Tax Administration: Nonfiling Among U.S. Citizens Abroad (Letter Report,
05/11/1998, GAO/GGD-98-106).

GAO was asked to review available data on the tax compliance of U.S.
citizens living abroad. GAO found the Internal Revenue Service (IRS) has
not estimated the overall prevalence of nonfiling abroad or the
resulting loss of tax revenue, and the data GAO identified in its review
were inadequate to support reliable quantified estimates. Data on the
number of U.S. taxpayers residing abroad and the number of returns they
file are of uncertain reliability, and the amount of taxes that
nonfilers would owe if they were to file is unknown. One recent IRS
initiative, however, focused on certain Mideast countries and identified
enough nonfilers and additional tax revenue that IRS believes there may
be benefit to looking for concentrations of nonfilers in other foreign
countries.

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

 REPORTNUM:  GGD-98-106
     TITLE:  Tax Administration: Nonfiling Among U.S. Citizens Abroad
      DATE:  05/11/1998
   SUBJECT:  Americans employed abroad
	     Income taxes
	     Tax administration systems
	     Tax returns
	     Voluntary compliance
	     Taxpayers
	     Tax nonpayment
	     Noncompliance

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************
GAO/GGD-98-106

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

Report to the Chairman Committee on Ways and Means, House of
Representatives

May 1998

TAX ADMINISTRATION - NONFILING
AMONG U.S.  CITIZENS ABROAD

GAO/GGD-98-106

Nonfiling Among U.S.  Citizens Abroad

(268973)

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

  AC - Assistant Commissioner
  IRC - Internal Revenue Code
  IRS - Internal Revenue Service
  OPM - Office of Personnel Management
  SSN - Social Security number

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

B-276745

May 11, 1998

The Honorable Bill Archer
Chairman, Committee on Ways and Means
House of Representatives

Dear Mr.  Chairman:

This report responds to your request that we review available data on
the tax compliance of U.S.  citizens residing in foreign countries,
particularly whether they file required U.S.  tax returns.  The
Committee was concerned that nonfiling among U.S.  citizens residing
abroad could result in lost tax revenue and undermine the
expatriation provisions enacted in August 1996 in the Health
Insurance Portability and Accountability Act (P.L.  104-191).  The
provisions increased the authority of the Internal Revenue Service
(IRS) to tax former U.S.  citizens who renounced their citizenship to
avoid paying U.S.  taxes.

As agreed with your office, our objectives for this report were to
(1) determine whether it is possible, given available data, to
estimate the prevalence and revenue impact of nonfiling among U.S.
citizens residing abroad; (2) identify factors that may limit IRS'
enforcement of the filing requirement or otherwise contribute to
nonfiling abroad; (3) describe IRS' recent initiatives to improve the
filing compliance in this population; and (4) contrast the Department
of the Treasury's study on the income tax compliance of U.S.
taxpayers residing abroad with our results in this study.

   BACKGROUND
------------------------------------------------------------ Letter :1

U.S.  citizens residing abroad are generally subject to the same
filing requirements as citizens residing in the United States.  In
particular, section 6012 of the Internal Revenue Code (IRC) requires
individuals to file tax returns if they meet certain gross income
thresholds, regardless of whether or not they owe taxes.  Individuals
residing abroad must file tax returns even if they think their income
is exempt from tax under the foreign earned income and housing
expense exclusions.  Without a return, IRS cannot verify a taxpayer's
interpretation of the rules limiting eligibility for the exclusions.

Under IRC section 911, U.S.  citizens or resident aliens may qualify
to exclude up to $70,000 per year of their foreign earned income
through 1997,\1 and an additional amount based on their housing
expenses if they meet certain foreign residency requirements.
Nonfilers detected by IRS before filing voluntarily lose their
eligibility for the exclusions in some circumstances.  (See app.  I
for additional information on the exclusions and related rules
affecting U.S.  citizens residing abroad.)

IRS' Office of the Assistant Commissioner (International)--AC
(International) is responsible for all international tax matters.  To
support its mission, AC (International) maintains about 13 full-time
personnel at 9 foreign posts of duty.  Additionally, some staff who
are normally based in the United States are available for temporary
tours of duty in foreign countries.

We have responded to two earlier congressional inquiries into
nonfiling by U.S.  citizens residing abroad.  In a 1985 testimony, we
noted that our analysis of filing among a limited sample of U.S.
citizens in selected countries indicated a potential nonfiling
problem.\2 As a result, Congress enacted IRC section 6039E:
Information Concerning Resident Status in the Tax Reform Act of 1986.
This section includes provisions requiring U.S.  citizens applying
for passports to provide their Social Security number (SSN), any
foreign country of residence, and other information that might be
prescribed by the Treasury Department.  The intent of section 6039E
was that IRS would use this information to identify nonfilers
residing abroad.

In May 1993, we reported on IRS' relevant compliance initiatives, the
lack of reliable data on U.S.  citizens abroad, and IRS' limited use
of passport application data as a compliance tool.\3

--------------------
\1 The Taxpayer Relief Act of 1997 increased the limitation on the
exclusion for foreign earned income from $70,000 to $80,000 in $2,000
increments each year beginning in 1998, and provides that the
limitation is indexed for inflation beginning in 2008.

\2 See U.S.  Citizens Residing in Foreign Countries and Not Filing
Federal Income Tax Returns testimony before the Subcommittee on
Commerce, Consumer and Monetary Affairs, House Committee on
Government Operations (May 8, 1985).

\3 See IRS Activities to Increase Compliance of Overseas Taxpayers, a
report to the Chairman of the Senate Finance Committee
(GAO/GGD-93-93, May 18, 1993).

   RESULTS IN BRIEF
------------------------------------------------------------ Letter :2

IRS has not estimated the overall prevalence of nonfiling abroad or
the resulting loss of tax revenue, and the data we identified in our
review were inadequate to support reliable quantified estimates.
Data on the number of U.S.  taxpayers residing abroad and the number
of returns they file are of uncertain reliability, and the amount of
taxes that nonfilers would owe if they were to file is unknown.  One
recent IRS initiative, however, focused on certain Mideast countries
and identified enough nonfilers and additional tax revenue that IRS
believes there may be benefits to looking for concentrations of
nonfilers in other foreign countries.

We were able to identify several factors that may limit IRS'
enforcement of the filing requirement or otherwise contribute to
nonfiling abroad.  Some of these factors are beyond IRS' control.
First, the income of U.S.  citizens residing abroad is generally not
subject to U.S.  tax withholding or information reporting if it is
derived from foreign employers or foreign financial investments.  IRS
data show that tax withholding and information reporting by employers
or other income providers resulted in much higher rates of tax
compliance than when neither system is in place.\4 Second, IRS
generally cannot collect unpaid taxes from assets that have been
transferred to a foreign country.  The enforcement actions, such as
liens, levies, and seizures, that IRS uses in the United States have
no legal standing in most foreign countries.\5

Some factors, however, are within IRS' control.  First, although IRS
obtains passport data from the State Department, it has made little
use of these data; and in recent years, IRS has not attempted to
penalize the large number of applicants who fail to furnish an SSN,
as the law provides.  Second, IRS has no systematic way of capturing
a passport applicant's country of residence and occupation, which
could provide demographic data on foreign concentrations of U.S.
citizens and help IRS distinguish them from tourists.  Third, the
instructions for filing Form 1040 are potentially misleading and may
cause some taxpayers residing abroad to erroneously conclude that
they have no obligation to file.

IRS' recent initiatives concerning nonfiling abroad include a special
project in the Middle East that was initiated as a result of events
related to the Desert Storm War and a data-gathering effort to
identify other potential concentrations of nonfilers residing abroad.
IRS estimates that the Mideast effort contributed to a 51-percent
increase in returns filed from the area and a total revenue increase
of about $76 million due to additional returns filed from Saudi
Arabia from 1992 through 1995.  In fiscal year 1997, IRS began to
gather foreign census and other demographic information on U.S.
citizens residing abroad to identify other countries where similar
compliance efforts may be beneficial.

