--mothers on behalf of an unborn child if there are no other children in the family; and

-- families in need because of an unemployed parent.

Two States in our review did not provide assistance to some of the groups. Indiana did not provide assistance to any of the three groups, and Maine did not provide aid on behalf of an unborn child or to families with an unemployed parent. QC directors from both States indicated that including one or more of these groups would increase their error rates because particular eligibility conditions require extensive monitoring, more recipients would be involved, and grant computations would be more complex, among other things, making these options difficult to administer without having errors.

Differences in QC review procedures

The QC reviewers in different States followed different procedures in determining the correctness of cases they reviewed. (See ch. 3.) QC reviewers in some States relied heavily on the statements of recipients for verifying the accuracy of information that was the basis for the grant, while those in other States made extensive investigations and collateral checks with neighbors, employers, landlords, and financial institutions.

While HHS has been conducting a project to standardize the QC reviews by State and HHS reviewers, that project has not determined to what extent differences in the review procedures are reflected in State error rates. Without knowing these effects, the proportion of a State's error rate that is due to administrative problems and the proportion due to QC review differences cannot be determined.

APPROPRIATIONS COMMITTEES' ACTION
NEEDED TO STOP FISCAL SANCTIONS

Current HHS efforts to implement a sanctioning system are based in part on a directive contained in a congressional conference report. During deliberations on a fiscal year 1979 supplemental appropriations bill (Public Law 96-38), the House-Senate conferees decided more ambitious error reduction efforts were needed. The conference committee directed the Secretary of HHS to issue new regulations requiring all States to reduce their AFDC payment error rate for eligible

but overpaid cases and ineligible cases to 4 percent by September 30, 1982, in three annual increments beginning with fiscal year 1980. Failure to meet the error rate target each year is to result in the loss of Federal matching funds associated with erroneous payment expenditures in excess of the target.

CONCLUSIONS

Sanctions place the Federal and State governments in an adversary role at a time when cooperation is needed for reducing errors. Also, because identifying too many errors can cost States Federal funding, sanctions can be a disincentive to error identification efforts. This is contrary to the intent of QC to improve the administration of AFDC, because better error identification would make more data available to managers for developing corrective actions to reduce

errors.

QC error rates are not sufficiently comparable to serve as the basis for sanctions. The statistical precision of the error rate estimates varies from State to State with some States being more likely to have error rate fluctuations due to chance. QC error rates can also be affected by both the QC review procedure differences and State program differences, which tend to reduce the value of QC as a comparative measure of administrative quality in State AFDC programs.

For these reasons, we believe efforts to sanction high error States based on QC error rates should be discontinued. Instead of sanctioning States, the Federal Government should provide more assistance in error reduction efforts. (See ch. 5.)

Because the current HHS effort to sanction States is based in part on a conference committee directive, the Appropriations Committees should play a role in discontinuing this effort.

RECOMMENDATION TO THE

APPROPRIATIONS COMMITTEES

We recommend that the House and Senate Appropriations Committees retract the conference committee directive for Federal fiscal sanctions against the States based on the AFDC QC error rates.

AGENCY COMMENTS AND OUR EVALUATION

While five 1/ of the six States we visited were in favor of such action, HHS officials did not agree that sanctions should be retracted. They said that some form of sanction was needed to assure continued management attention to AFDC quality. They favored modifying the present stringent error reduction requirements.

We agree that continued management attention to AFDC quality is essential and that the threat of sanctions in the past had a beneficial effect on the States' management of the program. The issue, as we see it, is whether the threat of sanctions is as valid a method of obtaining program improvement today as it was in the past. For the reasons stated earlier in this chapter, we believe not.

We believe that the States now have sufficient motivation to improve AFDC quality in that they pay about half the cost of AFDC benefits. Furthermore, as we indicated earlier, in recent years there has been considerably more public awareness of the cost of the welfare programs, including the high level of erroneous payments, and it appears that this awareness will continue. This awareness, and its impact on State legislatures who are concerned with keeping the cost of State programs as low as possible, would seem to add a further incentive to AFDC program officials to carefully manage their activities.

