Parents—especially mothers—find it hard to combine child-rearing with their endeavors in the larger world. For some, the tension is primarily psychological—they feel torn between family and career and never seem to find the right balance. But for many, the conflict has a harder edge: it pits economic necessity against a child’s need for conversation, for recreation, for closeness.

This problem has grown increasingly intractable because of changing social expectations about parenthood. In the not-so-distant past, child-rearing was an economic bargain as much as an emotional venture: children provided a small platoon of workers for the farm or factory and an economic hedge against illness, disability, and old age. But technological and social change have reduced the economic rewards of child-rearing while imposing steeply increased demands on parents. Today, society expects parents to do the intensive work of preparing children for life—to supervise a lengthy period of education and to give priority to their children’s needs for nearly two decades. As a result, parenthood has become an extraordinarily demanding social role, requiring a complete restructuring of economic and personal lives. These escalating demands have led some to call for a more-generous social policy in support of child-rearing.

Libertarians of course reject such proposals. Parents know—or ought to know—what they are getting into; they choose to get into it, and they should plan ahead. Because their burden results from their choices, it is not an occasion for special state support. For the past generation, this laissez-faire perspective has dominated American social-welfare policy. Republican presidents took steps to reduce the size of the welfare state and to require that families on welfare become economically self-sufficient; the Clinton administration’s 1996 welfare reforms embraced the same view.

But you do not have to be a libertarian to raise questions about subsidies for parenting. Some on the Left appreciate the affirmation of individual equality implicit in the libertarian claim. People make all sorts of decisions that limit what they can do with the rest of their lives. Why should a parent be able to claim a greater share of public resources than her childless peer?

But this assertion of equality between parents and non-parents overlooks society’s keen interest in parental performance. Good parents provide their children with continuity of care—the intensive, intimate, and ongoing relationship that human beings need to develop their capabilities. Without it, society’s institutions of education, health care, and even public order may fail.

Most parents will be responsible because they genuinely want to do right by their children, but because of the social importance of parenting, it makes sense to provide them with public support. After all, society too takes a hand in setting the responsibilities of parenthood. Parents choose to have children, but they do not choose the specific burdens society imposes on this choice; these are politically imposed, and therefore it is not unreasonable to compensate parents to improve their economic options and prospects.

Society’s Stake

Children need nurturing: that much is obvious. But why is it important that a parent continue in her role for the long term? What does continuity add to the idea of care?

Child psychologists explain that healthy emotional development requires a close and enduring relationship with one or more parent figures. Children rely on these “psychological parents” to create the stable environment that is essential for their development. For instance, consistency in parental praise and discipline helps children develop emotional control. Parents also represent lasting role models for older children to emulate—and for teenagers to reject, safe in the knowledge that the parent will not leave. Every child needs to be a member of a family in which he feels wanted and where he can express not only love but also anger and thereby learn to manage aggression.

Parents sometimes must exit: illness, accident, or another calamity may cut short a parent-child relationship. But psychologists emphasize the importance of recreating continuity for these children as soon as possible. Studies document the serious and lasting emotional harm suffered by children denied continuity of care for long periods or during formative stages.

While the philosophy of continuity of care emphasizes the loving bond between parent and child, parents need not devote every moment to their child or eschew paid child care. Nor does continuity of care endorse any narrow class of family structures. The key is the quality and endurance of the relationship—and not the parents’ sex, age, or biological tie to the child. Children can thrive in extended-family settings and in non-traditional arrangements of many kinds.

In addition to fostering children’s emotional development, continuity of care serves a second social function: parents who persist with their children for the long term become uniquely qualified to represent their interests in the health care system, the educational system, and other public settings. Virtually all our public institutions for children’s care rely on the expert proxy of parents.

Of course, parents do not ordinarily experience continuity of care as a command from the state. Good parents care for children out of love and a sense of moral obligation. But society uses both carrots and sticks to encourage parents to provide continuity of care. When parents persist with their children, the state rewards them with a sphere of privacy that is protected from state intervention. What lawyers call the doctrine of parental authority grants parents wide latitude to rear their children as they choose. Generally speaking, the state has the legal authority to challenge parents’ actions only if they threaten serious harm to the child. As long as parents provide their children with continuity of care, they scarcely feel the law’s supervision.

