Tax Administration: IRS' 1997 Tax Filing Season (Letter Report, 12/29/97,
GAO/GGD-98-33).

Pursuant to a congressional request, GAO assessed the Internal Revenue
Service's (IRS) performance during the 1997 tax filing season, focusing
on: (1) the ability of taxpayers seeking answers to questions to reach
IRS via the telephone; (2) the number of returns filed by means other
than the traditional paper method; (3) IRS' efforts to deal with returns
that have missing or incorrect Social Security Numbers (SSN); (4) the
use of banks, known as lockboxes, to process certain tax payments; and
(5) performance of the imaging system IRS uses to process certain tax
returns.

GAO noted that: (1) the IRS met or exceeded most of its 1997 filing
season related performance goals; (2) of particular note is the
substantial improvement in two important areas where GAO has criticized
IRS' performance in past filing seasons--telephone accessibility and the
use of alternative filing methods; (3) according to IRS data, telephone
accessibility increased from 20 percent during the 1996 filing season to
51 percent during the 1997 filing season; (4) the number of tax returns
filed by means other than the traditional paper method increased by 25
percent over the last year, with the number of returns filed by
telephone (TeleFile) showing the most significant increase--65 percent;
(5) although the revised tax package apparently contributed to an
increase in the use of TeleFile, it also apparently contributed to a
decrease in the performance of the Service Center Recognition/Image
Processing System (SCRIPS)--a document imaging and optical character
recognition system that IRS implemented in 1994 to process Forms 1040EZ
and certain other tax documents; (6) another major change during the
1997 filing season involved the procedures IRS used to process returns
that were filed with missing or incorrect SSNs; (7) in 1997, as
authorized by the Welfare Reform Act of 1996, IRS began treating missing
or incorrect SSNs as math errors rather than as issues that, in the
past, had to be resolved through a lengthy notice process; (8) as of
September 1, 1997, according to IRS, it had protected about $1.46
billion in revenue through the disallowance of claimed credits or
dependent exemptions in 1997, more than doubling the amount disallowed
using the procedures IRS followed in 1996; (9) one issue that GAO
discussed in a previous report that continues to be of concern is the
cost-effectiveness of IRS' use of lockboxes to process 1040 tax
payments; (10) additional information GAO obtained this year heightened
its concern by calling into question a key assumption IRS and the
Department of the Treasury's Financial Management Service (FMS) have
used to calculate the interest cost savings associated with this use of
lockboxes; (11) although FMS had planned a study to further assess
interest cost savings, those plans have been deferred, and there is no
assurance when such a study will be done.

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

 REPORTNUM:  GGD-98-33
     TITLE:  Tax Administration: IRS' 1997 Tax Filing Season
      DATE:  12/29/97
   SUBJECT:  Customer service
             Telephone
             Electronic forms
             Tax refunds
             Cash management
             Social security number
             Tax credit
             Information and referral services
             Tax returns
             Tax administration systems
IDENTIFIER:  IRS TeleFile Program
             IRS 1040PC Program
             IRS Service Center Recognition/Image Processing System
             World Wide Web
             
******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


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


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

December 1997

TAX ADMINISTRATION - IRS' 1997 TAX
FILING SEASON

GAO/GGD-98-33

IRS' 1997 Tax Filing Season

(268779)


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

  FMS - Financial Management Service
  FTD - Federal Tax Deposit
  IRS - Internal Revenue Service
  SCRIPS - Service Center Recognition/Image Processing System
  SSN - Social Security Number

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


B-276564

December 29, 1997

The Honorable Nancy L.  Johnson
Chairman, Subcommittee on Oversight
Committee on Ways and Means
House of Representatives

Dear Chairman Johnson: 

This report responds to your request that we assess the Internal
Revenue Service's (IRS) performance during the 1997 tax filing
season.  Besides providing data on various indicators that IRS uses
to measure its filing season performance, we discuss five areas that
have been problematic in past filing seasons:  (1) the ability of
taxpayers seeking answers to questions to reach IRS via the
telephone, hereafter referred to as telephone accessibility;\1 (2)
the number of returns filed by means other than the traditional paper
method; (3) IRS' efforts to deal with returns that have missing or
incorrect Social Security Numbers (SSN); (4) the use of banks, known
as lockboxes, to process certain tax payments; and (5) performance of
the imaging system IRS uses to process certain tax returns.  In March
1997, we testified before the Oversight Subcommittee on the interim
results of our work.\2


--------------------
\1 Accessibility, as we have traditionally defined it, is the total
number of calls answered divided by the number of call attempts,
which is the sum of the following:  (1) calls answered, (2) busy
signals, and (3) calls abandoned by the caller before an IRS assistor
got on the line. 

\2 Tax Administration:  IRS' Fiscal Year 1997 Spending, 1997 Filing
Season, and Fiscal Year 1998 Budget Request (GAO/T-GGD/AIMD-97-66,
Mar.  18, 1997). 


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

IRS met or exceeded most of its 1997 filing season related
performance goals.  Of particular note is the substantial improvement
in two important areas where we have criticized IRS' performance in
past filing seasons--telephone accessibility and the use of
alternative filing methods. 

According to IRS data, telephone accessibility increased from 20
percent during the 1996 filing season to 51 percent during the 1997
filing season.  That improvement was due to an increase in the number
of telephone calls IRS answered and a decrease in the number of calls
coming into IRS.  IRS was able to answer more telephone calls because
it devoted more staff to do so and revised its procedures for
answering questions on more complex tax issues.  IRS' ability to
answer more calls also contributed to the decrease in the number of
calls coming into IRS by reducing the number of busy signals and thus
the need for redials. 

The number of tax returns filed by means other than the traditional
paper method increased by 25 percent over last year, with the number
of returns filed by telephone (TeleFile) showing the most significant
increase--65 percent.  One factor that most likely contributed to the
increase in TeleFile was IRS' decision not to include a Form 1040EZ
in the tax package sent to taxpayers who appeared eligible to file
using TeleFile--the thinking being that if those taxpayers did not
receive a form that they could use to file on paper, they would be
more inclined to file using the telephone. 

Although the revised tax package apparently contributed to an
increase in the use of TeleFile, it also apparently contributed to a
decrease in the performance of the Service Center Recognition/Image
Processing System (SCRIPS)--a document imaging and optical character
recognition system that IRS implemented in 1994 to process Forms
1040EZ and certain other tax documents.  In that regard, IRS data for
the 1997 filing season show that the number of Forms 1040EZ processed
per hour on SCRIPS equipment declined from the number processed per
hour in 1996.  According to IRS, because taxpayers who were eligible
to use TeleFile did not get a Form 1040EZ that they could use to file
on paper, they also did not get a preprinted address label to affix
to a paper form.  Thus, if they chose to file on a Form 1040EZ, which
many of them did, they had to write in their names, addresses, and
SSNs.  The additional handwritten information on those returns
increased the amount of operator intervention needed to process those
returns through SCRIPS, which resulted in decreased productivity. 

Another major change during the 1997 filing season involved the
procedures IRS used to process returns that were filed with missing
or incorrect SSNs.  In 1997, as authorized by the Welfare Reform Act
of 1996, IRS began treating missing or incorrect SSNs as math errors
rather than as issues that, in the past, had to be resolved through a
lengthy notice process.\3 As of September 1, 1997, according to IRS,
it had protected about $1.46 billion in revenue through the
disallowance of claimed credits or dependent exemptions in 1997. 
That result more than doubled the amount disallowed using the
procedures IRS followed in 1996. 

One issue that we discussed in our report on the 1996 filing season\4
and that continues to concern us is the cost effectiveness of IRS'
use of lockboxes to process Form 1040 tax payments.  Additional
information we obtained this year heightened our concern by calling
into question a key assumption IRS and the Department of the
Treasury's Financial Management Service (FMS) have used to calculate
the interest cost savings associated with this use of lockboxes. 
Although FMS had planned a study to further assess interest cost
savings, those plans have been deferred, and there is no assurance
when such a study will be done. 


--------------------
\3 Section 6213(g)(2) of the Internal Revenue Code specifies those
conditions on a return that can be treated as math errors.  The
conditions include such things as computational mistakes and missing
or incorrect SSNs.  When IRS finds one of those conditions while
processing a return, it can immediately adjust the taxpayer's return
(by, for example, correcting the computation or disallowing the
dependent exemption, earned income credit, or child care credit
associated with the missing or incorrect SSN) and make appropriate
changes to the taxpayer's reported tax liability and refund, if any. 
IRS is to send the taxpayer a notice explaining the change. 

\4 IRS' 1996 Tax Filing Season:  Performance Goals Generally Met;
Efforts to Modernize Had Mixed Results (GAO/GGD-97-25, Dec.  18,
1996). 


