This is the accessible text file for GAO report number GAO-02-509 
entitled 'Tax Deductions: Further Estimates of Taxpayers Who May Have 
Overpaid Federal Taxes by Not Itemizing' which was released on March 
29, 2002. 

This text file was formatted by the U.S. General Accounting Office 
(GAO) to be accessible to users with visual impairments, as part of a 
longer term project to improve GAO products' accessibility. Every 
attempt has been made to maintain the structural and data integrity of 
the original printed product. Accessibility features, such as text 
descriptions of tables, consecutively numbered footnotes placed at the 
end of the file, and the text of agency comment letters, are provided 
but may not exactly duplicate the presentation or format of the 
printed version. The portable document format (PDF) file is an exact 
electronic replica of the printed version. We welcome your feedback. 
Please E-mail your comments regarding the contents or accessibility 
features of this document to Webmaster@gao.gov. 

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. 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 General Accounting Office: 
GAO: 

Report to the Honorable Dick Armey, Majority Leader, House of 
Representatives: 

March 2002: 

Tax Deductions: 

Further Estimates of Taxpayers Who May Have Overpaid Federal Taxes by 
Not Itemizing: 

GAO-02-509: 

United States General Accounting Office: 
Washington, DC 20548: 

March 29, 2002: 

The Honorable Dick Armey: 
Majority Leader: 
House of Representatives: 

Dear Mr. Armey: 

When computing their federal taxes, taxpayers either claim a standard 
deduction or itemize deductions, and these deductions are subtracted 
from adjusted gross income in determining taxable income. Taxpayers 
generally claim the type of deduction that is larger, because this 
minimizes their taxable income. In recent years, approximately 70 
percent of taxpayers have claimed the standard deduction, while the 
other 30 percent have itemized. 

This is our second report in response to your request for an estimate 
of the number of taxpayers who may have overpaid their taxes by 
claiming the standard deduction instead of itemizing their deductions, 
as well as of the amount of taxes that they may have overpaid. In our 
first report,[Footnote 1] we estimated that on about 510,000 returns 
for tax year 1998, taxpayers who claimed the standard deduction had 
deductible mortgage interest expense in excess of their standard 
deduction and overpaid their taxes by about $311 million. 

In this report, our estimates include payments not only for mortgage 
interest but also for mortgage points, state and local income taxes, 
charitable contributions, and real estate and personal property tax 
for tax year 1998. We developed our estimate for mortgage interest 
payments, mortgage points, and state and local income taxes by 
matching a sample of tax year 1998 returns for individuals who did not 
claim itemized deductions with information in other Internal Revenue 
Service (IRS) databases. This information was supplied by lending 
institutions and employers and consisted, respectively, of the 
mortgage interest and points paid by each taxpayer and the wages paid 
to each taxpayer. Our estimates for charitable contributions and real 
estate and personal property taxes are averages for various income 
classes based on Department of Labor data. We assigned (imputed) these 
class averages to each taxpayer in the same income class in our 
sample. In this report, we also present information on the proportion 
of returns prepared by a third party on which taxpayers may have 
overpaid their taxes. 

When the data allowed, we identified and excluded from our sample the 
most common situations, such as returns with no tax liability, on 
which taxpayers could not have itemized deductions. We also attempted 
to identify and exclude the most common instances in which a payment, 
such as a mortgage interest payment, would not be deductible. However, 
we could not exclude all nonitemizable payments. To the extent that 
our estimates included some nonitemizable deductions, they may be 
overstated. Likewise, data were not available for us to identify and 
include all itemizable deductions. To the extent that we were not able 
to include all the deductions that nonitemizers could have claimed, 
our estimates may be understated. A more detailed discussion of our 
possible over- and underestimates of deductible payments is contained 
in an appendix. 

Because our estimates are based on a random sample of tax returns, 
they are subject to sampling error, which indicates the extent to 
which random samples are likely to differ from the populations that 
they represent. Our estimates may also be subject to other types of 
error, called nonsampling error, which can occur, for example, when we 
impute values for missing data. Because we used imputed values for our 
estimates of charitable contributions and of real estate and personal 
property tax payments, and because these estimates may be affected by 
other sources of nonsampling error, these estimates are likely to have 
larger nonsampling errors and may be less reliable than our estimates 
of mortgage interest and state and local tax payments. The appendix 
contains more detail on our methodology and our assessment of the 
reliability of our estimates. 

As agreed with your office, we did not attempt to determine the 
reasons why taxpayers claimed the standard deduction when they might 
have paid less tax had they itemized deductions. 

