President Clinton says that we need to “end welfare as we know it.” The current welfare system is such a mess that it is hard to disagree. The difficult question is: What should take its place?

Though reform is not likely this year, there are already as many answers to that question as there are experts on training, drop-out rates, teen pregnancy, community development, and persistent poverty. But welfare reform is not simply a matter for policy experts. It raises fundamental issues about jobs, about families, and about what a decent society owes its members. To address those issues, we need something the policy analysts will not provide: an ideal picture of the kind of society we hope to live in and to leave to the next generation.

I want to present such a picture, a vision of a just and humane social-welfare system, rooted in the insights of feminism, and based on a recognition that dependency is a normal part of life and that caring for others is socially valuable work. This vision, to be sure, will not be realized in the immediate future. Still, by setting out long-term aspirations, it can serve now as an organizing tool and a critical yardstick for evaluating the policy proposals that are beginning to make the rounds.

What, then, would an ideal system look like? I will contrast two attractive answers, either of which would be an enormous advance over current US arrangements. Then I will offer a third alternative that brings out the best in both. Before getting to the proposals, though, we need to get straight about what is really wrong with welfare as we know it.

Beyond the World of the Family Wage

In the United States, “welfare” typically means Aid to Families with Dependent Children (AFDC). But AFDC is really only one part of a broader social-welfare terrain that — in the United States and other western countries — descends from the industrial era of capitalism. A guiding assumption in that era was that people were organized into male-headed households, living principally from the man’s wages. According to the ideal image, the male head of the household would be paid a “family wage” sufficient to support a wife-and-mother — who performed domestic labor without pay — and children. Of course countless lives never fit this pattern. Still, it provided an important social paradigm underlying the structure of most industrial-era welfare states.

Those states had three tiers, with social-insurance programs occupying the first rank. Designed to protect people from the vagaries of the labor market (and to protect the economy from shortages of demand), these programs replaced the breadwinner’s wage in case of sickness, disability, unemployment, or old age. Many countries also featured a second layer of programs, providing direct support for full-time female homemaking and mothering. A third level served what was called the “residuum.” Largely a holdover from traditional pre-industrial poor relief, these programs provided paltry, stigmatized, means-tested aid to needy people who had no claim to honorable support because they did not fit the family-wage mold.

Despite endless variations in organizational detail and generosity of support, welfare states in all western democracies fit this broad pattern. But two exceptional features of the US system merit brief notice.

First, social insurance was (and continues to be) provided through a “private welfare state.” A wide range of benefits — from pensions to health plans — are paid by firms as part of market compensation; they do not come from the government as basic entitlements for all citizens. In addition, the second and third tiers have been merged. AFDC is both the principal support for women’s childrearing and the principal program of poor relief. This merger explains the term “welfare mother,” and the associated images that circumscribe American debate about the welfare state.

Reform proposals must, of course, take these distinctive features of the American welfare system into account. But troubles in the welfare state are not peculiar to the United States, and no amount of tinkering with AFDC can resolve them. Serious ideas about reform must address the problems faced by welfare states quite generally. And the roots of those more general troubles lie in the collapse of the world of the family wage, and its central assumptions about labor markets and families.

In the labor markets of postindustrial capitalism, it is hard to find jobs that pay enough to support a family. Women’s employment is increasingly common, though it continues to be much less well paid than men’s. Moreover, many jobs are temporary or part-time and do not carry standard benefits. This growth of low-wage, temporary, and unstable employment generates deep fears about personal and family security. And because so many benefits in the United States are tied to employment, those fears are particularly deep here.

We all know about the changes in families: people are marrying less and later, and divorcing more and sooner. Growing numbers of women, both divorced and never married, struggle to support themselves and their families without access to a male breadwinner’s wage. And gender norms are highly contested: thanks in part to the feminist and gay-and-lesbian liberation movements, growing numbers are rejecting the male breadwinner/female homemaker model.

In short, a new world of economic production and social reproduction is emerging — a world of less stable jobs and more diverse families. No one can be certain about its ultimate shape, but this much is clear: the emerging world, like the world of the family wage, will require a welfare state that effectively insures people against uncertainty. In fact, the new forms of labor market and family increase the importance of such protection. It is clear, too, that the traditional welfare state is no longer suited to providing this insurance. We need something new, a postindustrial welfare state suited to radically new conditions of employment and reproduction. What should it look like?


To answer that question, let’s consider some of the values, ideals, and principles that ought to guide reform efforts. Here, then, are five objectives for refashioning the welfare state: 1. Anti-Poverty: The first objective of social-welfare provision is to prevent poverty. Because of the decline of the family wage, the special disadvantages of women in the labor market, and the growing numbers of single-mother families, this objective is now particularly pressing.