   OBJECTIVE, SCOPE, AND
   METHODOLOGY
------------------------------------------------------------ Letter :2

Our objective was to assess IRS' performance during the 1997 filing
season, with particular emphasis on those areas that were identified
as problematic in our reviews of past filing seasons.  To achieve our
objective, we

  -- interviewed IRS National Office officials and IRS officials in
     the Atlanta, Cincinnati, Fresno, and Kansas City service centers
     responsible for the various activities we assessed;\5

  -- analyzed filing season related data from various management
     information systems, including IRS' Management Information
     System for Top Level Executives;

  -- analyzed IRS data relating to its telephone assistance and
     conducted a test of IRS' telephone accessibility during the last
     2 weeks of the filing season (see app.  I for information on our
     test methodology);

  -- analyzed IRS data on alternative filing methods, including IRS
     surveys of TeleFile users and nonusers;

  -- visited two lockbox banks, one in Atlanta and one in St.  Louis,
     to review remittance processing procedures; \6

  -- interviewed staff from FMS, which is responsible for negotiating
     and administering lockbox contracts, about the use of lockboxes
     to process Form 1040 tax payments and analyzed cost/benefit data
     related to lockbox processing;

  -- interviewed officials from IRS' Taxpayer Advocate's Office about
     the impact of various filing season activities on taxpayers;

  -- analyzed activity data for IRS' Internet World-Wide Web site and
     forms distribution centers; and

  -- reviewed relevant IRS internal audit reports. 

We did our work from January through October 1997 in accordance with
generally accepted government auditing standards.  We obtained
written comments on a draft of this report from the Deputy
Commissioner of Internal Revenue.  Those comments are discussed at
the end of this letter and are reprinted in appendix II. 


--------------------
\5 We selected these locations because we had staff available to do
the work in those cities. 

\6 We selected these locations because we had staff available to do
the work in those cities. 


   FILING SEASON GOALS WERE
   GENERALLY MET
------------------------------------------------------------ Letter :3

IRS uses various indicators to measure its filing season performance. 
Those indicators relate to workload, like the number of returns
processed; timeliness, like the number of days needed to process and
issue refunds; and quality, like the accuracy of IRS' answers to
taxpayers' questions and the accuracy with which IRS processes
individual income tax returns and refunds.  As shown in table 1,
those indicators show that IRS met or exceeded most of its
performance goals for the 1997 filing season.\7



                                     Table 1
                     
                        IRS' Performance Goals and Related
                      Accomplishments for the 1996 and 1997
                                  Filing Seasons

                                   1996\a                      1997\a
                          ------------------------  ----------------------------
                                       Accomplishm
Indicator                 Goal         ent          Goal         Accomplishment
------------------------  -----------  -----------  -----------  ---------------
Accuracy of individual    Process 93%  94% were     Process 95%  95% were
income tax returns        accurately   processed    accurately   processed
processed by Code and                  accurately                accurately
Edit staff\b

Accuracy of individual    Process      94.2% were   Process      94.7% were
income tax returns        93.0%        processed    95.0%        processed
processed by data         accurately   accurately   accurately   accurately
transcribers

Service center            Process      12,174       Process      12,692 returns
individual income tax     10,000       returns      11,730       were processed
returns processing        returns per  were         returns per  per staff year
productivity\c            staff year   processed    staff year
                                       per staff
                                       year

Individual income tax     11 days      Various      Less than    Various types
returns processing cycle               types of     16 days      of 1040s ranged
time\d                                 1040s                     between 5 and
                                       ranged                    14 days
                                       between 8
                                       and 11 days

Timeliness of processing  Payments     All          Payments     All payments
tax payments submitted    received 4/  payments     received 4/  received 4/15/
with individual income    15/96 thru   received 4/  15/97 thru   97 thru 4/30/
tax returns               5/1/96 were  15/96 thru   4/30/97      97 were
                          to be        5/1/96 were  were to be   deposited by
                          deposited    deposited    deposited    5/2/97
                          no later     by           no later
                          than 5/1/    5/1/96       than 4/30/
                          96                        97

Accuracy of individual    Process 98%  99.6% were   Process      99.4% were
income tax refunds on     accurately   processed    99.3%        processed
paper returns                          accurately   accurately   accurately

Timeliness of refund      Issue        Issued       Issue        Issued within
check for individual      within an    within an    within an    an average of
income tax returns filed  average of   average of   average of   38 days as of
on paper\e                40 days      38 days as   40 days      5/97
                                       of 5/96

Timeliness of refund for  Issue        Issued       Issue        Issued within
individual income tax     within an    within an    within an    an average of
returns filed             average of   average of   average of   15 days as of
electronically\f          21 days      16 days as   21 days      5/97
                                       of 5/96

Level of access to        Provide      Provided     Provide      Provided 71%
taxpayer service          41.4% level  51% level    60.2% level  level of access
telephone system\g        of access    of access    of access

Accuracy of tax law       Answer 90%   91% were     Answer 92%   95% were
assistance                accurately   answered     accurately   answered
                                       accurately                accurately

Level of access to forms  Provide 70%  Provided     Provide 70%  Provided 76.7%
ordering telephone        level of     54.3% level  level of     level of access
system\h                  access       of access    access

Accuracy of processing    Process      97.3% were   Process      97.0% were
form orders               96.5%        processed    96.5%        processed
                          accurately   accurately   accurately   accurately
--------------------------------------------------------------------------------
\a Data are as of April 1996 and April 1997, unless otherwise noted. 

\b Code and Edit staff are to prepare returns for computer entry by,
among other things, ensuring that all data are present and legible. 

\c Returns processing productivity is based on the number of weighted
returns processed, which includes all returns (whether processed
manually, through scanning equipment, or electronically).  The
different types of returns are weighted to account for their
differing processing impacts.  For example, a paper Form 1040 has a
higher weighting factor than a paper Form 1040EZ, which in turn has a
higher weighting factor than electronically processed returns. 

\d Cycle time is the average number of days it takes service centers
to process returns for the entire filing season--January 1 through
mid-April. 

\e This indicator is based on a sample of paper returns and is an
average calculated starting from the signature date on the return to
the date the taxpayer should have received the refund, allowing 3
days after issuance for the refund to reach the taxpayer. 

\f This indicator is based on a sample of electronically filed
returns and is an average calculated from the date the return is
received to the date the taxpayer should have received the refund,
allowing 2 or 3 days after issuance (depending on whether the refund
is by check or direct deposit) for the refund to the reach the
taxpayer or the taxpayer's bank account. 

\g IRS defines this indicator as the number of calls answered divided
by demand.  Demand is the number of individual callers.  As discussed
later, we have traditionally used a different indicator to measure
IRS' performance in providing telephone service. 

\h IRS defines this indicator as the number of calls answered divided
by demand.  Demand is the number of individual callers. 

Source:  IRS data. 


--------------------
\7 The goals shown in table 1 are set by IRS; we did not assess their
appropriateness. 


   ACCESSIBILITY TO IRS' TELEPHONE
   ASSISTANCE IMPROVED
   SUBSTANTIALLY
------------------------------------------------------------ Letter :4

During each filing season, millions of taxpayers call IRS with
questions about the tax law, their refunds, or their accounts.  The
number of callers who get through to an IRS assistor is an important
indicator of filing season performance.  According to IRS data, as
shown in table 2, telephone accessibility, as we have defined it in
the past, increased substantially during the 1997 filing season. 
Results of our independent test also pointed to an improvement in
accessibility.  Despite the improvement, however, accessibility is
still low. 



                                Table 2
                
                    Accessibility of IRS' Telephone
                              Assistance\a

                               Number of      Number of
                                    call          calls
                                attempts       answered        Percent
Filing season\b            (in millions)  (in millions)  accessibility
-------------------------  -------------  -------------  -------------
1997                                  62             32           51\c
1996                                 114             23             20
1995                                 236             19              8
----------------------------------------------------------------------
\a The percent accessibility presented in this table is the
calculation we have traditionally used to measure IRS' performance in
providing telephone assistance.  The calculation is the total number
of calls answered divided by the total number of call attempts (the
sum of calls answered, busy signals, and calls abandoned by the
caller before an IRS assistor got on the line).  Because the IRS
indicator in table 1--level of access to taxpayer service telephone
assistance--is based on the number of callers, it shows a higher
level of performance than does our indicator, which is based on the
number of call attempts. 

\b These data are for January 1 through April 19, 1997, January 1
through April 20, 1996, and January 1 through April 15, 1995. 

\c Numbers do not compute to this percent due to rounding. 

Source:  GAO analysis of IRS data. 

To check whether accessibility had increased, we conducted an
independent test to measure taxpayer access to IRS' telephone system
from March 31 through April 15, 1997.  Our results,\8 compared with
the results of a similar test we conducted in 1995, showed that
accessibility had improved.  For example, during the 1997 test, we
had to make 584 calls to gain access to a live assistor 211 times--a
36-percent accessibility rate.  That was a significant improvement
over 1995, when we had to make 1,655 calls to gain access 98 times--a
6-percent accessibility rate.  Also, of the 584 calls placed during
the 1997 test, 288 resulted in busy signals--a 49-percent busy rate. 
That compares favorably with a 92-percent busy rate during the 1995
test.  Our test methodology and detailed results are described in
appendix I. 

