The federal tax system has undergone a profound transformation in the last century in the way it treats low-income households. When it was first instituted in 1913, the modern income tax was levied on only the top 1 percent of earners, and when the payroll tax was added in 1937, it started at a rate of only 2 percent. As a result, the tax system effectively ignored low-income households. But a steady broadening of the income tax base and increases in payroll taxes meant that, by the late 1960s, the tax system was adding substantially to poverty by requiring payments from households that pushed some of them under the poverty line and pressed others still deeper into poverty.

A series of legislative measures passed since the 1970s has reversed this trend. In 1975, the earned-income tax credit (EITC) was created; in 1997, the child tax credit became law. Since their creation, both have been extended, with an expansion included in almost every major tax bill since the mid-1970s. As a result, over the past century we have moved from a tax code that ignored the poor to one that exacerbated their condition to the one we have today that directly reduces poverty for households with children, while increasing incentives for work, education, and advancement.

Looking back at the history of poverty and the tax code in the last several decades reveals some important lessons for expanding opportunity and combating poverty going forward, including the value of having a pro-work, pro-family tax code. The most important new prospect in this area is expanding such an approach for households without children, a proposal that President Obama included in his 2015 budget, and an idea that is also being advanced across the political spectrum, from Senator Marco Rubio to Bush Administration economist Glenn Hubbard to Isabel Sawhill at the Brookings Institution.

This history also shows some of the limits in using the tax code—and government benefits more broadly—to create economically sustainable improvements in incomes. Ultimately, these approaches need to be part of a strategy that raises incomes widely and expands mobility and opportunity.

Trends in Poverty in the Last Five Decades

In January 1964, President Lyndon B. Johnson inaugurated an “unconditional war on poverty,” declaring that “[i]t will not be a short or easy struggle, no single weapon or strategy will suffice, but we shall not rest until that war is won.” What followed was a range of initiatives designed to improve the education, health, skills, and access to economic resources of those in need. Over the next several decades numerous other policies were added or reformed, with a particular focus on shifting to a system that better rewarded and encouraged work.

To measure the impact of the War on Poverty—as well as the broader impact of the economy on trends in poverty and inequality—we must reassess how poverty itself is measured. As part of President Johnson’s War on Poverty, an Official Poverty Measure was developed in the 1960s and subsequently adopted by the federal government. Unfortunately, while it originally provided an informative assessment of poverty, today it is woefully inadequate as a measure of individuals’ well-being, resulting in a distorted understanding of the level of poverty in the United States and, even more importantly, how poverty has changed over time.

The biggest limitation of the Official Poverty Measure is caused by its measure of family resources, which captures pre-tax income (wages and salaries) plus cash transfers (Social Security, unemployment insurance), but not the effects of taxes, tax credits (such as the EITC), or non-cash transfers (such as the Supplemental Nutrition Assistance Program, or SNAP, also called “food stamps”). Because of this, it excludes a significant portion of initiatives directed at the poor. Additionally, the Official Poverty Measure has been found to have several technical and methodological shortcomings in its measure of the costs of basic needs. Fortunately, these deficiencies did not go unnoticed by the Census Bureau, which, under the leadership of Acting Commerce Secretary Rebecca Blank, published for the first time in 2011 a Supplemental Poverty Measure that represents a significant improvement upon the official measure in its methodology for calculating poverty thresholds and family resources.

Unlike the Official Poverty Measure, the Supplemental Poverty Measure tabulates family resources after tax and transfers, including measures such as the EITC. It also subtracts medical out-of-pocket expenses from families’ resources. Using this more inclusive measure, Christopher Wimer, Liana Fox, Irwin Garfinkel, Neeraj Kaushal, and Jane Waldfogel at the Columbia Population Research Center have created an “anchored” version of the Supplemental Poverty Measure that allows them to extend the measure back before the earliest Census estimates in 2009 and adjust poverty thresholds only for inflation. When compared to the Official Poverty Measure, the Supplemental Poverty Measure depicts a profoundly different trajectory for the War on Poverty over the last 50 years. Comparing and adjusting these measures makes it possible to pinpoint what’s responsible for these differing trajectories.

