Robert M. Howard

Getting a Poor Return

Courts, Justice,
and Taxes

This page intentionally left blank.

GETTING A POOR RETURN

This page intentionally left blank.

GETTING A POOR RETURN

Courts, Justice, and Taxes

ROBERT M. HOWARD

State University of New York Press

CONTENTS

Published by

State University of New York Press, Albany

Printed in the United States of America

No part of this book may be used or reproduced in any manner whatsoever without
written permission. No part of this book may be stored in a retrieval system
or transmitted in any form or by any means including electronic, electrostatic,
magnetic tape, mechanical, photocopying, recording, or otherwise without the
prior permission in writing of the publisher.

For information, contact

State University of New York Press, Albany, NY

www.sunypress.edu

Production, Laurie Searl

Marketing, Anne M. Valentine

Library of Congress Cataloging-in-Publication Data

Howard, Robert M., 1956–
Getting a poor return : courts, justice, and taxes / Robert M. Howard.
p.

cm.

Includes bibliographical references and index.
ISBN 978-1-4384-2889-5 (hardcover : alk. paper)
1. Tax courts—United States. 2. Tax protects and appeals—United States. 3. Tax
collection—United States. 4. Justice, Administration of—United States. I. Title.

KF6324.H69

2009

343.7304'2—dc22

2009007364

10 9 8 7 6 5 4 3 2 1

PREFACE

To Taryn, Courtney, and Jordan with love

CONTENTS

This page intentionally left blank.

vii

PREFACE

CONTENTS

List of Figures and Tables

Acknowledgments xi

One

Tax Policy and the Pursuit of Justice

Courts and Justice

Taxes and Tax Policy

Courts, Fairness, and Taxes

Chapter

Overview

Two

Courts and the IRS

The Need for Tax Forums

Administrative Appeals and Trial Courts: Options

and

Odds

Conclusion

Three

Tax Litigation and Tax Forum Choice

Political Forces and Litigation

The Decision to Litigate and Forum Strategy

Hypotheses

Factors Infl uencing Litigation and Forum Choice

Results

Conclusion

Four

Tax Decision Making

Courts, Law, and Politics

Taxpayers,

Courts,

and

Outcomes

Hypotheses

Results

Conclusion

viii

CONTENTS

Five

Infl uences on the IRS and the Audits of Low-Income

Taxpayers

Bureaucratic Control and the IRS

The National Political Coalition, Courts, and

Bureaucratic

Control

Hypotheses

Testing

Infl

uences

Results

Conclusion

Six

Courts, Fairness, and the Creation and

Enforcement of National Tax Policy

Courts, Fairness, and National Governing Coalitions

Changing Tax Policy, Changing the Courts

Appendix Methodology and the Use of Panel Data

Estimation

Problems

Models for Chapter Three

Models for Chapter Four

100

Models for Chapter Five

103

Cases Cited

109

Notes

111

References

115

Index

123

PREFACE

LIST OF FIGURES AND TABLES

CHAPTER TWO: COURTS AND THE IRS

Table 2.1 Comparing Tax Trial Court Options

CHAPTER THREE: TAX LITIGATION AND TAX FORUM CHOICE

Figure 3.1 Total Tax Litigation Filings 1994–2000

Figure 3.2 Change in Ideology of Tax and District Courts

1992–2000

Figure 3.3 Change in the Ratio of Tax Court to District Court

Filings

1994–2000

Table 3.1 Model of Judicial Ideology

Table 3.2 Descriptive Statistics Filings and Forum Choice

1994–2000

Table 3.3 Determinants of Tax Litigation 1994–2000

Table 3.4 Determinants of Forum Choice 1994–2000

CHAPTER FOUR: TAX DECISION MAKING

Figure 4.1 President, Senate, and Tax Court Ideology 1983–1998

Table 4.1 Tax Court Judge Experience

Table 4.2 Descriptive Statistics U.S. Tax Court and

U.S.

District

Court

Table 4.3 Ideology and the Probability of IRS Winning

Table 4.4 Structure, Specialization, and Expertise and the

Probability of IRS Winning

CHAPTER FIVE: INFLUENCES ON THE IRS AND THE

AUDITS OF LOW-INCOME TAXPAYERS

Figure 5.1 Executive-Appointee Target—Veto-Override Point

Figure 5.2 Court-Conscious Executive Appointee Targets

Figure 5.3 Trial Court Ideological Array

Table 5.1 Judicial Ideological Array and Expected Value

Table 5.2 Descriptive Statistics—Audit Rates

FIGURES AND TABLES

Table 5.3 Infl uences on the Number of High-Income/

Low-Income

Audits

Table 5.4 Infl uences on the Number of Individual/

Corporate

Audits

APPENDIX: METHODOLOGY AND THE USE OF PANEL DATA

Table A.1 Determinants of Tax Litigation 1994–2000

(Chapter

Three)

101

Table A.2 Determinants of Tax Forum Choice 1994–2000

(Chapter

Three)

102

Table A.3 Probability of Outcome Favoring IRS (Chapter Four)

104

Table A.4 High-Income/Low-Income Audits (Chapter Five)

106

Table A.5 Individual/Corporate Audits (Chapter Five)

107

ACKNOWLEDGMENTS

There are many people to thank and acknowledge who have helped me
in my career as well as in the preparation of this manuscript. I would like
to thank Byron Nichols of Union College, my college mentor of long ago
and just recently retired. A truly great and inspiring teacher dedicated to
his students, he awakened in me, and in so many others, a lifelong love of
learning and a passion for the mind. He has remained an inspiration and
role model throughout the over thirty years of our friendship. I also want to
thank Jeff Segal, John Scholz, Rich Timpone, and Paul Teske for all their help
before, during, and after graduate school. I particularly want to thank John
for providing the initial tax data that allowed me to start my examination
of courts and taxes so many years ago. I thank them all for their teaching,
inspiration, and repeated kindnesses throughout the years.

I want to thank Scott Graves for his patience and help in answering

my many questions on methodology and Dave Nixon for allowing me the
almost too generous use of his ideas, data, measures, and patience. I thank
them both for their continued friendship. I thank the editors and staff at
SUNY Press for their warmth and patience throughout the production of
this manuscript. I also want to thank Joe Smith, Virginia Hettinger, Harold
Spaeth, and Steve Wasby for their comments, ideas, and suggestions for
various chapters. I owe my parents, Amy and Bernie Howard, more than I
can ever repay. To Courtney and Jordan, I apologize for the warped sense
of humor you both appear to have inherited, but no parent could ask for
better children. It is wonderful to have such great children with whom I
have so much in common. And to Taryn, the love of my life, who knows
what the next years have in store, but I hope we both continue fi ghting
over directions.

xii

CONTENTS

This page intentionally left blank.

TAX POLICY AND THE PURSUIT OF JUSTICE

ONE

TAX POLICY AND THE PURSUIT OF JUSTICE

Anyone may arrange his affairs so that his taxes shall be as low as possible;
he is not bound to choose that pattern which best pays the treasury. There
is not even a patriotic duty to increase one’s taxes . . . nobody owes any
public duty to pay more than the law demands.

—Judge Learned Hand

Taxes are the price we pay for a civilized society.

—Justice Oliver Wendell Holmes

COURTS AND JUSTICE

In the aftermath of Hurricane Katrina, congressional Republicans offered
a detailed list of cuts in the federal budget to offset the enormous cost
of recovery and reconstruction. Among the featured items was a proposal
to increase the audits of those claiming the Earned Income Tax Credit, a
tax credit designed to provide fi nancial assistance to working low-income
individuals. Those claiming the credit in 2005 had a median income of less
than $12,000. By separating the truly poor from those who were merely “non-
affl uent,” the Republican proposal claimed an estimated savings of $85 billion
over ten years. At the same time, the Republican Party continued to urge
the complete and total elimination of the Federal Estate Tax, which affects
roughly one percent of all taxpayers, and which even the offi cial estimate puts
as a revenue loss of $396 billion dollars over a similar ten-year period.

Public policy often favors one group over another. It is expected that

the dominant or governing political coalition will offer tax policies that
favor one group for a variety of reasons—from benefi ting political supporters
to notions of sound economic policy. The dominant or governing coalition

GETTING A POOR RETURN

can be defi ned by party or ideology. Generally, if one party controls the
elective branches of government, we can defi ne that party and its goals as the
dominant or governing coalition. Many times in American history the same
party has controlled the presidency and both houses, or at least one house,
of Congress. This assumes that each party is bound by certain core ideologi-
cal assumptions—for example, the nature of federalism and the respective
power of the state and federal government, or the extent and size of federal
regulatory authority over markets, working conditions, and income.

If the policy is improperly applied or enforced, the disfavored group

can always assert their rights in court. Audited low-income taxpayers claim-
ing the Earned Income Tax Credit can challenge the assessments by fi ling a
lawsuit against the Internal Revenue Service (IRS). The expectation is that,
as Justice Harlan wrote in his dissent in Plessy v. Ferguson,

“all citizens are

equal before the law. The humblest is the peer of the most powerful.”

One of those audited working poor seeking equality before the law is

Joy Anders, a day-care-center worker from Phoenix, Arizona. In one par-
ticular year, Joy earned $5,700, which she used to support herself and her
sixteen-year-old son and claimed an Earned Income Tax Credit of $1,500.
However, because Joy lived with her mother, whose sole source of income
was a pension paying less than $7,200 per year and therefore unearned,
the IRS rejected Joy’s claim to the Earned Income Tax Credit. The IRS
claimed that although Joy’s mother did not fi le a tax return, Joy’s son was
not a qualifying child because that child also lived with the grandmother
and the grandmother’s income was unearned. The IRS immediately assessed
an additional tax of $1,500, the amount of the credit, plus interest, and also
notifi ed Joy that penalties would accrue if the balance were not paid within
ten days. With the pro bono assistance of a local attorney, Joy fi led a claim
in Tax Court, and penalties and interest accrue as Joy tries to navigate
through the arcane meanings of a “qualifying child.”

Despite assertions of fairness and equality, evidence suggests that Joy

will have diffi culty proving her claim. Taxpayers do not do very well chal-
lenging assessments and in particular challenged Earned Income Credits
rarely survive court scrutiny. Perhaps this should not be surprising. The
dominant political coalition sees many earned income claims as fraudulent
and Robert Dahl argued many years ago that court rulings rarely confl ict
with the preferences of the majority political coalition. Modern research
largely supports this assertion. Recent scholarship has found, for example,
that the majority political coalition uses courts to accomplish goals that it
cannot achieve through the legislative process (see, e.g., Whittington 2005).
Despite this research, it is still the prevailing belief in this country that a
court is the one branch of government that protects the minority against
the power of the majority, and that courts exist to ensure “justice as fair-
ness” (Fogel and Hudson 1981).

TAX POLICY AND THE PURSUIT OF JUSTICE

This book examines these competing claims and beliefs about the

American legal system in the area of tax policy and tax enforcement. Taxes
and tax policy are perhaps the dominant domestic public policy issue of the
past twenty-fi ve years, taking only a backseat to post 9/11 national security
concerns, and there is a complex legal system with competing federal trial
courts that taxpayers can access to fi ght tax assessments. As one respected
policy analyst noted, throughout the 1980s there were more frequent and
detailed changes to the tax code during this decade than in any other pe-
riod in U.S. history (Steuerle 1991), and the pace of tax legislation has not
diminished with succeeding decades. Both Presidents Clinton and George
W. Bush’s initial major domestic policy proposals dealt with taxes. Taxes
and tax policy continued as a major focus of campaigns, policy debates, and
legislation. The role of courts is critical in ensuring fairness in tax policy
and acting as an institutional barrier against the power of government.

Taxpayers challenging tax assessments annually fi le over 30,000 cases.

Few tax cases are appealed and even fewer involve low-income taxpayers.
Most cases are disposed of at the trial level and there are competing courts
that litigants of all income levels use; the two most important are the Federal
District Court, the trial court of general jurisdiction within the federal courts
system, and the U.S. Tax Court, a specialized trial court created under the
Article I legislative power of Congress. These courts are the front line in
the guerilla war between taxpayers and the IRS and these are the courts
that low-income taxpayers like Joy Anders turn to for relief against the IRS
and the power of the political majority. Do these courts protect the rights
of the individual, particularly the low-income taxpayer, or do they enforce
dominant policy preferences? Do they infl uence the IRS to change its audit
behavior and focus less attention on the lowest income group of taxpayers?
Is there a difference in these courts in their decisions and in their relative
infl uence on the IRS?

This book attempts to answer these questions. I argue and demonstrate

that courts differ little from the national policy makers in their approach
to tax policy and tax enforcement and, in fact, the president and Congress
can use the courts to support their tax policy goals. Because of this, it is
unreasonable to expect low-income taxpayer to fare well in our court system.
To argue this premise, I examine the tax litigation process from the initial
decision of choosing the tax forum, through an analysis of the decision mak-
ing process in these competing courts, to an examination of the respective
infl uence and impact of these different tax forums.

The book shows that while fairness before the law might be a laudable

goal, the appointment process ensures that tax policy and tax enforcement
rulings by the courts refl ect the dominant political beliefs. Given the longer
tenure of the federal judiciary, it is quite possible that even if the national
coalition changes in the next few election cycles, which in turn could lead

GETTING A POOR RETURN

to changes in tax policy and tax enforcement, the judiciary will fall short of
ensuring the new coalition’s notions of fairness in the tax policy domain.

TAXES AND TAX POLICY

Taxes and the courts have always been intimately intertwined, and the state-
ments that begin this book from two famous judges represent the dichotomy
of the attitude that law, and, by extension, the judges who interpret and
rule on the law, have to taxes. Judge Learned Hand, one of the most re-
nowned of all appellate court judges, summarizes one attitude that the law,
and the public, has toward our tax burden: no one likes to pay taxes and
all taxpayers have the right to do whatever they legally can to minimize
their tax burden. Yet, as Oliver Wendell Holmes notes, taxes are the price
we pay for a civilized society. Minimization of one’s tax burden means fewer
resources for the government. Without income tax revenue, we would have
no army, navy, highways, or airports. We would have no FBI or CIA, or
even any disaster management agency. It is unrealistic to think that these
potentially divergent judicial attitudes are meaningless when it comes to
judicial rulings on taxes and tax policy or that judicial attitudes have not
played an important role in the formation and development of tax policy
almost since the inception of the income tax.

The fi rst income tax laws were enacted in 1861 and 1862 to fund

the Civil War.

These same acts created the IRS.

The Estate of Abraham

Lincoln was even issued a tax refund for an overpayment. After the need for
revenue to fund the Civil War diminished, these income tax laws expired.
The growth of the progressive movement, the increase in the size of the
national government, and the increase in U.S. involvement in international
affairs all lay in the future and the current customs duties were suffi cient to
fund the operations of the federal government.

Our story of the interaction of courts, national politics, and taxation

really begins in 1894, almost thirty years after the end of the Civil War.
Responding to dissatisfaction with high tariff rates and the need to raise
revenue, Congress relied on its powers granted by Section 8 of Article I
of the U.S. Constitution and passed a federal income tax for the fi rst time
since acts passed to fund the Union’s fi ght against the Confederacy. While
almost all in our modern world accept the constitutionality of the income
tax, the arguments for and against it echo much of the controversy one
hears today about taxation and these arguments have the same ideological
and partisan fi lters. Liberals and Democratic politicians supported the income
tax while Republicans and conservatives opposed it. For those in favor of
the income tax, it was seen as a crucial element of the progressive agenda,
which sought, among other things, fairness and equity in the collection
of revenue for the Unites States. Progressive advocates and Democratic

TAX POLICY AND THE PURSUIT OF JUSTICE

politicians attacked the tariffs as regressive and hurting the poor and lower
classes by increasing the costs of goods. Thus, Democrats viewed the income
tax as a progressive measure and a fairer way to raise revenue. Republicans,
however, argued that the income tax amounted to a socialist redistribution
of income. The income tax would sap initiative and punish the hard work
which leads to the accumulation of capital.

Unlike many in his party, Democratic President Grover Cleveland

personally opposed the imposition of the income tax on individuals (Witte
1985). However, because of the favorable and politically popular tariff
relief provisions attached to the legislation, Cleveland reluctantly signed
the income tax into law (Whittington 2005; Witte 1985). As part of the
tax legislation, banks were required to pay a two percent tax on income in
excess of $4,000. One such bank that had to pay an income tax was the
Farmers’ Loan Bank located in Massachusetts. In a contrived case, Charles
Pollock, a shareholder in the Farmers’ Loan Bank, sued to enjoin the bank
from paying the income tax. Eventually, the lawsuit, Pollock v. Farmers’ Loan
& Trust Co. (1895), reached the Supreme Court of the United States.

After two hearings, Chief Justice Fuller, writing for a scant 5–4 major-

ity, struck down the income tax as an unconstitutional direct tax. Among
other reasons, the Court focused on the part of the statute that included
rents from real estate as income. Land is subject only to direct taxation,
the opinion stated, and you cannot separate income from the land itself.
Therefore, the Court held, this provision violated the apportionment provi-
sions of the Constitution. One cannot apportion tax on land because such
a tax must be in proportion to the population.

Of course these bare and very questionable legal reasons hid the

underlying emotions of the case and views of the social desirability of the
income tax. Despite Chief Justice Fuller’s admonition in his majority opin-
ion that “we are not concerned with the question whether an income tax
be or be not desirable,” Joseph Choate, the lead attorney for the plaintiff,
attacked the law as “communistic” and “socialistic” (Hall, Finkelman, and
Ely, Jr. 2005, p. 385). Certainly other justices on both sides saw the decision
in ideological terms far outside of bare-bones legal reasoning, with some
commenting on the emotions and lack of logic in the respective opinions
of those with whom they disagreed.

Called the “most controversial case of its era” (Hall et al. 2005),

one scholar (Whittington 2005) has recently argued that the Pollock case
provides a strong example of how the dominant political coalition uses the
courts to achieve political goals that it cannot reach through legislation.
In this set of circumstances, because of the complications and unpopularity
of high tariffs, President Cleveland and the legislators in Congress opposed
to the income tax were forced into supporting compromise legislation that
enacted an income tax since it was the only political avenue open that

GETTING A POOR RETURN

would lead to a tariff reduction. Unhappy with the bundled income tax part
of the legislation, President Cleveland and those legislators opposed to the
income tax supported the effort to overturn the tax through the courts as an
unconstitutional tax. Thus, an attempt at progressive taxation was stymied
by the courts acting in concert with national political interests.

The decision met with great criticism and, even though the tariff

provisions were upheld, dissatisfaction with tariffs as a primary source of
revenue continued. Because of these circumstances, a coalition of newly
elected liberal Republicans and progressive Democrats began to seek legisla-
tive reenactment of an income tax. Under the guidance of then President
William Howard Taft, who was fearful of a constitutional crisis between
Congress and the Supreme Court, Congress agreed to a compromise. Congress
would propose a constitutional amendment specifi cally allowing an income
tax without direct apportionment based on population and Congress would
pass an excise tax on corporate profi ts but no tax on individual incomes.
With the compromise, Congress overwhelmingly passed an amendment to the
U.S. Constitution in 1909 permitting an income tax without apportionment
with a near unanimous vote in the House and a unanimous vote in the
Senate. State ratifi cation was slow, with mostly southern states supporting
it and many eastern states reluctant to vote for the amendment. Northern
states were far more industrialized, and the economy of the southern states
depended greatly on agriculture and farming. Northern states therefore feared
that the income tax burden would fall unevenly on the taxpayers of their
states and the southern states would pay much less.

Eventually, in 1913, the necessary thirty-sixth state ratifi ed the Sixteenth

Amendment, providing the constitutional basis for the right of Congress to
authorize the collection of a federal income tax.

The Sixteenth Amendment

overturned the 1895 Pollock decision, and this remains one of the few times
when a constitutional amendment has overturned a constitutional ruling
of the Supreme Court.

The ratifi cation coincided with major Democratic

victories in the 1912 election, including the election of President Woodrow
Wilson, the fi rst Democratic president since Grover Cleveland. This new
more progressive Congress then enacted income tax legislation in the same
year as ratifi cation of the Sixteenth Amendment, upending President Taft’s
compromise position.

This fi rst income tax was, by any modern standard, extremely mod-

est and of limited reach. The top rate for all taxpayers was six percent for
incomes over one-half of a million dollars ($500,000), an enormous fi gure in
1894 America and the equivalent income in excess of $11 million in today’s
dollars. Only about two percent of the labor force fi led tax returns during
the fi rst two years of the new statute’s existence (Friedman 2002). Even
under a revised income tax law enacted in 1916, the fi rst $3,000 ($4,000
for married couples) of income was exempt, an income equivalent today of

TAX POLICY AND THE PURSUIT OF JUSTICE

almost $54,000 for single people and over $71,000 for married couples. The
upper rate increased to thirteen percent, but only on incomes in excess of
$2,000,000 per year; only the very few wealthiest would ever pay this rate,
and most paid no tax at all.

Even so, this modest income tax led to signifi cant opposition. The

most affl uent Americans and the business community were the most out-
spoken in their opposition to the income tax. The wealthy saw the fairly
small tax as a stalking horse for far greater income redistribution. The fear
was that there would be a call for increased income taxation of wealth and
eventually, to paraphrase Joseph Choate, a progressive income tax would
lead to socialism.

While socialism has never occurred, not all these fears regarding increased

rates of taxation were misplaced. Just one year later, to fi nance World War
I, the government dramatically increased income tax rates and also imposed
an excess profi ts tax on business. The maximum rate increased to seventy-
seven percent, and from almost no infl uence on federal revenue prior to
1916, individual and corporate tax income accounted for sixteen percent of
federal revenue. Between 1917 and 1920, revenue from these income taxes
constituted almost sixty percent of all federal revenues. The tariff, once the
centerpiece of federal revenue, now saw its importance to the economy and
public policy decline in dramatic fashion. The revenue and importance of the
income tax dwarfed the money generated by, and the consequence of, tariff
collection. With this shift in importance came both signifi cant opposition
and tax evasion (or the more preferable term of tax avoidance), along with
of course the use of courts to settle taxpayer and government disputes.

With the end of World War I and a coming decade of Republican

presidents and Republican majorities in Congress, both tax rates and the
progressive nature of the income tax declined. Even Democrats led by outgo-
ing President Wilson questioned the need for such high maximum rates and
thus the effective top tax rate was reduced to twenty-four percent. Although
a push was made by many wealthy businessmen to eliminate the income tax,
Secretary of Treasury Andrew Mellon, no fan of disproportionate tax rates,
convinced enough Republican offi ceholders that some degree of progressive
taxation was socially responsible and benefi cial for political purposes.

With the support of a Republican administration, the income tax was

now a permanent fi xture of American political and economic life. Between
the turn of the century and 1925, total internal revenue collections from
income tax grew from $207 million to $3.2 billion. By comparison, customs
duties climbed only from $185 million to $464 million during the same
period (Chommie 1970; Witte 1985). The Bureau of Internal Revenue grew
along with its collections. The number of employees increased from 4,000
in 1913, a number that had remained almost constant for more than half
a century, to close to 16,000 by 1920.

GETTING A POOR RETURN

However, what the top and bottom rates should be, whether and to

what extent the tax should apply to corporations as opposed to individuals,
whether and when excess profi ts should be taxed, whether estates would be
taxed, and what constitutes income, among other questions, would remain
to be argued about and ruled on over the coming decades. In fact, just nine
years after the passage of the Sixteenth Amendment and the subsequent
constitutionally permissible income tax legislation, the fi rst book to help
taxpayers avoid excess taxation was published. It was entitled Minimizing
Taxes, and written by Wall Street lawyer John Sears. Foreshadowing Judge
Learned Hand, Sears wrote his book from the point of view of the tax-
payer, offering advice on how to minimize taxes legally, which in Sears’s
view amounted to the patriotic goal of “exposing evils in the system.”

paraphrase Sherlock Holmes, the game was afoot.

In the decades that followed, tax policy continued to be a function

of “revenue demands and ideology” (Witte 1985, p. 96). Even though the
critical election of 1932 brought Franklin D. Roosevelt and solid Demo-
cratic majorities power, there was no immediate push for more progressive
tax rates. Throughout the Depression, lower maximum income tax rates
meant that relatively few individuals paid any income tax, although the
1930s did see the imposition of the Social Security tax, bitterly fought
by business, but a regressive tax on individuals. The Supreme Court with
a narrow majority

upheld the Social Security tax. Income tax revenues

accounted for about forty percent of all federal revenue. Once again, war
came. Just as World War I changed assumptions about the need for govern-
ment revenue and the appropriate source for such revenue, World War II
would also lead to a signifi cant increase in the need for revenue and alter
the nature of income taxation.

With this need for revenue and full employment led by massive govern-

ment spending to both fi nance and support the war effort, tax rates increased
and more individuals were subject to the income tax than ever before. A
surtax was imposed on the income tax that started at a rate of thirteen
percent on the fi rst $2,000 of income and went up to eighty-two percent
on incomes over $200,000. In addition, the federal government imposed a
guaranteed collection mechanism that had the additional benefi t of ensuring
a steady stream of tax revenue throughout the year. Income tax withholding
was introduced for the fi rst time. More than one-third of Americans now
paid some form of income tax and the tax was by far the dominant source of
governmental income. By 1945, about 45 million American paid an income
tax out of a total population of some 132 million people, and income taxes
accounted for over seventy percent of all federal revenues.

In the immediate aftermath of the war, income and corporate taxes

were roughly equal; however, during the succeeding decades, the individual
income tax began to dominate corporate revenue, as more and more earners

TAX POLICY AND THE PURSUIT OF JUSTICE

became subject to the income tax. Although some attempts were made to
reduce taxation following World War II with a brief Republican takeover in
Congress, Truman’s reelection and the revenue needed to fi ght the Korean
War forestalled any signifi cant change in tax rates and the importance of
the income tax. President Eisenhower resisted tax reduction throughout his
term of offi ce and, with the passage of the Internal Revenue Code (IRC)
of 1955, tax rates stabilized.

Another change during these years was the increase of individual

income tax revenue compared to corporate tax revenue. From the 1920s
through World War II, tax collections from corporations and individuals
were equal. By 1955, however, individual income taxes provided $31.6
billion of income tax revenue, while corporations provided $18 billion.
The disparity between corporate and individual taxes continues to this day
(Witte 1985, p. 124).

The ensuing decades saw movement back and forth between cutting

taxes as a stimulus to the economy and raising them to fi nance the Viet-
nam War and new government social and economic programs. Considerable
government expansion led to both a much greater need for revenue and
collection and tax practices that signifi cantly increased federal revenue. By
1968, the IRS was collecting $78 billion from individuals and $30 billion
from corporations. By the mid 1990s, individuals paid almost $600 billion in
income taxes and corporations more than $150 billion and the top income
tax rate at one point climbed to fi fty percent.

Of course, expansion of government services and spending and the

subsequent expansion of income tax rates and income tax collections led
to increases in opposition to such government spending and tax collection
rates. By 1980, due to infl ation outpacing income gains, many American
taxpayers were paying increasing tax rates without any corresponding real
increase in income or buying power. A tax revolt was imminent and voters
in both Massachusetts and California passed initiatives that capped property
taxes. A similar movement was under way to reexamine and lower income
tax rates.

A rejection of high-income rates became one of the cornerstones of the

presidential campaign of the Republican candidate Ronald Reagan. With his
election came Republican control of the Senate, and in his subsequent presi-
dency Reagan proposed and Congress adopted a tax package that decreased
top rates soon after he took offi ce. This was not the end of his presidency’s
altering of tax rates. With bipartisan support, Congress enacted the 1986
Tax Reform Act, which, although revenue neutral, led to a large decline in
progressive tax brackets, reducing over fi fteen brackets to two.

However, once again income taxes and the controversy as to the

proper method and rate of collection resulted in additional changes during
the ensuing decade. Concern about large federal defi cits led both the George

GETTING A POOR RETURN

H. W. Bush and Bill Clinton administrations to once again increase tax
rates and tax collections. For example, a third tax bracket was created and
top rates climbed again to almost forty percent. George W. Bush ran, in
part, in opposition to the imposition of more taxes and promised to reduce
government spending and lower maximum tax rates. With his election,
some of his initial legislation dealt with taxes. Thus, in the second Bush
administration, maximum tax rates again declined and the Estate Tax was
signifi cantly reduced with a goal of eventual elimination.

COURTS, FAIRNESS, AND TAXES

Millions of tax returns are fi led each year and individual income tax col-
lection is by far the single most important source of all federal government
revenue. With so much money collected from so many people, signifi cant
confl ict between individuals and the U.S. taxing authority, the IRS, becomes
inevitable. In a typical year, taxpayers fi le over 100 million individual tax
returns. The individual taxpayer initially determines his or her tax liability
by calculating income and deductions and then fi ling a tax return to the
appropriate IRS offi ce, along with the amount of tax payment due to the
government.

The taxpayer is expected to comply with the tax laws and

honestly report income, exemptions, and deductions. This is our system of
self-assessment, one called “quasi voluntary” (Daily, 1992 p. 1/7; Freeland,
Lind, and Stevens 1977, p. 971).

Increasing rates and growing complexity lead to more incentives for

avoidance and evasion and a greater need for planning. Tax cheating, in the
form of outright noncompliance, is a severe problem. Many sources, including
scholars, the IRS, and journalists, estimated that, in the 1970s and 1980s,
noncompliance resulted in an unreported taxable income shortfall between
$61 and $80 billion per year (IRS 1979; Kurtz and Pechman 1982). By 1986,
Roth, Scholz, and Witte (1989) note that the IRS estimated that individuals
failed to report between $70 and $79 billion in income received, and that
the fi gure might in fact be closer to $100 billion. Philip Brand, the IRS’s
chief compliance offi cer, verifi ed the $100 billion fi gure for the tax year
1994.

This amount is nearly twenty percent of all reported income.

When income is underreported, the IRS collects less tax. The lack of

compliance represents a severe loss of revenue for the government. Because
of these problems, the IRS will closely examine a certain number of returns
in a procedure known as an audit. It is the most powerful policy tool the IRS
possesses (Burnham 1989; Roth, Scholz, and Witte 1989; Scholz and Wood
1998; Steuerle 1986). In its simplest form, an audit is a detailed examina-
tion of a taxpayer’s income tax return.

The audit seeks to determine if the

taxpayer is actually telling the truth in the claims made on the income tax

TAX POLICY AND THE PURSUIT OF JUSTICE

return. The IRS does not accuse the taxpayer of wrongdoing, but instead
seeks justifi cation and verifi cation of the listed income and expenditures.

Taxpayers seek to avoid paying excess taxes and the IRS seeks to

collect all taxes due and thus audits a selected number of returns both to
determine individual tax liability and to try to ensure overall compliance.
As one scholar of American law noted, “tax avoidance became a national
pastime . . . the Internal Revenue Service and the taxpayers became locked
in a kind of dance, a rhumba of avoidance and counteravoidance” (Fried-
man 2002, p. 395). Because of this dance of confl ict, the taxpayer and the
IRS often fail to resolve the matter to the satisfaction of one or the other,
usually the taxpayer. When this course and all appeals fail, the taxpayer
can turn to the courts to settle the disputes and courts become a major
player, not just in the individual dispute, but in the overall formation and
enforcement of U.S. tax policy.

Perhaps we can call the courts the dance instructor or the referee of

the rules of the dance. The Supreme Court has issued opinions on major
tax cases involving more than 200 tax disputes since the beginning of the
Warren Court (Epstein et al. 2003), more than any domestic policy domain.
The greatest percentage of Supreme Court cases during the Warren Court
years involved controversies over the IRC. During the Burger and Rehnquist
Courts, IRC controversies accounted for the third and second highest per-
centage of cases (Epstein et al. 1992, pp. 553–54).

Tax cases also represent

a very large part of the docket of the U.S. Court of Appeals. For example,
during the years 1997 to 2004, the Circuit Courts of Appeals handled, on
average, more than 220 new tax case appeals per year.

This heavy Supreme Court docket of tax cases is remarkable consider-

ing the assertion of one scholar that the Supreme Court is reluctant to take
such cases because of the justices’ lack of interest and the issue complexity
(Perry 1991). Among major cases, the Supreme Court has determined what
constitutes income,

acceptable business and personal deductions,

and taxable

consequences of divorce.

All of these represent claims by the IRS follow-

ing post audit assessments; they pitted one party against the IRS, and hence
against the government of the United States. Although the taxed parties were
concerned with their individual liability, the holdings of the Supreme Court
have signifi cantly shaped and determined tax law and tax policy.

