Skip to main content

Program Integrity for the Supplemental Nutrition Assistance Program

Testimony of Stacy Dean, Vice President for Food Assistance Policy,
Center on Budget and Policy Priorities,
Before the Subcommittees on Intergovernmental Affairs and Health Care, Benefits and Administrative Rules of the Committee on Oversight and Government Reform,
U.S. House of Representatives

Thank you for the opportunity to testify today.  I am Stacy Dean, Vice President for Food Assistance Policy at the Center on Budget and Policy Priorities (CBPP), an independent, non-profit, nonpartisan policy institute located here in Washington.  CBPP conducts research and analysis on a range of federal and state policy issues affecting low- and moderate-income families.  The Center’s food assistance work focuses on improving the effectiveness of the major federal nutrition programs, including the Supplemental Nutrition Assistance Program (SNAP).  I have worked on SNAP policy and operations for more than 20 years.  Much of my work is providing technical assistance to state officials who wish to explore options and policies to improve their program operations in order to more efficiently serve eligible households.  My team and I also conduct research and analysis on SNAP at the national and state levels.  CBPP receives no government funding.

My testimony today is divided into three sections: 1) SNAP’s role in our country as a federal nutrition program; 2) an overview of how SNAP addresses and maintains program integrity; and 3) an assessment of how the House Agriculture Committee farm bill, H.R. 2, would compromise SNAP’s effectiveness.

SNAP Plays a Critical Role in Our Country

To provide context for today’s hearing topic of SNAP’s program integrity, I think it is important to review some of SNAP’s most critical features.  The program is a highly effective anti-hunger program.   Much of its success reflects its entitlement structure, its consistent national benefit structure, and its food-based benefits.  It also imposes rigorous requirements on states and clients to ensure a high degree of program integrity.

SNAP helps more than 40 million low-income Americans afford a nutritionally adequate diet by providing them with benefits via a debit card that can be used only to purchase food.  Benefits average about $1.40 per person per meal.  One in eight Americans participate in SNAP — a figure that speaks both to the extensive need across our country and to SNAP’s important role in addressing it.

Policymakers created SNAP, then known as the Food Stamp Program, to help low-income families and individuals purchase an adequate diet.Policymakers created SNAP, then known as the Food Stamp Program, to help low-income families and individuals purchase an adequate diet.  Policymakers created it at a time when hunger and malnutrition were much more serious problems in this country than they are today.  A team of Field Foundation-sponsored doctors who examined hunger and malnutrition among poor children in the South, Appalachia, and other very poor areas in 1967 (before the Food Stamp Program was widespread in these areas) and again in the late 1970s (after the program had been instituted nationwide) found marked reductions over this ten-year period in serious nutrition-related problems among children.  The doctors gave primary credit for this reduction to the Food Stamp Program.  Findings such as this led then-Senator Robert Dole to describe the Food Stamp Program as the most important advance in the nation’s social programs since the creation of Social Security.

Consistent with its original purpose, SNAP continues to provide a basic nutrition benefit to low-income families, the elderly, and people with disabilities who cannot afford an adequate diet.  In some ways, particularly in its administration, today’s program is stronger than at any previous point.  By taking advantage of modern technology and business practices, SNAP has become substantially more efficient, accurate, and effective.  While many low-income Americans continue to struggle, this would be a very different country without SNAP.

Protecting Families From Hardship and Hunger

SNAP benefits are an entitlement, which means that anyone who qualifies under program rules can receive benefits.  This is the program’s most powerful feature:  it enables SNAP to respond quickly and effectively to support low-income families and communities during times of economic downturn and increased need.  For example, SNAP — aided by a temporary benefit increase from the 2009 Recovery Act — kept poverty and food insecurity (lack of consistent access to sufficient food) from rising during the Great Recession as much as they would have without the program.[1]

SNAP enrollment expands when the economy weakens and contracts when the economy recovers.  (See Figure 1.)  As a result, SNAP responds immediately to help families and to bridge temporary periods of unemployment.  It also can help individual families weather a short-term crisis, such as job loss or divorce.  A U.S. Department of Agriculture (USDA) study of SNAP participation over the late 2000s found that slightly more than half of all new entrants to SNAP participated for a year or less and then left the program when their immediate need passed.[2]

SNAP’s powerful response during the recession contrasts sharply with that of Temporary Assistance for Needy Families (TANF), whose block grant structure severely limits its ability to expand during economic downturns.  While the number of unemployed doubled in the Great Recession, TANF caseloads rose only modestly, by 13 percent from December 2007 to December 2009.

SNAP also acts as a first responder in the wake of natural disasters, providing critical food assistance to vulnerable households.  After disasters, USDA and states work together to provide quick, targeted assistance.  This can include replacing participants’ benefits to compensate for lost food, providing temporary Disaster SNAP (D-SNAP) benefits to non-participants who have suffered significant loss, and relaxing program requirements to ease access and relieve undue burden on staff.  In 2017, SNAP helped households affected by Hurricanes Harvey, Maria, and Irma in Texas, Louisiana, Florida, and the U.S. Virgin Islands put food on the table.

SNAP also acts as an automatic stabilizer:  caseloads rose as more households were eligible for and qualified for SNAP during the Great Recession and in subsequent years because of the slow recovery, and then fell as incomes rose.  During the recession, as the official poverty rate rose from 12.5 percent to 15.1 percent, SNAP enrollment rose to respond to this increase. As the effects of the economic recovery have been felt more broadly, poverty has since fallen to 12.7 percent in 2016 (the most recent year for which data are available) and SNAP caseloads have declined significantly.  Caseloads peaked in December 2012 and have since fallen faster every year, declining about 2 percent annually in 2014 and 2015, 3 percent in 2016, and almost 5 percent in 2017. Between their December 2012 peak and January 2018 (the most recent month for which data are available), caseloads fell by 7 million people.

