Samuel Bowles and Herbert Gintis make a compelling case that the calculating, narrowly self-interested Homo economicus is, if not exactly a fiction, a misleadingly incomplete understanding of human behavior. In laboratory experiments, they show, subjects are willing to incur a personal cost in order to achieve a more equitable overall distribution of rewards or punish a participant who is perceived to have acted unfairly. But what about the "real world"? Do most Americans really support egalitarian social policies, even when those policies conflict with their personal economic well-being?

The enormous popularity of Social Security and the unpopularity of welfare appear to have led most politicians, commentators, and academics to feel that middle-class Americans will only support welfare state programs from which they themselves benefit. Theda Skocpol and William Julius Wilson, for example, argue that means-tested anti-poverty programs are politically weak and that more can be done to help the poor through universal social programs that provide benefits to both poor and non-poor Americans.

Bowles and Gintis are on solid ground in rejecting this self-interested model of mass politics. First, as they report, polls show that most Americans think government should be doing more to help the poor, and a majority of the public expresses a willingness to pay higher taxes to fight poverty. Furthermore, when we compare public attitudes toward comparable means-tested and universal social programs, we find little evidence of greater public support for the latter. For example: Supplemental Security Income (SSI) and Medicaid are means-tested programs that benefit only the poor, while Social Security and Medicare provide similar benefits to Americans of all income levels. Yet surveys show equal levels of support for the two retirement/disability programs, and nearly equal levels for the two medical programs. (These findings and the survey evidence cited below are contained in my forthcoming book, Why Americans Hate Welfare.)

The changing tides of public opposition to welfare also conflict with the "middle-class self-interest" account of welfare politics. Both Wilson and Lester Thurow, for example, argue that Americans may be willing to share with the less fortunate, but only when the economic pie is growing. Recessions, this argument goes, stimulate the economic "self-preservation instincts" of the middle class. But survey data show just the opposite: when the economy turns sour, the public becomes more supportive of welfare spending and more sympathetic toward the poor. This pattern suggests that welfare politics turn not on judgments of economic self-interest, but on perceptions of the level of need or deservingness of welfare recipients.

Support for welfare ebbs and flows over time, but these changes represent only degrees of unpopularity. Bowles and Gintis believe that their model of Homo reciprocans can explain both the public's desire to help the poor and its opposition to welfare. I'm not so sure. "Egalitarian policies that reward people independent of whether and how much they contribute to society," they write, "are considered unfair and are not supported, even if the intended recipients are otherwise worthy of support." But public support hinges more on perceptions of recipients' need than on perceptions of their social contributions in any concrete sense–and the neediest recipients are typically those who are least able to contribute. (On the role of perceived need in shaping support for various social programs, see Fay Cook and Edith Barrett's Support for the American Welfare State.)

Bowles and Gintis's Homo reciprocans also has a dark side that is hard to reconcile with surveys of public attitudes. Although welfare is unpopular, and a majority of Americans think most welfare recipients could get along without it if they tried, there is little evidence that the public seeks to punish people on welfare. True, the polls do show strong support for welfare time limits and work requirements, but support is just as high–and in many cases higher– for increased spending to help welfare recipients become self-supporting. Large majorities of Americans want their government to provide welfare recipients (and poor people in general) with greater access to education, job training, childcare, medical care, housing, and transportation to work. Nor can this support be attributed to "enlightened self-interest." While some of these programs might reduce welfare costs in the long run (e.g., education or job training) others (medical care, housing assistance) would appear to benefit the poor with little prospect of any future "payoff" to the taxpayer.

As Bowles and Gintis point out, few Americans support any strong form of equality with regard to economic outcomes. Indeed, a certain degree of economic inequality is viewed as a natural and healthy condition in a society in which each is free to pursue his or her own version of the good life. But Americans are troubled by the growing inequality of the past 25 years, and they support measures to aid its victims–for example, the Earned Income Tax Credit and increases in the minimum wage. Equality is hardly passé, but neither is it the animating force behind public support for anti-poverty programs. I think Bowles and Gintis are on the right track with "basic needs generosity," which itself may be just one example of a broader humanitarian inclination to help those who suffer through no fault of their own.