Looking just at “market income” (a household’s income from work and other sources, but not counting taxes or benefits), the poverty rate actually increased slightly from 1967 to 2012, as shown in Figure 1. This is likely due to the fact that cash income growth has been hampered by a rise in inequality in recent years. Economic growth is typically an important antidote to poverty—as long as gains from growth are shared with those in the bottom of the income distribution. Unfortunately, this has not been the case for almost 30 years; rising inequality has left incomes at the bottom relatively unchanged, resulting in the observed increase in poverty as measured without tax credits and benefits. Part of this trend is due to the 25 percent decline in the real value of the minimum wage since 1967, a point this article will return to.

Figure 1 also shows the tremendous progress we’ve made in reducing actual poverty outcomes as experienced by individuals and families. In 1967, the net effect of the tax system and public programs together had a negligible impact on poverty, as benefits like Social Security for some families were effectively offset by taxes on other families that pushed them into poverty. By 2012, however, the net effect of public policies directly cut the poverty rate by more than a third. As a result, over the intervening period, the poverty rate was reduced by nearly 40 percent. Although the struggle persists for far too many, this progress is notable.

Much of this progress was the result of the creation and expansion of tax credits, as well as nutrition assistance, neither of which are included in the Official Poverty Measure. As such, the Official Poverty Measure fails to capture the significant progress in outcomes over this period. All told, public programs have lifted an average of 27 million people per year out of poverty over the past 45 years. This is particularly important whenever critics ask what we have gotten for what they claim are the trillions of dollars spent combating poverty in the last 50 years—we’ve gotten a total of 1.2 billion people-years cumulatively lifted over the poverty line, and higher incomes for many more.

Poverty, the Business Cycle, and the Great Recession

While anti-poverty programs have had beneficial effects on the level and trend in poverty, they have also succeeded in lessening the impact of the business cycle on poverty, especially for those in “deep poverty” (that is, below 50 percent of the poverty line). Counting just market income, deep poverty rises sharply in recessions and tends to fall as the economy recovers. After accounting for taxes and benefits, however, the deep-poverty rate barely registers the business cycle: It is largely protected from dramatic rises during recessions, although it also does not fall that much faster in recoveries. In contrast, using the standard poverty line, we still see a greater reflection of the business cycle in poverty—although that too is changing, as is especially clear in the most recent recession.

The recent financial crisis dealt a severe blow to American families, wiping out more than $13 trillion in household wealth, causing median household wealth to fall by 39 percent, and forcing eight million people out of their jobs. Without any tax or benefit policies, the poverty rate measured by market incomes would have risen by 4.5 percentage points from 2007 to 2010, as shown in Figure 2. That amounts to about 14 million more people, including many from the middle class, who would have fallen into poverty, just based on the economy. Instead, the comprehensive measure of poverty fully reflecting taxes and benefits went up only half a percentage point—about 1.5 million people. While that amount is certainly lamentable, and we should be doing our best to avoid even that outcome, it is a massive difference from the 14 million that would have fallen into poverty absent those policies.

Moreover, this improvement in poverty was the result of a combination of pre-existing policies and important expansions in the American Recovery and Reinvestment Act, which bolstered tax credits such as the EITC and child tax credit, temporarily expanded SNAP benefits, and extended and temporarily expanded unemployment insurance benefits, in addition to giving states incentives to undertake ongoing unemployment insurance reforms. All told, the expansions in 2009 and beyond were responsible for more than 40 percent of the total poverty reduction from tax credits and benefits.

The Direct Impact of the EITC and Child Tax Credit

The creation of the EITC precipitated a dramatic shift in poverty-reduction policy to focusing on promoting work through anti-poverty programs. Initiated under President Nixon and signed into law by President Ford in 1975, the EITC provides a refundable tax credit for working families and individuals that can either offset income taxes or, if larger than the family’s income tax liability, be given as a direct payment. President Reagan once called a bill that included an EITC expansion “the best anti-poverty, the best pro-family, the best job-creation measure to come out of Congress.” The credit equals a fixed percentage of earnings from the first dollar of earnings up until a certain threshold, at which point the credit then stays flat as earnings continue to rise, and then is eventually phased out. Credit amounts vary significantly by marital status and number of children. In 2014, the maximum EITC for a household with two children is $5,460, while for a childless household it’s $496.