Contributing to the caseload decline, at least 500,000 people — including some of the nation’s poorest — lost SNAP in 2016 due to the return in many areas of a three-month time limit on benefits for unemployed adults aged 18-49 who aren’t disabled or raising minor children.

As a result of the caseload decline, SNAP spending fell for the fourth straight year in 2017 as a share of gross domestic product (GDP).  SNAP spending fell by 7 percent in 2016 and another 8 percent in 2017, and CBO projects that it will return to its 1995 level as a share of GDP in the next few years.

Lessening Poverty and Unemployment

SNAP targets benefits on those least able to afford an adequate diet.  Its benefit formula considers a household’s income level as well as its essential expenses, such as rent, medicine, and child care.  A family’s total income is the most important factor affecting its ability to purchase food, but not the only one.  For example, a family spending two-thirds of its income on rent and utilities will have less money to buy food than a family that has the same income but lives in public or subsidized housing. While targeting benefits adds some complexity to the program and is an area where states sometimes seek to simplify, it helps SNAP provide the most assistance to the families with the greatest needs.

This makes SNAP a powerful tool in fighting poverty.  SNAP kept 8.4 million people out of poverty in 2015, including 3.8 million children, according to a CBPP analysis using the government’s Supplemental Poverty Measure (which counts SNAP as income) and correcting for underreporting of public benefits in survey data.  (See Figure 2.)  SNAP also lifted 2 million children above half of the poverty line in 2015, more than any other government assistance program.

These results reflect SNAP’s targeting of very low-income households.  Roughly 93 percent of SNAP benefits go to households with incomes below the poverty line, and 56 percent go to households below half of poverty.

SNAP also reduces food insecurity.  A study comparing SNAP participant households before and after six months of participation found that SNAP reduced food insecurity by up to ten percentage points and reduced “very low food security,” which occurs when one or more household members have to skip meals or otherwise eat less because they lack money, by about six percentage points.[3]

SNAP also protects the economy as a whole by helping to maintain overall demand for food during slow economic periods.  In fact, SNAP benefits are one of the fastest, most effective forms of economic stimulus because they get money into the economy quickly.  Moody’s Analytics estimates that in a weak economy, every $1 increase in SNAP benefits generates about $1.70 in economic activity.

Improving Long-Term Health and Self-Sufficiency

SNAP also brings important long-run benefits.  The available evidence suggests that SNAP is at least associated with, and may promote, better health and lower health care costs.  It is therefore plausible that reducing SNAP eligibility and benefits would harm health and raise health care costs. Substantial research on SNAP and related areas has shown:

  • SNAP is associated with improved current and long-term health.  SNAP participants are more likely to report excellent or very good health than low-income non-participants.[4]  Early access to SNAP among pregnant mothers and in early childhood improved birth outcomes and long-term health as adults.[5] Elderly SNAP participants are less likely than similar non-participants to forgo their full prescribed dosage of medicine due to cost.[6]  SNAP may also help low-income seniors live independently in their communities and avoid hospitalization.[7]

    A study examining what happened when the government introduced food stamps in the 1960s and early 1970s concluded that people who had access to food stamps in early childhood and whose mothers had access during their pregnancy had better health outcomes as adults than children born at the same time in counties that had not yet implemented the program.[8]  (See Figure 3.)

  • SNAP is linked with reduced health care costs.  Low-income adults participating in SNAP incur about $1,400, or nearly 25 percent, less in medical care costs in a year than low-income non-participants, according to a study that controlled for factors expected to affect spending on medical care.[9]  The difference is even greater for those with hypertension (nearly $2,700 less) and coronary heart disease (over $4,100 less).[10]

Given these impressive findings, we encourage Congress to consider ways to strengthen SNAP by improving its basic benefit and by extending the reach of the program to ensure that reaches a greater share of eligibles.

Supporting and Encouraging Work

In addition to acting as a safety net for people who are elderly, disabled, or temporarily unemployed, SNAP is designed to supplement the wages of low-income workers.

The SNAP benefit formula contains an important work incentive.  For every additional dollar a recipient earns, her benefits decline by only 24 to 36 cents — much less than in most other programs.  (See Figure 4.)  Families thus have a strong incentive to work longer hours or to search for better-paying employment.  States further support work through the SNAP Employment and Training program, which funds training and work activities for unemployed adults who receive SNAP.

Because SNAP is an important work support and states have made progress in reaching workers, the number of households that have earnings while participating in SNAP has more than tripled — from about 2 million in 2000 to 6.9 million in 2016.  The share of SNAP families that are working while participating in SNAP has also risen: while only about 28 percent of SNAP households with children and an able-bodied adult had earnings in 1990, 60 percent did so in 2016, according to Agriculture Department data.  (See Figure 5.)

SNAP helps workers both when they are between jobs and to supplement low earnings while they work. Millions of workers are in jobs that provide low pay, can have shifting schedules, and often lack key benefits such as paid sick leave.  These features can contribute to income volatility and job turnover: low-wage workers, including many who participate in SNAP, are more likely than other workers to experience periods when they are out of work or when their monthly earnings drop, at least temporarily.  These dynamics lead many adults to participate in SNAP temporarily. Others, such as workers with steady, but low-paying, jobs, or those unable to work, participate on a longer-term basis.  SNAP’s dual function as both a short-term support to help families afford food during a temporary period of low income and a support for others with longer-term needs is one of its principal strengths.

Because workers are more likely to turn to SNAP when they are out of work, looking at work status among SNAP participants at a given point in time substantially overstates their joblessness.  Many participants receive SNAP for short periods and work both before and after (or work in some months while on SNAP but are un- or underemployed in other months).  Participants are also more likely to receive SNAP when they are out of work and need more help affording food.

