Earned Income Credit: IRS' 1995 Controls Stopped Some Noncompliance, But
Not Without Problems (Letter Report, 09/18/96, GAO/GGD-96-172).

Pursuant to a congressional request, GAO reviewed the Internal Revenue
Service's (IRS) efforts to reduce Earned Income Credit (EIC)
noncompliance in calender year 1995.

GAO found that: (1) the up-front controls used by IRS in its Electronic
Filing System helped IRS reduce some EIC noncompliance and identify
about 1.3 million Social Security number (SSN) problems on
electronically filed tax returns in 1995; (2) IRS placed increased
emphasis on validating SSN on paper returns, since it identified about
3.3 million returns with missing or invalid SSN for EIC-qualifying
children; (3) although IRS identified 3.3 million returns with problems,
IRS only had the resources to follow up on 1 million cases; (4) IRS
delayed refunds on about 4 million EIC returns that did not have any SSN
problems to check for the use of duplicate SSN, but released almost all
of those refunds without checking for duplicate SSN; (5) IRS has taken
steps to better utilize its resources in 1996, such as attempting to
identify more productive cases and limiting the number of delayed
refunds; and (6) the overall impact on IRS efforts to reduce EIC
noncompliance cannot be assessed because IRS commingled 1995 data with
data from previous years.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  GGD-96-172
     TITLE:  Earned Income Credit: IRS' 1995 Controls Stopped Some 
             Noncompliance, But Not Without Problems
      DATE:  09/18/96
   SUBJECT:  Noncompliance
             Social security number
             Erroneous payments
             Tax credit
             Internal controls
             Tax administration
             Tax refunds
             Fraud
             Eligibility determinations
             Disadvantaged persons
IDENTIFIER:  IRS Questionable Refund Program
             Earned Income Tax Credit
             IRS Taxpayer Compliance Measurement Program
             IRS Electronic Filing System
             TCMP
             
******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


Cover
================================================================ COVER


Report to the Chairman, Committee on Finance, U.S.  Senate

September 1996

EARNED INCOME CREDIT - IRS' 1995
CONTROLS STOPPED SOME
NONCOMPLIANCE, BUT NOT WITHOUT
PROBLEMS

GAO/GGD-96-172

Earned Income Credit

(268692)


Abbreviations
=============================================================== ABBREV

  EIC - Earned Income Credit
  IRS - Internal Revenue Service
  RAL - refund anticipation loan
  SSN - Social Security number
  TCMP - Taxpayer Compliance Measurement Program

Letter
=============================================================== LETTER


B-261635

September 18, 1996

The Honorable William V.  Roth, Jr.
Chairman, Committee on Finance
United States Senate

Dear Mr.  Chairman: 

This report responds to your request for information on Earned Income
Credit (EIC) noncompliance.\1 Specifically, our objective was to
provide information on and our analysis of the results of the
Internal Revenue Service's (IRS) efforts to reduce noncompliance in
calendar year 1995.  This report updates preliminary data included in
other products we provided to you during the course of this
assignment.\2


--------------------
\1 Throughout this report we use the term "noncompliance" to include
erroneous EIC claims caused by taxpayer mistakes, negligence, or
fraud.  Determining whether an EIC claim is fraudulent requires
knowing the taxpayer's intent, which is difficult to prove. 

\2 Tax Administration:  Earned Income Credit--Data on Noncompliance
and Illegal Alien Recipients (GAO/GGD-95-27, Oct.  25, 1994); Earned
Income Credit:  Targeting to the Working Poor (GAO/GGD-95-122BR, Mar. 
31, 1995); Earned Income Credit:  Targeting to the Working Poor
(GAO/T-GGD-95- 136, Apr.  4, 1995); and Earned Income Credit: 
Noncompliance and Potential Eligibility Revisions (GAO/T-GGD-95-179,
June 8, 1995). 


   RESULTS IN BRIEF
------------------------------------------------------------ Letter :1

IRS took several steps to prevent and detect EIC noncompliance in
1995.  IRS' data indicate that those steps stopped some
noncompliance.  However, there were also some problems. 