The EITC is complemented by the child tax credit, which was originally established in 1997 to provide $500 per child. At its inception, the child tax credit had limited refundability for working families (defined as those in which at least one parent works) with three or more children. The child tax credit has since been expanded to $1,000 per child and is now refundable for 15 percent of earnings in excess of $3,000 through 2017.

The EITC was expanded as part of legislation for revenue neutral tax reform (1986), permanent tax increases on high-income households (1990 and 1993), permanent tax cuts (2001), and temporary tax cuts (2009). Similarly, the child tax credit was first created in a deficit reduction bill (1997), expanded in a tax cut bill (2001), and further expanded in temporary countercyclical bills (2008 and 2009). Most recently, the EITC and child tax credit expansions were extended through 2017 by the American Taxpayer Relief Act of 2012. At the same time, 25 states and the District of Columbia have created EITCs that piggyback on the federal EITC, with the largest benefits going to families with incomes between about $10,000 and $23,000.

The creation and expansion of these tax credits have served as a powerful demonstration that the old adage “a program for the poor is a poor program” need not always be true. That idea, as it was originally propounded by architects of Social Security under Franklin D. Roosevelt, suggested that a program that was not designed to broadly benefit the entire population would be unpopular, subject to threats, and erode over time. Indeed, there has been a long-standing torrent of fierce criticism of the EITC and the partially refundable child tax credit, including claims of fraud, criticism of beneficiaries who end up not paying any federal taxes (going so far as to call them “lucky duckies”), and strong resistance to extending or expanding the benefits from these credits. But due to the effectiveness of these programs, their expansion over time has generally been hailed on a bipartisan basis as a market- and work-oriented way of expanding opportunity and reducing poverty. The political success of these credits over the years is also likely a function both of the inherent work requirement and the fact that they are administered through the tax code, which is a universal system.

These policies have also succeeded partly because they are not just tax credits for one section of the population, but are also measures that provide broader insurance to a much wider set of beneficiaries over time. While in any given year 13 percent of people receive these tax credits, one study found that over an 18-year period, because of fluctuations in income, more than half of taxpayers benefitted from the EITC. (As an aside, a similar point applies to other programs like nutrition assistance or unemployment insurance that are targeted at a specific set of households by income or work status at any point in time but benefit a much wider range of households over time and provide an insurance value to an ever wider set of households.)

The relative size and scope of the tax credits compared to traditional means-tested programs underscores the extent to which poverty-alleviation programs now emphasize employment. Today, expenditures for the EITC and the partially refundable portion of the child tax credit total $79 billion annually, four times more than those for Temporary Assistance for Needy Families (TANF), the major welfare program that used to be known as Aid to Dependent Families and Children (AFDC), as shown in Figure 3. Moreover, programs like TANF have simultaneously been transformed to become more pro-work.

The EITC and partially refundable child tax credit (CTC) have dramatically altered the impact of the tax code on poverty. In 1967, a household at the poverty line paid about 12 percent of its income in federal taxes all told, including payroll taxes. Paying those taxes pushed millions of families below the poverty line, in turn raising the overall poverty rate by 3.2 percentage points, as shown in Figure 4. The impact of the tax system on poverty for nonelderly households with children was even more pronounced, raising the poverty rate by 3.9 percentage points largely because, for those households, the poverty line is somewhat higher to reflect the greater needs of larger families.

The tax system today is dramatically different, working not to increase but to reduce the overall poverty rate by 1.3 percentage points in 2012. Instead of exacerbating the poverty rate for families with children, it lowers it—by a total of 3.7 percentage points in 2012. But the tax system still taxes low-income childless households, raising their after-tax poverty rate.

Although the changes to the tax system since the 1990s have reduced its contribution to poverty among families with children, it has only been since Democrats insisted that the refundability of the child tax credit be expanded as part of the 2008 stimulus that the tax system stopped increasing overall poverty. In 2009, the Recovery Act further expanded the refundability of the child tax credit and made two critical enhancements to the EITC: reducing the marriage penalty that had dramatically cut down on the credit for some low-income people with children who married; and expanding the tax credit for families with three or more children to reflect both their greater expenses and higher poverty rates. Taken together, the anti-poverty policies under the Recovery Act reduced poverty rates by 2.6 percentage points for families with three or more children and 1.3 percentage points for families with one or two children. The EITC and child tax credit policies first enacted in the Recovery Act now benefit 16 million families a year by an average of $900 per family.

