Tax Administration: IRS Is Working to Improve Its Estimates of Compliance
Burden (Letter Report, 05/22/2000, GAO/GGD-00-11).

Pursuant to a congressional request, GAO provided information on the
status of the Internal Revenue Service's (IRS) efforts to improve its
estimates of taxpayer compliance burden, focusing on: (1) IRS' overall
strategy to improve its methodology for estimating compliance burden;
(2) what IRS has done to begin implementing its overall strategy and how
IRS expects this to improve its methodology; and (3) whether IRS expects
that its new methodology will be able to measure the burden associated
with the complex tax rules identified in IRS' forthcoming first annual
report on tax complexity.

GAO noted that: (1) to improve its methodology for estimating compliance
burden for all types of federal taxpayers, IRS is pursuing a multiphased
strategy; (2) initially, IRS is focusing on taxpayers who have only wage
and investment (W&I) income because they bear a large portion of the
overall compliance burden and because their burden may be easiest to
estimate; (3) later phases would develop estimates for other taxpayer
groups, such as small businesses and the self-employed and large- and
medium-sized businesses; (4) this incremental strategy should give IRS
and its contractors the flexibility to incorporate lessons learned from
early phases of the process into later efforts; (5) IRS has begun to
implement its strategy by contracting for the development of two models
that, when combined, should provide more reliable estimates of W&I
taxpayers' prefiling, filing, and postfiling compliance burdens; (6)
compared with IRS' current methodology, these models are designed to
produce more comprehensive estimates of federal income tax compliance
burdens and to provide IRS with a greater capacity to analyze the impact
of tax law and administrative changes on those burdens; (7) as with all
such modeling, the specific capabilities and precision of the new models
will depend on the quality of the underlying data; (8) IRS expects that
one of the new models, covering prefiling and filing activities, will
provide some assistance in estimating burdens associated with the
complex rules identified in IRS' forthcoming first annual report in tax
complexity; (9) for some rules, the model may be able to show both the
number of W&I taxpayers affected and the approximate size of their
burdens; (10) for other rules, IRS expects that the model will only be
able to show the potential number of taxpayers affected; and (11) the
other model, which is to estimate postfiling burden, is in the early
design stage, and its capabilities have not yet been determined.

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

 REPORTNUM:  GGD-00-11
     TITLE:  Tax Administration: IRS Is Working to Improve Its
	     Estimates of Compliance Burden
      DATE:  05/22/2000
   SUBJECT:  Tax administration systems
	     Voluntary compliance
	     Reporting requirements
	     Taxpayers
	     Tax law
	     Projections
	     Tax returns
	     Systems design

******************************************************************
** 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.                 **
**                                                              **
** 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.                         **
******************************************************************

United States General Accounting Office
GAO

Report to the Chairman

Committee on Small Business

U.S. Senate

May 2000

GAO/GGD-00-11

TAX ADMINISTRATION
IRS Is Working to Improve Its Estimates of

Compliance Burden

Ordering Copies of GAO Reports
The first copy of each GAO report and testimony
is free. Additional copies are $2 each. Orders
should be sent to the following address,
accompanied by a check or money order made out
to the Superintendent of Documents, when
necessary. VISA and MasterCard credit cards are
accepted, also. Orders for 100 or more copies to
be mailed to a single address are discounted 25
percent.
Order by mail:
U.S. General Accounting Office
P.O. Box 37050
Washington, DC 20013
or visit:
Room 1100
700 4th St. NW (corner of 4th and G Sts. NW)
U.S. General Accounting Office
Washington, DC
Orders may also be placed by calling (202) 512-
6000 or by using fax number (202) 512-6061, or
TDD (202) 512-2537.
Each day, GAO issues a list of newly available
reports and testimony. To receive facsimile
copies of the daily list or any list from the
past 30 days, please call (202) 512-6000 using a
touch-tone phone. A recorded menu will provide
information on how to obtain these lists.
Viewing GAO Reports on the Internet
For information on how to access GAO reports on
the INTERNET, send e-mail message with "info" in
the body to:
[email protected]
or visit GAO's World Wide Web Home Page at:
http://www.gao.gov

Reporting Fraud, Waste, and Abuse in Federal
Programs
To contact GAO FraudNET use:
Web site:
http://www.gao.gov/fraudnet/fraudnet.htm
E-Mail: [email protected]
Telephone: 1-800-424-5454 (automated answering
system)

United States
General Accounting Office
Washington, D.C. 20548-0001
Official Business
Penalty for Private Use $300
Address Correction Requested

                    Bulk Rate
               Postage & Fees Paid
                       GAO
                 Permit No. G100

(268893)

Contents
Page 161GAO/GGD-00-11 Taxpayer's Compliance Burden
Letter                                                                      1
Appendix I                                                                 18
Limitations of the
Arthur D. Little Model
Appendix II                                                                20
Details on the
Definition of Burden to
Be Used in the PwC
Survey
Appendix III                                                               24
Comments From the
Internal Revenue Service

Abbreviations

ADL       Arthur D. Little, Inc.
AMT       Alternative Minimum Tax
GPRA      Government Performance and Results Act
of 1993
IRS       Internal Revenue Service
OMB       Office of Management and Budget
PwC       PricewaterhouseCoopers
W&I       Wage and investment income

B-283788

Page 15 GAO/GGD-00-11 Taxpayer's Compliance Burden
B-283788

May 22, 2000

The Honorable Christopher S. Bond
Chairman, Committee on Small Business
United States Senate

Dear Mr. Chairman:

Taxpayers spend considerable time and money
complying with the federal income tax rules. While
much of this compliance burden is attributable to
the tax code, another portion can be linked to
Internal Revenue Service (IRS) actions. Reliable
estimates of the burdens imposed by such
compliance could assist Congress in reducing those
burdens, and assist IRS in developing measures of
its own performance, making better informed
resource allocation decisions within the agency,
and meeting reporting requirements under the
Paperwork Reduction Act. We have reported that the
estimation model that IRS currently uses for
compliance burden ignores important components of
burden and has limited capabilities for analyzing
the determinants of burden.1