IRS uses up-front controls (filters) in its Electronic Filing System
to identify electronic submissions with problems, such as invalid or
duplicate Social Security numbers (SSN).\3 If a problem is
identified, the submission is rejected, and the problem has to be
corrected before IRS will accept the electronic return.  IRS added
some new filters for 1995.  Those new filters, plus the ones already
in place, identified about 1.3 million SSN problems on electronic
submissions from persons who were claiming the EIC in 1995, compared
with about 600,000 problems in 1994. 

In 1995, for paper returns, IRS placed an increased emphasis on
transcribing and validating SSNs.  IRS identified about 3.3 million
paper returns (most claiming refunds) with missing or invalid SSNs
for EIC-qualifying children\4 and/or dependents.  In about 1 million
of those cases, IRS delayed refunds or otherwise held up the
processing of the returns while staff in IRS' Examination function
contacted taxpayers, asking them to prove their eligibility.  As of
June 30, 1996, those contacts had resulted in over $800 million in
either reduced refunds or additional tax assessments. 

Even though IRS' efforts in 1995 produced some favorable results, IRS
also encountered several problems in implementing those efforts.  For
example, IRS' procedures generated a workload that far exceeded IRS'
capabilities.  Although 3.3 million paper returns were identified
with missing or invalid SSNs, IRS had sufficient resources to follow
up on only about 1 million.  And, according to IRS' Internal Audit
Division, IRS' procedures for deciding which of the cases warranted
follow-up did not ensure selection of the most productive cases and
resulted in an inefficient use of resources.  IRS was also unable to
follow through on plans to check for duplicate use of SSNs.  In that
regard, IRS delayed refunds on about 4 million EIC returns that did
not have an SSN problem, primarily to give it more time to determine
whether other returns were filed using the same SSNs.  However, IRS
eventually released almost all of those refunds, after holding them
for several weeks, without checking for duplicate SSNs.  IRS took
steps to prevent the recurrence of these problems in 1996.  It
revised its procedures for selecting cases to review in an attempt to
identify more productive cases and to limit the number of delayed
refunds. 

IRS' efforts and the publicity surrounding them may have also had a
sizable deterrent effect, at least in the short term.  According to
IRS data, for example, over 2 million fewer EIC claims were filed in
1995 than IRS had expected. 

Although IRS' data provided some evidence of the results of its
efforts in 1995, the data were not sufficient to allow an overall
assessment of the impact of IRS' initiatives on EIC noncompliance. 
For example, (1) IRS has not yet released the results of an EIC
compliance study it did in 1995 and (2) data on the results of the
Examination function's SSN verification efforts were not reported in
a way that isolated tax year 1994 cases from prior years' cases or
distinguished between EIC cases and cases involving dependents. 


--------------------
\3 As used in this report, an invalid SSN is one that does not match
Social Security Administration records. 

\4 A qualifying child (1) is an EIC claimant's son, daughter, adopted
child, grandchild, stepchild, or foster child; (2) is under age 19,
or under age 24 and a full-time student, or any age and permanently
and totally disabled; and (3) lives in the claimant's home in the
United States for more than half of the year (or all of the year if a
foster child). 


   BACKGROUND
------------------------------------------------------------ Letter :2

The EIC is a refundable tax credit available to low-income working
taxpayers.  Congress established the EIC in 1975 to offset the impact
of Social Security taxes on low-income families and to encourage
low-income individuals with families to seek employment rather than
welfare.  EIC coverage and benefit amounts have expanded
significantly since 1975.  For example, provisions in the Omnibus
Budget Reconciliation Act of 1993 (1) raised the maximum credit for
families with two or more EIC-qualifying children to $2,528 for tax
year 1994 and (2) made certain taxpayers without qualifying children
eligible for the credit starting in tax year 1994.  For tax year
1994, about 18.9 million taxpayers received about $20.8 billion in
EIC benefits.  Almost all of those benefits ($20.1 billion) went to
families with EIC-qualifying children.  Other families who qualify
for the EIC may not be claiming it.  In that regard, researchers have
estimated that between 75 and 86 percent of all eligible families
actually claimed the EIC in 1990.\5


--------------------
\5 Yin et al., Improving the Delivery of Benefits to the Working
Poor:  Proposals to Reform the Earned Income Tax Credit Program,
American Tax Policy Institute, Feb.  1994. 


