Advance Earned Income Tax Credit: Low Use and Small Dollars Paid 
Impede IRS's Efforts to Reduce High Noncompliance (10-AUG-07,	 
GAO-07-1110).							 
                                                                 
The Advance Earned Income Tax Credit (AEITC) allows individuals  
to receive a portion of the Earned Income Tax Credit (EITC) in	 
their paychecks, instead of receiving all of it when filing their
year-end tax return. Limited research has been conducted on the  
AEITC since GAO last examined it in the early 1990s. GAO was	 
asked to determine (1) how many individuals received the AEITC	 
compared with the EITC in tax years 2002 through 2004, what	 
actions, if any, have been taken to increase use, and the	 
potential for increases in use in the future; (2) the extent of  
noncompliance, if any, associated with the AEITC; and (3) how	 
well the Internal Revenue Service's (IRS) procedures address the 
areas of noncompliance. To address these questions, GAO analyzed 
Forms W-2 and tax return data and interviewed IRS and Social	 
Security Administration (SSA) officials.			 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-07-1110					        
    ACCNO:   A74228						        
  TITLE:     Advance Earned Income Tax Credit: Low Use and Small      
Dollars Paid Impede IRS's Efforts to Reduce High Noncompliance	 
     DATE:   08/10/2007 
  SUBJECT:   Advance payments					 
	     Cost effectiveness analysis			 
	     Data integrity					 
	     Eligibility determinations 			 
	     Income taxes					 
	     Noncompliance					 
	     Program evaluation 				 
	     Reporting requirements				 
	     Social security number				 
	     Tax administration 				 
	     Tax credit 					 
	     Tax law						 
	     Tax return audits					 
	     Tax returns					 
	     Program goals or objectives			 
	     Earned Income Tax Credit				 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO 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.                                         **
**                                                              **
******************************************************************
GAO-07-1110

   

     * [1]Results in Brief
     * [2]Background
     * [3]AEITC Use Low and May Be Difficult to Increase

          * [4]AEITC Use Has Not Increased Significantly Despite Several Ta
          * [5]Future AEITC Growth Is Unlikely

     * [6]High Noncompliance Exists with AEITC Requirements

          * [7]About 20 Percent of AEITC Recipients Had an Invalid SSN
          * [8]Almost 40 Percent of AEITC Recipients Did Not File the Requi
          * [9]About Two-Thirds of AEITC Recipients Who Filed a Federal Tax
          * [10]AEITC Recipients Were Noncompliant with Other Program Requir

     * [11]IRS's AEITC Compliance Procedures Have Limited Effectiveness

          * [12]IRS's Procedures to Verify SSNs Have Limited Effectiveness f
          * [13]Many Taxpayers Underreported AEITC and IRS Worked on a Limit
          * [14]IRS Works on a Limited Number of AEITC Nonfiler Cases
          * [15]Collection Handles Few AEITC Cases
          * [16]IRS Audits and Criminal Investigation Identified a Small Num
          * [17]Soft Notices Have Increased Compliance in Situations Similar
          * [18]Federal On-Line SSN Verification Services Could Be Used to D
          * [19]IRS Does Not Require Submission of Forms W-5 from Employers
          * [20]Options to Reduce AEITC Noncompliance May Be Cost Effective
          * [21]Eligible Taxpayers Could Receive Full EITC Benefits If the A

     * [22]Conclusions
     * [23]Recommendations for Executive Action
     * [24]Agency Comments
     * [25]GAO Contacts
     * [26]Acknowledgments
     * [27]GAO's Mission
     * [28]Obtaining Copies of GAO Reports and Testimony

          * [29]Order by Mail or Phone

     * [30]To Report Fraud, Waste, and Abuse in Federal Programs
     * [31]Congressional Relations
     * [32]Public Affairs

Report to the Joint Committee on Taxation

United States Government Accountability Office

GAO

August 2007

ADVANCE EARNED INCOME TAX CREDIT

Low Use and Small Dollars Paid Impede IRS's Efforts to Reduce High
Noncompliance

GAO-07-1110

Contents

Letter 1

Results in Brief 3
Background 6
AEITC Use Low and May Be Difficult to Increase 9
High Noncompliance Exists with AEITC Requirements 15
IRS's AEITC Compliance Procedures Have Limited Effectiveness and Options
to Better Address Noncompliance Have Advantages and Disadvantages 21
Conclusions 44
Recommendations for Executive Action 45
Agency Comments 46
Appendix I IRS's Implementation of Recommendations from GAO's 1992 Report
on the Advance Earned Income Tax Credit 47
Appendix II Scope and Methodology 49
Appendix III Demographic Characteristics of Advance Earned Income Tax
Credit Recipients and Their Employers 56
Appendix IV Administrative and Legislative Changes to the Advance Earned
Income Tax Credit Since 1990 76
Appendix V Additional Analyses of Advance Earned Income Tax Credit
Noncompliance 78
Appendix VI Comments from the Internal Revenue Service 92
Appendix VII GAO Contacts and Staff Acknowledgments 96

Tables

Table 1: EITC Requirements for 2006 That All AEITC Recipients Must Also
Expect to Meet 8
Table 2: Percentage of AEITC Recipients Receiving an Amount of AEITC, Tax
Years 2002 through 2004 11
Table 3: Frequency of Individuals Electing the EITC and AEITC, Tax Years
2002 through 2004 14
Table 4: AEITC Recipients and Recipients with Invalid SSN on Form W-2,
Individuals and Dollars, Tax Years 2002 through 2004 17
Table 5: AEITC Recipients and Recipients Who Did Not File a Tax Return,
Individuals and Dollars, Tax Years 2002 through 2004 18
Table 6: AEITC Recipients, Recipients Who Did Not File a Federal Tax
Return, and Recipients Who Did Not File a Tax Return and Had an Invalid
SSN on the Form W-2, Individuals and Dollars, Tax Years 2002 through 2004
19
Table 7: AEITC Matches and Mismatches between Form(s) W-2 and Filed Tax
Returns, Tax Years 2002 through 2004 20
Table 8: IRS Implementation of Recommendations from GAO's 1992 Report on
the Advance Earned Income Tax Credit 47
Table 9: Number of Forms W-2, Number of Individuals, and the Dollars for
Each Subpopulation, Tax Years 2002 through 2004 51
Table 10: Number of Forms W-2 Received by Subpopulation, Tax Years 2002
through 2004 58
Table 11: AGI for AEITC and EITC Recipients, Number, Average, and Sum, for
Tax Year 2002 62
Table 12: AGI for AEITC and EITC Recipients, Number, Average, and Sum, for
Tax Year 2003 63
Table 13: AGI for AEITC and EITC Recipients, Number, Average, and Sum, for
Tax Year 2004 63
Table 14: Wages Reported for AEITC and EITC Recipients, Number, Wage
Amount, and Average Dollars Received for Tax Years 2002 through 2004 65
Table 15: Filing Status of AEITC and EITC Recipients, Number and
Percentage, Tax Years 2002 through 2004 66
Table 16: Age of AEITC and EITC Recipients, Tax Years 2002 through 2004 68
Table 17: Gender of AEITC and EITC Recipients, Number and Percentage, Tax
Years 2002 through 2004 70
Table 18: Number of Qualifying Children for AEITC and EITC Recipients,
Number and Percentage, Tax Years 2002 through 2004 72
Table 19: Filing Method of AEITC and EITC Recipients, Number and
Percentage, Tax Years 2002 through 2004 72
Table 20: Employer Size, Number of Employers, Number of Forms W-2
Employers Issued to AEITC Recipients and Total Dollars Reported on Forms
W-2, Tax Years 2002 through 2004 75
Table 21: Administrative Changes to the AEITC since 1990 76
Table 22: Number of Individuals and AEITC Dollars Received by AEITC
Recipients and Individuals Who Elected the AEITC Consecutively Using an
Invalid SSN and an Invalid SSN and Not Filing, Tax Years 2002 through 2004
79
Table 23: Number of AEITC Consecutive Recipients with Matches and
Mismatches between Forms W-2 and Filed Tax Returns, Tax Years 2002 through
2004 80
Table 24: Amount of AEITC Received Per Invalid Form W-2, Number and
Percentage, Tax Years 2002 through 2004 81
Table 25: Number of Individuals Receiving a Form(s) W-2, Total Number of
Form(s) W-2 Reporting AEITC, Total Associated and Average Dollars Received
Relative to Number of Form(s) W-2 Reporting AEITC for AEITC Recipients
with an Invalid SSN on the Form(s) W-2, Tax Year 2002 82
Table 26: Number of Individuals Receiving a Form(s) W-2, Total Number of
Form(s) W-2 Reporting AEITC, Total Associated and Average Dollars
Received, Relative to Number of Form(s) W-2 Reporting AEITC for AEITC
Recipients with an Invalid SSN on the Form(s) W-2, Tax Year 2003 83
Table 27: Number of Individuals Receiving a Form(s) W-2, Total Number of
Form(s) W-2 Reporting AEITC, Total Associated and Average Dollars
Received, Relative to Number of Form(s) W-2 Reporting AEITC for AEITC
Recipients with an Invalid SSN on the Form(s) W-2, Tax Year 2004 84

Figures

Figure 1: AEITC Process for Employees and Employers 7
Figure 2: Number of EITC Recipients, EITC Recipients Potentially Eligible
for the AEITC, AEITC Recipients, and Total Dollars They Received, Tax
Years 2002 through 2004 10
Figure 3: Percentage of AEITC Recipients Compliant and Noncompliant with
at Least One AEITC Requirement, Tax Years 2002 through 2004 15
Figure 4: Submission Processing Procedures for Paper and Electronic Tax
Returns Reporting AEITC 22
Figure 5: AUR Program Procedures to Resolve Discrepancies between AEITC
Amounts on Forms W-2 and Tax Returns 26
Figure 6: Average Amount of AEITC Received by Consecutive and AEITC
Recipients, Tax Years 2002 through 2004 61
Figure 7: Geographic Location of AEITC Recipients in the Valid
Subpopulation, Tax Years 2002 through 2004 74
Figure 8: Geographic Location of AEITC Recipients in the Invalid Name
Subpopulation, Tax Years 2002 through 2004 87
Figure 9: Geographic Location of AEITC Recipients in the Invalid Number
Subpopulation, Tax Years 2002 through 2004 89
Figure 10: Number of Forms W-2 Received with Amounts above the Yearly
AEITC Maximum, Tax Years 2002 through 2004 90
Figure 11: AEITC Dollars Reported on Form W-2s above the Yearly AEITC
Maximum, Tax Years 2002-2004 91

Abbreviations

AEITC Advance Earned Income Tax Credit
AGI Adjusted Gross Income
AUR Automated Underreporter
DM-1 Data Master File
EITC Earned Income Tax Credit
ERS Error Resolution System
IRS Internal Revenue Service
OBRA '93 Omnibus Budget Reconciliation Act of 1993
SSA Social Security Administration
SSN Social Security number
SSNVS Social Security Number Verification System
TIN Taxpayer Identification Number

This is a work of the U.S. government and is not subject to copyright
protection in the United States. It may be reproduced and distributed in
its entirety without further permission from GAO. However, because this
work may contain copyrighted images or other material, permission from the
copyright holder may be necessary if you wish to reproduce this material
separately.

United States Government Accountability Office
Washington, DC 20548

August 10, 2007

The Honorable Charles B. Rangel
Chairman
The Honorable Max Baucus
Vice Chairman
Joint Committee on Taxation

Beginning in 1979, eligible taxpayers could elect to receive the Earned
Income Tax Credit (EITC), a refundable credit available to low income
workers, as an advance payment--an option currently found in only one
other federal tax credit.^1 Individuals who elect this option receive a
portion of the EITC from their employer throughout the year with their
regular pay, instead of receiving only a lump sum refund or tax credit
when filing their year end federal income tax returns.^2 The Advance
Earned Income Tax Credit (AEITC) provides employees an immediate benefit
from the credit, which may help them meet their daily expenses by
increasing their take-home pay.

Little has been known about the individuals who receive the AEITC, the
employers who pay it, and compliance with its requirements. Neither the
Internal Revenue Service (IRS) nor the employer is required to confirm the
eligibility of those who elect the AEITC before they receive it. Thus, the
extent of compliance problems related to the AEITC may depend on factors
such as how knowledgeable individuals applying for the AEITC are about
eligibility requirements and how well the IRS addresses various kinds of
noncompliance that the agency may identify only after AEITC payments have
been made.

Limited research has been conducted on the AEITC since we last examined it
in the early 1990s. In 1992, we reported that the percentage of
individuals receiving the EITC in 1989 who also received the AEITC was low
and that many eligible workers were not aware of it.^3 We estimated that
almost half of those who received the AEITC in 1989 and filed a tax return
did not report receiving it and that 45 percent of people whom IRS records
show may have received the AEITC never filed a tax return, meaning that
IRS would have no way of determining their eligibility. We made six
recommendations to the Commissioner of Internal Revenue designed to
increase awareness of the AEITC among eligible individuals and improve
compliance among those who receive it. Appendix I discusses IRS's
implementation of the recommendations from our 1992 AEITC report.

^1 The other credit with an advance option is the Health Coverage Tax
Credit, which pays 65 percent of qualified health plan premiums for
eligible trade-affected workers and certain Pension Benefit Guaranty
Corporation benefit recipients.

^2 In tax year 2004, about 22 million individuals received about $40
billion in EITC payments from the IRS. Any portion of the EITC that an
individual did not obtain as an advance can be obtained when he or she
files a tax return, assuming qualifications are maintained.

To provide current information on the AEITC, the Committee asked us to
answer the following questions: (1) How many individuals received the
AEITC compared with the EITC and how much did they receive in tax years
2002 through 2004? What actions, if any, have been taken to increase use
since 1992 and what is the potential for significant increases in the
future? (2) What is the extent of noncompliance, if any, associated with
the AEITC? (3) How well do IRS's procedures address any areas of
noncompliance? In addition, the Committee asked us to provide basic
demographic characteristics of AEITC recipients and their employers and to
identify legislative and administrative changes made since our 1992
report, some of which were discussed by IRS in its response to that
report's recommendations. The information is presented in appendixes I and
III through V.

To answer these questions, we obtained a data file from IRS containing all
Form W-2 "Wage and Tax Statements" that reported AEITC in tax years 1999
through 2004. We performed data reliability tests on this file to
determine whether the data were sufficiently reliable for our intended
purposes and determined that they were. This process involved, among other
things, checking the validity of the Social Security number (SSN) on the
Form W-2, determining whether the AEITC amount was above the yearly
maximum limit, and creating subpopulations in which to conduct our
analyses. Using the Form W-2 data and data from other sources, such as
IRS's Individual Return Transaction File that contains tax return
information, we compiled relevant demographic characteristics data and
compared AEITC data with EITC data when possible. We also reviewed
administrative and legislative changes as well as relevant reports and
considered these along with our data analyses. To examine actions intended
to increase use and determine the potential for increases in future AEITC
use, we reviewed various reports and literature and interviewed IRS
officials and individuals we determined to be experts. We also used the
Form W-2 and other data to identify areas of noncompliance by comparing
the data to the AEITC requirements. In addition, we examined IRS's
procedures for processing individual income tax returns with AEITC and
interviewed IRS officials to determine how IRS addressed noncompliance.
Finally, we explored possible remedies for the noncompliance we identified
through literature searches and interviews with various Department of the
Treasury, IRS, and Social Security Administration (SSA) officials. It was
not within the scope of our work to fully evaluate the potential cost and
benefits of these options for reducing noncompliance. We conducted our
work in accordance with generally accepted government auditing standards
from December 2005 through July 2007. Appendix II provides more detail on
the scope and methodology we used to conduct our work.

^3 GAO, Earned Income Tax Credit: Advance Option Is Not Widely Known or
Understood by the Public, [33]GAO/GGD-92-26 (Washington, D.C.: Feb. 19,
1992).

Results in Brief

AEITC use and the amount paid out by employers was low in tax years 2002
through 2004--only about 3 percent of the EITC recipients potentially
eligible for the advance received it and about half of the recipients
received $100 or less per year. In total, approximately 514,000
individuals received about $146 million of AEITC in each of the 3 years.
Several federal efforts, including two legislative and several
administrative changes, have been aimed at increasing AEITC use in the
last approximately 15 years. For example, the Omnibus Budget
Reconciliation Act of 1993 (OBRA '93), directed IRS to send notices to
taxpayers during 2 tax years informing them that they were potentially
eligible for the advance. Similarly, IRS made changes to its forms and
increased speaker seminars to inform the public about the AEITC. Despite
these efforts, use has not substantially increased and, for several
reasons, it may be difficult to significantly increase it in the future.
For example, IRS officials, other experts, and our prior work suggest that
individuals often do not elect the AEITC because they prefer receiving the
entire EITC as a lump sum when filing their tax return. An IRS-funded
study revealed that low use is also attributed to concerns that
prospective recipients would receive more in advance than they would
ultimately be entitled to, thus owing the difference when filing their tax
return.

As many as 80 percent of recipients did not comply with at least one AEITC
requirement, while some individuals were noncompliant with more than one
requirement. Approximately 20 percent (more than 100,000) of all AEITC
recipients each year may not have been eligible for the advance because
they did not have a valid SSN, that is, a SSN that has been assigned by
the SSA or a SSN/name combination that could be matched to SSA records.
These individuals received $37 million to $39 million in each tax year
from 2002 through 2004. Almost 40 percent (about 200,000 recipients) did
not file the required tax return; they received between $42 million and
$50 million of AEITC for each of these years. When individuals do not file
a return, IRS cannot readily identify whether they were eligible for AEITC
and whether they owed IRS any of the amounts they received. Conversely, by
not filing a tax return, some individuals did not receive additional EITC
monies that they could only receive had they filed. Of the 60 percent of
AEITC recipients who did file a return, about 300,000 individuals in each
of the 3 years, two-thirds misreported the amount they received--the
majority did not report receiving any AEITC. As a result, these AEITC
recipients who claimed the EITC when filing their return collectively
received an excess amount of EITC of about $64 million for the 3 years.

Although IRS has processing and enforcement procedures in place to monitor
the AEITC, the agency has limited effectiveness in addressing AEITC
noncompliance, in large part because it targets most of its enforcement
efforts at more widely used programs that involve larger amounts of money.
For example, Automated Underreporter (AUR) staff worked only a fraction of
the AEITC underreporting cases because resource constraints and criteria
limited case selection. Moreover, AUR and other IRS enforcement programs
use revenue collection potential as a primary criterion in identifying
cases on which to work, and AEITC cases generally have low revenue
potential. Although all AEITC nonfiler cases are eligible to be worked on
by various enforcement programs, IRS officials said they work on only a
limited number for similar reasons.

One means to improve AEITC compliance might be to send "soft notices."
Recent IRS tests have shown that the agency has successfully reduced
subsequent noncompliance in situations that involve relatively small
amounts of money by sending soft notices that ask taxpayers to voluntarily
fix their misreporting by filing an amended return or not repeating the
action in the next year. However, high annual turnover in the AEITC
population could undermine this approach. Another means to improve AEITC
compliance would be to partially verify individuals' eligibility before
they receive the advance by verifying their SSNs. IRS operates the
Taxpayer Identification Number (TIN) Matching service and SSA maintains
the Social Security Number Verification System (SSNVS), both of which are
open to registered public users for the purpose of determining the
validity of an SSN. Both have unique advantages and disadvantages that
would come into play if employers were required to use them to verify the
SSNs of individuals seeking AEITC. SSA emphasized that verifying
eligibility for the AEITC is most appropriate for IRS because it is a tax
administration issue and therefore outside the scope of SSA's mission.
Separately, IRS could verify individuals' SSNs and take other enforcement
steps if it had a database of AEITC applicants. However, IRS cannot create
such as database because it does not require employers to send copies of
the Form W-5, "Earned Income Credit Advance Payment Certificate," to them
when an employee requests receipt of the AEITC. Despite some benefits, IRS
officials questioned the effectiveness of such a requirement due to (1)
the low anticipated return on investment of AEITC compared with other
areas with noncompliance problems and (2) their experience with the prior
Questionable W-4 program, which required employers to submit information
to IRS on taxpayers claiming more than 10 withholding allowances or
exemptions, that showed employers did not adhere to similar requirements
imposed by the agency.

We recommend that the Acting Commissioner of Internal Revenue analyze
whether it could cost effectively and significantly reduce AEITC
noncompliance by sending soft notices to potentially noncompliant AEITC
recipients, requiring employers to verify the SSN of employees seeking the
AEITC, or requiring employers to submit Forms W-5 to IRS to create an
AEITC database. To better identify the costs and implementation issues as
well as the likelihood for these or other options to reduce AEITC
noncompliance, where practical, the Acting Commissioner of Internal
Revenue should test these options to make a more fully informed judgment
about whether any would be worthwhile. If the Acting Commissioner of
Internal Revenue determines that none of these options would be cost
effective and that no other remedies are viable, then the Treasury
Secretary should inform the Congress of this and provide Treasury's
opinion about whether the AEITC should be retained.

In providing written comments on a draft of this report (see app. VI), the
Acting Commissioner of Internal Revenue agreed with our recommendation and
outlined the actions IRS would take to address that recommendation,
including conducting further analyses and possible testing of proposed
options for reducing AEITC noncompliance. He also stated that IRS will
conduct its cost-benefit analyses in conjunction with a congressional
requirement to study the impact of expanding eligibility of the AEITC to
all EITC recipients.^4 We also provided a draft of this report to the
Department of the Treasury and SSA and incorporated technical comments
where appropriate. SSA emphasized that verifying eligibility for the AEITC
is most appropriate for IRS because it is a tax administration issue and
outside the scope of SSA's mission.

