Skip to main content

Greenstein Testimony Before House Budget Committee on Poverty and the Safety Net

I appreciate the invitation to testify today on trends in poverty over the last 50 years, the current state of the safety net, and — most importantly — on some ways that policymakers on both sides of the aisle might work together to make progress going forward.

This testimony begins with a review of the data on poverty and income trends over the last 50 years, various factors that have contributed to significant progress over that period, and factors that have stymied further progress.  I then review data and research on the strengths and shortcomings of policies now in place that seek to reduce poverty and promote opportunity.  Finally, I discuss several ways to make progress going forward, with a particular emphasis on three policy areas that I believe are ripe for bipartisan cooperation and could make a meaningful difference in reducing poverty, increasing employment, and creating more stability and opportunity for poor children.

Conservatives and progressives may not agree on a common understanding of what went right and what went wrong over the last 50 years in our efforts to reduce poverty and promote opportunity.  But I think we all agree that today poverty is too high, children born to low-income families face a more difficult path to economic success, and we want to see more people succeed in the labor market.  What’s heartening is that among both progressive and conservative analysts, there is some consensus on steps we could take to make progress in addressing these issues.

1. Poverty and Income Trends in the 50 Years Since LBJ Declared War on Poverty

There is a quip that we fought a war on poverty and poverty won, but the data don’t support that simplistic assessment.  Poverty has fallen significantly over the last half-century, and key troubling poverty-related conditions have declined.  Since the mid-1960s, average incomes among the poorest fifth of Americans have risen significantly, infant mortality has dropped sharply, and severe child malnutrition has largely disappeared.

Nevertheless, poverty and hardship remain high, and too many of Americans experience difficulty putting sufficient food on the table throughout the month and keeping a roof over their heads.  Nearly 50 million Americans were poor in 2012, including 13 million children.  Some 16 million people lived below half of the poverty line.  And over 1 million children were considered homeless.  Moreover, large racial disparities remain, with child poverty much higher, and the share of African Americans with a college degree much lower, than among whites.  In addition, poverty in America is high compared to other wealthy nations.

The poverty story over the last half-century in the United States is mixed for several reasons.  A much stronger safety net along with factors such as rising education levels, higher employment among women, and smaller families with fewer children helped push poverty down.  At the same time, worsening labor market prospects for less-skilled workers, growing income inequality, and rising numbers of single-parent families have pushed in the other direction.

To measure poverty since the late 1960s by today’s living standards, researchers at Columbia University took the 2012 poverty line under the Census Bureau’s Supplemental Poverty Measure (which analysts generally view as a much more accurate measure of poverty than the “official” poverty measure) and adjusted it for inflation back to 1967.  Under this measure (known as the “anchored” Supplemental Poverty Measure), overall poverty fell from 26 percent to 16 percent between 1967 and 2012.  It fell among children from 29 percent to 19 percent, and among the elderly from 47 percent to 15 percent.[1]   (The researchers also measured poverty using a different poverty line that is based on what people spent for basic necessities such as food and shelter in different years, rather than based on what people spend today and adjusted back for inflation.  This method results in a significantly lower poverty line in 1964, reflecting the lower living standards of that time, and hence a lower poverty rate for 1967 than if a lower poverty line weren’t used.  Under this measure, while poverty still has fallen since 1967, the decline is more modest because the poverty rate for 1967 is set at a lower level.[2] )

The Columbia researchers’ analysis also shows that safety-net programs played a very large role in the reduction in poverty over this period.  To measure how assistance programs affect poverty, researchers can compare poverty rates using two measures of income: 1) a “pre-transfer” measure, which reflects a household’s cash income excluding all government benefits (both cash and non-cash) and also ignoring taxes paid; and 2) the Supplemental Poverty Measure, which is a “post-tax, post-transfer” measure that reflects a family’s income after government benefits (including both cash and non-cash benefits, such as SNAP) and the net effect of the tax system (including income and payroll taxes paid and receipt of tax benefits such as the Earned Income Tax Credit) are taken into account.

Figure 2 shows the share of children in poverty using these two measures.[3]   Under the “pre-transfer” measure, child poverty rose slightly between 1967 and 2012 (though it was modestly lower in 2007, before the Great Recession, than in 1967).  But under the “post-tax, post-transfer” measure that includes government benefits and taxes, child poverty fell significantly — from 29 percent in 1967 to 19 percent in 2012, as noted above.

In addition, public benefits proved highly effective in preventing a substantial increase in child poverty during the Great Recession and its aftermath.  As Figure 2 shows, the child poverty rate would have risen substantially between 2007 and 2012 if one looks only at pre-transfer income, but remained relatively flat under the broader income measure that includes government benefits and taxes.
  • It’s also instructive to look at trends in income over the last 50 years. The average income of the bottom fifth of households, when non-cash and tax-based benefits are included, rose by more than 75 percent between 1964 and 2012, after adjusting both for inflation and for the significant decline in household size over this period.  Both earnings and government assistance contributed substantially to this growth. The bottom fifth of households generally includes poor families and some people modestly above the poverty line.

If we measure from 1973 rather than from 1964, income growth for the bottom fifth has been less dramatic but still notable; the bottom fifth’s average income increased 18 percent between 1973 and 2007, years that are comparable because both were peaks of a business cycle. 

To be sure, incomes fell among the poorest fifth of households during the Great Recession.  Yet increased non-cash benefits like SNAP and unemployment benefits helped buffer the loss.

Figure 3 highlights the growing importance of non-cash and tax-based benefits such as SNAP (formerly food stamps) and the Earned Income Tax Credit (EITC).  If one counts only households’ cash income (including government cash benefits such as Social Security and cash welfare assistance through TANF) but excludes non-cash and tax-based benefits such as SNAP and the EITC, the purchasing power of low-income households was lower in 2012 than in 1973.  But when the non-cash and tax-based benefits are counted, the decline turns into a modest improvement, despite the fact that incomes were pushed down in 2012 by the weak economy.

Incomes grew for both African American and white families (and in more recent decades when data are available, among Latino families as well).  Still, large disparities remain.  Average cash incomes for the poorest fifth of African-American households rose by 41 percent between 1964 and 2012, from just under $7,000 to $9,800.  (As above, these figures adjust for inflation and the decline in household size over this period.)  Non-cash and tax-based benefits boosted the average incomes of this group; when they are taken into account, average income among these families rose by 122 percent — or more than doubled — over this period.  Nevertheless, the bottom fifth of African-American households continue to have very low incomes. 