In its May 4, 1998, report, Treasury suggested that the revenue
impact of nonfiling abroad may be limited by the foreign earned
income and housing expense exclusions and foreign tax credits.
Treasury also cited IRS' ongoing demographic project as a means of
identifying particular countries where additional compliance efforts
may be warranted.  The report did not recommend any additional IRS
actions for improving tax compliance abroad beyond those that IRS is
already planning.  By contrast, our review found that certain
additional IRS actions appear to be warranted.

--------------------
\4 See Taxpayer Compliance:  Analyzing the Nature of the Income Tax
Gap (GAO/T-GGD-97-35, Jan.  9, 1997).

\5 The United States has mutual collection assistance agreements with
Canada, France, Denmark, Sweden, and the Netherlands.

   SCOPE AND METHODOLOGY
------------------------------------------------------------ Letter :3

To explore the possibility of estimating the prevalence of nonfiling
abroad, we obtained State Department and foreign government estimates
of U.S.  citizens abroad and IRS data on returns filed from abroad.
We also obtained Social Security Administration data on the number of
Social Security beneficiaries and Office of Personnel Management
(OPM) data on the number of federal and military retirees residing
abroad.  We looked at data on the number of nonfilers abroad
identified through IRS' information matching program.  And, we
attempted to use data IRS received from the State Department to
assess the extent of filing among recent passport applicants who
cited foreign addresses.  The details of our scope and methodology
for this objective are discussed in appendix II.

Estimating the revenue impact of nonfiling requires information on
the average tax liability of nonfilers in addition to an estimate of
prevalence.  We identified little data bearing on the tax that
nonfilers abroad would owe if they were to file.  However, we did
obtain the average tax owed by those who file from abroad and the
taxes assessed in audits of nonfilers detected by IRS, but neither
can be reliably projected to nonfilers abroad in general.

To identify the factors that may limit IRS' enforcement of the filing
requirement or otherwise contribute to nonfiling abroad, we talked
with responsible officials in AC (International) and the nonfiler
program under AC (Collection) regarding relevant compliance
information and programs and their limitations.  We obtained IRS data
summarizing the results of its information matching and audit
programs for individual taxpayers abroad.  We also reviewed relevant
sections of the tax code and IRS regulations and obtained general
information on the enforcement tools available to IRS through U.S.
tax treaties or administrative agreements with other
nations.\6

To describe IRS' recent initiatives to address nonfiling abroad, we
talked to responsible officials in AC (International) and obtained
documentation describing the initiatives they cited.  We also talked
to them about the status of initiatives under way when we issued our
1993 report.\7

To contrast the Treasury study of noncompliance abroad with our
study, we reviewed its report in light of the information we gathered
in this review.  We also contacted Treasury Department and IRS
officials to clarify our understanding of the report.

We conducted our review from October 1997 through April 1998 in
accordance with generally accepted government auditing standards.  We
requested comments from IRS, the Treasury Department, and the State
Department and their oral comments are discussed at the end of this
report.

--------------------
\6 In particular, we reviewed IRC section 911:  Citizens or Residents
of the United States Living Abroad, IRC section 6039E:  Information
Concerning Resident Status, and IRS regulation 1.911-7 on procedures
for electing the foreign earned income and housing exclusions.

\7 GAO/GGD-93-93.

   DATA LIMITATIONS IMPEDE
   RELIABLE ESTIMATION OF
   NONFILING ABROAD
------------------------------------------------------------ Letter :4

U.S.  citizens, regardless of where they reside, are generally
required to file income tax returns.  Thus, U.S.  citizens abroad who
exceed certain annual income thresholds are generally required to
file tax returns.\8 Estimates of the numbers of citizens who are
required to file and those who did not could possibly be made if
there were reliable data on the total U.S.  population residing
abroad, related demographic characteristics, and the number of
returns they filed.  However, the data we obtained on the U.S.
population residing abroad--from State Department and foreign
government estimates--and the number of returns they filed are too
uncertain to support such estimates.

We did obtain some information concerning nonfiling abroad from a
recent IRS compliance project and IRS' information matching results.
The information is not definitive, but it does indicate that there
was a serious nonfiling problem in one region of the world (the
Middle East) in the early 1990s and that nonfiling could be
relatively prevalent abroad, compared with the general U.S.
population, among higher income taxpayers who are covered by
information reporting.

We also attempted to determine if the prevalence of nonfiling abroad
could be estimated by using passport application data IRS receives
from the State Department.  These data were not useful, however,
because many applicants did not provide an SSN on their passport
applications, as required by IRC section 6039E.  Generally, it is
difficult for IRS to match taxpayer information against its database
of filed tax returns without a valid SSN or other identification
number.

Given the limitations of available data, the total revenue impact of
nonfiling abroad cannot be reliably estimated.  Estimating revenue
impact would require reliable information concerning the number of
U.S.  citizens residing abroad, the number who would be required to
file tax returns, the extent of nonfiling, and the amount of tax
nonfilers would owe if they were to file.  IRS' most recent estimate
of the revenue lost to individual nonfilers residing in the United
States--$13.8 billion in 1992--illustrates the difficulty in deriving
reliable estimates of the revenue losses attributable to nonfiling.
According to an official in IRS' research division, (1) the estimate
is limited to nonfilers residing in the United States and
incorporates assumptions, necessitated by data limitations, about
taxes owed by nonfilers who could not be identified or located; and
(2) the statistical reliability of the estimate has not been
quantified.

--------------------
\8 For example, in 1997, the threshold amounts for individuals under
age 65 were $6,800 if single and $12,200 if married and filing
jointly.

      LIMITED DATA ON U.S.
      POPULATION ABROAD
---------------------------------------------------------- Letter :4.1

The State Department estimated the total population of U.S.  citizens
residing abroad at about 3.1 million in 1995, excluding active
military and current government personnel.  This number was based on
estimates derived by 221 U.S.  embassies and consulates, does not
include demographic breakdowns, and is not meant to be statistically
reliable.  The posts' estimates are intended only as rough population
indicators to be used in evacuation planning.

Officials at the 18 U.S.  embassies and consulates contacted during
our review reported that they used various sources of information in
deriving their estimates, such as data on the number of U.S.
citizens renewing passports or voluntarily registering at the post or
data obtained from the host country.  Data limitations required the
posts to use subjective judgment in deriving the estimates.  For
example, posts attempted to adjust their estimates to account for
certain limitations in the registration data, e.g., eight posts
estimated that the majority of U.S.  citizens residing in their
jurisdictions were not registered.  Also, some who do register may
remain on file even after they leave the country.

Many foreign governments collect data on the nationality of their
residents, sometimes by age group, including the number who are U.S.
citizens.  The foreign data are not comparable with the State
Department data because of differences in how U.S.  citizens are
defined.  For example, the estimates from many of the U.S.  embassies
and consulates we contacted included U.S.  citizens who are dual
nationals, particularly individuals who were born abroad but acquired
U.S.  citizenship by virtue of a parent's citizenship; while some of
the foreign estimates we obtained did not count such individuals as
U.S.  citizens.  The foreign estimates we obtained are also not
comparable across countries; for example, some countries count their
resident aliens based on country of birth and others based on
citizenship.  The latter approach would include some naturalized
citizens not born in the United States.

Different countries obtain their estimates in different ways.  For
example, some countries rely on census counts of individuals
intending to reside in the country for a certain time, while others
use data on immigrants granted permanent residence status, and some
countries exclude U.S.  citizens in certain age categories.

Also, different estimates for the same country can vary widely, and
it is not always clear who is being counted.  For example, a 1991
Italian Census report noted 15,031 U.S.  citizens residing in Italy
while Eurostat counted 62,066 in 1993.\9 Given the limited
methodological descriptions in the reports we obtained and the
translation difficulty, we could not determine exactly how the U.S.
population was defined in these cases.  Further, data from the Social
Security Administration indicated that about 14,000 U.S.  citizens
resided in Italy and received U.S.  Social Security benefits in 1996.

We did not contact foreign government officials about the reliability
of their data on U.S.  citizen populations because of resource
constraints and because limitations in IRS' data on returns filed
from abroad, discussed below, could limit the usefulness of
country-specific data.  Analysts in the U.S.  Census Bureau's
International Program Center told us that data from foreign censuses
in developed countries are generally reliable.  However, the Census
officials were not specifically knowledgeable about foreign estimates
of U.S.  citizens residing abroad.