Most adults in SNAP are either working while they participate or are temporarily out of work. (See Figure 6.)  For example, in 2016, about 15.2 million non-elderly adults not receiving disability benefits participated in SNAP in an average month, according to USDA data.  Of those, about 5.7 million had earnings in that month while they were participating in SNAP, leaving about 9.5 million adults without earnings.  Detractors of the SNAP often point to this figure as an indication that the program is not working, but this assertion is misleading.  When one looks closer at the details of this population, the numbers do not bear out the characterization that SNAP recipients are work avoidant.  Close to half of adults who were not working in a month when they were participating in SNAP worked within a year, CBPP analysis of Census data finds; this means that about 4.5 million of those 9.5 million adults are workers participating in SNAP during a period of unemployment.[11]  Of the remaining 5 million or so participants without earnings, over one-third are unable to work due to caregiving responsibilities (most of whom have working spouses); about one-quarter have a work-limiting disability or chronic health condition, despite not receiving disability benefits; close to one-third report either going to school or being unable to find work; and the rest report other reasons, such as temporary inability to work.[12]

When participants work, they typically work at least half time (at least 20 hours per week), and usually full time (at least 35 hours per week).  Of those non-disabled adults who participated in a given month in 2012 and worked in the year after that month, about half (52 percent) worked at least six months full time, and another 19 percent worked at least one month full time.[13]  Only 15 percent worked 20 hours per week for less than six months or worked fewer hours than that for any length of time.

Most non-disabled adults participate in SNAP for shorter periods, but even those who participate for longer periods are workers who use SNAP to supplement earnings and fill gaps when between jobs.  Close to two-thirds of the adults who ever participated in SNAP in a three-and-a-half-year period participated in the program for less than two years total.[14]  Regardless of how long these adults participated in SNAP, however, they worked in the majority of months in which they received SNAP assistance.

II. SNAP Prioritizes Program Integrity

SNAP cannot fulfill its primary purpose of helping struggling households afford a basic diet without maintaining strong program integrity.  USDA and states take their roles as stewards of public funds seriously and emphasize program integrity throughout program operations.  Moreover, the authorizing committees have mandated in SNAP some of the most rigorous program integrity standards and systems of any federal program.  They provide oversight of the program’s accuracy and fraud detection and prevention systems.  These strong systems ensure a high degree of integrity and accuracy in the program.

When a household applies for SNAP it must report its income and other relevant information; a state eligibility worker interviews a household member and verifies the accuracy of the information using third-party data matches, paper documentation from the household, and/or by contacting a knowledgeable party, such as an employer or landlord.  Households must reapply for benefits periodically, usually every six or 12 months, and between reapplications must report income changes that would affect their eligibility.

Numerous measures ensure the accurate assessment of household eligibility during the eligibility process, through ongoing checks and reassessment of eligibility.  The same is true with respect to the proper use of benefits, an area of fraud prevention and detection where USDA also plays a significant role.  These measures are designed to detect and prevent the occurrence of honest mistakes, careless errors, systemic mistakes, and the less frequent problem of intentional fraud.

They include extensive requirements that households applying for or seeking to continue receiving SNAP prove their eligibility, sophisticated computer matches to detect unreported earnings, the most rigorous quality control (QC) system of any public benefit program, and administrative and criminal enforcement mechanisms.

CRS’s recent report, “Errors and Fraud in the Supplemental Nutrition Assistance Program,” does an excellent and comprehensive job summarizing the issues.  It highlights a few key observations that bear repeating:

Errors are not the same as fraud. Fraud is intentional activity that breaks federal and/or state laws, but there are also ways that program stakeholders—particularly recipients and states—may inadvertently err, which could affect benefit amounts. Certain acts, such as trafficking, are always considered fraud, but other acts, such as duplicate enrollment, may be the result of either error or fraud depending on the circumstances of the case.

SNAP fraud is rare, according to all available data and reports. While this report discusses illegal or inaccurate activities in SNAP, they represent a relatively small fraction of SNAP activity overall.

There is no single data point that reflects all the forms of fraud in SNAP. The most frequently cited measure of fraud is a national estimate of retailer trafficking, which is a significant, but not the only, type of fraud in the program.

While retailer trafficking and retailer application fraud are pursued primarily by a single federal entity, recipient violations are pursued by 53 different state agencies. This leads to disparate approaches and disparate reporting.

The national payment error rate (NPER) is the most-often cited measure of nationwide SNAP payment accuracy, but it has limitations.[15]

Strong Eligibility and Payment Accuracy Backed Up by Quality Control System

SNAP has long had one of the most rigorous payment error measurement systems of any public benefit program.  When, under the leadership of this Committee, Congress enacted the Improper Payments Act in the early 2000s, SNAP was among the few programs to already meet the Act’s high standards.  Each year states take a representative sample of SNAP cases (totaling about 50,000 cases nationally) and thoroughly review the accuracy of their eligibility and benefit decisions.  Federal officials re-review a subsample of the cases to ensure accuracy in the error rates.  States are subject to fiscal penalties if their error rates persistently exceed the national average.

The percentage of SNAP benefit dollars issued to ineligible households or to eligible households in excessive amounts fell for seven consecutive years and stayed low in 2014 at 2.96 percent, USDA data show.[16]  The underpayment error rate also stayed low at 0.69 percent.  The combined payment error rate — that is, the sum of the overpayment and underpayment error rates — was 3.66 percent, low by historical standards. The amount of benefits that were not paid as a result of improper denials, i.e. benefits that should have been issued to eligible households who were denied eligibility, is not captured in this figure.

If one subtracts underpayments (which reduce federal costs) from overpayments, the net loss to the government that year from errors was 2.27 percent of benefits. 

The overwhelming majority of SNAP errors that do occur result from mistakes by recipients, eligibility workers, data entry clerks, or computer programmers, not dishonesty or fraud by recipients.  In addition, states have reported that almost 60 percent of the dollar value of overpayments and almost 90 percent of the dollar value of underpayments were their fault, rather than recipients’.  Much of the rest of overpayments resulted from innocent errors by households facing a program with complex rules.

