Additional Effort Is Needed to Prevent Taxpayers’ Personal
Exemptions and Tax Credits From Being Erroneously Denied
May 2003
Reference Number: 2003-40-124
This report has cleared the Treasury
Inspector General for Tax Administration disclosure review process and
information determined to be restricted from public release has been redacted
from this document.
May
30, 2003
MEMORANDUM FOR
COMMISSIONER, WAGE AND INVESTMENT DIVISION
FROM: Gordon C. Milbourn III /s/ Gordon C.
Milbourn III
Acting Deputy Inspector General for Audit
SUBJECT: Final Audit Report – Additional Effort Is
Needed to Prevent Taxpayers’ Personal Exemptions and Tax Credits From Being
Erroneously Denied (Audit # 200240025)
This
report presents the results of our review to determine whether the Internal
Revenue Service (IRS) accurately processed Tax Year (TY) 2001 individual income
tax returns identified as having errors with the spouse’s name or Social
Security Number (SSN). In a prior audit
report, we found that the IRS did not always correctly resolve these errors.
Our
overall results showed that the IRS accurately resolved an estimated 91 percent
of the TY 2001 tax returns filed containing a mismatch in the spouse’s name or
SSN. However, despite this accuracy
rate, we also determined that the IRS denied taxpayers an estimated $34.8
million in tax benefits by erroneously denying personal exemptions and Earned
Income Tax Credits (EITC) claimed on some returns.
The
impact on taxpayers was estimated using statistically valid methods to sample
from 523,259 individual income tax returns that the IRS had identified as
having a mismatch in either a spouse’s name or SSN. The estimated $34.8 million loss in taxpayer benefits was
statistically calculated from a sample taken from 85,171 of these 523,259
returns where the IRS had initially denied the spouses’ personal exemptions and
the EITC claims during the IRS’ original processing of the tax returns.
The
loss of these tax benefits occurred when IRS employees did not perform adequate
research to properly resolve the mismatch identified in a spouse’s name or SSN
and erroneously denied the personal exemption and EITC claims in 62 percent of
the returns sampled. This condition
also unnecessarily increased taxpayer burden because taxpayers had to address
the subsequent IRS tax assessment notices that were sent to them. Further research of taxpayer accounts found
that 50 percent of the affected taxpayers in our sample had responded to the
IRS to have their tax benefits properly restored. However, the remaining 50 percent of the taxpayers who have not
responded to their notice assessments represents an estimated IRS liability of
$14.3 million in outstanding taxpayer entitlements based on the additional tax
assessments erroneously issued to these taxpayers.
We
also determined that resources of $500,764 could have been more efficiently
used if the IRS had updated its computer records with spouse information by
assigning a Returns Processing Code (RPC) E. In our sample, 76 percent of the TY 2001 returns were assigned an
RPC E. The IRS had planned to review
these returns during Calendar Year 2002 to determine if the spouses’ personal
exemptions and EITC claims were accurately processed. If the results had indicated a high accuracy rate, the IRS would
have performed a one-time systemic update of the spouse information to its
computer records. However, this plan
was postponed because resources were needed to resolve Rate Reduction Credit
errors identified on TY 2001 returns.
In addition, in January 2002, the IRS eliminated a procedure that would
have updated the spouse’s name and SSN on its computer records if research
indicated the spouse information was accurate.
The IRS felt that implementing the procedure would have created capacity
problems for its Error Resolution System.
To
decrease the burden to taxpayers who have their personal exemptions and EITC
claims erroneously disallowed, we recommended the Commissioner, Wage and
Investment Division, ensure the controls are strengthened to accurately resolve
returns with mismatches and ensure computer records are updated with spouse
information if an RPC E is assigned to the return.
While management agreed that
a systemic solution that accurately updates IRS computer records would be
beneficial to the processing of mismatched returns and save staff resources,
they did not agree that their decision not to perform this update was an
inefficient use of resources. They
believe that the accuracy rate for the update of 92.5 percent would not have
been acceptable for this critical process, as this would have erroneously classified
approximately 30,000 taxpayer accounts as valid. Therefore, they believe it was a sound management decision to
postpone the update until the accuracy rate increases to an acceptable level. We concur with the IRS’ proposed suspension
and management’s assurances that they will monitor the accuracy rate and
institute a systemic update when that rate attains an acceptable level. Management’s complete response to the
draft report is included as Appendix VIII.