But when parents fail to provide continuity of care, the state revokes or curtails their parental prerogatives. Parents who undermine continuity of care may lose their opportunity to be parents at all. In a variety of circumstances, the law sends a stern message: parents who exit or who participate only intermittently in their children’s lives may well lose custody to others, who are prepared to provide children with the continuity of care that they need.

Our society’s commitment to continuity of care is grounded in a deep belief in human equality. An egalitarian society aspires to create institutions that ensure to every person the conditions of autonomy—the chance to develop the capabilities that she will need to formulate, choose, and pursue a vision of the good life. Although citizens may disagree about what these capabilities are and how best to foster them, there is a consensus on the basics: for instance, almost everyone would agree that the state should guarantee each child an education.

Children’s claim to the conditions of autonomy follows from their equal worth as individuals. Continuity of care is in this sense fundamental to an egalitarian society. It cannot be up to parents to decide whether they endorse continuity or would instead prefer to spend their time and money on other things. Children need continuity of care to flourish; short of science fiction, there is simply no acceptable substitute for parental care.

The Double Edge

From a parent’s perspective, continuity of care has a double edge: it promotes children’s interests and may provide moral and emotional satisfaction, but it also limits parents’ opportunities. Parents may, for example, be forced to compromise in their choice of jobs or homes; they may have to turn down higher pay if it would mean working the night shift, relying on a questionable babysitter, or traveling. For some, these adjustments would mean a lower standard of living and uncertain security in old age. For others, compromise would bring daily hardship. A low-income mother needs her job to pay the rent, but a lengthy commute, a long workday, or a rigid job schedule may make it difficult or impossible for her to give her children adequate supervision.

Social-science research is often equivocal, but on the cost of parenthood to mothers in particular a truckload of research exists to establish how it limits economic options in every class. Mothers work less, earn less, and achieve less in the marketplace than fathers and than childless women. Child-rearing takes up enormous amounts of time—especially mothers’ time. Taking into account child care, housework, and paid work, mothers work more hours than any other group. Even when both parents hold paid jobs, the mother usually takes primary responsibility for arranging child care, caring for the children during non-work hours and taking time off when children are sick or day-care arrangements fall through. Mothers also shoulder the greater burden among single-parent families. More than 80 percent of single- parent households are headed by mothers, and more than two thirds of children of divorced parents live primarily with their mothers.

The data on mothers’ work-force participation confirm the effect that child-rearing has on paid work. A high percentage of mothers (about 46 percent) either do not hold paid jobs or work only part-time, and mothers interrupt their paid employment at far higher rates than fathers. These job interruptions are consistent across class lines: in 2000, 20 to 35 percent of mothers in every income bracket did not hold paid jobs. Only 54 percent worked full-time. In contrast, 90 percent of fathers work full-time, and even fathers with access to family-friendly workplace benefits, such as paternity leave and part-time work, tend not to use them.

When mothers do remain in the work force, they earn less than other women. This gap persists even after controlling for age, education, work experience, and other attributes. Over a lifetime, mothers earn about five percent less per child than they would have earned otherwise.

Mothers’ economic disadvantage during their working years has repercussions into old age. Job interruptions and lower lifetime earnings reduce their private pensions and their Social Security benefits. Some married women can share their husbands’ (generally larger) Social Security pensions, but never-married women and many divorcées are on their own.

Although this situation reflects continuing gender discrimination, the economics of parenthood are not simply a byproduct of gender inequality. The fundamental problem is that child-rearing requires an irreducible minimum amount of work. Whatever the division of labor between mothers and fathers, continuity demands intensive time and energy from someone. Our (commendable) commitment to higher and higher standards for children’s care comes at a price measured in parents’ life chances.

No Exit

When society asks parents to provide continuity of care to their children, it makes an extraordinary demand. In effect, society says to parents: No Exit.

The No Exit obligation is deeply embedded in our understanding of parenthood, but from a broader perspective it represents an exceptional limitation on parents’ capacity to choose their way of life—a capacity that our society ordinarily seeks to protect.

