SNAP Failure The Food Stamp Program Needs Reform
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
No. 738 October 16, 2013 SNAP Failure The Food Stamp Program Needs Reform by Michael Tanner Executive Summary The Supplemental Nutrition Assistance Pro- liberate policy choices by both federal and state gram (SNAP), the food assistance program for- governments, which loosened eligibility stan- merly known as food stamps, has become Amer- dards and actively sought new participants. At ica’s fastest growing social welfare program. As the same time, evidence that the expansion of recently as 2000, just 17 million Americans par- SNAP has significantly reduced hunger or im- ticipated in the program at a cost of less than proved nutrition among low-income Americans $18 billion. Today, roughly 48 million Ameri- is scant at best. cans receive SNAP benefits, costing taxpayers SNAP is a deeply troubled program. It has more than $78 billion per year. Yet according high administrative costs and significant lev- to the U.S. Department of Agriculture (USDA), els of fraud and abuse. The program’s work re- nearly 18 million American households remain quirements are weak and frequently evaded at “food insecure.” the state level. The program increasingly breeds This many households living in hunger has greater dependence on government. It has little raised significant questions about whether the “bang for the buck.” growth of SNAP has been justified and whether The time has come to reform the food stamp it successfully addresses hunger in America. program by reducing its spending and enroll- The evidence suggests that much of the in- ment and, ultimately, by returning responsibil- crease was due not to the economy but to de- ity for its operation to the states. Michael Tanner is a senior fellow with the Cato Institute and author of The Poverty of Welfare: Helping Others in Civil Society (2003).
SNAP has been Introduction administration in 2000 to 26 million people one of this by 2006. The 2008 farm bill—the Food, Con- This country’s first food stamp program servation, and Energy Act of 2008—contin- country’s fastest was temporary, running from 1939 to 1943. ued this trend. It simplified the application growing social It allowed low-income Americans to pur- process further and made it easier still for chase food that the USDA considered sur- people to enroll, indexed the asset limits to welfare programs plus. In some ways, it was as much a farm inflation, and changed the name from food in recent years. price support program as an anti-poverty stamps to the Supplemental Nutrition As- one. At its peak, 4 million Americans partici- sistance Program. President Obama’s 2009 pated at a cost to the government of $262 stimulus bill further expanded the program, million.1 The program was phased out as increasing benefits by 13.6 percent and sus- food surpluses were depleted, and concerns pending the time limits for SNAP benefits were raised about the lack of congressional for Able-Bodied Adults without Dependents authorization and about highly publicized (ABAWDs). As a result, roughly 48 million instances of fraud and abuse. Americans receive SNAP benefits today, In the years following, several attempts thereby costing taxpayers more than $78 bil- were made to re-establish the program, but lion per year.3 it wasn’t until 1959 that Congress passed This rapid growth has forced a reassess- legislation authorizing the USDA to again ment of the program. This year, for the issue food stamps to low-income Americans. first time since the program began, the Even then, the Eisenhower administration House of Representatives voted to consider ultimately decided not to go forward with food stamps separately from the farm bill, implementation. The Kennedy administra- thereby breaking up the political coalition tion did implement several food stamp pi- that had traditionally backed the program.4 lot programs, but it wasn’t until Congress Congress is also expected to consider cuts in passed the Food Stamp Act of 1964 at the program spending as well as other reforms urging of President Lyndon B. Johnson that to the program. the modern food stamp program began. Therefore, it is important to consider Food stamps were always seen as fulfill- the causes of the recent growth in SNAP, to ing two separate goals—“improved levels of consider whether that growth has been justi- nutrition” and “strengthening the agricul- fied, and, perhaps more important, to con- tural economy.”2 The food stamp program sider whether the program has succeeded in has been this two-pronged mission that has addressing hunger in America. ensured bipartisan support over the years, with traditional liberal advocates for the poor joined by farm state conservatives in Explosive Growth backing the program. However, beginning under President As noted, SNAP has been one of this George W. Bush, and even more rapidly un- country’s fastest growing social welfare der President Barack Obama, the number of programs in recent years, and it is now the Americans receiving food stamps and the as- nation’s second most costly means-tested sociated cost of the program have risen dra- program behind Medicaid. Whereas the matically. For example, the 2002 farm bill— program has grown steadily since its incep- the Farm Security and Rural Investment Act tion, rapid expansion started under Presi- of 2002—expanded eligibility to noncitizens, dent Bush and then escalated even more increased benefits for large families, and rapidly under President Obama (Figure 1). made it easier for people to claim benefits. Since 2000, spending on SNAP increased As a result, participation increased from 17 from just $17 billion per year to more than million Americans at the start of the Bush $78 billion in 2012, a greater than fourfold 2
Figure 1 Cost and Enrollment Growth 2001–13 Cost, Billions of Dollars Participation, Millions 90 80 70 60 50 40 30 20 10 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Source: U.S. Department of Agriculture “Supplemental Nutrition Assistance Program Participation and Costs,” Food and Nutrition Services, http://www.fns.usda.gov/pd/SNAPsummary.htm. increase. The increased spending was driven rollment in food stamps increased by only both by an increase in the number of re- 635,000, and spending rose by just $124 cipients (a surge from 17 million in 2000 to million (in constant 2012 dollars). During more than 48 million today) and an average the 1990–92 recession and jobless recovery, benefit per person that has almost doubled. enrollment increased by 5.2 million, and Today, nearly one out of every six Americans spending rose by $9.1 billion. During the receives SNAP. most recent recession (over a comparable SNAP is jointly administered by federal three-year period), enrollment increased by and state governments, thereby giving states 12 million people, while spending increased more leeway to set eligibility rules than with by $30 billion.6 The Congressional Budget some other welfare programs. As a result, Office (CBO) estimates that about 35 per- the program’s