The Dynamic Effects of Tax Policies on Households

So far, this discussion of tax credits and other government benefits has focused on their direct effects. That is to say, it provides a static picture that takes a household’s market income as given and asks how adding to or subtracting from it would affect whether or not the household was above or below the poverty line. Academic research has found that this assumption is reasonable in aggregate. For example, research generally finds that nutrition assistance does not discourage work, and thus one can measure its impact on poverty just by looking at the direct benefits it provides in lifting people above the poverty line.

In the case of tax credits, however, this methodology may understate the impact tax credits and other benefits have on poverty, and it entirely misses the impact they have on mobility and intergenerational outcomes. A raft of economic research since the 1990s has found that expansions in the EITC have increased labor force participation among single mothers with children, with little effect on participation among single women without children. Bruce Meyer and Dan Rosenbaum found that it was not just a trivial positive impact; in a paper published in a leading economics journal, they found it was quantitatively very important, and the EITC could explain more of the very large increase in the participation rate for single mothers during the 1990s than the reduction of welfare benefits, welfare waivers, child care, and job training combined.

The most recent research has stressed that the benefits of tax credits are not limited to participation in the workforce but also extend to mobility and opportunity. Hilary Hoynes, Douglas Miller, and David Simon, economists at the University of California, Davis, have found that an increase in EITC income leads to a reduction in low birth weight for children, which is known to have important impacts on opportunity. Economists Gordon Dahl and Lance Lochner have shown that children in households that receive the EITC score higher on reading and math tests than their peers. Finally, work released recently by University of Texas economist Dayanand Manoli and Nicholas Turner of the Treasury Department’s Office of Tax Analysis finds a significant impact of the EITC and tax refunds on college enrollment. In particular, a $1,000 increase in tax refunds received in the spring of a student’s senior year of high school increases college enrollment the following fall by roughly 2 to 3 percentage points.

The child tax credit has been less studied by itself, but research by Raj Chetty and John Friedman of Harvard and Jonah Rockoff of Columbia analyzed its effects in conjunction with the EITC and found that a $1,000 tax credit increases a child’s test score by 6 percent of a standard deviation. (For comparison, high-quality teachers increase achievement by about 10 percent of a standard deviation.) This improvement in test scores in turn implies higher college attendance rates and higher lifetime earnings.

Expanding the EITC for Households without Children

The major changes described thus far apply almost exclusively to households with children. As a result, the federal tax rate for a married couple with two children and with income just at the poverty line has gone from 10 percent in 1967 to -16 percent in 2012, as shown in Figure 5a. But the tax rate for a married couple with no children at the poverty line has been practically constant, going from 12 percent in 1967 to 11 percent in 2012. The same divergent trends appear among single parents, as seen in Figure 5b (note that this assumes single workers with children are filing as head of households). Overall, the emphasis on families with children has been appropriate. The tax system used to do more to add to poverty for households with children than for households without. And even with these changes the poverty rate for nonelderly households with children is still 1.8 percentage points higher than it is for households without children, due to their often higher needs, which is why it would not make sense to expand the childless EITC at the expense of the EITC for households with children.

A small EITC for childless households was established in 1993. In 2015, it is expected to have a maximum value of $503 and is fully phased out for individuals making more than $14,790 ($20,290 for married couples). A childless adult with wages equal to the poverty line is expected to face a federal tax burden of $1,979—her total payroll and income tax burden minus an EITC gain of $170—driving her deeper into poverty and making childless workers the sole demographic for which that is the case.

The President’s proposal would double the childless EITC to be worth up to $1,005 and lower the age threshold from 25 to 21 to help more lower-income young people, while also increasing the upper age limit from 65 to 67 to align with scheduled changes to Social Security’s normal retirement age. The household at the poverty line would see its EITC expand from $170 to $842, more than eliminating its income taxes, although it would still pay net taxes on earnings when including payroll taxes. (Note that these workers would receive returns during retirement through Social Security and Medicare.) The proposal would benefit more than 15 million people by an average of $430, including lifting about half a million people above the poverty line and reducing poverty for ten million more. The EITC expansion would be fully paid for by closing loopholes that let some high-income professionals avoid income and payroll taxes, including the carried-interest loophole.

