Description
The purpose of this study is to utilize prior research in US Congressional politics, the accounting/state relationship,
and corporate political activity to analyze corporations’ political activities during the development and passage of the
United States’ Taxpayer Relief Act of 1997.Our study provides evidence consistent with the notion that large corporations
exercise considerable political power during the state’s formulation of new tax accounting laws.These
findings lead us to question the applicability of a strict pluralist model in accounting policy research and have
implications for future research in corporate political activity, corporate tax accounting, and the political economy of
accounting.
The politics of tax accounting in the United States:
evidence from the Taxpayer Relief Act of 1997
Robin W. Roberts*, Donna D. Bobek
School of Accounting, University of Central Florida, PO Box 161400, Orlando, FL 32816, USA
Abstract
The purpose of this study is to utilize prior research in US Congressional politics, the accounting/state relationship,
and corporate political activity to analyze corporations’ political activities during the development and passage of the
United States’ Taxpayer Relief Act of 1997. Our study provides evidence consistent with the notion that large cor-
porations exercise considerable political power during the state’s formulation of new tax accounting laws. These
?ndings lead us to question the applicability of a strict pluralist model in accounting policy research and have
implications for future research in corporate political activity, corporate tax accounting, and the political economy of
accounting.
#2003 Elsevier Ltd. All rights reserved.
A prominent debate continues over the extent to
which business interests dominate politics and
policymaking (Grier, Munger, & Roberts, 1994;
Suarez, 1998). This debate occurs on multiple the-
oretical levels (Quinn & Shapiro, 1991). In the
accounting literature, some researchers argue
that the heart of the debate rests on whether or
not neoclassical state theory is the appropriate
basis for accounting policy research (Cooper &
Sherer 1984; Tinker, 1984; Tinker, Merino, & Nei-
mark 1982). Of particular concern is the neoclassical
assumption of political voluntarism, its linkage to
the individualistic emphasis of interest group the-
ory, and its reliance on the notion of equal access
to the political marketplace. Tinker (1984, p. 62)
argues that interest group theory’s pluralistic view
of social relations is inadequate because, ‘‘while a
recognition of the self-seeking of individuals may
be necessary to understanding social behavior, it is
not su?cient. It is also important to understand
the type of social system in which individual self-
interest is situated.’’ Research signi?cance is
attached to understanding the social system, at
least in part, because the structure of the social
system in?uences the manner in which interest
groups develop and compete against each other
for favored legislation.
A signi?cant result of the adoption of strict
neoclassical assumptions in accounting policy
research is a failure to recognize the social and
political implications inherent in the public policy
arena (Cooper & Sherer, 1984; Tinker et al., 1982).
In this paper, we present a case study of US tax
accounting law formation to investigate the extent
to which tax policies are subject to the social
and political powers of corporate interests. By
examining the political activities of individual
0361-3682/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0361-3682(03)00036-9
Accounting, Organizations and Society 29 (2004) 565–590
www.elsevier.com/locate/aos
* Corresponding author. Tel.: +1-407-823-6727; fax: +1-
407-823-3881.
E-mail addresses: [email protected]
(R.W. Roberts), [email protected] (D.D. Bobek).
corporations and corporate coalitions we address
both the collective political in?uence of corporate
interests at a societal level and the strategic poli-
tical activities of corporations at the organiza-
tional level. Even though there have been popular
press accounts of corporate in?uence on political
decisions (e.g., Birnbaum & Murray, 1987), there
is a strong need for academic research that explores
the corporate/state relationship in this manner.
Hillman and Hitt (1999) discuss the paucity of
academic research on corporate political activity
and strongly encourage new empirical work, stres-
sing, ‘‘the speci?c behaviors that ?rms choose in
order to participate in the public policy process
have received relatively little attention’’ (Hillman
& Hitt, 1999, p. 827).
Speci?cally, in this study we undertake a close
examination of corporate political activities
during the development of the US Taxpayer
Relief Act of 1997 (the 1997 Act). We study
three cases of corporate political activity con-
cerning the 1997 Act, each examined within a
speci?c level of political competition or con?ict
(Epstein, 1969; Mitnick, 1993) during the 6-
month public policy formulation stage (i.e.
development) of the legislation (Mack, 1997).
We analyze the insurance industry’s corporate
political activity concerning legislation that
broadly a?ected their industry in order to demon-
strate inter-social interest (e.g., business versus
consumers) political con?ict. Our second case
involves the airline industry and examines the
political and economic debate surrounding revi-
sions to an airline ticket tax. The industry’s frag-
mented e?orts, which pitted the major airlines
against smaller, low-fare carriers, provide an
example of intra-industry political con?ict. In the
last case we illustrate individual corporations
competing with others for favored legislation (i.e.,
inter-?rm political con?ict). Our case of inter-?rm
political con?ict involves an examination of the
political activities of two corporations, Amway
Corporation and Sammons Enterprises, and the
passage of ri?e-shot rules (i.e. tax law changes
targeted to bene?t a very speci?c set of taxpayers)
that provided direct bene?t to their shareholders.
Although this classi?cation of levels of political
con?ict is imperfect, it provides an e?ective way to
organize our study.
1
We gathered data for each
case by systematically examining relevant tax pro-
visions, published articles, hearing transcripts,
corporate lobbying reports, corporate annual
reports, and campaign contribution data.
Our study is designed to accomplish two impor-
tant goals. First, given the debate over the extent
to which assumptions of pluralist theory can be
used to completely describe how US accounting
policy is formed, our study gathers empirical evi-
dence as a basis from which to challenge the strict
pluralist assumptions that support most account-
ing policy research. Our purpose is not to disprove
all variants of pluralist theory, but rather to shed
light on the presence of potential structural
inequities in the current system of US policy-
making and to highlight its consequences. The
intensity and extent of corporate political activity
that we document during the act’s 6-month policy
formulation stage raises signi?cant concerns
regarding pluralist assumptions. Speci?cally, plur-
alist theory does not address the observed systemic
inequities in participation and in?uence in the
development of accounting and tax policy.
Second, our study was undertaken to advance
the use of corporate political activity frameworks
in accounting research. By studying corporate
political activities in a tax accounting policy con-
text, we highlight a void in empirical tax research.
For example, a major segment of empirical tax
accounting research has been built upon the posi-
tive approach embedded in the Scholes–Wolfson
paradigm (Shackelford & Shelvin, 2001). This
paradigm utilizes a microeconomic-based, multi-
lateral contracting perspective to examine corpo-
rate tax planning activities as a part of overall
business strategy (Scholes & Wolfson, 1992).
However, this paradigm fails to recognize the
state as a negotiable contracting party and the
structure of tax accounting laws as endogenous
1
We use the term political con?ict to describe intense inter-
est group competition that may contain underlying structural
inequities (e.g., vastly di?erent wealth endowments). This is not
equivalent to class con?ict in a Marxist sense. However, later in
the paper we discuss Berg’s (1994) argument that many plural-
ist interpretations of US policy collapse into class structure
explanations because business interests hold such a dominating
power position during most policy debates.
566 R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590
to a corporation’s tax planning activities. Our
study shows that corporations not only structure
business transactions, they engage in political
activities in order to in?uence the structure of tax
accounting laws under which their corporations
must operate.
Given the study’s goals, we selected the context
of tax legislation for two primary reasons. First,
tax accounting is situated at a crucial intersection
of the accounting/state relationship. Cooper and
Sherer (1984, p. 218) purport, ‘‘The role of the state
is indeed central to an understanding of accounting
policy, for the latter is strongly interrelated with at
least one obvious element of state activity, namely
taxation.’’ Second, we believe that a study of the
development of tax accounting law provides a
compelling setting from which to examine the
interface of accounting with the claims of pluralist
theory, the distributive dimensions of wealth and
power, and their resulting implications for social
welfare. The social and political practice of
accounting (Burchell et al., 1980) and the proble-
matic nature inherent in the development of
accounting policy is displayed prominently in a
tax setting. Again, we quote Cooper and Sherer
(p. 208):
Not only is accounting policy political in that
it derives from the political struggle in society
as a whole but also the outcomes of account-
ing policy are essentially political in that they
operate for the bene?ts of some groups in
society and to the detriment of others.
We adopt multiple research methods in the hope
of providing convincing evidence to a broad audi-
ence of accounting researchers. We utilize descrip-
tive and inferential empirical analyses to aid in our
interpretation of the case materials. This blending
of methods follows recent works that do not use
strict ontological-method pairings, but instead
advocate multi-paradigm approaches as heuristics
bene?cial in developing an understanding of com-
plex social topics (Bourdieu & Wacquant, 1992;
Lewis &Grimes, 1999; Warsame, Neu, & Simmons
2002).
The remainder of our paper is structured as fol-
lows. First, we explain theories of US congres-
sional politics and the accounting/state
relationship and corporate political activity and
relate them to our study. Second, we explain our
method of analysis. Third, we present our three
case analyses of corporate political activity regarding
selected provisions of the 1997 Act. Finally, we
present our conclusions and o?er suggestions for
future research.
Theoretical development
In this section, we review and synthesize rele-
vant literature on U.S. congressional politics, the
accounting/state relationship, and corporate poli-
tical activity. We use this literature to guide our
examination of corporate political activity and it
provides a theoretical perspective from which to
partition and evaluate corporate activities during
the US public policy-making process.
US Congressional politics and the accounting/state
relationship
Published theoretical work concerning US con-
gressional politics re?ects the broader debate con-
cerning the construction of a civil society. Thus,
research on interactions between corporations and
the state in the establishment of tax accounting
laws is rooted (in varying degrees of explicit and
implicit underpinnings) in variants of three grand
theories of American politics—pluralist, institu-
tionalist, and Marxist (Berg, 1994). Our paper is
concerned with contrasting the manner in which
pluralist and Marxist theories may be utilized to
study tax accounting policy.
2
Pluralists hold that congressional outcomes are
the result of interest group pressures. These interest
2
Institutionalists argue that through years of customs and
practices members of Congress have attained a high degree of
autonomy (Berg, 1994) that insulates them from the pressures of
voters, other branches of government, interest groups, and
political parties (Berg, 1994). Institutionalists tend to focus on
the socialization processes used in congressional and bureau-
cratic settings to perpetuate the institution’s culture and/or gen-
erate self-protection (Berg, 1994). Because we focus on the
political activities of interest groups, we do not apply institu-
tional theory.
R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590 567
groups represent their constituencies before Con-
gress in order to in?uence legislation (Berg, 1994).
From a pluralist perspective, legislation and regula-
tion exist within a ideal market characterized by
self-interest motivated actors. Legislators seek
re-election, regulators desire larger budgets, and
interest groups demand favorable legislation/reg-
ulation. Legislation is determined as the outcome
of a ‘‘political marketplace’’ in which members of
Congress supply laws and regulations to con-
stituencies and interest groups in exchange for
political re-election support. A relatively large
body of accounting policy research adopts this
pluralist perspective through its embrace of neo-
classical economics (e.g., Benston, 1982; Freed &
Swenson, 1995; Watts & Zimmerman, 1978).