You asked us to report on the status of IRS'
efforts to improve its estimates of taxpayer
compliance burden. Accordingly, in this report,
our objectives were to (1) describe IRS' overall
strategy to improve its methodology for estimating
compliance burden, (2) describe what IRS has done
to begin implementing its overall strategy and how
IRS expects this to improve its methodology, and
(3) determine whether IRS expects that its new
methodology will be able to measure the burden
associated with the complex tax rules identified
in IRS' forthcoming first annual report on tax
complexity.2

Results in Brief
To improve its methodology for estimating
compliance burden for all types of federal
taxpayers, IRS is pursuing a multiphased strategy.
Initially, IRS is focusing on taxpayers who have
only wage and investment (W&I) income because they
bear a large portion of the overall compliance
burden and because their burden may be easiest to
estimate. Later phases would develop estimates for
other taxpayer groups, such as small businesses
and the self-employed and large and medium-sized
businesses. This incremental strategy should give
IRS and its contractors the flexibility to
incorporate lessons learned from early phases of
the process into later efforts.

IRS has begun to implement its strategy by
contracting for the development of two models
that, when combined, should provide more reliable
estimates of W&I taxpayers' prefiling, filing, and
postfiling compliance burdens. Compared with IRS'
current methodology, these models are designed to
produce more comprehensive estimates of federal
income tax compliance burdens and to provide IRS
with a greater capacity to analyze the impact of
tax law and administrative changes on those
burdens. As with all such modeling, the specific
capabilities and precision of the new models will
depend on the quality of the underlying data.

IRS expects that one of the new models, covering
prefiling and filing activities, will provide some
assistance in estimating burdens associated with
the complex rules identified in IRS' forthcoming
first annual report on tax complexity.3 For some
rules, the model may be able to show both the
number of W&I taxpayers affected and the
approximate size of their burdens. For other
rules, IRS expects that the model will only be
able to show the potential number of taxpayers
affected. The other model, which is to estimate
postfiling burden, is in the early design stage,
and its capabilities have not yet been determined.

IRS reviewed a draft of this report and agrees
with the content.

Background
IRS has been seeking to improve its compliance
burden estimation methodology for several reasons.
First, it would like a tool to assist policymakers
in understanding the sources of burden and
evaluating options for reducing that burden.
Second, IRS would like to know how its own
programs and activities affect compliance burden.
This information could be used to develop
additional performance goals and measures for
reducing burden and would assist managers in
targeting resources and developing new programs.
Third, the 1995 amendments to the Paperwork
Reduction Act of 1980 and the 1993 Government
Performance and Results Act (GPRA) impose new
reporting requirements that IRS' existing
methodology was not designed to meet.

IRS' current model for estimating taxpayer
compliance burden, developed by Arthur D. Little,
Inc. (ADL), was considered a substantial
improvement over the agency's previous methodology
when it was first introduced in the 1980s.
However, IRS and other observers have identified
several significant shortcomings of the model that
limit its usefulness and accuracy. Two important
limitations of the model are that it ignores
significant components of burden, as well as
important determinants of burden.4

The ADL model was designed to meet reporting
requirements arising from the Paperwork Reduction
Act. Pursuant to that act, the Office of
Management and Budget (OMB) established the annual
Information Collection Budget and required federal
agencies to provide estimates of the paperwork
burdens that they impose on the public.5 The
definition of burden that the ADL model uses to
meet this reporting requirement includes the
estimated time costs (burden hours) imposed by
information collections but excludes costs
associated with tax planning and postfiling
activities, such as preparing for an audit. It
also excludes most monetary expenses. For example,
paid preparer fees were converted into an
equivalent time amount, but expenditures on books,
software, and delivery services were excluded.
Amendments to the Paperwork Reduction Act in 1995
now require IRS and other agencies to include
these monetary expenses in their burden estimates.
IRS would like to use an even broader definition
of compliance burden-including the tax planning
and postfiling burdens-when developing performance
measures under GPRA or when evaluating the effects
of changes in tax laws and administrative
practices.

Past reviews of the ADL model have criticized the
approach that it uses to update estimates of
compliance burden from year to year as simplistic.6
These reviews noted that the model does not
account for changes in important determinants of
burden, such as tax preparation technology and
taxpayers' education and experience. These
shortcomings limit the model's usefulness as an
analytical tool and raise concerns about the
accuracy of its estimates.

In August 1998, an IRS study group identified the
type of burden estimation model that the agency
would like to have.7 Ideally, the model would be
able to

ï¿½    estimate the burden associated with all
prefiling, filing, and postfiling activities
undertaken to comply with federal income, estate
and gift, employment, and excise tax rules;
ï¿½    disaggregate total compliance burden by type
of tax, taxpayer, and activity;
ï¿½    disaggregate burden by origin of compliance
requirements (tax laws, regulations, and
administrative procedures);
ï¿½    estimate changes in burden associated with
potential tax law changes;
ï¿½    function in an integrated manner, allowing
users to see how certain tax changes affect
multiple taxpayer groups;
ï¿½    disaggregate burden by IRS function (e.g.,
submission processing and customer service) for
burdens associated with the requirements of, and
taxpayer interactions with, those functions;
ï¿½    estimate the impact on taxpayer burden of
alternative enforcement programs and techniques
providing customer service;
ï¿½    provide different types of estimates for
different purposes (e.g., the estimates required
for the Information Collection Budget are
different from those that IRS would use for
performance measurement under GPRA);
ï¿½    estimate burden in terms of dollars as well
as time; and
ï¿½    incorporate new data with sufficient ease so
that the model, itself, would not have to be
replaced in the near future.