Despite the significant increase in 1997, telephone accessibility is
still too low.  The National Commission on Restructuring IRS made
that point in its June 25, 1997, report.  After noting how
accessibility had improved to 51 percent, the Commission noted that
"the level of access continues to be unacceptable and inferior to
service performance in private sector service organizations."

As table 2 showed, the increase in IRS' telephone accessibility
between the 1996 and 1997 filing seasons was due to a combination of
more calls being answered and fewer calls coming in (i.e., "call
attempts"). 


--------------------
\8 Our test was a nonstatistical sample.  The results relate just to
our test calls and cannot be projected. 


      INCREASE IN NUMBER OF CALLS
      ANSWERED
---------------------------------------------------------- Letter :4.1

Two primary reasons for the increase in the number of calls answered
were (1) a revision to IRS' procedures for handling calls involving
complex tax issues and (2) more staff assigned to answer the
telephone, some of whom were detailed from other IRS functions.  IRS'
decision to detail staff from other functions resulted in some
opportunity costs because these staff were not available to perform
their normal duties, such as auditing tax returns. 

In an effort to increase the number of calls answered, IRS conducted
a study to analyze the subject and length of taxpayer telephone
calls.  According to IRS, the study showed that several areas of
complicated tax law involved 20- to 30-minute telephone conversations
and that an assistor could answer about 5 simpler calls within the
same amount of time.  Thus, for the 1997 filing season, IRS revised
its procedures so that callers with questions in certain complex tax
areas were automatically connected to a voice messaging system.\9
Those callers were asked to leave their name, telephone number, and
the best time for IRS to call back, and they were told that someone
would be calling back within 2 working days. 

According to IRS, it received 619,310 calls to the voice messaging
system during the filing season and contacted 451,051 taxpayers in
response to those calls.  IRS said that there are several reasons why
it may not have responded to a message.  For example, the message may
have been garbled, thus preventing IRS from deciphering the caller's
telephone number; callers may have failed to include an area code; or
IRS attempts to contact the caller may have gone unanswered.\10

To help return calls to the messaging system, IRS detailed staff from
the Examination function, which is the IRS organization primarily
responsible for auditing tax returns.  IRS data show that staff who
were detailed from the Examination function spent about 125
full-time-equivalent staff years returning calls received by the
messaging system.  A cognizant official in the Examination function
estimated that the use of Examination staff to answer taxpayer
questions resulted in about $55 million in foregone revenue because
those staff were not available to audit returns.  We did not assess
the validity of that estimate. 

Another factor that contributed to the increase in the number of
calls answered was IRS' decision to assign more staff to answer the
telephone.  Nationwide, according to data provided by IRS' Customer
Service function, IRS dedicated 2,546 full-time-equivalent staff
years to answer taxpayers' telephone calls between January 1 and
April 30, 1997.  This was an increase of 605 staff years over the
1,941 staff years dedicated during the same period in 1996.  In
addition, some field offices, including the three service centers we
visited, temporarily detailed staff to help answer the telephone,
some of whom came from functions other than Customer Service. 
According to IRS, some of these staff were used only as needed, while
others were detailed for a few months. 


--------------------
\9 Some of the complicated tax areas involved the sale of a
residence, self-employment income and tax, rental property, and
depreciation. 

\10 According to IRS, staff were to make at least two attempts to
contact the caller. 


      DECREASE IN CALL ATTEMPTS
---------------------------------------------------------- Letter :4.2

The increase in the number of calls answered contributed to the
decrease in the number of call attempts.  As IRS improves its ability
to answer the telephone, taxpayers should encounter fewer busy
signals.  Fewer busy signals reduce the need for taxpayers to redial,
which reduces the number of call attempts.  In that regard, IRS'
telephone data showed that the number of busy signals dropped from
86.0 million during the 1996 filing season to 22.7 million during the
1997 filing season and that the average number of call attempts per
taxpayer dropped from 2.5 during the 1996 filing season to 1.4 during
the 1997 filing season. 

IRS cited two other contributors to the decrease in call
attempts--the elimination of certain notices and the availability of
information through other IRS sources, such as the Internet.  Before
the 1997 filing season began, IRS eliminated 23 notices that it
deemed unnecessary, which, in turn, reduced the need for persons to
call IRS with questions about these notices.  IRS estimated that its
action eliminated the issuance of about 7.5 million notices, but IRS
could not estimate how many calls might have been eliminated because
every notice does not necessarily generate a telephone call to IRS. 
IRS has a World-Wide Web site on the Internet that was first
available during the 1996 filing season.  The Web site provides,
among other things, some interactive applications that answer tax
questions, IRS regulations with "plain English" summaries, answers to
the most frequently asked tax questions, and tax forms.  IRS data
showed a significant growth in the use of IRS' Web site in 1997.  For
example, taxpayers accessed the Web site about 117 million times
between January 1 and April 20, 1997, compared with about 102 million
accesses throughout 1996, and taxpayers downloaded about 6.3 million
files during the 1997 filing season compared with about 2.4 million
files for the same period in 1996. 


      MEASURING TELEPHONE
      ACCESSIBILITY
---------------------------------------------------------- Letter :4.3

As noted earlier, the data in table 2 reflect our traditional way of
measuring telephone accessibility.  Over the last few years, IRS has
used another indicator, which it calls "level of access," to measure
its performance in providing telephone assistance to taxpayers.  IRS
defines level of access as the number of calls answered divided by
the number of callers (i.e., the number of taxpayers seeking
assistance).  Because IRS' indicator is based on the number of
callers, it shows a higher level of performance than does our
indicator, which is based on the number of call attempts. 
Nonetheless, IRS' indicator, like ours, showed a significant
improvement in performance during the 1997 filing season.  IRS
reported its level of access as 71 percent through April 19, 1997,
compared with 51 percent during a comparable period in 1996. 

We have been working with IRS to establish one mutually agreeable
measure of telephone accessibility.  As a result, we have reached
agreement on a measure to be used in future filing seasons.  That
measure defines accessibility as the number of calls that get into
IRS' automatic call distribution system, including those that are
answered and those that are abandoned by the caller before getting
assistance, divided by the total number of call attempts, which would
consist of calls answered, calls that are abandoned, and calls that
receive a busy signal.  As part of that measure, IRS agreed to show,
for the calls that got into the automatic call distribution system,
how many were answered\11 and how many were abandoned by the caller
before receiving assistance. 

Using IRS data as of April 19, 1997, the new measure shows that
taxpayers calling IRS were able to gain access 64 percent of the time
(39.8 million calls that got into IRS' automatic call distribution
system divided by 62.4 million call attempts).  Of the 39.8 million
calls that got into IRS' system, 31.8 million (80 percent) were
answered and 8 million (20 percent) were abandoned by the caller
before getting assistance. 


--------------------
\11 For purposes of this measure, "answered calls" would include
calls to the voice messaging system that were subsequently answered
by IRS. 


   NUMBER OF RETURNS FILED THROUGH
   ALTERNATIVE METHODS INCREASED
------------------------------------------------------------ Letter :5

As of October 31, 1997, IRS had received 120.9 million individual
income tax returns, an increase of 1.8 percent compared to the 118.8
million received at the same time last year.  Although the increase
in the overall number of returns filed was small, the increase in the
number filed through alternative methods was significant--about 25
percent higher than last year.  IRS offers three
alternatives--electronic filing, TeleFile, and 1040PC--to the filing
of traditional paper returns.\12 Among other benefits, returns filed
through these alternatives involve fewer errors and are presumed to
be less costly for IRS to process.  As shown in table 3, of the three
alternatives, TeleFile had the largest percentage change, by far, in
1997. 



                                Table 3
                
                Number of Individual Income Tax Returns
                  Received Through Alternative Filing
                                Methods

                        Number    Number    Number    Number
                      filed in  filed in  filed in  filed in   Percent
                        1994\a    1995\a    1996\a    1997\a    change
                           (in       (in       (in       (in   in 1997
                      thousand  thousand  thousand  thousand      from
Type of return              s)        s)        s)        s)      1996
--------------------  --------  --------  --------  --------  --------
TeleFile                   519       680     2,840     4,694       65%
Electronic              13,510    11,144    12,140    14,457        19
Form 1040PC              4,193     2,917     7,042     8,427        20
======================================================================
Total                   18,222    14,741    22,022    27,578       25%
----------------------------------------------------------------------
\a Data are as of November 4, 1994, November 3, 1995, November 1,
1996, and October 31, 1997. 

Source:  IRS' Management Information System for Top Level Executives. 