We also believe, as discussed earlier, that sanctions tend to put the States and the Federal Government in an adversary role at a time when cooperation would seem more essential than ever because of the high costs and limited resources for such Federal-State programs.

1/officials from California and Indiana contacted during our review said they favored sanctions. (See p. 8.) However, in commenting on our draft report, California supported our recommendation to retract the sanctions directive. (See app. I.)

CHAPTER 3

HHS SHOULD DEVELOP

UNIFORM QC REVIEW PROCEDURES

To be an effective measurement tool and provide data that are comparable, QC reviews should follow uniform basic review procedures. We found several inconsistent QC proce

dures among States and HHS regions.

These differences existed because HHS had neither developed uniform QC review procedures nor implemented a system for making comprehensive evaluations of State and regional QC procedures. In our opinion, some States that limit verification may not be obtaining all the management information on incorrect payments available through the QC review process. However, other States may be performing unnecessarily extensive reviews.

HHS is attempting to increase standardization in the QC system through revisions of the QC manual. The purpose of the revision is to establish minimum verification procedures for all States and HHS regions to follow. In addition, HHS is conducting reviews and evaluations of regional QC programs on a pilot-test basis as an initial step toward developing an improved monitoring system.

UNIFORM PROCEDURES ARE Needed

TO ELIMINATE VARIATIONS

IN STATE QC PRACTICES

Federal guidance allows a great deal of flexibility in the way State QC cases are reviewed. The QC manual provides that reviewers will evaluate the eligibility and benefit amount factors on a case-by-case basis and determine how much and what kind of verification is needed. As a result, each State has been free to select the particular steps its QC reviewers follow to validate the correctness of the factors on which the AFDC eligibility and amount of payment are based.

In the six States in our review, we found several indications of how procedures for conducting QC reviews vary on important matters, such as the verification techniques used to determine the accuracy of AFDC payments and the conditions under which State QC reviewers drop certain cases from their review. Some procedures seemed unnecessarily complex, while others seemed to be too limited to discover all errors.

State verification procedures differ

The six States used a variety of practices to verify the accuracy of AFDC eligibility and payment amount factors. We compared QC review cases from each State to identify the different practices. The differences discussed below were identified in our comparison of California, Indiana, and New York QC case review practices. We believe the types of verification differences discussed below indicate that some States may be making unnecessarily extensive reviews, and others may be making limited reviews that are not identifying all errors.

Continued absence of a parent

To be eligible for AFDC, a child must be deprived of parental support or care because of the continued absence of a parent from the home, death of a parent, physical or mental incapacity of a parent, or in some States unemployment of a parent. Most AFDC cases are eligible due to the continued absence of a parent.

In California, of the 16 cases that we examined, 13 had an absent parent. The recipient was the source of verification in all 13 cases. Reviewers accepted recipients' verbal statements that the parent did not reside in the home as verification of the absence in 8 of the 13 cases. In the other five cases the recipient was asked to provide additional evidence.

In contrast to California, both Indiana and New York obtained additional collateral verification of the parent's absence in all 24 continued absence cases we reviewed. In Indiana, the number of collateral sources in addition to the recipient's statement for each case ranged from 1 to 3, and in New York, the number ranged from 2 to 11. Reviewers from these States made contacts with such collateral sources as the State division of motor vehicles, telephone companies, employment offices, and other third-party sources.

Recipient denial of bank account

Variations existed in State practices for validating recipients' claims that they had no bank accounts. In California, QC reviewers accepted the clients' statements that they had no bank accounts and made no collateral bank checks in the 13 cases we examined. In contrast, Indiana and New York reviewers made collateral bank checks in the 24 similar cases we examined.