Exit occupies a central place in most conceptions of individual autonomy. Theories differ, but most suggest that an autonomous individual should be able to choose a life plan that seems good to her and modify it over time as her values change. To be “the author of one’s own life,” in the philosopher Joseph Raz’s phrase, one must have a variety of meaningful options.

Ordinarily, a free society hesitates to require individuals to follow a particular way of life. My Yale Law School colleague Jed Rubenfeld frames the argument for abortion rights in just this way: denying women the opportunity to have abortions, he argues, would amount to requiring them to live a life of the state’s choosing. Banning abortions would be like imposing a No Exit obligation on pregnant women. As long as we do not consider a fetus to be a child entitled to continuity of care, such a No Exit obligation would be an unreasonable burden on women’s ability to determine their own way of life.

Although parents choose to bear (or adopt) children, the voluntary nature of the entry into parenthood cannot be used to justify unlimited regulation of parents’ lives. The state should not be able to severely restrict fundamental life activities and then excuse its action on the grounds that individuals may avoid that way of life. An abortion ban, for example, may not be dismissed as trivial because women could avoid unwanted pregnancies by not having sex.

To be sure, the state legitimately regulates individual conduct all the time. At first glance, the No Exit obligation seems to fit the familiar category of laws that prevent one person from harming another. Criminal laws prohibit assault, for example, and environmental laws prohibit pollution. We ordinarily understand such rules as protecting individual autonomy and not limiting it: we don’t suppose the law is infringing anyone’s “right” to assault or to pollute.

In form, the No Exit obligation seems similar: it prevents harm to children by restricting parental exit. But a legitimate regulation may still impose an unfair burden on individuals who pursue the regulated activity. If the ideal is for everyone to enjoy a lifelong ability to form, refine, and pursue her own vision of the good, then it is insufficient to focus on children alone: we owe some consideration to the autonomy of parents.

Of course, the law does not absolutely oblige parents to remain with their children. The state can encourage continuity, but it cannot compel intimacy. Parents determined to exit their children’s lives may do so if they are willing to renounce their parental prerogatives, and many do. Almost one third of American children live with just one parent, usually their mother. And although many children have warm relationships with non-custodial parents, many do not.

But we should not mistake the fact of parental exit for societal indifference to the value of continuity of care. Society has only limited means to secure continuity for children. It is no coincidence that most parents who exit do so in the context of a romantic or marital breakup. When parents become hostile or simply can no longer cooperate with each other, it may be impossible for them both to remain in a close relationship with their child. Nearly 80 percent of custody cases are uncontested, and in the great majority of these, the mother takes physical custody. True joint physical custody remains relatively rare; estimates suggest it occurs in only 10 to 30 percent of cases.

It may be that the state should try harder. Perhaps the law might do more to promote joint custody. Public policy could even restrict marital exit, as some experts advise. But to date, our society has not managed to discourage parents from breaking up or to prevent some parents from exiting their children’s lives when they do.

Still, most parents, and especially most mothers, provide continuity of care to their children. Somehow the combination of parental love, social norms, and legal incentives and disincentives induces most parents to do the right thing.

While most parents probably feel that their obligation to their children comes from within and not from society or the state, continuity of care is a social creation as well as a natural phenomenon. Parents’ choices are governed by the legal and social institutions that define parenthood, and the historical record reveals that parents have not always taken care of their children over the long term. For instance, a vivid study by the historican John Boswell notes that child abandonment has a long history in Europe, ending surprisingly recently. It was not only the hard-pressed poor who gave up their children but also better-off people who found them inconvenient.

Moreover, it is impossible to disentangle parents’ “pre-social” values from their values as shaped by society. Indeed, it is one measure of social progress that most people would probably endorse continuity of care as an integral part of parenthood.

Today the fundamental problem is not how to get parents to provide continuity of care to their children but whether the terms of modern parenthood are fair to those who meet their responsibilities. What does society owe these parents?