IRS classifies individual tax returns as being "international" if the
return cites a foreign mailing address or includes a Form 2555
claiming the foreign earned income or housing expense exclusions.
Returns reporting amounts in foreign denominations or attaching
foreign earnings reports are also classified as international
returns.  However, these data are of uncertain reliability as an
indicator of total returns filed by U.S.  citizens residing abroad.
IRS' classification generally has not captured returns filed by
individuals who lived abroad during the tax year but cited a domestic
address on their return and did not claim the exclusions.  IRS, too,
has found that its computer system continued to classify some
individuals as international filers even after the tax year when they
returned to the United States.

The reliability of IRS' data on returns filed from a particular
country is further limited because IRS' data do not track the filer's
country of residence in some cases.\10 And IRS' data on returns filed
include, but do not distinguish from other returns, returns from
permanent resident aliens of the United States who are living
abroad.\11 These individuals are not U.S.  citizens and therefore
would not be included in the State Department or foreign government
estimates of the U.S.  population abroad.

Table 1 summarizes data available on U.S.  citizens abroad and
returns filed from abroad in tax year 1995 in total and for the seven
countries in which State Department estimates indicated more than
100,000 U.S.  citizens reside.  The table illustrates the variations
in available estimates of the U.S.  population abroad and the lack of
comparable data across countries.  We note the number of tax returns
filed from a particular country as "unknown" because a large
percentage of the returns received from abroad are not differentiated
by country in IRS' database.

                                     Table 1

                     Estimates of the Number of U.S. Citizens
                       Abroad and the Number of Tax Returns
                                      Filed

                      Estimates
          ----------------------------------
                                                                   Returns filed
          State Department           Foreign   Year of foreign     (for tax year
Country           (1995)\a        government              data           1995)\b
--------  ----------------  ----------------  ----------------  ----------------
All              3,051,491               N/A               N/A           380,577
Canada           622,035\c           249,080              1991                \d
Germany            121,749           108,300              1994                \d
Israel             158,400               N/A               N/A                \d
Italy              147,000          62,066\e              1993                \d
Mexico             514,161           194,600              1990                \d
Philippi           123,000            19,529              1990                \d
 nes
United             216,000           143,484              1991                \d
 Kingdom
--------------------------------------------------------------------------------
Note 1:  U.S.  active military and current government personnel are
excluded from State Department estimates and from IRS data on returns
filed.  Their treatment could not be determined from available
foreign data.

Note 2:  N/A represents not available.

\a State Department estimates are limited to data from 221 of 247
U.S.  posts abroad.

\b IRS data, excluding returns from nonresident aliens.

\c Based on 7 of 8 U.S.  posts that reported data (excludes the
Halifax area).

\d Number of returns filed were unknown.

\e From data collected by Eurostat.

Sources:  Department of State; foreign government estimates obtained
from the U.S.  Census Bureau International Programs Center and
Eurostat; and IRS data on returns filed as obtained from the Midwest
Automated Compliance System.

The above data, even if reliable, would not provide the number or
proportion of actual nonfilers abroad because the number of
individuals required to file is unknown.  We explored whether the
number of nonfilers abroad--those who are required to file but do
not--might be roughly estimated by using the ratio of total
individual returns filed to total U.S.  population, about 0.45 in
recent years,\12 as a benchmark.  In particular, a ratio of returns
filed from abroad to U.S.  population abroad that is much smaller
than 0.45 might indicate proportionately more nonfilers in the
population abroad than in the general U.S.  population.  However,
available data on the U.S.  population abroad and the number of
returns they file is too uncertain to allow a reliable comparison
with the general population.  Such an analysis would also require
data on how characteristics related to the filing requirement compare
in the two populations, particularly the age and income
distributions.

We identified two other sources of information that, while not
definitive or indicative of the overall extent of the problem, imply
that nonfiling may be a problem in certain segments of the U.S.
population abroad.

  -- IRS estimates that its Mideast compliance project, described in
     more detail later in this report, was largely responsible for a
     51-percent increase in returns filed by U.S.  citizens residing
     in the region.  IRS does not know whether those results reflect
     that nonfiling was more or less prevalent among U.S.  citizens
     residing in Mideast countries compared with other areas of the
     world.  The region's representativeness depends in part on how
     it compares with other parts of the world in terms of the number
     of U.S.  citizens employed there by foreign corporations.  Most
     of the nonfilers IRS identified in the Middle East worked for
     foreign companies, which do not participate in U.S.  information
     reporting or tax withholding.  In general, IRS has found much
     higher rates of noncompliance among individuals not covered by
     these systems.

  -- IRS data on nonfilers identified through its information
     matching program, which we did not verify, indicate that
     nonfiling among those who have relatively high incomes and are
     covered by information reporting may be more common among U.S.
     citizens abroad than in the U.S.  population generally.\13 IRS
     relies on an automated system to select the potential nonfiler
     cases identified in its information matching program that may
     warrant subsequent enforcement action.  IRS' system identified
     21,852 individuals classified as residing abroad who were
     potential nonfilers for tax year 1995 and had sufficient income
     reported on information returns or met other criteria that cause
     IRS to issue a delinquency notice.  Using the same criteria, the
     system selected about 1.9 million individuals from the total
     U.S.  population for the same year.

Compared with the number of returns that were filed--about 935,000
returns classified as filed from abroad in 1995 versus 118 million
filed from the general population--the number of potential nonfilers
abroad who were selected to receive notices was about 40-percent
larger, proportionately, than the number identified in the general
U.S.  population.\14

--------------------
\9 Eurostat is a statistical organization that collects demographic
statistics from the 15 countries in the European Union.

\10 About 43 percent of individual returns classified as being
international in tax year 1995 were not classified by country in one
IRS computer system (the Midwest Automated Compliance System) because
they did not claim the foreign earned income or housing expense
exclusion.  About 33 percent of international returns for tax year
1993--the latest year for which data were available at the time of
our review--were not classified by country in the other computer
system used to track the number of international returns filed (the
Compliance Research Information System) because they did not cite a
foreign mailing address.

\11 Permanent resident aliens are required to file U.S.  tax returns
on their worldwide income, generally using the same forms as U.S.
citizens, regardless of where they reside.  Nonresident aliens are
required to file U.S.  tax returns on their U.S.  source income in
some circumstances, via Forms 1040-NR.  Forms 1040-NR are counted and
tracked separately.

\12 This represents 118 million returns divided by an estimated U.S.
population of 263 million in 1995.  The ratio was also 0.45 for 1994
and 1996.

\13 IRS identifies potential nonfiler cases by computer matching
taxpayers' returns with information returns IRS receives from
employers and other income providers.  Potential nonfilers are
identified when the match shows income but no corresponding tax
return.

\14 That is, 21,852/935,348 = 0.023, while 1,917,000/118,218,000 =
0.016; and 0.023 is 44 percent greater than 0.016.

      MISSING SSNS LIMIT
      USEFULNESS OF PASSPORT
      APPLICATION DATA
---------------------------------------------------------- Letter :4.2

We obtained passport application data to determine if they could be
matched against IRS' database of SSNs from filed tax returns to help
estimate the number of U.S.  citizens residing abroad who did not
file tax returns.  The data include an applicant's date of birth,
which might be useful in identifying adults who are more likely than
children to meet the filing requirement.  However, many of the recent
passport records IRS received from the State Department did not
include SSNs and so could not be readily matched against IRS'
database.  As a result, we could not reliably estimate the number or
proportion of passport applicants who did not file tax returns.

We analyzed 303,000 passport records that listed foreign mailing
addresses and were processed by the State Department in the last half
of 1995 and throughout 1996.\15 About 133,000, or 44 percent of these
records, did not contain SSNs and could not be readily matched.\16

Of about 170,000 records that did contain SSNs, the proportion of
individuals not filing returns, as either primary filers or secondary
filers on a joint return, did not differ dramatically from the
comparable proportion for the general U.S.  population.  In
particular, for tax year 1994, 41 percent of the applicants did not
file compared with 37 percent not filing from the general population.
For tax year 1995, 35 percent of the applicants did not file compared
with 36 percent in the general population.  However, the large number
of applications without SSNs preclude reliable estimation of the
percentage of the total population of passport applicants residing
abroad who did not file tax returns.  (Detailed results related to
the passport data analysis are provided in app.  III.)