growth has varied significant- cent of the program’s growth from 2007 to ly from one state to another. For example, 2011 was caused by non-economic factors. since 2008, the number of people receiving This CBO estimate suggests that much of SNAP benefits has doubled or more in Flori- the increase was due to deliberate policy da, Idaho, Nevada, Rhode Island, and Utah.5 choices.7 Of course, some of this increase could In addition, if one looks at food stamp The program’s be considered countercyclical because wel- receipt rates state by state, one finds little fare programs automatically expand during correlation between unemployment rates or growth economic downturns, such as in the recent poverty rates and the number of people on has varied recession. However, increases in both par- food stamps. Thus, poor states such as Loui- significantly ticipation and spending were bigger during siana and Mississippi have high food stamp this recession than in previous ones. For participation (above 20 percent), but so does from one state to example, during the 1980–82 recession, en- Oregon, where more than 21 percent of the another. 3
SNAP’s eligibility population receives food stamps in a state rest of the decade and beyond (Figure 3). In- requirements with an unemployment rate of just 8 per- deed, in 2023, SNAP will still cost taxpayers cent and an average wage above the national at least $73.2 billion, more than double the have been average.8 Tennessee has very high participa- total cost in 2007 and more than four times significantly tion, but Alabama, Georgia, and Kentucky, the cost of the program in 2000.11 with similar demographics, do not. Vermont relaxed. has high participation; New Hampshire has Three Factors of Expansion much lower participation.9 Aside from the recession, three factors Moreover, both enrollment and costs are have driven the expansion of SNAP. expected to remain high for the next several Relaxed Eligibility. First, and perhaps years, despite the economic recovery and de- most important, SNAP’s eligibility require- clines in unemployment. As Figure 2 shows, ments have been significantly relaxed. SNAP participation is expected to peak in 2013 is no longer a program targeted at the poor- at 47.7 million recipients, before starting est Americans who may need some tem- to decline slightly. However, participation porary help, but it has become part of an will remain above 40 million until 2019 and ever-growing permanent welfare state. To- would still be more than 34 million in 2023, day, nearly 17 percent of SNAP households double the number of recipients in 2000.10 have incomes above the poverty line.12 And Food stamp participation at levels far in ex- almost 4.5 million recipients are ABAWDs, a cess of historical averages will become the group that accounts for more than 10 per- new normal. cent of the beneficiaries.13 Program costs will also peak in 2013, at Low-income families can become eligi- $83 billion, before leveling off. However, ble for SNAP in several ways. For instance, SNAP will remain one of the most expen- households can qualify for SNAP benefits sive social welfare programs throughout the if they meet the program’s income and as- Figure 2 Projected Average Monthly SNAP Participation 2000–23 60 50 40 Millions 30 20 10 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 Source: Congressional Budget Office, “Supplemental Nutrition Assistance Program: May 2013 Baseline,” http:// www.cbo.gov/publication/44211. 4
Figure 3 SNAP Costs 2000–23 90 80 70 60 Billions ($) 50 40 30 20 10 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 Source: Congressional Budget Office, “Supplemental Nutrition Assistance Program: May 2013 Baseline,” http:// www.cbo.gov/publication/44211. set test: a gross income below 130 percent a brochure or referral to an “800” number of the poverty level and a net income below telephone hotline.14 100 percent of poverty, as well as less than Because the cash assistance component of $2,000 in assets (although there are some ex- TANF is more tightly targeted than its other emptions, such as the value of houses, a car, benefits, expanding categorical eligibility for and retirement accounts). Households with SNAP to include these other components a person over age 60 or who is disabled have means that some households with higher a higher asset threshold. income can qualify for benefits.15 Some of However, more often participants become this shift to more people qualifiying through eligible for SNAP because they are also eligi- broad categorical eligibility is due to the ef- ble for other government welfare programs. fectiveness of the TANF reforms in reducing Nearly two-thirds of households receiving welfare caseloads; because significantly fewer SNAP qualify through this type of categori- people receive cash assistance from TANF cal eligibility and were not subject to asset compared to its peak years, then naturally tests or certain income tests (although they fewer people qualify for SNAP through the remain subject to the net income test). receipt of cash assistance. But much of the Initially, categorical eligibility was limited shift is due to the spread of broad-based to those receiving TANF (then called Aid to categorical eligibility throughout the states Families with Dependent Children, but now in recent years. The number of states using USDA has changed to Temporary Assistance for Needy the broad categorical eligibility policies has been actively Families). However, the 1996 welfare reform expanded rapidly in the wake of the reces- allowed states to expand the definition of sion, increasing from 29 states in fiscal year encouraging TANF benefit far beyond the actual cash (FY) 2009 to 39 to 42 states by the end of states to expand payment to include things like childcare and FY2011.16 transportation, or even in some cases allow In recent years, USDA has been actively categorical enrollment from something as minimal as encouraging states to expand categorical eli- eligibility. 5
The federal gibility. As the Congressional Research Ser- ity has more than tripled since 2007, and government has vice reported: these households now account for almost two-thirds of SNAP households. Whereas also embarked The USDA’s Food and Nutrition many of these would qualify under the old on a massive Service has taken an official stance requirements as well, there is the danger encouraging states to use so-called that as the economy eventually improves outreach “categorical eligibility” authority to and some of these households see their in- program expand eligibility to significant num- come and assets increase, they will contin- designed bers of households by (1) increasing or ue to improperly receive benefits for some completely lifting limits on assets that months because they are subject to relaxed to increase eligible households may have and (2) tests. Concern exists that states will con- participation. raising dollar limits on households’ tinue to shift toward categorical eligibility gross monthly income.17 even as the economy improves because it potentially