match. The data exchange agreement would contain a provision to require adherence to Section 1603 of the IRS code. Section 1603 outlines strims nt requirements restricting access to wage and benefit records. Additionally, the proposed rulemaking would restrict state use of wage and benefit information solely to eligibility and benefit determinations.

Mr. WHITTEN. You state that states will receive an increased rate of federal cost-sharing for installation of a new computer system to encourage computer matching and other computer-related activities to reduce errors. What is the total estimated cost of this increased rate of federal cost-sharing for 1980, 1981 and 1982?

Mr. DICKEY. It is estimated that the federal share-at 75 percent-for the planning, design, development, and installation of new State systems meeting regulatory standards will approach $29 to $30 million for fiscal years 1981 and 1982. Regulations for this increased rate were not effective during fiscal year 1980.

FOOD STAMP ERROR RATES

Mr. WHITTEN. How are error rates determined in the Food Stamp Program?

Mr. DICKEY. The Food Stamp Act of 1977 and Food Stamp Program regulations published in August 1979 require that each state establish a food stamp quality control system. This quality control system is based upon reviewing a statistically selected sample of households receiving food stamp benefits and another valid sample of those households which have been denied food stamps or have had their benefits terminated. Cases selected in this sample are reviewed following established food stamp certification policy. States review each case selected in the latter sample to determine the validity of the action taken to deny or terminate benefits. The State then reports the percentage of cases with invalid decisions in the sample. The percentage of invalid decisions is synonymous with the negative case error rate reported by quality control.

States review each household in the sample of households receiving benefits to determine if the household is eligible and, if so, whether it is receiving the correct amount of food stamps. The state then reports the percentage of ineligible households, eligible households which received too many food stamps, and eligible households which received too few food stamps that it found in the sample. The state also reports the dollar loss associated with each type of error case identified above. From this information error rates are calculated for the percentage of ineligible cases, overissued cases, and underissued cases. The sum of these rates is the food stamp case error rate. The rates at which food stamps are misused are also calculated and represented as the underissuance, overissuance, and ineligible dollar error rates. The sum of these rates is referred to as the cumulative allotment error rate or payment error rate.

By virtue of using valid sampling techniques and adequate sample sizes, the quality control statistics are used to estimate the error rate in each state's caseload and for the nation. States report on the results of quality control reviews twice each fiscal year for the time periods October-March and April-September.

Mr. WHITTEN. Is there a uniform definition of error rates which applies to all states?

Mr. DICKEY. Yes. The food stamp quality control regulations establish definitions which are used by all states to determine the correctness of each case reviewed. The only inconsistency between states' error rates would result from differences in states' operational guidelines. These are approved by FNS and may contain deviations from standard program procedures to accommodate states' individual circumstances. To ensure consistent application of quality control error definitions, FNS reviews states' quality control operations periodically.

Mr. WHITTEN. On page 170 of the notes, you state that the 1980 amendments established an error rate sanction system whereby states with high error rates, which failed to reduce them below prescribed targets, will be liable for fiscal sanction. On page 168 of the notes, you stated that under the Act, states with high error rates will be liable to the Federal Government for the cost of their errors above a specified target, while states with low error rates will be rewarded with increased federal cost-sharing of their administrative expenses. Do these two statements, taken together, mean that if state agencies don't need federal assistance, then they will get more of it, and if they do need federal assistance, then they get less of it?

Mr. DICKEY. No. These two statements mean that if a state is successful in an effort to reduce food stamp error rates, the Department will provide the state with additional financial assistance to further decrease error rates or maintain already low rates. If a state does not reduce a relatively high error rate, the Department will not bear the full burden of the state's continuing high dollar loss alone. The Department intends to recognize and assist any state's efforts to reduce errors, and has developed regulations that will ensure that no state which is in the process of reducing its error rate as required will be interrupted by sanctions and removal of the necessary resources.

PHOTO IDENTIFICATION CARDS

Mr. WHITTEN. You state that in those areas in which the Secre tary, in consultation with the Inspector General, finds it useful. photo identification cards will be required. Has this been done yet? Do you anticipate that these identification cards will be required anywhere during 1981?

Mr. DICKEY. A proposed rule on Food Stamp Program photo identification requirements was published on January 23, 1981. The final rule should be published about mid-1981.

The Department has proposed that areas subject to the photo identification requirement issue all photo ID cards, and begin requiring presentation of photo ID cards at issuance points as a precondition of issuing coupons to households, no later than nine months after publication of the final rule. The nine-month timeframe was proposed to allow state agencies sufficient time to purchase and install photo ID equipment and to issue photo ID cards to affected recipients.

Therefore, in light of both the projected publication date of the final photo identification rule and the proposed nine-month imple

mentation timeframe, photo identification cards will in all likelihood be required in early 1982.

INCENTIVES TO REDUCE HIGH ERROR RATES

Mr. WHITTEN. You state that you have been working on efforts to lower state error rates by mobilizing key high participation states which have high error rates to take aggressive action to reduce those errors. How is this mobilization achieved? What incentives are available to get states to participate? Given the success of your current efforts in this area, which geographical areas will be targeted next?

