There is much with which to agree in Michael Tanner’s Liberty Forum essay summarizing the post-1996 history of welfare reform, its successes and failures, and where to go from here.
However, I view the Personal Responsibility and Work Opportunity Reconciliation Act as having succeeded to a greater extent than Tanner believes. Indeed, as I have argued elsewhere, together with the expansion of the Earned Income Tax Credit (EITC), welfare reform was the most successful policy effort in modern history to reduce poverty among nonelderly Americans. That said, it is clear that there are a great many improvements that could be made to our web of antipoverty programs, and Tanner’s proposals, by and large, are sensible ones.
The ultimate indicator of the success of welfare reform is whether or not it lowered poverty, and if so, by how much. The answer to these questions are elusive.
First, there are important technical issues related to poverty measurement. The official poverty rate—used to assess eligibility for various federal programs and for other government purposes—is based on relatively arbitrary thresholds that vary with the number of adults in a family, the number of children, and depending on whether or not the family head is elderly.
However, in assessing whether a family falls below the relevant threshold, income from three of the four largest federal programs for low-income Americans is ignored. Benefits from “cash transfers” are counted, so that incorporates income from the sizable Supplemental Security Insurance (SSI) program. Not counted is income from the refundable EITC, a tax-time payout to low-income taxpayers with earnings that was expanded substantially in 1993. Also ignored are benefits from the Supplemental Nutrition Assistance Program (SNAP, better known as food stamps) and from Medicaid. These programs deliver “in-kind” benefits in the form of food vouchers and health insurance that are not completely fungible as is cash. However, they are surely valuable to the families that receive them.
Poverty status is also determined on the basis of pre-tax income. But tax burdens have fallen substantially among low-income families. In part, but not entirely, due to the EITC, average federal tax rates for the poorest fifth of households fell from 6.4 percent in 1996 to 1.9 percent by 2011, according to the Congressional Budget Office. The additional disposable income available to low-income families by virtue of lower taxes over time is not reflected in poverty statistics.
Finally, the poverty thresholds are adjusted to take increases in the cost of living into account every year. But, as I have shown elsewhere, inflation is overestimated by the “price deflator” used by the Census Bureau. That means each year there is a higher bar for a low-income household to reach in order to escape “poverty.”
Note that all three of these technical problems make the progress we have made in reducing poverty look worse than reality. Benefits from the federal programs excluded from income have expanded dramatically since 1996. Leaving those benefits out of income in 1996 makes too many people in 1996 look poor, but leaving them out in 2016 is far bigger a problem. Similarly, falling taxes overstate poverty more over time. And if families in 1996 had to achieve the same standard of living implied by today’s poverty thresholds, more would be deemed poor than implied by official 1996 poverty rates.
Fortunately, recent research by Chris Wimer, Liana Fox, Irv Garfinkel, Neeraj Kaushal, and Jane Waldfogel of Columbia University addresses the issues related to in-kind benefits and taxes. If we were to use the official poverty rate for children, we would conclude that child poverty was higher in 1993 than ever before and that the 2012 rate was about the same. But that would ignore the rising income from in-kind benefits and falling taxes. The Columbia team’s estimates indicate that the child poverty rate was the same in 1993 as it was in 1967. But they show poverty falling dramatically thereafter—from 29 percent to 17 percent in 2009. In 2012, the child poverty rate stood at 19 percent—only a bit higher than its all-time low, despite the Great Recession.
The Columbia estimates update the poverty thresholds each year using a better cost-of-living adjustment than that used for the official poverty thresholds. But they do not use the best price deflator available, which shows lower inflation over time. In other words, the drop in poverty would be steeper using a more appropriate cost-of-living adjustment.
So it is not the case that the increase in antipoverty spending Tanner documents has failed to reduce poverty. Poverty fell a lot after welfare reform. The next question, of course, is how much credit should welfare reform get for this decline?
The strong 1990s expansion deserves some credit for the initial drop, from 1993 to 2000. But the decline in poverty, as measured by the Columbia researchers, was much steeper than in the expansions of the 1970s and 1980s. And if changes in the poverty rate were primarily about the state of the economy, the child poverty rate should have risen back to its previous levels—particularly during the Great Recession.
