Three Questions About the Economics of Relative Position: A Response to Frank and Sunstein

Abstract

Cost-benefit analyses typically ignore the importance of relative position. That is, they do not take into account the possibility that people value particular goods, services, or other determinants of well-being through comparisons with others. Robert Frank and Cass Sunstein have recently concluded that taking into account positional issues implies that the benefits of health and safety regulations may be twice as large as the levels commonly found in cost-benefit analyses. However, the effects of positional externalities on the valuation of safety and health regulations, and hence the correct modifications to cost-benefit analyses, are theoretically ambiguous. Frank and Sunstein assume that people like others to become worse off and that the incomes of others are more important for comparison than their health and safety on the job. Because different assumptions can lead to opposite conclusions about the value of additional regulations, this response addresses whether the evidence supports Frank and Sunstein's assumptions. The nature of relative position can be described as answers to three questions. First, what is the relevant group to which people compare themselves? Second, which characteristics of the comparison group matter? Third, how strongly do these comparisons affect people? This paper evaluates Frank and Sunstein's evidence on all three questions. People inclined to favor the model of positional externalities espoused by Frank and Sunstein may find their evidence convincing, but there are appealing alternative explanations. There is also direct evidence that only a minority of people act in the way they assume. They estimate that people should be willing to spend $6,000 of a $10,000 raise to prevent their coworkers from getting the same raise. Though some people might pay to reduce the salaries of their coworkers, others would surely pay to raise them. That many people display altruistic behavior in many situations is completely ignored by Frank and Sunstein. Particularly because of the variation in preferences across individuals, the evidence is at present too limited to permit the precise characterization necessary to evaluate the effects of policy. Under some plausible models, the policies suggested by Frank and Sunstein make people worse off; in others, better off. Though it is premature to incorporate positional externalities in policy analysis, such issues undoubtedly will become a formal part of policy analysis as our understanding improves.