Food Assistance: Reducing Food Stamp Benefit Overpayments and Trafficking
(Chapter Report, 06/23/95, GAO/RCED-95-198).

Pursuant to a congressional request, GAO reviewed fraud, waste, and
abuse in the Department of Agriculture's (USDA) food stamp program,
focusing on: (1) the causes of benefit overpayments; (2) whether USDA
controls and procedures are adequate to prevent retailer involvement in
the trafficking of food stamp coupons; and (3) what can be done to
reduce trafficking by food stamp recipients.

GAO found that: (1) state caseworkers' and recipients' errors cause food
stamp overpayments; (2) state officials believe that caseworkers' errors
stem from the complexity of the program's regulations, the different
eligibility criteria for food stamp and Aid to Families with Dependent
Children programs, and recipients providing inaccurate income and other
information; (3) the states that have reduced their error rates have
made a commitment to do so in response to fiscal sanctions and
incentives; (4) these states have improved their program administration,
caseworker training, accountability, and information analyses and
verification; (5) states use group recertifications, recipient contacts
to detect household changes, and shorter certification periods for
recipients with fluctuating incomes to prevent recipient errors and
reduce overpayments; (6) USDA and the states are simplifying program
regulations, reducing eligibility criteria differences, and using
waivers from certain program requirements to improve payment accuracy;
(7) Congress and USDA are considering changing the current system of
sanctions and incentives for rewarding and penalizing states' error rate
performance; (8) insufficient resources for site visits and other
monitoring activities hamper USDA controls and procedures for monitoring
food stamp trafficking; (9) although USDA has implemented a number of
initiatives to improve its retailer monitoring processes, the
initiatives do not include additional resources for retailer monitoring;
and (10) aggressively reducing the number of retailers who are
trafficking in food stamps would make it more difficult for recipients
to sell their benefits or use them for nonfood purchases.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  RCED-95-198
     TITLE:  Food Assistance: Reducing Food Stamp Benefit Overpayments 
             and Trafficking
      DATE:  06/23/95
   SUBJECT:  Food stamp programs
             Aid to families with dependent children
             Administrative errors
             Eligibility criteria
             Overpayments
             Program abuses
             Welfare recipients
             Retail facilities
             Federal/state relations
             Welfare benefits
IDENTIFIER:  AFDC
             USDA Food Stamp Quality Control System
             Louisiana
             Massachusetts
             New Mexico
             South Carolina
             Florida
             New York
             Texas
             Food Stamp Program Project Recall
             FNS Income and Eligibility Verification System
             FCS Store Tracking and Redemption System
             USDA Electronic Benefit Transfer System
             Personal Responsibility Act of 1995
             FCS State Exchange Program
             Special Supplemental Food Program for Women, Infants, and 
             Children
             WIC
             
**************************************************************************
* This file contains an ASCII representation of the text of a GAO        *
* report.  Delineations within the text indicating chapter titles,       *
* headings, and bullets are preserved.  Major divisions and subdivisions *
* of the text, such as Chapters, Sections, and Appendixes, are           *
* identified by double and single lines.  The numbers on the right end   *
* of these lines indicate the position of each of the subsections in the *
* document outline.  These numbers do NOT correspond with the page       *
* numbers of the printed product.                                        *
*                                                                        *
* No attempt has been made to display graphic images, although figure    *
* captions are reproduced. Tables are included, but may not resemble     *
* those in the printed version.                                          *
*                                                                        *
* A printed copy of this report may be obtained from the GAO Document    *
* Distribution Facility by calling (202) 512-6000, by faxing your        *
* request to (301) 258-4066, or by writing to P.O. Box 6015,             *
* Gaithersburg, MD 20884-6015. We are unable to accept electronic orders *
* for printed documents at this time.                                    *
**************************************************************************


Cover
================================================================ COVER


Report to the Committee on Agriculture, House of Representatives

June 1995

FOOD ASSISTANCE - REDUCING FOOD
STAMP BENEFIT OVERPAYMENTS AND
TRAFFICKING

GAO/RCED-95-198

Food Stamp Overpayments and Trafficking


Abbreviations
=============================================================== ABBREV

  AFDC - Aid to Families with Dependent Children
  EBT - electronic benefit transfer
  FCS - Food and Consumer Service
  GAO - General Accounting Office
  IEVS - Income and Eligibility Verification System
  OIG - Office of Inspector General
  QC - quality control
  STARS - Store Tracking and Redemption System
  USDA - Department of Agriculture

Letter
=============================================================== LETTER


B-261444

June 23, 1995

The Honorable Pat Roberts
Chairman
The Honorable E (Kika) de la Garza
Ranking Minority Member
Committee on Agriculture
House of Representatives

This is the second of two reports that you requested on fraud, waste,
and abuse in the U.S.  Department of Agriculture's (USDA) Food Stamp
Program.  The report focuses on the current coupon-based system of
delivering benefits, with specific emphasis on three issues:  (1) the
causes of benefit overpayments, (2) the adequacy of USDA's controls
and procedures to prevent retailers' involvement in trafficking, and
(3) the trafficking of food stamp coupons by program recipients. 

Copies of this report are being sent to the Secretary of Agriculture,
the Administrator of the Food and Consumer Service, and to other
interested parties.  Copies are available to others on request. 

This work was performed under the direction of John Harman, Director,
Food and Agriculture Issues, who can be reached at (202) 512-5138 if
you or your staff have any questions.  Major contributors to this
report are listed in appendix I. 

Keith O.  Fultz
Assistant Comptroller General


EXECUTIVE SUMMARY
============================================================ Chapter 0


   PURPOSE
---------------------------------------------------------- Chapter 0:1

The U.S.  Department of Agriculture's (USDA) Food Stamp Program is
one of the nation's largest welfare programs.  In fiscal year 1994,
about $24 billion in food stamps was provided to over 27 million
recipients, with the vast majority of the benefits provided in the
form of food stamp coupons.  There is ample evidence that the
coupon-based system for delivering benefits is vulnerable to fraud,
waste, and abuse, but there are no reliable data available to
precisely determine the full extent of the problem.  Available data
from USDA show that errors in determining recipients' eligibility and
benefit levels are resulting in nearly $2 billion in benefit
overpayments each year.  In addition, a large, but unquantifiable,
amount of food stamps is sold or used for nonfood purchases by
recipients--generally referred to as trafficking.  Because of these
problems, the House Committee on Agriculture asked GAO to determine
(1) why overpayment errors occur, (2) whether USDA's controls and
procedures are adequate to prevent retailer involvement in
trafficking, and (3) what can be done to reduce trafficking by food
stamp recipients. 


   BACKGROUND
---------------------------------------------------------- Chapter 0:2

The Food Stamp Program is administered by USDA's Food and Consumer
Service in partnership with the states.  The Food and Consumer
Service provides nationwide criteria for determining who is eligible
for assistance and the amount of benefits recipients are entitled to
receive.  The states are responsible for the day-to-day operation of
the program, including meeting with applicants and determining their
eligibility and benefit levels.  State caseworkers rely on
documentation provided by households and information obtained in
interviews with the applicants in making these decisions.  Recipients
are normally certified to receive benefits for several months but are
required to report changes in their income or household information
that occur during the certification period that could change their
benefit levels.  The states' accuracy in determining applicants'
eligibility and benefit levels is determined annually through the
Food and Consumer Service's Quality Control System.  States with
error rates above certain levels are subject to federal fiscal
sanctions, while states with error rates below certain levels are
eligible for enhanced administrative funding. 

The Food and Consumer Service, through its field offices, is
responsible for authorizing retailers to redeem food stamp coupons as
well as for monitoring program compliance by the 209,255 stores
authorized to redeem coupons.  The Food and Consumer Service and
USDA's Office of Inspector General are responsible for investigating
retailers suspected of violating program regulations--such as food
stamp trafficking (termed "retailer trafficking").  States have
primary responsibility for investigating "recipient trafficking."

GAO conducted its review in seven states that either have been
successful in reducing benefit overpayments or are in the process of
implementing programs to reduce overpayments.  GAO reviewed efforts
to reduce overpayments and combat trafficking in the three Food and
Consumer Service regions in which the seven states were located. 


   RESULTS IN BRIEF
---------------------------------------------------------- Chapter 0:3

Overpayments result from state caseworkers' and recipients' errors
that occur for a multitude of reasons.  State officials told GAO that
the complex Food Stamp Program regulations and the differences in
eligibility and benefit determination criteria between regulations
governing the Food Stamp and the Aid to Families with Dependent
Children Programs are the primary reasons for caseworkers' errors. 
State officials also said that recipient-caused overpayments are
mostly the result of recipients either inadvertently or fraudulently
providing inaccurate income and other information that affects their
benefit levels and that these overpayments are difficult to prevent. 
Despite these obstacles, states GAO visited that have successfully
reduced their error rates attribute their successes to the fact that
they made a commitment to reduce errors.  The desire to avoid fiscal
sanctions or to receive enhanced funding provided much of the
incentive for making this commitment.  The Food and Consumer Service
is working with states to help reduce errors and has taken a number
of steps to simplify program regulations and reduce the differences
between Food Stamp and Aid to Families with Dependent Children
regulations.  The Congress and the Food and Consumer Service are also
considering changing the current system of incentives and sanctions
for rewarding and penalizing states' error rate performance.  GAO
agrees that simplifying eligibility determination requirements and
providing greater incentives to states to reduce errors could improve
food stamp payment accuracy. 

Controls and procedures for authorizing and monitoring retailers that
participate in the Food Stamp Program have not deterred nor prevented
retailers from trafficking in food stamps.  Primarily because of
insufficient resources to make site visits to retailers, the Food and
Consumer Service has not been able to prevent ineligible retailers
from being approved to participate in the program.  Furthermore, the
Food and Consumer Service has not adequately monitored stores
authorized to redeem food stamps to detect those that may be
trafficking.  The Food and Consumer Service has various initiatives
under way to improve its authorization and retailer-monitoring
processes, many of which have been included in the House-passed
Personal Responsibility Act of 1995; however, these initiatives do
not include making additional resources available to monitor
retailers. 

Although data are not readily available to accurately estimate the
amount of recipient trafficking, federal and state officials involved
in policing the Food Stamp Program believe that trafficking is
pervasive and that food stamps are often traded on the streets as a
second currency.  If the Food and Consumer Service can reduce the
number of retailers that are trafficking food stamps through more
aggressive on-site monitoring, greater enforcement actions, and
stiffer penalties, it will be much more difficult for recipients to
sell their benefits or use them for nonfood purchases. 


   PRINCIPAL FINDINGS
---------------------------------------------------------- Chapter 0:4


      CAUSES OF OVERPAYMENTS AND
      EFFORTS TO REDUCE THEM
-------------------------------------------------------- Chapter 0:4.1

In fiscal year 1993, food stamp overpayments totaled over $1.8
billion.  Of these overpayments, 42 percent occurred because
caseworkers did not accurately determine applicants' eligibility or
the amount of food stamps that the individuals should receive. 
Federal and state program officials told GAO that caseworkers' errors
occur for a multitude of reasons, including large caseloads, high
turnover, inadequate training, and poor supervision.  While sharing
the view that a multitude of factors are at work, state program
officials believe that the primary causes of caseworkers' errors are
the complexity of the Food Stamp Program regulations and the
differences in eligibility and benefit determination criteria between
the Food Stamp Program and the Aid to Families with Dependent
Children Program--which caseworkers are also tasked with
administering.  State officials told GAO that caseworker-caused
overpayments could be significantly reduced if the Food Stamp Program
eligibility and benefit determination regulations were simplified and
made more congruous with the Aid to Families with Dependent Children
Program regulations. 