Scope and Methodology
To address the first two objectives of this report
we interviewed IRS officials about their ongoing
and planned efforts to develop new burden
estimation methodologies. We also reviewed
extensive documentation prepared by IRS'
contractor PricewaterhouseCoopers (PwC), which
described (1) the objectives of the new compliance
burden estimation models; (2) the conceptual and
operational designs of the model covering
prefiling and filing burdens and the conceptual
design of the model for postfiing burdens; and (3)
research that PwC completed to inform its design
decisions. We did not evaluate the effectiveness
or appropriateness of the designs for the new
models. We also reviewed documentation relating to
the ADL model, as well as existing evaluations of
that model by IRS and external analysts, in order
to identify differences between that model and the
new models.

To address the third objective we obtained
information from IRS about the expected content of
their forthcoming complexity report. We reviewed
PwC's description of how it expects the new
prefiling/filing burden model will operate and
inferred how that model could be used to estimate
the burden associated with certain aspects of the
tax rules discussed in IRS' report. We confirmed
with IRS officials that our understanding of how
the model could be used with respect to these
complex rules was consistent with IRS'
expectations of the model's capabilities.

We conducted our review from July 1999 through
April 2000 in accordance with generally accepted
government auditing standards. We requested
comments on a draft of this report from the
Commissioner of IRS. His comments are discussed
near the end of this letter and are reproduced in
appendix III.

IRS Is Pursuing a Multiphased Approach in
Developing New Burden Estimation Methods
To accomplish its goal of improving its
methodology for estimating compliance burden for
all types of federal taxpayers, IRS is pursuing a
multiphased strategy. Initially, IRS is focusing
on taxpayers who have only W&I income. Later
phases would develop estimates for other taxpayer
groups, such as small businesses and the self-
employed and large and medium-sized businesses.

IRS' incremental approach is intended to mitigate
the risks associated with developing a new
methodology for estimating compliance burden. The
approach should give IRS and its contractors an
opportunity to incorporate lessons learned from
early phases of the process into later efforts. It
also should give IRS the flexibility to
significantly modify or terminate the overall
development process or its relationship with a
particular contractor. IRS' Office of Program
Evaluation and Risk Analysis is responsible for
overseeing the contractor's work.

The first phase of the development process focuses
on the design and implementation of models for
estimating the federal income tax compliance
burden of W&I taxpayers-those that derive all of
their income from wages, pensions, interest,
dividends, and capital gains. This taxpayer group
was selected for the initial phase because it
accounts for a large share of IRS' overall
compliance burden and because its burden may be
the easiest to estimate. The conceptual definition
of compliance burden developed for this group,
along with the theoretical framework for measuring
that burden, are intended to build a foundation
for developing burden estimates for other taxpayer
groups.

In subsequent phases, estimation methodologies are
to be developed for other segments of the taxpayer
population, such as small businesses and self-
employed taxpayers and large and medium-sized
businesses. The timing of these later phases will
depend on resource availability and the
accumulation of experience. IRS expects to
contract out later this year for a study that
would develop a methodology for estimating the
time and money that self-employed individuals
spend preparing and filing their federal income
tax returns. The agency also expects to begin work
on employment taxes later this year and on small
corporate taxpayer burden in calendar year 2001.

New Estimation Models for W&I Taxpayers Are in
Development
IRS has begun to implement its strategy by
contracting for the development of two models
that, when combined, are to estimate W&I
taxpayers' prefiling, filing, and postfiling
compliance burdens. Compared with IRS' current
methodology, these models are designed to produce
more comprehensive estimates of federal income tax
compliance burdens and to provide IRS with a
greater capacity to analyze the impact of tax law
and administrative changes on those burdens. As
with all such modeling, the specific capabilities
and precision of the new models will depend on the
quality of the underlying data.

IRS' Contractor Has Begun Developing Two Models
PricewaterhouseCoopers, the contractor engaged to
assist IRS in the initial phase of its strategy,
has begun to develop two models covering W&I
taxpayers-one for estimating the prefiling and
filing burdens, the other for estimating the
postfiling burdens of those taxpayers. This work
began in September 1998. The model for estimating
prefiling and filing burdens is expected to be
delivered to IRS in May 2001. The postfiling
burden model also is expected to be designed by
then, but the operational model is to be delivered
at a later date. Before delivering either model,
PwC is to provide IRS with paperwork burden
estimates for this taxpayer group, using the new
methodology, for the agency's Information
Collection Budget due in December 2000. The
cumulative cost of this initial effort is expected
to reach approximately $5 million by the end of
fiscal year 2001. Expenditures for later years
have not yet been determined.

The New Models Are Intended to Produce More
Comprehensive Estimates of Burden
The two models are intended to produce more
comprehensive estimates of burden by including all
of the time costs covered by the ADL model, plus
time spent on tax planning and postfiling
activities. They are also to estimate the monetary
expenditures associated with federal income tax
compliance, such as the purchase of books and
software, which are not covered by the ADL model.8

PwC said it used focus groups of taxpayers and tax
preparers to identify a lengthy list of activities
and costs involved in the prefiling and filing
stages of the tax compliance process. It has
grouped these activities and costs into six broad
categories: recordkeeping, gathering tax
materials, using IRS services, tax planning, form
completion, and form submission. The initial plan
for the postfiling burden model is also to group
activities and burdens into six categories:
recordkeeping and retrieval, gathering tax
materials, using IRS services, form completion and
form submission, using a paid professional, and
interviewing.9

The definition of compliance burden for both
models excludes any activities or costs associated
with

ï¿½    the basic financial planning and
recordkeeping that individuals might do even if
there were no income tax,
ï¿½    requirements imposed by federal taxes other
than the income tax,
ï¿½    requirements imposed by state or local taxes,
ï¿½    criminal tax investigations, and
ï¿½    Tax Court proceedings.

According to PwC, its models will not distinguish
between "required" and "voluntary" compliance
activities because there is no clear benchmark of
"required" activities against which to measure
"voluntary" tax minimizing activities. PwC noted
that, although the expenditure of time and money
spent on tax minimizing activities, such as
itemizing deductions, is voluntary,10 some
taxpayers would incur the cost of higher taxes if
they did not make those expenditures.