      EIC NONCOMPLIANCE
---------------------------------------------------------- Letter :2.1

IRS data show high noncompliance rates associated with EIC claims. 
Some noncompliance involves mathematical errors and other obvious
mistakes made by taxpayers or their representatives in preparing the
returns.  Staff in IRS' 10 service centers are to review each paper
return when it is received to make sure it is accurate and complete
(e.g., includes required supporting schedules) and, for returns
claiming the EIC, contains basic eligibility information, such as the
age of qualifying children.\6

Information is then entered into computers.  The computers check for
math and qualifying errors, some of which could affect EIC
eligibility or the size of the EIC claim. 

EIC claims have for years been the source of many errors identified
during processing.  In 1995, for example, IRS identified 1.6 million
EIC-related errors involving about 8 percent of all returns with EIC
claims.  Of the 1.6 million errors, about 600,000 involved situations
where the taxpayer claimed the EIC but was found not to qualify and
about 1 million involved cases where the taxpayer erred in computing
the EIC. 

Other noncompliance involves mistakes that can be detected only
through an audit of the return.  In the past, IRS measured this kind
of noncompliance through its Taxpayer Compliance Measurement Program
(TCMP).\7 The last TCMP, involving audits of tax year 1988 returns,
found that about 42 percent of EIC recipients were not entitled to
some or all of the credit that they claimed--representing about 34
percent of the EIC dollars paid that year.  However, these TCMP
results may not represent current compliance levels because the EIC
has changed substantially since 1988.  For example, changes enacted
in 1990 included a major redesign and simplification of the EIC
eligibility rules that had accounted for many of the errors found in
the 1988 TCMP. 

More recently, IRS sampled EIC returns filed electronically during a
2-week period in January 1994.  Although the results can be
generalized only to returns filed in that 2-week period, IRS'
analysis of this limited EIC data showed that 39 percent of the
returns involved overstated EIC claims that represented 26 percent of
the dollars claimed.  IRS conducted a broader, more statistically
reliable study of EIC returns filed electronically and on paper in
1995.  As of August 26, 1996, IRS had not released the results of
that study. 

The most serious form of noncompliance involves deliberate attempts
to defraud the government through, for example, phony refund claims. 
The Questionable Refund Program, established in the 1970s, is IRS'
primary effort to identify fraudulent refund claims, including those
involving the EIC.  An IRS computer program analyzes all returns to
identify those that are potentially fraudulent.  Then, fraud
detection teams in the 10 service centers perform more in-depth
reviews and, if a return is considered fraudulent, attempt to stop
any refund before it is issued.  The number of returns identified by
IRS as containing fraudulent refund claims and the total dollar
amount of stopped refunds have increased significantly since 1990. 
From January 1 through December 31, 1995, the fraud detection teams
had identified about 62,000 fraudulent returns and stopped about $83
million in refunds.  About 72 percent of the returns had EIC claims. 
We do not know whether that large percentage reflects (1) the level
of EIC-related fraud compared with other types of fraud, (2) IRS'
emphasis on EIC-related fraud, or (3) the comparative ease of
identifying EIC-related fraud versus other types of fraud. 


--------------------
\6 These types of errors are avoided on electronic returns through
filters that check such things as mathematical correctness before the
submission is accepted. 

\7 TCMP involves audits of a random sample of tax returns in which
every line of every return is to be examined. 