--------------------
\15 According to State Department officials, passport applications
citing a foreign mailing address are generally from applicants who
are applying for passports at U.S.  embassies and consulates in
foreign countries and would include long-term residents abroad who
are applying for passport renewals.

\16 The State Department reviewed the actual applications for a
limited sample of 194 records without SSNs and found that 6, or about
3 percent, contained SSNs that were not transcribed when the State
Department converted the data to magnetic media.

      REVENUE IMPACT OF NONFILING
      ABROAD CANNOT BE ESTIMATED
---------------------------------------------------------- Letter :4.3

The revenue impact of nonfiling abroad cannot be estimated, primarily
because the prevalence of nonfiling and the income levels of the
nonfilers are unknown.  The impact could be relatively small or
substantial, depending on the assumptions used in the analysis.

If it were assumed that the U.S.  population abroad contains more
children and low-income individuals than the general U.S.
population, the potential number of nonfilers abroad and the
resulting revenue impact may be small.  Assuming that the foreign
earned income and housing expense exclusions and foreign tax credit
would generally eliminate much of a nonfiler's tax liability would
also tend to minimize the revenue impact.

By contrast, assuming that the State Department's estimate of the
U.S.  population abroad is generally accurate and the population does
not contain proportionately more children and low-income individuals
could imply a potentially large number of nonfilers abroad.  There
could be a substantial revenue impact if these nonfilers have income
characteristics similar to those who do file from abroad.  In 1995,
individuals filing from abroad, excluding military personnel and
nonresident aliens, had an average income tax liability of about
$6,700 despite available exclusions and credits.

Assuming that IRS' tax assessments against nonfilers that are
identified represent the amounts owed by those not identified would
also suggest a relatively large potential revenue impact.  IRS
assessed an average tax of $22,057 on 1,237 nonfilers residing abroad
who were audited in fiscal years 1995 and 1996.  It should be noted,
however, that IRS generally focuses its enforcement efforts on
nonfilers thought to have the highest incomes and largest unpaid tax
liabilities.  Further, IRS generally does not consider the effect of
the foreign income exclusions or foreign tax credits in making the
assessments.  However, the foreign earned income and housing expense
exclusions that could effectively lower overall tax liability are not
necessarily extended to certain nonfilers.\17

--------------------
\17 Under IRS regulation 1.911-7, taxpayers cannot claim the
exclusions unless (1) they file a Form 1040 and a Form 2555 electing
to take the exclusions before IRS discovers that they have not filed
these forms, and (2) they owe tax after the exclusions are
considered.

   FACTORS LIMITING IRS'
   ENFORCEMENT OF THE FILING
   REQUIREMENT OR OTHERWISE
   CONTRIBUTING TO NONFILING
   ABROAD
------------------------------------------------------------ Letter :5

IRS' enforcement of the filing requirement abroad is impeded by the
limited reach of U.S.  law in foreign countries.  In particular, IRS
has no authority to require tax withholding or information reporting
from foreign employers and little ability to enforce collection if a
taxpayer's assets have been transferred abroad.  IRS' enforcement
abroad may be further hampered by its limited use of the information
that is available, particularly the passport application records it
receives from the State Department.  Also, IRS' filing instructions
for individuals may lead some U.S.  citizens residing abroad to
erroneously conclude that they do not need to file tax returns.

      LIMITED INFORMATION
      REPORTING AND TAX
      WITHHOLDING ON U.S.
      CITIZENS ABROAD
---------------------------------------------------------- Letter :5.1

Information reporting and tax withholding from employers and other
income providers are the key tools available to IRS for identifying
nonfilers and reducing the resulting lost revenue, but they have
limited applicability to U.S.  citizens residing abroad who are
employed by foreign companies or derive investment income from
foreign sources.  IRS' tax-gap estimates indicate that those covered
by information reporting and tax withholding pay a far greater share
of their true tax liabilities than those who are not subject to
them.\18 U.S.  citizens residing abroad have generally not been
subject to tax withholding on income earned from foreign employers or
foreign investments, and IRS receives little third-party information
on such income.  U.S.  citizens working abroad for U.S.  employers
are covered by withholding and information reporting, and IRS uses
this information in its matching program to identify some nonfilers
abroad.\19

In recent years, IRS has routinely received information on the
foreign source income of U.S.  citizens only from 19 of the countries
with which the United States has information exchange agreements or
tax treaties.  Even in those countries, the information is limited to
whatever is collected under a foreign country's own tax system.  Most
information received from foreign countries pertains to the
investment income of individuals residing in the United States, while
only 731 of about 302,000 foreign information documents processed for
tax year 1993 pertained to the earned income of U.S.  citizens
employed abroad by foreign companies.  IRS officials believe that
foreign employers and financial institutions generally have not
identified U.S.  citizens who reside abroad or noted their
citizenship on information returns.\20 Additionally, IRS has had
difficulty processing and matching foreign information returns due to
computer system limitations and because most foreign returns do not
include the taxpayer's SSN or are received too late to be processed
as part of IRS' information matching program.

IRS noted that it may receive some additional information on U.S.
citizens abroad through Qualified Intermediary Agreements with
foreign financial institutions beginning in tax year 2000.  Qualified
Intermediary Agreements, introduced by IRS regulations under IRC
section 1441, generally relate to U.S.  withholding by foreign
financial institutions on U.S.  source income paid to foreign
persons; but, IRS expects the agreements will also require the
foreign institutions to report certain information on U.S.  citizens.

--------------------
\18 For example, IRS estimated that in tax year 1992, wage earners
whose wages are subject to information reporting and tax withholding
reported 99 percent of their wages to IRS, while self-employed
informal suppliers, who are not covered by tax withholding and are
unlikely to be covered by information reporting, reported only 19
percent of their business income.

\19 According to IRS Publication 54, Tax Guide for U.S.  Citizens and
Resident Aliens Abroad, U.S.  employers of Americans abroad have the
option of not withholding income taxes to the extent of the foreign
earned income and housing expense exclusions if they have good reason
to believe that an employee qualifies for the exclusions.

\20 Although IRS has received information on the foreign investments
of U.S.  citizens residing in the United States, IRS officials
believe they have received little such information on U.S.  citizens
residing abroad.  This is because an investor's presumed citizenship
in foreign information reporting systems is likely to have been based
on their mailing address.

      IRS LACKS COLLECTION
      AUTHORITY ABROAD
---------------------------------------------------------- Letter :5.2

The mechanisms provided to IRS under U.S.  law for collecting unpaid
taxes, including liens, levies, and seizures, generally cannot be
applied against assets that have been transferred to a foreign
country.  As a result, IRS generally cannot collect unpaid taxes from
assets that have been transferred to a foreign country, except for
the five countries that have entered into mutual collection
assistance agreements as part of tax treaties with the United
States--Canada, France, Denmark, Sweden, and the Netherlands.  Mutual
collection assistance agreements generally provide for each country
to use measures available within its own legal system to collect
taxes owed to its partner in the agreement.  The agreement with
Canada was ratified in 1995, and the others were ratified between
1939 and 1948.  According to IRS documentation on the program's
evolution, the 47-year hiatus between the last two agreements
occurred because the Senate indicated in 1948 that it did not favor
additional agreements of this type.

      IRS' LIMITED USE OF PASSPORT
      APPLICATION DATA
---------------------------------------------------------- Letter :5.3

IRC section 6039E was enacted in 1986 to provide IRS with data from
passport applications processed by the State Department for use in
identifying individuals residing abroad who do not file tax returns.
The law required passport applicants to provide their SSNs, foreign
country of residence, and other information to be prescribed by
Treasury, and established a penalty of $500 for each failure to
provide the required information.  However, IRS has made little use
of passport application data in identifying potential nonfilers
abroad, and some application records are difficult to use because
they lack SSNs, as noted previously.  Also, the State Department does
not capture the country of residence of some passport applicants who
reside abroad, and IRS has not prescribed occupation data among the
items it requires from passport applicants.  Passport applications
contain no income information for directly identifying nonfilers, but
they do contain age and occupation data, which could help IRS
identify individuals who are likely to have gross incomes above the
filing thresholds.