Background

The EITC was enacted in 1975 and was originally intended to offset the
burden of Social Security taxes and provide a work incentive for
low-income taxpayers.^5 It is a refundable federal income tax credit,
meaning that qualifying working taxpayers may receive a refund greater
than the amount of income tax they paid for the year. For tax year 2006,
the maximum amount of EITC a taxpayer could receive was $4,536.

Beginning in 1979, individuals could elect to receive the EITC in advance
payments from their employer during the year along with their regular
pay.^6 One purpose of the advance payment is to provide employees with an
immediate reward for their work effort rather than forcing them to
postpone receiving the credit until they file their tax returns. To get
the credit, at any time during the year, an employee would complete the
Form W-5 and provide it to his or her employer.^7

Upon receiving a completed Form W-5, the employer calculates the amount of
the AEITC payment to include with the employee's pay by considering (1)
the employee's wages, (2) whether the employee is married or single, and
(3) if married, whether the employee's spouse has a Form W-5 in effect
with an employer. The AEITC payment to the employee is considered to be
equivalent to the employer making a payment to IRS for employee income tax
withholding and employee and employer Social Security and Medicare tax.
When employers file their quarterly tax returns, they show the total
payments made to employees on the AEITC payment line on Form 941,
"Employer's Quarterly Federal Tax Return." This amount is then subtracted
from the total amount of tax the employer owes.

^4 U.S. Troop Readiness, Veteran's Care, Katrina Recovery, and Iraq
Accountability Appropriations Act of 2007, Pub. L. No. 110-28.

^5 26 U.S.C. Sec. 32.

^6 Pub. L. No.103-66 (1993). The Omnibus Budget Reconciliation Act of 1993
limited the amount of advance payments that eligible individuals can
receive to 60 percent of the EITC available with one qualifying child.

^7 The Form W-5, completed by the employee, expires at the end of each
calendar year. Therefore, employees must resubmit the form each year if
they want to receive the AEITC.

At the end of the calendar year, the employer indicates the total AEITC
payments the employee received on the employee's Form W-2. Employees are
then required to report this amount either on their Form 1040 or Form
1040A tax return. Assuming the employee qualified for the EITC, the AEITC
amount received should be reported on the tax return as other taxes,
which, in effect, subtracts the amount received from the total amount of
any EITC. If the employee did not qualify for the EITC, he or she is still
required to file a tax return, regardless of income level, and the AEITC
amount paid is added to any taxes owed. Figure 1 illustrates this process
and notes the major forms involved.

Figure 1: AEITC Process for Employees and Employers

An individual must meet certain requirements to qualify for the AEITC.
Specifically, an individual must expect to (1) be able to claim the EITC
for the current year (EITC requirements for 2006 are shown in table 1),
(2) have at least one qualifying child, and (3) have earned income and
adjusted gross income below a certain amount for that year. There are
additional requirements that an individual may only have one Form W-5 in
effect at a time and that he or she informs their employer if their spouse
also has a Form W-5 in effect.^8 AEITC recipients can receive up to 60
percent of the EITC benefits for one qualifying child. The maximum AEITC
amount for 2006 was $1,648.

Table 1: EITC Requirements for 2006 That All AEITC Recipients Must Also
Expect to Meet

                                      Second, an individual must                           
                                      meet all the rules in one of   Third, an individual  
First, an individual must meet all    these columns, whichever       must meet the rules   
the rules in this column.             applies.                       in this column.       
                                                       Rules if do                         
                                                       not have a                          
                                      Rules if have a  qualifying    Figuring and claiming 
         Rules for everyone           qualifying child child         the EITC              
1. Adjusted gross     2. Must have    8. Child must    11. Must      15. Earned income  
income (AGI) must     a valid         meet the         be at         must be less than: 
be less than:         Social          relationship,    least age                        
                         Security        age, and         25 but                o $36,348  
           o $36,348     number.         residency        under age             ($38,348   
           ($38,348      3. Filing       tests.           65.                   for        
           for           status          9. Qualifying    12. Cannot            married    
           married       cannot be       child cannot     be the                filing     
           filing        "Married        be used by       dependent             jointly)   
           jointly)      filing          more than one    of another            if more    
           if more       separately."    person to        person.               than one   
           than one      4. Must be a    claim the        13. Cannot            qualifying 
           qualifying    U.S. citizen    EIC.             be a                  child      
           child.        or resident     10. Cannot be    qualifying                       
                         alien all       a qualifying     child of              o $32,001  
           o $32,001     year.           child of         another               ($34,001   
           ($34,001      5. Cannot       another          person.               for        
           for           file Form       person.          14. Must              married    
           married       2555 or Form                     have lived            filing     
           filing        2555-EZ                          in the                jointly)   
           jointly)      (relating to                     United                if one     
           if one        foreign                          States                qualifying 
           qualifying    earned                           more than             child, or  
           child, or     income).                         half of                          
                         6.                               the year.             o $12,120  
           o $12,120     Investment                                             ($14,120   
           ($14,120      income must                                            for        
           for           be $2,800 or                                           married    
           married       less.                                                  filing     
           filing        7. Must have                                           jointly)   
           jointly)      earned                                                 if no      
           if no         income.                                                qualifying 
           qualifying                                                           child.     
           child.                                                                          

Source: Excerpt of IRS Publication 596, Earned Income Credit.

^8 When both spouses have a Form W-5 in effect, the amount each person can
receive is reduced to below 60 percent of the EITC maximum.

A change in an individual's personal circumstances after submitting the
Form W-5 could affect their eligibility for both the EITC and the AEITC.
For example, an individual who received the AEITC, separated from his or
her spouse during the year, and used the married filing separately filing
status would not be eligible for the AEITC (or EITC). In such cases, when
the individual files their tax return and reports the AEITC amount
received, the amount would be added to any taxes due or subtracted from
any refund.

AEITC Use Low and May Be Difficult to Increase

About 3 percent of the EITC recipients potentially eligible for the
advance, about 514,000 individuals on average, elected it in each year,
tax years 2002 through 2004, according to data employers reported on the
Form W-2. As shown in figure 2, about 21 million taxpayers received the
EITC each year and of these recipients, about 17 million were eligible for
the AEITC.^9

9 Individuals who receive the EITC without a qualifying child are not
eligible for the AEITC.

Figure 2: Number of EITC Recipients, EITC Recipients Potentially Eligible
for the AEITC, AEITC Recipients, and Total Dollars They Received, Tax
Years 2002 through 2004

Note: This figure includes the valid, invalid name, and invalid number
subpopulations. (see app. II).

In total, AEITC recipients received an average of about $146 million in
AEITC for each tax year from 2002 through 2004. Yet, those who elected it
often received relatively few dollars from their employers. As table 2
indicates, about half of all individuals who got AEITC received $100 or
less each year and about 75 percent received $500 or less. Even at $100
per year, this equates to about $8 a month, or $4 every 2 weeks, and even
at $500 per year, this equates to about $42 a month or $19 every 2 weeks.
The amounts most individuals received were significantly less than the
yearly maximum, and were consistent for the 3 years we reviewed.

Table 2: Percentage of AEITC Recipients Receiving an Amount of AEITC, Tax
Years 2002 through 2004

                            Percentage of AEITC recipients
Amount of AEITC received       2002      2003      2004 
$1--$100                         50        50        48 
$101--$250                       17        17        17 
$251--$500                       13        13        13 
$501--$750                        7         7         8 
$751--$1,000                      5         5         5 
$1,001--yearly maximum            7         8         8 

Source: GAO analysis of IRS data.

Note: The AEITC yearly maximum was $1,503, $1,528, and $1,563 for tax
years 2002, 2003, and 2004, respectively. This table includes the valid,
invalid name, and invalid number subpopulations.

The number of individuals who have elected the AEITC has remained low for
many years. For example, between 1990 and 1997, AEITC use never exceeded 2
percent of qualifying EITC recipients. IRS calculated use based on tax
returns showing receipt of the AEITC divided by the EITC population that
reported at least one qualifying child. Using this same methodology, use
for tax years 2002 through 2004 was relatively the same at an average of
0.8 percent. Our figures for 2002 through 2004 are higher than these prior
AEITC figures because our figures are based upon the Forms W-2 that
reported AEITC (see fig. 2). Historically, AEITC use has been based upon
the number of individual federal tax returns that reported an amount on
the AEITC line. The historic method excludes individuals who did not file
a federal tax return and individuals who filed a federal tax return but
did not report the AEITC.

Additional demographic data about individuals who elected the AEITC are
included in appendix III. These additional data represent new analysis
that has not been previously available, including each recipient's filing
method, age, and gender, and each employer's size.

AEITC Use Has Not Increased Significantly Despite Several Targeted Efforts

Use of the AEITC has remained low for many years despite several targeted
efforts to increase it. There have been several federal efforts targeted
to increase AEITC use over the last approximately 15 years, including both
legislative and administrative changes. (A full description of these
changes is included in app. IV.) One significant piece of legislation was
OBRA '93, which involved the AEITC in two ways: it (1) reduced the maximum
amount of EITC an individual could receive in advance and (2) required IRS
to conduct outreach directly to potentially eligible AEITC recipients.
First, OBRA '93 reduced the AEITC maximum from 100 percent to 60 percent
of the maximum credit available to a taxpayer with one qualifying child.
This change was made to improve compliance and lessen concerns that
recipients would owe the difference when filing their federal tax return,
which was thought to discourage AEITC use.

Second, OBRA '93 directed IRS to send notices to taxpayers who were likely
to be eligible for the AEITC for 2 years and directed the Secretary of the
Treasury to study the effect of the notice program on AEITC use. Only some
information is available about the first notice mailing, which occurred in
1994.^10 For the first notice mailing, IRS mailed Publication 1235,
"Advance Earned Income Tax Credit Brochure," and the Form W-5 to about
13.5 million taxpayers who were potential AEITC candidates during tax year
1993 informing them about the AEITC. AEITC use increased about 1 percent
following this effort; however, because other outreach efforts were
ongoing during this time, IRS could not conclude that the increase was
attributable to the notice or any other effort specifically.

In 1997, IRS mailed the notice to about 6 million taxpayers who claimed
the EITC in tax year 1996, but did not report receiving AEITC on their
federal tax return. With this second mailing, IRS created two groups, a
test group of about 60,000 taxpayers that received the notice and a
control group of about 60,000 taxpayers with similar characteristics who
did not. Results from the IRS report indicated that about 1.27 percent
(771 taxpayers) of the tax returns in the test group reported the AEITC
compared to 0.51 percent (309 taxpayers) of returns in the control group.
The summary report concluded that further efforts to increase AEITC use
substantially are unlikely to succeed.^11 Further, the study recommended
that notification of EITC recipients about the advance not be repeated.

Some of IRS's administrative changes include outreach to specific groups
and changes to publications. For example, after our 1992 report, the White
House, the Treasury Department, and IRS conducted extensive EITC and AEITC
outreach efforts, including a 1993 announcement of the AEITC by President
Clinton. Other outreach efforts included IRS contacts with charitable,
social welfare, and minority groups to encourage awareness of the EITC and
AEITC among their memberships. IRS also contacted a number of employer
organizations to encourage them to publicize the AEITC with their
memberships. IRS also made changes to its forms, developed print and video
products, and increased speaker seminars to inform the public about the
AEITC. For example, IRS developed publicity materials, such as grocery bag
and milk carton art, brochures, and posters; provided information in the
Small Business Taxpayer Education Program guide; and increased outreach
speaker seminar efforts.^12

^10For the first notice mailing, specific information on how the population
was selected, for example, is not known.

^11IRS, National Research Office, Advance Earned Income Tax Credit 1994
and 1997 Notice Study, A Report to Congress (Washington, D.C.: August
1999).

Presently, IRS continues to conduct outreach about the AEITC as part of
its EITC outreach efforts. IRS focuses its outreach to large employer
organizations or to specific large employers, which then promote the AEITC
to employers or employees.

Future AEITC Growth Is Unlikely

Increasing AEITC use in the future is unlikely for several reasons, but
perhaps primarily because of potential recipients' preferences and high
AEITC turnover. Interviews with IRS officials, other experts, and our
prior AEITC work suggest that those eligible for the AEITC prefer
receiving the EITC in a lump sum after they file their federal tax return
instead of receiving relatively small portions spread throughout the
year.^13 Another reason is that, despite the reduction in the yearly AEITC
maximum to 60 percent of the maximum credit available to a taxpayer with
one qualifying child, results from an IRS-funded study using focus groups
of EITC participants and interviews we conducted with experts indicated
that potential recipients continue to have concerns that they would
receive more AEITC than they were ultimately entitled to and that they
would owe the difference when filing their federal tax return.

In addition, AEITC growth is adversely affected by individuals who elect
the AEITC and fail to elect it again, i.e., turnover. As table 3
indicates, more than half of the individuals who elected the AEITC did so
for the first time since 1999 in either tax year 2002, 2003, or 2004. With
the overall AEITC use remaining relatively constant over the 3 years, this
indicates the large number of yearly first-time recipients was almost
equally offset by existing recipients forgoing the AEITC in a following
year.^14 The percentage of AEITC first-time recipients is much higher than
the percentage of first-time EITC recipients, which is slightly less than
one-third. Also, of the individuals who elected the AEITC for the first
time in 2002 or 2003, about 28 percent elected it again in the following
year, 2003 or 2004, respectively, while the remainder did not elect it
again in the subsequent year. Conversely, only about 98,000 (9 percent)
individuals elected the AEITC consecutively in all 3 years, 2002 through
2004.

^12The Small Business Taxpayer Education Program provides help to those
who are starting or already have a small business and need information in
such areas as taxes, recordkeeping, accounting practices, and completing
federal business and employment tax returns. Much of the assistance is
free.

^13We identified other studies with limited populations that also found
taxpayers preferred receiving a lump sum instead of smaller periodic AEITC
payments; for example, Jennifer Romich and Thomas Weisner, "How Families
View and Use the EITC: Advance Payment vs. Lump Sum Delivery," National
Tax Journal (December 2000).

Table 3: Frequency of Individuals Electing the EITC and AEITC, Tax Years
2002 through 2004

                                    Tax year 2002 Tax year 2003 Tax year 2004 
EITC recipients                     20,918,823    21,431,377    21,721,218 
First-time EITC recipients           6,077,104     5,313,089     4,842,922 
(percentage) ^a, b                                                         
                                            (29%)         (25%)         (22%) 
AEITC recipients                       531,799       503,382       507,957 
First-time AEITC recipients            271,600       259,579       259,282 
(percentage) ^a, c                                                         
                                            (51%)         (52%)         (51%) 
First-time AEITC recipients who        198,096       187,871           N/A 
did not elect the advance in the                                           
subsequent year (percentage) ^d          (73%)         (72%)               
First-time AEITC recipients who         73,504        71,708           N/A 
elected the advance in the                                                 
subsequent year (percentage) ^d          (27%)         (28%)               
Number of individuals who                         97,998
consecutively received the AEITC               
(percentage) ^e                                    (9%)

Source: GAO analysis of IRS data.

Note: N/A means not applicable. Tax year 2004 data are not applicable
because we did not look at any subsequent years. All AEITC figures include
the valid, invalid name, and invalid number subpopulations.

aFirst-time EITC recipients are based upon how often a primary or
secondary taxpayer filed a federal tax return receiving EITC for the first
year since 1999. First-time AEITC recipients are based upon the first year
an individual (defined by unique SSN) received AEITC since 1999. IRS
officials agreed this approach is a reasonable method to compare AEITC and
EITC first-time use.

bPercentage is out of EITC recipients.

cPercentage is out of AEITC recipients.

dPercentage is out of first-time AEITC recipients.

eNumber of individuals who consecutively received the AEITC are
individuals who received the advance each year in tax years 2002 through
2004. Percentage is of individuals who received the AEITC each year in tax
years 2002 through 2004 out of all recipients who received it at least
once during these same years (1,133,908 individuals). The percentage who
received it twice during these years is 19 percent (213,234) and once is
73 percent (822,676). Percentages may not add due to rounding.

^14 One reason an AEITC recipient may not elect the advance in a
subsequent year is because he or she no longer meets the eligibility
requirements.

High Noncompliance Exists with AEITC Requirements

Overall, as many as 80 percent of all AEITC recipients did not comply with
or made errors involving one of the three AEITC requirements that we
reviewed, and they received about $282 million when the 3 years, 2002
through 2004, are aggregated. Some taxpayers were noncompliant with more
than one requirement. Those requirements are having a valid SSN, filing a
federal tax return, and reporting the proper amount of AEITC received on
the tax return (see fig. 3).

Figure 3: Percentage of AEITC Recipients Compliant and Noncompliant with
at Least One AEITC Requirement, Tax Years 2002 through 2004

Note: The pie chart includes the valid, invalid name, and invalid number
subpopulations.

aThis includes the invalid name and invalid number subpopulations.

bThis includes the valid, invalid name, and invalid number subpopulations.

cThis is of individuals who filed a federal tax return in the valid,
invalid name, and invalid number subpopulations.

Specifically, in tax year 2002, we found that individuals were
noncompliant with at least one of three requirements we reviewed 79
percent of the time and they received about $93 million of AEITC. For tax
year 2003 and 2004, individuals were noncompliant with at least one
requirement 78 percent (about $91 million) and 79 percent (about $98
million) of the time, respectively.

Some of the noncompliance we identified could have resulted from IRS or
employer clerical errors or improper reporting by the taxpayer. Therefore,
some of the errors may be correctible or were corrected by filing an
amended return. IRS cannot readily identify the number of amended returns
specifically associated with the AEITC because such returns combine
several credits onto one line. The explanation attached to the amended
return would provide details on the change and any analysis of the
explanation would be a manual process.

About 20 Percent of AEITC Recipients Had an Invalid SSN

AEITC recipients are required to provide their employer with a valid SSN
for the Form W-2.^15 As table 4 illustrates, about 20 percent (more than
100,000) of AEITC recipients each year may not have been eligible for the
advance because they did not have a valid SSN on their Form W-2.
Collectively, these individuals received between $37 million and $39
million in AEITC each year. The data are consistent over the 3 years
reviewed.^16

^15 AEITC recipients must expect to be able to claim the EITC, which
requires that taxpayers and their spouses, if filing a joint return, have
valid SSNs issued by SSA to all U.S. citizens or individuals eligible to
work.

^16If any portion of the invalid SSNs is attributable to aliens not
currently authorized to work in the United States, the number of invalid
SSNs could be affected if the Congress changes policies regarding such
aliens. See Treasury Inspector General for Tax Administration, The
Internal Revenue Service's Individual Taxpayer Identification Number
Creates Significant Challenges for Tax Administration, 2004-30-023
(January 2004), for more information about aliens' use of invalid SSNs.

Table 4: AEITC Recipients and Recipients with Invalid SSN on Form W-2,
Individuals and Dollars, Tax Years 2002 through 2004

                     Tax year 2002       Tax year 2003       Tax year 2004
                  Individuals Dollars Individuals Dollars Individuals Dollars 
AEITC              531,799    $145     503,382    $143     507,957    $150 
recipients                 million             million             million 
AEITC              113,025     $37     104,856     $37     109,622     $39 
recipients                 million             million             million 
with an              (21%)               (21%)               (22%)         
invalid SSN on               (26%)               (26%)               (26%) 
Form W-2                                                                   
(percentage)^a                                                             

Source: GAO analysis of IRS data.

aPercentage is out of AEITC recipients. This analysis includes the invalid
name and invalid number subpopulations.

For purposes of this report, invalid SSNs include instances where the SSN
did not match SSA's records (i.e., the number was never assigned by SSA)
and the SSN/name combinations reported on the Form W-2 did not match SSA
records.^17 Some of these individuals were likely eligible for the AEITC.
For example, a name/SSN mismatch could include instances when a woman who
receives the AEITC marries and changes her name with her employer but not
with SSA. This could result in the employer issuing a Form W-2 in the new
name, but IRS and SSA only identifying her by the former name.^18

Individuals who file a federal tax return are required to include a valid
SSN on their return.^19 An individual who provides an invalid SSN on the
tax return is not compliant in meeting AEITC requirements and may also
violate the Social Security Act.^20 Such an individual is also required to
provide a valid SSN to their employer for income tax withholding purposes
and for purposes of certifying eligibility for the AEITC and could be
subject to a penalty for failure to do so.^21 Further, if the individual
does not file a valid SSN, IRS is unable to assess the recipient's federal
tax liability and SSA cannot credit the recipient for money withheld for
Social Security purposes.^22 Also, because taxpayers often do not report
receipt of the advance on their tax return, IRS cannot determine whether
the taxpayer owes money or deserves a refund.

^17There could be some instances where the SSN reported on the Form W-2
does not match the SSN the taxpayer used on the tax return. Even still,
IRS uses the SSN on the Form W-2 and other third party forms for document
matching to identify noncompliance such as wage underreporting.

^18 Of the more than 100,000 AEITC recipients with an invalid SSN, most
involved a SSN that did not match the individual's name on the Form W-2.

^19 IRC S 6109(a)(1).

^20 42 U.S.C. S 408.

^21 IRC S6723.

Almost 40 Percent of AEITC Recipients Did Not File the Required Federal Tax
Return

All AEITC recipients are required to file a federal tax return, regardless
of the amount of their income, which is generally the primary basis for
determining whether a return is required to be filed. Table 5 shows that
between 36 and 40 percent, about 200,000 AEITC recipients, did not file a
required federal tax return each year. Collectively, these individuals
received between $42 million and $50 million of AEITC benefits.^23

Table 5: AEITC Recipients and Recipients Who Did Not File a Tax Return,
Individuals and Dollars, Tax Years 2002 through 2004

                     Tax year 2002       Tax year 2003       Tax year 2004
                  Individuals Dollars Individuals Dollars Individuals Dollars 
AEITC              531,799    $145     503,382    $143     507,957    $150 
recipients                 million             million             million 
AEITC              192,857     $42     186,423     $44     200,706     $50 
recipients who             million             million             million 
did not file a       (36%)               (37%)               (40%)         
federal tax                  (29%)               (31%)               (33%) 
return                                                                     
(percentage)^a                                                             

Source: GAO analysis of IRS data.

aPercentage is of AEITC recipients. This analysis includes the valid,
invalid name, and invalid number subpopulations.