Recipient-caused overpayments, which accounted for 58 percent of all
overpayments made in fiscal year 1993, are difficult to prevent. 
Caseworkers do not normally have the means necessary, nor the time,
to verify applicant-furnished information prior to determining the
eligibility to receive food stamps.  Furthermore, many inadvertent
recipient-caused overpayments occur after eligibility and benefit
levels have been determined when recipients fail to properly report
changes in income or household information that affect benefit
levels.  GAO's analysis of the Food and Consumer Service's data shows
that in fiscal year 1993, nearly two-thirds of all recipient-caused
payment errors were linked to recipient income, and, in all seven
states that GAO visited, recipients' failure to report increases in
income after certification was a primary cause of overpayments. 

Federal and state officials told GAO that state agencies' top
management's commitment to reducing errors is essential to reducing
overpayments.  State agencies are using a variety of actions to
reduce errors, such as improved caseworker training, improved
supervision, and caseworkers' increased accountability for errors. 
States are also requiring certain recipients to be recertified more
frequently, in an attempt to reduce recipient-caused overpayments. 
For example, the seven states that GAO visited were certifying
recipients with fluctuating income for no more than 3 months.  State
officials that GAO met with said that if the Food and Consumer
Service would eliminate or change its requirement that food stamp
recipients report increases in monthly income exceeding $25 to a
higher dollar amount, overpayment errors would decrease
significantly.  About 20 percent of food stamp recipients have some
earnings from work, and sometimes they earn additional income after
their benefit levels have been determined.  If the additional income
exceeds $25 in any month, the recipients' benefits must be adjusted
to avoid an overpayment error.  According to state officials, this
reporting requirement was established in 1974 and has never been
adjusted for inflation.  Two of the states included in GAO's
review--South Carolina and Texas--have received the Food and Consumer
Service's waivers from the requirement.  With the waiver, recipients
no longer have to report a fluctuation in income--all they are
required to report to caseworkers are changes in employer, wage rate,
or employment status (e.g., changing from part-time to full-time). 
South Carolina officials told GAO that this waiver alone reduced the
state's payment error rate by 2 percentage points, and Texas
officials are expecting a similar reduction.  GAO notes that while
this waiver does help reduce error rates, it does so by redefining
what is an overpayment and does not necessarily reduce program
benefit costs.  The Food and Consumer Service is reviewing the cost
benefits of this waiver in the nine states that have been given this
authority before proposing a change in the regulations or approving
the waiver for additional states. 

The Congress, in the Personal Responsibility Act of 1995, passed by
the House on March 24, 1995, is currently contemplating fundamental
changes to the federal food assistance structure, which includes a
change in the method of determining sanctions, or penalties, against
states having high error rates and increases the sanction amounts. 
Under present circumstances, the formula in this provision would
subject to sanction those states that have error rates that exceed
10.3 percent.  In view of the 1993 national error rate of 10.8
percent, more states could be sanctioned in the future if this
provision is enacted unless states substantially improve their
payment accuracy performance. 


      MORE MONITORING OF RETAILERS
      COULD DETER TRAFFICKING
-------------------------------------------------------- Chapter 0:4.2

The Food and Consumer Service's procedures and process for
authorizing retail stores to redeem food stamps are not adequate to
prevent unscrupulous retailers from being authorized to participate
in the Food Stamp Program.  Furthermore, once stores are authorized,
the Food and Consumer Service does not adequately monitor them to
ensure that their owners do not traffick food stamps or violate other
program regulations.  Officials responsible for reviewing stores'
applications and authorizing them to redeem food stamps told GAO that
the single most effective deterrent to preventing ineligible
retailers from being authorized is a preauthorization on-site visit. 
Such visits seldom occur, however, because the Food and Consumer
Service's current procedures state that visits to stores prior to
authorization should be the exception, rather than the rule. 
Managers from the eight Food and Consumer Service field offices that
GAO visited said that, because of insufficient resources, their
offices make few, if any, visits prior to authorizing a store to
participate in the program.  For example, field offices in the
Southwest Region reviewed authorization applications from 9,564
stores in 1994, but only 269 on-site reviews were completed. 

Monitoring visits to the more than 209,000 authorized stores
nationwide to ensure that they are complying with program rules are
also the exception rather than the rule.  In addition, reports on
retailers' activities, such as total food sales and food stamp
redemptions, are often untimely or inaccurate and of limited utility
in identifying retailer trafficking.  With regard to resources, Food
and Consumer Service officials told GAO that there are only 46
investigators nationwide to conduct investigations of retailers
suspected of trafficking or violating other program regulations.  In
fiscal year 1994, the Food and Consumer Service investigated 4,300
stores suspected of trafficking food stamps or violating other
program rules.  USDA's Inspector General also investigates stores
suspected of trafficking food stamps and devotes about 45 percent of
its investigative staff years to these investigations. 

The Food and Consumer Service is pursuing several initiatives to
improve its authorizing and monitoring of retailers and thereby
reduce food stamp trafficking.  On March 1, 1995, the Secretary of
Agriculture issued a statement calling for legislative changes to
help identify unscrupulous store owners and prevent them from
defrauding the program.  The Personal Responsibility Act includes
provisions giving the Food and Consumer Service a number of the
authorities requested.  The proposed legislation also directs that
on-site visits be made to stores prior to their being authorized to
accept food stamps by either Food and Consumer Service or state or
local officials.  While GAO agrees that the proposed changes would be
beneficial, they do not address the resources needed to perform
on-site visits at the time of retailer's authorization or monitoring
visits after authorization.  According to the Food and Consumer
Service an additional $19 million would be needed to make at least
one annual on-site visit to nonsupermarkets that, according to Food
and Consumer Service officials, are considered to be more prone to
traffick food stamps. 


      REDUCING RETAILER
      TRAFFICKING SHOULD DETER
      RECIPIENT TRAFFICKING
-------------------------------------------------------- Chapter 0:4.3

Available law enforcement data indicate that a considerable amount of
the food stamps issued each year is sold or used by recipients to
purchase nonfood items.  Because of the nature of this misuse and the
fact that 27 million people receive food stamps each year, GAO was
not able to precisely estimate the dollar value of trafficked
benefits.  The Food and Consumer Service, Inspector General, and
state investigators were also unable to accurately estimate the
amount of trafficking, but the investigators told GAO that recipient
trafficking is profuse and that food stamps are often used as a
second currency. 

Inspector General and state investigators told GAO that it is not
cost-effective to target for criminal investigation those recipients
who misuse their food stamps; therefore, they target the retailers
and individuals who are suspected of trafficking large quantities of
food stamps.  GAO agrees that if the number of retailers that
traffick food stamps can be reduced, it will be more difficult for
recipients to sell their benefits or use them for nonfood purchases. 


   RECOMMENDATION
---------------------------------------------------------- Chapter 0:5

In view of proposed legislation, GAO recommends that the Secretary of
Agriculture direct the Administrator of the Food and Consumer Service
to determine the resources needed to incorporate on-site visits into
its processes for authorizing and reauthorizing stores to accept food
stamp benefits.  This analysis should include a determination of the
resources needed to effectively monitor stores once they are
authorized to accept benefits and to investigate stores suspected of
food stamp trafficking.  The Secretary should consider the results of
this analysis in the setting of departmental priorities and the
allocation of resources, and provide this information to the Congress
for its use in considering program reforms. 


   AGENCY COMMENTS
---------------------------------------------------------- Chapter 0:6

In commenting on a draft of this report, the Administrator of the
Food and Consumer Service did not believe that the resource analysis
that GAO recommended to help reduce trafficking was necessary. 
Rather, the agency believes that the smarter use of new
technology--especially the future use of electronic benefit transfer
systems--along with the passage of its proposed retailer-related
antifraud legislation will enable it to detect and remove violating
stores more efficiently.  Moreover, the agency believes that its
antifraud proposals will help ensure that only legitimate stores are
authorized to redeem food stamps. 

GAO agrees that USDA's initiatives should help deter retailer
trafficking of food stamps.  However, GAO believes that it is
important to consider that the effectiveness of the agency's
legislative proposals in reducing trafficking is unproven at this
time and will not be known until some time in the future.  Also, GAO
does not believe that the use of electronic benefit transfer systems
will necessarily eliminate the need to conduct store visits. 
Instead, it will serve as a valuable supplementary tool in the effort
to reduce trafficking.  Furthermore, the implementation of electronic
benefit systems is in its infancy, and nationwide implementation will
not be completed for several years.  In this context, GAO continues
to believe that on-site visits to retailers should be included in any
overall strategy to increase retailer's integrity and that the Food
and Consumer Service should take the necessary steps to ensure that
such visits are made. 


INTRODUCTION
============================================================ Chapter 1

The Food Stamp Program provides monthly benefits to millions of needy
people who meet specific income, asset, and employment-related
eligibility requirements.  With federal benefit payments of about $24
billion and an average monthly caseload of over 27 million
individuals in fiscal year 1994, the program ranks as one of the
nation's largest welfare programs and is by far the largest food
assistance program. 

There is abundant evidence that the coupon-based system used in
delivering and redeeming food stamp benefits is vulnerable to fraud,
waste, and abuse; however, the precise amount of losses that occur
each year is unknown.  Program data show that caseworkers' and
recipients' errors occurring during the process of determining
eligibility and benefit levels resulted in over $1.8 billion in
overpayments in fiscal year 1993.  This amount represents a
121-percent increase in overpayments since 1988.  In addition,
although there are no data available to develop a precise estimate,
the number of cases being prosecuted each year suggest that a
significant number of the food stamps issued each year are sold or
traded for nonfood items--referred to as trafficking--thereby
diverting benefits from their intended use. 


   FOOD STAMP PROGRAM
   ADMINISTRATION
---------------------------------------------------------- Chapter 1:1

The Food Stamp Program is administered by the Department of
Agriculture's (USDA) Food and Consumer Service (FCS) in partnership
with state and local governments.  The Congress established the
current Food Stamp Program in 1964 to improve the nutrition of
low-income individuals by increasing their food purchasing power. 
The program is federally designed and generally requires applicants
to apply in person at a local food stamp or welfare office and meet
numerous program requirements pertaining to their household
composition, financial assets, employment requirements, and income to
be eligible for monthly benefits.  Benefits are generally provided in
the form of food stamp coupons, which are issued to recipients in
person or by mail.\1 Recipients then use their coupons to purchase
food items at stores that have been authorized by FCS to redeem food
stamps.  Retailers deposit redeemed food stamps at financial
institutions and are credited for their sales.  Financial
institutions then forward the redeemed stamps for credit through the
banking system to the Federal Reserve--which destroys the coupons. 

The Food Stamp Program is the cornerstone of USDA's food assistance
programs, providing more benefits than all other federal food
assistance programs combined.  Although the household is the base
unit to which food stamp benefits are issued, most of the people who
benefit from the program are children or elderly persons.  In
addition, more than 20 percent of recipient households have some
earnings from work, making the program a critical source of
nutritional support for the nation's working poor. 


--------------------
\1 In some states, food stamp benefits are provided electronically
through electronic benefit transfer technologies or provided in the
form of cash. 


