Trapped in America's Safety Net

Even if she had had private health insurance, ultimately she would have needed to enroll in Medicaid: it’s the only realistic source—public or private—for the long-term services she continues to need as a disabled person. Because Medicaid is a program for the poor, Marcella and Dave had to spend down their assets to qualify; they now live on a near-poverty income. And to stay eligible, they will have to remain under these strictures, perhaps for the rest of their lives.

Just over half of Americans—55 percent—get their health insurance through their employer, while 30 percent are covered by a public program such as Medicare or Medicaid; 16 percent are uninsured.

The couple didn’t think of themselves as being in a precarious situation, but their modest wages, limited access to needed education and training programs, and lack of employer-provided benefits and comprehensive health insurance left them vulnerable. And, as with many Americans living with insecurity, a simple misstep, let alone a tragedy, would spell financial disaster.

while GDP grew 18 percent from 2000 to 2011, median income for working-age households fell by 12 percent. The reason? Most economic growth during this period accrued not to ordinary workers but to the top 1 percent of earners, who absorbed 65 percent of the nation’s income growth between 2002 and 2007.17

The opportunity gap has widened. Since 1980, the United States has experienced what MIT economist David Autor calls a “polarization” of skills and wages. Employment growth exhibits a U-shaped pattern, with the greatest growth among both high-skill occupations at one end of the spectrum and low-skill service jobs at the other. At the same time, the greatest wage increases have been concentrated at the high end. 23 Unequal access to education and training programs makes it very hard for individuals to upgrade their skills and to enhance their financial prospects.

The United States is just about the only industrialized country in the world that doesn’t guarantee health insurance for its citizens (Belarus and the former Balkan states are the other exceptions).

Bankruptcies tend to have multiple causes, so it is difficult to measure precisely the effect of medical debt. 34 But we do know that this pattern of events—illness, job loss, insurance loss, bankruptcy—is uniquely American. 35 In virtually all other economically advanced nations, citizens have health insurance regardless of their employment situation. The Obama administration health reform, the Affordable Care Act, will cut the number of uninsured and underinsured and will reduce some—but not all—disparities in health insurance coverage by age, income, and education, as chapter 6 explains. In other countries, however, these differences in insurance coverage across subgroups do not exist. People have health coverage by virtue of their citizenship.

Moreover, because Dave is employed, he and Marcella would be in a particular version of the program called “Share of Cost” Medi-Cal. It works this way: as a family of three with one disabled member, they are allowed to keep $ 2,100 of Dave’s $ 3,250 monthly earnings to live on. The rest of Dave’s earnings, $ 1,150, would go to Medi-Cal as the family’s share of cost. That is, any month in which Marcella incurred medical expenses, she and Dave must pay the first $ 1,150. To our surprise, if Dave earned more money, the extra amount would also go to Medi-Cal: the cost sharing is a 100 percent tax on Dave’s earnings. I figured out later that the $ 2,100 my brother and sister-in-law are to live on puts them at 133 percent of the federal poverty level for a family of three. Essentially, the way they meet the income test is for Medi-Cal to skim off Dave’s income until they are in fact poor. Brian noted that they are “lucky” that they are allowed to retain that much income; if Marcella weren’t disabled, the amount they’d be allowed to retain would be even lower than $ 2,100. And this is how things will be indefinitely. In order to get poor people’s health insurance, Dave and Marcella must stay poor, forever.

He and Marcella struggled to make ends meet on that near-poverty income. And then much later, at the time I was writing this book, we discovered that he probably didn’t need to reduce his pay; he and Marcella were probably not subject to a Share of Cost requirement even at Dave’s pre-accident income, a revelation to which I will return. But we didn’t know that for two years.

the asset “test” is an asset limit. Their house and one vehicle are exempt. Beyond those two items, they can possess only $ 3,150 in assets, total. They have to liquidate everything else4 and must put the resulting cash only into the house and the one vehicle. They can’t use the money to pay household bills, credit card bills, or Marcella’s student loans. They will have to save every receipt to prove how the money was spent.