saves them some administrative The effect of this explicit encouragement costs in oversight. Meanwhile, states bear can be seen in the drastic increase in the none of the costs of increased benefit pay- proportion of SNAP households qualifying ments. The spread of broad-based categori- through categorical eligibility in just the past cal eligibility has likely increased program few years. costs and could contribute to keeping them As can be seen, the share of households higher than they otherwise would be in the qualifying through categorical eligibil- coming years. Figure 4 Increase in Proportion of SNAP Households Qualifying Through Broad-Based and “Other” Categorical Eligibility 70 65.8 60 50 40 Percent 30 18.8 20 10 0 2007 2008 2009 2010 2011 Source: U.S. Department of Agriculture, “Characteristics of Supplemental Nutrition Assistance Program Households: Fiscal Year 2011,” Report No. SNAP-12-CHAR, Food and Nutrition Service, November 2012. Note: This figure displays the percentage of all recipients who qualify through categorical eligibility other than the more traditional, narrowly defined mechanism of qualifying through receipt cash assistance through Supplemental Security Income or TANF. This assistance includes Broad-Based Categorical Eligibility, receipt of noncash TANF benefits or services, and transportation subsidies, among others. 6
Figure 5 Method of Qualification for SNAP, 2011 70 60 Percentage of SNAP Participants 50 40 30 20 10 0 Broad-Based Categorical Cash Assistance Income and Asset Tests Eligibility Source: U.S. Department of Agriculture, “Characteristics of Supplemental Nutrition Assistance Program Households: Fiscal Year 2011,” Report No. SNAP-12-CHAR, Food and Nutrition Service, November 2012. In 2011, only a little more than one in even hired so-called food stamp recruiters ten households still qualified through the with monthly quotas of recipients to sign. traditional income and asset tests. Now al- For example, Florida SNAP recruiters have most two-thirds of SNAP households qual- a quota of 150 new participants per month, ify through expanded categorical eligibility, which may help account for that state’s tre- and a further quarter qualify through tra- mendous growth in food stamp receipt.21 ditional categorical eligibility.18 Although Rhode Island, another state that has seen many of these households would still qual- SNAP receipt double over the past five years, ify if they were subject to the traditional hosts SNAP-themed bingo games for the income and asset tests, this qualification is elderly. Alabama hands out fliers that read: not uniformly the case. “Be a patriot. Bring your food stamp money Increased Participation from Outreach. home.” A USDA brochure advises its field Second, at the same time that eligibility for offices to SNAP has been broadened, the federal gov- ernment has also embarked on a massive Throw a Great Party. Host social events Federal and state outreach program designed to increase par- where people mix and mingle. Make it governments ticipation by those who are eligible. As the fun by having activities, games, food CBO found in a 2013 report, between 2002 and entertainment, and provide infor- now spend and 2007, “greater participation mainly re- mation about SNAP. Putting SNAP more than sulted from outreach initiatives, including information in a game format like efforts to increase awareness of SNAP and BINGO, crossword puzzles . . . is fun $41.3 million streamline the application process.”19 and helps get your message across in a annually on Federal and state governments now spend memorable way.22 advertising and more than $41.3 million annually on adver- tising and outreach for food stamps, a six- The USDA also advertised food stamps outreach for food fold increase since 2000.20 Some states have in a 10-part serialized Spanish language ra- stamps. 7
Work dio novella, until public outcry forced them cipients with incomes above the poverty line participation to cancel the program.23 And, in one noto- have risen from 12 percent of SNAP house- rious incident, an official in Ashe County, holds to 16.7 percent, a nearly 40 percent in- by SNAP SC, received an award for developing a strat- crease in just 4 years. recipients egy to counteract what they described as The current SNAP program does have “mountain pride” and appealing to “those work requirements that, at first glance, seem remains low. who wished not to rely on others.”24 comparable to the TANF work requirements This outreach has resulted in a signifi- that were so successful in helping beneficia- cant increase in participation by those eligi- ries transition into the labor market and re- ble for SNAP. In 2010, 75 percent of people duce TANF caseloads. However, loopholes estimated to be eligible for SNAP received and selective enforcement have significantly benefits, a substantial increase from 66 per- limited their effectiveness in helping SNAP cent in 2008.25 Significantly, increased par- beneficiaries find jobs and transition out of ticipation has taken place primarily among the program. higher income households; lower income On paper, the program currently requires households have always had high participa- all nondisabled and nonelderly recipients tion rates. to register for the SNAP Employment and With categorical eligibility making it eas- Training Program, to accept a job if offered, ier for participants to avoid the program’s to search for work, or to meet other work re- strict income limits and higher income eligi- quirements that states impose. Able-bodied bles increasing their participation, the pro- adult recipients between the ages of 18 and portion of SNAP households with incomes 50 without children, must work, must par- above the poverty line has risen significant- ticipate in an employment and training pro- ly. As Figure 6 shows, since 2007, SNAP re- gram, or must participate in a SNAP “work- Figure 6 Percentage of SNAP Households with Gross Income Above 100 Percent Federal Poverty Level 18 16.7 16 14 12 12 10 Percent 8 6 4 2 0 2007 2008 2009 2010 2011 Sources: U.S. Department of Agriculture, “Characteristics of SNAP Households, 2009–2011,” http://www. fns.usda.gov/ora/MENU/Published/snap/FILES/Participation/2011Characteristics.pdf; USDA, “Program Participation Trends 2001–2008,” http://www.fns.usda.gov/ora/menu/Published/snap/FILES/Participation/ Trends2001-2008.pdf. 8