Results in Brief: 

We estimate that on about 948,000 tax returns for tax year 1998, 
taxpayers did not itemize their deductions yet had payments for 
mortgage interest and points and for state and local income tax that 
exceeded the standard deduction for their tax filing status.[Footnote 
2] When we impute charitable contributions and real estate and 
personal property tax payments, the estimate of the total number of 
taxpayers whose potential itemized deductions exceeded the claimed 
standard deduction could be as high as 2.2 million tax returns. 

We estimate that these taxpayers with unitemized mortgage interest, 
mortgage points, and state and local income taxes that exceeded the 
claimed standard deduction for their tax filing status are likely to 
have overpaid their taxes by about $473 million. When charitable 
contributions and real estate and personal property tax payments are 
included, the total overpayment of taxes could be as high as about 
$945 million. On the basis of this $945 million estimate, the average 
overpayment amount was an estimated $438. About 76 percent of these 
taxpayers may have overpaid by $500 or less, while the remaining 24 
percent may have overpaid their taxes by more than $500. We also 
estimate that a third party prepared about 50 percent of the tax 
returns on which taxes may have been overpaid. 

Background: 

The tax code allows taxpayers to subtract from gross income either a 
standard deduction or certain itemized deductions. The standard 
deduction is the sum of two components, the basic standard deduction 
and the additional standard deduction for taxpayers who are aged 65 
years or older, blind, or both. Both the basic and additional standard 
deductions vary in amount, depending on filing status.[Footnote 3] 
Basic and additional standard deduction amounts are set by Congress 
and are adjusted annually for inflation using the Consumer Price Index. 

Itemized deductions are specified personal and other expenses that 
Congress has chosen to allow as deductions in arriving at taxable 
income. Deductible personal expenses include certain of the taxpayer’s 
interest payments, such as mortgage interest and points, which are 
charges paid, or treated as paid, by a borrower to obtain a home 
mortgage; certain nonfederal taxes, such as state and local income 
taxes; gifts to charity; real estate taxes and personal property 
taxes; medical and dental expenses; and casualty and theft losses. 
Other deductible expenses include certain payments related to the 
production or collection of income and expenses related to the 
management of property held for the production of income. 

Generally, a taxpayer is allowed to claim either itemized deductions 
or the standard deduction, whichever is greater. To minimize their tax 
liability, taxpayers can compare their total standard deduction (the 
sum of their basic standard deduction and any additional standard 
deductions) with their total itemized deductions. If their itemized 
deductions are less than their standard deduction, taxpayers can 
compute taxable income using the standard deduction. If their itemized 
deductions exceed the standard deduction, taxpayers can compute 
taxable income by itemizing. 

Scope and Methodology: 

To estimate the number of taxpayers who may have overpaid their taxes 
by claiming the standard deduction instead of itemizing, and to 
estimate the amount of taxes that they may have overpaid, we used 
IRS’s Statistics of Income (SOI) data for tax year 1998 (the most 
recent year for which data were available). The SOI data consist of a 
random sample of about 164,000 individual returns that is 
statistically representative of the 124.8 million individual returns 
filed. We attempted to identify and exclude taxpayers who could not 
have claimed itemized deductions, such as those with no tax liability, 
from our estimate of potential overpayers. 

We matched the sampled returns with data from IRS’s Information 
Returns Master File, which contains data for mortgage interest and 
points from form 1098 information returns that were received by IRS 
from taxpayers’ lending institutions. We used this matched file to 
obtain mortgage interest payments and points for taxpayers in the SOI 
sample who did not itemize their deductions. We attempted to identify 
and exclude from our estimate of potentially deductible mortgage 
interest and points any of these payments that could not be claimed as 
itemized deductions. 

Our estimate for state and local income taxes was based on our 
analysis of a subsample of the SOI sample returns for individuals who 
reported wage income and did not claim itemized deductions. For these 
wage earners, we collected the state and local income taxes reported 
on their form W-2 and applied these values to the total SOI sample of 
returns on which taxpayers reported wages but did not itemize their 
deductions. 

Our estimates for charitable contributions and real estate and 
personal property taxes were based on the Department of Labor’s 
calendar year 1998 consumer expenditure survey (CES), which contains 
data on these expenditures by income level and geographic location. 
For estimates of the charitable contributions and the personal 
property taxes of nonitemizers, we calculated the average value of 
these payments by income class in the CES sample and imputed these 
class averages to each nonitemizer in the same income class in the SOI 
sample. For estimates of real estate taxes, we calculated the average 
payment by geographic location as well as by income class and imputed 
this class average only to nonitemizers in the SOI sample who also had 
mortgage interest payments. 

