The White House Council of Economic Advisers (CEA) just released a highly partisan analysis to justify new work requirements likely to be proposed by the Trump administration and Congress on recipients of Medicaid, the Supplemental Nutrition Assistance Program (formerly food stamps), and federal housing assistance.

Disguised in technical terms and seemingly serious research, the CEA analysis is a transparent attempt to claim that these welfare programs lead to shirking of work, poverty level incomes, and damage to children. Moreover, they claim that measured “properly,” poverty has been eliminated for all practical purposes, renouncing decades in which Republicans claimed anti-poverty programs failed.

Aside from the false claim that poverty has been essentially eliminated, this fusillade is a replay of the same arguments created and used in the 1980s and 1990s as part of the war on Aid to Families with Dependent Children (AFDC), the welfare program that was started during the Great Depression and expanded over subsequent decades. In this attack on AFDC, Republicans and conservative commentators widely and loudly claimed that cash welfare itself as well as food stamps was the cause of poverty, a culture of indolence, and in particular, a wave of unwed mothers. (Never mind that, during the height of this conversation, the birth rate for unwed mothers was falling sharply, starting in 1991.)

The policy result of this attack was the 1996 reform that replaced AFDC with Temporary Aid for Needy Families (TANF). TANF imposed stern work requirements and a time limit for benefits; AFDC had had neither of these. TANF, signed by Bill Clinton, was administered by the states, which, depending on their political attitudes, reduced the dollar amount of benefits and put up barriers for prospective recipients.

Now, with the release of this analysis, the Trump CEA claims TANF has been a major success, saying that it has put women to work in vast numbers and has reduced the welfare rolls sharply. And, citing this success, the CEA says Congress should now apply similar reforms to additional programs, because many Americans are still living on the dole.

Unfortunately, TANF has not been the success that the CEA claims. It is now a much smaller program, serving relatively few. Its benefits are low, and few people bother to sign up. To give an idea of how ungenerous welfare in America now is, even if an applicant meets the work and job search requirements, “every state’s TANF benefits for a family of three with no other cash income were at or below 60 percent of the poverty line,” according to the Center on Budget and Policy Priorities. “Most states’ benefits were below 30 percent of the poverty line.”

TANF has also stranded millions of very poor Americans without an adequate lifeline, other than food stamps. Furthermore, its work requirements have not created sustainable new jobs (as we shall see).

So, the evidence seems to suggest that the real “success” that work requirements had with regard to TANF was that they shrunk the program.

Turning the Attack from TANF to Bigger Targets

The argument being made by the CEA is particularly dangerous right now, since the GOP-controlled Congress is currently seeking major social spending cuts—not because they are good policy, as far as I can tell, but rather because they can help pay for the huge tax cuts that were just passed.

The Trump administration now wants Congress to attach work requirements similar to TANF’s to the remaining major welfare programs. (Note that while the Earned Income Tax Credit and the Child Tax Credit are large programs that provide substantial monetary benefits to the poor, recipients must have earnings.)

Let’s be clear about what is going on. Advocates of work requirements are going after much bigger game than they did in the 1990s. Back then, AFDC plus food stamps came to no more than 1 percent of GDP.

But now, Medicaid is the big target. In setting its sights on a program that cost $565.5 billion in 2016, or some 3 percent of GDP, the CEA is seeking to impose work requirements that will reduce access to medical insurance for poor men, women, and children. Also under fire is the Supplemental Nutrition Assistance Program (SNAP), which costs about $71 billion a year—less than half a percent of GDP. Housing assistance, about $35 billion, could also be affected.

Faulty Claims about Work Requirements

The CEA makes two major claims—which they can’t substantiate—in order to justify their new work requirements gambit. The first claim is that many nondisabled people are “shirking work” because they benefit from SNAP and Medicaid. They cite as evidence the numbers of benefit recipients who are not working, or working few hours. (Even the CEA admits the circular nature of this reasoning, because the programs are designed specifically for the poor, who typically are poor because they don’t have decent work.) The second claim the CEA makes is that TANF work requirements put people back on the job in big numbers, and these people wouldn’t be working without the push.

Here are a few facts to contradict these two CEA claims. Regarding Medicaid beneficiaries, the CEA reports that more than half of nondisabled adults work less than twenty hours a week. But the Kaiser Family Foundation finds that 62 percent have full- or part-time jobs, and an additional 18 percent lived in a family with a working adult. According to the New York Times, the CEA says Kaiser used poor data. Hardly likely.

Regarding SNAP, about three out of four SNAP recipients in any month work that very month or within a year of taking the benefits. In one study, half of SNAP recipients were working when they received the benefits. In total, recipients on average receive SNAP for less than two years. Moreover, about two out of three participants in SNAP are children, the elderly or disabled.

The CEA’s blithe claims that women found many more jobs after welfare reform are also wrong. The studies that the CEA cites in claiming success typically ended after the first few years of TANF’s launch in 1996—which means their findings were measuring the booming economy of the late 1990s, as well as the effects of a more generous EITC. (Studies show the EITC had more constructive influence on work than did TANF, even in the early years.)

Thirteen thorough studies of the consequences of the TANF programs administered by states found that five years later, the employment rates among women who were not subject to TANF work requirements were equal to those who were.

More relevant, the jobs TANF recipients took were not stable. In sum, they were short-lived, and workers had to jump to new jobs of varying and usually low pay. For example, only one out of five of those who left TANF in a Maryland program between 2001 and 2009 were employed in stable jobs.