However, despite these raw numbers and major cases, most matters are

disposed of at the trial level, and, unlike other legal areas, the tax domain
uses a system of competing courts. These courts allow litigants of all income
levels to assert their claims against the IRS. The two most important of these
tax trial courts are the U.S. Court, the trial court of general jurisdiction
within the federal court system, and the U.S. Tax Court, a specialized trial
court created under the Article I legislative power of Congress. These two

GETTING A POOR RETURN

courts, along with the U.S. Court of Federal Claims, are the courts that
decide the winners of the battles between the IRS and taxpayers and these
are the courts that low-income taxpayers turn to for relief against the IRS
and the power of the political majority. Given the particular prominence of
taxes and tax policy over the past three decades, where tax litigants choose
to sue, how these tax trial courts make decisions, and the impact of these
decisions are critically important if one wants to understand the dynamics
of not just tax policy, but public policy. Courts need litigants to bring a
matter to them before they have a chance to rule on any particular matter
and these lawsuits can change public policy by forcing the agency to expend
time and resources fi ghting the litigation and then having to comply with
the directives of these courts.

The balance between legal avoidance as advocated by Judge Learned

Hand and illegal tax evasion, those seeking to avoid what Oliver Wendell
Holmes called the “price we pay for a civilized society,” is the point when
courts step in and rule either in favor of the taxpayer or in favor of the
United States. Taxpayers have a right to pay as little as legally possible,
but if the taxpayer does not pay what is truly owed, “the price we pay for
a civilized society,” then all the other taxpayers bear this burden and the
noncompliant taxpayer becomes the free rider enjoying the benefi ts, but not
the burdens, that federal spending bestows.

Most cases that confront the trial courts present few diffi cult issues or

problems. Often a small number of facts are in dispute and the cases involve
few, if any, complex legal issues. However, many are not and the courts have
to pick a winner and a loser with often devastating consequences to the
taxpayer if the courts rule for the government. Often, however, regardless
of how the courts rule, these decisions have other consequences. If courts
consistently rule in favor of one type of one income level of taxpayer as
opposed to another income level, or rule in favor of corporations as opposed
to individuals, those decisions become part of the tax policy of the govern-
ment and infl uence the ideals and perceptions of justice, fairness, and the
notion that all citizens are equal before the courts. How courts decide these
matters is the fulcrum on which we attempt to answer these questions. How
they decide these cases has much to say about how and why we treat the
taxpayer who uses the Earned Income Tax Credit and the taxpayer who
employs the services of top accounting and law fi rms.

CHAPTER OVERVIEW

Those looking for debate and analysis of the legal arguments over the con-
stitutionality, interpretation, and meaning of tax law will be disappointed.
This is neither a textbook on the fundamentals of income taxation, nor a
demonstration on how to survive a tax audit, nor advice on how to prepare

TAX POLICY AND THE PURSUIT OF JUSTICE

and win a court challenge. How a court should interpret a tax law—and
the tax implications of the various interpretations—are properly the subjects
of income tax textbooks and the domain of tax professors, tax lawyers, and
those who provide income tax advice and assistance. These are obviously
important topics for analysis and investigation, but this book is about the
politics of courts and taxes, and whether courts deviate from the majority
or enforce dominant beliefs in their rulings.

After the introduction, the chapters proceed down the path of an

audited taxpayer. The chapters use the audit experiences of Joy Anders,
along with those of a large multinational corporation, a wealthy taxpayer
who invested in a tax shelter, and a middle-class couple whose deductions
were challenged as the counterpoints for the examinations and analyses to
follow. Chapter two provides an overview of the choices these types of tax-
payers have in challenging a post audit assessment by the IRS. This chapter
examines the trial options available to taxpayers, which are fi rst internal
appeals and then appealing the audit assessment either in the Tax Court,
the District Court, or the Court of Claims. While most of the chapter is
descriptive, I do provide an examination of the choices as well as the costs
and benefi ts of the particular trial forums. Ideally, the system should work
to ensure that all taxpayers, regardless of wealth, have a fair and impartial
forum to contest their assessments. However, arguably the least costly option
might offer the smallest chance of success, calling into question the fairness
of the options and ultimately the effect on compliance.

Chapters three, four, and fi ve are the heart of the book and contain

the analyses of this dance of courts, the IRS, and these taxpayers. Chapter
three analyzes in detail the decisions of each of the potential litigants in
choosing fi rst whether to sue and then selecting a forum. To do this, I fi rst
describe theories of litigation, including rational actor theories and informa-
tion asymmetries, and use these theories to examine both the decision to
sue the IRS and, once that choice is made, analyze forum choice. Then I
gather data from the years 1994 to 2000, which coincides with the Contract
for America campaign through the Republican takeover of Congress in 1995
and fi nishes with the last year of President Clinton’s term of offi ce.

The chapter argues that tax policy has been politicized and ideo-

logically divisive since the 1970s, with conservatives generally opposed to
greater tax collection and enforcement. Therefore, as the political majority
became more conservative over this time period, litigants were encouraged
to challenge the assessments and thus one should fi nd a corresponding in-
crease in tax litigation in both the Tax Court and the District Courts. In
addition, litigants should also be encouraged to choose the court that offers
the greatest chance of a conservative judge because the more conservative
the judge, the more likely the support for the taxpayer opposing the IRS.
The effect on the specifi c taxpayers is then examined.

GETTING A POOR RETURN

Chapter four describes the decision making of the two principal courts

in which to challenge the IRS, the Tax Court and the U.S. District Courts.
This chapter uses data from 1996 and 1997 to systematically examine the
differences, if any, in the decision making of these courts. The chapter scru-
tinizes the specialized Tax Court and its expertise and sees if that leads to
differences with the District Court in the use of ideological preferences in
the outcomes of the tax cases, specifi cally focusing on judicial independence
and congressional control. The chapter argues that the limited-tenure U.S.
Tax Court uses its expertise, and lack of realistic structural or hierarchical
constraints, to decide cases more in accordance with the Tax Court judges’
personal policy preferences than do the judges of the U.S. District Courts.
The chapter concludes with an assessment of each taxpayer’s probability of
success in each court.

Chapter fi ve examines the aftermath of these decisions and the infl u-

ence of courts on tax policy and tax enforcement. Specifi cally, beyond the
immediate impact on the taxpayer, do these aggregate decisions change IRS
behavior? Do they lead to greater auditing of Joy Anders and less examina-
tion of the large corporation? Do these decisions instead do the reverse and
change IRS audits to benefi t poorer taxpayers? Several studies have shown
that national agency policy can change as the ideology or partisanship of
the federal district or federal appellate court changes. Some agencies must
contend with different federal courts with overlapping jurisdiction, and cal-
culate responses to these different courts that have different judges, different
ideologies, and different agendas while trying to carry out the preferences
of the executive and legislative branches of government in a separation of
powers system. This chapter applies these theories of infl uence and impact by
examining the tax trial courts and the IRS. Using data from 1994 through
2000, and studying audit ratios of individual and corporate taxpayers, I
fi nd that the IRS pays attention to the preferences of the executive and
legislative branches of government and the preferences of the Tax Court,
but not the District Courts, in determining the ratio of audits of individuals
and corporations, with the IRS responding to more conservative courts and
more conservative executive and legislative preferences by shifting the audit
rate toward auditing more individual taxpayers. The smaller variation in Tax
Court ideology lessens the overall impact of the Tax Court.

Finally, chapter six offers a summation of the fi ndings of this book

and what these fi ndings mean for tax policy, public policy, and the role of
courts in a democracy. It is my hope that this book will accomplish two
goals. First, it will underscore and demonstrate the importance of tax trial
courts in setting and determining the nation’s tax policy. Then I hope to
add to the body of literature that shows that courts really represent and
make rulings consistent with the preferences of the majority. They are not

TAX POLICY AND THE PURSUIT OF JUSTICE

out of control, but they do not really constitute the last best hope for the
disenfranchised. In the end, perhaps both the independence and the non-
democratic or antimajoritarian nature of federal courts are overestimated by
friends and critics alike.

This page intentionally left blank.

COURTS AND THE IRS

TWO

COURTS AND THE IRS

Taxpayers have several choices to consider following a post audit tax as-
sessment. The choices range from administrative appeals, not paying and
then suing in the U.S. Tax Court, or paying and suing in either the U.S.
District Court or the U.S. Court of Federal Claims. The choice depends on
calculations of the cost and benefi ts of each particular forum and for that
analysis each taxpayer has to ask and answer several different questions.

For example, does a low-income taxpayer have a realistic chance

for success in her dispute with the IRS? For that matter, does this type of
taxpayer really have any choice in venue, given fi nancial circumstances? If
a taxpayer can afford to pay the assessment, is there still a reason to sue in
the Tax Court? What advantages, if any, does the Court of Federal Claims
have over the District Court? What alternatives are open to a large multi-
national corporation to challenge a post audit assessment? Does it make a
difference if the assessment is several million dollars or several thousand?
Does it make a difference if the challenge is premised on a questioned ac-
counting practice as compared to a challenged deduction? Does it make a
difference if one can retain counsel?

A poverty-level taxpayer challenging an audit of the Earned Income

Tax Credit and a large corporation seeking to defend a questionable ac-
counting practice represent two extremes. Most taxpayers and most audited
taxpayers fall between these two types. The IRS often challenges the deduc-
tions claimed by a taxpayer or the failure of a taxpayer to report taxable
income. From the point of view of most taxpayers, these claimed deductions
or omissions of reported income are undoubtedly legitimate methods of
avoiding unnecessary tax liability. For example, on the advice of his lawyer
and accountant, an affl uent taxpayer invests in a tax shelter to reduce his
tax liability. However, the IRS challenges the shelter. The investor sought
to lower his tax burden, but now faces a post audit assessment of over
$100,000. What about a middle-class or upper-middle-class couple who fi nds

GETTING A POOR RETURN

various deductions on their 1040 Schedule A itemization form challenged?
Imagine if this couple now owes $25,000 to the government. Do both types
of taxpayers sue? If so, what forum choices are available? Why do taxpayers
have so many forum choices and do all these choices have anything to do
with taxpayer compliance and acceptance of our tax system?

There are names one can attach to these taxpayers. Joy Anders is the

taxpayer who challenges the denial of her claim to the Earned Income Tax
Credit. Wal-Mart, the retail giant, defended its accounting practices in the
Tax Court as did tax shelter investors such as James William. Mr. and Mrs.
David Forrest fi led a claim in the U.S. District Court to uphold disallowed
deductions. While all have, at least theoretically, multiple options, each had
reasons for the forum choice. The analyses in the following chapters deal
with aggregate choice, decision, and impact. However, poor, middle-class, and
rich taxpayers, small business and large corporations, and estates and trusts
make decisions each and every day and must live with their consequences.
The accumulation of these individual preferences leads to the aggregate
impacts described in the succeeding chapters.

This chapter provides an overview of the choices these taxpayers and

others like them have in challenging a post audit assessment by the IRS.
Each choice has costs and benefi ts and, while each particular situation is
unique, one can examine some general factors that are taken into account by
each type of taxpayer. The chapter considers these choices from the initial
administrative appeal through each particular trial court. The conclusion
ponders the meaning of these choices in assisting fairness and justice. Ideally,
the forum choice system should work to ensure all taxpayers, regardless of
wealth, an appropriate forum. Ultimately, a sense of fairness can also lead to
greater compliance in a system that depends on voluntary reporting of tax
liability. Does the most affordable option offer the smallest chance of success?
If so, it calls into question the fairness of the choices and compliance.

THE NEED FOR TAX FORUMS

In a typical year, taxpayers fi le over 100 million individual tax returns under
a system of reporting called “quasi voluntary” (Daily, 1992 p. 1/7; Freeland,
Lind, and Stevens 1977, p. 971). This reporting system assumes that the vast
majority of taxpayers do in fact comply, but the quasi voluntary assessment
places a tremendous burden on the taxpayer and sometimes mistakes oc-
cur through error and sometimes through deliberate dishonesty. To counter
these mistakes, the IRS examines a certain number of returns for error in
a process known as an audit.

Trained IRS offi cials conduct the audits out of district offi ces located

throughout the fi fty states. The IRS uses numerous procedures and formulas
to determine whom to audit and many different types of audits. The most

COURTS AND THE IRS

common type is the correspondence audit. This type of audit usually occurs
because of computer matching discrepancies. For example, a taxpayer might
fail to include income that the IRS fi nds on a corresponding 1099 form.
Most often, the taxpayer will receive a notice of the adjustment and, unless
contested, there is little more for the taxpayer to do then sign the correspon-
dence form and remit the additional tax owed. The IRS can be wrong and
the taxpayer can send appropriate documents justifying the initial return.

The next type of audit is known as the offi ce audit. This type of au-

dit is usually conducted by less experienced examiners and again involves
discrepancies in the tax return. Similar to the correspondence audit, here
the taxpayer has to prove, through documentary evidence, reported income
and claimed deductions.

The fi eld audit is the most extensive audit and is conducted by the

most experienced tax examiners. In this procedure, the auditor goes to the
offi ce or home of the taxpayer to examine the books and records, although
the taxpayer can bring the requested documents to the IRS offi ce. Usually
this will result in greater tax liability.

The IRS uses many different processes for selecting a return for one of

these types of audits. Perhaps the most well known is the statistical analysis
known as the DIF scoring system. The DIF or discriminate function system
is computed by a comparison of tax returns of the same type or category
and specifi cally singles out returns likely to produce more revenue after the
audit (Johnston 1996). The IRS does not reveal the exact formula, but seeks
outliers in similar income categories in which the deductions do not match
or equal similar returns. For example, a sole proprietor who fi les a Schedule
C return and claims an inordinate number and amount of deductions might
fi t the DIF profi le. One researcher argues that the DIF is unlikely to activate
if a taxpayer claims less than forty-four percent of adjusted gross income in
Schedule A deductions or no more than fi fty-two percent of revenue for a
Schedule C fi ler (Johnston 1996). A supplemental program started in 2002
and known as UI DIF also calculates the probability of unreported income
on a tax return (Shafi roff 2004). In addition, the IRS uses system matching
systems, special programs that target particular occupations or businesses,
evidence of criminal activity and informant tips, tax returns that have had
prior trouble, and sometimes random audits.

The matching or information reporting program is another computer-

based program the IRS uses to select returns for audits based on discrepancies
between reported income and deductions and other information that the
IRS has received on the taxpayer that suggests irregularities. As suggested
above, a taxpayer might not report the income shown on W-2 and 1099
forms submitted to the IRS with the taxpayer’s Social Security number.

Other programs that lead to audits are tips from informants, criminal

activity, targeted programs, and random auditing. For example, if the police

GETTING A POOR RETURN

break up a drug ring, the IRS can examine returns of those involved and
assess tax liability for the unreported, albeit illegal, income. Certain profes-
sionals such as lawyers and professors are often targeted through programs
because of the likelihood of their use of tax avoidance. Finally, random
audits are also used. Through a program originally known as the Taxpayer
Compliance Measurement Program and since replaced with the National
Research Plan audits, randomly selected people and business are required to
prove every detail on their tax returns. The IRS then uses this information
to create its DIF score. If the audit shows a defi ciency, then the agency can
assess interest and penalties in addition to the defi ciency. All enforcement
agencies are prone to make errors. The IRS confronts almost countless daily
decisions on how to interpret the complexities of tax law and devise rules
and procedures to collect taxes. If their decisions are too lenient toward
potential tax evaders, the amount of tax collected would fall and reduce
the revenue needed to fund the operations of the U.S. government. If the
decisions are too stringent, they impose unnecessary burdens on taxpayers.
Where to draw this line is a matter of judgment and it is these judgments and
the subsequent court treatment that can affect the level of compliance.

Drawing this line is not easy, and evidence exists showing that a

signifi cant proportion of IRS post audit claims are incorrect. Congressional
hearings in 1997 and 1998 focused on several instances of IRS abuse in
auditing and collection, although many of the worst alleged abuses were
never proven (Johnston 2000). One individual who has studied the IRS
for several years, David Burnham (1989), argued that IRS auditors who
spend their lives working for the IRS develop a rather peculiar view of the
world and hold two particular convictions. The fi rst is that any individual
who comes to their attention is guilty. The second is the perception that
their duty is to assert and maintain the position of the IRS on all issues.
Given these perceptions, perhaps it is not surprising that a 1979 General
Accounting Offi ce study found IRS examiners made technical errors in more
than one-third of the cases, and that in over sixty percent of the cases, the
agents failed to follow agency procedures.

Improper auditing does not minimize the problem of tax compliance.

In fact, in subsequent hearings, the IRS argued that the rules instituted to
curb auditor abuse hampered collection and led to unwarranted dismissals
of agents (Johnston 2000). However, when improper auditing, actual or
perceived, is combined with the verifi ed lack of tax compliance and then
added to partisan and ideological confl ict over tax policy, the IRS–audited
taxpayer relationship is bound to be hostile and adversarial. The need to audit
combined with the potential for abuse and differing policy interpretations
inevitably lead to disputes between the agency and individual taxpayers.

When confronted with a post-audit assessment, taxpayers can pay, ap-

peal, or sue in response. Many take the last course. The IRC is a litigated

COURTS AND THE IRS

area. In 1992, for example, taxpayers fi led 2,296 cases in the Federal District
Courts and 30,345 in Tax Court (MacLachlan 1995 p. 16). A signifi cant
percentage of Supreme Court cases during the Warren Court years involved
controversies over the IRC. These controversies accounted for high per-
centages of all cases on the docket during the Burger and Rehnquist years,
respectively (Epstein et al. 1992, pp. 553–54). This heavy Supreme Court
docket of tax cases is remarkable considering the assertion of one scholar
that the Supreme Court is reluctant to take such cases because of the jus-
tices’ lack of interest and the issue complexity (Perry 1993). It is also the
heaviest litigation load of any federal agency. For example, in comparison,
the Securities and Exchange Commission (SEC) reported fi ling 1,605 ad-
ministrative complaints and 413 civil suits, for a total of 2,018 actions from
July 1, 1999, to June 30, 2000, while the Federal Trade Commission (FTC)
reported a total of 107 actions in the U.S. District Court for one quarter
of the 2006 fi scal year.

ADMINISTRATIVE APPEALS AND TRIAL COURTS:

OPTIONS AND ODDS

Internal Appeal

Prior to receiving a formal demand letter, the taxpayer can request an
informal conference to resolve the matter with the local IRS offi ce. If no
resolution is reached, then the IRS mails the taxpayer a thirty-day notice
to pay the tax due. The letter contains the agent’s report and the basis of
the proposed adjustments. Following receipt of the letter, the taxpayer has
thirty days to protest and request a conference in the district director’s of-
fi ce. If no resolution is reached at this level, then the taxpayer can request
an internal appeal through what is known as the appellate division of the
IRS. The taxpayer does have the right to skip the district conference and
go directly to the appellate conference. Often the IRS encourages this,
particularly if the matter is complex. The appellate conference offi cer has
direct authority from the commissioner of Internal Revenue to settle cases.
If there is no agreement following this appeal, the taxpayer is given a statu-
tory ninety-day notice of defi ciency.

If the internal appeals process fails to

produce a settlement, the next step is litigation.

Litigation Options

Each taxpayer then has the option of various trial courts—the “S” op-
tion, a small claims and lower cost option within the U.S. Tax Court,
the Tax Court, the U.S. District Court, and the U.S. Court of Federal
Claims. There are several structural and institutional differences between

GETTING A POOR RETURN

these courts. For example, District Court judges are given full judicial in-
dependence with lifetime tenure, and their salary is guaranteed while the
judge is on the bench.

In contrast, the Tax Court and Court of Federal

Claims judges have fi xed terms and lack a salary guarantee. The Court of
Federal Claims hears monetary claims against the United States. Congress
creates specialized courts, including the Tax Court, for several important
reasons—complexity of issues, a need for economic and business-related
consistency, similarity of issues, and sheer volume of cases. Each has its
particular strengths and weaknesses.

Tax Court

By far the most common forum to litigate is the Tax Court, a specialized
tribunal designed to handle defi ciency claims. After the passage of the
Sixteenth Amendment, the District Court was the sole venue for challeng-
ing tax assessments. However, in 1924, nine years after the passage of the
amendment, Congress created the Board of Tax Appeals to handle taxpayer
litigation. It was renamed the Tax Court in 1942. Originally an independent
part of the executive branch of government, Congress passed legislation in
1969 removing the court from the executive branch and establishing its
independence as an Article I constitutional court.

In 1974, the court moved

out of the space it shared with the IRS and received its own building.

The Tax Court is an example of a court with limited jurisdiction and

limited independence created through the legislative power of Congress under
Article I. There are nineteen full-time judges appointed by the president and
subject to confi rmation by the Senate. Each judge serves for fi fteen years,
although, as a practical matter, if they so desire, they may be reappointed
either to an additional term or allowed to remain on the court to decide
cases on senior status. There are also several special judges appointed by
the chief judge of the court. The full-time judges enjoy full federal pension
and retirement benefi ts and draw the same salary as District Court judges
(see generally Dubroff 1979, part 1).

To sue in Tax Court, a taxpayer fi les a claim in the local district

court offi ce, and then the Tax Court randomly assigns the case to one of
the judges sitting in Washington, DC. The assigned judge will then try the
case in the taxpayer’s local area or in Washington, DC, depending on the
request of the taxpayer. No prepayment is required, and all cases are tried
without a jury. The losing party can appeal the decisions to the Appellate
Circuit in the federal district within which the claim was initially fi led. The
proceedings are conducted under the federal rules of civil procedure and the
federal rules of evidence. The IRS is defended by attorneys who come from
the Offi ce of the Chief Counsel of the IRS.

COURTS AND THE IRS

If the claim is for less than $50,000, the proceeding is a small claims

or “S” case. The taxpayer fi les a claim stating the reasons for protesting the
assessment, serves it to the IRS, and is then notifi ed of a trial date. Costs
are at a minimum with low fi ling fees, limited discovery, and often pro se
(taxpayer representing him or herself without counsel) litigants. Trial dates
are typically within six months of fi ling, and decisions are rendered within
a year. The average trial lasts less than two hours (Daily 1992, pp. 104–5).
Procedural and evidentiary rules are simplifi ed, and the rules mandate that
the trial judge “ensure substantial justice.” There is no right of appeal, and
the decisions have no precedent value.

If the matter is more than $50,000, the taxpayer still does not pay

the disputed amount, but the backlog to hear such cases is signifi cant. It
can take much more than one year for a decision. Each side preserves the
right of appeal. Since formal procedural and evidentiary rules are used, ef-
fective presentation of the taxpayer’s case almost demands the services of a
skilled advocate. The trial can take several days, although presumably since
the Tax Court judge is tax specialist and because there is no jury, the par-
ties will spend less time educating the judge about the particular tax issue.
The specialized knowledge of the judge also means there is less information
asymmetry between the parties and the court in the Tax Court than in a
court of general jurisdiction.

U.S. Court of Federal Claims

The U.S. Court of Federal Claims is another Article I court. This court
consists of sixteen judges nominated by the president and confi rmed by the
Senate for terms of fi fteen years. The Court of Federal Claims is authorized
to hear money claims against the United States. Approximately one-quarter
of the cases before the Court involve tax refund suits.

In 1982, the Federal Courts Improvement Act established the U.S.

Court of Federal Claims, formerly the U.S. Claims Court, as the successor
to the trial division of the Court of Claims, which had been in existence
since 1855. The court’s headquarters are in Washington, DC, but cases are
heard at other locations convenient to the parties involved. The court has
nationwide jurisdiction over a variety of cases, including tax refunds, federal
taking of private property for public use, constitutional and statutory rights
of military personnel and their dependents, back-pay demands from civil
servants claiming unjust dismissal, persons injured by childhood vaccines,
and federal government contractors suing for breach of contract.

Similar to the procedure for claims fi led with the Federal District

Court, the taxpayer pays the disputed assessment and sues for a refund.
Because these judges are usually not tax specialists, they lack the specifi c

GETTING A POOR RETURN

subject matter expertise of the Tax Court judges. However, like the Tax
Court, there is no jury trial at the Court of Claims. Most often the case will
be tried in Washington, DC, and appeals go the U.S. Court of Appeals for
the Federal Circuit. The number of tax cases fi led varies considerably. From
1998 through 2005, there were several years when the number of lawsuits
ranged from seventy-eight to ninety-one. In other years, taxpayers fi led over
300 cases in the Court of Federal Claims.

U.S. District Court

Alternatively, the taxpayer can pay and sue by fi ling a refund claim in the
U.S. District Court. The pay and prestige of the U.S. District Court are
presumably greater than that of the Tax Court and Court of Claims. There,
a judge’s salary is guaranteed and the tenure is lifetime, dependent only on
good behavior. To sue in the District Court, the taxpayer must pay fi rst and
then fi le the claim in the taxpayer’s local district where the case is tried.
One judge presides over the case from the onset of the pleadings through
trial. The taxpayer can request a jury trial. Decisions of these courts can be
appealed to the circuit within which the court is located.

In addition to its powers to try civil cases involving tax refunds, the

District Court also tries criminal tax cases, and this court has equity power,
which allows the District Court to order certain behaviors. The IRS will
use the District Court to enforce a subpoena when it seeks information
about the taxpayer that is in the possession of a third party such as a bank.
The IRS is represented by the U.S. Attorney’s offi ce. The cases are usually
important, skillfully argued, and can have precedent value. Thus, although
only fi ve percent of all taxpayer suits (Daily 1992, p. 5/12) are initiated in
the District Court, its prestige, precedent value, equity, and criminal case
power make it an important part of the tax justice system.

Comparing the Courts

Despite a low chance of success at verdict or decision, taxpayers continue
to fi le cases in federal courts, and since many, if not most, cases settle,
simply by contesting the audit, the taxpayer often obtains some tax relief.
For example, in almost one-half of the cases fi led in the Tax Court, the
taxpayer obtained a reduction in taxes (Daily 1992, p. 95). Of the specifi c
options, the “S” option is the quickest and least costly, but presents the
most information asymmetries and the taxpayer often lacks expert advice
and protection. However, a decision to litigate in the Tax Court entails
substantial costs, even if the litigant does not have to pay the contested
amount in order to begin proceedings. Pre-trial pleadings and discovery, the
trial, and post-trial planning and maneuvering take considerable time and

COURTS AND THE IRS

money, and a decision by a Tax Court judge is often delayed for more than
one year; if the taxpayer loses, interest and penalties can accrue. Expert
advice and protection, however, are available, and information asymmetries
are lessened.

Taxpayers pay fi rst in both the Federal District Court and U.S. Court

of Claims. According to some, however, taxpayers have a greater probability
of winning in these courts than they do in the Tax Court (see Daily 1992,
p. 105). Both options are expensive because of the prepayment requirement
and the cost of expertise, representation, communication, and even op-
portunity because of lost time and money. Given the lack of specialization
on the part of the trial judge, there is a greater possibility for information
asymmetries and perverse incentives, particularly when the fact fi nders are
jury members in the Federal District Courts. However, swift discovery rules,
along with strict time schedules, lead to a shorter decision time than for
Tax Court cases.

For many reasons—complexity, discretion, need for consistency, and

similarity of issues—a specialized Tax Court and a low-cost speedy option
make sense for a taxpayer. For many other reasons, however, courts of more
general trial jurisdiction are preferable. The need for juries, for arbiters who
are generalists and thus can look at the broad picture, for some localized
control, and a swifter resolution all argue for litigating in a District Court.
Table 2.1 (next page) presents a comparison of the benefi ts and costs as-
sociated with each forum choice.

Clearly, the available court options allow at least the appearance of

fairness for the taxpayer. Having different courts with different costs and
different benefi ts promotes equity and a sense of fairness, and the taxpayer
can choose the forum most appropriate to the level of the claim involved.
Simple matters can be resolved swiftly and inexpensively. More complex
matters get a full hearing with the choice of an expert arbiter or a jury trial.
Even if success is low (MacLachlan 1995), the taxpayer at least obtains the
appearance of justice. The system design appears fair, and, as Tyler notes,
the fairness of the process itself is often more important than the outcome
(Tyler 1990)—and of course fairness is the yardstick by which courts are
usually evaluated.

However, while the appearance of fairness is important, this still begs

the question whether the system is actually helpful and fair. This is an
important issue because the judicial system should benefi t all taxpayers, not
just those who have no other option but to litigate in Tax Court because
prepayment is required in the other courts. The infl uence of courts extends
beyond the individual matter, and this strengthens the credible commit-
ment of government to the taxpayer. The fact that multinational corpora-
tions or wealthy tax shelter investors choose the Tax Court implies that
there is more to litigation than just the prepayment requirement. Wealthy

GETTING A POOR RETURN

able 2.1

Comparing T

ax T

rial Court Options

“S” Court

Claims Court

ax Court

District Court

Benefi

Specialized Arbiter

Level Playing Field

Specialized Arbiter

Right to Jury

Speed

Right of Appeal

Fewer Information

Level Playing Field

Low Communication

Greater Chance of W

inning

Assymetries

Right of Appeal

Opportunity & T

ime Costs

Right of Appeal

Fewer Opportunities for

Precedential and Policy

No Expertise Costs

Perverse Incentives

alue to Decisions

Few Information Assymetries

Right of Appeal

Greater Chance of

Lack of Perverse Incentives

Precedential and Policy

inning

alue to Decisions

Prestigious Court

Costs

No Right of Appeal

Information Assymetries

Entails T

ime, Communication,

Information Assymetries

Jurisdictional Limit

Lack of Specialized Arbiter

Communication and

Negotiation, Strategy

Low Chance of W

inning

Opportunity Costs

Communication &

Non-level Playing Field

Expertise Cost

Opportunity Costs

Lack of Formal Rules Inhibits

Long W

ait for a Decision

Expertise Cost

Complete Disclosure

Low Chance of W

inning

Greater Opportunity for

Lower Court Prestige

Lower Prestige

Perverse Incentives

Lack

Specialized

Arbiter

COURTS AND THE IRS

litigants have options and use them. Forum choice means little if you have
no realistic selection. A Joy Anders cannot weigh the costs and benefi ts
or the applicable precedent in each court that would allow her to choose
an option with a greater chance of success. While the forum choices allow
even poor taxpayers to have their “day in court,” they cannot equalize the
choices and chances for all taxpayers.

CONCLUSION

Although the tax laws are undoubtedly “revenue raisers” (MacLachlan 1995),
the choice of forums and the ease and relative effortlessness of fi ling claims
allow almost any aggrieved taxpayer to at least have a fair hearing. Low- and
middle-income taxpayers can fi le claims quickly and have them heard in an
informal, inexpensive setting. The appearance of fairness is achieved even
if most taxpayers have little actual choice of forum because of the cost of
litigation outside the Tax Court.

However, the courts serve a broader role of ensuring actual fairness.

We have a system of voluntary compliance at least to the extent that
the system depends on the individual taxpayers self-reporting. To ensure
voluntary observance of tax fi ling, taxpayers need a credible commitment
from the courts that justifi es their compliance. That is, we want the IRS
to be aggressive in the collection of tax revenue. The good of the entire
country depends on the agency’s collecting revenue in as effi cient a manner
as possible, albeit not at the expense of fairness or injustice to taxpayers.
In an ideal world, the IRS aggressively pursues tax cheaters and leaves the
honest citizen alone. We know this does not happen. An imbalance toward
aggression, toward effi cient collection, inevitably means that some honest
taxpayers will not be treated in a fair and just manner.

Thus, we also want to know that taxpayers will get a fair hearing if

they choose to contest the agency action. The appearance of fairness is
not enough. If certain forums offer advantages to the litigants over oth-
ers, then these choices, to be effective, must be open and available to all
litigants. In the end, compliance depends on taxpayers believing that they
have some sort of stake in the outcome. Why litigants choose to sue and
why they would select one particular forum over another are examined in
the next chapter.

This page intentionally left blank.

TAX LITIGATION AND TAX FORUM CHOICE

THREE

TAX LITIGATION AND TAX FORUM CHOICE

There are the Joy Anders’s of the world and there is Wal-Mart. In between
are middle-income taxpayers like David and Barbara Forrest and tax shelter
investors such as James William. Joy Anders is seeking to recover the $1,500
the IRS has assessed in its claim of improper use of the Earned Income
Tax Credit. Wal-Mart was disputing over $30,000,000 due to challenged
accounting practices. David and Barbara Forrest sought to recover $30,000
of defi ciencies for denied loss deductions, while James William was charged
with investing in an abusive tax shelter. Each has had a dispute with IRS,
and each has a choice of a tax trial court; three chose the Tax Court, while
one litigated in the District Court. Why did Joy Anders, David and Barbara
Forrest, and Wal-Mart select the Tax Court, while James William decided
on the District Court? Why would they even bother to litigate? What is the
incentive to spend time and money to recover any or all the assessment? Is
it possible that Congress can promote the use of one forum over another or
even advance litigation against the IRS?

Consider taxpayers such as Joy Anders. Congress can encourage the

IRS to audit her and similarly situated taxpayers while it can also encour-
age her to litigate her claim against the IRS. Thus, Congress can push for
audits of Joy and other poor taxpayers, defl ect the blame onto the IRS,

and offer hope to the taxpayer for a just solution. Although legislators and
scholars often outwardly decry policy making through the courts or the use
of litigation to solve social problems, this chapter explores how the politi-
cal majority can use litigation to move policy or settle issues in a manner
consistent with the preferences of the political majority. This can defl ect
attention away from the deliberate policy choices of the political majority.
Instead it allows taxpayers to direct their ire toward a bureaucratic agency,
particularly, the IRS.