Caretaker Resource Accounts

If a society expects parents to provide continuity of care to their children, it should in return give special consideration to the consequences for parents’ lives. There are many ways for the state to take practical action, but I want to outline just one. The program that I propose uses “caretaker resource accounts” to improve the long-term opportunities of parents who provide their children with continuity of care. (I refer to these as “caretaker parents” to emphasize that not every legal parent participates actively in the work of child-rearing. Later on, I consider the issues involved in identifying caretaker parents.)

With the creation of caretaker resource accounts, the caretaker parent of every child under age 13 would be given an annual grant of $5,000, which the parent could use to pay for child care, (his or her own) education, or (his or her own) retirement savings in the current year or in any future year. Each participant would receive an equal share of public resources per year, and each would decide for herself (or himself) how to divide the funds among the three alternative uses. The program would expand parents’ options to give them maximum freedom to shape their own lives as they think best. Parents differ in their values, talents, and aspirations, and they would use their caretaker resource accounts in different ways—which is precisely the point.

Suppose Abigail has a new baby. The government would establish an account in her name on the books of a government agency or private financial institution and would deposit $5,000 once a year. Abigail’s account would receive annual deposits as long as she remained the child’s caretaker parent and until the child reached an age at which most children need far less intensive care—say, between 13 and 18.

Each year Abigail could spend her $5,000 on child care or on tuition for herself, or she could make a deposit of $5,000 into a retirement account in her own name. She could split her $5,000 among the three options or spend it all on one. If she spent less than $5,000 in any given year, the remaining funds would accumulate with interest to be used in the future. The program would not permit Abigail to withdraw cash from her account or spend the money on day-to-day expenses, either for herself or for her family. This restriction reflects the aim of the program, which is to improve parents’ opportunities, not to underwrite general consumption.

Importantly, the program would permit parents to change their plans over time: Abigail might spend her annual allotment on child care during heavy work years; on education to qualify her for a promotion or a new job in a transitional year; and on retirement contributions other years to make up for the time spent in jobs with low pay or few benefits.

The child-care subsidy would be especially valuable to a moderate- or lower-income parent. A child-care subsidy of $5,000 may seem modest to an upper-middle-class family used to the price of a nanny or a high-end day care. But for many families, child care is difficult to afford but necessary for a mother to be able to work; and perhaps surprisingly, $5,000 exceeds the average annual cost of child care even for full-time workers and workers with small children.

The $5,000 voucher is significantly more generous than existing child-care subsidies. Tax subsidies for child care provide at most $1,500 to $1,800 per year, and they assist only a subset of families, excluding the poorest. They also vary by family income, sometimes arbitrarily. Working-class families can earn too much to qualify for the child-care subsidies for former welfare recipients and yet too little to benefit much from those for middle-class workers.

The flexibility of caretaker resource accounts would also prove valuable to parents with low and moderate incomes. Families’ need for paid child care is highly variable. Many parents of modest means minimize child-care costs by working split shifts or asking relatives to care for their children; only about half of two-earner families with children under age 13 pay for child care. Furthermore, since a substantial fraction of mothers in every income class do not hold paid jobs or work only part-time, they need less child care.

These patterns might change with the creation of caretaker resource accounts. The program would enable parents to pay for more or better child care if they wished to do so. But the program would open up other options as well. Parents who did not spend all their funds on child care could use the remainder of their $5,000 to supplement their retirement savings. That option would be especially valuable to low- and moderate-income parents, who may work in jobs without pension benefits. It would also improve old-age security for mothers who spend time out of the work force or in part-time jobs. In contrast, existing tax incentives for retirement savings are far less meaningful for mothers and families of modest means.

The education option should be valuable across the income spectrum. Low- and moderate-income caretakers often have less education to begin with and less disposable wealth for paying tuition. Because the caretaker resource account would supplement, not supplant, existing student-aid programs, it could make existing grants more effective and decrease students’ debt burden. To make the education option as useful as possible, the tuition subsidy should be available not only for four-year colleges but for GED programs, vocational and technical training, and community colleges. To be sure, studies suggest that education produces only modest gains in earnings for some workers; for this reason, and also to respect individual values, caretaker resource accounts expand access to education while also preserving alternative options for those who conclude that more education is not the right path.