All sample results in this report have been weighted to reflect the 
entire population and are subject to sampling error. These sampling 
errors reflect the extent to which random samples of these sizes and 
structures are likely to differ from the populations that they 
represent. Each of the estimates is surrounded by a 95 percent 
confidence interval, indicating that we can be 95 percent confident 
that the interval contains the actual population value. 

In addition to sampling error, our estimates may also be affected by 
other sources of error, such as errors in interpreting questions or 
providing information on the CES, errors in processing or recording 
data, and errors made when estimating missing data. These types of 
error are called nonsampling error. The estimates of charitable 
contributions and of real estate and personal property tax are likely 
to have larger nonsampling errors than our other estimates, because, 
unlike the estimates of mortgage interest and state and local tax 
payments, they are based on survey data and require procedures for 
estimating missing values. Because of these additional sources of 
nonsampling error, these estimates may be less reliable than the 
estimates of mortgage interest and state and local income tax payments 
by nonitemizers. 

The appendix contains more detail on our findings and data 
limitations. We did our work from July 2001 to March 2002 in 
accordance with generally accepted government auditing standards. 

We requested and received comments from IRS’ director of the Office of 
Research, Analysis and Statistics of Income on a draft of this report. 
In addition, we discussed our methodology with academic and other tax 
policy experts. They stated that the methodology we used for 
developing our estimates was reasonable. 

Itemized Deductions Exceeded the Standard Deduction for Some Taxpayers 
Who Did Not Itemize: 

For tax year 1998, taxpayers filed more than 124 million tax returns. 
On almost 86 million, or about 70 percent, of these returns, taxpayers 
claimed the standard deduction. On the remaining 38 million, or about 
30 percent, taxpayers itemized their deductions. We estimate that on 
about 948,000 of the returns claiming the standard deduction, 
taxpayers could have reduced their taxes because of mortgage interest, 
mortgage points, and state and local income tax payments that exceeded 
the standard deduction. After imputing charitable contributions and 
real estate and personal property tax payments to taxpayers, we 
estimate that the number of returns on which payments for itemized 
deductions exceeded the standard deduction could be as high as 2.2 
million returns. These estimates and their confidence intervals are 
reported in table 1. 

Table 1: Estimated Number of Returns on Which Taxpayers May Have 
Overpaid Taxes because Itemizable Deductions May Have Exceeded the 
Standard Deduction, and Estimates of Amounts Overpaid, Tax Year 1998 
(numbers in thousands): 

Return category: Returns on which mortgage interest and points and 
state and local income tax exceeded the standard deduction[A]; 
Estimated returns: 948; 
95% confidence interval: 607–1,411; 
Estimated overpayments: $473,048; 
95% confidence interval: $367,964–$578,133. 

Return category: Returns on which mortgage interest and points and 
state and local income tax plus imputed charitable contributions, real 
estate and personal property taxes exceeded the standard deduction[B]; 
Estimated returns: 2,157; 
95% confidence interval: 1,602–2,838; 
Estimated overpayments: $945,182; 
95% confidence interval: $792,542–$1,097,822. 

[A] Estimate based on SOI and sampled tax return data. 

[B] Estimate based on SOI and sampled tax return data as well as 
imputations from CES data. 

Source: GAO analysis of SOI and other IRS data and Department of Labor 
CES data. 

[End of table] 

In addition, as shown in table 1, we estimated that taxpayers who 
could have reduced their taxes by itemizing mortgage interest and 
points and state and local income tax payments may have overpaid their 
1998 taxes by about $473 million. We estimated that if charitable 
contributions and real estate and personal property tax payments are 
also imputed, taxpayers may have overpaid taxes by as much as $945 
million (or about 0.1 percent of all 1998 individual income tax paid). 
On the basis of this $945 million estimate, the average estimated 
overpaid tax amount is $438. As shown in table 2, on about 76 percent 
of returns, taxpayers may have overpaid by $500 or less, while on the 
remaining 24 percent of returns, taxpayers may have overpaid their 
taxes by more than $500. 

Table 2: Distribution of Returns on Which Taxpayers May Have Overpaid 
Taxes because Itemizable Deductions May Have Exceeded the Standard 
Deduction, Tax Year 1998 (numbers in thousands): 

Distribution of overpaid tax: $1–$100; 
Returns: 
Estimated number: 677; 
Percentage of total: 31%; 
95% confidence interval: 358–1,070. 

Distribution of overpaid tax: $101–$250; 
Returns: 
Estimated number: 591; 
Percentage of total: 27%; 
95% confidence interval: 310–950. 

Distribution of overpaid tax: $251–$500; 
Returns: 
Estimated number: 382; 
Percentage of total: 18%; 
95% confidence interval: 264–525. 

