The Distance Between the Statute and the Supermarket

A statute does not hand food to anybody. The Food and Nutrition Act, codified in title 7 of the United States Code, declares in Washington that a household meeting the statutory conditions holds a claim to nutrition assistance, sets the terms on which that claim arises, and appropriates federal money to pay for the benefits that follow from it. Between that declaration and a completed purchase at a supermarket checkout stand a long chain of institutions, decisions, documents, computer systems, and deadlines, almost none of which the statutory text describes in operational detail and almost all of which determine whether the claim written into federal law becomes food in a household kitchen. The Supplemental Nutrition Assistance Program is therefore two things at once. In legal form, the program is a single national entitlement, uniform in its core eligibility rules and its benefit formula. In lived operation, the program is fifty three separate administrative systems, each translating the same federal text through a different agency structure, a different eligibility computer platform, a different network of local offices, and a different set of choices about staffing, verification, and service. Administration is not a footnote to the program. Administration is the mechanism by which the program either reaches a household or fails to reach that household, and the distance between the statute and the supermarket is the distance that administration has to cross.

The scale of that crossing is large enough to make small administrative failures visible in national totals. According to data collected by the Food and Nutrition Service of the United States Department of Agriculture and reported in the Economic Research Service Food Assistance Landscape series, an average of 35.7 million people participated in the program in each month of fiscal 2019, and federal spending on the program totaled 60.4 billion dollars in that fiscal year. In fiscal 2018, the same data series recorded average monthly participation of 40.3 million people and federal spending of 65.0 billion dollars. Both figures stood well below the program peak recorded in fiscal 2013, when average monthly participation reached 47.6 million people and federal spending reached 79.9 billion dollars, totals that the Economic Research Service identified as the historical high for the program in the Food Assistance Landscape series. Behind each annual total stood tens of millions of individual administrative acts performed under state and county authority: applications taken, interviews conducted, pay records verified, benefit amounts computed, electronic benefit transfer accounts loaded, changes processed, and certifications renewed or closed. A program operating at that scale does not experience administration as overhead surrounding the real work. At that scale, administration is the real work, repeated monthly for every participating household, and the quality of that work decides whether national totals reflect the population Congress made eligible or only the portion of that population that a particular state system managed to process accurately and on time.

The institutional design that produces those totals rests on a deliberate division of labor between the federal government and the states. The federal side of the division holds the money for benefits and the authority over rules. The Food and Nutrition Service, the agency within the Department of Agriculture charged with the program at the federal level, issues the regulations that implement the statute, approves the state plans of operation under which each state agency runs the program, grants or denies waivers of federal requirements where the statute and regulations permit waiver authority, authorizes the retail food stores at which benefits may be redeemed, and operates the oversight systems that measure state performance. The state side of the division holds the contact with the household. No federal employee interviews an applicant in a county office. No federal employee decides whether a particular household has documented its income, its household composition, and its deductible expenses to the standard the regulations require. No federal employee issues the monthly benefit to the household electronic benefit transfer account or answers the telephone call when a benefit arrives late or a case closes after a missed recertification deadline. State agencies perform each of those functions, and county agencies perform them under state supervision in states with county administered systems. The applicant encounters the program as a state or county institution with a local office, a local telephone number, and a case record held in a state system. The entitlement in the file may be federal in origin, but the experience of claiming it is state and local from the first form to the final notice.

That division would matter less if it were only a division of tasks. The division also separates the funding of benefits from the funding of the machinery that delivers benefits, and that separation of funding streams is the central mechanism around which the administration of the program turns. Benefits are wholly federally funded. Every dollar of benefit that a state agency correctly issues to an eligible household comes from the federal Treasury, and a state that enrolls more eligible households draws down more federal benefit money without spending state funds on the benefits themselves. Administrative costs follow a different rule. The federal government reimburses states for roughly half of the allowable costs of administering the program, and the state bears the remaining share from state funds, subject to the detailed cost allocation rules that govern which expenses qualify for federal reimbursement. The two halves of the program therefore sit on different ledgers in a state budget. The benefit ledger belongs to Washington. The administrative ledger is shared, and the state portion of that ledger competes in a state legislature against education, transportation, public health, and every other claim on state revenue. The size of the administrative appropriation a legislature grants sets the outer boundary on how many caseworkers the agency employs, how many offices the agency keeps open, and how much verification the agency can afford to perform.

From that funding structure flows an asymmetry that shapes state behavior more reliably than any instruction in a federal manual. A state cannot save its own money by paying out fewer benefits. A decision that removes an eligible household from the rolls, or that deters an eligible household from applying, reduces federal spending and reduces the food purchasing power of a household inside the state, but the saving accrues to the federal Treasury rather than to the state budget that made the decision. The benefit dollars were never state dollars, so denying them frees no state funds for any other purpose. The only route by which a state saves its own money inside the program runs through the administrative ledger. A state saves state funds by employing fewer eligibility workers, by closing or consolidating local offices, by shortening office hours, by operating a smaller call center with longer waits, by postponing replacement of an aging eligibility computer system, or by reducing the staff time devoted to processing reported changes and renewals. Each of those economies is real from the standpoint of a state budget officer, because the state pays roughly half of every administrative dollar and pockets roughly half of every administrative dollar cut. The program therefore presents every state with the same tilted set of incentives: generosity or restrictiveness in benefit delivery costs the state nothing and saves the state nothing, while the capacity to deliver benefits accurately and accessibly costs the state money in every fiscal year, in direct competition with other state priorities.

The consequences of that tilt appear in the operational details that households experience as the character of the program itself. Staffing levels decide how large a caseload each eligibility worker carries, and caseload size decides how much time a worker can give to a complicated household with fluctuating earnings, shared housing, or income from several part time jobs. Office hours and office locations decide whether a working applicant can complete an interview without losing wages to attend in person during the working day. Call center capacity decides whether a household reporting a change of address or a lost job reaches a person able to record the change, or abandons the attempt after a long wait and leaves the case record wrong through no defect in household compliance. Verification intensity, the degree to which a state demands documentary proof of facts an applicant reports, sits at the intersection of the funding asymmetry and the integrity pressures described later in this article: documentation demands cost staff time to impose and to process, yet insufficient verification exposes the state to measured error, while excessive documentation demands shift the cost of proof onto applicants in time, travel, and repeated submissions, and push eligible households out of the program for procedural reasons unrelated to eligibility in substance. Computer systems bind all of those choices together. A state eligibility platform decides how notices are generated, how deadlines are tracked, and how errors surface or remain hidden, and a platform purchased or deferred under administrative budget pressure can govern household outcomes for a decade. None of those operational choices appears in the statutory definition of eligibility. All of them decide, in practice, who receives the entitlement the statute defines.

Against that background, the article adopts a distinction that the remainder of the analysis treats as foundational, stated as a namable claim: error is not fraud. The claim sounds elementary, and the persistence with which public argument about the program collapses the distinction gives the claim its analytical force. Error, as the program measures error, means a benefit amount that differs from the amount the rules required for that household in that month. An error may overpay a household or underpay a household. An error may originate in an agency act, as when a worker misapplies a deduction rule, enters income incorrectly, or acts late on a reported change, or in a household act, as when a participant misunderstands a reporting duty or reports a change after the required date. Error in that sense records an inaccurate outcome without any finding about intent, and a large share of measured error involves households that were eligible for a substantial benefit, paid in the wrong amount because a complex case moved through an overburdened system imperfectly. Fraud is a different category with a different legal character. Fraud involves an intentional violation, established through investigation and adjudication rather than through statistical measurement, and the program treats the diversion of benefits in exchange for cash, the practice called trafficking, as a separate measure from payment accuracy altogether. The Food and Nutrition Service trafficking estimates placed trafficking at about 1.0 percent of benefits for the period from 2006 to 2008 in the 2011 study by Mantovani and Wilson, and at 1.3 percent of benefits for the period from 2009 to 2011 in the August 2013 study by Mantovani, Williams, and Pflieger. Those figures measure a defined form of diversion during defined periods. Neither figure measures the payment error rate, and the payment error rate measures neither figure. Argument that cites an error statistic as evidence of fraud, or a trafficking statistic as evidence of general inaccuracy, borrows the authority of one measurement to make a claim about a phenomenon that measurement never examined.

The measurement system behind the error side of that distinction is the quality control system, and the quality control system exerts a force on state behavior that exceeds the force of any single eligibility rule in the statute. Under quality control, a sample of cases from each state undergoes review for the accuracy of the eligibility determination and the benefit amount, federal reviewers reexamine a portion of that work, and the results produce a payment error rate for each state and for the nation. The combined national payment error rate reached a record low of 3.20 percent in fiscal 2013, rose to 3.66 percent in fiscal 2014, went unpublished at the national level for fiscal 2015 and fiscal 2016 while the measurement system faced questions about state review practices, and resumed publication at 6.30 percent for fiscal 2017 and 6.80 percent for fiscal 2018, as recorded in the consistency figures carried through the Food and Nutrition Service quality control record. A rate of that kind is a statistic, but state agencies experience the rate as an incentive structure with fiscal and reputational consequences attached. Error rates feed federal fiscal sanctions and liability calculations, corrective action demands, and public comparison among states, and an agency director answers for the published rate in legislative hearings in a way that no director answers for an eligible household that never applied. The predictable response is defensive administration organized around the measured variable. States invest verification effort where quality control samples look hardest, manage cases with an eye to the review record a file creates, and weigh every policy option partly by the effect the option carries for the error rate, including options that would simplify enrollment for eligible households. A measurement system built to protect accuracy thereby becomes a governing force in its own right, pressing states toward demonstrable correctness in sampled cases with an intensity that no equivalent system presses states toward reaching eligible households outside the rolls. Error is counted, published, and priced. Nonparticipation by eligible households, for most of the program history covered in this article, was counted in separate participation rate research and assigned no fiscal price in the relationship between a state and the federal government.

The result is a structural gap between statutory eligibility and delivered benefits, visible from two directions that the remainder of the article holds together rather than treating as rival accounts. From the integrity direction, the gap appears as inaccuracy inside the caseload: households paid too much or too little, cases carried past a change in circumstances, and a state apparatus straining to keep a measured error rate down under caseload pressure and administrative budget limits. From the access direction, the gap appears as eligible households outside the caseload entirely, or pushed out at recertification and application stages by procedural demands, and that second appearance supplies the counter-reading that this article carries alongside the integrity account. The Food and Nutrition Service participation rate research, published in the Trends in Supplemental Nutrition Assistance Program Participation Rates series in a report dated July 2018, estimated that the program served 85 percent of all eligible individuals in fiscal 2016. The same research estimated participation at 45 percent among eligible older adults, defined in that series as individuals age 60 or older, and at 75 percent among eligible individuals in households with earnings in fiscal 2016. An overall rate of 85 percent means that roughly one eligible individual in seven received no benefit in an average month of fiscal 2016 despite meeting the federal eligibility rules, and the distribution of that shortfall was not random. Older adults participated at slightly more than half the overall rate. Working households, whose earnings placed them inside eligibility while complicating application, interview, and reporting duties around employment schedules, participated below the overall rate as well. The explanations that the participation research and the access literature associate with that pattern include lack of awareness of eligibility, the stigma attached to seeking means tested assistance, and the burden of the application process itself, from documentation gathering to interview scheduling to periodic recertification. Each explanation points back to administration. Awareness depends on outreach that administrative budgets fund or neglect. Burden is the accumulated weight of the procedural choices described throughout this section. Stigma grows or shrinks partly with the character of the offices a state maintains.

Holding both directions of the gap in view at once disciplines the conclusions available about the program. A state may post a low payment error rate while serving a sharply incomplete share of the eligible population, and the published accuracy figure alone cannot distinguish administrative excellence from administrative narrowness, because the error rate samples cases inside the program and says nothing about eligible households that the system never reached. A state may enroll a high share of eligible residents while carrying a higher measured error rate, in part because broader reach brings harder cases and larger processing volumes into the system. Neither pattern excuses inaccuracy, and neither pattern converts nonparticipation into a virtue. The patterns instead reveal that administration mediates between two statutory commitments that coexist in the Food and Nutrition Act: the commitment to pay the right household the right amount, and the commitment to pay every household that the statute makes eligible. The federal and state funding split, the asymmetry of state savings, and the quality control incentive structure all push administrative energy toward the first commitment, because the first commitment is measured case by case and priced in federal fiscal consequences. The second commitment depends on capacity, outreach, and procedural design that cost state administrative dollars and generate no matching fiscal reward of comparable force. The sections that follow trace that tension through the machinery of the program in sequence: the state plan and waiver framework, the application and certification process, the issuance system that moves benefits onto electronic benefit transfer cards, the reporting and recertification cycle, the quality control and fiscal accountability system, and the hearing and restoration procedures available when the machinery produces a disputed result. Each stage repeats the pattern established in this opening section. Federal law writes the entitlement in uniform national terms. State and county administration, funded on a divided ledger and governed by incentives that reward measured accuracy over unmeasured reach, decides how much of that entitlement arrives at the supermarket.

Two Governments in One Program

The Supplemental Nutrition Assistance Program runs on a deliberate division of labor between two levels of government, and nearly every feature of its administration follows from that division. Congress wrote the governing statute, the Food and Nutrition Act, and placed responsibility for the program in the Department of Agriculture. The department in turn assigned the program to its Food and Nutrition Service, the agency that writes the rules, pays the benefits, and oversees performance. The business of deciding who receives help, however, takes place somewhere else entirely, in state welfare offices, county buildings, call centers, and contracted processing sites across the country. No federal employee decides whether a particular household in Ohio or Oregon qualifies for an allotment, and no state legislature sets the national eligibility formula or the benefit schedule. Authority over the rulebook sits in Washington, while authority over the individual case sits in the states, and the program that households experience is produced where those two authorities meet.

On the federal side, the chain of authority runs from the Secretary of Agriculture through the Under Secretary for Food, Nutrition, and Consumer Services to the Administrator of the Food and Nutrition Service, where a dedicated program office develops policy, drafts regulations, issues guidance, and manages national systems. Budget documents for fiscal 2015 noted that the agency’s staff amounted to less than two percent of the total employment within the Department of Agriculture, a small structure charged with overseeing benefit spending that flowed through tens of thousands of state and local workers. Unable to operate the program directly, the Food and Nutrition Service governs by written rule, by funding condition, and by review, setting the terms on which fifty three separate administering agencies carry out a single national program.

The rulebook itself occupies a large block of title 7 of the Code of Federal Regulations, and its table of contents reads as a map of the federal state relationship. Section 271.1 of title 7 describes the scope of the parts that follow. Part 271 supplies general information and definitions. Part 272 sets the requirements for participating state agencies. Part 273 describes the eligibility criteria that state agencies apply and the processing standards for certification. Part 274 governs issuance of benefits. Part 275 establishes the performance reporting system, including management evaluations and quality control reviews. Part 276 establishes state agency liability and federal sanctions. Part 277 outlines procedures for payment of state administrative costs. Part 278 delineates the terms for participation by retail food stores and other outlets that redeem benefits, and part 279 establishes administrative and judicial review for retailers. Later parts cover emergency allotments, designation of Indian tribes as state agencies, and demonstration projects.

Rulemaking gives that structure its legal force. When the Food and Nutrition Service changes a regulation, it acts under the Food and Nutrition Act through the standard federal process of proposed and final rules published in the Federal Register, and the resulting text binds every state agency as a condition of participation. Beneath the regulations sits a second layer of federal direction in the form of policy memoranda, handbooks, and guidance documents that interpret the rules for state administrators. The arrangement concentrates interpretive authority in Washington. Two states reading the same regulation may reach different understandings of a verification question or a reporting rule, but the understanding that prevails is the one the Food and Nutrition Service issues, because the service controls plan approval, funding, and the findings of federal reviews. State flexibility inside the program is therefore real but delegated, exercised within boundaries that federal officials draw.

Between the national office and the states stands the service’s field structure. Food and Nutrition Service budget documents for fiscal 2015 described program operations managed through seven regional offices, each directed by a regional administrator, supplemented by field offices and satellite locations. The regional offices carry the working relationship with state agencies. Regional staff receive and review the components of state plans of operation, process waiver requests, conduct onsite management reviews of state operations, deliver technical assistance, and transmit national policy to the agencies in their territories. Identical regulations reach a large state agency in the Western region and a small one in the Northeast region through different regional offices, and the consistency of the program depends in part on how evenly those offices interpret and enforce the same text.