It should be noted that an overpayment is counted in a state’s error rate whether or not the overpaid benefits are collected back from households.  In fiscal year 2016, states collected about $402 million in overissued benefits.[17]

A USDA Office of Inspector General (OIG) report in 2015 drew attention to data quality issues with SNAP QC error rates in many states.  The report suggested, and USDA later confirmed, that many states’ review processes risked introducing bias into their measurement process.  This raised concerns that states were not measuring errors consistently across states and that some were taking extra steps to find a case correct.  As a result, USDA did not report national or state-level error rates for all states for 2015 or 2016.  Since 2015, USDA has conducted detailed reviews in all states and taken action to address the quality and consistency of the measure.  Administrator Lipps reported to the Senate Agriculture Committee in September 2017 that USDA expects reliable SNAP error rates for 2017 will be released in June 2018[18].  It is worth noting that in the states that USDA identified in its detailed reviews as not having any problems, error rates were still below 6 percent in all but one state and below 5 percent in seven states, including two large states, New York and Arizona.[19]

The problem of inconsistency and bias in the QC sample is quite serious and we appreciate all that the OIG, USDA, and states are doing to address it.  It’s worth noting, however, that this issue did not arise from lack of attention to program integrity.  SNAP’s QC system (including the level of resources required to staff it), the penalty structure associated with high error rates, and the public attention on states with high payment error rates all place enormous pressure on states to address payment errors.

I cannot overstate the importance that SNAP’s operational culture places on achieving and maintaining low payment error rates.  USDA and the states monitor SNAP error rates throughout the year and share best practices.   A significant number of federal and state personnel are assigned to program integrity.[20]  The error rate is the major performance measure for accountability at state and local SNAP offices and even for individual SNAP state eligibility workers and policy officials.

Fear of high error rates has sometimes driven states to adopt policies that reduced program access — particularly in the late 1990s, when the share of SNAP households with earnings began rising due to the strong economy and new welfare policies.  Low-income earners often experience sharp fluctuations in their monthly income, making household income difficult to predict accurately for SNAP benefit calculations.  Some states instituted administrative practices designed to reduce errors that had the unintended effect of making it harder for many working-poor parents to participate, largely by requiring them to take too much time off from work for repeated visits to SNAP offices at frequent intervals, such as every 90 days, to reapply for benefits.

This prompted many analysts and state policy officials from across the political spectrum to call for policy and quality control changes that would improve working families’ access to SNAP, and led both the Clinton and the Bush administrations to address this problem.  There was bipartisan consensus that requiring a family to be on welfare to receive food stamps would reduce work incentives and was contrary to welfare reform goals.  Congress enacted meaningful, although relatively modest, changes in 2002 and 2008 to lessen barriers to SNAP participation among the working poor without compromising program integrity.  As I discuss later in my testimony, I am concerned that some of the House Agriculture Committee’s proposed farm bill changes would reverse or undermine the key reforms adopted during this period, threatening the gains SNAP has made with respect to program access.  The group most impacted by the House bill would be workers with low wages.  They would face a dramatic increase in paperwork and verification — essentially a monthly redetermination of eligibility, as well as experience some of the biggest benefit cuts.

SNAP Provides for a Strong Anti-Fraud System

Fraud, while relatively rare, is taken seriously in the program.  Within the SNAP context, fraud is defined to mean occurrences where:

  • SNAP benefits are exchanged for cash.  This is called trafficking and it is against the law.  Trafficking involves two parties — typically a household and a SNAP retailer.
  • A household intentionally lies to the state to qualify for benefits or to get more benefits than it is supposed to receive.
  • A retailer previously disqualified from SNAP due to abuse lies on its application to rejoin the program.

States and USDA each play a role in pursuing these different kinds of fraud, dedicating significant resources and staff to pursing allegations of fraud and rooting it out when found.  My testimony will briefly cover two of these issues:  household fraud and trafficking.

Household Fraud

SNAP’s rigorous application process and eligibility review serves as the first line of fraud prevention.  The application process requires an interview with a caseworker and demands that, in addition to mandated verification (and often third-party data checks), any questionable information provided by the applicant be verified.  For example, if an individual claims that her rent is $1,000 a month but that she has no income, this raises the question of how she affords her rent.  If a caseworker were to accept such a statement without probing, the case could be in error, because the client was either confused about what counts as income (i.e., not counting support from a family member) or did not tell the whole truth.  The caseworker should follow up on this information at the interview and require additional verification from the individual before she is approved.  States can set their own filters on what provokes further follow-up based on individual circumstances.

Caseworkers feel appropriate pressure to ensure that benefits are issued accurately.  There is the formal QC review process, and many states conduct quality and accuracy reviews of staff work at the line manager level.  Managers review a certain number of cases from each worker each month, generally focusing on less experienced workers.

State agencies run database checks to match the information provided by applicants.  For example, if an application lists Social Security as a source of income, the caseworker would check with the Social Security Administration to verify the amount of the monthly payment.  In many instances a caseworker can reconcile information discrepancies on the application while talking with the applicant.  An area for program improvement would be for Congress to consider providing all states with this capacity.

Households determined eligible must remain eligible to continue to participate.  Households must report changes that would make them income-ineligible.  And many states run third-party matches throughout a household’s eligibility cycle to continue to check that external information confirms the household’s circumstances.  For example, Congress has mandated that states check with prison records and state vital statistics to ensure that no member of a SNAP household continues to receive benefits during incarceration or after death.