About 56,000 to 60,000 of the about 200,000 individuals who did not file
the required tax return (about 30 percent) had an invalid SSN on the Form
W-2 each year, as shown in table 6. Having a valid SSN is another AEITC
requirement, discussed previously, which means these individuals were
noncompliant or made an error with at least two AEITC requirements.

^22 IRS can file a substitute return for people who do not voluntarily
file. IRS assesses a tax liability and notifies the taxpayer of any tax
due based on the substitute tax return, which is created from available
Form W-2 and other information.

^23 There could be instances where our methodology did not detect a filed
tax return; for example, when a taxpayer's SSN on the Form W-2 might have
been incorrect. See our scope and methodology (app. II).

Table 6: AEITC Recipients, Recipients Who Did Not File a Federal Tax
Return, and Recipients Who Did Not File a Tax Return and Had an Invalid
SSN on the Form W-2, Individuals and Dollars, Tax Years 2002 through 2004

                     Tax year 2002       Tax year 2003       Tax year 2004
                  Individuals Dollars Individuals Dollars Individuals Dollars 
AEITC              531,799    $145     503,382    $143     507,957    $150 
recipients                 million             million             million 
AEITC              192,857     $42     186,423     $44     200,706     $50 
recipients who             million             million             million 
did not file a       (36%)               (37%)               (40%)         
federal tax                  (29%)               (31%)               (33%) 
return                                                                     
(percentage)                                                               
^a                                                                         
AEITC               59,607     $20      55,601     $20      58,370     $21 
recipients who             million             million             million 
did not file a       (31%)               (30%)               (29%)         
federal tax                  (48%)               (45%)               (42%) 
return and had                                                             
an invalid SSN                                                             
on Form W-2                                                                
(percentage)                                                               
^b                                                                         

Source: GAO analysis of IRS data.

aPercentage is of AEITC recipients. This analysis includes the valid,
invalid name, and invalid number subpopulations.

bPercentage is of AEITC recipients who did not file a federal tax return.
An unknown portion of these recipients may have corrected the SSN error
and subsequently filed a tax return. This analysis includes the invalid
name and invalid number subpopulations.

There are several reasons why a significant number of AEITC recipients
might not have filed a federal tax return. For example, depending on their
filing status, age, and type of income they receive, recipients may not
have had a filing responsibility other than for the AEITC and they may not
have remembered or understood they must file a return. In addition, AEITC
recipients may not have filed because they were not initially eligible or
they became ineligible for the AEITC because of a change in their personal
circumstances, and filing would require them to pay back the AEITC they
received.

When individuals are required to file a federal tax return and do not, IRS
cannot readily identify whether the individual was eligible for the
advance or whether they owed IRS any of the amounts they received.
Conversely, by not filing a federal tax return, some individuals did not
receive additional EITC monies that they could only receive had they
filed.

About Two-Thirds of AEITC Recipients Who Filed a Federal Tax Return Misreported
the Amount of AEITC They Received

All AEITC recipients are required to report on their federal tax return
the amount of AEITC they received according to the Form(s) W-2. Reporting
this amount allows the IRS to determine whether the taxpayer received too
much AEITC, and owes money back to the IRS, or whether the taxpayer is
entitled to additional amounts of the EITC.

Of the approximately 60 percent (about 300,000) AEITC recipients who filed
a federal tax return, two-thirds misreported the amount they received in
tax year 2002 through 2004, as shown in table 7. Misreported means that
the total amount of AEITC reported on the Form W-2 does not match the
AEITC amount reported on the federal tax return. Approximately one-third
of the federal tax returns correctly matched to the Form(s) W-2 AEITC
amount. Of those that misreported, the vast majority did not report
receiving any AEITC.

Table 7: AEITC Matches and Mismatches between Form(s) W-2 and Filed Tax
Returns, Tax Years 2002 through 2004

                     Tax year 2002        Tax year 2003        Tax year 2004
                   Number      Dollars  Number      Dollars  Number      Dollars 
Number of      342,841 $108,189,196 320,588 $101,078,453 309,724 $102,512,732 
individuals                                                                   
who filed                                                                     
federal tax                                                                   
returns^a                                                                     
Number of      112,421  $51,457,254 111,266  $52,246,111 106,980  $52,338,650 
returns that                                                                  
matched the      (33%)        (48%)   (35%)        (52%)   (35%)        (51%) 
Form(s) W-2                                                                   
AEITC amount                                                                  
(percentage)^b                                                                
Number of      230,420  $56,731,942 209,322  $48,832,342 202,744  $50,174,082 
mismatches                                                                    
between          (67%)        (52%)   (65%)        (48%)   (65%)        (49%) 
Form(s) W-2                                                                   
and filed                                                                     
federal tax                                                                   
return                                                                        
(percentage)^b                                                                
Number of      222,691  $53,343,102 202,152  $45,487,033 195,571  $46,806,223 
mismatches due                                                                
to               (97%)        (94%)   (97%)        (93%)   (96%)        (93%) 
nonreporting                                                                  
of AEITC                                                                      
(percentage)^c                                                                
Number of        4,996   $2,349,197   4,779   $2,389,854   4,881   $2,337,603 
mismatches due                                                                
to                (2%)         (4%)    (2%)         (5%)    (2%)         (5%) 
underreporting                                                                
of AEITC                                                                      
(percentage)^c                                                                
Number of        2,733   $1,039,643   2,391     $955,455   2,292   $1,030,256 
mismatches due                                                                
to                (1%)         (2%)    (1%)         (2%)    (1%)         (2%) 
overreporting                                                                 
AEITC                                                                         
(percentage)^c                                                                

Source: GAO analysis of IRS data.

Note: Percentages may not add due to rounding. This table includes
individuals who filed a tax return in the valid and invalid name
subpopulations.

aThe total number of filed federal tax returns is greater than the total
number of AEITC recipients who filed a return because there were some
dependents who received AEITC and were listed on more than one federal tax
return. The total number of AEITC recipients who filed a federal tax
return can be calculated using table 5 by subtracting the number of AEITC
recipients who did not file a federal tax return from the number of AEITC
recipients.

bWe considered a match to be anything plus or minus a dollar on the Form
W-2 in order to allow for taxpayer rounding. We tested the sensitivity of
this result by using a difference between the AEITC reported on the Form
W-2 and the tax return of up to $100 and the results were similar.
Percentage is of number of individuals who filed federal tax returns.

cPercentage is of number of mismatches between Form(s) W-2 and filed
federal tax returns.

Taxpayers may not report the amount of AEITC they received because they
either forget or do not know they are required to do so. Underreporting
can occur when there is a computation error involving multiple Forms W-2,
a taxpayer disagrees with the amount reported on the Forms W-2, or there
is willful noncompliance. IRS officials in the AUR program pursue some
AEITC cases where taxpayers either underreport or do not report receipt of
the AEITC. These procedures are discussed in detail in a following
section.

Due to the high number of mismatches, many of the AEITC recipients who
also claimed the EITC likely received excess benefits. For example, of the
222,691 taxpayers who did not report receipt of the AEITC on their tax
return in tax year 2002, almost half went on to claim the EITC. We
determined that those taxpayers received nearly $22 million in excess
AEITC benefits. For the 3 years, 2002 through 2004, taxpayers received a
total of about $64 million in excess AEITC benefits.

AEITC Recipients Were Noncompliant with Other Program Requirements

Additional noncompliance or errors existed with other program
requirements, such as receiving excess AEITC. Details and demographic
information on the noncompliant individuals are in appendix V.

IRS's AEITC Compliance Procedures Have Limited Effectiveness and Options to
Better Address Noncompliance Have Advantages and Disadvantages

IRS's Submission Processing is responsible for receiving, processing, and
archiving the nation's federal tax returns, payments, and information
returns. In the context of AEITC, Submission Processing attempts to catch
mismatches between both paper and electronic tax returns and Forms W-2
(see fig. 4).

Figure 4: Submission Processing Procedures for Paper and Electronic Tax
Returns Reporting AEITC

aAfter some initials checks, such as comparing the EITC/AEITC line of a
tax return to the Form W-2(s), all paper returns go to Submission
Processing staff, who enter select data from the Form 1040 into an
electronic database.

bEffective January 2, 2007, at our suggestion, Submission Processing
changed its selection criteria to include filing status.

Between 32 percent and 22 percent of tax returns reporting an AEITC amount
were filed on paper between tax years 2002 and 2004, respectively. If a
paper tax return has an entry on either the AEITC or EITC lines on the
Form 1040, Submission Processing tax examiners are required to ensure that
the AEITC amount reported on the return matches the amount in box 9 of the
Form(s) W-2, which records the amount of AEITC paid by the employer to the
employee.^24

If the amounts on the AEITC line of the tax return and Form W-2(s) match,
the tax examiner takes no further action. If the amounts differ (e.g., the
tax return reports a lesser amount than is reflected on the Form(s) W-2)
the examiner is required to adjust the entry on the return to equal the
total AEITC amount from the Form(s) W-2. All paper returns then go to
staff who enter data from the Form 1040 into an electronic database.

After an examiner makes an adjustment, IRS sends a letter to the taxpayer
explaining that an adjustment was made and it was based on the mismatch
between the tax return and the Form(s) W-2. IRS sent 282 and 220 such
letters in tax years 2003 and 2004, respectively.^25 Submission
Processing's role is to ensure that the return amount is consistent with
the Form W-2, not to determine which of the differing numbers accurately
reflects the amount of AEITC actually paid to the employee. If the
taxpayer disagrees with the adjustment, e.g., believes that the amount on
the Form(s) W-2 is incorrect, the taxpayer can dispute it.

Between 68 percent and 78 percent of returns reporting an AEITC amount
were filed electronically between tax years 2002 and 2004. Submission
Processing runs a computer check to find any mismatches between the
amounts on the electronic tax returns and the electronic Form(s) W-2. When
mismatches are found, Submission Processing rejects the return and sends
it to the taxpayer or preparer to correct and retransmit to IRS.

In most tax preparation software, once a user enters an amount from the
Form(s) W-2, the amount is automatically transferred to the appropriate
line of the tax return. Thus, if the user errs in entering the proper
amount from the Form W-2, the software would enter this erroneous number
on the tax return. As a result, Submission Processing rarely identifies
AEITC tax return/Form(s) W-2 mismatches because the original Form(s) W-2
from the employer(s) is rarely included in electronic filings.^26 IRS
rejected 172 electronically filed returns reporting AEITC in tax year 2004
and 147 in tax year 2003. Adding these mismatches to the paper return
mismatches, Submission Processing found 392 mismatches in tax year 2004
and 429 in tax year 2003.

^24 Because AEITC is an advance payment of the EITC, filers who claim the
EITC are required to reduce the EITC by any AEITC received. By examining
the Form(s) W-2 of EITC claimants for AEITC payments, the IRS attempts to
ensure that excess amounts of EITC are not claimed.

^25 Submission Processing did not have tax year 2002 data available.

^26 The electronic Form (s)W-2, which is an iteration of a user's entries
into the tax preparation software, is generally included in the electronic
filing submission.

Next, Submission Processing sends the return through its Error Resolution
System (ERS) when the AEITC amount on the Form 1040 meets certain
selection criteria.^27 If ERS finds that AEITC exceeds this amount, a tax
examiner matches the Form 1040 AEITC amount to the Form W-2(s). If they
match, no action is taken, and the return is posted in IRS's Masterfile,
the agency's central repository for taxpayer information. If there is a
mismatch, IRS adjusts the return to match the Form(s) W-2, and IRS sends a
letter to the taxpayer describing the error and the ERS correction.
Taxpayers who disagree with the change can dispute it. The ERS process
thus serves as a back-up check in case the earlier physical or electronic
processes missed these AEITC mismatches. ERS examined 3,380 tax returns
with AEITC in tax year 2004. IRS was not able to tell us how many of these
returns involved mismatches.

Prior ERS selection criteria would not have identified some returns for
individuals when the Form(s) W-2 had AEITC amounts that were above the
legal maximums, but below the ERS selection criteria. While interviewing
Submission Processing officials, we suggested that the criteria be
modified and associated with filing status. IRS made such changes to the
criteria, effective January 2, 2007.

IRS's Procedures to Verify SSNs Have Limited Effectiveness for AEITC Recipients
Due to Taxpayer Noncompliance in Reporting and Filing

Submission Processing's procedures have limited effectiveness in verifying
that AEITC recipients have a valid SSN because, as previously noted, many
individuals do not file the required tax return and, for those who do,
most do not report receipt of the advance.

For both paper and electronically filed returns, Submission Processing
checks IRS's Data Master (DM-1) file, which is a database that includes,
among other things, all validly issued SSNs and the individual's name
associated with each SSN. Submission Processing rejects electronically
filed tax returns with an invalid SSN on the Form 1040, including those
from taxpayers who received the AEITC, and sends back the tax return to
the taxpayer for correction before processing. For returns filed on paper
with an invalid SSN on the Form 1040, Submission Processing processes the
return, but disallows certain credits and exemptions, such as the EITC. If
the taxpayer's SSN is invalid and the AEITC is claimed on a paper return,
Submission Processing processes any AEITC reported. Since the AEITC is
reported as tax, it will either offset all or part of any refund due or,
if no refund is due, require the taxpayer to pay back the full AEITC
amount.

^27 ERS is the system for examining and correcting electronic and paper
individual income tax returns rejected due to taxpayer and processing
errors. ERS treats paper and electronic returns in the same way.

Thus, of the approximately 514,000 individuals who received the AEITC each
year between tax years 2002 through 2004, Submission Processing's
procedures would apply to about 118,000 taxpayers--those who filed and
reported receipt of the advance. For nonfilers, Submission Processing
cannot verify SSNs because there are no tax returns---the basis of its
examination. While Submission Processing performs a SSN verification for
those who file and do not report receipt of the AEITC, it is not effective
for the advance since the advance is not reported on the return.

Many Taxpayers Underreported AEITC and IRS Worked on a Limited Number of These
Cases, Some of Which Submission Processing Did Not Find

After a return has been processed, the next point when IRS might identify
and correct AEITC noncompliance is when it matches tax returns to other
documents it receives. Overall, each year AUR identifies about 14 million
discrepancies between taxpayer income and deduction information submitted
by third parties and amounts reported on individual income tax returns.
AUR has the resources to only work on a fraction of these cases each year
and uses criteria, such as revenue collection potential, for case
selection.

For AEITC, AUR compares the amount of AEITC that employers report on a
taxpayer's Form W-2 to the amount reported on an individual's tax return
(see fig. 5). AUR receives this information in separate databases from SSA
and undertakes these comparisons in August and December of each year--well
after Submission Processing has completed its review of the returns and
associated Forms W-2.

Figure 5: AUR Program Procedures to Resolve Discrepancies between AEITC
Amounts on Forms W-2 and Tax Returns

Generally, AUR does not take action when the AEITC amounts match the
amounts reported on the return or when the discrepancy does not meet IRS
case selection criteria. For example, for tax year 2003, our data show
that there were about 209,000 tax returns where the AEITC amount did not
match the amount on taxpayers' Forms W-2, primarily because the returns
failed to report any AEITC or underreported the amount. AUR identified
about 25,000 of those cases in tax year 2003 as potential cases on which
to work.

In tax year 2003, Submission Processing identified 429 instances where the
amount of AEITC reported on the tax return was less than the amount on the
Forms W-2, which was much fewer than the about 25,000 mismatches detected
by AUR that year. The mismatches involving individuals who filed paper
returns and reported receiving the AEITC should have been caught by
Submission Processing because its examiners are supposed to match the
amount reported on the return with the Form(s) W-2. However, Submission
Processing examiners review huge volumes of returns and it can be a
challenge to identify the relatively few with AEITC.

Because AUR does not break down its AEITC cases by filers who underreport
and filers who fail to report any AEITC amount or by filers of paper
versus electronic returns, it is not possible to determine exactly how
many AUR cases should have been caught by Submission Processing. AUR
officials told us that it would be difficult for them to routinely break
out this information in this way because this would require new computer
programming and the budget for new programming requests was reduced for
fiscal year 2007.

In accordance with return processing procedures, if Submission Processing
identifies an underreporter, it reduces the size of the refund that will
be sent to the taxpayer. Thus, using Submission Processing, rather than
AUR, better protects revenue because the erroneous EITC amounts are never
paid to the taxpayer. If AUR works on the case, it assesses the tax owed,
but further IRS action is required to collect the actual
overpayment--through a future refund offset or a current effort to collect
the refund paid erroneously.

Although catching AEITC underreporting in Submission Processing would
better protect revenue than do AUR processes, the benefits to improving
Submission Processing to catch AEITC errors may already be small and could
get smaller. This is because the number of EITC returns filed
electronically is more than 68 percent and has been growing and, as
previously noted, Submission Processing identifies few mismatches for
returns that are filed electronically because the electronic Form W-2 is
an iteration of the user's entries. In its July 2006 report, the Treasury
Inspector General for Tax Administration recommended that IRS reemphasize
the use of the current AEITC review procedures for paper returns, but
agreed with IRS that it could not implement additional procedures for
electronic returns because IRS is unable to fully verify the accuracy of
the Form(s) W-2 during electronic processing of returns.

AUR worked on about twice as many AEITC cases in tax year 2003 as in the
previous year. The total amount of tax assessed in these cases in tax year
2003 was more than three times the amount in the previous year, and the
amount assessed per case increased about 71 percent, from $555 to $947,
which is near the average AUR assessment of about $1,000.^28 These trends
were either holding steady or improving for the 84 percent of tax year
2004 cases for which we had data when we completed our review. The number
of cases in which the taxpayer fully agreed with the assessment rose from
about 29 percent in tax year 2002 to about 41 percent in tax year 2003.
The number of cases in which IRS withdrew the assessment, generally after
the taxpayer provided documentation that it was erroneous, dropped from
about 15 percent in tax year 2002 to about 4 percent in tax year 2003.

^28 AUR categorizes a case as an AEITC case if AEITC is the primary issue,
responsible for at least 80 percent of the increased tax liability. Thus,
the total amount assessed in each AEITC case may include other
underreported amounts. The total amount assessed is not necessarily the
total amount that IRS collects from the taxpayer.

IRS Works on a Limited Number of AEITC Nonfiler Cases

As previously noted, from tax years 2002 through 2004, about 40 percent of
the individuals who had a Form W-2 reporting that they received AEITC did
not file a tax return as required by law.

All nonfiler cases, including AEITC nonfiler and other lower dollar
nonfiler cases, are eligible to be worked on by IRS's Wage and Investment
and Small Business/Self Employed divisions. IRS's policy dictates that
some lower dollar cases and even some cases in which nonfilers are due a
refund may be worked on to ensure that all kinds of cases have the
possibility of being worked on. Still, a key criterion that IRS uses to
determine cases on which to work is potential revenue or the anticipated
net balance due. The higher this amount, the more likely the case will be
worked on. IRS does not track the total number of nonfiler cases worked on
or the kinds of income or credits taken by the nonfilers whose cases were
worked on. Because IRS selects cases based on these criteria, IRS
officials said they worked on few AEITC nonfiler cases.

Beyond AEITC nonfiler cases potentially being worked on by Wage and
Investment and the Small Business/Self Employed divisions, these cases are
also eligible to be worked on by other IRS programs. For example, the
automated collection system involves calls from IRS staff to taxpayers
asking them to file a return or explain why they believe filing is not
required.^29 IRS's automated substitute for return program involves
collection staff preparing a tax return on behalf of a nonfiler based on
third-party and other information that IRS has available. Once a return
has been created, IRS can act to collect any taxes due.

During 2005 through 2006, IRS conducted a test to determine whether
receipt of AEITC should be a criterion to determine nonfiler cases on
which to work. The test involved working on cases from tax years 2000
through 2003 for 433 taxpayers drawn from a sample of taxpayers who did
not file a return in tax year 2002, but did receive the AEITC that year,
and whose income was between $35,000 and $50,000 for that year.^30 The
cases were worked on in the same way that IRS works on other nonfiler
cases.

^29 The automated collection system primarily focuses on collecting taxes
from taxpayers who are delinquent in paying assessed taxes.

In interviews with IRS officials about the AEITC nonfiler test, we found
that the agency's test plan lacked documentation and detail, such as test
justification, likely costs and benefits, and implementation details. It
also lacked a rationale for some important decisions underlying the test
and some changes implemented after the test began were also not well
documented. In our report on three tests conducted by IRS in 2004 to
address leading sources of EITC errors, we noted that the lack of such
documentation hindered monitoring and oversight and did not foster a
common understanding of the tests among management and staff.^31 One of
our recommendations was that the rationale for key decisions on such tests
be documented, and the Commissioner of Internal Revenue agreed with that
recommendation.

Because IRS had not completed its evaluation of the test as of June 2007,
the agency has not decided whether AEITC should be a criterion to
determine nonfiler cases on which to work. Regardless of the agency's
decision, however, it is unlikely to significantly reduce the number of
AEITC nonfilers because any increase in cases worked on would likely
represent only a small number of such nonfilers, relative to both the
total number of AEITC nonfilers and all nonfiler cases worked on by IRS.

Collection Handles Few AEITC Cases

All AEITC cases are eligible to be worked on in IRS's collection program;
however, it generally does not work on cases involving AEITC because the
amounts involved are below selection criteria that determine which cases
to pursue.