The New Models Are Also Intended to Provide IRS
With Increased Analytical Capabilities
The two new estimation models are intended to
provide IRS with increased capabilities to analyze
burden, disaggregated by type of taxpayer
activity. In addition, the prefiling/filing burden
model, which is further along in development, is
intended to enable IRS to analyze the effects of
changes in tax laws, IRS programs and services,
and other important determinants of burden.

PwC determined that, to meet IRS' analytical
requirements, its prefiling/filing burden model
would have to take into account the influences of
many significant determinants of burden that the
ADL model ignores. PwC said it relied on its focus
groups of taxpayers and interviews with tax
preparers to identify important determinants of
burden and to gain a better understanding of the
complex relationships between these determinants
and burden. The insights obtained through the
focus groups and interviews have guided PwC's
model design, data gathering, and model
construction efforts.

PwC conceives the tax compliance process as a
series of decisions made by taxpayers that result
in a variety of activities, each of which imposes
some burden. For example, a taxpayer incurs
additional recordkeeping and reporting burdens if
he or she chooses to itemize deductions from
income, rather than claim the standard deduction.
These taxpayer decisions and activities are
influenced by such factors as tax system
characteristics (e.g., specific recordkeeping and
reporting requirements), demographic
characteristics (e.g., age, income, and tax-filing
experience of individual taxpayers), and others
(e.g., tax preparation and filing technologies).

One important insight from the focus groups was
that, in most cases, the activities a taxpayer
engages in change little from year to year.
Another important insight was that the most
significant changes in compliance activities are
due to life changes (e.g., marriage or birth of
dependents), changes in financial situation (e.g.,
the purchase of a house), and changes in the tax
code. According to PwC, changes in these factors
have particularly significant effects on
compliance burden when they lead taxpayers to
change their basic approach to tax preparation.
This is due to the fact that the types of
activities taxpayers engage in vary significantly,
depending on whether they prepare their return by
hand, use tax preparation computer software, or
hire a professional to prepare their tax return.
PwC also determined that the more experience
taxpayers gain in filling out returns, the less of
a burden it becomes. PwC expects that models that
account for these determinants of burden will
provide IRS with analytical capabilities it does
not possess now.

Realizing the Increased Capabilities Depends on
Complicated Data Collection and Analysis
There are three phases in the construction and
operation of the prefiling/filing burden model,
all of which must be successfully completed for
IRS to realize the increased analytical
capabilities it is seeking. First, data must be
collected that describe taxpayer characteristics,
decisions, activities, and burdens. Second,
statistical techniques must be used to estimate
the relationships between key outcomes, such as
taxpayer decisions and compliance burden levels,
and their determinants. Finally, the estimated
relationships must be applied to data from a
representative sample of taxpayers in order to
simulate taxpayer decisions and the resulting
level of burden.

Data Collection Will Be Challenging
PwC intends to meet many of its data requirements
by relying upon existing IRS administrative and
research data files. It also hopes to make use of
demographic and personal finance data that the
Census Bureau and other government agencies
collect from samples of households. However, PwC
and IRS have decided that they need to survey a
representative sample of taxpayers in order to
collect information on the compliance burdens
those taxpayers incur. Each telephone interview is
expected to cover up to 92 questions and last
approximately 20 minutes, depending on the
characteristics of the respondent. Data from this
survey are to be directly merged with data from
IRS' computer files for these same taxpayers. In
addition, statistical imputation techniques11 are
to be used to supplement the data record for each
taxpayer with information that is not available
from IRS files (such as information on the
taxpayer's educational background and spending
behavior).

The data collection and analysis tasks that PwC
must accomplish are ambitious for several reasons.
First, the taxpayers surveyed by PwC will be
relying on their memories to estimate the amounts
of time and money they spent preparing their 1999
tax returns. Second, for each of these taxpayers,
PwC will have to be able to collect or impute
accurate data on the numerous determinants of
their compliance burdens, including the specific
tax rules that apply to them. Third, PwC will have
to estimate statistical equations that accurately
isolate the individual effect that specific tax
rules have on the compliance burdens of specific
types of W&I taxpayers. While the data collection
and analysis tasks are ambitious, PwC believes
that the available data will enable it to estimate
equations that improve upon IRS' current
methodology.

The data most critical to the modeling
effort-those relating to the time and money that
taxpayers spend complying with tax rules-are
subject to measurement error of an unknown degree.
The accuracy of the burden data that PwC collects
in its survey will depend, in part, on each
respondent's ability to understand the definition
of burden used in the model. PwC recognizes that
there are limits to the amount of guidance it can
provide to respondents during a survey.12 For
example, it does not intend to provide guidance on
how to separate the travel time attributable to
purchasing tax software from the time attributable
to other errands completed during a multipurpose
shopping trip. Individual respondents will decide
for themselves whether all, or only a portion, of
the time spent on the trip should be considered a
burden.

The accuracy of the burden data that PwC collects
will also be limited by recall bias on the part of
respondents. The focus groups that PwC conducted
reconfirmed findings from earlier focus groups
conducted by ADL that taxpayers are not able to
recall the amount of time they spent on tax
compliance very accurately. ADL provided further
evidence of this problem when it used two
different methods to collect burden estimates from
taxpayers-a mail survey and a diary study (in
which taxpayers kept logs of their activities as
they prepared their returns). On average,
compliance time estimates made by respondents to
ADL's mail survey were 78 percent greater than the
estimates made by participants in its diary study
(see app. I). PwC recommended a telephone survey
for collecting its burden data because that
approach usually yields a higher response rate
than a mail survey and allows interviewers to
clarify questions. However, PwC acknowledges that
telephone surveys do not permit taxpayers to
consult documents when responding and that
respondents' recall may not be as accurate as in a
diary study.13

The accuracy of the model's underlying data for
several important determinants of burden, such as
taxpayers' education level and experience in
completing returns, will depend on the precision
of PwC's imputation techniques. Although PwC
intends to collect information on some of these
determinants in its taxpayer survey, it will have
to impute values for them in the model's
production file that is to be updated each year.