Copies of this
report are also being sent to the IRS managers who are affected by the report
recommendations. Please contact me at
(202) 622-6510 if you have questions or Michael R. Phillips, Assistant
Inspector General for Audit (Wage and Investment Income Programs), at (202)
927-0597.
Personal Exemptions and Earned Income Tax Credits Totaling
$34.8 Million Were Erroneously Denied
Resources Could Be More Efficiently Used by Updating Computer Records With Spouse Information
Appendix I – Detailed Objective, Scope, and Methodology
Appendix II – Major Contributors to This Report
Appendix III – Report Distribution List
Appendix IV – Outcome Measures
Appendix VIII – Management’s Response to the Draft Report
Generally, a taxpayer identification number refers to an individual’s Social Security Number (SSN) and is considered correct when it matches the information provided by the Social Security Administration (SSA). The Small Business Job Protection Act of 1996 gave the Internal Revenue Service (IRS) the authority to deny the personal exemption to taxpayers that fail to provide a correct taxpayer identification number. The Personal Responsibility and Work Opportunity Reconciliation Act of 1996 authorized the IRS to also deny the Earned Income Tax Credit (EITC) to taxpayers that do not provide a correct taxpayer identification number.
In the 2001 Filing Season, the IRS expanded its SSN verification process from the primary taxpayer to the spouse, also referred to as the secondary taxpayer, on jointly filed tax returns. We previously reviewed this process and estimated that 66,000 taxpayers had been erroneously denied the spouse’s exemption and EITC. This review was conducted to assess whether the IRS has effectively addressed this condition and reduced its reoccurrence.
We conducted the audit from June 2002 through January 2003
at the IRS’ Austin Campus. The audit
included TY 2001 individual income tax returns that were processed at the 10
IRS Submission Processing sites between January and May 2002, and was conducted
in accordance with Government Auditing
Standards.
Detailed information on our audit objective, scope, and methodology is presented in Appendix I. Major contributors to the report are listed in Appendix II.
Our overall results showed that the IRS accurately resolved an estimated 91 percent of the TY 2001 tax returns filed that contained a mismatch in the spouse’s name or SSN. These results were estimated using statistically valid methods to sample from 523,259 individual income tax returns that the IRS had identified as having this type of mismatch.
A mismatch or error occurs when the spouse’s name and SSN information does not match the IRS and SSA records. The IRS verifies this information during the processing of the return. These mismatches can occur for several reasons. For example, the spouse’s name may not match IRS and SSA records because:
· The taxpayer is newly married and has not informed the SSA of a name change.
· The spouse uses an incorrect SSN on the return.
· The IRS incorrectly transcribes the spouse’s name or SSN during processing.
When a return with a mismatch on the spouse’s name or SSN is identified, it is forwarded to an IRS employee for resolution. If the mismatch is not the result of an IRS transcription error and if additional documentation with the return indicates the SSN is correct, the employee is instructed to accept the information on the return. If there is no additional documentation to support the information on the return, the IRS will deny the spouse’s personal exemption and the EITC, if applicable.
Despite the IRS’ accuracy rate in resolving the returns with mismatches, we also determined that the IRS denied taxpayers an estimated $34.8 million in tax benefits by erroneously denying personal exemptions and EITC claims on some returns.
The estimated $34.8 million loss in taxpayer benefits was calculated from a statistical sample taken from the 85,171 returns where the IRS had initially denied the spouses’ personal exemptions and the EITC claims during the IRS’ original processing of the tax returns. See Appendix VII for details of our statistical sampling methods, sample results, and population projections.
When the IRS determines that the spouse’s name or SSN on a return is incorrect, it denies the spouse’s personal exemption and EITC, if applicable, and notifies the taxpayer. The tax effect for the loss of the personal exemption depends on the taxpayer’s tax rate bracket. For example, if the taxpayer is in the 15 percent tax rate bracket, losing the spouse’s $2,900 personal exemption amounts to $435 in additional tax.
The amount of the EITC a taxpayer will lose because of a mismatch in the spouse’s name or SSN will vary significantly between taxpayers because the EITC is based on the taxpayer’s earned income and family size. For example, married taxpayers with 2 children that meet all eligibility requirements and have earned income of $14,010 will qualify for an EITC of $3,811. Another married taxpayer with 1 child and $27,010 in earned income only qualifies for $201.