Pluralist language used to characterize the social
and political interactions of members of society
re?ects its assumption of political voluntarism and
its corresponding view that state policy is derived
from an aggregation of individual utilities (Tinker,
1984).
In contrast, Marxists characterize society by its
mode of production (e.g., capitalist, socialist). The
mode of production leads to a class structure that
determines who bene?ts most fromthe processes of
production. The class that holds economic power
also dictates the state’s dominant political ideology
and uses political power to further its own interests
(Berg, 1994). Class structure theories generally
reject the pluralistic assumption of political
voluntarism.
A critical stream of accounting research argues
that neoclassical economics-based accounting pol-
icy studies ignore the role of societal-level interests
in the determination of laws and regulations gov-
erning accounting (Armstrong, 1991; Burchell et
al., 1980; Cooper & Sherer, 1984; Tinker, 1984;
Tinker et al., 1982). This critical research ques-
tions pluralist theory, political voluntarism, and
the concomitant positive approaches to account-
ing policy research. As opposed to assuming that
interest groups are ‘‘parachuted in’’ to static
research models such as is normally the case in
accounting policy studies, critical researchers
contend that policy research should include an
analysis of relevant historical and institutional
factors. An understanding of these factors will
reveal the interest group con?ict surrounding
potential policy outcomes and lead to a more
informed analysis of the development and the con-
sequences of policy decisions (Cooper & Sherer,
1984; Tinker, 1984). Berg (1994) argues persuasively
that many pluralist explanations of US policy
have class structure underpinnings. In essence,
business interest groups possess such an imbalance
of power that the result is corporate hegemony.
The demarcation between pluralist and Marxist
lines of research is rather clear in accounting
policy research, principally due to accounting
researchers’ adherence to strict interpretations of
pluralist theory. Political scientists, on the other
hand, sometime develop theoretical perspectives
that draw from pluralist theory yet acknowledge
the overwhelming power advantage of corporate
interests. These studies tend to investigate relative
di?erentials in interest group strength, o?er expla-
nations why corporate interests do not always win
political contests, or explore how imbalances in
interest group power corrupt democratic processes
(Clawson, Neustadtl, and Weller 1998; Quinn and
Shapiro 1991; Useem 1984). The theoretical fram-
ing of these studies opens up a conceptual space
from which researchers can incorporate notions of
power and the abuse of power into their empirical
analyses yet retain the basic democratic notions of
governance.
Conceptual research on corporate political activity
Contemporary research streams concerning US
corporate political activity acknowledge the com-
plexities in US interest group power relationships
and recognize that corporate political con?ict
exists on several levels, occurs in a multitude of
legislative and regulatory arenas, and surfaces
with varying degrees of intensity. According to
corporate political activity theory, a corporation
may engage in political con?ict at the inter-?rm
and/or inter-social interest level (Mitnick, 1993).
3
Suarez (2000) views corporate political activity as
a supplement to the structural power of business.
3
Mitnick (1993) presents a detailed description of levels of
corporate political competition. Inter-?rm and inter-social
interests are most relevant to our study.
568 R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590
From her perspective, corporations tend to
become engaged directly in state politics when
they realize that their particular interests in the
business/state relationship are threatened by
speci?c policy deliberations. At this point corpora-
tions may choose to compete with other ?rms for
desired policy outcomes and/or join other ?rms in
collective action.
Inter-?rm con?ict may occur among ?rms in the
same industry (intra-industry) or among ?rms in
di?erent industries. For example, ?rms in the same
industry may use corporate political activities as a
competitive strategy when attempting to secure
governmental contracts (Mitnick, 1993). Firms in
di?erent industries may compete for government
subsidies (Suarez, 1998). Inter-?rm con?ict may
also occur directly between individual ?rms and
all other taxpayers, as is the case when seeking
ri?e-shot legislation (Hulse, 1996). Inter-social
interest con?ict exists between corporations and
other organized interests such as labor, consumer,
or environmental advocacy groups (Mitnick,
1993). This con?ict may be regarding speci?c leg-
islation (e.g., raising the minimum wage) or a
more general con?ict over the distribution of gov-
ernment support (or taxation).
Numerous arenas for political con?ict exist because
of the complexity of the US federal public policy-
making process. For this reason, public policy studies
often examine issues through a life cycle model com-
posed of three stages: public opinion formation, pub-
lic policy formulation, and public policy
implementation (Buchholz, 1988; Mack, 1997).
4
The
public opinion formation stage occurs when there
becomes a substantial di?erence between public
expectations and institutional performance such
that a de?nable issue is formed. A public policy
issue enters the public policy formulation stage
when the issue becomes politicized, is widely dis-
cussed, becomes a concern for a?ected interest
groups, and is formally introduced as speci?c leg-
islation. When legislation becomes public law the
public policy issue becomes bureaucratized and
enters the public policy implementation stage
(Buchholz, 1988).
Corporate political activity during the public policy
formulation stage
Our study focuses on corporate political activity
during the policy formulation stage of the 1997
Act. A federal tax policy issue typically enters the
public policy formulation stage through commit-
tee discussions within the US House of Repre-
sentatives’ Ways and Means Committee or the
Senate Finance Committee (Congressional Uni-
verse, 2000) and ends when the ?nal legislation is
passed (or defeated). The public policy formula-
tion stage of the 1997 Act lasted approximately 6
months (Congressional Universe, 2000).
Just as individual corporations develop strate-
gies to enhance their economic power, they also
select and execute strategies designed to increase
their political power, especially during the policy
formulation stage (Suarez, 1998). Once a corpora-
tion commits to political activity during this stage,
its speci?c actions depend mainly upon the sub-
stantive nature of the policy issue (Snyder, 1992)
and the level of political con?ict (Grier et al., 1994;
Suarez, 1998).
During the public policy formulation stage,
corporations advance their goals by undertaking
activities such as testifying at congressional hear-
ings, lobbying public policymakers, and making
political action committee (PAC) and individual
campaign contributions (Buchholz, 1988; Mack,
1997). Corporations testify at congressional hear-
ings and lobby public policymakers to communicate
their views to legislators regarding proposed legis-
lation. These activities provide a corporation with
the opportunity to provide legislators with their
reasons for adopting a particular position on pro-
posed legislation and to try and in?uence policy
outcomes (Hillman and Hitt, 1999; Mack, 1997).
Because legislators have limited amounts of time to
devote to policymaking, corporate political activity
research views campaign contributions as the pri-
mary activity that can be used to gain the access to
legislators that is required to in?uence policy.
In inter-social interest and some types of inter-
?rm con?ict, corporations may augment their
4
While the ?uidity of actual policy-making raises questions
concerning the discreteness of each stage, the life cycle model is
viewed as a useful tool in the analysis of corporate political
activity (Getz, 1993, p. 247).
R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590 569
political activities through coordination with trade
associations and/or advocacy coalitions (Hojnacki,
1998; Mack, 1997; Mizruchi, 1992; Suarez, 1998).
Trade associations typically represent a broad
spectrum of industry membership. These types of
associations (e.g., the American Institute of Certi-
?ed Public Accountants) are most likely to be
politically active during policy formulation when
its members have similar concerns over proposed
legislation (Mack, 1997; Hillman & Hitt, 1999).
Advocacy coalitions are most likely to form when
multiple industries and/or ?rms reach agreement
on the need for speci?c policies or legislation to be
passed (Mack, 1997; Suarez, 1998). For example,
during the debate over private securities litigation
reform both an inter-social interest and an intra-
industry coalition were formed. Public accounting
?rms, investment banking ?rms, and many cor-
porations generally agreed on the need for secu-
rities liability reform. Over 1400 organizations
joined an inter social-interest coalition named the
Coalition to End Abusive Securities Suits (Fritsch,
1995). Similarly, public accounting ?rms agreed on
the need for reform and an intra-industry coalition,
The Accountants Coalition, was also organized
(Journal of Accountancy, 1996).
Researchers have been particularly interested in
the potential link between campaign ?nancing and
legislative decision-making. The relationship is
di?cult to untangle because of the complexities
inherent in the campaign ?nance system and in the
legislative process (Grenzke, 1989a, 1989b). Prior
research has studied PAC contributions made
from a variety of industries including defense
(Fleisher, 1993), ?nancial services (Kroszner &
Stratmann, 1998; Romano, 1997), and trucking
(Frendreis & Waterman, 1985). These studies
model both legislators and interest groups using
pluralist assumptions. In this context, corpora-
tions and coalitions use PAC contributions strate-
gically in an attempt to in?uence the legislators
most important to them (Grier, Munger, & Tor-
rent, 1990). Prior empirical work that is grounded
in neoclassical economic theory has generally
found that corporate PAC contributions are more
likely to be given to legislators of the majority
party, higher-ranking legislators, and members of
committees with jurisdiction over the policy issues
of concern (Grier et al., 1990; Romano, 1997;
Kroszner & Stratmann, 1998).
Taxation and corporate political activity
Prior research shows that tax policy a?ects the
business decisions of corporations (Vines & Moore,
1996) and that corporations engage in inter-social
interest (Jacobs 1987, 1988; Quinn & Shapiro,
1991) intra-industry (Suarez, 1998), and inter-?rm
(Hulse, 1996; Suarez, 1998) political con?ict in
hopes of reducing their ?rms’ overall tax burden
(Mizruchi, 1992).
Quinn and Shapiro (1991) examined several
hypothesized relationships among taxation, redis-
tribution, politics, and business/class power. Using
four alternative empirical models, they found strong
support for the presence of inter-social interest
political con?ict between business and labor. Of
particular note, they found a signi?cant relation-
ship between increases in the general level of cam-
paign ?nancing from the corporate sector and
lower overall corporate tax burdens.
Freed and Swenson (1995) investigated cam-
paign contributions made by US corporations to
tax-writing members of the US Congress prior to
the passage of the Economic Tax Recovery Act
of 1981 and the Tax Reform Act of 1986. Their
study, developed within a marginalist economic
framework, hypothesized that a corporation’s
campaign contributions to tax writing legislators
was determined by the ?rm’s cost/bene?t analysis.
Using ?rm speci?c data for approximately 400 tax
paying ?rms for each tax act, Freed and Swenson
regressed two related measures of ?rm campaign
contributions against measures of ?rm size,
industry concentration, and expected bene?ts of
the tax legislation. Their results were supportive of
the model, with measures of industry speci?c tax
accounting rules related to the Tax Reform Act of
1986 having the largest impact on contributions.
Although not using the framework of corporate
political activity theory, Hulse (1996) investigated
the stock market reaction to ?rms’ successful tax
code restructuring e?orts. Hulse utilized an events
study methodology to estimate the stock market
reaction individual ?rms experienced when ri?e-
shot rules bene?ting their ?rms were included in
570 R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590
the Tax Reform Act of 1986. While Hulse (1996)
emphasized the secrecy surrounding these types of
provisions and the e?orts of lobbyists to obtain
tax code restructuring for their clients, he did not
investigate the political strategies used by cor-
porations in their restructuring e?orts.