Figure 7 Map of States with ABAWD Waiver for FY2013 Source: U.S. Department of Agriculture, “Supplemental Nutrition Assistance Program (SNAP)—Able Bodied Adults without Dependents Waivers for Fiscal Year 2013,” March 2012. Note: Able-Bodied Adults without Dependents (ABAWD) waivers are in effect for the shaded states. Kansas, Oklahoma, and Wisconsin have signaled that they will let the ABAWD waiver expire in FY2014. fare” program for at least 20 hours per week. ments were required to participate in SNAP Otherwise they can collect SNAP benefits employment and training programs, whereas for only 3 months in a given 36-month pe- only 3 percent of this population participat- riod.26 In addition, recipients cannot quit a ed voluntarily.29 The low voluntary partici- job or voluntarily reduce their hours of work pation could be because SNAP participants below 30 hours per week. (States can exempt view the employment and training programs up to 15 percent of able-bodied recipients as ineffective, or because they have little in- from work requirements for hardship or centive to find employment given weak work other reasons without a waiver).27 requirements. Yet actual work participation by SNAP The inherent weakness of SNAP work recipients remains low. In 2011, the most requirements was briefly made even worse recent year for which data are available, only as a result of the 2009 federal stimulus bill, Recently states 27.7 percent of nonelderly adult participants which suspended work requirements for were employed, while another 28.0 percent ABAWDs for one year. The blanket suspen- have begun to reported that they were looking for work. sion was not extended in 2010. However, use a loophole That means that fully 44 percent were nei- states with high unemployment rates were in the law to ther employed nor actively searching for still given greater flexibility to loosen work work. Looking specifically at working age, requirements. Today, waivers are still in increase the childless, able-bodied adults, more than effect for 44 states and the District of Co- amount of half, or 2 million SNAP households, had no lumbia, covering 92 percent of able-bodied earned income.28 adult recipients without dependents.30 benefits that Similarly, fewer than half of those adults No accurate data are now available on a family can not otherwise exempt from work require- which states are or are not meeting work re- receive. 9
Surprisingly quirements for SNAP. Indeed, states are not ing increased SNAP costs is an increase in little data even required to report such information. the benefits themselves. The 2009 federal However, given how low work participation stimulus boosted the maximum monthly exist about is for other welfare programs such as TANF, benefit in 2009 for a family of three from the program’s where only 42 percent of adults nationwide $463 to $526. As a result, CBO “estimated are engaged in even broadly defined work that 20 percent of the spending growth can effectiveness. activities, there is ample reason for con- be attributed to the benefit increases in the cern. Moreover, states with some of the low- ARRA.”32 As Figure 9 shows, the average est TANF work participation rates, such as benefit per person has almost doubled since Massachusetts, Missouri, and Rhode Island, 2000. are among those with waivers in effect. Recently states have begun to use a loop- As Figure 8 shows, both the number of hole in the law to increase the amount of able-bodied adults without children partici- benefits that a family can receive. Because pating in SNAP and the share of total par- SNAP benefits phase out with higher levels ticipants has grown significantly in recent of income, the lower the countable income years, with a particularly sharp uptick after of an individual, the higher that person’s work requirements were suspended in the benefits will be. If a person or family pays American Recovery and Reinvestment Act for utility costs separately from rent, they of 2009. In fact, able-bodied adults without can deduct both the utility cost and the rent children have been the fastest growing bene- from their countable income, which will al- ficiary group, more than tripling in the past low them to receive higher SNAP benefits. four years.31 The law currently allows states to assign the Increased Benefits. The third factor driv- higher utility allowance to anyone who re- Figure 8 Increase in ABAWDs 2008–11 Percent of SNAP Participants Number of SNAP Participants, Millions 12 10 8 6 4 2 0 2008 2009 2010 2011 Sources: U.S. Department of Agriculture, “Characteristics of Supplemental Nutritional Assistance Program Households, Fiscal Years 2008–2011,” Table A.1. Distribution of Participating Households, Individuals, and Benefits by Household Composition, Locality, Countable Income Source, and SNAP Benefit Amount, http:// www.fns.usda.gov/ora/MENU/Published/snap/FILES/Participation/2011Characteristics.pdf. 10
Figure 9 Average Benefit per Person 2000–11 160 140 120 Benefit per Person, Dollars 100 80 60 40 20 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Source: U.S. Department of Agriculture, “Supplemental Nutrition Assistance Program Participation and Costs,” Food and Nutrition Services, http://www.fns.usda.gov/pd/SNAPsummary.htm. ceives aid from the Low Income Energy As- Questionable Bang for sistance Program (LIHEAP), even if heating the Buck and air conditioning are already included in the individual’s rent or shelter deduction. Considering SNAP’s growing cost and en- In recent years, this deduction that LIHEAP rollment, surprisingly little data exist about helps more beneficiaries qualify for has be- the program’s effectiveness. And many of come the most claimed deduction besides the studies that do exist date from before the the standard one. Almost 72 percent of program’s recent expansion. households claimed it in 2011.33 The USDA, which administers SNAP, The federal government pays for both claims that the program is one of the federal SNAP benefits and LIHEAP, so states have government’s most effective for reducing an incentive to extend LIHEAP to as many poverty. One USDA study found “an aver- beneficiaries as possible. These residents can age decline of 4.4 percent in the prevalence then receive more benefits at no additional of poverty due to SNAP benefits, while the cost to the states. As such, some states send average decline in the depth and severity of a minimal amount of LIHEAP benefits, poverty was 10.3 and 13.2 percent, respective- The literature sometimes as low as $1, to SNAP partici- ly.”34 A second USDA study found that SNAP is inconclusive pants to allow them to automatically qualify reduces food insecurity by 5 to 10 percentage for the related allowance and the resulting points.35 regarding higher benefits. The program’s flawed struc- These results should not be especially sur- whether SNAP ture creates perverse incentives by which the prising. It would be next to impossible for the alleviates states bear no additional costs and actually federal government to spend $75 billion an- have incentive to inflate the benefits to their nually to provide food assistance (counting hunger and residents. just SNAP benefits) to low-income people malnutrition. 11
