Mr. FLOOD. What steps will be taken to carry out some of the recommendations? Mr. CARDWELL. We have added them up and find that on some 47 of the 90, we had already come to at least a similar conclusion and have some action already in progress. We will submit to this committee and to the substantive committees of interest an implementation plan for the report. I would mention, however, that a number of their recommendations are very far reaching in their criticism of the basic law. One of the toughest choices that we in the executive branch will have to make, and one of the toughest choices the Congress will have to make, is whether to entertain those particular recommendations. As a general proposition, they have concluded, as we did, that the program may well be too complex to ever fulfill the original idea that justified central administration, but that is something we will all have to struggle with. Mr. FLOOD. Over the past 9 months or so, there has been a continuing series of reports concerning overpayments. You recall last September you sent us a letter saying many of these reports you felt presented a distorted picture. Tell us briefly something about this situation of overpayments as you see it now. ERROR RATES Mr. CARDWELL. We sent an updated report to this committee and other committees of Congress just this last January. To summarize, that report showed that the so-called SSI error rates were not improving significantly when measured against the first 6 months, which we reported to the committee last year. In fact, the error rate turned out to be essentially the same, about 24.4 percent. Mr. FLOOD. Suppose you do this. Put a copy of that January report in the record. [The information follows:] DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE, SOCIAL SECURITY ADMINISTRATION, Hon. AL ULLMAN, Chairman, Committee on Ways and Means, DEAR MR. ULLMAN: In April, we advised the House Ways and Means Committee and the Senate Finance Committee about preliminary supplemental security income (SSI) quality assurance (QA) findings for the period July-December 1974, the second 6 months of operations under SSI. Under a contractual arrangement with those States for whom we administer State supplements, the results for that period provide a performance base for measuring the Federal Government's fiscal liability to those States due to inaccurate payments of State funds. The data presented last April were based on preliminary samples of cases for the base period. Those preliminary samples have been completed. and the purpose of this letter is to present the final results for that period as well as the final results of the QA reviews for the subsequent 6-month period, January through June 1975. The results are as follows: The above shows that while overpayments were reduced during the second period, payments to ineligibles and underpayments increased. We should note that the second period also shows a higher incidence of errors among new federally administered cases than the prior period, offset by a reduction in the rate of incorrect payments among cases taken over from the States. The net result of these pluses and minuses was a marginal decrease in the overall rate of payment inaccuracy. As previously indicated to the committee during testimony and other exchanges concerning SSI payment accuracy, we believe that a number of actions that have been taken will eventually result in improved payment efficiency, as measured by the QA system. As the above data indicate, those measures had not produced any significant improvement during the first 6 months of the 1975 calendar year. However, with about one-half of the sample complete, sampling for the subsequent 6-month period (July-December 1975) suggests that significant improvement will be shown when the final QA results are tabulated for the period. One of the reasons we are expecting improvements to show in the July to December period is because our new system for automated interface between the social security record and the SSI record became operational during that period. According to the QA results for the first 18 months of operations, this was one of the most significant sources of overpayments and payments to ineligibles. We will, of course, furnish these later results as soon as the review is complete, expected by mid-April 1976. HOW THE QA SYSTEM WORKS Perhaps it would be useful to describe the QA system and give some idea about the range of its coverage. As the Congress is aware, the system is a mirror image of the arrangement developed several years ago for application to the Federal/State public assistance program. The QA system was designed for two purposes: 1. To provide base data for determining any fiscal liability to the States for SSA-caused errors during the administration of State supplements. 2. To measure, in as much depth as possible, the performance of the program against the literal requirements of the statute. It should be emphasized that the QA system attempts to identify and measure overpayments and underpayments, some of which, particularly in the posteligibility category, are beyond the agency's control. In this regard, the QA system was established around the assumption that the most efficient administration of the program would produce a pavment accuracy level of no better than 92 percent of the cases. This produced the 8-percent tolerance level which, if met, would exonerate SSA from any fiscal liability to the States. This, of course, is the same level set for State administration of AFDC. To determine the level of payment performance, the QA system calls for a review of approximately 27,000 cases every 6 months. Each case is reviewed for two separate levels of accuracy: (1) total payment to the recipient; and (2) the State's share of such total payment. The first level identifies SSA accuracy and performance, and the case is considered to be without error if the recipient received the proper amount. The rates given on the first page of this letter are based on the total payment level—(1), above. For purposes of determining the Federal fiscal liability to a State, however, the State share must also be totally correct for the case to be considered error free. Following is an example of how this works. The recipient is due $200, $100 from SSA and $100 from the State. He receives a check for $200, but the SSA portion was $75 while the State portion was $125. For overall SSI accuracy purposes, this case is correct; however, for Federal fiscal liability purposes, it is in error. This type of case does not occur frequently. It is used, however, to illustrate the two different rates, SSI and Federal fiscal liability, that are provided by the QA system. Under our arrangement with the States, the base period-the second 6 months of the program-represents a starting point from which SSA performance will be measured against the 8-percent