The poorest fifth of white households saw their cash incomes rise by a somewhat greater amount — 64 percent — over the 1964-2012 period, from $14,800 to $24,300, and they also benefited from increases in non-cash and tax-based benefits.  Their total income including the non-cash and tax-based benefits rose by 86 percent.  While this is a smaller percentage increase than African-American households secured, the average income of the poorest fifth of white households rose by more in dollar terms than did that of African-American households.

The larger dollar, but smaller percentage, increase among white households reflects the stark differences in the incomes of the two groups in 1964, when the average income of the bottom fifth of African-American households, at about $7,000, was less than half the average income of the bottom fifth of white households.  Between 1964 and 2012, this disparity narrowed in percentage terms while growing further in actual dollars.  Today, the average income of the poorest fifth of African American households is $12,000 less than the average income of the poorest fifth of white households, a substantial gap.


Many changes in American society over the last 50 years have affected poverty, some helping to reduce it and others doing just the opposite.

Over this period, the safety net became much more effective at fighting poverty.  While it reduced the number of people who were poor by only 4 percent in 1967, by 2012 it lifted out of poverty 44 percent of those who would otherwise be poor.

Other trends, as well, helped drive poverty lower.  The share of adults finishing high school rose from 56 percent in 1964 to 88 percent in 2012, and the share of adults with a college degree also increased; these developments boosted economic growth and improved the earnings prospects of those who secured more education.  Families also got smaller, so income was shared among fewer people, which lowers poverty rates.  The share of families with children who had fewer than three children rose from 61 percent of families in 1964 to 80 percent in 2012.  This decline in family size occurred across the income spectrum.

In addition, women increasingly worked outside the home.  The share of women aged 18-64 who work rose over this period from 42 to 64 percent. 

But at the same time, other developments impeded progress and kept poverty higher than it otherwise would be.  One such factor was rising earnings inequality, which led to less of the benefits of economic growth going to those at the bottom.  Between 1964 and 2012, the share of national income going to the top 1 percent of households doubled, from 11 percent to 22 percent.  The share of income going to the poorest fifth of households fell between 1979 (the earliest year for which comprehensive income data are available) and 2012.  Poverty would be significantly lower if the benefits of economic growth had been more widely shared as they were during the 1950s, 1960s, and early 1970s.

Another factor is that less-skilled men faced an increasingly inhospitable labor market.  Over the period, the share of adult men (aged 18 to 64) with a job fell from 87 percent to 74 percent, with the reduction occurring largely among men with a high school diploma or less.  Furthermore, while the share of men working year-round whose earnings were less than the poverty line for a family of four fell markedly between 1964 and 1973, this trend then reversed.  Between 1973 and 2012, the share of men who work throughout the year but earn below-poverty wages rose from 10 percent to 14 percent. (This analysis only includes men who, at the time they are surveyed, are working full-time.)

And while parents had fewer children, the share of families with children headed by a single parent grew from 11 percent of such families in 1964 to 35 percent in 2012.  Single-parent families have much higher poverty rates, both because there is only one parent in the home who can work and because these parents tend to have less education and poorer employment prospects. 


Poverty and the Safety Net Today

Data indicate that today’s safety net cuts poverty nearly in half.And the safety net does more than ameliorate hardship in the short term.  A growing body of research shows that programs that help low-income families make ends meet can have positive longer term impacts on children, including improving their educational attainment, long-term health, and later labor-market success.  Too often the safety net is derided for hurting the long-term prospects of those it serves, but careful research does not bear that out.

Reducing Poverty, Providing Health Care

The safety net kept 41 million people, including 9 million children, out of poverty in 2012, according to the Census Bureau’s Supplemental Poverty Measure (SPM).  Specifically, if government benefits are excluded, today’s poverty rate would be 29 percent under the SPM; with those benefits, the poverty rate is 16 percent.  Most analysts view the SPM as a better poverty measure than the “official” measure because it is more comprehensive.  The SPM counts not only cash income but, unlike the official measure, also non-cash and tax-based benefits such as SNAP, the EITC, and rental vouchers.  Also unlike the official measure, it accounts for income and payroll taxes paid and out-of-pocket medical and child care expenses, and it adjusts the poverty line to reflect geographic differences in living costs.

The Census data show that together, the EITC and the Child Tax Credit kept 10.1 million people — including 5.3 million children — out of poverty in 2012.  Unemployment benefits also reduce poverty significantly, especially when unemployment is high:  they kept 2.5 million people (including 600,000 children) out of poverty in 2012.  (In 2009, when unemployment was higher and jobless benefits more generous, they kept 4.9 million people out of poverty, including 1.3 million children.)

In 2012, SNAP kept 4.9 million people — including 2.2 million children — out of poverty.  Moreover, SNAP targets most of its benefits on the poorest households and as a result is a bulwark against deep poverty for many families.  In 2012, SNAP cut the number of children with income below half of the poverty line from 4.9 million to 3.5 million.  Recent work by researchers at the University of Michigan and Harvard University demonstrates the importance of SNAP in keeping households with children out of very deep poverty.  This research shows that SNAP keeps more households with children out of “extreme poverty” — defined as having income of less than $2 per person per day — than any other government program.[4]

In addition to reducing poverty, the safety net helps seniors, people with disabilities, and low-income children and adults access affordable health care.  Medicare and Medicaid, created in 1965, provide health coverage to millions of low-income seniors, people with disabilities, parents, children, and starting in 2014 in states that adopt the Affordable Care Act’s Medicaid expansion, to poor childless adults as well. 

In its early years, Medicaid was primarily limited to families receiving cash welfare benefits and some poor seniors and people with disabilities, but it has since expanded.  Initially these expansions focused on low-income children mostly in working families, and to a lesser degree on parents in those families.  These expansions meant that children did not lose access to health coverage if their parents worked in low-wage jobs and didn’t receive cash welfare benefits.  The Affordable Care Act will extend Medicaid to adults with incomes up to 138 percent of the poverty line in states that elect to implement the expansion.  Health reform also provides near-poor and moderate-income people with subsidies to buy private coverage through insurance marketplaces.

Over the past 15 years, expansions in Medicaid and the creation of the Children’s Health Insurance Program (CHIP) have substantially reduced the share of children who are uninsured.

