United States Government Accountability Office Report to Congressional Requesters SUPPLEMENTAL December 2018 NUTRITION ASSISTANCE PROGRAM Actions Needed to Better Measure and Address Retailer Trafficking GAO-19-167 December 2018 SUPPLEMENTAL NUTRITION ASSISTANCE PROGRAM Actions Needed to Better Measure and Address Highlights of GAO-19-167, a report to Retailer Trafficking congressional requesters Why GAO Did This Study What GAO Found SNAP is the largest federally funded The U.S. Department of Agriculture (USDA) Food and Nutrition Service’s (FNS) nutrition assistance program, providing estimates of retailer trafficking—when a retailer exchanges Supplemental about $64 billion in benefits to over 20 Nutrition Assistance Program (SNAP) benefits for cash instead of food—have million households in fiscal year 2017. limitations, though they suggest trafficking has increased in recent years, to $1 FNS oversees SNAP at the federal billion each year from 2012 to 2014. One key limitation of the estimates is that level and is responsible for authorizing FNS has not evaluated the accuracy of its assumption about the percentage of and overseeing retailers. While most SNAP benefits trafficked. FNS assumes that, among stores that trafficked, 90 benefits are used as intended, some percent of the benefits redeemed in small stores, and 40 percent in large stores, retailers have engaged in trafficking, were trafficked. A former FNS official stated that this assumption is based on which represents fraud and diverts discussions with investigators in the 1990s when FNS first developed its federal funds from their intended use. GAO was asked to review FNS’s approach to estimate trafficking, and that they have not since evaluated it for efforts to address SNAP retailer accuracy. However, there are options available for evaluating this assumption, trafficking since GAO’s last report in such as reviewing SNAP transaction data from stores that are known to have 2006. trafficked. Statistical standards for federal agencies indicate that assumptions should be reviewed for accuracy and validated using available, independent This report examines (1) what is known information sources. By not evaluating this key assumption, FNS’s commonly about the extent of SNAP retailer cited estimates of SNAP fraud are potentially inaccurate. trafficking, and (2) the extent to which FNS has taken steps intended to Illustration of Retailer Trafficking in the Supplemental Nutrition Assistance Program (SNAP) improve how it prevents, detects, and responds to retailer trafficking. GAO reviewed relevant federal laws and regulations, FNS policies, and studies related to retailer trafficking; assessed FNS’s use of statistical standards for federal agencies and selected leading practices in GAO’s Fraud Risk FNS has generally taken steps to address retailer trafficking that align with Framework; and interviewed FNS and leading fraud risk management practices, but the agency has not pursued USDA Office of Inspector General additional actions to prevent and respond to trafficking. For example: officials and key stakeholders. • Although FNS assigns a risk level to each store when it applies to participate What GAO Recommends in SNAP, it is not currently using this information to target its reauthorization activities to stores of greatest risk. During reauthorization, FNS reviews GAO is making five recommendations, previously approved stores for continued compliance with program including that FNS improve its requirements. FNS currently reauthorizes all stores on the same 5-year trafficking estimates by, for example, cycle, regardless of risk, although its policy states that it will reauthorize evaluating the accuracy of its certain high-risk stores annually. FNS officials planned to reauthorize a assumption of the percentage of sample of high-risk stores each year, but said they did not follow through with benefits that are trafficked; assess the those plans. Officials also stated that they did not document an analysis of benefits and costs of reauthorizing a the benefits and costs of this practice, which would be consistent with leading sample of high risk stores more fraud risk management practices. As a result, FNS may be missing an frequently than others; and move opportunity to provide early oversight of risky stores and prevent trafficking. forward with plans to increase penalties for trafficking. FNS generally • The Food, Conservation, and Energy Act of 2008 gave USDA the authority to agreed with GAO’s recommendations. strengthen penalties for retailers found to have trafficked, but as of November 2018, FNS had not implemented this authority. FNS proposed a View GAO-19-167. For more information, related rule change in 2012 and indicated the change was necessary to deter contact Kathy Larin at (202) 512-7215 or retailers from committing program violations, but the rule was not finalized. firstname.lastname@example.org. By failing to take timely action to strengthen penalties, FNS has not taken full advantage of an important tool for deterring trafficking. United States Government Accountability Office Contents Letter 1 Background 3 FNS Estimates Suggest Retailer Trafficking Has Increased in Recent Years, but the Estimates Have Limitations 10 FNS Has Taken Some Steps That Generally Align with Leading Practices to Prevent, Detect, and Respond to Retailer Trafficking, but Has Not Pursued Others 17 Conclusions 30 Recommendations for Executive Action 31 Agency Comments and Our Evaluation 31 Appendix I GAO Contact and Staff Acknowledgments 34 Figures Figure 1: Number of Authorized Stores and Amount of Benefits Redeemed in the Supplemental Nutrition Assistance Program (SNAP), Fiscal Years 2007 – 2017 4 Figure 2: Examples of Legitimate and Retailer Trafficking Transactions in the Supplemental Nutrition Assistance Program (SNAP) 6 Figure 3: Food and Nutrition Service (FNS) Offices with Responsibilities Related to Supplemental Nutrition Assistance Program (SNAP) Retailer Oversight and Trafficking 8 Figure 4: Food and Nutrition Service’s (FNS) Methodology to Estimate Retailer Trafficking in the Supplemental Nutrition Assistance Program (SNAP) 9 Figure 5: Supplemental Nutrition Assistance Program (SNAP) Retailer Trafficking Estimates and Ranges, 2006 – 2014 12 Figure 6: Previous (as of 2006) and Current Supplemental Nutrition Assistance Program (SNAP) Food Stock Requirements for Authorized Retailers 19 Figure 7: Food and Nutrition Service (FNS) Photographs from Training Session on Detecting Retailer Trafficking 22 Figure 8: U.S. Department of Agriculture (USDA) Office of Inspector General (OIG) Retailer Trafficking Investigations Accepted and Declined for Prosecution, Fiscal Years 2007 – 2017 25 Page i GAO-19-167 Retailer Trafficking Figure 9: Permanent Disqualifications and Penalties for Supplemental Nutrition Assistance Program (SNAP) Retailer Trafficking, by Type of Case, Fiscal Years 2007 – 2017 26 Abbreviations ALERT Anti-Fraud Locator using Electronic Benefits Transfer Retailer Transactions EBT Electronic Benefits Transfer FNS Food and Nutrition Service MOU memorandum of understanding OIG Office of Inspector General OMB Office of Management and Budget SLEB state law enforcement bureau SNAP Supplemental Nutrition Assistance Program USDA United States Department of Agriculture This is a work of the U.S. government and is not subject to copyright protection in the United States. The published product may be reproduced and distributed in its entirety without further permission from GAO. However, because this work may contain copyrighted images or other material, permission from the copyright holder may be necessary if you wish to reproduce this material separately. Page ii GAO-19-167 Retailer Trafficking Letter 441 G St. N.W. Washington, DC 20548 December 14, 2018 Congressional Requesters The Supplemental Nutrition Assistance Program (SNAP) is the largest federally funded nutrition assistance program, providing approximately $64 billion in benefits to about 42 million people in over 20 million households in fiscal year 2017. SNAP recipients receive monthly cash benefits on an Electronic Benefits Transfer (EBT) card and redeem them for food at authorized retailers. Most SNAP benefits are used for the intended purpose, according to the U.S. Department of Agriculture’s (USDA) Food and Nutrition Service (FNS), which oversees SNAP at the federal level and is responsible for authorizing and overseeing retailers that participate in the program. However, program officials have long-standing concerns about retailer trafficking—a practice in which retailers exchange recipients’ benefits for cash instead of food, often taking a fraudulent profit. As we previously reported in 2006, FNS has faced challenges in addressing retailer trafficking. 1 You asked us to review steps FNS has taken in this area since our last report. We assessed (1) what is known about the extent of retailer trafficking in SNAP, and (2) the extent to which FNS has taken steps since 2006 intended to improve how it prevents, detects, and responds to retailer trafficking. To examine what is known about the extent of retailer trafficking and FNS’s trafficking estimates, we reviewed FNS’s three most recent reports on the Extent of Trafficking in the Supplemental Nutrition Assistance Program—covering data from calendar years 2006 through 2014. In order to assess the reliability of the estimates, we reviewed technical information on FNS’s methodology included in appendices to its trafficking reports and additional FNS studies regarding the methodology used to estimate trafficking. We also reviewed the Office of Management 1 GAO, Food Stamp Trafficking: FNS Could Enhance Program Integrity by Better Targeting Stores Likely to Traffic and Increasing Penalties, GAO-07-53 (Washington, D.C.: Oct. 13, 2006). Page 1 GAO-19-167 Retailer Trafficking and Budget’s statistical standards for federal agencies. 