IRS and PwC recognize the limitations of the data
on burden and its determinants that will be used
in the model. They have concluded that their plans
represent a cost-effective approach for collecting
the data needed to improve IRS' burden estimation
methodology.

Numerous Equations Describing Taxpayers' Decisions
and Burden Must Be Estimated
Once this initial database is constructed, PwC
plans to use statistical techniques to estimate
equations that describe the important
relationships needed for making simulations. One
set of equations is to show how the decisions that
each taxpayer makes during the tax compliance
process are related to specific factors, such as
that taxpayer's age and education, the particular
forms that the taxpayer must complete, and the
costs of alternative preparation methods. Another
set of equations is to show how the amount of
burden that a taxpayer bears is related to
specific characteristics of that taxpayer and the
particular compliance requirements that the
taxpayer faces. These relationships are to be
estimated separately for each of the six
categories of burden (recordkeeping, form
completion, etc.).

If the estimation efforts are successful, the
relationships described in both sets of equations
should enable IRS to regularly update its model
using data on taxpayer characteristics that it
already collects every year (or that can be
imputed from other available sources). IRS should
not need to collect new data on burden for each
update of this "production" data file.14

To estimate the aggregate prefiling and filing
compliance burden for tax years after 1999, IRS
will need to update the model's data file to
reflect changes in the taxpayer population and
enter information about any tax law changes made
since 1999. The model is to then process this
information in several stages. First, it is to
simulate the effect that the law changes have on
the compliance requirements faced by each taxpayer
represented in the data file. Second, the model is
to simulate taxpayer decisions to determine if the
changes in the tax laws, or changes in taxpayers'
circumstances from one year to the next, would
cause any of

the taxpayers to change their preparation or
submission methods.15 Finally, the model is to take
this updated information on tax requirements and
compliance methods for each taxpayer and, by
applying the relationships from the previously
estimated equations, estimate the time and money
that each taxpayer spent on each of the six
categories of burden activities.16

The Prefiling/Filing Burden Model, by Itself, Will
Not Provide Everything IRS Was Seeking
While IRS expects the new prefiling/filing burden
model to improve its burden estimation
capabilities, the model, by itself, will not
provide all of the capabilities that IRS was
originally seeking. It will not disaggregate
burden by specific IRS function (e.g., submission
processing and customer service), nor will it
disaggregate burden by origin of compliance
requirement (tax law vs. administrative
procedure). The model is supposed to be able to
accommodate these capabilities at a future date,
but it will be up to IRS to determine how these
disaggregations would be done.

In addition, PwC is not expected to statistically
estimate the effect that each and every tax rule
has on the average burdens of W&I taxpayers.
However, even in cases where PwC has not
statistically estimated the average burden of a
particular tax rule, the model may be able to
identify how many and what types of W&I taxpayers
are affected by the rule. IRS could then use some
independent method, such as a targeted research
project, to estimate the change in the average
amount of burden that a specific change in the tax
rule would cause. By combining this estimate with
the information on the number of taxpayers
affected, IRS would be able to estimate the
aggregate burden change that would result from the
tax rule change. At this time, PwC does not know
for how many, or which, tax rules it will be able
to make specific statistical estimates.

Fundamental Decisions on the Design of the
Postfiling Model Remain to Be Made
The primary functional requirements of the
postfiling burden model are the ability to

ï¿½    estimate the change in burden due to changing
factors, such as the number of audits IRS
undertakes;
ï¿½    disaggregate burden estimates by postfiling
segment17 and by taxpayer activity; and
ï¿½    link to the prefiling/filing burden model.

Fundamental decisions remain to be made regarding
the operational design of the postfiling model. A
report that presents PwC's broad conceptual design
for the model notes that PwC and IRS need to agree
on specific interpretations of the general
requirements that were set by the contract's scope
of work. The report also indicates that the
ultimate functionality of the model may be
constrained by the limitations of IRS' data
systems. For example, IRS has no database that
tracks the activities of individual taxpayers
throughout the entire postfiling process. This may
prevent the model from determining if the burdens
that taxpayers experience in one segment of the
postfiling process are influenced by their
experiences in a preceding segment.

The New Models Could Assist IRS in Estimating the
Burdens Associated With Some Aspects of Complex
Tax Rules
The prefiling/filing burden model may assist IRS
in estimating burdens associated with some aspects
of the complex rules identified in IRS'
forthcoming report on tax complexity. That report
will devote detailed attention to three areas of
concern for W&I taxpayers: the Alternative Minimum
Tax for individuals, the variety of definitions
that taxpayers must learn in order to determine
filing status and eligibility to claim dependents
and credits, and the estimated tax rules.18 The
data that the contractor plans to collect for the
prefiling/filing burden model will provide it with
the opportunity to directly estimate the burdens
associated with some, but not all, aspects of
those complex rules. IRS has not yet decided
whether the postfiling model should have any
ability to estimate burdens associated with
specific tax rules.

The current list of variables that PwC intends to
incorporate in its statistical equations includes
many that relate to the complex tax rules covered
in the Commissioner's report. Among these are
variables that indicate19

ï¿½    whether taxpayers spent any time dealing with
the AMT form, even if they ultimately did not have
to submit it;
ï¿½    whether taxpayers who did file the AMT form
filled in selected lines on that form;
ï¿½    which filing status, dependents, and credits
taxpayers actually claimed; and
ï¿½    whether taxpayers paid estimated taxes and
estimated tax penalties.

The availability of data for these variables may
enable PwC to make estimates, such as the average
amount of time that the existence of the AMT adds
to the total compliance time of different groups
of W&I taxpayers. In those cases where PwC is able
to estimate average burden effects, the model
should be able to use that information, along with
information on the number of taxpayers affected by
the tax rules in question, to estimate the
aggregate burdens attributable to those rule.
However, as previously noted, PwC is not expected
to directly estimate the average burden effect of
every aspect of these complex tax rules. For some
rules, IRS expects that the model will only be
able to show the number of taxpayers affected.

