that is consistent with this assumption, there seems to be another that contradicts it. Travelers on interstate highways leave tips for waitresses they will never see again. People vote in presidential elections and they walk away from profitable transactions whose terms they believe to be "unfair." Soldiers throw their bodies atop live grenades to save the lives of their comrades. In these and countless other ways, people do not seem to be maximizing utility functions of the egoistic sort.

   A second difficulty with the selfishness assumption is that our models of human behavior appear to mold the behavior of both the modelers themselves and those they teach. Thus, Gilovich, Regan and I found evidence that people with extensive training in economics are less likely than others to cooperate in social dilemmas, and that this difference is at least in part a consequence of their training.(1)

   In the end, the importance of other-regarding preferences must be assessed empirically. In this paper I employ the time-honored tradition of looking to the labor market for evidence on preferences. Just as wage differentials across jobs that carry different levels of risk tell us something about worker's preferences regarding safety, I will argue that wage differentials in other settings shed light on the strength of people's concerns about others. Indeed, I will argue that, for a large sample of college graduates, even crude measures of moral satisfaction on the job do more to explain wage differences than do the human capital variables traditionally used for this purpose. An second intriguing finding is that measures of moral satisfaction also seem to eliminate unexplained wage differentials between men and women. Women are more likely than men to choose jobs that afford high measures of moral satisfaction, but men who choose such jobs pay precisely the same wage penalty that women do.

2. Evidence from Existing Studies

   My survey of the existing literature uncovered several sources of evidence - mostly indirect or qualitative - in favor of the hypothesis that moral concerns figure prominently in career decisions. One early study presented the results of a national cross-section survey of the factors that influence the career choices of college students [17]. Participants in the survey were asked to "consider to what extent a job or career would have to satisfy each of [the following] requirements before you would consider it IDEAL" [17, 12].

   1. Provide an opportunity to use my special abilities or aptitudes.

   2. Provide me with a chance to earn a great deal of money.

   3. Give me a chance to exercise leadership.

   4. Give me social status and prestige.

   5. Give me an opportunity to work with people rather than things.

   6. Enable me to look forward to a stable, secure future.

   7. Leave me relatively free of supervision by others.

   8. Permit me to be creative and original.

   9. Provide me with adventure.

   10. Give me an opportunity to be helpful to others.

   The most highly ranked of these ten career values was "provide an opportunity to use my special abilities or aptitudes," which was rated as "highly important" by 78 percent of respondents. Next came "enable me to look forward to a stable, secure future" (61 percent), followed by "permit me to be creative and original" (48 percent). The career value of special interest for present purposes - "give me an opportunity to be helpful to others" - was ranked fourth on the list of ten. It was rated as highly important by 43 percent of respondents, and another 44 percent called it of "medium importance." Only 13 percent said it was of "little or no importance, irrelevant, or distasteful." Taken at face value, these responses suggest that an overwhelming majority of college students would be willing to trade pay or other desired working conditions in exchange for additional opportunities to help others on the job. These responses, of course, tell us nothing about the rate at which students would be willing to make such tradeoffs.

   A second study asked business executives about the effect of their firm's ethical conduct on its profitability [2]. Several respondents said that their firm's ethical posture affected its ability to recruit in the labor market. For instance, an executive of a retail firm whose founder had created a strong reputation for ethical dealing offered this description [2, 105].

   The top executive has created the kind of environment he wants; it is calculated, but it is good and promotes good ethics. Furthermore, our customers . . . like the courteous and kind treatment given by our sales people. Now we get sales ladies from the same pool of personnel that other retail stores in the area get their help, but our help is more courteous because of the reputation which the president has established. New sales ladies catch on; their better side shows through in their work And we don't pay any more than other stores around here either - probably we pay a little less. (Emphasis added.)

   Other executives reported that a firm's 'reputation for unethical dealings - even when based on behavior in the distant past - made recruiting difficult. As one railroad executive put it, "We still have people around who remember the old robber barons . . . These men were very unethical at the time and the result of their operation still hurts us today" [2, 107]. The same sentiment was echoed by an electrical industry executive in reference to that industry's highly publicized price-fixing scandal: "The students really quiz us since the indictments. They want to know whether collusion has stopped and what the company policy is" [2, 107]. Again, while such responses are suggestive, they tell us little about the rate at which workers are willing to trade pay for a working environment that emphasizes social responsibility.

   In a recent study, Scott Vitell and D. L. Davis [22] surveyed managers in the management information systems (MIS) field to discover the extent to which the ethical posture of their employers was related to job satisfaction. Job satisfaction was assessed using the Managerial Job Satisfaction Questionnaire developed by Cellucci and De Vries [4]. The ethical posture of the respondents' employers was measured by the extent to which respondents agreed with the statement, "MIS managers in my company often engage in behavior that I consider to be unethical." The correlation coefficient between responses to this statement and the index of job, satisfaction was -0.26, which is statistically significant at the 0.05 level.

   As in the case of the other studies reported, the Vitell and Davis finding suggests a willingness on the part of employees to trade pay and other valued working conditions in exchange for a more ethical working environment. But here again the authors have no basis upon which to estimate the rate at which such job characteristics might be exchanged. This issue is the focus of the remaining sections of this paper.

3. The Theoretical Framework for the Current Study

   Jobs differ in countless dimensions, one of which is the degree to which the worker contributes to the well-being of others. Consider two jobs identical along all dimensions except this one. (For example, one job might involve writing advertising copy for a product known to cause serious health problems, while the other involves writing advertising copy for the United Way.) If people derive satisfaction from engaging in altruistic behavior, it follows that if the wages in these two jobs were the same, there would be an excess supply of applicants to the second job, a shortage of applicants to the first. In equilibrium, we would therefore expect a compensating wage premium for the less altruistic job. A job applicant can occupy the moral high ground if he wants to, but only by accepting lower wages. This wage differential serves as a measure of the strength of the particular nonegoistic motive involved.

   More formally, suppose we modify the traditional utility function to be of the form

   U = U(X, SR), (1)

   where X is monetary income and SR the degree of social responsibility attributed to the job. In Figure 1, the curves [I.sub.1], [I.sub.2], and [I.sub.3] are part of the indifference map to which this utility function gives rise. Let M stand for ability, the endowment with which agents purchase...