2 We interviewed FNS officials about the methodology and data used to estimate trafficking and any steps the agency has taken or plans to take to address any limitations of the methodology. We also interviewed USDA Office of Inspector General (OIG) officials about their prior work related to the retailer trafficking estimates, as well as a former FNS official familiar with the development of the methodology and the contractors who produced the most recent estimates and studied limitations of the methodology. We determined that FNS’s trafficking estimates were sufficiently reliable for the purpose of describing general trends in retailer trafficking, but we also identified some limitations of the estimates, which we discuss in this report. To examine FNS’s steps intended to improve how it prevents, detects, and responds to retailer trafficking, we reviewed relevant federal laws and regulations and proposed rules related to retailer trafficking, as well as relevant FNS policies, standard operating procedures, and staff performance work plans, among other documents. We also reviewed relevant GAO and USDA OIG reports and determined the status of FNS’s responses to prior audit recommendations by interviewing FNS and OIG officials and obtaining relevant documents. We interviewed FNS officials about steps the agency has taken since 2006 to improve how it prevents, detects, and responds to retailer trafficking, and OIG officials about their role investigating SNAP retailer trafficking. We assessed FNS’s efforts to address retailer trafficking against selected leading practices in GAO’s A Framework for Managing Fraud Risks in Federal Programs (Fraud Risk Framework) related to designing and implementing fraud controls and monitoring and evaluating activities. 3 To obtain stakeholder perspectives, we interviewed representatives from two national associations representing independent grocery stores and convenience stores. We also interviewed officials from three states—Georgia, Pennsylvania, and 2 Office of Management and Budget, Standards and Guidelines for Statistical Surveys (Washington, D.C.: September 2006). 3 GAO, A Framework for Managing Fraud Risks in Federal Programs, GAO-15-593SP (Washington, D.C.: July 2015). This document describes leading practices that agencies can follow to proactively and strategically manage fraud risks. The Fraud Risk Framework describes leading practices in four components: commit, assess, design and implement, and evaluate and adapt. We selected leading practices within the design and implement and evaluate and adapt components because FNS had identified SNAP retailer trafficking as a fraud risk. The other components relate to organizational culture and risk assessments, which we considered to be broader than retailer trafficking. We did not evaluate the effectiveness of FNS’s controls for addressing retailer trafficking. Page 2 GAO-19-167 Retailer Trafficking Washington—about their experiences working with FNS to help investigate SNAP retailer trafficking. We selected these states as illustrative examples based on several criteria, including the rate at which retailers in each state received sanctions for SNAP violations and geographic dispersion. Finally, to provide context for our findings, we analyzed summary-level data from FNS and the USDA OIG on trends in SNAP participation and agency activities to address retailer trafficking from fiscal years 2007 through 2017. We assessed the reliability of these data through a review of FNS and OIG documents and interviews with knowledgeable officials. We found the data to be sufficiently reliable for the purpose of providing contextual information on trends in FNS and OIG retailer monitoring and oversight activities. We conducted this performance audit from January 2018 to December 2018 in accordance with generally accepted government auditing standards. Those standards require that we plan and perform the audit to obtain sufficient, appropriate evidence to provide a reasonable basis for our findings and conclusions based on our audit objectives. We believe that the evidence obtained provides a reasonable basis for our findings and conclusions based on our audit objectives. SNAP is intended to help low-income households obtain a more nutritious Background diet by providing them with benefits to purchase food from authorized retailers nationwide. 4 SNAP is jointly administered by FNS and the states. FNS pays the full cost of SNAP benefits and shares the costs of administering the program with the states. FNS is responsible for promulgating SNAP program regulations, ensuring that state officials administer the program in compliance with program rules, and authorizing and monitoring stores from which recipients may purchase food. 5 States are responsible for determining applicant eligibility, calculating the amount of their benefits, issuing the benefits on EBT cards—which can be used like debit cards to purchase food from authorized retailers—and investigating possible program violations by recipients. 4 Recipients may not purchase certain items, such as tobacco, alcohol, or hot food intended for immediate consumption. 5 For the purposes of this report, we use the term “retailer” generally or to refer to the entity that owns a store. We use the term “store” to refer to individual store locations. FNS authorizes store locations individually. Page 3 GAO-19-167 Retailer Trafficking Participation in SNAP has generally increased among recipients and retailers in recent years. Specifically, participation in SNAP increased from about 26 million recipients in fiscal year 2007 to 42 million in fiscal year 2017, leading to a corresponding increase in the amount of SNAP benefits redeemed. The number of stores FNS authorized to participate in SNAP also increased, from about 162,000 nationwide in fiscal year 2007 to more than 250,000 in fiscal year 2017. 6 Although there was particular growth in the number of small grocery and convenience stores, as well as “other” stores (which include independent drug stores, general merchandise stores like dollar stores, and farmers’ markets), the majority of SNAP benefits were redeemed at large grocery stores and supermarkets in each year (see fig. 1). Figure 1: Number of Authorized Stores and Amount of Benefits Redeemed in the Supplemental Nutrition Assistance Program (SNAP), Fiscal Years 2007 – 2017 6 The number of authorized stores—based on data provided by FNS—excludes meal services, which can also participate in SNAP. An example of a meal service is a dining facility in a building predominantly occupied by elderly people and recipients of Supplemental Security Income and their spouses. Page 4 GAO-19-167 Retailer Trafficking a Specialty food stores include stores that specialize in the sale of certain foods, such as bread or cereal products, fruits or vegetables, or meat products. b The “other” category includes stores that primarily sell general merchandise, but also sell a variety of food products. These stores include independent drug stores, dollar stores, general stores, as well as farmers’ markets. Retailer Trafficking According to FNS, most SNAP benefits are used for the intended purpose; however, as we have reported in prior work, FNS has faced challenges addressing trafficking—one type of program fraud. 7 In general, trafficking occurs when retailers exchange recipients’ SNAP benefits for cash, often taking a fraudulent profit. For example, a retailer might charge $100 to a recipient’s SNAP EBT card and give the recipient $50 in cash instead of $100 in food. The federal government reimburses the retailer $100, which results in a fraudulent $50 profit to the retailer. While this type of trafficking is a direct exchange of SNAP benefits for money, trafficking also can be done indirectly. For example, a retailer might give a recipient $50 in cash for the use of $100 in benefits on that recipient’s EBT card. The retailer could then use the EBT card to purchase $100 in products at another SNAP retailer (see fig. 2). In this instance, the retailer would profit because they paid $50 for $100 worth of products, and the retailer might also increase their profit by reselling the products at a higher price in their own store. 7 GAO-07-53. Page 5 GAO-19-167 Retailer Trafficking Figure 2: Examples of Legitimate and Retailer Trafficking Transactions in the Supplemental Nutrition Assistance Program (SNAP) Retailer Management and Among other things, FNS is responsible for authorizing and monitoring Oversight retailers who participate in SNAP to ensure program integrity. In order to participate in SNAP, a retailer applies to FNS and demonstrates that they meet program requirements, such as those on the amount and types of food that authorized stores must carry. FNS verifies a retailer’s compliance with these requirements, for example, through an on-site inspection of the store. If the retailer meets requirements, FNS generally authorizes it to participate for a period of 5 years. FNS then monitors retailers’ participation by analyzing data on SNAP transactions and conducting undercover investigations, among other activities. If FNS suspects a retailer is trafficking, it generally must notify the USDA OIG—which is responsible for investigating allegations of fraud and abuse in all of USDA’s programs, including SNAP—before opening a Page 6 GAO-19-167 Retailer Trafficking case. The OIG may choose to open its own investigation of the retailer for possible criminal prosecution, or allow FNS to pursue the case. 8 If FNS determines that a retailer has engaged in trafficking, FNS sanctions the store. 9 Generally, stores found to have engaged in trafficking are permanently disqualified from SNAP, but in limited circumstances, the owner may instead receive a civil monetary penalty. 10 Retailers who do not agree with the sanction assessed by FNS can file a written request to have FNS’s Administrative Review Branch review the decision, and, if not satisfied, file a complaint in the appropriate U.S. District Court. In 2013, FNS consolidated its retailer management functions, including those for authorizing stores and analyzing SNAP transaction data, into a single national structure known as the Retailer Operations Division (see fig. 3). 8 The OIG also may work with other federal law enforcement agencies, such as the Federal Bureau of Investigation, to pursue retailer trafficking. 