      IRS EFFORTS TO ADDRESS EIC
      NONCOMPLIANCE IN 1995
---------------------------------------------------------- Letter :2.2

Over the past few years, more attention has been placed on
determining whether the EIC is being paid to ineligible taxpayers. 
In 1995, IRS expanded its efforts to identify and stop incorrect
refunds.  Much of what IRS did involved verifying SSNs, with an
emphasis on returns claiming the EIC.  IRS was looking for SSNs that
did not match the Social Security Administration's records or that
had been used on another return filed that year, and returns that
were missing one or more required SSNs.  IRS warned taxpayers that
their refunds could be delayed if they submitted a return with a
missing or incorrect SSN.  On the cover of the instructions
accompanying Form 1040, for example, IRS warned taxpayers to check
their SSNs and explained that "incorrect or missing SSNs for you,
your spouse, or dependents may delay your refund." IRS also issued
several public service announcements to alert taxpayers to the need
for correct SSNs. 

Another important step IRS took in preparing for the 1995 filing
season was to eliminate the direct deposit indicator.  In conjunction
with the electronic filing program, the private sector offers what is
commonly referred to as a refund anticipation loan (RAL).  These
loans enable taxpayers, for a fee, to get their money more quickly
than if they were to wait for IRS to issue their refunds.  A taxpayer
repays the loan by arranging to have the refund deposited directly to
an account specified for repayment of the loan.  Although RALs are
contracts between the financial institution and the borrower, IRS
facilitated the process in the past by providing the direct deposit
indicator to the electronic return transmitter after the return was
received, acknowledging that the taxpayer's direct deposit request
would be honored.  IRS would not honor a request if the taxpayer had
a debt, such as unpaid child support or unpaid federal taxes, that
would be offset against the taxpayer's refund.  Because the
opportunity to get money quickly through RALs was seen as encouraging
electronic filing fraud and because a large number of EIC fraud
schemes identified by IRS in the past involved RALs, IRS did not
provide the direct deposit indicator in 1995. 


   OBJECTIVE, SCOPE, AND
   METHODOLOGY
------------------------------------------------------------ Letter :3

Our objective was to provide information on and our analysis of IRS'
efforts to reduce EIC noncompliance in 1995. 

To achieve our objective, we

  -- reviewed studies on EIC noncompliance,

  -- reviewed IRS' initiatives and procedures for preventing and
     detecting EIC noncompliance,

  -- analyzed IRS data on the results of its efforts to reduce EIC
     noncompliance,

  -- interviewed IRS National Office officials responsible for
     various EIC compliance initiatives, and

  -- interviewed Cincinnati and Fresno Service Center officials
     responsible for EIC-related studies or investigations. 

IRS took several steps in 1995 to address a growing problem with
refund fraud in general.  In March 1996, we reported on those
efforts.\8 Unlike that report, this report focuses, to the extent
possible, on IRS' efforts and results as they relate specifically to
EIC noncompliance.  As discussed later, our attempt to focus
specifically on EIC-related results was limited by the nature of IRS'
data. 

We relied on data provided in IRS' reports and did not verify the
data.  We did our work from January 1995 through June 1996 in
accordance with generally accepted government auditing standards. 

You also asked us to determine the extent of EIC noncompliance. 
Although this letter contains some data on noncompliance, a critical
piece of information needed to respond to that portion of your
request--the results of IRS' study of EIC returns filed in 1995--was
unavailable at the time we prepared this report.  That study was
designed to provide current and projectable data on the extent of EIC
noncompliance.  IRS said that it would provide a report on the
results of this study after completing its analysis of the data. 
After we receive the report, we will analyze the results and issue a
separate product. 

We requested comments on a draft of this report from the Commissioner
of Internal Revenue or her designee.  On July 30, 1996, we met with
the Assistant Commissioner for Criminal Investigations and other IRS
officials, who provided us with oral comments.  Those comments were
generally reiterated and expanded upon in an August 12, 1996, letter
from the Acting Chief Compliance Officer.  IRS' comments are
summarized and evaluated beginning on page 12 and the Acting Chief's
letter is reprinted in the appendix. 


--------------------
\8 Tax Administration:  IRS Efforts to Control Fraud in 1995
(GAO/GGD-96-96R, Mar.  25, 1996). 