In sum, they are barred from doing many of the things middle-class families are constantly advised to do: Save for retirement. Save for emergencies. Take advantage of tax-free college savings plans. Just $ 3,150 in total assets—that’s it.

In the early days, we had so many questions about Marcella’s eligibility for various programs. Medicare? No. Medi-Cal? Yes. SSDI? No. SSI? Yes. IHSS—what’s that? I thought Dave’s head would explode. It was impossible for any single person, even nondisabled, to navigate the entire, immensely complicated system alone. Marcella’s sister and sister-in-law divvied up the programs, got permission to act as her proxy, and began figuring everything out.

Just look at the calculators and figure it out, they advised. But I can’t—not with any certainty, to my enormous frustration. So much for helping my brother and sister-in-law navigate the system. Medi-Cal is a collection of over one hundred programs, each with its own income methodology and rules. A person familiar with Medi-Cal likened the program to the Winchester Mystery House, the San Jose mansion constructed continually over four decades by the odd widow of the Winchester rifle fortune: there is no master plan. “All the ‘rooms’ added on over the years makes it very difficult to see which rules apply to which groups and to follow them all the way through,” this observer told me. And even if Dave and Marcella could retain a bit more income to live on, they are still subject to the asset limit and all of Medi-Cal’s other strictures. They are still trapped in an eccentric’s mansion, where the stairways lead to ceilings and the doors open onto walls.

Marcella’s and Logan’s ICU stays up to that point: $ 474,000 for Marcella, $ 111,000 for Logan. The baby spent another week in the NICU, while Marcella was in the Mercy ICU for three more weeks, followed by the three-month rehabilitation hospital stay. With private insurance, even if they didn’t hit annual or lifetime caps, Marcella and Dave may well have had a 20 percent coinsurance requirement, which would have totaled hundreds of thousands of dollars. Either way, they would have had to have declared bankruptcy.

Between the ages of twenty-five and sixty-five, two-thirds of Americans will live in a household that receives food stamps, Medicaid, Supplemental Security Income (SSI), or Temporary Assistance for Needy Families (TANF) or other cash welfare. 2 In 2012, one in seven Americans received food stamps, including one in four children. Medicaid funds 40 percent of all births in the United States. 3 Social assistance programs aren’t marginal—they’re mainstream.

Every year, one in seven earners experiences a drop in earnings of 50 percent or more, while one in five experiences a 25 percent drop. 6 During these troughs, individuals and families may fall into social assistance eligibility. The consensus among most researchers is that income volatility has increased over time, particularly for family earnings.

Despite their neediness, many people who are eligible for means-tested programs don’t enroll. “Take-up” rates are very low. Over 60 percent of American children who lack health insurance are eligible for Medicaid or CHIP but aren’t enrolled. 15 Only 72 percent of those eligible for food stamps are enrolled (and only 34 percent of eligible senior citizens are). Complicated application procedures and confusing eligibility criteria are one barrier: many only seek information about these programs and fight their way in when they’re truly desperate, like Dave and Marcella. Outreach efforts are often modest in scope (and vary by state and by program). And the enduring stigma surrounding means-tested programs reduces take-up rates as well.

In 2011, 55.7 million people were enrolled in Medicaid, up from 23 million in 1990 and 34.5 million in 2000; over 70 million, or one in five Americans, were enrolled for at least one month during 2011.26

in 2013 the Republican-controlled House led efforts to reduce SNAP further and de-couple it from the farm bill, where food stamps had long coexisted with agricultural subsidies, backed by a bipartisan coalition of rural agricultural and urban interests.

In 2009, just two-thirds of workers aged eighteen to sixty-four had employer-provided health insurance; just half such workers had an employer that sponsored a retirement plan. 51 Furthermore, access to employer-provided benefits is highly skewed by income. Such benefits are most widely available to the middle class and especially the affluent in higher-paid jobs. If you sort US firms by the average wage they pay, most firms in the top quartile offer retirement benefits and paid sick leave, compared with only one-third of the firms in the bottom quartile.