9 Prior to assessing a permanent disqualification or penalty as a result of an undercover investigation, FNS generally presents the case to the OIG for possible escalation. Specifically, the OIG will decide whether to investigate the case further for possible prosecution by the U.S. Attorney’s Office or by state and local prosecutors, or allow FNS to complete the disqualification action. 10 Under current regulations, FNS assesses either a civil monetary penalty for trafficking or permanent disqualification, but not both. 7 C.F.R. § 278.6. At the retailer’s request, FNS can assess a civil monetary penalty for trafficking if a store demonstrates that it has an effective compliance policy and program in effect to prevent violations of relevant law and regulations and meets additional criteria. Page 7 GAO-19-167 Retailer Trafficking Figure 3: Food and Nutrition Service (FNS) Offices with Responsibilities Related to Supplemental Nutrition Assistance Program (SNAP) Retailer Oversight and Trafficking Note: FNS generally notifies the U.S. Department of Agriculture Office of Inspector General (OIG) before opening a case based on analysis of SNAP transaction data. In addition, if an FNS investigator is successful in trafficking with a retailer, FNS generally presents the case to the OIG for possible escalation. Estimating Retailer Since 1995, FNS has published periodic reports estimating the extent of Trafficking trafficking in SNAP as part of its efforts to monitor program integrity. These trafficking estimates are the most commonly cited measure of SNAP fraud, including in the news media and congressional testimony. FNS estimates retailer trafficking by adjusting a sample of stores known or suspected of trafficking to reflect the total population of SNAP- authorized stores. For each report, FNS uses 3 years of data on stores Page 8 GAO-19-167 Retailer Trafficking and SNAP transactions to estimate the amount and percentage of benefits that were trafficked and the percentage of stores engaged in trafficking (see fig. 4). For example, the most recent report—published in September 2017—analyzes data from 2012 through 2014. Figure 4: Food and Nutrition Service’s (FNS) Methodology to Estimate Retailer Trafficking in the Supplemental Nutrition Assistance Program (SNAP) Note: FNS’s Watch List is a list of stores that FNS has identified as having suspicious transaction patterns or other indications of potential trafficking. Page 9 GAO-19-167 Retailer Trafficking FNS Estimates Suggest Retailer Trafficking Has Increased in Recent Years, but the Estimates Have Limitations FNS Estimates Indicate an FNS’s data indicate an increase in the estimated rate of retailer trafficking Increase in Retailer in recent years. FNS reported in March 2011 that approximately $330 million in SNAP benefits (or 1 percent of all benefits redeemed) were Trafficking, but the Actual trafficked annually from 2006 through 2008, and that approximately 8.2 Extent of Trafficking Is percent of all authorized stores engaged in trafficking. 11 In its most recent Uncertain report from September 2017, FNS reported that approximately $1 billion in SNAP benefits (or 1.5 percent) were trafficked annually from 2012 through 2014, and that approximately 11.8 percent of all authorized stores engaged in trafficking. 12 Although FNS produces the trafficking estimates with accepted statistical methods, its reports do not clearly convey the level of uncertainty introduced by the approach used to calculate the estimates. Throughout each report, FNS presents its estimates as precise numbers. However, uncertainty is introduced when extrapolating from a smaller sample—in this case, an investigative sample that solely includes stores known to have trafficked or suspected of trafficking—to the full population of SNAP- authorized stores because the extent to which the sample reflects the 11 U.S. Department of Agriculture, Food and Nutrition Service, The Extent of Trafficking in the Supplemental Nutrition Assistance Program: 2006-2008 (Alexandria, VA: March 2011). 12 U.S. Department of Agriculture, Food and Nutrition Service, The Extent of Trafficking in the Supplemental Nutrition Assistance Program: 2012-2014 (Alexandria, VA: September 2017). Page 10 GAO-19-167 Retailer Trafficking broader population of stores is unknown (see sidebar). 13 According to the Office of Management and Budget’s (OMB) statistical standards for federal agencies, possible variation in estimates should be noted, such as by reporting the range of each estimate. 14 While FNS discusses some limitations of its trafficking estimates in the body of each report, only the report’s appendices include information that can be used to assess the level of uncertainty around the estimates. Using information contained in these appendices, we found widely varying trafficking estimates. 15 For example, although FNS reported that approximately $1 billion in SNAP benefits were trafficked annually from 2012 through 2014, information in the report’s appendices indicates that the amount trafficked could have ranged from about $960 million to $4.7 billion. In other words, the total value of SNAP benefits that were trafficked each year from 2012 through 2014 could have been approximately $40 million less or $3 billion more than FNS reported (see fig. 5). 13 FNS’s sample is not a random sample drawn from the population of all SNAP-authorized stores. Instead, it is the subset of stores FNS suspected of trafficking or found trafficking. As a result, any estimates made from the sample to the population need to be adjusted to reduce the bias introduced because the sample only includes stores meeting these criteria. In a random sample, every SNAP-authorized store would have a known chance of being selected for the sample, regardless of whether they were suspected of trafficking. Statistical theory ensures that such a sample would produce estimates for the population of SNAP-authorized stores that are accurate to a specified level of certainty. 14 Office of Management and Budget, Standards and Guidelines for Statistical Surveys (Washington, D.C.: September 2006). 15 Two of the appendices in FNS’s reports include the results from tests that look at how much the estimates could change if different stores were included in the sample. The first test varies the cases that are chosen from the current investigative sample. FNS refers to the results of the test as “estimate intervals.” The second test is a sensitivity analysis that assesses how the estimates could change if the sample were defined differently (e.g., if FNS assumed that a certain percentage of cases on its Watch List that were closed without further action were actually trafficking). The lower end of the range we report in figure 5 is the lowest estimate from either appendix, and the upper end of the range is the highest estimate from either appendix. Although these appendices use different approaches, FNS reported that both should be considered to indicate the possible extent of trafficking. These analyses are included in appendices H and I to FNS’s August 2013 and September 2017 trafficking estimates reports and appendices I and L to the March 2011 report. Page 11 GAO-19-167 Retailer Trafficking Figure 5: Supplemental Nutrition Assistance Program (SNAP) Retailer Trafficking Estimates and Ranges, 2006 – 2014 Note: The estimates presented in the figure above are annualized, based on 3 years of data. Information on the ranges of the estimates was obtained from two appendices to FNS’s The Extent of Trafficking in the Supplemental Nutrition Assistance Program reports. The two appendices report the results of separate tests that look at how much the estimates could change if different stores were included in the sample. For example, one appendix includes the results of a sensitivity analysis that assessed how the estimates could change if the sample were defined differently. FNS officials stated the agency has not considered and does not intend to consider changes to how it reports its trafficking estimates in the next report. 16 According to an FNS official, FNS would like the reports to continue to provide non-technical information that is comparable to prior years. However, as shown in the figure above, it is possible to compare estimates over time when estimates are presented with ranges. Further, reporting the level of uncertainty with each estimate increases transparency. According to a recent Congressional Research Service report, these estimates are the most-often cited measure of fraud in SNAP. 17 The estimates have been cited in the news media and 16 FNS officials told us that they awarded the contract for the next trafficking estimates study in September 2018. 17 Congressional Research Service, Errors and Fraud in the Supplemental Nutrition Assistance Program (SNAP) (Washington, D.C.: Mar. 30, 2018). Page 12 GAO-19-167 Retailer Trafficking congressional testimony, and FNS officials stated the estimates can help quantify the outcomes of FNS’s efforts to prevent, detect, and respond to retailer trafficking. By not clearly reporting the level of uncertainty around these commonly cited estimates of SNAP retailer trafficking, FNS’s reports are potentially providing misleading information to Congress and the public regarding the extent of fraud in SNAP. FNS Evaluated Ways to FNS has acknowledged limitations with its current approach to estimating Address Some Limitations retailer trafficking and evaluated ways to address some of those limitations. As previously noted, FNS selects a non-random sample of in the Trafficking stores known to have trafficked or suspected of trafficking when Estimates, but Does Not calculating its estimates, which may introduce bias into those trafficking Plan to Take Further Steps estimates (see sidebar). For example, the sample could overestimate the extent of retailer trafficking if the stores in the sample that are targeted for investigation are more likely to traffic. Conversely, if FNS’s detection methods do not capture all instances of trafficking—such as retailers who only traffic with people they know—the sample could lead to an underestimate of trafficking among all SNAP-authorized stores. Recognizing that the trafficking estimates provide important information on program fraud, FNS evaluated ways to address limitations in the estimates. In 2013, FNS convened a technical working group of experts to discuss alternate ways to estimate retailer trafficking. That group made various recommendations to improve the estimates, some of which FNS pursued through additional analyses. For example, to address limitations introduced by the sample FNS uses to estimate trafficking, the agency conducted a study to assess the feasibility of calculating its estimates using a national random sample of stores. 