      FEDERAL ROLE
-------------------------------------------------------- Chapter 1:1.1

The federal government funds all food stamp benefits and shares with
the states in the administrative costs of distributing benefits to
eligible recipients.  The federal government pays all costs for
printing the stamps, distributing the stamps to the states, and
destroying the stamps after they are used.  The government also pays
all costs associated with approving and monitoring food stores
authorized to redeem food stamps.  In addition, the government pays
roughly 50 percent of the costs incurred by the states in
administering the program.  These costs include certifying recipients
for eligibility, distributing benefits to recipients, and conducting
quality control activities. 

FCS prepares Food Stamp Program regulations and provides technical
assistance to the states.  In addition, as part of its oversight
responsibilities, FCS has established a Quality Control (QC) System
to monitor states' determinations of individuals' eligibility for
benefits and the level of those benefits.  Under the QC System,
states are required to review a sample of their food stamp cases to
test the accuracy of eligibility and benefit level determinations
made by their caseworkers.  The states' error rate findings, a
combination of overpayments and underpayments, are provided to FCS. 
FCS then reviews a subsample of each state's sample to ensure the
quality of their efforts.  After disagreements with the states'
reported error rates are resolved, FCS determines the official error
rate for each state.  If a state's combined error rate (both
overpayments and underpayments) exceeds a national tolerance level,
calculated by FCS, then the state can be required to reimburse the
federal government for a portion of the erroneous payments.  On the
other hand, if a state's combined error rate is less than 6 percent,
the state may be eligible for enhanced federal funding. 

FCS is also responsible for authorizing stores to redeem food stamps
and for monitoring stores to ensure that they comply with Food Stamp
Program regulations.  Store owners suspected of engaging in food
stamp trafficking or abusing other program regulations are referred
to FCS' Compliance Branch for investigation.  In addition to FCS'
oversight activities, USDA's Office of Inspector General (OIG)
investigates retailers involved in illegally using food stamp
coupons. 


      STATE ROLE
-------------------------------------------------------- Chapter 1:1.2

States have primary responsibility for the day-to-day administration
and operation of the Food Stamp Program.  Generally, a state welfare
agency operates the program either directly or by delegating these
responsibilities to a local government agency.  Using FCS-established
eligibility and benefit determination regulations, state and local
government caseworkers take applications from households seeking food
stamp benefits and, through face-to-face discussions and review of
information provided by the applicants, determine recipient
households' eligibility for food stamp benefits and the amount of
benefits to which they are entitled. 

Normally, a recipient is certified to receive food stamps for a
specified period of time--usually between 3 and 12 months--which
means that a recipient does not need to reapply for benefits each
month.  However, recipients must report changes in income, in the
composition of their household, or other pertinent information as
changes occur during the certification period.  On the basis of
reported changes, caseworkers make adjustments to the benefit amounts
provided to the recipients. 

The states also share in responsibilities regarding the integrity of
the Food Stamp Program.  Specifically, states are responsible for
investigating and prosecuting individuals suspected of falsifying
information in order to obtain food stamps and misusing their
benefits--such as selling their benefits for cash or trading them for
other nonfood items.  Also, states sometimes work with FCS or the OIG
in investigating retailer abuse in the program. 


   PRIOR GAO AND OFFICE OF
   INSPECTOR GENERAL REPORTS
---------------------------------------------------------- Chapter 1:2

On numerous occasions, we and the USDA OIG have reported on fraud,
waste, and abuse in the Food Stamp Program.  Since 1977, we have
completed over 50 reviews of the program and on several occasions
pointed out that waste and abuse were costing the federal government
billions of dollars in food stamp overpayments.  In 1990, we reported
that reducing certain major types of overpayment errors, such as
those associated with recipients' income, was very difficult for the
states because state caseworkers did not have access to the
information necessary to prevent these errors.\2

In our most recent report and testimony on food stamp fraud before
the House Agriculture Committee, we pointed out that the rate of
overpayment errors had increased significantly in recent years and
that by fiscal year 1993 overpayments exceeded $1.8 billion.\3 We
also noted that 42 percent of the overpayments resulted from errors
made by state and local caseworkers in determining recipients'
eligibility and benefit levels.  We also reported that 58 percent of
the overpayments were the result of recipients' errors in reporting
information to caseworkers at the time of application for benefits,
or in updating information on changes in a recipients' household
during the certification period.  Of the fiscal year 1993
recipient-caused overpayment errors, about 60 percent were judged by
state QC officials to have been inadvertent errors while about 40
percent were judged to have been deliberate, or suspected fraudulent,
misrepresentations by the recipient in reporting household
information. 

Furthermore, we reported that, according to FCS' QC System data,
about two-thirds of the overpayments occurring in 1993--regardless of
whether the overpayment was caseworker- or recipient-caused--were
linked to errors related to recipient income.  The next largest
source of overpayments, about 14 percent, involved errors associated
with the composition of recipients' households, such as number, ages,
and relationships of household members. 


--------------------
\2 Food Stamp Automation:  Some Benefits Achieved; Federal Incentive
Funding No Longer Needed (GAO/RCED-90-9, Jan.  24, 1990). 

\3 Food Assistance:  Potential Impacts of Alternative Systems for
Delivering Food Stamp Program Benefits (GAO/RCED-95-13, Dec.  16,
1994) and Food Assistance:  Reducing Benefit Overpayments in the Food
Stamp Program (GAO/T-RCED-95-94, Feb.  1, 1995). 


   OBJECTIVES, SCOPE, AND
   METHODOLOGY
---------------------------------------------------------- Chapter 1:3

The Chairman and Ranking Minority Member of the House Committee on
Agriculture asked us to determine (1) why benefit overpayments occur,
(2) whether USDA controls and procedures are adequate to prevent food
retailer involvement in trafficking, and (3) what can be done to
reduce trafficking by food stamp recipients. 

In addressing these objectives, our first step was to determine how
the Food Stamp Program currently works by reviewing applicable laws
and regulations that govern how recipients' eligibility and benefit
levels are determined, how food stores are approved to redeem food
stamps, and how food stamps are to be redeemed by recipients.  We
then interviewed FCS and state officials to determine how the program
was being implemented on a day-to-day basis.  In accomplishing this
step, we met with FCS officials at headquarters and at three of its
seven regional offices.  We also met with FCS field office officials
in each of the three regional offices. 

In addition to the FCS officials, we met with state and local agency
officials responsible for operating the Food Stamp Program in seven
states--all located within the three FCS regions we visited.  Four of
the states selected--Louisiana, Massachusetts, New Mexico, and South
Carolina--were selected for review because they have been extremely
successful in reducing error rates in recent years, and we wanted to
learn how these states had lowered their error rates.  South Carolina
reduced its overpayment error rate from 11.66 percent in fiscal year
1986 to 7.87 percent in fiscal year 1993.  Massachusetts reduced its
overpayment error rate from 10.88 percent in fiscal year 1990 to 4.45
percent in fiscal year 1993.  Louisiana's overpayment rate was 9.48
percent in fiscal year 1989, but by fiscal year 1993 the state had
lowered the rate to 6.88 percent.  New Mexico reduced its overpayment
rate from 10.28 percent in fiscal year 1987 to 7.61 percent in fiscal
year 1993. 

We also visited three states that have historically had high
overpayment error rates--Florida, New York, and Texas--to learn why
overpayment errors have remained high.  While all three states have
experienced high error rates, they have undertaken recent initiatives
to reduce errors. 

At all the locations we visited, we reviewed pertinent documents,
reports, and error-reduction plans.  We also reviewed USDA OIG
reports on why overpayment errors occur and what is being done to
reduce these errors.  We discussed overpayment issues with
appropriate officials and obtained their opinions on why food stamp
overpayments occur and what can be done to reduce overpayment errors. 
We also reviewed various state performance reporting data, including
food stamp quality control reviews, management evaluations, and
assessed how successful the states have been in reducing their
overpayments.  Additionally, we discussed food stamp overpayments and
error-reduction activities with representatives from the American
Public Welfare Association because of their interest in the Food
Stamp Program. 

In Florida, New York, and Texas, we visited two local offices
responsible for operating the program in each state.  In Florida, we
visited two state district offices--Miami and Orlando--having high
food stamp caseloads and overpayment error rates.  New York has two
separate error-reduction plans--one for upstate and one for New York
City--we visited a local food stamp office in Albany and New York
City.  In Texas, we visited the two regional offices that serve the
Dallas/Ft.  Worth and south Texas areas because these offices have
high food stamp caseloads and overpayments. 

With regard to our overpayment objective, we relied heavily on data
from FCS' QC System to determine the amounts of overpayments that
result from various types of errors.  While the QC System provides
extensive data on the types of overpayment errors that occur and who
makes overpayment errors, it does not specifically identify why these
errors occur. 

With regard to our second and third objectives, which focuses on
efforts to curtail retailer and recipient trafficking, we met with
FCS, OIG, and state/local program and law enforcement officials to
discuss the extent of food stamp trafficking and what is being done
to prevent this illegal activity.  We reviewed FCS' controls and
procedures for authorizing and monitoring retail stores that
participate in the Food Stamp Program.  This effort involved visits
to not only the three regional offices responsible for the states
included in our review but also eight FCS field offices within these
regions.  The field office is where applications from stores are
reviewed and approved and some retailer monitoring activities are
carried out.  We also discussed the extent of trafficking with
investigators from FCS' Compliance Branch and the OIG and reviewed
the results of their investigations. 

We conducted our review from November 1994 through May 1995 in
accordance with generally accepted government auditing standards. 
USDA's comments on a draft of this report appear in appendix I. 


STATE MANAGEMENT COMMITMENT IS KEY
TO REDUCING OVERPAYMENTS
============================================================ Chapter 2

Caseworkers' and recipients' errors cost the federal government over
$1.8 billion in overpaid food stamp benefits in fiscal year 1993, the
most current year for which data were available.  Caseworkers'
errors, which accounted for 42 percent of the overpayments, occur for
a multitude of reasons, but state officials said that the complex
Food Stamp Program regulations and the differences between the Food
Stamp Program and the Aid to Families with Dependent Children (AFDC)
regulations are the primary contributing factors.  States have taken
a number of actions to reduce caseworkers' errors.  Recipients'
failure to accurately report household information, particularly
income, cause 58 percent of the overpayment errors.  While some
states have reduced recipient-caused overpayments, it is more
difficult for the states to prevent these errors from occurring. 

States that have been successful in reducing overpayments attribute
their success to the fact that top management is committed to
overcoming the causes of the overpayments.  The desire to avoid
fiscal sanctions or to receive enhanced funding provided much of the
incentive for making this commitment.  FCS is also taking a number of
actions to improve payment accuracy.  The impact of FCS' actions on
reducing overpayments is uncertain, but the positive experiences of
the states that we visited in reducing error rates indicate that
considerable improvement in reducing overpayments could be achieved
nationwide. 


   MULTITUDE OF FACTORS CONTRIBUTE
   TO CASEWORKERS' ERRORS
---------------------------------------------------------- Chapter 2:1

While the Food Stamp Program QC System provides extensive data on the
types of caseworker errors that result in overpayments, it does not
identify "why" caseworkers make these errors.  Officials in the
states that we visited believe that the complexity of the Food Stamp
Program and differences with the AFDC regulations are the largest
hurdles to improving caseworkers' payment accuracy. 


      REASONS FOR CASEWORKERS'
      ERRORS
-------------------------------------------------------- Chapter 2:1.1

Caseworkers' errors accounted for $763 million in overpayments in
fiscal year 1993.  We spoke with numerous FCS and state officials
regarding the reasons caseworkers make errors when determining
applicant's eligibility for benefits and the amount of benefits
applicant's are entitled to receive. 