Such litigation, separate and apart from the actual decisions of the

judges and their policy preferences, can move an agency like the IRS to

GETTING A POOR RETURN

alter policy choices. To show this, the chapter examines the infl uences
behind these choices to litigate and, once that choice is made, what forum
to choose. Although this chapter will examine the individual decisions that
lead to litigation and forum choice, the emphasis will be on the factors
that increase aggregate tax litigation and the reasons behind the decisions
of litigants to select one tax forum over another.

As observed in chapter two, while every taxpayer who sues the IRS has

a theoretical choice of litigation or settlement, given fi nancial circumstances,
taxpayers like Joy Anders have few options but to sue and a limited choice
of forums. However, wealthier litigants and those with greater resources do
have a real choice. The cumulative effects of these choices lead to changes
in IRS behavior. To demonstrate the effects of these choices, I examine
national policy, court preferences, and litigation patterns in the Tax Court
and District Court from 1994 through 2000.

Arguably, tax policy has been politicized and ideologically divisive since

the late 1970s, with conservatives generally opposed to greater tax collection
and enforcement. As noted in chapter one, tax cuts were prominent platform
positions of the Republican Party in presidential elections from 1980 onward.
Ideological positions on public policy issues in Congress can translate into
increasing or decreasing litigation trends. By encouraging or discouraging
litigation through the use of standing, a national governing coalition can
push policy toward its favored direction. There are many ways that Congress
can do this. For example, through expanding or contracting the standing
of those who can sue, Congress can enlarge or shrink the potential pool of
litigants (Smith 2006) and thereby enlarge or shrink the number of lawsuits.
Then, through the appointment and confi rmation process, the president and
Congress can choose judges more or less sympathetic to the claims of the
litigants and push decisions in favor of a particular policy point of view.

In several policy domains, courts have a played a prominent and perhaps

decisive role in determining public and social policy. In many famous cases,
court rulings have signifi cantly changed policy. These range from desegregation
through major Supreme Court decisions such as Brown v. Board of Education
(1954), the right to privacy through Griswold v. Connecticut (1968), abortion
through Roe v. Wade (1973), and even state-level education policy through
several state high court decisions (Bosworth 2001).

Given the prominent role of courts in setting and determining policy

and the particular prominence of taxes and tax policy over the past three
decades, why and where tax litigants choose to sue is critically important
if one wants to understand the dynamics of not just tax policy, but public
policy. Courts need litigants to bring a matter to them before they have
a chance to rule on any particular matter and these lawsuits can change
public policy by forcing the agency to expend time and resources fi ghting

TAX LITIGATION AND TAX FORUM CHOICE

the litigation and then having to comply with the directives of these courts.
Without litigants, courts cannot change policy.

To show the importance of litigation and litigation choice, this chap-

ter fi rst reviews the divergent literature on litigation and policy and then
examines individual and aggregate causes of litigation. These analyses are
then applied to the causes of aggregate increases or decreases in tax litigation
and the reasons litigants would choose one tax forum over another. From
this discussion, testable hypotheses are derived and presented to determine
the causes of both tax litigation and tax forum choice. The next section
describes the data, methodology, and models. I then explain and interpret
the results and in the conclusion offer some suggestions as to the broader
meaning of litigation and law forum choice and policy.

POLITICAL FORCES AND LITIGATION

Politics matters to litigation because politics is about policy choice and
policy change. Litigation can lead to policy change by compelling judges
and juries to pick one party over the other. Decisions to litigate—the idea
of whether or not to sue and, if that choice is made, where to sue—are not
made in a vacuum. These choices depend on larger economic and politi-
cal circumstances in addition to an individual grievance. Most analyses of
litigation and forum choice examine precedent and performance of litigants
in each respective court to calculate the likelihood of winning. Little atten-
tion has been paid to the dominant national coalition’s ability to encourage
litigation or the use of one court over another through legislation and an
appointment process that nominates and selects judges more favorable to one
policy point of view. If national policy encourages taxpayers to dispute the
IRS and if courts preferences are such that they are more likely to rule in
favor of the taxpayer against the IRS, then these become additional factors
to consider in calculating forum choice and the dominant elected coalition
uses this information to encourage the use of one particular forum. The
political, social, and economic circumstances infl uence the choices made
by Joy Anders, Wal-Mart, and the Forrest’s.

Scholarship has paid considerable attention to litigation and its

infl uence on social and bureaucratic policy (Cortner 1968; Olson 1990;
Shipan 2000; Smith 2006; Unah 2003). One area of investigation focused
on interest group litigation. Early explanations centered on the theory that
disadvantaged groups unable to change policy through the legislative process
used litigation to achieve policy goals (Cortner 1968). This political disad-
vantage theory seemed particularly pertinent for those who examined the
National Association for the Advancement of Colored People (NAACP) and
the organization’s deliberate strategy to use the courts and the Fourteenth

GETTING A POOR RETURN

Amendment to achieve civil rights goals that a Congress dominated by
southern conservative gatekeepers on key congressional committees continu-
ally thwarted (Cortner 1968; see also Epstein 1985).

While theoretically appealing and enforcing the notion that courts help

achieve justice and fairness, later research discounted this “disadvantaged
group” theory. Subsequent research showed that interest groups on all sides
of the political spectrum resorted to the courts and that even powerful busi-
ness and other interest groups used the courts to achieve policy objectives
(Olson 1990; Unah 2003). Advantaged as well as disadvantaged groups sued
to attain policy goals to supplement, enforce, or even just maintain legisla-
tive gains. For example, business interests and conservative interest groups
as well as traditional liberal and politically disadvantaged groups have used
the courts either by participating as parties to the litigation or through fi ling
of amici briefs (Caldeira and Wright 1988; Unah 2003). As Unah (2003,
p. 66) noted, pursuing rights in court is not something that is restricted to
disadvantaged groups.

Scholars have also shown that Congress encourages individual litigation

as way to achieve policy goals. Although politicians and certain scholars
might decry the use of the courts to achieve social and political goals, politi-
cal institutions can use the courts to achieve policy goals and regulatory
control through encouraging litigation, both by organized interests and by
individuals. Congress does so through promoting judicial review and legislation
design that allows citizen suits as an additional measure of regulatory control
(Shipan 2000; Smith 2006). For example, Congress can encourage citizen
suits as a way of obtaining regulatory control over an agency that cannot be
achieved through the legislative process. The lawsuits can act as a monitor-
ing device on agency action and save considerable legislative time and cost
(Smith 2006). Lawsuits can prevent “bureaucratic drift” away from political
control and ensure statutory construction in accordance with congressional
preferences (McCubbins, Noll, and Weingast 1989). The dominant coalition
is most likely to favor such lawsuits when the court that exercises review is
ideologically closer to the elected policy makers (Shipan 2000).

Applying this to tax litigation means that the elected coalition

through implict legislative signals and the judicial appointment process can
encourage the use of tax litigation in an era of tax cutting legislation and
antigovernment spending rhetoric. Tax litigation is both a way to control
the IRS and another forum for disaffected taxpayers to fi nd tax relief short
of legislation. The same logic should apply to forum choice. If one forum
offers a greater opportunity to control the IRS, then the coalition should
encourage the use of that particular forum. In the following section, I detail
these arguments as to why and how litigation and forum choice would be
encouraged by Congress.

TAX LITIGATION AND TAX FORUM CHOICE

THE DECISION TO LITIGATE AND FORUM STRATEGY

Rational Actors and the Political and Economic Environment

Individuals and groups have goals and decide to go to court to support those
goals. As part of this calculation, scholars have theorized and found evidence
for the concept that litigants act rationally when deciding to proceed to trial
or even appeal a decision to U.S. Supreme Court (Bebchuk 1984; Landes
1971; Posner 1973; Reinganum 1988; Reinganum and Wilde 1986; Songer,
Cameron, and Segal 1995). Early formal models showed that the decision to
litigate was a simple calculation of costs and benefi ts (Landes 1971; Posner
1973). Applying these insights to some of the individuals in this book, Joy
Anders then proceeds to court because she has nothing to lose. The costs
are minimal and the potential benefi t is enormous.

Later game theoretic models showed that the litigation calculus includes

potential economic gain versus the ability to afford court costs, signals from
the opposing party, and potential liability (Bebchuk 1984; Reinganum 1988;
Reinganum and Wilde 1986). Empirical research extended litigation calcula-
tions to appellant decisions. Appeals to the Supreme Court are more likely
if the appellant calculates that there is a greater chance of reversal of the
adverse appeals court decision (Songer, Cameron, and Segal 1995). If there
is a greater likelihood of winning at the appellate level than there is of los-
ing, an individual would be more likely to proceed to trial.

While costs, benefi ts, and signals help determine individual decisions

to proceed to trial, those who have examined aggregate decision making
fi nd that cumulative litigation rates rise and fall in relation to broader envi-
ronmental factors. These causes include such factors as the economy, the
political environment, and changes in the political and economic environ-
ment (Clark 1990; McIntosh 1990; Stookey 1990). Economic growth and
political change encourage litigation (McIntosh 1990; Stookey 1990).This
means that implict signals from Congress and the executive branch as well
as explicit tax legislation should matter. There is a strong ideological dimen-
sion attached to tax policy and tax enforcement. If conservatives oppose
taxation and rigid enforcement of tax policy, then the more conservative
the political coalition, the greater the likelihood of litigation contesting tax
assessment. If the political leadership attacks existing tax policy and tax col-
lection, this should increase lawsuits against the IRS. It is a clear signal to
challenge tax assessments. Litigation should also increase if Congress enacts
legislation that makes it easier to sue or challenge the IRS.

Conservatives have adopted an antitax posture over the previous three

decades. It has become a conservative position to oppose taxes and, by exten-
sion, the IRS. Conservatives believe in lower taxes and less enforcement.

GETTING A POOR RETURN

Arguably, this has occurred because of the antigovernment stance adopted
by conservatives and the pro-government posture adopted by liberals during
this time period and the corresponding linkage with support for spending,
support for government, and political liberalism (McClosky and Zaller
1984). Government spending depends on tax revenue and collection of
revenue; therefore, opposition to government spending means opposition
to the collection of revenue that supports such spending, while support for
government spending means support for the collection of revenue. Trends
in polling data seem to bear this out.

Following this, one should then expect to see an increase in tax litiga-

tion after Republican legislative gains or the introduction of legislation that
makes it easier to sue the IRS and as the judiciary becomes more conserva-
tive. An examination of data from 1994 through 2000 seems to bear this
out. For example, Figure 3.1 shows aggregate tax litigation fi lings in the
U.S. District Courts and the U.S. Tax Courts during this period. We see a
gradual increase after the election of Republican majorities in both houses
of Congress in 1994 and another gradual increase 1997 following critical
hearings about the IRS and subsequent taxpayer bill of rights legislation.
A decrease then occurs after the legislation curtailing the ability of the
IRS to audit.

5000

10000

15000

20000

25000

30000

35000

1994

1996

1998

2000

Year

Filings

Figure 3.1 Total Tax Litigation Filings 1994–2000

35,000

30,000

25,000

20,000

15,000

10,000

5,000

TAX LITIGATION AND TAX FORUM CHOICE

While aggregate political trends are important, they are only one

factor in a calculation to litigate. The likelihood of success can depend on
other factors. For example, a key consideration in the likelihood of success
is existing precedent. Practitioner tax manuals stress the importance of rel-
evant precedent (Berall 1992; Garbis and Schwait 1971; Taylor et al. 1990).
Favorable precedent makes litigation more likely, while unfavorable precedent
would discourage challenging a post audit assessment. Thus, precedent acts
as an infl uence on whether or not to sue the IRS and forum choice.

The Logic of Forum Choice

In addition to encouraging litigation, political coalitions have other means
to control agency policy. One method is to encourage the use of one forum
over another. Though often derided as forum shopping, forum choice is not
an isolated procedural issue. In many areas of law, litigants have a choice of
venue. Plaintiffs, or their lawyers, often select where they bring a lawsuit.
For example, if a party is injured in an accident and the prospective plaintiff
and defendant are from different states, the plaintiff can sue in a state court
or use diversity jurisdiction to bring the case in federal court.

While diversity jurisdiction is a constitutional right, many of these

forum choices are predicated on legislation, and a leading example of forum
choice is federal tax litigation, as shown in chapter two. All else being equal,
if one forum offers a greater chance of success than another, one would
expect that taxpayers would choose that particular forum. In tax litigation,
defendants have a choice of three forums: the U.S. District Court, the
Article III court of general trial jurisdiction in the federal system, and the
U.S. Tax Court, a specialized court created by Congress under its Consti-
tutional Article I legislative power. The third is the U.S. Court of Federal
Claims. This court’s jurisdiction is limited to taxpayer lawsuits against post
audit assessments by the IRS.

While there is signifi cant existing research on the reasons for litiga-

tion, there is much less research on choices litigants have made once they
decided to sue. There has been little investigation of the costs and benefi ts
or the political, economic, and social factors that lead to the choice of one
forum or another. Why would Wal-Mart choose to litigate in the Tax Court
when the District Court is often viewed as a more favorable forum for the
taxpayer? Why does anybody opt for the more costly District Court? This
scholarly omission is surprising. Litigants have a choice of forum in several
areas. For example, multistate or multinational corporations often have suf-
fi cient ties to several states or even several nations that allow these entities
a choice of several different forums in which to fi ght unfavorable laws or
defend corporate action from individual litigation. Individual plaintiffs can
use diversity jurisdiction to bring state law claims in federal court, and estates

GETTING A POOR RETURN

and bankruptcy petitioners can employ asset location to select a favorable
venue for their respective claims.

Forum choice is often referred to derisively as “forum shopping.”

Forum shopping has the potential to denigrate the ideal of courts standing
for fairness and impartiality (Eisenberg and LoPucki 1999). Forum shop-
ping is viewed as perversion of the ideals of justice. It is regarded as the
unfair exploitation of jurisdiction or venue rules to affect the outcome of
the lawsuit (Juenger 1990, p. 677). Different jurisdictions or venues can
treat the same class or type of litigant differently and therefore justice can
differ by location. Examining litigation decision making or the reasons for
such litigation without examining the forum chosen to bring a court case
leaves at best an incomplete picture of much of the litigation process and
the potential impact on justice and fairness.

Forum choice rarely occurs accidently. In several domains, forum choice

is a function of deliberate legislative design (Eisenberg and LoPucki 1999).
Federal tax litigation is one such area. For audited taxpayers fi ghting an
IRS post audit, the “choice of forum is perhaps the single most important
decision in planning for tax litigation” (Berall 1992, p. 75).

There are many practical reasons that Congress created an alternative

specialized forum for tax disputes, among them the complexity of the issues
and the potential for tax litigation to overwhelm the existing court system.
However, encouraging tax litigation forum choice also can serve political
purposes. Congress has a much greater ability to change the policy preferences
of the shorter-tenured Tax Court than the lifetime-tenured District Court.
This can be seen by the an examination of Figure 3.2, which tracks the

0.05

0.1

0.15

0.2

0.25

0.3

0.35

1992

1993

1994

1995

1996

1997

1998

1999

2000

Year

US District Court Median

US Tax Court Median

Figure 3.2 Change in Ideology of Tax and District Courts 1992–2000

U.S. Tax Court Median

U.S. District Court Median

TAX LITIGATION AND TAX FORUM CHOICE

median ideology of the Tax Court as compared to the median ideology of
the District Court from 1992 to 2000. I start with 1992 because this allows
me to indicate changes since the election of President Clinton.

Figure 3.2 shows that the median ideology of the Tax Court changed

more than the median ideology of the District Court during this period. In
fact, a simple analysis shows that the ideology on average for the Tax Court
changes more than twice as much each year than for the District Court
(.008 as compared to .003). While both courts became more liberal dur-
ing the Clinton years, a logical result of a Democratic president’s power of
appointment, the median tax court judge remained more conservative than
the median district court judge, except for a few years. Given conservative
and Republican disenchantment with taxes and tax collection, it follows
that the Republican majority would encourage litigation against the IRS
as an additional measure of agency control over tax collection and would
encourage taxpayers to use the Tax Court if that court was more conserva-
tive than the District Court.

Following this logic, just as there was an increase in tax litigation

following Republican electoral victories in Congress, one should then also
see an increase in litigation in the Tax Court compared to litigation in the
District Courts, and again examination of data from 1994 through 2000
seems to bear this out. Figure 3.3 compares aggregate tax litigation fi lings
in the U.S. District Courts and the U.S. Tax Courts during this period.

Figure 3.3 Change in the Ratio of Tax Court to

District Court Filings 1994–2000

1994

1995

1996

1997

1998

1999

2000

Year

Filing Ratio

GETTING A POOR RETURN

What we see is a signifi cant in spike in Tax Court fi lings compared

to District Court fi lings following the Republican victory in the midterm
elections of 1994 and again after the series of hearings on IRS abuse and the
introduction of legislation favoring the taxpayer in disputes with the IRS.

Just as costs and benefi ts become part of the calculus of litigation

decisions, they should also factor into tax forum choice. Because of the
prepayment requirement in the District Court, far more taxpayers choose
to fi le a claim in the Tax Court. If cost were the only concern, then no
taxpayer would ever choose the more expensive option of the District Court.
However, the probability of winning is part of the calculation, and a taxpayer
has to offset the potential cost of litigating versus the likelihood of success.
Practitioner manuals make clear that an important factor in both bringing
suits, and choice of forum, is precedent (Berall 1992; Garbis and Schwait
1971; Taylor et al. 1990). Precedent acts as an infl uence on whether to sue
the IRS as well as on forum choice. The more favorable the precedent in a
particular forum, the more likely taxpayers will sue in that forum. Winning
because of favorable precedent in the District Court is a benefi t that could
outweigh the cost.

In addition to precedent, another factor that should matter in the

likelihood of prevailing on the merits, and therefore forum choice, is ideology
and, in particular, judicial ideology. Ideology matters because, as previously
argued, tax policy is an issue charged with signifi cant ideological dimensions
(see, e.g., Scholz and Wood 1998, 1999; Stevenson 1997; Wiseman 1997a,
1997b). An earlier study found a signifi cant positive correlation between
judicial liberalism and support for the IRS (Howard and Nixon 2002). If
the judges on a particular court are more conservative and more opposed
to the IRS, a rational tax litigant should consider that a more conservative
court will offer the tax litigant a greater probability of winning.

As previously noted, Tax Court nominees are appointed by the president

and confi rmed by the Senate. Although the nominees have considerable
prior experience and expertise in tax law, they often have signifi cant politi-
cal background and experience (Tidrick 2004). In this they are similar to
nominees of the other specialized courts (see Unah 1998) and to nominees
of more generalized Article III courts. Thus, the more conservative the
nominees that are appointed to the Tax Court, the greater the likelihood
that litigants will choose the Tax Court over the District Court while the
more conservative the appointees to the District Court, the greater the
likelihood that the taxpayer will choose the District Court instead of the
Tax Court. In addition, the Tax Court judge has a fi fteen-year tenure on
the Tax Court. This is in contrast to the lifetime tenure of the District
Court judge. Because of this, it should be easier to shift the ideology of
the Tax Court; as the ideology of the national governing coalition becomes
more conservative, then the Tax Court should show a greater shift toward

TAX LITIGATION AND TAX FORUM CHOICE

conservatism than the life-tenured District Court. This should in turn lead
litigants to prefer the Tax Court to the District Court.

According to research on aggregate litigation, economic and political

factors and changes in the economic and political environment also should
affect forum choice. This means that implict signals from the national
governing coalition and explicit tax legislation both should matter. Given
the ideological dimension attached to tax policy and tax enforcement the
more conservative the dominant political coalition, the greater the likeli-
hood of litigation. If the political leadership attacks existing tax policy and
tax collection, then this should increase lawsuits against the IRS, and if
legislation is enacted that makes it easier to sue or challenge the IRS, this
should increase litigation. Since contesting assessments in the Tax Court is
a less expensive than contesting them in the District Courts, an increase
in tax litigation should lead to a greater increase in Tax Court litigation
compared to District Court litigation. In addition, as previously mentioned,
the national coalition can change the ideology of the Tax Court much more
quickly than the District Court. As the ideology of the national coalition
becomes more conservative, this should send a signal to taxpayers to prefer
litigation in the Tax Court and avoid the District Court.

Economic factors should also infl uence litigation and forum choice.

Wealth allows greater freedom in pursuing more costly alternatives and thus
greater wealth should lead to greater use of the more costly District Court.
In addition to income, an indication of wealth is the actual post audit
assessment of additional taxes and penalties imposed by the IRS. The greater
the assessment, the greater the likelihood that the taxpayer has suffi cient
income to sue in the more expensive District Court and thus higher post
audit assessments should lead to an increase in District Court litigation.
Conversely, unemployment means fewer resources and less ability to prepay
and therefore increased unemployment should lead to greater use of the Tax
Court. In the next section, I offer hypotheses to test these assertions.

HYPOTHESES

If, as argued, rational taxpayers will seek to maximize benefi t and minimize
cost, then the greatest determinant for this would be the likelihood of win-
ning. Thus, the more favorable the precedent is to taxpayers, the greater the
probability that taxpayers will sue the IRS. Derived from this, an increased
probability of winning in one particular forum will move litigants to choose
one forum over another. In particular, precedent and ideology will determine
forum choice, leading to the fi rst hypotheses:

Hypothesis 1: The more favorable the precedent is to the taxpayer,
the greater the likelihood litigants will sue the IRS.

GETTING A POOR RETURN

Hypothesis 1a: The more favorable the precedent is to a particular
forum, either the Tax Court or the District Court; the greater the
likelihood litigants will choose that particular forum.

Given the ideological nature of tax policy and tax enforcement, the

more conservative the judiciary, the greater the likelihood that taxpayers will
sue the IRS. In addition, taxpayers will choose the judicial forum with more
conservative jurists because that will increase the likelihood of prevailing
in their claim against the IRS. This leads to the second set of hypotheses:

Hypothesis 2: The more conservative the ideology of the judiciary, the
greater the likelihood that litigants will sue the IRS.

Hypothesis 2a: The more conservative the ideology is in the Tax
Court or the District Court, the greater the likelihood that litigants
will choose that particular forum.

Politics matter in that the national governing coalition sends signals

encouraging tax litigation and the use of one forum over another. The more
conservative the national coalition, the more taxpayers should want to sue,
particularly in the Tax Court, which will more quickly refl ect the national
political consensus. In addition, Congress can enact specifi c legislation that
promotes taxpayers to sue and promotes the use of the Tax Court over the
District Court. This leads to the next hypotheses:

Hypothesis 3: As the Congress becomes more conservative, the greater
the likelihood that litigants will sue the IRS.

Hypothesis 3a: As the Congress becomes more conservative, the greater
the likelihood that litigants will sue in the Tax Court.

Hypothesis 4: Legislation favoring the taxpayer should lead to an
increase in tax litigation.

Hypothesis 4a: Legislation favoring the taxpayer should lead to an
increase in Tax Court litigation compared to litigation in the District
Court.

Finally, the economy should determine both tax litigation and tax

forum choice, leading to the fi nal hypotheses:

Hypothesis 5: Wealthier taxpayers will have a greater likelihood of
suing the IRS.

TAX LITIGATION AND TAX FORUM CHOICE

Hypothesis 5a: Wealthier taxpayers will have a greater likelihood of
litigating in the District Court.

Hypotheses 6: The greater the post audit assessments imposed by the
IRS, the greater likelihood of suing the IRS.

Hypotheses 6a: The greater the post audit assessments imposed by the
IRS, the greater the likelihood of taxpayers litigating in the District
Court.

Hypothesis 7: Poorer taxpayers will be less likely to litigate.

Hypothesis 7a: Poorer taxpayers will have a greater likelihood of liti-
gating in the Tax Court.

FACTORS INFLUENCING LITIGATION AND FORUM CHOICE

Following the approach used by earlier analyses (e.g., Scholz and Wood
1998, 1999), I examine state variation for this study. I created what is
known as a panel, or cross-sectional time series, featuring observations of
the fi fty states over time—specifi cally, one observation for each state for the
years 1994 through 2000 (n = 350).

The cases were then aggregated by

state per year, with the state determined by the domicile of the taxpayer.
I used two different dependent variables for the two examinations to test
fi rst aggregate litigation and then forum choice. The fi rst simply totals the
number of tax lawsuits fi led in both the Tax Court and the District Court per
state per year. The second dependent variable creates a ratio of the number
of lawsuits fi led in the Tax Court by dividing that fi gure by the number of
lawsuits fi led in the District Courts per state per year. The ratio increases
in each state as more lawsuits are fi led in the Tax Court and decreases as
more lawsuits are fi led in the District Court. I used the same factors to test
each dependent variable and the following paragraphs detail the process of
selecting the factors to examine, based on the arguments of the preceding
sections, and how each factor is measured for the analyses.

For example, practitioners and scholars alike agree that precedent

matters to litigation and forum choice. However, precedent is a diffi cult
concept to measure, particularly when examining aggregate data. It is dif-
fi cult to point to and then incorporate any one particular case and show
that the case infl uences aggregate decision making and aggregate choice.
In addition, there are two types of precedent that need measurement. The
fi rst is the concept of horizontal precedent. Horizontal precedent is a case
decided by a hierarchically similar court that in theory binds that particular
court at that same level to follow that ruling. Thus, a Tax Court ruling on

GETTING A POOR RETURN

a particular item of tax law acts as horizontal precedent on the Tax Court,
but would not bind an appellate court.

The second is the concept of vertical precedent. Vertical precedent

is a case decided by a hierarchically superior court that is binding on lower
courts. For example, a Court of Appeals decision for a particular circuit is a
vertical precedent binding on all trial courts within that circuit, as well as
a horizontal precedent for that particular circuit. Thus, I needed to create
both types of precedent applicable to aggregate analyses.

For horizontal precedent, my solution was to create a “win percent-

age” score for each particular tax trial court. Cases were coded so that
each individual case was listed as either a win for the IRS or a win for the
taxpayer. A ruling for the petitioner counts as a win for the taxpayer, while
a win for the commissioner of the Internal Revenue Service counts as a
win for the IRS. Often each side wins some issues but loses in some other
areas. When this happens in the Tax Court, the cases are decided under
Rule 155.

This is the Tax Court opinion concluding that the decision is

neither a complete victory nor a total defeat for either side (Dubroff 1979;
Keir, Argue, and Seery 1981). Often a Rule 155 case involves some minor
concessions by the IRS. So the issue becomes one of fi tting these types of
cases into the coding scheme. I decided to follow a pattern established by
previous scholars who have looked at taxpayer litigation. Thus, I followed
the coding scheme outlined by Daniel Schneider (2001, 2002) in two of his
articles concerning tax decisions. I examined the major issue or issues under
consideration and coded the case as a win for the petitioner or respondent if
the taxpayer won some tax relief in the decision. Minor or subsidiary issues
were discounted. Although the U.S. District Court has no Rule 155, I used
the same coding rules for the decisions of this court. That is, I looked at
the major issue in dispute, and whether the taxpayer won some relief in
the decision over that particular issue.

For a vertical measure of precedent, I used the ideology of the appro-

priate Federal Court of Appeals, reasoning that the lower courts, both Tax
Court and District Courts, would follow the rulings of the appropriate Court
of Appeals. The more conservative the appropriate Court of Appeals, the
more likely the lower court would rule in favor of the taxpayer and the
more likely the taxpayer would litigate.

Admittedly, neither measure is a perfect substitute for precedent. At

best, each functions as a proxy—something that approximates the concept
I am trying to measure—and each rests on certain assumptions. One is the
assumption that aggregate decisions that show trends of taxpayers or the
agency winning or losing are related to legal trends. Taxpayers or the IRS
win or lose more often because the courts have ruled on a point of law in
prior cases and the facts of particular cases are similar to the facts of these
prior cases. The logic is intuitive in that court decisions, as shown by win

TAX LITIGATION AND TAX FORUM CHOICE

percentages, are related to a taxpayer winning or losing a case. The verti-
cal precedent measure rests on the assumptions that lower courts follow
the rulings of the higher courts and that the ideology of the higher court
captures the trend of the rulings.

This second measure of precedent, which relies on ideology, begs the

question of how to evaluate and determine political attitudes and beliefs.
Ideology, like precedent, is another concept that is diffi cult to measure.
While we all accept that certain political fi gures (e.g., Ted Kennedy) are
liberal, while others are conservative (e.g., Jessie Helms), it is diffi cult to
measure the precise degree that a Ted Kennedy is more liberal than a Jes-
sie Helms. For elected offi cials, the easiest way to calculate ideology is a
voting record. This is the approach of the group known as Americans for
Democratic Action (ADA). The ADA publishes scores that measure the
percentage of liberal votes on a select number of issues. The ADA then
calculates liberalism on a 0 to 100 scale, with 0 being the most conservative
and 100 the most liberal.

However, that begs the questions as to what is a liberal or conservative

issue and what the vote should be to determine the liberal or conservative
side. In addition, these measures use a preordained idea of liberalism or
conservatism to determine if someone votes the correct way, and then in
turn considers these votes to determine liberalism and conservatism. These
ideology scores are premised on a small number of votes (McCarty, Poole,
and Rosenthal 2006). This also leaves unanswered the question of measur-
ing judicial preferences. Judges do not “vote” on policy issues and therefore
one cannot calculate an ideological score based on voting. While some
judges might have held prior elective offi ce and one could use those scores,
presumably the majority did not and thus have no voting record. Using one
measure for elected offi cials and another for judges would allow the potential
problem of having different measures, with different scales meaning different
things in defi ning liberalism and conservatism.

To avoid these problems and to test the prevailing elected offi cial

and judicial preferences, I fi rst used the Poole and Rosenthal median
W-nominate scores of the U.S. Congress from 1994 through 2000 (Poole and
Rosenthal 1997). These are static (unchanging) measures based on spatial
roll call voting. They are predicated on the entire spectrum of votes and
on “who votes for whom and how often” (McCarty, Poole, and Rosenthal,
2006, p. 5). With the Republican takeover of both houses of Congress in
the midterm elections of 1994, the Congress became signifi cantly more con-
servative. Beginning in 1998, this Republican-dominated Congress enacted
several measures to provide taxpayer relief. Among these changes was the
taxpayer bill of rights, which shifted the burden in contesting audits from
the taxpayer to the IRS. To measure the impact of these legislative enact-
ments, I created a dummy (two potential choices) variable for the year 1998

GETTING A POOR RETURN

(i.e., 1 for the year 1998 and 0 for all other years) to represent the impact
of the legislation from this pivotal year. I did not include any measure of
presidential ideology because Bill Clinton was the president for all the years
of this study and therefore no signifi cant measurable change occurred in
presidential ideology.

That brings us to measuring federal judicial ideology, a complex and

diffi cult task especially when it comes to judges who have no actual voting
record and do not espouse overt partisan statements or beliefs. Several methods
have been employed in previous studies. A commonly accepted measure of
Supreme Court ideology, the Segal-Cover scores (1991), was developed by
coding newspaper editorials about each justice at the time of the justice’s
nomination. Such a measure is not practical for lower court judges because
their nomination is far less salient and often accompanied solely by a news
release by the court or the White House at the time of nomination.

Thus, judicial politics scholars have developed several other methods

for measuring lower court ideology. For example, a judge’s own partisan
affi liation

and the ideology of a judge’s appointing president have often

been employed as substitutes of judicial attitudes (Segal and Spaeth 1993).
But focusing on judicial partisanship restricts possible ideology indicators to
one of two values, assumes partisanship equals ideology, and fails to account
for the great variation of judicial attitudes. Scholars have sometimes ignored
the ideology of the judge, and inferred his or her ideology from that of the
appointing president. Tate and Handberg (1991) proposed an ordinal measure
of the ideology of the appointing president: –1 for ideologically conservative
presidents, 0 for nonideological presidents, and 1 for ideologically liberal
presidents. This measure may be attributed to every judge on the circuits,
but the data range is not much better than judicial partisanship. In this
measure, there are only three possible values of ideology; it again rests on
two assumptions—that presidential ideology equals judicial ideology and
that partisanship equals ideology.

Segal, Timpone, and Howard (1999) improved on the Tate and Hand-

berg ranking of presidential ideology by surveying presidential scholars and
establishing an interval scale for each president since FDR. Scholars were
asked to rate presidential liberalism on a 0 to 100 scale and then the scales
were combined. The Segal, Timpone, and Howard economic liberalism scores
for judges range from 17.6 (for appointees of Reagan, the most conservative
president) to 82.5 (for appointees of FDR, the most liberal president).

Using

this approach, the data range for judge ideology is at least theoretically bet-
ter, but no rankings are available for presidents (and their appointees) prior
to FDR, and all judges appointed by the same president receive the same
score regardless of individual variation.

