Supporting Workers by Accounting for Care

By Noah D. Zatz

Unlike food, clothing, shelter, and so on, child care makes no appearance in the basket of goods and services that means-testing policies use to determine what a household needs to maintain a mini­mally decent standard of living. Photo by peoplesworld.

Introduction

What is an essay about child care, welfare, and poverty doing in a sym­posium devoted to “Jobs and the American Worker”? An answer begins with a question: Who is “the” American Worker? Consider the dispute over the types of jobs created by the Obama Administration’s recent stimulus bill. When some observed the predominance of traditionally male-dominated oc­cupations in manufacturing and construction, a predictable dispute broke out among feminist commentators. Some argued for redirecting resources, and respect, toward “women’s work,” especially in human services; others coun­tered that this objection simply reinforced gender stereotypes when the real problem was women’s exclusion from “good jobs.”[1] This dustup high­lighted the interaction between public policy priorities and the gendered di­vision of labor—sex segregation, less politely—inside the labor market.

An even more basic question probes not just “which jobs,” but also the underlying choice to make jobs within the labor market the foundation for social policy promoting Americans’ economic well-being. That choice, ide­ological in the sense of politics hardened into truism, implicates the broader gendered division of labor between paid employment and unpaid efforts, especially caregiving, within families. Again, many feminists have objected that targeting benefits at market workers unfairly excludes the work done outside the labor market—work done predominantly by women and cultur­ally associated with them. And again come the rebuttals, noting that, in any one year, most women, indeed most mothers of young children, are in the labor market and that the pressing task is to support those efforts. This might be done by widening access to child care that is high quality, afforda­ble, and provided by people who get paid to do it.

One could be excused for thinking that the social policy primacy of labor market participation had been resolved definitively in the 1990s with welfare reform. President Clinton and the Democratic Party embraced the mantra of “if you work, you shouldn’t be poor,” including its implication that if you don’t work, you’re out of luck. Henceforth, questioning this con­sensus has seemed the province of dreamers and cranks.

Yet the issue will not disappear. One reason is money. Child care— ignored or disparaged when low-income parents, especially women of color and especially single mothers, do it themselves—is enormously expensive when you need to hire someone to do it reasonably well. In 2006, subsi­dizing care for low-income preschoolers cost over $400 per month per child,[2] more than the typical maximum cash welfare benefit for a family with two children.[3] And so it was that one could read the following head­line: L.A. County Officials Offer a Novel Idea to Save Millions: Supervisors Suggest Putting Unemployed Parents to Work Caring for Their Own Chil­dren.[4] Turning the usual rhetoric of “dependency” on its head, pushing some parents into the labor market actually increases their reliance on tax­payer support if the government pays for child care. Los Angeles County’s response could be good news to some who prefer to care for their young children themselves, but for others it represents lost opportunity to access the labor market. Another headline tells us Cuts to Child Care Subsidy Thwart More Job Seekers.[5]

In this article, I argue that two pathologies in contemporary antipoverty policy—inadequate child-care assistance and failure to value parental caregiving as work—are flip sides of one coin. That coin is “child-care invisibility.” Child-care invisibility is a ubiquitous feature of “means test­ing,” a basic technique that targets assistance to low-income households in programs like Temporary Assistance for Needy Families (TANF), Supple­mental Nutrition Assistance Program (SNAP, formerly known as Food Stamps), and the Earned Income Tax Credit (EITC). Means testing identi­fies who is poor, in the sense of lacking the “means” to meet basic needs, and who therefore requires assistance. Unlike food, clothing, shelter, and so on, child care makes no appearance in the basket of goods and services that these policies use to determine what a household needs to maintain a mini­mally decent standard of living.[6] Child care again goes unseen when the government determines whether those needs are being met or, instead, whether the household requires assistance to make ends meet. This absence from the two sides of means testing—needs and resources—is child-care invisibility.

Child care’s invisibility within means testing also implicates child care’s status as work. This connection arises from two legacies of 1990s welfare reform that more generally have made antipoverty policy into “work” policy as well. First, work requirements became central to benefits for low-income families with children, most notably when TANF replaced the New Deal-era Aid to Families with Dependent Children (AFDC). AFDC had not initially included work requirements, and eventually it became the primary target of efforts to “end welfare as we know it.” Now, poverty relief depends on laying claim to “worker” status. Second, massive new benefits have been directed to low-income families with children and a working adult. Pro­grams like the EITC and the Child Care and Development Fund (CCDF) provide “work supports” to the “working poor.” No longer does policy assume that “workers” can be left to live off market earnings alone. Whether old benefits are withdrawn from those who do not work or new benefits are granted to those who do, access to poverty relief now requires passing a work test in addition to a means test. That is why it matters so much that these policies do not see family caretaking as a form of work.

To end child-care invisibility, I propose that we begin accounting for care. First, as I argue below in Part I, we should fully incorporate child care into the needs and resources prongs of means testing. Like food or shelter, child care should count among the basic needs that any household with chil­dren ought to have the resources to meet. In contrast, existing approaches treat child care as a means to the end of parental employment, not an end in itself.[7] When these child-care needs are met through nonmarket care, we should count as a resource the household’s savings from not having to spend income to hire third-party care, which in turn lowers its need for government assistance.

Second, as I argue below in Part II, accounting for care also transforms the work-testing component of antipoverty policy. Nonmarket caretaking should satisfy work requirements when it meets child-care needs that other­wise would be met through government transfers. This conclusion follows directly from applying a standard theory of work requirements that treats work as a form of self-support that limits the need for government assis­tance. This theory has been thought to privilege market employment and exclude nonmarket care as nonwork, but that result no longer follows once child care is made visible within means testing. Thus, while I build on long­ standing feminist arguments for recognizing caretaking as valuable work,[8] my approach breaks new ground by relying on a theory of work tailored to means-tested benefits.

Tailoring my argument to antipoverty policies allows me to meet objec­tions that have greater force against broader programs of supporting child care. It also counters a tendency for conversations about gender, work, and family to revolve around care in relatively high-income (or potentially high-income) households headed by a cross-sex couple.[9] There, basic subsistence is not at stake and the relationship between earnings potential and child-care costs may be quite different. In complementary fashion, analyses of anti­poverty policy too often focus solely on single mothers, thereby missing double standards for different family structures and divisions of labor,[10] an issue I emphasize here.

I. Accounting for Care When Assessing Financial Need

This Part analyzes how child-care invisibility functions within the means-testing component of antipoverty policies. This aspect of child-care invisibility disadvantages employed single parents relative to those who care for their own children. Among two-parent households, it disadvantages dual-earner couples relative to those with a breadwinner/caretaker division of la­bor; within the latter, it disadvantages the caretaker. These problems arise because means-testing systems were built on a “family wage” model of so­cial policy, one that assumes a breadwinner/caretaker division of labor and focuses exclusively on cash transactions. Truly fixing these problems re­quires recognizing that child care is needed, and provided, regardless of whether all parents are employed or whether there is an at-home caregiver; the differences concern only shifts between nonmarket and market forms of care. In contrast, conventional solutions utilize the half-measure of treating child care as an issue only for employed parents; when a parent stays at home, child care becomes invisible again.

A. How Child-Care Invisibility Works, and Who Gets Hurt

A means test compares what a household needs with its resources avail­able to meet those needs—its means. A “standard of need” represents the minimum income on which a household can be expected to live, and from this one subtracts available income. A cash transfer makes up any differ­ence. A longstanding concern about means-tested benefits is that they re­duce the financial incentive to increase earnings. The standard economic model treats means testing as an implicit tax because when earnings rise, transfers decline. The additional earnings essentially get split between re­ducing government transfers and going into the earner’s pockets.