18 However, FNS determined it would be infeasible to use a random sample because of the costs and resources that would be involved. According to FNS officials, it cost the agency approximately $67,000 to produce the September 2017 trafficking estimates report. By comparison, FNS estimated that using a national random sample could cost between $11.5 million and $38 million, depending on the specific 18 FNS contracted for a feasibility study in response to a recommendation from the USDA OIG. U.S. Department of Agriculture, Office of Inspector General, Analysis of FNS’ Supplemental Nutrition Assistance Program Fraud Prevention and Detection Efforts, Audit Report 27002-0011-13 (Washington, D.C.: September 2012). Some of the members of the technical working group in 2013 also recommended FNS explore using a random sample of stores. Page 13 GAO-19-167 Retailer Trafficking sample selection method. This is because, among other factors, taking this approach would require investigative staff to visit stores suspected of trafficking as well as those that are not suspected of trafficking. Doing so would require a significant number of additional investigators, according to the feasibility study. Also in response to a recommendation by the technical working group, FNS contracted for a study in November 2017 that reviewed the five factors the agency uses to make adjustments to reduce the bias in its sample of stores (see sidebar). FNS began using these five factors—such as the type of store—more than 20 years ago when it initially developed its approach to estimating trafficking. The study evaluated the continued relevance of the five factors, as well as the relevance of additional factors related to store characteristics and neighborhood demographics. 19 The study did not make recommendations, and the expert who conducted the study told us that based on the analysis, the original five factors remain relevant. As a result, FNS officials stated the agency would continue to use these factors to reduce bias in the sample. However, FNS has not evaluated whether factors the agency currently uses to identify stores for possible investigation could help reduce bias in the sample and improve the trafficking estimates. Specifically, FNS analyzes data on SNAP transactions and looks for suspicious patterns or other indications of potential trafficking. Based on the results of these analyses, FNS assigns a numeric score to each store, and stores with scores above a certain threshold are added to FNS’s Watch List for further review. 20 Several experts have suggested to FNS that including this score or other related factors when adjusting the investigative sample could help reduce the bias in the sample and improve the trafficking 19 FNS’s contractor explored the current factors and possible other factors as part of a larger proof-of-concept study on the use of statistical models—instead of the current methodology—to calculate trafficking estimates. Prior to this, the technical working group in 2013 recommended FNS re-examine the current factors to ensure they are still predictors of trafficking. 20 The numeric scores are based on the results of a logistic regression model. Stores also must meet other criteria, such as having monthly SNAP transactions over a certain level, to be included on the Watch List. FNS analysts and investigators prioritize the stores on this list, among other factors, for further review. We discuss FNS’s data analytics to detect trafficking in further detail later in this report. Page 14 GAO-19-167 Retailer Trafficking estimates, 21 yet FNS has not evaluated the use of these factors for this purpose. 22 FNS officials said that stores’ numeric scores and the factors related to the Watch List are not public information, and the agency’s preference is to be transparent about the methodology used to produce the trafficking estimates. However, FNS already describes its Watch List and the use of a numeric score threshold in an appendix to its trafficking reports. According to OMB’s statistical standards, federal agencies should take steps to maximize the objectivity of the statistical information they produce. Objectivity refers to whether the information is accurate, reliable, and unbiased. 23 Without evaluating the usefulness of the Watch List factors for adjusting the sample, FNS may miss an opportunity to reduce the bias in the sample and improve the accuracy of its trafficking estimates. In addition, FNS has not evaluated the accuracy of its assumption of the percentage of SNAP benefits trafficked by different types of stores, which FNS developed over 20 years ago from anecdotal information. Among stores that engaged in trafficking, FNS assumes that 90 percent of benefits redeemed in small stores and 40 percent of benefits redeemed in large stores were trafficked (see sidebar). A former FNS official who helped develop the agency’s approach for estimating trafficking stated that the assumption was based on conversations with investigators in the 1990s—deemed to be the best source of information at the time. He noted that the investigators who were consulted generally disagreed on the percentage of benefits that were trafficked, as the actual percentage could vary widely based on whether, for example, one employee had trafficked or the entire store was a front for trafficking. However, the investigators generally agreed that 90 percent and 40 percent would overestimate trafficking by retailers in small and large stores, respectively. According to FNS officials, in the absence of other data, they preferred to use an overestimate, rather than an underestimate, of the percentage of benefits trafficked in stores found to have trafficked. 21 The contractors who produced the most recent retailer trafficking estimates and studied limitations of the methodology, as well as the technical working group FNS convened in 2013, both discussed the use of a numeric score or related factors with FNS. 22 Specifically, using factors to adjust the investigative sample that are related to a store’s likelihood of being investigated by FNS could help reduce selection bias in the trafficking estimates. 23 Office of Management and Budget, Standards and Guidelines for Statistical Surveys (Washington, D.C.: September 2006). Page 15 GAO-19-167 Retailer Trafficking Despite an increase in the availability of data on retailer trafficking over the last 20 years, FNS officials have not evaluated the accuracy of this key assumption and said that they have no plans to do so. FNS officials noted that they do not believe there are available data that indicate whether the assumption is accurate and, as such, any evaluation would require new data collection. However, according to contractors and a former official we spoke with who had studied the methodology as well as USDA OIG officials, data are currently available that may help FNS evaluate the accuracy of this assumption. 24 For example, they suggested FNS could analyze the transaction data of stores that trafficked to identify the percentage of all redeemed SNAP benefits that were consistent with known indicators of trafficking. Currently, OIG officials told us that they use a similar approach to calculate the amount of benefits trafficked for a store whose owner is being prosecuted. According to OMB’s statistical standards, assumptions should be reviewed for accuracy and validated using available, independent information sources. 25 Without FNS evaluating its key assumption of the percentage of SNAP benefits trafficked, the estimates it reports on the extent of program fraud are potentially inaccurate. 24 We interviewed a former FNS official familiar with the development of the methodology and the contractors who produced the most recent estimates and studied limitations of the methodology. One of the contractors also was a member of FNS’s technical working group in 2013. 25 Office of Management and Budget, Standards and Guidelines for Statistical Surveys (Washington, D.C.: September 2006). Page 16 GAO-19-167 Retailer Trafficking FNS Has Taken Some Steps That Generally Align with Leading Practices to Prevent, Detect, and Respond to Retailer Trafficking, but Has Not Pursued Others FNS Has Taken Steps to Address Retailer Trafficking, but Has Not Pursued Certain Prevention and Response Activities Preventing Trafficking in the FNS has taken some steps to prevent retailer trafficking that align with Retailer Authorization Process leading fraud risk management practices and our prior recommendations, but has not pursued some opportunities for early oversight. 