When a caseworker suspects that a client is seeking to deceive the program, the case is referred to the state’s fraud unit for investigation.  Members of the public and other state agencies may similarly report any suspected fraud.  Most states prominently display fraud hotlines on their main webpages or take other steps to make it easy for the public to report fraud. [21]

Many investigations do not result in a fraud finding.  Of the roughly 964,000 fraud investigations in fiscal year 2016, 48 percent of the cases were determined not to be fraud.[22]  If investigators determine that fraud has occurred, there is typically a hearing to review the facts and enable clients to respond to the allegations.  This helps prevent innocent participants who made unknowing mistakes from being disqualified.

Persons found guilty of fraud lose SNAP eligibility, and the state pursues the improperly issued benefits for repayment via SNAP’s claims process.  States are eligible to retain a share of mis-issued benefits that they collect as an incentive for them to pursue the claims.

Fraud Violation Penalty
First fraud/intentional program violation 12-month disqualification period
Second fraud/intentional program violation 24-month disqualification period
Third fraud/intentional program violation Permanent disqualification from SNAP
False statement with respect to identity or place or residence in order to receive multiple SNAP benefits simultaneously 10-year disqualification period

In fiscal year 2016, 56,000 individuals were disqualified from SNAP for fraud, up from 47,000 the previous year.[23]


Another area of program integrity in which SNAP has a strong systems and has made considerable improvements is trafficking. Trafficking is the sale or exchange of SNAP benefits for something other than food and typically involves both a retailer and a recipient. USDA monitors SNAP transactions for patterns that may suggest abuse; federal and state law enforcement agencies are then alerted and investigate.  Retailers or SNAP recipients who defraud SNAP by trading their benefit cards for money or misrepresenting their circumstances face criminal penalties. Based on the most recent data, USDA has cut trafficking by three-quarters over the past 15 years.

Over the years, USDA has sanctioned thousands of retail stores for not following federal requirements.  In fiscal year 2017, USDA permanently disqualified approximately 1,700 SNAP retailers for program violations and imposed sanctions, through fines or temporary disqualifications, on another 900 stores.[24]  Another possible area for exploration on trafficking prevention is for FNS to consider its store authorization and monitoring process.

According to the September 2017 USDA-FNS Retailer Trafficking Study[25] and summarized in CRS’s recent report on SNAP Error and Fraud,[26] the national retailer trafficking rate for 2012-2014 was 1.5 percent, up from 1.34 percent in the 2009-2011 study.  This means that, during this period, USDA-FNS estimates that 1.5 percent of all SNAP benefits redeemed were trafficked at participating stores.  Additionally, this study estimated that 11.82 percent of SNAP-authorized retailers engaged in retailer trafficking at least once during this period.

The study found that the increase in retailer trafficking was caused by increased participation in SNAP by smaller stores, which have a much higher rate of retailer trafficking than other stores.  The number of stores newly authorized to participate in the program over the last ten years (FY2007-FY2016) was about 93,000, and about 63 percent of this growth came from convenience stores.  As of FY2016, convenience stores constitute about 46 percent of all stores in the program, up from 36 percent in FY2007.  Convenience stores account for about 5 percent of total SNAP redemptions but about 57 percent of retailer trafficking over the 2012 to 2014 period, according to the study.  About 20 percent of all SNAP benefits used at authorized convenience stores are trafficked and about 20 percent of all authorized convenience stores are engaged in trafficking.

These rates are significantly higher than the national rates for all stores.  The increase in SNAP participation by smaller stores appears to correlate to an overall increase in retailer trafficking, according to USDA-FNS.  USDA’s finding that increased participation by smaller stores appears to correlate to an increase in retailer trafficking is an important area for further exploration for program improvement.  Perhaps USDA’s store screening and monitoring process ought to be different for the types of stores with a higher incidence of trafficking than others.

SNAP Administration Is Efficient

SNAP has low administrative overhead.  About 93 percent of federal SNAP spending goes to providing benefits to households for purchasing food.  About 6.5 percent goes for state administrative costs, including eligibility determinations, employment and training and nutrition education for SNAP households, and anti-fraud activities.  Less than 1 percent goes for federal administrative costs.  In addition to SNAP, the SNAP budget funds other food assistance programs, including a block grant for food assistance in Puerto Rico and American Samoa, commodity purchases for the Emergency Food Assistance Program (which helps food pantries and soup kitchens across the country), and commodities for the Food Distribution Program on Indian Reservations.

PART III:  The House Agriculture Committee Farm Bill, H.R. 2

The House is likely to vote on significant legislation that would impact SNAP in the coming weeks as a part of the 2018 farm bill.  The farm bill is the legislation that, among other things, reauthorizes SNAP, which provides an opportunity to strengthen the program by building upon its successes and addressing important oversight issues.  The House Agriculture Committee passed on April 18 its version of the 2018 farm bill, H.R. 2.  Traditionally, the legislation is bipartisan and represents meaningful compromise and collaboration within the Committee to improve SNAP.  Unfortunately, the House Agriculture Committee bill did not follow such a path.  Its bill, if enacted, would increase food insecurity and hardship and end or cut benefits for a substantial number of low-income people.

My organization has produced numerous analyses of the bill’s impact,[27] including its program integrity provisions.  Here I will provide a brief overview of our assessment and then discuss two issues in greater depth.


Overall, the Agriculture Committee’s proposal would reduce SNAP’s effectiveness and put large numbers of families and individuals at increased risk of hardship.

  • The bill contains changes that would cause more than 1 million low-income households with more than 2 million people — particularly low-income working families with children — to lose their benefits altogether or have them reduced.  The Committee would use these benefit cuts, in part, to pay for a few modest benefit enhancements. But the net effect of all these provisions would still be a significant benefit cut, and a substantial number of people would lose benefits altogether.  The remaining savings from the eligibility and benefit cuts would go to expanding state and federal bureaucracies and financing various grant programs outside of SNAP, at the expense of low-income families and individuals whose basic food assistance would end or shrink.
  • In particular, the plan includes sweeping, aggressive new work requirements that would likely prove unworkable and do substantially more harm than good, fueling increases in hunger and poverty.  These provisions would force states to develop large new bureaucracies, but research suggests they would do little to increase employment.  This expensive, risky approach runs counter to evidence-based policymaking, particularly since the results from work pilots for SNAP recipients that the 2014 farm bill established, which are well underway, aren’t yet available.  Moreover, experience suggests that the bill’s work requirements would leave substantial numbers of low-income people who have various barriers to employment — such as very limited skills or mental health issues like depression — with neither earnings nor food assistance.
  • The plan would also impose significant new state mandates and roll back numerous areas of state flexibility that were designed on a bipartisan basis in prior farm bills to streamline and modernize program operations and make the program easier for states to administer and for eligible households, particularly working families, to navigate.