--------------------
\12 Under electronic filing, returns are transmitted over
communications lines through a third party (such as a tax return
preparer or electronic return transmitter) to an IRS service center,
where the data are automatically edited and processed.  Under
TeleFile, certain taxpayers who are eligible to file a Form 1040EZ
are allowed to file using a toll-free number on Touch-Tone
telephones.  Once the return is filed, it is processed like an
electronic return.  Under the Form 1040PC method, a taxpayer or tax
return preparer uses personal computer software that produces paper
tax returns in an answer-sheet format.  The Form 1040PC shows the tax
return line number and the data (dollar amount, name, etc.) on that
line.  Only lines on which the taxpayer has made an entry are
included on the Form 1040PC. 


      TELEFILE
---------------------------------------------------------- Letter :5.1

There were three changes to TeleFile in 1997 that most likely
contributed to the large increase in filings:  the eligibility
criteria were expanded to include certain married persons filing
joint returns, persons using TeleFile could request that any refund
be directly deposited to their bank account, and IRS changed the tax
package sent eligible TeleFile users in an attempt to encourage their
use of the system.  IRS data show that about 191,000 TeleFile returns
were filed jointly by married couples, thus accounting for about 10
percent of the growth in 1997.  The amount of growth due to the other
two changes could not be quantified. 

IRS' decision to change the TeleFile tax package was the subject of
some disagreement within IRS.  In past years, IRS sent taxpayers who
appeared eligible to use TeleFile a package that included not only
TeleFile materials but also a Form 1040EZ and related instructions. 
Thus, taxpayers who could not or did not want to use TeleFile had the
materials they needed to file on paper, assuming they were still
eligible to file a Form 1040EZ.  For the 1997 filing season, IRS
eliminated the Form 1040EZ and related instructions from the package
sent to taxpayers who were apparently eligible to use
TeleFile--hoping that more taxpayers would be inclined to use
TeleFile if they received only the TeleFile materials. 

Officials from the Taxpayer Advocate's Office said that they did not
agree with IRS' decision.  They said that they were originally led to
believe that IRS would be sending the revised TeleFile package only
to persons who had used TeleFile in 1996 and to a sample of other
taxpayers.  By the time they learned that IRS was going to send the
package to all apparently eligible TeleFile users, it was too late to
effect a change.  According to the officials, their concern centered
on the extra burden the revised package would impose on taxpayers who
wanted a Form 1040EZ, as well as the extra costs IRS might incur in
filling additional mail and telephone orders for Form 1040EZ from
those taxpayers.  Internal Audit expressed similar reservations in
communications with IRS management before the start of the filing
season.  Management responded by saying that (1) their intent was to
increase the use of TeleFile, which would actually reduce taxpayer
burden for those who used it and (2) they expected few of the
affected taxpayers to contact IRS' form distribution centers for
copies of Form 1040EZ. 

Officials from the Taxpayer Advocate's Office told us that they did
not receive many complaints from taxpayers and found no evidence that
the number of taxpayer orders for Form 1040EZ was significantly
higher than in past years.  Nonetheless, they said that they continue
to be concerned about this procedure, which IRS has indicated will
remain unchanged for the 1998 filing season. 

For the 1997 filing season, IRS sent about 26 million TeleFile tax
packages to taxpayers who, based on the tax returns they filed in
1996, would be eligible to use TeleFile in 1997.  After allowing for
the fact that some of those taxpayers might no longer be eligible to
use TeleFile because they no longer met the qualifying criteria, IRS
estimated that about 15.6 million of the taxpayers would be eligible
to use TeleFile in 1997.  As of October 31, 1997, about 4.7 million
taxpayers had filed their returns using TeleFile (about 30 percent of
the number IRS estimated to be eligible).  Assuming the validity of
IRS' estimate of eligible users, about 10.9 million of those
taxpayers chose not to use TeleFile in 1997. 

IRS conducted three TeleFile surveys in 1997--one electronic and one
written survey of users and one written survey of nonusers--that shed
some light on taxpayers' reactions to the revised tax package and the
reasons why more people did not use TeleFile. 

Results of the electronic user survey showed that 30.3 percent of the
TeleFile users in 1997 were repeat customers, while the rest were
using it for the first time.  The results also show that 84.5 percent
of the users were able to complete their filing with one call to IRS,
and 98.8 percent would use TeleFile again.  When questioned about the
new tax package, 85.2 percent of the respondents said that the
package "encouraged" them to use TeleFile, 2.9 percent said that it
"frustrated" them, and 2.5 percent said that it forced them to use
TeleFile.  Results of the written user survey showed that 88 percent
of the users were very satisfied with TeleFile and another 10 percent
were somewhat satisfied.  Also, 97 percent of the users said they
would use TeleFile again if they could, and about 96 percent said
that they were very satisfied or somewhat satisfied with the new
TeleFile tax booklet. 

Results from the nonuser survey are critical, in our opinion, if IRS
is to identify and effectively deal with barriers that are preventing
eligible taxpayers from using TeleFile.  In past surveys, IRS learned
that most nonusers preferred filing on paper.  But IRS did not
solicit more specific information on the reasons for that preference. 
In our report on the 1996 filing season, we recommended that IRS
conduct a survey of nonusers during the 1997 filing season that
included some specific questions on why they prefer to file on
paper.\13 The questionnaire IRS used for the nonuser survey in 1997
solicited more specific data on why taxpayers did not use TeleFile,
which, we believe, make the results more useful than earlier surveys. 

From a list of several potential reasons provided on the
questionnaire, respondents were asked to identify the main reason
they did not use TeleFile.  The main reasons they cited were: 

  -- filed a Form 1040 or 1040A, which made them ineligible to use
     TeleFile (25 percent);

  -- did not receive the TeleFile tax package (17 percent);\14

  -- used a tax preparer or accountant (15 percent);

  -- got help from a friend or family member in filing their return
     (15 percent); and

  -- preferred a paper copy of the return for their records (12
     percent). 

In response to questions about the revised TeleFile package and its
impact, 12 percent of the nonusers said that the package caused a
great deal of inconvenience.  Remembering the Taxpayer Advocate's
concern that the absence of a Form 1040EZ in the revised package
might significantly increase the number of mail and telephone orders
for that form, IRS also asked nonusers who prepared their own returns
where or how they got the tax form.  Only 2 percent said that they
called or wrote IRS.  The vast majority (about 82 percent) said that
they got the form from a post office, library, or bank.  The
remaining (about 16 percent) mentioned other methods, such as
visiting an IRS walk-in site or downloading a form from IRS' Internet
World-Wide Web site. 

IRS plans few changes to TeleFile for the 1998 filing season.  For
example, the TeleFile package for 1998 will again not include a Form
1040EZ.  However, one change that might eventually make TeleFile more
attractive to taxpayers is a pilot program with Indiana and Kentucky
that will allow TeleFile users to submit their state returns at the
same time they file their federal return.  In that regard, responses
to the TeleFile nonuser survey showed that about 44 percent of the
nonusers might be encouraged to use TeleFile if they could also use
it to file their state tax returns. 


--------------------
\13 GAO/GGD-97-25. 

\14 As of the date we completed our audit work, it was unclear why
these persons would not have received the TeleFile package. 


      ELECTRONIC FILING
---------------------------------------------------------- Letter :5.2

Electronic filing began as a pilot test in 1986, and the number of
individual income tax returns filed electronically continued to grow
each year until a drop in 1995.  IRS attributed that drop to the
various steps it took to deal with refund fraud.  As shown in figure
1, electronic filing recovered somewhat in 1996 and continued to grow
in 1997, establishing a new high of about 14.5 million returns as of
October 31, 1997. 

   Figure 1:  Number of Individual
   Income Tax Returns Filed
   Electronically

   (See figure in printed
   edition.)

Source:  IRS data. 

One impediment to even more growth in electronic filing is the fact
that the method is not completely paperless.  Taxpayers must still
send IRS their W-2s and a signature document (Form 8453) after their
returns have been electronically transmitted.  IRS must then manually
input these data and match them to the electronic returns. 

In an attempt to eliminate the paper associated with electronic
returns, IRS began testing the use of digitized signatures at three
locations during the 1996 filing season.  IRS planned to expand the
test to seven locations in 1997.  The seven locations included three
private tax return preparation offices and four sites that were part
of IRS' Volunteer Income Tax Assistance Program.  Because of some
technical problems with the software, however, IRS delayed its
distribution to the seven test sites until April 1, 1997.  Because
one of the four volunteer sites prepared very few returns after April
1, it did not participate in the test. 

The test consisted of preparers offering eligible taxpayers the
option of signing with a stylus "pen" on an electronic signature pad
in place of signing a Form 8453.\15 The electronic signature would
then be attached to the taxpayer's electronic return and both would
be transmitted to IRS.  From April 1 through April 17, 1997, the test
generated 435 returns that were submitted with electronic signatures. 
IRS did not collect information on the number of taxpayers who were
offered the chance to participate in the test but declined. 
According to IRS, the six participating locations provided feedback
that was overwhelmingly positive, including the reduced cost or
burden from not having to store the Forms 8453 and not having to pay
someone to batch, mail, and track the forms.  IRS plans to conduct
the test again in 1998 at the same seven locations. 