I would need to pay for personal care assistants out of pocket both before and after the age of 65. Twelve hours of care per day, at $ 20 per hour, would total $ 87,600 per year. Even at $ 12 per hour—the lowest rate I could find in Massachusetts—the annual bill would come to $ 52,560. And the bill could be much higher: a bioethics professor in New York whose husband is paralyzed from the shoulders down spends nearly $ 250,000 per year for his twenty-four-hour-per-day personal care.

How you are treated in times of need and what kinds of help you can expect to receive depend almost entirely on your status as a worker—and not only whether you are employed but for how long and for which employer.

American means-tested programs aren’t principally about pulling people out of poverty. Instead, they’re intended to provide a safety net that is absolutely minimal (given its taxpayer funding), goes to the “deserving” rather than the undeserving, and doesn’t deter holding a job. To achieve these goals, means-tested programs are tightly targeted on those at the very bottom of the socioeconomic ladder, with low income and asset ceilings. Tight targeting addresses the main fear, which has always been that programs for the poor deter working. Thus from its inception, American social assistance was organized around the principle of “less eligibility” that originated in the British poor laws of the nineteenth century. 2 The idea back then was to deter paupers from work-houses by making conditions there worse than the worst job on the outside. In contemporary America, this idea persists: means-tested programs have to remain inferior to the alternative—the worst jobs at the worst wages. If benefits were easy to get, stigma free, and generous, it would be too hard to find people willing to take bottom-ofthe-barrel jobs. There would be pressure on low-wage employers to improve pay and benefits. Instead, policy makers try to make means-tested programs as miserable as possible to discourage would-be applicants, minimize enrollment, and incentivize work in low-level jobs. And as conditions in the low-end job market have deteriorated over time, with shrinking real wages, an increase in part-time work, and the near disappearance of employer-provided benefits, means-tested programs have gotten worse in tandem. The problem is that narrow targeting and fear of replacing work have perverse consequences: the programs’ corresponding design features also prevent people from leaving poverty. The extreme eligibility criteria that tight targeting requires—the low income and asset ceilings—mean that as people leave assistance for work, their benefits first diminish and then disappear altogether. The cost of working is steep indeed, as the recipient loses cash assistance, housing, food assistance, child care subsidy, and medical insurance, one after another. And there’s little on the outside to replace these lost benefits.

Consider Temporary Assistance for Needy Families: in no state do welfare benefits come close to lifting recipient families out of poverty. In the most generous TANF state, New York, the maximum benefit is less than half the federal poverty level. In the least generous state, Mississippi, the maximum TANF benefit is 11 percent of the poverty level. For a family of three, that’s $ 2,040 for a year—a little more than the annual household income in Nigeria.

In New York City, where 5,000 to 6,000 Housing Authority apartments become available each year, it would take thirty-eight years to get through the quarter-million-household waiting list.

Most means-tested programs in most states have an asset test, which requires that an applicant not only be cash poor but also possess very minimal assets to qualify for assistance. Most of these asset test amounts haven’t been raised in years, making the poor even worse off. As the Suze Ormans and Dave Ramseys of the world attest, the first step to financial stability is having an emergency fund, followed by saving for other needs: your next car, retirement, college. But most social assistance programs prohibit such savings, forcing the poor to stay poor. Without savings, not only can you not get ahead, but you can easily fall further behind.