Longer-Term Positive Effects of the Safety Net

Beyond reducing poverty, alleviating hardship, and giving millions of Americans access to health care, the safety net generates other achievements.  Recent research demonstrates that programs such as SNAP and the EITC have long-term positive educational and health benefits for children. 

A recent study of what happened when food stamps (now called SNAP) gradually expanded nationwide in the 1960s and early 1970s found that disadvantaged children who had access to food stamps in early childhood and whose mothers had access during pregnancy had better health and educational outcomes as adults than children who didn’t have access to food stamps.[5]

Among other things, children with access to food stamps were less likely in adulthood to have stunted growth, be diagnosed with heart disease, or be obese.  They also were significantly more likely to graduate from high school. 

Numerous studies have also found that young children in low-income families do better in school if their families receive additional income from the EITC or similar supports.[6]   For example, one study found that raising family income through refundable tax credits (primarily the EITC) makes it more likely that children in the family will attend college and have higher earnings as adults.  The authors conclude that “a substantial fraction of the cost of tax credits may be offset by earnings gains in the long run.”[7]

Another study found that children in low-income families that received an annual income boost of $3,000 (in 2005 dollars) between the children’s prenatal year and fifth birthday earned an average of 17 percent more as adults, and worked 135 hours more annually, than similar children whose families did not receive the added income.  The additional 135 hours of work is nearly a third of the gap in adult work hours between children raised in poor families and children raised in families with incomes exceeding twice the poverty line.[8]


The EITC’s positive impacts aren’t limited to children; by helping to “make work pay,” the credit also acts as a powerful work incentive for adults.  The EITC expansions enacted in the 1990s “appear to be the most important single factor in explaining why female family heads [of households] increased their employment over 1993-1999,” according to a well-regarded study.[9]


Poverty and Hardship Remain High and Large Racial Disparities Remain

While the safety net reduces poverty and has positive longer term outcomes, no one should be satisfied with the status quo.  Some 49.7 million people were poor in 2012 under the SPM, including 13 million children.  Moreover, 16 million people, including nearly 3.5 million children, lived below half of the poverty line.

U.S. poverty is high compared to other wealthy nations, largely because our safety net does less to lift people out of poverty than other Western countries do.  An OECD study that used a common international definition of poverty found that in the mid-2000s, the poverty rate in the United States before public benefits are taken into account was roughly in line with poverty in many other wealthy nations, but that the less-generous safety net here lifted fewer people out of poverty than the safety nets in other wealthy nations — with the result that poverty is higher here than in most other such nations.


Importantly, many poor families face material hardships.  Nearly six in ten (58 percent) poor children in the United States live in households that experience at least one of four serious hardships during the course of the year — problems affording adequate food, overcrowded housing, falling behind on rent or mortgage payments, or having utilities cut off. 


While average incomes have grown significantly among the poorest fifth of households, as noted above, much of that growth occurred between 1964 and 1973.  During the 34-year period from 1973 to 2007 (before the Great Recession), the average incomes of these households grew a rather modest 19 percent.

In addition, large racial disparities persist.  The child poverty rate among African Americans (29 percent) and Latinos (30 percent) was about 20 percentage points higher than the rate for non-Hispanic white children in 2012.  The share of people aged 25-34 in 2013 who had completed a four-year college degree stood at 41 percent for non-Hispanic whites, but 23 percent for African Americans and 15 percent for Latinos.  Infant mortality, while down substantially over the last 50 years, remains significantly higher among African Americans than among whites.

The Safety Net and Work

Some argue that the safety net impedes progress against poverty by influencing individual behavior in adverse ways, such as by reducing work effort.  Several leading researchers who reviewed all of the available research found, however, that even after accounting for what the research indicates are relatively modest behavioral effects, the safety net still lowered the poverty rate by 14 percentage points in 2014, keeping 40 million people out of poverty.

Economist Robert Moffitt of Johns Hopkins University, one of the nation’s leading and most widely respected experts on the incentive effects of social programs for the poor, (and part of the research team for the aforementioned study) recently reviewed the literature on work disincentives in light of the large increases in safety-net spending during the Great Recession.  Moffitt found “little evidence that the expenditure increases reported here are attributable to work disincentives to any significant degree.”[10]

Often overlooked in this discussion is the way in which the safety net has become much more supportive of work and working families over the last four decades.  Before 1975, little cash or non-cash assistance was available for low-income working families, and families frequently faced a choice between sub-poverty wages and sub-poverty welfare.  That changed gradually with the creation of the EITC in 1975 and the Child Tax Credit in 1997 and subsequent expansions of both credits in 2001 and 2009, and the extension of Medicaid to children in low-income working families.  Federal and state child care assistance for low-income working parents also expanded in the 1990s, and in the past decade, SNAP has become much more focused on serving working families.  The U.S. safety net has become much more of a work-based safety net, providing significantly more assistance to low-income working families with children than in earlier decades and significantly less assistance to poor families without earnings. 

Even so, cross-country comparisons conducted by economists Janet Gornick and Markus Jäntti show that the U.S. benefit and tax system does less to reduce the child poverty rate of working single-mother families than the support systems in each of the 13 other comparably well-off countries they studied, including Canada, Great Britain, Ireland, Australia, and countries in continental Europe and Scandinavia.[11]  The United States also had the weakest safety net for single mothers with little or no earnings; the strengthening of government assistance in recent decades for working families with children was accompanied by a weakening of the safety net for non-working families.

The shift to a more work-based safety net has had costs as well as benefits — in particular, an increase in the number of extremely poor families with children receiving neither earnings nor cash assistance — since the 1996 welfare reform law was enacted.  Addressing the needs of these families in ways that continue to incentivize work but provide a more adequate safety net for very vulnerable children remains an important task.

Where Do We Go From Here?

Regardless of one’s view of the trends over the last 50 years, most people would agree that poverty is too high and that children born to low-income parents have a more difficult path to success than children born to middle- and high-income parents.  The challenge should be to identify ways to make further progress, while maintaining the critical supports that careful research has shown both reduce poverty and improve longer-term outcomes, particularly for children.

First, we need to “do no harm.”  To ensure that poverty doesn’t get worse in the near term, policymakers should temporarily extend federal jobless benefits to long-term unemployed workers.  Although the unemployment rate has fallen to 6.7 percent, much of that decline reflects abnormally slow growth in the labor force due to limited job opportunities.  The more telling metric is that which shows the percentage of people aged 16 and over who have jobs.  This figure fell markedly in the recession and has recovered only modestly since.  It remains only modestly above the level to which it fell at the recession’s bottom.