--------------------
\15 Eligible test participants whose returns did not include any
attachments that had to be submitted on paper were also granted a
waiver from the requirement to submit W-2s.  According to an IRS
official, tax return preparers participating in the test were
instructed to review the paper W-2s and not prepare electronic
returns for taxpayers whose W-2s looked fraudulent. 


   NEW PROCEDURES ENABLE IRS TO
   PROTECT MORE REVENUE
------------------------------------------------------------ Letter :6

An important change for the 1997 filing season involved IRS'
implementation of new procedures for handling returns filed with
missing or incorrect SSNs.  The amount of revenue protected as a
result of these new procedures greatly exceeded the amount protected
under the previous procedures. 

Correct SSNs help ensure that taxpayers are entitled to the credits
and dependency exemptions they claim.  While missing or incorrect
SSNs are often the result of honest taxpayer errors, they have also
been linked to fraudulent attempts to reduce tax liabilities and
obtain refunds and/or Earned Income Credits.  Accordingly, over the
last few years, IRS has become more vigilant in checking SSNs. 

During the last few filing seasons before 1997, when IRS identified a
missing or incorrect SSN, it was to delay the taxpayer's refund and
correspond with the taxpayer to resolve the issue.  This procedure
often required multiple correspondence and months to resolve.  As we
reported in 1996, IRS did not have enough resources to pursue all of
the cases involving missing or incorrect SSNs and ended up releasing
many of the refunds associated with those cases.\16

IRS' SSN error procedures changed in 1997 as a result of a provision
in the Welfare Reform Act of 1996.  That provision authorized IRS to
treat missing or incorrect SSNs as math errors, similar to the way it
has historically handled computational mistakes.  Under the new
procedures, if IRS identifies a missing or incorrect SSN while
processing a return, it can immediately adjust the return.  For
example, if a taxpayer claims one dependent and the child care
credit, but lists an incorrect SSN for the dependent, IRS is to
increase the taxable income by the personal exemption amount claimed
for the dependent and not allow the child care credit.  IRS then is
to adjust the taxpayer's tax liability and reduce the taxpayer's
refund, if any.  The taxpayer is to receive a notice explaining the
change to his or her tax liability and/or refund.  The standard
notice IRS used in 1997 provided a special toll-free telephone number
that taxpayers could call if they wanted to discuss IRS' changes
and/or provide corrected information to support their claims. 
Taxpayers could also write to IRS to resolve the issue.  If taxpayers
do not respond to IRS' notice, there is to be no further
correspondence unless they fail to pay any additional tax that was
assessed as a result of IRS' change. 

In planning for this new procedure, IRS estimated that it would send
about 2.4 million notices to affected taxpayers in 1997 and that
those notices would generate about 1.68 million responses (telephone
calls or letters) from taxpayers.  As of September 1, 1997, IRS had
sent about 2.2 million notices, which generated about 876,000 calls
and letters.  IRS said that based on those responses, it subsequently
allowed some of the claims it had originally disallowed.\17 As of
September 1, after netting out adjustments made in response to
taxpayers' calls and letters, IRS reported that it had protected
about $1.46 billion in revenue (i.e., claimed refunds or credits not
paid and additional taxes assessed).  That is about 150 percent more
than the amount of revenue IRS reported as having been protected as a
result of the procedures used in 1996.  That year, according to IRS,
it sent out about 629,000 notices that resulted in the protection of
about $590 million. 

We asked officials in the Taxpayer Advocate's Office whether the new
SSN error procedures posed any problems for IRS and/or taxpayers. 
They said that they did have concerns about the procedures, which
they voiced to IRS management before the start of the filing season. 
They were concerned, for example, that (1) the procedures may lead to
an unmanageable workload for IRS and (2) the notices were not clear. 
According to the officials, as a result of their input, some changes
were made before the filing season began.  They also told us that
they had not received a significant number of complaints from
taxpayers nor had there been an increase in the number of problem
resolution program cases or hardship requests for refunds. 

Even though significant problems did not arise, officials of the
Taxpayer Advocate's Office believed that some additional changes are
needed.  For example, they said that some individuals have problems
obtaining an SSN or some other taxpayer identification number either
because of religious affiliations or questionable alien status. 
Officials also think the notices should be revised to provide the
taxpayer with specific information about the error. 


--------------------
\16 Earned Income Credit:  IRS' 1995 Controls Stopped Some
Noncompliance, But Not Without Problems (GAO/GGD-96-172, Sept.  18,
1996) and IRS' 1996 Tax Filing Season:  Performance Goals Generally
Met; Efforts to Modernize Had Mixed Results (GAO/GGD-97-25, Dec.  18,
1996). 

\17 If a taxpayer provided IRS with a missing SSN or corrected an
inaccurate SSN, IRS would adjust its change.  However, not every
taxpayer response resulted in an adjustment.  In some cases, the
taxpayer may only have wanted IRS to explain the notice or may have
decided, after discussing the law and the facts with IRS, that IRS'
position was correct. 


   LOCKBOX PROCESSING FOR FORM
   1040 REMITTANCES MAY NOT BE
   COST EFFECTIVE
------------------------------------------------------------ Letter :7

In an effort to improve remittance processing and deposit tax
receipts more timely, IRS has been using lockboxes to process tax
payments, including the payments associated with individual income
tax returns (Forms 1040).\18 IRS and FMS assume that the use of
lockboxes is beneficial to the government because, in general, banks
can get the payment processed and the money deposited to a Treasury
account quicker than service centers can.  This means that Treasury
would not have to borrow as much to pay government obligations,
thereby avoiding interest charges. 

In our report on the 1996 tax filing season, we expressed our concern
about the way IRS was using lockboxes for Form 1040 payments.\19 Our
concern then was not with the processing of the payments but with
IRS' decision to have taxpayers send their tax returns along with
their tax payments to the lockboxes and to have the banks sort those
returns before shipping them to IRS service centers for processing. 
Information we received from FMS, which has been paying the lockbox
fees, and IRS indicated that having banks sort and ship tax returns
increased the cost of the lockbox service by about $4.7 million
during the first 8 months of 1996.  For example, FMS said that it
paid the banks an average of 92 cents per return to sort the 7
million returns received during those 8 months--a function that,
according to IRS, service centers performed at an average cost of 37
cents per return.  FMS also paid the banks 13 cents per return to
ship the tax returns to IRS for processing.\20

Our concern about the Form 1040 lockbox program has intensified since
last year.  We are no longer concerned only about having lockboxes
receive and sort tax returns but about the use of lockboxes to
process the Form 1040 payments themselves.  Information we obtained
this year called into question a key assumption used to calculate the
interest cost avoidance figures that IRS and FMS have cited to
support the use of lockboxes to process those payments. 


--------------------
\18 Under the lockbox concept, taxpayers are to mail payments to a
lockbox, which is a postal rental box serviced by a commercial bank. 
The bank is to process the payments and transfer the funds to a
federal government account, record the payment and payer information
on a computer tape, and forward the tape to IRS for use in updating
taxpayers' accounts.  In addition to Form 1040 payments, tax
remittances processed at lockbox banks include estimated tax
payments; various business payments, such as Federal Unemployment tax
and Social Security tax; and certain vehicle use taxes. 

\19 GAO/GGD-97-25. 

\20 According to IRS, the banks handled about 10 million returns
through the first 11 months of fiscal year 1997.  According to FMS,
it paid the banks an average of 86 cents per return to sort the
returns and 13 cents per return to ship the returns to IRS (compared
to averages of 92 cents and 13 cents, respectively, in 1996). 


      IRS' DECISION TO HAVE
      TAXPAYERS SEND THEIR RETURNS
      TO LOCKBOXES IS BASED ON
      INCONCLUSIVE EVIDENCE
---------------------------------------------------------- Letter :7.1

For the last several years, IRS has been testing the use of lockboxes
to process Form 1040 remittances.  Those test results and various
studies done for IRS led to the decision to have certain taxpayers
send their returns and tax payments to lockboxes.  Under the current
procedure, many taxpayers receive a tax package with one envelope and
two differently colored mailing labels.  If their return involves a
payment, they are to use one label that directs their return and
payment to a lockbox.  If their return does not involve a payment,
they are to use the other label that directs their return to an IRS
service center. 

In explaining the decision to have persons who were making payments
send their returns along with their payments to a lockbox, IRS
officials responsible for the lockbox program said that they believed
an increase in taxpayer burden would result if taxpayers were
required to separate their payments from their returns and mail each
to a different address.  They cited the results of taxpayer surveys
done in 1993 and 1994, which, they said, showed that taxpayers
preferred to keep their payment and return together.  IRS interpreted
this preference as an indicator that asking taxpayers to separate
their return from their payment would impose a burden. 