Distorting the History of Poverty

As for CEA’s assertion that poverty no longer exists in America, it is a cruel distortion of claims that many more thoughtful economists have been making for some time to show that social programs have been effective. Indeed, the CEA even admits in its analysis that the reason poverty rates are down are the very social programs they want to limit.

Columbia University researchers had analyzed the history of poverty rates since the 1960s based on the relatively new if still unofficial Supplemental Poverty Measure (SPM). By creating a historical series, they find the poverty rate has been reduced by more than one third from much higher levels fifty years earlier. But the fact is today’s poverty rate, based on either the official poverty measure (OPM) or the SPM, is still higher than in any other large rich nation. For adults, using the newer SPM, poverty rate is 14.0 percent. For children under 18, it is 15.2 percent. Hence, centrist and liberal economists have been arguing that there has been serious progress on reducing poverty due to new social programs, a claim Republicans and conservatives have generally denied—until now. Ronald Reagan’s famous quote—“We fought a war on poverty and poverty won”—is being abandoned by the Trump administration.

The CEA argues that the proper measure of poverty reduces it to roughly 3 percent of the population—based not on income earned, like the OPM or the SPM, but rather on measures of actual consumption. To say the least, this is contentious, but the CEA treats it as uncontroversial. Consumption data are especially difficult to compile accurately. A study published in the prestigious American Economic Review adjusted for such measurement inaccuracies found that both income- and consumption-measured inequality followed the same trajectory, for example. Consumption measures, in other words, did not show less inequality.

More glaring, the CEA argue that, with scant evidence, material deprivation is down—the poor have met their needs. Material deprivation measures housing, health care, nutrition, the ability to pay the bills, and so on; few serious analysts agree it is down. One 2015 study whose authors derive a material deprivation index found that it actually rose. In another example, the Robin Hood Foundation found that last year 63 percent of households in New York City were deprived. As Robin Hood writes, “For some, deprivation took the form of generalized low income or what is traditionally defined as ‘poverty.’ For others, deprivation took the form of specific material hardships, such as: hunger, eviction, cutoff of phone or electricity. Still others suffered from a chronic, debilitating illness.” Indeed, a nation with around 25 million still having no health insurance—even after the considerable success of the ACA—does not suggest deprivation is a non-issue.

An analysis by Luke Shaefer and Josh Rivera also finds that several conventional measures of material deprivation have risen in recent years, contrary to the offhand assertions of the CEA. More to our point, they find that material deprivation rose even as the consumption measure of poverty widely used fell sharply. They point out that, judging by the consumption poverty measure, you would come to the ridiculous conclusion that “households were markedly better off at the height of the Great Recession [2010] than in what is largely considered the very strongest year of the historic 1990s economic boom [1999].”

Using Bad Measures

One serious point should be made, though it is a somewhat technical one. The consumption measure uses a poverty threshold, the OPM, that few believe is credible. It is based on a food budget from fifty years ago, multiplied by three, and adjusted for inflation. The authors the CEA cites for the consumption measure, Bruce Meyers and James Sullivan, then use a different inflation measure, which reduces the poverty line, and they add back welfare benefits such as the EITC and food stamps (they also add imputed rental income for homeowners).

The newer SPM makes roughly these same adjustments on the income side, but it uses an updated and relative rather than unchanging poverty threshold—roughly one third of average American spending on food, housing, electricity, and transportation over the most recent five years—not the irresponsible original OPM absolute threshold based on an ancient food budget. This is a higher poverty threshold, thus implying more poverty. The SPM, devised over years of research that originated at the National Academy of Sciences, also deducts expenses for child care, work, and medical from the income of the poor. Neither the OPM nor the consumption measures make these deductions or use a more sensible threshold. Meyers and Sullivan stick with the OPM poverty threshold, as primitive and misleading as that is.

Thus, the SPM is higher than the OPM, even with addition of social benefits (which, as a reminder, the OPM does not include). The consumption budget is similar to an analysis by Christopher Jencks in The New York Review of Books based on the OPM. Jencks also changed the inflation computation, added back social spending, but did not adopt the SPM’s new threshold or deduct medical, work, or child care expenses. He found a poverty line of about 5 percent based on income (which didn’t include imputed housing income for homeowners), not all that different than the consumption line. The SPM—now at 14.5 percent—is far closer to the true state of poverty in America than either the consumption line or Jencks’ income construction. But all evidence shows that the poverty rate has fallen due to social programs, which even the Trump CEA admits is probably the case even as they try to limit them with the addition of work requirements.

The CEA also denies, without any serious evidence, the findings of Luke Shaefer and Katherine Edin that, since the end of TANF, members of some 1.5 million families have to live on cash income of $2 a day before adding in food stamps. Nobel laureate Angus Deaton has found that 5.3 million Americans live on only $4 a day, once government assistance is included. Analyst Arloc Sherman found that AFDC had lifted more than 1.2 million children above half of the poverty line in 2005, but TANF lifted only 635,000 in 2010. The Trump CEA claims the very poor are not doing as badly as such research shows.

One last point: there is now a veritable mountain of research that shows that children living in families beneath the income poverty line suffer cognitive, emotional, and health damage. This is the clearest testimony to the usefulness of income measures of poverty.

The issues that the Trump CEA attempts to address are complex ones. But the administration’s economists produced a callous advocacy report, not a sharp and comprehensive piece of analysis on a critical subject. And they issued it at a time when wages for those at the bottom of the income distribution are especially low. If the report has influence, it will be a sad commentary on the state of America’s interest in the nation’s poor.