A number of scholars have suggested combinations of existing measures

of ideology. Humphries and Songer (1999) constructed a more differentiated

TAX LITIGATION AND TAX FORUM CHOICE

measure of judicial ideology based on a logit analysis of judicial voting in
economic cases, with a North/South dummy and the Tate and Handberg
measure of appointing president’s ideology as predictors (Humphries and
Songer 1999). The scores range from 0 (conservative) to 1 (liberal), and
can be computed for every appeals judge who has ever served on the modern
circuit bench. However, this still leaves little actual variation. In practice,
judges are assigned only one of six possible scores. In addition, these mea-
sures are based on different scales than the W-nominate scores, so we do
not know if a liberal under any of these judicial ideology measures would
be the same liberal under the W-nominate scores.

Another group of scholars have explicitly used the nominate scores to

calculate a measure of judicial ideology for lower court nominees. Michael
Giles and colleagues Virginia Hettinger and Todd Pepper (2001, 2002) used
the nominate scores of the home state senators or of the nominating president,
depending on the political circumstances of the appointment, in calculating
the ideology of Appellate Court judges. Political circumstances that preclude
using the home state senator’s nominate score and instead substituting the
president’s ideology score occur when there was no home state senator of
the same party as the president, or when there was an appointment to the
D.C. Circuit Court of Appeals, which has no home state senator.

This creates more variation in the scores than ordinal measures rank-

ing presidential ideology, but the ideology of any individual judge is never
different from the patron senator or the president, and there can be no
variation between judges appointed from the same state, or by the same
president, even if, as sometimes happens, they are appointed from different
political parties.

In addition, because of the calculation of the nominate

scores for the president, which are based on public pronouncements, the
president usually has a more extreme liberal or conservative nominate score
than most senators. This in turn makes a judge appointed by the president
in the absence of senatorial courtesy often have a far more extreme liberal
or conservative ideological score than is warranted.

To avoid these problems, for this and the next two chapters, a measure

of the personal ideology of each federal District Court judge and Tax Court
judge is calculated, which, like the Giles, Hettinger, and Pepper score, is
strictly comparable to the W-nominate scores but assigns an essentially unique
ideology score to each appointed federal judge (Howard 2007; Howard 2008;
Howard and Nixon 2003). These scores are constant throughout the judicial
career. Many measures of ideology remain constant throughout the career of
the judge, executive, or legislator; among these are the widely used Segal-
Cover scores for Supreme Court ideology, W-nominate scores for legislative
ideology, and Wright, Erickson, and McIver scores for state-level ideology.
Based on these constant measures, it is reasonable to assume that these
attitudes generally remain constant throughout a political career and that

GETTING A POOR RETURN

these attitudes will not change if an individual moves from one branch of
government to another—for example, from Congress to the judiciary. That
is the premise and insight behind this measure of judicial ideology (Howard
2008; Howard and Nixon 2003; Nixon n.d.).

In fact, there have been sixty-three federal judges since 1938 who have

also served in Congress. These members of Congress have held positions on
almost every part of the federal bench, from Supreme Court to the Federal
Courts of Appeals to the U.S. District Courts, to the U.S. Tax Court and
also other specialized courts. Initial evidence for consistency can be shown
from Segal-Cover scores for Supreme Court justices, which again remain
constant throughout the career of the justice. Although there are only
four Supreme Court justices with prior congressional experience—Minton,
Burton, Vinson, and Black—the correlation between their fl oor voting and
Segal-Cover’s ideology measure is extremely strong (Pearson = –.946***).
For the thirteen circuit court judges for whom a legislative career exists, the
correlation between their fl oor voting and bench behavior (based on Songer’s
Appeals Courts Database) is also very strong (Pearson = –.621***).

Nixon (n.d.) assembled ninety-fi ve executive appointees who have also

served in Congress since 1938 and demonstrated that their common-space
W-nominate scores are predicted well by SOP considerations at the time of
their appointment. Nixon, along with Nixon and Howard (2003), extended
Nixon’s insight by assembling the collection of the sixty-three federal judi-
cial appointees who have also served in Congress since 1938 (see Howard
and Nixon 2003 for additional information) to create a model that predicts
common space W-nominate scores for all judicial nominees. The idea behind
the measurement is to create a set of variables that predicts the ideology
of these sixty-three judges that can be used to ascertain the ideology of all
other federal judges. Table 3.1 represents that effort; it is a predictive model
that serves as the basis for judicial ideology measures used throughout the
book. The model was developed by David Nixon (n.d).

Specifi cally, the ideology of a federal judge is calculated using a com-

bination of the party of the judge, the party of the appointing president,
region, unifi ed government, and whether there was a war, and ideology of
the home state of the appointee with various weights assigned to each of
these categories. Interestingly, what are often thought of as the two primary
determinants of judicial ideology—separation of powers constraints and
senatorial courtesy—do not make a signifi cant contribution to the model
(F

2,54

= 2.8, n.s.). Most likely, this is because some of the independent vari-

ables, such as judicial and presidential party, serve as proxies for separation
of powers constraints while the regional and state ideology variables act
as proxies for senatorial courtesy. Unlike many other lower court ideology
measures, this formula allows and provides for substantial variation between
judges and appears to predict lower court judicial voting quite well. In addi-

TAX LITIGATION AND TAX FORUM CHOICE

tion, since the scores are scaled in the same issue space and on the same
metric as W-nominate scores, this allows for direct comparison across institu-
tions. Bailey and Chang (2002) have demonstrated the pitfalls of failing to
employ strictly comparable ideology measures in cross-institutional models
of politics. This measure of ideology is used for this chapter and the other
chapters, with judicial ideology as an explanatory variable.

Precedent, attitudes, and national political factors are not the only

aspects that infl uence litigation and forum choice. Previous studies have
shown that the economy will infl uences litigation and forum choice. To
account for the economy, I used several different measures. Included were
per capita income, average IRS post audit assessments, and state unemploy-
ment. Per capita income and post audit assessments are measures of wealth.
The post audit assessment used in this study was limited to the assessment
from a single tax year, not an aggregation of prior year assessments. The
logic of using a measure of wealth is that the wealthier the state, the more
it is likely that the taxpayers in the state, on average, can afford to litigate
in the more expensive District Court.

Unemployment is a proxy for lack of wealth. Unemployment should

be greater in poorer states, and the higher the unemployment, the more
likely that taxpayers will litigate since there will be more Earned Income Tax
Credit (EITC) audits. Increases in EITC audits should lead the Joy Anders’s
of the country to challenge the loss of this credit. This in turn will lead
to increased litigation in the Tax Court as opposed to the more expensive
District Court since the Tax Court will be the only forum available for poor

Table 3.1 Model of Judicial Ideology

Variables Coeffi

cients

Constant –.0306
Judge’s own party (–1 Dem., 0 ind., + 1 Rep.)

.2371

Appointing president’s party (–1 Dem., +1 Rep.)

.0448

Unifi ed gov’t at time of appointment (0—no, 1—yes)*

.0249

Wartime appointment (0—no, 1—yes)

.0694

Southern Democrat (0—no, 1—yes)

.1285

Northeastern Republican (0—no, 1—yes)

–.0151

Wright/Erikson/McIver State Ideology

.3999

= .75.

*Positive values for Republican presidents, negative values for Democratic presidents
F test against inclusion of SOP Constraint and Senatorial Courtesy:
F = ((e*’e* – e’e)/m) / (e’e/(n–k))

m n k

restricted full

2.804, n.s.

(e*’e*’) (e’e’)

1.0054

.9108

GETTING A POOR RETURN

taxpayers. This should increase the Tax Court/District Court ratio. Finally,
I added a few other control variables. Beyond politics, ideology, and the
economy, certain demographic factors and state characteristics will infl uence
litigation and forum choice. Two such items are state population and the
number of audits per state. The greater the population and the greater the
number of audits, the greater the likelihood of litigation and litigation in
the Tax Court litigation (given the low cost to the taxpayer), all else being
equal. Since taxpayers choosing either court can fi le in the local District
Court I did not include a variable of distance from a courthouse.

Table 3.2 provides a complete list of the descriptive statistics for

the variables with means, standard deviations, minimums, and maximums.
Filings vary greatly by state and the numbers also confi rm the popularity
of the Tax Court as the preferred forum choice, with more than thirty-six
times more lawsuits fi led in Tax Court than District Court on average per
state, although the numbers range from an eighteen-to-one ratio to more
than fi fty-to-one. The ideology of the Tax Court, the District Court, and
the Appellate Court is moderately conservative during this time period,
although the District Court shows greater variation, refl ecting a much
more diverse court. The Courts of Appeals and the Congress are also
conservative, although there is signifi cant variation between the circuits.
Congress, beginning with the Republican takeover in January 1995, is
more conservative than the Tax Court and District Court. The win rates
for both the Tax Court and the District Court are the same for the years
of this study—seventy-nine percent—although there is greater variation in
the District Court win percentage.

Table 3.2 Descriptive Statistics Filings and Forum Choice 1994–2000

Variable Mean

St.

Dev.

Min

Max

Total Petitions

2015

101

3523

Tax Ct./D. Ct. Ratio

36.69

10.54

18.22

51.21

Appellate Court Ideology

.12

.19

–.34

.46

Tax

Court

Ideology

.10 .02 .07 .15

District Court Ideology

.09

.08

–.08

.38

Congressional

Ideology

.41 .21 –.08 .54

Tax Court Win Pct.

.79

.21

D. Ct. Win Pct.

.79

.32

Assessments 34.79

17.54

126.50

Audits

23414 21098 4111 195239

Population 5427

5927

502

34000

Per Capita Income

24540

4186

16392

41446

Unemployment

Rate

4.6 1.3 2.1 9.3

TAX LITIGATION AND TAX FORUM CHOICE

RESULTS

Litigation Rate Analysis

The results

are presented in Table 3.3 and show the change in the prob-

ability of litigation occurring from one standard deviation below the mean
for each specifi c factor (or variable) to one standard deviation above the
mean. In the case of 1998, the table shows the specifi c infl uence of the
legislation for the year 1998 (“1”) compared to not showing this infl uence
(“0”). Thus, for continuous independent factors or variables, the probability
represents a shift of two standard deviations, while for 1998 the “dummy”
factor or variable measures the effect of moving from 0 to 1. An asterisk
represents whether the infl uence of the factor showed statistical probability
of occurring by chance at fi ve percent or less.

As hypothesized, potential litigants responded to the more conserva-

tive Congress, conservative judges, and the legislation in deciding whether
to challenge IRS assessments. Politics dominate the process. The shift from
Democratic control of Congress to Republican control, which corresponds
to the two standard deviation shift in ideological change, almost doubled
(ninety-four percent) the probability of litigation. As the Tax Court becomes
more conservative using the same two standard deviation shift, litigation
increases by almost two-thirds. Two other measures also led to an increase
in litigation. The 1998 legislation and the measure of vertical tax court
precedent, the win rate, both increased litigation by more than one-third.

Table 3.3 Determinants of Tax Litigation 1994–2000

Variable Probabilities

(percent)

Lag of Tax Filings

34%

Tax Court Ideology

65%

District Court Ideology

–12%

Congressional Ideology

93%

Tax Court Win Pct.

34%

D. Ct. Win Pct

.05%

Assessments –1%
1998 36%
Audits 2%
Citizen Ideology

Population 0%
Per Capita Income

–2%

Unemployment Rate

Appellate Court Ideology

GETTING A POOR RETURN

Except for unemployment, other economic and demographic factors showed
little infl uence on tax litigation. However, unemployment means that many
people are out of work and relying on such things as the EITC in their returns.
Given that, an increase in unemployment leads to greater litigation.

Forum Choice Analysis

These results are reported in Table 3.4 and, similar to the litigation analy-
sis, they show the change in the probability of forum choice with a shift
from one standard deviation below the mean for each specifi c factor (or
variable) to one standard deviation above the mean. In the case of 1998,
the table indicates the specifi c infl uence of the legislation for the year 1998
(“1”) compared to not showing this infl uence (“0”). Thus, for continuous
independent factors or variables, the probability represents a shift of two
standard deviations, while for 1998 the “dummy” factor or variable measures
the effect of moving from 0 to 1. Again, an asterisk represents whether the
infl uence of the factor showed statistical probability of occurring by chance
at fi ve percent or less.

The results are very similar to the results for litigation. The results show

that litigants do act rationally in choosing a forum and that the national
coalition is successful in encouraging forum shopping. As the ideology of
the Tax Court becomes more conservative, making it more favorable to the
taxpayer, more lawsuits are fi led in the Tax Court compared to the District
Court. Similarly, the conservative Republican takeover of Congress in the
1994 midterm elections led to signifi cantly more Tax Court litigation, as
did the legislative enactments of 1998 and 1999.

Table 3.4 Determinants of Forum Choice 1994–2000

Variable Probabilities

(percent)

Lag of Tax Filings

84%

Tax Court Ideology

165%

District Court Ideology

–.002%

Congressional Ideology

44%

Tax Court Win Pct.

District Ct. Win Pct.

–.05%

Assessments –.07%
1998 121%
Audits .06%
Population 2%
Per Capita Income

–.04%

Unemployment Rate

Appellate Court Ideology

.04%

TAX LITIGATION AND TAX FORUM CHOICE

The effect of the ideology of the Tax Court on the ratio of the Tax

Court fi lings to District Court fi lings is considerable. As one moves a stan-
dard deviation below the mean to a standard deviation above it, there is
a corresponding increase in Tax Court fi lings of 165 percent. This means
that, in a typical state, as the Tax Court becomes more conservative, the
ratio moves from about eighteen Tax Court fi lings to every District Court
filing to more than eighty-three Tax Court filings for every District
Court fi ling. The District Court coeffi cient, while not statistically distin-
guishable from zero, was in the expected direction. As the District Court
becomes more conservative, taxpayers shift fi lings to the District Court and
away from the Tax Court.

Although the impact of a more conservative Congress and the legisla-

tion of 1998 and 1999 are not as pronounced as the infl uence of the ideology
of the Tax Court, these variables cause signifi cant shifts in the ratio toward
litigants using the Tax Court. As congressional ideology moves in a two
standard deviation shift from liberal to conservative, the ratio changes by
forty-four percent, or a shift from twenty-eight Tax Court fi lings for every
District Court fi ling to more than forty-four Tax Court fi lings compared to
District Court fi lings. The legislation of 1998 and 1999 had similar impacts
on Tax Court fi lings, increasing the ratio by more than one hundred percent.
On average, the 1998 and 1999 legislative changes increased the ratio to
about seventy-three Tax Court fi lings for every District Court fi ling.

Unlike the litigation measures, precedent, at least as measured by

win percentages, and appellate court ideology of the preceding year did not
infl uence the ratio. The control variables were in the expected direction
but, apart from the important lagged variable, were not statistically distin-
guishable from zero. State poverty or state wealth seems not to matter in
forum choice.

CONCLUSION

Joy Anders does not have much choice in her decision of where to litigate,
and once help was given, if affordable, she has little to lose by suing the
IRS, although if she does lose, she will have to pay all accumulated penalties
and interest. As the aggregate numbers in this chapter show, the Tax Court
was her only alternative. She had neither the means nor the wherewithal
to litigate in the District Court. Litigating itself was diffi cult for Joy Anders,
but by suing the IRS she can place her blame on the agency rather than
the legislation and congressional backing that created such confusing rules
regarding the earned income tax credit and encouraged the IRS to audit
EITC claims like hers.

Instead of blaming Congress and the political majority for increas-

ing IRS attention to needy individuals such as herself, Joy can blame the

GETTING A POOR RETURN

IRS and receive some relief from sympathetic judges. The analysis in this
chapter also provides some insight as to why a multinational corporation
would forego the seemingly easier-to-win District Court and instead sue in
the Tax Court. Even without calculating precedent, the results suggest that
Wal-Mart has a greater likelihood of winning in the Tax Court.

Political institutions can use litigation and forum choice to attain

policy goals. The likelihood of winning, along with signals from dominant
political coalition as to the possibility of winning in a particular forum, lead
low-income taxpayers in particular to litigate and to choose the Tax Court
over the District Court. Taxes and tax policy have signifi cant ideological
dimensions, with conservatives favoring not just fewer taxes but greater
constraints on the ability of the primary tax enforcement agency, the IRS,
to collect taxes and enforce tax policy. When tax litigants have a choice
of forum, it is rational for them to select the more conservative forum
because that will be the one most favorable to the taxpayer and least likely
to support the IRS.

With the shorter tenure of judges on the Tax Court compared to the

lifetime tenure of the District Court, it is easier for the governing national
coalition to move the ideology of the Tax Court much more quickly than
that of the Article III District Court. In addition, the national coalition
can implicitly and explicitly signal its preferred choice through appoint-
ments and legislation.

Critics deride policy making through litigation, and forum choice is

often dismissed as “forum shopping.” It is ridiculed because the idea that
litigants and their lawyers can play games with our justice system to create
favorable outcomes goes against ingrained notions of justice and fairness and
that, in the words of Justice Harlan, “all citizens are equal before the law”
(dissent in Plessy v. Ferguson 163 U.S. 537, 559–60). The ability to choose
a forum means that not all citizens are equal before the law because the
law differs from one venue to another. However, at least in tax litigation,
Congress has made the explicit decision to allow taxpayers to “game” the
system and, certainly in the 1990s, seemed to encourage taxpayers to choose
a forum that offers them the greatest likelihood of winning against the IRS.
So it appears that litigation and forum shopping are only bad until they help
the governing national coalition achieve political and policy goals.

TAX DECISION MAKING

FOUR

TAX DECISION MAKING

Although individual cost benefi t analyses and aggregate political forces
push taxpayers to challenge post audit assessments and lead to selection
of one tax forum over another, the question remains whether there are
actual differences in judicial decision making in these different tax forums.
Obvious differences exist in structure and membership. Do these differences
lead to different outcomes? Is the theoretically less independent Tax Court
biased in favor of the IRS? Is a more conservative court biased in favor of
the taxpayer?

These are intriguing questions because they call into question the

independent nature of the tax judiciary. It is almost a universal given that
an independent judiciary is essential to democratic governance (Russell
and O’Brien 2001). Montesquieu, for example, noted that the separation of
powers between the judicial and executive branches was far more important
than any separation between the legislative and the executive associated
with presidential democracies. Montesquieu reasoned that the most harm
that can be caused by governments is perpetrated by executives; the only
check on them is a resolute judiciary in which ordinary citizens ideally can
protect their civil rights.

Alexander Hamilton echoed these ideas. He suggested that the very

purpose of judicial independence was to counterbalance majoritarian will.
Hamilton argued that “the complete independence of the courts of justice
is peculiarly necessary . . . [because] the courts were designed to be an
intermediate body between the people and the legislature, in order, among
other things, to keep the latter within the limits assigned to their author-
ity” (Rossiter 1961, p. 457). The independence of the judiciary operated
as a “safeguard against the effects of occasional ill humors in the society”
(Rossiter 1961, p. 470).The framers recognized this importance by creating
an independent judiciary in Article III of the U.S. Constitution, with salary
and tenure guarantees.

GETTING A POOR RETURN

So, given that many taxpayers, such as Joy Anders, have no real choice

where they litigate, do they confront a stacked deck favorable to the IRS
with little chance of taxpayer success or is the Tax Court a less costly venue
that decides cases in a manner similar to that of the independent Article
III court? If in fact the District Court offers a more favorable venue to the
taxpayer, then that represents another barrier to the poor and even middle
class from obtaining relief from an improper assessment. Does Wal-Mart or
James William stand a better or worse chance of winning their case in the
Tax Court compared to the District Court?

In answer, this chapter compares the decision making of the fi xed term

Tax Court and the District Courts. Using data from 1996 and 1997, the
chapter examines the differences in decision making of these courts, specifi -
cally focusing on the meaning and terms of independence in determining
outcomes in these two courts.

COURTS, LAW, AND POLITICS

The Tax Court is not the only Article I court. Despite the acknowledged
value of judicial independence on several occasions, Congress has used its
Article I legislative powers to create specialized courts. Judges who serve on
Article I courts sit for a fi xed, as opposed to a lifetime, term and lack the
constitutional protection of a salary guarantee while the judge serves on the
bench. For example, the Bankruptcy Court and the U.S. Court of Federal
Claims are courts of limited jurisdiction and tenure.

As shown in chapter three, Congress can use litigation as a method

of policy change and control. Arguably, Congress can exert similar change
and control through court creation. By limiting the independence of a court
designed to hear tax appeals, Congress should be able to exert greater control
over the outcomes of this court than the more independent and much larger
court of general trial jurisdiction, the U.S. District Court. Most obviously,
a shorter tenure means that judicial replacement occurs at a quicker pace
than at a lifetime-tenured federal court.

Perhaps it is not surprising that Congress seeks greater control over taxes

than other policy issues that reach federal courts because Congress microman-
ages the IRS to a greater extent than other regulatory agencies (Howard and
Nixon 2002). While political science has not ignored this policy domination
(see, e.g., Quinn and Shapiro 1991; Scholz and Wood 1998, 1999), relatively
little attention has been paid to the specifi cs of tax decision making by the
courts, and whether one group of taxpayers receives different treatment in one
type of court as opposed to another. Is there greater control by, or ideological
concordance with, the elected branches of government with the decisions of
the Tax Court than with the decisions of the District Court?

TAX DECISION MAKING

Scholars have long acknowledged the importance of ideology in judicial

decision making for the U.S. Supreme Court and for lower courts (Hettinger,
Martinek, and Lindquist 2004; Segal and Spaeth 2002), to the extent that it
is now a scholarly given (Rowland and Carp 1996). The ability of a justice
to use ideology is premised on judicial independence, with the guarantee of
lifetime tenure and the lack of any realistic constraints (Segal and Spaeth
2002). However, unlike Supreme Court rulings, decisions of lower courts
are a function of ideology and additional constraints such as hierarchical
precedent and institutional structure. For example, Songer, Cameron, and
Segal (1995) show that federal appellate courts responded to conservative
decisions of the U.S. Supreme Court with increased conservative decisions,
even controlling for appellate ideology. Studies of state supreme courts show
how institutional structure and design infl uence decision making, with
elected judges and judges facing retention or reelection less likely to vote
their sincere preferences (Brace and Hall 1990; Hall 1992).

As noted, specialized courts are another example of courts within

the federal hierarchy. Unlike the courts created under the judicial power
of Article III, Congress creates Article I courts to deal with certain specifi c
subject areas. Usually these matters involve complex, frequently litigated
issues (Baum 1990; Hansen, Johnson, and Unah 1995). While expertise
is undoubtedly a consideration in their creation, the failure to guarantee
lifetime tenure or salary allows greater legislative and executive control
over outcomes in comparison to the outcomes of a more general, lifetime-
tenured, independent Article III court (see generally McCubbins, Noll, and
Weingast 1987).

The Tax Court deals with a large docket of post-audit assessment

cases. During the mid-1990s, on average more than 25,000 cases were fi led
per year. These cases often involve technically complex issues calling for
signifi cant judicial discretion in meaning and interpretation. Few would
argue that the Internal Revenue Code is simple. One former IRS revenue
agent and author of tax preparation manuals urged his readers not to use
the Internal Revenue Code as a reference source for tax law because of its
sheer complexity and lack of accessibility (Wade 1986). For example, the
concept of “adequate documentation” of business expenses involves a large
amount of discretion beyond the extremes of no documentation or requiring
contemporaneous written evidence for every claimed deduction. The evalu-
ation of such claims demands some area expertise in addition to accounting
or bookkeeping skills. Complexity, caseload, and limited tenure imply greater
expertise in the specialized courts as well as potentially greater control of the
agency through the use of this expert judiciary and greater control of the
specialized court through a much shorter appointment time frame. It appears
to be a trade-off—there is greater expertise, but less independence.

GETTING A POOR RETURN

Scholars have examined specialized courts in other areas of law and

policy. Studies of the International Court of Trade and the Patent Court
(whose duties are now divided between the Patent and Trademark Offi ce
and Court of Appeals for the Federal Circuit) show that these courts behave
as theoretically expected. The specialized courts are less deferential to agen-
cies than the nonspecialized federal courts of general jurisdiction (Hansen,
Johnson, and Unah 1995; Unah 1997). Because the courts have expertise in
these areas of law they do not have to rely on agency interpretation. This
expertise extends to their relationship with hierarchically superior courts.
One scholar examining the now replaced Court of Customs and Patent
Appeals found that this court was signifi cantly less likely than the Federal
Circuit Court of Appeals to rely on Supreme Court authority when making
decisions (Baum 1995).

These fi ndings are consistent with theories of congressional design

and bureaucratic control. Creating a specialized court that has the ability to
disregard agency expertise could serve political as well as policy goals (see,
e.g., McCubbins, Noll, and Weingast 1987, 1989) by controlling the agency
when Congress has neither the time nor resources to do so. This presupposes
that the specialized court will enforce the policy goals of the democratic
branches of government. A specialized Article I court of limited tenure can
be aligned with the policy preferences of the legislature and the executive
in a much shorter time frame than can the more independent Article III
courts. Independent, lifetime-tenured Article III courts should be less subject
to control than specialized courts and therefore less likely to control the
agency in concordance with congressional and executive preferences.

TAXPAYERS, COURTS, AND OUTCOMES

This research on specialized courts and the bureaucratic control literature
both suggest that taxpayers have a greater probability of winning at the Tax
Court than at the District Court. The District Court judge should show a
greater reliance on IRS expertise as well as on relevant courts of appeals
decisions. The trade-off for litigating in the District Court and losing the
expertise of the Tax Court would be the greater variation in decision mak-
ing of the District Court due to the lifetime tenure of the independent
Article III court. Decisions should show a greater variance from legislative
and executive preferences.

However, despite the fi ndings of the previous chapter, many others

who have examined raw numbers argue that the taxpayer does better in the
Article III District Court than in the specialized Article I Tax Court. Some
studies show that the taxpayer wins only fi ve percent of the time in Tax
Court compared to twenty percent to thirty percent in the District Court,
with other studies acknowledging at least a twenty percent differential (Geier

TAX DECISION MAKING

1991, p. 998). The data used for the years of this study do show a twelve
percent differential, with taxpayers winning twenty percent of the time in
Tax Court and thirty-two percent of the time in the District Court. This
would seem to imply that the less well-off are at a distinct disadvantage.
Only those who can afford to prepay their tax liability can afford to litigate
in the more favorable District Court.

Of course, this begs the question of why there are different rates of

success. Some argue that the Tax Court is, contrary to research on other
specialized courts and bureaucratic control literature, biased in favor of the
IRS and less subject to democratic control than its limited tenure Article I
status would suggest (Kroll 1996; but see Maule 1999). One court scholar
notes that it is this potential for bias that has led Congress to resist creat-
ing other specialized courts (Baum 1990), and another suggests that it is
the reason Congress has not made the Tax Court an independent Article
III court (Dubroff 1979). While scholarship has failed to develop a coher-
ent theory for why such bias exists, posited reasons include such factors
as ideology, institutional design and structure, prior IRS work experience,
the type of litigant, attorney representation, and even social and personal
characteristics of the presiding judge.

The failure to develop a coherent theory stems in large part from a

focus on bias as a product of structural constraints and agency dominance
rather than a focus on bias due to policy and democratic preference. If
there is a bias in favor of the agency at any given time, it is because the
elected branches of government want such a bias; if there is a bias in favor
of the taxpayer, or class of taxpayer, it is because that is what the elected
branches desire.

One preliminary way to answer these questions is to examine some

descriptive information about Tax Court judges. Tax Court judges, like other
federal court nominees, are appointed by the president and confi rmed by the
Senate. Although the nominees have expertise in tax law and considerable
prior experience in the tax fi eld (often with the IRS), the nominees can
also have signifi cant political background and experience (Tidrick 2004).
In Table 4.1, I list the initial appointment date and the background of all
current full-time and senior judges on the Tax Court. Senior judges are
retired judges sitting by designation and are denoted by an asterisk next
to their names.

After the initial appointment date, the next column shows whether

the judge worked in some capacity for the IRS at some time prior to the
appointment. The third column displays whether the judge had a special-
ized skill background, either obtaining an LLM in taxation or having an
accounting background in addition to a law degree. The fourth column
lists whether the judge had private tax experience, which means the judge
at some point worked as a tax attorney for a private law fi rm. The fi nal

GETTING A POOR RETURN

column lists governmental experience exclusive of the IRS. For example,
the judge might have been a counsel to a congressional tax, fi nance, or
ways and means committee. The bottom three rows list the percentages for
each column for all the judges. I then divide the percentages into senior
judges and regular judges.

Several interesting details emerge. First, and contrary to common per-

ception, on the modern Tax Court only a minority of judges have actual IRS
or government experience. Most judges have a private practice background,
and most also have a specialized background in taxation. However, when

Table 4.1 Tax Court Judge Experience

Initial

IRS

Specialized

Private

Govt.

Judge

Appt.

Date Experience Background Experience Experience

John O. Colvin

9/1/88

Yes

Mary Ann Cohen

9/24/92

Yes

Maurice

Foley

4/9/95 Yes Yes No Yes

Joseph H. Gale

2/6/96

Yes

Joseph

Goeke

4/2/03 Yes No Yes No

Harry A. Haines

4/22/03

Yes

James S. Halpern

7/3/90

Yes

Mark V. Holmes

6/30/03

Yes

Diane L. Kroupa

6/13/03

Yes

L. Paige Marvel

4/6/98

Yes

Stephen J. Swift

12/1/00

Yes

Michael

Thornton

3/8/98 No Yes Yes Yes

Juan F. Vasquez

5/1/95

Yes

Thomas B. Wells

10/13/86

Yes

Robert

Wherry

4/23/03

No Yes Yes No

Renato

Beghe*

3/26/91

No No Yes No

Herbert

Chabot*

4/3/78 No Yes Yes Yes

Carolyn

Chiechi*

10/1/92

No Yes Yes Yes

How.

Dawson*

8/21/62

Yes No Yes No

Joel Gerber*

6/18/84

Yes

Julian I. Jacobs*

3/30/84

Yes

David

Laro* 11/2/92

No Yes Yes No

Arthur Nims, III*

6/29/79

Yes

Robert P. Ruwe*

11/20/87

Yes

Laur.

Whalen*

11/22/87

Yes Yes Yes No

Percent

All

44% 60% 88% 40%

Percent

Regular

33% 53% 93% 53%

Percent

Senior

60% 70% 20% 20%

* Senior Judge

TAX DECISION MAKING

one examines the differences between the senior and regular judges, one
observes much greater disparity and this examination shows a change in the
background of the judges over time. The majority of the senior judges do
have IRS experience, but only a small percentage worked in private practice.
That is, on average, the older judges have a very different background from
the newer judges. The newer judge is unlikely to have worked for the IRS,
but instead has both private fi rm experience and has worked for a congres-
sional committee. This would suggest that the newer judges would not have
a bias to the IRS and, if anything, would be more sympathetic to both the
taxpayer (because of their private fi rm background) and to the preferences
of Congress (because of their governmental background).

The appointment process and background mark Tax Court judges as

similar to nominees of the other specialized courts (see Unah 1998). While
expertise is important and necessary, the process of appointment is political.
Republicans want Republican and conservative judges and Democrats want
Democratic and liberal judges. This makes the nomination and confi rmation
process of advice and consent for Tax Court judges the same as it is for
other federal judges (Epstein and Segal 2005). As previously mentioned, an
earlier study found a signifi cant positive correlation between judicial liberal-
ism and support for the IRS (Howard and Nixon 2002). Given their lack
of independence with the concomitant fewer number of judges and shorter
tenure, in the time frame of this study one would expect the Tax Court to
decide cases on ideology and personal policy preferences at least as much
as the longer-tenured, larger, more independent District Court.

This can move beyond just conservative ideology favoring the taxpayer;

it can also mean that certain judges are far more likely to favor wealthy
taxpayers. This preference would be in accordance with tax policy, which,
at least since 1981, seems to favor the wealthy at the expense of the less
well-off (Johnston 2003). Given the prominence of antitax and anti-IRS
rhetoric from 1981 onward, one would expect ideological and pro wealthy
Tax Court IRS rulings to increase, with liberals supporting tax collection
efforts and conservatives opposing such efforts, and court rulings in general
supporting wealthy litigants.

Because there are more judges, and these judges have a longer tenure,

the independent Article III District Court should not exhibit the same
ideological tendencies in its decisions. The shorter tenure of the Tax Court
allows greater control of the court outcomes than the longer time-framed,
much larger District Court. The ability of the dominant political coalition
to align the judiciary with its policy preferences was observed many years
ago by Robert Dahl (1958), who noted that presidents appoint about two
justices every four years to the U.S. Supreme Court. Because of this, even
with lifetime tenure, the very small, nine-member Supreme Court is “inevi-
tably” part of the dominant political coalition (1958, p. 293).

