For more than one hundred years, economists have discussed the concept of positional goods—that the utility conferred by many, perhaps even most, goods depends not only on the amount the individual consumes, but also on the amount others consume (e.g., Thorstein Veblen, 1899; James Duesenberry, 1949; John K. Galbraith, 1958; Robert Frank, 1985). While economic interest in “positional goods” is increasing (e.g., Mark Pingle and Mike Mitchell, 2002; Kenneth Arrow et al., 2004; Ed Hopkins and Tatiana Kornienko, 2004), the literature remains largely theoretical rather than empirical. If status concerns affect all items in the utility function equally (leisure as well as goods), the positional effect would operate like a lump-sum tax, reducing well-being without changing the allocation of time or money (Arrow et al., 2004). However, if positional concerns are stronger for some things than for others, then in order to understand how people can become better off in well-being, not simply in wealth, we must investigate how interpersonal competition interacts with material gains (Frank, 1997). Evidence about what goods are more positional than others is essential for correct policy recommendations (Gregory Besharov, 2002), but little is actually known about the relative positional rankings of items in the typical consumer’s utility function. Four hypotheses proposed in the literature are:

We welcome Bradley J. Ruffle and Orit Tykocinski’s (1999) attempt to reconcile the difference in findings between our paper (Solnick and Hemenway, 1996) and Joel Waldfogel’s (1993). In both studies, subjects were asked to estimate the cost and the value to them of gifts they had received. Waldfogel found that gift recipients typically valued their gifts at only 87 percent of their estimated cost to the givers. By contrast, our sample of adults—from train stations and airports and from faculty, staff, and graduate students at Harvard—reported valuing their gifts at more than twice the estimated cost. Ruffle and Tykocinski examine the wording of the question used to elicit the valuation of the gift. Waldfogel’s subjects were economics students; hence he could ask them directly what amount of cash would make them “indifferent” between the gift and a cash alternative. Since we wanted to question people without economics training, we asked respondents to name the amount of money that would make them “equally happy” as receiving the gift. Ruffle and Tykocinski find that the change from Waldfogel’s wording to ours results in a 15to 20-percent increase in estimated value while leaving accompanying cost estimates unchanged. The wording of the question clearly matters, but to pose a question that only those with training in economics can answer limits the scope of inquiry. In the study by Ruffle and Tykocinski, 10 percent of the psychology students raised their hands to ask for a definition of “indifferent.” We wonder how this concept was explained to them, what those who asked and those who did not ask made of the term, and what set of emotions might have been activated. Had Waldfogel used our question, rather than finding any deadweight loss (a yield of 87 percent), he would have reported gifts being valued very close to cost (100.0 or 104.4 percent). Thus Ruffle and Tykocinski take us partway towards an understanding of the difference in our findings and Waldfogel’s. Yet a wide gulf remains. Ruffle and Tykocinski reject our claim that sample selection may be largely responsible for the difference between our results and Waldfogel’s. It does appear that results do not vary a great deal with the type of undergraduate questioned—whether a student is enrolled in a psychology or an economics class, or in a more or less prestigious institution. But the real issue of sample selection has not been adequately addressed; ours remains the only study to use adults, living independently, as subjects. We note that Ruffle and Tykocinski’s subjects performed poorly in estimating costs. For example, psychology students estimated the cost of the mask at 603 shekels and economics students, at 375 shekels. Actual retail price was 800 shekels. This incidental finding supports our belief that college students may not be full-fledged participants in many market or gift-exchange processes. It is certainly possible that the average gain in gift exchange may be lower than the 214 percent measured in our survey, but we remain convinced that even examining solely the material value to the recipient, gift-giving can, and often does, result in a gain rather than a deadweight loss.