Passport data are included in IRS' matching program, but have rarely
been used to identify potential nonfilers abroad.  The criteria IRS
used in recent years to select potential nonfilers to be contacted
emphasized the total amount of income reported on information
returns.  One low-priority criterion applied to mismatches where IRS
received passport or green-card records, but no corresponding tax
return.  However, only 21 of 21,852 potential nonfilers abroad
selected to receive delinquency notices in 1995 were selected based
on that criterion.  And most of the passport records IRS received
from the State Department cited U.S.  rather than foreign mailing
addresses.\21 Applications that cite foreign mailing addresses are
not flagged or analyzed separately in IRS' returns matching program.
IRS officials said that in the future they plan to obtain passport
data routinely only for those applicants who cite foreign mailing
addresses.  IRS expects that this will reduce the cost of obtaining
the data and make it easier to use in identifying nonfilers abroad.

IRS has not attempted to penalize passport applicants in recent years
for failure to provide their SSNs.  As previously noted, IRS has
difficulty matching records that do not contain SSNs.  IRS officials
believe the penalty program was dropped in 1993 because IRS had
difficulty determining the SSNs of applicants who did not furnish one
on the application.  At that time, IRS generally did not send
inquiries or penalty notices for missing SSNs unless the individual's
SSN could be determined from another source.  IRS officials said that
it is administratively difficult to track penalty cases without
taxpayers' SSNs, but there is currently no rule that requires them to
obtain the applicant's SSN before inquiring about missing
information.  The officials said they are exploring ways of
reinstating the penalty program, possibly by sending correspondence
to the mailing address cited on the application without attempting to
determine the applicant's SSN from another source.

Passport application forms include a statement noting that an SSN
must be provided if the applicant has received one, subject to a $500
penalty.  However, the State Department does not deny passports to
applicants who do not provide an SSN, as it relies on other proofs of
an applicant's citizenship.  Whether it could do so is unclear.
Denying a passport to a U.S.  citizen for failure to provide an SSN
could raise a constitutional issue, based on our review of relevant
court cases.  In particular, the Supreme Court held that the right to
travel is a fundamental liberty and government restrictions on it
must conform to the due process provisions of the 5th amendment.\22

IRS has not collected complete information on the country of
residence and has not obtained occupation data on passport applicants
residing abroad.\23 The data IRS has received has been limited to the
applicant's name, mailing address, date of birth, and SSN if the
applicant provided one.  The applicant's country of residence is
currently not required on passport applications.  According to State
Department and IRS officials, country of residence can be obtained in
some cases from mailing addresses on passport applications, primarily
when a U.S.  citizen residing abroad applies for a passport renewal,
or when U.S.  citizens born abroad apply for passports, although
applicants are not required to cite a foreign address even in these
cases.  Passport application forms do not contain a field for
capturing the country of residence of those applying for a passport
in this country and intending to live or work abroad.\24 Passport
applications do contain a field for the applicant's occupation, but
IRS has not obtained this information routinely or prescribed that
applicants provide it.

According to State Department officials, the cost of capturing
occupation data would include data transcription costs of about 6
cents per record and other costs to revise the computer programs used
to store and retrieve the data.  State Department officials also
believe that the passport application form would need to be revised
to capture the country of residence and to provide additional
instructions to the applicant.  The officials said that they have not
estimated the cost of modifying the relevant computer programs or
revising the application form.  IRS officials noted that certain IRS
computer programs would also need to be modified to process the
additional data, and, based on a preliminary estimate, this could
require the equivalent of about 2 staff years at the GS-12/13 level
and $10,000 for related equipment and software upgrades.

IRS proposed regulations on section 6039E in 1993 that would have
required applicants to provide their country of residence, address
within the country of residence, occupation, and other information.
The Office of Chief Counsel is working to finalize the regulations in
1998.  An official in IRS' Office of Chief Counsel said that one
reason the proposed regulations were not finalized earlier is that
section 6039E already provides IRS with the authority to prescribe
the information required from passport applicants without specifying
the requirements in regulations.

In the 1960s and 1970s, U.S.  citizens residing abroad and applying
for passports or registering at U.S.  consular posts abroad were
asked to complete an IRS Form 3966:  Identification of U.S.  Citizen
Residing Abroad.  U.S.  citizens were asked to voluntarily provide
their foreign mailing address, occupation, date of last filed tax
return, and other identifying information.  When they learned that
completing the form was voluntary, many citizens declined to do so.
For this reason, and because some complained that the form
constituted an invasion of their privacy, IRS discontinued the form
in 1979.

--------------------
\21 Only about 62,000, or less than 4 percent of the approximately
1.8 million passport records IRS received for the period of April
through June 1996, contained a foreign mailing address, according to
an IRS analysis of records received in that period.  Similar results
were obtained in an earlier IRS analysis of 1992 data.

\22 Relevant Supreme Court cases are Kent v.  Dulles, 357 U.S.  116
(1958) and Zemel v.  Rusk, 381 U.S.  1 (1965).

\23 IRS' Memorandum of Understanding with the State Department has
allowed IRS to obtain hard copies of selected passport applications
as requested on a case-by-case basis.  The hard copies include the
occupation field among other entries.  Also IRS tabulated the country
of residence of recent passport applicants citing foreign mailing
addresses, as part of an analysis we requested.

\24 The State Department generally will not mail passports to foreign
addresses if the application is taken in this country.

      CERTAIN IRS FILING
      INSTRUCTIONS MAY BE SUBJECT
      TO MISINTERPRETATION
---------------------------------------------------------- Letter :5.4

IRS' instructions for Form 1040 and related guidance\25 may
contribute to misinterpretation of the filing requirement among
individuals who think they qualify for the foreign earned income or
housing expense exclusions.  The instructions state that only gross
income that "is not exempt from tax" should be considered in
determining whether the filing threshold is met.  However, income
qualifying for the foreign earned income or housing exclusions must
be included in applying the threshold, as is clarified in Publication
54:  Tax Guide for U.S.  Citizens and Resident Aliens Abroad, even
though the income is "exempt from tax" under section 911.  IRS
generally revises its instructions and publications annually to
reflect statutory changes and to clarify potentially confusing
language.

--------------------
\25 Including IRS' Publication 17:  Tax Guide for Individuals.

   IRS' RECENT INITIATIVES TO
   ADDRESS NONFILING ABROAD
------------------------------------------------------------ Letter :6

IRS has initiated some actions in recent years to improve filing
compliance abroad, but has not yet developed global information on
the prevalence or impact of the problem or the countries where the
problem may be particularly severe.  In particular, IRS initiated a
multiyear compliance project in 1991 aimed at U.S.  citizens working
in the Middle East.  IRS believes that the project resulted in the
recovery of a substantial amount of tax revenue, and is now
attempting to gather foreign census and other demographic data that
might reveal other concentrations of nonfilers abroad with tax
liabilities.  IRS officials cited several other recent or ongoing
projects focused on compliance problems other than nonfiling among
certain categories of U.S.  citizens residing abroad, such as one on
nonreporting of scholarship and grant income among those studying or
teaching abroad and another on highly paid executives claiming tax
deferrals on nonqualified foreign pension plans.

IRS estimates that the Mideast project was largely responsible for a
51-percent increase in the number of returns filed from the
region--from 13,686 in 1991 to 20,647 in 1995.  IRS also estimated
that the increased returns filed from Saudi Arabia from 1992 through
1995 resulted in a total revenue increase of about $76 million.\26
The project was initiated late in 1991 after IRS noticed that many
civilians who returned to the United States during Operation Desert
Storm filed tax returns for the first time in years.  Also, IRS
believed that the potential increase in tax revenue would justify the
compliance resources expended because these countries had no income
tax.  U.S.  taxpayers in these countries therefore could not reduce
their tax liabilities by claiming foreign tax credits.