Both FCS and the states describe a multitude of factors that
contribute to the errors.  These reasons include the following: 

  There has been a substantial increase in program recipients in
     recent years.  Between July 1989 and March 1994, the number of
     individuals participating in the Food Stamp Program rose by 51
     percent.  The caseload growth has occurred at a time when states
     have been unable to hire additional staff to handle the
     increased workload.  Stretched resources have caused the
     workload of caseworkers to increase drastically. 

  A high turnover of caseworkers causes state and local offices to
     have less experienced staff.  High caseloads, stress, and low
     salaries have led to attrition. 

  Complex Food Stamp Program regulations contribute to confusion in
     determining eligibility, which leads to errors. 

  There are differing eligibility and benefit requirements for the
     various welfare programs caseworkers are tasked with
     administering.  For example, caseworkers often authorize
     recipients for multiple programs, including food stamps, AFDC,
     and Medicaid. 

  Inadequately trained caseworkers and poor supervision lead to
     errors. 

  FCS officials believe that some states place a lower priority on
     the Food Stamp Program than on other programs they administer
     and commit less effort to reducing food stamp overpayments. 
     This is because the federal government pays the entire amount of
     food stamp benefits, whereas in some other programs--such as
     AFDC--costs are shared.  FCS and state officials also told us
     that competing demands of other social programs often affect a
     state's ability to focus adequate people and resources on Food
     Stamp Program administration. 

Of all these factors, state officials that we visited cited (1) the
complexity of the Food Stamp Program regulations and (2) the
differences between the Food Stamp Program regulations and the AFDC
regulations as the primary causes of caseworkers' errors.  The
officials believe that the program in itself is one of the more
complex federal welfare programs and that caseworkers are prone to
make mistakes when determining applicants' eligibility and benefit
levels. 

The officials explained that in many states, the same caseworkers
handle both AFDC and food stamp cases.  They said that the
probability of errors increases when caseworkers are responsible for
processing applicants for both food stamp benefits and other benefit
programs--particularly AFDC because of the differences in the
eligibility rules and benefit levels between programs.  For example,
caseworkers must take into account liquid assets, household income,
and household size when determining eligibility and benefits for both
the Food Stamp Program and AFDC.  However, the treatment of this
basic information differs for both programs.  With regard to liquid
assets, AFDC allows a family to possess assets of $1,000, whereas the
Food Stamp Program sets the maximum asset limit at $2,000 per
household ($3,000 for households with an elderly member).  The
treatment of vehicles also is different.  AFDC calculates the equity
value of a vehicle and excludes the first $1,500 from assets;
however, the Food Stamp Program calculates the fair market value and
excludes the first $4,550.  While both programs allow vehicles to be
exempted from the eligibility and benefit determination process, the
criterion for exempting vehicles is not the same for both programs. 

Also, the definition of a food stamp beneficiary unit and an AFDC
beneficiary unit differ.  The basic food stamp beneficiary unit is
the household.  A food stamp household can either be an individual
living alone or several individuals living together who customarily
purchase food and prepare meals together.  Individuals who are
members of the same household must apply together, and the income,
expenses, and assets of all members are counted in determining the
household's eligibility and benefit allotment.  Conversely, the AFDC
beneficiary unit is defined as being made up of at least one
dependent child, the child's parents and siblings (by blood or
adoption), and possibly other caretaker relatives living with the
child and/or family.  For example, a mother, her two children, and a
nonfamily-related companion live and prepare meals together.  For the
Food Stamp Program, all four people are included in the household,
and the household's total expenses and income, including the
companion's, are counted.  The AFDC assistance unit would be the
mother and her two children.  The companion's income would not be
included in determining AFDC benefits. 

In the Food, Agriculture, Conservation and Trade Act of 1990 (P.L. 
101-624), the Congress authorized a Welfare Simplification and
Coordination Advisory Committee to review the federal government's
four major assistance programs--food stamps, AFDC, Medicaid, and
housing.  In its report, one of the committee's objectives was to
examine the difficulties experienced by program administrators in
providing timely benefits efficiently to all who are qualified to
receive them.  In its June 1993 report to the Congress,\4 the
committee recommended establishing uniform rules and definitions to
be used by all needs-based programs in making their eligibility
determinations, including:  common definitions of countable income,
allowable deductions, resources, and household composition.  The
report identifies 57 differences between the Food Stamp and AFDC
Programs that were compiled by program administrators with the help
of the American Public Welfare Association.  According to the
committee report, some of the differences are rooted in statute and
only congressional action can eliminate them.  Others are disparities
in regulations and could be addressed by the federal agencies. 


--------------------
\4 Time for A Change:  Remaking the Nation's Welfare System, (June
1993). 


      STATES' INITIATIVES TO
      REDUCE CASEWORKERS' ERRORS
-------------------------------------------------------- Chapter 2:1.2

The seven states we visited have used a variety of initiatives to
reduce error rates.  The most important factor in reducing errors
seems to be a commitment by state officials to aggressively pursue
reductions in the error rate.  State officials told us that their
motivation to reduce error rates stems from their desire to either
reduce the possibility of being sanctioned for having high error
rates or to receive enhanced funding for reducing their error rates
below 6 percent.  States have demonstrated their commitment to
reducing food stamp benefit errors in a variety of ways, including

  restructuring state administration of the Food Stamp Program,

  holding management and caseworkers more accountable for error
     reduction,

  conducting more detailed analyses of local error data,

  targeting supervisory reviews on error-prone cases, and

  improving caseworker training. 

FCS and state officials told us that because states use a variety of
initiatives to lower overpayments, and because so many factors can
effect a state's performance, there is no way to determine precisely
the impact any one of these initiatives has in reducing caseworkers'
overpayment errors. 


         RESTRUCTURING STATE
         PROGRAM ADMINISTRATION
------------------------------------------------------ Chapter 2:1.2.1

Louisiana, Massachusetts, and South Carolina recently restructured
their state Food Stamp Program in ways that give the state more
control and ability to address overpayment problems.  For example,
Louisiana reorganized the welfare office in New Orleans, which had
high error rates.  Louisiana also realigned its field and program
staff positions to more effectively cover the needs of the whole
state.  Massachusetts reduced the number of food stamp offices by
almost half--from 80 offices to 44.  South Carolina officials told us
that they restructured the administration of their program by
requiring local program offices to begin reporting to state-level
program offices rather than to local welfare boards. 


         MORE ACCOUNTABILITY FOR
         ERROR REDUCTION
------------------------------------------------------ Chapter 2:1.2.2

Some states that we visited are emphasizing accountability and
holding officials more responsible for overpayments as a means to
reduce caseworker overpayments.  For example, Texas, in September
1994, implemented a process that establishes accountability for food
stamp payment accuracy at the regional management level.  It is now
part of the regional managers' performance expectations to lower the
region's errors, and according to state program managers, their jobs
could be in jeopardy if the region's errors are not lowered. 
Louisiana, Massachusetts, and South Carolina are also emphasizing
payment accuracy by identifying staff who have made errors and
emphasizing to them during formal discussions the consequences of
their mistakes.  The management commitment in these three states has
been very effective and has contributed to these states reducing
their error rates in fiscal year 1994 to 6 percent or less--the level
at which they may be eligible for enhanced funding. 


         ADDITIONAL ANALYSIS OF
         LOCAL OFFICE ERROR RATES
------------------------------------------------------ Chapter 2:1.2.3

Four states that we visited, Florida, Massachusetts, New York, and
Texas, have devoted special staff to reviewing additional recipient
cases in local offices to gather more data on the cause of
caseworkers' errors at the local level.  The statewide QC process
does not provide local offices with sufficient data to determine the
error rate performance of their caseworkers.  The additional local
review process provides information on local office error rates and
data on the types of errors occurring at each location.  These
special staff also help the local offices develop corrective actions
for the errors that are occurring in their location. 

According to Massachusetts officials, the reviews (1) help
state-level management officials know which local offices are having
problems; (2) establish a healthy competition among offices to
achieve lower rates; and (3) encourage workers to be more careful,
knowing that their cases could be selected for review.  New York City
officials reviewed approximately 13,000 cases in fiscal year 1994,
which allowed them to provide error data to all 53 city offices.  In
addition to determining error rate data, New York City has a team of
organizational research analysts that helps offices operate more
efficiently.  The analysts assist the offices in developing
office-specific corrective action plans based on their findings.  New
York City officials said that these and other initiatives have been
quite successful in reducing and maintaining the city's error rate
during a time of staff shortages and restricted budgets.  The city's
combined error rate was reduced from a high of 15 percent in fiscal
year 1989 to 10 percent in fiscal year 1993. 


         TARGETING ERROR-PRONE
         CASE TYPES FOR
         SUPERVISORY REVIEW
------------------------------------------------------ Chapter 2:1.2.4

Both New Mexico and South Carolina have enhanced their case review
processes by targeting cases with characteristics that have the
highest probability of error for supervisory review.  In New Mexico,
over 50 percent of the dollar errors involve earned income.  For that
reason, all supervisory reviews in the state are targeted to
recipient households having earned income.  Supervisors in South
Carolina are examining certain error-prone aspects of the case file
such as earned income and household composition.  These reviews, in
both states, stress caseworker accountability, identify weaknesses,
and correct possible overpayments.  Supervisors can identify error
trends and problems in the application of program regulations, as
well as help determine the training needs of individual workers. 


         TRAINING CASEWORKERS TO
         AVOID ERRORS
------------------------------------------------------ Chapter 2:1.2.5

Several states that we visited have emphasized training as a method
of reducing caseworkers' errors.  Louisiana recently revised its
training courses for both new and experienced workers.  For example,
the state requires experienced workers to be retrained every 2 years
on program regulations that are difficult to apply or lead to
caseworker errors.  New Mexico similarly developed a refresher
training course that focuses on deficiencies identified through QC
reviews.  South Carolina is also providing training on error-prone
areas for its experienced workers. 


   RECIPIENTS' ERRORS ARE
   DIFFICULT TO PREVENT
---------------------------------------------------------- Chapter 2:2

Our review of FCS' QC System data for fiscal year 1993 shows that
over half of the overpayments, about $1 billion, were caused by
recipients either failing to report information to caseworkers or
reporting incorrect information to caseworkers.  Most
recipient-caused overpayments involve the misreporting of income.  In
the states that we visited, recipients not reporting changes in their
income or withholding sources of income from caseworkers was one of
the leading causes of all overpayments.  Recipients' errors are more
difficult for states to control than caseworkers' errors because
caseworkers are dependent upon the accuracy of information reported
by the recipient.  The information available to caseworkers does not
enable them to discover all types of unreported income or other
resources, such as motor vehicles, or to always accurately determine
household composition, which includes establishing the living and
eating arrangements of all household members.  State officials
conducting QC reviews, on the other hand, have the time necessary to
do more extensive reviews of the information provided by
recipients--including detailed income verifications and home visits
to verify household information. 


      RECIPIENT-CAUSED ERRORS
      MOSTLY INVOLVE ERRONEOUS
      INCOME REPORTING
-------------------------------------------------------- Chapter 2:2.1

FCS' QC System data classifies the type of errors made by recipients
into the following five categories: 

  income reporting, such as wages and salary information;

  nonfinancial reporting, such as household composition;

  resources reporting, such as bank accounts;

  deduction reporting, such as expenses for medical care and shelter
     costs that can be deducted from household income in determining
     benefits; and

  other errors. 