If, however, increasing employment also means losing the ability to provide child care oneself, then not only do transfers decline, but the parent also must make alternate child-care arrangements. If those arrangements cost money, then wage-earning parents can easily end up financially worse off than non-earners, despite a significantly higher cash income. This prob­lem is especially acute for low-wage workers, because child-care costs can easily approach or exceed earnings from even full-time work. Research in the early 1990s showed that these and other “costs of working” meant that single mothers receiving public assistance could not break even financially unless they got full-time jobs paying one-and-a-half to two times the mini­mum wage.[11] For the vast majority of those able to make ends meet with wages alone, a crucial reason was that they spent zero money on child care because they had older children, access to free care, or the ability to match work and school hours.[12]

This counterintuitive result of earning more but having less arises from two related ways in which child care is structurally invisible within conven­tional means-tested transfer programs. First, child care is not included when setting the standard of need. Consequently, regardless of whether house­holds contain only employed parents or instead include an at-home caregiver, existing policies treat them all as having the same financial needs, which do not include child care. Second, employed parents are treated as necessarily having higher incomes available to meet these fixed needs, not accounting for how a household with an at-home caregiver obtains valuable child care without having to pay for it in cash.

In combination, these two aspects of conventional means testing lead to overstating economic well-being and therefore understating the need for transfers. This dynamic disadvantages families that have to pay for child care relative to those with a family caretaker. Among single-parent house­holds, the result is a clear disadvantage for employed parents.

Similarly, among two-parent households, child-care invisibility disad­vantages dual-earner couples relative to those with breadwinner/caretaker divisions of labor. The former have a higher cash income, but their higher child-care expenses remain invisible. More subtly, this result may be bad individually for caretakers in two-parent households, notwithstanding the benefit at the household level.[13] Holding the couple (and the breadwinner’s employment) constant, the means test conditions transfers on the caretaker remaining out of the labor market, as with a single caretaker. Additionally, holding caretaking constant, the work test conditions her access to transfers on remaining coupled to the wage-earner, for reasons described below.[14]

B. Child-Care Invisibility as a Legacy of the Family Wage

Child care’s invisibility is no accident. To the contrary, it reflects the integration of poverty measurement and poverty relief into a broader social policy regime constructed in the early 20th century to support a family wage system.[15] This regime rests on an idealization in which citizens dwell in families headed by a married cross-sex couple, one of whom (guess which) is (at least primarily) a market wage-earner while the other (guess which) is (at least primarily) their children’s caretaker.[16] The “breadwinner” is the “American Worker” whose market production brings resources into the household where they are consumed by his “dependents.”[17]

Because economic distress is treated as a function of the breadwinner’s wage, social policy responds by bolstering that wage. This may be done directly through efforts to create jobs and raise wages, or indirectly through wage-based social insurance like workers’ compensation, social security, and unemployment insurance. The latter efforts are structured to mimic insur­ance purchased by the wage-earner to provide for his dependents in the event he can no longer bring home the bacon.

“Welfare” has been the residual response when a household lacked ac­cess to a breadwinner’s wage or wage-based insurance. Even here, however, the payments have been figured as substitutes for the absent wage.[18] Thus, AFDC was explicitly designed to enable “deserving” white mothers to pro­vide care and remain outside the labor market, thereby fulfilling their role within the family wage system.[19] Eligibility required both the absence of a male wage-earner and the mother’s maintenance of a “suitable home.”[20]

The influence of family wage norms dovetails with a conception of household economic well-being grounded in cash purchases in consumer markets. Households are poor when they lack the money to buy things. This understanding of poverty reflects broader tendencies to identify eco­nomic activity with markets and to identify social citizenship with participa­tion in a “consumer’s republic.”[21] If child care is assumed to be provided by the mother, then child care is excluded from household needs measured in cash. And because caretaking (and housework generally) produces no cash for use in consumer markets, it adds nothing to a cash measure of resources available to meet household needs.

Fingerprints of this consumerist family wage model lie all over the ori­gins of current poverty measurement. When the first federal poverty thresh­olds were developed in the early 1960s, officials based them on surveys of actual expenditure patterns.[22] Child care was virtually absent from these family budgets because employment was a relative rarity among mothers, especially married ones with young children,[23] and even then care almost always was provided without charge by a relative.[24] Since then, official pov­erty measures have been adjusted only for inflation, not for changes in the composition of household spending.

In this way, the work/family patterns of the 1950s literally remain built into contemporary poverty measurement. As a result, a household of three adults has the same poverty line as a household of one adult and two chil­dren, and the latter has the same poverty line regardless of whether the adult is available as a caretaker. Variations in child-care costs among these three-person households are invisible to official determinations of economic need.[25]

Child care’s invisibility in poverty measurement carries over into as­sessments of financial need that determine eligibility for antipoverty pro­grams. Federal poverty levels themselves often are the basis for eligibility thresholds,[26] and other, independent metrics used to evaluate financial need have similar structures that ignore child care. For instance, TANF financial eligibility standards typically aim to establish whether a household has “suf­ficient income to meet their most basic needs” or the like.[27] Yet child care never appears in any state’s explanation of what those basic needs include,[28] even when they provide detailed lists including not only housing, clothing, food, utilities, and health care, but also “recreation”[29] or even “scout uniforms.”[30] Similarly, a current federal regulation characterizes TANF pay­ments as “designed to meet a family’s ongoing basic needs (i.e., for food, clothing, shelter, utilities, household goods, personal care items, and general incidental expenses).”[31] Again, child care is nowhere to be seen.

C. Child Care as a Basic Need

Fresh approaches to policy redesign open up once we trace means test­ing’s pathological treatment of child care to its grounding in a family wage model. The solutions lie in rejecting the assumptions both that child care is provided outside the market and that markets are the exclusive determinant of a household’s economic well-being. This section argues for taking the first of these two steps by treating child care as a basic need of all house­holds with children. In contrast, the more familiar, but less justified, ap­proach treats child care as a “work expense” for employed parents.

1. Basic Needs Budgeting for All Families

Imagine if poverty measurement and antipoverty policy began with a household in which all adults are employed full time. Child care would seem indistinguishable from other purchased necessities like food, clothing, and shelter. That is precisely my proposal: simply treat child care as some­thing that all households with children need.

Others have taken a step in this direction with “basic needs budgets” or “self-sufficiency standards.”[32] These incorporate employed parents’ child­care expenses into calculations of how much income a household needs to meet its basic needs. Of course, these expenses vary by the number, age, and health status of the household’s children, and so the amount added varies accordingly. These proposals fall short, however, because they remain anchored in the cash economy of market earnings and market spending.[33]

Child care becomes visible only when all adults are in the labor market or otherwise systematically unavailable to provide care. With a nonmarket caretaker present, child care again disappears from the analysis.

This market-centered approach to child care mirrors an early poverty measurement approach to subsistence food production by farm families. Be­cause farm families purchased less food, they were deemed to have lower household needs, reflected in lower poverty thresholds that persisted until 1981.[34] By analogy to child care, one might treat food purchases as a “work expense.” After all, only when farmers stop farming and enter the labor market do they need to buy groceries for dinner, and so only then is food included in the tally of household needs. By this logic, families that farm do not need to eat, just like families that include a nonmarket caretaker do not need to ensure their kids are cared for. This is silly. What varies in these cases is not a household’s economic needs, but only whether those needs are met through nonmarket production or market consumption.[35] Accordingly, I would extend the basic needs budgeting approach by incorporating child care into the standard of need for all households, not only those that lack a nonmarket caretaker. Whether or not you grow your own, we all need to eat.

2. Against the “Work Expense” Model

Alongside work-based welfare reform, a consensus emerged that lack of child care could justify a transfer recipient’s failure to work and that, to avoid this situation, public resources should be devoted to providing access to child care. The precise contours of these principles remain controver­sial,[36] but even the Heritage Foundation concedes that “if low-skilled single mothers are moved into the labor force, child care assistance must be pro­vided.”[37] Funding for child-care assistance to low-income households has expanded massively in tandem with the growing emphasis on work.[38]

This work-expense framework yields two kinds of institutional re­sponses. The first deducts some or all child-care expenses from gross earn­ings; the resulting net income becomes the measure of household resources for the purpose of poverty measurement or means testing.[39] The second, now dominant, approach leaves means-tested cash transfers untouched but creates a separate child-care benefit tied to employment. Examples include child-care subsidies to TANF recipients participating in mandatory work ac­tivities, “transitional child care” to former recipients who become income-ineligible due to increased earnings, and child-care assistance to low-income workers through CCDF and related programs. Rather than providing addi­tional cash to cover child-care expenses, direct child-care assistance relieves parents from needing to spend out-of-pocket.