Collection does not keep track of the specific types of income and credits
claimed by the individuals whose cases they handle. Still, collection
officials stated that they worked on few, if any, cases involving the
AEITC. Instead, these cases go into "deferral status," which means any
refund the taxpayer is due will be reduced until the balance due has been
paid off. Like other taxpayers with an outstanding tax liability, these
taxpayers would also get a notice stating how much they owe.

^30 IRS officials said they decided to work on the cases for all 4 years,
even if they did not involve the taxpayer receiving the AEITC in some of
those years so as to bring taxpayers in these cases into full compliance.

^31 GAO, Earned Income Tax Credit: Implementation of Three New Tests
Proceeded Smoothly, but Tests and Evaluation Plans Were Not Fully
Documented, [34]GAO-05-92 (Washington, D.C.: Dec. 30, 2004).

Collection had an effort under way to work on cases under its selection
criteria by setting up automatic monthly installment arrangements and
notifying taxpayers that they were expected to begin paying what they
owed. Officials said that too many taxpayers ignored the arrangements, and
as a result, the program was determined to be cost prohibitive. The
program was therefore discontinued.

IRS Audits and Criminal Investigation Identified a Small Number of AEITC Cases

IRS annually audits about 500,000 of the more than 21 million tax returns
that claim the EITC. With only about 3 percent of EITC recipients
potentially eligible for the advance receiving it, only a small number of
the audited returns are likely to involve the AEITC.^32 When IRS audits
tax returns claiming the EITC, an examiner is required to determine if the
taxpayer was eligible for it, whether the taxpayer took the AEITC and, if
so, whether the taxpayer reported the correct amount. If the examiner
determines that the taxpayer was not eligible for the EITC, then the AEITC
is disallowed.

In addition, Criminal Investigation, which investigates potential criminal
violations of the tax code and related financial crimes, has identified
six cases associated with the AEITC since 2001. Five of the six cases
involved refund fraud based on individuals creating fake businesses in
order to obtain the AEITC. The other case involved a business owner
attempting to evade employment tax by falsely signing up employees for the
AEITC, but not including the credit in their paychecks.

Soft Notices Have Increased Compliance in Situations Similar to AEITC, but
Turnover and Difficulty Locating Recipients May Limit Effectiveness for AEITC

Because IRS's enforcement resources cannot fully cover all areas of
noncompliance, including AEITC noncompliance, the agency has tried to cost
effectively increase voluntary compliance in some areas that involve
relatively small amounts of money by mailing taxpayers soft notices. Soft
notices are letters that ask taxpayers to comply with a certain
requirement in the future or, if the notice informs them that they are not
entitled to a benefit that they received, to file an amended return.

^32 IRS does not track the number of AEITC cases that are audited.

While IRS officials said there have not been any soft notices specifically
targeting AEITC noncompliance, we reviewed the results of three tests that
involved taxpayers who were not filing accurately--many of whom would not
otherwise be subject to an enforcement action. Additionally, IRS has
modified one soft notice test that includes cost estimates. Each of the
completed soft notice efforts show some benefits in improving compliance,
however, they may be less effective for AEITC recipients.

The First Soft Notice Test: The first soft notice--called the "Duplicate
TINs" test--involved different taxpayers claiming the same, or a
duplicate, TIN for a dependent or qualifying child in order to obtain an
exemption, the EITC, or child tax credit benefits. In tax year 2002, IRS
identified a total of about 2.4 million taxpayers who used a duplicate
TIN. IRS sent soft notices to about 820,000 taxpayers.

In November 2005, IRS reported that after receiving the soft notice, 11.4
percent of the population amended their tax year 2002 returns.^33 Other
results focused on taxpayers who received the notice for tax year 2002 and
whether they repeated the use of a duplicate TIN on their tax years 2003
and 2004 tax returns. The results were as follows:

           o 84.9 percent did not repeat their behavior in either of the
           ensuing years;
           o 7.7 percent repeated the behavior in 2003, but not again in
           2004;
           o 4.0 percent did not repeat the behavior in 2003, but did so in
           2004; and
           o 3.4 percent repeated the behavior for both ensuing years.

Although IRS did not report the costs associated with this test, it did
estimate the revenue that would have been lost without the soft notices.
IRS reported that it protected a total of $218.3 million using the
Duplicate TINs test. Due to limitations in the research design, such as
not using a control group, IRS reported that it was uncertain whether
these results were solely influenced by the receipt of a soft notice or if
other factors may have contributed to the change in taxpayer behavior and
subsequent revenue protected. IRS no longer considers this a test and
continues to send out soft notices for Duplicate TINs issued each year.

^33 IRS, Wage and Investment Research Group, Soft Notice for Duplicate
TINS for Tax Years 2002, 2003, and 2004, 6-05-12-2-030E (Nov. 29, 2005).

The Second Soft Notice Test: The second soft notice test--called the "AUR
Soft Notice" test--involved filers who underreported small amounts of
certain categories of income, such as wages, unemployment insurance, or
sales of securities. In December and January 2004 and 2005, respectively,
IRS sent 500 soft notices to randomly selected taxpayers who underreported
income on their tax year 2003 returns. IRS also randomly selected a
control group of 500 taxpayers who underreported small amounts, but did
not send notices to this group.

An outside consultant that IRS hired to determine the effectiveness of the
test reported in October 2005 that (1) soft notices appeared to have a
beneficial result in reducing repeat behavior and (2) IRS resources were
not overburdened by the notices.^34 Their conclusion was based on several
results. First, after receiving the soft notice, 71 out of the 500
taxpayers (14.2 percent) filed an amended return. Second, only 33
taxpayers (6.6 percent) who received the notice repeated the
underreporting the following year. In contrast, 174 taxpayers in the
control group (34.8 percent) repeated their underreporting. Third, the
consultants did not consider IRS resources to be burdened because only 45
of 500 taxpayers (9 percent) called IRS with questions. Similarly, the
study found there was limited undeliverable mail--only for 3 taxpayers
(0.6 percent). An additional test was conducted for fiscal year 2006 and
had similar positive results. IRS is in the process of determining whether
it will send out soft notices for AUR in the future.

The Third Soft Notice Test: The third soft notice test--called the
"Dependent Database" test--involved cases selected for three EITC related
issues, including qualifying child, filing status, and Schedule C, "Profit
or Loss from Business," errors. IRS found that 2.4 million taxpayers
appeared to have had errors on their tax returns. In November 2005, IRS
selected about 12,500 taxpayers to determine the impact of soft notices on
taxpayers' behavior when filing their tax year 2005 return. About another
12,500 taxpayers were selected as a control group not to receive the
notice.

In its October 2006 report, IRS found that, although there was a
difference between the test group of taxpayers who received the soft
notice and the control group that did not, the direct relationship between
receiving a soft notice and taxpayers' subsequent filing behavior was
weak.^35 Specifically, the report cited that 88 percent of the test group
and 86 percent of the control group changed their subsequent tax year
filing behavior, including not breaking the same rule, amending the prior
year return, or not filing a 2005 return. Specific noteworthy results
were:

^34 Booz Allen Hamilton, AUR Soft Notice Test Results (Oct. 6, 2005).

           o 84 percent of the test group and 83 percent of the control group
           filed a return in the subsequent year;
           o 46 percent of the test group and 44 percent of the control group
           broke no rules at all;
           o 26 percent of the test group and 25 percent of the control group
           broke a different rule;
           o 12 percent of the test group and 14 percent of the control group
           repeated their behavior the next year by breaking the same rule;
           and
           o 1 percent of the test group and 0.4 percent of the control group
           amended their prior-year return.

Also in October 2006, IRS modified the Dependent Database test in both the
Wage and Investment and Small Business/Self Employed divisions to target
notices to another population, i.e., noncustodian person(s) claiming a
child. IRS prepared a preliminary cost analysis for this soft notice test
based on a sample of 300,000 taxpayers. It estimated the total costs of
sending out 300,000 soft notices to be about $533,000, which included
$449,000 for the labor to process amended returns and answer telephone
calls and $84,000 for mailing. Additional information, including the
results of this test, was not available as of mid-June 2007.

Although IRS did not develop criteria for these soft notice tests about
what would constitute a success, such as a self-correction percentage, an
IRS official knowledgeable about the tests said the agency considers the
three completed tests a success, despite the few shortcomings. The first
and second tests were considered successes because they led to noteworthy
changes in taxpayer behavior. The third test was considered a success
because, although taxpayer behavior did not change significantly,
officials considered it a cost-effective way to have an enforcement
presence among these taxpayers. Officials thought that a soft notice test
could be beneficial for reducing AEITC noncompliance as well, particularly
since the amount of money involved with AEITC is low and the noncompliance
might not otherwise be addressed by IRS.

^35 IRS, Wage and Investment Research Group, Dependent Database Soft
Notice Report for Tax Year 2005, 6-06-12-2-0280E (Oct. 31, 2006).

Although soft notices may have some potential to address certain AEITC
noncompliance, characteristics of the AEITC population might make such
notices less effective or more costly than for the test populations for
two reasons. First, AEITC turnover is high. In each tax year 2002 through
2004, more than half of the individuals were first time recipients.
Moreover, about 73 percent of first-time AEITC recipients in tax years
2002 and 2003 did not elect the AEITC the following year and, thus, would
not repeat noncompliance related to the AEITC.^36 Second, almost 40
percent of AEITC recipients do not file a tax return, which means that IRS
may not have a current address for those taxpayers. If IRS were to send
soft notices to AEITC nonfilers using the last known address, a
significant number of individuals may no longer reside there. This means
IRS might not be able to locate them or it might spend additional
resources trying.

Federal On-Line SSN Verification Services Could Be Used to Determine AEITC
Eligibility, but IRS and SSA Have Raised Concerns

More than 100,000 AEITC recipients had invalid SSNs and reported receiving
millions of dollars in total benefits for each tax year 2002 through 2004
without any substantial check of their eligibility. Because of the
low-dollar amounts involved per taxpayer, IRS worked on only a small
number of these cases. IRS does not have an up-front control or procedure
in place to require employers to verify that an employee seeking the AEITC
has a valid SSN, which could address this noncompliance. Two federal
on-line services have the potential to be used to implement such controls.
Although the services could be used for this purpose, IRS and SSA
officials raised several concerns about implementing such a requirement.

The TIN Matching service and SSNVS are federal on-line services that some
private organizations may use voluntarily to verify whether federal
records show that the name and SSN provided by an individual match.^37 TIN
Matching is a pre-return filing service offered by IRS that allows those
payers whose income is subject to backup withholding, who submit any of
six Form 1099 information returns (e.g., financial institutions), to match
the TIN of the 1099 payee against IRS records.^38 It is one of several
e-service products offered by IRS. The goal of TIN Matching is to improve
the accuracy of Form 1099 data and reduce subsequent inappropriate
penalties and error notices. SSNVS is a service offered by SSA that allows
registered users (i.e., employers or, in certain instances, their
third-party representatives) to verify the names and SSNs of employees
against SSA records.^39 The AEITC is outside the scope of SSA's
responsibilities, and SSNVS is a voluntary service that is currently used
only to increase the accuracy of wage reporting on Forms W-2.

^36 There are different ways to calculate turnover. We considered it to be
first time use within a 6-year period. To provide additional information,
we also compared the number of first-time AEITC recipients who did not
elect the advance in tax year 2003 with the total number of first-time
AEITC recipients in tax year 2002. We made the same comparisons for the
following year.

IRS and SSA officials identified a number of challenges that the agencies
and employers may face if the TIN Matching service or SSNVS were used to
verify AEITC eligibility.

Accuracy: Both the TIN Matching service and SSNVS are based on SSA records
and have high rates of accuracy in terms of determining whether submitted
names and SSNs match. Still, IRS and SSA officials had concerns about
whether these rates were high enough for purposes of verifying AEITC
eligibility. IRS officials told us that the TIN Matching service is about
98 percent accurate. A December 2006 report by the SSA Office of Inspector
General sampled more than 2,000 determinations by SSA's Numident file--the
database upon which SSNVS is based--and found a name and SSN match
accuracy rate of more than 99 percent.^40 Still, SSA officials said that
if SSA records were used to verify AEITC eligibility, they might want to
subject an employee's name and SSN to more "routines"--procedures such as
correcting for transposed numbers that SSA uses to increase the likelihood
of a match--than is currently done by SSNVS. IRS officials told us that
when the agency was informally considering a proposal to charge a fee for
using the TIN Matching service, several IRS officials knowledgeable about
the database opposed the idea because they did not believe it was accurate
enough that users should have to pay for it.

^37 The Department of Homeland Security operates a voluntary program known
as Employment Eligibility Verification, formerly known as Basic Pilot,
through which participating employers enter employee information, such as
name and SSN, into the Web site to verify newly hired employees' work
authorization status. The program then attempts to match that information
against SSA and, if necessary, Department of Homeland Security databases.
We did not examine Employment Eligibility Verification as part of this
report because we have identified several weaknesses in the program's
implementation that could adversely impact increased use. See GAO,
Immigration Enforcement: Weaknesses Hinder Employment Verification and
Worksite Enforcement Efforts, [35]GAO-06-895T (Washington, D.C.: June 19,
2006).

^38 Financial institutions and other businesses that make certain payments
must file a Form 1099 with IRS. Generally, these payments are not subject
to withholding unless certain conditions exist, one of which is an invalid
taxpayer identification number.

^39For the purposes of this report, SSN validity with regard to SSNVS
includes instances where a name matches its SSN as verified by SSA
records.

Accuracy concerns, however, do not preclude IRS or SSA from using SSA
records to make an initial determination about whether individuals who may
claim credits or benefits have demonstrated that they are entitled to
them. For example, as previously noted, IRS rejects electronically filed
tax returns with a name/SSN mismatch and returns them to the taxpayer for
correction. For paper returns reporting AEITC receipt that have a name/SSN
mismatch, IRS processes the returns, but since the AEITC is reported as
tax, it will either offset all or part of any refund due or, if no refund
is due, require the taxpayer to pay back the full AEITC amount.

Similarly, SSA instructs its claims representatives that if the identity
of a claimant for Social Security benefits or Supplemental Security Income
benefits remains questionable because the individual has not provided
sufficient proof to establish his or her identity, the claim will be
denied even if other factors of eligibility are met.^41

Additional employer responsibilities: IRS and SSA officials said a major
concern about employers using either service was whether a name/SSN
mismatch would create new responsibilities for employers beyond informing
the employee and denying the AEITC. IRS officials also expressed concern
that requiring employers to use the services for AEITC would discourage
them from promoting the AEITC and perhaps encourage them to dissuade
employees from seeking it. Under current procedures for using the TIN
Matching service and SSNVS, employers are not required to take any action
based on the results they receive.

^40 SSA, Office of Inspector General, Accuracy of the Social Security
Administration's Numident File A-08-06-26100 (December 2006). When SSA
assigns an SSN to an individual, the agency creates a master record in its
Numident file, which contains information about the number holder,
including the person's name. Thereafter, the Numident record for each
number holder identifies any changes to the original information provided
by the number holder, including name changes.

^41 Supplemental Security Income is a federal program administered by SSA
designed to provide cash for food, clothing, and shelter to aged, blind,
and disabled people who have little or no income.

IRS officials expressed concern that a June 2006 regulation proposed by
the Department of Homeland Security could expressly list employer receipt
of a "no match" letter from SSA as possible evidence that the employer
knew or should have known that it was employing an individual not
authorized to work in the United States.^42 Under the proposed regulation,
if the employer fails to take reasonable steps to resolve the discrepancy
after receiving the letter, the Department of Homeland Security may find
that the employer had such knowledge and assess civil monetary penalties
against the employer.

The proposed Department of Homeland Security regulation describes "safe
harbor" procedures that the employer can follow in response to the letter.
Those steps include the employer promptly checking its records for
clerical errors and obtaining required documentation by working with the
employee, SSA, and the Department of Homeland Security. If the name/SSN
mismatch cannot be resolved, the employer would have to choose between
terminating the employee or facing the risk that the Department of
Homeland Security may find that the employer knew that the employee was
not authorized to work and, by continuing to employ the individual,
violated the law.

While the proposed Department of Homeland Security regulation covers only
no match letters, IRS and SSA officials expressed concern that the
regulation could be expanded to include name/SSN mismatches disclosed by
the TIN Matching service or SSNVS. Under current law, existing and
prospective users who are concerned that matching could be used for
purposes beyond improving the accuracy of Form 1099 data or wage reporting
can voluntarily cease using, or not start using, the systems. IRS and SSA
officials noted, however, that employers would no longer have this choice
if they were required to use one of these services to verify an AEITC
applicant's SSN.

^42 As of June 2007, this proposal was still outstanding. SSA sends "no
match" letters to employers who submit more than 10 Forms W-2 to SSA or
when more than 0.5 percent of these wage reports consist of a name and SSN
combination that do not match SSA records.

In addition, existing Department of Homeland Security guidance for
employers on the interaction between antidiscrimination laws and legal
requirements for verifying employment eligibility states that employers
must treat all employees in the same manner. Employers cannot set
different employment eligibility verification standards or require that
different documents be presented by different groups of employees. If
mandatory verification of AEITC applicants' SSNs created additional
employer responsibilities under employment eligibility verification
requirements, this result could be inconsistent with efforts to ensure
that verification procedures apply to all employees.

SSA's position on SSNVS is that a name/SSN mismatch does not make any
statement about an employee's immigration status and should not be a basis
for taking any adverse action against the employee. SSA officials also
expressed reservations about SSNVS results being used to terminate
employees.

Capacity and User Access: IRS and SSA officials said changes to the
capacity and user access of the TIN Matching service or SSNVS would either
be unnecessary or minor if employers used them to verify SSNs of employees
seeking AEITC, although the officials said their agencies might favor
creating a new service for this purpose instead. They told us that their
systems have the capacity to handle the increased volume of requests that
would result from this expanded use. Because both services already are
used to verify SSNs, IRS and SSA said any changes to how the services are
accessed and used would also probably not be extensive.

Both SSNVS and TIN Matching are Web-only services.^43 To use one of the
services, the employer designates one or more employees or third-party
representatives to register on behalf of the employer. The initial
registration for both the TIN Matching service and SSNVS is handled in a
similar way and may take as long as 4 weeks:

           o Registrants go to the agency's Web site and provide information
           about themselves and their employers on a form, which they send
           electronically to the agency.
           o The agency sends the employer of the registrant a unique code.
           The letter directs the employer to provide that code to the
           registrant.
           o After receiving the code, the registrant can go back to the
           agency Web site and input the code to activate use of the service.

^43 Both IRS and SSA have toll-free telephone numbers listed on their Web
sites that those looking for help with the particular service can call to
speak with agency staff and get help.

TIN Matching registrants who do not use it for 6 months must reregister,
primarily to receive a new password and update any of the information
provided during the registration. SSNVS requires registrants to change
their password once a year to keep it from expiring, which also requires
reregistration.

IRS and SSA officials told us that the great majority of users of their
respective services generally report that they are not difficult to use
for either the registration process or ongoing use. IRS officials said
users access TIN Matching voluntarily and that some, particularly from
smaller organizations, appear more likely to find it burdensome than users
from larger organizations. We found that more than half the employers that
provided the AEITC in tax years 2002 through 2004 were small businesses or
self- employed and about one-quarter were tax exempt and government
entities. In addition, despite IRS's efforts to outreach to large
employers, fewer than one-fifth were large and midsize employers (see
table 20 in app. III). IRS officials said TIN Matching service users who
found the registration process burdensome were generally those who
reported to IRS that they were not use to filling out forms on-line and
creating and using passwords. The officials also said that some TIN
Matching users reported being uncomfortable having to provide personal
information to register.

SSA officials said SSNVS is used mostly by larger employers, and a
relatively small number of them reported that they found the service
burdensome. SSA officials did say, however, that SSA received about 89,000
calls through June 2007 from individuals about registration. And, an SSA
official said that small business representatives with whom he has
recently spoken expressed frustration with the overall number of tasks
that the federal government was already requiring them to perform and,
therefore, might be reluctant to verify SSNs.

IRS officials also told us that the TIN Matching service is not programmed
to track users, which likely would be useful to IRS for enforcement
purposes. Still, IRS officials added, that the service could be modified
to track employers that used the services for AEITC purposes and report on
the results. SSNVS tracks its users to determine whether the service is
being properly used.

Additional resources: IRS and SSA officials said that if employers were
required to begin using their respective services to verify the SSNs of
employees seeking the AEITC, the agencies would need additional resources.
For example, officials from both agencies cited the need to handle
questions from new users, particularly in the first year when all
individuals seeking the AEITC would be required to have their names and
SSN's matched.^44

When we told IRS officials that our data showed that about 50,000
employers had at least one AEITC recipient, they said such an increase in
the number of registrants could require IRS to increase the number of
staff available to answer user questions. But they said they could not
estimate how many more staff would be necessary.

IRS officials also said they were trying to make the reregistration
process easier because they received a substantial number of calls from
users who needed to reregister. This would be important for AEITC-related
use of TIN Matching because employers with only one or two employees
seeking the AEITC would only need to use the service once or twice a year,
making it likely that they would have to reregister.

SSA officials told us that if the agency was considering or was directed
to make its SSN records available for the purpose of verifying AEITC
eligibility, SSA would have to devote additional resources to conduct a
comprehensive assessment to determine the changes necessary to SSNVS to
properly achieve this goal, including possibly creating a different
service for assessing AEITC eligibility and buying a new database server
to handle the increased volume of users. Again, SSA officials could not
estimate how many more staff would be necessary.

New federal legislation: Enactment of federal legislation would be needed
for employers to begin using the TIN Matching service to verify the SSNs
of their employees seeking the AEITC.^45 In 2000, the Department of the
Treasury recommended to the Congress that TIN verification be expanded to
include other payers subject to an IRS reporting requirement, such as
employers who file Forms W-2.^46

^44 The higher volume of questions in the first year assumes that employees
whose name and SSN matched and who sought the AEITC from the same employer
in a subsequent year would not be required to undergo a second match.