Our analysis of the QC System data for fiscal year 1993 shows that
nearly two-thirds of the erroneous payments caused by recipients were
related to an error in reporting household income.  According to the
QC System, sometimes these errors occurred at the time a recipient
applied for benefits, and sometimes they were the result of a
recipient not reporting household income changes after they were
certified to receive benefits.  For example, changes in income
exceeding $25 a month are generally required to be reported to
caseworkers, who then recalculate the recipients' benefit levels. 

Because of the high incidence of income-related errors, we focused
our review primarily on the reasons recipients made income-reporting
errors and what is being done to reduce these errors.  Using the QC
System, we divided our analysis into two parts:  (1) errors that were
judged to have been committed inadvertently by the recipient and (2)
errors that were judged to be deliberate income misrepresentations by
the recipient. 


      MOST INADVERTENT ERRORS DUE
      TO RECIPIENTS NOT REPORTING
      CHANGES AFTER CERTIFICATION
-------------------------------------------------------- Chapter 2:2.2

The greatest amount of inadvertent recipient overpayments occur as a
result of recipients not reporting changes in their households after
they are certified to receive benefits--especially changes in income. 
Fiscal year 1993 FCS QC System data show that 62 percent of all
overpayments caused by inadvertent recipient errors occurred after
certification.  According to Food Stamp Program regulations,
households are required to report the following changes in their
households within 10 days of the date the change becomes known to the
household: 

  changes in the sources of income, or in the amount of gross monthly
     income, that exceed $25;

  changes in household composition, such as the addition or loss of a
     household member;

  changes in the household's residence and in shelter costs;

  the acquisition of a motor vehicle; and

  increases in household assets that result in the total household
     assets reaching or exceeding $2,000. 

When a recipient reports any of the above changes, caseworkers must
determine if the change affects their eligibility and recalculate the
recipient's benefit levels.  The above events can result in a change
in the amount of benefits that recipients are eligible to receive. 
Thus, when recipients do not report this information, payment errors
may occur, and these errors are reflected in a state's error rate. 
According to program regulations, caseworkers are to explain these
reporting requirements to recipients at the time they are certified
to receive benefits.  Recipients are given a form for reporting
changes.  The form outlines the civil and criminal penalties for
hiding or providing false information. 

State and local government officials that we talked with said that
recipients do not report changes in their households for a variety of
reasons.  For example, some recipients have difficulty in
understanding the Food Stamp Program requirements.  Furthermore,
recipients receiving both food stamp and AFDC benefits often confuse
the different reporting requirements for each program.  Second, some
recipients wait until their recertification interviews to report
changes to caseworkers.  Third, the recipients do not have time to
report changes to the office or are unable to get through to a
caseworker when they try to report changes by telephone.  Finally,
some recipients intentionally do not report changes in their
households so that their benefits will not be decreased.  State
program officials told us that it is often difficult to determine the
recipients "intent" for not reporting post-certification changes, so
these cases are usually documented as inadvertent recipient error. 

According to state officials and our analysis of state QC data, the
failure to report changes in income of $25 or more is a major cause
of inadvertent recipient overpayments in all of the states that we
visited.  State officials said that the $25 reporting requirement is
not practical for recipients who generally have fluctuating income. 
A recipient need work only a few extra hours during a month to earn
an additional $25 or more.  The officials said that the reporting
requirement should be eliminated or at least raised to a more
realistic amount.  Furthermore, state officials pointed out that the
$25 amount was established in 1974 and has never been adjusted for
inflation. 


      STATES' EFFORTS TO REDUCE
      INADVERTENT RECIPIENT ERRORS
-------------------------------------------------------- Chapter 2:2.3

The states that we visited are attempting a variety of efforts to
control inadvertent recipient errors.  These efforts include: 

  conducting group recertifications to ensure that all recipients are
     informed of the requirement to report changes,

  contacting recipients during their certification period to
     determine if there have been any changes in their households,
     and

  certifying recipients with fluctuating earnings for shorter periods
     to shorten the amount of time that overpayments can occur. 

In addition to states' efforts, FCS can and has waived the
requirement that recipients report fluctuations in their income of
$25 or more in nine states.  FCS is considering a change in the
regulations to eliminate the $25 reporting requirement and is
evaluating information received from some of the states with the
waiver to determine the extent to which program costs might be
affected. 


         GROUP RECERTIFICATIONS
------------------------------------------------------ Chapter 2:2.3.1

Local offices in some states are using group recertifications. 
Generally, caseworkers provide recipients information on the
reporting requirements on an individual basis.  With group
recertifications, a caseworker provides the reporting requirements to
a group of recipients.  New York City officials believe that group
recertification better ensures that all recipients are receiving the
same information--thereby eliminating the possibility that a
caseworker did not provide the reporting requirements to a recipient. 
State officials also told us that group recertification helps
caseworkers reduce the time they must spend with recipients. 


         CONTACTING RECIPIENTS
         DURING THE CERTIFICATION
         PERIOD
------------------------------------------------------ Chapter 2:2.3.2

Local offices in various states are conducting "Project Recall"
activities where caseworkers telephone recipients during the
certification period to determine if there have been any changes in a
recipient's household or income.  The officials said that Project
Recall reminds recipients that they need to report any changes and
that their case is being followed.  For example, officials in a Texas
office contacted over 6,500 recipients in fiscal year 1994.  The
telephone calls resulted in about 1,500 households having their
benefits lowered or terminated. 


         SHORTER CERTIFICATION
         PERIODS REDUCE
         OVERPAYMENTS
------------------------------------------------------ Chapter 2:2.3.3

States are trying to reduce recipient overpayments by having shorter
certification periods for recipients with earnings.  The longer the
certification period, the more likely that the recipient will have an
income change during the certification time frame that will affect
benefit levels.  FCS regulations require that households be assigned
certification periods based on the predictability of the household's
circumstances.  Households with unstable incomes are to be certified
for no more than 3 months, while households with stable incomes can
be certified for 6 months or longer.  Both Texas and Florida have had
high error rates because their certification periods were too long
for recipients with unstable incomes.  These states were certifying
such recipients for 6 months or longer. 

Of the seven states that we visited, all now have 3-month
certification periods for recipients with unstable incomes. 
According to state officials, the shorter certification period means
that the caseworker has more frequent contact with the
recipients--four recertification interviews a year--and can obtain
updated information from recipients with unstable incomes that do not
report changes when they occur.  By discovering the recipients'
changes sooner, overpayments will be decreased.  Texas officials
estimate that 3-month certifications for recipients with unstable
incomes will reduce the state's error rate by 1.5 percent. 


         WAIVER REDEFINES INCOME
         REPORTING REQUIREMENTS
------------------------------------------------------ Chapter 2:2.3.4

As mentioned earlier, fluctuations in a recipient's income of $25 or
more that is not reported is a major factor in overpayments.  Two
states that we visited, South Carolina and Texas, requested and
received permission from FCS to waive this program requirement and
replace it with an alternate reporting requirement.  With this
waiver, recipients are required only to report a change in their
source of employment, wage rate, or employment status (e.g., changing
from part-time to full-time).  Thus, unreported recipient income of
$25 or more during a certification period is no longer considered an
overpayment. 

According to South Carolina officials, this waiver lowered the
state's error rate by 2 percentage points.  Texas implemented this
waiver in September 1994, and state officials likewise anticipate
that the waiver will reduce the state's error rate by 2 percentage
points. 


      SOME RECIPIENTS DEFRAUD THE
      PROGRAM
-------------------------------------------------------- Chapter 2:2.4

According to FCS' QC data, 24 percent of the overpayments in fiscal
year 1993, totaling $424 million, were a result of recipients
intentionally misrepresenting household information to caseworkers. 
Examples of intentional misrepresentation include withholding
information on employment or bank accounts.  Because the Food Stamp
Program relies on self-reported information, it is hard for
caseworkers to detect these violations at the time of certification. 

The best way for states to decrease recipient fraud is to verify more
recipient-reported information before certifying them to receive
benefits.  However, this is a timely and expensive effort.  Four of
the seven states that we visited conduct investigations prior to
certification, including home visits, of some recipients.  These
states, Florida, New York, South Carolina, and Texas conducted over
58,000 front-end investigations in fiscal year 1993.  These
investigations can be cost-effective.  For example, Texas conducted
3,000 front-end investigations in fiscal year 1994 and, as a result,
did not issue $4.8 million in benefits that it otherwise would have. 

We also found that after certification, some counties in various
states conduct home visits and telephone recipients to verify
information as part of their error-reduction initiatives.  For
example, one Texas region conducted over 5,500 home visits in fiscal
year 1994.  As a result, about 900 households had their benefits
denied and 640 had their benefits lowered.  Program officials told us
that these efforts correct overpayments that may be occurring and
deter client fraud in the area by establishing a presence in the
community. 

According to Food Stamp Program regulations, each state is to
institute an automated system that helps caseworkers verify reported
information by matching the data with other sources.  This system is
called the Income and Eligibility Verification System (IEVS).  The
system contains earned and unearned income information, such as
unemployment compensation and Internal Revenue Service information,
maintained by federal and state agencies.  FCS requires state
agencies to periodically compare their food stamp case income
information with information contained on IEVS.  The data-matching
process helps verify reported income, and it helps identify income
that the applicant failed to report. 

State officials told us that while the system enables them to match
recipient's names with wage reports, its use is reduced because the
information is not timely.  Under the system, employers report
earnings information to the states on a quarterly basis.  Thus,
employment information from the IEVS is usually 3 to 6 months old at
the time the applicant applies for food stamps; accordingly,
caseworkers are not always able to prevent errors at the time of
certification.  However, caseworkers are able to use the IEVS to
detect discrepancies in income later in the certification period and
avoid further overpayments. 

Some states are working to provide caseworkers with better
information to help them verify recipient information.  For example,
in 1993, Massachusetts began requiring all employers to report to the
state all new employees within 14 days of their hiring.  Unreported
employment was the state's biggest single error leading to
overpayments.  With this requirement, food stamp officials are now
matching their recipient data base with the state's Department of
Revenue employment information.  Because the computer matches are
done after certification, the matching does not prevent recipients
who do not report employment from receiving benefits.  However, it
does detect the unreported employment quicker thereby lowering the
length of time recipients get overpayments and the state's error
rate.  As a result of the state's effort, recipients' errors caused
by misreported wages and salaries dropped from 8.6 percent in fiscal
year 1992 to 2.7 percent in fiscal year 1993.\5

Some states have better access to recipients' bank records for
verifying recipient information.  For example, New York and
Massachusetts both require banks to provide the state with
recipients' bank records at no cost.  New York caseworkers contact
several local banks to verify recipient asset information. 
Massachusetts has been able to conduct computer matches of
recipient-reported data with bank records to verify asset information
provided by recipients.  We noted in Massachusetts that there were no
reported asset overpayments in the first 9 months of fiscal year
1994.  Officials in South Carolina told us that they do not have the
authority to obtain recipients' bank records at no cost.  As a
result, caseworkers do not routinely investigate recipients' bank
records because banks charge as much as $25 for each inquiry.  The
officials said that recipients misreporting assets is a significant
cause of errors. 


--------------------
\5 This statistic is for food stamp recipients only.  Massachusetts
has separate statistics for recipients receiving both food stamps and
AFDC. 