GETTING A POOR RETURN

Tax Court judges also know the law and do not have to rely on

hierarchically superior courts, the agency, or lawyers for interpretation
and meaning. This allows Tax Court judges greater freedom in the range
of acceptable decision making. This would be in keeping with political
objectives of the dominant political coalition through the ability of the
court to control IRS behavior. Tax policy is controversial and liberals and
conservatives would want to control that policy by controlling court out-
comes, which in turn can control the IRS. Arguably, Congress has designed
a court with enough freedom and expertise to be independent of the IRS,
but also one that can be controlled by the dominant coalition through
the quicker time frame appointment process if the ideology is not to the
coalition’s political preferences.

There are really no obvious institutional constraints restricting the

ideological choices of the Tax Court that should lead to a bias in favor of
the IRS (Posner 1996, p. 268; see also Easterbrook 1990). Tax Court judges
enjoy full pension protection, usually are reappointed or assume senior status,
and have the same salary as District Court judges. In practice, Tax Court
judges enjoy the same protections as the Article III judges (see Dubroff 1979;
Maule 1999) with the exception of special judges, who are appointed by the
current chief judge of the Tax Court, and, of course, with the exception of
lifetime tenure. While, in the past, many judges came out of the IRS before
assuming Tax Court judgeships, that is not the case with the modern court
and many have had experience as counsel to various congressional tax and
fi nance committees. Thus, preferences of the Tax Court should align more
quickly with the preferences of the Congress and the president.

An examination of some data that measure the ideology of the Senate,

the Tax Court, and the president supports this argument. Figure 4.1 presents
the ideology of the president, the ideological fi libuster point of the Senate,
and the average ideology of the Tax Court from 1983 to 1998. The senate
and presidential scores are the fi rst-dimension “common space” W-nominate
scores (Poole 1998), and indicate economic conservatism of the presidents
and members of the Senate.

This time frame not only corresponds to ideological polarization on tax

policy, but also the opposite party control of the presidency and the Senate.
Not surprisingly, the mean ideology of the Tax Court lies in between that of
the executive and the Senate over this time period, becoming more liberal
since the reelection of President Clinton. Given the partisan change at the
executive and legislative level, one would expect that appointees of Reagan
and Bush would be far more supportive of the taxpayer and appointees
of President Clinton more supportive of the IRS. The appointees of the
president would be subject to the constraint of Senate confi rmation, with
the appointee needing to survive any fi libuster attempt. The mean result
is the slightly conservative line, with more liberal appointees bringing the

TAX DECISION MAKING

average down during the last few years of the Clinton administration as
depicted in Figure 4.1.

To provide additional evidence of democratic control and concordance

with dominant political coalition preferences, I calculated the Pearson cor-
relation between the economic conservatism scores of the Senate fi libuster
point, the president, and the average for the Tax Court. I also derived
average ideological scores for District Court judges using the comparable
W-nominate measure detailed in chapter two during this time period and
calculated the Pearson correlation between the economic conservatism
scores of the Senate fi libuster point, the president, and the average for the
District Court. For the Tax Court and the president the correlation is .780
and signifi cantly different from zero at the .01 level; for the Tax Court and
the Senate fi libuster point the correlation is .573 and signifi cantly different
from zero at the .05 level. These correlations indicate substantively and
statistically positive associations between the higher economic conservatism
of the president, the Senate, and the Tax Court.

In contrast, the Pearson correlations between the economic conser-

vatism scores of the Senate fi libuster point, the president, and the averages
for the District Court judges show no such relationship. For the District
Court and the president, the correlation is .132, while the correlation for
the Senate fi libuster point and the District Court is –.094. Neither correla-
tion is signifi cantly different from zero at any meaningful level. The more
independent Article III court shows greater independence from the executive
and legislative branches than the Article I court.

Figure 4.1 President, Senate, and Tax Court Ideology 1983–1998

-1

-0.7

-0.4

-0.1

0.2

0.5

0.8

1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998

President

Senate

Tax Court

GETTING A POOR RETURN

Previous examinations of ideological decision making by the Tax Court

demonstrate either a weak or nonexistent correlation between ideology and
tax decision making (King and Lazarus 2003; Schneider 2002, 2001). These
fi ndings, in all likelihood, stem from measurement error. Consistent with
previous research, the Schneider studies, along with the King and Lazarus
studies, use the partisanship of the appointing judge and the partisanship
of the judge, respectively, as their ideology measures. While a judge’s own
partisan affi liation and the ideology of a judge’s appointing president have
often been employed as useful surrogates of judicial attitudes (Segal and Spa-
eth 1993), focusing on the partisanship of judges restricts possible ideology
indicators that fail to account for the subtlety and diversity of attitudes on
the bench, and does not take into account other factors such as the control
of the Senate (Maule 1999), region, and home state senator ideology.

Previous studies of court tax decisions also have failed to account in

any signifi cant way for the infl uence of facts and issues, besides separating
out litigants into three main types—businesses, individuals, and estate and
trusts—on the theory that businesses and estates have more complex tax
returns than the tax returns of individuals. Specifi c issues or facts can tem-
per or trigger attitudes (see Segal and Spaeth 1993). Cases involving the
meaning of what constitutes income, or what are proper deductions or the
valuation of stock, can trigger attitudinal responses in a judge leading to a
greater or lower probability of fi nding for the taxpayer.

In addition, both courts deal with salient, ideologically charged mat-

ters. For example, a percentage of cases that come before the Tax Court
involve tax protestors. These individuals assert that the income tax is
unconstitutional and that the IRS has no authority to collect it. Another
series of cases involve the assertion of tax frauds. In these cases, individu-
als or businesses are accused of illegally hiding income or asserting patently
false deductions. This calls for the judge to fi nd the line between legitimate
tax avoidance and illegal tax evasion. Tax shelters have also come under
increased scrutiny. A shelter is a fi nancial arrangement used to reduce the
investor’s tax liability, with some designed to lose money for potentially
greater tax savings.

Highly relevant to comparisons of the Tax Court and District Courts

is that, although some matters appear in both courts, the two courts at
times deal with different issues. For example, District Courts often consider
jurisdictional questions. This occurs when litigants have failed to pay their
tax assessment and thus the District Court declines to hear the case, forcing
taxpayers to have their claim heard by the Tax Court. The District Court
also litigates withholding issues. Employers are required to withhold taxes
from their employees and submit these taxes on a periodic basis. The District
Court hears cases when an employer fails to deposit income tax withheld
from employees or tries to avoid withholding taxes by claiming that the

TAX DECISION MAKING

workers were independent contractors, not employees. In contrast, the Tax
Court often deals with innocent spouse issues. This situation occurs when
a spouse signed a joint return, but claims lack of knowledge of the contents
so as to not be held responsible for penalties and interest. In addition, the
court addresses valuation issues, particularly for gifts, estates, and trusts.

There are other differences between the courts besides hearing different

types of issues. Even though the Tax Court can try a case in the jurisdic-
tion of the taxpayer and the decisions can be appealed to the appropriate
regional Federal Court of Appeals, it is a national court, with Tax Court
judges assigned to the local area. The District Courts are not monolithic.
They meet in very different geographic areas with different rules of proce-
dure and are very culturally diverse. The District Courts are located within
each state, and the judges are appointed from that particular state. Previous
research has shown that District Court decision making is subject to regional
and political effects (Rowland and Carp 1996; see also Giles and Walker
1975). In particular, the District Court should be more infl uenced by the
state political and economic climate.

HYPOTHESES

Given the expected prevalence of ideological decision making, as well as
the differences in independence and structure, I derive several hypotheses,
which I categorize into two subgroups: independence hypotheses and spe-
cialization hypotheses.

Independence

Since tax policy, along with support or opposition to the IRS is ideological,
one would expect judges to exhibit ideological bias in their rulings. Tax
Court judges are subject to the same nomination and confi rmation process
as judges of the District Court; therefore, one should see similar ideological
rulings for both courts, regardless of their relative independence. Thus, the
fi rst hypothesis:

Hypothesis 1: The more liberal the judge, the greater the support for
the IRS; the more conservative the judge, the greater the support
for the taxpayer.

However, tax policy and its interpretation are is complex and demand-

ing, and there are many fewer Tax Court judges and their tenure is limited.
With shorter tenure, the ideological positioning inherent in tax policy would
be more likely to manifest itself in the smaller Tax Court than in the larger
District Court where ideological change would take place at a much slower

GETTING A POOR RETURN

pace over a much longer time. In addition, judges of the general jurisdiction
courts need to rely more on litigants, lawyers, the IRS, and other courts for
the meaning and proper construction of the Internal Revenue Code. This
reliance on outside interpretation will restrict the use of ideology in the
rulings by the judges of the District Court. The expertise of the Tax Court
judges, and their concomitant lack of reliance on lawyers, litigants, the IRS,
or other courts, means that the Tax Court judges have greater freedom to
use their ideology in their rulings, leading to hypothesis 2:

Hypothesis 2: Ideology will have a greater infl uence in the Tax Court
than in the District Court.

Of all the issues and facts confronting the judges of both courts, the

greatest challenge to the collection of taxes comes from those who question
the very notion of the process and the right of the state to collect taxes
either through fraud, deception, or abuse. The most fl agrant challenges to
the tax collection system come from those who commit fraud, create or use
abusive tax shelters, or engage in protests against the right of the state to
assess and collect revenue. More than anything else, these assertions should
lead to signifi cant divergence between liberal and conservative judges, with
liberal judges being the most opposed to these assertions and arguments and
conservative judges being more supportive. Thus, hypothesis 3:

Hypothesis 3: The more liberal the judge, the greater the support
for the IRS when the issues concern tax protesters, tax fraud, and
tax shelters.

Specialization

Since prior research has shown that other specialized courts are less likely
to follow precedent imposed by hierarchically superior courts than courts of
general jurisdiction, there is no reason to expect the Tax Court and District
Court to behave differently; hence, hypothesis 4:

Hypothesis 4: The regional federal courts of appeals will have a greater
infl uence on the decisions of the District Court than on the decisions
of the Tax Court.

Given the expertise of the Tax Court compared to the District Court,

one should see that attorneys have a greater infl uence on District Court
judges than on Tax Court judges. District Court judges need to rely on the
presentation and arguments of legal experts more than Tax Court judges,
leading to hypothesis 5:

TAX DECISION MAKING

Hypothesis 5: Representation by an attorney will have a greater infl u-
ence on the District Court than on the Tax Court.

Tax Court special judges do not have the same pension and reappoint-

ment advantages as a regular judge of that court. Because of this, special
judges should show greater deference to the agency. Following prior literature,
a judge’s prior IRS experience will also infl uence the decisions of special
judges and regular judges. One should expect that special judges would be
more likely to decide cases in favor of the IRS. Therefore, all Tax Court
judges with IRS experience, special or regular, will be more likely to decide
cases in favor of the IRS. Thus, hypotheses 6 and 7:

Hypothesis 6: Special judges will be more likely to decide cases in
favor of the IRS.

Hypothesis 7: Tax Court judges with IRS experience will be more
likely to decide cases in favor of the IRS.

Prior research on District Courts demonstrates the impact of local

conditions on the District Court. Because of this, the District Court deci-
sions should show the infl uence of state-specifi c factors. This leads to the
following hypotheses:

Hypothesis 8: District court judges will be more infl uenced by state-
level political and economic factors.

Finally, given the lack of expertise in the District Court, business and

estate and trust litigants should have greater success in the District Court
than in the Tax Court. This leads to the fi nal two hypotheses:

Hypothesis 9: Tax Court judges and District Court judges will be more
likely to decide cases in favor of business or estate and trust litigants
than individual litigants.

Hypothesis 10: Business and estate and trust litigants will be more
successful in the District Court than in the Tax Court.

To test the hypotheses, I collected both Tax Court and District Court

data, including all decisions of the Tax Court for the year 1996 and for all
tax cases from the District Court for the years 1996 and 1997.

Each case

was coded for the names and type of taxpayer, home state of the taxpayer,

case number, docket number, date of decision, the name of the judge, type
of judge, issue being contested, and if counsel was present. Similar to the

GETTING A POOR RETURN

procedure used in other chapters, each case was coded as either a win for
the IRS or a win for the taxpayer. For both courts, the docket number was
used as the unit of analysis. This coding led to an initial database of 681
decisions for the Tax Court and 207 decisions for the District Court.

Most of the other variables came from these cases. Litigant type

depended on who opposed the IRS: individual, business, estate or trust, and
partnership. For the litigants, a business was any incorporated entity, while
an estate and trust was any estate or trust or any individual who challenged
a gift tax determination. Each variable was coded as a separate dichotomous
variable, with a coding of 1 for the type of litigant, 0 otherwise. The cat-
egory of partnership was excluded to provide the baseline. Thus, all results
of litigation are in comparison to the omitted category of partnership.

I followed the same pattern for the issue section, with previously

mentioned issue differences for the Tax Court and District Court. Both
courts dealt with several of the same issues, including income, deductions,
tax shelters, tax protesters, and fraud. I combined the tax protestor/tax fraud
into one variable since both represent an assault on the legitimate collection
of taxes. Thus, for both courts, tax protestor/fraud, income, deductions, and
tax shelters were coded 1 if that was the main issue, 0 otherwise. I followed
the same coding rules for the different issues with which each court deals,
with 1 for all jurisdictional issues confronted by the District Courts, 0 oth-
erwise, and 1 for all withholding issues, 0 otherwise. For the Tax Court, I
coded 1 for a valuation issue, 0 otherwise, and 1 for an innocent spouse,
0 otherwise.

To account for the structural differences of the Tax Court, I created

a dichotomous variable for each case decided by a special judge (1, 0 oth-
erwise), and a variable to indicate if the judge had IRS experience (1, if
so, 0 otherwise). The assumption is that the more constrained special judge
will be more likely to rule in favor of the IRS. For both courts, I used a
variable to indicate whether an attorney represented the taxpayer (1, if an
attorney, 0 otherwise).

Other variables were then merged into this dataset. For the state

control variables I used the Berry et al. measure of state governmental
liberalism and a measure of state per capita income. State-level political
and economic measures have been used in prior studies of taxes and politi-
cal responsiveness (Scholz and Wood 1998, 1999), and the same logic of
responsiveness should apply to the District Court. Since they live and work
in the state, the judges of the District Courts should be more responsive to
the political environment (Giles and Walker 1975). A more liberal state,
as measured by Berry et al., should lead to a greater probability of a ruling
by the District Court in favor of the IRS. Similarly, a more wealthy state,
as measured by per capita income, should have more audits of wealthier
individuals, increasing the probability of more judges deciding in favor of

TAX DECISION MAKING

the IRS. To measure individual judicial ideology, again I used the measure
of the personal ideology of each District Court judge and Tax Court judge
comparable to Poole’s (1998) common-space W-mominate scores.

When a special judge renders a decision and a regular judge approved

that decision, I used the latter’s ideology score. This follows Tax Court Rule
182 (c), which states, “The Judge to whom . . . the case is assigned may adopt
the Special Trial Judge’s report or may modify it or may reject it in whole or
in part” and as the Supreme Court case of Ballard v. Commissioner of Internal
Revenue 125 S.Ct. 1270 (2005) makes clear, such reports can be signifi cantly
modifi ed. When there was no regular judge, I used the incumbent president
as the nominating president for the special judges, because the president
appoints the chief judge and chief judge appoints the special judges. If a District
Court judge approved a magistrate’s decision, I used the ideology score of the
approving District Court judge. In some jurisdictions, a magistrate judge can
hear civil tax cases with the consent of the parties, and appeals go directly
to the Court of Appeals. In other districts, the cases are assigned to both the
magistrate and a District Court judge and appeals go to the District Court
judge. Because of this, I coded the District Court judge’s ideology score for
decisions of the magistrate judge when both judges are listed.

If there was no approving judge, the magistrate’s ideology score was

used. The same process applied to the appellate court measure of ideology,
with the measure pegged to the median ideology of the relevant appellate
court. I then created interaction variables using these ideology scores and the
specifi c issue codings. Each of these was included in the respective analyses,
with the exception of the excluded variables serving as baselines.

The variables with means and standard deviations are listed in Table

4.2 (next page). Although the range of litigants and issues for each court
has many similarities, including mean ideology, with both being slightly
conservative, some differences stand out. One is the much greater number
of taxpayers who forego counsel in the Tax Court (thirty-six percent) as
compared to the District Court (ten percent). Another difference is that
more than twice as many tax shelter and tax protester cases are heard in
the Tax Court (twenty-one percent) compared to the District Court (ten
percent). Finally, the IRS wins twelve percent more often in the Tax Court
than in the District Court.

RESULTS

The estimated probabilities of an IRS win are reported in two separate
tables. The appendix contains the complete results. What emerges from this
comparison is confi rmation that the shorter-tenured, less independent Tax
Court is more ideological in tax rulings than the District Court. The lack
of independence through limited tenure and smaller size has allowed the

GETTING A POOR RETURN

appointment of more ideological judges, thus resulting in rulings premised
on ideology in keeping with dominant coalition political preferences.

Independence

The ideology and independence probabilities are listed in Table 4.3. As
hypothesized, the more liberal the judge, the greater the likelihood of support
for the IRS, while the more conservative the judge, the greater the likelihood
of support for the taxpayer. Also as hypothesized, the shorter-tenured, less
independent Tax Court judges rely on personal policy preferences more than
the independent lifetime-tenured District Court judges. The hypothesis that
liberal judges would react more negatively and conservatives judges more
positively to a tax protestor or tax fraud issue was confi rmed. However, con-
trary to the hypothesis, conservative judges are more likely to decide cases
against the taxpayer when there is a tax shelter than are liberal judges.

Since these are all continuous variables, the change represents an

increase of one standard deviation below the mean to one standard deviation
above the mean. Moving from one standard deviation below the mean to one
standard deviation above on the ideology score decreases the probability of an
IRS win by .16. Given a mean of .10, this means moving from a moderately
liberal judge to a very conservative judge. The tax protestor/fraud interaction
shows that moving from a moderately liberal judge to a very conservative
judge result is almost a sure bet that the IRS will lose the decision (–.88).

Table 4.2 Descriptive Statistics U.S. Tax Court and U.S. District Court

Tax

Court

District

Ct.

Mean

St. Dev.

Mean

St. Dev.

IRS Win

.80

.40

.68

.47

Per Capita Income

24,776.75

2755.22

24,776.75

2755.22

State Government

38.63

20.03

38.63

20.03

IRS Experience

.55

.50

Special Judge

.33

.47

Attorney .64

.48

.90

.30

Business .17

.38

.25

.43

Estate or Trust

.07

.25

.10

.30

Individual .70

.46

.62

.49

Income Issue

.27

.45

.22

.41

Deduction Issue

.27

.45

.19

.40

Tax Shelter

.16

.37

.07

.26

Appellate Court

.17

.04

.17

.04

Ideology .10

.23

.10

.26

TAX DECISION MAKING

This appears to indicate that the sentiments about the lack of legitimacy of
the tax collection process expressed by the most conservative legislators are
shared by the most conservative Tax Court judges, in contrast to the lack of
any fi nding in this area for the more independent District Court judges.

Specialization

The probabilities for specialization are listed in Table 4.4. The results confi rm
the expectations derived from prior research that general court judges are

Table 4.3 Ideology and the Probability of IRS Winning*

Variable Tax

Court

District

Court

Ideology –.19

.04

Income * Ideology

.10

.007

Deduction * Ideology

.14

–.05

Valuation * Ideology

.06

Tax Shelter * Ideology

.18

–.19

Fraud/Protester * Ideology

–.88

–.43

Withholding * Ideology

–.48

*For continuous variables, the probability represents an increase of one standard deviation
below the mean to one standard deviation above the mean. For dichotomous variables, it
represents a change from 0 to 1.

Table 4.4 Structure, Specialization, and Expertise and the Probability of
IRS Winning*

Variable Tax

Court

District

Court

Appellate Court

–.004

.14

Per Capita Income

–.003

.14

Attorney –.07

–.09

Business –.13

–.22

Estate or Trust

–.01

–.31

Income Issue

–.01

.001

Deduction Issue

.0001

–.10

Tax Shelter

.05

.15

Fraud/Tax protester

.19

.10

Valuation Issue

.02

Withholding

.23

*For continuous variables the probability represents an increase of one standard deviation
below the mean to one standard deviation above the mean. For dichotomous variables, it
represents a change from 0 to 1.

GETTING A POOR RETURN

more likely to rely on precedent than specialized court judges. The District
Court, but not the Tax Court, is strongly infl uenced by the appropriate
Federal Court of Appeals. In addition, special judges decide issues differently
than regular court judges. However, prior IRS experience does not matter.
Curiously, the presence of an attorney appears to help the taxpayer in the
more specialized Tax Court, but not in the District Court. This could be an
artifact of sample since counsel represented most taxpayers in the District
Courts, while more than one-third of the Tax Court litigants appeared pro
se. It could also represent the ability of the Tax Court judge to appreciate
and understand the arguments of counsel, whereas the more generalized
District Court judge relies on the agency. The state-specifi c variables had
no infl uence on the Tax Court decisions, while the wealth of the state did
seem to have some infl uence on the District Court judge.

The District Court judge was more sensitive to the type of litigant, with

a District Court judge having a greater likelihood of supporting a business
or a trust or estate than a Tax Court judge, although the Tax Court judge
had a greater probability of supporting a business litigant than an individual
litigant, with the effect smaller than that of the District Court. Finally, the
Tax Court judge had little tolerance for the tax fraud/tax protester while this
specifi c issue had no infl uence on the District Court. Moving from a circuit
with a moderate appeals court to a circuit with a much more conservative
court leads to a .14 increase in the probability of the District Court judge
ruling in favor of the IRS, while a moving from a relatively poor state to a
more affl uent state leads to the same .14 probability increase in the District
Court ruling in favor of the IRS. The type of litigant has a much stronger
infl uence on the District Court than the Tax Court. An estate or trust leads
to a –.31 decline in the probability of a decision in favor of the IRS, as
compared to –.01 for the Tax Court, while a business litigant leads to a –.22
decline in the probability, compared to a –.13 decline for the Tax Court.
Finally, the presence of a tax fraud or tax protester in the Tax Court leads
to a .19 increase in the probability of the IRS winning.

CONCLUSION

The results presented in this chapter show that decisions can be different for
independent courts with lifetime tenure compared to specialized courts with
greater expertise, but more limited tenure and independence. The Congress
and the president can use the Article I specialized court’s limited tenure and
reduced numbers to change the ideology of the court in a much quicker time
frame than a more general, independent Article III court. Controlling the
ideology of the specialized court can in turn lead to greater control of the
agency. Congress and the president can push the agency to produce policy in
line with congressional and executive preferences because they have greater

TAX DECISION MAKING

control over the decision making of the specialized court. Almost fi fty years
after Dahl (1957) noted that the Supreme Court is inevitably part of the
dominant political coalition, Congress and the executive can use Article I
courts to keep judicial preferences in agreement with the dominant coali-
tion preferences by aligning the ideology of the specialized court with the
ideology of the coalition.

This seems to deviate from the ideal that a specialized court is supposed

to use its expertise to decide issues without reference to ideology. However,
this ideal ignores the practical matter that the collection and distribution
of revenue are the single most important and politically charged issue that
any government must confront. Who or what should pay and how much,
and who or what should receive this revenue and how much, are inescap-
ably political questions charged with ideological overtones (see Johnston
2003). Despite repeated calls for simplifi cation and ease, any system of
taxation must have complex laws, rules, and regulations; its interpretation
of complicated questions of assessment and collection requires discretion
and technical expertise; and any result necessitates opinion and judgment,
and such opinion and judgment cannot be divorced from basic views about
the collection and allocation of scarce resources. This elaborate system of
appeals allows control of the IRS and ultimately tax policy.

Almost since the inception of the Tax Board of Appeals, there have

been various proposals to consolidate all tax cases in one court and to cre-
ate a U.S. Tax Court of Appeals to handle all cases from the lower courts
(Geier 1991; Posner 1996). While many reasons have been put forward to
oppose such proposals, an important consideration has been the fear that
such specialized courts would be no more than extensions of the IRS and
thus likely to exhibit bias in favor of the agency. However, while the Tax
Court might issue more rulings in favor of the IRS than the District Court,
the reason might be as much due to congressional design and executive
appointment to a more limited independent court, compared to an Article
III court, than it is due to bias in favor of an agency.

This page intentionally left blank.

INFLUENCES ON THE IRS

FIVE

INFLUENCES ON THE IRS AND THE

AUDITS OF LOW-INCOME TAXPAYERS

It was asserted in the beginning of this book that courts rarely confl ict with
the governing national coalition and instead are often used by the domi-
nant political coalition to accomplish national goals that cannot otherwise
be achieved through legislation or executive action. Joy Anders and oth-
ers in similar situations get audited because the dominant coalition wants
to audit those who claim the EITC. If Dahl and others are correct, then
court preference and resulting infl uence should show little difference from
national political preferences and resulting infl uence, and actually enforce
the preferences of the majority.

As the country’s tax policy moved to favor lower taxes for wealthier

individuals, tax enforcement should follow suit. In fact, this has occurred. In
1999, audits of low-income tax payers began to exceed audits of the wealthy
(Crenshaw 2002; Tracfed press release 2006), confi rming a decade-long
trend of increased IRS attention to the tax returns of low-income taxpayers
(Johnston 2003; TracIRS 2005). More recent news reports allege that IRS
investigations of tax fraud among the poor have resulted in signifi cant delays
in obtaining income tax refunds for many innocent low-income taxpayers
(Associated Press 2006), and that taxpayers with incomes lower than $25,000
per year are more than six times as likely to be audited as taxpayers earning
more than $200,000 per year (Tracfed press release 2006). A Joy Anders
getting audited is no accident and is instead the result of deliberate policy
choices. Congressional and executive public policy resulted in the increase
in audits of low-income taxpayers.

However, although we often see public policy and policy change as the

result of legislative and executive action, courts have long been acknowl-
edged as policy players (McCann 1994; Moe 1985; Segal and Spaeth 2002).
Most, but not all, of these examinations concern social policy, although

GETTING A POOR RETURN

there has been emerging research on the impact of courts on bureaucratic
and agency policy (Howard and Nixon 2002, 2003; Moe 1985; Ringquist
and Emmert 1999).

This chapter specifi cally examines claims about the role of courts

and how they impact and infl uence tax policy and tax enforcement. Court
protection of the poor is especially relevant due to these recent tax policy
and tax enforcement trends that have shown increased concentration on
the tax fi lings of the poorest taxpayers. Some members of Congress have
expressed concern over this development, implying that the agency is pur-
suing its own agenda (Crenshaw 2002). A policy that targets low-income
taxpayers says much about national mores and a commitment to equity
and fairness beyond the mere collection of revenue. The role of courts is
potentially critical in ensuring fairness and protecting the poor from the
power of government. However, if Dahl and others are correct, then courts
would push enforcement in the same direction as Congress. In other words,
Joy Anders and those in her income and tax bracket have little reason to
expect courts to ensure fairness.

In asserting court and policy maker alignment, Dahl was writing

about the U.S. Supreme Court. However, in examining tax policy and tax
enforcement, focusing on the Supreme Court or even the Federal Court of
Appeals causes one to miss the larger picture of courts, tax policy, and tax
compliance. Very few tax cases are appealed and higher courts review even
fewer cases involving low-income taxpayers. Most tax cases are disposed of
at the trial level by the District Court and the Tax Court. These courts, not
appellate courts, interact the most with the IRS and taxpayers and these
are the courts that low-income taxpayers turn to for relief against the IRS.
Do these courts protect the rights of the individual, particularly the low-
income taxpayer, or do they enforce dominant policy preferences? Do they
infl uence the IRS to change its audit behavior and focus less attention on
the lowest income group? Is there a difference in these courts in their rela-
tive infl uence? Little attention has been paid to these different trial courts
and their role in infl uencing tax policy. Perhaps this lack of attention is due
to the scholarly debate over the impact of courts on social and economic
policy (see, e.g., McCann 1994; Rosenberg 1991) or perhaps it is because
tax laws and tax policy are so complex that it is diffi cult to parse out the
infl uence of any one particular court on the IRS.

I analyze this by examining the separate infl uences of the national

political coalition, the Tax Court, and the District Courts on the number of
state-level audits of the lowest income class of taxpayers from 1982 through
2000. Understanding the enforcement of national policy is as important as
understanding its creation. Given these two courts and their institutional
differences, an examination of the IRS response to these courts provides a
unique opportunity to demonstrate how an enforcement agency approaches

INFLUENCES ON THE IRS

potentially competing claims of different courts with different structure, out-
look, and independence while attempting to carry out the policy preferences
of the executive and legislative branches. This chapter offers insight into
courts and whether their commitment to equality differs from the national
political commitment.

In the following section, I review the literature on bureaucratic control

and control of the IRS. After that I present the argument of how courts
within a national political framework can infl uence and control bureaucratic
behavior, and why that control will augment the preferences of the dominant
national coalition. I then offer two primary testable hypotheses and progress
to explain the construction of the dataset, development of the model, and
the methodology before presenting the results. In the last section, I discuss
the meaning of the fi ndings and suggestions for the future.

BUREAUCRATIC CONTROL AND THE IRS

Bureaucratic literature often emphasizes that agencies with enforcement
responsibility have an incentive to focus on legally sound but politically
inconsequential cases to augment their perceived successes (Katzmann 1980).
For example, police and prosecutors may believe that compiling arrest and
prosecution records are more important than societal values such as fairness
or the administration of justice because arrests and successful guilty pleas
lead to greater prestige and larger budgets.

Many argue that the IRS often exceeds other agencies in its disdain

for broad political and social values. Shapiro (1988) observes that the IRS
goal of maximum tax yield could subvert a broad tax policy aim of equity
and fairness. Long (1980) and others (Burnham 1989; Crouch 1996; Stras-
sels 1979) concur that the IRS has its own idiosyncratic goals independent
of justice and fairness and that a particular concern is maximizing revenue,
while MacDonald (1994) alleges that the IRS rewards regions and districts
that collect the most in revenue with larger budgets. Congressional hear-
ings in the late 1990s buttressed many of these assertions when witnesses
testifi ed about capricious and arbitrary agent behavior during both audit
and post audit activity. In short, the claim is that the IRS values allocative
effi ciency—how much money it can collect for the lowest cost—regardless
of the social harm or the damage to political and policy goals of the gov-
erning coalition.

Yearly, state and regional audit variation supports these claims. A citi-

zen of the Middle Atlantic region has a twenty percent greater probability
of getting audited than a citizen of the Midwest. Almost three percent of
corporate returns were audited in 1993, but the number dropped to less
than one percent by 2004. About one percent of individual federal returns
were audited in 1994, but the number increased to more than one and one

GETTING A POOR RETURN

half percent in 1995 and 1996 before falling to a low of less than one-half
of one percent in 2000 (TracIRS 2005). Corporate tax income represented
more than twenty percent of all tax dollars in 1995, but less than fourteen
percent by 2001, and, as previously noted, the number of audited low-income
returns has increased.

One explanation for the variation is the decentralized administrative

structure of the IRS (Burnham 1989; Scholz and Wood 1998). Due to a
series of scandals in the 1950s, the IRS decentralized its organization and
for most of the post–World War II period it operated out of sixty districts
within state boundaries in a larger set of seven national regions. Some of
this administrative structure has changed and currently the IRS audits out
of thirty-three districts encompassing the fi fty states in four national regions.
Each region is under the control of a regional director. Thus, each region
or state operates outside of centralized political control, with regional direc-
tors free to pursue their own preferred audit policy. Auditing low-income
taxpayers who cannot afford skilled counsel or perhaps lack education to
successfully fi ght an audit might mean that few agency resources need be
expended to recoup tax dollars from faulty tax returns. Because of this, the
cost of low-income audits might be relatively inexpensive and the IRS could
demonstrate a successful record before Congress.

Competing courts provide another reason for this lack of control. The

IRS is subject to the jurisdiction of three very different federal trial courts,
the U.S. District Court, the court of general trial jurisdiction in the federal
system, the U.S. Tax Court, and the U.S. Court of Federal Claims.

These

competing courts might allow the agency to “play” one court off the other
and ignore the consequences of any one particular court ruling.

Despite these theorized reasons and the evidence of audit variation,

scholars have demonstrated the political and appellate court control of the
IRS. In one study, Scholz and Wood (1998) use several state and federal
political control variables to examine the impact of partisanship and political
variation to show that changes in presidential administrations and congres-
sional committee membership led to changes in IRS audit rates. President
Reagan’s election, for example, led to an increase in individual audits and a
decrease in corporate audits. Howard and Nixon (2002, 2003) observe IRS
responsiveness to appellate courts and illustrate that regional variation in
audit policy is explained in part by ideological differences in these regional
federal appellate courts.

However, preceding any action of appellate courts, litigants fi rst chal-

lenge the IRS in a trial court and few of the trial court decisions will ever
be appealed. For example, in 2001, a total of 149 tax cases with the United
States as a defendant were fi led in the U.S. Federal Court of Appeals and
there were only two tax decisions by the U.S. Supreme Court during the
same year (Epstein et al. 2003).