2. Anti-Exploitation: Social welfare should aim to protect vulnerable people from exploitation. And needy women with no other way to feed themselves and their children are especially vulnerable — to abusive husbands, sweatshop foremen, and pimps. Welfare policy should help prevent such exploitation by enabling subordinates to leave destructive relationships through provision of alternative sources of income. The point of increasing “exit options” is not necessarily to break up relationships, but, where possible, to improve them: a wife who knows she could support herself and her children outside her marriage has more leverage within it (the same is true for the low-paid nursing-home attendant in relation to her boss).

3. Equality: A welfare system that prevents women’s poverty and exploitation might still tolerate severe inequality between men and women. A further aim, then, is to reduce such inequality.

This is in part a matter of reducing income inequality: women’s earnings are less than 70% of men’s, much of women’s labor is not compensated at all, and many women suffer from “hidden poverty” because of inequality within families. Income equality does not mean identical incomes for all. But it rules out unequal pay for equal work, the wholesale undervaluation of women’s labor and skills, and vast differences in post-divorce income.

Social policy should also promote a more equal distribution of leisure time. Many women, but only a few men, do both paid work and the unpaid labor of caring for children, parents, and other relatives; and women suffer disproportionately from “time poverty.” So we want to avoid welfare arrangements that equalize incomes while requiring women to work an extra shift.

4. Anti-Marginalization: A welfare system could meet all the preceding objectives and still push women to the margins of society. By limiting support to generous “mothers’ pensions,” for example, it could render women independent, well-provided for, and respected, but isolated in a separate domestic sphere, cut off from the larger society. Social policy should combat such marginalization. It should promote women’s full participation on a par with men in employment, politics, and the associational life of civil society.

5. Anti-Androcentrism: To enjoy levels of well-being comparable to men’s, women should not have to live traditionally male lives, or fit into institutions designed for men. Policy should aim to ensure that human beings who can give birth and often care for relatives and friends are not treated as departures from a norm, but as equal participants. And that means changes for men as well as for women.

Not everyone would endorse all five objectives, and people are certain to disagree about their relative importance. Still, each has considerable force, and they all should be accorded some weight. To arrive at a maximal vision, then, we need to imagine a postindustrial welfare system that promises to address all of them.

Universal Breadwinner

One vision of welfare reform proposes a path from the world of the Family Wage to the world of the Universal Breadwinner. This project, whose guiding assumptions are implicit in the political practice of most American feminists and liberals, aims to ensure labor market equality for women — to ensure that women can support themselves and their families through their own wage earning. The idea is to get everyone into breadwinner role, so that women can take their place alongside men as citizen-workers. Universal Breadwinner offers an ambitious vision. It would require extensive day care and elder care to free women from their current, unpaid domestic responsibilities; workplace reforms to remove sex discrimination, sexual harassment, and other obstacles to equal opportunity; public efforts to break the cultural association of breadwinning with masculinity; and new policies to help change socialization, reorienting women’s aspirations from domesticity to employment, and men’s expectations toward acceptance of women’s new role. None of this would work, however, without macroeconomic policies to create permanent, full-time, high-paying jobs for women, and (re)training programs to ensure the skill required for those jobs. With such jobs, women would have full access to both the private welfare state of firm-based benefits and the public system of social-insurance entitlements.

How would Universal Breadwinner organize carework? The bulk of this work would shift out of the family and into the market and the state, where employees would perform it for pay (sufficient to support a family). Universal Breadwinner, then, requires a policy of comparable worth to redress the current undervaluation of skills and jobs currently coded as “feminine” and/or “non-white.”

Many benefits would continue to be linked to employment and distributed through social insurance; their levels would vary with earnings. In this respect, the model resembles the industrial-era welfare state. It differs in that many more women’s careers would look like men’s. Not all adults can be employed, however. Some will be unable to work for medical reasons, and others will be unable to get jobs. Some, too, will have responsibilities for other people that they are unable or unwilling to shift elsewhere; and most of these will be women. To provide for these people, Universal Breadwinner must preserve a tier of social welfare that provides a need-based, means-tested alternative to wages.

Universal Breadwinner is far removed from present realities. And the idea of creating massive numbers of jobs that pay enough to support a family single-handedly runs precisely counter to current trends of increasingly unstable conditions of employment. There may be no great gain in ensuring equal access of men and women to a shrinking pool of jobs and the private welfare state benefits that come with them. Assume, however, that a combination of tight labor markets and new programs of training generated an expanded pool of well-paying jobs. Let’s now ask whether we would get what we want.

On the generous assumptions now in place, Universal Breadwinner would do well at preventing poverty and exploitation. Secure breadwinner-quality jobs for all employable women and men — plus the services that would enable women to take such jobs — would keep most families out of poverty; and decent levels of transfers would take care of the rest. Assured of decent incomes, moreover, women would be able to leave unsatisfactory relationships, reducing their vulnerability to exploitation.