Many of Agriculture Committee Chairman Michael Conaway’s stated themes for the nutrition package — funding job training, modernizing benefit delivery, strengthening program integrity, and supporting healthy eating — are priorities that Committee members on both sides of the aisle broadly share and could have been the basis for a meaningful bipartisan bill. In some of these areas, the bill contains modest, useful proposals.  But the Committee broke with longstanding bipartisan tradition when it passed a package with support from Republicans alone that’s unbalanced, untested, and likely unworkable in key areas like its sweeping work requirements — and that would put the food security and well-being of many low-income families, including working families, at risk.

Nutrition Title of House Agriculture Committee Farm Bill Includes More Than $20 Billion in SNAP Benefit Cuts Over Ten Years
  CBO 10-Year Cost Estimate
SNAP benefit cuts -$23.1 billion  
SNAP benefit improvements $5.8 billion  
New administrative costs and work programs $15.0 billion  
SNAP benefit delivery and other program changes -$0.6 billion  
Non-SNAP grants $2.8 billion  
Total Nutrition Title* $0 billion  
Details do not add to total due to rounding. * Title IV total including $463 million in direct spending and $465 million in increased revenue. Sources: Congressional Budget Office (CBO), H.R. 2, Agriculture and Nutrition Act of 2018, as ordered reported by the House Agriculture Committee on April 18, 2018, May 2, 2018,

Program Integrity:  A Flawed Approach to the National Accuracy Clearinghouse

Improving program integrity and program operations has long been an important goal of the farm bill.  States and USDA often provide the Agriculture Committees with ideas for, or examples of, important innovations.  Moving food stamps to paper coupons to an electronic benefit transfer system, shifting applications online, supporting electronic document management systems, and funding innovative approaches to fraud detection are examples of ideas that have come from this collaboration.  Traditionally, these efforts have been non-partisan in nature.

An example of a program integrity provision that could have garnered bipartisan support concerns the National Accuracy Clearinghouse (NAC).  Section 4001 of the Committee’s bill would establish a duplicative enrollment database by transforming the NAC into a nationwide, mandatory effort.  Now implemented under a state-initiated contract between Mississippi and LexisNexis, the NAC lets five participating southern states (Alabama, Florida, Georgia, Louisiana, and Mississippi) disclose data on their SNAP participants as a means to identify and prevent SNAP participation in more than one state. An evaluation of the NAC found that fewer than 0.2 percent of SNAP participants were dual participants (i.e., more than 99.8 percent were not).[28]  H.R. 2 proposes expanding the NAC nationwide within two years, which CBO estimates would save $588 million over ten years.

Expanding the NAC could strengthen SNAP’s program integrity and is a sound concept, but the Committee’s proposal is flawed.  States have indicated that the two-year timeframe to implement a system across all 50 states is ambitious given the limitations of their computer systems and the need to build business processes that allow them to work with other states to determine if an applicant is truly attempting to draw benefits twice or if the home state did not take necessary action to disenroll the individual.  A more serious concern relates to the privacy of current and past SNAP participants. The bill would require USDA to build an enormous database of all SNAP participants that would include income and other personal information (including Social Security numbers and income and asset information) that’s well beyond what’s needed to limit duplicate participation.  The data, which states would share with USDA each month, would be stored in the database for many years, if not indefinitely.  When all states participate, the NAC would include personal information on every SNAP recipient ― some 40 million people at any point in time and potentially tens of millions more over time as others participate in SNAP and then leave the program.  This raises concerns about data security, privacy, and other potential uses of the information.  There are also important questions about under what circumstances USDA would be permitted to transfer or disclose the data to other government agencies and for what purposes.

The proposal does not specify USDA’s costs to implement security and privacy protections that match those of the Social Security Administration (SSA) and Internal Revenue Service for the similar information they store about millions of Americans.  Nor does it appear to provide any dedicated new funding for such measures, or to authorize USDA to slow or pause implementation under the aggressive timeframe required, if the Agriculture Secretary believed that USDA could not protect the personal data of Americans.  While I have tremendous respect for the agency staff at FNS and believe they would do their best to secure and protect this information, this would be a significant undertaking for the agency.  Moreover, my understanding is that a federal database such as this would be subject to Privacy Act obligations, including potential liability for breaches that may result.  As a result, significant security would be essential.  Providing the necessary level of security and privacy requires deep expertise, detailed planning, significant investments, and dedicated staffing.  It’s worth noting that SSA’s cybersecurity budget, which includes both staff and information technology costs, is over $100 million per year.[29]  While the SNAP participant database would be smaller than SSA’s, the provision would result in FNS, like SSA, holding and securing private information for tens of millions of individuals in perpetuity.

Moreover, this approach was unnecessary and does not reflect current best practices for large data matches.  To avoid becoming a target for hackers and to protect private information, data security experts recommend that a match like the one required to prevent dual participation under the NAC “ping” — that is, send an individual query for whether an individual is already enrolled in SNAP — to the various individual state datasets, rather than have all the data aggregated in one enormous database.  The Federal Data Services Hub, which the Department of Health and Human Services oversees under the Affordable Care Act, facilitates matches between states and federal agencies and private vendors rather than gathering all the data together.  This provision is one of several examples of how the bill includes sweeping, untested and unnecessary proposals that would create risks for low-income Americans if the government does not execute them well.  The NAC could have been designed in such a way that it garnered widespread support and did not raise serious concerns.