Mr. DICKEY. This mobilization is achieved through a combination of efforts, depending on the severity of the problem and the degree of progress and cooperation from states. The reduction of error rates in key high participation states is a high priority. The preferred approach is to achieve success through early detection and active involvement in providing technical assistance for the development of targeted corrective action to reduce error rates to acceptable levels. Following implementation of approved corrective action, the impact of such action is closely monitored by both the states and the Food and Nutrition Service. Successes are usually achieved through a spirit of cooperation and good faith efforts by states.

The primary mechanism which is available to motivate states to voluntarily participate in reducing error rates is incentive funding. Under this provision of Public Law 96-249, the standard 50 percent federal share of administrative funding could be increased up to a maximum of 65 percent for achieving and maintaining an error rate below a specific level or for a predetermined acceptable rate of error reduction. Beginning with the rates reported for the October 1980 through March 1981 quality control review period, states which achieve dollar error rates below 5 percent shall receive 65 percent funding. For rates between 5 and 8 percent, or the national standard-whichever is lower-states shall receive 60 percent funding. Those states which reduce their error rates by 25 percent or more in one year's time shall receive 55 percent funding. The national standard for the October 1980 through March 1981 period has been established at 12.6 percent. Thus, states would need to achieve a rate of between 5 and 8 percent for that period to be eligible for 60 percent funding.

If these incentives and attempts at full cooperation do not succeed in reducing error rates to acceptable levels, we can impose two types of fiscal sanctions against states. First, we will require a state to remit payment for the dollar equivalent of the error rate in excess of a predetermined goal. For example, if a state's goal for error rate reduction is set at 10 percent for a six month period and its payments error rate is determined to be 12 percent, the state would be liable for 2 percent of its total issuance during that six month period if there was not good cause for exceeding their goal. In addition to this mandatory repayment of program losses, we may also suspend and/or cancel a portion of the federal share of administrative costs for the certification of households.

The key states currently being targeted for special mobilization effects are New York, Illinois, Pennsylvania and Michigan. These

states ranked among the top for both total dollars issued and error rates above the national standard of 12.6 percent for the period October 1979 through March 1980. Although the District of Columbia ranks only 39th highest for total dollars issued, we are also targeting them for special efforts as they reported an error rate of 21.6 percent for the same period. They are tied with New York for the highest error rate in the Nation. Similarly, Nebraska, Rhode Island, Maryland, New Mexico, Iowa and the Virgin Islands rank among the top seven highest error rates, although they rank among the lowest in total dollars issued.

ROUNDING PROCEDURES

Mr. WHITTEN. You state that new rounding procedures in food stamp income calculations will save $50 million in fiscal year 1981. What a comment it is on the magnitude of this program, that a change in rounding procedures can account for a $50 million savings. Please describe this change in some detail.

Mr. DICKEY. Under the old rounding procedures, all figures included in the computation of net income, except for individual shelter expenses, were rounded down to the closest dollar. For example, income of $206.11 was handled as $206.00. The same was true of income of $206.97. This system, while administratively simple, was costly and was not always the system used in the Aid to Families with Dependent Children-AFDC-Program. It was costly because, through a number of calculations, a household's net income could be underestimated by several dollars. It was contradictory to AFDC because many state agencies used other rounding techniques; for example, rounding up, rounding to the nearest dollar, not rounding at all.

The new rounding procedure offers to each state agency two options: rounding as it does in AFDC, and rounding down from $.49 and up from $.50. Most state agencies round under the latter procedure, which more accurately estimates the net incomes of all households. By lowering the benefits of many households, there is a sayings of about $50 million per year.

PARTICIPATION RATES

Mr. WHITTEN. You state that in fiscal year 1981, an average of 21,960,000 persons are expected to participate in the program monthly and receive bonus coupons worth an average of $39.85 each month. What are your comparable estimates for 1982, both under existing legislation and the proposed revisions?

Mr. DICKEY. Under existing legislation, an average of 21,440,000 would participate in the program in fiscal year 1982 and benefits would average $46.37 per person per month. Enactment of the proposed changes in the Food Stamp Program would reduce participation levels to an average of 18,660,000 persons per month and monthly benefits to $40.92 per person.

GRANTS FOR STATE ADMINISTRATION

Mr. WHITTEN. Please submit for the record a table which will indicate, by state, the distribution of matching grants for state administration for 1980 actuals and 1981 estimates, indicating

which percentage rate of cost sharing this represents for each state.

Mr. DICKEY. We will submit a table for the record detailing the distribution of matching grants for state administration for fiscal years 1980 and 1981. Final regulations increasing funding for investigations, prosecutions and fraud hearings were issued on December 30, 1980. To date, 21 states are receiving 75 percent reimbursement for these costs, but because the total dollar amount is very small it has very little impact on the total percentage of state costs matched by the Federal Government.

The final regulations implementing enhanced funding for states which reduce their error rates below specified levels were issued on January 23, 1981 and the proposed regulations for 75 percent funding for the costs of developing and installing computerized systems will be published shortly. Both of these regulations will impact the matching for state administrative costs retroactive to October 1. It is, however, too early to predict on a state-by-state basis how much these incentive payments will increase overall costs above the basic 50 percent.

[The information follows:]

FOOD STAMP PROGRAM-MATCHING GRANTS FOR STATE ADMINISTRATION, FISCAL YEAR 1980 AND