Universal Breadwinner does much less well, by contrast, at ensuring equality of leisure time. It makes the unrealistic assumption that virtually all the work associated with childbearing, parenting, and other family responsibilities can be shifted to the market and/or the state. But it cannot, short of universal surrogacy and other presumably undesirable arrangements. Some housework, of course, such as cooking and (some) cleaning, could — if we are prepared to live in collectives or eat out all the time. Even those tasks that are shifted, however, do not disappear without a trace, but lead to burdensome new tasks of coordination.

Women’s chances for equal leisure, then, depend on whether men will do their fair share of this work. But here the Universal Breadwinner model does not inspire confidence. It offers no disincentives to male free-riding; moreover, by celebrating paid work, it implicitly denigrates unpaid work and actually fuels the motivation for men to shirk responsibilities. The increasing numbers of women with children and without partners would in any case be on their own. And those in lower-income households would be less able to pay for help with housework.

Universal Breadwinner also performs poorly in overcoming androcentrism. Men provide the norm; women are expected to fit in. Traditionally female carework is treated as second-rate — what you must slough off to become a breadwinner. The model citizen is the breadwinner, now nominally gender-neutral. But that status is implicitly masculine; it is the male half of the old breadwinner/homemaker couple, now required of everyone. The female half of the couple has simply disappeared. Her distinctive virtues and capacities are not preserved for women, let alone extended to men.

Universal Breadwinner’s record on the remaining objectives is fair — but no better than that. Take, for example, income equality. Although it reduces wage inequalities between men and women, it is not otherwise egalitarian. It contains a basic social fault line dividing breadwinners from others, to the considerable disadvantage of those others, most of whom are likely to be women.

Universal Breadwinner, then, has its attractions. But they are greatest for childless women and for those women whose domestic responsibilities can easily be shifted to social service — women whose lives most closely resemble the male half of the old family-wage ideal couple.

Caregiver Parity

According to a second vision, reform should move us from the family wage to Caregiver Parity. This is the picture suggested by the political practice of most Western European feminists and social democrats. Rather than emphasizing labor market equality, the idea is to support the work of caring for children, parents, and other relatives in the household. The point is to enable women with significant domestic responsibilities to support themselves and their families either through carework alone or through carework plus part-time employment. Instead of making women’s lives the same as men’s, it would put caregivers and breadwinners on an equal footing: eliminating the penalties caregivers currently incur and making the difference between the two kinds of lives costless. This model, too, is ambitious. It assumes that many (though not all) women will follow current US female practice — alternating spells of full-time employment, full-time carework, and mixtures of part-time carework and part-time employment. The role of social policy will be to eliminate the considerable burdens now faced by women who follow this life-course. This will require a program of caregiver allowances to compensate childbearing, childraising, housework, and other forms of domestic labor, with allowances sufficiently generous to support a family — equivalent to the wages of breadwinners. Workplace reforms are also needed to make it easier to combine carework and part-time employment. The key here is flexibility. That means a generous program of mandated pregnancy and family leave so that caregivers can exit and enter employment without losing security or seniority, and a program of retraining and job search for those not returning to old jobs. Mandated flex-time is also essential to permit caregivers to shift their hours to accommodate their responsibilities, including shifts between full- and part-time employment. Finally, in the wake of all this flexibility, there must be programs to ensure continuity of all the basic social-welfare benefits, including health, unemployment, disability, and retirement insurance.

This model organizes carework very differently from Universal Breadwinner. The bulk of such work would remain in the household, supported with public funds. To assure continuous social-insurance coverage for people alternating between carework and employment, benefits attached to both must be integrated into a single system. In this system, part-time jobs and supported carework must be covered on the same basis as full-time jobs. Thus, a woman finishing a spell of supported carework — taking care of an elderly parent, for example — would be eligible for unemployment insurance benefits on the same basis as a recently laid off employee in the event she could not find a suitable job. If she became disabled, she would receive disability payments on the same basis as a disabled employee. In establishing eligibility for pensions, years of supported carework would count on a par with years of employment. Benefit levels would ensure equivalent treatment for carework and employment.

Caregiver Parity also requires a residual tier of social welfare. Some adults will be unable to do either carework or waged work, including some without prior work records of either type; most of these will probably be men. So the model must offer means-tested wage-and-allowance replacements. Caregiver Parity’s residual tier should be smaller than Universal Breadwinner’s, however, because of the near-universal coverage by the integrated breadwinner-caregiver system of social insurance.

Caregiver Parity, too, is far removed from current US arrangements. It requires large outlays of public funds to pay caregiver allowances, hence major tax reform and a sea-change in political culture. Assume though, for the sake of argument, that it could be established under favorable conditions. Let’s consider whether we would get what we want.