Federal money follows federal rules. Under the Food and Nutrition Act as it stood on December 1, 2019, the federal government paid the full cost of benefits issued to households, matched against shared funding for administration. Section 16(a) of the act authorized federal payment of fifty percent of state administrative costs, the share that applied to the ordinary work of taking applications, determining eligibility, and issuing benefits. Associate Administrator Jessica Shahin summarized the division in testimony before the House Committee on Agriculture on July 6, 2016: the department established the rules, monitored state administration, paid the full cost of benefits, and paid half of state administrative expenses. The funding split is the program’s central fiscal fact. The federal treasury bears the entire risk of benefit spending, which rises automatically when caseloads grow, while the cost of the labor that decides those cases is divided evenly between the two levels of government.

The fifty percent figure describes the general rule rather than every function. Employment and training services operated under a different structure set out in section 16(h) of the act. States received federal grants that paid one hundred percent of approved employment and training administrative costs up to an allocated amount distributed by formula, and a state that spent beyond its grant allocation could claim federal reimbursement of fifty percent of the additional allowable costs, including reimbursements owed to participants for expenses such as transportation and dependent care. Administration on Indian reservations followed another exception. Under section 16(a) of the act and section 281.9 of title 7, the Food and Nutrition Service could pay seventy five percent of approved administrative costs for a state agency or an Indian tribal organization acting as a state agency in administering the program on a reservation. Where Congress wanted to encourage a particular activity, or where the administering government had a limited tax base, the federal share rose above the standard half, and the reimbursement rate became an instrument of policy.

On the state side, the unit of administration is the state agency, a single agency designated under state law and accepted by the Food and Nutrition Service as the body answerable for the program within that jurisdiction. The Congressional Research Service counted fifty three state agencies operating the program, comprising the fifty states, the District of Columbia, Guam, and the United States Virgin Islands. In most states the designated agency is the state’s human services or social services department, often the same department that administers cash assistance and Medicaid eligibility, which allows a household to apply for several programs through one office. The state agency answers to the Food and Nutrition Service for the accuracy of eligibility decisions, the timeliness of application processing, the conduct of fair hearings, and the integrity of issuance, whether the work is performed by state employees in a central office, by county staff in a local office, or by a private firm under contract.

The daily work of the state agency is the individual case. State agency staff, or the county and contracted staff acting in the agency’s name, accept applications in person, by mail, by telephone, and through online portals; conduct the certification interview; verify income, household composition, and deductible expenses against federal standards; compute the allotment from the federal benefit formula; and set the certification period after which the household must reapply. The same agency handles reported changes, issues notices of adverse action, conducts fair hearings for households that contest a decision, pursues claims after an overissuance, and investigates suspected participant fraud. Issuance moved from paper coupons to electronic benefit transfer, and nationwide electronic issuance was completed in June 2004 under the mandate set by the welfare statute of 1996. Under part 274 of title 7, state agencies issue benefits through electronic benefit transfer systems that credit a household’s account on a scheduled date each month, and households draw on the account with a card at authorized retailers. The state agency selects and contracts with the processor that operates its issuance system, answers for the system’s accuracy, and reconciles issuance records against the federal accounts from which benefits are drawn.

In a minority of states, the state agency supervises rather than performs the casework. A 2018 study by the Center for Law and Social Policy, titled Ten Degrees of Decentralization, examined the ten states in which administration was county run under state supervision: California, Colorado, Minnesota, New Jersey, New York, North Carolina, North Dakota, Ohio, Virginia, and Wisconsin. The National Association of Counties reported the same count of ten in its program materials. In those states, application processing and eligibility determinations are conducted by county employees in county run local offices, and counties contribute local funds toward the nonfederal share of administrative costs in most of them. The single unifying characteristic identified by the Center for Law and Social Policy study was county controlled hiring. Beyond that point the ten states diverge, delegating different degrees of control over workflow, technology, and office organization to their counties. Federal rules bind the state agency, the state agency binds its counties, and a household’s experience of a national program can vary between counties inside the same state.

The instruments that bind the levels together begin with the state plan of operation, required by section 272.2 of title 7. No state agency operates the program on the strength of the statute alone. Each agency maintains a plan whose basic components the regulation names as the federal state agreement, the budget projection statement, and the program activity statement, together with attachments that include an organizational outline, agreements with other agencies that perform program functions, and a quality control sampling plan. The federal state agreement is signed by the governor or an authorized designee and remains in effect until terminated. The budget projection statement is submitted annually no later than August 15 and projects administrative costs for the coming federal fiscal year, while the program activity statement is submitted quarterly unless the Food and Nutrition Service directs otherwise. Each component goes to the appropriate regional office for approval, and the regulation gives the service thirty days to approve, deny, or request additional information before a submitted document stands approved. The plan converts the obligations of the statute into a specific, examinable description of how one agency will organize offices, staff functions, and spend money in the year ahead, and federal approval of the plan is the threshold act that keeps a state inside the program.

Within the approved plan, state agencies exercise discretion through two distinct channels that are often confused. The first channel is the state option. The statute and regulations build choices into the program itself, on matters such as reporting requirements for households with earnings, transitional benefits for families leaving cash assistance, and methods of treating utility costs in the benefit calculation, and the Food and Nutrition Service periodically published a State Options Report cataloging which options each of the fifty three agencies had adopted. An option requires no special permission, because the federal rulebook itself offers the choice, although the agency records its selection in its plan and applies the choice consistently. The second channel is the waiver. Under section 272.3(c) of title 7, the Food and Nutrition Service may authorize a state agency to deviate from a specific regulatory provision where the provision cannot be implemented because of extraordinary temporary situations, where the service determines that the waiver would produce more effective and efficient administration, or where unique geographic or climatic conditions inside a state preclude effective implementation. A review of waiver processing by the Department of Agriculture’s Office of Inspector General described administrative waivers under that section as the type most commonly requested by states. Waivers are time limited, rest on federal discretion, and reach only regulatory provisions; no waiver suspends the statute itself. Demonstration projects, governed by part 282 of title 7, form a narrower third channel for testing changes to program design on an evaluated basis.

Oversight flows downward through two review systems under part 275 of title 7, which establishes the performance reporting system. The first system is the management evaluation. State agencies divide their operations into management units, geographic or organizational subdivisions that the service must approve under section 275.6 of title 7, and each unit receives management evaluation reviews on a cycle that reflects its size. State reviewers examine local operations against program requirements, identify deficiencies, and require corrective action, while federal reviewers conduct their own management evaluations of state operations. The second system is quality control. Under section 16 of the Food and Nutrition Act and subpart C of part 275, each state agency draws a statistically valid sample of its caseload each year, covering both participating households, called active cases, and households whose participation was denied, suspended, or terminated, called negative cases. State reviewers reconstruct each sampled case against program standards to determine whether the household was eligible and received the correct allotment, and the Food and Nutrition Service re-reviews a federal subsample to validate the state findings. The regulations support that process with an arbitration procedure for disagreements between state and federal reviewers and a good cause procedure for cases in which an unusual event outside the agency’s control distorted its results. The product of the system is the state payment error rate, the measure on which federal judgments about state administrative performance principally rest.

The distinction embedded in that measure deserves emphasis, because public argument about the program so often loses it. Error is not fraud. The payment error rate generated by quality control reviews measures administrative accuracy in both directions, counting benefits issued above the correct amount and benefits issued below it, and it captures mistakes by caseworkers and by households alike, from a misreported wage to a missed deduction. The rate says nothing direct about trafficking, the exchange of benefits for cash, which federal studies measured separately. A state can post a high payment error rate with very little intentional abuse anywhere in its caseload, and the fiscal consequences that follow from the rate fall on the state agency rather than on any individual wrongdoer. Those consequences give the oversight system its force. Part 276 of title 7 establishes state agency liability for payment errors and sets out federal sanctions, part 277 governs how administrative funds are paid and how the service may disallow costs that fall outside approved purposes, and the quality control provisions tie liability to error performance measured against national standards, subject to good cause relief. Alongside liability, section 16 of the act provided for performance bonuses that rewarded agencies for strong results on measures such as payment accuracy and application timeliness. A state agency therefore administers a federal program with its own budget exposed at several points, which is precisely the incentive structure the design intends.

One major federal function sits inside the state administered program without passing through state hands at all: the authorization and policing of the retailers at which benefits are spent. Under part 278 of title 7, a retail food store applies to the Food and Nutrition Service, not to any state agency, and the service authorizes the store if it meets the federal criteria, which turn on the variety of staple foods stocked or on the share of total sales derived from staple foods. A store authorized in one state redeems benefits issued by every state, because the electronic benefit transfer systems through which households shop settle against federal accounts rather than state ones. The service’s Retailer Operations Division monitored the authorized store population, which Associate Administrator Shahin testified in July 2016 numbered more than 260,000 retailers under department authorization, monitoring, and oversight. Federal investigators and analysts mine transaction data for patterns that suggest trafficking, conduct store investigations, and impose sanctions that range from civil money penalties and periods of disqualification to permanent disqualification for trafficking, with administrative and judicial review available under part 279 of title 7. States, meanwhile, investigate fraud by participants in their own caseloads. The two integrity systems meet at the store counter but answer to different governments, and a trafficking case can require coordination between a federal retailer investigation and a state participant investigation that neither government controls outright.

The seam between the rulebook and the case produces most of the program’s administrative friction, and the points of strain are structural rather than accidental. A change in federal regulation obliges fifty three agencies to revise forms, retrain staff, and reprogram eligibility systems, and major state computer changes require advance planning documents submitted for federal approval before development proceeds, so the pace of federal policy is set in part by the slowest state system conversion. A state that reads a verification rule more strictly than its neighbor produces different outcomes for similar households under a statute that promises national standards, and the management evaluation and quality control systems exist largely to detect and narrow that divergence after the fact. Fiscal pressure runs along the same seam. Because benefits are fully federal and administration is half state, a caseload surge raises federal benefit spending automatically while the state share of the added casework competes inside a state budget that the surge itself may be straining. Each tension is the predictable cost of a program that seeks national uniformity of entitlement through governments that Congress does not command directly, and the administrative law of the program, its plans, options, waivers, reviews, and fiscal levers, is the machinery built to manage that cost year after year.

How the Benefit Amount Is Built

The amount credited to a household in a given month is the product of a single statutory formula applied to a set of federally defined inputs, and almost every administrative dispute about a benefit level is, on inspection, a dispute about one of those inputs rather than about the formula itself. The formula has two sides. On one side stands the maximum allotment, a dollar ceiling set by household size and derived from the cost of a federally priced food plan. On the other side stands the household’s net monthly income, a figure produced by passing gross income through a sequence of deductions that Congress, rather than any state agency, prescribed. The benefit is the distance between the two, reduced by a fixed share of net income that the statute treats as the household’s own expected contribution toward food. Because each input is defined nationally, the same arithmetic governs in every state, even though state agencies perform the calculation and gather the documentation behind it. The calculation must be followed in the order the regulations impose, because a deduction taken at the wrong step changes every figure downstream of it.

The starting point is the Thrifty Food Plan, a market basket of foods priced by the Department of Agriculture as the lowest cost of the department’s four food plans. The plan specifies quantities of food groups and food items that, prepared at home, would provide a nutritionally adequate diet for a reference household at minimal cost. The department prices the basket using food price data and consumption information, and the resulting monthly cost becomes the anchor for the entire allotment schedule. The reference household for the plan is a household of four, and the maximum allotments for other household sizes are derived from that anchor through adjustments that reflect economies of scale in food purchasing and preparation. A one person household receives more per person than the four person reference household, because a single person cannot exploit bulk purchasing or shared preparation to the same degree, and a household larger than four receives less per person, because the shared kitchen spreads fixed costs across more meals. The schedule is therefore not a simple multiplication of a per person rate.

The plan carries a history that explains its level. The Thrifty Food Plan replaced the earlier Economy Food Plan in 1975, and it did so on a cost neutral basis, meaning that the new basket was constructed to cost the same in real terms as the basket it displaced rather than to raise the purchasing power behind the benefit. The Food Stamp Act of 1977 then codified the thrifty plan as the basis of the maximum allotment, and the plan served from the late 1970s onward as the measure from which allotments were set. The cost neutral design had a durable consequence. Periodic revisions of the basket could substitute one food for another or reflect changed dietary guidance, but a revision conducted under a cost neutrality instruction could not, by itself, make the underlying diet more expensive in real terms. The maximum allotment therefore rose with food price inflation, as measured through the annual adjustments described below, without acquiring additional real purchasing power from the plan revisions themselves. The ceiling was built to hold a fixed real value derived from a minimal cost basket, not to expand toward the cost of an average grocery budget.

For fiscal 2020, under the Food and Nutrition Service memorandum titled SNAP Fiscal Year 2020 Cost of Living Adjustments, effective October 1, 2019, the maximum monthly allotment in the 48 contiguous states and the District of Columbia was 194 dollars for a household of one, 355 dollars for a household of two, 509 dollars for a household of three, and 646 dollars for a household of four. The same schedule set the maximum at 768 dollars for a household of five, 921 dollars for a household of six, 1,018 dollars for a household of seven, and 1,164 dollars for a household of eight, with 146 dollars added for each additional person beyond eight. Separate schedules applied in Alaska and Hawaii at higher levels, reflecting the higher food costs the department measured in those states. A household with no countable net income received the full maximum for its size, and every other eligible household received the maximum reduced by the income based subtraction at the center of the formula. The maximum is thus best understood as a starting figure from which the calculation works downward, rather than as a typical payment, and average benefits in program statistics for any fiscal year fall well below the maximum for most household sizes precisely because most households bring some countable income into the calculation.

Those dollar figures were not permanent. The Food and Nutrition Act requires an annual cost of living adjustment, and the Food and Nutrition Service implements that adjustment at the start of each federal fiscal year, on October 1. The fiscal 2020 figures quoted above took effect on October 1, 2019 and governed allotments through September 30, 2020. The annual adjustment reaches beyond the maximum allotments themselves. It also resets the standard deduction amounts, the cap on the shelter deduction, the income eligibility standards, and the minimum benefit, so that the whole arithmetic of the program shifts together on a single date each year. The adjustment is tied to the cost of the Thrifty Food Plan as priced for the preceding June, which means that the allotment schedule for a fiscal year reflects food prices measured several months before the fiscal year opens. A household certified under one schedule and recertified under the next could see its benefit change even though its income, its rent, and its composition had not changed at all. The calendar, acting through the adjustment mechanism, is itself one of the inputs into the benefit amount.

Against the maximum allotment, the calculation sets net income, and net income in the program is a constructed figure rather than a figure any household would recognize from a pay stub or a bank statement. The construction begins with gross monthly income, meaning the household’s total countable income from earnings, unearned sources such as Social Security benefits or unemployment compensation, and other recurring sources the statute counts, before any deduction. Excluded income, which the statute removes from consideration altogether, never enters the figure. The deductions then apply in a prescribed sequence, and the sequence is not an administrative convenience. Because the shelter deduction, the largest and most variable of the deductions, is computed from income remaining after the earlier deductions have been taken, the order in which deductions apply determines the size of the shelter deduction and therefore determines net income itself. A payment error in a benefit calculation very often traces to a step performed out of order or to documentation that placed an expense at the wrong step.

The first deduction in the sequence is the earned income deduction, set at 20 percent of the household’s gross earnings. Twenty percent of wages or self employment earnings is disregarded before any other deduction applies, so a household supported by earnings enters the rest of the calculation with only four fifths of those earnings counted. The deduction applies to earned income only. Unearned income, such as a retirement benefit or a cash assistance payment, receives no equivalent percentage disregard. Two households with identical gross monthly income therefore receive different treatment at this first step if one household’s income arrives as wages and the other’s arrives as unearned benefits, and the difference created at this step persists through every later step. The earned income deduction functions in practice as the program’s principal built in work incentive within the benefit formula itself, separate from any work requirement imposed as a condition of eligibility, because earnings always leave a household better off in benefit terms than the same amount of unearned income.