It is uncertain whether IRS could require employers to use SSNVS to verify
the SSNs of employees seeking the AEITC. SSA officials said they would
need to determine whether SSA's disclosure of SSN data is compatible with
the reason it collected the information and, if so, whether verifying SSNs
via SSNVS for purposes of AEITC eligibility is consistent with SSA's legal
obligations. Both Treasury and SSA officials said their agencies would
strongly prefer enactment of legislation before requiring employer or any
verification of SSNs of employees seeking the AEITC.

Employee appeal of mismatch: One difference between using the TIN Matching
service or SSNVS would occur when an employee claimed that a name and SSN
mismatch was inaccurate. IRS and SSA officials said employees who
questioned a SSN mismatch would presumably contact IRS, which would send
them to SSA to resolve the issue because the TIN Matching service is based
on SSA records. IRS officials said that, in contrast, employees
questioning a name and SSN mismatch generated by SSNVS would presumably go
directly to SSA.

Agency mission: SSA officials also said that verifying eligibility for the
AEITC is most appropriate for IRS because it is a tax administration issue
and is therefore outside the scope of SSA's mission. However, it is not
unusual for agencies to assist other federal agencies in carrying out
their mission. SSA officials also said that, regardless of which service
was used, IRS would have full administrative responsibility for overseeing
a program for employers to verify AEITC applicants' SSNs.

^45 Internal Revenue Code Section 6103 prohibits the disclosure of
information on tax returns, including SSNs, unless permitted by a specific
exception. Because there is no exception for the verification of SSNs by
employers, legislation would be needed to use the TIN Matching database
for this purpose.

^46 Office of Tax Policy, U.S. Department of the Treasury, Scope and Use
of Taxpayer Confidentiality and Disclosure Provisions October 2000. Such a
provision was included in S. 1321, 109 Cong. Sec. 509 (2006).

IRS Does Not Require Submission of Forms W-5 from Employers for Employees
Seeking the AEITC

IRS does not require employers to submit a Form W-5 when an employee
requests receipt of the AEITC. Several advantages and disadvantages exist
if IRS creates a Form W-5 database to use in monitoring AEITC
noncompliance issues.

IRS could require employers to submit a Form W-5 when an employee requests
receipt of the AEITC. In turn, IRS could use the Forms W-5 to create a
database to monitor the AEITC. The database could be used to ensure that
the SSN provided on the Form W-5 is valid and that it matches the
individual's name. Doing such a check could have prevented more than
100,000 individuals from receiving as much as between $37 million and $39
million each year in AEITC to which they were potentially not entitled
because of not meeting the valid SSN requirement. Such a database could
also allow IRS to know which individuals received the AEITC and provide
the agency with an opportunity to send recipients a notice at the start of
the next filing season reminding them to file a federal tax return. A
reminder to file notice could likely reduce noncompliance for up to about
200,000 individuals who received between $42 million and $50 million each
year in AEITC without filing a federal tax return. Similarly, IRS
officials could use a W-5 database to verify other AEITC requirements,
such as ensuring that each recipient has only one Form W-5 in effect at a
time. This check could reduce the probability that individuals would
receive more than the yearly AEITC maximum.

While acknowledging that potential advantages exist to developing and
maintaining a Form W-5 database, IRS officials said that the disadvantages
could outweigh these and any other advantages. Although IRS officials said
it was too early in the proposal process to calculate the database's
potential costs and subsequent return on investment, they said it very
likely would be substantially lower than the return on investment for
either existing or anticipated future noncompliance programs. For example,
IRS estimates the current return on investment for EITC Examination
noncompliance is between $17 to $1 and $19 to $1. Although these amounts
only include labor and do not include overhead such as facilities,
equipment, and supplies, officials felt confident that the EITC return on
investment would far exceed that for the AEITC. Their opinion was largely
based on the few dollars involved with the AEITC, especially compared to
other noncompliance programs.

In addition, IRS officials expressed concerns that employers would not
submit Forms W-5 to IRS. Officials raised an analogy between this proposal
and the prior Questionable W-4 program.^47 As we reported in 2003, about
75 percent of the large employers with 1,000 or more employees in IRS's
Large and Medium-Size Business and Small Business/Self Employed divisions
who filed tax returns in tax year 2001 did not send IRS any questionable
Forms W-4.^48 After our report, IRS discontinued the Questionable W-4
program. Additionally, officials noted that requiring employers to submit
Forms W-5 may discourage them from participating because if the employer
was notified that the SSN on a Form W-5 was invalid or that it did not
match the employee's name, employers would likely have a responsibility to
discuss the matter with the employee, creating yet another new burden
employers would not want to accept. Finally, employers may not participate
because if the employee left the employer during the year, the employer
would again have to contact IRS so the Form W-5 database could be updated.

Options to Reduce AEITC Noncompliance May Be Cost Effective

Although we do not know how successful the various options we have
identified for improving AEITC compliance may be if implemented or what
the full cost of implementation would be, IRS may be able to achieve a
return on investment somewhat comparable to that for EITC examinations. We
found an average of about $94 million a year in AEITC noncompliance for
recipients in tax years 2002 through 2004.^49 If a compliance effort could
reduce AEITC noncompliance by one-quarter, that is, $24 million per year,
IRS could spend about $1.3 million each year to do so and achieve a $19 to
$1 return on investment.^50 Alternatively, IRS estimated that it would
cost about $533,000 to send soft notices to 300,000 taxpayers during the
Dependent Database test. If IRS were to test sending soft notices for
AEITC and it cost IRS about the same amount to send notices to 300,000
noncompliant AEITC recipients, IRS would only need to reduce AEITC
noncompliance by about 11 percent (about $10 million) to achieve a $19 to
$1 return on investment.

^47 The Questionable W-4 program required employers to submit information
to IRS on taxpayers who claimed more than 10 withholding allowances or
exemptions.

^48 GAO, IRS's Questionable Form W-4 Program, [36]GAO-04-79R (Washington,
D.C.: Nov. 6, 2003).

^49 The number includes AEITC amounts received by nonfilers, filers who
misreported the amount received, and individuals with invalid SSNs.

^50 This estimate is illustrative only and many variables could affect how
successful any compliance effort would be and the cost of implementing an
effort.

Eligible Taxpayers Could Receive Full EITC Benefits If the Advance Option Were
Discontinued

If the advance option were discontinued, eligible AEITC recipients could
still receive the full benefits of the EITC as a lump sum after filing
their tax return. In addition, improper AEITC payments to ineligible or
noncompliant individuals would be eliminated. The exact amount of revenue
that could be saved is not known. However, in determining an amount, IRS
officials said they would consider the following: the amount of AEITC
disbursed compared with the amount shown on filed tax returns, the cost of
administering the AEITC (e.g., forms and publications, processing,
compliance activities), and any amount currently recovered through
compliance activities.

Conclusions

The AEITC has never achieved a significant participation rate, the amount
recipients received in the period we reviewed was quite low, and
noncompliance was high. However, policymakers may judge that the goal of
providing funds to low-income workers during the year, as opposed to a
lump sum that they could get as part of the EITC when filing their taxes,
remains important and should continue to be allowed. If so, IRS needs to
pursue potentially cost-effective measures to address AEITC noncompliance.

Each of the three options we identified for improving AEITC
compliance--using soft notices, having up-front verification of AEITC
applicants' SSNs by employers, or requiring employers to submit copies of
the Form(s) W-5 and creating a database to monitor AEITC--appears to have
potential to improve compliance, but their full benefits and costs need to
be evaluated and, if possible, tested. Soft notices have improved
compliance in other tax programs but they could be somewhat less effective
in improving AEITC compliance, in part, due to its high turnover rate. The
TIN Matching service and SSNVS have the potential to reduce AEITC
noncompliance by enabling employers to verify workers' SSNs before
providing them the AEITC. Differences exist between the two services, but
either could likely be used for AEITC SSN verification. Both services are
based on SSA records that are already deemed accurate enough such that SSA
and IRS make decisions based on them to disallow certain exemptions and
credits until eligibility has been properly demonstrated. Significant
concerns exist, however, such as the need for legislation authorizing
either the use of TIN Matching or SSNVS for AEITC. This and other issues
would need to be further explored as the costs and benefits of employers
verifying employees' SSNs are fully identified. If IRS required employers
to submit copies of the Form W-5 when an employee requests the AEITC, IRS
could create a database to better monitor and address all three of the
noncompliance problems we analyzed. However, imposing additional
responsibilities on employers for both the SSN verification option and the
Form W-5 database option have the potential to adversely affect the
AEITC's already low participation rate if employers avoid providing the
AEITC due to increased responsibilities on their part.

Due to the relatively small size of the AEITC overall, combined with the
low dollar amounts per taxpayer, IRS officials are concerned that
addressing AEITC noncompliance may provide less return on IRS's
enforcement efforts than would addressing other noncompliance issues.
However, IRS may be able to achieve returns on AEITC enforcement that
would not be significantly out of line with returns on other enforcement
work. For example, it cost IRS about $533,000 to send soft notices to
300,000 taxpayers in the Dependent Database test. If IRS were to test
sending soft notices for AEITC and it cost IRS about the same amount to
send notices to 300,000 noncompliant AEITC recipients, IRS would need to
reduce noncompliance by about 11 percent (about $10 million) to achieve a
$19 to $1 return on investment.

Recommendations for Executive Action

The Acting Commissioner of Internal Revenue should analyze whether any of
the following options could cost effectively and significantly reduce
AEITC noncompliance:

           o sending potentially noncompliant AEITC recipients soft notices,
           such as to nonfilers whose Forms W-2 show that they received AEITC
           and filers who misreported the amount they received or whose SSN
           and name do not match;
           o requiring employers to verify the SSN of employees seeking
           AEITC; or
           o requiring employers to submit Form W-5 to IRS and IRS creating
           and maintaining a database for these forms.

To better identify the costs and implementation issues as well as the
likelihood for these or other options to reduce AEITC noncompliance, where
practical, the Acting Commissioner of Internal Revenue should test these
options to make a more fully informed judgment about whether any would be
worthwhile.

If the Acting Commissioner of Internal Revenue determines that none of
these options would be cost effective and no other remedies are viable,
then the Treasury Secretary should inform the Congress of this and provide
Treasury's opinion about whether the AEITC should be retained.

Agency Comments

The Acting Commissioner of Internal Revenue provided written comments in a
July 18, 2007 letter. He agreed with our recommendation and outlined the
actions IRS would take to address that recommendation, including
conducting further analyses and possible testing of proposed options for
reducing AEITC noncompliance. He also stated that IRS will conduct its
cost-benefit analyses in conjunction with a congressional requirement to
study the impact of expanding eligibility of the AEITC to all EITC
recipients. We also provided a draft of this report to the Department of
the Treasury and SSA and incorporated technical comments where
appropriate. SSA emphasized that verifying eligibility for the AEITC is
most appropriate for IRS because it is a tax administration issue and
therefore outside the scope of SSA's mission.

As agreed with your offices, unless you publicly announce its contents
earlier, we plan no further distribution of this report until 30 days
after its date. At that time, we will send copies of this report to the
Secretary of the Treasury, the Commissioner of the Social Security
Administration, the Acting Commissioner of Internal Revenue, appropriate
Congressional committees, and other interested parties. This report is
available at no charge on GAO's web site at [37]http://www.gao.gov .

If you or your staffs have any questions, please contact me at (202)
512-9110 or [38][email protected] .

Michael Brostek
Director, Tax Issues Strategic Issues Team

Appendix I: IRS's Implementation of Recommendations from GAO's 1992 Report
on the Advance Earned Income Tax Credit

After analyzing IRS's responses to our recommendations in our 1992 Advance
Earned Income Tax Credit (AEITC) report, we determined that IRS has
implemented five of the six recommendations to the Commissioner of
Internal Revenue from our 1992 report on the AEITC and partially
implemented the remaining one (see table 8).

Table 8: IRS Implementation of Recommendations from GAO's 1992 Report on
the Advance Earned Income Tax Credit

Number Recommendation                Status      IRS implementation        
1      Include information on AEITC  Implemented Developed publicity       
          in employee outreach                      materials.                
          materials and programs.                                             
2      Notify taxpayers who receive  Implemented Revised Form W-2 and      
          the Earned Income Tax Credit              Notice 797 to contain     
          (EITC) about the advance                  information on how to     
          option.                                   apply for the AEITC and   
                                                    developed radio and       
                                                    television announcements  
                                                    on the AEITC.             
3      Encourage employers to notify Implemented Added text to the face of 
          employees who have no income              the Form W-4 instructions 
          tax withheld of the advance               advising employees to     
          option.                                   consider the AEITC.       
4      Clarify instructions on AEITC Implemented Clarified employers'      
          in "Circular E, Employer's                responsibilities and      
          Tax Guide."                               liabilities in advancing  
                                                    the EITC in the 1993      
                                                    "Circular E, Employer's   
                                                    Tax Guide."               
5      Send, to individuals who do   Partially   Included information on   
          not file tax returns, a       Implemented the AEITC in a reminder   
          notice explaining their                   to file notice until 1997 
          requirements to file.                     and added a separate      
                                                    AEITC box on Form W-2.    
6      Explore ways to identify      Implemented Added a line on the Form  
          those individuals who receive             1040 for taxpayers to     
          the credit advance, but do                report AEITC received and 
          not report it, so as to                   redesigned the Form W-2.  
          prevent them from receiving                                         
          the credit a second time.                                           

Source: GAO analysis of IRS data.

Our first recommendation was for the Commissioner of Internal Revenue to
include information on AEITC in employee outreach materials and programs.
IRS implemented this recommendation primarily by developing publicity
materials (i.e., grocery bags, milk carton art, brochures, posters) and
distributing them to the public.

For the second recommendation, IRS stated that it did not have the
approximately $2 million in funding that the agency said would have been
required to notify all taxpayers who receive the EITC, but did not elect
the advance option. Instead, IRS took other actions including revising the
Form W-2 in 1992 to contain information on how to apply for the AEITC and
IRS Notice 797, "Possible Federal Tax Refund Due to the Earned Income
Credit," to include information on how to apply for the AEITC.

For our third recommendation, IRS noted that it could encourage employers
to make such notifications, but that there are no statutory sanctions on
employers who fail to do so. Beginning in 1992, IRS placed text on the
face of the Form W-4 instructions advising employees to consider filing a
Form W-5 with their employer to obtain the advance through lower
withholding.

For our fourth recommendation to clarify instructions on the AEITC in
"Circular E, Employer's Tax Guide" IRS did so through the inclusion of new
examples explaining to employers how they should make advance payments to
employees and how employers can report these amounts.

Our fifth recommendation was for the IRS Commissioner to send individuals
who received the AEITC and do not file tax returns a notice explaining the
requirement to file. IRS partially implemented this recommendation by
including information on advance payment in a reminder to file notice and
adding a separate AEITC box on Form W-2. IRS did not track the number of
AEITC nonfilers who received the notice. The reminder to file notice was
only sent until 1997 and IRS officials were uncertain why that notice was
discontinued.

The last recommendation was for exploring ways to identify those
individuals who receive the credit in advance but do not report it. IRS
pointed out that its systems were not geared to detecting unreported AEITC
payments at the time the returns are processed and the best approach to
preventing noncompliance by AEITC payment recipients is a proactive one
that recognizes the filing of correct returns. IRS implemented this
recommendation by providing a separate line on Form 1040 on which to
report AEITC payments and redesigning the Form W-2, for tax year 1993,
which it believed would increase the accuracy of the AEITC payment
information reported on Form W-2. Our work in this report demonstrates a
continuing need to explore additional compliance initiatives aimed at
those who receive the AEITC, but do not report it on their tax return.

Appendix II: Scope and Methodology 

To answer the first and second objectives:

           o how many individuals received the Advance Earned Income Tax
           Credit (AEITC) compared with the Earned Income Tax Credit (EITC)
           and how much did they receive in tax years 2002 through 2004; what
           actions, if any, have been taken to increase use since 1992; and
           what is the potential for significant increases in the future; and
           o what is the extent of noncompliance, if any, associated with the
           AEITC;

we obtained a data file of all Forms W-2, "Wage and Tax Statement," for
tax years 1999-2004 indicating AEITC payments as shown by an amount
greater than $0 in box 9 of the Form W-2 from the Internal Revenue Service
(IRS).^1 We used these tax years because they were the most current
available at the time we started our review. The Form W-2 identified key
information, including the AEITC recipient's name, address, Social
Security number (SSN), and amount of AEITC dollars paid, as well as the
employer's name and address.

To determine the number of individuals who received the AEITC, we used
Forms W-2 instead of tax returns, which IRS has historically used to
estimate AEITC use. We used this alternate approach because we believe the
Forms W-2 provide results that are more accurate and complete.^2 For
example, using Forms W-2 would include in the population of AEITC
recipients those who received AEITC, but did not file a return, and those
who filed a return, but did not report any AEITC. Using tax returns would
not capture these individuals or related noncompliance issues. In
addition, using tax returns counts instances where both spouses receive
the AEITC and file jointly on one return as opposed to two individuals.
IRS's Research, Analysis, and Statistics and EITC program office officials
agreed with our methodology.

We performed data reliability tests on the data file to determine whether
the data were sufficiently reliable for our intended purposes. We did this
testing, in part, by conducting preliminary analyses, which identified
certain data irregularities or anomalies. We identified two noteworthy
anomalies in the data file: (1) excessive AEITC dollar amounts and (2)
invalid AEITC recipient SSNs. First, many Forms W-2 showed that employees
received amounts over the allowable limits. A few even showed individuals
each receiving about $1 million in AEITC--amounts clearly above AEITC
legal limits and which IRS officials said would be improbable, potentially
resulting from transcription errors. Second, we also found some instances
where the SSN and/or name on the Form W-2 were invalid, which means that
the number was never issued by SSA or that the name and number on the Form
W-2 did not match the listed name for that same SSN in official records
maintained by IRS. We compared the number and name information on the Form
W-2 to the National Account Profile to evaluate the validity of that
information and to identify any possible subsequent corrections.^3

^1The AEITC amount on the Form W-2 is what an employer reported paying to
an employee. While this is not necessarily evidence that the employee
actually received that amount, IRS considers the Form W-2 the official
wage and tax statement and makes adjustments to the tax return based on
the Form W-2 when discrepancies exist.

^2An individual who had more than one Form W-2 reporting AEITC for a
particular year was counted as only one individual and the amount of AEITC
paid was totaled.

To address these data anomalies, we separated the Form W-2 file into four
subpopulations using the following three criteria: whether (1) the SSN on
the Form W-2 was valid, according to Data Master File (DM-1); (2) the SSN
and the recipient's name on the Form W-2 matched, according to DM-1; and
(3) the amount of AEITC received was in excess of the yearly maximum.^4
Each of the subpopulations had a unique profile, as follows:^5

           1. Valid subpopulation: This group of Forms W-2 represents all
           individuals (1) that had a valid SSN, meaning that it was a number
           issued by SSA, (2) whose name matches the SSN, and (3) that had an
           AEITC amount within the yearly maximum. More than 75 percent of
           the Forms W-2 on average during tax years 2002 through 2004 were
           in this subpopulation.
           2. Invalid name subpopulation: This group of Forms W-2 represents
           all individuals that had (1) a valid SSN (2) a SSN that did not
           match the individual's name and (3) the AEITC amount was within
           the yearly maximum. About 17 percent of the Forms W-2 fell in this
           subpopulation for each of the 3 years we reviewed.
           3. Invalid number subpopulation: This group of Forms W-2
           represents all Forms W-2 that had an invalid SSN and an AEITC
           amount that was within the yearly maximum. About 7 percent of the
           Forms W-2 during tax years 2002 through 2004 were in this
           population.^6 
           4. Dollar limit subpopulation: This group of Forms W-2 represents
           all instances where the AEITC amount was above the yearly maximum,
           regardless of whether the SSN was invalid or if the individual's
           name matched the SSN. This represented less than 1 percent on
           average of all Forms W-2 in each of the 3 years we reviewed.
           Because IRS officials told us these data were likely erroneous, we
           excluded it from most of our analyses, and IRS officials agreed.

^3 The National Account Profile incorporates the DM-1 from SSA, which is
the primary source file for all SSNs. The National Account Profile also
includes additional SSN change information available to IRS to show the
most current and complete taxpayer information.

^4 The AEITC maximum limits for tax years 1999-2004 were: $1,387, $1,412,
$1,457, $1,503, $1,528, $1,563, respectively. These amounts are limited if
two married individuals both elect the AEITC.

^5 While each subpopulation contains a unique set of Form W-2s, during tax
years 2002-2004 there were 856 Form W-2s that appear in both the valid and
invalid name subpopulations. This could occur due to variations in how the
name is spelled or reported by the employer and/or due to timing
differences in reporting and updating name changes to SSA.

Table 9: Number of Forms W-2, Number of Individuals, and the Dollars for
Each Subpopulation, Tax Years 2002 through 2004

                       Tax year 2002                    Tax year 2003                    Tax year 2004
               Number                           Number                           Number                          
                   of                               of                               of                          
                Forms   Number of                Forms   Number of                Forms   Number of              
                  W-2 individuals      Dollars     W-2 individuals      Dollars     W-2 individuals      Dollars 
Valid         488,007     418,774 $107,230,369 463,285     398,526 $106,075,016 469,234     398,335 $111,098,330 
subpopulation                                                                                                    
                (74%)       (78%)        (57%)   (76%)       (79%)        (70%)   (76%)       (78%)        (68%) 
(percentage)                                                                                                     
Invalid name  114,819      80,524  $25,734,033 101,646      74,234  $25,480,158 101,140      77,747  $26,819,235 
subpopulation                                                                                                    
                (17%)       (15%)        (14%)   (17%)       (15%)        (17%)   (16%)       (15%)        (16%) 
(percentage)                                                                                                     
Invalid        49,432      32,501  $11,582,604  44,244      30,622  $11,650,727  43,696      31,875  $12,166,702 
number                                                                                                           
subpopulation    (8%)        (6%)         (6%)    (7%)        (6%)         (8%)    (7%)        (6%)         (7%) 
                                                                                                                 
(percentage)                                                                                                     
Dollar limit    6,408       6,360  $43,703,747   2,692       2,677   $7,486,726   2,768       2,757  $12,917,134 
subpopulation                                                                                                    
                 (1%)        (2%)        (23%)   (less        (1%)         (5%)   (less        (1%)         (8%) 
(percentage)                                      than                             than                          
                                                   1%)                              1%)                          
Total         658,666     538,159 $188,250,753 611,867     506,059 $150,692,627 616,838     510,714 $163,001,401 

Source: GAO analysis of IRS data.