When Congress let federal unemployment benefits expire last month, an estimated 1.3 million people who have been looking for work for over six months saw their unemployment insurance (UI) benefits end abruptly, and the economic recovery lost a valuable source of purchasing power.  CBO estimates that the economy will have up to 300,000 more jobs in the fourth quarter of 2014 if the federal UI benefits are extended than if Congress doesn’t reinstate them.

In short, while Emergency Unemployment Compensation must be a temporary program, economic conditions haven’t yet improved enough to end it.  Indeed, the long-term unemployment rate — the percentage of the labor force that has been out of work more than six months — is nearly twice as high today as the highest it ever was when any of the emergency federal UI programs that policymakers enacted in each of the previous seven major recessions expired, as the chart below shows.

Moving beyond the immediate need to extend jobless benefits, we know that to make substantial progress in reducing poverty, we need more people to do better in the labor market — that is, we need to raise both employment rates and earnings levels.  One part of the answer should be raising the minimum wage.  Today’s minimum wage is 22 percent below its late 1960s peak, after adjusting for inflation.  Increasing the minimum wage to the $10-an-hour range would help to offset some of the unfavorable trends facing low-wage workers, including stagnant or falling real wages, too little upward mobility, and a deficit of bargaining power that leaves them solidly on the “have-not” side of the inequality divide.

The question of whether raising the minimum wage reduces employment for low-wage workers is one of the most extensively studied issues in empirical economics.  The weight of the evidence is that for minimum wage levels in the range now being discussed, such impacts are small, and that minimum-wage increases that are of the size that’s been enacted in the past — and would occur under the proposals now being discussed — are a net benefit to low-wage workers as a group.  Raising the minimum wage also would modestly lower poverty to some degree[12] and help push back against rising inequality.

Some opponents of raising the minimum wage argue that it would primarily benefit teenagers working for extra money, but the large majority of those who would benefit are adults, most of them women.  Indeed, the average worker who would benefit brings home half of the family earnings.  This reflects the fact that the low-wage workforce has gotten older (and more educated) in recent decades:  the share of low-wage workers (those earning less than $10 per hour in 2011 dollars) who are between the ages of 25 to 64 grew from 48 percent of low-wage workers in 1979 to 60 percent in 2011.  The share with at least some college education grew from 25 percent to 43 percent.

But while we strongly support an increase of this dimension in the minimum wage, that’s but one step.  If we are serious about reducing poverty and increasing opportunity, there are a number of other policy and program changes we should pursue, including extending high-quality early education to more low-income children and providing help paying for child care to more low-income parents so they can look for and accept jobs and make ends meet.[13]   Another area where there are new promising results is in funding ways to  help low-income students not only attend, but successfully complete, two-year and four-year college degrees that translate into better jobs and higher earnings.[14]   And just as it is important to improve children’s preparation for school and what happens after high school, continued efforts to help low-income children succeed in elementary, middle, and high school are important as well.  Education will not solve every problem.  But it certainly can make a large difference in children’s future prospects.

Education, education reform, and child care are important, but not my areas of expertise.  So today, I’d like to focus on three other areas where there is some support among both conservative and progressive analysts for policy improvements and where I think real bipartisan progress could be made.

Strengthening the EITC for Childless Adults

Policymakers have made substantial progress in recent years in “making work pay” for low-income families with children by strengthening the EITC and Child Tax Credit.  But low-income workers not raising minor children receive little or nothing from the EITC.  For example, a childless adult working full time at the minimum wage is ineligible for the EITC, because his earnings exceed the very low income limit for the tiny EITC for workers not raising minor children.  Partly a result, childless workers are the sole group of workers in American whom the federal tax system taxes into — or deeper into — poverty. 

Moreover, all childless workers under age 25 are flatly ineligible for the EITC, so young people just starting out receive none of the EITC’s proven benefits, such as promoting work,[15] alleviating poverty, and supplementing low wages.

The average credit for those eligible childless workers who do qualify for the credit is very small at just $270, or one-tenth the average $2,790 credit for filers with children.  In addition, the childless workers’ EITC begins phasing out when earnings exceed $7,970, or just 55 percent of full-time, minimum-wage earnings.  As a result:
  • A childless adult working full time throughout the year at the minimum wage — and earning $14,500 — receives no EITC.  This worker has a federal income and payroll tax burden of $2,669 in 2013, a large tax burden for someone with income this low.[16]  
  • And a childless adult with wages equal to the Census Bureau’s poverty line (projected at $11,905 in 2013) faces a federal income and payroll tax burden of $1,826 (including the employer share of the payroll tax) but receives an EITC of just $186.  Such workers are literally taxed into poverty. 

Providing a more adequate EITC to low-income childless workers and lowering the eligibility age so younger workers can qualify would have several important benefits beyond raising these workers’ incomes and helping offset their federal taxes.  Some leading experts from across the political spectrum believe that an expanded credit would help address some of the challenges that less-educated young people (including young African American men) face, including low and falling labor-force participation rates, low marriage rates, and high incarceration rates.

On the employment side, the more robust EITC for families with children has been shown in multiple studies to have a substantial positive impact on employment rates among parents, particularly single mothers.[17]   As discussed above, one well-regarded study concluded that the EITC expansions in the 1990s were more important for increasing employment rates among single mothers than the changes to cash welfare programs that came about as part of welfare reform.  Because the EITC is only available to people who work, and — for those with the lowest earnings — is larger the more that a person works, an expanded EITC for childless adults would create a positive incentive for these workers to work (or to work in the above-ground economy).

Some analysts think that by increasing the returns to work, an expanded EITC for childless adults can also have a positive impact on marriage rates and reduce involvement in criminal activity.  A 2009 study found that three-quarters of low-income, unwed survey respondents cited financial concerns as an obstacle to marriage.[18]   Work by the eminent scholar William Julius Wilson focuses on the lack of “marriageable men” in poor communities as a significant factor that increases the share of children who grow up in single-parent families.  If more low-income men are employed and gain a toehold in the economy, some of those concerns could lessen. 