The second deduction is the standard deduction, a fixed monthly amount subtracted for every household, regardless of its actual spending, in recognition of the basic nonfood expenses that no household avoids. For fiscal 2020, effective October 1, 2019, the Food and Nutrition Service set the standard deduction in the 48 contiguous states and the District of Columbia at 167 dollars for households of one to three persons, 178 dollars for a household of four, 209 dollars for a household of five, and 240 dollars for a household of six or more, with higher amounts under the Alaska and Hawaii schedules. The standard deduction requires no receipts and no documentation of particular expenses, which makes it the simplest step in the calculation and the step least likely to generate a dispute between a household and an agency. After the earned income deduction and the standard deduction, the regulations allow further deductions for particular categories of expense, including dependent care costs incurred so that a household member can work or pursue training or education, medical expenses for household members who are age 60 or older or who have a disability, above a threshold amount the statute sets, and legally obligated child support payments, treated as a deduction in states that elected that treatment. Each such deduction depends on documentation, and each is a point at which an eligible expense that goes unreported or unverified produces a smaller benefit than the household’s circumstances would support.

The final and most consequential deduction is the excess shelter deduction. Shelter costs in the program include rent or mortgage payments, property taxes and insurance on the home, and utility costs, which state agencies commonly measure through standard utility allowances approved by the Food and Nutrition Service rather than through a household’s individual bills. The calculation takes the household’s income after every deduction described above has been applied, computes one half of that remaining figure, and compares the household’s total shelter costs against that half. Shelter costs count as a deduction only to the extent that they exceed one half of the household’s adjusted income, and only that excess is subtracted. The deduction therefore responds to housing burdens that are heavy relative to income rather than to housing costs as such. A household paying modest rent relative to its income receives no shelter deduction at all, while a household paying the same rent on a smaller income may receive a large one. For fiscal 2020, effective October 1, 2019, the Food and Nutrition Service capped the excess shelter deduction at 569 dollars a month in the 48 contiguous states and the District of Columbia. The cap did not apply to a household containing a member who was age 60 or older or who had a disability, and such a household received the full amount of its excess shelter costs as a deduction.

Once the deductions have all been taken, the remainder is net monthly income, and the benefit follows from a single subtraction. The allotment equals the maximum allotment for the household size minus 30 percent of net monthly income. The 30 percent figure rests on a statutory expectation, present in the program’s design since the 1977 act set the allotment at the cost of the Thrifty Food Plan reduced by 30 percent of household income, that a household devotes roughly 30 percent of its own resources to food. The benefit, in other words, is calibrated as a supplement that fills the gap between what the Thrifty Food Plan costs and what the household is expected to contribute toward that cost from its net income. A household with net income of zero receives the full maximum, because the expected contribution is zero and the gap is the entire cost of the plan. As net income rises, the benefit falls by 30 cents for each additional dollar of net income, a reduction rate far gentler than a dollar for dollar offset. Where 30 percent of net income equals or exceeds the maximum allotment, the formula yields no positive benefit, and a household of three or more in that position receives no allotment, subject to the minimum benefit exception that protects the smallest households.

A complete illustration, built entirely on the fiscal 2020 schedule effective October 1, 2019 for the 48 contiguous states, shows how the steps combine. Consider a household of four whose gross monthly income consists of 2,000 dollars in earnings and nothing else. The earned income deduction removes 20 percent of those earnings, which is 400 dollars, leaving 1,600 dollars. The standard deduction for a household of four in fiscal 2020, 178 dollars, then applies, leaving 1,422 dollars. In this illustration the household claims no dependent care deduction, no medical deduction, and no child support deduction, so the calculation moves directly to shelter. The household’s monthly shelter costs, rent and utilities combined, come to 900 dollars. One half of 1,422 dollars is 711 dollars, and the shelter costs exceed that half by 189 dollars, so the excess shelter deduction is 189 dollars, an amount below the fiscal 2020 cap of 569 dollars and therefore allowed in full. Subtracting 189 dollars from 1,422 dollars produces net monthly income of 1,233 dollars. Thirty percent of 1,233 dollars is 369.90 dollars, treated as 370 dollars in whole dollar terms. The maximum allotment for a household of four is 646 dollars, and 646 dollars minus 370 dollars yields a monthly allotment of 276 dollars. The household in the illustration is then expected, under the statute’s own assumption, to direct roughly 370 dollars of its net income toward food, with the 276 dollar benefit closing the distance between that expected contribution and the 646 dollar cost assigned to the Thrifty Food Plan for a household of that size.

The illustration also exposes the most persistent misunderstanding of the program’s arithmetic, which is the expectation that the benefit is designed to cover the full cost of a household’s food. It is not, and it never has been under the formula as codified. The household in the illustration receives 276 dollars against a federally priced minimal food budget of 646 dollars, and the difference is not an error, a state level reduction, or a partial payment of an entitlement to the full amount. The difference is the formula working exactly as written, on the stated premise that households contribute a share of net income to food purchases and that the federal benefit supplements that contribution up to the level of a thrifty, minimal cost diet. Even the maximum allotment itself represents only the cost of the lowest cost food plan the Department of Agriculture publishes, a basket that assumes food is prepared at home and that allows very little margin for waste, spoilage, or meals eaten away from home. Public argument about benefit adequacy, in either direction, that treats the allotment as a failed attempt to pay for a household’s entire food spending misdescribes the instrument Congress built, whatever conclusion the argument reaches about whether the benchmark or the 30 percent expectation ought to change.

At the bottom of the schedule, a minimum benefit protects the smallest households from a formula result so small that issuance would cost more to administer than the benefit is worth. For fiscal 2020, effective October 1, 2019, the Food and Nutrition Service set the minimum monthly benefit in the 48 contiguous states and the District of Columbia at 16 dollars, and the minimum applied to eligible households of one or two persons. A one or two person household that satisfied the eligibility rules received at least 16 dollars even if the formula, maximum minus 30 percent of net income, produced a smaller positive amount or no amount at all. The guarantee did not extend to households of three or more, and a larger household whose 30 percent subtraction consumed the entire maximum received no allotment. The floor bears on participation among older adults living alone or in pairs, a group among whom a calculated benefit in single digits might otherwise discourage an application altogether, although the 16 dollar amount remains modest and its value lies partly in maintaining the connection to the program that participation establishes.

Two adjustments complete the picture of how a calculated figure becomes an issued figure. The first is the annual adjustment already described, which means that the allotment a household receives is always the product of the schedule in force for that fiscal year, and a benefit can rise or, in a year when a schedule component moves the other way, fall at the October 1 boundary without any change in the household’s own circumstances. The second is proration in the initial month of participation. A household does not ordinarily receive a full month’s allotment for the month in which it applies, because eligibility in that first month runs from the date the application is filed rather than from the first day of the month. The agency calculates the full monthly allotment through the ordinary formula and then prorates that amount according to the portion of the month remaining after the filing date, so that a household applying late in a month receives a correspondingly small allotment for that month and the full calculated allotment from the following month onward, provided eligibility continues. Proration explains a common source of confusion at the start of participation, where a first payment far below the expected amount reflects the filing date rather than a recalculation of need, and it places administrative weight on the accurate recording of the application date itself.

The structure of the calculation resolves a question that aggregate program statistics often raise, which is why two households reporting the same income can receive markedly different benefits, and why a household with higher income can occasionally receive a larger benefit than a household with lower income. Income alone never determines the allotment. Household size sets the maximum, and a larger household starts from a higher ceiling against which the same income subtraction operates. The source of income matters, because the 20 percent earned income deduction reduces the countable value of wages but not of unearned income. The standard deduction varies by household size. Documented expenses matter at every later step: dependent care costs, qualifying medical expenses, and child support payments each reduce net income, and shelter costs above one half of adjusted income reduce it further, up to the cap for most households and without a cap for households with a member who is age 60 or older or who has a disability. State level variation enters through standard utility allowances, which differ among states and which feed directly into the shelter figure, so that identical households with identical rents in different states can produce different excess shelter deductions where the applicable utility allowances differ. The deductions, taken together, do the work of individualizing a national formula, and they are also the points at which the calculation is most sensitive to documentation, to agency accuracy, and to a household’s own reporting. A benefit figure detached from the deduction record behind it cannot be evaluated, because the same number can represent a correct application of the formula to one set of circumstances and an error when applied to another.

The Eligibility Tests and the Doors Around Them

Eligibility for the Supplemental Nutrition Assistance Program was decided by a sequence of tests applied to a unit that the statute defined before any income was counted. The Food and Nutrition Act did not measure the means of an individual applicant in isolation, and it did not measure the means of everyone who shared an address. It measured the means of a household, a term defined by the way food was bought and eaten. Under the definition carried in the Food and Nutrition Service guidance in force during fiscal 2019, a household consisted of persons who lived together and customarily purchased food and prepared meals together. A person who lived alone formed a household of one. A person who lived with others but purchased food and prepared meals separately could form a separate household. A group that lived together and shared the purchasing and preparation of meals formed a single household, whether or not its members were related to one another, and the income and resources of the group were counted together in a single determination.

Against that general rule of shared meals, the statute imposed a set of mandatory groupings that no arrangement of separate shopping or separate cooking could dissolve. Spouses who lived together were a single household. Parents and their children under the age of 22 who lived together were a single household, including stepchildren, and including the spouse and children of such a child where those persons lived in the same home. A child under the age of 18 who lived under the parental control of an adult household member was counted with that adult, foster children excepted. The effect of the mandatory groupings was administrative as much as it was financial. A caseworker first reconstructed the economic unit to which the applicant belonged, attributed the income of every member of that unit to the whole, and tested the combined figure against limits that rose with household size.

The composition rules contained one carefully bounded door of their own. An elderly or disabled person who was unable to purchase and prepare meals separately because of a permanent disability could, together with a spouse, form a separate household from the other persons in the home if the gross income of those other persons did not exceed 165 percent of the federal poverty guidelines. The Food and Nutrition Service fiscal 2020 income eligibility standards, effective October 1, 2019, set that 165 percent figure for a household of four in the 48 contiguous states and the District of Columbia at 3,541 dollars per month. The exception recognized that disability could separate a person’s food economy from the food economy of the surrounding home even where affection and shelter were shared, and it priced that recognition with an income test applied to the surrounding household rather than to the applicant.

Once the household was fixed, the regular federal rules applied two income tests in sequence, and the order mattered. The first was the gross monthly income test. A household without an elderly or disabled member was required to show gross monthly income, meaning total nonexcluded income before deductions, at or below 130 percent of the federal poverty guidelines for its size. The Food and Nutrition Service fiscal 2020 standards, dated October 1, 2019, set that limit for a household of four in the 48 contiguous states and the District of Columbia at 2,790 dollars per month, with higher schedules for Alaska and Hawaii reflecting the higher poverty guidelines in those states. Households containing a member who was age 60 or older or who met the program definition of disability were exempt from the gross test altogether under the regular rules, a distinction that reflected the different expense structure the deductions were designed to recognize for such households. A household without such a member that failed the gross test was denied without reaching the arithmetic that followed, however large its rent or its child care bill happened to be.

A household that passed the gross test, or that was exempt from it, moved to the second and more intricate stage, in which a defined stack of deductions converted gross income into net income. The first deduction was the standard deduction, a flat amount allowed to every household by size, intended to represent basic unavoidable expenses that the program did not require any household to document. Under the Food and Nutrition Service fiscal 2020 cost of living adjustments, effective October 1, 2019, the standard deduction in the 48 contiguous states and the District of Columbia was 167 dollars per month for households of one to three members, 178 dollars for a household of four, 209 dollars for a household of five, and 240 dollars for a household of six or more. The second deduction rewarded earnings directly. Twenty percent of gross earned income was disregarded, so that wages and self employment earnings entered the net income calculation at four fifths of their face value, while unearned income such as Social Security benefits, unemployment insurance, and pensions entered in full. It reduced the rate at which benefits fell as earnings rose, and it meant that two households with identical gross income faced different net income figures if one lived on wages and the other lived on a pension of the same amount.

The remaining deductions were expense based, and each carried its own verification burden. The dependent care deduction allowed the actual cost of care for a child or for another dependent, to the extent the care was necessary to allow a household member to work, to seek work, or to pursue training or education. Legally obligated child support payments made to a person outside the household were also deductible. The medical expense deduction was narrower in its reach and larger in its consequence. It was available only for the out of pocket medical expenses of elderly or disabled members, only to the extent those expenses exceeded 35 dollars per month, and only for expenses that no insurer or third party reimbursed. Prescription costs, medical and dental care, health insurance premiums, and attendant care could all enter the calculation once the 35 dollar threshold was passed, and for an older household with several recurring prescriptions the deduction could move net income by an amount large enough to decide the case. The distribution of the deduction was uneven in practice, because it depended on a household knowing that unreimbursed medical costs counted and gathering proof of them.

The final deduction addressed shelter, and it was calculated last because its formula depended on everything before it. The excess shelter deduction allowed shelter costs, including rent or mortgage payments, taxes and insurance on the home, and utility costs, to the extent those costs exceeded one half of the household income remaining after every other deduction had been taken. For households without an elderly or disabled member, the deduction was capped. The Food and Nutrition Service fiscal 2020 cost of living adjustments, effective October 1, 2019, set the cap in the 48 contiguous states and the District of Columbia at 569 dollars per month, an increase of 17 dollars from the fiscal 2019 cap recorded in the same adjustment series. For households with an elderly or disabled member, the excess shelter deduction was uncapped, so that the full amount by which shelter costs exceeded half of adjusted income reduced net income. A working household with children and a high rent therefore received shelter relief up to a fixed ceiling and no further, while an older household paying the same rent on a similar income could deduct without a ceiling.

Net monthly income, the figure produced when the full stack had been subtracted, was then tested against 100 percent of the federal poverty guidelines. The Food and Nutrition Service fiscal 2020 standards, dated October 1, 2019, set that limit for a household of four in the 48 contiguous states and the District of Columbia at 2,146 dollars per month. Every household under the regular rules, including households with elderly or disabled members that had skipped the gross test, was required to pass the net test. Net income performed double duty in the program. It was an eligibility threshold, and it was also the input to the benefit formula, under which the household was expected to devote 30 percent of net income to food and the allotment filled the gap between that expected contribution and the maximum allotment for the household size. The deduction stack therefore decided two questions at once. It decided whether a household received anything, and, for households that qualified, it decided how much, with each verified dollar of dependent care, medical expense, or shelter cost raising the benefit by thirty cents at the margin until the maximum allotment was reached.

The third regular test concerned resources rather than income. Under the federal rules in fiscal 2019, a household was required to hold countable liquid resources below 2,250 dollars, or below 3,500 dollars if the household included an elderly or disabled member, according to the Congressional Research Service report on categorical eligibility. The Food and Nutrition Service fiscal 2020 cost of living adjustments, effective October 1, 2019, recorded both limits as unchanged for fiscal 2020, at 2,250 dollars for other households and 3,500 dollars for households with at least one member age 60 or older or disabled. The test counted resources that could be converted to cash for food, such as bank balances and cash on hand, and it excluded much of what a household owned. The value of the home and lot was excluded. Retirement and educational savings were excluded. A portion of the value of a household vehicle, up to 4,650 dollars of fair market value under the federal rule described by the Congressional Research Service, was excluded, and federal law separately allowed a state to apply the vehicle exclusion used in its Temporary Assistance for Needy Families program, under which many states excluded the value of one vehicle in full. Its dollar limits sat far below the value of the savings that financial counselors advised low income households to accumulate against emergencies.

Around each of the three tests, the statute and the regulations built categorical doors, and the doors decided more cases than the tests did. The oldest door was traditional categorical eligibility. A household in which every member received Supplemental Security Income, Temporary Assistance for Needy Families cash assistance, or state general assistance was categorically eligible for the Supplemental Nutrition Assistance Program without a separate application of the income and resource tests, on the reasoning that such a household had already passed a means test at least as demanding in another program and that a second full determination wasted administrative money in both directions. Traditional categorical eligibility was a coordination device between programs, and limited in its reach, because the cash assistance caseloads it keyed to had contracted sharply in the decades after the 1996 welfare statute replaced the former cash aid structure with the Temporary Assistance for Needy Families block grant.