Note: Percentages may not add due to rounding.

^6 Because the SSN was invalid, it is not possible to determine whether
the individual's name matched.

We conducted additional data reliability tests for each of the databases
we used to obtain information about the AEITC, including IRS's Individual
Returns Transaction File, for return and filing information, which came
from the Compliance Data Warehouse; National Account Profile/DM-1, for
IRS's SSN and name reference information, which also came from the
Compliance Data Warehouse; Automated Underreporter (AUR), for IRS's
third-party information return data; and Taxpayer Identification Number
(TIN) Matching and SSA's Social Security Number Verification System
(SSNVS), for alternative SSN and name reference information used by
employers.^7 After completing our data reliability assessments, we
determined the AEITC data to be sufficiently reliable for analysis and our
reporting objectives.

We also developed a comprehensive analysis plan that included our
researchable issues, planned analysis, data sources, and limitations. We
shared our plan with IRS and others and incorporated their feedback.
Because IRS's workload precluded them from providing information related
to employers/AEITC payers within our time frames, we were able to conduct
only limited analyses of employers who paid AEITC.

Using the analysis plan for each subpopulation, we conducted multiple
analyses to develop relevant demographic, characteristic, and compliance
data. Because each population had different criteria, certain
characteristics or compliance data could not be developed or compared
across the subpopulations. For example, the only characteristics data that
could be developed for the invalid SSN subpopulation came from the Form
W-2 (e.g., amount of AEITC, geographic location) because it is the only
available source. Similarly, tax return data, such as filing status, was
not available for those who did not file a tax return.

All data pertaining to filed tax returns came either from returns that
reported receipt of the AEITC on the appropriate line or from a
"constructed tax return," which IRS officials created using the SSN on the
Form W-2 and matching it to an SSN in the primary, secondary, or dependent
position on a filed return. The location of the SSN in one of these
positions is relevant due to the way IRS manages its data files. There
could be instances when a tax return was filed but it was not detected
using our methodology. For example, a taxpayer's SSN on the Form W-2 might
have been incorrect and the taxpayer reported the correct number on the
tax return (Form 1040).

^7 The Compliance Data Warehouse houses excerpts of data from various IRS
systems for easy retrieval.

To report on data pertaining to the EITC, we relied on published EITC data
provided by IRS research and program office officials, including the EITC
database, the EITC Database Year to Year Comparison Report, and EITC Fact
Sheet.^8 When possible, our analysis compares individuals who received the
AEITC with individuals who received the EITC. IRS defines EITC recipients
by the number of federal tax returns that received EITC. In addition, EITC
data are based upon the primary TIN of all taxpayers who received an
amount of EITC. We determined the number of individuals who receive the
AEITC based on the number of Forms W-2 reporting AEITC per unique SSN. IRS
officials agreed that even though these populations are not identical, it
is reasonable to make a comparison between them.

We frequently consulted with IRS officials on the data and our analyses;
they generally agreed with both the approach and the accuracy of the
results. For the analyses that IRS conducted, we agreed with both the
approach and the accuracy of the results.

In addition, we reviewed legislative and IRS administrative changes to the
AEITC since 1992 and discussed them with IRS and other officials,
including Department of the Treasury officials. We reviewed reports on
IRS's implementation of some of our prior recommendations pertaining to
the EITC and discussed them with IRS officials, including the National
EITC Director. We also coordinated this work with the Treasury Inspector
General for Tax Administration.

Finally, we also identified and interviewed 11 individuals we determined
to be experts to provide us a fuller understanding on the potential for
significant increases or improvement to AEITC use and noncompliance and
included academics, researchers, practitioners, and individuals
representing the areas of tax policy, low-income individual issues, and
compliance issues.^9 We chose these individuals based on our knowledge of
their areas of expertise and our research that indicated they were
knowledgeable about the EITC.

^8 The EITC Database is an extraction of EITC data from IRS's Individual
Returns Transaction File. The database is located on the Compliance Data
Warehouse. These data met our reliability standards.

To address the third objective, how well do IRS's procedures address any
areas of noncompliance, we examined portions of IRS's Internal Revenue
Manual and interviewed IRS Wage and Investment and Small
Business/Self-Employed division officials to determine procedures for
processing returns that reported receipt of the AEITC. We examined how
IRS's enforcement procedures and operations, including Submission
Processing, AUR, Nonfiler, Collection, Examination, Criminal
Investigation, and Taxpayer Advocate, address certain kinds of potential
noncompliance.

We explored various options to improve AEITC compliance. This involved
conducting literature searches and interviews with IRS and SSA officials.
We reviewed and discussed the results of soft notice tests with IRS
officials, including the National EITC Director, and discussed the
applicability of soft notices for addressing AEITC noncompliance. We also
reviewed and analyzed documents and reports about IRS, SSA, and Department
of Homeland Security databases about whether they could be used by
employers to verify the SSN of an employee seeking the AEITC before the
employer begins paying it. We also interviewed knowledgeable officials at
IRS and SSA about the advantages and disadvantages of such systems when
considering the AEITC. Further, we interviewed IRS officials from various
offices, such as EITC Program, Modernization and Information Technology
Services, and Stakeholders, Partnership, Education and Communication,
about the advantages and disadvantages of creating a database for the
Forms W-5. It was not within the scope of our work to fully evaluate the
potential cost and benefits of these options for reducing noncompliance.

We also reviewed prior GAO, IRS, Treasury Inspector General for Tax
Administration, and other reports on the AEITC and EITC.

^9We spoke with Leonard Burman, Urban Institute; Leslie Book, Villanova
University School of Law; Jonathan Forman, University of Oklahoma College
of Law; William Gale, The Brookings Institution; Robert Greenstein, Center
on Budget and Policy Priorities; Teresa Hinze, Community Tax Aid, Inc.;
Julie Kruse, Center for Economic Progress; Diana Leyden, University of
Connecticut School of Law; David Marzahl, Center for Economic Progress;
John Karl Scholz, University of Wisconsin Department of Economics; and
Eugene Steuerle, Urban Institute.

We conducted our work primarily in Washington, D.C., and Atlanta, Ga.,
from December 2005 through July 2007 in accordance with generally accepted
government auditing standards.

Appendix III: Demographic Characteristics of Advance Earned Income Tax
Credit Recipients and Their Employers

We identified basic demographic characteristics of Advance Earned Income
Tax Credit (AEITC) recipients and their employers in tax years 2002
through 2004. Specifically, we identified the following 11
characteristics: (1) number of Forms W-2 received; (2) average amount of
AEITC received by consecutive recipients; (3) average adjusted gross
income (AGI) for AEITC recipients; (4) average wages; (5) filing status;
(6) age; (7) gender; (8) number of qualifying children; (9) filing method;
(10) geographic location of AEITC recipients; and (11) employer size,
number of employees, and number of Forms W-2 with AEITC issued to
employees. Each of the characteristics represents an analysis and provides
additional objective information about AEITC recipients not previously
discussed. Where possible, we compared AEITC recipients to Earned Income
Tax Credit (EITC) recipients. Further analyses may provide information to
better target IRS enforcement efforts. For example, IRS's information on
EITC associated with gender differs from our AEITC results in table 17 on
gender.

The characteristics are organized by a declining AEITC population size.
For example, table 10, "Number of Forms W-2 Received by Subpopulation, Tax
Years 2002 through 2004," includes the valid, the invalid name, and
invalid number subpopulations, as described in the scope and methodology
(see app. II), while table 15, "Filing Status of AEITC and EITC
Recipients, Number and Percentage, Tax Years 2002 through 2004," includes
only the valid and invalid name subpopulations. The invalid number
subpopulation was not included in the table about filing status because
that subpopulation contains only Form W-2 data and not tax return data.
Thus, information such as filing status, which comes from the tax return,
is not available.

Number of Forms W-2 with AEITC that each individual received: Most
individuals who received the AEITC only had one Form W-2 reporting its
receipt (see table 10). Having more than one Form W-2 does not necessarily
indicate noncompliance because an individual may have more than one job
during the year and receive the AEITC from more than one employer. The
data did not enable us to analyze whether any of these individuals had
more than one Form W-5 in effect at one time. Presently, IRS's
administrative procedures do not enable it to identify whether the
taxpayer has more than one Form W-5 in effect at one time.

Table 10: Number of Forms W-2 Received by Subpopulation, Tax Years 2002
through 2004

                                        Tax year 2002
Number of                                Invalid name       Invalid number 
Forms W-2    Valid subpopulation        subpopulation        subpopulation 
1                        370,350               59,660               22,812 
2                         35,821               13,275                6,233 
3                          8,056                4,438                1,954 
4                          2,671                1,829                  874 
5                          1,040                  704                  332 
6                            408                  327                  142 
7                            202                  138                   74 
8                            109                   65                   37 
9                             48                   32                   10 
10                            33                   18                    7 
11                            16                   11                    7 
12                             5                    8                    3 
13                             4                    7                    4 
14                             2                    1                    1 
15                             2                    3                    1 
16                             2                    0                    2 
17                             2                    2                    1 
18                             2                    1                    0 
19                             1                    1                    2 
20                             0                    1                    1 
21                             0                    0                    0 
22                             0                    0                    0 
23                             0                    0                    2 
25                             0                    1                    0 
26                             0                    0                    0 
27                             0                    1                    0 
28                             0                    0                    1 
30                             0                    0                    0 
31                             0                    0                    0 
33                             0                    1                    0 
34                             0                    0                    0 
38                             0                    0                    0 
643                            0                    0                    0 

              Tax year 2003                             Tax year 2004
                                  Invalid                                   Invalid 
        Valid  Invalid name        number         Valid  Invalid name        number 
subpopulation subpopulation subpopulation subpopulation subpopulation subpopulation 
      353,284        56,863        22,474       349,544        62,422        24,853 
       33,440        11,599         5,424        35,518        10,515         4,776 
        7,645         3,514         1,660         8,568         3,040         1,452 
        2,417         1,299           600         2,735         1,078           483 
          907           512           259         1,032           382           187 
          394           238            94           410           149            57 
          196            99            41           258            77            27 
          104            38            22           107            33            13 
           63            18            17            65            17             8 
           33            22             6            42             8             4 
           17             5             4            19             5             5 
           14             6             5            19             4             1 
            6             5             5             8             1             2 
            5             2             3             2             0             0 
            1             3             1             1             3             1 
            0             3             1             1             5             0 
            0             3             1             0             1             0 
            0             2             0             2             1             0 
            0             0             0             0             0             1 
            0             0             1             2             3             1 
            0             0             0             0             1             0 
            0             0             0             1             0             1 
            0             1             0             0             0             0 
            0             0             0             1             1             0 
            0             0             1             0             0             1 
            0             1             0             0             0             0 
            0             0             1             0             0             0 
            0             1             0             0             0             0 
            0             0             1             0             0             0 
            0             0             0             0             0             0 
            0             0             0             0             1             0 
            0             0             0             0             0             1 
            0             0             1             0             0             0 

                                        Tax year 2002
Number of                                Invalid name       Invalid number 
Forms W-2    Valid subpopulation        subpopulation        subpopulation 
935                            0                    0                    1 
1088                           0                    0                    0 
Totals                   418,774               80,524               32,501 

              Tax year 2003                             Tax year 2004
                                  Invalid                                   Invalid 
        Valid  Invalid name        number         Valid  Invalid name        number 
subpopulation subpopulation subpopulation subpopulation subpopulation subpopulation 
            0             0             0             0             0             0 
            0             0             0             0             0             1 
      398,526        74,234        30,622       398,335        77,747        31,875 

Source: GAO analysis of IRS data.

Note: This table includes the valid, invalid name, and invalid number
subpopulations. About 20 percent of the combined invalid name and invalid
number subpopulations were noncompliant (see table 4).

Average amount of AEITC received by consecutive recipients: About 98,000
individuals received the AEITC consecutively during tax years 2002 through
2004. These 98,000 individuals received a higher average amount of AEITC
than the entire AEITC population (see fig. 6).

Figure 6: Average Amount of AEITC Received by Consecutive and AEITC
Recipients, Tax Years 2002 through 2004

Note: This figure includes the valid, invalid name, and invalid number
subpopulations.

Adjusted gross income (AGI) for AEITC and EITC recipients: The maximum
amount of AGI a taxpayer could have in tax years 2002 through 2004 and
receive the AEITC and/or EITC was $34,178, $34,692 and $35,458,
respectively. As noted in tables 11, 12, and 13, most individuals who
received the AEITC and filed a tax return reported an AGI of $1--$20,000.
Some taxpayers had an AGI above the allowable limits. However, because an
individual's personal circumstance may have changed during the year, for
example the individual may have gotten a higher paying job, as long as the
same amount of AEITC received as shown on the Form W-2 was reported on the
tax return, AGI outside the limit for AEITC recipients is permissible and
the taxpayer is considered compliant. By reporting the correct amount on
the tax return, the AEITC would increase the tax due or reduce any refund.

Table 11: AGI for AEITC and EITC Recipients, Number, Average, and Sum, for
Tax Year 2002

                                                   Tax year 2002
                                     AEITC                                             EITC
              Valid subpopulation         Invalid name subpopulation              
Adjusted                                                                                                    
Gross                                                                                                       
income    Number Average            Sum Number  Average            Sum     Number  Average              Sum 
Less       2,328 -$4,602   -$10,713,330    851 -$38,560   -$32,814,646    135,881 -$40,267  -$5,471,560,820 
than or                                                                                                     
equal to                                                                                                    
$0^a                                                                                                        
$1        97,641  $5,459   $533,029,760  7,946   $5,363    $42,617,768  7,360,626   $5,905  $43,461,774,192 
-$10,000                                                                                                    
$10,001   92,795 $14,768 $1,370,422,047  9,236  $14,920   $137,799,876  7,238,561  $14,676 $106,230,907,584 
-$20,000                                                                                                    
$20,001   51,787 $24,053 $1,245,652,310  7,867  $24,847   $195,471,529  5,395,193  $24,549 $132,448,552,653 
-$30,000                                                                                                    
Greater   38,765 $56,298 $2,182,391,376 33,625  $88,847 $2,987,478,924    788,562  $31,806  $25,081,179,444 
than                                                                                                        
$30,000                                                                                                     
Total    283,316 $18,780 $5,320,782,163 59,525  $55,952 $3,330,553,451 20,918,823  $14,425 $301,750,853,053 

Source: GAO analysis of IRS data.

aNegative dollar amounts are possible with an AGI less than or equal to $0
because these returns likely have eligible deductions (such as alimony
payments) that exceed the AGI amount, thus resulting in a negative AGI.
This table includes the valid and invalid name subpopulations.

Table 12: AGI for AEITC and EITC Recipients, Number, Average, and Sum, for
Tax Year 2003

                                                    Tax year 2003
                                     AEITC                                             EITC
              Valid subpopulation         Invalid name subpopulation              
Adjusted                                                                                                     
gross                                                                                                        
income    Number Average            Sum Number  Average            Sum     Number  Average               Sum 
Less       1,402 -$8,980   -$12,589,922    790 -$82,807   -$65,417,790    159,145 -$37,491   -$5,966,471,268 
than or                                                                                                      
equal to                                                                                                     
$0^a                                                                                                         
$1        89,053  $5,485   $488,423,673  7,115   $5,369    $38,201,515  7,497,336   $5,895   $44,194,704,408 
-$10,000                                                                                                     
$10,001   86,420 $14,768 $1,276,286,596  8,436  $14,899   $125,687,374  7,340,828  $14,611 $1,072,555,29,694 
-$20,000                                                                                                     
$20,001   50,197 $24,168 $1,213,157,369  7,290  $24,794   $180,748,114  5,498,853  $24,634  $135,457,333,696 
-$30,000                                                                                                     
Greater   38,894 $57,039 $2,218,488,004 30,991  $94,505 $2,928,794,298    935,215  $31,988   $29,915,739,209 
than                                                                                                         
$30,000                                                                                                      
Total    265,966 $19,490 $5,183,765,720 54,622  $58,731 $3,208,013,511 21,431,377  $14,505  $107,255,529,694 

Source: GAO analysis of IRS data.

aNegative dollar amounts are possible with an AGI less than or equal to $0
because these returns likely have eligible deductions (such as alimony
payments) that exceed the AGI amount, thus resulting in a negative AGI.
This table includes the valid and invalid name subpopulations.

Table 13: AGI for AEITC and EITC Recipients, Number, Average, and Sum, for
Tax Year 2004

                                                    Tax year 2004
                                     AEITC                                              EITC
               Valid subpopulation         Invalid name subpopulation              
Adjusted                                                                                                     
gross                                                                                                        
income    Number  Average            Sum Number  Average            Sum     Number  Average              Sum 
Less         641 -$22,631   -$14,506,202    686 -$61,661   -$42,299,493    177,820 -$28,625  -$5,090,160,585 
than or                                                                                                      
equal to                                                                                                     
$0^a                                                                                                         
$1        86,180   $5,578   $480,670,518  6,988   $5,391    $37,671,007  7,497,833   $5,896  $44,206,887,726 
-$10,000                                                                                                     
$10,001   83,792  $14,716 $1,233,045,113  8,270  $14,914   $123,334,680  7,355,688  $14,562 $107,113,059,413 
-$20,000                                                                                                     
$20,001   49,165  $24,204 $1,189,979,660  7,139  $24,846   $177,374,704  5,519,610  $24,710 $136,390,977,111 
-$30,000                                                                                                     
Greater   34,225  $54,623 $1,869,459,016 32,638 $101,396 $3,309,361,834  1,170,267 $32, 204  $37,687,855,307 
than                                                                                                         
$30,000                                                                                                      
Total    254,003   12,735 $4,758,648,105 55,721  $64,705 $3,605,442,732 21,721,218  $14,746 $320,308,618,972 

Source: GAO analysis of IRS data.

aNegative dollar amounts are possible with an AGI less than or equal to $0
because these returns likely have eligible deductions (such as alimony
payments) that exceed the AGI amount, thus resulting in a negative AGI.
This table includes the valid and invalid name subpopulations.

Wages for AEITC and EITC recipients: The yearly wage limits for AEITC and
EITC recipients were $34,178 for tax year 2002, $34,692 for tax year 2003,
and $35,458 for tax year 2004. The average wages for AEITC recipients in
the valid subpopulation were about $18,000, while they were about $47,000
for the invalid name subpopulation.^1 This compares with about $13,000 for
EITC recipients (see table 14). Some wages are outside the allowable
limits. However, because an individual's personal circumstance may have
changed during the year, for example, the individual may have gotten a
higher paying job, as long as the same amount of AEITC received as shown
on the Form W-2 was reported on the tax return, wages outside the limit
are permissible and the taxpayer is considered compliant. By reporting the
correct amount on the tax return, the AEITC would increase the tax due or
reduce any refund.

^1 The large difference in wages between the AEITC valid and invalid name
subpopulations could be a result of women who married during the year and
changed their name with their employer, but not SSA. These women likely
reported combined wages with their spouse.

Table 14: Wages Reported for AEITC and EITC Recipients, Number, Wage
Amount, and Average Dollars Received for Tax Years 2002 through 2004

                                AEITC                                            EITC
          Valid subpopulation        Invalid name subpopulation              
Tax                                                                                                   
year  Number    Wage amount Average Number    Wage amount Average     Number      Wage amount Average 
2002 283,316 $4,926,169,548 $17,388 59,525 $2,696,220,392 $45,296 20,918,823 $273,167,081,421 $13,058 
2003 265,966 $4,829,307,231 $18,158 54,622 $2,545,765,228 $46,607 21,431,377 $280,828,962,882 $13,104 
2004 254,003 $4,409,967,145 $17,362 55,721 $2,759,417,493 $49,522 21,721,218 $290,119,218,681 $13,356 

Source: GAO analysis of IRS data.

Notes: This table includes the valid and invalid name subpopulations.

Filing status of AEITC and EITC recipients: As shown in table 15, about
half of AEITC recipients in the valid subpopulation and most individuals
who received the EITC used the Head of Household filing status. This
compares to AEITC recipients in the invalid name subpopulation who most
frequently used the Married Filing Jointly filing status. About 2 percent
of the tax returns that reported receiving AEITC used the Married Filing
Separate filing status, which is not allowed. However, an individual's
personal circumstance may have changed during the year, for example the
individual may have separated from their spouse. As long as the same
amount of AEITC received as shown on the Form W-2 was reported on the tax
return, this situation is permissible and the taxpayer is considered
compliant. By reporting the correct amount on the tax return, the AEITC
would increase the tax due or reduce any refund.