Distribution of overpaid tax: $501–$1,000; 
Returns: 
Estimated number: 275; 
Percentage of total: 13%; 
95% confidence interval: 185–387. 

Distribution of overpaid tax: $1,001–$5,000; 
Returns: 
Estimated number: 224; 
Percentage of total: 10%; 
95% confidence interval: 147–322. 

Distribution of overpaid tax: More than $5,000; 
Returns: 
Estimated number: 8; 
Percentage of total: 0; 
95% confidence interval: 3–20. 

Distribution of overpaid tax: Total; 
Returns: 
Estimated number: 2,157; 
Percentage of total: 100%; 
95% confidence interval: 1,602–2,838. 

Note: Totals may not sum because of rounding. 

Source: GAO analysis of SOI and other IRS data and Department of Labor 
CES data. 

[End of table] 

We also estimated the number of returns with overpayments that were 
prepared by third parties. Most of these third-party preparers were 
paid preparers; less than 1 percent of the returns were prepared by 
IRS and other noncommercial preparers. On the basis of the data in 
table 3, we estimate that for about 50 percent of the taxpayers who 
might have overpaid their taxes, the tax returns were prepared by a 
third party. 

Table 3: Distribution of Returns Prepared by a Third Party on Which 
Taxpayers May Have Overpaid Taxes because Itemizable Deductions May 
Have Exceeded the Standard Deduction, Tax Year 1998 (numbers in 
thousands): 

Distribution of overpaid tax: $1–$100; 
Estimated returns: 677; 
Estimated returns prepared by a third party: 327; 
95% confidence interval: 70–823. 

Distribution of overpaid tax: $101–$250; 
Estimated returns: 591; 
Estimated returns prepared by a third party: 339; 
95% confidence interval: 88–792. 

Distribution of overpaid tax: $251–$500; 
Estimated returns: 382; 
Estimated returns prepared by a third party: 155; 
95% confidence interval: 98–231. 

Distribution of overpaid tax: $501–$1,000; 
Estimated returns: 275; 
Estimated returns prepared by a third party: 138; 
95% confidence interval: 87–206. 

Distribution of overpaid tax: $1,001–$5,000; 
Estimated returns: 224; 
Estimated returns prepared by a third party: 104; 
95% confidence interval: 62–162. 

Distribution of overpaid tax: More than $5,000; 
Estimated returns: 8; 
Estimated returns prepared by a third party: 8; 
95% confidence interval: 3–20 

Distribution of overpaid tax: Total; 
Estimated returns: 2,157; 
Estimated returns prepared by a third party: 1,071; 
95% confidence interval: 701–1,442. 

Note: The confidence intervals for the estimated returns are shown in 
table 2. The upper ends of the confidence intervals for estimated 
returns prepared by a third party can exceed estimated returns because 
the latter also are estimates. 

Source: GAO analysis of SOI and other IRS data and Department of Labor 
CES data. 

[End of table] 

As agreed with your office, we plan no further distribution of this 
report until 30 days from the date of this letter, unless you publicly 
announce its contents earlier. At that time, we will send copies to 
Representative William Thomas, chairman, and Representative Charles B. 
Rangel, ranking minority member, House Committee on Ways and Means; 
Senator Max Baucus, chairman, and Senator Charles E. Grassley, ranking 
member, Senate Finance Committee; Representative Amo Houghton, 
chairman, and Representative William J. Coyne, ranking minority 
member, Subcommittee on Oversight, House Committee on Ways and Means; 
and the Honorable Charles O. Rossotti, commissioner of internal 
revenue. We will make copies available to others on request. 

If you have any questions regarding this report, please contact me at 
(202) 512-9110 or Ralph Block at (415) 904-2150. The major 
contributors to this report were John Mingus, Kevin Daly, and Anne 
Stevens. 

Sincerely yours, 

Signed by: 

James R. White: 
Director, Tax Issues: 

[End of section] 

Appendix: Methodology for Estimating the Number of Returns on Which 
Taxpayers Overpaid Taxes by Not Itemizing Deductions: 

Our estimates of the number of taxpayers who potentially overpaid 
their federal taxes by claiming the standard deduction instead of 
itemizing and of the amount of taxes they may have overpaid are based 
on estimates of nonitemizing taxpayers’ payments for mortgage interest 
and points, state and local income taxes, charitable contributions, 
and real estate and personal property taxes for tax year 1998. 
[Footnote 4] Data were not available for us to identify and include 
all itemizable deductions. To the extent that we were not able to 
include all the deductions that nonitemizers could have claimed, our 
estimates may be understated. 