This asset policy makes more sense when you consider its real purpose: not poverty reduction, but making sure that only the neediest and most deserving enroll in these assistance programs. The asset limits are intended to ensure that no one with a fat bank account is receiving aid. They are part of the effort to tightly target these programs, in other words. But once people are in these programs, asset limits become a straitjacket. 12 It’s very expensive to be poor,

Some states have made changes to their asset rules. Twenty-four have eliminated their asset tests for Medicaid, partly for administrative simplicity and cost savings: it’s cheaper simply to cover poor people than to hound them about their assets and then deny care. 15 Forty-seven states have eliminated their asset test for the Children’s Health Insurance Program. And in order to promote personal savings and greater financial security, President Obama’s fiscal 2011 budget proposed raising asset tests for federally funded means-tested programs such as TANF, food stamps, and the Low-Income Home Energy Assistance Program to a minimum of $ 10,000. But in the polarized Congress this proposal went nowhere, and it wouldn’t have helped the many poor people in state programs that still have asset tests. The Affordable Care Act eliminates the Medicaid asset test for those newly eligible for Medicaid, but this doesn’t help Dave and Marcella, who are still under the state asset test because of her being in a previously eligible group, the disabled. If only California would follow the lead of half the states and lift its Medi-Cal asset test, it would make a huge difference in their lives.

Medicaid and CHIP, have an eligibility cliff: earn one more dollar in income, and you lose eligibility for public health insurance altogether. This is the essence of the “means test” imposed by all social assistance programs: as your means rise, you fail the test. These benefit phaseouts and falloffs are essentially huge marginal taxes on those exiting social assistance for work. They provide a powerful disincentive, because the effective cost of working is so very high.

these phaseouts, cliffs, and huge marginal tax rates aren’t failures of means-tested programs alone. Recipients are trapped in poverty also because of failings at the boundaries between social assistance, other social supports, and jobs. If jobs or other social programs reliably provided health insurance or child care, transitions out of social assistance wouldn’t impose such steep costs to the affected individuals. The problem is that on one side, we try diligently to minimize assistance programs by keeping their scope limited and clearly bounded. But this tight bounding means that people often exit means-tested programs—and do so abruptly—when they are still poor. On the other side, we have social programs and jobs that fail to provide the working poor with adequate wages or needed protections. If only one or the other existed, people would exit social assistance without the steep costs described here. But because both poor wages and meager protections exist together, leaving social assistance often means falling into a chasm.

why get married? EITC, CHIP, and sometimes Medicaid benefits extend relatively far up the income ladder. However, when two modest-income people marry, their combined incomes can push them out of eligibility, and their increased taxes and lost benefits combine for the high effective marginal tax rates described above. The Medicaid and CHIP health insurance benefits are particularly valuable, since many low-and moderate-income people have no other means of getting health coverage. So why marry? As economist Eugene Steuerle puts it, “Not getting married is the major tax shelter for low-and moderate-income households with children.” 20

In the perverse world of means-tested programs, work doesn’t pay.

Universal policies recognize that nearly everyone, not just the poor, needs help with the big-ticket items in life. They give support to all with these needs, providing considerable financial relief to the middle class while helping lower-income people in a stigma-free way. In addition, they make it easier to leave poverty, because they don’t have the benefit cliffs and phaseouts that means-tested programs do. Additional areas in which such a program design might be used include universal child care and preschool, universal paid family leave, and universal health insurance. To this list I would add a social insurance program for long-term care so that the disabled of any age don’t have to impoverish themselves to get assistance.

In a number of other states, they’d be allowed to save for retirement, and Marcella wouldn’t have had to liquidate her 401( k) savings to qualify for public health insurance. Indeed, twenty-four states use no asset test for Medicaid. In these states, Marcella and Dave could do it all: save for retirement, college, and emergencies, and Dave could drive a safe car to work. But not in California, where they happen to live. The experience of the poor in means-tested programs varies tremendously across states.

States have always played an important role in social assistance. 3 Before the Great Depression and World War II, which greatly expanded the size and role of the federal government, state and local governments collected far more tax revenue and were responsible for more functions than the government in Washington, DC. Indeed, a number of states had workers’ compensation programs, old-age pensions, and mothers’ pensions (which supported poor single mothers with children at home) long before the New Deal instituted programs for the aged, blind, poor, and unemployed. Because these earlier programs were entirely state run, states had a great deal of discretion over eligibility and benefits, and were loath to give up these powers to the federal government. In particular, southern states used the differential application of assistance programs to reinforce an economic system based on black agricultural labor. For example, mothers’ pensions weren’t available in counties with the highest concentrations of African Americans in order to maximize the number of available workers.