We reviewed the taxpayer surveys and considered the results to be
inconclusive as they relate to burden--45.9 percent of the taxpayers
surveyed said that they felt uneasy about mailing their checks and
returns in separate envelopes while 41.2 percent said that they did
not feel uneasy (the other 12.9 percent did not know).  Even for
those respondents who said they felt uneasy, it was unclear whether
they considered the use of separate envelopes an unreasonable burden
when weighed against the extra cost to the government associated with
sending returns to lockboxes. 

We realized, in preparing our report on the 1996 filing season, that
it was too late to do anything to change IRS' lockbox plans for the
1997 filing season.  Thus, we recommended that IRS take action that
would be effective for filing seasons after 1997.  Specifically, we
recommended that if the government was unable to negotiate lockbox
fees that were more comparable to service center costs and in the
absence of more compelling data on taxpayer burden, IRS should either
discontinue having returns sorted by the banks or reconsider the
decision to have taxpayers send their tax returns to the lockboxes
along with their tax payments.  As noted earlier, the combined fee
paid banks for sorting and shipping tax returns dropped only slightly
from 1996.  And, as discussed below, IRS still does not have
conclusive data on taxpayer burden. 

In May 1997, an IRS/FMS task force that had been formed to identify a
solution to this issue for 1998 and beyond recommended that IRS have
taxpayers separate their returns from their payments--mailing the
former to an IRS service center and the latter to a lockbox.  While
recognizing the extra burden on taxpayers (e.g., the extra postage
associated with mailing two envelopes and possible confusion over
which envelope to use), the task force said that such a procedure
would minimize lockbox costs and would enable the banks to deposit
remittances faster because they would no longer have to handle tax
returns. 

Despite the task force's recommendation, IRS decided that lockboxes
would continue to receive and sort tax returns in 1998.  The IRS
official responsible for the lockbox program told us that IRS
continues to believe that an increase in taxpayer burden would result
if taxpayers were required to separate their payments from their
returns and mail each to a different address, a view shared by
representatives from the Taxpayer Advocate's Office.  To support its
position, IRS cited the results of several focus groups that became
available after the task force had completed its work. 

IRS held 8 focus groups in 4 cities involving a total of 29 taxpayers
who prepared their own federal income tax returns and 31 tax
practitioners.  According to IRS, "even though there was not a
dominant trend from the [focus groups], taxpayers noted the cost of
two stamps and the confusion of two envelopes as burden issues."
Although focus groups are useful in providing insight on a particular
issue, they are not statistically representative of the population
and should not, in and of themselves, provide the basis for
far-reaching conclusions.  Given that and after reviewing transcripts
of the focus groups and a July 7, 1997, summary report on the focus
group results, we believe that IRS still does not have conclusive
evidence that the additional taxpayer burden that may be caused by
requiring the use of two envelopes would outweigh the millions of
dollars in additional costs the government is incurring to have banks
sort and ship tax returns. 

For example, although the report noted that participants were
concerned about the extra postage associated with using two
envelopes, it went on to say that taxpayers participating in the
focus groups viewed the extra cost "as something that would be
accepted" and that "some taxpayers were willing to accept additional
burden so that IRS could operate more effectively." In that regard,
focus group participants were not told about the amount of additional
cost being incurred by the government to have banks sort and ship the
tax returns.  The report also said that "several participants voiced
concern about the check being separated from the return prior to
receipt by the IRS." However, there is no evidence that participants
were told how the two-envelope procedure compares to the current
procedure and that, even under the current procedure, the tax return
and check get separated.  Under the current procedure, even though
returns and payments are mailed in one envelope to one location (the
lockbox), they are separated at the bank.  The return is shipped to
IRS for processing while the bank processes the payment. 

In a July 15, 1997, memorandum to the then Acting Commissioner of
Internal Revenue, Treasury's Fiscal Assistant Secretary provided his
views on the processing of Form 1040 tax payments.  He noted that he
had sought IRS' support for having taxpayers mail their returns and
payments separately but that IRS had rejected that option because of
the perceived taxpayer burden.  That left only two viable options in
the Assistant Secretary's opinion--continue the current arrangement
or return the processing of Form 1040 tax payments to IRS' service
centers.  The Assistant Secretary noted, however, that it was his
understanding that equipment, personnel, and space issues and the
lack of sufficient planning time made it infeasible to move
processing back to the service centers for fiscal year 1998.  Thus,
he concluded that it would be in the best interest of the government
to continue the current lockbox arrangement for at least 1 more year. 
He said that this issue should be reviewed in March/April 1998 to
make decisions about fiscal year 1999 and that IRS should continue to
seek ways to reduce the cost of this program, by either changing
processing procedures or by continuing its search for a resolution on
how to direct the Form 1040 returns to the service centers without
causing significant taxpayer burden. 


      IRS STUDY CALLS INTO
      QUESTION THE VALIDITY OF
      INTEREST COST SAVINGS
---------------------------------------------------------- Letter :7.2

Our concerns about the Form 1040 lockbox program were heightened this
year by new information relating to the interest cost avoidance
figures that IRS and FMS have used to show the program's cost
effectiveness.  This new information calls into question not just the
decision to have tax returns sent to the banks but the more basic
decision to use lockboxes to process Form 1040 payments. 

As part of its review, the lockbox task force compared how much
various procedural options for processing Form 1040 remittances would
cost IRS and FMS in 1998.  Assuming a volume of 11,373,133 items, the
task force estimated that the current lockbox procedure would cost
about $23.3 million, compared with about $14.5 million if two
envelopes were used\21 and about $12.8 million if IRS decided to stop
using lockboxes to process Form 1040 remittances and return that
function to the service centers.  Although this comparison would seem
to argue against the use of lockboxes, IRS and FMS assume that having
lockboxes process Form 1040 remittances generates savings, in the
form of interest cost avoidance, that more than offset the increased
IRS and FMS costs. 

According to FMS, for example, lockboxes processed about 9.7 million
Form 1040 tax payments from October 1996 through July 1997, which
resulted in an interest cost avoidance of about $23.8 million.  As in
past years, the interest cost avoidance was calculated on the basis
of a general assumption that lockboxes can process and deposit tax
payments an average of 3 days faster than IRS service centers during
peak workload periods.  The validity of that assumption is critical
because, according to the lockbox task force, if lockboxes are not
processing payments at least 2 days faster than service centers, the
amount of interest cost avoidance would be insufficient to offset the
additional costs associated with having lockboxes handle tax returns. 

In that regard, the results of an IRS-commissioned study, issued in
March 1997, show that, on average, lockboxes processed payments only
about 1 day faster than service centers.  However, that comparison
covered peak and nonpeak workload periods; there was not a similar
comparison just for the peak workload period (April 13 to May 1,
1995), when most of the Form 1040 payments are received and when
differences in processing times might be more pronounced.  Although
the reported study results are insufficient to make an informed
judgment, they do raise questions about the assumption that lockboxes
process payments 3 days faster than service centers.  FMS had planned
to commission another study to assess the comparative processing
times for lockboxes and service centers.  However, those plans have
been deferred, and FMS could give us no assurance when such a study
would be done. 


--------------------
\21 If the analysis were broadened to reflect the extra cost that
taxpayers would have to incur to mail a second envelope, the cost of
this option would increase by $3.6 million (11,373,133 envelopes
times 32 cents postage) to a total of about $18.1 million. 


   SCRIPS PROCESSING RATES DECLINE
------------------------------------------------------------ Letter :8

One of IRS' major business objectives is to move away from a
labor-intensive tax return processing system that relies on thousands
of employees transcribing data from paper tax returns and move to an
electronic system that reduces processing costs and eliminates
transcription errors.  One strategy for achieving that objective is
to reduce the number of paper returns by increasing the number of
returns filed electronically.  We discussed IRS' progress in that
area earlier in this report.  For returns that will continue to be
filed on paper, IRS planned to achieve its objective through document
imaging and optical character recognition systems.  SCRIPS is one
such system. 

IRS uses SCRIPS, which was implemented in 1994, to process tax
returns filed on Form 1040EZ, Federal Tax Deposit (FTD) coupons, and
information returns (e.g., Forms 1099).  In January 1997, we reported
that one of the major problems with SCRIPS was slow processing rates
(i.e., the number of documents processed per hour).\22 As the data in
table 4 show, this problem intensified with respect to Forms 1040EZ
and FTD coupons in 1997. 



                                Table 4
                
                SCRIPS Processing Rates as of September
                           30, 1996, and 1997

                                  SCRIPS processing rate
                                   (documents per hour)
                              ------------------------------
                                      Jan. 1          Jan. 1
                                     through         through
                               September 30,   September 30,   Percent
Form type                               1996            1997    change
----------------------------  --------------  --------------  --------
Form 1040EZ                               67              57       -15
Information returns                      145             145         0
FTD coupons                              722             619       -14
----------------------------------------------------------------------
Source:  GAO computations based on IRS data. 

