Activity Related to the Suspense File
- A-03-96-31001 and A-03-96-31002 - 9/15/97
This management advisory report presents the results of work
performed to date by the Office of the Inspector General (OIG)
on Suspense File related activities and the efforts of the Deputy
Commissioner for Operations to reduce the size of the Suspense
File. It also provides OIG’s observations on the value of
certain procedures available for use in reinstating suspended wage
items and allegations made about the earnings system.
The OIG informed the Social Security Administration (SSA)
that it would conduct two reviews concerning Suspense File activities.
These reviews were entitled "Use of Internal Data to Post
Wage Items from the Suspense File" (A-03-96-31001) and "Review
of Computerized Methods for Identifying Postable Suspense Items" (A-03-96-31002).
During the testing phase of these reviews, SSA informed us it was
conducting pilot studies on ways to reduce the Suspense File. SSA’s
studies evolved from a task force project sponsored by the Deputy
Commissioner for Operations. Our reviews are being deferred pending
completion of SSA’s studies. We will review the results of
SSA’s pilot studies and progress toward implementing the results
of those studies later this month. From our preliminary work, however,
we have made some observations which you may wish to consider.
We determined that procedures utilized by the San Bernardino
District Office (SBDO) have value in reinstating suspended wage
items and that the pilot studies should experiment with SBDO procedures.
However, SSA pilot studies indicate that the operation could provide
a better internal control and quality assurance environment if
conducted and controlled in a single location.
Further, in conducting our preliminary work, we considered
the allegations that SSA failed to reinstate suspended wage items
when ownership of the earnings was apparent from its own records
and that many workers were disadvantaged, especially women and
minorities. This is because name and Social Security number (SSN)
matches are required for wages to be posted to an individual’s
Master Earnings File (MEF). We found that SSA does not routinely
cross match wage items against other information files, other than
the Numident file and a file of previously reinstated wages, to
identify ownership prior to discontinuing the reviews. We do not
believe, however, that the system discriminates against women and
minorities. The Numident validation system includes several processes
which increase the likelihood that wages for women and individuals
with compound surnames will be posted. Our review of the earnings
system also found that it has special validation rules for both
women and wage earners with compound surnames which increase the
chances that their wages will be posted.
BACKGROUND
Employers and self-employed individuals are required to submit
wage and self-employed data to SSA on an annual basis. This information
includes the amount of wages paid or self-employment income earned
and the amount of taxes withheld on wages and self-employment income;
the names and SSNs of wage earners; and the employers’ identification
numbers. All wage items and self-employed data are subject to SSA’s
overnight validation procedures. The purpose of the procedures
is to identify ownership of wage items using a 3-step process to
allow posting to an individual’s MEF record. The MEF contains
a detailed record of an individual’s earnings (Detailed Earnings
Record) and also a summary record of earnings for each individual
for each year worked (Summary Earnings Record). Through Calendar
Year (CY) 1995, there were 375.2 million MEF records containing
wages of $42.6 trillion.
The first step of validation compares the SSN and name to
the Numident file. The Numident file is SSA’s primary identification
record and contains information such as the first and last name
of the number holder, SSN, and other identifying data. Each name
can be processed against as many as 16 validation rules. The first
rule requires that the first seven positions of the last name contained
in the Numident file be identical to the first seven positions
of the last name contained on the wage report submitted by an employer;
about 90 percent of the data pass this rule. If the name does
not pass the first rule, the data is then processed against up
to 15 other rules which manipulate the name in an attempt to reach
a match. Good matches are posted and nonmatches are passed to the
second validation step. In the second validation step, the SSN
and name are run against prior year files that were reinstated
to determine if the same combination was previously reported and
reinstated. In the third step, the SSN is assumed to be wrong and
is run against a program which rearranges the SSN in 89 different
combinations seeking a match.
SSA’s Suspense File is the alternate repository for
annual and quarterly wage items which are processed through SSA’s
earnings system but fail to match SSA name and SSN records. SSA
places mismatched wages in an electronic Suspense File. The electronic
Suspense File offers orderly accounting, methodical and systematic
reference abilities, and a means to apply both electronic and manual
efforts toward movement of wage items from the Suspense File to
the MEF. The Suspense File contains over 203 million wage items
and represents over $217 billion in covered wages accumulated
from the inception of the program through February 1997.