   FCS HAS INITIATED EFFORTS TO
   REDUCE OVERPAYMENTS
---------------------------------------------------------- Chapter 2:3

FCS has recently embarked on a number of initiatives to reduce errors
that cause both overpayments and underpayments.  These initiatives
include

  focusing FCS management attention on error reduction,

  reducing the differences between the food stamp and AFDC programs,

  granting states waivers from program requirements, and

  evaluating new incentives to offer states to reduce their error
     rates. 

FCS' emphasis on error-rate reduction represents a clear signal to
the states that the accuracy of program eligibility and benefit
levels must be improved.  However, some of its initiatives may simply
be modifying or removing regulatory requirements that lead to
erroneous payments and may not result in reductions in program costs. 


      FOCUSING MANAGEMENT
      ATTENTION ON ERROR REDUCTION
-------------------------------------------------------- Chapter 2:3.1

FCS increased its management attention and focused on error reduction
during fiscal year 1995.  FCS has taken a number of actions
demonstrating commitment to reducing errors in the Food Stamp
Program.  For example, the FCS Administrator has assembled a core
team at the headquarters level to work exclusively on the development
and coordination of payment accuracy issues.  In addition, a national
conference was held on November 16 and 17, 1994, with state food
stamp officials to discuss benefit payment issues.  At the
conference, the FCS Administrator set a national goal of lowering the
payment error rate by 1 percent in fiscal year 1995.  Also, FCS has
set aside $1 million in fiscal year 1995 for state and federal
error-reduction activities in addition to the $379,000 it has
earmarked for its State Exchange Program.  The State Exchange Program
provides funds that allow states and local agencies to travel to
other localities to observe and share information on methods proven
to reduce overpayments.  Last, FCS regional offices have been
instructed to work more aggressively with states to reduce their
error rates.  In response to this effort, each of the seven FCS
regional offices have developed error-reduction plans for the states
in their jurisdiction. 


      REDUCING THE DIFFERENCES
      BETWEEN THE FOOD STAMP AND
      AFDC PROGRAMS' REGULATIONS
-------------------------------------------------------- Chapter 2:3.2

FCS is working to eliminate some of the differences between the food
stamp and AFDC regulations by approving waivers and changing food
stamp regulations.  By increasing the consistency between food stamp
and AFDC eligibility and benefit determination regulations, FCS staff
believe the chances of caseworker error will be reduced.  FCS has
worked with the Department of Health and Human Services to change
many of the differences in the programs.  FCS has published several
proposed or final rules that reduce inconsistent requirements between
the programs.  For example, FCS has simplified regulations regarding
recipient residency; certification periods; self-employment income
received from taking in boarders or providing daycare services;
student income and eligibility; utility expense reimbursements; and
reporting changes in medical expenses. 


      GRANTING WAIVERS FROM
      PROGRAM REQUIREMENTS
-------------------------------------------------------- Chapter 2:3.3

In October 1994, FCS offered states the opportunity to apply for
waivers from 27 separate parts of the Food Stamp Program regulations
to help them reduce time-consuming procedural requirements.  FCS has
approved workload and error-reduction waivers for all states and
Guam.  Since October, FCS has revised regulations that will make many
of the waivers permanent. 

According to FCS, all of the waivers could have an indirect impact on
error rates.  As mentioned earlier, one of these waives the
requirement for recipients to report certain monthly income changes
of $25 or more during the period for which they are certified to
receive benefits.  FCS has also granted waivers to 16 states that
decrease the number of face-to-face recertification interviews with
recipients from as many as four times a year to only one meeting. 
Reducing the number of interviews will free caseworker time needed to
plan and conduct interviews and should allow caseworkers more time to
focus on reducing their errors as well as recipient errors. 

These changes can affect error rates because they change the
requirements states must abide by in managing the program.  However,
the error-rate reductions attributed to these regulatory changes may
not directly translate into reduced benefit payments or a reduction
in program costs.  For example, waiving the $25 postcertification
income reporting requirement may lower a state's error rate by
eliminating a leading cause of overpayments.  However, there may not
necessarily be a corollary decrease in the total benefits issued to
households.  It may be that the administrative costs associated with
enforcing this requirement and recalculating benefits may exceed the
savings in benefit costs that would result from complying with this
regulation.  In this regard, FCS is awaiting results from the nine
states that have been granted this waiver and will conduct a cost
analysis of this change before proposing regulatory changes or
approving additional waivers. 


      EVALUATING INCENTIVES FOR
      STATES TO REDUCE THEIR ERROR
      RATES
-------------------------------------------------------- Chapter 2:3.4

FCS is preparing a legislative proposal to offer states additional
incentives to reduce their error rates.  The incentives would provide
additional money to states that reduce their error rates below the
national average and in future years maintain error rates below the
national average. 


   CONGRESSIONAL ACTIONS TO REDUCE
   PAYMENT ERRORS
---------------------------------------------------------- Chapter 2:4

Incentives for states to improve the accuracy of food stamp benefit
payments were first legislated in the Food Stamp Act of 1977.  In
this act, states reducing their error rates below 5 percent were
rewarded with additional administrative funding.  In the early 1980s,
the Congress revisited the issue and added incentives to reduce
errors but also mandated that a state repay some benefits to the
federal government if its error rate exceeded the national average. 

Since the Food Stamp Act Amendments of 1980, the Congress, on four
different occasions, has changed the food stamp statutes to address
state payment accuracy.  With each revision, different combinations
of incentives and sanctions to reduce errors, as well as different
approaches to determine sanctionable error rates, have been adopted. 
Sanctions have included a state returning a portion of the benefits
issued or reducing a state's reimbursements for administering the
program.  Incentives have included the opportunity for states to earn
additional, and differing, percentages of administrative funds for
low error rates--depending on the level of error rate achieved.  In
addition to changing incentives and sanctions, the method for
determining error rates has changed.  For example, previously only
overpayment errors were used to calculate error rates, later, the
Congress directed that both overpayment and underpayment errors be
used to determine state error rates. 

The Congress has varied the way in which sanctionable error rates are
determined.  In the early 1980s, the Congress established specific
numerical target error rates that states were to achieve to avoid
sanctions.  Later, the Congress allowed the sanctionable error rate
to be determined by averaging the error rate experiences of the
states.  In 1988, the Congress used the lowest annual national error
rate ever achieved plus 1 percent as the baseline for determining
sanctions.  States exceeding this tolerance level were required to
repay a percentage of benefits issued equal to the percentage amount
that they exceeded the tolerance level.  Later, in 1993, the Congress
changed the sanctionable error rate from the lowest national annual
error rate ever achieved to an error rate based on the national
average for the current year.  A state sanction is calculated based
on the degree the state exceeds the sanction.  There is no 1
percentage point tolerance; however, the determination of the
sanction is based on a sliding-scale calculation.  This sliding-scale
method of calculating the sanction has the effect of reducing the
amount of sanctions for states near the national tolerance level
while increasing the sanction for states exceeding the tolerance by a
greater margin. 

FCS data indicate that congressional efforts to reduce errors have
produced limited results.  For example, few states have been able to
achieve error rates that would qualify them for enhanced funding. 
According to FCS, no more than eight states received enhanced funding
in any year during the 14-year period between 1980 and 1993.  The
states that did qualify for enhanced funding received about $56
million in additional funds. 

During this same period, almost all states were sanctioned for high
error rates, totaling about $857 million.  Under current statutes,
sanctioned states are given the choice of making payments to the
federal government for the sanctioned amounts or reinvesting these
funds in initiatives to improve their payment accuracy.  However, as
a result of legal challenges and congressional action, states have
ultimately been held accountable for a reduced portion of the
sanctioned amounts.  For example, only about $51 million of the
sanctions have been collected or reinvested by states in improving
the food stamp payment accuracy. 

The Congress, in the Personal Responsibility Act of 1995 (H.R.  4),
passed by the House of Representatives on March 24, 1995, is
currently contemplating fundamental changes to the federal food
assistance structure, which includes a return to the 1988 standard
for determining state sanctions for food stamp payment errors.  Under
current circumstances, the formula in this provision would make
subject to sanctions states that exceed a 10.3-percent error rate. 
In view of the national error rate for 1993 of 10.8 percent, more
states will probably be sanctioned in the future, and the amount of
the sanctions will likely increase, unless states substantially
improve their payment accuracy performance. 


   CONCLUSIONS
---------------------------------------------------------- Chapter 2:5

The experiences of the states that we visited, some with historically
high error rates, clearly indicate that food stamp error rates can be
reduced.  The primary factor in lowering error rates in these states
appears to be the willingness of states to focus on reducing
overpayments and making a commitment to do so.  The major motivation
for states to reduce overpayments seems to be linked to the desire to
either avoid federal sanctions for high overpayments or to "earn"
additional funding for their states by having low error rates. 

FCS' increased commitment to reduce payment errors should send a
clear signal to states that the accuracy of food stamp benefit
payments must be improved.  While these initiatives are steps in the
right direction to improving payment accuracy, it is difficult to
determine if they will be sufficient to motivate states to do a
better job in determining recipients' eligibility and benefit
payments and the bottom-line impact they will have on reducing
overpayments.  For example, states that are historically not above
the sanction level or reasonably close to the enhanced funding level
may not be willing to make the necessary commitment.  In these
instances, a different incentives/sanctions package may be needed. 

With regard to FCS initiatives to simplify and grant waivers to Food
Stamp Program regulations, while they may be beneficial in reducing
payment errors, caution should be exercised to ensure that these
regulatory changes are the result of sound business decisions aimed
at (1) removing nonessential or burdensome administrative regulatory
processes or (2) updating program requirements to reflect current
conditions and are cost-effective in terms of program benefit and
administrative costs. 

FCS is considering a legislative proposal to provide additional
incentives to states that lower their error rates.  At the same time,
the Congress is considering changes to Food Stamp Program legislation
that would tighten sanctions against states with high error rates. 
It is difficult to predict the impact these changes could have on
state payment accuracy, if enacted, because of the past experience in
enforcing sanctions.  However, the changes are options that could
have positive impacts on reducing payment errors and are worth
exploring. 


MORE MONITORING OF RETAILERS COULD
HELP DETER TRAFFICKING
============================================================ Chapter 3

The precise extent of food stamp trafficking by retailers and
recipients is unknown, but law enforcement officials believe that
billions of dollars of food stamps are trafficked each year. 
Existing FCS controls and procedures have not proven effective in
reducing retailer involvement in food stamp trafficking.  Stores that
do not meet eligibility criteria are being authorized to redeem food
stamps, and once admitted, FCS' monitoring process is inadequate to
detect retailers that violate program regulations during their period
of certification.  The major weakness in FCS' process is its lack of
resources to make site visits to stores to ensure that they meet
program eligibility criteria and comply with basic program
regulations.  FCS' resources are insufficient to investigate all of
the retailers suspected of trafficking.  FCS has various initiatives
underway to improve its authorization and monitoring of food stores
and is proposing additional actions.  However, these initiatives do
not include providing additional resources to make site visits. 

With regard to recipient trafficking or other recipient misuses of
program benefits, it would be difficult and expensive to eliminate
these program abuses.  As long as food stamp coupons are used for
delivering benefits and are available for use as a second currency,
recipient misuse will likely continue.  FCS efforts to reduce the
number of violating retailers, however, should make it more difficult
for recipients to use their benefits for nonfood purchases. 