Agency Comments and Our Evaluation
We received written comments from the Commissioner
of IRS in a letter dated May 3, 2000. In his
letter, the Commissioner agreed with the contents
of our report and emphasized two points that we
had mentioned: (1) in contrast to the ADL model,
the new model will incorporate tax preparation
method as a key determinant of burden, and (2) IRS
expects to be able to keep the new model up to
date without unduly burdening itself or taxpayers.

In addition to the written comments, IRS provided
technical comments on the report, which we
incorporated where appropriate. The Commissioner's
letter is reproduced in appendix III.

As agreed with your office, unless you announce
its contents earlier, we plan no further
distribution of this report until 30 days from the
date of this letter. We will then send copies to
Senator William V. Roth, Jr., Chairman, and
Senator Daniel P. Moynihan, Ranking Minority
Member, Senate Committee on Finance; Senator John
F. Kerry, Ranking Minority Member, Committee on
Small Business; Representative Bill Archer,
Chairman, and Representative Charles B. Rangel,
Ranking Minority Member, House Committee on Ways
and Means; the Honorable Lawrence H. Summers,
Secretary of the Treasury; the Honorable Charles
O. Rossotti, Commissioner of Internal Revenue; and
the Honorable Jacob J. Lew, Director, Office of
Management and Budget. We will also make copies
available to others on request.

Please contact me or James A. Wozny at (202) 512-
9110 if you have any questions. Marvin G. McGill
and Margarita Vallazza made key contributions to
this report.

Sincerely yours,

James R. White
Director, Tax Policy and
   Administration Issues

_______________________________
1Small Business Tax Compliance Burden (GAO/GGD-99-
96R, May 5, 1999).
2This report was mandated by the IRS Restructuring
and Reform Act of 1998 (P.L. 105-206, July 22,
1998).
3The IRS report will focus on the Alternative
Minimum Tax (AMT) for individuals, the variety of
definitions that taxpayers must learn in order to
determine their filing status and eligibility to
claim dependents and credits, and the estimated
tax rules.
4See appendix I for more detail on these and other
limitations of the model.
5The Information Collection Budget is the means by
which the federal government, through OMB,
measures and controls the number of hours
individuals, businesses, state and local
governments, and others must spend complying with
federal reporting requirements. The Information
Collection Budget is prepared annually based upon
the prior fiscal year's experience and current
estimates of the "burden hours" imposed by
individual forms, surveys, and other information
collections.
6See, e.g., IRS, "Report of the Taxpayer Burden
Study Group: Roadmap for a New Measure and Pilot
Study for Individual Non-Business Taxpayers,"
Request for Proposal (Aug. 19, 1998); Marsha
Blumenthal, Burden Reduction Research and
Analysis, report submitted to the Analysis and
Studies Division, IRS (May 1996); and Henry Beale,
untitled draft report on Arthur D. Little
Compliance Burden Model, submitted to the Analysis
and Studies Division, IRS (1996).
7IRS, "Report of the Taxpayer Burden Study Group."
8Neither of the new models is to  estimate the
costs of employers, banks, and other third parties
that are required to provide information returns
to individual taxpayers. Those costs are to be
included when the total tax compliance costs of
those entities are estimated at a later date.
9Even though some of the category names are the
same for the two models, the activities covered by
the two models do not overlap. For the postfiling
model, e.g., "form completion and submission"
covers only submissions that may have to be made
after the initial filing of the taxpayer's return.
10Taxpayers have the option of claiming the
standard deduction, which is less burdensome than
itemizing.
11Imputation techniques are commonly used to
combine data from separate sources when it is not
possible to use a common identifier, such as a
Social Security Number, to ensure that the data
being linked from each source pertain to the same
individual. One could use, e.g., detailed data
from a survey to estimate a regression equation
showing the relationship between variables, such
as income and area of residence on the one hand,
and level of education on the other. With the
estimate of this relationship, one could then
"impute" the level of education of individuals
represented in a different database, as long as
the second database contained data on income and
area of residence.
12PwC has drafted a questionnaire to be used in a
telephone and mail survey of taxpayers. Appendix
II summarizes the instructions from that
questionnaire.
13PwC intends to send an advance letter to
respondents, telling them the general nature of
the questions they will be asked. PwC also intends
to mail a written version of the questionnaire to
taxpayers for whom they cannot find telephone
numbers.
14The core of the production data file is to
consist of detailed data that IRS' Statistics of
Income Division collects annually from a large
sample of individual tax returns.
15Although the model is to simulate changes in
compliance-related behavior, as currently
designed, it would not, by itself, simulate other
types of behavioral changes that affect aggregate
burden. For example, it would not simulate how an
increase in documentation requirements for
charitable contributions would affect the number
of taxpayers who claim deductions for such
contributions. The model, however, is to allow
users to make "off-model" estimates of those
behavioral changes and then adjust the data in the
model to capture the effects of such changes.
16If appropriate data become available, the model
would have the capability to translate time spent
by taxpayers into a monetary cost. The methodology
for monetizing taxpayer time is to be based on
forthcoming guidance from OMB. A PwC subcontractor
has drafted a "white paper" on monetizing
compliance burden that summarizes existing
research and identifies best practices. OMB is to
review that paper before developing its guidance.
OMB has also solicited input to its guidance from
the broader public.
17PwC proposes to divide the postfiling process
into five segments: clarification (relating to IRS
requests for further information), collection,
audit, appeals, and amendment (relating to the
filing of amended returns).
18IRS selected these three areas after reviewing
over 40 reports and testimonies published since
1989 and meeting with both internal and external
stakeholders interested in tax
complexity/simplification.
19PwC notes that its list of variables is "dynamic"
and that some will be dropped and new ones added
as the development of the model proceeds.