The IRS is required to perform research on the mismatches. However, while the mismatch returns are forwarded for resolution, the IRS does not always perform adequate research. The loss of these tax benefits occurred when IRS employees did not perform adequate research to properly resolve the mismatch identified in a spouse’s name or SSN and erroneously disallowed the personal exemption and EITC claims; this happened in 62 percent of the returns sampled. This also unnecessarily increased taxpayer burden because taxpayers had to address the subsequent IRS tax assessment notices that were sent to them.
Half of the affected taxpayers in our sample did not receive $14.3 million in taxpayer entitlements
When the IRS erroneously denied the spouses’ personal exemptions and EITC claims on returns in our sample, we performed an additional analysis to determine if a subsequent correction had been made. In some instances, taxpayers contacted the IRS when their personal exemption and EITC were denied. When they provided additional information, the IRS made corrections to allow the personal exemption and EITC, if applicable.
Further research of
taxpayer accounts found that 50 percent of the affected taxpayers in our sample
had responded to the IRS to have their tax benefits properly restored. However, we estimate that the remaining 50
percent of the taxpayers who have not responded to their assessment notices
represents an estimated IRS liability of $14.3 million in outstanding taxpayer
entitlements based on the additional tax assessments erroneously issued to
these taxpayers.
Overall, the IRS
did well in accurately resolving tax returns filed with a mismatch in the
spouse’s name or SSN. However, we
believe the conditions identified demonstrate the need for additional effort
when processing personal exemptions and tax credits on individual income tax
returns. These additional efforts will
decrease the burden to taxpayers who have their personal exemptions and EITC
claims erroneously disallowed.
1. The Commissioner, Wage and Investment (W&I) Division, should strengthen existing controls to ensure more mismatches of spouses’ names and SSNs are accurately resolved.
Management’s Response: The Internal Revenue Manual (IRM) 3.12.3 was revised on April 21, 2003, to provide additional and more concise procedures to resolve mismatches of spouses’ names and SSNs.
When an IRS employee determines that the spouse’s information is valid on a mismatch return, the employee assigns a Return Processing Code (RPC) E to the return, indicating that the spouse’s name and SSN were determined to be correct. This code instructs the Error Resolution System (ERS) to accept the spouse’s name and SSN on the return and allow the spouse’s personal exemption and EITC claim, if applicable.
The IRS had planned to review TY 2001 returns assigned an RPC E during Calendar Year 2002 to determine if the spouses’ personal exemptions and EITC claims were accurately processed. If the results of its review had indicated a high accuracy rate, the IRS would have performed a one-time systemic update of the spouse information to its computer records. However, this plan was postponed because resources were needed to resolve the large volume of Rate Reduction Credit errors identified on TY 2001 returns.
In addition, in January 2002, the IRS eliminated a procedure that would have updated the spouse’s name and SSN on IRS computer records if research indicated the spouse information was accurate. The IRS felt that implementing this procedure would have created capacity problems for the ERS and delayed the date the returns would be processed.
During our review, we determined that 76 percent of the TY
2001 returns in our sample were assigned an RPC E and that all of these returns were processed
accurately. We also found that an RPC E
was assigned to several taxpayers’ TY 2000 returns that showed the same
spouses’ names and SSNs as reported on the TY 2001 returns. If these same taxpayers file a joint return
for TY 2002, we estimate that 399,963 errors could be avoided if the IRS had
gone forward with its plan to update its computer records. This would result in a potential staff year
savings of $500,764 by more efficiently using existing IRS resources.
The IRS’ current plan is to review a sample of TY 2001 returns assigned an RPC E and perform the one-time systemic update if the sample results indicate a high accuracy rate. If a low accuracy rate is found, the IRS will revise its procedures for updating the spouse’s name and SSN data to its computer records, and employees will continue to work these types of cases.
We believe that if the IRS had updated its computer records with the spouse’s name and SSN information when an RPC E was assigned to the prior year return, then the TY 2001 return would not have been identified as a mismatch, as will likely be the case with many TY 2002 returns also. More importantly, the IRS would not have incurred the costs of researching the spouse’s name or SSN when the return was processed.