We can stipulate quickly that Caregiver Parity would do a good job of preventing poverty and exploitation, helping those who are now most vulnerable. It would reduce non-employed wives’ dependence on husbands by providing them with a direct source of income. And it would keep the families of single mothers out of poverty during spells of part-time employment or unemployment. Because those spells would carry the basic social-insurance package, moreover, women with “feminine” work patterns would have considerable security.

Caregiver Parity performs quite poorly, however, on income equality. Although the system of allowances and wages provides the equivalent of minimum breadwinner income, it also institutes a “mommy track” in employment. Most of the jobs on this track will pay considerably less (even at the full-time rate) than comparable breadwinner-track jobs. Two-adult families will have an economic incentive to keep one partner on the breadwinner track rather than to share spells of carework between them; and given current labor market inequalities, heterosexual couples will generally “choose” a male breadwinner. Given current culture, socialization, and reproductive biology, moreover, men are generally unlikely to choose the mommy track in the same proportions as women. So the two tracks will carry traditional gender associations. And those associations are likely in turn to produce discrimination against women in the breadwinner track. Caregiver Parity may make difference cost less, then, but it will not make difference costless.

Caregiver Parity also performs poorly in preventing women’s marginalization. By supporting women’s informal carework, it reinforces the view of such work as women’s work and preserves the current division of domestic labor between men and women. By consolidating separate labor markets for breadwinners and caregivers, moreover, the model marginalizes women within the employment sector. By reinforcing the association of caregiving with women, finally, it may also impede women’s participation in other spheres of life, such as politics and civil society.

On the remaining objectives, the performance of Caregiver Parity is fair. It helps promote equality of leisure time by making it possible for all women to avoid the double shift if they opt for full- or part-time supported carework at various stages in their lives. (Currently, this choice is available only to a small percentage of privileged American women.) This choice, however, is not truly costless. Some women with families will not want to forego the benefits of breadwinner-track employment and will try to combine it with carework. Those without a partner on the caregiver track will suffer significant disadvantage with respect to leisure time. Men, in contrast, will largely be insulated from this dilemma.

Caregiver Parity is also only fair at combating androcentrism. It does treat caregiving as intrinsically valuable, not as a mere obstacle to employment. And it accommodates traditionally “feminine” life-patterns, thereby rejecting the demand that women assimilate to “masculine” patterns. Still, it leaves something to be desired. Caregiver Parity stops short of affirming the universal value of activities and life-patterns associated with women. It does not even ask men to pursue caregiving, much less provide incentives or encouragement for them to do so.


Universal Breadwinner and Caregiver Parity are both extremely ambitious programs. But neither one — even in a highly idealized form — gives us everything that we want. If they were the only possibilities, we would face a hard set of tradeoffs. Suppose, however, we reject this unhappy choice and try to develop a third — equally ambitious — alternative that combines the best of both, while jettisoning the worst features of each. The nub of the third idea is to induce men to become more like women are now. That one master stroke would do wonders for both projects: If men were to do their fair share of informal carework, Universal Breadwinner would come much closer to equalizing leisure time and eliminating androcentrism. Similarly, if they were to do their fair share of supported carework, Caregiver Parity would better approximate income equality and reduce women’s marginalization.

In its ideal form, then, the postindustrial welfare state is a feminist welfare state that integrates breadwinning and caregiving. Unlike Caregiver Parity, this “Integration” model would not divide employment into two different tracks; all jobs would assume workers who are caregivers, too; all would have a shorter work week than full-time jobs have now; and all would have services that make it easier to be both a worker and a family member. Unlike Universal Breadwinner, however, employees would not be assumed to shift all carework to social services; some informal carework would be publicly supported and integrated on a par with paid work in a single social-welfare system. Some but not all of this work would be performed in households by relatives and friends. Other supported carework would be located in civil society. In state-funded but locally organized institutions, childless adults, older people and others without kin-based responsibilities would join parents and others in democratic, self-managed carework activities. In this way, the Integration model would not only promote women’s equal participation in employment, it would also promote women’s and men’s equal participation in civil society.

The key to integration is to develop policies that discourage free-riding. Contra conservatives, the real free-riders in the current system are not poor single mothers who shirk employment, but men of all classes who shirk carework and domestic labor, and corporations who free-ride on the labor — underpaid and unpaid — of working people.

The best statement of the ideal of Integration comes from the Swedish Ministry of Labor: “To make it possible for both men and women to combine parenthood and gainful employment, a new view of the male role and a radical change in the organization of working life are required.” The trick is to imagine a social world in which the lives of citizens integrate wage-earning and caregiving with community activism, political participation, and involvement in the associational life of civil society — while also leaving time for some fun. This world is not likely to issue from any of the reform proposals that will appear on the table in the upcoming debate. But it would be a good world to live in, and unless we are guided by this vision now, we will never get any closer to achieving it.