Wage items that remain in the Suspense File are not available
for use in computing an individual’s benefit amount, unless
a number holder questions or disputes earnings when applying for
benefits. Thus, reinstatement of wages contained in the Suspense
File is essential. After other clerical, electronic, and correspondence
attempts to reinstate Suspense File wage items fail, they are reviewed
in Operation 30. At the Wilkes-Barre Data Operations Center (WBDOC),
key operators compare Suspense File items with corresponding Numident
records which are system-generated on a computer screen. The system
generates up to five name options for use in the comparisons. A
match identifies the wage item for posting.
Since March 1996, there has been much attention paid to the
subject of SSA’s earnings Suspense File by the Congress, SSA
management, and OIG. The genesis for the interest was a number
of allegations made by the Manager of SBDO that SSA had been remiss
in correcting workers’ earnings records. In response, the
Deputy Commissioner for Operations formed a task force in March
1996 to develop recommendations aimed at both preventing wage and
self-employment items from entering the Suspense File and reducing
the number already in suspense. Also, in July 1996, OIG initiated
two reviews related to the Suspense File. Our objectives were to
evaluate the procedures used by SBDO in posting open suspense items
to the MEF using internal records and to evaluate the effectiveness
of SSA’s computerized methods for identifying postable wage
items with incorrect names and/or SSNs.
Specifically, the SBDO Manager alleged that: SSA failed to
reinstate earnings from the Suspense File when ownership of the
earnings was apparent from its own records; and the failure to
reinstate earnings due to the requirement of matching name and
SSN was especially disadvantageous to women and minorities. Women
often marry without notifying SSA of their name changes and employers
often mishandle ethnic surnames of minority groups. SBDO based
these allegations on its study of the Suspense File and experience
in performing 117,000 wage corrections over a 15-month period.
The allegations surfaced in March 1996, and in April 1996 were
widely circulated in the form of a white paper prepared by SBDO.
SBDO had, for some time, been conducting a self-initiated
pilot test it called the "earnings repair study." Instead
of waiting for number holders to come into the office inquiring
about their earnings, SBDO researched the Suspense File looking
for employers with a high number of open wage items that SBDO determined
had high potential for matches and subsequent postings of earnings.
The research efforts were not always successful. When a candidate
was found, SBDO would then utilize various SSA records to match
enough data related to suspended wage items to make a reasonable
inference about the identity of the wage earners. When satisfied
that an identity had been determined, SBDO then posted the suspended
wages to the account holder’s earnings record. The SBDO Manager
had actively recommended that the procedure utilized in the study
be adopted by Headquarters as a nationwide methodology for correcting
workers’ earnings records. The SBDO procedures for correcting
earnings records were not in accordance with SSA’s current
policy, which requires reviewing the earnings record with the number
holder or his/her survivor to be sure that the earnings posted
are accurate and consistent with the individual’s work history.
This policy applies to pre-claims, claims, and post adjudicative
claims situations.
Our initial proposal was to conduct two reviews. Both reviews
were to utilize data extracts from SSA’s files to assist us
in our determination of whether wages contained in the Suspense
File could be posted to the MEF. The objective of the first review
was to examine the effectiveness and accuracy of procedures utilized
by SBDO and determine the proper point in the process for implementing
SBDO validation procedures. This review was to test the accuracy
of a sample of the corrections made by SBDO. The objective of the
second review was to identify additional profiles for incorporation
into the validation rules. This review was to analyze an extract
of the Suspense File to identify additional, acceptable profiles
for use as validation rules.
The testing phase for both reviews was not accomplished.
However, while waiting for SSA to extract the requested data, we
asked SBDO to perform a study for our review and understanding.
The employer chosen for the review was selected from a pre-screened
list of employers that SBDO had determined had a large number of
open wage items with a high potential for matches in its Suspense
File. This employer’s 1980 Suspense File contained 131 wage
items which were not posted because of incorrect names or SSNs.
The amount of the suspended wages was $1.8 million. We reviewed
the 131 wage items.