   IRS' EFFORTS TO REDUCE EIC
   NONCOMPLIANCE IN 1995:  SOME
   POSITIVE RESULTS, SOME PROBLEMS
------------------------------------------------------------ Letter :4

IRS took several steps in 1995 to combat a growing problem with
refund fraud in general and, more specifically, EIC noncompliance. 
Most significantly, IRS (1) expanded the up-front controls in its
Electronic Filing System, (2) placed increased emphasis on its
efforts to verify SSNs on paper returns, and (3) held up the refunds
on millions of EIC returns with valid SSNs to allow IRS time to check
for duplicate SSN usage.\9 IRS' efforts had some positive results
(e.g., over $800 million in reduced refunds and additional tax
assessments) but also had some problems (e.g., an inability to follow
through on plans to check for duplicate SSNs) that limited their
effectiveness.  IRS' efforts and the publicity surrounding them also
may have had a sizable deterrent effect.  For example, they may have
contributed to the receipt of many fewer EIC claims in 1995 than IRS
had expected. 


--------------------
\9 These changes were made possible because IRS transcribed more SSNs
from tax returns in 1995, thus allowing for more computer matching. 


      SOME NONCOMPLIANCE AVOIDED
      BY UP-FRONT CONTROLS IN
      ELECTRONIC FILING SYSTEM
---------------------------------------------------------- Letter :4.1

For the past several years, IRS has included up-front controls
(filters) in its Electronic Filing System to identify submissions
that had data problems, such as missing or invalid SSNs or an SSN
that had already been used on another return filed for the same tax
year.  If a problem was identified, IRS refused to accept the
electronic submission until the problem was corrected.  In 1994, IRS'
electronic filters identified about 1 million SSN problems, about
600,000 of which involved the EIC.\10 In 1995, IRS added more
electronic filters to prevent multiple use of an SSN on a return or
EIC schedule and identified 4.1 million SSN problems, of which about
1.3 million involved the EIC.\11 Those EIC-related problems included
instances where the SSN, name, and date of birth for an
EIC-qualifying child did not match Social Security Administration
records and instances where the SSN had been previously used on
another return claiming an EIC-qualifying child. 

There is no way of knowing how many of those problems involved
intentional noncompliance, as opposed to honest mistakes or IRS
database problems.  Also, IRS does not routinely track electronic
filing rejections and thus does not know whether the problems were
eventually resolved or whether the rejected returns were ever
resubmitted (either electronically or on paper).  However, evidence
suggests that some taxpayers whose electronic submissions were
rejected because of an SSN problem were able to avoid the problem by
filing on paper.  IRS reviewed 395 electronic submissions that were
rejected because of duplicate SSNs and found that in 113 cases (29
percent) the taxpayers subsequently filed on paper, using the same
problem SSNs, and received their refunds.  The results of this test,
which involved cases from two of the five IRS service centers that
receive electronic returns, are not projectable to all electronic
filers. 


--------------------
\10 Because an electronic submission can be rejected for more than
one reason, the number of problems identified does not equal the
number of submissions rejected. 

\11 Of the 700,000 additional EIC-related problems identified by the
electronic filters in 1995 compared with 1994, about 150,000 were due
to the 1 new EIC-related filter IRS added in 1995. 


      IRS IDENTIFIED SIGNIFICANT
      NONCOMPLIANCE ON PAPER
      RETURNS BUT ALSO ENCOUNTERED
      PROBLEMS
---------------------------------------------------------- Letter :4.2

IRS also set up controls to better identify noncompliance on paper
returns.  Starting in 1994, IRS identified certain returns that had
missing or invalid SSNs for EIC-qualifying children and delayed
refunds to give Examination staff in IRS' 10 service centers enough
time to validate EIC eligibility.  Those validation efforts in 1994
showed that about 300,000 of the EIC claimants were ineligible for
some portion of the credit. 

IRS expanded its SSN validation efforts in 1995 to include dependents
with a problem SSN and identified 3.3 million paper returns with 1 or
more missing or invalid SSNs for EIC-qualifying children and/or
dependents.\12 About 3 million of those returns involved requests for
refunds.  Examination had enough resources to review only about 1
million of the questionable returns.  In those 1 million cases, IRS
sent notices to taxpayers telling them (1) that a problem had been
identified with their returns; (2) what they had to do to resolve the
problem; and (3) that their refund, if they had claimed one, was
being delayed while IRS checked for noncompliance.  For the other 2.3
million returns, all of which involved refunds, IRS delayed the
refunds but did not refer the returns to Examination for follow-up. 
The taxpayers were told that their refunds were being delayed but
were not told that IRS had identified a problem on their returns. 
IRS subsequently released the refunds after holding them for several
weeks. 