Revenue agents and other personnel from AC (International) traveled
to the region to conduct informational seminars for U.S.  individuals
concerning their tax filing obligations and possible adverse
consequences from not filing, such as losing eligibility for the
foreign earned income and housing expense exclusions under Treasury
Regulations section 1.911-7.  The seminars were focused on companies
employing a large number of U.S.  citizens, which IRS identified
through the financial news media and information obtained from the
Department of State, the Department of Labor, and other sources.

One foreign employer of about 5,000 U.S.  citizens agreed to provide
IRS with information on its U.S.  employees' income as requested on a
case-by-case basis and also issued a letter to its U.S.  employees
outlining their need to file and pay U.S.  taxes.  Also as part of
this project, IRS mailed delinquency letters to all potential
nonfilers in selected locations, including a warning that they could
lose their right to claim the foreign earned income and housing
exclusions if they did not file voluntarily.  IRS generally sends
such delinquency notices only to potential nonfilers meeting certain
selection criteria based on the amount of income reported on
information returns and other factors.

IRS did not know, at the time of our review, whether other
geographical areas could offer compliance improvement opportunities,
particularly for increased filing of required tax returns, similar to
or greater than those discovered in its Mideast effort.  Early in
fiscal year 1997, IRS began a project to identify countries or
regions where additional compliance projects similar to the Mideast
project might be warranted.  The project is attempting to obtain
demographic data on the number, location, age stratification, and
likely income levels of U.S.  citizens residing abroad.

IRS' sources of information for the project include its own data on
returns filed, population estimates from foreign governments, and
data from the Social Security Administration and OPM on the number of
Social Security beneficiaries and federal retirees residing abroad.
IRS had obtained at least some foreign data from 10 countries as of
December 1997, including some relatively detailed demographic
information obtained directly from foreign governments.  However, IRS
had not obtained data from Canada, Mexico, the United Kingdom,
Israel, Germany, Italy or the Phillipines--the countries where, in
each case, more than 100,000 U.S.  citizens resided in 1995,
according to State Department estimates.  IRS expects to obtain and
analyze data for the countries accounting for about 80 percent of
U.S.  citizens abroad and to release a draft report on the results in
the summer of 1998.  IRS officials believe that the information will
be complete and reliable enough to identify any countries where
additional compliance efforts appear to be warranted.

--------------------
\26 The countries included were Saudi Arabia, Kuwait, the United Arab
Emirates, Bahrain, Qatar, and Egypt.  IRS only estimated the revenue
impact from Saudi Arabia, which accounted for most of the increase in
returns filed.

   TREASURY'S REPORT ON
   NONCOMPLIANCE ABROAD
------------------------------------------------------------ Letter :7

In the Health Insurance Portability and Accountability Act of 1996,
Congress required Treasury to study and report on issues related to
the income tax compliance of U.S.  citizens and resident aliens
residing abroad.  In its report, Treasury discussed the current law
regarding the taxation of U.S.  citizens and permanent residents
residing abroad and the difficulty of administering tax code
provisions affecting expatriates--those who have relinquished their
U.S.  citizenship.  The report included information on IRS'
initiatives to improve compliance among U.S.  taxpayers abroad and
some factors currently limiting these efforts.  It also discussed the
extent to which the Department of State and the Immigration and
Naturalization Service collect information that could help IRS
determine and improve compliance.

Treasury suggested that the revenue impact of nonfiling abroad may be
limited by the foreign earned income and housing expense exclusions
and foreign tax credits.  While available exclusions and credits
would tend to reduce the revenue impact of nonfiling abroad, we note
that the impact would not necessarily be rendered insignificant.
Some nonfilers lose eligibility for the exclusions, and the average
tax liability of those who did file from abroad was about $6,700 in
1995, despite available exclusions and credits.  Also, the IRS
studies that Treasury cited as evidence of limited impact involved a
small number of taxpayers\27 and cannot be used to estimate the
impact of nonfiling abroad because of serious data limitations, as
noted in our 1993 report.\28

IRS' ongoing demographic study is highlighted as an initiative that
will allow IRS to identify the countries where certain compliance
improvement strategies may be warranted.  We could not assess the
effectiveness of this initiative because it was not complete at the
time we performed our work.

The Treasury study cited several factors beyond IRS' control as
inhibiting its efforts to improve compliance levels in the U.S.
population abroad.  These included limitations on information
reported from foreign sources and IRS' authority to enforce
collection in foreign countries, factors which are also noted in our
report.  Our report also cites IRS' limited use of passport data and
potentially unclear filing instructions as factors related to
nonfiling abroad that are within IRS' control.  The Treasury report
discussed the factors that it believes limit the usefulness of
passport data, including limitations in the mailing address as a
means of identifying and locating applicants residing abroad, and the
large number of records received without SSNs.  The report also
suggests that attempting to penalize applicants who do not provide
SSNs could entail more administrative cost than is warranted and
notes that most applicants who do not provide SSNs appear to be under
20 years old.

By contrast, we have recommended that IRS explore certain ways of
obtaining better information from passport applicants and attempt to
enforce the information requirements of section 6039E.  We note that
it is not necessary for IRS to obtain an applicant's SSN from another
source--a high cost factor, according to IRS--because inquiries can
be sent to the mailing address cited on the passport application.
And the applicant's date of birth, included in the data IRS receives,
might allow IRS to focus its efforts on adult applicants.  Finally,
while most of the applicants we analyzed who did not provide SSNs
were under age 20, a significant percentage were adults.  In
particular, 24 percent were at least 30 years old.  And, the age
distribution of the applicants we analyzed is not a reliable
indicator of the age distribution among all applicants residing
abroad because IRS' information on applicants who reside abroad is
incomplete, as noted above.  Due to this limitation, our analysis
excluded U.S.  citizens who applied for their passports in the United
States before moving abroad, but included passports issued to
children who were born abroad to U.S.  citizens.

The Treasury report did not recommend any additional IRS actions to
improve tax compliance abroad, beyond IRS' ongoing demographic
project and planned follow-up.  Treasury noted that State Department
data on U.S.  citizens registered at U.S.  consular posts may be of
some usefulness to IRS, although the Privacy Act could restrict IRS
from obtaining them.  We have not recommended that IRS obtain
registration data because State Department officials believe that
many U.S.  citizens residing abroad do not register, and those who do
register may remain on file even after they have left the country.

The report also noted that modifying U.S.  laws that define when U.S.
citizenship is lost for tax purposes--so that the loss does not occur
until the individual notifies the State Department--could close an
existing loophole.  The loophole might allow some individuals to
avoid U.S.  taxes by claiming a retroactive loss of U.S.
citizenship.

--------------------
\27 The largest study involved a sample of 437 taxpayers suspected of
being nonfilers.  However, as noted in our 1993 report, IRS was
unable to locate or contact 60 percent of the taxpayers in the
sample.  IRS officials further qualified the study because they were
unable to travel abroad to do the necessary work.

\28 GAO/GGD-93-93.

   CONCLUSIONS
------------------------------------------------------------ Letter :8

The extent and impact of nonfiling abroad remain largely unknown, due
to uncertainties in the data we identified on the U.S.  population
abroad and returns filed from abroad.  However, some evidence
suggests that nonfiling may be relatively prevalent in some segments
of the U.S.  population abroad.  And the revenue impact, while
unknown, could be significant even though it would be reduced by
available exclusions and credits.

IRS' ability to identify and collect taxes from nonfilers residing
abroad is restricted by the limited reach of U.S.  law in foreign
countries, particularly U.S.  laws on tax withholding, information
reporting, and IRS' authority to collect taxes through liens, levies,
and seizures.  However, IRS has not fully explored the usefulness of
passport application data as a means of identifying potential
nonfilers abroad and gauging the extent of the problem.  Also, some
of IRS' filing instructions may confuse some taxpayers and cause them
to erroneously believe they are not required to file.