Both approaches shift child-care costs from employed parents to the state. But if not to meet basic needs, then why? The usual rationale is some variant on “supporting work” or enabling transfer recipients to comply with work requirements. These rationales lose track of why work is promoted within a means-tested transfer system. Instead, antipoverty policy slides to­ward making employment an end in itself.[40]

The simple problem is that child-care assistance is enormously expen­sive. For a household with a toddler and a school-age child, a typical state CCDF program would authorize child-care subsidies totaling about $11,500 per year.[41] This roughly equals the median maximum household benefit that a household of three without any income would get from TANF and SNAP combined.[42] It also roughly equals full-time, full-year earnings at the mini­mum wage, net of federal payroll taxes.[43]

Linking child-care assistance to employment collides with the theory that employment reduces reliance on transfers. What is the point of these child-care expenses? To support work. And why is work important? Be­cause it reduces transfers in favor of self-support.[44] Except that work may not reduce transfers once child care becomes visible and work triggers new transfers. This contradiction is masked by keeping a separate set of books for cash assistance (which employment reduces) and child-care assistance (which it increases).[45]

These tensions also contribute to practical failings of our existing child­care system. Treating child-care subsidies as an instrument to promote em­ployment, not as part of a floor beneath living standards, weakens parents’ claims to these benefits. Indeed, nationally only about fifteen percent of households financially eligible for child-care subsidies receive them, largely because of lengthy waiting lists.[46] Assistance tends to be targeted at those on the margin between seeking or maintaining employment and receiving TANF transfers.

A focus on altering employment status neglects low-wage workers who never passed through TANF, even if child-care costs leave them similarly squeezed financially.[47] If the goal is to promote employment, that neglect is perfectly justified; indeed, assisting those who would hold down jobs any­way would confer an ineffective “windfall.”[48] Yet we do not consider it a windfall if transfers enable an adult to live in better housing, or allow a parent to reallocate rent money to food or clothing, even if these things make no difference to her employment. That is because housing is considered a basic need, and transfers aim to allow people to consume at least a certain amount of it without having to skimp on other basic needs, like food or clothing. The same seems true when a parent somehow manages to hold down a job without child-care assistance, yet receiving it would allow her to upgrade child-care arrangements or free up funds to spend on other needs. We should see in the depth and breadth of support for child-care assistance to low-income parents an implicit acceptance of child care as a basic house­hold need.

D. How Means-Testing Can Account for Care

Treating child care as a basic need alters both components of means testing: setting a standard of need and measuring the resources available to meet those needs. This section sketches how both components can account for care.

1. Setting the Standard of Need

Treating child care as a basic need is straightforward with regard to the household needs side of means testing. Instead of one cash assistance pro­gram that ignores child care and a separate child-care assistance program for some employed parents, there should be a single program with a comprehen­sive standard of need, child care included. In practical terms, this means integrating what are now separate systems for providing income support through TANF and child-care assistance through CCDF.

Integrating cash and child-care assistance can be accomplished by tak­ing existing standards of need and adding to them the reimbursement rates already authorized for child-care assistance. Schematically, if existing cash assistance programs assume that a household needs $10,000 per year, exclu­sive of child care, and existing child-care assistance programs authorize gov­ernment expenditures of $6,000 per year to provide care for a low-income worker’s child, then the two can be combined to yield a $16,000 standard of need. Existing reimbursement rates vary with children’s age, special needs, and types of care, and so too would the new standards of need. They would also reflect how needs vary with the number of children. In the example above, a second child would raise the standard of need to $22,000. Of course, there’s nothing magical about existing standards of need and reim­bursement rates. They could be too high or too low, but those concerns are distinct from the question of integration.

Because it builds on existing programs, this change alone would not alter the financial bottom line for any household receiving cash or child-care assistance. Instead, its practical effect would be to expand child-care assis­tance to households who today would be waitlisted despite financial eligibil­ity. That is the import of putting child care on par with food, shelter, and other basic needs. In this regard, it would be equivalent to fully funding CCDF, as many have called for.[49] Important as that change would be, the real innovation lies in the interaction between changing the standard of need and other policy elements, to which I now turn.

2. Measuring Available Resources

The next step addresses the household resources side of means testing. Here, the challenge arises from implementing the proposition that child care does not vanish as a basic need when a family caretaker is present. Instead, that need is met by the caretaker directly, not through purchasing care as a consumer service. Correspondingly, the means-testing calculus must widen its net beyond market transactions in order to capture the household’s nonmarket resources for meeting its needs.

Failing to account for this nonmarket resource would generate a wind­fall in transfers to households that include a nonmarket caretaker. For those households, measuring resources in cash alone would generate a transfer large enough to pay for child care and other needs. However, with a care­taker present who spends her time, but not her money, on child care, the portion of the transfer designated for child care ends up left over, even after all other basic needs are paid for. Such windfalls arise whenever resources that the household can use to satisfy its needs are not counted in the means test’s measurement of income. Understating resources leads to overstating the need for assistance.

That windfalls result from unmeasured income provides a clue toward preventing them. Consider the more familiar example of food. Like $200 in cash, a $200 gift certificate to a grocery store should count as “income” because it substitutes for $200 in transfers otherwise needed to purchase food.[50] Moving further from cash toward in-kind resources, the same analy­sis applies to a gifted shopping cart full of $200 in essential groceries. Not counting any of these as income would create a $200 windfall relative to what the transfer is designed to provide.

Windfalls to nonmarket caretakers can be avoided using the same basic techniques that capture in-kind income. Like a full shopping cart and food expenses, nonmarket caretaking allows the household to avoid child-care ex­penses. Therefore, the solution is to credit the household with “imputed income,” an accounting device that treats avoided expenses as equivalent to income used to pay the expense.[51] In this case, the amount of imputed in­come from caretaking would equal the amount designated for child care in the standard of need.

Placing a value on nonmarket care in this way follows from defining “income” for means-testing purposes as resources available to satisfy desig­nated needs. The resulting figure does not imply any abstract claim about the “true value” of the care provided, nor even its market value. Although it can be used to obtain services with a market value of $200, a $200 gift certificate to a pet store would count as zero income because it does not help meet any recognized basic needs, assuming pets are not included.[52] More subtly, a shopping cart full of caviar and filet mignon might cost $1000 but count only as $200 in income if it could satisfy only $200 worth of a basic food budget. Similarly, imputed income from caretaking will equal the max­imum amount the government would otherwise be willing to pay for third-party care, even if the parental care provided is arguably better or if third-party care is arguably undervalued in the market.

Imputing income in this way avoids any rigid commitment to treating caretaking and employment interchangeably. Such a commitment character­izes an opportunity cost approach to valuation that sets imputed income from caretaking equal to whatever could have been earned in employment instead.

That method, however, takes no account of what the caretaker actually does. The opportunity costs are the same whether keeping an eye on one teenager or juggling the diapering, feeding, and napping of three pre-schoolers. In­deed, they are the same for caretaking and any activity, no matter how triv­ial, that competes for time with employment.

Instead, by anchoring imputed income to the standard of need, my ap­proach aims to capture realistically the trade-off between devoting time (but not money) to child care versus devoting time to employment and its earn­ings to child care. For a household with one sixteen-year-old who has no special needs, caretaking would yield no imputed income because existing child-care programs would not fund any care at all (thirteen is the usual age cutoff)—reflecting the judgment that adult supervision outside school hours is not necessary, even if it may be desirable. In contrast, the imputed income from caring for three pre-schoolers might well exceed the parent’s opportu­nity costs in the labor market. It all depends on the number, age, and health status of the children, and the price and quality of acceptable child care. How this compares to full-time employment further depends on a parent’s earnings capacity. Generally speaking, for low-wage workers, wage income will be roughly equivalent to imputed income from caring for two preschool-age children without special needs.