Why $5,000? Any particular sum would be arbitrary, but $5,000 would be large enough to expand parents’ economic options yet still manageable from the perspective of the public budget. Five thousand dollars would help close the motherhood wage gap. It could fund a year’s tuition and books at the average public four-year college, and if saved for retirement, it would exceed the average annual contribution by 401(k) plan participants. The median family income is just $50,000, making the caretaker grant a significant addition to parents’ economic resources. For poorer parents, the relative impact would be even greater: the median income for single mothers, for example, is under $25,000.

The total annual cost of the program would be about $100 billion, taking into account savings in current programs that would be rendered duplicative. While this is not a small commitment, neither is it outsized for a major social initiative, especially when the annual federal budget is more than $2 trillion. Compare this cost to that of other large-scale programs for vulnerable groups: Social Security retirement benefits and public education each cost more than $400 billion a year. Consider, too, the current president’s successful effort to enact tax cuts worth more than $1 trillion over ten years.

Dynamic Effects

Parents’ opportunities would be expanded by caretaker resource accounts in an obvious financial sense: parents would have more dollars to spend. But would the program meaningfully improve parents’ life options? Some may suggest that caretaker grants are too small to make a difference; others may worry that they would do too much, perhaps encouraging mothers to drop out of the paid work force, leaving them better off in the short term but more vulnerable over the long term, as some feminists have argued. But this objection has been directed primarily at income-support programs, and the dynamic effects of caretaker resource accounts are likely to be quite different.

Would parents work more, or less, in response to caretaker resource accounts? We know that men’s, and fathers’, work behavior is relatively inelastic; as for mothers, theory predicts that some would remain in the work force longer or work more hours than otherwise. But other mothers would work less, perhaps feeling more relaxed about taking a part-time job or a stint out of the work force, relying on their caretaker accounts to give them an economic boost later on. For example, a parent who chooses to save her grant for retirement might feel somewhat less pressure to take a paid job for long-term financial security. The education option might lead some mothers to spend more time in school and less in the workplace.

Theory cannot definitively predict how many mothers will fall into each group. Researchers found that pre-1996 welfare programs reduced the hours recipients worked, but only slightly (contrary to some conservative rhetoric). In contrast, child-care subsidies have substantial positive effects on both single and married mothers’ presence in the work force. Caretaker resource accounts would probably fall somewhere in the middle, encouraging work to a greater extent than income support but somewhat less than pure child-care subsidies. Unlike income support, caretaker grants do not directly support time away from work: parents cannot use their grants to pay the grocery bill. But unlike traditional child-care subsidies, caretaker resource accounts do not have a use-it-or-lose-it feature; parents could stop working without losing benefits.

The long-term effects of the program would be especially complex and interesting. Some mothers would use their caretaker grants to re-enter the labor force sooner than otherwise or with higher levels of education. The program could support mid-life career changes. And improving the economic options of the caretaker parent might even encourage greater sharing of child-rearing work between fathers and mothers.

It is important to bear in mind that caretaker resource accounts are not designed to maximize the labor supply. Their aim is to expand parents’ life options.

Why Not Income Support?

The caretaker resource account program may appear to provide an unduly limited menu of options. Why shouldn’t parents use the money to pay the rent or the grocery bill?

All three options support parents’ futures. Child care helps parents who wish to preserve their skills and opportunities by holding paid jobs. Education helps parents improve or refresh their marketable skills. Retirement contributions improve old-age financial security. In more abstract terms, caretaker resource accounts represent a (mildly) paternalistic initiative. The objective is not to subsidize parents’ consumption but to support forward-looking investments.

Compare this strategy to alternative proposals for income support such as children’s allowances or monthly cash payments to parents. Although these have lost political ground in the United States in the last decade, they retain a distinguished constituency in the academy and follow a model adopted by some European countries.

I have considerable sympathy for income-support proposals. Like caretaker resource accounts, these programs tend to reflect normative commitments to individual freedom and the dignity of unpaid work. But I have come to believe that caretaker resource accounts better address parents’ needs. The crucial difference is that income support aids family consumption, while caretaker resource accounts concentrate funds on parents’ own life options. Of course, contributing to basic family expenses is a classic function of anti-poverty programs, and children’s allowances might be a vital part of such an agenda: too many parents earn too little to support their children properly. I would point out simply that parental opportunity and child poverty are distinct, if related, problems. Not all parents who struggle for autonomy are poor; and not all those laboring under the burden of poverty are parents.