The usefulness of passport data in identifying nonfilers abroad has
been limited because IRS has not (1) enforced the requirement for
applicants to provide their SSNs and other information and (2)
obtained data on the applicant's occupation or, in some cases,
country of residence.  While passport applications contain no income
information, the occupation and age data could help identify
individuals residing abroad who are more likely to have income above
the filing thresholds, provided IRS could reliably distinguish
applicants residing in foreign countries from those who are merely
tourists.  The cost of obtaining additional data elements on
occupation and country of residence would be offset to some degree by
savings from the reduced volume of data processed if IRS carries out
its plan to restrict the data to applicants residing abroad and
exclude tourists who now account for the bulk of the data IRS
receives.

IRS had difficulty enforcing the requirement for applicants to
provide SSNs and could find it difficult to enforce requirements for
additional information on the applicant's occupation and country of
residence.  However, IRS said some of the difficulty in enforcing the
SSN requirement, before abandoning such efforts, stemmed from its
self-imposed constraint of not sending inquiries to applicants unless
their SSN could be determined from other sources.

Another factor that could contribute to nonfiling abroad is the
ambiguity in IRS' filing instructions for Forms 1040 and related
guidance, such as Publication 17.  The current language could be
misinterpreted to mean that income qualifying for the foreign earned
income or housing expense exclusions does not need to be considered
in determining the filing requirement.

IRS has undertaken an initiative--the Mideast Project--to improve
filing compliance among U.S.  citizens residing in one region abroad
and is now attempting to identify other geographical areas where such
efforts may be beneficial.  As of December 1997, IRS had obtained
foreign data from 10 countries, but these did not include the 7
countries where the State Department estimated that the largest U.S.
populations reside.  IRS officials expect to obtain data on about 80
percent of the U.S.  population abroad and release a draft report on
their results in the summer of 1998.  IRS has not analyzed passport
application data to help identify countries where nonfiling among
U.S.  citizens may be particularly severe, and missing SSNs currently
limit the usefulness of the data for this purpose.  While our review
was under way, IRS began efforts to make greater use of passport data
from individuals residing abroad and is exploring ways of reinstating
a program to penalize applicants who do not provide their SSNs.

In its May 4, 1998, report, Treasury suggested that the revenue
impact of nonfiling abroad may be limited by the foreign earned
income and housing expense exclusions and foreign tax credits.  We
note that, while the revenue impact is unknown, it is not necessarily
rendered insignificant by available exclusions and credits.  The
report did not recommend any IRS actions for improving tax compliance
abroad, but it noted that IRS' ongoing demographic project may
identify countries where additional compliance efforts are warranted.
The report also discussed several factors limiting the usefulness of
passport application data.

   RECOMMENDATIONS
------------------------------------------------------------ Letter :9

To obtain better data on the filing compliance of the U.S.
population residing abroad and to promote their understanding of
their filing requirements, the Commissioner of Internal Revenue
should ensure that IRS

  -- assesses the usefulness of country of residence and occupation
     data, in addition to data IRS currently receives from passport
     applicants, as a means of identifying potential nonfilers abroad
     and supplementing IRS' other sources of demographic data on U.S.
     citizens abroad.  The assessment might include reviewing a
     limited random sample of currently available information.

  -- estimates the cost of obtaining the additional data routinely
     for passport applicants residing abroad, including those who
     apply in the United States.  If the estimated costs appear to be
     justified, IRS should (1) prescribe that passport applicants
     provide the additional items and (2) routinely obtain and
     analyze the additional data elements.

  -- undertakes additional efforts to enforce the information
     requirements of IRC section 6039E, including the requirement for
     applicants to provide their SSNs.  One potential effort would be
     to contact a random sample of adult applicants who did not
     provide an SSN, using the mailing address provided on their
     passport application.

  -- revises the instructions for Form 1040 and related guidance,
     such as Publication 17, to clarify that income that qualifies
     for foreign earned income exclusions must be considered in
     determining whether one's gross income exceeds the filing
     threshold.

   AGENCY COMMENTS AND OUR
   EVALUATION
----------------------------------------------------------- Letter :10

We requested comments on a draft of this report from the Commissioner
of Internal Revenue, the Secretary of the Treasury, and the Secretary
of State, or their designated representatives.  In an April 1, 1998,
meeting, responsible Treasury and IRS officials, including IRS'
Deputy Assistant Commissioner (International), provided oral comments
and suggested clarifications, which we have incorporated where
appropriate.  IRS indicated that it generally agreed with the draft
report and two of its four recommendations--on estimating the cost of
obtaining additional types of passport data and revising relevant
filing instructions--but questioned the cost efficiency of
implementing two of the recommendations.  IRS interpreted our
recommendation on assessing the usefulness of certain additional
passport application data as implying that it pay for and routinely
obtain the additional data before knowing if the associated costs are
justified.  We revised the recommendation to reflect that the
assessment could be based on a sample of data currently available to
IRS.

IRS also interpreted our recommendation on attempting to enforce the
information requirements of IRC section 6039E as implying that it
launch a full-scale enforcement program without first testing the
program's cost and feasibility.  We revised the recommendation to
specify that the effort could be limited to a random sample of
applicants who did not provide SSNs.  We believe that such a test
would constitute additional effort to enforce the requirements as
suggested in our recommendation, provided that IRS evaluates the test
and continues or modifies the approach as the results warrant.

The State Department provided written comments dated April 6, 1998,
that suggested clarifications and additional information, which we
have incorporated in this report where appropriate.  In particular,
the State Department noted that providing the additional passport
information suggested in our report would not prove burdensome, but
the Department would be concerned if IRS sought to require passport
applicants to answer extensive questions on their income and its
sources.  The State Department also commented that the draft seemed
to imply that a statutory provision denying a passport to an
applicant who failed to provide an SSN would be successfully
challenged on constitutional grounds.  Our intent was only to note
that such a policy would raise a significant constitutional issue,
and we modified the wording in this report to avoid any unintended
implication as to how a legal challenge would be decided.

--------------------------------------------------------- Letter :10.1

As agreed with your staff, unless you announce its contents earlier,
we plan no further distribution of this report until 30 days from the
date of this letter.  At that time, we will send copies to the
Ranking Minority Member of the House Ways and Means Committee; the
Chairman and Ranking Minority Member of the Subcommittee on
Oversight, Committee on Ways and Means; various other congressional
committees; the Secretary of the Treasury; the Commissioner of
Internal Revenue; and other interested parties.  We also will make
copies available to others upon request.

Please contact me at (202) 512-9110 if you or your staff have any
questions.  The major contributors to this report are listed in
appendix IV.

Lynda D.  Willis
Director, Tax Policy
 and Administration Issues

MAJOR TAX CODE PROVISIONS FOR
INDIVIDUALS RESIDING ABROAD
=========================================================== Appendix I

In general, the foreign earned income exclusion allows taxpayers
meeting specific foreign residency requirements to exclude up to
$70,000 of their earned income, as of tax year 1997.\1 The excludable
amount is to be increased incrementally to $80,000 by 2002 per
modifications to Internal Revenue Code (IRC) section 911 enacted in
1997.

Excludable income is generally limited to amounts earned for services
performed abroad, including salaries and wages (except wages from the
U.S.  government), and does not include income derived from capital,
such as interest, dividends, capital gains, or pension and IRA
distributions.  The foreign housing exclusion generally allows
taxpayers meeting the residency requirements to exclude a portion of
their housing expenses if they are employed abroad.\2 Income
qualifying for the foreign earned income exclusion is reduced by the
amount of the housing exclusion.\3

The foreign income tax credit is available to taxpayers who owe taxes
to foreign governments on their foreign source income.  To claim the
credit, taxpayers must file a Form 1116, which provides for separate
calculation of the credit amount for each of eight different income
categories.

Also, P.L.  104-191, enacted in August 1996, included modifications
to the tax treatment of expatriates and a requirement for the
Treasury Department to report within 90 days on the income tax
compliance of U.S.  taxpayers residing abroad.  The legislative
history indicates that the Treasury report was mandated because of
past difficulties in determining when a U.S.  citizen had committed
an expatriating act with a tax avoidance purpose and thus must
continue to pay U.S.  taxes on their worldwide income.\4

--------------------
\1 Foreign residency in this context requires meeting either a bona
fide residency test, based on the intent and nature of one's stay in
a foreign country for an entire tax year, or a physical presence
test, requiring the taxpayer to be in a foreign country or countries
for at least 330 days of any consecutive 12-month period.  Questions
of bona fide residency are determined by IRS on a case-by-case basis.
To qualify for the exclusions, U.S.  citizens abroad must also
establish that their ï¿½tax home,ï¿½ generally defined as the location of
one's place of business or employment or where one regularly lives,
is located in a foreign country throughout the period of bona fide
residency or physical presence.