Appendix I
Limitations of the Arthur D. Little Model
Page 19 GAO/GGD-00-11 Taxpayer's Compliance Burden
IRS and Other Reviewers Have Identified a Number
of Limitations
To measure the paperwork dimension of compliance
burden, IRS sponsored a study in 1984 that has
been the basis of its burden estimation
methodology since that time. Arthur D. Little,
Inc. (ADL), an IRS contractor, released the
results of the paperwork burden study in 1988.
Since then, IRS and other reviewers of ADL's
paperwork burden estimation model have criticized
it for number of reasons.1 The reported
limitations of the model and its estimates can be
categorized into four areas: (1) the age of the
underlying survey data, (2) the exclusion of
certain components of burden, (3) the model's
simplistic treatment of the determinants of
burden, and (4) the questionable statistical
validity and poor documentation of the estimates.

Underlying Data Are Outdated
Several reviewers of the ADL model, including IRS,
have criticized the age of the burden model's
survey data. They have observed that the survey
data reflect 1983 information and lack relevance
to today's tax environment because of the many
significant changes in tax law and tax preparation
technologies that have taken place since then.

Reviewers have also expressed concern over the
accuracy of the burden estimates taxpayers
provided when they were surveyed. To obtain the
information about burden, ADL used two collection
methods-a mail and a diary study. The burden
estimates yielded by the two methods varied
significantly. On average, diary respondents'
estimated burden, by return, was 8.32 hours, while
mail survey respondents' estimated burden was
14.82 hours-78 percent higher. ADL could not
determine which estimate was more accurate, so it
decided to use the average of the two estimates.2

Certain Components of Burden Are Excluded
Past reviewers have noted that the definition of
burden used in the ADL model excludes important
components of burden.  The ADL model does includes
the time costs (burden hours) imposed by
information collections but excludes costs
associated with tax planning and postfiling
activities, such as preparing for an audit. It
also excludes most monetary expenses. For example,
paid preparer fees are converted into an
equivalent time amount, but expenditures on books,
software, and delivery services are excluded.

The Determinants of Burden Are Simplistically
Incorporated
The ADL model has been criticized for the
simplistic nature of the equations it uses to
represent the relationship between burden and its
determinants. In its report, ADL acknowledged the
simplicity of its equations but explained that IRS
decided it would be operationally infeasible to
use more detailed and precise models.

One past reviewer noted that the only determinants
of burden used in the final model were measures of
return size (numbers of forms and attachments),
form size (number of words or number of line
items), the number of references to the tax code
in forms and instructions, and the number of line
items requiring records.3 As a consequence, many
important determinants were omitted from the
model. For example, the model does not allow for
alternative filing methods, such as electronic
filing software programs, and it also does not
differentiate levels of burden between simple and
complex types of forms. The model does not
differentiate burden between simple and complex
line items on a form, and it does not separate
burden within the form and identify parts of the
form that represent the greatest burden. In some
cases the model erroneously estimates the impact
of efforts to simplify forms, instructions, or
procedures. For instance, if new lines are added
to a form's worksheet to simplify computations,
the ADL model reports an increase in burden.

The Statistical Validity of the Model Has Been
Questioned
Reviewers have also called into question the
statistical validity of the model. IRS indicates
that errors were made in the sample weighting
techniques. In addition, the model's methodology
was poorly documented. IRS noted that the ADL
burden estimates are inconsistent because they
yield results that assign unrealistically high
levels of burden to certain forms (e.g., Form
4789, Currency Transaction Report, and Form 8300,
Report of Cash Payments Over $10,000 Received in a
Trade or Business). At the same time, IRS noted
that the ADL burden estimates for many other forms
did not include the amount of time taken to mail
them, even though those forms had to be mailed to
IRS. Researchers have also questioned the validity
of the business burden estimates because the ADL
model yields an estimate of burden greater than
that found in the survey data on which the model
is based.

_______________________________
1IRS, "Report of the Taxpayer Burden Study Group:
Roadmap for a New Measure and Pilot Study for
Individual Non-Business Taxpayers," Request for
Proposal (Aug. 19, 1998); Henry Beale, Report on
Arthur D. Little Taxpayer Compliance Burden Model,
untitled draft report submitted to the Analysis
and Studies Division, IRS (1996); Marsha
Blumenthal, Burden Reduction Research and
Analysis, Report submitted to the Analysis and
Studies Division, IRS (May 1996);
PricewaterhouseCoopers, Compliance Burden
Literature Review, draft report submitted to IRS
(June 16, 1999).
2ADL made this correction by multiplying the
burden estimates that would have been obtained if
it had relied exclusively on the mail survey
results by 0.78. The corrected estimate equals the
average of the two estimates that would have been
obtained by using the diary results and the mail
survey results separately.
3Beale, Report on Arthur D. Little Taxpayer
Compliance Burden Model.

Appendix II
Details on the Definition of Burden to Be Used in
the PwC Survey
Page 23 GAO/GGD-00-11 Taxpayer's Compliance Burden
Draft Questionnaire on Prefiling/Filing Burden
PwC has drafted the questionnaire that it plans to
use to gather information on prefiling and filing
from a sample of wage and investment taxpayers.
This appendix summarizes the parts of the
questionnaire that instruct respondents on the
types of costs to include in their estimates of
compliance burden and ask respondents to estimate
the specific components of their burden. The full
draft questionnaire includes over 90 questions. In
addition to the questions listed below, the
questionnaire asks respondents for selective
demographic information that is not available from
IRS files and information on tax preparation
practices. Additional questions are designed, in
part, to prompt respondents' memories about other
activities that they may have undertaken in
preparing their tax returns.1

Recordkeeping
The recordkeeping category includes maintaining
documents needed as proof of income or expenses
for federal income tax purposes. Relevant
documents include receipts, W-2 forms, and 1099
statements of interest or dividends. Time and
money spent balancing checkbooks, paying bills, or
keeping records for state or local tax purposes
are not to be included. The draft questions
include the following:

ï¿½    How much time did you personally spend on
recordkeeping activities related to your 1999
federal income tax return? Include
ï¿½    Time spent retrieving and organizing your
 records in preparation for
    completing your return.
ï¿½    All the time spent on activities you did
related to your federal income
   tax return even if you also needed the
information for other purposes.
ï¿½    How much time, if any, did your spouse,
relatives, or friends spend in assisting you with
these recordkeeping activities?
ï¿½    How much money, if any, did you spend on
these recordkeeping activities? Include
ï¿½    Expenses such as overnight delivery fees,
 appraisal costs, and faxing
    or copying fees.