Of course, like any in-kind or voucher program, caretaker resource accounts would be vulnerable to the phenomenon of cash equivalence. Any caretaker parent who would, in the absence of the program, buy any of the three services would find that caretaker resources free up extra funds in the family budget. For example, if a parent who would ordinarily purchase child care used her caretaker grant to defray her regular child-care expenses, she would find herself with extra money in her budget. The result may be that a parent is richer but does not increase her expenditures on child care, education, or retirement.

How prevalent would cash equivalence be? Empirical studies find some substitution in the case of child care and higher education but conclude that subsidies tend to increase net purchases. Traditional retirement subsidies in the tax code are less successful, tending to reward individuals for earmarking savings for retirement. But caretaker resource accounts would extend retirement savings opportunities to a younger and less affluent group, making it more likely that parents who would have saved nothing for retirement may now save something.

The larger point is that cash equivalence dampens the behavioral impact of caretaker resource accounts but would not fundamentally undermine the program because it would ensure that every caretaker parent would spend at least $5,000 on investments for the future.

Who Is a Caretaker Parent?

Caretaker resource accounts raise a number of familiar administrative issues. For instance, qualifying for the program might require a means test or not. I tend to think the program would work more effectively without one.

The use of income testing presumes that income reflects parents’ long-term economic opportunities. But family income can overstate a caretaker parent’s access to resources. Studies suggest that middle- and upper-middle-class mothers may have limited power to direct resources for their own use and bargain for a change in their role over time—precisely because they have focused on child-rearing at the expense of paid work opportunities. A universal program, without an income test, would recognize that caretaker parents, whatever their income level, are different from non-parents because they bear the cost of the No Exit obligation.

Practical problems compound the costs of income testing. Means-tested programs are expensive to administer and error-prone, and they make planning difficult for individuals, who may not know until the end of the year whether they will meet the income requirements for next year’s grant. In contrast, a universal program provides a predictable grant every year, allowing caretaker parents to plan ahead. Income testing also tends to worsen the problem of high marginal tax rates on low earners; even families of modest means already find that every extra dollar of earnings costs them roughly 45 cents in actual taxes and lost government benefits.

While there is no need to nail down every administrative detail at this stage, one administrative question merits more sustained attention, even now: how should the program identify caretaker parents eligible to participate in the program? In principle, caretaker resource accounts should assist parents who provide continuity of care to their children. But in practice the state may not be able to determine accurately which parents meet that criterion. Furthermore, because predictability is important, the program will need a relatively simple rule. Caretaker parents should be certain about their claim to resources and not anxious about legal nuances that might, upon adjudication, force them to repay money already spent.

In some cases, a simple rule would suffice. A mother whose husband has long ago abandoned the family could easily be identified as the caretaker parent by her co-residence with the children. Similarly, a divorced or separated parent who has sole physical custody of his children could reasonably be deemed the caretaker parent. Although noncustodial parents can certainly play important roles in their children’s lives, studies suggest that custodial parents take on the greatest responsibility.

But how can the state identify caretaker parents when both live with their children (either because the whole family lives together or because parents live apart but share joint physical custody)? Sociological studies suggest that some couples share child-rearing work equally; others follow a traditional (gendered) division of labor; and still others engage in some middle course, with fathers taking on more than the traditional paternal role, but less than the mother’s commitment of time and energy. The gendered, traditional model is waning, but equal sharing has not become dominant.

Two simple solutions come to mind, neither entirely satisfactory. The program might award the entire grant to just one parent, identifying the “primary” caretaker parent by his or her lower earnings. That rule would benefit the mothers who take primary responsibility for child-rearing and typically earn less than fathers. But it would be unfair to parents who share responsibilities relatively equally—or to a parent who is both the higher earner and the primary caretaker.