As hard-hit as the South was by the Depression—many of its states had to scale back their social assistance programs greatly—federal provision of Unemployment Insurance or mothers’ pensions posed a profound threat. These funds would provide a stream of monetary aid to black agricultural workers and domestics outside the control of the white planter elite, thereby threatening the underlying racial and class structure of the southern economy and society. 5 Southern congressmen used their dominance of key congressional committees to maintain the racial order and to reinforce state sovereignty and control. The social insurance programs of the Social Security Act of 1935—Old Age Insurance (what we think of today as “Social Security”) and Unemployment Insurance—initially covered only workers in “commerce and industry,” thus excluding occupations where African Americans were concentrated, such as farm laborers and domestics. 6 These racially discriminatory effects gradually diminished as more occupations were brought into the system. By the 1970s, Social Security in particular became a virtually “universal” program for which most retired workers were eligible, one that helped alleviate poverty in old age among the majority of income and ethnic/ racial groups.

the Social Security Act also contained two social assistance programs: Old Age Assistance (cash payments for the impoverished elderly) and Aid to Dependent Children (ADC, later AFDC and then replaced by TANF—that is, cash “welfare” payments to poor families with children, the successor to mothers’ pensions). Unlike Old Age Insurance, Old Age Assistance and ADC were not purely federal programs, at the insistence of southern congressmen. Instead, the federal government would pay a portion of program costs, but states would retain operational control and set eligibility criteria and benefit levels. These representatives further insisted that a provision that ADC payments provide “a reasonable subsistence compatible with decency and health” be stripped from the legislation. 7 Although Old Age Assistance was eventually federalized as Supplemental Security Income (SSI) in 1972, the hybrid federal-state design of ADC persisted. The phenomena that we observe today—state discretion over social assistance, widely varying program parameters, and the failure of many social assistance programs to meet underlying needs—hark back to this founding era and the legacy of racial politics.

When the nation’s public health insurance programs were created in 1965, the same dichotomies between insurance and assistance and between federal and joint federal-state responsibility were adopted. Medicare, for senior citizens, would be a federal-level social insurance program with nationally uniform eligibility criteria and benefits. In contrast, Medicaid, for the poor, would adopt AFDC’s hybrid federal-state structure: the federal government would pay part of the cost and stipulate minimum eligibility criteria and benefits, but states could determine the scope of benefits and eligible populations beyond those minima as well as set provider reimbursement levels (which ultimately affected recipients’ access to health care). 8 Program funding differed as well. Medicare financing came entirely from federal and recipient sources: a payroll tax for Part A hospital insurance, and general federal tax revenues and monthly premiums for Part B supplemental medical insurance. In contrast, states were on the hook for part of Medicaid’s financing, which they shared with the federal government. That is, states had to spend their own dollars in order to receive federal matching money. Despite a matching formula that provided a greater federal contribution in poorer states, the robustness of the program would nonetheless come to depend on each state’s fiscal capacity. To this day, means-tested social assistance programs bear the marks of their birth in the crucible of American racial politics and states’ rights. 9 For our purposes, what matters is that the states, both then and now, play a central role in these programs. Today, state fiscal capacity and to a lesser extent ideology are what shape social assistance provision. But the result is the same: vast interstate differences in policies and outcomes.

Southern states adopted Medicaid slowly, and Arizona waited until 1982. Eligibility varies dramatically as well, meaning that an individual at a given income or asset level qualifies for a program in one state but not another. And what counts as income and assets differs too. Moreover, the size and scope of benefits vary dramatically, from the size of monthly checks in the cash programs to the scope of covered medical benefits in Medicaid. Take-up rates—the proportions of eligible people actually enrolled in each program—also vary across states. The result of all these differences is considerable variation in adequacy: in the proportion of poor people who actually receive benefits and in the ability of those benefits to meet basic human needs.