An IRS official in the SCRIPS project office attributed the decline
in the processing rate for Forms 1040EZ, at least in part, to IRS'
decision, as discussed earlier, to issue TeleFile packages that did
not have a Form 1040EZ.  Because the package contained no form that
the taxpayer could use to file on paper, it also contained no
preprinted label for the taxpayer to affix to a paper form. 
Successful optical character recognition operations depend, in part,
on the kind of clearly printed data provided by a label.  In that
regard, IRS estimated that about 95 percent of the Forms 1040EZ
processed through SCRIPS in 1997 did not have the scannable
preprinted address label, compared with about 50 percent in 1996,
causing a significant increase in the amount of data IRS had to
manually transcribe. 

IRS attributed the decrease in the processing rate for FTD coupons to
a problem associated with the way in which blocks of data are
transferred throughout the system.  In those instances where
characters on the document cannot be identified correctly by the
scanner, the electronic block of work that contains those documents
is sent to a workstation operator.  That operator retrieves the block
of work, reviews the image of the document to determine what
corrections are needed, then updates the file, which is sent back to
the file server.  Each time a block of work is moved from one
component of SCRIPS to another, a time delay results.  According to
IRS, the contractor provided a software solution to this problem and,
since then, processing times have improved, except on those days when
the service centers have to process the largest volumes of FTD
coupons. 


--------------------
\22 Tax Systems Modernization:  Imaging System's Performance
Improving but Still Falls Short of Expectations (GAO/GGD-97-29, Jan. 
16, 1997). 


   CONCLUSIONS
------------------------------------------------------------ Letter :9

IRS made noteworthy progress in several critical areas during the
1997 filing season.  It achieved significant increases in telephone
accessibility and alternative filings, and it implemented a major
change in dealing with missing or incorrect SSNs, all without any
noticeable major problems.  A couple of these successes involved
trade-offs.  By detailing staff to answer telephone calls, IRS
improved accessibility but, according to IRS, the detailing of that
staff caused it to forgo some enforcement revenue.  By not including
a Form 1040EZ in the tax package sent potential TeleFile users, IRS
apparently encouraged some taxpayers to file their returns by
telephone.  But, in doing so, IRS imposed some burden on recipients
of the TeleFile package who needed or wanted a Form 1040EZ and caused
a reduction in SCRIPS processing rates.  Such trade-offs are
inevitable, given the fact that IRS does not have unlimited
resources, and we saw nothing to indicate that either trade-off was
inappropriate. 

We are concerned, however, about the cost effectiveness of IRS' use
of lockboxes rather than service centers to receive and process Form
1040 tax payments.  On the basis of the data currently available, we
do not believe IRS is in a position to make an informed decision on
whether to continue to use lockboxes for that purpose. 

The results of an IRS-commissioned study suggest that the interest
cost avoidance figures IRS and FMS have cited to support the use of
lockboxes to process Form 1040 payments may not be valid.  Although
FMS had planned another study to further assess the comparative
processing times and costs for lockboxes and service centers, those
plans have been deferred, and it is unclear when such a study will be
done.  As a result, it is unclear whether the government actually
realizes savings, and if so how much, through the use of lockboxes. 

Because IRS has decided to continue the use of lockboxes for the 1998
filing season and it is too late to alter that decision, it seems
that IRS and FMS have an opportunity during that filing season to
develop the definitive data needed to make more informed decisions on
the future use of lockboxes.  Such data would include the average
amount of time needed by both banks and IRS to process Form 1040 tax
payments during both peak and off-peak periods and the average
interest costs to the government of borrowing during those periods. 
It seems that such data would be readily available and there is still
time to make any arrangements that might be needed to capture these
data during the 1998 filing season. 

We are equally concerned that IRS' decision to continue having
taxpayers send both tax returns and payments to lockboxes is also
based on inconclusive evidence.  Another option would be to have
taxpayers mail their payments to a lockbox and their tax returns to
IRS.  An IRS/FMS task force studied these two options and recommended
the latter because it would save the government millions of dollars
in payments to banks for processing the tax returns.  Although
Treasury's Fiscal Assistant Secretary also supported this
recommendation, IRS decided against this change because of the added
taxpayer burden, including mailing cost. 

However, our review indicated that IRS did not have persuasive
evidence on the amount of taxpayer burden, how this burden would
compare with the government's savings, or whether taxpayers
considered the burden to be unreasonable.  The new evidence available
to IRS in making its decision for the 1998 filing season was from
taxpayer focus groups.  The number of taxpayers participating in the
focus groups was small, and the information they provided on burden
was inconclusive.  In that regard, the participants were not provided
all of the information needed to make an informed response about the
burden associated with mailing tax returns in one envelope and tax
payments in another.  Specifically, they were not told that mailing
both tax returns and tax payments to lockboxes is costing the
government, and thus taxpayers, millions of extra dollars.  Had the
focus group participants been informed about all the relevant
factors--the additional postage and burden involved in separating
returns from payments and mailing them in two envelopes versus the
savings that would accrue to taxpayers overall in the form of savings
to the government if this were done--we believe that they would have
been in a better position to assess the trade-offs involved in
deciding on the reasonableness of the burden involved. 

IRS has three basic options concerning the use of lockboxes for Form
1040 tax payments:  (1) continue the existing practice, (2)
discontinue the use of lockboxes altogether, or (3) revise the
existing practice to have taxpayers send their returns to service
centers and their payments to lockboxes.  In deciding which option to
select, IRS faces the following two basic issues.  First, in deciding
between options 1 and 2, IRS needs to know whether using lockboxes to
process Form 1040 tax payments generates a net savings to the
government.  Second, if the use of lockboxes generates significant
savings, which would remove option 2 from the equation, IRS needs to
know, in choosing between options 1 and 3, whether the additional
savings to the government of having taxpayers send their tax returns
to IRS service centers outweigh the taxpayer burden associated with
taxpayers sending two different envelopes to two locations.  Although
these issues would involve different analyses, the results and
related decisions are intertwined.  That is, if the analysis for the
first issue shows that using lockboxes for Form 1040 tax payments
does not result in a significant net savings to the government and
IRS decides to stop using lockboxes, the second issue would become
moot. 


   RECOMMENDATIONS TO THE
   COMMISSIONER OF INTERNAL
   REVENUE
----------------------------------------------------------- Letter :10

We recommend that the Commissioner of Internal Revenue require the
appropriate IRS officials to conduct, during the 1998 tax filing
season, the analyses necessary to determine (1) whether there are net
savings to the government attributable to the use of lockboxes to
process Form 1040 tax payments and (2) whether the potential savings
of requiring affected taxpayers to mail their tax returns to IRS and
their tax payments to lockboxes in separate envelopes outweigh the
estimated additional cost and other burden that this could be
expected to cause taxpayers.  In doing these analyses, the officials
should collect definitive data on (1) the actual time and interest
cost differences between sending tax payments to lockboxes and
sending them to IRS during peak and off-peak periods and (2) whether
taxpayers believe, given the processing cost savings to the
government, that it would cause them an unreasonable burden to mail
tax returns and tax payments to different locations.  If the analyses
indicate that using lockboxes does not produce a net savings to the
government, we recommend that the Commissioner take steps to have IRS
service centers process all Form 1040 payments starting with the 1999
tax filing season.  If the analyses indicate that the use of
lockboxes produces savings and that taxpayers would support the
practice of mailing returns and payments to different locations, we
recommend that the Commissioner change the current lockbox procedures
as soon as possible and instruct taxpayers to send their returns to
IRS and their payments to lockboxes. 


   AGENCY COMMENTS AND OUR
   EVALUATION
----------------------------------------------------------- Letter :11

We obtained written comments on a draft of this report from the
Deputy Commissioner of Internal Revenue (see app.  II).  He said that
IRS generally agreed with our findings and recommendations and that
the ultimate determination of the benefit of using lockboxes to the
taxpaying public requires information and analysis by both IRS and
FMS.  He said that IRS would pursue a study with FMS during fiscal
year 1998 to reach a long-term decision about lockbox processing and
that IRS would welcome the opportunity to assist FMS in analyzing the
savings to the government attributable to the use of lockboxes to
process Form 1040 tax payments. 

The Deputy Commissioner also said that (1) it would be impossible to
determine conclusively whether any savings from having taxpayers mail
their returns and payments in separate envelopes outweighed the
additional cost and other burden to taxpayers because "the perceived
burden of this new way of paying and filing may involve intangibles
which cannot be measured (e.g., a negative reaction to changes in
procedures)" and (2) a broad-based study to determine taxpayers'
perception of burden and their willingness to accept that burden
would be necessary to satisfy our recommendation.  In addition, the
Acting Chief of Customer Service, at a meeting to discuss IRS'
comments, told us that IRS would always come down on the side of
reduced burden. 

