frequent sources of error-work registration and determination of income. Each accounted for about 40 percent of the cases in which errors were noted. Since OIG and GAO have noted errors in compliance with work registration requirements in many States, some explanation is in order. Failure of nonexempt recipients to register for work is normally a caseworker problem rather than a matter of recipient abuse. The problem usually stems from the caseworker failing to require an applicant to complete a work registration form.

Two offices we visited in Puerto Rico did not require members of households to register for work because the offices did not have the required forms. Noncompliance can also result from an incorrect determination that a household is exempt from the work registration requirement, and we found some instances of this as well. These errors in work registration do not generally result in overissued benefits, because compliance with the work registration requirement does not produce savings unless the household member actually finds employment. Given the high unemployment rate in Puerto Rico, actual savings would probably not be large. However, work registration is a legal requirement and there is no excuse for not imposing it.

We also found problems in the treatment of income in 40 percent of the cases in which errors were noted. These included computation of errors, differences in the income reported to the caseworker compared with earnings disclosed during our home visits, and acceptance of recipients' statements about self-employment without documentation. If work registration errors are excluded but other errors in certification are included for those with work registration omitted, total coupons issued to improperly certified households amounted to $15.3 million.

RECONCILIATION

Prior to the elimination of the food stamp purchase requirement States were required to reconcile the authorization cards redeemed with food stamps issued and the amount of cash collections turned over to the Federal government. When we used all available information to reconcile cash collections and coupon issuances, shortages of $12.1 million in cash and $3.7 million in coupons were indicated for the period from July 1974 through December 1978. The Food and Nutrition Service has billed the Commonwealth of Puerto Rico for the full amount of the shortage. The actual amount to be collected will not be determined until some time in the future while the Commonwealth continues to search for documentation or reconstruct past transactions to account for the shortages. To date about $2.3 million of the cash shortage has been explained. OIG is continuing to review Puerto Rico's documentation.

The Commonwealth has developed a corrective action plan to deal with the other problems that we pointed out. That plan is now being reviewed by the Food and Nutrition Service. Work has already begun on the correction of some deficiencies. Last May, Public Law 96-249, which requires a special financial audit review by OIG of any "State" where participation in the Food Stamp Program exceeds 60 percent of the population, was enacted. In effect, this provision applies to Puerto Rico only since no other State is close to a 60-percent participation level. We maintain audit and investigative staffs in Puerto Rico and have monitored the program almost continuously since it began. We will continue to review program operations on a frequent basis even if the 60-percent threshhold is not reached because of the size of the program in Puerto Rico and the history of problems in administration.

MEMPHIS AND NASHVILLE, TENN.

We have recently completed our audit fieldwork on food stamp operations in Memphis and Nashville, Tennessee. At both locations, we utilized computer matching techniques to compare income reported on the food stamp application to that reported to the State Employment Security Bureau. In March 1980, there were approximately 59,000 households in Memphis and 22,000 households in Nashville participating in the Food Stamp Program. Our matching results show that about 1,570 households underreported their income for food stamp purposes.

About 75 of these cases are being prosecuted through the Federal Court system. These cases resulted in large dollar losses, not only to the Food Stamp Program, but also to Health and Human Services (HHS) programs, such as the Aid to Families with Dependent Children (AFDC), Supplemental Security Income (SSI) and Medicaid. The remaining cases are being handled by State investigators and will be prosecuted in the State court.

In addition to the affect on the Food Stamp Program of underreporting income, we also reviewed its impact on the AFDC, Medicaid, SSI, and Housing and Urban Development (HUD) housing assistance program.

Of the 1,570 identified food stamp fraud cases, about 500 of these households participated in AFDC and Medicaid programs; another 100 households were receiving SSI and Medicaid, and about 125 were receiving HUD housing assistance. State and local investigators have now completed their work to determine program losses on about 329 of the 1,570 cases. From their results, we estimate that the 1,570 fraud cases resulted in losses of about $3 million to the Food Stamp Program, and about $800,000 to the AFDC program. The Department of Health and Human Services, Office of Inspector General, is presently working to determine the losses to the Medicaid and SSI programs. HUD officials estimate an average loss of $2,000 per case for the 125 HUD housing assistance cases.

DADE COUNTY, Miami, Fla.