Similarly, while there are myriad reasons that individuals engage in criminal activity, researchers have found that lower wages for less-educated people are associated with higher crime rates.[19]   By boosting the effective wage of low-paid workers, an expanded EITC could encourage more people to take low-wage jobs in the above-ground economy and reduce the returns to crime.

It is important to note that an expanded EITC for childless adults would be good policy even if these marriage and crime-related impacts did not materialize.  Creating a positive work incentive and reducing poverty among low wage workers are important goals by themselves.

In part because an expanded EITC for childless adults is pro-work and pro-marriage, it has gained substantial support among a growing number of conservative analysts as well as among centrist and progressive analysts.  For example:

  • Noting that the EITC “promotes work as it reduces poverty,” former George W. Bush economic advisor Glenn Hubbard wrote recently, “Increasing the credit for childless workers to an amount closer to that for families with children would augment the direct work incentive and help counter poverty among the working poor.”[20]
  • “Don’t forget the poor,” Lori Sanders and Eli Lehrer of the R Street Institute (a conservative think tank) recently urged Weekly Standard readers, explaining that “Properly structured work incentives would build on … the Earned Income Tax Credit, which remains decidedly modest.  For a single worker without children living at home, the EITC refunds less than $425 per year.  Introducing and expanding similar wage supplements . . . would further encourage a life of work as preferable to welfare or life in the underground economy.”[21]
  • The American Enterprise Institute’s (AEI) Michael Strain noted that the EITC “gives very little help to childless workers” and called for amending the EITC “to offer more support to childless workers.”[22]
  • David Neumark, a University of California economist whom conservatives often cite on minimum wage issues, asked “whether we are providing a sufficient income floor for childless low-wage adults, who have to a large extent been left out in the cold by our income-support and safety net programs.”[23]

In addition, Senator Marco Rubio (R-FL) recently highlighted[24] why the federal government should do more to subsidize the wages of low-income workers, including childless adults (though he proposed replacing the EITC with a new, untested wage subsidy that would be very difficult for the IRS to administer fairly and accurately and is an ill-advised approach to achieving a very laudable goal).[25]  

Such bipartisan interest in the EITC isn’t surprising; the credit has enjoyed broad bipartisan support over the years.  President Ford signed it into law; President Reagan lauded it as one of our best anti-poverty programs and proposed and signed a major EITC expansion, because the credit helps low-income people struggling to make ends meet while encouraging work and personal responsibility. 

As Richard Burkhauser of Cornell University and AEI said recently, “I’m not exaggerating when I say, look, I’ve been doing public policy since the 1970s, and this program worked.” [26]  The Nobel laureate Gary Becker of the University of Chicago, a renowned conservative economist, has made similar, highly laudatory comments about the EITC, noting that it “increases the labor force participation and employment of people with low wages.”[27]

Subsidized Employment

Policymakers should also look at creating subsidized jobs, primarily in the private sector.  The Recovery Act provided modest funding to states that they could use for several purposes, including subsidized jobs.  Thirty-nine states and the District of Columbia established subsidized jobs programs for jobless low-income parents and youth, with the states that did so representing a real cross-section of the country.  About half of the states that established subsidized jobs programs had Republican governors; the other half had Democratic governors.

States worked with private and non-profit employers, as well as government agencies, to create these subsidized job positions — but most of the job placements were in the private sector.  Many of the programs worked directly with private employers and required them to contribute to the costs of providing a subsidized job placement.  For example, South Carolina paid participants’ wages up to the minimum wage, with employers required to cover any additional wage costs as well as related payroll costs.  Oklahoma and Mississippi provided wage subsidies that started out as full wage subsidies but declined over time.  States adopted a variety of approaches regarding the maximum wage level that could be subsidized and how long the wage subsidy could last per employee.[28]   Typically, employers created positions that low-income parents or youth filled for temporary periods such as 6 months to a year. 

The program proved highly successful.  Over a 1½ year period, these states succeeded in placing 260,000 low-income parents and young people in subsidized jobs.  Moreover, the Economic Mobility Corporation (EMC) studied what happened to participants in these subsidized jobs programs and found the programs did exactly what they were supposed to do — help disadvantaged jobless individuals find work during hard economic times.  The study also provides evidence that the jobs programs improved some participants’ chances of finding unsubsidized jobs when their time in the subsidized job position came to an end.[29]   And the study indicated that the long-term unemployed benefitted most.

The EMC study looked at subsidized jobs programs established in five sites — Los Angeles, San Francisco, Wisconsin, Mississippi, and Florida — and produced the following findings:[30]

  • Participation in subsidized employment led to significant increases in employment and earnings. Participants in four of the five programs studied were much more likely to have an unsubsidized job in the year after working in a subsidized job than in the year before joining the program.  The findings from Florida are especially noteworthy, because researchers could compare participants there with applicants who were eligible for the program but didn’t receive a subsidized job before the program ended.  (The study confirmed that the subsidized job participants and those in the comparison group were largely similar on key demographic characteristics.)  Participants in Florida earned an average of $4,000 more in the year after the program than in the year before it, compared to a $1,500 increase for people in the comparison group.
  • The programs were especially effective for the long-term unemployed. In Mississippi and Florida, average annual earnings of long-term unemployed individuals rose by about $7,000 after participating.  In Los Angeles and Wisconsin, they rose by about $4,000.  In these four sites, earnings rose much more among the long-term unemployed than among people who had been unemployed for shorter periods.
  • Employers reported hiring more workers than they would otherwise have hired and hiring workers with less experience than their usual hires.
  • Most participating employers reported multiple benefits from the program, including expanding their workforces, serving more customers, and improving their productivity.

Because of its success, state policymakers from across the political spectrum praised the program.  For example, Governor Haley Barbour of Mississippi noted that, “Mississippi STEPS [the name of the subsidized jobs program in Mississippi] is unique in that it is a program specifically designed to benefit both the employee and employer… provid[ing] much-needed aid during this recession by enabling businesses to hire new workers, thus enhancing the economic engines of our local communities.”[31]   Governor Barbour said of the program, “It’s welfare to work.”