However, the child poverty rate barely budged upward during the two post-1990s recessions. Further, the employment trends differ notably between single mothers and other women, suggesting that something other than the business cycle was at work. While employment rates among single mothers rose after 1993, fewer single childless women worked over the next ten years, and employment among married mothers was unchanged. The era through 1993 and the era from 2000 forward are divided by seven years in which a permanent shock occurred.
There were actually two such shocks. The expansion of the EITC in 1993 was one reason for the strong labor supply response after 1993. That is, apart from pushing down the poverty rate directly through lower taxes and bigger refunds, the EITC reduced poverty indirectly by subsidizing work and thereby encouraging it.
But it is unlikely that the EITC expansion by itself would have produced such a large drop in child poverty. The share of children who were poor fell in 1994 and in 1995, but the decline abruptly stalled in 1996. Similarly, employment rates among unmarried mothers jumped in 1994 and rose again the next year, but not in 1996. From 1996 to 2000, however, employment among single mothers rose steadily and child poverty dropped.
Welfare reform was the second shock to child poverty rates. Most states began implementing the reform legislation in 1997, coinciding with the 1996-to-2000 decline in child poverty. Indeed, some of the earlier 1993-to-1995 drop should properly be attributed to welfare reform or anticipation of reform. Waivers to federal welfare requirements had been granted in most states by 1996, candidate Bill Clinton’s end-welfare-as-we-know-it rhetoric on the 1992 campaign trail heralded the inevitability of reform, the introduction of President Clinton’s reform proposal in 1994 made his pledge concrete, and Congressional Republicans passed two welfare reform bills in 1995 (both vetoed by Clinton).
The drop in the welfare rolls was without precedent. In 1994, at 58 percent, welfare receipt as a share of single-mother households was only down 13 percentage points from its 1972 high, 22 years earlier. By 1998, four years later, it had fallen 24 percentage points. It fell another 17 points over the next ten years; in 2008 TANF receipt as a share of single-mother households approached 17 percent.
The 1990s boom, the EITC expansion, and welfare reform worked together to finally reduce child poverty. The EITC encouraged single mothers to take advantage of the strengthening job market, whereas in the past, falling unemployment had drawn fewer of them into the work force. Welfare reform made it unappealing to try to get by without working. Once the boom was over, the EITC made work pay more generously than in the past, and the shriveling of the welfare rolls provided further incentive to find and keep a job.
Welfare reform, then, promoted opportunity—both by reducing poverty and by increasing the number of children with a breadwinner to look up to. It may have even, as I have argued, arrested the rise in out-of-wedlock birthrates, by making it more expensive to raise a child alone and by mandating that teen recipients live with a guardian or be enrolled in school or a training program. Births per 1,000 single women stopped rising in 1995, after half a century, and the birthrate fluctuated with the economy after that (falling during downturns, rising during expansions). Strikingly, births per 1,000 married women rose after the early 1990s, reversing a 25-year decline.
What is more, the impacts on birthrates were largest among groups of women most dependent on welfare benefits. Welfare receipt is highest among younger women, and it is higher among African American women than among white women. Trends in out-of-wedlock births showed the most improvement among teens, and improved less impressively in each successive older group. Within each age group, the improved trends were more impressive among black women than among white women.
Despite this successful record, there is, as Tanner suggests, ample room for improvement in our antipoverty policy. The success of welfare reform was only secondarily due to effective state efforts to strengthen and expand the skills of beneficiaries and get them into jobs. More important was the way that time limits and work requirements made welfare less attractive than working.
Rather than stay on the rolls, after 1996, many women who could always have worked decided to do so. They did not need elaborate job training schemes to find employment. Further, the sharp drop in welfare rolls reflected not only an increase in outflow from TANF, but a reduction in inflow onto the rolls. One important consideration—for the upcoming reauthorization of welfare reform and for reforms of other safety-net programs—is whether more effort should be devoted to preparing those with employment barriers for jobs or whether simply making receipt of public benefits unappealing should be the primary mechanism for shrinking the rolls and increasing work.
A strong case can be made that time limits and work requirements with generous exemptions will ensure that most would-be beneficiaries will choose work instead and be better off for it. And this may also be the approach that is best for those who will always struggle to maintain independence. A key element of welfare reform was that states could count declines in the welfare caseload toward the requirement that a minimum percentage of beneficiaries be engaged in work activities. Furthermore, sizable numbers of beneficiaries were exempted from the calculation of work-participation rates.
