Verification of earned income

Some States have computerized earnings clearance systems which provide reviewers with information for verifying whether recipients had any earned income. Of States we reviewed, California and Indiana had earnings clearance systems. Reviewers in those States used the system for 24 of the 25 cases that we examined in which a recipient denied having earned income. During the period from which we selected QC cases for our review, New York did not have an earnings clearance system. 1/ Reviewers in New York used a variety of other sources, including unemployment insurance information, the State tax department, and SSA, to verify whether recipients had income.

Other procedural differences

We found that State procedures for verifying whether recipients had social security numbers (SSNs) varied as did their practices concerning home visits and collateral contacts, as well as dropping sample cases from the QC review process.

SSNS

Federal QC policy directs reviewers to verify that the local agency has an SSN on file for each person in the AFDC family or that an application for an SSN has been made. A recipient's failure to have an SSN or to have applied for one will result in the recipient's ineligibility.

In Maryland, State QC reviewers consider a case to be in error if a copy of the recipient's SSN or SSN application is not in the case file. In the January-June 1978 review period, Maryland reviewers found SSN-related errors in 54 of the 1,200 cases in the sample.

QC reviewers in Maine also check a recipient's case file for the SSN. However, they also verify that the recipient has an SSN during the home visit. If the recipient does not have an SSN, the reviewer determines if the recipient has applied for one. Only if the recipient has not applied is the

/New York implemented an earnings clearance system in January 1979, which was after the January-June 1978 QC review period from which we selected cases for our review.

case considered to be in error.

Maine had only 3 SSN-related

errors of the 603 cases in the January-June 1978 review period.

Home visits and collateral contacts

it

The States' approaches toward visits to recipients homes during the QC reviews also varied. In California, was QC policy to schedule visits to recipients' homes in advance. Indiana prefers that home visits to recipients be unannounced. California QC reviewers are required to obtain a signed consent form from the recipient for each collateral contact they wish to make (excluding those involving public records). In contrast, New York reviewers use a general recipient authorization statement that is part of the original AFDC application form, rather than obtaining a consent form for each collateral contact.

Policy on dropping sample cases

Under Federal QC guidelines, States need not complete reviews of selected cases and can drop them from the QC sample. Two circumstances where cases can be dropped are when the recipient is unwilling to give information and when the recipient cannot be located. The following are examples of cases that were dropped even though they were likely to be in error:

--In a case dropped by the State, the Federal QC reviewer
made four visits to the home of the recipient. The
recipient was at home, but was unwilling to talk to
the reviewer. The case was dropped from the sample
even though (1) there were indications of unreported
employment, (2) the recipient had three SSNs, and
(3) reports that the client was married and that her
husband was living at home conflicted with the recip-
ient's claim of being single.

--A State reviewer dropped a case because he was unable
to locate the recipient even though the State's earn-
ings clearance system indicated unreported earnings
for the recipient. The Federal reviewer found the
recipient and made a home visit. However, the case
was dropped at the Federal level when she refused
to sign an authorization form allowing verification
of earnings.

The following table shows the number of cases State QC offices dropped from samples in the January-June 1978 sample period because the client refused to cooperate with QC or because reviewers could not locate recipients.

The procedures followed by the QC reviewers before dropping cases from the QC sample varied among the States in our review. For example, California seldom used third-party collateral sources for locating recipients in the "unable to locate" category. We did not review any dropped cases in Indiana, but State officials told us that reviewers used a variety of third-party collateral sources, including the post office, the school attended by the dependent children, and the recipient's landlord or neighbors.

Dropping cases when the reviewer is unable to contact the recipient or the recipient is unwilling to cooperate may result in valuable management information on incorrect payments being lost because reviewers may be dropping error-prone Recipients who cannot be contacted may have moved or may be on extended visits, both of which could affect their payment status. Recipients who are unwilling to cooperate with reviewers may be withholding information for computing the correct grant. If the incidence of error in the dropped cases is higher than the error incidence in completed cases, then overall error information would be understated.

Because of problems we identified, HHS should study dropped cases to ascertain how much error information is being lost. If this information could materially affect errors, HHS should consider tightening up QC policies that allow cases to be dropped from QC samples for these two

reasons.

UNIFORM FEDERAL QC REVIEW

PROCEDURES NEEDED

HHS regional reviewers examine a subsample of State QC cases to assure that State reviewers are accurately identifying and reporting the correctness of AFDC payments. There are inconsistencies, however, between HHS regional offices in how they carry out these re-reviews. In the absence of clear guidelines defining what is appropriate, the regions have developed differing QC methodologies. Differences among regions included the (1) extent to which reviewers contacted the recipients and (2) degree of additional verification they perform.

One difference between HHS regional QC re-reviews is the proportion of contacts that the regions make with recipients. Because Federal reviewers have been given only broad guidance on when to contact a recipient, the policies of the regional offices vary. HHS reviewers contact recipients to obtain additional information or leads that would be useful in determining the correctness of cases. Among the five regions in our review, the extent to which Federal reviewers contacted recipients ranged from 0 to 30 percent for the January-June 1978 review period. A 1978 HHS study on OFA by the SSA Division of Management Services reported a nationwide range from 4 to 60 percent in regional contacts with recipients. Some regions indicated that they considered a home visit with a recipient essential, while others considered it only as a last resort.

In addition, regions performed varying amounts of additional verification when making their re-reviews. The 1978 HHS report on OFA expressed similar concern about variations among regions regarding what constitutes adequate verification of different elements of eligibility.

There are differences among the regional QC practices and differences in the adequacy of State QC reviews. We believe that HHS regions should have consistent standards for making the QC re-reviews, particularly because of the variations among State QC procedures discussed earlier.

STANDARDIZATION EFFORTS IN HHS

QC is a measurement tool for improving AFDC program administration. As such, it has been used for tracking a State's ability to pay eligible persons correct amounts of assistance over a period of time. It has also been used for

comparing one State's performance with another to identify effective error reduction techniques. A uniform QC review procedure would help make sure that error rates are comparable over time and between States.

In December 1978, HHS began revising the QC manual used by State and HHS reviewers to increase the uniformity of QC reviews. The new manual, which provides specific prioritized primary and secondary verification sources that are to be used for each item being validated, instructs QC reviewers to perform sufficient verification to confirm or disprove the correctness of the case.

For example, if the basis for eligibility is that the father does not reside with the recipient, and the father can be located, the revised manual requires the reviewer to obtain two of four recommended primary sources of verification of absence, such as the father's driver's license or rent receipts. If it is not possible to get the primary sources, the manual requires that two additional pieces of evidence be gathered from a prioritized list of eight secondary sources. This list includes such things as post office records and telephone directories. If the father cannot be located, the manual sets out similar procedures with a slightly different list of verification sources.

Although some differences will continue to exist in State QC practices because of differences in State programs, the revised QC manual requires that reviewers document why they cannot obtain primary verification. The revised manual has been reviewed by HHS regional QC offices and will be reviewed by the States before it is finalized. HHS anticipates that the revised manual will first be used by the States for the October 1980-March 1981 review period.

HHS SHOULD IMPROVE ITS SYSTEM
FOR EVALUATING AND MONITORING
STATE AND FEDERAL QC PERFORMANCE

In addition to the QC case re-review process, the HHS regions are required to monitor State QC operations through procedural reviews. An improved system for monitoring and evaluating the States is needed to insure the integrity and validity of the AFDC-QC system. HHS currently requires its regional offices to make only limited annual assessments of State QC systems. In addition, HHS has no systematic approach for monitoring its regions' QC efforts. HHS' ability to insure consistent QC reviews.

This reduces