Accounting for Care When Assessing Work

Part I proposed accounting for care within the means test that deter­mines financial eligibility for transfers. This Part extends the analysis to work requirements. The two are joined at the hip because income is an input to the means test and an output of work.[53] Just as traditional means testing focuses on spending cash in consumer markets, traditional work require­ments focus on earning cash in labor markets. Likewise, incorporating nonmarket income and consumption into means testing provides the founda­tion for incorporating nonmarket production into work requirements.

A. The Invisible Work of Family Caretakers

Exactly what activities satisfy current work requirements in TANF and related programs is quite a bit messier than often assumed.[54] Nonetheless, paid employment clearly provides the paradigm, both as work requirements are formalized in policy and as they are discussed in scholarship and public debate. It is equally clear that policymakers have invoked “work” to justify denying transfers to family caretakers. With only the narrowest and most tenuous of exceptions, caring for a member of one’s own household never counts as work in contemporary welfare policy.

In its raw form, the charge against caretakers is that they are sitting at home doing “nothing.” Nothing in what sense? The most common formu­lations invoke doing nothing to make the household more “self-sufficient,” or nothing “productive” that “contributes to society.” Either way, this “nothing” charge dovetails with child care’s economic invisibility in the ma­chinery of means testing and poverty measurement. A cash accounting sys­tem attributes zero income to the family caretaker, in contrast to the thousands that market employment would generate. The EITC does this quite literally by using “earned income” as its method of identifying work.[55] Thus, the family caretaker is no different than the proverbial Malibu surfer who looms large in philosophical debates over linking redistribution to work.[56] Neither brings in cash, neither produces anything valued in the mar­ket, and so, neither works.

B. Defining Work As Meeting Basic Needs

Assessing the “doing nothing” charge requires a theory of work re­quirements.[57] Many virtues are attributed to employment, but which provide the “something” that makes it work and not “nothing”? This section offers an answer distilled from leading arguments for work requirements and struc­tural features of actual policies. Rather than justifying work requirements from the ground up,[58] I aim to challenge proponents to embrace the implica­tions of their views or, if not, go back to the drawing board.

I begin with a simple, familiar rationale for means-tested transfers: so­ciety owes support to people who otherwise could not meet their basic needs through no fault of their own. Faultlessness includes having made reasona­ble efforts at self-support through one’s own work, where possible.[59] This framework justifies transfers both to low-wage workers and to the involunta­rily unemployed.

An expectation of reasonable self-support integrates means testing and work testing at both theoretical and practical levels. A means test identifies those in need (and its extent) by comparing resources to some threshold. The relevant concept of “need” necessarily incorporates some level of re­sponsibility. It will not do to plead poverty if, despite a lack of any current income, one could simply go to the bank and withdraw oodles of cash. This basic point is reflected in asset tests on means-tested transfers.

Work requirements capture the idea that individuals also can “spend” their time on activities that help meet their economic needs. Someone who could work but does not is someone who has the means to satisfy his basic needs, notwithstanding low income. This point helps explain why work tests are so specific to means-tested transfers and not, say, voting. Similarly, work tests are stronger and more deeply entrenched than other kinds of be­havioral requirements that some might include within a broad definition of responsible citizenship.[60]

This connection between work and need operates automatically within a system of cash accounting. When need is reduced to cash purchasing, work can be reduced to generating cash income. The only way to meet needs is to acquire cash, and all cash can be used to meet the relevant needs.

As noted above, however, means tests have always sought to capture at least some noncash resources. And when in-kind receipts should count as income, the activity producing them should count as work. If an employee gets paid in free rent rather than cash, she still has earnings and still works,[61] though the issue rarely arises because minimum wage law generally man­dates cash payment. Also consistent with this principle, Vermont law ex­plicitly provides that self-employment activity may satisfy TANF work requirements if the participant “earn[s] in-kind income” and “the in-kind income reduces the family’s . . . financial assistance grant.”[62]

This analysis applies with equal strength when a household member produces goods or services that are consumed within the household. Obvi­ous examples include hunting, farming, and knitting or sewing clothes. There is no functional difference between employment paid with a $200 gro­cery store gift certificate and subsistence farming that produces $200 worth of food. Along these lines, Alaska includes “subsistence activities” as TANF work,[63] as do several Native American TANF programs.[64]

This approach to identifying nonmarket work draws on, but is narrower than, techniques for identifying nonmarket work based on its substitution for market production. In the means-testing context, what counts is not all eco­nomic production but only production of these resources that the worker’s household can use to meet its basic needs. A volunteer firefighter substitutes for a paid firefighter and thereby would contribute to a comprehensive GDP, but this does not put food on her table. Similarly, walking one’s own dog substitutes for hiring a dog walker; it thereby contributes both to GDP and to household well-being, but it does nothing to meet basic household needs unless these include having and maintaining a pet. Accordingly, unlike the farmer, neither the volunteer nor the dog walker would satisfy work require­ments on means-tested transfers, even though all three are in some sense engaged in nonmarket work that substitutes for market production.

Contribution to self-support thus provides a criterion for identifying policy-relevant work that neither obeys nor obliterates a distinction between market and nonmarket work.[65] Instead, what matters is something that can be found systematically in market work yet is not found exclusively there. Nor is it found willy-nilly wherever there is productive activity. What matters is the specific way that work is woven into the household economy, something irreducible to cash transactions.

C. How Work Requirements Can Account for Care as Self-Support

Treating nonmarket caretaking as work under work requirements rises or falls with treating child care as a basic need under means testing. By providing care herself, a nonmarket caretaker substitutes her work for what a market child-care provider would otherwise be hired to produce.[66] Such market/nonmarket substitution supports characterizing caretaking as eco­nomically productive work, but that does not yet tell us whether it is the kind of work demanded by work requirements.[67] That requires knowing whether nonmarket production helps meet household needs and thereby substitute for transfers. If, as the previous Part argued, we should treat child care as a basic household need, then we should treat its nonmarket production as work for the same reason we treat it as a source of (imputed) income.

Tying work status to self-support offers an important variation on more familiar feminist theories of nonmarket care work as a basis for transfer receipt. These theories generally rely on the point that nonmarket care is socially valuable, even though the caretaker does not get paid.[68] Producing a social contribution triggers a claim to share in the fruits of this labor (or to relief from the burdens of producing them).[69] On that analysis, the resulting transfer is analogous to a worker’s wage, a share of what she produced. For this reason, social contribution theories typically point toward economic claims by all caretakers, regardless of their poverty status.[70]

In contrast, my argument is that low-income caretakers are members of the working poor: they lack the resources to meet basic needs despite rea­sonable efforts at self-support. The caretaker covers child care directly but then needs cash transfers with which to buy other necessities. The low-wage worker generates cash to purchase those same necessities, but at the cost of needing yet more cash to obtain child care. Either way, “work alone is not enough,”[71] and the government steps in to make up the difference with trans­fers. Such transfers to a low-wage employee do not compensate her for the value of her work, nor do they do so for the low-income caretaker. To insist otherwise mistakenly reduces all economic transactions to “nothing but” im­plicit bargains.[72]

Using self-support as the basis for treating nonmarket caretaking as work avoids a series of difficulties associated with broader social contribu­tion arguments. Amy Wax has rightly argued that treating caretaking as work risks opening the floodgates to a ridiculously broad definition of work. Avoiding that result requires a principle that can distinguish a caretaker from an amateur who “works hard” producing third-rate paintings in his garage, even if the painter genuinely values the product and would otherwise have spent money purchasing substitute art.[73] A theory of self-support provides that principle.[74] Providing one’s own child care is not like providing one’s own art, or walking one’s own dog. If it were, then we should not have means-tested transfers designed to allow low-income parents to obtain child care, just like we do not have transfers designed to allow them to buy art, or hire dog walkers. Self-support also distinguishes child care from other un­paid but productive work that contributes to society but does not substitute for transfers to the worker,[75] and it avoids the need to interrogate either the magnitude of these contributions or the sufficiency of any compensating pri­vate gains.[76] In a delightful twist on gender stereotypes, the caretaker is best compared not to the selfless volunteer but to the yeoman farmer.