There is an without having some impact on poverty. The pation and food insecurity is difficult to see. inherent danger real question is whether marginal increases Of course, Figure 10 does not show wheth- in SNAP expenditures have a substantial er the problem of food insecurity might have that SNAP can effect on poverty or nutrition, or whether been even worse in the absence of SNAP. make recipients SNAP provides more benefits than would Here, the limited academic research is am- private charitable alternatives. biguous. For example, according to the Gov- dependent on Despite the massive increase in SNAP ernment Accountability Office (GAO) “the government. spending and participation since 2000, the literature is inconclusive regarding whether federal government currently classifies 17.9 SNAP alleviates hunger and malnutrition for million American households as “food in- low-income households.”38 The GAO found secure,” that is, “uncertain of having, or un- that participants in SNAP and other federal able to acquire, enough food to meet the food programs “tend to be more food inse- needs of all their members because they had cure compared to others that are eligible for insufficient money or other resources for programs but do not participate.”39 However, food” at some point during the year.36 Of as the USDA warns, there is bit of a chicken- these, 39 percent, or 6.8 million households and-egg problem in drawing a causal rela- were considered to have “very low food secu- tionship between SNAP and food insecurity rity,” where “normal eating patterns of one because “households that have recently expe- or more household members were disrupted rienced unusually high levels of food insecu- and food intake was reduced.”37 rity” could be more likely to enroll in SNAP.40 Indeed, as Figure 10 shows, a direct rela- The Urban Institute estimates that food tionship between SNAP spending or partici- stamps have a modest effect on severe food Figure 10 Food Insecurity and SNAP since 2000 80 70 60 50 40 30 20 Food Insecure Individuals, Millions 10 Participation, Millions Expenditure, Billions of Dollars 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Sources: U.S. Department of Agriculture, “Food Security in the U.S., Key Statistics and Graphics,” Economic Research Service, http://www.ers.usda.gov/topics/food-nutrition-assistance/food-security-in-the-us/key-statis tics-graphics.aspx#foodsecure; U.S. Department of Agriculture, “National Level Annual Summary” Food and Nutrition Services, http://www.fns.usda.gov/pd/SNAPsummary.htm. 12
insecurity and a larger one on the broader tive costs expended” and did not provide any measure of food insecurity.41 However, re- insight into outcomes.48 search by Huffman and Jensen was unable At the very least, therefore, the research to find any effect, concluding that “[n]ei- calls into question claims about the pro- ther participation in the FSP [food stamp gram’s effectiveness and suggests that addi- program] nor the size of the FSP benefit re- tional study should precede any increase in duced the probability of the household’s be- spending. ing food insecure.”42 Looked at in terms of nutrition and health, SNAP’s effect is again difficult to The Dependency Trap determine. Evidence suggests that SNAP may indeed make healthy food more avail- As with all welfare programs, there is an able to low-income Americans, but it may inherent danger that SNAP can make recipi- not actually increase the consumption of ents dependent on government. As shown in nutritional food by recipients. For example, Figure 11, a surprising proportion of SNAP a 2008 report by Nancy Cole and Mary Kay recipients remains on the program for long Fox for the USDA concluded that for nearly periods of time, almost 56 percent for lon- all vitamins, minerals, and macronutrients ger than five years.49 This fact suggests that SNAP is an assessed, the dietary intake among SNAP SNAP is not being used as a temporary safe- inefficient, participants was comparable to that of ty net, but rather as a permanent source of fraud-ridden, nonparticipants.43 This finding appears to income. confirm older research that SNAP partici- The limited scholarship available on the and deeply pation does “not lead to greater intake of issue suggests that SNAP does result in some troubled food energy or vitamins and minerals over- level of government dependence. For exam- all.”44 A more recent review of the literature ple, a study by scholars at Virginia Polytech- program. by the GAO similarly concluded that “re- nic and Agricultural University found that search finds little or no effect on the dietary “SNAP participation in the previous period or nutrient intake of individuals.”45 Per- increases the propensity to participate in the haps this conclusion is because SNAP sub- SNAP in the current period by 30 percent.” sidizes unhealthy food as much as healthy As a result, they concluded “[a] major policy foods. For example, SNAP subsidizes “soft implication of the results is that FSP poli- drinks, candy, cookies, snack crackers, and cies aimed at reducing the initial entrance ice cream.”46 into the FSP through changes in benefit A study from the Institute of Medicine levels or certification requirements can have and the National Research Council also long-term benefits in terms of reducing de- raised questions about food stamp effective- pendence among low-income households.” ness, for example, about the program’s slow Conversely, a study for the Journal of Sociol- response to changing prices.47 ogy and Social Welfare agreed that dependency Little evidence exists to support claims did increase in the short term for mothers that the program provides a stimulus for with children on SNAP, that is, over the first the economy as a whole. Indeed, the agricul- five years, but it concluded that the effect ture department’s own Inspector General disappeared over the long term.50 concluded that the department was unable The evidence is much stronger that re- to determine whether the additional dollars ceipt of food stamps reduces work effort. in the stimulus were in any way effective in For example, a study by Hilary Hoynes and meeting the 2009 Recovery Act’s goals be- Diane Schanzenbach in the Journal of Pub- cause three of the four performance mea- lic Economics found “modest reductions in sures “reflected outputs, such as the dollar employment and hours worked after food amount of benefits issued and administra- stamp benefits are introduced.”51 This study 13
Figure 11 SNAP Cumulative Exit Rate 60 50 40 Percent 30 20 10 0 Six Months 1 Year or 2 Years or 3 Years or 4 Years or 5 Years or Longer or Less Less Less Less Less Less Source: U.S. Department of Agriculture, “Dynamics of Supplemental Nutrition Assistance Program Participation in the Mid-2000s,” Food and Nutrition Service, Table II.17, September 2011. found significant differences by gender, with unemployment programs pay more, men showing very little change in labor par- the sacrifices that jobs require do not ticipation, but a nearly “50 percent reduction disappear. The commuting hassle is in total hours worked” for women.52 And a still there, the possibility for injury on study by Udaya Wagle of Western Michigan the job is still there, and jobs still take University confirmed the work disincentive time away from family, schooling, effect of food stamps, concluding “that a hobbies, and sleep. But the reward to one percent increase in FSP receipts reduces working declines, because some of the work hours for families by about one half of money earned on the job is now avail- one percent.” Moreover, Wagle