2.
The Commissioner, W&I Division, should ensure that
an assessment of the accuracy rate for processing TY 2001 returns assigned an
RPC E is completed. This assessment
should include either a one-time systemic update of its computer records for
the spouse’s name and SSN data or a revision to the IRS procedures to ensure
computer records are updated with the spouse’s name and SSN data at the time
the RPC E is assigned.
Management’s Response: While management agreed that a systemic solution that accurately updates IRS computer records would be beneficial to the processing of mismatched returns and save staff resources, they did not agree that their decision not to perform this update was an inefficient use of resources. They believe that the accuracy rate for the update of 92.5 percent would not have been acceptable for this critical process, as this would have erroneously classified approximately 30,000 taxpayer accounts as valid. Therefore, they believe it was a sound management decision to postpone the update until the accuracy rate increases to an acceptable level.
The IRS will perform a similar analysis of TY 2002 RPC E returns to determine the success of the IRM procedural revision and to see if the accuracy level is high enough for a systemic update. IRS management does not plan to issue procedures to update the records when the RPC E is assigned because systemic updates are more efficient and allow them to ensure the accuracy of changes to the Master File.
Auditor’s Comment:
We concur with the IRS’ proposed suspension and management’s assurances
that they will monitor the accuracy rate and institute a systemic update when
that rate attains an acceptable level that is in the best interest of the
taxpayer and the IRS. Once this
systemic update is performed, we believe resources will be used more
efficiently.
Appendix I
Detailed Objective, Scope,
and Methodology
The overall objective of this review was to determine if the Internal Revenue Service (IRS) accurately processed Tax Year (TY) 2001 individual income tax returns identified as having errors with the spouse’s name or Social Security Number (SSN). The purpose of the review was to follow up on issues identified in our review of the IRS’ 2001 Filing Season. In that report, we estimated the IRS erroneously denied spouses’ personal exemptions and Earned Income Tax Credits (EITC) totaling approximately $36.5 million for 66,000 taxpayers. To accomplish our objective, we:
I.
Determined if the IRS
accurately processed TY 2001 individual income tax returns identified as having
spouse name or SSN mismatches.
A. Obtained assistance from the Treasury Inspector General for
Tax Administration’s (TIGTA) Electronic Data Processing (EDP)
Group. Identified
a population of 523,259 returns with spouse name or SSN mismatches that were processed
to the Individual Master File (IMF) between January and May 2002.
B. Used statistical sampling methods and selected a random
sample of 196 returns to determine if the IRS accurately processed the returns
with spouse name or SSN mismatches.
1.
Researched spouse information
on IRS and Social Security Administration (SSA) computer records and researched
return information posted to the IMF.
Determined that the IRS accurately processed 178 (91 percent) of the 196
returns in the sample in accordance with IRS Internal Revenue Manual (IRM)
procedures.
2.
Based on the sample results,
calculated the number of returns that were accurately processed for the
population. See details of our
statistical sampling methodology, sample results, and population projections in
Appendix VI.
C.
Computer analyzed the
population of 523,259 returns and identified 399,963 returns that were assigned
a Return Processing Code (RPC) E.
D. Interviewed IRS Error Resolution System (ERS) officials at
one Submission Processing site about potential staff year savings associated
with returns assigned an RPC E. Based
on their comments, we estimated the potential staff year savings for the
399,963 returns assigned RPC E.
Discussed our estimate with Submission Processing officials during the
review.
II.
Determined if the IRS
accurately processed TY 2001 individual income tax returns that were assigned
Taxpayer Notice Codes (TPNC) 648 or 701.
A. Obtained assistance from the TIGTA EDP Group and identified
a population of 85,171 returns assigned TPNC 648 or 701 that were processed to
the IMF between January and May 2002.
B. Used statistical sampling methods and selected a random
sample of 382 returns to determine if the IRS accurately processed returns
assigned TPNC 648 or 701.
1.
Researched spouse information
on IRS and SSA computer records and researched return information posted to the
IMF. Determined that the IRS
erroneously assigned TPNC 648 or 701 and erroneously denied the spouses’
personal exemptions and EITC claims, if applicable, to 237 of the 382 returns
in the sample.
2.