Northcutt and Vines (1998) studied changes in
?nancial accounting methods used by corpora-
tions that were characterized by advocacy groups
as ‘‘corporate freeloaders’’. Their sample con-
sisted of corporations that were singled out and
criticized by the Citizens for Tax Justice because
of the corporations’ extremely low e?ective tax
rates. Northcutt and Vines (1998) found that cor-
porations responded to this political scrutiny in
later reporting periods by choosing income-
decreasing accruals with low book-tax conformity,
thereby increasing their reported e?ective tax rate.
By examining the interplay between advocacy
groups and corporate interests, this study provides
an example of inter-social interest level research.
Suarez (1998) reported the ?ndings of her case
study of the US possessions’ tax credit (e.g., credits
for maintaining manufacturing facilities in Puerto
Rico). She analyzed relevant public information
regarding political action committee contribu-
tions, congressional hearings, and company reports.
Suarez argued that when a tax policy issue is
mutually bene?cial to a group of ?rms or indus-
tries, coalitions and trade associations coordinate
political activities in order to maximize their
e?ectiveness. However, when a tax policy issue
a?ects these groups di?erently, coalition and trade
association alliances recede and fragmented lob-
bying emerges. Suarez concluded from her case
work that corporations used collective and frag-
mented political alliances in attempts to protect
the possessions’ tax. In particular, she asserted
that the introduction of proposed changes that had
di?erential impacts on the electronic and pharma-
ceutical industries collapsed coalition e?orts and
resulted in an increase in inter-?rmpolitical con?ict.
Implications for the case study
The two sections that followdescribe our research
method and the results of our case analysis con-
cerning corporate political activities during the
6-month policy formulation stage of the Taxpayer
Relief Act of 1997. Our study of prior research on
the accounting/state relationship, corporate poli-
tical activity, and tax accounting helped us iden-
tify several contributions that our research can
make to our collective understanding of the devel-
opment of US tax accounting laws. First, we are
interested in identifying and investigating cor-
porations that undertook political activities
regarding the 1997 Act. We argue that these cor-
porations view the state as a contracting party
over which they can exert in?uence. Thus, con-
trary to the microeconomic-based, strict pluralist
assumptions common in empirical tax accounting
research, tax accounting rules are not necessarily
exogenous to the ?rm. In other words, corporate
interests play a substantive role in the develop-
ment of tax accounting policy.
Second, in addition to addressing pluralist
assumptions, we also are able to extend research
on corporate political activity. As recommended
by Hillman and Hitt (1999), we investigate and
document the speci?c behaviors that ?rms use to
in?uence legislation. Further, we document and
analyze corporate behaviors at three di?erent
levels of political competition.
Third, evidence that corporations exert in?uence
in the development of tax policy is not necessarily
inconsistent with pluralist theory. In order to
assess the applicability of strict pluralist assump-
tions, we must examine the intensity and extent of
corporate political activities. Intense and extensive
political activities made by corporations during
the six-month period bolster arguments that all
interest groups do not have similar opportunities
to in?uence accounting policy. The more that
lobbying and campaign contributions allow cor-
porate interests to dominate the public policy for-
mulation process, the less applicable strict
pluralist assumptions are in the study of policy
development (Clawson et al., 1998). Thus, this case
study provides a basis from which to re?ect on the
appropriateness of any pluralist-based theory to
accounting policy research. The nature and levels of
political con?ict that we analyze can shed light on
the importance of incorporating political economy
more prominently in the study of tax accounting
policy.
R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590 571
Research method
Our study followed a four-step empirical case-
based research approach (Ryan et al., 1992, pp. 113–
129). First, we analyzed the provisions of the 1997
Act, and publications, press releases and other
documentation associated with the 1997 Act.
These data were gathered through a repetitive
search of electronic databases (ABI-Inform, Busi-
ness Index, Congressional Universe, Firstsearch,
Wall Street Journal Index, and multiple search
engines for the World Wide Web). Our search
produced a wide variety of public documents
related to the 1997 Act, including published trade
articles, transcripts of congressional hearings,
legislative histories of the proposed law, popular
press articles, and newspaper columns.
From this search, we identi?ed three sets of
taxpayers for further analysis. These taxpayers
were selected because each corresponded most
closely with a speci?c level of political con?ict
(Buchholz, 1988; Mack, 1997). The insurance
industry was concerned with numerous provisions
in the 1997 Act. Their uni?ed e?orts provided an
example of inter-social interest political con?ict
because the industry attempted to in?uence a
number of tax law changes within a constrained
resource setting. Thus, to the extent they were
successful, their e?orts a?ected the overall dis-
tribution of the tax burden. The airline industry
displayed intra-industry political con?ict over
the airline tax provisions. Major airline carriers
and regional carriers strongly disagreed over the
structure of the tax. Amway Corporation and
Sammons Enterprises were selected as glaring
examples of inter-?rm political con?ict because
these corporations sought speci?c, narrowly tar-
geted, tax relief.
In the second step of our research process we
gathered lobbying reports and political action
committee contribution data for the taxpayers
under study for the period relevant to the policy
formulation stage of the 1997 Act. The data were
collected to provide additional primary source
information regarding the political activities of
these taxpayers and to corroborate the informa-
tion we gathered from secondary sources. The
lobbying and political contribution activities rele-
vant to the public policy formulation stage of the
1997 Act occurred between the times the 1997 Act
was originally introduced (as the Revenue Recon-
ciliation Act) during January 1997 until its passage
on August 5, 1997 (Congressional Universe, 2000).
Lobbying reports are ?led with the Clerk of the
House of Representatives and the Secretary of the
Senate on a semi-annual basis. Therefore, we gath-
ered reports for Amway Corporation, Sammons
Enterprises and for all insurance and airline rela-
ted ?rms and associations for the January–June
1997 reporting period. These mandatory ?lings are
required of all organizations engaged in lobbying
e?orts. The ?lings specify the legislation being
lobbied, the cost of lobbying activities, whether or
not the ?ling was made by a lobbying ?rm, and
the organization ?nancing the lobbying activity.
Next we gathered PAC contribution data for
each ?rm that reported lobbying activity asso-
ciated with the 1997 Act. Lobbying and making
?nancial contributions are ?rms’ and associations’
most visible and e?ective activities during the pol-
icy formulation stage of proposed legislation
(Hillman & Hitt, 1999; Mack, 1997). Political
action committee contributions are reported to the
Federal Election Commission and compiled by the
Center for Responsive Politics (2000). These con-
tributions are reported by election cycle. We
gathered contribution data for the 105th Con-
gressional election cycle because the 1997 Act was
passed by this congress.
The third step in our research was to use the
framework provided by our theoretical work to
analyze and interpret the information that was
gathered in steps one and two (Grier et al., 1994;
Mitnick, 1993; Suarez, 1998). This analysis pro-
duced evidence of the taxpayers’ interests in the
1997 Act and their motivation for undertaking
lobbying and political campaign contribution
activities. Political campaign contribution data
reported ?rms’ contribution amounts that were
donated to each senator and house member but,
of course, did not provide direct information on the
reason for the contribution. Therefore, we appealed
to prior research on PAC contributions to guide
our empirical work (Romer & Snyder, 1994;
Snyder, 1992; Wright, 1985). In the two cases in
which multiple ?rms were involved (insurance and
572 R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590
airlines), we developed and tested an empirical
model to explain the campaign contributions
made to Senate and House members by these
?rms’ political action committees.
As we mentioned in the previous section, prior
research models ?rms’ PAC contribution decisions
through a neoclassical economics perspective and
hypothesizes that ?rms allocate contributions to
legislators who are in the best positions to lead
e?orts to have Congress adopt the corporation’s
favored legislation (Grier et al., 1994). According to
prior studies, these are most likely to be legislators
who hold leadership positions and/or who are
members of committees with jurisdictional author-
ity over relevant legislation. Thus, the political con-
tributions that were particularly relevant to the
1997 Act were contributions given to congres-
sional leaders and/or members of the House Ways
and Means Committee or the Senate Finance
Committee.
Prior PAC research also states that legislators
who hold in?uential positions may not be equally
likely to attempt to provide the desired legislation
(Grier et al., 1994; Snyder, 1992). This di?erential
in legislators’ willingness to support a corporation’s
speci?c policy preferences is because legislators are
ultimately concerned with re-election. Therefore,
their ability to support legislation that is important
to PAC contributors is constrained by their own
political ideology and their voting constituents’
legislative preferences (Grier et al., 1994; Snyder,
1992).
Our PAC contribution regressions controlled
for political ideology, constituent preferences, and
industry-related committee membership in order
to improve the reliability of our tests of the rela-
tionship between ?rm PAC contributions and leg-
islator in?uence over tax policy. We performed four
regressions, two regressions for each of the cases
with multiple ?rms (insurance and airlines) for each
legislative body (House and Senate). The general
speci?cations of the empirical models we tested
were:
PAC
house
¼ b
0
þ b
1
VOTE þ b
2
LEADTEAM
þ b
3
INDCOMMþ b
4
TAXCOMM
PAC
senate
¼ b
0
þ b
1
VOTE þ b
2
LEADTEAM
þ b
3
INDCOMMþ b
4
TAXCOMM
þ b
5
ELECT
þ b
6
ELECT
Ã
INDCOMM
þ b
7
ELECT
Ã
TAXCOMM
Where:
PAC= a legislator’s total receipt of PAC
contributions from either
insurance or airline industry
PACs that lobbied the 1997
Act (100 observations for the
Senate and 437 for the House
of Representatives).
VOTE= the legislators’ cumulative voting
rating by the U.S. Chamber
of Commerce (range=0-100).
The higher the rating, the more
favorable their view of the
legislator’s voting record.
LEADTEAM=1 if the legislator holds a
leadership position; 0 otherwise.
INDCOMM= 1 if the legislator is a member of
a committee that has jurisdiction
over legislation related
speci?cally to a ?rm’s industry
(not tax-related); 0 otherwise.
TAX COMM=1 if the legislator is a member of
the legislative committee that has
jurisdiction over tax legislation.
ELECT= 1 if the legislator is a Senator and
is up for reelection during the
next election cycle; 0 otherwise.
This variable is used only in the
Senate regressions.
TAX COMM is our primary test variable
because its signi?cance shows that a relationship
R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590 573
exists between corporate PAC contributions and a
legislator’s ability to a?ect tax accounting changes
that are preferred by those corporations. All of the
other variables controlled for non-tax-related fac-
tors that may explain variance in contributions
made by insurance or airlines ?rms to legislators
during the 105th congressional election cycle.
VOTE was included in the model to control for
political ideology and constituent voting pre-
ferences. As expected, VOTE is highly correlated
with legislators’ political party membership (i.e.,
Republican legislators possess higher pro-business
voting ratings than Democrats). LEADTEAM
and INDCOMM were included to control for
other legislator attributes that could explain PAC
giving. For LEADTEAM to be set equal to one, a
legislator must have served the 105th Congress as
Majority Leader, Minority Leader, Majority
Whip, Minority Whip or Chair of the applicable
tax or industry committee.