found “these able even when not working.54 are consistent across all families and families with children.”53 These findings really should not be a sur- Fraud, Waste, and Abuse prise. As Casey Mulligan of the University of Chicago testified before Congress, Even setting aside the growing cost and The federal doubtful effectiveness, SNAP is an ineffi- government Both the 2008 Farm Bill and the 2009 cient, fraud-ridden, and deeply troubled pro- currently funds ARRA increased the amount of the gram. SNAP benefits paid to eligible house- For example, SNAP’s administrative costs 21 different holds, and thereby increased job accep- are considerably higher than those of most programs tance penalty and layoff subsidy rates other social welfare programs. In 2013, the . . . . Earning income requires sacri- program’s total administrative expenses at providing food or fices, and people evaluate whether the both the federal and state level are expected food-purchasing net income earned is enough to justify to top $7 billion, more than 9 percent of assistance. the sacrifices. When the food stamp or program costs.55 The federal share of admin- 14
istrative expenses alone is more than $4.5 overseas relatives of recipients.59 Such abus- The current billion, and that is expected to increase to es should be kept in perspective—they are welfare system almost $6 billion by 2023. tiny in comparison to the program’s overall SNAP also has a high rate of fraud and costs—but they still undermine faith in the has done a poor abuse. Officially, the USDA puts program government’s ability to successfully conduct job of reducing fraud at around $858 million last year, its operations. which would amount to just a bit more than poverty. 1 percent of SNAP expenditures.56 But this calculation only refers to direct fraud, such SNAP in the Context of the as trafficking in benefits. It does not include American Welfare State roughly $2.2 billion annually in erroneous payments to individuals who were not prop- One of the biggest mistakes in evaluating erly eligible for participation or who received SNAP is to consider the program in isolation benefits in excess of the amount to which rather than as part of a much larger U.S. wel- they should have been entitled.57 The erro- fare state. For example, the average benefit neous payments raise the total fraud and amounts to just $4.50 per day, not a great abuse rate to more than 3.9 percent, making deal with which to feed a family. This daily SNAP one of the most frequently abused dollar amount has led to much grandstand- non–health care social welfare programs. ing on the part of program supporters, fre- As bad as this fraud rate is, it does repre- quently taking the “food stamp challenge,” sent an improvement over the past few years. in which they demonstrate how difficult it The error payment rate has actually fallen is to eat on $4.50 per day. However, as Wash- from 5.64 percent in FY2007 to 3.80 per- ington Post reporter and “The Fact Checker” cent in FY2011.58 USDA officials and pro- columnist Glenn Kessler pointed out, “the gram advocates credit this improvement to name of the program refers to ‘supplemen- the transition from physical food stamps to tal’ assistance.”60 the Electronic Benefit Transfer (EBT) cards. Fewer than 20 percent of SNAP recipients However, part of the decline is an indirect rely solely on the program for their income. result of the program’s eligibility changes. More than 30 percent of SNAP households Because, as described above, the vast major- have some form of earned income, that is, ity of recipients now qualify through cat- they are working, whereas 60 percent of egorical eligibility, fewer beneficiaries are SNAP households had unearned income, subject to the program’s asset tests. Some such as cash assistance from TANF, general states have virtually eliminated asset tests assistance, Supplemental Security Income or altogether. Many payments that previously another related program, or more than one would have been considered improper are of these. no longer flagged as such because without Indeed, SNAP may not even be the only the asset test there is no way to verify that food assistance program that they receive. the payments are improper. Thus, at least The federal government currently funds 21 part of the apparent decline in erroneous different programs providing food or food- payments is due to moving the foul lines purchasing assistance, administered by 3 rather than to greater program efficiency. different federal departments and 1 inde- And none of this fraud includes highly pendent agency. This funding is just part of publicized abuses in recent years, which, a massive structure of 126 often overlapping while not technically fraudulent, neverthe- federal anti-poverty programs, 72 of which less violate the spirit of the program. This provide either cash or in-kind benefits to abuse includes use of SNAP EBT cards at individuals (the remainder fund communi- liquor stores and even strip clubs, as well as ty-wide programs for low-income neighbor- the shipment of SNAP-funded foodstuffs to hoods), at a cost of more than $668 billion 15
annually.61 State, county, and municipal Still, given the level of welfare spending, governments operate additional welfare pro- we seem to be having little success at getting grams. Of course, no individual or family people out of poverty over the long term, receives benefits from all 72 programs, but or making them less dependent on govern- many recipients do receive aid from a num- ment. This lack of success should suggest ber of them at any given time. that we are doing something wrong. That For an individual receiving major welfare “something” is unlikely to be remedied sim- programs such as TANF; Medicaid; public ply by spending more on food stamps. housing assistance; Women, Infants, and Children; LIHEAP; and free commodities, in addition to SNAP, total benefits would Proposals for Reform range from $49,175 in Hawaii to $16,984 in Mississippi. Nationwide, in a medium-level SNAP is unusual among government welfare state benefits total $28,500.62 SNAP programs in that it combines aspects of both therefore becomes a component of a large an entitlement and a discretionary program and growing welfare system and needs to be subject to appropriation. On the individual evaluated in that context. level, SNAP functions as an entitlement: if a Able-bodied The current welfare system has done a person meets eligibility standards, he or she food stamp poor job of reducing poverty. Since President receives benefits. In addition, authorization recipients should Johnson declared “war on poverty” in 1965, for SNAP and its funding levels remain in the United States has spent more than $15 place unless otherwise changed by legisla- be required trillion on anti-poverty programs.63 Yet the tion. Traditionally, Congress has also set 5- to work or to poverty rate has remained relatively constant or 10-year funding totals as part of periodic since 1965, despite rising welfare spending. farm bills. actively seek work In fact, the only