Information on the results of Examination's review of the 1 million
cases was not readily available from IRS.  Although IRS routinely
reports on the disposition of such reviews, the data are not reported
in a way that aligns results with specific tax years.  Instead,
results are reported on the basis of the fiscal year a case is
closed.  Thus, data reported for fiscal year 1995 represented the
results of all cases closed in 1995, no matter what the tax year. 
While that kind of reporting has value, it was not useful for
assessing the results of IRS' expanded SSN-verification procedures in
1995 because the results of that year's cases were commingled with
the results of prior years' cases. 

To determine the results of IRS' fiscal year 1995 efforts, we
requested a special breakdown of Examination's case closure data that
aligned results by tax year.  We analyzed the data and provided a
copy to IRS' Office of Refund Fraud.\13 We concluded, and officials
of the Office of Refund Fraud agreed, that tax year results are
helpful in assessing program initiatives.  According to IRS
officials, these data could be provided on a regular basis at minimal
cost. 

Our analysis of the special breakdown of Examination's case closure
data showed that 986,000 of the 1 million tax year 1994 cases had
been closed as of June 30, 1996.  Of the closed cases, 500,000 (51
percent) were closed with no change to the reported tax liability or
refund because the taxpayers were able to prove that they were
entitled to claim the dependent or the EIC.  The other 486,000 cases
were closed with changes (either in reduced refunds or additional
taxes assessed) amounting to about $808 million.  Even with our
special breakdown, we do not know how much of this money related to
EIC claims because IRS' data did not distinguish between cases
involving dependents and those involving EIC-qualifying children. 

According to data reported in IRS' Data Book for fiscal years 1993
and 1994 (the most recent reported data), the 51-percent no-change
rate for cases with missing or invalid SSNs for EIC-qualifying
children or dependents was more than double the 24 percent no-change
rate for all service center audits done in fiscal year 1994.  The
high no-change rate can be attributed to IRS procedures that,
according to IRS' Internal Audit Division, did not adequately ensure
selection of the most productive cases and thus resulted in an
inefficient use of Examination resources and an undue burden on
thousands of taxpayers.  IRS changed its procedures for 1996 in an
attempt to target its resources on the most egregious cases and
minimize the burden on taxpayers. 


--------------------
\12 IRS does not know the number of paper returns filed in 1995 with
missing or invalid SSNs for EIC-qualifying children and dependents
because (1) IRS only identified returns that met certain minimum
dollar criteria and (2) computer problems caused some returns not to
be selected for SSN verification when they should have been.  The
computer problems occurred primarily because IRS did not have enough
time to design and test program changes. 

\13 The Office of Refund Fraud was established on October 1, 1994, to
monitor IRS' programs to stop refund fraud and noncompliance and
suggest needed changes. 


      IRS DID NOT FOLLOW THROUGH
      ON PLANS TO CHECK FOR
      DUPLICATE SSN USE
---------------------------------------------------------- Letter :4.3

Because IRS studies have shown a high risk of noncompliance with
returns claiming the EIC, IRS decided to delay about 4 million EIC
refunds in 1995 even though IRS had identified no missing or invalid
SSNs on those returns.\14 The 4 million returns included returns
filed electronically and on paper with EIC claims above a certain
dollar amount.  IRS stated that one of its goals in doing so was to
allow additional time to identify any returns that might be filed
later using one or more of the same SSNs as the delayed returns, with
the expectation that Examination staff would research the duplicate
SSN usage and stop inappropriate refunds.  IRS was not able to
realize this potential because it did not have enough resources to
research many questionable cases.  After holding the refunds for
several weeks, IRS released almost all of them without checking for
duplicate SSNs.  For the 1996 filing season, IRS revised its
procedures so as not to delay refunds on returns with valid SSNs. 