To make these estimates, we used IRS’s Statistics of Income (SOI) data 
for tax year 1998. Our estimate for mortgage interest and mortgage 
points was based on SOI sample returns for individuals who did not 
claim itemized deductions. For these returns, we obtained tax year 
1998 data on mortgage interest payments and points from IRS’s 
Information Returns Master File. Our estimate for state and local 
income taxes was based on our analysis of a subsample of SOI sample 
returns for individuals who reported wage income and did not claim 
itemized deductions. For these wage earners, we collected the state 
and local income taxes reported on their form W-2. Our imputations of 
charitable contributions and real estate and personal property taxes 
were based on 1998 consumer expenditure survey data for the SOI sample 
of taxpayers who did not itemize their deductions. The sample results 
in our report have been weighted to reflect the entire population and 
are subject to sampling and other types of error. 

In our report, we estimate federal income tax overpayments by 
taxpayers who do not itemize. We do not include the effect on state 
tax liabilities of itemizing at the federal level. For most states, 
itemizing on the federal tax return has no effect on state tax 
liabilities. However, in two states-—Rhode Island and Vermont—-where 
the state tax liability is a percentage of the federal tax liability, 
taxpayers who overpaid by not itemizing at the federal level would 
also overpay state taxes. In three states—-Georgia, Virginia and Utah-—
and the District of Columbia, where taxpayers who itemize on their 
federal returns must also itemize on their state tax returns and not 
all federal itemizable deductions are deductible, taxpayers who 
overpaid by not itemizing at the federal level may have increased or 
decreased tax liabilities at the state level. In these jurisdictions, 
the effect on state tax liabilities of itemizing on the federal return 
depends on the size of the state standard deduction, state tax rates, 
and the deductions allowed at the state level. 

Exclusions from the SOI Sample: 

We attempted to identify and exclude from our sample the most common 
situations in which taxpayers could not have itemized their 
deductions. Except as noted, these are the same exclusions as in our 
first report. We excluded: 

* Returns with no tax liability. Because these filers had no tax 
liability, there was no potential gain to them in itemizing deductions. 

* Returns where the filing status was “married filing separately.” 
Taxpayers in this filing status are required to use the same type of 
deductions as their spouse (i.e., if one spouse claimed the standard 
deduction, the other must also). Some of these taxpayers may have been 
required to claim the standard deduction because their spouse did so. 
We did not have data on the filing status of the spouses, so we 
excluded these taxpayers from our analysis. In addition, we would not 
have been able to allocate the itemized deductions between the two 
returns. 

* Returns of taxpayers that were in the phaseout range for itemizing 
deductions and who would not have lowered their taxes by itemizing. 
Certain itemized deductions are subject to being phased out, beginning 
at certain levels of adjusted gross income. For tax year 1998, the 
income levels are $124,500 for filing single, for married filing 
jointly, for qualified widow(er) with dependent child, and for head of 
household; and $62,250 for married filing separately. The itemized 
deductions subject to phaseout are home mortgage interest, including 
points; state and local income taxes; gifts to charities; real estate 
and personal property taxes; unreimbursed employee expenses; and some 
other itemizable expenses. 

We excluded those taxpayers in the 1998 SOI sample who had claimed a 
standard deduction at the time the sample was drawn but who later 
itemized their deductions on an amended return. In our first report, 
we simply excluded all taxpayers in the 1998 SOI sample who later 
filed an amended return. 

Mortgage Interest Payments: 

We matched data from IRS’s Information Returns Master File, which 
contains form 1098 data on mortgage interest and points, provided to 
IRS by mortgage lending institutions, to the SOI sampled returns to 
obtain mortgage interest payment amounts for taxpayers in the sample 
who did not itemize their deductions. We attempted to identify and 
exclude from our sample the most common situations in which mortgage 
interest or points could not be claimed as an itemized deduction. 
Again, except as noted, these are the same exclusions as in our first 
report. We excluded: 

* Returns on which taxpayers reported their entire mortgage interest, 
on their Schedule C: Profit or Loss From Business; Schedule E: 
Supplemental Income and Loss; or Schedule F: Profit or Loss From 
Farming; or on some combination of Schedules C, E, and F. For those 
returns where the mortgage interest on the forms 1098 exceeded the 
amounts reported on the schedules C, E, and F, we assumed that the 
excess mortgage interest could be personal expenses. In our first 
report, we simply excluded returns with these types of business 
interest deductions and did not attempt to estimate the personal 
mortgage interest of these taxpayers. 