A second simple rule would split the grant equally between co-resident parents. But while this solution is initially attractive, it would be inequitable for parents who do not share responsibilities equally (as most do not). Today it is mothers and not parents as a class who sacrifice economic opportunity in favor of child-rearing. While the program would still serve its objective by directing greater resources to (true) caretaker parents, the equal-splitting rule would dilute its effectiveness.

Technocrats could design intermediate options; for example, half the grant could be split between parents while the other half goes to the lower-earning parent. But the important point is that the administrative problem mirrors the gap between the ideal family and the reality. Ideally, child-rearing work should be freely chosen and evenly shared, and mothers and fathers should negotiate from positions of equality. But a combination of socialization, discrimination in the workplace, and other factors typically leave mothers at a disadvantage.

The lack of a perfect solution should not be unduly discouraging. Any large-scale public program encounters similar administrative difficulties when it uses simple rules to approximate complex social facts. States generally permit children to drive at around age 16, ignoring the variability of temperament and maturation among individual teens. The welfare system sets its benefits based on low income, even though poverty is complex in its causes and effects. The Social Security system awards retirement benefits at age 65, regardless of individuals’ ability to work and their need for assistance. Similarly, caretaker resource accounts would inevitably adopt some rough but workable eligibility rule.

Workplace Regulation

It may seem curious that caretaker resource accounts do not attempt any direct restructuring of paid employment. The program could affect working conditions indirectly, but it takes initial market conditions as given.

In contrast, many scholars and policymakers concerned about parents’ economic plight urge the state to actively regulate workplace conditions. Present law offers the Family and Medical Leave Act, a three-month program of unpaid leave for medical emergencies and childbirth. But the law might do more, granting workers the right to paid child-care leave, flex time, and comp time, and even part-time work at proportional full-time wages.

These initiatives could help nudge the workplace in a family-friendly direction. American and Scandinavian experience teaches that more mothers than fathers would likely take advantage of these arrangements. Still, mothers would be better off than they are today, when many employers seek an “ideal worker” who has no dependents at all.

The family-friendly workplace has undeniable appeal, and yet I remain uneasy. There are too many unanswered questions about the desirability and feasibility of legally mandated workplace benefits of this type. For example, family-friendly workplace programs by definition increase flexible-work options for parents. But in doing so they tend to exclude parents who wish to structure their lives differently. Family-friendly benefits provide, in effect, wage subsidies for working parents: they increase the net compensation package (understood as money plus time) that the market now offers. But like all wage subsidies, family-friendly benefits reward those who take paid jobs, and they provide the greatest rewards to workers who take the most heavily subsidized jobs—in this case, part-time and flex-time work, and so on. These policies offer little to parents who prefer jobs with traditional terms and to those who intend to take time out of the workforce.

Caretaker resource accounts reflect a more inclusive, and individual-centered, vision. They would provide equal funds to every caretaker parent, including those who wish to work full-time, those who prefer to work part-time, and those who choose to remain out of the labor force. Caretaker grants would not (at least directly) change the relative economic positions of these options.

The underlying ideal is that the state should remain neutral toward the role of paid employment in parents’ lives. Some caretaker parents work (or want to work) full-time, for financial security and for the dignity they find in that endeavor. Others forgo income and job prospects to take time away from paid work for child-rearing. These parents may be motivated by religion, or tradition, or a personal conviction that parental care is best. Some have children with special needs, who require more intensive care than even a flexible paid job will permit. Some parents find that their skills are not highly valued in the marketplace and qualify them only for routine jobs they dislike, while child-rearing may give them chance to make a difference in their children’s lives. I do not mean to romanticize the life of an at-home parent: many find it isolating and stressful. Still, a pluralist approach would enhance each caretaker parent’s capacity to decide for herself which life to lead.

Some may feel that women’s choices have been so distorted by the system of gender roles that neutrality and choice are impractical ideals and that the state ought to orient mothers firmly toward paid employment. But we can recognize that gender affects women’s preferences without concluding that public policy must urge paid employment for parents. Caretaker resource accounts adopt a subtler approach: they encourage caretaker parents to make investments in their own futures while leaving room for individual values and aspirations. Caretaker resource accounts could also complement other efforts to promote gender equality, including measures to fight workplace discrimination and the influence of gender in education and early socialization, in medical care, and in the administration of justice.