These and other loopholes have allowed states to report impressive-sounding rates without actually having many beneficiaries engaged in work activities. The effect has been to shelter some of the most vulnerable families from the tougher aspects of welfare reform. Arguably, this end result is preferable to expectations that a large share of those who do not choose work will be able to find jobs and remain in them successfully. Placing beneficiaries with difficult challenges will require additional funding, and to the extent that work-participation programs are effective, they might actually attract more families onto welfare rolls to take advantage of them.
There is, however, some evidence that welfare reform caused a small number of families with the greatest challenges to become worse off. As Tanner notes, the evidence on this point is far from clear, but even if deep poverty has been stable over the past 20 years, we ought to be concerned about the kinds of families described in Kathryn Edin’s and Luke Shaefer’s 2015 book, $2 a Day: Living on Almost Nothing in America. There are several million children who are living in misery, often because of the poor decisions of their parents, but often because of bad luck, low parental skills, or impoverished family social capital.
To the extent that states and localities are dissuading the most vulnerable families from receiving TANF benefits when they are eligible or making it difficult to access the program, they should be discouraged from doing so. A relatively unobtrusive way to achieve that goal would be to require states to meet a caseload adequacy requirement tied to a minimum ratio of TANF recipients to families with children in deep poverty (that is, families under 50 percent of the poverty line.)
Another set of children who have been ill-served by efforts to minimize the TANF caseloads are those in families that have been diverted to the SSI rolls. Because SSI funding is overwhelmingly federal and because it has no work requirements or time limits, diverting would-be TANF beneficiaries to SSI is appealing to states. Richard V. Burkhauser and Mary C. Daly show (in their 2011 book), that it may seem appealing to short-sighted parents, too, since it is more secure than TANF.
But because children receiving SSI must have physical or mental disabilities, many of the children from these diverted families are eligible for benefits on the basis of questionable diagnoses, including learning disabilities or anxiety. If these children show signs of recovery, that jeopardizes the family’s eligibility for benefits, creating perverse incentives in terms of school performance and progress. And many child SSI recipients “graduate” at eighteen into the adult SSI program, never accumulating work experience or jobs on that first rung of the career ladder. Many children would be much better off within TANF.
None of this should be taken to suggest a crisis big enough to overturn the conclusion that welfare reform was a great success. In particular, calls for large public jobs programs should be rejected. The extremely poor have profound employability issues. A public jobs program would be expensive and might expand over time. And it is unnecessary to the scale of the demand-side labor market problem. Welfare reform showed that there is sufficient labor demand for additional low-skilled workers.
Time limits, work requirements, generous exemptions from these strictures, and well-funded work supports like the EITC and the Child Tax Credit (CTC)—this was the formula for welfare reform’s success. It reduced material hardship substantially. Because of it, poverty stayed lower than it had been, even after the 1990s boom and even during the Great Recession.
As Tanner emphasizes, welfare reform only affected a small portion of the safety net. The evidence that reforming cash welfare reduced poverty suggests that the model established by reform should be ported to other safety-net programs like SNAP and public housing assistance. Several worthy proposals also call for consolidating these programs into a single block grant to streamline administration and increase state flexibility.
Reforming disability programs—SSI and the Social Security Disability Insurance (SSDI) program—to promote work and independence among beneficiaries with relatively minor physical or mental ailments should be another goal of antipoverty policy. They have become, in addition to vital programs supporting children and adults with relatively severe disabilities, long-term unemployment programs for a growing number of able-bodied but less-skilled adults. These programs are not optimal for them even if we ultimately decide that we need long-term unemployment policies for adults with few marketable skills.
Finally, Tanner is right to emphasize that simply giving money to poor families is unlikely to expand their opportunities. Ultimately, we should be thinking more about upward mobility than about poverty, and increasing mobility will require effective policies to promote economic growth and child opportunity. Welfare reform has demonstrated what one component of this upward mobility agenda should look like, but a war on immobility will require additional approaches.
So two-and-three-quarters cheers for welfare reform, and a Happy 20th Birthday to the most important antipoverty policy (in conjunction with the EITC expansion) of the past generation.