The Recovery Act funding for the subsidized jobs program expired in 2010, but there is growing support among analysts across the political spectrum for this type of strategy, as it helps to address several fundamental problems — too few jobs (especially in the current economy) for less educated workers, a substantial number of workers who have been out of work a long time, and the lack of sufficient work experience among significant parts of the low-income population.  Those recommending such an approach include my fellow panelist today, Ron Haskins,[32] and Kevin Hassett, Director of Economics at the American Enterprise Institute.[33]

I recommend creating a subsidized employment program of this nature that would provide modest funding to states to create ongoing programs targeted on groups that face particular difficulties finding jobs.  Such a program could be structured to target hard-to-employ low-income individuals, such as long-term jobless workers and disadvantaged young adults.  It could provide funding for subsidized jobs for people who have difficulty finding employment and getting a toehold in the labor market even during good economic times — which could help them to overcome those barriers and get on a path to increased labor-market success.  Funding for these jobs could then be increased during recessions in order to temporarily increase the number of subsidized jobs that are created during periods when the economy is shedding jobs in large numbers.

Any such ongoing program should be coupled with an evaluation to identify its overall impact and, in particular, the approaches that are most successful.  MDRC is currently conducting two evaluations of subsidized employment programs that will increase our knowledge of what works.  Providing new funding for subsidized jobs programs coupled with evaluation funding now would create much-needed job opportunities in the near term while the labor market remains difficult for many workers and also provide a larger set of programs to evaluate and learn from.

Reforming and Expanding Housing Choice Vouchers

My final recommendations involve increasing the number of low-income families that receive Housing Choice Vouchers for private rental housing and reforming some aspects of the voucher program.  Many poor children would have a better chance of moving up the economic ladder if they were able to live in stable housing in safe neighborhoods with access to well-performing schools.  (This is especially true of the more than 1.1 million children and youth who were homeless during the 2011-2012 school year, according to the Department of Education[34] and the additional homeless children too young to be counted in this school-based survey.)

Expanding and improving the Housing Choice Voucher program could enhance poor children’s prospects in several ways.  We know, for example, that housing vouchers reduce homelessness; a rigorous evaluation involving families with children poor enough to be eligible for welfare assistance found that the availability of housing vouchers reduced the incidence of homelessness among these families by 75 percent.[35]  Numerous studies have also shown that living in stable housing, rather than having to move repeatedly, is linked to better long-term educational and health achievement among poor children.[36]

Housing vouchers can also give families and children access to better opportunities by enabling them to move from high-crime areas with few jobs and failing schools to areas with more job opportunities, better schools, and less crime.  For example, recent research shows that families that had the opportunity to use a housing voucher to move to a less-poor neighborhood were less likely to suffer from extreme obesity and diabetes,[37] and a growing body of evidence indicates that the stress that children can experience from growing up poor, particularly in neighborhoods of concentrated poverty, can adversely affect their health, education, and long-term economic prospects.[38]   For children living in particularly violent neighborhoods, using a voucher to move to a less-poor, safer neighborhood appears to lead to an increase in their test scores.[39]  And where housing policies have enabled low-income children to attend high-performing, economically mixed schools over time, the students have scored higher on math and reading tests than comparable children who attended higher-poverty schools.[40]

Results such as these have helped the Housing Choice Voucher program generate bipartisan support.  But the program falls short in two important ways. 

First, its reach is quite limited.  Only about one in four low-income families eligible for rental assistance receives it; waiting lists for assistance are at least several years long in most parts of the country.  This problem has only become more acute as gaps between rental charges and what poor families can afford have widened.  More than 8 million low-income households who don’t receive any federal housing assistance paid more than half of their income for rent and utilities in 2011, a 43 percent increase since 2007.[41]   Rents have been increasing in recent years at a faster pace than overall inflation.[42]   Meanwhile, the number of families able to receive rental vouchers fell last year due to sequestration.  (The 2014 funding levels will restore only about half of the vouchers that were lost.) 

Second, a core feature of the program — that families are supposed to be able to use their vouchers to move to where there are more jobs, better  schools, and/or less crime — hasn’t worked as well as it should.  Families often experience difficulty in trying to use their vouchers to move to significantly less-poor neighborhoods.[43]   Enhancing the potential of housing vouchers to enable families to move to areas with more jobs and/or better schools could increase economic mobility.[44]

There’s consequently a need both to provide vouchers to more families (which entails an increased level of funding for this voucher program) and to institute reforms to better enable the program’s “choice” function to work as Congress intended.

In the reform area, we recommend that Congress encourage HUD to modify various policies in the voucher program to increase the ability of families to use vouchers to live in safer communities with better access to jobs and well-performing schools.  For example, in metropolitan areas where vouchers are concentrated in poorer neighborhoods, HUD could set the Fair Market Rents on which the voucher subsidy levels are based at the zip code level, rather than using the same rent limits across an entire metropolitan area, which ignores the fact that rents can vary widely across a metro area.  Such a reform would increase housing opportunities in “opportunity neighborhoods” without increasing federal costs.   

In addition, as part of a long-overdue overhaul of how it measures local agency performance in administering the voucher program, HUD could reward agencies that succeed at enabling more families to use their vouchers in high-opportunity areas.  The recent report of the Bipartisan Policy Center’s Housing Commission called on HUD to adopt such an outcome-based performance measurement system that promotes the de-concentration of poverty and access to neighborhoods with more opportunity.[45]

Even with such changes, however, the administration of vouchers by a plethora of separate local agencies within the same metro area often hinders families’ ability to use their vouchers to move to areas with greater opportunities.  HUD contracts with 2,300 separate public housing agencies (PHAs) across the country to administer the Housing Choice Voucher program.  Some metropolitan areas include as many as 60 separate PHAs — and hence 60 individual voucher programs — each with its own administrative apparatus.  The sheer number of PHAs undermines the effectiveness of the program at promoting mobility and other goals, and also increases costs. 