The decisive door was built later, out of a different feature of that block grant. Federal regulations allowed a household to be categorically eligible not only upon receipt of cash aid, but also upon receipt of, or authorization to receive, a noncash benefit or service funded by Temporary Assistance for Needy Families dollars or by state funds counted toward the federal maintenance of effort requirement. Because Temporary Assistance for Needy Families funds could finance services directed at the block grant statutory purposes, and because two of those purposes, reducing out of wedlock pregnancies and encouraging the formation and maintenance of two parent families, were not confined by the statute to families below a stated income in the way the cash assistance purposes were, a state could finance a modest service aimed at those purposes, set the income limit for that service as high as 200 percent of the federal poverty guidelines under the framework the Food and Nutrition Service described in its categorical eligibility guidance, and confer categorical eligibility on every household that received the service or was authorized to receive it. In the form most states adopted, the service was deliberately slight. The Congressional Research Service report on categorical eligibility described households made eligible through receipt of a brochure or a referral to a social services telephone number, funded with Temporary Assistance for Needy Families or maintenance of effort dollars and directed at a statutory purpose of the block grant. The brochure cost almost nothing to provide, the eligibility it conferred cost the state nothing in benefit dollars because benefits were federally funded, and the determination it replaced was one of the most error prone in the state agency workload, the verification and valuation of household assets.

The practice acquired a name, broad based categorical eligibility, and by the end of the second decade of the program under that name its geography was extensive. The Center on Budget and Policy Priorities, in a report updated July 30, 2019, counted more than 40 states using the policy to apply less restrictive income and asset tests. The Congressional Research Service, reporting Department of Agriculture figures, recorded that in July 2019 a total of 42 jurisdictions operated broad based categorical eligibility to make most or all households with whom the state welfare office came into contact eligible through that route. The Congressional Research Service table of state policies in October 2018 showed the internal variety behind the totals. Of the jurisdictions using the policy, 36 applied no asset test at all to households made eligible through it, six states, identified as Idaho, Indiana, Maine, Michigan, Nebraska, and Texas, applied an asset test to all such households, and 33 jurisdictions set a gross income limit above the regular 130 percent of poverty. State gross income limits under the policy ranged upward to the regulatory ceiling, with states including California, Colorado, Delaware, Florida, and Hawaii, and the District of Columbia, setting the limit at 200 percent of the federal poverty guidelines in the October 2018 table. Two jurisdictions in that count, New Hampshire and New York, limited the policy to certain household types rather than extending it to all family types.

What broad based categorical eligibility changed, in mechanical terms, was which test a household faced. For a household made eligible through the door, the state deemed the household to satisfy the gross income test and the resource test of the program whose benefit or service the household had received, substituting the gross income limit and the asset rule of the Temporary Assistance for Needy Families funded service for the regular federal figures. Where the service had no asset limit, the household faced no asset test. Where the service carried a gross income limit of 185 or 200 percent of poverty, the household faced that limit in place of 130 percent. The net income test retained its force in the benefit calculation, because the allotment was computed from net income in the ordinary way, and a household whose deductions left net income above 100 percent of poverty received no payable benefit in practical terms, a constraint the Center on Budget and Policy Priorities emphasized in its July 30, 2019 account of the policy. The door therefore widened entry chiefly for households whose gross income exceeded 130 percent of poverty but whose deductible expenses were heavy enough to pull net income to or below the poverty line, a profile that described working households paying market rent and market child care prices, and for households whose income was plainly low enough but whose savings, a vehicle, or a modest bank balance accumulated for a security deposit would have failed the resource test in its regular form.

The administrative consequences followed from the same mechanics. Asset verification had required caseworkers to value vehicles, to examine bank statements, and to classify accounts, and each valuation was a point at which an agency could err and a household could be denied for a balance held briefly for rent. Removing the asset test for most households in most states removed a class of errors and a class of denials together. The Center on Budget and Policy Priorities argued in its July 30, 2019 report that the policy supported working families whose earnings had risen past the regular gross limit while housing and child care costs consumed the increase, and that it allowed families, older adults, and people with disabilities to hold modest savings without losing food assistance. Congressional critics of the policy, including the House Agriculture Committee in its materials for the Agriculture and Nutrition Act of 2018, argued that eligibility conveyed by a brochure or a hotline number severed the connection between food assistance and any meaningful test of need conducted by another program, and proposed to confine categorical eligibility to households receiving cash assistance or substantial ongoing services. The statute as it stood on December 1, 2019 contained the broader door, and state agencies administered eligibility through it at the scale the Congressional Research Service and the Center on Budget and Policy Priorities counts recorded.

Beyond the categorical doors, the eligibility structure excluded defined classes of persons on grounds that had nothing to do with income. A person who lived in an institution that provided most meals was generally ineligible, on the reasoning that the institution already met the need the benefit addressed, with exceptions recognized for residents of federally subsidized housing for older adults, drug and alcohol treatment facilities, group living arrangements, shelters for battered women and children, and shelters for homeless persons. A boarder, a person who paid reasonable compensation for lodging and meals in another household, could not form a separate household and could participate only as a member of the household providing the board, at that household request. A student enrolled half time or more in an institution of higher education was ineligible unless the student fit an exemption, a rule that made enrollment itself a disqualifying status and then readmitted defined categories of enrolled students. The exemptions in force during fiscal 2019 included a student under age 18 or age 50 or older, a student physically or mentally unfit, a student working in paid employment an average of 20 hours per week, a student participating in a state or federally financed work study program, a student responsible for the care of a dependent household member under the age of 6, a single parent enrolled full time who was responsible for a child under the age of 12, a student receiving Temporary Assistance for Needy Families benefits, and a student placed in higher education through an employment and training program.

The eligibility machinery, assembled whole, presented two faces. In the statute and in the federal tables, the tests were uniform and national. One gross limit, expressed as 130 percent of the poverty guidelines, one net limit at 100 percent, one pair of resource figures, and one deduction stack applied from Maine to California, varied only by the Alaska and Hawaii schedules and by the annual cost of living adjustments each October 1. In the state office, the effective tests were a state product. A household with gross income at 175 percent of poverty and 4,000 dollars in a savings account was eligible in a state that had set a 200 percent gross limit and no asset test through broad based categorical eligibility, provided its deductions brought net income to the poverty line, and the same household was ineligible at the application screen in a state that administered the regular tests or that kept an asset limit inside its categorical policy. The difference lay not in the household need, which was identical on both sides of the state line, but in a state plan choice about a brochure, a hotline, or a referral service financed with block grant dollars. Administration of eligibility was therefore federal in its funding, federal in its printed rules, and state by state in its operative meaning.

The eligibility to issuance pipeline

Stage	Actor	Rule	Timeline
Application	Household and state agency	Application filed in person, by mail, by telephone, or online; the filing date starts the determination period and governs proration of the first month	Filing date recorded on receipt
Interview	State agency eligibility worker and household	Certification interview completed before eligibility is decided	Before determination, inside the application period
Verification	Household and state agency	Income, household composition, and deductible expenses verified against program standards	Before determination
Eligibility determination	State agency	Gross income, net income, and resource tests applied, with categorical eligibility where it applies	Within 30 days of application; 7 days for expedited issuance to qualifying households
Benefit calculation	State agency	Maximum allotment for the household size reduced by 30 percent of net income, after deductions taken in the prescribed order	At determination and at each recertification
Issuance	State agency through its electronic benefit transfer processor	Benefits credited to the household electronic benefit transfer account under Part 274 of Title 7	Monthly on the state staggered schedule, with no more than 40 days between allotments
Recertification	Household and state agency	Household reapplies, completes an interview, and verifies circumstances again before the certification period expires	At the end of each certification period, ordinarily 6 to 12 months, and up to 24 months for households of older adults or adults with disabilities

The Accuracy Regime That Runs the Program

No other measurement in the Supplemental Nutrition Assistance Program shaped state behavior as powerfully as the quality control system. Eligibility rules defined what a correct decision was, but the quality control system defined what a state agency was judged by. Each year the system converted thousands of individual case decisions into a single published percentage for each state and a single combined percentage for the nation. That percentage determined fiscal liability, governed eligibility for performance bonuses, triggered corrective action, and supplied the ranking by which a state administration was judged. Administrators therefore organized training, verification policy, computer systems, and office workflow around the error rate, and applicants experienced that organization directly, in the documents demanded of them and the friction of the process.

The mechanics were deliberately two-tiered, because the state agency whose work was being measured performed the first measurement. Under the Food and Nutrition Act and the regulations at Part 275 of Title 7 of the Code of Federal Regulations, each state agency drew a random statistical sample of its active caseload each month. According to the Food and Nutrition Service fact sheet on the payment error rate issued in June 2018, the monthly samples added up to about 50,000 cases a year across the nation. State quality control reviewers, organizationally separate from the eligibility workers who had handled the sampled cases, reopened each case, interviewed the household, verified income, household composition, and deductible expenses, and decided what the household had been entitled to receive in the sample month. A difference between the amount issued and the amount the reviewer determined to be correct became a variance, and a variance beyond the permitted tolerance became an error charged against the state.

Food and Nutrition Service staff then selected a subsample of the state reviewed cases, about 25,000 cases a year according to the June 2018 fact sheet, and re-reviewed those cases independently. Where federal and state findings disagreed, a resolution and arbitration process settled the final finding. Errors were also corrected household by household: an overpaid household faced a claim for repayment, and an underpaid household received a supplemental issuance. Federal analysts weighted the sample results by caseload and combined them into state payment error rates and the national rate announced each year.

The payment error rate had an exact meaning, and much public argument about the program went wrong by ignoring it. The rate measured the share of benefit dollars issued incorrectly, and its numerator combined two components that were added together rather than netted against each other. The overpayment rate captured dollars issued to households that were ineligible and dollars issued to eligible households in amounts above their entitlement. The underpayment rate captured dollars that eligible participating households should have received but did not, because the agency calculated the allotment too low. The combined rate was the sum of the two components, not a net loss figure, because underpayments offset overpayments in arithmetic only if the two were subtracted, and the quality control system deliberately refused that subtraction.

Two further features of the definition mattered as much as the headline formula. The rate counted only errors above a tolerance threshold set in statute, designed to keep state attention on errors large enough to matter. The Agricultural Act of 2014 set the threshold at 37 dollars for fiscal 2014, with annual adjustment tied to the Thrifty Food Plan, after a period in which the threshold had stood at 50 dollars. In GAO-16-708T, issued in July 2016, the Government Accountability Office reported that errors below the threshold in fiscal 2013 accounted for 38 percent of the dollars actually paid in error in that year. Cases in which the household was determined ineligible were counted in full regardless of the threshold. Defined categories of cases also sat outside the review frame. Households certified under disaster procedures authorized by the Food and Nutrition Service were not subject to review under Section 275.12 of the regulations, and the same exclusion applied to households that had not participated in the sample month and households under active investigation for an intentional program violation.

A companion measurement covered the cases in which the agency had said no. Alongside the active sample, states drew a sample of negative actions, meaning applications denied and cases suspended or terminated, and reviewers assessed whether the action had been correct, whether the stated reason had been the right reason, and whether the household had received a correct, complete, and timely notice. The results produced the case and procedural error rate, which replaced the older negative error rate. The Food and Nutrition Service expanded the negative case review beginning in fiscal 2012, holding states more accountable for the timeliness, accuracy, and clarity of notices, as the Center on Budget and Policy Priorities described in its July 2015 analysis of the fiscal 2014 rates. That analysis, drawing on data released by the Department of Agriculture, reported that in fiscal 2014 more than one quarter of state actions denying or terminating benefits were found improper, and that the case and procedural error rates in eight states approached or exceeded 50 percent. The case and procedural error rate was not a dollar rate and could not be added to the payment error rate. It captured a harm the payment rate could not reach, because a household wrongly denied benefits received no payment at all.

The article framework that error is not fraud rested on this measurement structure. The payment error rate did not measure fraud, and it did not attempt to measure fraud. A finding of error required no showing of intent by anyone, and a case counted whether the discrepancy came from a caseworker mistake, a data entry error, a household misunderstanding a reporting rule, or a household deliberately concealing income. Intent belonged to a different legal apparatus of intentional program violation determinations, disqualifications, and prosecutions, which the quality control sample neither administered nor estimated. In testimony before the House Committee on Agriculture on July 6, 2016, Jessica Shahin, the Associate Administrator for the Supplemental Nutrition Assistance Program at the Food and Nutrition Service, stated that most errors stemmed from unintentional mistakes on the part of the state agency or the household rather than fraud, and that the majority of the errors, 62 percent, were state agency errors, while 38 percent were client errors. The Center on Budget and Policy Priorities, analyzing the Food and Nutrition Service quality control annual reports, reported that states themselves were responsible for about half of overpayment errors and for almost 90 percent of underpayment errors in the years covered by those reports.

The published series gave the system its drama. After a long decline, the combined national payment error rate reached a record low of 3.20 percent in fiscal 2013, as the Food and Nutrition Service announced in June 2014, and then rose to 3.66 percent in fiscal 2014. Even that modest rise illustrated how sensitive the series was to measurement choices. The Center on Budget and Policy Priorities attributed roughly a quarter to a third of the fiscal 2014 increase to the Agricultural Act of 2014 itself, which lowered the error tolerance threshold from 50 dollars to 37 dollars and thereby moved a band of small errors from the excluded category into the counted category without any change in what caseworkers had actually done. The same analysis pointed to operational changes in several large states and to the reinstatement of the three month time limit for able-bodied adults without dependents in a group of states in fiscal 2014, a complex policy that made the affected cases unusually error prone.

Publication then stopped. The Department of Agriculture identified concerns about the quality of quality control data in 2014, and the Food and Nutrition Service suspended national error rate reporting for fiscal 2015 and fiscal 2016 while federal teams reviewed the quality control systems of all fifty states, the District of Columbia, the Virgin Islands, and Guam. The June 2018 Food and Nutrition Service fact sheet reported that the review found data integrity issues requiring corrective action in 42 states. The Office of Inspector General of the Department of Agriculture, in a report issued in September 2015, had identified the structural vulnerability: accurate reporting exposed states to penalties while skillful mitigation of findings earned them bonuses, and federal oversight had been inadequate to police the conflict. Federal officials described the techniques in congressional testimony during 2017. States had used third party consultants and error review committees to work on sampled cases after selection, supplying missing information or mitigating individual findings before federal reviewers saw the cases, rather than improving the original determinations the sample was supposed to test. The Department of Justice obtained False Claims Act settlements in fiscal 2017 from three state agencies accused of falsifying quality control data and unlawfully obtaining federal bonuses, as the Congressional Research Service recorded in its March 2018 report on errors and fraud in the program.

When publication resumed, the numbers arrived on a new footing. The combined national payment error rate stood at 6.30 percent for fiscal 2017, announced by the Food and Nutrition Service in June 2018, and at 6.80 percent for fiscal 2018. The Food and Nutrition Service rejected that reading in the announcement itself, stating that the apparent increase in fiscal 2017 resulted from the improved reporting process rather than from an actual increase in improper payments, and that the fiscal 2017 rate was more accurate than the rates that had preceded the suspension. Several reasons supported the warning. The first was the bias removed during the suspension: if state mitigation practices had suppressed findings in the earlier period, part of the decline to 3.20 percent had measured skill at managing the review rather than skill at determining eligibility, and the resumed rates were produced under new controls against statistical bias, including a management evaluation process for examining state quality control procedures. The second reason was the exclusion and counting rules defining the measured universe. The tolerance threshold stood at 50 dollars in fiscal 2013 and at 37 dollars in fiscal 2014, before adjustment carried it to 38 dollars in fiscal 2017 and back to 37 dollars in fiscal 2018, so identical casework produced a higher counted rate in the resumed years. The Government Accountability Office, in GAO-16-708T, identified the threshold changes and federal policy changes excluding certain income and resources from eligibility determinations as changes that likely affected the rate directly, because each exclusion removed a potential source of countable error. Case mix shifted the comparison as well: households certified under disaster procedures sat outside the review frame, and cases administered under waivers or simplified reporting options were judged against the rules as modified for those cases, so rates moved as waivers expired and policies such as the time limit for able-bodied adults without dependents returned to force. The third reason concerned the negative side of the ledger, where the case and procedural error rate that replaced the negative error rate after fiscal 2012 rested on expanded review of notices and procedural correctness, so the fuller accuracy picture in the resumed years incorporated a more demanding assessment of denials and terminations. None of these reasons implied that the resumed rates were wrong; the Food and Nutrition Service presented them as the more trustworthy figures.