Forty years ago, four in five poor children received welfare; now four in five poor children do not.

There is also considerable variation in the size of TANF benefits (fig. 5.2). Among the lower forty-eight states, the maximum monthly TANF cash benefit in July 2011 for a family of three varied from $ 788 in New York, $ 714 in California, and $ 640 in Vermont to just $ 185 in Tennessee and $ 170 in Mississippi. 11 In twenty-two states, the maximum benefit provided an income less than one-fourth of the poverty line; in a few states at the bottom, the most a recipient could get from TANF was barely over 10 percent of the poverty line. Per person, that’s $ 2 per day—the World Bank’s definition of poverty in developing countries.

TANF rules and benefits differ widely across states, even across near neighbors. A stepparent’s income is included in the income test in Iowa but not in Kansas; a grandparent’s income is included in New Hampshire but not in Vermont. A family with $ 2,000 in assets would be eligible for TANF in Oregon but not in Washington. Idaho requires applicants to be engaged in a job search to receive benefits, while Montana does not. In Mississippi, the highest sanction for failing to comply with the work requirement is total loss of benefits and permanent ineligibility for TANF, while in Arkansas the case is merely closed until the family is in compliance for two weeks. A family’s experience depends entirely on the state in which it happens to live.

Twenty-four states have no asset test for Medicaid;

One reason for such wide variation in eligibility and benefits is that Medicaid is a very expensive program to run, and poorer states can afford less. States split the cost with the federal government; the feds provide more funding to the poorer states, as measured by per capita income. A poor state receives $ 2.85 in federal funds for every dollar it spends of its own funds, while richer states only receive $ 1.00.30 But getting the federal match still means spending a state dollar, which poor states are less likely to have. 31 The inverse matching formula is meant to incentivize states to enroll more of their low-income groups in Medicaid. But in reality, poorer states often can’t afford to be more generous, even with the greater federal match. The result is that even for health insurance, the experience of the poor varies dramatically across states and across time with states’ fluctuating fiscal circumstances.

The working parent of a Medicaid-eligible child in Arkansas can get Medicaid for herself only if her income is less than 16 percent of the federal poverty level (that’s Nigeria-level income), while in Wisconsin she can get Medicaid if her income is up to 200 percent of the poverty line (that’s one-third of American households). An adult Medicaid recipient can get eyeglasses in Texas but not in Oklahoma; Tennessee provides one pair of glasses after cataract surgery, while in Utah the only Medicaid adults who can get eyeglasses are pregnant women. In a Massachusetts Medicaid family, a child can get a filling in any tooth, but his mom can get a filling only in her twelve front teeth, not in her molars. Between 2010 and 2012, she would have been out of luck entirely, because state budget cuts eliminated most adult dental benefits altogether. 33 And on and absurdly on. As with TANF, how you fare is an accident of time and place.

CHIP provides health insurance for children whose families lack coverage but whose incomes are too high for Medicaid. It is funded jointly by the federal and state governments (using the same matching rate as Medicaid, plus a 15 percent sweetener to encourage state participation), but states determine eligibility and other program parameters. By 2012, nearly all states—forty-seven—had eliminated their asset test for CHIP, 34 but they still set income thresholds, with the income cutoff for eligibility varying from just 160 percent of the federal poverty level in North Dakota to 400 percent in New York.

Thirty-eight states have a waiting period, a length of time a child is required to be uninsured before allowed to enroll in CHIP, which varies from one to twelve months. 37 States impose waiting periods because they fear that parents will drop employer-provided insurance to enroll their children in CHIP. There is little evidence that parents engage in such behavior, however.