INFLUENCES ON THE IRS

These low numbers compare to the over 29,000 fi lings in the District

Courts and Tax Courts in the 2000 calendar year, and over 700 total deci-
sions rendered by both courts during that same period. Trial courts remain
the principal courts for taxpayers to interact with the federal system and
one would expect that these competing federal tax trial courts will induce
regional and yearly variation in agency policy because the agency must
continually interact and deal with these competing courts to a much greater
extent than they interact with any appellate level court.

THE NATIONAL POLITICAL COALITION, COURTS,

AND BUREAUCRATIC CONTROL

Courts and the National Political Coalition within a
Separation of Powers Framework

The argument of the power of these competing courts to infl uence agency
policy within the national political coalition is premised on a separation of
powers (SOP) model. Admittedly, this is not the only theory of infl uence on
agency policy. Another theory showing how agencies respond to competing
principals is that of “common agency” (Dixit 1995; Dixit et al. 1997). Specifi -
cally, it is a theory of a “multilateral relationship in which several principals
simultaneously try to infl uence the actions of an agent” (Dixit et al. 1997).

For our purposes, this formulation better describes the process herein

of multiprincipals with common goals acting sequentially. The IRS regional
directors are appointed by the IRS commissioner, an appointee of the president
subject to confi rmation by the Senate. I assume that the regional directors
will carry out the policy preferences of the president because the president
has the power to fi re and replace a wayward director.

The regional directors should use their ability to shift audits between

states and districts to most effectively carry out the president’s policy, subject
to the constraint of the legislative branch. As Figure 5.1 (next page) shows,
the SOP model suggests that a “core” of legislative and executive branch
preferences constrain agency policy choices (Hammond and Knott 1996;
Hammond and Miller 1987). The fi gure implies that in a liberal-conservative
dimension, agencies may propose any policy within a “legislative-executive
core” because that policy will not be overturned through statutory means
(Hammond and Knott 1996). If the president, P, is relatively extreme, one
boundary of the legislative-executive core is defi ned by the median House
member, H

, or the median senator, S

, whoever is furthest from the presi-

dent. The crucial veto-override legislator in the House, H

, or Senate, S

whoever is closer to the president, defi nes the other boundary of the core.
Boundaries defi ned by this core represent the preferences of the dominant
national coalition.

GETTING A POOR RETURN

An agency, or regional director, that establishes a policy outside the

core will be overturned through a statutory revision at some point within the
core. If the SOP model is accurate, executive agencies will establish policies
at the boundary of the legislative-executive core closest to the president,
shifting as it shifts (Shipan 1998).

Based on this “core” dynamic, each regional director ought to establish

policy at the exact same point. However, there is still a need to control for
nonpolitical factors such as regional wealth disparities, or even regional or
local political factors. For example, IRS policy can directly impact a state’s
economy and local politicians would ignore the effect of IRS audits at their
peril. Therefore, other political actors, such as governors, may play a role
in the calculations of regional directors.

Courts and the Legal Set within the Separation of Powers Framework

Courts should also induce variation in agency policy through a “legal set”
that may or may not overlap with the legislative-executive core. If the legal

Liberal

P = President

= 290th Most Liberal House Member (key veto-override representative)

= 67th Most Liberal Senator (key veto-override senator)

= Median Representative

= Median Senator

Executive Appointee Target

Legislative-Executive Core

Conservative

Figure 5.1 Executive-Appointee Target—Veto-Override Point

INFLUENCES ON THE IRS

set does not overlap with the core, then judicial review presents an addi-
tional constraint on agency choice. A regional director may not establish a
policy at the boundary of the core because the director risks a court-ordered
policy anywhere within the legal set. Instead, the regional director should
establish a policy at the boundary of the legal set closest to the president,
if it lies within the core. Such a policy will survive any legislative efforts to
overturn it, and will also pass judicial review. It represents the best policy
the president can obtain. For a regional director of the IRS, there are two
primary trial court legal sets as demonstrated in Figure 5.2. Here TC repre-
sents the ideal policy of the Tax Court, and DC represents the ideal policy
of any particular federal district court.

Of course, how does the agency determine the policy preferences or

legal set for a trail court? Several court scholars examining appellate court
decision making have established the median justice as pivotal in calculating
court preferences (Hausegger and Haynie 2003; Martin, Quinn, and Epstein
2004). The median justice is crucial on multimember appellate courts where
every justice has a vote, with a majority decision determining the outcome.
There are fewer comparable analyses of agency policy and the legal set
established by trial courts (but see Smith 2006). It is diffi cult to reconcile
the logic of the median justice for a trial court, which is composed of an
identifi able fi xed number of justices, with each judge having an equal prob-
ability of assignment and each making a separate decision. Setting policy to
the median judge could leave the agency in a very disadvantageous position
if a judge very distant from the median is selected to try the case.

To arrive at a solution, the random assignment of judges may be thought

of as a calculation of probabilities and expected values. For example, consider
Figure 5.3’s (next page) illustration of a small trial court of fi ve judges, J

1–5

whose ideological preferences are arrayed on a liberal-conservative scale,
with random assignment of each judge to a case.

Figure 5.2 Court-Conscious Executive Appointee Targets

Legal Sets

Liberal

Conservative

GETTING A POOR RETURN

There are fi ve distinct assignments among this hypothetical collec-

tion of fi ve court members. If each possible judge has an equal probability
of assignment (1/5 in this example), then a specifi c judge may choose to
overturn an agency policy even if the agency establishes that policy within
the broad bounds of a potential court median. For example, if a regional
director establishes a policy at J

, but draws a J

, or J

judge, the IRS policy

is vulnerable to judicial reversal. A rational IRS director ought to set policy
based on a calculation of the probability of the assignment of each judge
and, therefore, the expected value among all possible assignments. This
ought to minimize the risk of loss and maximize the potential for victory
for the IRS.

One way to think about this is that each judge has an expected

ideological value, which is derived from the judge’s position on the ideo-
logical array multiplied by the probability of their selection. Thus, assume
that the ideologies for the fi ve judges in Figure 5.3 are on a 0 to 1 scale,
with 1 being the most liberal and 0 being the most conservative. Further
assume that J

has an ideology score of .1, J

has an ideology score of .2,

a score of .3, J

a score of .7, and J

a score of .8. Since the probability

of any judge being assigned a case is 1/5 or .2, the expected ideological
value of each judge is the ideology score multiplied by .2. These fi gures are
presented in Table 5.1.

In this table, the expected ideological values range from .02 to .16.

If one assumes J

to be the median judge, and the IRS places its policy

at that position, as already noted, drawing J

or J

would put the IRS at

a serious disadvantage. To avoid this, and to minimize loss and maximize

Liberal

Conservative

Figure 5.3 Trial Court Ideological Array

Table 5.1 Judicial Ideological Array and Expected Value

Judge Ideology Expected

Value

.1 .02

.2 .04

.3 .06

.7 .14

.8 .16

INFLUENCES ON THE IRS

gain, the IRS should add up all the calculated expected values that would
give the agency the optimum policy position for the maximum expected
outcome, which is in all cases such as this the equivalent of the average
of the ideology scores, or .42. Thus, the best possible strategy for the IRS
in dealing with a randomly assigned fi nite number of trial court judges is
to place policy at the average of the ideological values for all the justices
on the court (Smith 2006).

However, this still begs the question of comparative response to the

different courts with different mean ideological legal sets. While the District
Court is an Article III court with lifetime tenure and judicial independence
compared to the Article I Tax Court, many have argued that, in practice,
these other constraints are “largely theoretical” (Posner 1996, p. 268; see
also Easterbrook 1990). Tax Court judges enjoy full pension protection,
usually are reappointed or assume senior status, and have the same salary as
District Court judges. Thus, in practice, Tax Court judges enjoy the same
protections as the Article III judges (Dubroff 1979; Maule 1999) with the
exception of explicitly protected lifetime tenure.

The Tax Court is, in contrast to the regionally specifi c District Courts,

a national court. Tax Court rulings have national jurisdiction. Although
the Tax Court under the case of Golson v. Commissioner (1970) voluntarily
follows the rulings of the appropriate circuit, given the national scope of
its rulings, and its acknowledged expertise in tax matters, the IRS must
change policy at least regionally, and sometimes nationally, in response to
Tax Court decisions.

The jurisdiction of the District Court judge extends only to the

geographic limits of the specifi c district. District Court decisions are not
national, or even regional, but district specifi c, or at best state specifi c if the
district covers the entire state. This district-wide jurisdiction is considerably
smaller than the regional or national jurisdiction of the Tax Court. Given
the broad scope of the rulings, it is realistic to posit that the IRS, subject
to competing jurisdictions, will pay closer attention, and be more likely to
respond to, rulings of the national court than the district and state-specifi c
federal court.

HYPOTHESES

No enforcement tool used by the IRS produces more attention than the
audit, and the degree to which audits focus on various groups of taxpayers
based on their earnings is politically relevant to competing concerns along
a liberal-conservative continuum. One would expect more liberal executives
and legislatures and more liberal courts to want the IRS to audit fewer
low-income individuals, while one would expect more conservative politi-
cal actors to favor greater attention to the audits of the poor, particularly

GETTING A POOR RETURN

if it shifts attention away from the wealthy. In addition, the Tax Court is a
national court with national jurisdiction. In comparison, the District Courts
only have jurisdiction over their respective districts. Thus, the primary
hypotheses are as follows:

Hypothesis 1: As both the courts and the executive-legislative coalition
move in a conservative direction, the IRS will audit more low-income
individuals. The national political coalition and the courts should both
move IRS audit policy in the same direction.

Hypothesis 2: Because the Tax Court is a national court with national
jurisdiction and because the differences in the independence of the
District Court is more theoretical than actual, the IRS will respond more
to the national Tax Court than the state-specifi c District Court.

TESTING INFLUENCES

Data on the IRS, including low-income and high-income audit rates, audit
rates, and percentage of returns changed after audit, come from a database
compiled by TRAC IRS, a Syracuse University-based center that compiles
data on various federal agencies. Following the earlier approaches in audit
variation (Scholz and Wood 1998, 1999), I examine state variation for this
study because IRS districts and regions have changed over this period and
state-level analysis provides greater stability. I used the years 1994 to 2000.
I derive the dependent variable by dividing the audits per state per year of
the highest income class (> $100,000) by the total number of audits per
state per year of the lowest income class of taxpayers (< $25,000). This
ratio increases when the IRS shifts audit attention away from low-income
taxpayers toward high-income taxpayers. The measure correlates with modern
political and economic liberalism and conservatism. If courts and the national
political coalition become more conservative, one should see an increase in
the audits of low-income taxpayers. Conversely, if the courts protect minor-
ity interests, there should be a divergence of court and national coalition
infl uence, regardless of court ideology.

As prior research has shown, audits are affected by more than just court

or national political considerations. Echoing Shapiro, one should fi nd that
if the IRS values allocative effi ciency, then state demographic and political
factors should impact collection. States with high per capita income are
more likely to have a greater number of middle-class and wealthy taxpay-
ers to audit, lessening the need to audit the poor. Conversely, states with
high urbanization might make it more effi cient to audit the poor because
taxpayers live much closer together and therefore one tax or revenue agent
can audit many more individuals in one specifi c area than one who audits

INFLUENCES ON THE IRS

in areas of greater population dispersion. States with high unemployment
should also affect low-income audits, although not in the direction one
would expect. That is, ordinarily one would expect greater unemployment
to lead to fewer audits of the poor—why audit someone who is unemployed?
However, during the 1990s, Congress enacted, and the president signed,
various legislative measures that restructured the IRS and made changes
to various tax and trade provisions. Although the legislation was primarily
designed to provide taxpayer relief, the measures also made some changes
to qualifying for the earned income tax credit (EITC). Unlike most other
tax credits, if the EITC credit exceeds tax liability, the taxpayer receives the
difference in cash from the IRS. The tax law changes called for stepped-up
IRS monitoring of returns claiming this low-income credit. An increase in
unemployment then should lead to an increase in those qualifying for the
EITC, which should in turn lead to an increase in auditing returns of the
lowest income classes.

Because of this, I added the control variables of per capita income

(in thousands of dollars), unemployment rate, population, urbanization
(percentage of population living in urban areas); as a fi nal gauge of alloca-
tive effi ciency, I use a variable that measures the percentage of low-income
returns changed after audit. The greater the percentage of low-income returns
changed after audit, the more likely the IRS will continue to increase audits
of low-income taxpayers because it is cost effective to audit this type of return.
To measure the impact of local government, as previously discussed, I added
a measure of state government liberalism developed by Berry et al. (1998).
To test the national coalition, I use the point in the W-nominate scale of
the pivotal separation of powers veto-override member in the Congress (see
Krehbiel 1998). W-nominate scores are positive for conservatives, negative
for liberals, and have been tested and shown to be static (Poole 1998).
Judicial ideology was again measured through the comparable nominate
scores based on the political and geographical circumstances surrounding
the appointment of each judge.

I also used time-point dummy variables for each year, omitting the

last year as the baseline. One reason for the use of these variables was
methodological and that is explained in the appendix. The use of dummy
variables for fi xed effects models is often atheoretic—that is, they are
employed strictly as a control mechanism. However, in this case, the use of
time-point dummies makes sense for this particular model. Many of the years
studied had particular political importance. For example, 1994 represented
the Republican takeover of Congress and 1995 their fi rst year of power. The
year 1996 was the presidential election and 1997 and 1998 saw signifi cant
tax reform efforts, including seeking stepped-up audit enforcement of those
claiming the EITC. The year 1998 was also a midterm election, which
saw a slight Democratic gain in the House of Representatives, while 1999

GETTING A POOR RETURN

was the fi rst full year following passage of the increased audit attention to
EITC recipients.

While the focus of this book is on the varying treatments of different

income classes, as an additional test of the dominant ideology and court
treatment, I also analyzed the differences in tax enforcement of individuals
as compared to corporations. The degree to which audits focus on individual
and corporate taxpayers is politically relevant to competing concerns along
a liberal-conservative continuum and this analysis would provide additional
evidence of court treatment conforming to the preferences of the dominant
national coalition. The hypotheses posited for the movement of low-income
audits should be the same for the movement of individual and corporate
audits. Using the same data and independent variables, I derive this second
dependent variable by dividing the audits per state per year of the individual
taxpayers by the audits per state per year of corporate taxpayers. This ratio
decreases when the IRS shifts audit attention away from individual taxpay-
ers, toward corporate taxpayers, and increases when the IRS audits more
individuals and fewer corporations. The explanation is in the appendix.

With the judicial ideology measures included, Table 5.2 provides a

complete list of the descriptive statistics for the variables with means and
standard deviations for the data used in this chapter. On average, the IRS
audits low-income taxpayers more than two and half times as much as high-
income taxpayers per state per year, but there is signifi cant variation, with
the ratio increasing as much as over nine to one, and in some states more
high-income taxpayers are audited. The Tax Court for the years of this study
is, on average, slightly more conservative than the District Court, but the
District Court exhibits greater variation. The core shows a slight conserva-
tive orientation, although there is a signifi cant range over this time, given
the shift in control of Congress from 1995 onward.

Table 5.2 Descriptive Statistics—Audit Rates

Variable Mean

Std.

Dev.

Individual/Corp. Audit Rate

11.07

6.85

High/Low-income Audit

4.22

2.18

Urbanization Percent

67.41

16.99

Unemployment Rate

5.99

2.24

Density 174.94

238.38

Population 5,427

5,927

Percent Change after Audit

90.42

3.95

Core .06

.31

Tax Court Ideology

.13

.03

District Court Ideology

.12

.16

INFLUENCES ON THE IRS

RESULTS

The results presented in Table 5.3 and Table 5.4 (next page) show the
impact of the coeffi cient as the variable moves from one-half a standard
deviation below the mean to one-half a standard deviation above the
mean. The appendix contains the complete tables with coeffi cients and
standard errors.

These results provide strong evidence to confi rm the primary hypotheses.

The courts and national coalition as represented by the core measure impact
IRS audit behavior, and court preferences closely track the preferences of
the national political coalition. In a more conservative political climate,
both the courts and the political branches push the IRS to increase audits
of the lowest income level.

Institutional ideology is a powerful explanation for the increased

attention to low-income tax returns. Tax Court ideology, District Court
ideology, and the national core coalition are substantively important. The
IRS alters the ratio of low-income to high-income audits in response to the
policy preferences of the Tax Court, the District Court, and the national
coalition. As the ideology of the average Tax Court judge, average District
Court judge, and the national political coalition moves in a conservative
direction, more low-income taxpayers are audited.

The Tax Court is a major and important actor in infl uencing agency

policy. A standard deviation shift in its average ideology leads to twenty-fi ve
percent more low-income audits. Also as hypothesized, the Tax Court coef-
fi cient was substantively much greater than the District Court coeffi cient.
While the infl uence of the District Court is less than that of the Tax Court,
the IRS still fashions its audit policy in relation to the District Court aver-
age ideology. A standard deviation shift from a more liberal District Court
to a more conservative District Court results in sixteen percent more audits
of the lowest income class in relation to the highest.

As the executive and legislative preferences grow more conserva-

tive, there is a shift toward low-income audits that parallels the infl uence
of the courts. The IRS regional directors audit low-income taxpayers in
accordance with the wishes of the executive branch of government, sub-
ject to the constraint of the pivotal veto-override member of Congress.
The standard deviation move from a more liberal to a more conservative
national governing coalition shifts audits to the lowest income class by
twenty-three percent. Although the coeffi cient is smaller than that of the
Tax Court, this shift is nearly equal to the corresponding shift of the Tax
Court because the ideology of elected offi cials changes much more rapidly
than judges appointed for life or appointed for fi fteen years with the prob-
ability of reappointment.

GETTING A POOR RETURN

Table 5.3 Infl uences on the Number of High–Income/Low–Income Audits*

Variable Impact

State Gov. Ideology

.11

Per Capita Income

–.30

Population Density

–.23

Urbanization –.06
Unemployment –.02
Population –.10
Tax Court Ideology

.25

District Court Ideology

.16

Core .23
Pct Change after Audit

.21

1994 .02
1995 .09
1996 .26
1997 .03
1998 .56
1999 .18

*Impact represents a one standard deviation change in the independent variable, moving

from

one-half standard deviation below the mean to one-half standard deviation above the mean.

Table 5.4 Infl uences on the Number of Individual/Corporate Audits*

Variable Impact

State Gov. Ideology

–.03

Per Capita Income

–.30

Population Density

–.05

Urbanization .03
Unemployment –.03
Population .01
Tax Court Ideology

.15

District Court Ideology

.11

Core .23
Pct Change after Audit

–.02

1994 1.25
1995 .94
1996 2.70
1997 1.22
1998 1.11
1999 .71

*Impact represents a one standard deviation change in the independent variable, moving

from

one-half standard deviation below the mean to one-half standard deviation above, or a change
of 0 to 1 for a dichotomous variable.

INFLUENCES ON THE IRS

Effi ciency does matter. The percentage change in audits, a measure of

how successful the low-income audit was in collecting additional tax due,
does lead to a twenty-one percent shift in the ratio of low-income audits.
The greater wealth of the state, as measured by per capita income, leads
to a corresponding shift of thirty percent away from low-income taxpayers
and toward high-income taxpayers. Urbanization leads to a six percent shift
from low-income audits; however, urbanization is correlated with per capita
income and greater urbanization might just be an additional measure of the
wealth of the state.

The time-point dummies all were substantively and statistically signifi -

cant, albeit with differences in impact. The election cycle years of 1996 and
1998 showed dramatic infl uence on the audit rate. The year 1996 resulted
in a shift toward greater audits of low-income taxpayers by twenty-six per-
cent. The year 1998, which was also the year of tax reform and the agreed
on increase of audits of EITC recipients, saw a fi fty-six percent shift from
high-income taxpayers toward low-income taxpayers.

The second dependent variable shows similar infl uences, but with

important differences. The IRS altered the individual/corporate audit ratio
in response to the policy preferences of the average judge of the Tax Court.
As the ideology of the average Tax Court judge moves in a conservative
direction, a greater percentage of individual taxpayers are audited. As
the average ideology moves in a liberal direction, the IRS targets more
corporate returns for audit. A one standard deviation shift in the average
ideology of the Tax Court leads to more than fi fteen percent individual
audits compared to corporate audits. Of course, over the time period of
this study the Tax Court ideology ranged only eight percent in this sample,
compared, by contrast, to over forty-seven percent variation for the Dis-
trict Court, limiting the impact of the shifts in the audit ratio caused by
shifts in the Tax Court ideology. Also as hypothesized, the impact of the
Tax Court was greater than that of the District Court, which showed an
eleven percent change toward more individual audits as the District Court
became more conservative.

Political circumstances, in comparison to judicial ideology, appear to

be very important to the corporate individual audit rate, far more so than
for the variation in individual audits. Here the standard deviation shift
in the core leads to almost twenty-fi ve percent more audits of individu-
als. A more liberal political coalition emphasizes corporate auditing more
than individual auditing. Undoubtedly, this is confi rmation of Scholz and
Wood’s (1998) fi ndings of audit changes due to the changes in presidential
administrations. These fi ndings are buttressed by the results of the individual
years, which represent substantial political change; most of the years show
signifi cant substantive impact.

GETTING A POOR RETURN

CONCLUSION

Low-income and individual audits increase because of political and judicial
considerations confronted by the IRS. In addition to the governing national
coalition, the Tax Court, a national specialized court specifi cally designed
to deal with audited taxpayers and resolve their disputes with the IRS, is a
major player in setting and determining agency policy. To a lesser extent,
the District Court also infl uences tax policy. The Tax Court and District
Court establish clear “legal sets.” Changes in judicial ideology change the
audit behavior of the IRS. A study encompassing even more years would
show greater variation in Tax Court ideology and therefore greater change
due to its shifting ideology.

Contrary to assertions by political leaders that the IRS is different

from other agencies, that it is an agency “out of control,” there is substantial
political and judicial infl uence on the IRS. This book demonstrates that all
three branches of government work to control the agency, and the increasing
conservatism of the Tax Court, District Court, and the governing national
political coalition has led to an increase in the number of low-income and
individual audits. Increased attention to the returns of low-income taxpayers
is the result of deliberate policy choices by the political branches of govern-
ment and judicial preferences.

In the past two decades, there has perhaps been no domestic public

policy that has so dominated American politics as taxes, and the interre-
lated questions of who should pay and how much. There is no reason that
a judiciary nominated and confi rmed through a political process should not
have policy preferences or that the IRS should not follow these preferences
in addition to following the preferences of other political actors. Robert
Dahl (1957) noted many years ago that the Supreme Court is rarely out
of political alignment with majoritarian elected institutions because of the
appointment power of the president. More recently, Whittington (2005)
observed that an emerging literature argues that Supreme Court doctrine fi ts
within the broader political regime (see, e.g., Gillman 2002; Pickerill and
Clayton 2004), and “structural characteristics of the political systems such
as the United States encourage cooperation between judges and political
leaders to obtain common objectives” (Whittington 2005, p. 584).

This appears to be true for IRS enforcement policy and its relation-

ship to the preferences of lower federal courts, including specialized courts.
Future research should examine the infl uence of other trial courts and other
specialized courts on other agencies and other public policy domains to see
if the courts differ from the national political coalition in both the size and
direction of their infl uence. It appears that the protection of minority rights
and the commitment to equality depend on the membership of the court,
and not solely on its assumed institutional role.

THE CREATION AND ENFORCEMENT OF NATIONAL TAX POLICY

SIX

COURTS, FAIRNESS AND THE

CREATION AND ENFORCEMENT

OF NATIONAL TAX POLICY

COURTS, FAIRNESS, AND NATIONAL GOVERNING COALITIONS

Through the preceding pages of facts, illustrations, data, analyses, and argu-
ments, I have tried to demonstrate the importance of tax trial courts in
setting and determining the nation’s tax policy and tax enforcement in two
ways. First, I have attempted to show that Tax Courts play an important
part in setting and determining tax and tax enforcement policy. Who has to
pay and how much and whose tax return will be audited and why are not
choices solely made by elected offi cials and their delegated agents. Courts
as well as Congress and the president have an important and infl uential
role in these decisions. Second, I have endeavored to prove that the infl u-
ence of the courts is not independent of the broader directives and policy
preferences of the dominant political coalition. While courts may exist to
ensure fairness and allow tax litigants their “day in court,” the preferences
of the judges on these courts show little difference from the preferences of
the majority coalition. The result is that the rulings favor the taxpayer that
the coalition favors.

That is, the directional infl uence that courts have on this process differs

little from the choices preferred by elected offi cials and their agents. In our
federal system, the judges are nominated by the president and confi rmed by
the Senate. While there are many individual judges remaining on the bench
who have been appointed by presidents now out of power and approved by
a Congress very different from the one that sits during the later trials and
subsequent judicial decisions, over time the dominant coalition can change
the judiciary to resemble the ideology, values, and preferences of this political

GETTING A POOR RETURN

coalition. This is particularly so in those courts where judges sit for a fi xed
term and thus do not enjoy lifetime tenure.

This does not mean that every individual judge or even the majority

of decisions of a specifi c court will always coincide with the preferences of
the political majority. One personal opinion might not always represent the
interests of, or issue rulings consistent with, the preferences of the majority.
However, one ruling or even several rulings that are contrary to dominant
policy preference do not mean that the courts always protect the minority
against the wishes of the dominant majority. The aggregate picture reveals
that, on average and over time, court decisions mirror the preferences of the
majority coalition. When the dominant majority favors less tax enforcement,
court rulings will usually be less favorable to the tax enforcement authority, the
IRS. When the dominant majority favors less enforcement of wealthy taxpayers,
the IRS follows suit and court rulings reinforce and buttress this trend.

Courts are often criticized as activist, where the term is used in a

pejorative sense to mean that the court rulings oppose the preferences of the
democratic majority. Thus, courts are condemned as antimajoritarian and,
by implication, undemocratic. It is true that rulings can be antimajoritarian.
While often normatively decried as undemocratic, this can be a positive
element of a democratic system. Antimajoritarianism allows courts to be the
“last best hope” of the disenfranchised. However, most often courts mirror
and uphold majority preferences.

Each of the taxpayers portrayed here represents some of the various

types of U.S. taxpayers. One is rich, another is poor. One is a middle-
income couple seeking to uphold deductions, while another is a wealthy
individual attempting to uphold a challenged tax shelter. One taxpayer is
a large corporation, while the rest are individuals. Each taxpayer alleged
some inequity in enforcement of tax policy and chose to seek redress in the
U.S. court system. Each taxpayer made a choice to sue and then selected
a particular forum. The tax policies themselves were deliberate choices.
The audits they contested were themselves the product of choices made by
agents of the IRS. The accumulation of these individual choices led to the
analyses of this book.

Political science defi nes itself by positing theories of causation, then

developing testable hypotheses, and fi nally gathering information to learn
about political reality. We examine larger trends and try to determine what
constitutes average behavior or median positions of all the actors involved
in political process. We step back to examine the larger picture. In con-
sidering this broad view, we sometimes forget that each data point of some
trend represents someone who has gained or lost, suffered harm or been
rewarded. It is not unlike when law students realize that each case they
read in a textbook might represent an important principle of law, but also

THE CREATION AND ENFORCEMENT OF NATIONAL TAX POLICY

represents someone who has gained or lost something of value. A criminal
defendant might go to jail or be freed because of a court ruling. A plaintiff
might receive compensation resulting from a severe injury or fail to obtain
much needed relief because of the negative outcome of a particular case.
The same is true of tax litigation. While they are not life or death issues,
the outcomes of particular cases can have enormous consequences for indi-
vidual litigants as well as teach us something about the larger trends in tax
policy and tax enforcement.

Given the number of noncompliant taxpayers, no one can seriously

doubt the need for taxpayer audits. The government, and, by extension,
we the people, need the revenue collected from the income tax. However,
whether one is cheating or actually in compliance, whether one owes addi-
tional tax payments or not, it is unquestionable that undergoing an audit
and then contesting the audit means a nontrivial level of anguish for the
taxpayer, as well as a signifi cant investment of time and often money. Each
individual case is unique and some will win big, while others will lose. A
business entity the size of Wal-Mart will recover from an adverse ruling
regarding its accounting practice. The denial of the practice might result
in a loss for one quarter of a fi scal year. Perhaps James William will suffer a
signifi cant economic loss with a denial of a tax shelter, but he might have
enough assets to weather a short-term loss. However, in certain cases, the
loss can be devastating. Joy Anders and the Forrest’s would likely have to
deal with crippling economic losses. Few middle-income couples can pay an
additional $25,000 plus potential penalties and interest and almost no low-
income individual can withstand the denial of 1,500 worth of tax credits,
which might represent twenty-fi ve percent of yearly income. The court
decisions in these latter two cases have enormous impact on these taxpay-
ers beyond any aggregate story of audit shifts, forum choice, and chances
of winning in a tax trial court.

It is not wrong to focus on the aggregate numbers to gain greater insight

into causation and trends. These greater trends ultimately move public policy
and create societal change—and social science wants to know why change
has occurred. However, understanding what causes changes in tax policy
and tax enforcement does not mean individual taxpayers do not deal with
the consequences of these changes. It means a Joy Anders gets audited and
might confront an adverse court decision while a high-income individual
might not. It means a James William once audited might fi nd a court more
favorably inclined to uphold a tax shelter than a use of the unearned income
tax credit. It means a corporation using a questionable accounting practice
might fi nd that it does not confront an audit the following year, or, if it
does, it is likely that its tax position will be upheld by the tax trial court.
Consider the main fi ndings of chapters two, three, four, and fi ve.

GETTING A POOR RETURN

Chapter Two

In chapter two, I argued that the tax laws are meant to be “revenue rais-
ers.” Interpretation and burden of proof have usually been on the taxpayer.
Because of the purpose of these tax laws is to raise revenue and because of
this burden of proof, taxpayers lose far more often than they win in the Tax
Court, the Court of Federal Claims, and the District Court. The law favors
the government. That being said, even if the taxpayer loses, the choice of
forum and the ease and relative effortlessness of fi ling claims allow almost
any aggrieved taxpayer to at least have his or her “day in court.” Less affl uent
taxpayers can fi le claims in a short period of time and have them heard in
an informal and relatively inexpensive setting. Of course the term “forum
choice” is misused in this instance because of the prepayment option. That
is, both the Court of Claims and the District Court require prepayment and
only a few taxpayers can afford to prepay their tax obligation and subsequently
sue for a tax refund. Therefore, most taxpayers, including Joy Anders and
the Forrest’s, have no choice but to sue in Tax Court.

However, the courts have a broader role than acting as a forum to

settle tax disputes—they must ensure actual fairness. We have a system of
voluntary compliance at least to the extent that the system depends on
individual taxpayers self-reporting. To ensure voluntary observance of tax
fi ling, taxpayers need a credible commitment from the courts that justifi es
their compliance. That is, we want the IRS to be aggressive in the collection
of tax revenue. The good of the entire country depends on the agency’s col-
lecting revenue in as effi cient a manner as possible, but not at the expense
of fairness or injustice to taxpayers. In an ideal world, the IRS aggressively
pursues the tax cheaters and leaves the honest citizen alone. We know this
does not happen. An imbalance toward aggression, toward effi cient collec-
tion, inevitably means some honest taxpayers will not be treated in a fair
and just manner. Some tax shelters are legitimate, but many are not.

Thus, we also want to know that taxpayers get a fair hearing if they

choose to contest the agency action. The appearance of fairness is not enough.
If certain forums offer advantages to the litigants over other forums, then
these choices, to be effective, must be open and available to all litigants. In
the end, compliance depends on taxpayers believing that they have some
sort of stake in the outcome.

Chapter Three

Here I found evidence that political institutions use litigation and forum
choice to attain policy goals. Litigants calculate the likelihood of winning
their tax suit. If they are likely to win, they proceed with the litigation.
However, there is more to the calculation than individual costs and potential

THE CREATION AND ENFORCEMENT OF NATIONAL TAX POLICY

benefi ts. Litigants also receive signals from the dominant political coalition.
A more conservative coalition will encourage taxpayers to challenge IRS
post audits claims because the dominant coalition sees litigation as additional
measure of agency control and policy change.

In addition, political signals can also lead taxpayers to favor one forum

over another. It is rational for the litigants to choose the more conservative
forum because that will be the one most favorable to the taxpayer and least
likely to support the IRS. It is easier for the governing national coalition to
change the ideology of the Tax Court much more quickly than the ideology
of the District Court.

Thus, litigation and forum choice are political acts as much as they

are rational calculations on the costs and benefi ts of challenging an audit.
Forum shopping is usually derided because it appears that litigants and their
lawyers play games with our justice system to create favorable outcomes.
However, at least in tax litigation, Congress has made the explicit decision
to allow taxpayers to use forum choice and, at least in the 1990s, seems to
encourage taxpayers to choose a forum that offers them the greatest likeli-
hood of winning against the IRS. So it appears that litigation and forum
shopping are only bad until they help the governing national coalition
achieve political and policy goals.