appreciable decline occurred Congress is now debating a farm bill as a condition in the 1990s, a time of state experimentation again this year. As a result, there is a rare for receiving with tightening welfare eligibility, culminat- opportunity to reform the program. Much ing in the passage of national welfare reform of the debate so far has revolved around benefits. (the Personal Responsibility and Work Re- funding levels. Republicans in the House of sponsibility Act of 1996). Since 2006, poverty Representatives have passed legislation that rates have risen despite a massive increase in returns SNAP funding to 2010 levels, a re- spending.64 duction of roughly $39 billion over 10 years Of course, this increase in poverty rates from the projected baseline.67 Whereas such does not mean that anti-poverty spending cuts would be welcome in light of the overall has had no impact. Certainly one could ar- budget picture, simply changing appropria- gue that without such spending poverty tion levels do little to change the program’s levels would be even higher. As noted above, fundamental flaws. the USDA suggests that without SNAP, pov- More comprehensive reforms, some of erty rates could be four to eight percentage which are included in the House legislation, points higher than they are.65 Welfare spend- would also make the following five changes: ing may also reduce the severity of that pov- 1. Change the incentive structure for states. erty. According to the alternative poverty Currently, the federal government provides measure, for instance, taking into account bonus funding to states, which makes it the full range of welfare benefits received easier for people to enroll in food stamps. reduces the number of Americans living in In 2011, states received $48 million in bo- extreme poverty—that is below 50 percent nus payments. Bonus payments encourage of the poverty level—from 6.2 percent to 5.4 states to expand their programs beyond percent.66 These people remain poor, but less the truly needy. Notably, states such as Or- poor than before. egon, which have received bonus payments, 16
are among those with the largest program higher SNAP benefits through their receipt growth in the past five years. of LIHEAP. One reform proposal would seek The problem is not the modest amount to prevent states from providing token LI- that such bonuses cost but the incentives HEAP payments by requiring them to pro- that they establish. The participation rates vide LIHEAP benefits of at least $10 to qual- for the most needy families are much higher ify for the exemption. This proposal would than those near the margin of eligibility, in save about $4.5 billion over 10 years and part because these neediest families likely would affect roughly 500,000 SNAP house- already receive cash assistance from other holds.73 Because states do not bear any of the government programs and are thus categor- cost for LIHEAP, they could simply increase ically eligible. But also these families with LIHEAP benefits to whatever the new thresh- little earned income qualify for higher ben- old is as much as their block grant permits, efits, further creating incentives for them to thereby allowing their residents to continue participate in the program. to receive higher SNAP benefits without in- At the same time, the federal government curring additional state costs. fails to adequately penalize states for errone- A second proposal, therefore, would re- ous payments. Now, a state is only penalized peal the entire LIHEAP–SNAP link so that if erroneous payments total more than 6 per- families who receive the allowance would cent.68 In 2012, only six states exceeded that need to show their utility bills to receive the ceiling.69 If all erroneous payments had had utility portion of the deduction. CBO esti- to be repaid, the federal government would mates that ending automatic qualification have saved more than $2.5 billion in 2010.70 would save roughly $1.5 billion annually. Some states could be hit with repayments of It would affect approximately 1.3 million as much as $250 million.71 More important, SNAP households and lower their benefits requiring repayment may encourage states by roughly $90 per month.74 Families would to become more efficient in determining eli- still receive SNAP benefits, but they would gibility and avoiding mistaken payments. be more in line with the level of benefits that 2. End broad-based categorical eligibility. they should be eligible for. As discussed previously, most SNAP ben- 4. Strengthen work requirements. Most eficiaries qualified for the program through Americans would agree that able-bodied broad-based categorical eligibility and there- food stamp recipients should be required fore were not subject to the same income to work or to actively seek work as a con- and asset tests as other program partici- dition for receiving benefits. The current pants. Whereas the vast majority of categori- SNAP program has work requirements that cal eligibles would likely continue to qualify look strong on paper but fail to achieve their under the traditional asset and income tests, goals in practice. The fact that states do not at least some would not. CBO estimates that even report their rates of work participa- eliminating categorical eligibility and sub- tion should be especially troubling. There- jecting all SNAP applicants to traditional fore any reform of the food stamp program asset and income tests would reduce SNAP should prioritize work as a condition for participation by 1.8 million annually, at a participating in SNAP. savings of $1.2 billion per year.72 Ending For example, there should be clear statu- Any work categorical eligibility would also serve the tory standards for the percentage of non- valuable function of refocusing SNAP ben- exempt adults receiving SNAP who must requirement efits to those individuals most truly in need, engage in a specified number of hours of should allowing spending to be reduced without work-related activity each week. Exemptions strengthen the harming those who truly need assistance. should be narrowly tailored to cover only 3. End LIHEAP loophole. More and more those who are aged, disabled, or parenting definition of beneficiaries in recent years are qualifying for very young children. To enforce this require- work. 17