   THE EXTENT OF FOOD STAMP
   TRAFFICKING IS UNKNOWN
---------------------------------------------------------- Chapter 3:1

Due to the difficulty in collecting data, neither federal nor state
officials know the extent to which recipients and retailers are
trafficking or misusing food stamp benefits.  As cited in our
December 1994 report (see ch.  1), estimates of trafficking run as
high as 10 percent of benefits issued--or over $2 billion annually. 
This estimate has been widely reported in the media, but we were
unable to corroborate the estimate.  Program officials and law
enforcement officials involved in policing the Food Stamp Program
believe that recipient trafficking is pervasive, and food stamps are
often traded on the streets as a second currency.  However, the
officials have not conducted any definitive studies on trafficking. 
At the same time, it is clear that trafficking is occurring.  For
example in fiscal year 1993, FCS found 841 retailers involved in
trafficking food stamps. 


   FCS CONTROLS ARE INADEQUATE TO
   PREVENT INELIGIBLE STORES FROM
   BEING AUTHORIZED TO ACCEPT FOOD
   STAMPS
---------------------------------------------------------- Chapter 3:2

FCS has a comprehensive process for collecting information on stores
applying for authorization to redeem food stamps.  However, this
process relies almost totally on the integrity of applicant retailers
to accurately report information on their stores and business
activities.  While FCS' authorization process allows on-site visits
to corroborate retailer-reported information, such visits are the
exception rather than the rule.  Based on our visits to FCS field
offices and review of OIG reports, it appears that stores are not
visited prior to authorization primarily because FCS has assigned a
relatively low priority and devoted few resources to retailer
monitoring. 


      RETAILER AUTHORIZATION IS
      PRINCIPALLY A PAPER PROCESS
-------------------------------------------------------- Chapter 3:2.1

The authorization of retail stores to accept food stamps is delegated
to FCS' seven regional offices.  The regional offices have in turn
delegated this responsibility to its field offices.  According to FCS
regional officials, the process of authorizing and reauthorizing
stores for participation in the Food Stamp Program is similar at
every FCS field office.  A five-part FCS application is provided to
retailers requesting authorization to redeem food stamps.  The
application calls for the retailer to provide detailed information on
the store and its business activity.  For example, the retailer must
provide information on the location and address of the store;
estimated or actual sales; types of foods offered for sale; and
ownership information, such as employee identification number and
owner social security number. 

Field office personnel review the applications for completeness and
conformity to program regulations and determine if supplemental data
are required from the retailer.  At the eight FCS field offices we
visited, personnel generally ask for more information than required
by the program regulations, such as picture identification, sales and
ownership records, and naturalization and passport documentation. 
Other supplemental information the offices request can include such
items as operating and/or beverage licenses, health inspection
certificates, and income tax statements.  Incomplete applications are
returned to the retailer with instructions for providing the missing
information.  Field office personnel will discuss the incomplete
application with the applying retailer over the telephone.  If the
additional information is not forthcoming, the field office personnel
will deny the application.  The retailer may reapply at any time. 
When an application is determined to be complete, field office
personnel have 30 days in which to process and approve or disapprove
it. 

Retailers must also be given training on the program.  In some field
offices, the retailer is instructed to attend a 1- to 2-hour training
session given by field office personnel on program requirements and
procedures, such as how to redeem coupons, products that cannot be
purchased with food stamp coupons, and penalties for inappropriate or
illegal uses of the coupons.  In other field offices, retailer
authorization training occurs by telephone or through the mail or by
individual instruction.  In either event, the owner is not certified
to accept food stamps until the training is completed. 

Stores participating in the program must periodically reapply to
update their authorization certificate by making a new application
and providing updated sales data and other store information for
retail monitoring.  The frequency of FCS' reauthorization process
varies by the type of stores.  Major supermarkets are reauthorized
every 3 years.  All other stores--convenience, small grocery, and
privately owned neighborhood stores--are reauthorized every 2 years. 
According to FCS officials, smaller retailers are reauthorized more
frequently because FCS and OIG officials have identified them as the
ones more likely to violate program regulations than larger
supermarkets. 

The reauthorization process mirrors the process used to initially
authorize stores.  Retailers seeking reauthorization submit an
application form to the field office, which reviews the form for
completeness and the appropriateness of the store to continue to be
authorized to accept food stamps.  In addition, field offices check
the information in the reauthorization application with information
on the store contained in FCS' Store Tracking and Redemption System
(STARS).  The STARS is the major repository of information on each
store that is authorized to redeem food stamps.  The system contains
information submitted by the retailer in the initial application as
well as information on the redemption history of the store.  Field
office officials compare the sales reported by the retailer against
the redemption history of the store.  This analysis allows them to
identify stores whose redemptions are high in relationship to total
food sales.  FCS officials told us that in such cases the retailer
might be contacted by field office personnel to provide an
explanation as to why the disparity exists between their sales and
food stamp redemptions.  Failure to provide requested information may
result in the withdrawal of the store's approval to participate in
the program. 

FCS data show that 32,815 stores, or 16 percent of the authorized
stores, were withdrawn from the program as a result of the
reauthorization process in fiscal years 1992 and 1993.  According to
FCS officials, data on withdrawn stores includes stores that are no
longer open for business, stores that fail to meet program criteria,
and stores that no longer wish to participate.  FCS does not maintain
information specifically on how many of the withdrawn stores failed
to meet program criteria. 


      ON-SITE VISITS ARE NOT
      NORMALLY A PART OF THE
      AUTHORIZATION AND
      REAUTHORIZATION PROCESSES
-------------------------------------------------------- Chapter 3:2.2

According to FCS officials responsible for authorizing stores, they
rely almost exclusively on self-reported information provided by the
retailers and information in STARS in making decisions to authorize
or reauthorize a store to redeem food stamps.  While FCS procedures
provide for on-site visits to retailer stores to verify provided
information, field office officials told us that they make few
on-site visits.  We found that neither FCS headquarters nor two of
the three FCS regions that we visited had any overall data on the
number of on-site visits made to stores prior to authorization.  We
did find information on retailer visits in FCS' Southwest Region,
however.  Regional records indicated that out of 9,564 applications
received in 1994, 269 on-site reviews were completed. 

FCS officials told us that on-site reviews are the most effective way
to prevent inappropriate stores from being authorized to accept food
stamps.  Through on-site reviews, officials are able to visually
verify that stores are legitimate retailers with ample food
inventories.  However, the officials said that they do not have
sufficient personnel and resources to make store visits.  In the
1970s, FCS' policy required periodic on-site visits of retailers. 
FCS officials told us that at that time approximately 1,500 staff
years were devoted to this effort.  Since that time, the policy has
changed to no longer require on-site visits, and the associated
resources have been reduced. 

According to FCS officials, there are 323 staff responsible for all
field office activities.  In addition to authorizing and
reauthorizing retailers, these staff have other responsibilities. 
For example, they are tasked with conducting management evaluations
of state food stamp programs; monitoring and evaluating the school
feeding programs; conducting quality control reviews of recipient
case files; monitoring tribal commodities programs; completing
administrative processing of retailers that the FCS Compliance Branch
has confirmed violated program regulations; and various other duties
assigned to them by the individual regions. 

Field office officials in all three regions that we visited told us
that additional personnel alone would not be sufficient to enable
them to make on-site visits.  Additional funds for travel to stores
would also be required.  For example, the Dallas field office is
responsible for monitoring about 5,700 stores, spanning 136 counties
from the Louisiana border in the east to El Paso in far west Texas. 
The office receives between 100 and 130 new applications monthly. 
The workload is managed by five full-time employees and two part-time
students.  The Dallas field office also reviews child and adult day
care programs, oversees food distribution warehouses, monitors
compliance for the food assistance programs, and investigates civil
rights complaints. 

Two states that we visited--New Mexico and South Carolina--have
approved FCS pilot programs allowing them, rather than FCS, to
authorize retailers.  At the time of our visit, New Mexico had a
part-time employee conducting its retailer authorizations.  In
addition to the part-time employee, the state is relying heavily on a
private contractor installing point-of-sale terminals in retail
stores that are to be used in its new Electronic Benefit Transfer
(EBT) system for delivering food stamp benefits to detect problem
stores.  Furthermore, New Mexico officials told us that they did not
intend to continue the project unless they receive 100-percent
federal funding.  While South Carolina officials said that they would
like to continue authorizing stores after the pilot project ends in
December 1995, they are not prepared to fund the program sufficiently
to do on-site visits. 


   STORES ARE NOT ADEQUATELY
   MONITORED TO ENSURE THAT THEY
   COMPLY WITH PROGRAM RULES
---------------------------------------------------------- Chapter 3:3

FCS' primary methods for monitoring stores once they are authorized
consists of reviewing computer-generated reports showing the
activities of the stores and on-site investigations by the FCS
Compliance Branch.  Our discussions with field office personnel
disclosed that they depend most heavily on three computer reports
which, in addition to STARS, (1) identify potential trafficking by
analyzing the ratio of food stamp redemptions to total food sales;
(2) list stores that have been disqualified from the program, but
have continued to redeem food stamps; and (3) identify various
characteristics that are commonly found in stores that have
trafficked in food stamps and/or committed other serious program
violations. 

Our review indicates that all of these reports have limitations as
tools for monitoring retailers.  Also, in March 1992, an OIG audit of
FCS' retailer monitoring system found that the computer reports used
most often by field offices were inaccurate, unverified, or
outdated.\6 Field office personnel basically agree with the OIG's
findings.  However, the personnel maintain that because no other
information is available they continue to use these exception reports
for monitoring stores, even though they are often untimely and in
some cases inaccurate. 

FCS officials also told us that there are problems with STARS that
slows its response time and makes it cumbersome to use.  The
officials added that because the STARS report organizes stores by
type of store in geographical locations, not alphabetically, it is
difficult and time-consuming to use.  FCS and OIG personnel believe
that existing computer reports are inadequate for identifying a
reportedly large, but unknown, number of stores trafficking in food
stamps that are not authorized by FCS to accept food stamps.  FCS and
OIG personnel know that these stores exist because they are
occasionally discovered in investigations of other recipients and
stores, as well as from informants. 

The OIG also found in its 1992 audit that field office personnel
seldom performed monitoring visits.  FCS does not maintain data on
the number of on-site visits made to monitor stores after they are
authorized.  In discussing this issue with field office officials, we
were told that few on-site visits are made because they do not have
the time and resources necessary to make visits.  The FCS and OIG
officials that we spoke with told us that the lack of federal
presence in the field through on-site visits compromises the
integrity of the program. 

We asked FCS regional and field office officials what additional
resources would be needed to conduct more monitoring visits of
authorized stores.  The officials told us that they had not developed
estimates of the additional personnel and travel funds needed to
perform such visits.  In discussing the cost to make on-site retailer
visits with FCS headquarters officials, we were told that about $19
million would be needed to make an annual visit to nonsupermarket
retail stores.  Nonsupermarket stores are considered by FCS officials
to be more prone to traffic food stamps than supermarkets.  This
estimate is based on:  visiting 200,000 stores annually; with an FCS'
official visiting 3 stores per day; an average work year of 221 days;
at an average cost of $62,000 per person, which would include salary,
benefits, and travel. 


--------------------
\6 Food Stamp Program-Authorizing and Monitoring of Retailers, USDA
OIG, Audit Report 27600-0008-Ch (Mar.  31, 1992). 