Chapter Four

In the chapter that examines court decision making, I showed how the
Congress and the president can use the Article I specialized court’s limited
tenure and reduced numbers to change the ideology of the court much more
quickly than the independent Article III court of general trial jurisdiction.
Just as the coalition can use litigation to control agency policy, it also uses
the ideology of the specialized court as a method of agency control. Con-
gress and the president can push the agency to produce policy in line with
congressional and executive preferences because they have greater control
over the decision making of the specialized court.

Of course this fi nding differs from the ideal of a specialized court using

its expertise to decide issues without reference to ideology and free from
agency infl uence. However, the ideal ignores the practical matter that the
collection and distribution of revenue are politically charged issues. Who
or what should pay and how much, and who or what should receive this
revenue and how much, are inescapably political questions with ideological
overtones Despite repeated calls for simplifi cation and ease, any system of
taxation must have complex laws, rules, and regulations. The interpretation
of complicated questions of assessment and collection requires discretion
and technical expertise, and any result requires opinion and judgment;
such opinion and judgment cannot be divorced from basic views about the

GETTING A POOR RETURN

collection and allocation of scarce resources. This elaborate system of appeals
allows control of the IRS and ultimately tax policy.

There have been numerous calls for, and studies of, proposals to con-

solidate all tax cases in one court and to create a U.S. Tax Court of Appeals.
While many reasons have been put forward to oppose such proposals, an
important consideration has been the fear that such specialized courts would
be no more than extensions of the IRS and thus likely to exhibit bias in
favor of the agency. However, while the Tax Court might issue more rulings
in favor of the IRS than the District Court, the reason might be as much
due to congressional design and executive appointment to a more limited
independent court, as compared to an Article III court, than it is to bias
in favor of an agency.

Chapter Five

In this chapter, I found that there is substantial political and judicial infl uence
on the IRS. All the branches of government work to control the agency,
and the increasing conservatism of the Tax Court, District Court, and the
governing national political coalition is the reason for the increase in the
number of low-income audits. Increased attention to the returns of low-
income taxpayers is the result of deliberate policy choices by the political
branches of government and judicial preferences.

There is no reason that a judiciary nominated and confi rmed through

a political process should not have policy preferences, or that the IRS
should not follow these preferences in addition to following the preferences
of other political actors. Whittington (2005) observed that Supreme Court
doctrine fi ts within the broader political regime and “structural characteristics
of the political systems such as the United States encourage cooperation
between judges and political leaders to obtain common objectives” (Whit-
tington 2005, p. 584). This appears to be true for IRS enforcement policy
and its relationship to the preferences of lower federal courts, including
specialized courts.

CHANGING TAX POLICY, CHANGING THE COURTS

Clearly, federal trial court rulings, as well as appellate court decisions, refl ect
dominant political preferences. Judges are nominated by the president and
confi rmed by the Senate. Many judges are loyal party workers who diligently
work for the local, state, and national party to secure a nomination to the
federal bench. Even if the judge is not a committed and loyal Republican or
Democrat, he or she has a patron either in the nominating senator, approv-
ing governor or representative, or some other high-ranking party offi cial. In
short, the nominating and confi rmation route is a political process. Political

THE CREATION AND ENFORCEMENT OF NATIONAL TAX POLICY

leaders with policy preferences nominate and appoint judges who, for the
most part, share those preferences and are committed to seeing them enacted.
There is no reason to believe that the judiciary will have markedly different
political beliefs than the dominant political majority.

Having said this, there are some caveats to this now not so startling

idea of the alignment of judicial and majoritarian preferences. First, policy
and politics do not always determine outcomes, particularly for trial court
judges. Law, facts, precedent, and legislative intent all are important in case
outcomes. Trial courts, even federal trial courts, do not have the institutional
features of the U.S. Supreme Court that allows that court to have virtually
unlimited freedom in case outcomes. Lower courts must adhere to precedent
from both the Supreme Court and the appropriate Court of Appeals, have
little docket control, and presumably seek appointment to higher courts. At
the trial level, facts can often lead to only one outcome, and particularly so
in tax matters where the taxpayer has the burden of proof and most often
fails to reach it. All of these features will temper the ideological direction
of their rulings. In short, law does matter.

Second, as Dahl and others have noted, there are times when judicial

preferences signifi cantly differ from the majority of the elected offi cials.
Prominent among those times were the 1930s with the election of Franklin
Roosevelt and his New Deal coalition. The court, as Dahl showed, struck
down much of the New Deal legislation, at least until the “switch in time
that saved nine” in the mid 1930s and the ability of Roosevelt to replace
many of these more conservative justices with his own appointees. Another
scholar has argued that we are likely to see these misalignments between
elected and judicial preference following critical realigning elections such as
1800 and 1828 as well as 1932 (Funston 1975; but see Beck and Funston
1976 and Canon and Ulmer 1976 disputing Funston’s fi ndings).

Dahl argued that a “dead hand” of the previous coalition could reach

out through the judiciary and stymie the achievement of the goals of the
new electoral majority. While empirical evidence supporting Dahl’s specifi c
claims has been somewhat lacking, clearly the longer tenure of the unelected
judiciary means that it takes more time to change the median or average
preferences of the federal judiciary. Even the Article 1 Tax Court judges sit
for a term of fi fteen years—almost twice that of a reelected president, more
than twice the term of a senator, and more than seven times as long as the
term of a member of the House of Representatives.

Thus, a federal judge sits on the bench for a long time, potentially

much longer than the coalition that fi rst nominated and confi rmed him or
her to the judiciary. This means that long after majority preferences change
and a new coalition comes to power, the judiciary will have many judges
whose preferences do not match those of the now dominant coalition.
Although public policy might favor changes in tax rates and tax enforcement,

GETTING A POOR RETURN

it might take much longer for the judiciary to adopt and adhere to those
preferences. While the “dead hand” of the past will eventually change, and
change more quickly in the shorter tenure Tax Court than in the District
Courts, individual rulings will be inconsistent with majority preferences, at
least for the fi rst few years after electoral change.

Thus, although a new Congress and a new president might favor fewer

examinations of individuals claiming the Earned Income Tax Credit, a Joy
Anders might still confront an unfriendly judiciary partial to enforcement
against low-income tax payers. Conversely, although a new administration
might favor greater enforcement against tax shelters, James Williams might
still fi nd judges more sympathetic to the use of such shelters than would be
suggested by the current political climate.

This will change over time and the longer the coalition stays in power,

the greater the concordance between judicial preferences and elected pref-
erences. Even when judicial preference and electoral preference differ, this
does not mean that courts are out of control. However, nor does it mean
that the courts really represent the last best hope for the disenfranchised.
The courts follow the preferences of the majority, but since courts are made
up of judges appointed over time, not all judges will always support the
dominant preferences.

METHODOLOGY AND THE USE OF PANEL DATA

APPENDIX

METHODOLOGY AND THE

USE OF PANEL DATA

One can read and understand this book with a minimal appreciation of
quantitative methodology and data analysis. Basic concepts of central ten-
dencies and probability are suffi cient to appreciate and follow the statistical
fi ndings. This appendix is for those with a greater knowledge of statistical
methodology and analysis. It offers the complete tables, including coeffi cients
and standard errors, for those who want to understand more about how the
results were achieved. Of course, the presentation of the complete results
also allows those with methodological training to criticize or disagree with
the various methodologies chosen. I think the fi ndings are quite robust, but I
welcome any exchange with those seeking greater information. The data and
methodological analyses employed in this study do present several interesting
issues and potential hazards. For example, there are potential problems with
the Gauss-Markov assumptions of multicollinearity, heteroskedasticity, and
auto/serial correlation. This appendix is a brief discussion of those issues
and how I dealt with the associated problems.

Data used for most of the study are longitudinal panel data. That data

was gathered from fi fty states and I then proceeded to analyze the informa-
tion over a period of years. It is a combination of data gathered from both
a cross section (states) and over a period of time, or a time series (years).
Such data are very useful. They allow the examination of issues and ques-
tions that cross-sectional or time series data alone do not. Their usage also
increases the degrees of freedom in any given analysis. Such data, however,
also have the potential to exhibit all the problems associated with cross-
sectional and time series data in addition to problems that are unique to
panel data. The exact number of years depended on the specifi c analysis.
Some of the data was much more diffi cult and time consuming to collect. For
example, collecting and coding each and every judicial decision on lawsuits

GETTING A POOR RETURN

brought by taxpayers are labor intensive and thus the data for chapter four
was limited to two years.

In addition, the use of these types of data and the subsequent pres-

ence of these problems make the use of regression or Ordinary Least Squares
(OLS) problematic. OLS is optimal if one assumes fi rst that the error pro-
cesses have the same variance (homoskedasticity), and that the errors from
any given time point and any given unit are unrelated to the errors for
any other time point or unit point (no serial/auto correlation). Beck and
Katz (1995, p. 636) call these errors uncomplicated or “spherical.” Those
assumptions are diffi cult to make with data examining states over time. It
is diffi cult, if not impossible, to assume homoskedasticity and no auto/serial
correlation. States, for example, are of unequal size. A large state will have
a different infl uence than a small state, meaning there is probably hetero-
skedasticity. In addition, it is reasonable both to assume that what happens
in any one state infl uences what happens in another and what happens in
one year infl uences what happens in another. That is, there is undoubtedly
auto/serial correlation. Because of this, OLS estimation will not make the
most effi cient use of the data. This means that the standard errors will be
incorrect. Before discussing alternatives to OLS, I want to detail the prob-
lems with the data and estimation and the tests and methods available to
deal with these problems.

ESTIMATION PROBLEMS

Heteroskedasticity

Heteroskedasticity is unequal variance of the error term (Downs and Rocke
1979). This is a violation of a Gauss-Markov regression assumption, which
means that the use of ordinary least squares can lead to ineffi cient estimates
(Gujarati 1995, pp. 365–66). Each estimate has a greater probability of
being “off target” (Berry and Feldman 1985, p. 77). The t and F tests of
signifi cance are likely to give inaccurate results. In this sample, states are
obviously of unequal size within each panel. Before using any particular
method to test for heteroskedasticity, I transformed the data in an attempt
to reduce, if not eliminate, the heteroskedasticity caused by the unequal size
of states—for example, by using a percentage or a logarithmic transforma-
tion whenever possible.

Given the transformation, I then tested for heteroskedasticity. There

are no hard and fast rules for detecting heteroskedasticity, only some rules
of thumb. I performed basic tests on all equations in the model, including
graphing the estimated squared residuals (u

) against the estimated Y

, and

against several of the explanatory variables. I also performed White’s general
heteroskedasticity test on the equations. This involves running an auxiliary

METHODOLOGY AND THE USE OF PANEL DATA

regression of the estimated squared residuals against the original X variables,
interactions of the variables, and polynomials of the variables. Then I obtained
an R

from the regressions. The chi square values obtained were all below

the critical chi square value at the .05 level of signifi cance.

Auto/Serial Correlation

Time series and panel data, by contrast, are data arranged in chronologi-
cal order (Gujarati 1995). Analyses of time series and panel data involve
examining changes in variables over time. In using such data, care must be
taken to avoid the problem of auto/serial correlation. The diffi culty with
observations of the same variable through time is that the observations
are correlated to some extent with each other. The error terms are thus
correlated (or auto correlated), meaning there is a defi nite pattern, which
violates a Gauss-Markov regression assumption.

Closely related to autocorrelation is serial correlation. Many now treat

the terms synonymously (Gujarati 1995, p. 410). There are several methods
both to detect and correct autocorrelation/serial correlation. The most often
used is the Durbin-Watson statistic (Greene, p. 591). The test is inaccurate
in this instance since this a pooled sample of years. Stimson suggests using a
pooled Durbin-Watson to test for such correlation. There are also different
methods to correct for autocorrelation. One accepted method is to use a
lagged dependent variable (Beck and Katz 1996; Scholz and Wood 1998),
and that method was used here.

Since each chapter used slightly different models, the next sections

review the methods chosen for each model in each chapter.

MODELS FOR CHAPTER THREE

Because one of the dependent variables is a count and the other is a ratio,
I used different methods for to test the hypotheses. The dependent variable,
litigation rate, simply counts the number of tax litigation fi lings by year. It is
a discrete outcome conforming roughly to a Poisson distribution. However,
data are produced by fi fty separate states, each with a range of qualities
likely to produce heterogeneous variation in event production greater than
the conditional population mean. The negative binomial model modifi es
the Poisson regression model by introducing an additional stochastic term
to the conditional mean, typically assumed to have a gamma distribution,
to refl ect individual heterogeneity often observed in cross-sectional data
(Cameron and Trivedi 1998, pp. 100–101). The distribution itself is similar
to Poisson. Because the negative binomial model reduces to Poisson when
␣ = 0, a test of this hypothesis was conducted to assess whether the data
are better described by a Poisson or negative binomial distribution, and the

100

GETTING A POOR RETURN

results of this test show that the data are overdispersed as expected from the
crosssection of American states over this extended period. Because some of
the variation in the data is national and thus constant per state per year I
cluster the data on the states and use robust standard errors. Finally, I use a
lagged dependent variable in the model to control for error autocorrelation.
I test the hypotheses based on the following equation:

(Litigation Rate

Lag) = _

+ `

(Litigation Rate) + `

(1998) + `

(1999) + `

(Per Capita Income) + `

(Audits)

+ `

(Unemployment)

(Assessments) + `

(Tax Court Ideology) + `

District Court

Ideology) + `

(Tax Court IRS Win Pct.) + `

(District Court IRS

Win Pct.) + `

(Congressional Ideology) + ¡.

The full results are reported in Table A.1.

For the second analyses of this chapter, the dependent variable is a

ratio, which reduces the risk of heteroskedasticity due to variation in state
size and other methodological problems associated with studies over time
such as nonstationarity. For the test of the “a” hypotheses, I use a regression
analysis with panel corrected standard errors and I test the hypotheses using
this panel design based on the following equation:

(Tax Court/District Court Ratio

Lag) = _

+ `

(Tax Court/District

Court Ratio) + `

(1998) + `

(1999) + `

(Per Capita Income) + `

(Audits)

+ `

(Unemployment) +

(Assessments) + `

(Tax Court

Ideology) + `

District Court Ideology) + `

(Tax Court IRS Win

Pct.) + `

(District Court IRS Win Pct) + `

(Congressional Ideol-

ogy) + ¡.

The full results are reported in Table A.2 (page 102).

MODELS FOR CHAPTER FOUR

The dependent variable for this chapter takes on two values: a value of “0”
when the tax payer wins, and “1” when the IRS wins. I estimate the model
using probit analysis in STATA 8, clustering the data on the judge for both
the Tax Court and District Court analyses.

Positive coeffi cients indicate a

greater likelihood of an IRS win. After dropping decisions due to lack of
obtainable data for all the independent variables, there were 631 decisions
for the Tax Court analysis and 185 decisions for the District Court analysis.
Thus, I model the likelihood that the IRS will win the case. I rely on a
probit estimation procedure and report robust standard errors, clustered on
judges, to control for the potential infl uence of any outlying observations

101

METHODOLOGY AND THE USE OF PANEL DATA

Table A.1 Determinants of Tax Litigation 1994–2000 (Chapter Three)

Variable Coeffi

cient

Lagged DV

.00002***

(.0000018)
Tax Court Ideology

12.60***

(.59)
District Court Ideology

–.08

(.09)
Congressional Ideology

1.56**

(.19)
Tax Court Win Pct.

.09**

(.03)
District Court Win Pct.

.01

(.03)
Assessments –.0003
(.0007)
1998 .30***
(.02)
Audits .0000006
(.0000005)
Citizen Ideology

.0006

(.0005)
Population .000000082
(.000000056)
Per Capita Income

.000003

(000002)
Unemployment Rate

.03***

(.006)
Appellate Court Ideology

.00002

(.002)
Constant 7.37***
(.15)

Wald Chi Square

3922.92***

N = 300

* = p < .05 (two tailed). ** = p < .01 (two tailed). *** = p < .001 (two tailed).

(Western 1995). I use a time-point dummy to control for the effect of any
particular year.

Positive coeffi cients indicate a greater likelihood of an IRS win. Since

the IRS wins most often, other scholars have used Rare Events logistic
regression to model this data (King and Lazarus 2003; see King and Zeng
2001a, 2001b), on the theory that conventional dichotomous regressions

102

GETTING A POOR RETURN

techniques will produce biased coeffi cients since the observance of a “1”
is particularly rare. However in the datasets used here, while the IRS wins
most often, the taxpayer still won twenty percent of the time in the Tax
Court, and thirty-two percent in the District Court. While the IRS wins
much more, it is hard to argue that a taxpayer win is a rare event, thus

Table A.2 Determinants of Tax Forum Choice 1994–2000 (Chapter Three)

Variable Coeffi

cient

Lagged DV

.86***

(.08)
Tax Court Ideology

1664.63***

(131.65)
District Court Ideology

–.23

(1.2)
Congressional Ideology

37.85***

(14.67)
Tax Court Win Pct.

.22

(.28)
District Court Win Pct.

–.45

(.28)
Assessments –.02
(.17)
1998 40.49***
(2.72)
1999 42.49***
(2.36)
Audits .0000094
(.000015)
Population .000000082
(.000000056)
Per Capita Income

–.0000017

(000016)
Unemployment Rate

.04

(.14)
Appellate Court Ideology

.12

(.48)
Constant –179.36***
(15.99)

Wald Chi Square

2269.46***

N = 300

p < .05 (two tailed).

** = p < .01 (two tailed).
*** = p < .001 (two tailed).

103

METHODOLOGY AND THE USE OF PANEL DATA

ordinary probit is appropriate.

After dropping decisions due to magistrate

decisions, or lack of obtainable data for all the independent variables, there
were 631 decisions for the Tax Court analysis and 185 decisions for the
District Court analysis.

The model has the following functional forms:

U.S. Tax Court:

Y(IRS Win)

= _ + `

(Per Capita Income) + `

(State Govern-

ment) + `

(IRS Experience) + `

(Special Judge) + `

(Attorney)

+ `

(Business) + `

(Estate or Trust) + `

(Income Issue) +

(Deduction Issue) + `

(Tax Shelter) + `

(Fraud/Tax Pro-

testor) + `

(Valuation Issue) + `

(Appellate Court Ideology)

+ `

(Ideology) + `

(Income*Ideology) + `

(Deduction*Ide

ology) + `

(Valuation*Ideology) + `

(Tax Shelter*Ideology) +

(Fraud/Tax Protestor*Ideology) + ¡

U.S. District Court:

Y(IRS Win)

= _ + `

(Per Capita Income) + `

(State Govern-

ment) + `

(Attorney) + `

(Business) + `

(Estate or Trust) +

(Income Issue) + `

(Deduction Issue) + `

(Tax Shelter) +

(Fraud/Tax Protestor) + `

(Issue) + `

(Appellate Court

Ideology) + `

(Ideology) + `

(Income*Ideology) + `

eduction*Ideology) + `

(Withholding *Ideology) + `

(Tax

Shelter*Ideology) + `

(Fraud/Tax Protestor*Ideology) + ¡

The coeffi cients are reported in Table A.3 (next page).

MODELS FOR CHAPTER FIVE

The analysis is similar to that used in chapter three. Since the model is a
panel of fi fty states over seven years, I used a fi xed effects OLS regression
with panel corrected standard errors to test the model. Specifi cally, I used
time-point dummy variables for each year, omitting the last year as the base-
line. Although the use of dummy variables for fi xed effects models is often
atheoretic, use of time-point dummies makes sense for this particular model.
Many of the years used in this model had particular political importance.
For example, 1994 represented the Republican takeover of Congress, and
1995 their fi rst year of power. The year 1996 was the presidential election
and 1997 and 1998 saw signifi cant tax reform efforts, including seeking
stepped-up audit enforcement of those claiming the EITC. The year 1998
was also a midterm election year that saw a slight Democratic gain in
the House of Representatives, while 1999 was the fi rst full year following

104

GETTING A POOR RETURN

Table A.3 Probability of Outcome Favoring IRS (Chapter Four)

Variable

Tax Court

District Court

Per Capita Income

–.000006

.00007

(.00003)

(.00004)

State Government

.004

(.005)

(.006)

IRS Experience

.022

(.137)
Special Judge

.464*

(.192)
Attorney –.334*

–.295

(.139)

(.511)

Business –.534**

–.607**

(.183)

(.243)

Estate or Trust

–.056

–.819**

(.275)

(.327)

Income Issue

–.042

.004

(.200)

(.273)

Deduction Issue

.0005

–.283

(.184)

(.303)

Tax Shelter

.243

.497

(.253)

(.576)

Fraud/Tax protester

5.11**

.325

(1.78)

(.509)

Valuation Issue

–.075

(.269)
Withholding Issue

.797*

(.358)

Appellate Court

–.325

5.18*

(2.112)

(2.42)

Ideology –1.13*

.358

(.522)

(.647)

Income * Ideology

.991

.037

.761

(.986)

Deduction * Ideology

1.33*

–.147

.623

(1.09)

Valuation * Ideology

.706

(.921)
Withholding * Ideology

–3.78***

(1.15)

Tax Shelter * Ideology

2.05*

–1.26

(.928)

(2.03)

Fraud/Protester * Ideology

–14.06**

–4.35*

4.61

(2.11)

Constant 1.06

–1.77

.917

(1.27)

N =

638

N =

185

Chi sq

= 342.43***

Chi sq = 45.78***

Pseudo

= .12

Pseudo R

= .14

+ = p < .01 (two tailed) * = p < .05 (two tailed) ** = p < .01 (two tailed)
*** = p < .001 (two tailed)

105

METHODOLOGY AND THE USE OF PANEL DATA

passage of the increased audit attention to EITC recipients. Several of the
variables control for the specifi c effect of the states. In addition, I ran both
models with fi xed effects, controlling fi rst for states by creating fi fty dummy
variables and leaving out the last state as the baseline and then with seven
year dummies. In neither model did the key ideology variables change in
the level of signifi cance or substantially change in impact.

Since states are of different sizes, the use of a ratio as the dependent

variable reduces the risk of heteroskedasticity due to variation in state size,
and the use of a ratio also reduces the risk of other methodological problems
associated with studies over time such as nonstationarity (Scholz and Wood
1998). Both a q norm and a Jacque-Bera test showed that the dependent
variables were skewed, so I logged the dependent variables to achieve a
normal distribution. With the log of the dependent variable, and the time-
point dummies, the models then become the following:

(Log of the Income Audit Ratio) = _

+ `

(Tax Court Ideology) + `

(State Specifi c District Court Ideology) +`

(District Court Ideological

Variation) + `

(Local Government) + `

(Per Capita Income) + `

(Unemployment)

+ `

(Urbanization) +

(Population Density) +

(Population) + Time-point dummies + ¡.

(Log of the Corporate Individual Audit Ratio) = _

+ `

(Tax Court

Median Ideology) + `

(State Specifi c District Court Median Ideology)

(District Court Ideological Variation) + `

(Local Government) +

(Per Capita Income) + `

(Unemployment)

+ `

(Urbanization) +

(Population Density) +

(Population) + Time-point dummies + ¡.

The coeffi cients for the low-income high-income ratio results are reported
in Table A.4 (next page) and the individual corporate ratio results in Table
A.5 (page 107).

106

GETTING A POOR RETURN

Table A.4 High-Income/Low-Income Audits (Chapter Five)

Variable Coeffi

cient

State Gov. Ideology

.005***

(.001)
Per Capita Income

–.00005***

(.000005)
Population Density

–.0002

(.00012)
Urbanization –.004***
(.0001)
Unemployment –.009
(.02)
Population 6.11E-09
(4.92E-09)
Tax Court Ideology

8.23***

(1.03)
District Court Ideology

1.00***

(.09)
Core .73***
(.10)
Pct. Change after Audit

.03**

(.01)
1994 .16***
(.04)
1995 .46***
(.02)
1996 .64***
(.03)
1997 .35***
(.05)
1998 1.14***
(.04)
1999 .61***
(.03)
Constant 6.45***
(.90)

N = 350
R

= .52

Chi Square = 1081.26***

* = p < .05 (two tailed)
** = p < .01 (two tailed)
*** = p < .001 (two tailed)

107

METHODOLOGY AND THE USE OF PANEL DATA

Table A.5 Individual/Corporate Audits (Chapter Five)

Variable Coeffi

cient

State Gov Ideology

–.022*

(.009)
Per Capita Income

–.00004

(.00009)
Population Density

–.002

(.002)
Urbanization .006
(.005)
Unemployment –.40
(.21)
Population –4.23E-09
–3.01E-08)
Tax Court Ideology

365.79**

(151.11)
District Court Ideology

6.79***

(1.76)
Core 19.92*
(8.72)
Pct. Change after Audit

–.25**

(.10)
1994 14.92***
(4.45)
1995 10.89***
(3.89)
1996 20.98***
(3.14)
1997 14.50***
(3.06)
1998 11.08*
(5.54)
1999 7.25
(4.72)
Constant –28.46
(22.32)

N = 350
R Squared = .53
Chi Square = 121.31***

* = p < .05 (two tailed)
** = p < .01 (two tailed)
*** = p < .001 (two tailed)

This page intentionally left blank.

109

METHODOLOGY AND THE USE OF PANEL DATA

CASES CITED

Brown v. Board of Education I 347 U.S. 483 (1954)
Ballard v. Commissioner of Internal Revenue 125 S.Ct. 1270 (2005)
Commissioner v. Glenshaw Glass Co. 348 U.S. 426 (1955)
Golson v. Commissioner 54 T.C. 742 (1970)
Griswold v. State of Connecticut 381 US 479 (1968)
Oregon v. Mitchell 400 U.S. 112 (1970)
Plessy v. Ferguson 163 U.S. 537 (1896)
Pollock v. Farmers’ Loan & Trust Co. 157 U.S. 429 (1895)
Roe v. Wade 410 US 113 (1973)
Steward Machine Co. v. Davis 301 U.S. 548 (1937)
United States v. Davis 370 U.S. 65 (1962)
Welch v. Helvering 290 U.S. 111 (1933)

109

This page intentionally left blank.

NOTES

CHAPTER ONE: TAX POLICY AND THE PURSUIT OF JUSTICE

1. Plessy v. Ferguson (163 U.S. 537, 559–560, 1896).

2. The fi rst act was passed July 6, 1861, although no income tax revenue

was collected under this act. The second and more important act was passed July
1, 1862. By this time, Congress knew the Civil War was going to last more than
ninety days, creating a need for revenue.

3. The offi ce of Commissioner of Internal Revenue as part of the Bureau of

Internal Revenue (Chommie 1970) was created to collect the income tax. Although
successful in fi nancing the Civil War, the income tax act expired in 1872. The
Bureau of Internal Revenue remained, with almost 4,000 employees to administer
and collect such other taxes as existed. That fi gure was to remain constant for the
next fi fty years.

4. The amendment states that: “Congress shall have power to lay and collect

taxes on incomes, from whatever source derived, without apportionment among the
several states, and without regard to any census or enumeration.”

5. Constitutional amendments have overturned four Supreme Court rulings.

The last was the Twenty-sixth Amendment, ratifi ed in 1971, that allowed eighteen
year olds the right to vote, overturning the Supreme Court decision of Oregon v.
Mitchell 400 U.S. 112 (1970).

6. Sears wrote: “It is intended that Part I shall be frank and devoid of that

mock patriotism which preaches from the taxing power’s vantage point and secretly
takes advantage of any loophole in the law. No tricks are advocated or alluded to
in this book, except to illustrate the fundamental differences between avoidance
and evasion” preface, p. iii.

7. Steward Machine Co. v. Davis 301 U.S. 548 (1937)
8. IRC §§6012 (a), 6072 (a), 6151 (a).
9. Quoted by Joy Mann, “We All End Up Paying for Deadbeats,” Washington

Post, January 6, 1995, p. E3. As evidence of the lack of compliance, Brand noted
that when the law was changed to require the listing of Social Security numbers for
dependents, “Seven million dependents disappeared in the United States.”

10. Chapter two contains much greater detail on audit processes and

procedures.

11. Specifi cally, the numbers are Warren—122, Burger—77, and Rehnquist—

28.

12. Commissioner v. Glenshaw Glass Co. 348 U.S. 426 (1955).

111

13. Welch v. Helvering 290 U.S. 111 (1933).
14. United States v. Davis 370 U.S. 65 (1962).

CHAPTER TWO: COURTS AND THE IRS

1. Offi ce of General Counsel, Federal Trade Commission, “Quarterly Federal

Court Litigation Status Report,” June 30, 2006.

2. IRC § 6213 (a).
3. According to Rule 80, a federal judge who serves for fi fteen years before

reaching the age of eighty can retire at full salary and thus at this point the salary
is guaranteed for life.

4. Congress has the power to create courts under the Constitution. The Tax

Court is a constitutional court by Congress under Article I as opposed to Article
III. Article III, Section 1 of the U.S. Constitution states: “The judicial Power of
the United States shall be vested in one supreme Court, and in such inferior Courts
as the Congress may from time to time ordain and establish.” These judges have
lifetime tenure and salary protection. By contrast, Article I legislative courts have
fi xed terms and lack the salary guarantee (Ducat 1996).

5. Criminal tax cases in the Federal District Court are prosecuted in a much

shorter time frame. A criminal defendant has a constitutional right to a speedy trial
before a jury.

CHAPTER THREE: TAX LITIGATION AND TAX FORUM CHOICE

1. For example, the late Paul Coverdall, Republican senator from Georgia,

noted the high incidence of audits on southerners, particularly poor southerners,
lamenting that “the emphasis on attacking the poor is unconscionable” (Pace
1998).

2. I exclude the U.S. Court of Federal Claims, a third alternative, from this

analysis, since fewer cases were litigated in that court (see Daily 1992) during the
years of this study.

3. This represents over 4,000 decisions for the Tax Court and over 700

decisions for the District Court.

4. Rules of Practice and Procedure, United States Tax Court, 2003. Title

XV, Rule 155, p. 96.

5. The party of each judge to have served on one of the modern circuits is

available in the Appellate Biographical Database (Zuk, Barrow, and Gryski 1997).

6. Segal, Timpone, and Howard also constructed a social liberalism score

(see Segal, Timpone, and Howard 1999).

7. In New York, for example, when Senator Alphonse D’Amato and Senator

Daniel Patrick Moynihan were colleagues, the junior Republican Senator D’Amato
allowed the senior Democratic Senator Moynihan one nomination in four when
Reagan and Bush held offi ce. During Clinton’s term, Moynihan followed the same
custom in reverse.

8. There has been an ongoing debate about whether the Tax Court has a pro

IRS bias because several Tax Court judges have IRS experience and because earlier

112

NOTES

113

NOTES

studies argued that taxpayers appear to do better in the District Courts than they
do in the Tax Court. However, the examination of the judges on the court did not
seem to support this. See chapter four for more information.

9. The appendix provides complete tables and details on the datasets and

methodology used for chapters two, three, and four.

CHAPTER FOUR: TAX DECISION MAKING

1. One journalist studying the tax system stated that the Ronald Reagan

appointee, Judge Cohen (chief justice during the examined time), had a “well
established reputation as a real softie for . . . tax dodgers who appeared before her”
(Johnston 2003).

2. Data for the U.S. Tax Court and the U.S. District Courts were compiled

through Westlaw (www.westlaw.com) with the keyword “tax” and the appropriate
year added to the search.

3. For the District Court, it was the state of the appropriate district. For the

Tax Court, in almost all cases, the opinion cited the state of the taxpayer. When
that did not occur, the original docket fi ling was examined to determine the state
where the case was fi led.

CHAPTER FIVE: INFLUENCES ON THE IRS AND THE

AUDITS OF LOW-INCOME TAXPAYERS

1. As in chapter three, the Court of Claims is omitted from this analysis.

APPENDIX: METHODOLOGY AND THE USE OF PANEL DATA

1. I use event history analysis with probabilities because I am interested in

the likelihood or probability of a decision occurring as opposed to the timing or
duration of time before an event occurs. The latter can be determined through a
Cox proportional hazards duration model (Box-Steffensmeier and Jones 2004). Simply
put, I wanted to know the probability of an event occurring, not when it would
occur. However, I also tested the model using a Cox duration model. The model
performed similarly to the event history analysis, with most coeffi cients in the same
direction, with similar, but slightly smaller, chi squares. Given that I was looking for
probabilities, not timing, and given that the event history model provides a better
fi t than the Cox model, I present the results of the event history model.

2. The analyses are premised on the assumption that the rulings between

judges are independent, but not the rulings by each particular judge. I did run both
analyses using Rare Events logit and found no change in statistical signifi cance from
the probit analysis. Results from the Relogit are available on request. I also ran
both models with the logged defi ciency amount as an independent variable. While
for many cases the information is readily available, for many partnership matters it
is not, since the income/loss is reported on the individual return, and the District
Courts often failed to be precise in stating the amount being contested. When I
included the defi ciency, the key independent variables showed little change. Given

114

NOTES

the loss of degrees of freedom with little increase in explanatory power, I omitted
this variable from the analyses presented here.