Money stood behind the measurement from the beginning, in two directions. Under the liability system that operated before the reforms of the Farm Security and Rural Investment Act of 2002, a state whose combined error rate exceeded the national average faced fiscal sanctions, and because roughly half of all states necessarily sit above any average, roughly half the states stood exposed to penalty annually even in years when national accuracy improved. States with low rates received enhanced federal funding for administration, with a combined rate below 6 percent serving as the standard for that recognition before 2003, as the Center on Budget and Policy Priorities documented in its July 2015 analysis. Critics of that structure, including state agencies and the advocacy organizations in the consensus described in the same analysis, argued that a sanction line drawn at the average punished states for serving households whose cases were hardest to determine correctly, above all working households, whose wages fluctuated with hours, overtime, and job changes and whose allotments were therefore far easier to get wrong than the allotments of households living on fixed benefits. A state that discouraged such households protected its rate.

The 2002 farm bill rebuilt both sides of that ledger. On the liability side, the new system narrowed monetary penalties to states with persistently high rates. A state faced liability only when a 95 percent statistical probability existed that its combined payment error rate exceeded 105 percent of the national performance measure for a second consecutive year, with the first year operating as a warning year, and with the penalty amount tied to the portion of the rate above 6 percent, so that a state over the 105 percent line but under 6 percent owed no payment. The Agricultural Act of 2014 later removed the authority of the Secretary of Agriculture to waive such a liability. On the incentive side, Congress replaced the older enhanced funding arrangement with a fixed annual pool of 48 million dollars in high performance bonuses, paid to the top performers and the most improved performers across four measures. As the Center on Budget and Policy Priorities set out in its July 2015 analysis, 24 million dollars of the pool rewarded payment accuracy, 6 million dollars rewarded performance on the case and procedural error rate, 6 million dollars rewarded application processing timeliness against the federal standards of seven days for expedited cases and thirty days for regular cases, and 12 million dollars rewarded performance on the program access index, the measure of how successfully a state reached its eligible low income population. Accuracy retained the largest share, but access, timeliness, and correct handling of denials each carried a price. The Agricultural Act of 2014 added a condition that bonus funds be reinvested in the program, and the Agriculture Improvement Act of 2018 then eliminated performance bonus payments altogether, effective for bonuses that would have been paid in fiscal 2019 for fiscal 2018 performance, while leaving the liability provisions in place, as the Food and Nutrition Service advised state agencies in its summary of the 2018 act.

Between the distant prospect of liability and the annual contest for bonuses sat the corrective action plan, the instrument through which the error rate governed state management in ordinary years. Federal regulations at Section 275.16 of Title 7 required a state agency to develop and implement a corrective action plan when its payment error rate reached 6 percent or more, when its case and procedural error rate stood above the national average, or when a significant share of its sampled cases could not be completed. A plan required a root cause analysis classifying errors by type and by whether the agency or the client had caused each error, followed by remedies with owners and dates, from refresher training on earned income to redesign of a notice or added supervisory review of error prone case types. The categories named in a plan became the categories supervisors checked and trainers taught in local offices.

Measured against the scale of the program, the money at stake in these mechanisms was modest. A state share of a 48 million dollar national bonus pool amounted to a small fraction of one percent of the benefit dollars that state issued in a year, and liabilities fell on few states in any period. Yet administrators treated the error rate as the central fact of program management, and the record of the suspension years showed agencies spending real resources, and in a few cases crossing legal lines, in pursuit of favorable findings. The explanation lay in the sanction that publication itself imposed. The Food and Nutrition Service released state rates together, in a single table, and the table functioned as a ranking regardless of the statistical cautions attached to it. A state at the bottom faced legislative hearings and hostile press coverage, while a state at the top collected a public record of good management. The Center on Budget and Policy Priorities observed in its July 2015 analysis that the error rate served as the major performance measure for accountability at local and county offices and even for individual workers, so reputational pressure transmitted downward through every level of the organization. The same pressure also made the structure dangerous, because a ranking that determines professional reputations invites management of the ranking by every available means, legitimate and otherwise. Larry Goolsby of the American Public Human Services Association described the bonuses in 2013 as a significant factor in achieving low error rates, an assessment from within the administrative community that the structure worked.

States managed to the error rate, and that management explained much of what applicants experienced. In its beneficial forms, management produced simplification and investment. States adopted simplified reporting options that reduced the number of interim changes households had to report and workers had to process, invested in data matching, rebuilt eligibility computer systems, and directed intensive review toward the case types that generated the most error dollars, particularly households with fluctuating earnings. Participation among eligible people rose from 54 percent in 2002 to 83 percent in 2012, on the figures reported by the Center on Budget and Policy Priorities, during the same period in which error rates fell to their record low, supporting the claim of the defenders of the 2002 reforms that accuracy and access could improve together when the incentive system rewarded both.

In its harder forms, the same management shifted cost onto applicants, and the added burden fell hardest on households with irregular work, unstable housing, or difficulty assembling records, which overlapped heavily with the households whose cases were most likely to generate errors. A state protecting its rate had reason to verify aggressively, to demand documentary proof where a household statement might have sufficed under a less anxious administration, and to scrutinize the error prone cases of working households more heavily than the stable cases of households on fixed incomes. The case and procedural error rate existed in part to discipline the opposite temptation, since a state that resolved doubtful cases by denying or terminating them exposed itself on the negative side of the system, and the fiscal 2014 finding that more than a quarter of negative actions were improper showed how often the pressure toward refusal produced decisions the record could not support.

By December 1, 2019, the accuracy regime had thus produced a measurement of genuine value and a set of incentives of genuine ambivalence. The payment error rate, at 6.80 percent for fiscal 2018, told the public something real about how often the program paid the wrong amount, in both directions, for reasons that federal officials attributed mainly to agency and household mistakes rather than fraud. The same rate, as a published ranking attached to money and reputation, told state agencies what their most important task was. The eligibility rules stated the law. The error rate stated the priority, and state practice followed the priority.

What the Integrity Figures Actually Measure

Two figures dominate public argument about the Supplemental Nutrition Assistance Program, and the two figures measure different things. One figure, the payment error rate, records how accurately state agencies determined eligibility and calculated benefit amounts in sampled cases. The other figure, the trafficking rate, estimates how much of the benefit stream was diverted from food purchase into cash or ineligible goods through exchanges between recipients and retailers. The distinction carries the central claim of the administration record: error is not fraud. A payment error is a finding about a case file, reached through quality control review, and the finding holds whether the mistake helped the household or harmed the household, whether a caseworker caused the mistake or a household caused the mistake, and whether anyone intended anything at all. Trafficking is a finding about conduct, requiring an exchange, a willing recipient and a willing retailer, and a purpose outside the statute. Adding the two rates together, or quoting one rate as though the rate described the other phenomenon, produces a number that corresponds to no measurement any agency performed.

The payment error rate comes out of the quality control system that the Food Stamp Act of 1977 established. Under that system, each state agency drew a monthly sample of participating households and rechecked eligibility and benefit level against the program rules. The Food and Nutrition Service then reviewed a subsample of the state reviews to validate the findings, and the combined results produced state error rates and a national rate. An error entered the rate when a household received more than the household was entitled to receive, an overpayment, or less than the household was entitled to receive, an underpayment, above a small dollar threshold set in regulation. Errors in denied applications and terminated cases fell outside the headline rate. The Congressional Research Service, in a report dated September 24, 2018, stated the consequence directly: unlike the retailer trafficking rate, the national payment error rate is not a measure of fraud, because the rate reflects improper payments without identifying the cause. A caseworker who misread a pay stub, a household that failed to report a wage change on time, and a deliberate misstatement all landed in the same statistical category when the dollar result was wrong.

The rate itself had a documented history. The national combined rate reached a record low of 3.20 percent in fiscal 2013 and stood at 3.66 percent in fiscal 2014, according to Food and Nutrition Service quality control releases. The Congressional Research Service decomposition of the fiscal 2014 rate assigned 2.96 percentage points to overpayments and 0.69 percentage points to underpayments, a split that showed the smaller component of the headline figure running against households rather than against the Treasury. Publication of a national rate was suspended for fiscal 2015 and fiscal 2016 during a dispute over the integrity of state review data, and publication resumed at 6.30 percent for fiscal 2017 and 6.80 percent for fiscal 2018. The Congressional Research Service cautioned readers in September 2018 that the apparent jump from the fiscal 2014 rate to the fiscal 2017 rate could mislead, because the department had applied new controls to the fiscal 2017 data collection that had not governed the earlier estimates. The series therefore supported statements about accuracy within a stated method and a stated year, and the series did not support a claim that a stated percentage of benefits in any of those years was stolen, diverted, or obtained by deception.

The content of the errors reinforced the same point. Analyses of quality control data by the Food and Nutrition Service and by the Congressional Research Service attributed most errors to the mechanics of administering a program whose benefit formula turned on fluctuating earnings, household composition, shelter costs, and utility allowances, verified through documents that households and agencies exchanged under reporting deadlines. Agency errors included miscalculated expenses, misapplied policy, and data entry failures. Household errors included late or incomplete reports of income changes, often involving small amounts and irregular work schedules. Underpayments mattered to the interpretation of the rate in a way that fraud language could not accommodate: an underpayment was an error in which the government shorted an eligible household, counted in the same headline figure as an overpayment. Dottie Rosenbaum of the Center on Budget and Policy Priorities, in an analysis dated July 6, 2015, reported that state data attributed about half of overpayments and almost 90 percent of underpayments in the years the analysis covered to agency fault, with most of the remainder traced to household mistakes rather than deception. The recovery rules applied regardless of cause: an agency that identified an overpayment established a claim against the household and collected the excess, and an agency that identified an underpayment owed the household the shortfall.

Fraud by a recipient required a separate and narrower finding, called an intentional program violation. Federal regulations defined the violation around intent: a false or misleading statement, a concealed fact, a traded benefit, or another act committed knowingly to obtain or increase benefits. Intent separated the violation from the far larger volume of inadvertent household error, and procedure enforced the separation. A state agency pursuing an intentional program violation had to give notice and prove the charge at an administrative disqualification hearing before a hearing officer, unless the individual waived the hearing in writing or a court convicted the individual of the underlying conduct. Only after one of those routes produced a finding did disqualification follow. The scale of that enforcement track appeared in department reporting for fiscal 2011: state agencies completed nearly 798,000 recipient fraud investigations in that year and obtained just over 46,000 disqualifications. The investigation total counted inquiries opened, not violations proven, and the disqualification total counted the subset in which a state established intent through the required process or a court entered a conviction.

Trafficking occupied a different legal category again. Federal law defined trafficking as the buying, selling, stealing, or otherwise effecting an exchange of benefits for cash or for consideration other than eligible food, and the Food and Nutrition Act made the exchange a federal crime for the recipient who surrendered the benefit and for the retailer who redeemed the benefit. The standard transaction involved a discount: a store credited a benefit account for a stated amount of food that never changed hands, paid the cardholder a smaller amount in cash, and kept the difference when the department reimbursed the full redemption. Variants exchanged benefits for alcohol, tobacco, or other ineligible items rather than cash. Regulations the Food and Nutrition Service published in February 2013 expanded the regulatory definition to reach schemes that the older cash centered wording had missed, including the purchase of a product with benefits for the purpose of discarding the contents and returning the container for a cash deposit, and the resale of food bought with benefits. In each variant the defining element was conversion: value issued for food left the food economy entirely.

The department measured that conversion through periodic studies rather than through the quality control sample, and the two mandated estimates anchored the series at the vantage date. Mantovani and Wilson, in a study the Food and Nutrition Service published in 2011, estimated trafficking at about 1.0 percent of benefits for 2006 to 2008, diverting an estimated 330 million dollars annually. Mantovani, Williams, and Pflieger, in a study the Food and Nutrition Service published in August 2013, estimated trafficking at 1.3 percent of benefits for 2009 to 2011, diverting an estimated 858 million dollars annually against average redemptions of 63.7 billion dollars a year across that period. The same studies estimated the share of authorized stores that engaged in trafficking at 8.2 percent for 2006 to 2008 and 10.5 percent for 2009 to 2011. The gap between the store figure and the benefit figure carried information of its own: trafficking stores tended to redeem small volumes, so roughly one store in ten in the later estimate accounted for roughly one benefit dollar in seventy seven. The 2013 study also framed the fiscal meaning of diversion with care, observing that trafficking did not add to federal outlays in the manner of an overpayment, because the benefit had already been issued; the loss was a loss of purpose, as funds appropriated to purchase food purchased something else instead.

Earlier estimates placed the electronic era figures in a longer arc. The August 2013 study, presenting the series on a consistent exhibit, recorded a trafficking rate of 3.8 percent for 1993 and 3.5 percent for 1996 to 1998, periods in which benefits circulated as paper coupons across much of the country. A paper coupon passed from hand to hand without generating a record of the transaction, and a coupon redeemed at a store entered the banking system as a deposit largely detached from any individual purchase. Detection in the coupon era depended heavily on covert in store investigations, and the earliest national estimates rested on that investigative record alone. The decline from 3.8 percent in 1993 to about 1.0 percent in the 2006 to 2008 study period coincided with the replacement of coupons by electronic issuance, a transition that state programs completed nationwide in June 2004 under the deadline structure the 1996 welfare statute had set.

Electronic benefit transfer changed both the crime and the capacity to measure the crime. Each electronic transaction generated a record identifying the card, the store, the amount, and the time, and redemption patterns that had been invisible in the coupon system became legible in aggregate. A store whose redemptions arrived in repeated round dollar amounts inconsistent with grocery pricing, or whose benefit volume vastly exceeded the benefit volume of comparable stores in the same market, presented a pattern that analysts could flag without any investigator entering the premises. A 1996 statutory amendment gave the Food and Nutrition Service authority to disqualify a retailer on the basis of electronic transaction evidence alone, subject to the store opportunity to offer a legitimate explanation, and the Government Accountability Office recorded that authority in a report dated August 2014. The department organized the data flow through the Anti-Fraud Locator Using Electronic Benefits Transfer Retailer Transactions system, known by the acronym ALERT, which screened redemption data for patterns associated with trafficking and directed investigators toward high risk stores. The trafficking studies then combined three evidence streams: covert in store investigations, investigations by the department Office of Inspector General and state law enforcement agencies, and administrative determinations drawn from suspect transaction patterns. Because investigated stores were selected for suspicious behavior rather than drawn at random, the studies adjusted the investigative outcomes statistically to represent the full population of authorized stores, and the published rates rested on that adjustment rather than on a raw count of proven cases.

The pattern data also located the conduct. In the 2009 to 2011 study, small stores, a category composed largely of small and medium groceries and convenience stores, accounted for about 15 percent of all redemptions and an estimated 85 percent of trafficking redemptions. Large stores showed the opposite profile in the 2006 to 2008 study, accounting for 87.3 percent of redemptions and about 5.4 percent of trafficking. Trafficking appeared more often among privately owned stores than publicly owned stores and more often in higher poverty neighborhoods than in lower poverty neighborhoods, according to both studies. Those distributions explained why a store violation rate near one in ten coexisted with a benefit trafficking rate near one in a hundred.

Retailer authorization supplied the fulcrum on which that enforcement rested. A store could redeem benefits only after the Food and Nutrition Service authorized the store, on an application that required the store to meet stocking requirements for staple foods or to derive a majority of sales from staple foods, and the Government Accountability Office reported in May 2018 that authorizations generally ran for five years before reauthorization, across a population of more than 263,000 authorized retailers in fiscal 2017. Authorization created leverage that household enforcement lacked. The department dealt with each retailer directly, held the retailer redemption data centrally, and could end retailer participation by administrative action, whereas recipient enforcement ran through state agencies case by case. A retailer found to have trafficked faced permanent disqualification, with a civil money penalty available in lieu of permanent disqualification only in defined circumstances, including circumstances in which the department found that the firm had established effective compliance policies and the ownership had neither known of nor benefited from the violation. Lesser violations, such as selling ineligible items without a trafficking exchange, drew term disqualifications, fines, or warnings scaled to the violation. Store ownership answered for violations committed by employees, a rule the authorization application stated expressly.