Performed research to determine
if the IRS made subsequent adjustments to correct the erroneously denied
spouses’ personal exemptions and EITC claims on the 237 returns identified
above. Determined that 119 (50 percent)
of the 237 returns were not subsequently corrected.
3.
Based on the sample results,
calculated the number of returns in the population that were erroneously denied
spouses’ personal exemptions and EITC claims and the corresponding tax
dollars. Also calculated the number of
returns in the population that were not subsequently corrected and the
corresponding tax dollars. See details
of our statistical sampling methodology, sample results, and population
projections in Appendix VII.
Appendix II
Major Contributors to This Report
Michael R. Phillips, Assistant Inspector General for Audit
(Wage and Investment Income Programs)
Gary Young, Acting Director
Tina Parmer, Acting Audit Manager
Sharon Buford, Senior Auditor
Cari
Fogle, Senior Auditor
Steve Vandigriff, Senior Auditor
Lawrence White, Senior Auditor
Glory
Jampetero, Auditor
John
Ojeda, Auditor
Layne Powell, Computer Specialist
Appendix III
Commissioner N:C
Deputy
Commissioner N:DC
Director, Customer
Account Services W:CAS
Director, Strategy
and Finance W:S
Director,
Submission Processing W:CAS:SP
Chief Counsel CC
National Taxpayer Advocate
TA
Director, Legislative Affairs CL:LA
Director, Office of
Program Evaluation and Risk Analysis
N:ADC:R:O
Office of Management Controls N:CFO:AR:M
Audit Liaison:
Program/Process Assistant Coordinator, Wage and
Investment Division W:HR
Appendix IV
This appendix presents detailed information on the measurable impact that our recommended corrective actions will have on tax administration. These benefits will be incorporated into our Semiannual Report to the Congress.
Type and Value of Outcome Measure:
· Taxpayer Rights and Entitlements – Potential; 52,841 taxpayer accounts were erroneously denied the spouses’ personal exemptions and Earned Income Tax Credits (EITC) totaling $34.8 million; (see page 3).
Methodology Used to Measure the Reported Benefit:
Using computer analysis, we identified a population of 85,171 Tax Year (TY) 2001 individual income tax returns processed to the Individual Master File (IMF) between January and May 2002 where a Taxpayer Notice Code (TPNC) 648 or 701 was assigned. A statistical sample of 382 returns was reviewed to determine whether the Internal Revenue Service’s (IRS) denials of the spouses’ personal exemptions and EITC claims were appropriate. Review of the sample determined that 237 (62 percent) of the 382 returns contained erroneously denied spouses’ personal exemptions and EITC claims. For these 237 returns, we found that the spouses’ names or Social Security Numbers (SSN) on the returns were valid and that the IRS should have allowed the spouses’ personal exemptions or EITC claims.
Based on the sample results, we statistically projected that the IRS erroneously denied the spouses’ personal exemptions and EITC claims on 52,841 +/- 4,140 returns. The tax effect associated with the erroneously denied spouses’ personal exemptions and EITC claims amounted to $34.8 +/- $4.1 million. See Appendix VII for details of our statistical sampling methodology, sample results, and population projections.
Appendix V
Overview of the Internal Revenue Service’s Verification of
Returns With Spouse Name or Social Security Number Mismatches
This appendix presents an overview of how the Internal Revenue Service (IRS) processes returns with spouse name or Social Security Number (SSN) mismatches. It also presents detailed information on the tests we performed to determine if the IRS accurately processed these returns. The following diagram shows the three major processes and populations of returns that we tested during the review:
The diagram was
removed due to its size. To see the
diagram, please go to the Adobe PDF version of the report on the TIGTA Public
Web Page.
Population A
When a return is processed, the IRS verifies the spouse’s name and SSN on the return by performing a computer match against IRS and Social Security Administration (SSA) records. A “mismatch” occurs when the spouse’s name or SSN on the return does not agree with IRS or SSA records. Some examples include: 1) when a spouse marries and uses the married name on the tax return, but does not inform the SSA of the name change; 2) when the spouse uses an incorrect SSN on the return; or 3) when the IRS incorrectly transcribes the spouse’s name or SSN from the return during processing.