In our recently completed audit in Dade County, we matched the income of households participating in the Food Stamp Program in September 1979 to the State Employment Security Bureau income records. Of the approximate 86,200 participating households, 32,000 had some type of error. We estimated that the total dollar loss was about $1,002,530 for this one-month period.

Of the errors creating program losses, 81 percent were due to the recipient underreporting income. Our estimate of the dollar loss attributable to underreporting was $813,000 of the total $1 million plus dollar loss.

Florida is now in the process of refining their computer system. By August 1981, the State will have the capability in Miami to compare income reported by the recipient to the income reported to the State Economic Security Bureau when the initial certification is being processed. This will enable State officials to detect recipients underreporting income before benefits are issued. It should also act as a deterrent to recipient underreporting.

Senator LEAHY. What percentage of the households is that, approximately?

Mr. PETERSON. I am not sure I have that, Senator. I could present it for the record.

Senator PRYOR. Just to clarify the record, it is my understanding that about 57 percent of the population of Puerto Rico is today on food stamps, and roughly 1.8 million people, or about 10 percent of the food stamp budget goes to Puerto Rico, is this correct?

Mr. PETERSON. That is correct.

The CHAIRMAN. Thank you very much, gentlemen.1

The next witness is Mr. Henry Eschwege, Director, Community and Economic Development Administration, General Accounting Office.

Mr. Eschwege, as you are aware, you have a 10-minute time limitation.

Your entire statement will be included in the record.

STATEMENT OF HENRY ESCHWEGE, DIRECTOR, COMMUNITY AND ECONOMIC DEVELOPMENT DIVISION, GENERAL ACCOUNTING OFFICE

Mr. ESCHWEGE. Yes; I will do that. I would like to introduce my colleagues.

On my immediate left is Mr. Brian Crowley, Associate Director in our office, and on his left is Stan Sargol, who is our Group Director.

I will summarize the statement, and it also has a rather long appendix on the workfare program, which may be of interest, but I will just talk about that briefly."

See p. 366 for written questions from Senator Helms and Mr. Magee's answers thereto. 2 See p. 369 for the prepared statement of Mr. Eschwege.

We are here at your invitation to provide information on the results of some of our reviews relating to the Department of Agriculture's food stamp program.

We also want to offer our views on the President's proposed budget reductions affecting this particular program.

There are 13 major Federal domestic programs, costing many billions of dollars a year that provide food.

The President has proposed eliminating the overlap between food stamps and school lunch subsidies.

We agree that substantial savings would be possible by eliminating this overlap. This subject was included in our June 1978 report, which discussed food program benefit overlaps and gaps, and differences in program eligibility.

Our review showed at that time that some low-income families participated simultaneously in as many as six different programs providing food assistance.

We found that for households whose program benefits exceeded 100 percent of thrifty food plan costs, multiple program combinations most frequently involved food stamps, aid to families with dependent children, school lunch, school breakfast, and special milk.

In our report we estimated the additional costs to the Federal Government based on the assumptions we described in that report. We noted, however, that our estimate was intentionally conservative, because we sought to avoid overstating potential savings. Different assumptions would result in different estimates.

We have a new updated estimate. It would now be about $566 million a year.

The legislation and regulations governing the various Federal food assistance programs contained major differences in basic eligibility criteria and procedures. Also, although all of the programs are either totally or partially directed to the needy, there is no single uniform definition of the term needy, which applies to all of the programs.

Most of the programs have income standards against which potential participants are measured, but such standards have their limitations, and allowable income exclusions and deductions are not consistent among the programs.

Such inconsistencies can create inequities and administrative complications and confusion. There is still a need to act on our earlier recommendations.

Another area of potential savings that was discussed was the use of individualized food stamp allotments. By regulation and law, the uniform stamp allotments are based on thrifty food plan costs for a model four-person household consisting of two adults, a man and woman, and two children.

Uniform allotments for households of other sizes are computed from the four-person household allotment level with adjustments for economy of scale, because nutritional needs and thrifty food plan costs varied depending on the specific sex and age makeup of the household. Use of uniform food stamp allotments could enable some families with low nutritional needs, such as households consisting of a mother and three young children to receive more

benefits than would be indicated by their individual thrifty food plan costs.

Now, in contrast, benefits based on uniform allotments would probably be less than indicated for a four-person household in which the children were all, let's say, teenage boys.