Additionally, because work is linked to qualifying for anti-poverty transfers, there is no need to defend the controversial normative claim that workers always deserve a share of their product, even when they work vol­untarily without promise of such a share. This point also sheds light on the common and potent objection that public support for low-income nonmarket caretakers foists onto taxpayers the consequences of caretakers’ decisions to have children.[77] On that view, parents have chosen to create the “need” for child care, and so they should not get credit for meeting that need; child care is more like walking one’s dog than growing one’s food. Crucially, this ob­jection is not specific to counting caretaking as work. Instead, it applies equally when taxpayers foot the bill for hiring third-party child care. Em­bracing means-tested child-care assistance (like food assistance but unlike pet assistance) means already crossing that bridge.[78]

How Much Work?

Programs that support the working poor provide transfers when work falls short of satisfying all household needs. Current policies accept full-time, minimum wage employment as a sufficient effort at self-support. They do not insist on sixty-hour weeks or obtaining a $12 hourly wage. For in­stance, TANF states work requirements in weekly hours, typically in the thirty- to forty-hour range. To address the possibility of someone just killing time to satisfy work requirements, it mandates time- and performance-moni­toring measures for unpaid activities that qualify as work.[79] For paid em­ployment, the system relies on the employer’s own monitoring and obligation to pay the minimum wage.

Nonmarket caretaking presents obvious challenges to this temporal ac­counting system, but the methods to tackle them have been developed to address similar difficulties with self-employment. Like caretaking, self-em­ployment lacks the hierarchically imposed time discipline and separation of “working time” from “personal time” characteristic of employment.[80] To overcome these challenges, TANF rules permit imputation of time from in­come. These rules attribute to the self-employed individual hours of work equal to her net business income divided by the minimum wage.[81] Under that standard, self-employment that yields an annual net income of roughly $12,000 would be treated as full-time work. This method makes it unneces­sary to disentangle various uses of time, so long as the transfer recipient gets the job done, that is, earns the money that offsets transfers. Inversely, time and effort count for nothing without economic results. Someone who spends sixty hours a week on a business that earns only $6,000 a year would receive credit for only twenty hours per week of work.

Applying these methods, nonmarket caretaking can be integrated into hours-based work requirements. Hours worked would equal imputed in­come divided by the minimum wage. If caretaking substitutes for $12,000 in annual transfers to cover child-care expenses, it is equivalent to full-time work. But if it substitutes for only $6,000, the caretaker would receive credit only for half-time work. In this way, the techniques for quantifying income carry over to attributing hours of work. As a result, caretaking satis­fies work requirements to the same extent as a minimum wage job with an equal contribution to self-support.

At this point we can revisit the objection that allowing caretakers to collect transfer payments creates a perverse incentive to have children. As noted above, this objection seems inconsistent with supporting child-care assistance; that, too, shifts to the public some of the costs of parenthood, thereby arguably distorting any cost/benefit analysis toward having (more) children. But perhaps what really animates the objection is the suspicion that low-income parents do a poor job of parenting.[82] That concern could distinguish parental from third-party care, and it would provide a particularly pointed version of the accusation that caretakers are “sitting at home doing nothing.” Rather than representing a general denial that nonmarket caretak­ing can contribute to self-support like market employment can, this objection relies on a specific denigration of the care provided by low-income parents who receive government assistance. In terms of the preceding discussion, the objection would be that the care provided personally by low-income par­ents is systematically of lower quality than that delivered by third-party care funded by child-care assistance programs, and, furthermore, that this justi­fies minimizing the former on child welfare grounds.

There are legions of reasons to be skeptical of, or even hostile to, this objection. First, it reeks of the long history of race and class bias against the child-rearing practices of poor women and women of color.[83] Perhaps for that reason, it is rarely among the first volley of arguments against treating caretaking as work, arguments that I have met forcefully above. Second, it has no basis in evidence. Research has not found any significant differences in parenting behavior between those who rely on government aid and low-income parents relying on employment income.[84] Nor is there evidence that when work requirements induce shifts from caretaking to employment, any positive effects on child development result, except those mediated by higher income.[85] Finally, the objection is inconsistently applied because, as discussed below, current work requirements are structured to encourage caretaking by one parent in a two-parent couple, so the same women deemed incompetent when single become happy homemakers if only they marry.

Two-Parent Couples

Accounting for care within work requirements has some counterintui­tive effects when applied to two-parent families. In the simple case where both parents are required to work, the effects track those for single-parent households. Nonmarket caretaking shifts from being a basis for disqualifica­tion by work requirements to one acceptable method of satisfying them.

What complicates matters is that work requirements generally do not apply to both parents. Instead, both TANF and the EITC primarily utilize what I have labeled elsewhere a “breadwinner priority” structure.[86] They require only one adult per household to work, even when there are two or more adults. Contrary to the widespread interpretation that welfare reform required low-income mothers to substitute employment for caretaking, in­stead it has required them to be part of “working families” and permitted a breadwinner/caretaker division of labor. My proposal to account for care would continue to permit caretaking within a two-parent family, but it would treat caretakers as satisfying work requirements rather than exempted from them.[87] The household is eligible for transfers either way, but the shift is consequential in several respects.

First, treating caretaking as work provides a basis for reconciling as­pects of the existing breadwinner priority policy with a self-support theory of work requirements. Allowing a married parent to be a caretaker and re­ceive need-based transfers is difficult to justify if caretaking is not work when single parents do it. On that view of work, in both cases a caretaker is an adult who could bring in resources that would reduce transfers but fails to do so. In contrast, my analysis treats both the caretaker and the wage-earner as contributing to self-support, regardless of whether they are married.

Second, a self-support analysis implies that sometimes both parents should be expected to enter the labor market, something the breadwinner priority model never does. If a family’s only child is sixteen, the state will not provide child-care assistance even if both parents have jobs; therefore, staying home as a caretaker does not further self-support. Allowing it none­theless suggests an inappropriate thumb on the scale in favor of a (likely gendered) breadwinner/caretaker division of labor.

Finally, integrating caretaking into work requirements may expand the opportunities of parents, most likely mothers, whose caretaking in part re­flects labor market disadvantage. Subjection to work requirements is closely connected to eligibility for work supports, not only child care but also earn­ings incentives, education and training, and so on. Currently, potential “sec­ondary earners” in two-parent families risk being cut off from such supports when there is no expectation that they work at all.[88]

Conclusion: What’s Left of Work Requirements?

My proposal obviously calls for a substantial change in the way work requirements currently are structured in major programs designed to assist low-income parents. Given the political and academic popularity of those work policies,[89] skepticism, even scorn, is predictable and understandable. That granted, my approach has three features that differentiate it from a sim­ple recycling of old debates about work and care, features that could create some space for new thinking. First, it integrates caretaking into work re­quirements in a principled fashion; it does not reject work requirements or propose exemptions from them.[90] Second, it treats caretaking as work for reasons that are theoretically and practically intertwined with supporting child-care assistance for employed parents.[91] Third, it integrates the treat­ment of work and care in both single- and two-parent households, creating the potential to leverage policies and politics that already support caretaking in the latter.[92] These features yield a policy design that levels up, maintains meaningful work requirements for all while toughening them for some, and is driven by child care.

My proposal levels up by providing equivalent support to caretakers and low-wage workers, both of whom work hard to support their families but still need assistance to make ends meet. Doing so meets the politically and morally powerful demand for “class parity” with regard to work and care—that is, for implementing antipoverty policies that do not give transfer recipients life options unavailable to those with too much income to qualify for assistance. Existing work requirements arguably fulfill this demand by “leveling down,” driving down support for low-income family caretakers to match the stinginess of child-care assistance for low-income employed par­ents. I agree that support for caretaking and child-care assistance should be in sync, but by providing both rather than neither.

Because of their grounding in a principle of self-support, work require­ments still have teeth under my proposed approach. Caretaking does not automatically qualify as work, and even when it does it may not fully satisfy work requirements. The particulars will depend on the number of children, their age, and special needs. If caretaking contributes less to self-support than employment would, additional work activities would be necessary to satisfy work requirements. In contrast, current policies often neither require nor support any work at all from one adult in a married couple, regardless of the intensity of her caretaking responsibilities or the possibility that employ­ment would contribute more to self-support.