For the House of Representatives, INDCOMM
represented membership on the Commerce Com-
mittee in the insurance regression and it repre-
sented membership on the Transportation
Committee in the airline regression. For the Sen-
ate, INDCOMM represented membership on the
Banking, Housing, and Urban A?airs Committee
in the insurance regression and it represented
membership on the Commerce, Science, and
Transportation Committee in the airline regression.
These committees have jurisdiction over legisla-
tion that is most vital to each respective industry
(Congressional Universe, 2000).
5
The ELECT variable is added to the Senate
regressions because US senators serve 6-year,
staggered terms, resulting in only one-third of the
Senate incumbents being up for reelection each
two-year election cycle. This variable is important
because PAC contributions to individual senators
tend to increase during their election year and,
therefore, should be controlled for in the regressions.
All US House of Representatives incumbents are
up for reelection every 2 years, thus the ELECT
variable is not needed. Because of the potential
interaction between ELECT and TAXCOMM and
between ELECT and INDCOMM, we include
interaction terms in the Senate regressions.
We argue that a statistically signi?cant relation-
ship between PAC and TAXCOMM, after con-
trolling for these other factors, provides evidence
that corporate interests view the state as a negoti-
able contracting party when developing tax stra-
tegies. In this particular case, statistical signi?cance
of TAXCOMM will support our contention that
insurance and airline ?rms used their resource
endowments to their advantage to make PAC
contributions that enabled them to have access to
and in?uence over the determination of tax
accounting law during the policy formulation
stage of the 1997 Act. Statistical signi?cance does
not disprove pluralist theory, but it does provide
support for increasing the attention given to the
political economy aspects of accounting policy.
We report parameter estimates to indicate the
dollar e?ect of each independent variable on the
dependent variable.
6
5
Given our de?nitions for LEADTEAM, TAXCOMM and
INDCOMM, the chairs of TAXCOMM and INDCOMM were
included in the LEADTEAM variable. We performed sensitiv-
ity analyses of our results leaving these chairs out of LEAD-
TEAM. The results presented in the paper were not changed
when omitting committee chairs from the LEADTEAM
variables.
6
We recognize that the corporate taxpayers we are studying
could be lobbying Congress for reasons other than the 1997
Act. We thoroughly explored Congressional Universe for com-
mittee reports, testimony and bills of interest to the airline and
insurance industry before the House Ways and Means com-
mittee and/or the Senate Finance Committee during the time
period in question. Our search did not turn up any other bills
pertinent to the airline industry. However, Medicare legislation
was before the House Ways and Means committee at the same
time as the tax legislation. As a sensitivity check, we split the
House Ways and Means Committee into two variables. One
variable was coded ‘‘1’’ for House Ways and Means committee
members who were on the Health Subcommittee, the other
variable was coded ‘‘1’’ for House Ways and Means members
who were not on the Health Subcommittee. Our reasoning is
that if the PAC contributions were being made only to in?u-
ence Medicare legislation, we would expect to see a signi?cant
result for the members who were on the Health Subcommittee
versus. those House Ways and Means members who were not
on the Health subcommittee. This analysis showed that both
variables were highly signi?cant (P-value=0.000), and the
parameter estimates were similar. We suggest that this is evi-
dence consistent with our argument that these PAC Contribu-
tions were made, at least in part, to in?uence the tax bill.
574 R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590
Fourth, after completion of all of our analyses,
we organized our ?ndings in order to present a
structured examination of the corporate political
activities of each set of taxpayers. The analyses are
presented in the following section. First, we discuss
the tax issues and provisions of interest to the set
of taxpayers. Second, we discuss our ?ndings
regarding lobbying and political campaign con-
tribution activities.
Corporate political activity and the Taxpayer
Relief Act of 1997
Inter-social interest corporate political activity: the
insurance industry
Tax issues
The insurance industry engaged in corporate
political activity on a number of complex tax issues
that either directly a?ected the tax burden of the
industry or a?ected demand for their products. Our
search of industry publications and press releases
enabled us to identify the major 1997 Act provisions
of interest to the industry as well as its position on
these provisions. Table 1 summarizes these issues.
The insurance industry favored a change in the
treatment of foreign sourced investment income.
Subpart-F rules require ?nancial services companies
(e.g. insurance, banks and securities ?rms), but
not manufacturing ?rms, to immediately recognize
foreign-earned investment income even if the
funds have not been repatriated to the parent cor-
poration (Brosto?, 1997b). The insurance indus-
try, along with a coalition of the other a?ected
industries argued that these rules impeded their
global competitiveness (Brosto?, 1997f).
7
Insurance industry publications documented their
economic and political interest in the highway tax
provisions included in the 1997 Act. A 4.3-cent-
per-gallon tax had originally been enacted in 1993
as a de?cit reduction measure. However, a coalition
of industries (including insurance and construc-
tion) lobbied to have the revenue re-directed to the
Highway Trust Fund (Winston, 1997). Increased
funding for federal highway programs would
reduce automobile accidents and insured losses
(Brosto?, 1997d).
8
The insurance industry’s coali-
tion strategies concerning the gasoline tax and the
Table 1
Insurance industry 1997 Tax Act provisions
Tax issue Insurance
industry position
1997 Act change
Sub-part F Rules required tax on non-repatriated foreign
source investment income for ?nancial services ?rms
FAVORED
ELIMINATION OF TAX
Included one year reprieve
to the Sub Part F rules
Reallocation of 4.3 cent federal gas tax from de?cit reduction
to the Highway Trust Fund
FAVORED THIS CHANGE Provision was included in Act
Restriction of general borrowing interest expense based on
the existence of unborrowed cash values on Corporate
Owned Life Insurance (COLI)
OPPOSED RULE, HOWEVER
SOUGHT CLARIFYING
LANGUAGE TO LIMIT
REACH OF RULE
Limitation on interest expense
from general borrowing was
restricted, but it only applied to
mortgage-owned life insurance
Proposed changes in NOL carryover rules from
3 back and 15 forward to 2 back and 20 forward
OPPOSED RULE New NOL carryover rules were
enacted. Special exceptions were
allowed for presidentially declared
disaster areas
7
The 1997 Act passed by both houses of Congress included
a one-year reprieve from Subpart-F rules, however this provi-
sion was one of two vetoed by President Clinton under his line-
item veto privileges (Congressional Universe, 2000). The
Supreme Court later ruled the line-item veto law unconstitu-
tional; and the rule change was enacted (Congressional Uni-
verse, 2000).
8
Their e?orts were deemed successful because an amend-
ment to the 1997 Act was added to divert the 4.3-cent-per-gal-
lon fuel tax from de?cit reduction to the Highway Trust Fund
(Brosto?, 1997d).
R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590 575
Subpart-F rules is consistent with expectations
derived from corporate political activity research
because of the inter-social interest nature of these
provisions (Mack, 1997; Mizruchi, 1992; Suarez,
1998).
The insurance industry was also concerned with
the 1997 Act provisions a?ecting corporate owned
life insurance (Kerley, 1997b). While insurance
premiums are not deductible (if the payee is the
bene?ciary) and death bene?ts are not taxable,
any interest expense related to general borrowing
of funds to purchase these policies was deductible
prior to the 1997 Act. Congress’ interest in closing
this loophole caused two major concerns. First,
some insurance companies (along with Fannie
Mae) were planning to exploit the interest deduct-
ibility and opposed changes to the tax rule
(Kerley, 1997b). Second, other companies feared
that the rule changes would be drafted too
broadly.
9
Finally, the insurance industry was also opposed
to a change to the net operating loss (NOL) car-
ryover rules. Prior to the 1997 Act, a NOL could
be carried back 3 years and forward 15 years. The
1997 Act changed the carry back period to only 2
years, while the carry forward period was exten-
ded to 20 years. The insurance industry, both the
Alliance of American Insurers and the American
Insurance Association, argued that this change
would be particularly harmful to property and
casualty insurers that su?er catastrophic losses
(Brosto?, 1997e). While the industry was unable
to halt the rule change, an exception that allowed
the old rule to apply was made for losses that
occurred in presidentially declared disaster areas.
10
Lobbying and political action committee activities
Through a manual search of lobbying reports,
we found 112 reports associated with insurance
lobbying activity regarding the 1997 Act. These
reports were ?led either by insurance corporations
or associations or by lobbying ?rms that repre-
sented the insurance industry. Our analysis of the
reports documents that the 50 insurance compa-
nies that lobbied this bill spent approximately
$15.2 million on lobbying expenses in the ?rst 6
months of 1997. In addition, 16 insurance industry
associations reported an additional $6.4 million of
lobbying expenditures during this period.
These lobbying expenditures were supplemented
by campaign contributions to legislators (Center for
Responsive Politics, 2000) and ‘‘soft money’’ con-
tributions made directly to political parties (Com-
mon Cause, 2000). In total, the insurance industry
members who lobbied during the formulation stage
of the 1997 Act made PAC contributions of $5.0
million to the campaigns of US House members
and $1.9 million to the campaigns of US Senators
(Center for Responsive Politics, 2000). In addition,
‘‘soft money’’ contributions of $2.8 million were
made by the insurance industry during this time
period (Common Cause, 2000). For this time per-
iod, the insurance industry’s soft money contribu-
tions were more than that of any other industry
(Common Cause, 2000). Table 2 summarizes the
corporate political activities of the insurance indus-
try. In total, we documented over $30 million of
corporate political activity expenditures by this
industry.
Although soft money contributions cannot be
directly linked to individual legislators, PAC con-
tributions are made to individual congressional
campaigns. We examined the relationship between
insurance PAC contributions and legislator mem-
bership on either the House Ways and Means
Committee or the Senate Finance Committee by
developing and testing an empirical model that
controlled for other factors that could in?uence
10
The insurance industry also expressed concerns over pro-
posed changes that a?ected individual taxpayers. This was
because the provisions, if adopted, could have business reper-
cussions. These issues included privatization of social security
(Brosto?, 1997a), increased deductibility of health insurance
premiums for self-employed taxpayers (Kerley, 1997a), changes
in estate taxes (Kerley 1997a; Brosto?, 1997c; King, 1997), and
decreases in the capital gains tax rate (Kerley, 1997a). While
not explored further in our study, this further demonstrates the
extent to which tax-related issues a?ect corporate business and
tax-related political activities.
9
The industry was able to obtain clarifying language in the
provision that limited this provision to non-employees. Further
the tax rule was written such that it does not apply to coverage
of former employees, o?cers and directors (Brosto?, 1997b).
Thus, interest expense for general borrowing will not be dis-
allowed as a result of unborrowed policy cash values, except for
the case of mortgage-owned life insurance.
576 R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590
insurance PAC giving. The regression results for the
House of Representatives are shown in Table 3
and the results for the Senate are shown in Table 4.
