Has the Safety Net Greatly Expanded over the Last Five Years?
An Examination of Heritage Foundation Claims
March 7, 2006
The Heritage Foundation recently issued a paper by Brian Riedl claiming that programs for the poor have expanded greatly on President Bush's watch and have done very well in the budget changes of recent years. This claim has been promoted in several op-eds by Riedl and repeated in various media interviews.
Riedl often cites one key statistic as proof for this claim -- that federal expenditures for health care, housing, food, and cash assistance programs for low-income households rose 39 percent between fiscal years 2001 and 2005. The 39 percent increase statistic is becoming a frequently repeated sound-bite.
In fact, however, the 39 percent figure is misleading, and careful examination of the data produces a different picture than the one that Riedl paints.
- To generate the 39 percent figure, Riedl compared expenditures for these programs in fiscal years 2001 and 2005 without adjusting for inflation. Yet prices for food, health care services, medications, rent and other necessities all rose over this four-year period.
- Riedl also failed to adjust for the large increase in the number of poor people that occurred over this period, when about five million more people fell into poverty.
Once one adjusts for inflation (using the Consumer Price Index) and the rise in poverty, Riedl's 39 percent increase turns into an increase in expenditures for these programs of just 8 percent. In other words, spending for these programs rose eight percent per poor person over this period in inflation-adjusted terms (for an average increase of 1.9 percent per year). Moreover, as explained below, much of this 8 percent increase simply reflects increases in health care costs that rose faster than inflation in the public and private sectors alike. The largest anti-poverty program -- Medicaid -- actually was made less generous over this period, rather than more so, even though rising health care costs drove Medicaid expenditures up.
This 8 percent increase when inflation and the rise in poverty are taken into account is essentially attributable to growth in three programs: Medicaid, food stamps, and refundable tax credits. The increase in these programs is discussed below. The combined cost of all other low-income programs that Riedl examined shrank by 6 percent per poor person over this period, after adjusting for inflation. This paper concludes with a discussion of why, over the long run, low-income families are likely to be large net losers as a result of the policies adopted in recent years, which is a very different outcome than that which Riedl implies.
What Does the 8 Percent Increase Signify?
The fact that inflation-adjusted expenditures increased 8 percent per poor person over this period should not be taken to mean that the generosity of safety-net programs increased by 8 percent. To understand what a figure such as this does -- and does not -- signify, one must examine factors that affected the costs of major safety-net programs over this period.
Below we briefly review three major anti-poverty programs that essentially account for all of the real increase in expenditures per poor person that occurred. The review finds that the 8 percent per-poor-person increase in spending did not result primarily from decisions to make these programs more generous. Indeed, in the program where the largest increase in spending per poor person occurred -- Medicaid -- policy decisions made the program less generous overall.
- Medicaid: Increases in Medicaid costs are the single largest cause of the 8 percent rise in expenditures per poor person in inflation-adjusted terms. This rise in Medicaid expenditures primarily reflects large increases in health care costs throughout the U.S. economy; these costs have been rising considerably faster than inflation for many years. The increase in Medicaid expenditures also reflects the transfer of health insurance costs for several million people in low-income working families from private employers to Medicaid as employer-based coverage contracted. What the increase in Medicaid costs does not signify is an increase in the generosity of Medicaid.
To the contrary, during this period, numerous states pared back their Medicaid programs by lowering eligibility limits and/or scaling back the health services that Medicaid covers. As one example, between 2001 and 2005, Missouri lowered its Medicaid income eligibility limit for working-poor parents from 100 percent of the poverty line to 22 percent of the poverty line (from $16,600 for a family of three in 2006 dollars to about $3,700). Even so, Missouri's Medicaid expenditures rose by 39 percent in nominal terms between fiscal years 2001 and 2005. Medicaid costs everywhere, even in states like Missouri, were considerably higher in 2005 than in 2001 because doctors, hospitals, and other medical providers charged significantly more and because Medicaid enrollment climbed as some employers dropped coverage (and as the number of poor people increased). To cite figures on the increase in Medicaid expenditures and imply that this indicates Medicaid has been made more generous, or has not been made more restrictive, is misleading. (Moreover, changes in Medicaid made in the budget reconciliation legislation signed into law on February 8, 2006, are likely to lead to reductions in Medicaid benefits in numerous states in the years ahead, through increases in co-payments and premiums that various categories of beneficiaries will be required to pay and reductions in the health care services for which they are covered.)