An April 1996 report by IRS' Internal Audit Division provided some
indication of the level of noncompliance associated with the
duplicate use of SSNs on EIC claims.  In that report, Internal Audit
estimated that the number of duplicate SSN occurrences\15 on returns
filed in 1995 ranged from 233,000 to 449,000 and that the revenue
impact ranged from $283 million to $545 million. 


--------------------
\14 Because IRS' primary concern on these returns centered on the
validity of the EIC claim, it delayed only that part of the refund
attributable to the EIC. 

\15 "Occurrences" are the number of times an SSN was used more than
once.  For example, if the same SSN was used on three returns filed
in 1995, the number of occurrences would be two.  Similarly, if the
same SSN was used for two different children on the same return, the
number of occurrences would be one. 


      IRS' EFFORTS MAY HAVE HAD A
      DETERRENT EFFECT
---------------------------------------------------------- Letter :4.4

IRS' efforts to better control EIC noncompliance in 1995 and the
publicity surrounding them may have had a significant deterrent
effect. 

The number of EIC claims filed by persons with qualifying children
had increased steadily over the past 10 years.  For 1995, IRS'
Research Division had estimated that the number of such claims would
increase by about 2.2 million.  IRS data showed, instead, that
persons with qualifying children made about 100,000 fewer EIC claims
in 1995 than in 1994.  In that regard, the Congressional Budget
Office, in its August 1995 Economic and Budget Outlook update,
decreased anticipated EIC outlays by $2 billion to $3 billion a year. 
In doing so, it stated that EIC spending "has been lower than
expected this year, possibly as a result of a recent crackdown by the
Internal Revenue Service on fraudulent claims."

According to the Director of IRS' Office of Refund Fraud, another
indication of the deterrent effect of IRS' efforts in 1995 was the
drop in identified fraud by IRS' fraud detection teams.  In calendar
year 1995, the detection teams identified $132 million in fraudulent
refunds on 62,309 returns compared with $161 million on 77,781
returns in 1994.  Likewise, about 73 percent of the fraudulent
returns identified in 1995 involved EIC claims, down from 91 percent
in 1994.  Although the numbers went down, there is no way of knowing,
from available data, whether the decrease reflects a decline in the
incidence of fraud or just a decrease in the amount of fraud
identified by IRS.  For example, elimination of the direct deposit
indicator, which the Director said was one of the most effective
actions taken by IRS for 1995, may have contributed to a decrease in
fraud.  But such a direct cause and effect relationship is difficult,
if not impossible, to prove. 


   CONCLUSIONS
------------------------------------------------------------ Letter :5

EIC noncompliance has been an ongoing concern of Congress and IRS. 
To meet the challenge, IRS expanded its controls in 1995 to better
prevent taxpayers from receiving EIC benefits to which they were not
entitled.  A successful compliance program requires that IRS
effectively balance taxpayer burden against the program's revenue
protection benefits.  In implementing its fiscal year 1995 controls,
however, IRS delayed significantly more EIC refunds than it was able
to review and did not select the most productive cases to review. 
While we agree that IRS needs to delay EIC refunds in order to follow
up with taxpayers on questionable claims, we believe that IRS would
have achieved better results if it had better targeted its efforts to
those cases most in need of review.  For the 1996 filing season, IRS
decided to delay only those cases that it had the time and resources
to review and revised its procedures in an attempt to select the most
egregious cases to review. 

Although IRS data indicated that its controls for 1995 identified and
prevented some noncompliance, including that associated with the EIC,
IRS did not compile data in such a manner as to allow for a
meaningful analysis of those controls.  The results of IRS' SSN
validation efforts on paper returns were reported in a way that (1)
did not distinguish between dependent claims and EIC claims and (2)
commingled the results of IRS' efforts in 1995 with the results of
efforts in prior years.  The ultimate impact of the up-front filters
in the Electronic Filing System is unknown because IRS does not track
the resolution of problems identified by the filters.  If IRS does
not have adequate data to assess its efforts, it is less likely to
make informed decisions about continuing, expanding, or revising
those efforts. 