tolerance. During the base period, SSA is liable only for errors identified on a case-by-case bases. Liability for the first 6 months of operations is to be determined by separate audits subject to negotiations with the States. Under the QA system, Federal fiscal liability begins to run in the third 6-month period of operations, with three equal 6-month step-down periods until the final 8-percent tolerance level is reached. By that time, beginning with the January to June 1976 period, the permanent tolerance level of 8 percent will prevail, and the Federal fiscal liability will accrue for any errors over and above that rate. DOLLAR VALUE OF PAYMENTS There are three different measurements of dollar overpayments used by SSA, and we want to first explain what they are: 1. Quality Assurance Extrapolation Amounts This figure is established based on sample data projected to the dollars in the universe of all SSI cases. "Errors" are determined through both a lengthy and comprehensive interview with the individual at his home, as well as total verification (to the extent possible) of negative allegations. From the sample findings, the dollars in error in individual cases within the sample are divided by the total dollars in the total sample to create a payment error rate. This payment error rate is multiplied by the total SSI expenditures during the 6-month sample period to estimate total dollars in error. In many ways, the QA extrapolation highlights the difference between a perfect payment program and what occurs during actual administration. It does not apply a tolerance such as the AFDS 8-percent rate. Estimated dollars in error are related directly to their causes so that corrective action may be taken. In short, the magnitude of specific types of errors are tabulated so that the largest problems are identified and worked on in priority order. 2. Federal Fiscal Liability (FFL) Overpayment Amounts This figure represents the dollars SSA owes the States under the terms of our contracts with the States. As already noted, a separate measurement is made of errors in the payment of State supplements. Within the sample, dollars in error are extrapolated to the universe of State supplement dollars paid out by SSA, and a predetermined tolerance is applied to that figure. FFL overpayments, therefore, are the projected State dollars paid out in error reduced by the tolerance. 3. Automated System Overpayment Tally This figure represents a running record of dollar errors identified during dayto-day case processing by SSA. Such amounts are recorded and controlled by the automated payment system, and our district office personnel undertake recovery action in accordance with established criteria. SUMMARY OF DOLLAR VALUE OF OVERPAYMENTS Following are the dollar value of overpayments for the QA system compared to the estimated fiscal liability to the States under the QA system: As of December 31, 1975, the automated tally carries a cumulative amount of $547 million. This shows cumulative errors recorded for the full 2-year period from January 1974 through December 1975, of which $419 million remains unsettled and/or uncollected. As previously reported to the Congress, recovery of this amount is expected to be limited, based on AFDC experience. IMPROVEMENTS AND CURRENT PROBLEMS Two significant corrective actions taken since these QA results, but not reflected in the results, include: 1. Automated systems were put in operation on June 1, 1975, which will correct regular social security benefits on to the SSI records. Analysis indicates that this action should reduce the number of errors resulting from incorrect social security benefits being used in the SSI benefit computation. It should be noted that this factor was a major cause of error when the aged, blind, and disabled program was administered by the States. 2. Steps have been taken to improve the redetermination process through more training of personnel and improved interviewing. The major problems that still exist which we must correct are: 1. We know that federally taken cases-many of which were taken during the early months of the program-are showing a higher than expected error rate due in part to unreported changes. We have begun to redetermine these cases and are attempting to step up that process, especially with regard to claims taken at the beginning of the program. We know that early start-up problems are involved in the erroneous payments appearing in these cases. 2. There is a significant amount of error in unreported income and resource changes. We need to make beneficiaries more aware of their responsibilities, and, if possible, we need to see people more often than once a year, at least in high risk areas. We are taking steps to try to do this. 3. Non-SSA pension income continues to be a major payment deficiency. Work on an automated exchange with other agencies, primarily the VA, needs to move ahead as quickly as possible. I would note, however, that this is likely to still require a good deal of further time and effort, perhaps many more months. We believe that progress toward improved payment accuracy has been made and will continue to be made. We will continue to keep you informed. A similar letter has been sent to the Honorable Russell Long, to the Committees on Appropriations, and others in Congress. Enclosed is information on a State-by-State basis and other details about the QA results. sincerely yours, Enclosures. JAMES B. CARDWELL, Commissioner of Social Security. For purposes of tables I-A and I-B, a case error is recorded only if that portion of the SSI payment attributable to the State supplement is incorrect. These rates will be used in determining Federal Fiscal Liability (FFL). Only converted recipients (those on the State welfare rolls in December 1973) are eligible for mandatory State supplemental payments while all recipients (new as well as converted) are eligible for optional State supplemental payments. Relatively few converted recipients resident in a State which pays an optional payment qualify for a mandatory payment because of the high level of the optional payment. Mandatory payments are complex to determine and thus such cases are much more error prone than the optional payment case. It is for these reasons that the table has been separated into an "A" and "B" table in order to differentiate between the two types of supplements. In addition, the number of recipients eligible for mandatory payments has been declining steadily and at an accelerated rate due to at least three factors: Termination of SSI eligibility (primarily because of death); rise in the SSI benefit level because of the cost-of-living increases; and increases in the optional payment level. Mandatory payment errors as a factor of overall FFL payment accuracy should play a decreasing part in future reporting periods, thus resulting in improved payment accuracy. |