Table 15: Filing Status of AEITC and EITC Recipients, Number and
Percentage, Tax Years 2002 through 2004

                                                Tax year 2002
                                             AEITC                  EITC      
                                          Valid   Invalid name                
                                  subpopulation  subpopulation                
Head of household                    140,021         11,180     11,326,654 
                                                                              
                                          (49%)          (19%)          (54%) 
Single                                77,347         13,457      4,495,201 
                                                                              
                                          (27%)          (23%)          (21%) 
Married filing jointly                60,153         33,837      5,068,608 
                                                                              
                                          (21%)          (57%)          (24%) 
Married filing separate                5,546            988          3,042 
                                                                              
                                           (2%)           (2%) (less than 1%) 
Widow(er) with children                  180             40         25,318 
                                                                              
                                 (less than 1%) (less than 1%) (less than 1%) 
Married filing separate (with             69             23            N/A 
spouse not required to file)                                               
                                 (less than 1%) (less than 1%)                
Total                                283,316         59,525     20,918,823 

               Tax year 2003                          Tax year 2004
              AEITC               EITC               AEITC               EITC    
           Valid  Invalid name                    Valid  Invalid name            
subpopulation subpopulation            subpopulation subpopulation            
         133,985        10,255 11,532,582       134,964        10,551 11,659,521 
                                                                                 
           (50%)         (19%)      (54%)         (53%)         (19%)      (54%) 
          68,795        12,355  4,681,467        63,685        12,606  4,914,789 
                                                                                 
           (26%)         (23%)      (22%)         (25%)         (23%)      (23%) 
          58,133        31,043  5,190,469        51,053        31,633  5,119,517 
                                                                                 
           (22%)         (57%)      (24%)         (20%)         (57%)      (24%) 
           4,815           903          0         4,052           862          0 
                                                                                 
            (2%)          (2%)       (0%)          (2%)          (2%)       (0%) 
             187            42     26,859           193            52     27,391 
                                                                                 
      (less than    (less than (less than    (less than    (less than (less than 
             1%)           1%)        1%)           1%)           1%)        1%) 
              51            24        N/A            56            17        N/A 
                                                                                 
      (less than    (less than               (less than    (less than            
             1%)           1%)                      1%)           1%)            
         265,966        54,622 21,431,377       254,003        55,721 21,721,218 

Source: GAO analysis of IRS data.

Notes: EITC totals are the number of taxpayers who received EITC. N/A
means not applicable. This table includes the valid and invalid name
subpopulations.

Age of AEITC and EITC recipients:  Most individuals who received the
AEITC, as well as most EITC recipients, were between the ages of 26 and 64
(see table 16). IRS officials noted that recipients whose age fell into
the `Over 100' category are the result of a probable error, such as a
transcription error. Alternatively, an AEITC recipient in this category
may have used a SSN of a deceased individual.

Table 16: Age of AEITC and EITC Recipients, Tax Years 2002 through 2004

                                        Tax year 2002
                                     AEITC                          EITC      
                                                  Invalid name                
                 Valid subpopulation             subpopulation                
Age 25 and                 82,810                    29,010      3,259,332 
under                                                                      
                               (20%)                     (36%)          (16%) 
Age 26 to 64              333,519                    31,202     17,345,338 
                                                                              
                               (80%)                     (39%)          (83%) 
Age 65 to 100               2,409                    14,275        242,838 
                                                                              
                                (1%)                     (18%)           (1%) 
Over 100                       35                     6,037         71,315 
                                                                              
                      (less than 1%)                      (8%) (less than 1%) 
Total                     418,773                    80,524     20,918,823 

               Tax year 2003                          Tax year 2004
              AEITC               EITC               AEITC               EITC    
           Valid  Invalid name                    Valid  Invalid name            
subpopulation subpopulation            subpopulation subpopulation            
          78,215        26,035  3,289,587        77,096        26,499  3,348,578 
                                                                                 
           (20%)         (35%)      (15%)         (19%)         (34%)      (15%) 
         317,997        29,300 17,613,301       318,683        31,908 17,844,506 
                                                                                 
           (80%)         (40%)      (82%)         (80%)         (41%)      (82%) 
           2,278        13,033    261,347         2,514        13,270    274,428 
                                                                                 
      (less than         (18%)       (1%)          (1%)         (17%)       (1%) 
             1%)                                                                 
              35         5,866    267,142            42         6,070    253,706 
                                                                                 
      (less than          (8%)       (1%)    (less than          (8%)       (1%) 
             1%)                                    1%)                          
         398,525        74,234 21,431,377       398,335        77,747 21,721,218 

Source: GAO analysis of IRS data.

Notes: This table includes the valid and invalid name subpopulations. IRS
calculated age by subtracting the birth year from the analysis year.
Neither IRS nor GAO could identify why the tax years 2002 and 2003 AEITC
valid Forms W-2 SSN total is one below the total number of individuals who
received a Form W-2 reporting AEITC.

Gender of AEITC and EITC recipients: More males than females received the
AEITC during tax years 2002 through 2004. In contrast, more females than
males received the EITC during this same period (see table 17).

Table 17: Gender of AEITC and EITC Recipients, Number and Percentage, Tax
Years 2002 through 2004

                                   Tax year 2002
                               AEITC                           EITC      
           Valid subpopulation Invalid name subpopulation                
Female              165,694                     36,805     11,087,655 
                                                                         
                         (40%)                      (46%)          (53%) 
Male                252,922                     39,908      9,822,392 
                                                                         
                         (60%)                      (50%)          (47%) 
Unknown                 158                      3,811          8,776 
                                                                         
                (less than 1%)                       (5%) (less than 1%) 
Total               418,774                     80,524     20,918,823 
                                                                         
                        (100%)                     (100%)         (100%) 

               Tax year 2003                          Tax year 2004
              AEITC               EITC               AEITC               EITC    
           Valid  Invalid name                    Valid  Invalid name            
subpopulation subpopulation            subpopulation subpopulation            
         159,681        34,019 11,277,324       165,480        35,904 11,525,243 
                                                                                 
           (40%)         (46%)      (53%)         (42%)         (46%)      (53%) 
         238,683        36,762 10,148,364       232,690        38,472 10,190,319 
                                                                                 
           (60%)         (50%)      (47%)         (58%)         (49%)      (47%) 
             162         3,453      5,689           165         3,371      5,656 
                                                                                 
      (less than          (5%) (less than    (less than          (4%) (less than 
             1%)                      1%)           1%)                      1%) 
         398,526        74,234 21,431,377       398,335        77,747 21,721,218 
                                                                                 
          (100%)        (100%)     (100%)        (100%)        (100%)     (100%) 

Source: GAO analysis of IRS data.

Notes: This table includes the valid and invalid name subpopulations. We
reported gender information only for individuals who filed a federal tax
return. Percentages may not add due to rounding.

Number of qualifying children for AEITC and EITC recipients: About half of
the individuals who reported receiving the AEITC had two qualifying
children. Results were similar for EITC recipients. However, about 8
percent of this population did not report having any qualifying children,
which is not allowed (see table 18). However, an individual's personal
circumstance may have changed during the year, for example, the individual
may have separated from their spouse or divorced. As long as the same
amount of AEITC received as shown on the Form W-2 was reported on the tax
return, the taxpayer is considered compliant. By reporting the correct
amount on the tax return, the AEITC would increase the tax due or reduce
any refund. In contrast, most individuals who received the EITC also had
two qualifying children. There is no qualifying child requirement to
receive the EITC.

Table 18: Number of Qualifying Children for AEITC and EITC Recipients,
Number and Percentage, Tax Years 2002 through 2004

                       Tax year 2002      Tax year 2003      Tax year 2004
                       AEITC       EITC   AEITC       EITC   AEITC       EITC 
Number of           7,946  3,765,072   7,456  4,022,684   7,786  4,181,382 
taxpayers with no                                                          
qualifying           (8%)      (18%)    (7%)      (19%)    (8%)      (19%) 
children                                                                   
(percentage)                                                               
Number of          43,066  8,198,533  43,129  8,269,050  40,379  8,354,539 
taxpayers with 1                                                           
qualifying child    (41%)      (39%)   (41%)      (39%)   (40%)      (38%) 
(percentage)                                                               
Number of          53,721  8,955,218  55,558  9,139,643  52,448  9,185,297 
taxpayers with 2                                                           
qualifying          (51%)      (43%)   (52%)      (43%)   (52%)      (42%) 
children                                                                   
(percentage)                                                               
Total             104,733 20,918,823 106,143 21,431,377 100,613 21,721,218 
(percentage)                                                               
                      (100%)     (100%)  (100%)     (100%)  (100%)     (100%) 

Source: GAO analysis of IRS data.

Note: This table is based on the number of tax returns in the EITC
database that had AEITC greater than $0 and zero, one, or two qualifying
children.

Filing method of AEITC and EITC recipients: As noted in table 19, most tax
returns that showed receiving an amount of AEITC were filed
electronically, as were most tax returns which reported receipt of the
EITC. For both AEITC and EITC, about 70 percent of recipients filed
electronically for the 3 years we examined.

Table 19: Filing Method of AEITC and EITC Recipients, Number and
Percentage, Tax Years 2002 through 2004

                Tax year 2002      Tax year 2003      Tax year 2004
                AEITC       EITC   AEITC       EITC   AEITC       EITC 
Electronic  93,291 14,030,564 100,085 15,322,251  98,592 16,269,437 
                                                                       
                (68%)      (67%)   (72%)      (71%)   (78%)      (75%) 
Paper       44,477  6,888,259  38,885  6,109,126  28,228  5,451,781 
                                                                       
                (32%)      (33%)   (28%)      (29%)   (22%)      (25%) 
Total      137,768 20,918,823 138,970 21,431,377 126,820 21,721,218 
                                                                       
               (100%)     (100%)  (100%)     (100%)  (100%)     (100%) 

Source: GAO analysis of IRS data.

Note: Tax years 2002 and 2003 include 5 and 4 duplicate Taxpayer
Identification Numbers (TIN), respectively. The AEITC totals shown above
do not equal the total number of individuals who filed a federal tax
return for that year. This table only includes the number of tax returns
that had an AEITC amount greater than $0. IRS did not identify the filing
method used by individuals who filed a tax return but failed to report
receipt of the AEITC. The total number of individuals who filed a federal
tax return was 338,942, 316,959, and 307,251, for tax years 2002 through
2004 respectively. The EITC total represents the total number of taxpayers
claiming EITC.

Geographic location of AEITC recipients in the valid subpopulation: Of the
individuals in the valid subpopulation, use of the AEITC varied widely
across the country. In all 3 tax years, Florida and Illinois had the most
AEITC recipients (see fig. 7).

Figure 7: Geographic Location of AEITC Recipients in the Valid
Subpopulation, Tax Years 2002 through 2004

Note: AEITC geographic location is based upon information reported on the
tax return. This figure includes the valid subpopulation.

aThese geographic locations had at least 1 but fewer than 10 individuals
who received the AEITC.

bU.S insular areas include the three major U.S. territories, which are the
U.S. Virgin Islands, American Samoa, Guam, and the two Commonwealths,
which are the Commonweath of Puerto Rico and the Commonwealth of Northern
Mariana Islands.

cOther includes Federated States of Micronesia, Marshall Islands, Palau,
and Armed Forces mostly located outside the U.S.

dMissing refers to instances where the geographic location was not
available.

Employer size, number of employers, number of forms W-2 employers issued
to AEITC recipients and Total AEITC Dollars Reported on Forms W-2:
Slightly more than 50,000 employers reported paying at least one employee
AEITC in each tax years 2002 through 2004. Most of these employers were
classified by IRS as small business/self employed and they issued more
than half of all the Forms W-2 with AEITC (see table 20).

Table 20: Employer Size, Number of Employers, Number of Forms W-2
Employers Issued to AEITC Recipients and Total Dollars Reported on Forms
W-2, Tax Years 2002 through 2004

                       Tax Year 2002                  Tax Year 2003                  Tax Year 2004
                          Number                         Number                         Number              
                              of                             of                             of              
                           Forms                          Forms                          Forms              
                             W-2  Total AEITC               W-2  Total AEITC               W-2  Total AEITC 
                          issued      Dollars            issued      Dollars            issued      Dollars 
               Number of    with  Reported on Number of    with  Reported on Number of    with  Reported on 
Employer size  employers   AEITC    Forms W-2 employers   AEITC    Forms W-2 employers   AEITC    Forms W-2 
Small             30,004 368,112  $72,779,073    29,562 347,942  $73,120,770    30,296 351,710  $77,689,921 
business/self                                                                                               
employed           (57%)   (56%)        (50%)     (57%)   (57%)        (51%)     (59%)   (57%)        (52%) 
Tax exempt and    13,416  62,869  $29,103,769    12,976  57,856  $25,830,428    12,609  45,301  $24,715,548 
government                                                                                                  
entities           (26%)   (10%)        (20%)     (25%)   (10%)        (18%)     (24%)    (7%)        (16%) 
Large and          9,058 227,029  $42,551,520     8,715 202,459  $43,615,497     8,517 219,684  $47,612,524 
midsize                                                                                                     
business           (17%)   (35%)        (29%)     (17%)   (33%)        (30%)     (17%)   (36%)        (32%) 
Unknown or            89     656     $112,644       174   3,610     $639,206        72     143      $66,274 
unidentifiable                                                                                              
                   (less   (less   (less than     (less    (1%)   (less than     (less   (less   (less than 
                than 1%)    than          1%)  than 1%)                  1%)  than 1%)    than          1%) 
                             1%)                                                           1%)              
Totals            52,567 658,666 $144,547,006    51,427 611,867 $143,205,901    51,494 616,838 $150,084,267 

Source: GAO analysis of IRS data.

Note: Employers are defined by unique Employer Identification Numbers
(EIN) and the classification of employer size is based on IRS criteria.
Number of employers includes the valid, invalid name, and invalid number
subpopulations. Number of Forms W-2 issued with AEITC includes all Forms
W-2 issued that year reporting an amount of AEITC. Percentages may not add
due to rounding.

Appendix IV: Administrative and Legislative Changes to the Advance Earned
Income Tax Credit Since 1990

IRS has made several administrative changes to the Advance Earned Income
Tax Credit (AEITC) since the beginning of 1990 (see table 21). IRS
described these changes in responses to recommendations in our 1992 AEITC
report and a 2003 Treasury Inspector General for Tax Administration
report.^1

Table 21: Administrative Changes to the AEITC since 1990

Year      Change                                                           
1990      Added separate lines to the 1990 Forms 1040 and 1040A for        
             reporting AEITC received.                                        
1991      Revised Form W-4 to advise employees of the availability of the  
             AEITC and to refer them to the Form W-5.                         
1991-1992 Revised Form W-2 and IRS Notice 797 to contain information on    
             how to apply for the AEITC.                                      
1992      Developed publicity materials and distributed them to the        
             public; provided credit information in the Small Business        
             Taxpayer Education Program; and increased outreach speaker       
             seminar efforts, in cooperation with federal, state, and local   
             officials and private organizations.                             
1992      Developed print and audiovisual products to inform the public    
             about the AEITC.                                                 
1992      Expanded instructions on AEITC in "Circular E, Employer's Tax    
             Guide."                                                          
1992      Added examples to "Circular E, Employer's Tax Guide," explaining 
             how to advance the EITC to employees and how to report amounts   
             in advance.                                                      
1992      Clarified employer's responsibilities and liabilities in         
             advancing the Earned Income Tax Credit (EITC) in the 1993        
             "Circular E, Employer's Tax Guide."                              
1993      Expanded 1993 Form W-2 instructions concerning AEITC.            
             Specifically, revised the Form W-2 to place a bold outline       
             around the AEITC box of the employee's copy.                     
2003      Updated instructions that require tax examiners to compare the   
             amount of AEITC shown on the tax return to the amount listed as  
             AEITC on the Form W-2 and adjust the amounts when needed.        
2004      Created an error condition in return processing when the amount  
             of AEITC reported exceeds the maximum allowed for the year.      
2005      Created an error condition in return processing for whenever     
             AEITC is listed on the return without a corresponding entry for  
             the EITC.                                                        

Source: GAO analysis.

^1 [39]GAO/GGD-92-26 and Treasury Inspector General for Tax
Administration, Taxpayers Were Assessed Additional Tax for Advance Earned
Income Credit Payments Not Received, 2003-40-126 (June 2003).

Notes: Changes made between 1990 and 1993 were a result of GAO's 1992
report and changes between 2003 and 2005 were a result of the Treasury
Inspector General for Tax Administration's 2003 report. Some of the
changes made between 1990 and 1993 occurred before our report was issued.

There have been two laws enacted since our 1992 report that include
specific changes to the AEITC.^2 First, the Omnibus Budget Reconciliation
Act of 1993 (OBRA '93), did the following.

           1. Limited the amount of advance payment allowable in a taxable
           year to 60 percent of the maximum credit available to a taxpayer
           with one qualifying child.
           2. Directed the IRS to notify taxpayers with qualifying children
           who receive a refund on account of the EITC that the credit may be
           available on an advance basis. The conference report accompanying
           OBRA '93 stated that after these notifications had been made for 2
           taxable years, the Treasury Secretary was directed to study their
           effect on utilization of the advance payment mechanism and, based
           on the results of the study, the Secretary may recommend
           modifications to the notification program.^3

Second, the U.S. Troop Readiness, Veteran's Care, Katrina Recovery, and
Iraq Accountability Appropriations Act of 2007, which was enacted in late
May 2007, calls for a study of AEITC use.^4 The study is to be conducted
by the Secretary of the Treasury for the Congress and is to include the
benefits, costs, risks, and barriers to workers and to businesses (with a
special emphasis on small businesses) if the AEITC included all recipients
of the EITC (i.e., individuals without qualifying children). It also asks
what steps would be necessary to implement such an inclusion.

^2 We only included laws that made specific changes to the AEITC or its
administration that were independent of changes to the EITC. Thus, we did
not include laws whose only change to the AEITC occurred because of
changes to the broader EITC program, such as changes in EITC benefit
amounts or eligibility rules.

^3 H.R. Conf. Rep. No. 103-213 (1993).

^4 Pub. L. No. 110-28.

Appendix V: Additional Analyses of Advance Earned Income Tax Credit
Noncompliance

We identified additional areas of noncompliance for AEITC recipients
during tax years 2002 through 2004. We also examined demographic
characteristics of our noncompliant subpopulations, including the invalid
name, invalid number, and dollar limit subpopulations. Our analyses
revealed the following: (1) some consecutive AEITC recipients had an
invalid SSN, (2) most consecutive AEITC recipients filed a tax return, but
did not report the correct AEITC amount on the tax return, (3) AEITC
recipients with an invalid SSN received little money, (4) most AEITC
recipients with an invalid SSN received one to two Forms W-2, (5) AEITC
recipients in the invalid subpopulations lived in various geographic
locations, and (6) AEITC was paid in excess of yearly maximum limits. As
with the data in Appendix III, further analyses of this data may provide
information on noncompliance characteristics potentially useful for IRS
enforcement efforts.

Each of the tables provides additional information about AEITC recipients
not previously discussed. Where possible, we compared AEITC recipients to
EITC recipients. The characteristics are organized by a declining AEITC
population size, similar to the organization in appendix III. Most
analyses include the invalid name and invalid number subpopulations. The
subpopulations are described in the scope and methodology section (see
app. II).

Some consecutive AEITC recipients had an invalid SSN: About 98,000
individuals received the AEITC consecutively in each of the 3 years, tax
years 2002 through 2004, and received an average of about $56 million.
About a quarter of these individuals had an invalid SSN and received an
average of approximately $16 million in AEITC. Additionally, nearly 15
percent of consecutive users had an invalid SSN and did not file a federal
tax return, receiving an average of about $9 million in AEITC (see table
22).

Table 22: Number of Individuals and AEITC Dollars Received by AEITC
Recipients and Individuals Who Elected the AEITC Consecutively Using an
Invalid SSN and an Invalid SSN and Not Filing, Tax Years 2002 through 2004

                    Tax year 2002            Tax year 2003            Tax Year 2004
               Individuals      Dollars Individuals      Dollars Individuals      Dollars 
AEITC              531,799 $144,547,006     503,382 $143,205,901     507,957 $150,084,267 
recipients                                                                                
AEITC               97,998  $53,272,152      97,998  $60,849,730      97,998  $52,698,947 
consecutive                                                                               
recipients^a         (18%)        (37%)       (19%)        (42%)       (19%)        (35%) 
AEITC               23,175  $14,837,041      23,175  $17,217,171      23,175  $14,798,025 
consecutive                                                                               
recipients           (24%)        (28%)       (24%)        (28%)       (24%)        (28%) 
with invalid                                                                              
SSN                                                                                       
(percentage)^b                                                                            
AEITC               13,181   $8,593,857      13,153   $9,972,631      13,348   $8,722,592 
consecutive                                                                               
recipients           (13%)        (16%)       (13%)        (16%)       (14%)        (17%) 
with invalid                                                                              
SSN and did                                                                               
not file                                                                                  
federal tax                                                                               
return                                                                                    
(percentage)^b                                                                            

Note: The number of AEITC consecutive recipients represents the total
number of individuals who received the AEITC over each of the 3 years. The
associated dollars represent the amount those individuals received each
year.

aPercentage is out of AEITC recipients and dollars. This analysis includes
the valid, invalid name, and invalid number subpopulations.

bPercentage is out of AEITC consecutive recipients and dollars. This
analysis includes the invalid name and invalid number subpopulations.

Most consecutive AEITC recipients filed a tax return, but did not report
the correct AEITC amount on the tax return: About 98,000 individuals
received the AEITC consecutively in each of the 3 tax years, 2002 through
2004. More than half of these individuals filed a tax return. Of those who
filed, about half reported the same AEITC amount on the tax return as
shown on the Form W-2 (i.e., matched). Of the mismatches, the majority did
not report receipt of the AEITC (see table 23).