26 As we note in our Fraud Risk Framework, while fraud control activities can be interdependent and mutually reinforcing, preventative activities generally offer the most cost-effective investment of resources. FNS officials told us that the agency tries to prevent trafficking through its policies and procedures for authorizing stores to participate in SNAP. Since our 2006 26 The Fraud Reduction and Data Analytics Act of 2015, enacted in June 2016, requires the Office of Management and Budget to establish guidelines for federal agencies to create controls to identify and assess fraud risks and design and implement antifraud control activities and incorporate leading practices in GAO’s Fraud Risk Framework. According to the Act, the controls should include: (1) conducting an evaluation of fraud risks and using a risk-based approach to design and implement financial and administrative control activities to mitigate identified fraud risks; (2) collecting and analyzing data from reporting mechanisms on detected fraud to monitor fraud trends and using that data and information to continuously improve fraud prevention controls; and (3) using the results of monitoring, evaluation, audits, and investigations to improve fraud prevention, detection, and response. Further, agencies are required to annually report on their progress in implementing the required controls for each of the first three fiscal years after the Act’s enactment. Pub. L. No. 114-186, § 3, 130 Stat. 546, 546, codified at 31 U.S.C. § 3321 note. Page 17 GAO-19-167 Retailer Trafficking report, 27 FNS has taken some steps to amend retailer authorization policies to address vulnerabilities that we identified. For example: • Increasing requirements for food that retailers must stock to participate in SNAP: In 2006, we found that FNS had minimal requirements for the amount of food that retailers must stock, which could allow retailers more likely to traffic into the program. At that time, FNS officials said that they authorized stores with limited food stock to provide access to food in low-income areas where large grocery stores were scarce. In 2006, retailers were generally required to stock a minimum of 12 food items (at least 3 varieties in each of 4 staple food categories), but FNS rules did not specify how many items of each variety would constitute sufficient stock. 28 We recommended that FNS develop criteria to help identify the stores most likely to traffic, using information such as the presence of low food stock. 29 In 2016, FNS promulgated a final rule increasing food stock requirements and, in January 2018, issued a policy memorandum to clarify these requirements. 30 FNS officials told us that the new requirements are designed to encourage stores to provide more healthy food options for recipients and discourage trafficking. According to the memorandum, retailers are now generally required to stock at least 36 food items (including stocking at least 3 varieties in 27 GAO-07-53. 28 Alternatively, retailers could meet a second criterion for authorization, which required a store to have more than 50 percent of its sales in staple foods. According to FNS, staple foods are the basic food items that make up a significant portion of an individual’s diet and are usually prepared at home. Staple food varieties are different types of food within a staple food category. For example, oranges, bananas, and broccoli are three distinct staple food varieties in the vegetables or fruits category. 29 GAO-07-53. We recommended that FNS develop criteria to help identify and target its monitoring resources to stores most likely to traffic, including those with low food stocks. FNS stated that it generally agreed with our recommendations. 30 The final rule, “Enhancing Retailer Standards in the Supplemental Nutrition Assistance Program (SNAP),” published on December 15, 2016, implemented provisions of the Agricultural Act of 2014 (known as the 2014 Farm Bill), and made other changes. 81 Fed. Reg. 90,675. For example, changes were made to food stock requirements, including the number and types of items stores must stock (known as “depth of stock” requirements). Subsequently, the Consolidated Appropriations Act, 2017, enacted on May 5, 2017, prevented implementation of certain provisions of FNS’s December 2016 final rule and directed USDA to continue utilizing the requirements in place prior to the 2014 Farm Bill until FNS promulgates specific regulatory amendments. Pub. L. No. 115-31, § 765, 131 Stat. 135, 179. FNS published a policy memorandum in January 2018 to clarify the requirements, as those regulatory amendments had not yet been made. Page 18 GAO-19-167 Retailer Trafficking each of 4 staple food categories, and 3 items of each variety). 31 See figure 6 for a comparison of the previous (as of 2006) and current (reflecting the January 2018 memorandum) requirements. Figure 6: Previous (as of 2006) and Current Supplemental Nutrition Assistance Program (SNAP) Food Stock Requirements for Authorized Retailers Note: Alternatively, under both the previous and current requirements, retailers can meet a second criterion for authorization, which requires a store to have more than 50 percent of its sales in staple foods. According to FNS, staple foods are the basic food items that make up a significant portion of an individual’s diet and are usually prepared at home. a “Current requirements” refers to the requirements as clarified in FNS’s January 2018 policy memorandum. This memorandum was in response to a 2017 federal law that prevented the implementation of certain provisions of a 2016 FNS final rule. 31 Retailers can also meet a second criterion for authorization, which requires a store to have more than 50 percent of its sales in staple foods. Page 19 GAO-19-167 Retailer Trafficking • Assessing retailer risk levels: Also in 2006, we found that FNS had not conducted analyses to identify characteristics of stores at high risk of trafficking and to target its resources accordingly. For example, we reported that some stores may be at risk of trafficking because one or more previous owners had been found to be trafficking at the same location. At that time, FNS did not have a system in place to ensure that these stores were quickly targeted for heightened attention. We recommended that FNS identify the stores most likely to traffic and provide earlier, more targeted oversight to those stores. 32 In 2009, FNS established risk levels for stores: high, medium, and low. For example, high-risk stores are those with a prior permanent disqualification at that location or nearby. 33 In January 2012, FNS revised its policy for authorizing high-risk stores. The policy requires high-risk retailers to provide specific documentation to ensure that the owners listed on the application have not been previously disqualified or do not have ties to a previously disqualified owner, such as a letter from the bank listing the authorized signers on the store’s accounts. 34 Although FNS took these steps to identify risk levels for stores and target its initial authorization activities accordingly, the agency is not currently using this information to target its reauthorization activities to stores of greatest risk. During reauthorization, FNS reviews previously approved 32 GAO-07-53. FNS stated that it generally agreed with our recommendations, and the agency implemented this recommendation by establishing risk levels for stores and targeting its detection efforts to stores most likely to traffic. 33 According to FNS’s standard operating procedure for processing store applications, stores are considered high risk if the store is located in a ZIP code with a high concentration of permanently disqualified stores. Stores also can be considered high risk if a person associated with the applicant store is also associated with a permanently disqualified store, among other reasons. Medium-risk stores include those with temporary disqualifications or other prior program violations, according to FNS. 34 In 2013, the USDA OIG found that some permanently disqualified retail store owners circumvented their disqualification by transferring ownership of their stores to a spouse, relative, co-habitant, prior employee, or manager. The OIG found that this occurred because FNS had not established adequate guidance and management controls to ensure that staff reviewed authorization and compliance documentation. In addition, while FNS had procedures to deny applicants who attempted to circumvent disqualification, the OIG found that those procedures did not require a supervisory review to ensure documents obtained were adequately evaluated by staff prior to authorization. In response, FNS instituted supervisory reviews of a random sample of authorization decisions to ensure that staff followed high-risk and other policies and procedures. U.S. Department of Agriculture, Office of Inspector General, FNS: Controls for Authorizing Supplemental Nutrition Assistance Program Retailers, Audit Report 27601-0001-31 (Washington, D.C.: July 2013). Page 20 GAO-19-167 Retailer Trafficking stores for continued compliance with program requirements. Although the agency’s policy and website both state that certain high-risk stores will be reauthorized annually, FNS is currently reauthorizing all stores on the same 5-year cycle, regardless of risk. 35 FNS reauthorized most high-risk stores under this policy one time in fiscal year 2013, but officials told us that they then discontinued annual reauthorizations after an in-depth assessment of the benefits and costs of this practice. 36 For example, FNS staff reported collecting more than 150,000 documents as part of the fiscal year 2013 reauthorization cycle and found that collecting these documents annually is ineffective and burdensome to FNS and the retailer. FNS instead decided to annually reauthorize a sample of high-risk retailers as a result of its assessment of the fiscal year 2013 cycle, but did not follow through with those plans. Specifically, the agency decided to pursue annual reauthorizations of a sample of stores at the greatest risk of program violations—those at the same address as a store that had been previously permanently disqualified. However, FNS officials did not have documentation that the approach was ever implemented or that they assessed the benefits and costs of reauthorizing this sample of high-risk retailers. More frequent reauthorization of certain high-risk stores is consistent with federal internal control standards, which suggest that agencies should consider the potential for fraud when determining how to respond to fraud risks. 37 Considering the benefits and costs of control activities to address identified risks is a leading practice in GAO’s Fraud Risk Framework. By not assessing the benefits and costs of reauthorizing certain high-risk stores more frequently than other stores, FNS may be missing an opportunity to provide early oversight of risky stores and prevent trafficking. Detecting Retailer Trafficking The steps FNS has taken to improve how it detects retailer trafficking generally align with fraud risk management leading practices for 35 FNS, “What Is FNS Doing to Fight SNAP Fraud.” https://www.fns.usda.gov/fraud/what- fns-doing-fight-snap-fraud. Accessed September 28, 2018. 