Each table provides descriptive statistics for each
variable in the model and ordinary least squares
regression results.
11
Panel A of each table presents
descriptive statistics as well as the t-test results of
the di?erences in the mean PAC contribution for
each level of the dichotomous independent vari-
ables. The reported means indicate the dollar
impact of each of the independent variables. Panel
B of each table presents the multivariate OLS
regression results for the PAC contribution model.
We also present descriptive data regarding legisla-
tors’ political party a?liation but this variable is
not included in the regressions because it is highly
correlated with the legislators’ voting records. The
results provide strong evidence that the insur-
ance industry allocated its PAC contributions to
Table 2
Insurance industry summary of corporate political activities
1/1/97–6/30/97 1997 Tax Act lobbying expenditures
a
Insurance Companies (n=50) $ 15,189,740
Insurance Industry Associations (n=16) 6,362,646
Total Lobbying Expenditures by Insurance
Industry related to 1997 Tax Act, for ?rst
Six months of 1997 $ 20,512,993
PAC contributions to individual house and senate members by insurance companies and
insurance industry associations who lobbied for 1997 Tax Act: 1998 election cycle
b
House Members $ 5,005,851
Senate Members $ 1,942,667
Total PAC Contributions $ 6,948,518
‘‘Soft money’’ contributions to political parties by insurance industry
c
To the Democratic Party $ 902,161
To the Republican Party $ 1,942,667
Total ‘‘soft money’’ contributions $ 2,841,860
a
Lobbying reports are ?led with the Secretary of the Senate and the Clerk of the House of Representatives as required by the
Lobbying Disclosure Act. These reports include information about the speci?c bills that were being lobbied. We only included
amounts from lobbying reports that included the 1997 Tax Act as one of the lobbying targets. Note: in some cases the reported
amount was ‘‘ 70 seats
$1/seat on aircrafts with
The purpose of this study is to utilize prior research in US Congressional politics, the accounting/state relationship,
and corporate political activity to analyze corporations’ political activities during the development and passage of the
United States’ Taxpayer Relief Act of 1997.Our study provides evidence consistent with the notion that large corporations
exercise considerable political power during the state’s formulation of new tax accounting laws.These
findings lead us to question the applicability of a strict pluralist model in accounting policy research and have
implications for future research in corporate political activity, corporate tax accounting, and the political economy of
accounting.
The politics of tax accounting in the United States:
evidence from the Taxpayer Relief Act of 1997
Robin W. Roberts*, Donna D. Bobek
School of Accounting, University of Central Florida, PO Box 161400, Orlando, FL 32816, USA
Abstract
The purpose of this study is to utilize prior research in US Congressional politics, the accounting/state relationship,
and corporate political activity to analyze corporations’ political activities during the development and passage of the
United States’ Taxpayer Relief Act of 1997. Our study provides evidence consistent with the notion that large cor-
porations exercise considerable political power during the state’s formulation of new tax accounting laws. These
?ndings lead us to question the applicability of a strict pluralist model in accounting policy research and have
implications for future research in corporate political activity, corporate tax accounting, and the political economy of
accounting.
#2003 Elsevier Ltd. All rights reserved.
A prominent debate continues over the extent to
which business interests dominate politics and
policymaking (Grier, Munger, & Roberts, 1994;
Suarez, 1998). This debate occurs on multiple the-
oretical levels (Quinn & Shapiro, 1991). In the
accounting literature, some researchers argue
that the heart of the debate rests on whether or
not neoclassical state theory is the appropriate
basis for accounting policy research (Cooper &
Sherer 1984; Tinker, 1984; Tinker, Merino, & Nei-
mark 1982). Of particular concern is the neoclassical
assumption of political voluntarism, its linkage to
the individualistic emphasis of interest group the-
ory, and its reliance on the notion of equal access
to the political marketplace. Tinker (1984, p. 62)
argues that interest group theory’s pluralistic view
of social relations is inadequate because, ‘‘while a
recognition of the self-seeking of individuals may
be necessary to understanding social behavior, it is
not su?cient. It is also important to understand
the type of social system in which individual self-
interest is situated.’’ Research signi?cance is
attached to understanding the social system, at
least in part, because the structure of the social
system in?uences the manner in which interest
groups develop and compete against each other
for favored legislation.
A signi?cant result of the adoption of strict
neoclassical assumptions in accounting policy
research is a failure to recognize the social and
political implications inherent in the public policy
arena (Cooper & Sherer, 1984; Tinker et al., 1982).
In this paper, we present a case study of US tax
accounting law formation to investigate the extent
to which tax policies are subject to the social
and political powers of corporate interests. By
examining the political activities of individual
0361-3682/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0361-3682(03)00036-9
Accounting, Organizations and Society 29 (2004) 565–590
www.elsevier.com/locate/aos
* Corresponding author. Tel.: +1-407-823-6727; fax: +1-
407-823-3881.
E-mail addresses: [email protected]
(R.W. Roberts), [email protected] (D.D. Bobek).
corporations and corporate coalitions we address
both the collective political in?uence of corporate
interests at a societal level and the strategic poli-
tical activities of corporations at the organiza-
tional level. Even though there have been popular
press accounts of corporate in?uence on political
decisions (e.g., Birnbaum & Murray, 1987), there
is a strong need for academic research that explores
the corporate/state relationship in this manner.
Hillman and Hitt (1999) discuss the paucity of
academic research on corporate political activity
and strongly encourage new empirical work, stres-
sing, ‘‘the speci?c behaviors that ?rms choose in
order to participate in the public policy process
have received relatively little attention’’ (Hillman
& Hitt, 1999, p. 827).
Speci?cally, in this study we undertake a close
examination of corporate political activities
during the development of the US Taxpayer
Relief Act of 1997 (the 1997 Act). We study
three cases of corporate political activity con-
cerning the 1997 Act, each examined within a
speci?c level of political competition or con?ict
(Epstein, 1969; Mitnick, 1993) during the 6-
month public policy formulation stage (i.e.
development) of the legislation (Mack, 1997).
We analyze the insurance industry’s corporate
political activity concerning legislation that
broadly a?ected their industry in order to demon-
strate inter-social interest (e.g., business versus
consumers) political con?ict. Our second case
involves the airline industry and examines the
political and economic debate surrounding revi-
sions to an airline ticket tax. The industry’s frag-
mented e?orts, which pitted the major airlines
against smaller, low-fare carriers, provide an
example of intra-industry political con?ict. In the
last case we illustrate individual corporations
competing with others for favored legislation (i.e.,
inter-?rm political con?ict). Our case of inter-?rm
political con?ict involves an examination of the
political activities of two corporations, Amway
Corporation and Sammons Enterprises, and the
passage of ri?e-shot rules (i.e. tax law changes
targeted to bene?t a very speci?c set of taxpayers)
that provided direct bene?t to their shareholders.
Although this classi?cation of levels of political
con?ict is imperfect, it provides an e?ective way to
organize our study.
1
We gathered data for each
case by systematically examining relevant tax pro-
visions, published articles, hearing transcripts,
corporate lobbying reports, corporate annual
reports, and campaign contribution data.
Our study is designed to accomplish two impor-
tant goals. First, given the debate over the extent
to which assumptions of pluralist theory can be
used to completely describe how US accounting
policy is formed, our study gathers empirical evi-
dence as a basis from which to challenge the strict
pluralist assumptions that support most account-
ing policy research. Our purpose is not to disprove
all variants of pluralist theory, but rather to shed
light on the presence of potential structural
inequities in the current system of US policy-
making and to highlight its consequences. The
intensity and extent of corporate political activity
that we document during the act’s 6-month policy
formulation stage raises signi?cant concerns
regarding pluralist assumptions. Speci?cally, plur-
alist theory does not address the observed systemic
inequities in participation and in?uence in the
development of accounting and tax policy.
Second, our study was undertaken to advance
the use of corporate political activity frameworks
in accounting research. By studying corporate
political activities in a tax accounting policy con-
text, we highlight a void in empirical tax research.
For example, a major segment of empirical tax
accounting research has been built upon the posi-
tive approach embedded in the Scholes–Wolfson
paradigm (Shackelford & Shelvin, 2001). This
paradigm utilizes a microeconomic-based, multi-
lateral contracting perspective to examine corpo-
rate tax planning activities as a part of overall
business strategy (Scholes & Wolfson, 1992).
However, this paradigm fails to recognize the
state as a negotiable contracting party and the
structure of tax accounting laws as endogenous
1
We use the term political con?ict to describe intense inter-
est group competition that may contain underlying structural
inequities (e.g., vastly di?erent wealth endowments). This is not
equivalent to class con?ict in a Marxist sense. However, later in
the paper we discuss Berg’s (1994) argument that many plural-
ist interpretations of US policy collapse into class structure
explanations because business interests hold such a dominating
power position during most policy debates.
566 R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590
to a corporation’s tax planning activities. Our
study shows that corporations not only structure
business transactions, they engage in political
activities in order to in?uence the structure of tax
accounting laws under which their corporations
must operate.
Given the study’s goals, we selected the context
of tax legislation for two primary reasons. First,
tax accounting is situated at a crucial intersection
of the accounting/state relationship. Cooper and
Sherer (1984, p. 218) purport, ‘‘The role of the state
is indeed central to an understanding of accounting
policy, for the latter is strongly interrelated with at
least one obvious element of state activity, namely
taxation.’’ Second, we believe that a study of the
development of tax accounting law provides a
compelling setting from which to examine the
interface of accounting with the claims of pluralist
theory, the distributive dimensions of wealth and
power, and their resulting implications for social
welfare. The social and political practice of
accounting (Burchell et al., 1980) and the proble-
matic nature inherent in the development of
accounting policy is displayed prominently in a
tax setting. Again, we quote Cooper and Sherer
(p. 208):
Not only is accounting policy political in that
it derives from the political struggle in society
as a whole but also the outcomes of account-
ing policy are essentially political in that they
operate for the bene?ts of some groups in
society and to the detriment of others.
We adopt multiple research methods in the hope
of providing convincing evidence to a broad audi-
ence of accounting researchers. We utilize descrip-
tive and inferential empirical analyses to aid in our
interpretation of the case materials. This blending
of methods follows recent works that do not use
strict ontological-method pairings, but instead
advocate multi-paradigm approaches as heuristics
bene?cial in developing an understanding of com-
plex social topics (Bourdieu & Wacquant, 1992;
Lewis &Grimes, 1999; Warsame, Neu, & Simmons
2002).
The remainder of our paper is structured as fol-
lows. First, we explain theories of US congres-
sional politics and the accounting/state
relationship and corporate political activity and
relate them to our study. Second, we explain our
method of analysis. Third, we present our three
case analyses of corporate political activity regarding
selected provisions of the 1997 Act. Finally, we
present our conclusions and o?er suggestions for
future research.
Theoretical development
In this section, we review and synthesize rele-
vant literature on U.S. congressional politics, the
accounting/state relationship, and corporate poli-
tical activity. We use this literature to guide our
examination of corporate political activity and it
provides a theoretical perspective from which to
partition and evaluate corporate activities during
the US public policy-making process.