Do not include

ï¿½    The cost of any accounting software, such as
Quicken, unless you use it only for federal income
tax record keeping purposes.
ï¿½    Any costs associated with helping someone
else.

Gathering Tax Materials
The gathering tax materials category covers time
and money spent gathering materials needed to
complete the taxpayer's federal income tax return
or for tax planning purposes. These materials may
include any tax forms, tax-related publications,
books, or guides collected from IRS or other
sources and tax preparation software or 1999
upgrade.  Some of the draft questions are the
following:

ï¿½    How much time did you spend gathering the
materials we just discussed?
ï¿½    How much time, if any, did your spouse,
relatives, or friends spend in gathering your
materials?
ï¿½    How much money, if any, did you spend
obtaining the materials we just discussed?

Using IRS Services
The using IRS services category covers the time
and money spent using services, such as the Toll-
Free Tax Assistance telephone line, Walk-in
Customer Assistance site, Volunteer Income Tax
Assistance site, Tax Counseling for the Elderly,
IRS Web site, or Tele-Tax Telephone line. Some
questions for taxpayers include the following:

ï¿½    In total, how much time did you spend using
these IRS services, including the time reaching an
assistor, being put on hold, waiting, and using
these services?
ï¿½    How much time, if any, did your spouse,
relatives, or friends spend using these IRS
services related to your return, including the
time reaching an assistor, being put on hold,
waiting, and using these services?
ï¿½    Of the [time amount], how much time was spent
actually using the service after waiting to see
someone or holding on the telephone?
ï¿½    How much money, if any, did you spend, using
these IRS services?

Paid Professionals
The paid professional category covers time and
money spent by taxpayers when consulting with
professionals for tax planning purposes or to
prepare and complete their tax returns, including
the time spent finding and selecting the
professional. It does not include time spent
obtaining general investment advice that was not
directly related to the taxpayer's federal income
tax; nor does it include costs for consulting on
someone else's behalf. The category includes the
professionals' fees for consultation and services,
such as overnight delivery fees, appraisal costs,
and faxing and copying. Some draft and survey
questions are the following:

ï¿½    How much time did you spend consulting with
paid professionals?
ï¿½    How much time, if any, did your spouse,
relatives, or friends spend consulting with paid
professionals about your federal income taxes?
ï¿½    How much did you pay the professionals with
whom you consulted?
ï¿½    Was any of the [dollar amount] paid in order
to file your return electronically?
ï¿½    How much did you spend to file your return
electronically?
ï¿½    Was any of the [dollar amount] paid in order
to receive an accelerated or instant refund?
ï¿½    How much did you spend to get an accelerated
refund?
ï¿½    Was any of the [dollar amount] paid in order
to complete your state or local tax return?
ï¿½    How much did you spend to get the state or
local tax return completed?

Tax Planning
The tax planning category covers the time and
money spent considering the federal income tax
implications of activities, such as contributing
to an Individual Retirement Account, including
withdrawals and distributions; participating in a
401(k) plan or other employer-sponsored retirement
plan; taking advantage of tax-exempt or tax-
preferred investments; realizing capital gains or
capital losses on any investments; exercising
stock options; and planning charitable
contributions. A few of the draft questions are
the following:

ï¿½    How much time did you spend on federal income
tax planning activities?
ï¿½    How much time (in hours or minutes), if any,
did your spouse, relatives, or friends spend in
activities related to federal income tax planning
for your federal income taxes?
ï¿½    How much money, if any, did you spend on
these activities, including the cost of classes,
seminars, or workshops attended for tax planning
purposes? Do not include the costs of a paid
professional or costs related to tax planning for
someone else.

Form Completion
The form completion category covers the time and
money spent completing forms, whether they were
submitted or not; consulting instructions,
publications, or other materials while filling out
forms; or completing your federal income tax and
related returns. It also includes the time
reviewing and checking the returns no matter how
they were prepared and completing nonreturn forms
such as the W-4 and 1040ES. The category may also
include the time spent filing for an extension or
installing or updating tax preparation software.

The form completion category does not include time
submitting the returns; preparing state, local, or
someone else's tax returns; or recordkeeping. It
also does not include the costs of a paid
professional or costs related to completing
someone else's return. Draft questions include the
following:

ï¿½    How much time did you spend completing your
1999 federal income tax return?
ï¿½    How much time if any, did your spouse,
relatives, or friends spend completing your 1999
federal income tax return?
ï¿½    How much time did you spend on forms or
worksheets that you did not submit?
ï¿½    How much money, if any, did you spend to
complete your 1999 federal income tax return?

Form Submission
The form submission category may include
photocopying and mailing charges and cost of
electronic filing, including service fees.  Draft
questions include the following:

ï¿½    How long did it take you, your spouse,
relatives, or friends to submit your return [and
complete the form 8453], including the time spent
photocopying and assembling your return?
ï¿½    How much money did you spend to submit your
return?

_______________________________
1Examples of such questions are: "Did you maintain
any documentation related to the household
employee tax?" and "Did you attend any classes,
seminars, or workshops primarily to complete your
tax return last year?"

Appendix III
Comments From the Internal Revenue Service
Page 25 GAO/GGD-00-11 Taxpayer's Compliance Burden

*** End of Document ***