3. There is a potential for bias in a study comparing decisions in the Tax

Court and the District Courts in that there is a possibility that there are distinctly
different types of litigants in each court, with the District Court wealthier on average,
and therefore that this wealth causes some systematic differences in treatment. Con-
trolling for type of litigant and attorney should systematically control for wealthier
people, who would be more likely to be represented by an attorney, and would have
more complex claims.

4. I ran an analysis, not reported here, that controlled for assessment, a direct

measure of wealth, and this measure did not alter the results. All of these variables
control for systematic bias in the type of litigants and therefore systematic bias in
treatment. I added a dummy variable for 1996 to account for any time-specifi c effects
for the district court (Beck, Katz, and Tucker 1998). For ease of comparison, the
results omit this control variable.

115

REFERENCES

Associated Press. 2006. “IRS Urged To Restructure Fraud Program.” MSNBC.com.

January 19.

Bailey, Michael, and Kelly H. Chang. 2002. “Comparing Presidents, Senators, and

Justices: Interinstitutional Preferences.” Journal of Law, Economics and Orga-
nization 17: 477–506.

Baum, Lawrence. 1990. American Courts: Process and Policy. Boston: Houghton

Miffl in.

Baum, Lawrence. 1995. “Specialization and Authority Acceptance: The Supreme Court

and Lower Federal Courts.” Political Research Quarterly 47(3): 693–703.

Bebchuk, Lucian. 1984. “Litigation and Settlement Under Imperfect Information.”

Rand Journal of Economics 15: 404–415.

Beck, Nathaniel, and Jonathan N. Katz. 1995. “What To Do (and Not To Do) with

Times-Series Cross-Section Data in Comparative Politics.” American Political
Science Review 89: 634–647.

Beck, Nathaniel L., Jonathan Katz, and Richard Tucker. 1998. “Taking Time Seri-

ously: Time-Series-Cross-Section Analysis with a Binary Dependent Variable.”
American Journal of Political Science 42(4): 1260–1288.

Beck, Paul Allen, and Richard Funston 1976. “Letters and Responses.” American

Political Science Review 70: 931–932.

Berall, Frank S. 1992. “Litigating with the IRS: Choosing Your Forum.” The Practical

Tax Lawyer 6(2): 75.

Berry, William D., and Stanley Feldman. 1985. Multiple Regression in Practice. Sage

university paper series on quantitative applications in the social sciences,
series no. 07-050. Beverly Hills: Sage Publications.

Berry, William D., Evan J. Ringquist, Richard C. Fording, and Russell L. Hanson.

1998. “Measuring Citizen and Government Ideology in the American States.”
American Journal of Political Science 42: 337–346.

Bosworth, Matthew H. 2001. Courts as Catalysts: State Supreme Courts and Public

School Finance Equity. Albany: State University of New York Press.

Box-Steffensmeier, Janet M., and Bradford S. Jones. 2004. Event History Modeling: A

Guide for Social Scientists. Cambridge and New York: Cambridge University
Press.

Brace, Paul, and Melinda Gann Hall. 1990. “Neo-Institutionalism and Dissent in

State Supreme Courts.” Journal of Politics 53: 54–70.

Burnham, David. 1989. A Law Unto Itself: Power, Politics and the IRS. New York:

Random House.

115

116

REFERENCES

Caldeira, Gregory A., and John R. Wright. 1988. “Organized Interests and

Agenda-Setting in the Supreme Court.” American Political Science Review
82: 1009–1127.

Cameron, A. Colin, and Pravin K. Trivedi. 1998. Regression Analysis of Count Data.

New York: Cambridge University Press.

Canon, Bradley, and Sidney Ulmer. 1976. “The Supreme Court and Critical Elections:

A Dissent.” American Political Science Review 70: 1215–1218.

Chommie, John C. 1970. The Internal Revenue Service. New York: Praeger.
Clark, David S. 1990. “Civil Litigation Trends in Europe and Latin America Since

1945: The Advantage of Intracountry Comparisons.” Law & Society Review
24: 549–569.

Cortner, Richard C. 1968. “Strategies and Tactics of Litigants in Constitutional

Cases.” Journal of Public Law 17: 287–307.

Crenshaw, Albert B. 2002. “IRS Audits Increase for Low-Income Taxpayers.” Wash-

ingtonpost.com. April 15, p. A02.

Crouch, Holmes F. 1996. Going into Tax Court: Try Hard; Be Alert It’s a Stacked

Deck. Saratoga, CA: Allyear Tax Guides.

Dahl, Robert A. 1957. “Decision-Making in a Democracy: The Supreme Court as a

National Policy-Maker.” Journal of Public Law 6: 284–295.

Daily, Frederick W. 1992. Stand Up to the IRS. Berkeley: Nolo Press.
Dixit, Avinash K. 1995. The Making of Economic Policy: A Transaction Cost Politics

Perspective. Cambridge: MIT Press.

Dixit, Avinash K., Gene M. Grossman, and Elhanan Helpman. 1997. “Common

Agency and Coordination: General Theory and Application to Government
Policy Making.” Journal of Political Economy 105(4): 752–769.

Downs, George W., and David M. Rocke. 1979. “Interpreting Heteroscedasticity.”

American Journal of Political Science 23: 816–828.

Dubroff, Harold. 1979. The United States Tax Court: An Historical Analysis. Chicago:

Commerce Clearing House.

Ducat, Craig. 1996. Constitutional Interpretation: Powers of Government. Volume I,

6th ed. Minneapolis: West Publishing.

Easterbrook, Frank H. 1990. “Success and the Judicial Power.” Indiana Law Journal

65: 277–282.

Eisenberg, Theodore, and Lynn M. LoPucki. 1998–1999. “Shopping for Judges: An

Empirical Analysis of Venue Choice in Large Chapter 11 Reorganizations.”
Cornell Law Review 84: 967–1003.

Epstein, Lee. 1985. Conservatives in Court. Knoxville: University of Tennessee

Press.

Epstein, Lee, and Jeffrey A. Segal. 2005. Advice and Consent: The Politics of Judicial

Appointment. Oxford: Oxford University Press.

Epstein, Lee, Jeffrey A. Segal, Harold J. Spaeth, and William G. Walker. 2003.

The Supreme Court Compendium: Data, Decisions and Developments. 3rd ed.
Washington, DC: CQ Press.

Fogel, David, and Joe Hudson. 1981. “Justice as Fairness: An Overview,” in Justice

as Fairness: Perspectives on the Justice Model. David Fogel and Joe Hudson,
eds. Ottawa: Anders Publishing.

Freeland, James J., Stephen A. Lind, and Richard B. Stevens. 1977. Fundamentals of

Federal Income Taxation. 2d. ed. Mineola, NY: Foundation Press.

117

REFERENCES

Friedman, Lawrence M. 2002. American Law in the 20

Century. New Haven and

London: Yale University Press.

Funston, Richard. 1975. “The Supreme Court and Critical Elections.” 69 American

Political Science Review 69: 795–811.

Garbis, Marvin, J., and Allen L. Schwait. 1971. Tax Refund Litigation. Philadelphia:

American Law Institute.

Geier, Deborah A. 1991. “The Tax Court, Article Three, and the Proposal Advanced

by the Federal Courts Study Committee: A Study in Applied Constitutional
Theory.” Cornell Law Review 76: 985–1035.

Giles, Michael W. 2008. “Commentary on “Picking Federal Judges: A Note on

Policy and Partisan Selection Agendas.” Political Research Quarterly 61(1):
53–55.

Giles, Michael W., and William G. Walker. 1975. “Judicial Policy-Making and

Southern School Segregation.” Journal of Politics 37(4): 917–936.

Giles, Micheal W., Virgina A. Hettinger, and Todd Peppers. 2001. “Picking Federal

Judges: A Note on Policy and Partisan Selection Agendas.” Political Research
Quarterly 54: 623–641.

Giles, Micheal W., Virgina A. Hettinger, and Todd C. Peppers. 2002. “Measuring

the Preferences of Federal Judges: Alternatives to Party of the Appointing
President.” Emory University typescript.

Gillman, Howard. 2002. “How Parties Can Use the Courts to Advance Their

Agendas: Federal Courts in the United States, 1875–91.” American Political
Science Review 96: 511–524.

Green, William H. 1997. Econometric Analysis. 3d ed. Englewood Cliffs, NJ: Prentice

Hall.

Gujarati, Damodar N. 1995. Basic Econometrics. 3rd ed. New York: McGraw Hill.
Hall, Kermit L., Paul Finkelman, and James W. Ely, Jr. 2005. American Legal History.

3rd ed. New York: Oxford University Press.

Hall, Melinda Gann. 1992. “Electoral Politics and Strategic Voting in State Supreme

Courts.” Journal of Politics 54: 427–446.

Hammond, William H., and Jack H. Knott. 1996. “Who Controls the Bureaucracy?

Presidential Power, Congressional Dominance, Legal Constraints, and Bureau-
cratic Autonomy in a Model of Multi-Institutional Policy-Making.” Journal
of Law, Economics and Organization 12: 119–166.

Hammond, William H., and Gary J. Miller. 1987. “The Core of the Constitution.”

American Political Science Review 81: 1155–1174.

Hansen, Wendy L., Renee J. Johnson, and Isaac Unah. 1995. “Specialized Courts,

Bureaucratic Agencies, and the Politics of U.S. Trade Policy.” American Journal
of Political Science 39(3): 529–557.

Hausegger, Lori, and Stacia Haynie. 2003. “Judicial Decisionmaking and the Use

of Panels in the Canadian Supreme Court and the South African Appellate
Division.” Law and Society Review 37: 635.

Hettinger, Virginia A., Wendy L. Martinek, and Stephanie A. Lindquist. 2004.

“Comparing Attitudinal and Strategic Accounts of Dissenting Behavior on
the United States Courts of Appeals.” American Journal of Political Science
48(1): 123–137.

Howard, Robert M. 2007 “Controlling Forum Choice and Controlling Policy: Con-

gress, Courts and the IRS.” Policy Studies Journal 35(1): 109–123.

118

REFERENCES

Howard, Robert M. 2008. “Courts, Justice and Governing Coalitions and the Audits

of Low-Income Taxpayers.” Journal of Theoretical Politics 20(2): 181–200.

Howard, Robert M., and David C. Nixon. 2002. “Regional Infl uences Within a

Separation of Powers Framework: Courts, Ideological Preferences and IRS
Local Policymaking.” Political Research Quarterly (4): 55–907–922.

Howard, Robert M., and David C. Nixon. 2003. “Local Control of the Bureaucracy:

Federal Appeals Courts, Ideology, and the Internal Revenue Service.” Wash-
ington University Journal of Law & Policy 13: 233–356.

Humphries, Martha Anne, and Donald R. Songer. 1999. “Law and Politics in

Judicial Oversight of Federal Administrative Agencies.” Journal of Politics
61: 207–220.

Internal Revenue Service. 1979. Estimates of Income Unreported on Individual Income

Tax Returns. Publication 1104 (9–79). Washington, DC: Department of the
Treasury.

Johnston, David Cay. 1996. Your Taxes; Some New Tricks to Help Filers Avoid an

Old Audit Trap. New York Times, February 25. http://query.nytimes.com/gst/
fullpage.html?res=9500E1D71039F936A15751C0A960958260.

Johnston, David Cay. 2000. “Inquiries Find Little Abuse by Tax Agents.” New York

Times, August 15. http://query.nytimes.com/gst/fullpage.html?res=9D05E1DA
113FF936A2575BC0A9669C8B63&scp=6&sq=irs%20abuse&st=cse

Johnston, David Cay. 2003. Perfectly Legal: The Covert Campaign to Rig Our Tax System

to Benefi t the Super Rich—And Cheat Everybody Else. New York: Penguin.

Juenger, Friedrich K. 1990. “Forum Shopping, Domestic and International.” Tulane

Law Review 63: 553.

Katzmann, Robert A. 1980. “Federal Trade Commission.” in The Politics of Regulation,

ed. James Q. Wilson. New York: Basic Books.

Keir, Lloyd, Douglas W. Argue, and Brian J. Seery. 1981. Tax Court Practice. 6th ed.

Philadelphia: American Law Institute.

King, Chad M., and Ellen Lazarus. 2003. “Decision Making on the United States

Tax Court.” Paper presented at the 2003 Annual Meeting of the Midwest
Political Science Association, Chicago, April 2003.

King, Gary, and Langche Zeng, 2001a. “Logistic Regression in Rare Events Data.”

Political Analysis 9: 137–163.

King, Gary, and Langche Zeng, 2001b. “Explaining Rare Events in International

Relations.” International Organization, 55: 693–715.

Krehbiel, Keith. 1998. Pivotal Politics. Chicago: University of Chicago Press.
Kroll, Glenn. 1996. “Are Tax Court Judges Partial to the Public?” Oil & Gas Tax

Quarterly. 45: 135.

Kurtz, J., and J. A. Pechman. 1982. “Tax Fraud and Hyperbole.” New York Times,

July 12, p. 15.

Landes, William M. 1971. “An Economic Analysis of the Courts.” Journal of Law

and Economics 14: 61–107.

Long, Susan. 1980. The Internal Revenue Service: Measuring Tax Offenses and Enforce-

ment Response. Washington, DC: National Institute of Justice.

MacDonald, Elizabeth. 1994. “How the IRS Targets You.” MONEY 23: 132–139.
MacLachlan, Claudia. 1995. “The Tax Bench: A Code Apart.” National Law Journal,

September 27, p. 1.

119

REFERENCES

Mann, Joy. 1995 “We All End Up Paying For Deadbeats.” Washington Post, p. E3,

Friday January 6th.

Martin, Andrew, Kevin M. Quinn, and Lee Epstein. 2004. “North Carolina Law

Review: Symposium on Locating the Constitutional Center—Centrist Judges
and Mainstream Values: A Multidisciplinary Exploration.” North Carolina Law
Review 83: 1275–1317.

Maule, James Edward. 1999. “Instant Replay, Weak Teams, and Disputed Calls:

An Empirical Study of Alleged Tax Court Bias.” Tennessee Law Review 66:
351–426.

McCann, Michael. 1994. Rights at Work: Pay Equity Reform and the Politics of Legal

Mobilization. Chicago: University of Chicago Press.

McCarty, Nolan, Keith T. Poole, and Howard Rosenthal. 2006. Polarized America: The

Dance of Ideology and Unequal Riches. Cambridge and London: MIT Press.

McClosky, Herbert, and John Zaller. 1984. The American Ethos: Public Attitudes Towards

Capitalism and Democracy. Cambridge: Harvard University Press.

McCubbins, Matthew, Roger Noll, and Barry Weingast. 1987. “Administrative Pro-

cedures as Instruments of Control.” Journal of Law Economics and Organization
3: 243–277.

McCubbins, Matthew D., Roger G. Noll, and Barry R. Weingast. 1989. “Symposium

on the Law and Economics Of Bargaining: Structure and Process, Politics and
Policy: Administrative Arrangements and the Political Control of Agencies.”
Virginia Law Review 75: 431–482.

McIntosh, Wayne V. 1990. The Appeal of Civil Law: A Political-Economic Analysis of

Litigation. Urbana and Chicago: University of Illinois Press.

Moe, Terry. 1985. Control and Feedback in Economic Regulation: The Case of the

NLRB. American Political Science Review 79(4): 1094–1116.

Nixon, David C. “Measuring Judicial Ideology” (n.d.) www2.hawaii.edu/~dnixon/

pimp/judicial.pdf.

Offi ce of General Counsel. Federal Trade Commission. 2006. “Quarterly Federal

Court Litigation Status Report.” June 30.

Olson, Susan. 1990. “Interest Group Litigation in Federal District Court: Beyond the

Political Disadvantage Theory.” Journal of Politics 52: 854–882.

Pace, David. 1998. “Study: IRS Audits Poor Most.” Associated Press, February 28.
Perry, H. W. 1991. Deciding to Decide. Cambridge: Harvard University Press.
Pickerill, J. Mitchell, and Cornell W. Clayton. 2004. “The Rehnquist Court and the

Political Dynamics of Federalism.” Perspectives on Politics 2 (June): 233–248.

Poole, Keith T. 1998. “Recovering a Basic Space from a Set of Issue Scales.” American

Journal of Political Science 42: 954–993.

Poole, Keith T., and Howard Rosenthal. 1997. Congress: A Political-Economic History

of Roll Call Voting. New York: Oxford University Press.

Posner, Richard A. 1973. “An Economic Approach to Legal Procedure and Judicial

Administration.” Journal of Legal Studies 2: 399–458.

Posner, Richard A. 1996. The Federal Courts: Challenge and Reform. Cambridge:

Harvard University Press.

Quinn Dennis P., and Robert Y. Shapiro. 1991. “Economic Growth Strategies: The

Effects of Ideological Partisanship on Interest Rates and Business Taxation in
the United States.” American Journal of Political Science 35(3): 656–685.

120

REFERENCES

Reinganum, Jennifer F. 1988. “Plea Bargaining and Prosecutorial Discretion.” American

Economic Review 78: 713–727.

Reinganum, Jennifer F., and Louis L. Wilde. 1986. “Settlement, Litigation, and the

Allocation of Litigation Costs.” Rand Journal of Economics 17: 557–567.

Ringquist, Evan J., and Craig F. Emmert. 1999. “Judicial Policy-Making in Published

and Unpublished Decisions: The Case of Environmental Civil Litigation.”
Political Research Quarterly 52: 7–38.

Rosenberg, Gerald N. 1991. The Hollow Hope: Can Courts Bring About Social Change?

Chicago: University of Chicago Press.

Rossiter, Clinton, ed. 1961. The Federalist Papers: Alexander Hamilton, James Madison,

and John Jay. Federalist 78. New York: Mentor.

Roth, Jeffrey A., John T. Scholz, and Ann Dryden Witte, eds. 1989. Taxpayer Compli-

ance, Volume I. Philadelphia: University of Pennsylvania Press.

Rowland, C. K., and Robert A. Carp. 1996. Politics and Judgment in Federal District

Courts. Lawrence: University of Kansas Press.

Rules of Practice and Procedure. 2003. United States Tax Court. http://www.

ustaxcourt.gov/notice.htm.

Russell, Peter H., and David M. O’Brien, eds. 2001. Judicial Independence in the Age

of Democracy: Critical Perspectives from Around the World. Charlottesville and
London: University of Virginia Press.

Schneider, Daniel M. 2001. “Empirical Research on Judicial Reasoning: Statutory

Interpretation in Federal Tax Cases.” New Mexico Law Review 31: 325–352.

Schneider, Daniel M. 2002. “Assessing and Predicting Who Wins Federal Tax Trial

Decisions.” Wake Forest Law Review 37: 473–519.

Scholz, John T., and B. Dan Wood. 1998. “Controlling the IRS: Principals, Prin-

ciples and Public Administration.” American Journal of Political Science 42:
141–162.

Scholz, John T., and B. Dan Wood. 1999. “Effi ciency, Equity and Politics: Demo-

cratic Controls over the Tax Collector.” American Journal of Political Science
43: 1166–1188.

Sears, John H. 1922. Minimizing Taxes. Kansas City: Vernon Law Book Co.
Segal, Jeffrey A., and Harold J. Spaeth. 1993. The Supreme Court and the Attitudinal

Model. Cambridge and New York: Cambridge University Press.

Segal, Jeffrey A., and Harold J. Spaeth. 2002. The Supreme Court and the Attitudinal

Model Revisited. Cambridge and New York: Cambridge University Press.

Segal, Jeffrey A., Richard C. Timpone, and Robert M. Howard. 2000. “Buyer Beware?

Presidential Infl uence Through Supreme Court Appointments.” Political
Research Quarterly 53: 557–573.

Shafi roff, Ira L. 2004. Internal Revenue Service Practice and Procedure. New York:

Practicing Law Institute

Shapiro, Martin M. 1988. Who Guards the Guardians: Judicial Control of Administrative

Acts. Athens: University of Georgia Press.

Shipan, Charles R. 1998. “Political Infl uence, Regulatory Regimes, and the FDA.”

Paper presented at the Annual Meeting of the Public Choice Society. New
Orleans, LA.

Shipan, Charles R. 2000. “The Legislative Design of Judicial Review: A Formal

Analysis.” Journal of Theoretical Politics 12(3): 269–304.

121

REFERENCES

Smith, Joseph L. 2006. “Judicial Procedures as Instruments of Political Control:

Congress’s Strategic Use of Citizen Suits.” Legislative Studies Quarterly 31(May):
283–305.

Songer, Donald R., Charles M. Cameron, and Jeffrey A. Segal. 1995. “An Empiri-

cal Test of the Rational Actor Theory of Litigation.” Journal of Politics 57(4)
1119–1129.

Steuerle, Eugene. 1986. Who Should Pay for Collecting Taxes? Washington, DC:

American Enterprise Institute.

Steuerle, Eugene. 1991. The Tax Decade: How Taxes Came to Dominate the Public

Agenda. Washington, DC: Urban Institute Press.

Stevenson, Richard W. 1997. “Leading Republican Plans Bill to Shift Burden of

Proof in Tax Disputes.” Infoseek Business, October 20.

Stookey, John A. 1990. “Trials and Tribulations: Crises, Litigation, and Legal Change.”

Law & Society Review 24: 497–520.

Strassels, Paul N. 1979. All You Need to Know About the IRS: A Taxpayer’s Guide.

New York: Random House.

Tate, C. Neil, and Roger Handberg. 1991. “Time Binding and Theory Building in

Personal Attribute Models of Supreme Court Voting Behavior.” American
Journal of Political Science 35: 460–480.

Taylor, Marshall, W., Karen J. Simonson, Marc J. Winter, and Brain J. Seery. 1990.

Tax Court Practice. Philadelphia: American Law Institute.

Tidrick, Donald. 2004. “Inside the U.S. Tax Court: An Interview with the Honor-

able Judge Juan F. Vasquez, C.P.A., U.S. Tax Court.” The CPA Journal. www.
nysscpa.org/cpajournal/2004/104/text/p.20.htm.

Tracfed. 2006. http://trac.syr.edu/tracirs/latest/current/.
TracIRS. 2005. http://trac.syr.edu/tracirs/index.html.
Truman, David B. 1971. The Governmental Process: Political Interests and Public Opinion.

2nd ed. New York: Knopf.

Tyler, Tom R. 1990. Why People Obey the Law. New Haven: Yale University Press.
Unah, Isaac. 1997. “Specialized Courts of Appeals’ Review of Bureaucratic Actions and

the Politics of Protectionism.” Political Research Quarterly 50(4): 851–887.

Unah, Isaac. 1998. The Courts of International Trade: Judicial Specialization, Expertise,

and Bureaucratic Policymaking. Ann Arbor: University of Michigan Press.

Unah, Isaac. 2003. “Explaining Corporate Litigation Activity in an Integrated Frame-

work of Interest Mobilization.” Business and Politics 5(1): 65–94.

Wade, Jr., Jack Warren. 1986.

Audit-Proofi ng Your Tax Return. New York: Macmillan.

Western, Bruce. 1995. “Concepts and Suggestions for Robust Regression Analysis.”

American Journal of Political Science 39: 786–817.

Whittington, Keith E. 2005. “ ‘Interpose Your Friendly Hand’: Political Supports

for the Exercise of Judicial Review by the United States Supreme Court.”
American Political Science Review 99(4): 583–596.

Wiseman, Paul. 1997a. “IRS May Not be the Monster Critics Say It Is.” USA Today,

November 5, section A, p.17.

Wiseman, Paul. 1997b. “House Set to Vote Today on Overhaul of Tax Agency.”

USA Today, November 5, section A, p. 17.

Witte, John. 1985. The Politics and Development of the Federal Income Tax. Madison:

University of Wisconsin Press.

122

REFERENCES

Zuk, Gary, Deborah J. Barrow, and Gerard S. Gryski. Multi-User Database on the

Attributes of United States Appeals Court Judges, 1801–2000 [Computer fi le].
ICPSR06796-v2. Ann Arbor, MI: Inter-university Consortium for Political
and Social Research [distributor], 2009-02-03. doi:10.3886/ICPSR06796.

123

INDEX

123

Anders, Joy, 2, 3, 13, 14, 18, 27, 29,

30, 31, 33, 47, 51, 54, 73, 74, 91,
92, 96

Argue, Douglas, W., 42

Bailey, Michael, 47
Ballard v. Commissioner of
Internal Revenue, 47
Barrow, Deborah J., 112 n.5
Baum, Lawrence, 55, 56, 57
Bebchuk, Lucian, 33
Beck, Nathaniel, 98, 99, 114
Beck, Paul Allen, 95
Beghe, Renato, 58
Berall, Frank S., 35, 36, 38
Berry, William D., 66, 83, 98
Bosworth, Matthew H., 30
Box-Steffensmeier, Janet M., 113 n.1
Brace, Paul, 55
Brown v. Board of Education, 30
Burnham, David, 10, 20, 75, 76
Bush, George H.W., 10, 60, 112 n.7
Bush, George W., 3, 10

Caldeira, Gregory A., 32
Cameron, A. Colin, 99
Cameron, Charles M., 33, 55
Canon, Bradley, 95
Carp, Robert A., 55, 63
Chabot, Herbert, 58
Chang, Kelly H., 47
Chiechi, Carolyn, 58
Chommie, John C., 7, 111 n.3
Clark, David S., 33
Clayton, Cornell W., 88
Cleveland, Grover, 5, 6

Clinton, William, 3, 10, 37, 44, 60,

112 n.7

Cohen, Mary Ann, 58
Colvin, John O., 58
Commissioner v. Glenshaw Glass
Co., 111 n.12
Cortner, Richard C., 31, 32
Crenshaw, Albert B., 73, 74
Crouch, Holmes F., 75

Dahl, Robert A., 2, 59, 71, 73, 74, 88,

Daily, Frederick, W., 10, 18, 23, 24, 25,

112 n.2

Dawson, How., 58
Dixit, Avinash, K., 77
Downs, George W., 98
Dubroff, Harold, 22, 42, 57, 60, 81
Ducat, Craig, 112 n.4

Easterbrook, Frank H., 60, 81
Eisenberg, Theodore, 36
Eisenhower, Dwight, 9
Ely, Jr., James W., 5
Emmert, Craig, F., 74
Epstein, Lee, 11, 21, 32, 59, 76, 79

Feldman, Stanley, 98
Finkelman, Paul, 5
Fogel, David, 2
Foley, Maurice, B., 58
Fording, Richard C., 66, 83
Forrest, David and Barbara, 18, 29, 31,

91, 92

Freeland, James J., 10, 18

124

INDEX

Friedman, Lawrence M., 7, 11
Funston, Richard, 95

Gale, Joseph H., 58
Garbis, Marvin J., 35, 38
Geier, Deborah A., 56, 71
Gerber, Joel, 58
Giles, Michael W., 45, 63, 66
Gillman, Howard, 88
Goeke, Joseph R., 58
Golson v. Commissioner, 81
Griswold v. Connecticut, 30
Green, William H., 99
Grossman, Gene M., 77
Gryski, Gerard S., 112 n.5
Gujarati, Damodar N., 98, 99

Handberg, Roger, 44, 45
Haines, Harry A., 58
Hall, Kermit L., 5
Hall, Melinda Gann, 55
Halpern, James S., 58
Hamilton, Alexander, 53
Hammond, William, H., 77
Hand, Learned, 1, 4, 8, 12
Hansen, Wendy L., 55, 56
Hanson, Russell L., 66, 83
Harlan, John Marshall, 2, 52
Hausseger, Lori, 79
Haynie, Stacia, 79
Helpman, Elhanan, 77
Hettinger, Virginia A., 45, 55
Holmes, Mark V., 58
Holmes, Oliver Wendell, 1, 4, 12
Howard, Robert M., 38, 44, 45, 46, 54,

59, 74, 76, 112 n.6

Hudson, Joe, 2
Humphries, Martha Anne, 44, 45

Internal Revenue Service, 2, 3, 4, 9,

10, 11, 12, 13, 14, 17, 18, 19, 20,
21, 22, 23, 24, 27, 29, 30, 31, 32,
33, 34, 35, 36, 37, 38, 39, 40, 41,
42, 43, 47, 49, 51, 52, 53, 54, 55,
56, 57, 58, 59, 60, 62, 63, 64, 65,
66, 67, 68, 69, 70, 71, 73, 74, 75,

76, 77, 78, 79, 80, 81, 82, 83, 84,
85, 87, 88, 90, 92, 93, 94, 100, 101,
102, 103, 104, 112 n.8

Jacobs, Julian, I., 58
Jones, Bradford S., 113
Johnson, Renee, J., 55, 56
Johnston, David Cay, 19, 20, 59, 71,

73, 113 n.1

Juenger, Friedrich K., 36

Katz, Jonathan, 98, 99, 114
Katzmann, Robert A., 75
Keir, Lloyd, 42
King, Chad M., 62, 101
King, Gary, 101
Knot, Jack H., 77
Krehbiel, Keith, 83
Kroll, Glenn, 57
Kroupa, Diane L., 58
Kurtz, J., 10

Landes, William M., 33
Laro, David, 58
Lazarus, Ellen, 62, 101
Lind, Stephen A., 10, 18
Lindquist, Stephanie A., 55
Long, Susan, 75
LoPucki, Lynn M., 36

MacDonald, Elizabeth, 75
MacLachlan, Claudia, 21, 25, 27
Mann, Joy, 111 n.9
Martin, Andrew, 79
Martinek, Wendy L., 55
Marvel, L. Paige, 58
Maule, James Edward, 57, 60, 62, 81
McCann, Michael, 73, 74
McCarthy, Nolan, 43
McClosky, Herbert, 34
McCubbins, Matthew, 32, 55, 56
McIntosh, Wayne V., 33
Miller, Gary J., 77
Moe, Terry, 73, 74

Nims III, Arthur, 58

125

INDEX

Nixon, David C., 38, 45, 46, 54, 59,

74, 76

Noll, Roger G., 32, 55, 56

O‘Brien, David M., 53
Olson, Susan, 31, 32
Oregon v. Mitchell, 111 n.5

Pace, David, 112 n.1
Pechman, J.A., 10
Peppers, Todd C., 45
Perry, H.W., 11, 21
Pickerill, J. Mitchell, 88
Plessy v. Ferguson, 2, 52, 111 n.1
Pollock v. Farmers‘ Loan & Trust Co.,

5, 6

Poole, Keith T., 43, 60, 83
Posner, Richard A., 33, 60, 71, 81

Quinn, Dennis P., 54
Quinn, Kevin M., 79

Reagan, Ronald, 9, 44, 60, 76, 112 n.7,

113 n.1

Reinganum, Jennifer F., 33
Ringquist, Evan J., 74
Rocke, David M., 98
Roe v. Wade, 30
Roosevelt, Franklin D., 8, 44, 95
Rosenberg, Gerald N., 74
Rosenthal, Howard, 43
Rossiter, Clinton, 53
Roth, Jeffrey A., 10
Rowland, C. K., 55, 63
Russell, Peter H., 53
Ruwe, Robert P., 58

Schneider, Daniel M., 42, 62
Scholz, John T., 10, 38, 41, 54, 66, 76,

82, 87, 99, 105

Schwait, Allen L., 35, 38
Sears, John H., 8, 111 n.6
Seery, Brian J., 35, 38, 42
Segal, Jeffrey A., 33, 44, 45, 46, 55,

59, 62, 73, 112 n.6

Shafi roff, Ira L., 19

Shapiro, Martin M., 75, 82
Shapiro, Robert Y., 54
Shipan, Charles R., 31, 32, 78
Simonson, Karen J., 35, 38
Smith, Joseph L., 30, 31, 79, 81
Songer, Donald R., 33, 44, 45, 46,

Spaeth, Harold J., 55
Steuerle, Eugene, 3, 10
Stevens, Richard B., 10, 18
Stevenson, Richard W., 38
Steward Machine v. Davis, 111 n.7
Stookey, John A., 33
Strassels, Paul N., 75
Swift, Stephen J., 58

Taft, William, 6
Tate, C. Neil, 44, 45
Taylor, Marshall, 35, 38
Thornton, Michael, 58
Tidrick, Donald, 38, 57
Timpone, Richard C., 44, 112 n.6
Tucker, Richard, 114
Trivedi, Pravin K., 99
Tyler, Tom R., 25

Ulmer, Sidney, 95
Unah, Isaac, 31, 32, 38, 55, 56, 59
United States v. Davis, 112 n.14

Vasquez, Juan F., 58

Wade, Jr., Jack Warren, 55
Walker, William G., 63, 66
Warren Court, 11, 21
Warren, Earl, 111 n.11
Weingast, Barry, 32, 55, 56
Welch v. Helvering, 112 n.13
Wells, Thomas B., 58
Western, Bruce, 101
Whalen, Laur., 58
Wherry, Robert, 58
Whittington, Keith E., 2, 5, 88, 94
Wilde, Louis L., 33
William, James, 18, 29, 54, 91, 96
Winter, Marc J., 35, 38