RESULTS OF REVIEW ACTIVITY
About the time the testing for our audits was to begin, we
were briefed by the Director of Earnings and Adjustments of the
Office of Central Records Operations (OCRO) on the intended directions
of the pilot studies. We believed the objectives of these studies
would produce similar results as those planned in our audit. As
a result, we deferred our review to allow SSA the opportunity to
implement its pilot studies. We will review the results and progress
in implementing the pilot studies later this month. Sufficient
work has been done, however, to provide some insight into SBDO
procedures and assertions and the accuracy of the SBDO Manager’s
allegations.
Task Force Results
In September 1996, the Office of Operations’ sponsored
task force report recommended action in eight aspects of Suspense
File management which should aid in reducing the size of the Suspense
File and materially reduce the number of wage items going into
suspense. The recommended actions utilized a variety of approaches
to the problem and would:
complete already established, scheduled Suspense File
reduction functions (such as Operation 30) on a tighter, more
regular time schedule and investigate and pilot other Suspense
File reduction methods and business changes;
improve systems software to make the Suspense File easier
to reference and assist searchers with their efforts to find
items in suspense;
lower the rate of error acceptability in the magnetic
media annual wage reporting process (currently, if 70 percent
of the items on a wage report are in error, the system accepts
the report for processing) and offer specific examples of the
most common causes of suspense in the instructions provided to
employers;
conduct outreach efforts with "problem" industries
(in 1990, three industries accounted for 46.5 percent of all
suspense items; agriculture, services, and bars/restaurants)
and States (over a 6-year period, nine States account for 70
percent of all suspense items; California alone accounts for
35 percent);
institute a graduated approach of informing employers
of suspense problems, offering help, and where not successful,
work with the Internal Revenue Service to facilitate invoking
penalties for repeat offenders;
support verification of name/SSN information with employers,
aggressively, systematically, and through modern methods and
techniques;
measure the acceptability of changing suspense software
to include mother’s maiden name as part of validation, eliminating
sex code as a validation consideration, and programming given
name derivatives as an expansion of current matching criteria;
and
develop, staff, and conduct a proactive, continuing SSA
designed suspense evaluation program.
SSA Pilot Studies
Two pilot studies deal with expanding Operation 30 procedures
and the SSN/name validation procedures. The first pilot study involving
Operation 30 is to consist of experimentation with additional procedures
in an attempt to post wages which failed to validate. The study
initially will use suspended wage items for Tax Year 1980. Operation
30, as now designed, is performed periodically by WBDOC staff.
It utilizes system-generated Numident records (up to five when
wage earner has multiple surnames) for suspended wage items. If
the name for a suspended wage item agrees with a name option, the
name is corrected to the matched name and posted to MEF. When there
is no agreement, the wage item remains in the Suspense File. The
pilot study will create a third category of action in addition
to "corrected" and "not corrected." The third
category will apply to wage items where some doubt exists about
whether there is a match. These items would not be acted on at
the time, but set aside for further review by skilled technicians
in OCRO.
As described to us by Office of Operations’ personnel,
some of the additional procedures will be similar to those used
by SBDO and will utilize SSA records other than the Numident file
to identify ownership of suspended wage items. If the efficacy
of these test procedures is proven, we would endorse their placement
in a central location rather than in SSA’s 1,300 field offices.
Placement in a single location with qualified technicians offers
a more desirable internal control and quality assurance environment.
The objective of the second pilot study, involving the SSN/name
validation process, is to identify additional profiles for incorporation
into the Numident validation rules. This pilot study includes testing
of expanded criteria for the Numident validation rules, which could
prevent more wage items from going into suspense. Presently, the
Numident validation rules require a matching relationship with
the surname, the exception being when the wage earner is identified
as a female. Then the rules will allow wages to be posted to the
MEF based on a match of the first name and the middle initial only.
The pilot study would consider applying the more liberal female
rules to males. Other tests would expand the matching criteria
to include given name derivatives, and incorporating the mother’s
and father’s surname.
As part of our review, we intended to test the accuracy of
the postings performed by SBDO during its "earnings repair
study" and to evaluate the merits of SBDO’s allegations.