The second troublesome question about family-friendly workplace reforms is whether they would work. Both theory and research suggest that state regulations intended to improve mothers’ work options can, perversely, lower women’s wages, reduce women’s employment, and reinforce gender discrimination in the workplace. The relevant theory is complex, and the research is far from conclusive—happier outcomes are possible. But the evidence to date suggests caution.

The problem arises because employers could shift the costs of family-friendly workplace regulation to workers, and especially to female workers, through lower wages or lower employment. Crucially, this analysis applies only when the law requires employers to furnish family-friendly working conditions and not when employers voluntarily adopt family-friendly benefits to attract or retain workers.

To understand the problem, bear in mind that restructuring work to meet family-friendly regulations would be costly to the employer. Of course, family-friendly working conditions can also be beneficial for employers, improving worker morale and reducing absenteeism. But to the extent that regulations require employers to adopt family-friendly practices that reduce firms’ profits, there is the potential for perverse effects as employers attempt to recoup their costs by reducing wages or employment.

In effect, mandated benefits act like a tax on the employment of workers they are meant to benefit. And the effects are likely to be felt most keenly by female workers. Employers perceive (accurately) that mothers are more likely than fathers to take advantage of family-friendly benefits. Thus, mothers would become more expensive to employ than other groups, and firms would try to hire fewer of them, or pay them less. Although employers cannot usually identify mothers with precision, they can generally identify women.

This simplified explanation illustrates the problem, but it cannot capture all the relevant factors. The economist and law professor Christine Jolls has pointed out the critical impact of anti-discrimination laws. If the law could prevent employers from cutting women’s wages or reducing their employment, then the regulatory “tax” would fall on all workers and not primarily on women. But studies have found that employers have reduced women’s wages in response to legally mandated pregnancy protections and other benefits.

The point is not that family-friendly workplace regulations would inevitably produce these effects. There are better, and worse, ways to structure such programs; for instance, explicit taxation and income support are probably less prone to side effects than comparable workplace mandates. There is much we do not know empirically, and more research is needed. In the meantime, however, it seems best to concentrate on programs that can more reliably enhance parents’ ability to participate in the marketplace.

What Kind of Society?

Parents make a private decision to have children, but when they do so, they step into a public role. A fair society should expect parents to care for their children and to sacrifice time and opportunity if necessary, but it should also help parents to preserve a reasonable range of life options during and after their years of care.

Our current neglect of parents is far from benign. When we treat parental obligation as a purely private matter, we pit children’s interests against their parents’: we tell parents, in effect, Do the right thing for society, but don’t expect its help.

It is highly unlikely that pro-marriage reforms will save the day. The law cannot easily reverse the underlying social trends that have rendered marriage marginal to many children’s and parents’ lives. The proportion of divorced and unmarried parents remains high: a full third of American children are now born to unmarried parents.

Furthermore, the potential gains from post-divorce arrangements are limited. In 1997, only 56 percent of custodial parents had been awarded child support, and of those, only 67 percent received any payment at all. Better enforcement may not dramatically improve the situation. For poor and working-class families, the grim fact is that men’s real wages have fallen over the long term while women’s wages have remained low, leaving little surplus to divide between two households when parents divorce (or fail to marry). For middle-class couples, the long-term trend is that wealth is changing its form, with easily visible (and divisible) physical and financial capital being replaced by human capital—earning power—which is far more difficult to divide. The increasing number of stepfamilies and second families also limits the possibilities for greater financial support from absent parents.

These realities suggest that it is both impractical and unwise to rely on marriage as the primary source of security for children and those who care for them. But what is the alternative?

Caretaker resource accounts represent a first step: they would help parents provide continuity of care while pursuing lives of their own. But caretaker resource accounts should not be the end of the story. The special challenges confronting poor parents deserve further attention, as does the situation of parents who care for children with disabilities. For them, the No Exit obligation weighs especially heavily.

The goal should not be to render child-rearing costless. It would be impossible, and unwise, to attempt to erase the imprint of parenthood from parents’ lives. The goal should be to lighten the burden.