US Congressional politics and the accounting/state
relationship
Published theoretical work concerning US con-
gressional politics re?ects the broader debate con-
cerning the construction of a civil society. Thus,
research on interactions between corporations and
the state in the establishment of tax accounting
laws is rooted (in varying degrees of explicit and
implicit underpinnings) in variants of three grand
theories of American politics—pluralist, institu-
tionalist, and Marxist (Berg, 1994). Our paper is
concerned with contrasting the manner in which
pluralist and Marxist theories may be utilized to
study tax accounting policy.
2
Pluralists hold that congressional outcomes are
the result of interest group pressures. These interest
2
Institutionalists argue that through years of customs and
practices members of Congress have attained a high degree of
autonomy (Berg, 1994) that insulates them from the pressures of
voters, other branches of government, interest groups, and
political parties (Berg, 1994). Institutionalists tend to focus on
the socialization processes used in congressional and bureau-
cratic settings to perpetuate the institution’s culture and/or gen-
erate self-protection (Berg, 1994). Because we focus on the
political activities of interest groups, we do not apply institu-
tional theory.
R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590 567
groups represent their constituencies before Con-
gress in order to in?uence legislation (Berg, 1994).
From a pluralist perspective, legislation and regula-
tion exist within a ideal market characterized by
self-interest motivated actors. Legislators seek
re-election, regulators desire larger budgets, and
interest groups demand favorable legislation/reg-
ulation. Legislation is determined as the outcome
of a ‘‘political marketplace’’ in which members of
Congress supply laws and regulations to con-
stituencies and interest groups in exchange for
political re-election support. A relatively large
body of accounting policy research adopts this
pluralist perspective through its embrace of neo-
classical economics (e.g., Benston, 1982; Freed &
Swenson, 1995; Watts & Zimmerman, 1978).
Pluralist language used to characterize the social
and political interactions of members of society
re?ects its assumption of political voluntarism and
its corresponding view that state policy is derived
from an aggregation of individual utilities (Tinker,
1984).
In contrast, Marxists characterize society by its
mode of production (e.g., capitalist, socialist). The
mode of production leads to a class structure that
determines who bene?ts most fromthe processes of
production. The class that holds economic power
also dictates the state’s dominant political ideology
and uses political power to further its own interests
(Berg, 1994). Class structure theories generally
reject the pluralistic assumption of political
voluntarism.
A critical stream of accounting research argues
that neoclassical economics-based accounting pol-
icy studies ignore the role of societal-level interests
in the determination of laws and regulations gov-
erning accounting (Armstrong, 1991; Burchell et
al., 1980; Cooper & Sherer, 1984; Tinker, 1984;
Tinker et al., 1982). This critical research ques-
tions pluralist theory, political voluntarism, and
the concomitant positive approaches to account-
ing policy research. As opposed to assuming that
interest groups are ‘‘parachuted in’’ to static
research models such as is normally the case in
accounting policy studies, critical researchers
contend that policy research should include an
analysis of relevant historical and institutional
factors. An understanding of these factors will
reveal the interest group con?ict surrounding
potential policy outcomes and lead to a more
informed analysis of the development and the con-
sequences of policy decisions (Cooper & Sherer,
1984; Tinker, 1984). Berg (1994) argues persuasively
that many pluralist explanations of US policy
have class structure underpinnings. In essence,
business interest groups possess such an imbalance
of power that the result is corporate hegemony.
The demarcation between pluralist and Marxist
lines of research is rather clear in accounting
policy research, principally due to accounting
researchers’ adherence to strict interpretations of
pluralist theory. Political scientists, on the other
hand, sometime develop theoretical perspectives
that draw from pluralist theory yet acknowledge
the overwhelming power advantage of corporate
interests. These studies tend to investigate relative
di?erentials in interest group strength, o?er expla-
nations why corporate interests do not always win
political contests, or explore how imbalances in
interest group power corrupt democratic processes
(Clawson, Neustadtl, and Weller 1998; Quinn and
Shapiro 1991; Useem 1984). The theoretical fram-
ing of these studies opens up a conceptual space
from which researchers can incorporate notions of
power and the abuse of power into their empirical
analyses yet retain the basic democratic notions of
governance.
Conceptual research on corporate political activity
Contemporary research streams concerning US
corporate political activity acknowledge the com-
plexities in US interest group power relationships
and recognize that corporate political con?ict
exists on several levels, occurs in a multitude of
legislative and regulatory arenas, and surfaces
with varying degrees of intensity. According to
corporate political activity theory, a corporation
may engage in political con?ict at the inter-?rm
and/or inter-social interest level (Mitnick, 1993).
3
Suarez (2000) views corporate political activity as
a supplement to the structural power of business.
3
Mitnick (1993) presents a detailed description of levels of
corporate political competition. Inter-?rm and inter-social
interests are most relevant to our study.
568 R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590
From her perspective, corporations tend to
become engaged directly in state politics when
they realize that their particular interests in the
business/state relationship are threatened by
speci?c policy deliberations. At this point corpora-
tions may choose to compete with other ?rms for
desired policy outcomes and/or join other ?rms in
collective action.
Inter-?rm con?ict may occur among ?rms in the
same industry (intra-industry) or among ?rms in
di?erent industries. For example, ?rms in the same
industry may use corporate political activities as a
competitive strategy when attempting to secure
governmental contracts (Mitnick, 1993). Firms in
di?erent industries may compete for government
subsidies (Suarez, 1998). Inter-?rm con?ict may
also occur directly between individual ?rms and
all other taxpayers, as is the case when seeking
ri?e-shot legislation (Hulse, 1996). Inter-social
interest con?ict exists between corporations and
other organized interests such as labor, consumer,
or environmental advocacy groups (Mitnick,
1993). This con?ict may be regarding speci?c leg-
islation (e.g., raising the minimum wage) or a
more general con?ict over the distribution of gov-
ernment support (or taxation).
Numerous arenas for political con?ict exist because
of the complexity of the US federal public policy-
making process. For this reason, public policy studies
often examine issues through a life cycle model com-
posed of three stages: public opinion formation, pub-
lic policy formulation, and public policy
implementation (Buchholz, 1988; Mack, 1997).
4
The
public opinion formation stage occurs when there
becomes a substantial di?erence between public
expectations and institutional performance such
that a de?nable issue is formed. A public policy
issue enters the public policy formulation stage
when the issue becomes politicized, is widely dis-
cussed, becomes a concern for a?ected interest
groups, and is formally introduced as speci?c leg-
islation. When legislation becomes public law the
public policy issue becomes bureaucratized and
enters the public policy implementation stage
(Buchholz, 1988).
Corporate political activity during the public policy
formulation stage
Our study focuses on corporate political activity
during the policy formulation stage of the 1997
Act. A federal tax policy issue typically enters the
public policy formulation stage through commit-
tee discussions within the US House of Repre-
sentatives’ Ways and Means Committee or the
Senate Finance Committee (Congressional Uni-
verse, 2000) and ends when the ?nal legislation is
passed (or defeated). The public policy formula-
tion stage of the 1997 Act lasted approximately 6
months (Congressional Universe, 2000).
Just as individual corporations develop strate-
gies to enhance their economic power, they also
select and execute strategies designed to increase
their political power, especially during the policy
formulation stage (Suarez, 1998). Once a corpora-
tion commits to political activity during this stage,
its speci?c actions depend mainly upon the sub-
stantive nature of the policy issue (Snyder, 1992)
and the level of political con?ict (Grier et al., 1994;
Suarez, 1998).
During the public policy formulation stage,
corporations advance their goals by undertaking
activities such as testifying at congressional hear-
ings, lobbying public policymakers, and making
political action committee (PAC) and individual
campaign contributions (Buchholz, 1988; Mack,
1997). Corporations testify at congressional hear-
ings and lobby public policymakers to communicate
their views to legislators regarding proposed legis-
lation. These activities provide a corporation with
the opportunity to provide legislators with their
reasons for adopting a particular position on pro-
posed legislation and to try and in?uence policy
outcomes (Hillman and Hitt, 1999; Mack, 1997).
Because legislators have limited amounts of time to
devote to policymaking, corporate political activity
research views campaign contributions as the pri-
mary activity that can be used to gain the access to
legislators that is required to in?uence policy.
In inter-social interest and some types of inter-
?rm con?ict, corporations may augment their
4
While the ?uidity of actual policy-making raises questions
concerning the discreteness of each stage, the life cycle model is
viewed as a useful tool in the analysis of corporate political
activity (Getz, 1993, p. 247).
R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590 569
political activities through coordination with trade
associations and/or advocacy coalitions (Hojnacki,
1998; Mack, 1997; Mizruchi, 1992; Suarez, 1998).
Trade associations typically represent a broad
spectrum of industry membership. These types of
associations (e.g., the American Institute of Certi-
?ed Public Accountants) are most likely to be
politically active during policy formulation when
its members have similar concerns over proposed
legislation (Mack, 1997; Hillman & Hitt, 1999).
Advocacy coalitions are most likely to form when
multiple industries and/or ?rms reach agreement
on the need for speci?c policies or legislation to be
passed (Mack, 1997; Suarez, 1998). For example,
during the debate over private securities litigation
reform both an inter-social interest and an intra-
industry coalition were formed. Public accounting
?rms, investment banking ?rms, and many cor-
porations generally agreed on the need for secu-
rities liability reform. Over 1400 organizations
joined an inter social-interest coalition named the
Coalition to End Abusive Securities Suits (Fritsch,
1995). Similarly, public accounting ?rms agreed on
the need for reform and an intra-industry coalition,
The Accountants Coalition, was also organized
(Journal of Accountancy, 1996).
Researchers have been particularly interested in
the potential link between campaign ?nancing and
legislative decision-making. The relationship is
di?cult to untangle because of the complexities
inherent in the campaign ?nance system and in the
legislative process (Grenzke, 1989a, 1989b). Prior
research has studied PAC contributions made
from a variety of industries including defense
(Fleisher, 1993), ?nancial services (Kroszner &
Stratmann, 1998; Romano, 1997), and trucking
(Frendreis & Waterman, 1985). These studies
model both legislators and interest groups using
pluralist assumptions. In this context, corpora-
tions and coalitions use PAC contributions strate-
gically in an attempt to in?uence the legislators
most important to them (Grier, Munger, & Tor-
rent, 1990). Prior empirical work that is grounded
in neoclassical economic theory has generally
found that corporate PAC contributions are more
likely to be given to legislators of the majority
party, higher-ranking legislators, and members of
committees with jurisdiction over the policy issues
of concern (Grier et al., 1990; Romano, 1997;
Kroszner & Stratmann, 1998).
Taxation and corporate political activity
Prior research shows that tax policy a?ects the
business decisions of corporations (Vines & Moore,
1996) and that corporations engage in inter-social
interest (Jacobs 1987, 1988; Quinn & Shapiro,
1991) intra-industry (Suarez, 1998), and inter-?rm
(Hulse, 1996; Suarez, 1998) political con?ict in
hopes of reducing their ?rms’ overall tax burden
(Mizruchi, 1992).