Balancing Program Integrity with Program Access

Apart from the harm that individual provisions would cause or the concerns that they raise, the bill as a whole would sweep away recent progress in reaching more eligible people.  For more than 15 years, since the early 2000s ― under Republican and Democratic administrations ― Congress, USDA, and states have streamlined and modernized SNAP in ways that enhanced state flexibility and balanced strong program integrity with improved access for those who qualify under program rules.  These efforts have resulted in more efficient and effective administration, as demonstrated by higher participation among eligible households and strong payment accuracy.

Those efforts began as a response to the unanticipated problems stemming from the 1996 welfare law, which not only radically restructured cash assistance by creating the Temporary Assistance for Needy Families (TANF) program, but also included deep benefit cuts in SNAP (then called food stamps) and a three-month limit on SNAP participation for certain unemployed childless adults.  In the first years after the law’s implementation, SNAP participation and costs plummeted far more than federal policymakers had anticipated, due in large part to a substantial drop in the share of eligible families receiving SNAP.  Because of problems in state administrative systems in the first years of the welfare law, many families moving from welfare to work and joining the ranks of the working poor were cut off SNAP when they left welfare, even though they remained eligible for SNAP. 

Aggravating this problem, some states instituted administrative practices in those years that had the unintended effect of making it harder for many working-poor parents to participate, largely by requiring them to take too much time off from work for repeated visits to SNAP offices at frequent intervals (such as every three months) to reapply for benefits.  This prompted many analysts and state policy officials from across the political spectrum to call for reforms that would improve access to SNAP for low-income working families.

As noted above, both the Clinton and the Bush administrations acted to address this problem, reflecting a bipartisan consensus that to make it difficult for families to receive SNAP assistance if they left welfare for low-wage work would reduce work incentives and contradicted welfare reform goals.  Congress enacted meaningful, although relatively modest, changes in 2002 and 2008 to lessen barriers to SNAP participation among the working poor.  Critical to this effort were changes that reduced both the number of trips that low-wage working households had to make to the SNAP office and the reporting and documentation they needed to provide of non-essential changes to their circumstances.  As a result, the percentage of eligible individuals in low-income working families that receive SNAP rose by more than two-thirds: from 43 percent in 2002 to more than 70 percent in recent years.  These gains were maintained during and after the recent deep recession.

The House farm bill would move the program backwards with respect to access by mandating and incenting states to dramatically increase paperwork, verification, and likely office visits.  These changes would surely compromise program access, particularly among working families.  Beginning in 2021, an estimated 7 million people, including 1.5 million adults working more than half time, would have to prove every month that they met the new 20-hour-a-week work requirement or were exempt.  These individuals are in some 30-40 percent of SNAP households.  This extraordinary increase in requirements, affecting both participants and states, would put millions of families at risk of losing benefits due to bureaucratic error.  Any clerical slip-up, by workers or caseworkers, could result in lost benefits.  The bill includes numerous other provisions that would demand more documentation and verification.

In addition, the bill would set SNAP’s allowable payment error threshold at $0.  As noted above, SNAP has a QC process to ensure the accuracy of household eligibility and benefit amounts.  States must sample a representative number of cases each month and state QC staff thoroughly review the accuracy of the original eligibility determination and benefit level.  USDA further reviews a subset of these reviews and, based on these reviews, annually reports state overpayment and underpayment error rates; states can be assessed significant penalties for above-average error rates.

To encourage states to focus their payment integrity efforts on the costliest types of errors, SNAP’s QC rules have long had a “tolerance” level below which error amounts do not count toward the state’s error rate.  The current level, set by the 2014 farm bill, is $37, with annual adjustments for inflation.  From 2000 to 2009, the QC tolerance level was set at $25 with no adjustments for inflation, but in 2009 Congress temporarily raised the tolerance level to $50 in the Recovery Act to acknowledge states’ concern that the Recovery Act’s temporary benefit increase would raise states’ error rates without indicating fault in program administration.  The Committee’s proposal would eliminate the error threshold altogether and set the amount at zero.

By definition, this would increase SNAP’s payment error rate by including more small errors in the calculation.[30]  It would also impede states from focusing on larger errors by including any small error that results from a minor mistake.  Some states would almost certainly respond by requiring more paperwork, such as six weeks of pay stubs rather than the typical four weeks in order to verify income, and by imposing other rules (such as more frequent interviews) that aim to reduce errors but would likelier create administrative burden and deter access to the program or reduce benefits.

It’s also unclear how this change would interact with the bill’s monthly assessment of eligibility relative to required work or training.  If a SNAP participant reported increased work hours in a given month, would the state then need to demand earnings verification and recalculate her benefits in order to avoid a payment error?  We are deeply concerned that the combination of these two provisions will result in states feeling obligated to demand a signifcant increase in paper documentation of changes in household circumstances, a problem Congress resolved years ago.  Program operators learned in the late 1990s the difficulty of issuing benefits in precisely accurate amounts to workers.  If program rules do not accommodate the very natural fluctuations in income and circumstances that households — particularly working households — have, states would feel significant pressure to require workers to document both their hours and earnings each month.


SNAP is efficient and effective.  It alleviates hunger and poverty and has positive impacts on recipients’ long-term outcomes.  And, it has exacting standards with respect to eligibility determination and overall program integrity.

Over the many years that I have worked on this program, Congress, USDA, and states have endeavored to maintain SNAP’s successful structure and design while also changing the program to better meet the needs of underserved groups (such as working families and seniors) and testing or implementing new ideas to improve program efficiency without compromising effectiveness.  Unfortunately, the House Agriculture Committee proposal would weaken SNAP’s ability to meet low-income households’ food needs, increasing hunger and hardship.  Given the mounting evidence of SNAP’s critical role in improving health and well-being, I urge Congress to look at ways to strengthen SNAP, not compromise its ability to meet the basic nutrition needs of struggling Americans.