- Food Stamps: Food stamp participation, and hence food stamp costs, rose substantially between 2001 and 2005, another point that Riedl emphasizes. He fails to explain that this increase represented a recovery from the partial collapse in food stamp participation that occurred between 1995 and 2000, and that participation in 2005 -- rather than representing some new historic high -- still remained below the mid-1990s level. Monthly food stamp participation averaged 26.6 million people in 1995. It then plummeted to 17.2 million in 2000 and 17.3 million in 2001. USDA studies show that the percentage of eligible households receiving benefits fell sharply during this period, which appears to have occurred in part as a result of serious initial problems in adapting the food stamp program to the sweeping changes being made in the welfare system. In 2005, after food stamp participation had largely recovered, average monthly participation stood at 25.7 million people. (The rebound was due in modest part to reforms enacted in 2002 that were designed to help address the unintended decline in the percentage of working-poor families receiving benefits and also to soften cuts enacted in 1996 that disqualified many legal immigrants from food stamps; these changes increased food stamp costs by less than three percent, according to Congressional Budget Office estimates. The increase in the number of people in poverty also played a significant role here.) Riedl's figures start with food stamp participation at the low point in this decline-and-recovery cycle and thus show large cost increases for the 2001-2005 period.
Another, smaller factor contributing to the rise in food stamp costs over this period is that poor households became poorer, on average, between 2001 and 2005, as Census data show. When low-income households become poorer, their food stamp benefits rise; when they become less poor, their benefits decline.
Tax Credits: Tax credits for low-income working families became more generous during this period. In addition to raising the child tax credit to $1,000 per child, the 2001 tax cut law made the credit partially refundable for low-income working families with children that have incomes over $11,000 (in 2005 dollars). It also modestly increased the Earned Income Tax Credit for married families with incomes above $14,370, to reduce the marriage penalties the EITC can create. Yet these expansions in low-income tax credits will constitute only 4 percent of the total costs of the 2001 and 2003 tax-cut laws when those laws are fully in effect; this is a significant point, because if the 2001 and 2003 tax laws are made permanent, it is likely they eventually will be financed at least in part through substantial budget reductions (since the tax cuts cannot be deficit-financed forever). As a result, despite the significant increases in low-income tax credits, the tax-cut laws enacted in 2001 and 2003 are likely to lead ultimately to a net reduction in assistance for low-income families rather than a net increase, as the full costs of the tax cuts come due.
In short, the notion that poor families have been big winners in the budget decisions made since 2001 does not withstand scrutiny. To be sure, programs for poor families have not been cut deeply at this point. But difficult budget decisions will have to be made over the long run, and the fiscal profligacy of recent years -- especially on the tax side of the budget but also with respect to the prescription drug benefit, because its costs were not offset -- will make the corrective measures that ultimately have to be taken harsher and more painful than otherwise would have been necessary. History suggests that when the budget-cutting knife is unsheathed, programs for poor families tend to have weaker protection than programs that serve more powerful constituencies. Over the course of time, low-income programs and the people they assist are likely to be among the largest losers when the full consequences of the budget and tax decisions of recent years are felt.
 Census data show that the number of poor people increased by 5.4 million, or 17.1 percent, between calendar year 2000 and calendar year 2004. Poverty data are not available by fiscal years, which start on October 1. Eligibility for means-tested programs often is based on income in previous months or the previous year.
 This does not mean that expenditures per poor person declined in real terms for every one of the other programs, but that expenditures for these programs declined in the aggregate.
 Census data show that the amount by which the average poor person fell below the poverty line increased by more than $200 (or 8 percent) between 2000 and 2004, after adjusting for inflation, and that the percentage of people living below half of the poverty line (measured both as a proportion of all Americans and as a portion of poor Americans) rose over these years. In addition, Agriculture Department data show that the average real income of food stamp households fell during this period.
 Based on estimates from the Joint Committee on Taxation and the Congressional Budget Office, when the 2001 and 2003 tax cuts are fully in effect, the costs of the EITC expansion and the new refundable Child Tax Credit enacted in 2001 will account for four percent of the costs. This reflects costs in 2011, and assumes that the tax cuts are extended. The cost of the tax cuts, as used here, includes the additional cost of relief from the Alternative Minimum Tax that results from the 2001 and 2003 tax cuts. If the AMT cost is not included, the low-income tax credit expansions will account for 5 percent of the total cost of the 2001 and 2003 tax cuts when those tax cuts are fully in effect.
 See William G. Gale, Peter R. Orszag, and Isaac Shapiro, "The Ultimate Burden of the Tax Cuts: Once They are Paid For, Low- and Middle-Income Households Likely to be Net Losers, on Average," Urban Institute-Brookings Institution Tax Policy Center and Center on Budget and Policy Priorities, June 2, 2004.