There is little ment, the program would have to authorize states could spend on the program, but the proof that stronger penalties, including complete dis- states would have vastly increased authority qualification from the program, for non- over administering the program, such as set- the expansion compliance with work requirements. ting their own eligibility criteria and benefit of SNAP has An example can be found in TANF, levels. States would also be free to allocate which requires states to meet a 50 percent funds between traditional food benefits and significantly participation standard. There is no reason other programs such as job training or edu- reduced hunger that a similar requirement should not be cation. For example, since the 1996 welfare or improved established for SNAP. However, TANF also reform, a significantly higher proportion shows the problems that can develop if the of TANF funds have been used on ancillary nutrition. definition of work activities is too broad or benefits such as transportation assistance if states fail to cooperate. In fact, despite and other benefits designed to focus on em- TANF’s stated work requirements, partici- ployment.76 pation varies widely by state, ranging from Congress could also fix the amount of the nearly 88 percent in Idaho to just 17 percent block grant in either nominal or inflation- in Missouri. Nationwide, barely 42 percent adjusted dollars, thereby preventing the pro- of TANF recipients are actually taking part gram from growing in the future. States that in “work activities.” Moreover, work activity wished to expand their program by either frequently does not mean work. Only about increasing participation or benefits would one in five TANF recipients works in an un- be free to do so, but they would have to use subsidized job. Others engaged in work ac- their own funds, thereby making themselves tivities are really taking part in education or answerable to taxpayers and voters. If eco- training programs or are participating in job nomic conditions changed in the future, search activities (i.e., looking for work).75 whether for better or worse, Congress would Therefore, any work requirement should be free to revisit the level of funding. strengthen the definition of work to ensure Block grants, however, are simply a partial that more SNAP participants are actually step toward state control. Notably they split working. In addition, to ensure that states responsibility for funding and administering implement and enforce work requirements, the programs, meaning that it is often dif- those states that fail to meet their work re- ficult to hold any party responsible for the quirements should lose a portion of their program’s failures. In addition, the history SNAP funding. And finally, states should be of block grants suggests that states may in- given greater freedom to shift SNAP funds terpret block grant funds as free money, en- from individual benefits to the administra- couraging waste and inefficiency rather than tion of employment and training programs. experimentation. Moreover, states become 5. Convert SNAP to block grant. Whereas dependent on federal largess in the same each of the previous four reforms would way that individuals do. After all, today the improve the program in some way, none of federal government spends more than $561 them are sufficient; they either unnecessar- billion on grants and other aid to the states, ily limit state flexibility or would not have fully 16 percent of federal spending.77 a large enough effect on program costs. Ul- Nor should we forget that federal funds, timately, as with other social welfare pro- aside from that which is borrowed, come grams, responsibility for the program—and from taxpayers in the 50 states. Simply mov- authority over it—should be turned back to ing money to Washington and then back the states. to the states with various redistributional One approach would be to follow the suc- formulas does not improve the efficiency of cessful model of welfare reform and convert programs. Therefore, over the longer term, SNAP into a block grant. The federal gov- both operational control and the responsi- ernment would appropriate an amount that bility for financing SNAP (along with other 18
social welfare programs) should be returned ment-subsidized purchases of their states’ to the states. That would mean phasing out products. But “bipartisan policy” and “good federal funding in its entirety, leaving the policy” rarely mean the same thing. decision about what types of food and nu- Because Congress will be debating the re- trition programs to pursue directly to the authorization of the farm bill this year, there states and their taxpayers. is an opportunity to reexamine SNAP and begin to reform this expensive, bloated, and inefficient program. Conclusion SNAP has been growing rapidly since at Notes least 2000, and today it is the second largest 1. U.S. Department of Agriculture, Food and means-tested social welfare program in the Nutrition Service, “A Short History of the Food United States. The evidence suggests that Stamp Program,” http://www.fns.usda.gov/snap/ much of this growth is the result of both fed- rules/Legislation/about.htm. eral and state policy decisions, rather than 2. Ibid. economic conditions. Both states and the federal government have loosened eligibility 3. U.S. Department of Agriculture, “Monthly standards and increased outreach efforts in Data—National Level” Food and Nutrition Servic- es, http://www.fns.usda.gov/pd/34SNAPmonthly. an attempt to enroll more participants. In htm; U.S. Department of Agriculture, “National addition, work requirements are far weaker Level Annual Summary” Food and Nutrition than they appear on paper, and many states Services, http://www.fns.usda.gov/pd/SNAPsum have availed themselves of waivers and other mary.htm. maneuvers to weaken them still further. The 4. Federal Agriculture Reform and Risk Man- result is that today roughly one of every sev- agement Act of 2013, H.R. 2642. 113th Cong. en Americans receives food stamps. (2013). Yet there is little proof that the expansion 5. U.S. Department of Agriculture, “Supple- of SNAP has significantly reduced hunger mental Nutrition Assistance Program: Average or improved nutrition among low-income Monthly Participation (Persons),” Food and Nu- Americans. In the absence of much stronger trition Services, http://www.fns.usda.gov/pd/15 research, continued expansion of the pro- SNAPpartpp.htm gram seems to be based more on faith than 6. Ibid. evidence. Although SNAP does not seem to im- 7. Congressional Budget Office, “The Supple- prove nutrition, there is evidence that SNAP, mental Nutrition Assistance Program,” April 2012, p. 1. like other welfare programs, can increase de- pendence and undermine the work ethic. 8. Food Research and Action Center, “Supple- Food stamps are intended to be short-term mental Nutrition Assistance Program: Share of assistance, and too many recipients are re- Population Participating,” http://frac.org/pdf/ 2013_08_07_snap_june2013.pdf. maining on the program for far too long. However, SNAP represents just one portion 9. U.S. Department of Agriculture, “Supple- of a much larger welfare state that includes mental Nutrition Assistance Program: Average 21 federal food and nutrition programs. Monthly Participation (Persons),” Food and Nu- trition Services, http://www.fns.usda.gov/pd/15/ Backers of SNAP argue that food stamps SNAPpartpp.htm. have had a long history of bipartisan sup- port. Indeed they have. Liberal Democrats 10. Congressional Budget Office, “Supplemen- have unsurprisingly backed an expansion tal Nutrition Assistance Program: May 2013 Base- line,” http://www.cbo.gov/publication/44211. of the welfare state, while farm-state Re- publicans have been happy to have govern- 11. Ibid. 19
You can also read