Quinn and Shapiro (1991) examined several
hypothesized relationships among taxation, redis-
tribution, politics, and business/class power. Using
four alternative empirical models, they found strong
support for the presence of inter-social interest
political con?ict between business and labor. Of
particular note, they found a signi?cant relation-
ship between increases in the general level of cam-
paign ?nancing from the corporate sector and
lower overall corporate tax burdens.
Freed and Swenson (1995) investigated cam-
paign contributions made by US corporations to
tax-writing members of the US Congress prior to
the passage of the Economic Tax Recovery Act
of 1981 and the Tax Reform Act of 1986. Their
study, developed within a marginalist economic
framework, hypothesized that a corporation’s
campaign contributions to tax writing legislators
was determined by the ?rm’s cost/bene?t analysis.
Using ?rm speci?c data for approximately 400 tax
paying ?rms for each tax act, Freed and Swenson
regressed two related measures of ?rm campaign
contributions against measures of ?rm size,
industry concentration, and expected bene?ts of
the tax legislation. Their results were supportive of
the model, with measures of industry speci?c tax
accounting rules related to the Tax Reform Act of
1986 having the largest impact on contributions.
Although not using the framework of corporate
political activity theory, Hulse (1996) investigated
the stock market reaction to ?rms’ successful tax
code restructuring e?orts. Hulse utilized an events
study methodology to estimate the stock market
reaction individual ?rms experienced when ri?e-
shot rules bene?ting their ?rms were included in
570 R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590
the Tax Reform Act of 1986. While Hulse (1996)
emphasized the secrecy surrounding these types of
provisions and the e?orts of lobbyists to obtain
tax code restructuring for their clients, he did not
investigate the political strategies used by cor-
porations in their restructuring e?orts.
Northcutt and Vines (1998) studied changes in
?nancial accounting methods used by corpora-
tions that were characterized by advocacy groups
as ‘‘corporate freeloaders’’. Their sample con-
sisted of corporations that were singled out and
criticized by the Citizens for Tax Justice because
of the corporations’ extremely low e?ective tax
rates. Northcutt and Vines (1998) found that cor-
porations responded to this political scrutiny in
later reporting periods by choosing income-
decreasing accruals with low book-tax conformity,
thereby increasing their reported e?ective tax rate.
By examining the interplay between advocacy
groups and corporate interests, this study provides
an example of inter-social interest level research.
Suarez (1998) reported the ?ndings of her case
study of the US possessions’ tax credit (e.g., credits
for maintaining manufacturing facilities in Puerto
Rico). She analyzed relevant public information
regarding political action committee contribu-
tions, congressional hearings, and company reports.
Suarez argued that when a tax policy issue is
mutually bene?cial to a group of ?rms or indus-
tries, coalitions and trade associations coordinate
political activities in order to maximize their
e?ectiveness. However, when a tax policy issue
a?ects these groups di?erently, coalition and trade
association alliances recede and fragmented lob-
bying emerges. Suarez concluded from her case
work that corporations used collective and frag-
mented political alliances in attempts to protect
the possessions’ tax. In particular, she asserted
that the introduction of proposed changes that had
di?erential impacts on the electronic and pharma-
ceutical industries collapsed coalition e?orts and
resulted in an increase in inter-?rmpolitical con?ict.
Implications for the case study
The two sections that followdescribe our research
method and the results of our case analysis con-
cerning corporate political activities during the
6-month policy formulation stage of the Taxpayer
Relief Act of 1997. Our study of prior research on
the accounting/state relationship, corporate poli-
tical activity, and tax accounting helped us iden-
tify several contributions that our research can
make to our collective understanding of the devel-
opment of US tax accounting laws. First, we are
interested in identifying and investigating cor-
porations that undertook political activities
regarding the 1997 Act. We argue that these cor-
porations view the state as a contracting party
over which they can exert in?uence. Thus, con-
trary to the microeconomic-based, strict pluralist
assumptions common in empirical tax accounting
research, tax accounting rules are not necessarily
exogenous to the ?rm. In other words, corporate
interests play a substantive role in the develop-
ment of tax accounting policy.
Second, in addition to addressing pluralist
assumptions, we also are able to extend research
on corporate political activity. As recommended
by Hillman and Hitt (1999), we investigate and
document the speci?c behaviors that ?rms use to
in?uence legislation. Further, we document and
analyze corporate behaviors at three di?erent
levels of political competition.
Third, evidence that corporations exert in?uence
in the development of tax policy is not necessarily
inconsistent with pluralist theory. In order to
assess the applicability of strict pluralist assump-
tions, we must examine the intensity and extent of
corporate political activities. Intense and extensive
political activities made by corporations during
the six-month period bolster arguments that all
interest groups do not have similar opportunities
to in?uence accounting policy. The more that
lobbying and campaign contributions allow cor-
porate interests to dominate the public policy for-
mulation process, the less applicable strict
pluralist assumptions are in the study of policy
development (Clawson et al., 1998). Thus, this case
study provides a basis from which to re?ect on the
appropriateness of any pluralist-based theory to
accounting policy research. The nature and levels of
political con?ict that we analyze can shed light on
the importance of incorporating political economy
more prominently in the study of tax accounting
policy.
R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590 571
Research method
Our study followed a four-step empirical case-
based research approach (Ryan et al., 1992, pp. 113–
129). First, we analyzed the provisions of the 1997
Act, and publications, press releases and other
documentation associated with the 1997 Act.
These data were gathered through a repetitive
search of electronic databases (ABI-Inform, Busi-
ness Index, Congressional Universe, Firstsearch,
Wall Street Journal Index, and multiple search
engines for the World Wide Web). Our search
produced a wide variety of public documents
related to the 1997 Act, including published trade
articles, transcripts of congressional hearings,
legislative histories of the proposed law, popular
press articles, and newspaper columns.
From this search, we identi?ed three sets of
taxpayers for further analysis. These taxpayers
were selected because each corresponded most
closely with a speci?c level of political con?ict
(Buchholz, 1988; Mack, 1997). The insurance
industry was concerned with numerous provisions
in the 1997 Act. Their uni?ed e?orts provided an
example of inter-social interest political con?ict
because the industry attempted to in?uence a
number of tax law changes within a constrained
resource setting. Thus, to the extent they were
successful, their e?orts a?ected the overall dis-
tribution of the tax burden. The airline industry
displayed intra-industry political con?ict over
the airline tax provisions. Major airline carriers
and regional carriers strongly disagreed over the
structure of the tax. Amway Corporation and
Sammons Enterprises were selected as glaring
examples of inter-?rm political con?ict because
these corporations sought speci?c, narrowly tar-
geted, tax relief.
In the second step of our research process we
gathered lobbying reports and political action
committee contribution data for the taxpayers
under study for the period relevant to the policy
formulation stage of the 1997 Act. The data were
collected to provide additional primary source
information regarding the political activities of
these taxpayers and to corroborate the informa-
tion we gathered from secondary sources. The
lobbying and political contribution activities rele-
vant to the public policy formulation stage of the
1997 Act occurred between the times the 1997 Act
was originally introduced (as the Revenue Recon-
ciliation Act) during January 1997 until its passage
on August 5, 1997 (Congressional Universe, 2000).
Lobbying reports are ?led with the Clerk of the
House of Representatives and the Secretary of the
Senate on a semi-annual basis. Therefore, we gath-
ered reports for Amway Corporation, Sammons
Enterprises and for all insurance and airline rela-
ted ?rms and associations for the January–June
1997 reporting period. These mandatory ?lings are
required of all organizations engaged in lobbying
e?orts. The ?lings specify the legislation being
lobbied, the cost of lobbying activities, whether or
not the ?ling was made by a lobbying ?rm, and
the organization ?nancing the lobbying activity.
Next we gathered PAC contribution data for
each ?rm that reported lobbying activity asso-
ciated with the 1997 Act. Lobbying and making
?nancial contributions are ?rms’ and associations’
most visible and e?ective activities during the pol-
icy formulation stage of proposed legislation
(Hillman & Hitt, 1999; Mack, 1997). Political
action committee contributions are reported to the
Federal Election Commission and compiled by the
Center for Responsive Politics (2000). These con-
tributions are reported by election cycle. We
gathered contribution data for the 105th Con-
gressional election cycle because the 1997 Act was
passed by this congress.
The third step in our research was to use the
framework provided by our theoretical work to
analyze and interpret the information that was
gathered in steps one and two (Grier et al., 1994;
Mitnick, 1993; Suarez, 1998). This analysis pro-
duced evidence of the taxpayers’ interests in the
1997 Act and their motivation for undertaking
lobbying and political campaign contribution
activities. Political campaign contribution data
reported ?rms’ contribution amounts that were
donated to each senator and house member but,
of course, did not provide direct information on the
reason for the contribution. Therefore, we appealed
to prior research on PAC contributions to guide
our empirical work (Romer & Snyder, 1994;
Snyder, 1992; Wright, 1985). In the two cases in
which multiple ?rms were involved (insurance and
572 R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590
airlines), we developed and tested an empirical
model to explain the campaign contributions
made to Senate and House members by these
?rms’ political action committees.
As we mentioned in the previous section, prior
research models ?rms’ PAC contribution decisions
through a neoclassical economics perspective and
hypothesizes that ?rms allocate contributions to
legislators who are in the best positions to lead
e?orts to have Congress adopt the corporation’s
favored legislation (Grier et al., 1994). According to
prior studies, these are most likely to be legislators
who hold leadership positions and/or who are
members of committees with jurisdictional author-
ity over relevant legislation. Thus, the political con-
tributions that were particularly relevant to the
1997 Act were contributions given to congres-
sional leaders and/or members of the House Ways
and Means Committee or the Senate Finance
Committee.
Prior PAC research also states that legislators
who hold in?uential positions may not be equally
likely to attempt to provide the desired legislation
(Grier et al., 1994; Snyder, 1992). This di?erential
in legislators’ willingness to support a corporation’s
speci?c policy preferences is because legislators are
ultimately concerned with re-election. Therefore,
their ability to support legislation that is important
to PAC contributors is constrained by their own
political ideology and their voting constituents’
legislative preferences (Grier et al., 1994; Snyder,
1992).
Our PAC contribution regressions controlled
for political ideology, constituent preferences, and
industry-related committee membership in order
to improve the reliability of our tests of the rela-
tionship between ?rm PAC contributions and leg-
islator in?uence over tax policy. We performed four
regressions, two regressions for each of the cases
with multiple ?rms (insurance and airlines) for each
legislative body (House and Senate). The general
speci?cations of the empirical models we tested
were:
PAC
house
¼ b
0
þ b
1
VOTE þ b
2
LEADTEAM
þ b
3
INDCOMMþ b
4
TAXCOMM
PAC
senate
¼ b
0
þ b
1
VOTE þ b
2
LEADTEAM
þ b
3
INDCOMMþ b
4
TAXCOMM
þ b
5
ELECT
þ b
6
ELECT
Ã
INDCOMM
þ b
7
ELECT
Ã
TAXCOMM
Where:
PAC= a legislator’s total receipt of PAC
contributions from either
insurance or airline industry
PACs that lobbied the 1997
Act (100 observations for the
Senate and 437 for the House
of Representatives).