In certain cases, we were unable to exclude mortgage interest that may 
not have been itemizable. To the extent that these cases included 
nonitemizable mortgage interest payments, our estimates may be 
overstated. Home mortgage interest payments that we identified as not 
being fully itemizable but could not exclude from our estimates are as 
follows: 

* Interest payments on mortgage balances over $1 million when the 
mortgage(s) was used to acquire (i.e., buy, build, or improve) the 
residence(s). Although we could not exclude them, it is unlikely that 
our sample included many such large mortgages. Less than one percent 
of the returns that we projected to have overpaid taxes had associated 
mortgage interest payments of $50,000 or more, which would be the 
interest payment on a $1 million mortgage at an interest rate of 5 
percent. 

* Interest payments on home equity debt, where the mortgage totaled 
more than the lesser of $100,000 or the fair market value of the home, 
reduced by acquisition indebtedness. 

* Interest payments on a third home. Taxpayers who meet this 
nondeductible category generally would have to have three non-
incomeproducing homes. If any of the three homes were used for 
business purposes, including rental property, the mortgage interest 
payments would have been reported on the taxpayers’ Schedule C, E, or 
F. If so, they were already excluded from our sample. 

Our estimates may be affected by the accuracy of the form 1098 data 
that we used. We assumed that the data were accurate for both the 
amount of the mortgage interest payments and the taxpayer that made 
the interest payments. If there are any errors in the data, our 
estimates may be overstated to the extent that the mortgage interest 
amounts may be overstated. Conversely, our estimates may be 
understated to the extent that the mortgage interest amounts may be 
understated. Also, if the forms 1098 include the names and social 
security numbers of any taxpayers not legally entitled to claim a 
mortgage interest deduction, our estimates may be overstated. 

State and Local Income Taxes: 

To estimate amounts of state and local income taxes, we obtained a 
subsample of 900 tax returns in the SOI sample from IRS and reviewed 
the form W-2s attached to the returns. We selected this subsample by 
identifying nonitemizing taxpayers with wages who lived in states with 
a state income tax. To ensure that our subsample was representative of 
the taxpayers in the SOI sample with wages, we stratified the SOI 
sample into categories based on the amount of wages and mortgage 
interest for a given taxpayer. We assigned weights to each taxpayer in 
the subsample, so that it could be used to project the number of 
nonitemizing taxpayers in the population whose state and local income 
tax would have exceeded their standard deduction. Our estimates of 
both the number of overpayers and the taxes overpaid may be 
understated, because we are not able to identify the state and local 
taxes paid by taxpayers that were not withheld from their wages. 

Charitable Contributions, Real Estate, and Personal Property Taxes: 

Our estimates for charitable contributions and real estate and 
personal property taxes were based on our analysis of data from the 
Department of Labor’s consumer expenditure survey (CES). The CES 
contains data on the expenditures of consumer units (roughly 
equivalent to households) by income level and geographic location that 
include charitable contributions and payments of real estate and 
personal property taxes.[Footnote 5] The CES data that we used are 
from a random sample that represents all U.S. consumer units in 
calendar year 1998. 

To estimate charitable contributions for nonitemizers, we calculated 
the average charitable contribution of households in the CES sample by 
income class and imputed this class average to each nonitemizer in the 
same income class in the SOI sample. A limitation of this “mean value 
imputation” method is that it does not fully reflect the distribution 
of charitable giving among all taxpayers. The method imputes some 
contribution to all nonitemizers and imputes the same contribution to 
each nonitemizer in an income class. According to the CES data, most 
households in most income classes make no contributions or make 
contributions that are less than the average contribution. To the 
extent that nonitemizers did not make any charitable contribution or 
made contributions that were less than the average, our method may 
overstate the number of taxpayers who overpaid their taxes.[Footnote 
6] Also, to the extent that nonitemizers made contributions greater 
than the average, the method may understate the number of taxpayers 
who overpaid their taxes. 

However, we believe that our estimates of charitable contributions are 
likely to produce reasonably accurate estimates of overpaid taxes 
despite this limitation. The average contribution does not exceed the 
standard deduction for any filing status. (The contribution amounts 
range from $135 for those with less than $5,000 income to $1,348 for 
those with more than $70,000 income.) Only when these imputed 
contributions are added to other deductions can the total exceed the 
standard deduction. Our analysis of itemizers shows that taxpayers 
with the deductions that we did not impute, the mortgage interest and 
state and local tax payments, are more likely to make some charitable 
contribution. Furthermore, the amount of their contribution does not 
vary consistently with the amount of these other deductions. To the 
extent that charitable contributions are related to other deductions 
in the same way for nonitemizers, imputing the same average 
contribution to each taxpayer is likely to produce a reasonably 
accurate estimate of the effect on overpayment of taxes resulting from 
this additional deduction. 