Returns
with spouse name or SSN mismatches are forwarded to the IRS’ Error Resolution
System (ERS) for further analysis. A
tax examiner reviews the return and its attachments for documentation that
confirms that the spouse’s name and SSN on the return are valid; for example, a
Wage and Tax Statement (Form W-2), marriage license, or driver’s license. The tax examiner also verifies whether the
spouse’s name and SSN were transcribed correctly and researches IRS records to
determine if an incorrect name or SSN was reported on the return.
Using computer analysis, we identified 523,259 Tax Year (TY) 2001 individual income tax returns with spouse name or SSN mismatches that were processed to the Individual Master File (IMF) between January and May 2002 (Population A). We reviewed a statistical sample of 196 returns in the population to determine if the IRS accurately processed the returns (details of our statistical sampling methodology are presented in Appendix VI).
For each of the 196 returns, we researched the IRS and SSA computer records to determine if the spouse information on the return was valid. If the spouse information was valid, we verified that the IRS allowed the spouse’s personal exemption and Earned Income Tax Credit (EITC) claim. If the spouse information was incorrect, we verified that the IRS denied the exemption and the EITC claim.
If an IRS tax examiner determines that the spouse’s name or SSN on the return is incorrect, the tax examiner enters a Taxpayer Notice Code (TPNC) 648 or 701. This instructs the ERS to deny the spouse’s personal exemption and EITC claim, if applicable, and issue a notice to the taxpayer that explains the changes made to the return. If the taxpayer subsequently contacts the IRS or SSA and provides information that resolves the mismatch, the IRS will adjust the taxpayer’s account to allow the personal exemption and EITC claim, if applicable.
In a prior audit, we reviewed the IRS’ verification process and found that the IRS might have erroneously denied exemptions and EITC claims totaling approximately $36.5 million for 66,000 taxpayers. We performed this test to determine if the condition still exists.
Using computer analysis, we identified 85,171 TY 2001 individual income tax returns assigned a TPNC 648 or 701 that were processed to the IMF between January and May 2002 (Population B). We reviewed a statistical sample of 382 returns to determine if the IRS’ denials of the spouses’ personal exemptions and EITC claims were appropriate (details of our statistical sampling methodology are presented in Appendix VII).
Population C
If a tax examiner determines that the spouse’s name and SSN on the return are valid, the tax examiner assigns a Return Processing Code (RPC) E to the return. The RPC E instructs the computer to allow the spouse’s personal exemption and EITC claim, if applicable. The Internal Revenue Manual describes various situations for assigning an RPC E to a return. For example, an RPC E should be assigned to a return when the taxpayer provides documentation that confirms the spouse’s name and SSN are correct as reported on the return. In addition, an RPC E should be assigned when IRS computer records indicate the same spouse (e.g., wife) filed a joint return with the same primary taxpayer (e.g., husband) in a prior year. Finally, an RPC E should be assigned when research of IRS or SSA computer records indicate that the spouse is using a married surname on the return, but did not notify the SSA of the name change.
When we reviewed the initial sample of returns for Population A, we found that an RPC E was assigned to 152 (78 percent) of the 196 returns in our statistical sample and that all 152 returns were accurately processed. Using computer analysis, we determined that 399,963 (76 percent) of the 523,259 returns in Population A were assigned an RPC E.
Appendix VI
Sampling Methodology for
Tax Year 2001 Returns With Spouse Name or Social Security Number Mismatches
This appendix presents details of our statistical sampling methodology
and sample results for Tax Year (TY) 2001 returns identified as having spouse
name or Social Security Number (SSN) mismatches.
Sampling Methodology
Review a statistical sample of TY 2001 individual income tax returns
identified as having spouse name or SSN mismatch and determine if the Internal
Revenue Service (IRS) accurately processed the spouse’s personal exemption and
Earned Income Tax Credit (EITC) claim.
Using computer analysis, we identified 523,259 returns with
a spouse name or SSN mismatch that were processed to the Individual Master File
(IMF) between January and May 2002.
These returns had a filing status of “married filing joint” with a
Secondary (spouse) National Account Profile (NAP) Access Indicator that was not
equal to a blank.
A sample size of 196 was determined using the following criteria:
·
Population of Returns 523,259
·
Confidence level 95
percent
·
Expected Rate of Occurrence (error rate) 15 percent
·
Precision Rate (sample reliability) +/- 5 percent
We used a random number table and the random number generation tool in
Microsoft Excel to randomly select the sample of 196 returns from the
population.