Finally, my approach is grounded in the deep support for child-care assistance, an area where costly expansions in need-based transfers have re­mained politically popular.[93] So, too, is the idea that parents should get to control basic child-rearing decisions. Combining the two creates an opening for the idea that low-income parents should be able to choose between get­ting child-care assistance while they earn money and getting money while they provide child care.

Real-world policy initiatives support this analysis. For instance, during the 2008 presidential campaign, then-Senator Hillary Clinton proposed “to allow qualifying, low-income parents who want to stay at home rather than place their children in child care to receive child care subsidies through the Child Care Development Block Grant”; she argued that doing so was part of “giving every parent a real choice about whether to stay home and care for their children full time or work outside the home.”[94] Clinton never provided specifics, but she appears to have been building on a program known as At­ Home Infant Caregiver (AHIC), which provides cash to low-income, at-home caretakers of infants in lieu of child-care assistance. Minnesota, Mon­tana, and New Mexico all have established AHIC programs, and authoriza­tion for a federally funded pilot program was included in a TANF reauthorization bill that was approved by the Senate Finance Committee in 2002 with bipartisan support, though it did not become law.[95]

Although quite limited,[96] AHIC contains the seeds of a more ambitious restructuring of antipoverty policy along the lines I advocate. In particular, it suggests an institutional path that begins with the existing, popular child­care programs and grows from there, rather than one that begins with the highly stigmatized “welfare” program and attempts to revitalize it.[97]

Essential guideposts along this path are framing the issue in terms of parental choice, linking caretaking to child-care assistance for low-wage workers, and integrating treatment of single- and two-parent households. Among other things, this framing may help disrupt the toxic racial and cul­tural politics triggered by the association of “welfare” with single mothers of color. Those politics will never go away, of course, but people of good conscience should neither shy away from that fight nor wage it unarmed. Progressives too often have done so by ceding the discourse of work, re­sponsibility, and self-reliance to the right. In this way, I hold out some hope that we can embrace the cause of the working poor yet fend off its evil twin, abandonment of those who work outside conventional labor markets.

---

[asterisk]Professor of Law, UCLA School of Law. This paper derives from one written and presented while I was a fellow at the Program in Law and Public Affairs at Princeton Univer¬sity, and I gratefully acknowledge the material and intellectual support LAPA and its commu¬nity provided. I also have benefitted greatly from comments on related work from Robin Lenhardt and Viviana Zelizer and from workshop participants at UCLA School of Law, the Washington University, St. Louis workshop on Work, Family, and Public Policy, and the 2009 Emerging Family Law Scholars & Teachers Conference, as well as from the editors of the Harvard Law & Policy Review. Invaluable research assistance was provided by Elizabeth Adams, Kelly Hoehn, and Melissa Rose.[/asterisk]