There is no sound policy justification for administering separate tenant-based rental assistance programs for every individual town within a metropolitan area.  Larger programs operating on a metro-wide basis would not only be more cost-effective but, if well-managed, could also expand voucher holders’ access to neighborhoods that offer greater economic opportunities and better schools. [46]   

Congressional action would be required for such a significant restructuring of program administration.  If such a restructuring could not be achieved due to political resistance from local officials, policymakers could still adopt several significant measures to move in this direction, including:

  • Encouraging the formation of consortia of PHAs in a geographic area.  If PHAs in an area would at least form a consortium that has a single funding contract with HUD, and families could use their vouchers to move seamlessly within the cities and towns in the consortium, that would be an important step forward.  (Consolidation of separate housing agencies to form a single metro-wide PHA would potentially have greater benefits, but also poses greater political hurdles; for many PHAs, the ability to retain their independent identity is a paramount concern.  This makes it more likely that PHAs would join a consortium to achieve administrative economies of scale than to formally consolidate with other agencies.)  Under HUD’s current rules, however, agencies have little incentive to form consortia.[47]   HUD could be asked to revise various rules to reduce administrative burdens for agencies that form consortia and to provide incentives for such action, and Congress could provide modest funds to help agencies with the transition costs to operating as consortia.
  • Ceasing to pay for inefficiency.  Under current funding formulas, HUD gives smaller housing agencies higher subsidies per-housing-unit for their voucher-program administrative costs.  This provides an incentive for agencies to remain small rather than to form consortia or consolidate.  We estimate the voucher program could save about $40 million per year just by eliminating the higher per-voucher fee that agencies receive if they administer 600 or fewer vouchers.[48]  
  • Placing poor performing agencies on a shorter leash.  HUD has the authority to take over administration of programs that are in substantial default or are “troubled” and fail to improve satisfactorily.  Among the remedies Congress has empowered HUD to use are to consolidate the poorly performing agency with a willing, well-managed PHA or to appoint another PHA or private management entity to manage a poorly performing agency’s programs.  HUD could strengthen its performance assessment tools and use available remedies in response to poor performance to foster the formation of larger, more effective and efficient local programs.

End Notes

[1] Christopher Wimer et al., “Trends in Poverty with an Anchored Supplemental Poverty Measure,” Columbia Population Research Center, December 2013,

[2] Liana Fox, Irwin Garfinkel, Neeraj Kaushal, Jane Waldfogel, and Christopher Wimer, “Waging War on Poverty: Historical Trends in Poverty Using the Supplemental Poverty Measure,” Columbia Population Research Center, December 2013,

[3] The two measures reflected in Figure 2 are based on the anchored SPM.

[4] H. Luke Shaefer and Kathryn Edin, “Rising Extreme Poverty in the United States and the Response of Federal Means-Tested Transfer Programs,” University of Michigan National Poverty Center Working Paper 13-06, May 2013,

[5] Hilary W. Hoynes, Diane Whitmore Schanzenbach, Douglas Almond, “Long Run Impacts of Childhood Access to the Safety Net,” National Bureau of Economic Research Working Paper No. 18535, November 2012,

[6] For a more in-depth discussion of this research literature, see “Various Supports for Low-Income Families Reduce Poverty and Have Long-Term Positive Effects on Families and Children” Center on Budget and Policy Priorities, July 2013,

[7] Raj Chetty, John N. Friedman, and Jonah Rockoff, “New Evidence on the Long-Term Impacts of Tax Credits,” Statistics of Income Paper Series, November 2011,

[8] Greg J. Duncan, Kathleen M. Ziol-Guest, and Ariel Kalil, “Early-Childhood Poverty and Adult Attainment, Behavior, and Health,” Child Development (January/February 2010), pp. 306-325.)  The $3,000 figure is in 2005 dollars, equivalent to approximately $3,530 in 2012.  The 17 percent figure appears as 19 percent but is a typo; a revised figure of 17 percent appears in Duncan and Magnuson (2011).

[9] Jeffrey Grogger, “The Effects of Time Limits, the EITC, and Other Policy Changes on Welfare Use, Work, and Income among Female-Head Families,” Review of Economics and Statistics, May 2003.  Using different data, in another study, Grogger reaches similar conclusions.  Jeffrey Grogger, “Welfare Transitions in the 1990s: the Economy, Welfare Policy, and the EITC,” NBER Working Paper No. 9472, January 2003,

[10] Robert A. Moffitt, “The Great Recession and the Social Safety Net,” The Annals of the American Academy of Political and Social Science, 2013, 650: 143,

[11] Markus Jäntti and Janet Gornick, “Child Poverty in Comparative Perspective: Assessing the Role of Family Structure and Parental Education and Employment,” Paper No. 570, LIS Cross-National Data Center, Luxembourg, September 2011,

[12] See, “Minimum Wages and the Distribution of Family Incomes,” Arindrajat Dube, December 2013,

[13] Most analysts and program administrators agree that more needs to be done to raise the bar on quality in early education programs.  Some are overly pessimistic, however, about Head Start.  While studies often show that the positive impacts which Head Start (and other early childhood programs) have on measures of academic achievement diminish over time, longer term research on Head Start nevertheless finds significant long-term positive effects on high school completion, college enrollment, health status, and being either employed or in school.  See, David Deming, “Early Childhood Intervention and Life-Cycle Skill Development: Evidence from Head Start,” American Economic Journal: Applied Economics, 2009,

[14] See, for example, recent findings from an evaluation of the ASAP program through the City University of New York.  The evaluation, conducted by MDRC, showed significant gains in retention and graduation.

[15] For a summary on research on the EITC, see Chuck Marr, Jimmy Charite, and Chye-Ching Huang, “Earned Income Tax Credit Promotes Work, Encourages Children’s Success at School, Research Finds,” Center on Budget and Policy Priorities, Revised April 9, 2013,

[16] This figure includes the employer and employee shares of the payroll tax.  Economists generally believe that employees ultimately bear the employer share of the tax in the form of lower wages then they would otherwise receive (although this may not be the case for people paid the minimum wage).  This figure is $1,109 if only the employee share of the payroll tax is included.

[17] For a review of this research, see, “Earned Income Tax Credit Promotes Work, Encourages Children’s Success at School, Research Finds,” Center on Budget and Policy Priorities, April 2013,

[18] Kathryn Edin and Timothy Nelson, “Why Do Poor Men Have Children? Fertility Intentions Among Low-Income Unmarried US Fathers,” The ANNALS of the American Academy of Political and Social Science, July 2009, Vol. 624, No. 1, pp. 99-117.