Sample Results
Our test results indicated the IRS accurately processed the spouse’s
personal exemption and EITC claim for 178 (91 percent) of the 196 sample
returns.
Population
Projections
Based on the sample results, we are 95 percent confident the IRS
accurately processed between 453,998 and 496,410 returns in the
population. The midpoint is 475,204
returns, and the standard error at a 95 percent confidence level is +/- 21,206
returns.
Formula and Data Used for
Population Projection
We used the formula Np +/- N Z σp and the following data to calculate the
population projection:
·
N
(universe or population) 523,259
·
n
(sample size)-- 196
·
f
(fraction sampled, e.g., 196 divided by 523,259) .0003745
·
p
(sample accuracy rate, e.g., 178 divided by 196) .9081632
·
Z
(precision at 95 percent confidence level)-- 1.96
·
σp = (one standard error of p)
The formula for σp is the square root of [p(1-p)(1-f) / n-1].
· Np (estimated total number of accurate returns for the population)
·
N σp
(one standard error of the estimated total
number of accurate returns for the population)
Solutions for the formula were:
Np = 475,204 (midpoint of estimated total number of accurate returns in the population)
N Z σp = 21,206 (standard error of estimated total number of accurate returns in the population at a 95 percent confidence level)
Np +/- N Z σp = 453,998 to 496,410
Appendix VII
Sampling Methodology for
Tax Year 2001 Returns Where the Spouses’ Personal Exemptions and Earned Income
Tax Credits Were Denied
This appendix presents details of our sampling methodology for Tax Year
2001 returns assigned a Taxpayer Notice Code (TPNC) 648 or 701 to deny the
spouse’s personal exemption and Earned Income Tax Credit (EITC) claim.
Sampling Plan
Review a statistical sample of TY 2001 individual income tax returns
assigned a TPNC 648 or 701 and determine if the Internal Revenue Service’s
(IRS) denials of the spouse’s personal exemption or EITC claim were
appropriate. For each return where the
spouse’s personal exemption or EITC claim was erroneously denied, determine if
the IRS made a subsequent adjustment to correct the taxpayer’s account.
Using computer analysis, we identified 85,171 TY 2001
individual income tax returns that were processed to the Individual
Master File (IMF) between January and May 2002 where a TPNC 648 or 701 was assigned.
A sample size of 382 was determined using the following criteria:
·
Population of Returns 85,171
·
Confidence level 95
percent
·
Expected Rate of Occurrence (error rate) 50 percent
·
Precision Rate (sample reliability) +/- 5 percent
We initially used an expected error rate of 15 percent and computed a
sample size of 195. However, during our
review of the sample returns, we periodically computed the actual error rate
and found that it was greater than 50 percent.
As a result, we increased the expected error rate to 50 percent and
recomputed the sample size of 382.
We used a random number table and the random number generation tool in
Microsoft Excel to randomly select the sample of 382 returns from the
population.
Sample Results
Our test results indicated the IRS erroneously assigned TPNC 648 or 701
to 237 (62 percent) of the 382 returns in the sample. The spouses’ personal exemptions and EITC claims, if applicable,
were erroneously denied on the 237 returns.
The IRS subsequently corrected 118 (50 percent) of the 237 returns with
erroneously denied spouses’ personal exemptions and EITC claims. However, we determined that no subsequent
adjustments were made to 119 (50 percent) of the 237 returns to correct those
erroneously denied spouses’ personal exemptions and EITC claims.
Population
Projections
Based on the sample results, we are 95 percent confident that the IRS
erroneously denied the spouses’ personal exemptions and EITC claims on 48,701
to 56,981 returns in the population.
The midpoint is 52,841 returns, and the standard error at a 95 percent
confidence level is +/-
4,140 returns. The tax effect
associated with the erroneous disallowances is between $30.7 and $38.9 million.
The midpoint is $34.8 million, and the
standard error at a 95 percent confidence level is +/- $4.1 million. The tax dollars reflect IRS changes to spouses’ exemptions and
EITC claims only. We did
not consider other IRS adjustments that affected the tax dollars.