Some of the data discussed in this report, such as the disaggregation
of Examination results by tax year, would seem inexpensive to
compile.  Other data, such as the tracking of electronic rejections,
might be more costly.  Only IRS knows what such efforts would cost
and whether compilation of the data is feasible given the cost and
the level of effort IRS expects to devote to EIC noncompliance in the
future. 


   RECOMMENDATION TO THE
   COMMISSIONER OF INTERNAL
   REVENUE
------------------------------------------------------------ Letter :6

We recommend that IRS consider cost-effective ways to compile the
kind of data needed to better assess the effectiveness and direction
of its efforts to combat EIC noncompliance.  In doing so, IRS should
consider (1) routinely reporting data, by tax year, on the results of
Examination efforts to validate eligibility for benefits; (2)
tracking what happens to returns rejected by the Electronic Filing
System; and (3) distinguishing the results relating to EIC-qualifying
children from the results relating to dependents. 


   AGENCY COMMENTS AND OUR
   EVALUATION
------------------------------------------------------------ Letter :7

We requested comments on a draft of this report from the Commissioner
of Internal Revenue or her designee.  On July 30, 1996, we met with
the Assistant Commissioner for Criminal Investigations and other IRS
officials, who provided us with oral comments.  Those comments were
generally reiterated and expanded upon in an August 12, 1996, letter
from the Acting Chief Compliance Officer (see app.). 

In response to our suggestion that IRS consider reporting
Examination's results by tax year, IRS agreed that such information
is important in assessing program effectiveness and said that it is
available when needed by querying an automated management information
system.  Our report acknowledges that such information exists. 
However, it is important not only to have the information available
but also to use it.  As we pointed out earlier, IRS was not using
tax-year specific data to assess program results until we
specifically asked for it.  To clarify our intent, as discussed with
IRS officials at our July 30 meeting, we reworded our recommendation
to say that IRS should consider routinely "reporting" data by tax
year. 

IRS also said that while it seems reasonable to track what happens to
returns rejected by the Electronic Filing System, certain legal
ramifications have to be explored first.  Those ramifications center
on the question of whether any files of rejected electronically filed
returns that IRS might have to compile for tracking purposes would
constitute a system of records under the Freedom of Information Act. 
The officials said that the legal issues had been referred to IRS'
Office of Chief Counsel.  We agree that any possible legal issues
should be resolved before designing and implementing a system to
track rejected returns. 

With respect to our suggestion that IRS consider distinguishing the
results relating to EIC-qualifying children from the results relating
to dependents, IRS said that it wanted to defer any decision while
two pieces of legislation, that would have a bearing on how IRS
handles missing/invalid SSN conditions in the future, were
pending.\16 According to IRS, the new procedures called for in the
legislation would not automatically provide the sort of data we
envisioned and that complicated systems programming would be required
to capture the data systemically.  IRS said that until its fiscal
year 1997 appropriation is approved, it is unable to determine if
resources will be available to make the programming changes.  We
agree with IRS' position. 


--------------------
\16 The legislation referred to by IRS was signed into law on August
20 and August 22, 1996. 


---------------------------------------------------------- Letter :7.1

We are sending copies of this report to the Ranking Minority Member
of the Senate Finance Committee, the Chairman and Ranking Minority
Member of the House Committee on Ways and Means, various other
congressional committees, the Secretary of the Treasury, the
Commissioner of Internal Revenue, the Director of the Office of
Management and Budget, and other interested parties.  Copies will be
made available to others upon request. 


If you have any questions, please contact me on (202) 512-5594. 
Major contributors to this report were David J.  Attianese, Assistant
Director, and William H.  Bricking, Evaluator-in-Charge. 

Sincerely yours,

James R.  White
Associate Director, Tax Policy
 and Administration Issues




(See figure in printed edition.)Appendix
COMMENTS FROM THE INTERNAL REVENUE
SERVICE
============================================================== Letter 



(See figure in printed edition.)

*** End of document. ***