The proposal is also designed to make the childless EITC more salient, with the raise to $1,005 helping to make more low-wage, part-time, or part-year workers aware of it. The greater salience should induce some of the same behavioral impacts that have been demonstrated in the context of the EITC for households with children, including expanding participation in the workforce and higher paths for incomes. Research by Chetty, Friedman, and economist Emmanuel Saez has found that knowledge of the EITC varies geographically, with many regions not yet taking full advantage of the credit. The authors estimate that if average knowledge of the EITC were equal to that of the current top 10 percent of informed areas, there would be a threefold increase in favorable behavioral responses, such as take-up of the credit.

In addition, by making work more lucrative, an expanded credit could encourage those who are considering whether to replace employment with income support programs to remain in the labor force, thereby reducing expenditures on other social programs, and maybe even lower spending on the criminal justice system by making people less likely to turn to crime. Higher incomes could also encourage family formation, which is historically associated with lower poverty rates. Moreover, notwithstanding its name, many of those who would benefit from expanding the “childless” worker EITC are noncustodial parents, especially fathers. Encouraging these fathers to participate in the labor market and supplementing their incomes will likely benefit their children as well.

The proposed expansion will particularly target several demographics with low or declining labor-force participation rates, including workers with a disability that limits their work capacity, who currently account for about 14 percent of childless EITC recipients; African-American men, whose labor force participation rate has fallen from 74 percent to 64 percent since 1972; and non-college-educated young people at the beginning of their careers. (At the same time, the proposed expansion will maintain protections that also prevent the childless-worker EITC from benefitting full-time students.)

Expanding the childless EITC has increasingly been advocated by members of both parties. Senator Marco Rubio called for an expanded wage supplement for workers without children. Michael Strain of the American Enterprise Institute has said the program should give “more support to childless workers.” Glenn Hubbard, who chaired the Council of Economic Advisers under President George W. Bush, has written that “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.” These arguments have also received support from other conservatives, including Nobel Prize-winning economist Gary Becker and New York Times columnists David Brooks and Ross Douthat. On the other side of the aisle, Isabel Sawhill and Quentin Karpilow at the Brookings Institution recently proposed an EITC reform that would provide a “significant benefit” to childless single individuals. Likewise, the Center for Budget and Policy Priorities and economist John Karl Scholz have long proposed expanding the childless EITC.

What’s Next?

The history recounted here shows just how much has been accomplished through public policy, especially through the tax code. Protecting and preserving these highly effective measures most immediately means making permanent the changes to the EITC and child tax credit that are set to expire in 2017 and also protecting key programs like nutrition assistance. A key next step is to reform and expand these programs. One of the best opportunities with the most current potential is expanding the childless EITC.

However, none of the above should be construed to mean that we should ignore the stagnant market incomes of low-income households. If we are to make significant progress that is economically and politically sustainable, we also need to focus on raising wages and increasing incomes.

This is one key motivation for the President’s support for legislation to raise the minimum wage for all Americans to $10.10 an hour. Doing so in 2016 would raise a family of four with one full-time worker above the poverty line (as shown in Figure 6). As the President noted in his 2014 State of the Union address, “Americans overwhelmingly agree that no one who works full-time should ever have to raise a family in poverty.”

The goal of raising market wages and incomes is the motivation for much of the rest of the President’s economic agenda, including investments in education, infrastructure, and research as well as business tax reform and trade agreements and other policies designed to expand economic growth and to ensure that the benefits from growth are widely shared.

In 2064, when we look back on the War on Poverty at its one hundredth anniversary, if America still faces the same level of “market” poverty we do today, regardless of what further successes we have in reducing poverty through our tax and benefit system, we will have failed as an economy and a society. But this does not mean we should turn our back now on reforming and expanding measures like the childless EITC that boost post-tax incomes. In fact, encouraging work and mobility measures like the EITC are part of the broader strategy to raise pre-tax incomes—and working in tandem these efforts can help achieve a more just and more equal America.