For our real estate tax estimates, in addition to separating 
households in the CES sample into income classes, we separated the 
households into two classes depending on whether or not they lived in 
a metropolitan statistical area. We calculated the average real estate 
tax for these combined income and location classes and imputed this 
class average to each nonitemizer in the same income and location 
class in the SOI sample who also had paid mortgage interest. A 
limitation of this method is that, to the extent that nonitemizers 
without mortgage interest owned real estate and paid real estate taxes 
on this property, our estimates of both the number of overpayers and 
the amount overpaid may be understated. 

For estimates of personal property tax payments, we also used the CES 
sample to calculate average personal property tax by income class and 
imputed this class average to nonitemizers in the same income class in 
the SOI sample. This method has the same limitations as the method for 
estimating charitable contributions. However, for the reasons that we 
described regarding charitable contributions, we believe that this 
method produces reasonably accurate estimates of the effect of this 
additional deduction on overpayment of taxes. 

The estimates of charitable contributions and of personal and real 
estate taxes are less reliable than the estimates of the other 
deductions, because they are likely to have larger nonsampling errors. 
Unlike the mortgage interest and the state and local income tax 
estimates, which are based on tax return information, these estimates 
are based on the CES data and on a procedure for imputing missing data 
to nonitemizers. Our use of survey data and our imputation of missing 
data provide additional sources of nonsampling error that are not 
present with the other estimates. 

To assess the importance of nonsampling error from our use of survey 
data, we reviewed studies of the reliability of the CES data. In the 
studies, the CES estimates were judged more reliable to the extent 
that they agreed with estimates from other sources. The CES estimates 
of property taxes were found in general agreement with estimates from 
another source. However, we found no study that compared CES estimates 
of charitable contributions with estimates from other sources. The 
studies indicated that the CES might underreport income, particularly 
at the lower end of the income distribution. To the extent that this 
is true, imputations based on the CES data may overstate the number of 
taxpayers in lower income categories who overpaid their taxes. 

To assess the importance of nonsampling error from our use of 
imputations, we used an alternative method to impute missing data. As 
described above, a limitation of the imputation method used in our 
report is that in imputing the same average value by income class to 
all nonitemizers, it does not reflect the fact that many households in 
the CES reported making no charitable contributions or personal 
property tax payments. To test the effect of this limitation on our 
estimates, we randomly imputed zero values for charitable 
contributions and personal property tax payments to nonitemizers in 
the same proportion by income class as the proportion of households in 
the CES that reported making no contributions or payments. We imputed 
to the remaining nonitemizers the average charitable contribution and 
personal property tax payment of only those households that reported 
making such contributions and payments. We found that using this 
alternative imputation method had very little effect on our estimates. 
The estimated number of returns with overpayments increased by about 2 
percent and the amount overpaid increased by about 8 percent above the 
estimates in our report. 

The net effect on overpayment of taxes of our imputed estimates of 
charitable contributions and real estate and personal property taxes 
depends on how these three deductions are related to the other 
itemizable deductions that we estimated. The sum of the largest 
possible values for each of these imputed estimates does not exceed 
the standard deduction for any filing status. Only when these 
imputations are added to the other deductions can the total exceed the 
standard deduction. 

Components of Overpayment: 

The components of our estimate of overpaid taxes are shown in table 4. 
Beginning with the component of overpaid taxes for which our data are 
most reliable—-mortgage interest and points-—we identified more than 
half a million taxpayers whose estimated mortgage interest and points 
alone exceeded their standard deduction. The addition of state and 
local income tax withholding adds another approximately 430,000 
taxpayers whose estimated mortgage interest and points plus state and 
local income taxes exceeded their standard deduction. 

When imputed itemizable deductions are added to our estimated 
deductions, another approximately 1,209,000 taxpayers are found to 
have estimated mortgage interest and points, state and local income 
taxes, charitable contributions, and real estate and personal property 
taxes that exceeded their standard deduction. 

Table 4: Estimated Number of Returns on Which Taxpayers May Have 
Overpaid Taxes because Itemizable Deductions May Have Exceeded the 
Standard Deduction, Tax Year 1998 (numbers in thousands): 

Return category: Returns on which mortgage interest and points 
exceeded the standard deduction; 
Estimated number of returns: 518; 
Estimated additional returns: 518; 
Percentage of total additional returns: 24%; 
95% confidence interval: 440–611. 

Return category: Returns on which mortgage interest and points and 
state and local income tax exceeded the standard deduction; 
Estimated number of returns: 948; 
Estimated additional returns: 430;
Percentage of total additional returns: 20%; 
95% confidence interval: 607–1,411. 