[1]See, e.g., Nat’l Women’s Law Ctr., How The Senate American Recovery and Re­investment Plan Addresses Women’s Needs 4 (2009); Linda R. Hirshman, Op-Ed., Where Are the New Jobs for Women?, N.Y. Times, Dec. 9, 2008, at A35; Jennifer Barrett, More Stimulating, Slate, Feb. 5, 2009, http://www.slate.com/id/2210535 (on file with the Harvard Law School Library).
[2]Staff of H. Comm. on Ways & Means, 110th Cong., 2008 Green Book 9-43, chart 9-7 (Comm. Print 2008), available at http://democrats.waysandmeans.house.gov/singlepages.aspx?NewsID=10490.
[3] Id. at 7–46.
[4] Molly Hennessy-Fiske, L.A. County Officials Offer a Novel Idea to Save Millions, L.A. Times, June 17, 2009, at A9; see also Erik Eckholm, California’s Zigzag on Welfare Rules Worries Experts, N.Y. Times, Oct. 7, 2009, at A1.
[5] Peter S. Goodman, Cuts to Child Care Subsidy Thwart More Job Seekers, N.Y. Times, May 24, 2010, at A1.
[6] Nancy Folbre, Valuing Children: Rethinking The Economics Of The Family 45–64 (2008).
[7] See, e.g., Nat’l Research Council, Measuring Poverty: A New Approach 207 (1995); Diana May Pearce, The Self-Sufficiency Standard: A New Tool for Evaluating Anti-Poverty Policy, 10 Poverty & Race 2 (2001); Trudi J. Renwick & Barbara R. Bergmann, A Budget-Based Definition of Poverty With an Application to Single-Parent Families, 28 J. Hum. Resources 1 (1993).
[8] See, e.g., Martha Albertson Fineman, The Neutered Mother, The Sexual Family And Other Twentieth Century Tragedies (1995); Sylvia A. Law, Women, Work, Welfare, and the Preservation of Patriarchy, 131 U. PA. L. Rev. 1249 (1983).
[9] See, e.g., Dorothy E. Roberts, Welfare Reform and Economic Freedom: Low-Income Mothers’ Decisions About Work at Home and in the Market, 44 Santa Clara L. Rev. 1029 (2004); Michael Selmi & Naomi Cahn, Women in the Workplace: Which Women, Which Agenda?, 13 Duke J. Gender L. & Pol’y 7 (2006).
[10] Noah D. Zatz, Revisiting the Class Parity Analysis of Welfare Work Requirements, 83 Soc. Serv. Rev. 313 (2009).
[11] See Kathryn Edin & Laura Lein, Making Ends Meet: How Single Mothers Sur­vive Welfare and Low-Wage Work 81, 230 (1997).
[12] Id. at 94, 123.
[13] Zatz, supra note 10, at 316.
[14] See infra note 86 and accompanying text.
[15] See generally Linda Gordon, Pitied But Not Entitled: Single Mothers And The History Of Welfare (1994); Alice Kessler-Harris, In Pursuit Of Equity: Women, Men, And The Quest For Economic Citizenship In 20th-Century America (2001); Theda Skocpol, Protecting Soldiers And Mothers: The Political Origins Of Social Policy In The United States (1992).
[16] Ann Shola Orloff, Gender and the Social Rights of Citizenship: The Comparative Anal­ysis of Gender Relations and Welfare States, 58 Am. Soc. Rev. 303, 308 (1993).
[17] See generally Nancy Fraser & Linda Gordon, A Genealogy of Dependency: Tracing a Keyword of the U.S. Welfare State, 19 Signs 309 (1994).
[18] Gordon, supra note 15, at 12.
[19] African American women were often expected to work in the labor market, but this mandate was primarily expressed through exclusion from the transfer system rather than by conditioning transfers on work. See Gordon, supra note 15, at 293–94; Suzanne Mettler, Dividing Citizens: Gender and Federalism in New Deal Public Policy 172 (1998).
[20] In early practice, this primarily meant white widows, a reflection of both marital privi­lege and the racialization of sexual morality and gendered labor market norms. See Gordon, supra note 15, at 4–7; Dorothy E. Roberts, The Value of Black Mothers’ Work, 26 Conn. L. Rev. 871, 875 (1994).
[21] See Felicia Kornbluh, The Battle for Welfare Rights 40 (2007); see also Lizabeth Cohen, A Consumer’s Republic: The Politics Of Mass Consumption In Post­war America (2003).
[22] See Gordon M. Fisher, The Development and History of the Poverty Thresholds, 55.4 Soc. Security Bull. 3, 4 (1992), available at http://www.ssa.gov/policy/docs/ssb/v55n4/v55 n4p3.pdf; Mollie Orshansky, Counting the Poor: Another Look at the Poverty Profile, 28.1 Soc. Security Bull. 3, 8 (1965), available at http://www.ssa.gov/policy/docs/ssb/v28n1/v28 n1p3.pdf.
[23] Bureau Of Labor Statistics, U.S. Dep’t of Labor, Labor Force Statistics De­rived From The Current Population Survey: Bulletin 2307, 801, tbl. C-12 (1988).
[24] Sonya Michel, Children’s Interests/Mothers’ Rights: The Shaping Of America’s Child Care Policy 64 (1999).
[25] See Folbre, supra note 6, at 45, 52.
[26] They also exercise indirect influence when used as benchmarks for assessing other mea­sures. See, e.g., V. Joseph Hotz & John Karl Scholz, The Earned Income Tax Credit, in Means-Tested Transfer Programs in the United States 141, 146 (Robert A. Moffitt ed., 2003) (explaining how EITC was designed to lift minimum wage workers above the poverty line).
[27] ME. Rev. Stat. tit. 22, § 3769-C (2003). For the formulation “reasonable subsistence compatible with decency and health,” see ALA. CODE § 38-4-1 (1975); Kan. Stat. Ann. § 39­709 (Westlaw through 2010 Sess.); Miss. Code Ann. § 43-17-5 (Westlaw through 2010 Sess.); Tex. Hum. Resources Code Ann. § 31.003 (Westlaw through 2010 Sess.); Wyo. Stat. Ann. § 42-2-103 (Westlaw through 2010 Sess.).
[28] This statement is based on my review of the statutes and regulations governing the TANF programs in all states and the District of Columbia, as they stood on Aug. 1, 2008.
[29] Cal. Welf. & Inst. Code § 11452 (West, Westlaw through 2009 Reg. Sess.).
[30] Conn. Dep’tof Soc. Serv., Uniform Pol’y Manual § 4510.05.
[31] 45 C.F.R. § 260.31(a)(1) (2010).
[32] See, e.g., Pearce, supra note 7; Renwick & Bergmann, supra note 7.
[33] Folbre, supra note 6, at 57–59.
[34] Thomas A. Carlin et al., U.S. Dep’t of Agric., The Farm Differentiation in the Poverty Threshold: Should It Be Changed?, in Aspects of Welfare and Poverty in Rural America: Two Issue Briefs 11 (1979), available at http://eric.ed.gov/PDFS/ED1875 10.pdf; Fisher, supra note 22, at 10.
[35] Noah D. Zatz, What Welfare Requires From Work, 54 UCLA L. Rev. 373, 398–400 (2006).
[36] Lawrence M. Mead, The New Politics of Poverty: The Nonworking Poor in America 121–22 (1992).
[37] Brian Riedl, The Myth of a Child Care Crisis, The Heritage Found., May 2, 2003, http://www.heritage.org/Research/Reports/2003/05/The-Myth-of-a-Child-Care-Crisis (on file with the Harvard law School Library).
[38] Staff of H. Comm. on Ways & Means, supra note 2, at 9-2 to 9-3, 9-11.
[39] Nat’l Research Council, supra note 7, at 207–10; Jonathan Zasloff, Children, Fami­lies, and Bureaucrats: A Prehistory of Welfare Reform, 14 J.L. & POL. 225, 235–36 (1998) (describing “child care disregard”).
[40] Zatz, supra note 35, at 404–05, 424–25.
[41] Author’s calculation of population-weighted average maximum state CCDF reimburse­ment rates for family day care, based on Child Care Bureau, U.S. Dep’tof Health & Hum. Servs., Child Care and Development Fund: Report of State and Territory Plans, FY 2010–2011, at 15, tbl. 6 (2010), available at http://nccic.acf.hhs.gov/files/resources/sp1011full-report.pdf. For an infant and a toddler, the figure would be almost $14,000.
[42] This benefit would total $11,350 per year. Author’s calculation, based on Liz Schott & Ife Finch, Ctr. on Budget & Pol’y Priorities, Tanf Benefits Are Low and Have Not Kept Pace With Inflation 16–17 app.4 (2010), available at http://www.cbpp.org/files/10­14-10tanf.pdf. Note that a full-time, full-year minimum wage worker would still receive some of these benefits, though at a much reduced level, but would also be eligible for the maximum EITC in the vicinity of $5000. Integration with the EITC is theoretically attractive but raises additional issues. Cf. Nancy C. Staudt, Taxing Housework, 84 GEO. L.J. 1571 (1996) (arguing for inclusion of imputed income from household labor within the income tax base and consid­ering interactions with the EITC).
[43] The full year earnings would be calculated as follows: (50 weeks/year * 35 hours/week * $7.25/hour ) * (100% – 7.65% FICA) = $11,700.
[44] See discussion infra Part II.C.
[45] Extending the time frame could resolve the contradiction in conjunction with rapid earnings growth, but the available evidence is discouraging. See Zatz, supra note 35, at 405–18.
[46] Chris M. Herbst, Who Are the Eligible Non-Recipients of Child Care Subsidies?, 30 Child. & Youth Servs. Rev. 1037, 1039 (2008). Many states, including very large ones like California, Florida, New York, and Texas, do not fund their programs at levels adequate to provide assistance to all applicants who meet their eligibility standards. See Karen Schulman & Helen Blank, Nat’l Women’s Law Ctr., State Child Care Assistance Policies 2008: Too Little Progress For Children And Families 16, tbl.2 (2008), available at http://www.nwlc.org/sites/default/files/pdfs/StateChildCareAssistancePoliciesReport08.pdf.
[47] See Goodman, supra note 5 (discussing predicament of an employed single parent who could only access child-care assistance if she left her job and applied for public assistance).
[48] See Mead, supra note 36, at 119–24; Riedl, supra note 37.
[49] E.g., Suzanne W. Helburn & Barbara R. Bergmann, America’s Child Care Prob­lem: The Way Out 212–15 (2002).
[50] Zatz, supra note 35, at 395–400; see also James Tobin et al., Is a Negative Income Tax Practical?, 77 Yale L.J. 1, 12 (1967). In some contexts certain in-kind receipts are excluded from income calculations because of administrative difficulties in valuation, see, for example, Maisonet v. N.J. Dep’t of Hum. Servs., 643 N.J. Super. 1038 (App. Div. 1994), but these administrative exceptions illustrate the underlying rule.
[51] See Staudt, supra note 42, 1576 n.18, 1618–38; Katharine G. Abraham & Christo­pher D. Mackie, Beyond the Market: Designing Nonmarket Accounts for the United States 56 (2005); Folbre, supra note 6, 122–24.