Advance Earned Income Tax Credit Advance Earned Income Tax Credit Advance
Earned Income Tax Credit Advance Earned Income Tax Credit Advance Earned
Income Tax Credit Advance Earned Income Tax Credit Advance Earned Income
Tax Credit Advance Earned Income Tax Credit Advance Earned Income Tax
Credit Advance Earned Income Tax Credit Advance Earned Income Tax Credit
Advance Earned Income Tax Credit Advance Earned Income Tax Credit Advance
Earned Income Tax Credit Advance Earned Income Tax Credit Advance Earned
Income Tax Credit Advance Earned Income Tax Credit Advance Earned Income
Tax Credit

Table 23: Number of AEITC Consecutive Recipients with Matches and
Mismatches between Forms W-2 and Filed Tax Returns, Tax Years 2002 through
2004

                       Tax year 2002      Tax year 2003      Tax year 2004
                     Number     Dollars Number     Dollars Number     Dollars 
Number of         65,864 $42,750,223 63,189 $43,205,414 60,257 $36,916,220 
consecutive AEITC                                                          
recipients who                                                             
filed federal tax                                                          
returns                                                                    
Number of returns 29,017 $21,611,779 29,606 $24,369,871 28,478 $20,706,538 
that matched the                                                           
Form(s) W-2 AEITC  (44%)       (51%)  (47%)       (56%)  (47%)       (56%) 
amount for                                                                 
consecutive                                                                
recipients                                                                 
                                                                              
(percentage)^a                                                             
Number of         36,847 $21,138,444 33,583 $18,835,543 31,779 $16,209,682 
mismatches                                                                 
between Form(s)    (56%)       (49%)  (53%)       (44%)  (53%)       (44%) 
W-2 and filed tax                                                          
returns for                                                                
consecutive                                                                
recipients                                                                 
                                                                              
(percentage)^a                                                             
Number of         34,400 $19,751,401 31,035 $17,182,551 29,675 $14,922,931 
mismatches due to                                                          
nonreporting of     (93%       (93%)  (92%)       (91%)  (93%)       (92%) 
AEITC for                                                                  
consecutive                                                                
recipients                                                                 
                                                                              
(percentage)^b                                                             
Number of          1,567    $956,176  1,688  $1,164,494  1,352    $843,205 
mismatches due to                                                          
underreporting of   (4%)        (5%)   (5%)        (6%)   (4%)        (5%) 
AEITC for                                                                  
consecutive                                                                
recipients                                                                 
                                                                              
(percentage)^b                                                             
Number of            880    $430,867    860    $488,498    752    $443,546 
mismatches due to                                                          
overreporting       (2%)        (2%)   (3%)        (3%)   (2%)        (3%) 
AEITC                                                                      
                                                                              
(percentage)^b                                                             

Source: GAO analysis of IRS data.

Note: Percentages may not add due to rounding. This table includes
individuals who filed a tax return in the valid and invalid name
subpopulations.

aWe considered a match to be anything plus or minus a dollar on the Form
W-2 in order to allow for taxpayer rounding. We tested the sensitivity of
this result by using a difference between the AEITC reported on the Form
W-2 and the tax return of up to $100 and the results were similar.
Percentage is of number of consecutive AEITC recipients who filed federal
tax returns.

bPercentage is of number of mismatches due to nonreporting of AEITC for
consecutive recipients and number of consecutive AEITC recipients who
filed federal tax returns.

AEITC recipients with an invalid Social Security number (SSN) received
little money: Most AEITC recipients who had a Form(s) W-2 with an invalid
SSN obtained $100 or less of AEITC (see table 24). These data are
consistent with the overall AEITC population, as previously noted, where
about half of all recipients received less than $100 and 80 percent
received $500 or less for the year.

Table 24: Amount of AEITC Received Per Invalid Form W-2, Number and
Percentage, Tax Years 2002 through 2004

                          Tax year 2002     Tax year 2003     Tax year 2004   
AEITC amount         Count(percentage) Count(percentage) Count(percentage) 
$1- $100                        14,004            12,658            13,292 
                                                                              
                                    (43%)             (41%)             (42%) 
$101- $200                       4,075             3,818             3,817 
                                                                              
                                    (13%)             (12%)             (12%) 
$201 - $300                      2,651             2,471             2,605 
                                                                              
                                     (8%)              (8%)              (8%) 
$301- $400                       2,006             1,888             1,881 
                                                                              
                                     (6%)              (6%)              (6%) 
$401 - $500                      1,521             1,360             1,433 
                                                                              
                                     (5%)              (4%)              (4%) 
$501- $600                       1,172             1,147             1,218 
                                                                              
                                     (4%)              (4%)              (4%) 
$601- $700                         952               953               929 
                                                                              
                                     (3%)              (3%)              (3%) 
$701- $800                         848               809               891 
                                                                              
                                     (3%)              (3%)              (3%) 
$801- $900                         820               727               804 
                                                                              
                                     (3%)              (2%)              (3%) 
$901- $1,000                       705               718               710 
                                                                              
                                     (2%)              (2%)              (2%) 
$1,001 - $1,100                    649               690               691 
                                                                              
                                     (2%)              (2%)              (2%) 
$1,101 - $1,200                    657               634               624 
                                                                              
                                     (2%)              (2%)              (2%) 
$1,201 - $1,300                    634               665               652 
                                                                              
                                     (2%)              (2%)              (2%) 
$1,301 - $1,400                    621               709               731 
                                                                              
                                     (2%)              (2%)              (2%) 
$1,401 - yearly                    577               955             1,242 
maximum                                                                    
                                     (2%)              (3%)              (4%) 
Above maximum ^a                   609               420               355 
                                                                              
                                     (2%)              (1%)              (1%) 
Total                           32,501            30,622            31,875 

Source: GAO analysis of IRS data.

Note: This table includes the invalid number subpopulation.

aAbove the maximum is possible because this table presents aggregated
Forms W-2 and if an individual received more than one Form W-2, the total
received could be above the yearly maximum.

Most AEITC recipients with an invalid SSN received 1 to 2 Forms W-2: Most
recipients who had an invalid SSN received 1 to 2 Forms W-2 reporting
AEITC. For example, in tax year 2002, 6,223 individuals received two Forms
W-2 that reported AEITC. These resulted in a total of 12,466 Forms W-2
equaling $3,261,327 in AEITC, with an average of $262 in AEITC per Form
W-2 (see tables 25, 26, and 27).

Table 25: Number of Individuals Receiving a Form(s) W-2, Total Number of
Form(s) W-2 Reporting AEITC, Total Associated and Average Dollars Received
Relative to Number of Form(s) W-2 Reporting AEITC for AEITC Recipients
with an Invalid SSN on the Form(s) W-2, Tax Year 2002

              A                    B             C            D         E     
                                  Number of                                   
                                individuals Total number       Total  Average 
Number of Forms W-2 an     receiving the of Forms W-2  associated  dollars 
individual received      number of Forms    reporting  dollars on per Form 
reporting AEITC                      W-2      AEITC^a   Forms W-2    W-2^b 
1                                 22,812       22,812  $5,588,671     $245 
2                                  6,233       12,466  $3,261,327     $262 
3                                  1,954        5,862  $1,290,955     $220 
4                                    874        3,496    $689,595     $197 
5                                    332        1,660    $287,912     $173 
6                                    142          852    $138,384     $162 
7                                     74          518     $84,688     $163 
8                                     37          296     $46,962     $159 
9                                     10           90     $15,429     $171 
10                                     7           70      $9,601     $137 
11                                     7           77     $12,577     $163 
12                                     3           36      $6,073     $169 
13                                     4           52     $11,805     $227 
14                                     1           14      $6,246     $446 
15                                     1           15      $5,156     $344 
16                                     2           32      $8,271     $258 
17                                     1           17      $3,893     $229 
19                                     2           38     $10,695     $281 
20                                     1           20      $6,978     $349 
23                                     2           46      $8,101     $176 
28                                     1           28      $6,074     $217 
935                                    1          935     $83,211      $89 
Total Forms W-2 with              32,501       49,432 $11,582,604     $234 
invalid SSNs                                                               

Source: GAO analysis of IRS data.

Note: This table includes the invalid number subpopulation.

aThis number is derived by multiplying column A by column B.

bThis amount is derived by dividing column D by column C.

Table 26: Number of Individuals Receiving a Form(s) W-2, Total Number of
Form(s) W-2 Reporting AEITC, Total Associated and Average Dollars
Received, Relative to Number of Form(s) W-2 Reporting AEITC for AEITC
Recipients with an Invalid SSN on the Form(s) W-2, Tax Year 2003

              A                    B             C           D          E     
                                  Number of                                   
                                individuals Total number      Total   Average 
Number of Forms W-2 an     receiving the of Forms W-2 associated   dollars 
individual received      number of Forms    reporting dollars on per Forms 
reporting AEITC                      W-2      AEITC^a  Forms W-2     W-2^b 
1                                 22,474       22,474 $6,787,458      $302 
2                                  5,424       10,848 $2,689,186      $248 
3                                  1,660        4,980 $1,076,219      $216 
4                                    600        2,400   $478,420      $199 
5                                    259        1,295   $222,288      $172 
6                                     94          564   $107,704      $191 
7                                     41          287    $46,869      $163 
8                                     22          176    $32,510      $185 
9                                     17          153    $30,123      $197 
10                                     6           60    $14,293      $238 
11                                     4           44     $9,769      $222 
12                                     5           60    $14,794      $247 
13                                     5           65    $20,860      $321 
14                                     3           42    $10,857      $259 
15                                     1           15     $1,798      $120 
16                                     1           16     $4,378      $274 
17                                     1           17    $10,284      $605 
19                                     0            0         $0        $0 
20                                     1           20     $4,468      $223 
22                                     0            0         $0        $0 
23                                     0            0         $0        $0 
26                                     1           26     $8,584      $330 
28                                     1           28     $9,513      $340 
31                                     1           31     $9,404      $303 
38                                     0            0         $0        $0 
643                                    1          643    $60,948       $95 
935                                    0            0         $0         0 
1,088                                  0            0         $0         0 
Total                             30,622       44,244 11,650,727      $263 

Source: GAO analysis of IRS data.

Note: This table includes the invalid number subpopulation.

aThis number is derived by multiplying column A by column B.

bThis amount is derived by dividing column D by column C.

Table 27: Number of Individuals Receiving a Form(s) W-2, Total Number of
Form(s) W-2 Reporting AEITC, Total Associated and Average Dollars
Received, Relative to Number of Form(s) W-2 Reporting AEITC for AEITC
Recipients with an Invalid SSN on the Form(s) W-2, Tax Year 2004

              A                    B             C            D         E     
                                  Number of                                   
                                individuals Total number       Total  Average 
Number of Forms W-2 an     receiving the of Forms W-2  associated  dollars 
individual received      number of Forms    reporting  dollars on per Form 
reporting AEITC                      W-2      AEITC^a   Forms W-2    W-2^b 
1                                 24,853       24,853  $7,602,238     $306 
2                                  4,776        9,552  $2,663,173     $279 
3                                  1,452        4,356  $1,022,262     $235 
4                                    483        1,932    $385,252     $199 
5                                    187          935    $178,214     $191 
6                                     57          342     $72,679     $213 
7                                     27          189     $47,036     $249 
8                                     13          104     $30,029     $289 
9                                      8           72     $15,628     $217 
10                                     4           40      $9,060     $227 
11                                     5           55      $9,027     $164 
12                                     1           12      $3,178     $265 
13                                     2           26      $9,049     $348 
14                                     0            0          $0       $0 
15                                     1           15      $5,077     $338 
16                                     0            0          $0       $0 
17                                     0            0          $0       $0 
19                                     1           19      $5,604     $295 
20                                     1           20      $9,567     $478 
22                                     1           22      $5,076     $231 
23                                     0            0          $0       $0 
26                                     1           26      $6,837     $263 
28                                     0            0          $0       $0 
31                                     0            0          $0       $0 
38                                     1           38      $9,575     $252 
643                                    0            0          $0       $0 
935                                    0            0          $0       $0 
1,088                                  1        1,088     $78,141      $72 
Total                             31,875       43,696 $12,166,702     $278 

Source: GAO analysis of IRS data.

Note: This table includes the invalid number subpopulation.

aThis number is derived by multiplying column A by column B.

bThis amount is derived by dividing column D by column C.

AEITC recipients in the invalid subpopulations lived in various geographic
locations: Use of the AEITC varied widely across the country for
individuals in the invalid number and invalid name subpopulations. For the
invalid name subpopulation, in all 3 tax years, California and Illinois
had the most Forms W-2 reporting AEITC and for the invalid number
subpopulation, in all 3 tax years, Florida and Illinois had the most (see
figs. 8 and 9).

Figure 8: Geographic Location of AEITC Recipients in the Invalid Name
Subpopulation, Tax Years 2002 through 2004

Note: AEITC geographic location is based upon information reported on the
tax return. This figure includes the invalid name subpopulation.

aThese geographic locations had at least one but fewer than 10 individuals
who received the AEITC.

bU.S insular areas include the three major U.S. territories, which are the
U.S. Virgin Islands, American Samoa, Guam, and the two Commonwealths,
which are the Commonweath of Puerto Rico and the Commonwealth of Northern
Mariana Islands.

cOther includes Federated States of Micronesia, Marshall Islands, Palau,
and Armed Forces mostly located outside the U.S.

dMissing refers to instances where the geographic location was not
available.

Figure 9: Geographic Location of AEITC Recipients in the Invalid Number
Subpopulation, Tax Years 2002 through 2004

Note: AEITC geographic location is based upon information reported on the
tax return. This figure includes the invalid number subpopulation.

aThese geographic locations had at least 1 but fewer than 10 individuals
who received the AEITC.

bU.S insular areas includes the three major U.S. territories, which are
the U.S. Virgin Islands, American Samoa, Guam, and the two Commonwealths,
which are the Commonweath of Puerto Rico and the Commonwealth of Northern
Mariana Islands.

cMissing refers to instances where the geographic location was not
available.

AEITC was paid in excess of yearly maximum limits: A total of almost
12,000 Forms W-2, reporting about $64 million, showed AEITC paid above the
yearly maximum between tax years 2002 through 2004 (see figs. 10 and 11).
Specifically, in tax year 2002 there were 6,408 Forms W-2 above the yearly
maximum reporting almost $44 million; 2,690 in tax year 2003, reporting
over $7 million; and 2,768 in tax year 2004, reporting almost $13 million.
As noted in figure 10, most Forms W-2 above the yearly maximum were
between $1 above the limit and $5,000. An individual receiving the AEITC
was eligible to obtain a maximum yearly amount of $1,503 in tax year 2002,
$1,528 in tax year 2003, and $1,563 in tax year 2004.

Figure 10: Number of Forms W-2 Received with Amounts above the Yearly
AEITC Maximum, Tax Years 2002 through 2004

Note: This excludes 2 Forms W-2 in 2003 that when totaled, reported about
$1 billion paid in AEITC. This figure includes the dollar limit
subpopulation.

Figure 11: AEITC Dollars Reported on Form W-2s above the Yearly AEITC
Maximum, Tax Years 2002-2004

Note: This excludes 2 Forms W-2 in 2003 that when totaled, reported about
$1 billion paid in AEITC. This figure includes the dollar limit
subpopulation.

Appendix VI: Comments from the Internal Revenue Service

Appendix VII: GAO Contacts and Staff Acknowledgments

GAO Contacts

Michael Brostek, (202) 512-9110 Libby Mixon, (404) 679-1900 Joanna
Stamatiades, (404) 679-1900

Acknowledgments

In addition to those named above, Blake Ainsworth, Frances Cook, James
Cook, Rebecca Gambler, Evan Gilman, George Guttman, Donna Miller, Cheryl
Peterson, Michael Rose, Steve Sebastian, Daniel Schwimer, Richard Stana,
James Ungvarsky, Michael Volpe, and Paul Wright made key contributions to
this report.

(450619)

GAO's Mission

The Government Accountability Office, the audit, evaluation and
investigative arm of Congress, exists to support Congress in meeting its
constitutional responsibilities and to help improve the performance and
accountability of the federal government for the American people. GAO
examines the use of public funds; evaluates federal programs and policies;
and provides analyses, recommendations, and other assistance to help
Congress make informed oversight, policy, and funding decisions. GAO's
commitment to good government is reflected in its core values of
accountability, integrity, and reliability.

Obtaining Copies of GAO Reports and Testimony

The fastest and easiest way to obtain copies of GAO documents at no cost
is through GAO's Web site ( [40]www.gao.gov ). Each weekday, GAO posts
newly released reports, testimony, and correspondence on its Web site. To
have GAO e-mail you a list of newly posted products every afternoon, go to
[41]www.gao.gov and select "Subscribe to Updates."

Order by Mail or Phone

The first copy of each printed report is free. Additional copies are $2
each. A check or money order should be made out to the Superintendent of
Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more
copies mailed to a single address are discounted 25 percent. Orders should
be sent to:

U.S. Government Accountability Office 441 G Street NW, Room LM Washington,
D.C. 20548

To order by Phone: Voice: (202) 512-6000
TDD: (202) 512-2537
Fax: (202) 512-6061

To Report Fraud, Waste, and Abuse in Federal Programs

Contact:

Web site: [42]www.gao.gov/fraudnet/fraudnet.htm
E-mail: [43][email protected]
Automated answering system: (800) 424-5454 or (202) 512-7470

Congressional Relations

Gloria Jarmon, Managing Director, [44][email protected] (202) 512-4400 U.S.
Government Accountability Office, 441 G Street NW, Room 7125 Washington,
D.C. 20548

Public Affairs

Paul Anderson, Managing Director, [45][email protected] (202) 512-4800
U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, D.C. 20548

[46]www.gao.gov/cgi-bin/getrpt?GAO-07-1110 .

To view the full product, including the scope
and methodology, click on the link above.

For more information, contact Michael Brostek at (202) 512-9110 or
[email protected]

Highlights of [47]GAO-07-1110 , a report to the Chairman and Vice Chairman
of the Joint Committee on Taxation

August2007

ADVANCE EARNED INCOME TAX CREDIT

Low Use and Small Dollars Paid Impede IRS's Efforts to Reduce High
Noncompliance

The Advance Earned Income Tax Credit (AEITC) allows individuals to receive
a portion of the Earned Income Tax Credit (EITC) in their paychecks,
instead of receiving all of it when filing their year-end tax return.
Limited research has been conducted on the AEITC since GAO last examined
it in the early 1990s.

GAO was asked to determine

(1) how many individuals received the AEITC compared with the EITC in tax
years 2002 through 2004, what actions, if any, have been taken to increase
use, and the potential for increases in use in the future; (2) the extent
of noncompliance, if any, associated with the AEITC; and (3) how well the
Internal Revenue Service's (IRS) procedures address the areas of
noncompliance. To address these questions, GAO analyzed Forms W-2 and tax
return data and interviewed IRS and Social Security Administration (SSA)
officials.

[48]What GAO Recommends

GAO recommends that the Acting Commissioner of Internal Revenue analyze
options to reduce AEITC noncompliance such as implementing a "soft notice"
test. If these options are deemed ineffective, and no other options are
viable, the Treasury Secretary should provide an opinion to the Congress
as to whether the AEITC should be retained. IRS agreed with our
recommendation and outlined the actions IRS plans to take.

AEITC use was low--only about 3 percent of EITC recipients potentially
eligible for the advance received it in tax years 2002 through 2004, or
about 514,000 of the 17 million potentially eligible individuals each
year. About half of all recipients received $100 or less in AEITC and 75
percent received $500 or less for the year, with a total benefit paid of
about $146 million each year. Several efforts have been aimed at
increasing use over the last approximately15 years, such as sending
notices to individuals informing them that they were potentially eligible
for the AEITC and making changes to IRS forms. Despite these efforts, use
did not substantially increase and, for several reasons, it may be
difficult to increase it in the future. For example, IRS officials, other
experts, and prior GAO work suggests that individuals often do not elect
the AEITC because they prefer receiving the entire EITC as a lump sum
after filing their tax return.

As many as 80 percent of AEITC recipients did not comply with at least one
of the program requirements GAO reviewed, and some were noncompliant with
more than one during the 3 years we reviewed. In tax years 2002 through
2004, about 20 percent, or more than 100,000 AEITC recipients, may not
have been eligible for the AEITC because they had an invalid Social
Security number (SSN). These individuals received a total of $37 million
to $39 million each year. Almost 40 percent (about 200,000 recipients) did
not file the required tax return; these individuals received $42 million
to $50 million each year. Of the about 60 percent (more than 300,000)
AEITC recipients who did file a return, about two-thirds misreported the
amount received.

AEITC Use, Dollars, and Compliance in Tax Years 2002-2004

IRS's procedures have limited effectiveness in addressing AEITC
noncompliance. For example, Automated Underreporter (AUR) staff worked on
only a fraction of AEITC cases because of resource constraints and
criteria limiting case selection. IRS could address AEITC noncompliance by
sending "soft notices" to recipients, requiring employers to verify
employee SSNs before providing the AEITC, or creating a Forms W-5, "EITC
Advance Payment Certificate," database. Each of these options have
advantages, however, they also have potential disadvantages that could
limit their effectiveness.

References

Visible links
  33. http://www.gao.gov/cgi-bin/getrpt?GAO/GGD-92-26
  34. http://www.gao.gov/cgi-bin/getrpt?GAO-05-92
  35. http://www.gao.gov/cgi-bin/getrpt?GAO-06-895T
  36. http://www.gao.gov/cgi-bin/getrpt?GAO-04-79R
  37. http://www.gao.gov/
  38. mailto:[email protected]
  39. http://www.gao.gov/cgi-bin/getrpt?GAO/GGD-92-26
  40. http://www.gao.gov/
  41. http://www.gao.gov/
  42. http://www.gao.gov/fraudnet/fraudnet.htm
  43. mailto:[email protected]
  44. mailto:[email protected]
  45. mailto:[email protected]
  46. http://www.gao.gov/cgi-bin/getrpt?GAO-07-1110
  47. http://www.gao.gov/cgi-bin/getrpt?GAO-07-1110
*** End of document. ***