Recipient penalties operated on the individual rather than the household and escalated by offense under the disqualification schedule the Food and Nutrition Act prescribed. A first intentional program violation brought twelve months of ineligibility for the individual who committed the violation, a second brought twenty four months, and a third brought permanent ineligibility. Court convictions for the gravest trafficking conduct carried heavier consequences in the same schedule, including permanent ineligibility for a conviction involving benefits with a value of 500 dollars or more and for exchanging benefits for firearms, ammunition, or explosives. Criminal liability ran on a separate track: prosecutors in federal or state court could pursue fines and imprisonment scaled to the value trafficked, and a conviction could support disqualification without a separate administrative hearing. The schedule drew its boundary at intent throughout. A household that received too much because a caseworker misapplied a deduction rule, or because a wage report arrived late without deceptive purpose, owed repayment through a claim and faced no disqualification period at all, because no tribunal had found an intentional violation.

Program officials and federal oversight bodies presented the integrity case at full strength on that record. The Department of Agriculture argued, in testimony and program materials across the period the studies covered, that a rate near one percent applied to a program redeeming 63.7 billion dollars a year in the 2009 to 2011 period meant 858 million dollars a year diverted from food purchase, a sum that no steward of appropriated funds could treat as trivial, and that visible enforcement against trafficking stores protected the legitimacy of benefits for households that used benefits as directed. The Government Accountability Office, in testimony delivered in September 2010, framed the same arithmetic from the oversight side, observing that payment errors and trafficking together placed more than a billion dollars a year at issue in a program that had paid about 50 billion dollars in benefits in fiscal 2009, and the office pressed the department to strengthen retailer monitoring and penalty use even as the office acknowledged that both rates had declined. On that view, the movement from 3.8 percent in 1993 to about 1.0 percent in 2006 to 2008 demonstrated that sustained enforcement worked.

The access case, presented at full strength by analysts who defended the program design, answered that the same record showed diversion to be rare relative to the scale of lawful use and showed the accuracy system to be misdescribed when error language borrowed the vocabulary of crime. The Center on Budget and Policy Priorities, in the July 2015 analysis by Rosenbaum, described the quality control system as among the most rigorous payment accuracy regimes in the federal government and attributed the overwhelming majority of errors to honest mistakes by recipients, eligibility workers, and data entry clerks, with states themselves at fault for about half of overpayments and almost 90 percent of underpayments in the data the analysis cited. On that view, a combined error rate that counted a caseworker underpayment of an eligible household as program failure in the same figure as an overpayment could not support rhetoric about recipients taking money that did not belong to them, because a material share of the figure recorded the government failing households rather than households deceiving the government. The trafficking studies supported the rarity claim on their own terms: 1.3 percent of benefits for 2009 to 2011 meant that 98.7 percent of benefits in that period moved through authorized stores without diversion, and the concentration of trafficking in a small store sector showed a retailer conduct problem more than a household conduct pattern.

The two positions drew on the same published measurements, and the measurements constrained both. The error rate answered a question about administrative accuracy in sampled cases, in both directions, for a stated fiscal year. The trafficking rate answered a question about diversion of issued benefits, for a stated multi year period, through a method built on investigations and transaction patterns. Neither rate measured the other phenomenon, the trafficking estimate did not enter the payment error rate, and the payment error rate contained no finding about intent. A debate conducted inside those boundaries could dispute enforcement budgets, penalty severity, or tolerance for administrative error. A debate conducted outside those boundaries, in which a 6.80 percent error rate for fiscal 2018 became a claim that 6.80 percent of benefits were trafficked, described a program that no federal measurement had ever reported.

The Three-Month Clock and Its Waivers

The time limit known in administration as the ABAWD rule entered the food assistance statute in the Personal Responsibility and Work Opportunity Reconciliation Act of 1996, enacted on August 22, 1996. Congress placed the rule in section 6(o) of the Food Stamp Act, carried forward in the Food and Nutrition Act of 2008 and codified at 7 U.S.C. 2015(o). Section 6(o) added a calendar to the ordinary eligibility tests. For a defined class of participants, eligibility itself expired after a counted number of months unless a work condition was met, an exception applied, a waiver covered the place of residence, or a state agency spent one of a limited pool of individual exemptions. Administration of the Supplemental Nutrition Assistance Program therefore included, for this one class, the management of time as an eligibility fact.

The class was defined by age, household composition, and capacity for work. In the terms of section 6(o) as in force on December 1, 2019, the limit applied to an individual who was at least 18 years of age and not yet 50, who was not medically certified as physically or mentally unfit for employment, who was not a parent or another member of a household with responsibility for a dependent child, who was not residing in a household with a member under 18 years of age, who was not pregnant, and who was not otherwise exempt from the general work requirements. Program usage shortened the description to able-bodied adult without dependents, abbreviated as ABAWD. Household composition was tested against the household as certified, and capacity for work against the medical certification standards at 7 C.F.R. 273.24. A change in any element, such as a child joining the household, a pregnancy, or a medical certification, moved an individual out of the counted class for the months in which the circumstance applied.

The operative rule, stated in section 6(o)(2), barred participation by an individual who had received benefits for three months, consecutive or otherwise, within a 36-month period during which the individual did not meet the work requirement attached to the limit, did not receive benefits under a waiver or an individual exemption, and was not within an exception. Food and Nutrition Service guidance, including the agency’s May 2018 clarification memorandum on work requirements, restated the limit as three full months in a 36-month period. The word full carried weight. A month counted only where the individual received benefits for the full month while subject to the limit and out of compliance. A month in which benefits were prorated, as at application in the middle of a month, did not count, and neither did a month in which the individual met the requirement or a month covered by an exception, an area waiver, or a discretionary exemption. Case records therefore distinguished a countable month from every other month of receipt. At the close of the third countable month, a participant who remained in the counted class and out of compliance lost eligibility as a member of the household, although other household members retained whatever eligibility the household rules gave them.

The work requirement was measured in hours. Under section 6(o)(2), an individual satisfied it by working 20 hours or more per week, averaged monthly, by participating in and complying with a work program for 20 hours or more per week as determined by the state agency, by participating in and complying with a workfare program under section 20 of the act, codified at 7 U.S.C. 2029, or a comparable state or local program, or by a combination of work and program participation reaching the same total. The Food and Nutrition Service expressed the monthly equivalent as 80 hours per month. Work included paid employment and self-employment, and program guidance recognized unpaid and in-kind work. A work program, as defined in section 6(o)(1), included programs under federal workforce legislation, state or local employment and training programs meeting standards approved by the governor, veterans’ employment and training programs approved by the Secretary of Agriculture, and the employment and training program under section 6(d)(4) of the act itself. Workfare stood on a separate footing: compliance with an assignment for the hours set under a formula tied to the benefit amount and the applicable minimum wage satisfied the requirement for that month.

Loss of eligibility was not permanent within the 36-month period. Under section 6(o)(5), an individual who had lost eligibility regained it by meeting the work requirement during a 30-day period, by complying with a workfare assignment for a month, or by becoming excepted from the limit, and eligibility then continued for as long as the individual met the requirement month by month or remained excepted. The 30-day route was measured in total hours, so 80 hours compressed into a shorter span within the 30 days satisfied the test. Regaining eligibility through work carried a further consequence stated in section 6(o)(5) and elaborated at 7 C.F.R. 273.24. An individual who regained eligibility in that manner and then ceased to meet the requirement became eligible for one additional period of up to three consecutive months without meeting the requirement. That period was available once in a 36-month period, its months ran consecutively once the period began, and prorated months did not consume it. After the additional period ended, participation in a month depended on meeting the requirement in that month, on an exception, on a waiver, or on a discretionary exemption. No second additional period existed within the same 36 months.

The statute kept the ABAWD rule and the general work requirements separate in structure, population, and consequence, and public discussion persistently merged the two. The general requirements, stated in section 6(d) and the regulations at 7 C.F.R. 273.7, applied to most participants from 16 years of age through 59 years of age who were not exempt. A work registrant had to register for work at application and at intervals set by the state agency, accept suitable employment if offered, participate in an employment and training program or a workfare assignment when the state agency made such an assignment, furnish information sufficient for the agency to determine employment status and availability for work, and refrain from voluntarily quitting a job or reducing work effort below 30 hours per week without good cause. Failure to comply without good cause produced a period of disqualification, graduated by prior disqualifications, after which compliance restored eligibility. The general regime imposed no limit on the total number of months a registrant could receive benefits: registration, job search, and acceptance of suitable work were continuing obligations, not a ration of months.

The ABAWD rule was layered on top of that structure for the narrower class in section 6(o), and every ABAWD was also a work registrant unless an exemption removed the individual from both regimes. The age bands differed, running from 16 to 59 for registration and from 18 to 49 for the time limit, and the hour tests differed, since the time limit demanded an affirmative 80 hours per month from an individual who might have no job and no assigned program slot. The consequences differed most sharply: noncompliance with the general requirements disqualified an individual for a set period tied to the violation, while exhaustion of three countable months ended eligibility for the balance of the 36-month period subject only to the regaining and additional-month provisions. A waiver of the time limit removed the clock for residents of the waived area, and Food and Nutrition Service guidance stated expressly that such a waiver left the general requirements in force, because the authority to waive section 6(o) did not extend to section 6(d).

The employment and training structure gave the general requirements their institutional form and gave some ABAWDs a route to the 80-hour test. Under section 6(d)(4), each state agency operated a SNAP employment and training program under a plan approved by the Food and Nutrition Service, with components that included job search, workfare, work experience, education, and vocational training, supported by federal grants and matching funds. Participation was mandatory for registrants whom the state agency required to participate, and states retained discretion over the scale of mandatory participation and over voluntary places. Participation in a component that qualified as a work program counted toward the ABAWD hour total, so an assignment measured at 20 hours per week satisfied both regimes at once. The statute did not guarantee an ABAWD a program slot as a condition of the clock, and a separate provision that rewarded states pledging to offer a qualifying work opportunity to ABAWDs in the last month of the three-month period operated as an incentive in the funding formula rather than an entitlement held by the individual.

Exceptions to the time limit were stated at two levels of the statute. Section 6(o)(3) excepted an individual under 18 or over 49 years of age, an individual medically certified as physically or mentally unfit for employment, a parent or other member of a household with responsibility for a dependent child, a pregnant woman, and an individual otherwise exempt under section 6(d)(2). The regulations at 7 C.F.R. 273.24(c) restated the exceptions for administration, including residence in a household in which a member was under 18 years of age, and defined medical certification by reference to receipt of temporary or permanent disability benefits from governmental or private sources, obvious physical or mental unfitness as determined by the state agency, or, where unfitness was not obvious, a statement from a physician or other medical personnel designated in the regulation. State agency guidance expressed the medical test in hours, as physical or mental inability to work at least 80 hours a month. Because section 6(o)(3) incorporated the exemptions in section 6(d)(2), the general exemption list also determined ABAWD status. That list, restated at 7 C.F.R. 273.7(b), covered a person caring for a dependent child under six years of age or for an incapacitated person, a person complying with the work requirements of another program such as unemployment compensation or the Temporary Assistance for Needy Families program, a person who had applied for or was receiving unemployment compensation, a regular participant in a drug addiction or alcoholic treatment and rehabilitation program, a person employed or self-employed for at least 30 hours per week or with weekly earnings at least equal to the federal minimum wage multiplied by 30 hours, and a student enrolled in a recognized school, training program, or institution of higher education on at least a half-time basis.

The waiver system addressed place rather than person. Under section 6(o)(4), on the request of a state agency, the Secretary of Agriculture could waive the time limit for any group of individuals in a state where the Secretary determined that the area in which the individuals resided had an unemployment rate of over 10 percent or did not have a sufficient number of jobs to provide employment for the individuals. The regulations at 7 C.F.R. 273.24(f) converted the two statutory tests into evidence standards. For the unemployment test, a state agency could show a 12-month average unemployment rate over 10 percent for the area, a three-month average unemployment rate over 10 percent, or a historical seasonal unemployment rate over 10 percent. For the insufficient jobs test, the regulation recognized designation of the area as a Labor Surplus Area by the Employment and Training Administration of the Department of Labor, a determination by the Department of Labor that the area qualified for extended unemployment benefits, a low and declining employment to population ratio, a lack of jobs in declining occupations or industries, a description of the area in an academic study or other publication as an area lacking jobs, and a 24-month average unemployment rate at least 20 percent above the national average for the same 24-month period. Requests were framed for areas as states defined them for this purpose, including counties and groups of counties, cities, towns, reservations, and areas grouped with reference to Labor Market Areas as delineated for federal labor statistics, with the evidence computed for the area as grouped. Approvals ran for a stated period, ordinarily up to 12 months, and renewal required a fresh showing under the same standards.

Alongside area waivers, section 6(o)(6) gave each state a limited pool of individual exemptions, allotted as a percentage of the state’s covered individuals as estimated under the statutory formula. One exemption covered one individual for one month. The Balanced Budget Act of 1997, enacted on August 5, 1997, established the allotment at 15 percent, and exemptions at that level applied through fiscal 2019, which ended on September 30, 2019. Section 4005 of the Agriculture Improvement Act of 2018, enacted on December 20, 2018, reduced the allotment to 12 percent of covered individuals for fiscal 2020 and each fiscal year thereafter, and fiscal 2020 began on October 1, 2019. The Food and Nutrition Service communicated state allotments by memorandum each fiscal year, and states reported use on federal reporting forms. State agencies chose the grounds on which exemptions were spent, and agency guidance described uses that included individuals in places with weak labor markets outside waiver boundaries and individuals with barriers to work that did not meet an exception category. Exhaustion of the pool in a fiscal year left the clock, the exceptions, and the waiver map as the only sources of relief.

Administration of the clock was a case management and records problem. A state agency had to identify every participant who met the ABAWD definition, screen for exceptions at application and recertification and on reported changes, record hours from pay records, employer statements, and program attendance systems, decide for each month whether the month was countable, track the running total against the 36-month period, apply waivers according to the participant’s address, and record each discretionary exemption spent. The statute fixed the period at 36 months without prescribing a single national method for fixing its start, and Food and Nutrition Service guidance recognized three state methods: a fixed statewide period running on the same dates for every participant, a fixed individual period beginning at application or at the date the participant became subject to the limit, and a rolling individual period recalculated month by month by looking back over the preceding 36 months. Movement between states compounded the task, since the receiving state had to verify countable months used in another state under the certification rules at 7 C.F.R. 273.2 and reconcile those months with its own method. The rule’s bite therefore depended on state tracking systems and waiver geography: a participant in a waived county accrued no countable months, however few hours the participant worked, while a participant with an identical work history outside the waiver map exhausted eligibility in three months.

The historical operation of the rule divided into distinct phases marked by national unemployment. From implementation under the 1996 statute, the clock operated with area waivers concentrated in places meeting the regulatory evidence standards. The recession that began in December 2007 changed the scale of coverage as states qualified for extended unemployment benefits, a regulatory route to statewide waivers. Congress then suspended the time limit itself in section 101(e) of the American Recovery and Reinvestment Act of 2009, enacted on February 17, 2009, which prohibited application of section 6(o) to limit participation during the period from April 1, 2009 through September 30, 2010, with a qualification for an individual who failed to comply with a qualifying work program or workfare opportunity that a state agency offered. Food and Nutrition Service guidance on the Recovery Act confirmed that statewide waivers were unnecessary during the suspension because the limit was inoperative on a national basis, and that, from October 1, 2010, state agencies were to disregard periods of benefit receipt before that date in applying section 6(o), which restarted the count for participants whose months would otherwise have carried across the suspension.

The return of the clock after September 30, 2010 was gradual and state by state. For fiscal 2011 through fiscal 2013, eligibility for extended unemployment benefits kept statewide waivers available to the great majority of states, and most state agencies took them, so the limit remained inoperative across most of the country in practice even though the statutory suspension had ended. As extended benefit triggers expired, the number of states qualifying for statewide waivers contracted during fiscal 2014 and fiscal 2015, and waiver maps within states narrowed toward counties and grouped areas that met the 10 percent test or one of the insufficient jobs standards. Some state agencies also declined waivers for areas that qualified, restoring the clock by choice. Fiscal 2016 marked the broadest reinstatement, as the end of extended benefit eligibility removed the statewide route for most states at once and state agencies rebuilt tracking, notices, and worker training for a rule that had been dormant, in many places, since 2009. By fiscal 2017, the Food and Nutrition Service waiver status reports showed a divided map of states with statewide waivers, states with waivers for part of the state, and states with no waivers, a pattern that persisted in the reports through fiscal 2019.