SNAP may be federal, but it’s local and state social workers who inform people about their eligibility and sign them up—or not. In 2006, 67 percent of those eligible for food stamps participated in the program, but the participation rate varied from 98 percent in Missouri to just 50 percent in California

if states responded to need—then more generous programs and more spending would occur where and when poverty is greater. In reality, the exact opposite patterns emerge. Poor states do the least to help their poor residents, not the most. And during economic downturns, states often reduce rather than increase spending on means-tested programs, just when the need is greatest. Across states, a negative relationship exists between need and social assistance provision that undercuts the responsiveness argument. Poor states spend less (as do conservative states).

during the Great Recession that began in 2007, Arizona not only closed off CHIP enrollment to new children but also reduced the time limit for TANF from sixty months to thirty-six. Massachusetts cut adult dental care in its Medicaid program, as noted earlier. The State of Washington reduced cash grants in its Disability Lifeline program for the physically and mentally disabled. Pennsylvania reduced its state SSI supplement. Mississippi slashed its mental health budget. South Carolina reduced Medicaid hospital payments by 7 percent and physicians’ reimbursements by 10 percent. 62 The list of social assistance cuts goes on and on. 63 Such reductions reinforce downturns, making recessions even worse, whereas policy responsiveness would dictate greater, countercyclical spending to meet increased need.

Many state policy makers think the poor have a choice, and so they refuse to offer generous social assistance benefits for fear of becoming a “welfare magnet.” While empirical examinations of the welfare magnet phenomenon are inconclusive, what matters is that state lawmakers think the effect exists, and make policy accordingly. 75 But in fact, for many the choice of location is an illusion. An important study of poor single mothers found that while the generosity of welfare benefits did influence state residential choice somewhat, many move back to their home state, where their family network and connections are.

most Americans will eventually be eligible for Medicare, a federal program. If some of them are in poorer health by the time they reach the age of sixty-five because they were uninsured as adults due to state policy (such as narrow Medicaid eligibility), then the national program is worse off: those people will end up being more costly than healthier individuals from a state that provides more expansive social assistance. Beyond these pragmatic matters is the fact that the national government is supposed to promote “the general welfare.” Dramatic cross-state differences in public policy arguably conflict with this tenet of American citizenship.

A greater federal role is evidenced in CHIP as well. The program has a higher federal match than Medicaid does, to encourage states to cover as many low-income children as possible. Furthermore, when it was reauthorized in 2009, funding was increased substantially by raising the federal tobacco tax. The reauthorization also included provisions intended to maximize enrollment: states must use their allotted CHIP funds within a certain period or lose them, and they receive annual performance bonuses for exceeding their enrollment targets.

Some groups are exempted from the individual mandate to have insurance: those whose incomes fall below the tax-filing threshold; Native Americans; incarcerated individuals; and those with financial hardship. Illegal immigrants are both exempted from the mandate and prohibited from buying insurance on an exchange, even with their own money. Most significantly, millions more will receive no assistance for health insurance coverage, as they are ineligible for either Medicaid or subsidies to buy private insurance on the exchanges. The Supreme Court ruled in June 2012 that the Medicaid expansion is optional for states, and as a result about half the states—home to more than half the nation’s uninsured—initially announced they would not broaden Medicaid eligibility.

“In states that do not expand Medicaid, some of the neediest people will not get coverage. But people who are just above the poverty line or in the middle class can get subsidized coverage. People will be denied assistance because they don’t make enough money. Trying to explain that will be a nightmare.”

As written by Congress, the ACA states that people who have access to “affordable” insurance through their employer cannot get subsidies to buy health insurance on the exchanges instead (the intent was to prevent people from leaving employer-provided coverage for the exchanges). The ACA defines “affordable” coverage as costing no more than 9.5 percent of family income. But the IRS ruling tied “affordability” to the cost of health coverage for an individual worker, typically around $ 5,600, not the cost of family coverage, which costs about three times as much, around $ 15,700. Therefore, if an employer is unwilling to pay a portion of family premiums, some family members may remain uninsured, unable to afford the full premium on their own and ineligible for the subsidies.