   STAFF SHORTAGES RESULT IN
   SUSPECTED VIOLATORS AVOIDING
   INVESTIGATION
---------------------------------------------------------- Chapter 3:4

In addition to authorizing stores to redeem food stamps and
monitoring their performance in complying with program regulations,
FCS is responsible for investigating stores suspected of trafficking
food stamps or violating other program regulations.  FCS has only 46
investigators nationwide to conduct these investigations. 
Investigations are initiated by field office personnel based on
citizen or retailer complaints and FCS monitoring of reports.  Field
office personnel refer complaints about stores with suspicious
activity to FCS' Compliance Branch for investigation.  Once
investigated, Compliance Branch personnel determine whether the case
is to be administratively processed or forwarded to the OIG for
criminal investigation.  Table 3.1 shows the number of stores that
the Compliance Branch investigated during the past 5 years. 



                          Table 3.1
           
           FCS Compliance Branch Investigations for
                  Fiscal Years 1990 to 1994

                        1990    1991    1992    1993    1994
--------------------  ------  ------  ------  ------  ------
Stores investigated    5,447   4,984   4,848   4,644   4,300
Stores found to be       358     531     763     841     902
 trafficking\
------------------------------------------------------------
As table 3.1 shows, there has been a decline in the number of stores
investigated during this 5-year period.  Compliance Branch officials
told us that staff limitations have prevented them from investigating
all retailers suspected of program violations.  However, table 3.1
shows that the number of stores found to be trafficking increased
from 358 in fiscal year 1990 to 902 in fiscal year 1994. 

In instances where large amounts of food stamps are suspected of
being trafficked, the Compliance Branch refers these cases to the
USDA OIG for criminal investigation.  The OIG also receives referrals
from citizens and local law enforcement officials.  OIG officials in
two regions told us that they investigate about 10 percent of the
cases referred to them by the Compliance Branch and other sources. 
OIG information shows that its food stamp investigation caseload has
fluctuated between 1990 and 1994, but the amount of OIG time spent
investigating these cases has steadily increased from 35 to 45
percent of their investigative workload over the same time period. 

FCS and OIG officials told us that civil and criminal investigations
are very resource-intensive and that, due to a shortage of
investigators, many suspect stores are not investigated for program
violations such as trafficking.  Furthermore, field office personnel
tell us that they refer only the most significant cases for
investigation to the Compliance Branch.  According to field office
officials, no records are kept of the number of potential cases not
referred for investigation.  Instead of referring these cases to the
Compliance Branch, field office personnel give the stores a warning. 


   USDA IS PURSUING NUMEROUS
   INITIATIVES TO IMPROVE RETAILER
   INTEGRITY
---------------------------------------------------------- Chapter 3:5

On March 1, 1995, the Secretary of Agriculture issued a statement on
a number of new initiatives to reduce food stamp trafficking.  The
statement called for several legislative changes to help identify
unscrupulous retailers during the authorization process and keep them
from defrauding the program.  While these proposed legislative
changes should help improve retailer integrity in the program, they
do not address the problem of insufficient staff to authorize and
monitor stores and pursue investigations of retailers suspected of
violating program regulations. 

Some of the major legislative changes requested include: 

  requiring retailers to provide sales and income tax filing
     documents when applying for authorization along with permission
     for FCS to verify such information with other agencies;

  requiring a 6-month waiting period before reapplication by a store
     that does not meet initial eligibility criteria;

  allowing FCS to set time limits for retailer authorization,
     including shorter authorization periods for questionable
     retailers;

  suspending stores charged with trafficking while their cases are
     pending administrative or judicial review thereby preventing
     them from continuing to abuse the program during the appeals
     process;

  suspending stores charged with trafficking on the basis of
     electronic benefit transfer transaction data;

  expanding forfeiture authority to allow the seizure of property
     used by traffickers in felony food stamp transactions of $5,000
     or more, including property gained with the proceeds of illegal
     transactions;

  permitting permanent disqualification of retailers who
     intentionally falsify their applications;

  increasing the penalties for recipient trafficking; and

  disqualifying stores from participating in the Food Stamp Program
     that have been disqualified from participating in the Special
     Supplemental Nutrition Program for Women, Infants, and Children. 

In addition to seeking legislative changes, FCS plans to expand
retailer monitoring capabilities for states using EBT technologies to
deliver food stamp benefits.  FCS is currently testing the use of
automated EBT exception reports to identify suspicious transaction
patterns by retailers and recipients.  By the end of 1995, FCS
expects to have in place an automated system for analyzing EBT
transaction data and producing exception reports of suspected
violators for all states with EBT operations.  FCS is also providing
states that use EBT a checklist for evaluating retailer's eligibility
while installing EBT terminals in the store.  Also, FCS is requesting
authority to share retailer-provided social security and employer
identification numbers with state investigative and law enforcement
agencies in order to detect illegal retailer activity. 

The Personal Responsibility Act (H.R.  4) includes provisions giving
FCS a number of the authorities it has requested.  While these
authorities should help FCS' enforcement capabilities, on-site visits
to retailers applying for authorization to redeem food stamp benefits
is crucial to preventing unscrupulous retailers from being authorized
to accept food stamps.  The proposed act also provides for on-site
visits of food stores by FCS or state or local officials before
stores are authorized to redeem food stamps.  However, neither the
act, nor FCS, address the issue of monitoring retailers once stores
are authorized to participate in the program nor does either address
the need for additional resources to investigate retailers suspected
of violating program regulations. 

FCS has indicated that it has plans to increase the monitoring of
retailers, but it has not specifically said how these plans will be
carried out nor what resources will be needed to implement them.  To
the extent that FCS may be required to make preauthorization store
visits, and/or decide to increase its efforts to monitor and
investigate retailers, it will need to reassess its resource needs
and allocations. 


   IMPROVEMENTS IN THE MONITORING
   OF RETAILERS SHOULD HELP REDUCE
   RECIPIENT TRAFFICKING
---------------------------------------------------------- Chapter 3:6

Detecting and preventing recipient trafficking is normally the
responsibility of the states under Food Stamp Program regulations. 
State and federal officials told us that with 27 million recipients
in 11 million households currently receiving food stamp benefits, the
costs of preventing individual recipients from trafficking exceeds
the cost of the benefits being trafficked.  The officials believe
that recipient food stamp trafficking is widespread, but they have no
data to verify this belief.  Because of the difficulties in detecting
recipient trafficking--and the cost involved to do so--state
officials told us that they generally do not pursue recipient
trafficking.  Some state program officials prefer to focus their
limited resources on preventing, detecting, and prosecuting violating
stores, which they believe will have a greater impact on reducing
total recipient trafficking than by attempting to pursue violators
individually.  While reducing the number of violating stores will not
completely eliminate recipient trafficking, making it more difficult
to find sources for trafficking food coupons will discourage the
illegal use of food stamps. 


   CONCLUSIONS
---------------------------------------------------------- Chapter 3:7

Regardless of whether food stamp benefits are issued in the form of
coupons, electronically through electric benefit transfer
technologies, or in the form of cash, some unknown number of the
approximately 27 million recipients are going to use their benefits
for other than the intended purpose of buying food.  Controlling the
misuse of benefits can be a costly undertaking requiring a
significant amount of resources at the federal and state levels. 

An effective approach for controlling trafficking is to root out
retailers that redeem food stamps for other than legitimate food
items.  The 209,000 stores authorized to redeem food stamps represent
the only gateways through which food stamp coupons can enter the
banking system and be redeemed by the federal government.  Without
these gateways, recipient traffickers as well as other third-party
traffickers have no means to "redeem" their coupons for nonfood
purposes.  Thus, reducing the number of retailers that traffic food
stamps would close the gateways by which illegally used benefits make
their way back to the banking system for redemption.  Furthermore,
monitoring and controlling the 209,000 retailers is easier than
trying to monitor and control the behavior of approximately 27
million recipients. 

FCS is pursuing several initiatives to improve its authorizing and
monitoring of retailers, including requesting additional legislative
authorities, and thereby reduce food stamp trafficking.  The Personal
Responsibility Act includes provisions giving FCS a number of the
authorities requested.  While the proposed changes would be
beneficial, they do not address the resources needed to perform
on-site visits at the time of retailer authorization or for
monitoring visits after authorization.  Such visits are considered to
be an effective mechanism in preventing retailer involvement in
trafficking. 


   RECOMMENDATION
---------------------------------------------------------- Chapter 3:8

In view of proposed legislation, we recommend that the Secretary of
Agriculture direct the Administrator of FCS to determine the
resources needed to incorporate on-site visits into its processes for
authorizing and reauthorizing stores to accept food stamp benefits. 
This analysis should include a determination of the resources needed
to effectively monitor stores once they are authorized to accept
benefits and to investigate stores suspected of food stamp
trafficking.  The Secretary should consider the results of this
analysis in the setting of departmental priorities and the allocation
of resources, and provide this information to the Congress for its
use in considering program reforms. 


   AGENCY COMMENTS AND OUR
   EVALUATION
---------------------------------------------------------- Chapter 3:9

In commenting on a draft of this report, the Administrator of the
Food and Consumer Service did not believe that the resource analysis
that we recommended to help reduce trafficking was necessary. 
Rather, the agency believes that the smarter use of new
technology--especially the future use of electronic benefit transfer
systems--along with the passage of its proposed retailer-related
antifraud legislation will enable it to detect and remove violating
stores more efficiently.  Moreover, the agency believes that its
antifraud proposals will help ensure that only legitimate stores are
authorized to redeem food stamps. 

We agree that USDA's initiatives should help deter retailer
trafficking of food stamps.  However, we believe that it is important
to consider that the effectiveness of the agency's legislative
proposals in reducing trafficking is unproven at this time and will
not be known until some time in the future.  Also, we do not believe
that the use of electronic benefit transfer systems will necessarily
eliminate the need to conduct store visits.  Instead, it will serve
as a valuable supplementary tool in the effort to reduce trafficking. 
Furthermore, the implementation of electronic benefit systems is in
its infancy, and nationwide implementation will not be completed for
several years.  In this context, we continue to believe that on-site
visits to retailers should be included in any overall strategy to
increase retailer integrity and that the Food and Consumer Service
should take the necessary steps to ensure that such visits are made. 




(See figure in printed edition.)Appendix I
AGENCY COMMENTS AND OUR EVALUATION
============================================================ Chapter 3



(See figure in printed edition.)



   GAO'S COMMENTS
--------------------------------------------------------- Chapter 3:10

1.  FCS suggested a number of technical changes to update and clarify
certain information provided in the report.  We have incorporated the
agency's comments where appropriate. 

2.  After consideration of agency comments, we modified the latter
part of this recommendation to recommend that the Secretary provide
the results of the Department's resource analyses to the Congress for
its use in considering program reforms. 

3.  We agree that USDA's initiatives should help deter retailer
trafficking of food stamps.  However, we believe that it is important
to consider that the effectiveness of the agency's legislative
proposals in reducing trafficking is unproven at this time and will
not be known until some time in the future.  Also, we do not believe
that the use of electronic benefit transfer systems will necessarily
eliminate the need to conduct store visits.  Instead, it will serve
as a valuable supplementary tool in the effort to reduce trafficking. 
Furthermore, the implementation of electronic benefit systems is in
its infancy, and nationwide implementation will not be completed for
several years.  In this context, we continue to believe that on-site
visits to retailers should be included in any overall strategy to
increase retailer integrity and that the Food and Consumer Service
should take the necessary steps to ensure that such visits are made. 


MAJOR CONTRIBUTORS TO THIS REPORT
========================================================== Appendix II

Robert A.  Robinson, Associate Director
James A.  Fowler, Assistant Director
James G.  Cooksey, Project Leader
Syrene Mitchell, Evaluator
Patricia Sari-Spear, Evaluator
Leigh M.  White, Evaluator