We developed some preliminary observations in these two areas prior
to discontinuing the audits.
Our observations on the accuracy of SBDO postings are based
on a review of a demonstration performed by SBDO at our request.
Initially, we intended to obtain a data extract to review a random
sample of the approximately 117,000 wage items that SBDO posted
to MEF records during its "earnings repair study." The
purpose of the demonstration was to provide an understanding of
the procedures pending receipt of an extract of SBDO posted items.
However, we did not perform a random review because our audit was
suspended. Although the results of the demonstration cannot be
projected, we believe they provide insight into the value of SBDO
procedures.
SBDO selected an employer previously identified through its
research efforts from the Suspense File for CY 1980. This employer
had 131 open wage items which were not posted because of incorrect
names or SSNs. The amount of the suspended wages was $1.8 million.
SBDO determined that this particular suspense list had the potential
for a high percentage of matches and subsequent postings. The list
showed many unreadable surnames that contained ampersands and double
letters in combination, e.g., 0@GG0. Many showed high wages in
the $20,000 range. To perform its matches, SBDO relied on first
name and middle initial matches to SSA files other than the Numident,
such as the Detailed Earnings Query Report (DEQR), the Summary
Earnings Query Report, and the Master Beneficiary Record (MBR).
When matches were established that identified the owner of the
wages, SBDO posted the wages to MEF records.
SBDO concluded that 105, or 80 percent, of the 131 wage items
in the demonstration had sufficient matching data to other SSA
files to warrant posting. The 105 postings for this employer reduced
the Suspense File by $1.62 million. The average wage posted was
$15,470.
We reviewed SBDO’s determinations in the 131 demonstration
items for accuracy. Our verification procedures were similar to
SBDO’s procedures. We considered matches as valid only when
a clear and obvious relationship existed between the reported first
name and middle initial and the data on SSA’s Numident files
and DEQR files. For example, the Numident first name and middle
initial matched with Suspense File data. After the Numident match
was established, we reviewed DEQR for additional evidence of identity.
Our results differed from SBDO’s. Of the 131 wage items,
we believe that 107, or 81.7 percent, contained sufficient
matching data to Numident and DEQR files to warrant posting to
MEF records. We differed with SBDO’s determination on 16 cases
(see Appendix A for the details of the 16 cases.) Neither our review
nor SBDO’s review could find any relationship for 24 items
(including 10 items that did not have SSNs). Based on our
results, we believe there is merit to SBDO’s procedures and
that these procedures should be incorporated into SSA’s pilot
studies.
SBDO provided us with its comments on the 16 cases. Of the
seven SBDO matches that we determined to be nonmatches, SBDO agreed
with our determination for three accounts and stated that the three
earnings records have been corrected and no incorrect benefits
have been paid on these accounts. On the other four accounts with
which we disagreed, SBDO believes the matches are correct because
the open wages formed an exact match to the individuals’ earnings
pattern. We noted that one of the four wage earners was just 11
years old in our review year of 1980, based on date of birth data
in the Numident file.
Of the nine SBDO "no matches" that we determined
to be matches, SBDO agreed that all nine should have been posted
as matches. SBDO stated that five accounts were posted at the time
the pilot demonstration was performed for us and one account was
posted prior to the demonstration. For the remaining three errors,
SBDO had made the proper determination, but posted the wages to
the wrong account. However, the Suspense File used in our review,
dated several months after the SBDO demonstration, showed these
six accounts as open items. The SBDO Manager has stated that he
is in the process of performing a follow-up review and correction
of the nine accounts.
We noted that in five of the cases, SBDO made the proper
determination to post wages from the Suspense File to the individual’s
earnings record; however, the wages were posted to the wrong account.
Three of these errors were also included in SBDO’s nine "no
matches" discussed above. The errors caused an overstatement
of earnings for five individuals and an understatement of earnings
for five other individuals. These clerical errors were brought
to the attention of the SBDO Manager who concurred with our observations
and initiated the process of correcting the earnings records.
For the seven OIG "no matches," SBDO has corrected
the three accounts with which we are in agreement and has not corrected
the four with which we disagree. For those nine OIG matches and
the accounts involving clerical errors, only three of the records
have been corrected to date. We were informed that the remainder
are in the process of being corrected.