End Notes

[1] For example, one study found that food insecurity did not rise between late 2008 and late 2009 for low-income participants likely to receive SNAP, but did rise for those with slightly higher incomes.  Mark Nord and Mark Prell, “Food Security Improved Following the 2009 ARRA Increase in SNAP Benefits,” U.S. Department of Agriculture, Economic Research Service, ERR-116, April 2011.

[2] U.S. Department of Agriculture, “Dynamics and Determinants of Supplemental Nutrition Assistance Program Participation from 2008 to 2012,” December 2014,

[3] U.S. Department of Agriculture, “Measuring the Effect of Supplemental Nutrition Assistance Program (SNAP) Participation on Food Security,” August 2013,

[4] Christian A. Gregory and Partha Deb, “Does SNAP Improve Your Health?” Food Policy, 50:11-19, 2015,

[5] Douglas Almond, Hillary Hoynes, and Diane Schanzenbach, “Inside the War on Poverty: The Impact of Food Stamps on Birth Outcomes,” Review of Economics and Statistics, 93(2), May 2011,; and Hilary Hoynes, Diane Whitmore Schanzenbach, and Douglas Almond, “Long-Run Impacts of Childhood Access to the Safety Net,” American Economic Review, 106(4):903–934, April 2016,

[6] Mithuna Srinivasan and Jennifer A. Pooler, “Cost-Related Medication Nonadherence for Older Adults Participating in SNAP, 2013–2015,” American Journal of Public Health, published online December 21, 2017,

[7] See Sarah L. Szanton et al., “Food assistance is associated with decreased nursing home admissions for Maryland’s dually eligible older adults,” BMC Geriatrics, 17(1), 162, 2017,; and Laura J. Samuel et al., “Does the Supplemental Nutrition Assistance Program Affect Hospital Utilization Among Older Adults? The Case of Maryland,” Population Health Management, 2017,

[8] Hilary W. Hoynes, Diane Whitmore Schanzenbach, and Douglas Almond, “Long Run Impacts of Childhood Access to the Safety Net,” American Economic Review, 106(4): 903–93, April 2016.

[9] Seth A. Berkowitz et al., “Supplemental Nutrition Assistance Program (SNAP) Participation and Health Care Expenditures Among Low-Income Adults,” JAMA Internal Medicine, November 2017.

[10] Seth Berkowitz, Hilary K. Seligman, and Sanjay Basu, “Impact of Food Insecurity and SNAP Participation on Healthcare Utilization and Expenditures,” University of Kentucky Center for Poverty Research Discussion Paper Series, DP2017-02, 2017.

[11] This share is from an analysis of data from the Census Bureau’s Survey of Income and Program Participation (SIPP) from 2011-2013 and looks at individuals participating in SNAP in a given month in mid-2012 and their work within one year of that month. For a full explanation of this data, see Brynne Keith-Jennings and Raheem Chaudhry, “Most Working-Age SNAP Participants Work, But Often in Unstable Jobs,” Center on Budget and Policy Priorities, March 15, 2018,

[12] These are shares from SIPP data applied to Agriculture Department data, since this information is not available in Agriculture Department data.

[13] These individuals may also have worked some months part time, but here we first looked at workers with full-time work. Here, “full time” means at least 35 hours per week and “half time” means at least 20 hours a week but fewer than 35.

[14] This analysis looks at the group of non-disabled adults who received SNAP at any point in a roughly 3.5-year period from mid-2009 through mid-2013.

[15] Daniel R. Cline and Randy Alison Aussenberg, “Errors and Fraud in the Supplemental Nutrition Assistance Program (SNAP),” Congressional Research Service, March 30, 2018,

[16] 2014 is the most recent year for which national data are available.

[17] SNAP State Activity Report, Fiscal Year 2016, p. 35,

[18]  Statement of Brandon Lipps, Acting Deputy Under Secretary, Food, Nutrition and Consumer Services Administrator, Food and Nutrition Service, United States Department of Agriculture, Before the Senate Committee on Agriculture, September 14 2017,

[19] Prior to revisions in the 2002 farm bill, states were eligible for enhanced funding for superior performance if their error rates were below 6 percent.

[20] SNAP benefits are federally funded.  States and the federal government share SNAP’s administrative costs, including certifying eligibility, issuing benefits, and ensuring program integrity.

[21] USDA also has a fraud hotline that the public can use to report suspected fraud to the agency.

[22] SNAP State Activity Report Fiscal Year 2016.

[23] SNAP State Activity Report Fiscal Year 2016, p. 2.

[24] USDA, “SNAP Retailer Management Annual Report, 2017,

[25] U.S. Department of Agriculture, “The Extent of Trafficking in the Supplemental Nutrition Assistance Program: 2012–2014,” August 2013,

[26] Daniel R. Cline and Randy Alison Aussenberg, “Errors and Fraud in the Supplemental Nutrition Assistance Program (SNAP),” Congressional Research Service, March 30, 2018,

[27] Ed Bolen et al., House Agriculture Committee’s Farm Bill Would Increase Food Insecurity and Hardship,” Center on Budget and Policy Priorities, updated May 1, 2018,  Other CBPP analyses available at;; and

[28] See “National Accuracy Clearinghouse (NAC) Evaluation, Final Report,” PCG, October 2015, p. 9, .

[29] SSA, FY2019 Congressional Justification, page 187,

[30] Government Accountability Office, “Supplemental Nutrition Assistance Program: Policy Changes and Calculation Methods Likely Affect Improper Payment Rates, and USDA Is Taking Steps to Help Address Recipient Fraud,” July 6, 2016, GAO-16-708T,