\2 The excludable amount is based on total housing expenses less an
amount equal to 16 percent of the salary of a GS-14, step 1
government employee ($9,426 in tax year 1997).  A foreign housing
deduction is available to those meeting the residency requirements
who have self-employment income abroad.

\3 In effect, this provision reduces the total exclusion for earned
income and housing expenses unless the taxpayer's earned income
exceeds the total of $70,000 plus the excludable housing amount.

\4 The expatriation provisions included new criteria, based on an
individual's income and net worth, for determining when expatriation
occurs for tax avoidance purposes.  Those meeting the tax avoidance
criteria are liable for U.S.  taxes on their worldwide income for 10
years after they expatriate.

DESCRIPTION OF METHODOLOGY
========================================================== Appendix II

We obtained data on U.S.  taxpayers residing abroad from the State
Department and from foreign census or immigration reports collected
by the U.N.  Demographic Statistics Section; the International
Programs Center Library of the U.S.  Census Bureau; and Eurostat, a
statistical organization of the European Union.  We contacted
officials at 21 U.S.  consulates and embassies--those reporting more
than 40,000 U.S.  citizens in their jurisdictions--regarding the
information used in developing the State Department's estimates, and
received written responses from 18 of the 21.  We discussed the
reliability of foreign government data with IRS and U.S.  Census
Bureau officials and cross-checked some of the data against estimates
collected by Eurostat and against U.S.  data on the number of Social
Security beneficiaries and federal retirees residing in a given
foreign country.  We found that the reliability of both the State
Department and foreign government estimates is uncertain, as
discussed in our findings.

We obtained IRS data on returns filed from abroad for tax year 1995.
IRS classifies returns as international if filers cite a foreign
mailing address, attach a Form 2555 claiming the foreign earned
income or housing exclusions, or provide other indications of a
foreign residence, such as by reporting their income in foreign
currencies.  We discussed the data's reliability with IRS officials
and found that its reliability is uncertain, for the reasons noted in
our letter.

We also analyzed data on the number of potential nonfilers identified
abroad through IRS' Information Matching Program in 1995 relative to
the number of returns that IRS classified as being filed from abroad
in 1995.  We compared that proportion with the same proportion
calculated for the general U.S.  population in 1995.  This approach
was limited by the uncertainty of IRS' data on returns filed from
abroad and the lack of quantified IRS data on the number of potential
nonfilers who were nonresident aliens.  We included returns from
nonresident aliens in the number of returns filed from abroad for
1995, even though IRS officials believe that nonresident aliens
account for relatively few potential nonfiler cases identified
through information matching.  Excluding nonresident aliens in the
returns filed data would have made the proportion for nonfilers
abroad appear even larger relative to the proportion of nonfilers in
the general U.S.  population.

We also attempted to assess the prevalence of nonfiling abroad by
matching selected passport application records against IRS' database
of SSNs from filed tax returns.  In particular, we asked IRS to
extract foreign addressed passport records from all passport data it
had retained on magnetic media--which included applications processed
by the State Department and forwarded to IRS in the last half of 1995
and in 1996.  We asked IRS to match the SSNs in these passport
records against its database of SSNs from returns filed in tax years
1994 to 1996 to determine the proportion of applicants not filing tax
returns each year, by age category.  However, 44 percent of the
application records did not include SSNs, and so they could not be
readily matched.  This rendered the results inconclusive, as noted in
our findings, because the nonfiling rate found in the cases with SSNs
cannot be projected to the missing SSN cases.  Also, the match
against tax year 1996 returns did not provide useful data because it
did not include some unknown number of returns filed late under a
4-month filing extension available to U.S.  individuals residing
abroad.

DETAILED DATA TABLES
========================================================= Appendix III

This appendix presents the detailed data related to our analysis of
the passport application data IRS receives from the State Department.
Table III.1 shows the percent of individuals not filing income tax
returns among the passport applicants we analyzed who provided SSNs,
compared with the percent not filing from the general U.S.
population.  Those not filing are not necessarily required to
file--that is, those with gross income below the filing thresholds
and, in some circumstances, children whose income exceeds the
thresholds but is reported on their parents' returns are not required
to file.  Table III.2 shows the age stratification of the general
U.S.  population compared with passport applicants with and without
SSNs, as of 1995.

                                   Table III.1

                       Percent of Individuals Not Filing as
                     Primary Filers or Secondary Filers on a
                      Joint Return--Passport Applicants Who
                     Provided SSNs v. General U.S. Population

                            Taxpayer       Number in    Number not   Percent not
Tax year      Age range     category    population\a      filing\a        filing
------------  ------------  ----------  ------------  ------------  ------------
1994          All ages      Passport             170            69            41
                            applicants

                            General          260,372        96,077            37
                            U.S.
                            population

              30 to 59      Passport              82            22            27
                            applicants

                            General          104,669           N/A           N/A
                            U.S.
                            population

1995          All ages      Passport             170            59            35
                            applicants

                            General          262,890        95,692            36
                            U.S.
                            population

              30 to 59      Passport              83            16            19
                            applicants

                            General          106,634           N/A           N/A
                            U.S.
                            population
--------------------------------------------------------------------------------
Note:  N/A represents not available.

\a Numbers in thousands.

Source:  GAO analysis of IRS data and the general population
estimates from the U.S.  Bureau of the Census, Population Division,
release PPL-57.

                                   Table III.2

                     Age Distribution of Passport Applicants
                      Compared With General U.S. Population

    Applicants not providing    Applicants providing
              SSNs                      SSNs            Total U.S. population\b
    ------------------------  ------------------------  ------------------------
Ag
e\
a        Number      Percent       Number      Percent     Number\c      Percent
--  -----------  -----------  -----------  -----------  -----------  -----------
Un           14          0.0           28          0.0            0          0.0
 k
 n
 o
 w
 n
0        54,005         40.6       19,416         11.4       38,727         14.6
 to
 9
10       33,466         25.2       23,717         14.0       37,643         14.1
 to
 1
 9
20       13,067          9.8       16,114          9.5       36,567         13.8
 to
 2
 9
30       10,678          8.0       30,584         18.0       43,938         16.6
 to
 3
 9
40        7,242          5.5       31,226         18.4       39,252         14.8
 to
 4
 9
50        5,218          3.9       22,100         13.0       25,296          9.5
 to
 5
 9
60        4,357          3.3       14,794          8.7       19,891          7.5
 to
 6
 9
70        3,507          2.6        8,970          5.3       15,651          5.9
 to
 7
 9
80        1,199          0.9        2,726          1.6        6,951          2.6
 to
 8
 9
90          128          0.1          344          0.2         1310          0.5
 to
 9
 9
10            0            0            0            0           57            0
 0
 o
 r
 m
 o
 r
 e
================================================================================
To      132,881          100      170,018          100      265,284          100
 t
 a
 l
 \
 d
--------------------------------------------------------------------------------
\a Age of passport applicants as of the end of 1996.

\b Estimated U.S.  population distribution in 1996.

\c Number in thousands.

\d Percentages do not add due to rounding

Source:  GAO analysis of data from IRS (passport records processed in
the last half of 1995 and in 1996) and the U.S.  Census Bureau.

MAJOR CONTRIBUTORS TO THIS REPORT
========================================================== Appendix IV

GENERAL GOVERNMENT DIVISION,
WASHINGTON, D.C.

Joseph Jozefczyk, Assistant Director, Tax Policy and Administration
Issues
Robert Floren, Evaluator-in-Charge
Pamela Pavord, Evaluator
Elizabeth W.  Scullin, Communications Analyst
Don Phillips, Computer Specialist
Shirley Jones, Senior Attorney

*** End of document. ***