Delivery in Practice and the People the Program Misses

Approval did not end the administrative relationship between a household and the Supplemental Nutrition Assistance Program. Approval began a repeating cycle of issuance, reporting, and renewal, and the design of that cycle determined whether benefits arrived steadily or stopped for reasons that had little connection to need. The statute defined eligibility in the abstract. Delivery in practice depended on a calendar, a plastic card, a set of reporting duties scaled to the stability of a household, and a recertification process that required the household to prove eligibility again at fixed intervals. A household could remain poor, remain within every income and resource limit, and lose benefits because a form did not arrive, an interview was missed, or a verification document reached the agency after a deadline. The administrative record through December 1, 2019, treated that pattern as a central fact of program operation rather than as an exception. The same record showed a second gap at the front door. A substantial share of the people the statute made eligible never entered the program at all, or entered late, and the Food and Nutrition Service measured that gap each year in its participation rate series.

Monthly issuance ran through electronic benefit transfer. By June 2004, every state agency had completed the conversion from paper coupons to electronic issuance required under the timetable set by the Personal Responsibility and Work Opportunity Reconciliation Act of 1996, and after that conversion the monthly allotment reached the household as a credit in an electronic benefit transfer account, accessed with a card at authorized retailers. The credit did not arrive on a single national date. Each state agency set an issuance schedule, and most state agencies staggered issuance across a span of days within the month, assigning households to an issuance date by a case number, a Social Security number digit, an application date, or a similar administrative key. Staggering served operational purposes that the Food and Nutrition Service recognized in its guidance. A single issuance day concentrated retailer demand and overloaded agency telephone systems in the first days of a month, while a staggered calendar spread food purchasing and agency workload across the month. Federal regulation bounded the discretion. Under the issuance rules codified at section 274.2 of Title 7 of the Code of Federal Regulations, a state agency could not allow more than 40 days to elapse between allotments to a household participating for longer than two consecutive months, so the staggered date had to remain a dependable monthly rhythm rather than an irregular payment stream.

The length of time an approval lasted was the certification period, and federal regulation tied its length to the predictability of household circumstances. Section 273.10 of Title 7 required the state agency to assign the longest certification period possible based on that predictability, stated that households should receive certification periods of at least six months unless household circumstances were unstable or the household contained an adult subject to the time limit for able bodied adults without dependents, and capped the ordinary period at 12 months. The same regulation allowed a period of up to 24 months for a household in which all adult members were elderly or had a disability, with at least one agency contact in each 12 month span. In state practice under that framework, six month and 12 month periods accounted for most of the caseload. Households with earnings that fluctuated, with irregular shelter arrangements, or with other unstable circumstances drew shorter periods, because the agency expected the facts on which eligibility rested to change and set an earlier date to test them again. Households composed of older adults or adults with disabilities, with fixed unearned income and no earnings, drew the longest periods the regulation permitted, because the agency expected the underlying facts to hold. The choice was an administrative judgment with distributive effects. A longer period reduced the number of times a household had to reestablish eligibility and reduced the number of agency actions required to keep a stable case open. A shorter period gave the agency more frequent control points and gave the household more frequent opportunities to fail a procedural step.

Between approval and expiration, a household carried reporting duties, and by the vantage date of December 1, 2019, most households carried those duties under simplified reporting. The Farm Security and Rural Investment Act of 2002 authorized the simplified reporting option, and state agencies adopted the option across the program in the years after 2002. Under simplified reporting, a household reported mid period only in defined circumstances, principally when total gross monthly income rose above 130 percent of the federal poverty guideline for the household size, and when the work hours of a member subject to the able bodied adult time limit fell below 20 hours per week. A household certified for longer than six months under simplified reporting also filed a periodic report, an interim accounting of household circumstances due at an interval the state agency set between four and six months into the period, on which the agency acted to adjust the allotment prospectively. Households that remained under change reporting carried a heavier duty. A change reporting household reported specified changes in income, household composition, residence, and resources within 10 days of the change, and the agency acted on the report during the certification period rather than waiting for a scheduled accounting. Monthly and quarterly reporting systems, which had required reports on a fixed short cycle, survived in limited settings by 2019, and the contrast explained the appeal of the simplified design. Simplified reporting traded continuous surveillance for a small number of consequential reports, reducing both agency workload and the count of deadlines a household could miss, while preserving an agency duty to act on information the household supplied.

Recertification stood at the end of each period, and in substance the process resembled a fresh application. As expiration approached, the state agency sent a notice of expiration, and the household filed an application for recertification, completed an interview, and supplied verification of income, identity, residence, and deductible expenses sufficient for the agency to redetermine eligibility and compute a new allotment. A household that filed and cooperated on schedule received the first allotment of the new period on the normal issuance cycle, without a gap. A household that filed late, missed the interview, or left verification incomplete faced closure at the end of the period, even where the underlying facts would have supported continued eligibility. The distinction between a denial on the merits and a closure for a procedural reason mattered in the administrative data. A merits denial recorded a determination that the household failed an eligibility rule. A procedural closure recorded a failure to complete the process, and the record often did not establish whether the household had remained eligible on the facts. Procedural reasons stood prominent among closures and application denials in the administrative record, including failure to provide verification, failure to complete an interview, and failure to return a periodic report or recertification application by its deadline. The household bore the burden of curing the failure, ordinarily by filing again as an applicant, and the agency bore the cost of processing a new application for a household served in the prior month.

Researchers gave that cycle of exit and return a name. Churning described a case that left the program and reentered within a short span, defined by the Food and Nutrition Service for study purposes as reentry within four months or less. In a study prepared for the Food and Nutrition Service by Gregory Mills and colleagues at the Urban Institute, published in 2014 under the title Understanding the Rates, Causes, and Costs of Churning in the Supplemental Nutrition Assistance Program, the estimated annual churn rate for fiscal 2011 ranged from 17 percent to 28 percent across the six states studied, measured as the share of households receiving benefits at any point in the year that experienced a churn spell. The Mills study found that most exits associated with churn occurred at scheduled recertifications or required interim reports, and Mills testified to the Subcommittee on Nutrition of the House Committee on Agriculture on February 26, 2015, that the causes of churn rested primarily in procedural difficulties experienced by participants rather than in fluctuations in earnings. The study also priced the cycle. Each churn episode imposed forgone benefits on households that remained eligible during the spell off the program, time and expense on households that had to reenter, and added administrative cost on federal and state governments, because processing a reapplication cost an agency more than completing a timely recertification. Churn therefore converted a paperwork failure into a measurable interruption of food assistance and a measurable inefficiency, and state agencies that lengthened certification periods, simplified interim reports, or improved recertification notices acted on churn as an administrative defect within the control of the delivery system rather than as evidence about household need.

The delivery system also carried a surge mechanism for disasters, the Disaster Supplemental Nutrition Assistance Program, known as Disaster SNAP. Authority for disaster food assistance rested on the Food and Nutrition Act and on the Robert T. Stafford Disaster Relief and Emergency Assistance Act, and operation followed a presidential disaster declaration for individual assistance. A state agency requested Food and Nutrition Service approval, demonstrated that commercial food distribution channels in the affected area had been restored, and opened a short application period for households that had lived or worked in the disaster area and had suffered disaster related expenses or income losses. Disaster SNAP applied disaster specific income limits and deductions rather than the ordinary eligibility rules in full, reached households that did not participate in the regular program, and ordinarily issued a benefit set at the maximum allotment for the household size for a defined period. The Department of Agriculture announcement of September 22, 2017, approving Disaster SNAP for Florida counties after Hurricane Irma, stated that eligible households would receive two months of benefits equivalent to the maximum amount for a household of that size, and Texas opened Disaster SNAP applications on September 13, 2017, for counties covered by the federal disaster declaration after Hurricane Harvey. Operations after Hurricanes Harvey, Irma, and Maria in the 2017 Atlantic hurricane season ran the mechanism at a scale the regular monthly cycle never faced, enrolling disaster households through temporary sites in compressed periods while the regular caseload in the same states received replacement benefits for food lost in the storms and waivers of ordinary rules, including permission to purchase hot foods. In a disaster, the same administrative system prized speed and breadth of intake over the verification sequence that governed ordinary certification, on the judgment that delayed food assistance after a disaster failed the purpose of the authority.

The participation rate measured the people the front door missed. In the report Trends in Supplemental Nutrition Assistance Program Participation Rates, Fiscal Year 2010 to Fiscal Year 2016, issued by the Food and Nutrition Service in July 2018, the program served 85 percent of eligible individuals in fiscal 2016, on an average monthly estimate of 47 million eligible individuals and nearly 40 million participants. The aggregate rate concealed sharp differences among groups defined in the same report. The participation rate for eligible individuals in households with earnings stood at 75 percent in fiscal 2016, and the rate for the eligible working poor, a group the Food and Nutrition Service defined by income in relation to poverty among people in working households, also stood at 75 percent in fiscal 2016. The rate for eligible older adults stood at 45 percent in fiscal 2016, a figure the report presented in contrast with the overall rate and consistent with the pattern of lower older adult participation in each year of the series from fiscal 2010 forward. A rate of 45 percent meant that more than half of the individuals aged 60 or older whom the federal income and asset rules made eligible received no benefit in an average month of fiscal 2016. State rates for all eligible persons in fiscal 2016 ranged from 56 percent to 100 percent in the companion Food and Nutrition Service report Reaching Those in Need, a spread that demonstrated how much of the gap closed or widened under state administration of a single federal statute.

The research literature on nonparticipation, reviewed by Craig Gundersen and colleagues in a 2011 review of the economics of food insecurity and documented in surveys conducted for the Food and Nutrition Service by Susan Bartlett and colleagues in 2004, organized the reasons into a small number of recurring categories. Stigma operated in more than one form. Some eligible people held a personal distaste for receiving food assistance, some anticipated a negative reaction from agency staff or from other shoppers, and some treated reliance on benefits as a loss of independence, a reason surveys of older adults recorded with particular frequency. In the survey evidence summarized by Bartlett and colleagues, only a minority of eligible nonparticipants named stigma as the single most important reason for staying out, while nearly half answered positively to at least one question about stigma perceptions, a combination that placed stigma as a background condition on participation rather than as a sole explanation. Transaction costs operated directly. Applying required time away from work, travel to an office, gathering documents, completing forms, and waiting for an interview, and participation required repeating parts of that effort at reporting and recertification points. For a household weighing those costs, the expected benefit decided whether the effort appeared worthwhile, and eligible households near the income limits often qualified for a small allotment. The Food and Nutrition Service participation series recorded the gradient in fiscal 2016. Individuals in households with incomes above the poverty line but within eligibility limits participated at far lower rates than individuals in households with no income or with incomes deep below poverty, and a household eligible only for the minimum benefit faced the weakest arithmetic case for applying. Lack of awareness formed the third category. In the Bartlett survey of eligible nonparticipants conducted in 2000 and 2001, 69 percent of respondents stated that the respondent would apply upon knowing of eligibility, a finding that located a large share of the gap in information rather than in refusal. Office access compounded each category in particular places. Distance to a local office, office hours that conflicted with work schedules, limited transportation, and difficulty reaching an agency by telephone raised the cost of every step, and older adults faced added barriers that Gabor and colleagues identified in a 2002 study of older adult participation, including misinformation about eligibility, transportation limits, and communication barriers for people with limited English proficiency.

State agencies worked on the gap with levers the statute and regulations placed in state hands, and the Food and Nutrition Service treated those levers as program access policy rather than as local custom. Online applications and online case access allowed a household to file and to check a case without travel to an office during office hours, and telephone application and interview options, including call centers that handled applications, changes, and inquiries for a county or a state, separated the application from a single physical location. Outreach operated under state outreach plans that the Food and Nutrition Service approved, with federal reimbursement available for allowable outreach costs at the administrative matching structure that governed state administrative spending, and outreach partnerships with community organizations carried eligibility information to places where eligible nonparticipants already went, including food pantries, senior centers, and health clinics. Longer certification periods and simplified reporting, adopted as workload policies, functioned as participation policies in effect, because each removed deadlines at which eligible households left the rolls. Broad based categorical eligibility operated as the most structural of the levers. Under that option, a state agency conferred categorical eligibility on households receiving a service funded by the Temporary Assistance for Needy Families program or a state maintenance of effort program, and most states exercising the option set a gross income limit above the ordinary 130 percent level and eliminated the asset test for households covered by the policy. The option widened the doorway for working households with earnings above the ordinary gross limit, for households with modest savings, and for older adults whose resources would otherwise have required an asset determination, and the participation differences among states in fiscal 2016 reflected, in part, different combinations of these choices operating on similar populations.

The synthesis joined the two movements. Accuracy dominated public argument about administration, and the payment error rate supplied a national measure of whether agencies paid the right amount to the right household. Delivery supplied a second measure that argument addressed less often. A program that paid with perfect accuracy to the households that completed every procedural step, while losing eligible households at recertification in the range the Mills churn study documented and reaching 45 percent of eligible older adults in fiscal 2016, administered the statute to a population smaller and differently composed than the population the statute defined. Administration decided both questions at once. The length of a certification period, the burden of an interim report, the clarity of an expiration notice, the availability of an online application, the staffing of a call center, and the breadth of categorical eligibility each moved the realized caseload toward or away from the eligible population, without any change in the benefit formula or the income limits written in the Food and Nutrition Act. Error described whether a payment was correct. Delivery described whether a payment happened, for the full period of eligibility, to the people the law counted as eligible, and the administrative record through 2019 showed the second description carrying as much of the program in practice as the first.

The Assessment: Administration as the Operative Law

Administration, rather than the statutory text alone, determined what the Food and Nutrition Act delivered in the period ending December 1, 2019. The statute fixed the eligibility tests, the deduction sequence, and the benefit formula in national terms, and an average of 35.7 million people participated in each month of fiscal 2019 at a federal cost of 60.4 billion dollars. Between the text and those totals stood fifty three state agencies, county offices in ten county administered states, contracted issuance processors, and a quality control system that priced accuracy while assigning no comparable fiscal price to reach. The assessment that follows from the record is therefore institutional. Where delivery failed, the failure traced to funding ledgers, verification choices, certification calendars, and measurement incentives more often than to any sentence in the statute, and where delivery succeeded, the same machinery carried the entitlement across the distance between Washington and the supermarket.

What does the federal government pay for?

The federal government pays the full cost of every benefit issued to an eligible household and reimburses roughly half of allowable state administrative costs under section 16 of the Food and Nutrition Act. Benefits therefore draw on the federal Treasury alone, while the offices, workers, and systems that determine eligibility sit on a shared ledger.

What does a state gain by denying a benefit?

A state gains no state money. Benefits were federal dollars, so a denial or a deterred application saves the federal Treasury and removes food purchasing power from a household inside the state, while the state budget keeps nothing. State savings arise only on the administrative ledger, through fewer workers, fewer offices, or deferred computer systems.

That division was a product of legislative history rather than administrative habit. A reader tracing the entitlement to its statutory origin can follow the Food Stamp Act history from the pilot era into the national program, while the time limit for able bodied adults and the mandate for electronic issuance entered the program through the 1996 welfare reform statute. The authorizing statute itself lived inside the farm bill cycle, and the farm bill history and structure explains why a nutrition title moved through Congress alongside commodity programs. Each layer of that history deposited a rule that state administration then had to operate, and the operative law a household encountered was the accumulated product of those deposits, filtered through a state plan, a state computer system, and a local office.

What does the error rate count?

The error rate counts benefit dollars issued in the wrong amount in sampled cases, in both directions. Overpayments and underpayments are added together rather than netted, errors below a tolerance threshold are excluded, and no finding about intent enters the figure. The combined national rate stood at 6.80 percent for fiscal 2018.

What does trafficking measure?

Trafficking measures diversion after issuance, the exchange of benefits for cash or ineligible goods between a recipient and a retailer. Department studies estimated trafficking at about 1.0 percent of benefits for 2006 to 2008 and at 1.3 percent for 2009 to 2011. The estimate describes conversion of issued benefits, not accuracy in determining eligibility.