We agree that burden cannot be measured conclusively.  However, it is
not necessary to conclusively measure burden before deciding whether
to use two envelopes.  In fact, IRS has made other decisions
involving new ways of paying and filing without measurable data on
burden.  For example, the decision to not include a Form 1040EZ in
the TeleFile tax package involved a change in procedures that caused
additional taxpayer burden by requiring taxpayers who could not or
did not want to use TeleFile to find a Form 1040EZ elsewhere.  IRS
made that decision although it could not measure the impact of the
additional burden on taxpayers.  In the TeleFile case, IRS did not
come down on the side of reduced burden but decided, instead, that
the additional burden, though unmeasured, was acceptable given the
expected benefits.  In fact, IRS plans to send out the same kind of
abbreviated TeleFile tax package for the 1998 filing season even
though 12 percent of the TeleFile nonusers indicated that the package
IRS sent out for the 1997 filing season caused a great deal of
inconvenience.  We believe that asking taxpayers to use two envelopes
might be another instance where IRS might accept a small amount of
burden, if the potential savings to the government are significant. 

It is unclear what, if anything, IRS intends to do to get better
information on burden.  However, if it decides to obtain more
definitive data, it is important that it structure its data
collection to get specific input from taxpayers on their willingness
to assume the additional burden of dealing with two envelopes in
light of the savings to the government.  Among other things, that
would mean asking them to compare the two-envelope approach with the
current approach and making sure, for both approaches, that they know
the government's costs and benefits. 

With respect to our second recommendation, the Deputy Commissioner
indicated that necessary studies could not be done and analyzed in
time to make changes for the 1999 filing season, given the lead time
needed in awarding contracts for the printing of tax packages.  We do
not understand why this would be the case.  It seems to us that at
least the first stage of the analysis sought by our
recommendation--whether the government saves money by using
lockboxes--can be made in time.  Our prior analyses of IRS data
indicate that the key data needed to make that determination--the
comparative time it takes lockboxes and service centers to process
and deposit Form 1040 payments--should be readily available and can
be compiled and analyzed relatively quickly.  There is no reason we
are aware of that such analysis could not be done using data from the
1997 filing season, rather than waiting for new data from the 1998
filing season.  Thus, if the analysis shows that the government is
not benefiting from the use of lockboxes, IRS should have time to
make the necessary changes to the tax packages for 1999. 

The second part of the analysis, which considers burden, would only
be needed if the first part shows that using lockboxes to process
Form 1040 payments saves the government money.  If the analysis shows
that it does not save the government money, IRS could change the
procedure for handling Form 1040 tax payments without further
analyzing burden.  The revised procedure would still involve one
envelope, but the envelope would be mailed to a service center rather
than a lockbox.  Since that change would result in all returns
(remittance and nonremittance) being mailed to IRS, it would not only
avoid the additional burden of having taxpayers deal with two
envelopes but also reduce existing burden by negating the need for
taxpayers to deal with two mailing labels. 

We revised our second recommendation to recognize the different
decisions that will confront IRS depending on the results of the
analyses called for in our first recommendation.  As part of that
revision, the reference to the 1999 filing season now applies only if
the first part of the analysis shows that the government is not
saving money by using lockboxes to process Form 1040 payments. 

In commenting on this report, the Deputy Commissioner identified a
number of actions IRS took during the 1997 filing season and asked
that we include the information in our report (see app.  II).  Some
of the actions identified by the Deputy Commissioner, such as those
relating to electronic tax administration, level of access, and
complex tax law questions, are discussed in our report.  However, we
did no audit work on several other actions mentioned by the Deputy
Commissioner and thus cannot comment on their effectiveness. 


--------------------------------------------------------- Letter :11.1

We are sending copies of this report to the Subcommittee's Ranking
Minority Member, the Chairmen and Ranking Minority Members of the
House Committee on Ways and Means and the Senate Committee on
Finance, various other congressional committees, the Secretary of the
Treasury, the Commissioner of Internal Revenue, the Director of the
Office of Management and Budget, and other interested parties. 

Major contributors to this report are listed in appendix III.  Please
contact me on (202) 512-9110 if you have any questions. 

Sincerely yours,

Lynda D.  Willis
Director, Tax Policy and
 Administration Issues


TOLL-FREE TELEPHONE ACCESSIBILITY
TEST
=========================================================== Appendix I

To assess the ability of taxpayers to reach IRS by telephone to ask a
question about the tax law or their accounts, we conducted a
nonstatistical test of IRS' toll-free telephone assistance system. 
Our results relate just to the test calls; they cannot be projected. 
To conduct the test, we placed telephone calls at various times
during each workday from March 31 through April 15, 1997.  We made
our calls from five metropolitan areas--Atlanta, Chicago, Kansas
City, San Francisco, and Washington, D.C.  Each attempt to contact
IRS consisted of up to five calls spaced 1 minute apart.  If we
reached IRS during any of the five calls and made contact with an
assistor, we considered the attempt successful.  If we reached IRS
during any of the five calls but were put on hold for more than 7
minutes without talking to an assistor, we abandoned the call, did
not dial again, and considered the attempt unsuccessful (abandoned). 
If we received a busy signal, we hung up, waited 1 minute, and then
redialed.  If after four redials (five calls in total) we had not
reached IRS, we considered the attempt unsuccessful.  In conducting
our test, we did not ask questions of the assistors because it was
not our intent to assess the accuracy of their assistance. 

We attempted to contact IRS 330 times.  Of 330 attempts to contact an
assistor, 211 (64 percent) were successful--162 on the first call, 22
on the second call, and 27 after 3 to 5 calls.  When the 16 calls
that resulted in access to IRS' voice messaging system were added,
accessibility to IRS assistance increased to 69 percent.  In another
69 cases (21 percent), we accessed IRS' system but were put on hold
more than 7 minutes and thus hung up before making contact with an
assistor.  The remaining 34 attempts (10 percent) were aborted after
we received busy signals on each of our 5 dialing attempts.  Our 330
attempts to contact an assistor required a total of 584 calls to IRS'
toll-free telephone number.  Of those 584 calls, we succeeded in
contacting an IRS assistor 211 times--a 36-percent accessibility
rate. 

We followed the above methodology to conduct our 1995 test, but we
placed telephone calls from two additional metropolitan areas
(Cincinnati and New York) and for two separate 2-week periods
(January 30 through February 11, 1995, and April 3 through April 15,
1995).  Results of the 1995 test cited in the body of this report are
only for the 2-week period from April 3 through April 15, 1995. 




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

in the report text appear at the end of this appendix. 



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)


The following are GAO's comments on IRS' letter dated November 26,
1997. 

GAO COMMENTS

1.  IRS says that lockbox payments of $213 billion were deposited
during fiscal year 1997.  That figure covers all tax payments
processed by the lockbox banks.  The lockbox discussion in our report
focuses only on the processing of Form 1040 tax payments. 

2.  IRS says that the SCRIPS sites succeeded in processing over 90
percent of the Forms 1040EZ through SCRIPS.  However, only 5 of IRS'
10 service centers have SCRIPS.  The 90-percent figure cited by IRS
means that SCRIPS was used to process 90 percent of the Forms 1040EZ
filed at those 5 centers.  The other five centers used the
traditional keypunching system to process the Forms 1040EZ they
received.  Also, as noted in our report, while the 5 SCRIPS centers
may have processed 90 percent of the Forms 1040EZ they received, they
did so at a slower rate than in 1996. 

3.  IRS says that the number of telephone calls answered increased
from 99.2 million in fiscal year 1996 to 103.9 million in fiscal year
1997.  These numbers differ from the numbers cited in table 2 of our
report because (1) our numbers are for the filing season (January 1
through mid-April) while IRS' numbers are for the fiscal year
(October 1 through September 30) and (2) our numbers include just
those calls answered by IRS assistors, while IRS' numbers include
calls answered by assistors and by automated systems, such as TeleTax
(a system that has prerecorded information on about 150 topics). 

4.  IRS cites an initial contact resolution rate of above 95 percent. 
However, as IRS says, that rate only covers walk-in contacts and
correspondence.  It does not reflect the extent to which telephone
inquiries are resolved with one contact. 


MAJOR CONTRIBUTORS TO THIS REPORT
========================================================= Appendix III

GENERAL GOVERNMENT DIVISION,
WASHINGTON, D.C. 

David J.  Attianese, Assistant Director, Tax Policy and
Administration Issues
Robert L.  Giusti, Senior Evaluator
Monika R.  Gomez, Evaluator
Christopher E.  Hess, Senior Evaluator
John Lesser, Senior Evaluator

KANSAS CITY OFFICE

Royce L.  Baker, Issue Area Manager
Doris J.  Hynes, Evaluator-in-Charge
Marvin G.  McGill, Evaluator

ATLANTA OFFICE

Jyoti Gupta, Evaluator
Kim Rogers, Evaluator

SAN FRANCISCO OFFICE

Sharon K.  Caporale, Evaluator
Suzy Foster, Senior Evaluator


*** End of document. ***