Return category: Returns on which mortgage interest and points, state 
and local income tax, and real estate tax exceeded the standard 
deduction; 
Estimated number of returns: 1,319; 
Estimated additional returns: 371; 
Percentage of total additional returns: 17%; 
95% confidence interval: 964–1,760. 

Return category: Returns on which mortgage interest and points, state 
and local income tax, real estate tax, and charitable contributions 
exceeded the standard deduction; 
Estimated number of returns: 1,899; 
Estimated additional returns: 580;
Percentage of total additional returns: 27%; 
95% confidence interval: 1,403–2,512. 

Return category: Returns on which mortgage interest and points, state 
and local income tax, real estate tax, charitable contributions, and 
personal property tax exceeded the standard deduction; 
Estimated number of returns: 2,157; 
Estimated additional returns: 258; 
Percentage of total additional returns: 12%; 
95% confidence interval: 1,602–2,838. 

Note: Totals may not sum because of rounding. 

Source: GAO analysis of SOI and other IRS data and Department of Labor 
CES data. 

[End of table] 

[End of section] 

Footnotes: 

1. U.S. General Accounting Office, Tax Deductions: Estimates of 
Taxpayers Who May Have Overpaid Federal Taxes by Not Itemizing, 
[hyperlink, http://www.gao.gov/products/GAO-01-529] (Washington, D.C.: 
Apr. 12, 2001). 

2. The five filing statuses are single, married filing jointly, 
qualified widow(er) with dependent child, head of household, and 
married filing separately. The standard deduction amounts vary by 
filing status. For tax year 1998, the basic standard deduction was 
$4,250 for filing single, $7,100 for married filing jointly or as a 
qualified widow(er), $6,250 for head of household, and $3,550 for 
married filing separately. 

3. For example, the basic standard deduction for single filers was 
$4,250 in 1998 with an additional standard deduction of $1,050 for 
filers who were aged 65 years or older, blind, or both, for a 
potential total standard deduction of $6,350. The basic standard 
deduction for married filing jointly was $7,100 with an additional 
standard deduction of $850 for each spouse who was aged 65 or older, 
blind, or both, for a potential total standard deduction of $10,500. 

4. We selected these five itemizable payments for our estimates in 
part because, among those who did itemize their deductions in tax year 
1998, mortgage interest and deductible mortgage points, taxes paid, 
and charitable contributions were the largest itemized deductions, 
together accounting for 90 percent of all itemized deductions. 

5. The charitable contributions in the CES are contributions of cash, 
stocks, and bonds to religious, educational, and other charitable 
organizations. To the extent that nonitemizers made contributions of 
other types of property that are not included in the CES but that may 
be deductible on tax returns, our method may understate the number of 
taxpayers who overpaid their taxes. 

6. The CES data do not allow us to identify the charitable 
contributions made by nonitemizers. In 1986, when taxpayers who did 
not itemize could fully deduct their charitable contributions, about 
45 percent of nonitemizers made some charitable contribution, compared 
with about 63 percent of all taxpayers. The nonitemizers gave an 
average of $474, while the average for all contributors was $1,036. To 
the extent that these relative giving practices still applied in 1998, 
nonitemizers would be, like most households in the CES, less likely to 
make contributions and more likely to contribute less than the average 
amount. 

[End of section] 

GAO’s Mission: 

The General Accounting Office, the 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 the Internet. GAO’s Web site [hyperlink, 
http://www.gao.gov] contains abstracts and fulltext files of current 
reports and testimony and an expanding archive of older products. The 
Web site features a search engine to help you locate documents using 
key words and phrases. You can print these documents in their entirety, 
including charts and other graphics. 

Each day, GAO issues a list of newly released reports, testimony, and 
correspondence. GAO posts this list, known as “Today’s Reports,” on its 
Web site daily. The list contains links to the full-text document 
files. To have GAO e-mail this list to you every afternoon, go to 
[hyperlink, http://www.gao.gov] and select “Subscribe to daily E-mail 
alert for newly released products” under the GAO Reports heading. 

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. General Accounting 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: [hyperlink, http://www.gao.gov/fraudnet/fraudnet.htm]: 
E-mail: fraudnet@gao.gov: 
Automated answering system: (800) 424-5454 or (202) 512-7470: 

Public Affairs: 

Jeff Nelligan, managing director, NelliganJ@gao.gov: 
(202) 512-4800: 
U.S. General Accounting Office: 
441 G Street NW, Room 7149:
Washington, D.C. 20548: 

[End of document]