36 FNS officials told us that they were unable to reauthorize all high-risk stores in 2013, and that they issued a waiver to allow reauthorization of the remaining fiscal year 2013 stores to continue into fiscal year 2014. 37 GAO, Standards for Internal Control in the Federal Government, GAO-14-704G (Washington, D.C.: September 2014). Page 21 GAO-19-167 Retailer Trafficking designing and implementing control activities to detect fraud. 38 For example, FNS’s website shows how to report SNAP fraud, including retailer trafficking, through the USDA OIG’s fraud hotline. According to our Fraud Risk Framework, reporting mechanisms help managers detect instances of potential fraud and can also deter individuals from engaging in fraudulent behavior if they believe the fraud will be discovered and reported. Increasing managers’ and employees’ awareness of potential fraud schemes can also help managers and employees better detect potential fraud. To that end, FNS has developed fraud awareness training for staff in each of the branches in its Retailer Operations Division—the office primarily responsible for oversight of SNAP-authorized retailers. This includes training related to retailer trafficking for new staff and refresher training for experienced staff. Some of the training materials employ identified instances of trafficking to improve future detection and response activities. See figure 7 for photographs from a store investigation that were featured in an April 2017 training session. Figure 7: Food and Nutrition Service (FNS) Photographs from Training Session on Detecting Retailer Trafficking FNS also uses data analytics, another leading practice in our Fraud Risk Framework, to identify potential trafficking and prioritizes its investigative resources to the stores most likely to be trafficking. Specifically, FNS scans about 250 million SNAP transactions per month through its Anti- Fraud Locator using EBT Retailer Transactions (ALERT) system to identify certain patterns indicative of trafficking. ALERT assigns a numeric score to each store based on the likelihood of trafficking. Stores with 38 As noted previously, we did not assess the effectiveness of FNS’s activities in these areas. Page 22 GAO-19-167 Retailer Trafficking scores above a certain threshold are added to FNS’s Watch List, and FNS analysts and investigators prioritize the stores on this list for review based on factors such as average transaction amounts that are excessive for that type of store. 39 In addition, FNS’s analysts conduct their own data mining and review complaints and fraud tips from the OIG’s hotline to add stores to the Watch List. FNS also has explored ways and taken steps to improve its data analytics through internal workgroups and external studies. Using the results of monitoring and evaluations to improve fraud risk management activities is a leading fraud risk management practice. For example, staff in the Retailer Operations Division participate in a workgroup that uses findings from FNS’s trafficking investigations to improve the Division’s detection efforts. This collaborative effort has led to improvements such as using store ZIP codes to compare transactions at stores suspected of trafficking with similar stores nearby. According to FNS, its staff can use this information to substantiate charges against retailers by establishing what typical transaction patterns look like, compared to trafficking patterns, for similar stores. In addition, FNS commissioned studies in fiscal years 2014 and 2015 to evaluate the effectiveness of its data analytics to monitor stores and identify areas for improvement. For example, one of the studies identified and recommended new ways that FNS could analyze SNAP transaction data to detect emerging trafficking schemes—such as indirect trafficking at super stores and supermarkets, where more than 80 percent of SNAP benefits are redeemed. FNS officials reported in August 2018 that they examined the recommendations and implemented those they determined were feasible with current resources and would add value to their efforts. For example, they decided to analyze data over shorter periods of time (monthly instead of a 6-month period) to more quickly identify stores that may be trafficking. Officials also reported that they are continuing to assess the effectiveness of their data analytics. Responding to Retailer FNS’s efforts to respond to retailer trafficking generally align with leading Trafficking practices for fraud risk management. Consistent with our Fraud Risk Framework, FNS has established collaborative relationships with external stakeholders to respond to identified instances of potential fraud. For example, to amplify its own efforts, FNS has agreements (known as state law enforcement bureau, or SLEB, agreements) with 28 states. Through 39 Although FNS analyzes transactions of all store types, FNS does not include large stores (e.g., supermarkets or super stores) on its Watch List. Officials said that these types of stores already have antifraud measures in place to prevent trafficking. Page 23 GAO-19-167 Retailer Trafficking these agreements, FNS allows state and local law enforcement agencies to use SNAP EBT cards in their own undercover investigations of retailers. According to the most recent available FNS data, participating states opened 1,955 cases from fiscal year 2012 to fiscal year 2018 under SLEB agreements. These cases resulted in a total of 139 retailers being permanently disqualified from the program. 40 Within USDA, FNS and the OIG also said they recently updated a memorandum of understanding (MOU) that outlines, among other things, how the two entities will coordinate on retailer trafficking investigations. Under the MOU, FNS investigates retailers with average monthly SNAP redemptions below a certain dollar threshold without first obtaining clearance from the OIG to pursue the case. FNS and OIG officials said that this provision of the MOU allows FNS to more quickly investigate suspicious behavior and pursue administrative action, such as permanent disqualification, against retailers found to be trafficking. Previously, according to OIG officials, FNS needed to clear most cases against retailers suspected of trafficking through the OIG. As we noted in our 2006 report, due to the time it takes to develop an investigation for prosecution and the costs associated with doing so, a natural tension exists between the goal of disqualifying a retailer as quickly as possible to prevent further trafficking and seeking prosecution of the retailer to recover losses and deter other traffickers. 41 The MOU is also designed to strengthen collaboration between FNS and the OIG in identifying the situations that warrant criminal investigations. Since our 2006 report, OIG and FNS both generally increased the number of actions taken against SNAP retailers found to be trafficking. Specifically, the OIG reported an increase in the number of trafficking cases that it successfully referred for federal, state, or local prosecution (see fig. 8). The OIG also reported increases in the number of convictions resulting from its investigations, from 79 in fiscal year 2007 to 311 in fiscal year 2017. 42 40 In addition, SLEB investigations had other outcomes, including temporary disqualifications and warning letters. Some cases are still pending final action. Most of the investigations were closed with no further action (69 percent, or 1,352 cases). 41 GAO-07-53. 42 According to the OIG, these results are recorded as of the fiscal year they are entered into the data system, rather than the year the case was closed. Page 24 GAO-19-167 Retailer Trafficking Figure 8: U.S. Department of Agriculture (USDA) Office of Inspector General (OIG) Retailer Trafficking Investigations Accepted and Declined for Prosecution, Fiscal Years 2007 – 2017 Note: The data represent closed cases with results that the USDA OIG referred for federal, state, or local prosecution. The OIG reported that administrative or other actions were taken in many of the cases that were declined for prosecution. FNS also generally increased the number of retailers sanctioned for trafficking, though few received a monetary penalty. From fiscal year 2007 to fiscal year 2017, the number of permanent disqualifications resulting from FNS’s trafficking investigations nearly doubled (see fig. 9). 43 In lieu of a permanent disqualification, FNS sometimes imposes a monetary penalty on a retailer found to be trafficking. 44 However, FNS imposed few monetary penalties for trafficking in lieu of permanent 43 At the same time, since the number of authorized stores also increased over this period, the number of permanent disqualifications in each year represented less than 1 percent of all authorized stores. 44 Under current regulations, FNS assesses either a civil monetary penalty for trafficking or permanent disqualification, but not both. Page 25 GAO-19-167 Retailer Trafficking disqualification during this period. 45 From fiscal year 2007 to fiscal year 2017, FNS assessed a total of 40 such penalties, totaling $1.5 million (for an average of about $38,000 each). Figure 9: Permanent Disqualifications and Penalties for Supplemental Nutrition Assistance Program (SNAP) Retailer Trafficking, by Type of Case, Fiscal Years 2007 – 2017 Note: The data in this figure are reported as of the year the case was opened. As a result, cases in later years may have final results that were not captured as of March 2018, when the data were generated. Further, FNS officials told us that additional factors may have contributed to a decrease in permanent disqualifications after fiscal year 2015, such as an increase in the number of retailer appeals of FNS’s decisions. The data include the results of FNS cases and do not include permanent disqualifications or penalties resulting from other investigations, including those by the U.S. Department of Agriculture Office of Inspector General or state and local law enforcement entities. In our 2006 report, we found that FNS’s penalties for retailer trafficking may be insufficient to deter traffickers. We noted that trafficking will continue to be lucrative for retailers as long as the potential rewards 45 FNS might assess a civil monetary penalty for trafficking if a store demonstrates that it has an effective compliance policy and program in effect to ensure that the store is operated in a manner consistent with SNAP regulations and policy and meets additional criteria. Officials said that there is a purposefully high bar to obtaining a civil monetary penalty because, under current regulations, retailers can stay in the program and continue to redeem SNAP benefits after paying the penalty. Page 26 GAO-19-167 Retailer Trafficking outweigh the penalties and recipients are willing to exchange their benefits for cash. We recommended that FNS develop a strategy to increase penalties for trafficking. 