 Prior to 1982, there were also different thresholds for male and female and farm and non-farm family heads. The thresholds originally were derived from research in the mid-1960s and intended to approximate the 1963 cost of purchasing a “food plan” specified by the Department of Agriculture. Individual-level food costs (by age and sex) were grouped into family-level costs for families of different sizes. Those costs were then multiplied by three to arrive at poverty thresholds, reflecting the finding that food costs took up one-third of the income of the average family. A lower multiplier was used for families of two, and the threshold for a single person was set at 80 percent of the threshold for two-person families. Thresholds for families on farms were set as a percentage of the nonfarm thresholds. This set of thresholds was then simply updated for inflation over time. The different thresholds for male and female heads and for farm and non-farm families were eliminated in late 1981. Gordon M. Fisher, “The Development and History of the Poverty Thresholds,” Social Security Bulletin 55:4 (1992), 3-14.
 These cash transfer programs include the old Aid to Families with Dependent Children (AFDC) welfare program and its successor, Temporary Assistance for Needy Families (TANF), as well as the Supplemental Security Income (SSI) program, unemployment insurance, workers’ compensation, and the Social Security retirement, disability, and survivor benefit programs.
 Other in-kind benefit programs include various housing assistance programs, the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC), and school breakfasts and lunches.
 Statement of Heather Boushey, U.S. Equal Employment Opportunity Commission, Meeting of April 17, 2007, Perspectives on Work/Family Balance and the Federal Equal Employment Opportunity Laws. Estimates are for women age 25 to 45.
 Scott Winship and Christopher Jencks, “How Did the Social Policy Changes of the 1990s Affect Material Hardship Among Single Mothers? Evidence from the CPS Food Security Supplement,” Faculty Research Working Paper, John F. Kennedy School of Government, Harvard University, June 2004 (RWP04-027). https://research.hks.harvard.edu/publications/getFile.aspx?Id=127; Winship (2016), op cit.
 Research indicates that the EITC played a bigger role during the 1990s than welfare reform in affecting employment and poverty among single mothers and their children, but it is doubtful that any of the research strategies involved can disentangle the effects from each other. See Hilary W. Hoynes and Anjur J. Patel (2015). “Effective Policy for Reducing Inequality? The Earned Income Tax Credit and the Distribution of Income.” National Bureau of Economic Research Working Paper 21340; Jeffrey Grogger, “The Effects of Time Limits, the EITC, and Other Policy Changes on Welfare Use, Work, and Income Among Female-Headed Families,” Review of Economics and Statistics 85:2 (2003), 394-408. The Grogger study has been mischaracterized by welfare reform critics to argue that PRWORA was unimportant. See Scott Winship, “How Would Paul Ryan’s Opportunity Grants Affect Poverty?”, Forbes.com, July 28, 2014. http://www.forbes.com/sites/scottwinship/2014/07/28/how-would-paul-ryans-opportunity-grants-affect-poverty/#4377abe42c38.
 Because of under-reporting of income in household surveys, using a $2-a-day benchmark is inadvisable.
 In places where there is little demand, such as the Mississippi Delta, something like residential mobility assistance to move families to opportunity would be more effective than trying to create useful work in those places. On residential mobility and economic mobility, see Scott Winship, “When Moving Matters: Residential and Economic Mobility Trends in America, 1880-2010,” Manhattan Institute, e21 Report No. 2, November 10, 2015. https://www.manhattan-institute.org/html/when-moving-matters-residential-and-economic-mobility-trends-america-1880-2010-8048.html.
 For complementary economic growth and child opportunity agendas, see Scott Winship, “Modernizing Federal Fiscal Policy for Economic Growth,” in New Entrepreneurial Growth Agenda (Kansas City: Ewing Marion Kauffman Foundation, 2016) http://www.kauffman.org/neg/section-7#modernizingfederalfiscalpolicyforeconomicgrowth; and Scott Winship, “Up: Expanding Opportunity in America,” in Policy Options for Improving Economic Opportunity and Mobility (New York: Peter G. Peterson Foundation 2015). http://www.pgpf.org/sites/default/files/grant_cbpp_manhattaninst_economic_mobility.pdf