In our 1978 report, we estimated the savings achievable at that time by using individualized rather than uniform allotments. If households receiving less than their individualized thrifty food plan costs were given enough extra food stamps to bring their allotments up to their thrifty food plan costs, the increased monthly costs would be about $12.7 million.

Correspondingly, if households whose food stamp allotments were more than their thrifty food plan costs had their allotments reduced to their thrifty food plan level, the monthly savings would be about $60.1 million, the net savings in food stamps to the Federal Government at the time of our 1978 report, of about $47.4 million a month, or about $570 million annually.

I might say that the Congressional Budget Office in their report in 1977 also dealt with this particular issue. Without an extensive update of our earlier study, we can only offer an educated guess of potential savings.

At this time total program benefits have increased about 92 percent since we analyzed allotments for the 1978 report, and on this basis, annual savings from individualized allotment might well approach the $1 billion mark.

Now, in a review completed in 1977, we assessed efforts to identify and recover overissuances of food stamp benefits. At that time we estimated that the Government was losing over a half a billion dollars annually, because of errors, misrepresentation, and suspected fraud by recipients, and because of errors by local food stamp offices.

For every $100 of the more than $5 billion in benefits, issued nationally in fiscal year 1976, overissuances accounted for about $12, only about 12 cents of that $12 had been recovered at that point in time.

The local projects we reviewed were doing little to identify and recover the value of overissuances.

Since then there have been a number of legislative and administrative changes to deal with fraud and other overissuances.

According to recent Agriculture data, State recovery of overissuances have improved somewhat.

For every $100 of the nearly $8.7 billion in benefits issued nationally in fiscal year 1980, overissuances accounted for about $10.20.

About 46 cents of that $10.20, or about 4.6 percent, had been recovered. Overall however as program outlays increased overissuances increased and reached an estimated $887 million that year. We are talking here about 1980.

A few days ago we reported to you, Mr. Chairman, on the cost of the food stamp benefits to households containing a striker, and on fluctuations in striker participation in the program.

From historical Agriculture and Labor data, we estimated for fiscal year 1980 households containing a striker received about $37 million in food coupons.

About 24,000 strikers participated in the program monthly. There did not seem to be any consistent relationship between total number of strikers and number of strikers participating in the food stamp program.

This may be because the amount and immediacy of strike benefits vary greatly among unions and because many strikers may be disqualified because of the program's resource eligibility criteria.

A recent legislative proposal by the administration would require that food stamp benefits be based on income received in a prior period. This is known as retrospective accounting.

Since most strikes last less than a month, Agriculture officials told us that the effect of such a change would be that in many cases strikers would not be eligible for food stamps.

Now, a report by the investigative staff of the Senate Appropriations Committee released earlier this year recommended closing several loopholes in food stamp regulations and improving the verification of eligibility information.

I understand that more specific details of the report will be presented to you today in separate testimony; however, we will be considering the issues raised in the report in an ongoing review in which we are seeking ways to bring about improvement in program eligibility and benefit determination procedures.

Now, the President's budget proposals advanced a number of measures affecting the food stamp program. In addition to eliminating overlapping school lunch benefits-discussed earlier in my statement-the proposals included such changes as the retrospective accounting of income, proration of first-month benefits, establishment of gross income limits, and improving overall program management.

On October 9, 1978, we testified that retrospective accounting was preferable because it uses actual rather than estimated income for making benefit determinations.

Also we have consistently and strongly advocated improved program management and simplified eligibility procedures. Although we have not previously addressed the use of gross income limits and benefit proration in our work, their logical effect would be to reduce program costs.

Turning now briefly to our ongoing review of the workfare demonstration program, this as you know is the program where food stamp recipients will be required to work on public service jobs for the value of their stamp benefits.

We reviewed the operating results of the seven demonstration sites, problems in measuring work benefits and costs, and the need for legislative and administrative changes that will help provide a more effective and efficient workfare operation.

We have in mind such changes as eliminating some of the currently allowed exemptions, eliminating unnecessary waiting periods, strengthening program sanctions, and improving administration at the Federal and local levels.

An extended phase of the workfare demonstration involving 14 project sites is now ongoing and scheduled to end by September 30, 1981.

Just turning to one more point on the workfare program, on page 14 of my statement, Agriculture and Labor did not include in