46 Using the results of evaluations, such as audits, to improve fraud risk management activities is a leading practice in GAO’s Fraud Risk Framework. The Food, Conservation, and Energy Act of 2008 (known as the 2008 Farm Bill) gave USDA authority to impose higher monetary penalties, as well as authority to impose both a monetary penalty and program disqualification on retailers found to have violated relevant law or regulations (such as those found to be trafficking). Although USDA was granted this authority a decade ago, the department has not finalized regulatory changes to strengthen penalties against retailers found to be trafficking. In August 2012, FNS proposed regulatory changes to implement this authority from the 2008 Farm Bill, including assessing a new trafficking penalty in addition to permanent disqualification. 47 The penalty would have been based on the store’s average monthly SNAP redemptions and was intended to recoup government funds diverted from their intended use. 48 In proposing these changes, FNS stated that they were necessary to improve program integrity and deter retailers from committing program violations. FNS also estimated that it would assess an additional $174 million per year in these new trafficking penalties—a significant increase from the amounts FNS currently assesses in penalties for trafficking (less than $100,000 in fiscal year 2017). 49 However, FNS did not finalize this rule, and, as of spring 46 GAO-07-53. FNS stated that it generally agreed with our recommendations, and USDA requested strong penalties to strengthen enforcement, according to a July 2007 report from the House Committee on Agriculture. 47 Supplemental Nutrition Assistance Program: Farm Bill of 2008 Retailer Sanctions, 77 Fed. Reg. 48,461 (Aug. 14, 2012). 48 FNS stated that its proposed rule would have increased other types of penalties to up to $100,000 per violation, including the penalty for transferring ownership of a store after an owner is disqualified from the SNAP program for trafficking or other violations. Specifically, FNS imposes a “transfer of ownership” penalty if the disqualified owner sells his or her store before the expiration of the disqualification period (and in the case of permanent disqualification, the penalty is double that for a 10-year disqualification period). 49 FNS also noted that the majority of retailer penalties are turned over to the Department of the Treasury and never collected. Page 27 GAO-19-167 Retailer Trafficking 2018, the rule was considered “inactive.” 50 At that time, FNS officials told us that they had not finalized the rule because other rulemaking had taken priority in the intervening 6 years. 51 More recently, in August 2018, FNS officials told us that they plan to revise the previously proposed rule to increase penalties and submit it for the spring 2019 regulatory agenda. 52 In November 2018, FNS officials indicated that they were beginning to draft the proposed rule but could not provide us with documentation of this effort because the regulatory action was still pending. Increasing penalties for retailer trafficking would serve as an important tool to deter trafficking and safeguard federal funds. FNS Has Not Established FNS measures the effectiveness of many of its trafficking detection and Performance Measures to response activities, but lacks measures to evaluate its prevention activities. Measuring outputs, outcomes, and progress toward the Assess its Retailer achievement of fraud risk objectives is a leading practice in our Fraud Trafficking Prevention Risk Framework. At the agency level, FNS has a priority plan for fiscal Activities 50 Several other SNAP program integrity-related changes have also not yet moved forward. For example, in response to a 2013 OIG recommendation, FNS agreed to propose a rule to define “benefit” so that store owners could demonstrate that they did not personally benefit from the trafficking that occurred at their store and, therefore, be eligible for a civil monetary penalty in lieu of permanent disqualification. FNS submitted a regulatory work plan and secured a regulation identification number in fall 2015 but the regulatory action was considered inactive as of spring 2018. In addition, in response to another OIG recommendation, FNS agreed to submit a regulatory work plan in fiscal year 2016 to revise regulations to permanently disqualify retail store owners who own multiple stores at all authorized locations if the retailer was found to be trafficking at one and did not meet the criteria for a trafficking civil monetary penalty. Current policy allows owners found to be trafficking in one location to continue to accept SNAP benefits in other locations. U.S. Department of Agriculture, Office of Inspector General, FNS: Controls for Authorizing Supplemental Nutrition Assistance Program Retailers, Audit Report 27601- 0001-31 (Washington, D.C.: July 2013). In August 2018, FNS officials told us that they were planning to include both provisions in a revised retailer sanctions proposed rule, and that they would include these changes in the spring 2019 regulatory agenda. 51 We discussed the status of this proposed rule in our May 2018 testimony before the Subcommittees on Healthcare, Benefits and Administrative Rules and Intergovernmental Affairs of the House Committee on Oversight and Government Reform. GAO, Supplemental Nutrition Assistance Program: Observations on Employment and Training Programs and Efforts to Address Program Integrity Issues, GAO-18-504T (Washington, D.C.: May 9, 2018). 52 The Unified Agenda of Federal Regulatory and Deregulatory Actions is published semiannually and generally includes regulatory actions, such as notices of proposed rulemaking and final rules, that executive agencies plan to issue within the next 12 months. Page 28 GAO-19-167 Retailer Trafficking year 2018 that includes a goal of reducing the SNAP trafficking rate through retailer- and client-focused activities. At the program level, FNS’s Retailer Operations Division has an internal scorecard that tracks performance measures related to retailer oversight activities, but none of these focuses on prevention of trafficking. For example, the scorecard measures the outputs and outcomes of activities designed to detect and respond to trafficking, such as the total number of sanctions implemented against retailers and the percentage of undercover investigations that result in a permanent disqualification. However, the scorecard does not have any measures related to preventing trafficking through the retailer authorization process—a key area for prevention activities. The scorecard includes one output measure related to this process, but the measure (the percentage of retailer authorization requests processed within 45 days) focuses on how quickly retailers gain access to the program, rather than preventing trafficking. Although FNS officials have acknowledged that their program compliance efforts begin with the retailer authorization process, they said that they had not considered establishing measures related to preventing trafficking. They added that their supervisory review process may help ensure that staff who process retailer applications in the Retailer Operations Branch do not overlook evidence of potential fraud, but this review includes a small sample of approved store applications (typically 5 cases per staff member monthly). 53 Although FNS has not established measures to assess its trafficking prevention activities, the agency has data that it could leverage for this purpose. For example, FNS collects data on the number of applications that were denied because FNS found that the retailer lacked business integrity, such as applicants previously found to be trafficking or with ties to a prior owner who had trafficked. 54 Such data could be used to develop measures related to the number and percentage of retailer applications denied for business integrity. FNS officials acknowledged that these data could be used to develop performance measures for its trafficking prevention activities. Establishing such measures would enable FNS to more fully assess the effectiveness of its retailer oversight activities and 53 According to officials, FNS’s supervisory review process includes a sample of approvals and 100 percent of denials of authorization. 