[19] For a discussion of recent research and estimation issues, see David Mustard, “How Do Labor Markets Affect Crime? New Evidence on an Old Puzzle,” IZA Discussion Paper 4856, March 2010,

[20] Glenn Hubbard, “Tax reform is the best way to tackle income inequality,” Washington Post, January 10, 2014,

[21] Lori Sanders and Eli Lehrer, “Don’t Forget the Poor,” Weekly Standard, August 26, 2013,

[22] Michael Strain, “More than the Minimum Wage,” National Review, December 11, 2013,

[23] David Neumark, “The Minimum Wage Ain’t What It Used to Be,” New York Times, December 9, 2013,

[24] Senator Marco Rubio, “Reclaiming the Land of Opportunity: Conservative Reforms for Combatting Poverty,” remarks as prepared for delivery, January 8, 2014,

[25] Sharon Parrott, “Rubio Proposal to Replace EITC Would Likely Come at Expense of Working-Poor Families with Children,” Off the Charts blog, January 9, 2014,

[26] Richard Burkhauser, “Burkhauser: The earned income tax credit actually works,” American Enterprise Institute, March 15, 2013,

[27] Gary Becker, “How to End Welfare ‘As We Know it’ — Fast,” Business Week, June 2, 1996,

[28] For more information on how these programs were structured and the different approaches that states adopted, see “Creating Subsidized Employment Opportunities for Low-Income Parents: The Legacy of the TANF Emergency Fund,” by the Center on Budget and Policy Priorities and the Center for Law and Social Policy, February 2011,

[29] Ann Roder and Marc Elliot, “Stimulating Opportunity: An Evaluation of ARRA Funded Subsidized Employment Programs,” Economic Mobility Corporation, September 2013,

[30] In four of the five sites, these findings compare participants’ employment and earnings in the year before they were placed in a subsidized job to their employment and earnings in the year after the subsidy ended.  In one site, Florida, the study’s researchers were able to compare the employment and earnings of program participants to the employment and earnings of similar individuals who applied for the program but weren’t placed before the program ended.  This latter approach provides stronger evidence that participation in a subsidized job had a positive impact on participants’ employment and earnings after the program ended.

[31] “STEPS: Stimulus funds spent on jobs?”  Clarion Ledger, February 18, 2010,

[32] Ron Haskins, “No Way Out:  Dealing with the Consequences of Changes in Family Composition,” Washington, DC:  Brookings Institution, forthcoming.

[33] Kevin A. Hassett, Testimony Before the Joint Economic Committee, “Long-Term Unemployment: Consequences and Solutions,” American Enterprise Institute, April 24, 2013,

[34] National Center for Homeless Education, “Education for Homeless Children and Youths Program Data Collection Summary, October 2013.  Children are considered homeless in this survey if they are living in homeless shelters or motels or hotels, doubled up with other families, or sleeping in cars or abandoned buildings or on the street.

[35] Michelle Wood, Jennifer Turnham, and Gregory Mills, “Housing Affordability and Well-Being: Results from the Housing Voucher Evaluation,” Housing Policy Debate 19:367-412 (2008).

[36] Diana Becker Cutts, MD, “US Housing Insecurity and the Health of Very Young Children,” American Journal of Public Health, August 2011, Vol. 101, No. 8, p. 1508; Lorraine E. Maxwell, 2003. “Home and School Density Effects on Elementary School Children: The Role of Spatial Density,” Environment and Behavior 35(36438), (2003), pp. 566–78.  Dalton Conley, “A Room With a View or a Room of One’s Own?? Housing and Social Stratification.” Sociological Forum 16(2), (2001), pp. 263–80. 

[37] Jens Ludwig et al., “Neighborhoods, Obesity, and Diabetes — A Randomized Social Experiment,” New England Journal of Medicine, 365:16, October 2011,

[38] Jack P. Shonkoff, MD, et al., “The Lifelong Effects of Early Childhood Adversity and Toxic Stress,” PEDIATRICS, Vol. 129 No. 1, January 1, 2012, pp. e232 -e246,; Patrick Sharkey, Stuck in Place: Urban Neighborhoods and the End of Progress toward Racial Equality, Chicago and London: The University of Chicago Press, 2013.

[39] Julia Burdick-Will et al., “Converging Evidence for Neighborhood Effects on Children’s Test Scores: An Experimental, Quasi-Experimental, and Observational Comparison,” in Greg Duncan and Richard Murnane, eds., Whither Opportunity? Rising Inequality, Schools, and Children's Life Chances, pp. 255-276, Russell Sage Foundation, September 2011,  Similar neighborhood effects have not been shown in rigorous demonstrations where the crime levels in the neighborhoods did not differ greatly.

[40] Heather Schwartz, “Housing Policy is School Policy,” The Century Foundation, 2010,  

[41] U.S. Department of Housing and Urban Development, Worst Case Needs 2011:  Report to Congress, Summary, February 2013.

[42] Joint Center for Housing Studies of Harvard University, “America’s Rental Housing---Evolving Markets and Needs,” December 2013. 

[43] Martha M. Galvez, “What Do We Know About Housing Choice Voucher Program Location Outcomes,” Urban Institute, August 2010,  Ingrid Gould Ellen and Keren Mertens Horn, “Do Federally Assisted Households Have Access to High Performing Public Schools?”  PRRAC, November 2012,

[44] Huiping Li, Harrison Campbell and Steven Fernandez, “Residential Segregation, Spatial Mismatch and Economic Growth across US Metropolitan Areas,” Urban Studies, 50(13) 2642–2660, October, 2013; Raj Chetty et al., “Where is the Land of Opportunity? The Geography of Intergenerational Mobility in the United States,” National Bureau of Economic Research, January 2014,

[45]   Bipartisan Policy Center Housing Commission, “Housing America’s Future: New Directions for National Policy,” February 2013, p. 99.

[46] Regional voucher administration was most recently recommended by urban policy experts Bruce Katz (of the Brookings Institution) and Margery Austin Turner (of the Urban Institute) in their paper “Invest but Reform: Streamline Administration of the Housing Choice Voucher Program,” Brookings Institution, September 2013,

[47]   A recent Government Accountability Office (GAO) study reports that since 1998, when Congress modified the U.S. Housing Act to permit consortia, only four consortia have been formed (involving nine housing agencies that administer vouchers).  U.S. General Accountability Office, “Housing Choice Vouchers: Options Exist to Increase Program Efficiencies,” GAO-12-300, March 2012.   

[48] A majority of the PHAs that administer fewer than 600 vouchers are in metropolitan areas or the same non-metro county as another small PHA.  In a number of states, a state-level agency administers the program in the state’s rural areas.  About 500 PHAs that administer fewer than 600 vouchers, however, are the only voucher administrator in their county. HUD has the authority to increase administrative fees for agencies that serve widespread geographic areas, irrespective of the number of vouchers the agencies administer.  HUD could use this authority to boost administrative fees appropriately for these agencies, for which forming a consortium may not make practical sense, as well as others.