There’s one thing Dave could do to get out from under Marcella’s Medi-Cal restrictions: divorce her. This would separate Marcella and Dave’s income and assets. Only Marcella would then be forced into impoverishment; Dave could keep his half of the assets. And he and Logan would be free to live on as much income as Dave could make, to accept help from family members, to establish a 529 college fund, to have emergency money put away, to save for Dave’s retirement. In other words, they could go back to their middle-class life. And unattached to a spouse, Marcella would be eligible for means-tested programs without the “deeming rules” that count a spouse’s income and assets. Such “Medicaid divorces” are common among the elderly. When one member of a couple becomes disabled and needs to move to a nursing home, the couple divorces to separate their assets. Then the disabled former partner “spends down” his or her assets on nursing home fees until the state’s asset limit is reached, and then goes on Medicaid. In this manner, some assets for the spouse remaining in the community are protected.

At the age of sixty-five, the likelihood of needing home health care at some point in one’s remaining years is 72 percent; the likelihood of needing nursing home care is 49 percent. 33 As disability law expert Sam Bagenstos notes, “If we live long enough, we’re all going to have disabilities. That’s one of the things about the disability community. We’re all going to be a part of it, if we’re lucky.” 34 And yet in the United States, the only way for us to get public help with these expensive needs is to spend down our assets to the poverty level and Medicaid eligibility.

While many social protections are conferred on a near-universal basis in other rich democracies, in the United States more generous benefits are given to groups deemed deserving, such as retired workers. Those perceived as less deserving, such as the non-working poor, are seemingly held responsible for their plights and given only meager help. As Lynch says, the welfare policies of western Europe and Canada reflect “the principle of inclusion with elements of universal entitlement based on need and adequacy,” while the American welfare state instead “prioritizes personal responsibility, help for the deserving only, and the principle of less eligibility,” thereby making public benefits less desirable than work, as we saw in chapter 4. For the poor in America, the resulting orientation of the government is less about assistance than about skepticism, begrudging and meager help, concern about fraud, and even punishment. 41 These policies reflect public attitudes. As disability activist Dennis Heaphy says, “Americans have a punitive view of poverty and who is poor and why they are poor.” 42 Many are very concerned that some people might get more than they deserve. Heaphy adds, “We live in a mind-set of scarcity, and everyone is afraid that someone is getting away with something or taking away from them.” In the policy regime that arises from these attitudes, social workers function as gatekeepers rather than facilitators, and scarce dollars are devoted to sorting the deserving from the undeserving. The inefficiency of these practices is demonstrated by the decision of many states to simply drop the Medicaid asset test rather than try to ferret out the tiny level of resources most Medicaid applicants have.

Incredibly, the United States is one of only four nations—out of 173 for which data are available—that doesn’t offer paid maternity or parental leave for women (the other three are Liberia, Papua New Guinea, and Swaziland). 44 There is no national paid leave for family caregiving, either—just twelve weeks of unpaid leave for which only half are eligible and most can’t afford to take. 45 Also unlike many other advanced-economy countries, there is no universal public day care or preschool: only 58 percent of American three-to five-year-olds attend preschool, compared with 90 to 100 percent in much of Europe.

Fidelity Investments has calculated that a 65-year-old couple retiring in 2011 with an average life span (20 more years for the wife, 17 for the husband) would need $ 240,000 to cover their share of health care costs in retirement. 57 But only one in five people aged 55 to 64 has saved that much; nearly one-third have saved less than $ 10,000.58

a reform proposal that crops up perennially would turn the federal share of Medicaid into a block grant to the states. If the TANF block grant experience is any guide, this reform would have devastating effects, seriously eroding health insurance for the poor.

Raising taxes by just a few percentage points of its gross domestic product would still leave the United States a very low-tax country and wouldn’t harm its economic growth, 75 but would make a world of difference for needy Americans.

I worry that those who make social policy fail to comprehend what the requirements, rules, and regulations they create mean for the ordinary people who live under them. I hadn’t fully—and studying social policy is my job. But until my family fell down the rabbit hole and experienced these things firsthand, I hadn’t really understood or fully appreciated the ramifications of the design of the American welfare state.