54 FNS regulations lay out a number of criteria for a store to be found lacking the necessary business integrity and reputation to further the purposes of the program. See 7 C.F.R. § 278.1(k)(3). Page 29 GAO-19-167 Retailer Trafficking better balance retailer access to the program with preventing retailer fraud. FNS must continue to balance its goal of program integrity with its Conclusions mission to provide nutrition assistance to millions of low-income households. During a period in which SNAP retailer participation has markedly increased, FNS has made progress in addressing SNAP retailer trafficking by identifying high-risk stores and increasing the number of stores disqualified for trafficking. It is critical that FNS maintain progress and momentum in these areas, particularly since FNS’s own data suggest that trafficking is on the rise. To its credit, FNS has already evaluated some ways to improve how the agency measures and addresses retailer trafficking, yet, at the same time, the agency has missed opportunities to strengthen these areas. For example, since FNS has not taken steps to clarify and improve its retailer trafficking estimates—one of the only available SNAP fraud measures— questions remain regarding the accuracy of the estimates and the extent of fraud in SNAP. In addition, prevention and early detection of retailer trafficking are particularly important and deserve continued attention, especially since retailers can quickly ramp up the amount they redeem in federal SNAP benefits, potentially by trafficking. However, because FNS is reauthorizing all stores once every 5 years, the agency may be missing an opportunity to prevent trafficking through more frequent oversight of risky stores. Further, until FNS strengthens its response to trafficking by increasing penalties, the agency will continue to miss an opportunity to improve program integrity and deter retailers from committing program violations. Finally, FNS directs a significant amount of staff resources to authorizing and monitoring retailers who participate in SNAP. Ensuring that those staff understand the importance of addressing fraud is key for program integrity. FNS has taken steps to make that clear through the inclusion of relevant performance measures for the branches responsible for fraud detection and response, yet the agency has not developed such measures for its trafficking prevention activities. Until FNS establishes performance measures for these activities, it will be unable to fully assess the effectiveness of its overall efforts to address retailer trafficking. In addition, such measures would assist FNS in balancing its efforts to ensure retailer access with those to prevent retailer fraud. Page 30 GAO-19-167 Retailer Trafficking We are making the following five recommendations to FNS: Recommendations for Executive Action The Administrator of FNS should present the uncertainty around its retailer trafficking estimates in future reports by, for example, including the full range of the estimates in the report body and executive summary. (Recommendation 1) The Administrator of FNS should continue efforts to improve the agency’s retailer trafficking estimates by evaluating (1) whether the factors used to identify stores for possible investigation could help address the bias in its sample, and (2) the accuracy of its assumption of the percentage of SNAP benefits that are trafficked by different types of stores. (Recommendation 2) The Administrator of FNS should assess the benefits and costs of reauthorizing a sample of high-risk stores more frequently than other stores, use the assessment to determine the appropriate scope and time frames for reauthorizing high-risk stores moving forward, and document this decision in policy and on its website. (Recommendation 3) The Administrator of FNS should move forward with plans to increase penalties for retailer trafficking. (Recommendation 4) The Administrator of FNS should establish performance measures for its trafficking prevention activities. (Recommendation 5) We provided a draft of this report to USDA for review and comment. On Agency Comments December 3, 2018, the Directors of the Retailer Policy & Management and Our Evaluation Division and the Retailer Operations Division of FNS provided us with the agency’s oral comments. FNS officials told us that they generally agreed with the recommendations in the report. Officials also provided technical comments, which we incorporated as appropriate. Regarding the recommendation to present the uncertainty around the retailer trafficking estimates, FNS officials told us that they plan to include the estimate intervals and results of sensitivity analyses in the body of their next report, rather than in appendices. This is the information we used to determine the range around the trafficking estimates. Making this change would address our recommendation, as we continue to believe that reporting the level of uncertainty around each estimate would increase transparency and provide Congress and the public with better information on the extent of fraud in SNAP. Page 31 GAO-19-167 Retailer Trafficking In addition, regarding the recommendation to assess the benefits and costs of reauthorizing a sample of high-risk retailers more frequently, FNS officials noted that while reauthorizations currently occur at least once every 5 years, monitoring for potential violations occurs on an ongoing basis regardless of risk level. Low-, medium-, and high-risk stores are continually scanned by FNS’s ALERT system. FNS officials added that, in fiscal year 2017, FNS imposed sanctions (e.g., fines or temporary disqualifications) on 862 stores found to be violating program rules, and disqualified permanently 1,661 stores for trafficking SNAP benefits or falsifying an application. FNS officials noted that this is a 26 percent increase in the number of stores sanctioned, compared to fiscal year 2013. We agree that ongoing monitoring is important, and we discussed these and other FNS efforts to detect and respond to retailer trafficking in our report. We nevertheless believe, and FNS officials agreed, that assessing the value of earlier oversight of risky stores through the reauthorization process is warranted, and could enhance efforts to prevent trafficking. As agreed with your offices, unless you publicly announce the contents of this report earlier, we plan no further distribution until 30 days from the report date. At that time, we will send copies to the Secretary of the USDA, congressional committees, and other interested parties. In addition, this report will be available at no charge on the GAO website at www.gao.gov. If you or your staff have any questions about this report, please contact me at (202) 512-7215 or email@example.com. Contact points for our Offices of Congressional Relations and Public Affairs may be found on the last page of this report. GAO staff who made key contributions to this report are listed in appendix I. Kathryn A. Larin Director, Education, Workforce, and Income Security Issues Page 32 GAO-19-167 Retailer Trafficking List of Requesters The Honorable Trey Gowdy Chairman Committee on Oversight and Government Reform House of Representatives The Honorable Mark Meadows Chairman Subcommittee on Government Operations Committee on Oversight and Government Reform House of Representatives The Honorable Jim Jordan Chairman Subcommittee on Healthcare, Benefits, and Administrative Rules Committee on Oversight and Government Reform House of Representatives The Honorable Gary J. Palmer Chairman Subcommittee on Intergovernmental Affairs Committee on Oversight and Government Reform House of Representatives Page 33 GAO-19-167 Retailer Trafficking Appendix I: GAO Contact and Staff Appendix I: GAO Contact and Staff Acknowledgments Acknowledgments Kathy Larin, (202) 512-7215 or firstname.lastname@example.org. GAO Contact In addition to the contact named above, Rachel Frisk (Assistant Director), Staff Rachael Chamberlin (Analyst-in-Charge), and Swati Deo made significant Acknowledgments contributions to this report. Also contributing to this report were James Bennett, Thomas Cook, Alex Galuten, Lara Laufer, Olivia Lopez, Jean McSween, Jessica Orr, Philip Reiff, Almeta Spencer, Jeff Tessin, Matthew Valenta, and Erin Villas. (102523) Page 34 GAO-19-167 Retailer Trafficking The Government Accountability Office, the audit, evaluation, and investigative GAO’s Mission arm of Congress, exists to support Congress in meeting its constitutional responsibilities and to help improve the performance and accountability of the federal government for the American people. GAO examines the use of public funds; evaluates federal programs and policies; and provides analyses, recommendations, and other assistance to help Congress make informed oversight, policy, and funding decisions. GAO’s commitment to good government is reflected in its core values of accountability, integrity, and reliability. The fastest and easiest way to obtain copies of GAO documents at no cost is Obtaining Copies of through GAO’s website (https://www.gao.gov). Each weekday afternoon, GAO GAO Reports and posts on its website newly released reports, testimony, and correspondence. To have GAO e-mail you a list of newly posted products, go to https://www.gao.gov Testimony and select “E-mail Updates.” Order by Phone The price of each GAO publication reflects GAO’s actual cost of production and distribution and depends on the number of pages in the publication and whether the publication is printed in color or black and white. Pricing and ordering information is posted on GAO’s website, https://www.gao.gov/ordering.htm. Place orders by calling (202) 512-6000, toll free (866) 801-7077, or TDD (202) 512-2537. Orders may be paid for using American Express, Discover Card, MasterCard, Visa, check, or money order. Call for additional information. Connect with GAO on Facebook, Flickr, Twitter, and YouTube. Connect with GAO Subscribe to our RSS Feeds or E-mail Updates. Listen to our Podcasts. Visit GAO on the web at https://www.gao.gov. Contact: To Report Fraud, Website: https://www.gao.gov/fraudnet/fraudnet.htm Waste, and Abuse in Automated answering system: (800) 424-5454 or (202) 512-7700 Federal Programs Orice Williams Brown, Managing Director, WilliamsO@gao.gov, (202) 512-4400, Congressional U.S. Government Accountability Office, 441 G Street NW, Room 7125, Relations Washington, DC 20548 Chuck Young, Managing Director, email@example.com, (202) 512-4800 Public Affairs U.S. Government Accountability Office, 441 G Street NW, Room 7149 Washington, DC 20548 James-Christian Blockwood, Managing Director, firstname.lastname@example.org, (202) 512-4707 Strategic Planning and U.S. Government Accountability Office, 441 G Street NW, Room 7814, External Liaison Washington, DC 20548 Please Print on Recycled Paper.
Supplemental Nutrition Assistance Program: Actions Needed to Better Measure and Address Retailer Trafficking
Published by the Government Accountability Office on 2019-01-15.
Below is a raw (and likely hideous) rendition of the original report. (PDF)