[52] See infra note 73 and accompanying text.
[53] See Noah D. Zatz, Welfare to What?, 57 Hastings L.J. 1131, 1184 (2006).
[54] See id.; Zatz, supra note 35, at 384.
[55] Zatz, supra note 53, at 1178–79; see also Dennis J. Ventry, Jr., Welfare by Any Other Name: Tax Transfers and the EITC, 56 AM. U. L. REV. 1261, 1266 (2007).
[56] See Philippe Van Parijs, Real Freedom For All 89 (1995); Kirk J. Stark, Enslaving the Beachcomber: Some Thoughts on the Liberty Objections to Endowment Taxation, 18 Can J. L. & Juris. 47 (2005).
[57] See generally Zatz, supra note 35.
[58] For arguments that work requirements, even if ideally justified, are inappropriate in contemporary practice, see Stuart White, The Civic Minimum: On The Rights and Obli­gations of Economic Citizenship 137–47 (2003); Jurgen De Wispelaere & Lindsay Stirton, The Public Administration Case Against Participation Income, 81 Soc. Serv. Rev. 523, 524–25 (2007).
[59] See Amy L. Wax, Disability, Reciprocity, and “Real Efficiency”: A Unified Approach, 44 WM. & Mary L. Rev. 1421, 1446 (2003).
[60] Ron Haskins, Liberal and Conservative Influences on the Welfare Reform Legislation of 1996, in For Better and for Worse: Welfare Reform and the Well-Being Of Children and Families 9 (Greg J. Duncan & P. Lindsay Chase-Lansdale eds., 2001).
[61] Indeed, in 1973, Congress amended the Food Stamps statute specifically to count em­ployer-provided housing as income, thereby “equalizing the food stamp benefits of a person who receives his earnings entirely in the form of wages and a person who actually has the same income but receives part of it in kind rather than in cash.” Anderson v. Butz, 428 F. Supp. 245, 253 (D.C. Cal. 1975).
[62] 13-170-003 VT. Code R. § 2364.43 (2005).
[63] Alaska Admin. Code tit. 7, §§ 45.260(i)(1), (j)(2) (2010).
[64] U.S. Gen. Accounting Office, GAO-02-768, Welfare Reform: Tribal Tanf Al­lows Flexibility to Tailor Programs, But Conditions on Reservations Make It Diffi­cult to Move Recipients Into Jobs (2002).
[65] See Noah D. Zatz, The Impossibility of Work Law, in The Idea of Labour Law (Guy Davidov & Brian Langille eds., forthcoming 2011).
[66] Folbre, supra note 6, at 121–38.
[67] See Noah D. Zatz, Prison Labor and the Paradox of Paid Nonmarket Work, in Eco­nomic Sociology Of Work 369, 376 (Nina Bandelj ed., 2009).
[68] See, e.g., Fineman, supra note 8, at 9; Nancy Folbre, Children as Public Goods, 84 Am. Econ. Rev. 86, 87–88 (1994). My argument is more akin to Staudt’s argument for taxing housework based on its contribution to household consumption, though she shies away from fully equalizing nonmarket and wage income for the purpose of access to redistributive trans­fers like the EITC. See Staudt, supra note 42, at 1618–22, 1627–29, 1637–38.
[69] See, e.g., Anne L. Alstott, No Exit: What Parents Owe Their Children and What Society Owes Parents 58–61 (2004); Folbre, supra note 6, at 178–83.
[70] See Alstott, supra note 69, at 75–76; Eva Feder Kittay, Love’s Labor: Essays on Women, Equality And Dependency (Thinking Gender) 142–43 (1999); Gwendolyn Mink, Welfare’s End 135 (1998).
[71] Robert Haveman, When Work Alone Is Not Enough, in One Percent For The Kids40 (Isabel V. Sawhill ed., 2003).
[72] See Viviana A. Zelizer, The Purchase Of Intimacy 21, 29 (2005); see also Viviana A. Zelizer, Payments and Social Ties, 11 Soc. F. 481, 492 (1996).
[73] Amy L. Wax, Something for Nothing: Liberal Justice and Welfare Work Requirements, 52 Emory L.J. 1, 31–32 (2003).
[74] The same basic point applies to delimiting the relevant sources of imputed income for the purpose of measuring resources in the means test. A standard objection to imputed income techniques is that, once one begins with specific cases like child care, there is no principled reason not to attribute “income” to every practice valued by the person doing it: if walking the dog substitutes for hiring a pet sitter or playing solitaire substitutes for buying theater tickets, why not impute income to each at the value of their market substitutes? See Thomas Chancel­lor, Imputed Income and the Ideal Income Tax, 67 OR. L. Rev. 561, 561–62 (1988). The means-testing context provides an answer: child care is included in the basic needs budget but pet sitting and theater tickets are not.
[75] For arguments equating nonmarket care and volunteering, see Stuart White, Fair Reci­procity and Basic Income, in Real Libertarianism Assessed: Political Theory After Van Parijs 136 (Andrew Reeve & Andrew Williams eds., 2003); Iris Marion Young, Autonomy, Welfare Reform, and the Meaningful Work, in The Subject of Care: Feminist Perspectives on Dependency 40 (Eva Feder Kittay & Ellen K. Feder eds., 2003). Different lines of argu­ment may justify volunteering as a form of job training or a demonstration of willingness to work.
[76] See, e.g., Mary Anne Case, How High the Apple Pie? A Few Troubling Questions About Where, Why, and How the Burden of Care for Children Should Be Shifted, 76 Chi.-Kent L. Rev. 1753, 1775 (2001); Katherine M. Franke, Theorizing Yes: An Essay on Feminism, Law, & Desire, 101 Colum. L. Rev. 181, 191–92 (2001).
[77] Eric Rakowski, Equal Justice 153–55 (1991). But see Anne L. Alstott, What Does a Fair Society Owe Children—and Their Parents?, 72 Fordham L. Rev. 1941 (2004).
[78] This does not imply that the public always should bear costs associated with childrear­ing. For child care, as for food, clothing, and housing, means-tested transfers adopt the inter­mediate position that these goods are to be provided for oneself in the first instance but by a social safety net when personal resources fall short without fault.
[79] See 45 C.F.R. §§ 261.2, 261.60–.65 (2008).
[80] Zatz, supra note 35, at 431.
[81] 45 C.F.R. § 261.60(d) (2008). Similarly, the EITC uses earnings as a proxy for work effort by maximizing transfers at earnings levels approximating those from full-time, mini­mum wage employment.
[82] See P. Lindsay Chase-Lansdale & Laura D. Pittman, Welfare Reform and Parenting: Reasonable Expectations, 12 The Future of Child. 167, 171–76 (2002).
[83] See, e.g., Roberts, supra note 20; Jill Elaine Hasday, Parenthood Divided: A Legal History of the Bifurcated Law of Parental Relations, 90 Geo. L.J. 299 (2002).
[84] See Greg J. Duncan et al., How Different Are Welfare and Working Families? And Do These Differences Matter for Children’s Achievement?, in For Better and for Worse, supra note 60, at 104.
[85] See P. Lindsay Chase-Lansdale et al., Mothers’ Transitions from Welfare to Work and the Well-Being of Preschoolers and Adolescents, 299 SCI. 1548, 1549–50 (2003); Pamela A. Morris et al., Effects of Welfare and Employment Policies on Young Children: New Findings on Policy Experiments Conducted in the Early 1990s, 19 Soc. Pol’y Rep. 3, 11–12 (2005).
[86] Zatz, supra note 10, at 321. This predominates in state TANF rules, too, but some states require both adults to work.
[87] The two-parent context raises difficulties in attributing carework to individual parents. Alstott, supra note 69, at 181–84; Staudt, supra note 42, at 1623–24. However, because means-tested programs generally calculate and pay benefits on a household basis, one could require eighty hours of work in the aggregate, not necessarily forty by each parent. Similarly, we could attribute hours of care (and its imputed income) based on eighty hours minus the greater of forty hours or both parents’ aggregate time in wagework.
[88] This argument further requires a principle of occupational choice or other basis for insisting that caretakers have opportunities to fulfill work requirements in other ways.
[89] See, e.g., Haskins, supra note 60, at 9; Hugh Heclo, The Politics of Welfare Reform, in The New World of Welfare 169 (Rebecca M. Blank & Ron Haskins eds., 2001).
[90] For two examples of arguments against work requirements, based in part on their exclu­sion of caretaking, see Carole Pateman, The Patriarchal Welfare State, in Democracy and the Welfare State 231 (Amy Gutmann ed., 1988); Anne L. Alstott, Work vs. Freedom: A Liberal Challenge to Employment Subsidies, 108 YALE L.J. 967 (1999).
[91] Cf. Alstott, supra note 69 (proposing caretaker resource accounts that would benefit both employed parents and family caretakers).
[92] See Zatz, supra note 10, at 321.
[93] Sandra K. Danziger, The Decline of Cash Welfare and Implications for Social Policy and Poverty, 36 Ann. Rev. Soc. 523, 524, 537–38 (2010).
[94] Press Release, Hillary Clinton for President, Hillary Clinton’s Agenda for Working Families: Helping Parents Balance Work & Family (Oct. 16, 2007) (on file with the Harvard Law School Library).
[95] Nat’l P’ship for Women & Families, At-Home Infant Care (Ahic): A Side-By­-side Comparison of Federal and State Initiatives (2005), available at http://www.nationalpartnership.org/site/DocServer/AHICchartOct05.pdf?docID=1048; Work, Opportunity, and Responsibility for Kids Act of 2002, H.R. 4737, 107th Cong. (as reported by S. Comm. on Finance, July 25, 2002).
[96] AHIC allows conversion of child-care assistance into income support only for caretak­ers of infants, not older children. It also lacks any direct provision for basic needs other than child care.
[97] Similar lessons emerge from the recent Caregivers and Veterans Omnibus Health Ser­vices Act of 2010, which establishes a program of comprehensive assistance for family caregivers. 38 U.S.C.A. § 1720G(a) (West 2010). Family caregivers of injured veterans are eligible to receive, among other things, a cash stipend “not less than the monthly amount a commercial home health care entity would pay an individual . . . to provide equivalent per­sonal care services to the eligible veteran.” § 1720G(a)(3)(C)(ii).