The accuracy regime supplied the strongest incentive in the entire administrative structure, and its history showed why administrators organized around the published rate. The combined national rate reached a record low of 3.20 percent in fiscal 2013 and 3.66 percent in fiscal 2014, national publication was suspended for fiscal 2015 and fiscal 2016 while federal teams reviewed state quality control practices, and publication resumed at 6.30 percent for fiscal 2017 and 6.80 percent for fiscal 2018. Public argument repeatedly collapsed the two integrity measures into a single accusation of theft, a collapse examined directly in the welfare reform myths examination. The program also operated beside other federal nutrition programs governed by different statutes and different delivery systems, set out in the Child Nutrition Act guide. The benefit calculation and the quality control system formed standard examinable material in social policy and public administration coursework, and students preparing on that material could practice and revise US government and civics material on ReportMedic. Researchers assembling the longer statutory record could keep your statute notes, citations, and case chronologies together free on VaultBook.

Who counts as an ABAWD?

An ABAWD is an able bodied adult without dependents, aged 18 to 49 on December 1, 2019, who lives in a household with no member under 18, holds no responsibility for a dependent child, is not pregnant, and holds no medical certification of unfitness and no other exemption from the general work requirements.

Why do eligible households stay out?

Participation research grouped the reasons as lack of awareness of eligibility, stigma attached to means tested assistance, and the burden of application, interview, documentation, and recertification. Eligible older adults participated at 45 percent in fiscal 2016, and eligible individuals overall at 85 percent, a pattern that administration, outreach, and office design each influenced.

The series thesis holds across the whole record. Congress authorized a uniform entitlement, defined by national tests and a national formula, and administration delivered a program whose realized caseload was smaller and differently composed than the eligible population, because fifty three state systems translated the entitlement under a federal accuracy regime that measured and priced payment correctness while reach carried no equivalent price. Error is not fraud. The payment error rate measured how accurately the machinery paid the households inside the caseload, in both directions, and the trafficking rate measured how much of the issued benefit stream left the food economy, and neither measurement described the households the machinery never reached. The gap between statutory eligibility and delivered benefits was therefore not a drafting failure or a single agency failure. The gap was the predictable output of a funding split that made delivery capacity a state expense, a measurement system that rewarded demonstrable correctness in sampled cases, and procedural calendars that placed the burden of continuity on households. Administration was the operative law, and the distance between the statute and the supermarket was where that law was written in practice.

Frequently Asked Questions

Q: Who pays for SNAP benefits and who runs the program?

The federal government pays the full cost of benefits, and the federal government and the states share administrative costs. Section 16(a) of the Food and Nutrition Act authorized federal payment of fifty percent of state administrative costs, the share covering application taking, eligibility determination, and issuance, as Associate Administrator Jessica Shahin summarized in testimony before the House Committee on Agriculture on July 6, 2016. The Food and Nutrition Service writes the regulations, approves state plans of operation, authorizes retailers, and runs federal oversight. The Congressional Research Service counted fifty three state agencies, comprising the fifty states, the District of Columbia, Guam, and the United States Virgin Islands, and state or county staff conduct interviews, verify cases, compute allotments, and issue benefits.

Q: What income limits apply to SNAP?

Under the regular federal rules, a household without an elderly or disabled member faced a gross monthly income test at 130 percent of the federal poverty guidelines and every household faced a net monthly income test at 100 percent. The Food and Nutrition Service fiscal 2020 standards, dated October 1, 2019, set the gross limit for a household of four in the 48 contiguous states and the District of Columbia at 2,790 dollars per month and the net limit at 2,146 dollars per month. Countable resources were limited to 2,250 dollars, or 3,500 dollars for a household with an elderly or disabled member. Broad based categorical eligibility replaced those figures in most states, with state gross limits reaching 200 percent of poverty.

Q: How is a SNAP benefit amount calculated?

The allotment equals the maximum allotment for the household size minus 30 percent of net monthly income. For fiscal 2020, effective October 1, 2019, the maximum for a household of four in the 48 contiguous states and the District of Columbia was 646 dollars. Net income is built through deductions taken in a prescribed order: 20 percent of gross earnings, the standard deduction of 178 dollars for a household of four in fiscal 2020, dependent care costs, qualifying medical expenses, legally obligated child support where a state elected that treatment, and excess shelter costs above one half of adjusted income, capped at 569 dollars for most households in fiscal 2020. In the worked illustration in the benefit calculation section, a household of four with 2,000 dollars in earnings reached net income of 1,233 dollars and an allotment of 276 dollars.

Q: What is the Thrifty Food Plan in SNAP?

The Thrifty Food Plan is a market basket of foods priced by the Department of Agriculture as the lowest cost of the four food plans published by the Department of Agriculture, constructed to provide a nutritionally adequate diet prepared at home for a reference household of four at minimal cost. Maximum allotments for other household sizes derive from that anchor through adjustments reflecting economies of scale in purchasing and preparation. The plan replaced the Economy Food Plan in 1975 on a cost neutral basis, and the Food Stamp Act of 1977 codified the thrifty plan as the basis of the maximum allotment. The Food and Nutrition Service resets allotments each October 1 through a cost of living adjustment tied to the cost of the plan as priced for the preceding June.

Q: What is the SNAP quality control error rate system?

Under Part 275 of Title 7, each state agency draws a random monthly sample of its caseload, and state reviewers separate from the eligibility workers reopen each sampled case, interview the household, and verify income, composition, and deductible expenses. The Food and Nutrition Service fact sheet of June 2018 put the state samples at about 50,000 cases a year nationally and the federal re review subsample at about 25,000 cases. A variance above the tolerance threshold becomes an error, and overpayment and underpayment dollars are added together into the combined payment error rate. That rate reached 3.20 percent in fiscal 2013 and 3.66 percent in fiscal 2014, national publication was suspended for fiscal 2015 and fiscal 2016, and publication resumed at 6.30 percent for fiscal 2017 and 6.80 percent for fiscal 2018.

Q: How common is SNAP trafficking?

Trafficking is the exchange of benefits for cash or for consideration other than eligible food. Mantovani and Wilson, in a Food and Nutrition Service study published in 2011, estimated trafficking at about 1.0 percent of benefits for 2006 to 2008, diverting an estimated 330 million dollars annually. Mantovani, Williams, and Pflieger, in a Food and Nutrition Service study published in August 2013, estimated trafficking at 1.3 percent of benefits for 2009 to 2011, diverting an estimated 858 million dollars annually against average redemptions of 63.7 billion dollars a year. The same studies estimated that 8.2 percent of authorized stores engaged in trafficking in the earlier period and 10.5 percent in the later period, because trafficking stores tended to redeem small volumes.

Q: What is an ABAWD time limit in SNAP?

Section 6(o) of the Food and Nutrition Act limits an able bodied adult without dependents, aged 18 to 49 as the rule stood on December 1, 2019, to three full months of benefits in a 36 month period unless the individual works or participates in a qualifying program for 80 hours per month, meets an exception, lives in a waived area, or receives a discretionary exemption. A prorated month does not count. An individual who lost eligibility could regain the benefit by meeting the work requirement during a 30 day period, and one additional period of up to three consecutive months was available once in a 36 month period. Discretionary exemptions stood at 15 percent of covered individuals through fiscal 2019 and fell to 12 percent for fiscal 2020 under the Agriculture Improvement Act of 2018.

Q: How many eligible people actually get SNAP?

The Food and Nutrition Service participation rate research, published in the Trends in Supplemental Nutrition Assistance Program Participation Rates series in a report dated July 2018, estimated that the program served 85 percent of all eligible individuals in fiscal 2016. Participation stood at 45 percent among eligible older adults, defined in that series as individuals age 60 or older, and at 75 percent among eligible individuals in households with earnings in fiscal 2016. State rates for all eligible persons in fiscal 2016 ranged from 56 percent to 100 percent, so the national figure concealed wide differences in how completely state administration reached the eligible population. The research associated nonparticipation with lack of awareness of eligibility, the stigma attached to means tested assistance, and the burden of the application process itself.

Q: How does the SNAP application and interview process work?

A household applies in person, by mail, by telephone, or through an online portal, and the filing date starts the determination period and governs proration of the first month allotment. State agency staff then conduct a certification interview, ordinarily before eligibility is decided, covering income, household composition, residence, and deductible expenses. The federal timeliness standards applied in the performance bonus structure described in the quality control record allowed thirty days for a regular determination and seven days for expedited issuance to qualifying households. A household that disputes a decision may request a fair hearing, and the state agency that made the decision conducts that hearing. An approval assigns a certification period and a monthly issuance date, and a denial must state the reason in a notice to the household.

Q: What verification and documentation does a SNAP household have to provide?

A household must document the facts on which eligibility and the benefit amount rest, principally income, household composition, identity, residence, and deductible expenses such as dependent care costs, qualifying medical expenses, and shelter costs. State agency staff verify those facts against program standards, and quality control reviewers later reconstruct sampled cases against the same record. The verification burden carried a documented tradeoff in the administrative record. Insufficient verification exposed a state agency to measured payment error, while heavy documentation demands shifted the cost of proof onto applicants in time, travel, and repeated submissions, and pushed eligible households out of the program for procedural reasons. Failure to provide verification stood among the prominent procedural reasons for application denials and case closures in the administrative record through December 1, 2019.

Q: How do deductions work in practice for a SNAP household?

In practice, a deduction has value only when a household reports the expense and the agency verifies the expense. The standard deduction requires no receipts, so every household receives that step. Dependent care costs, medical expenses of elderly or disabled members above 35 dollars per month, legally obligated child support payments, and shelter costs above one half of adjusted income each depend on documentation gathered by the household. The Food and Nutrition Service fiscal 2020 adjustments, effective October 1, 2019, capped the excess shelter deduction at 569 dollars for households without an elderly or disabled member and left the deduction uncapped for households with such a member. Each verified dollar of deduction reduced net income by a dollar and raised the benefit by thirty cents at the margin, until the maximum allotment was reached.

Q: How does categorical eligibility work for SNAP?

Traditional categorical eligibility makes a household eligible without a separate application of the income and resource tests when every member receives Supplemental Security Income, Temporary Assistance for Needy Families cash assistance, or state general assistance. Broad based categorical eligibility extends the door through a noncash benefit or service funded by Temporary Assistance for Needy Families dollars or by state funds counted toward the federal maintenance of effort requirement. The Congressional Research Service described households made eligible through receipt of a brochure or a referral to a social services telephone number. For a household made eligible through that door, the state deems the gross income and resource tests satisfied under the service limits, while net income governs the benefit calculation in the ordinary way. The Congressional Research Service recorded 42 jurisdictions operating the policy in July 2019.

Q: What happens to a state agency with a high SNAP error rate?

Federal regulations at Section 275.16 of Title 7 require a corrective action plan when a state payment error rate reaches 6 percent or more, when the case and procedural error rate stands above the national average, or when a significant share of sampled cases cannot be completed. Under the liability system built by the Farm Security and Rural Investment Act of 2002, a state faced a monetary liability only when a 95 percent statistical probability existed that its combined rate exceeded 105 percent of the national performance measure for a second consecutive year, with the penalty tied to the portion of the rate above 6 percent. A fixed bonus pool of 48 million dollars rewarded top and most improved performers until the Agriculture Improvement Act of 2018 eliminated performance bonus payments, effective for bonuses that would have been paid in fiscal 2019.

Q: How does a store become authorized to accept SNAP benefits?

A retail food store applies to the Food and Nutrition Service rather than to a state agency, and the service authorizes the store if the store meets the federal criteria, which turn on the variety of staple foods stocked or on the share of total sales derived from staple foods. The Government Accountability Office reported in May 2018 that authorizations generally ran for five years before reauthorization, across a population of more than 263,000 authorized retailers in fiscal 2017. Federal analysts screened redemption data through the Anti-Fraud Locator Using Electronic Benefits Transfer Retailer Transactions system, known as ALERT, and investigators examined flagged stores. A retailer found to have trafficked faced permanent disqualification, with a civil money penalty available in lieu of permanent disqualification only in defined circumstances, and administrative and judicial review operated under Part 279 of Title 7.

Q: How are SNAP benefits issued on an EBT card and replaced when a card is lost?

Nationwide electronic issuance was completed in June 2004 under the timetable set by the welfare statute of 1996. Each state agency contracts with a processor that operates the issuance system, credits the household account on a scheduled date each month under a staggered state calendar, and reconciles issuance records against federal accounts. Federal regulation at Section 274.2 of Title 7 barred more than 40 days between allotments for a household participating longer than two consecutive months. The household draws on the account with a card at authorized retailers. When a card is reported lost or stolen, the state system deactivates the reported card and issues a replacement through the state agency or its processor, so access attaches to the household account rather than to the physical card alone.

Q: What happens at SNAP recertification and why do cases churn?

As a certification period expires, the state agency sends a notice of expiration, and the household files an application for recertification, completes an interview, and supplies verification sufficient for a fresh determination. A household that files and cooperates on schedule receives the first allotment of the new period without a gap, while a late filing, a missed interview, or incomplete verification produces closure at the end of the period even where the facts would have supported continued eligibility. Gregory Mills and colleagues at the Urban Institute defined churning, in a 2014 study prepared for the Food and Nutrition Service, as exit and reentry within four months or less, and estimated an annual churn rate for fiscal 2011 from 17 percent to 28 percent across the six states studied, with most churn exits occurring at recertifications or required interim reports.

Q: How does Disaster SNAP work after a disaster?

Disaster SNAP rests on the Food and Nutrition Act and the Robert T. Stafford Disaster Relief and Emergency Assistance Act, and operation follows a presidential disaster declaration for individual assistance. A state agency requests Food and Nutrition Service approval, demonstrates that commercial food distribution channels in the affected area were restored, and opens a short application period for households that lived or worked in the disaster area and suffered disaster related expenses or income losses. Disaster procedures apply disaster specific income limits and deductions and ordinarily issue a benefit set at the maximum allotment for the household size. The Department of Agriculture announcement of September 22, 2017, approving Disaster SNAP for Florida counties after Hurricane Irma, stated that eligible households would receive two months of benefits equivalent to the maximum amount for a household of that size.

Q: How does SNAP work registration differ from the ABAWD time limit?

The general work requirements in section 6(d) apply to most participants from 16 years of age through 59 years of age who are not exempt. A work registrant must register for work, accept suitable employment if offered, participate in an assigned employment and training program or workfare assignment, furnish information on employment status, and refrain from voluntarily quitting a job or reducing work effort below 30 hours per week without good cause. Noncompliance produces a graduated disqualification period, and the general regime imposes no limit on total months of receipt. The ABAWD rule in section 6(o) layers a calendar on the narrower class aged 18 to 49: three full countable months in a 36 month period unless the 80 hour monthly test, an exception, a waiver, or an exemption applies. A waiver of the time limit leaves the general requirements in force.

Q: Why do so few eligible older adults participate in SNAP?

The Food and Nutrition Service participation series, in a report dated July 2018, recorded eligible older adult participation at 45 percent in fiscal 2016, against an overall rate of 85 percent in that fiscal year. Survey research conducted for the Food and Nutrition Service by Susan Bartlett and colleagues in 2004 organized the reasons as stigma, transaction costs, and lack of awareness, and older adult surveys recorded reliance on benefits as a loss of independence with particular frequency. Gabor and colleagues, in a 2002 study of older adult participation, identified misinformation about eligibility, transportation limits, and communication barriers for people with limited English proficiency. A small calculated benefit weakened the arithmetic case for applying, and the minimum benefit for fiscal 2020, effective October 1, 2019, stood at 16 dollars for eligible households of one or two persons.

Q: How are SNAP administrative costs shared between the federal government and the states?

The general rule under section 16(a) of the Food and Nutrition Act pays fifty percent of allowable state administrative costs from federal funds, with the state bearing the remaining share. Two exceptions raise the federal share. The Food and Nutrition Service could pay seventy five percent of approved administrative costs for administration on an Indian reservation under section 16(a) and Section 281.9 of Title 7. Employment and training services operated under section 16(h), with federal grants paying one hundred percent of approved costs up to a formula allocated amount and fifty percent reimbursement beyond that allocation. The state share competes inside a state legislature against education, transportation, and public health, and that competition sets the outer boundary on caseworker staffing, office networks, call center capacity, and eligibility computer investment in each state.