The SBDO white paper alleged deplorable conditions of neglect
in the correction of suspended wage items for women and minorities.
It strongly implied that the system discriminated against them.
We do not believe, however, that the system discriminates against
women and minorities. The earnings system has special validation
rules for both women and minorities which increase the chances
of posting wage items for females with surname changes and minorities
with compound names.
With respect to women, the Numident validation rules include
three rules specific to females. These rules are more liberal than
the other rules in that matches are based on first names and middle
initials only, thereby, allowing SSA to post wages for those women
who fail to report their married name changes. In contrast, the
other rules require some matching relationship with the last name.
As for minorities, the Numident validation rules include
several that specifically address compound surnames that are common
to certain cultures and minority groups, e.g., Hispanics, Orientals,
Asians, and Arabs. The rules allow posting of wage items where
parts of the reported surname match the Numident. For example,
when a reported compound surname contains a hyphen or a blank,
the Numident validation rules will allow a match if either the
first part or the second part of the compound surname equals the
Numident surname.
CONCLUSIONS
The task force report recommended actions which should reduce
the size of the Suspense File and the number of wage items going
into suspense. Two of the pilot studies emanating from the task
force efforts have near identical objectives to the OIG reviews.
Therefore, we deferred our reviews pending completion of the SSA
pilot studies. We will review SSA’s progress in implementing
actions being studied under the pilots at that time. Our preliminary
work, however, indicates that the procedures used by SBDO in posting
suspended wage items have value in identifying ownership of suspended
wage items. From our perspective, the actions recommended by the
task force should be implemented. The pilot tests now underway
with Operation 30 and the validation procedures are especially
appealing and should be completed. The pilot test should experiment
with the procedures used by SBDO, i.e., establishing ownership
of suspended wage items by matching them with identifying data
and employment history contained on internal SSA records other
than the Numident.
The evidence, such as names and/or initials, provided in
the Suspense File for these seven wage items, in our opinion, is
insufficient to warrant a match with the account number’s
Numident file and Detailed Earnings Query Report.
a) For these nine wage items, we believe the evidence provided
in the Suspense File, such as names and/or initials and an earnings
history with this particular employer, is sufficient to warrant
a match. In many cases, the work history included the years prior
to and after the year of suspense.
b) SBDO made the right determination for 5 wage items but
posted them to the wrong accounts, resulting in errors affecting
10 earnings records. Three of these erroneous accounts are also
included in b) above.
SUMMARY OF SBDO CASES REVIEWED
(a) Seven SBDO Postable to MEF/OIG Not Postable to MEF:
OIG CASE NO. COMMENT
#38 Suspense File showed two initials only. Only middle
initial matched.
#42 Suspense File provided only a first name initial. No
work history with this employer.
#49 Suspense File showed only a first name initial.
#80 Suspense File name did not show any relationship to
Numident name. No work history.
#90 Due to age of wage earner, 11 years old, we believe
this case should be classified as a no match pending further
review.
#117 Suspense File provided only a first name initial.
#119 No name association with Numident and no work history
with this employer.
(b) Nine SBDO Not Postable to MEF/OIG Postable to MEF:
OIG CASE NO. COMMENT
#55 First name: four of seven letters match; prior and
post work history with this employer.
#58 First name: five of seven letters match; middle initial
match; prior work history with this employer.
#63 First name initial and middle initial match; prior
and post work history with this employer.
#74 First name: five of seven letters match; middle initial
match; prior and post work history with this employer.
#95 First name and middle initial match. Prior and post
work history with this employer.
#101 First name: five of seven letters match; middle initial
match; prior work history with this employer.
#105 First name: four of five letters match; middle initial
matches last name initial; prior work history with this employer.
#106 First name: five of six letters match; middle initial
match; prior work history with this employer.
#110 First name initial and middle initial match; prior
and post work history with this employer.
APPENDIX B
Office of the Inspector General
Roger Normand, Director, Program Audits (North)
Emil Mallek, Team Leader
Louis Faiola, Senior Auditor
Frank Cassidy, Auditor
Frank Trzaska, Auditor
www.socialsecurity.gov