Furthermore, we are 95 percent confident that the IRS has not made
subsequent adjustments on 22,581 to 30,483 returns, to allow spouses’
exemptions and EITC claims that were erroneously denied. The midpoint is 26,532 returns, and the
standard error at a 95 percent confidence level is +/- 3,951 returns. The tax dollars that have not been adjusted and credited to
taxpayers’ accounts range from $11.9 to $16.7 million. The midpoint is $14.3 million, and the standard
error at a 95 per cent confidence level is +/- $2.4 million.
Formulas and Data Used for
Population Projections
We used the formula Np +/- N Z σp and the following data to calculate the
population projection of “48,701 to 56,981 returns:”
·
N
(universe or population) 85,171
·
n
(sample size)-- 382
·
f
(fraction sampled, e.g., 382 divided by 85,171) .004485
·
p
(sample accuracy rate, e.g., 237 divided by 382) .6204188
·
Z
(precision at 95 percent confidence level)-- 1.96
·
σp = (one standard error of p)
The formula for σp is the square root of [p(1-p)(1-f) / n-1].
· Np (estimated total number of error returns for the population)
·
N σp
(one standard error of the estimated total
number of error returns for the population)
Solutions for the formula were:
Np = 52,841 (midpoint of estimated total number of error returns in the population)
N Z σp = 4,140 (standard error of estimated total number of error returns in the population at a 95 percent confidence level)
Np +/- N Z σp = 48,701 to 56,981
We used the formula N*ybar +/- Z*N*[s/sqrt(n)]*sqrt(1 - f)
and data below to calculate the population
projection of “$30.7 to $38.9 million” tax dollars:
·
N (universe or population) 85,171
·
n (sample size)-- 382
·
f (fraction sampled, e.g., 382 divided by 85,171) .004485
·
ybar (average of error dollars from the sample) $408.27
If the case was
properly processed, the error dollars for the case is 0.
·
s (sample standard deviation of error dollars for
one standard deviation)
The formula for s
is the square root of [sum of (Y – Ybar) squared / (n-1)].
·
Z (reliability coefficient at 95 percent confidence
level)-- 1.96
Solutions for the formula were:
N*ybar = $34.8 million (estimated total error)
Z*N*[s/sqrt(n)]*sqrt(1 - f) = $4.1 million (precision)
N*ybar +/- Z*N*[s/sqrt(n)]*sqrt(1 – f) = $30.7 to $38.9 million
We used the formula Np +/- N Z σp and the following data to calculate the
population projection of “22,581 to 30,483 returns:”
·
N
(universe or population) 85,171
·
n
(sample size)-- 382
·
f
(fraction sampled, e.g., 382 divided by 85,171) .004485
·
p
(sample accuracy rate, e.g., 119 divided by 382) .3115183
·
Z
(precision at 95 percent confidence level)-- 1.96
·
σp = (one standard error of p)
The formula for σp is the square root of [p(1-p)(1-f) / n-1].
· Np (estimated total number of error returns for the population)
·
N σp
(one standard error of the estimated total
number of error returns for the population)
Solutions for the formula were:
Np = 26,532 (midpoint of estimated total number of error returns in the population)
N Z σp = 3,951 (standard error of estimated total number of error returns in the population at a 95 percent confidence level)
Np +/- N Z σp = 22,581 to 30,483
We used the formula N*ybar +/- Z*N*[s/sqrt(n)]*sqrt(1 - f)
and data below to calculate the population
projection of “$11.9 to $16.7 million” tax dollars:
·
N (universe or population) 85,171
·
n (sample size)-- 382
·
f (fraction sampled, e.g., 382 divided by 85,171) .004485
·
ybar (average of error dollars from the sample) $168.38
If the case was
properly processed, the error dollars for the case is 0.
·
s (sample standard deviation of error dollars for
one standard deviation)
The formula for s
is the square root of [sum of (Y – Ybar) squared / (n-1)].
·
Z (reliability coefficient at 95 percent confidence
level)-- 1.96
Solutions for the formula were:
N*ybar = $14.3 million (estimated total error)
Z*N*[s/sqrt(n)]*sqrt(1 - f) = $2.4 million (precision)
N*ybar +/- Z*N*[s/sqrt(n)]*sqrt(1 – f) = $11.9 to $16.7 million
Appendix VIII
Management’s Response to
the Draft Report
The response
was removed due to its size. To see the
complete response, please go to the Adobe PDF version of the report on the
TIGTA Public Web Page.