VOTE= the legislators’ cumulative voting
rating by the U.S. Chamber
of Commerce (range=0-100).
The higher the rating, the more
favorable their view of the
legislator’s voting record.
LEADTEAM=1 if the legislator holds a
leadership position; 0 otherwise.
INDCOMM= 1 if the legislator is a member of
a committee that has jurisdiction
over legislation related
speci?cally to a ?rm’s industry
(not tax-related); 0 otherwise.
TAX COMM=1 if the legislator is a member of
the legislative committee that has
jurisdiction over tax legislation.
ELECT= 1 if the legislator is a Senator and
is up for reelection during the
next election cycle; 0 otherwise.
This variable is used only in the
Senate regressions.
TAX COMM is our primary test variable
because its signi?cance shows that a relationship
R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590 573
exists between corporate PAC contributions and a
legislator’s ability to a?ect tax accounting changes
that are preferred by those corporations. All of the
other variables controlled for non-tax-related fac-
tors that may explain variance in contributions
made by insurance or airlines ?rms to legislators
during the 105th congressional election cycle.
VOTE was included in the model to control for
political ideology and constituent voting pre-
ferences. As expected, VOTE is highly correlated
with legislators’ political party membership (i.e.,
Republican legislators possess higher pro-business
voting ratings than Democrats). LEADTEAM
and INDCOMM were included to control for
other legislator attributes that could explain PAC
giving. For LEADTEAM to be set equal to one, a
legislator must have served the 105th Congress as
Majority Leader, Minority Leader, Majority
Whip, Minority Whip or Chair of the applicable
tax or industry committee.
For the House of Representatives, INDCOMM
represented membership on the Commerce Com-
mittee in the insurance regression and it repre-
sented membership on the Transportation
Committee in the airline regression. For the Sen-
ate, INDCOMM represented membership on the
Banking, Housing, and Urban A?airs Committee
in the insurance regression and it represented
membership on the Commerce, Science, and
Transportation Committee in the airline regression.
These committees have jurisdiction over legisla-
tion that is most vital to each respective industry
(Congressional Universe, 2000).
5
The ELECT variable is added to the Senate
regressions because US senators serve 6-year,
staggered terms, resulting in only one-third of the
Senate incumbents being up for reelection each
two-year election cycle. This variable is important
because PAC contributions to individual senators
tend to increase during their election year and,
therefore, should be controlled for in the regressions.
All US House of Representatives incumbents are
up for reelection every 2 years, thus the ELECT
variable is not needed. Because of the potential
interaction between ELECT and TAXCOMM and
between ELECT and INDCOMM, we include
interaction terms in the Senate regressions.
We argue that a statistically signi?cant relation-
ship between PAC and TAXCOMM, after con-
trolling for these other factors, provides evidence
that corporate interests view the state as a negoti-
able contracting party when developing tax stra-
tegies. In this particular case, statistical signi?cance
of TAXCOMM will support our contention that
insurance and airline ?rms used their resource
endowments to their advantage to make PAC
contributions that enabled them to have access to
and in?uence over the determination of tax
accounting law during the policy formulation
stage of the 1997 Act. Statistical signi?cance does
not disprove pluralist theory, but it does provide
support for increasing the attention given to the
political economy aspects of accounting policy.
We report parameter estimates to indicate the
dollar e?ect of each independent variable on the
dependent variable.
6
5
Given our de?nitions for LEADTEAM, TAXCOMM and
INDCOMM, the chairs of TAXCOMM and INDCOMM were
included in the LEADTEAM variable. We performed sensitiv-
ity analyses of our results leaving these chairs out of LEAD-
TEAM. The results presented in the paper were not changed
when omitting committee chairs from the LEADTEAM
variables.
6
We recognize that the corporate taxpayers we are studying
could be lobbying Congress for reasons other than the 1997
Act. We thoroughly explored Congressional Universe for com-
mittee reports, testimony and bills of interest to the airline and
insurance industry before the House Ways and Means com-
mittee and/or the Senate Finance Committee during the time
period in question. Our search did not turn up any other bills
pertinent to the airline industry. However, Medicare legislation
was before the House Ways and Means committee at the same
time as the tax legislation. As a sensitivity check, we split the
House Ways and Means Committee into two variables. One
variable was coded ‘‘1’’ for House Ways and Means committee
members who were on the Health Subcommittee, the other
variable was coded ‘‘1’’ for House Ways and Means members
who were not on the Health Subcommittee. Our reasoning is
that if the PAC contributions were being made only to in?u-
ence Medicare legislation, we would expect to see a signi?cant
result for the members who were on the Health Subcommittee
versus. those House Ways and Means members who were not
on the Health subcommittee. This analysis showed that both
variables were highly signi?cant (P-value=0.000), and the
parameter estimates were similar. We suggest that this is evi-
dence consistent with our argument that these PAC Contribu-
tions were made, at least in part, to in?uence the tax bill.
574 R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590
Fourth, after completion of all of our analyses,
we organized our ?ndings in order to present a
structured examination of the corporate political
activities of each set of taxpayers. The analyses are
presented in the following section. First, we discuss
the tax issues and provisions of interest to the set
of taxpayers. Second, we discuss our ?ndings
regarding lobbying and political campaign con-
tribution activities.
Corporate political activity and the Taxpayer
Relief Act of 1997
Inter-social interest corporate political activity: the
insurance industry
Tax issues
The insurance industry engaged in corporate
political activity on a number of complex tax issues
that either directly a?ected the tax burden of the
industry or a?ected demand for their products. Our
search of industry publications and press releases
enabled us to identify the major 1997 Act provisions
of interest to the industry as well as its position on
these provisions. Table 1 summarizes these issues.
The insurance industry favored a change in the
treatment of foreign sourced investment income.
Subpart-F rules require ?nancial services companies
(e.g. insurance, banks and securities ?rms), but
not manufacturing ?rms, to immediately recognize
foreign-earned investment income even if the
funds have not been repatriated to the parent cor-
poration (Brosto?, 1997b). The insurance indus-
try, along with a coalition of the other a?ected
industries argued that these rules impeded their
global competitiveness (Brosto?, 1997f).
7
Insurance industry publications documented their
economic and political interest in the highway tax
provisions included in the 1997 Act. A 4.3-cent-
per-gallon tax had originally been enacted in 1993
as a de?cit reduction measure. However, a coalition
of industries (including insurance and construc-
tion) lobbied to have the revenue re-directed to the
Highway Trust Fund (Winston, 1997). Increased
funding for federal highway programs would
reduce automobile accidents and insured losses
(Brosto?, 1997d).
8
The insurance industry’s coali-
tion strategies concerning the gasoline tax and the
Table 1
Insurance industry 1997 Tax Act provisions
Tax issue Insurance
industry position
1997 Act change
Sub-part F Rules required tax on non-repatriated foreign
source investment income for ?nancial services ?rms
FAVORED
ELIMINATION OF TAX
Included one year reprieve
to the Sub Part F rules
Reallocation of 4.3 cent federal gas tax from de?cit reduction
to the Highway Trust Fund
FAVORED THIS CHANGE Provision was included in Act
Restriction of general borrowing interest expense based on
the existence of unborrowed cash values on Corporate
Owned Life Insurance (COLI)
OPPOSED RULE, HOWEVER
SOUGHT CLARIFYING
LANGUAGE TO LIMIT
REACH OF RULE
Limitation on interest expense
from general borrowing was
restricted, but it only applied to
mortgage-owned life insurance
Proposed changes in NOL carryover rules from
3 back and 15 forward to 2 back and 20 forward
OPPOSED RULE New NOL carryover rules were
enacted. Special exceptions were
allowed for presidentially declared
disaster areas
7
The 1997 Act passed by both houses of Congress included
a one-year reprieve from Subpart-F rules, however this provi-
sion was one of two vetoed by President Clinton under his line-
item veto privileges (Congressional Universe, 2000). The
Supreme Court later ruled the line-item veto law unconstitu-
tional; and the rule change was enacted (Congressional Uni-
verse, 2000).
8
Their e?orts were deemed successful because an amend-
ment to the 1997 Act was added to divert the 4.3-cent-per-gal-
lon fuel tax from de?cit reduction to the Highway Trust Fund
(Brosto?, 1997d).
R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590 575
Subpart-F rules is consistent with expectations
derived from corporate political activity research
because of the inter-social interest nature of these
provisions (Mack, 1997; Mizruchi, 1992; Suarez,
1998).
The insurance industry was also concerned with
the 1997 Act provisions a?ecting corporate owned
life insurance (Kerley, 1997b). While insurance
premiums are not deductible (if the payee is the
bene?ciary) and death bene?ts are not taxable,
any interest expense related to general borrowing
of funds to purchase these policies was deductible
prior to the 1997 Act. Congress’ interest in closing
this loophole caused two major concerns. First,
some insurance companies (along with Fannie
Mae) were planning to exploit the interest deduct-
ibility and opposed changes to the tax rule
(Kerley, 1997b). Second, other companies feared
that the rule changes would be drafted too
broadly.
9
Finally, the insurance industry was also opposed
to a change to the net operating loss (NOL) car-
ryover rules. Prior to the 1997 Act, a NOL could
be carried back 3 years and forward 15 years. The
1997 Act changed the carry back period to only 2
years, while the carry forward period was exten-
ded to 20 years. The insurance industry, both the
Alliance of American Insurers and the American
Insurance Association, argued that this change
would be particularly harmful to property and
casualty insurers that su?er catastrophic losses
(Brosto?, 1997e). While the industry was unable
to halt the rule change, an exception that allowed
the old rule to apply was made for losses that
occurred in presidentially declared disaster areas.
10
Lobbying and political action committee activities
Through a manual search of lobbying reports,
we found 112 reports associated with insurance
lobbying activity regarding the 1997 Act. These
reports were ?led either by insurance corporations
or associations or by lobbying ?rms that repre-
sented the insurance industry. Our analysis of the
reports documents that the 50 insurance compa-
nies that lobbied this bill spent approximately
$15.2 million on lobbying expenses in the ?rst 6
months of 1997. In addition, 16 insurance industry
associations reported an additional $6.4 million of
lobbying expenditures during this period.
These lobbying expenditures were supplemented
by campaign contributions to legislators (Center for
Responsive Politics, 2000) and ‘‘soft money’’ con-
tributions made directly to political parties (Com-
mon Cause, 2000). In total, the insurance industry
members who lobbied during the formulation stage
of the 1997 Act made PAC contributions of $5.0
million to the campaigns of US House members
and $1.9 million to the campaigns of US Senators
(Center for Responsive Politics, 2000). In addition,
‘‘soft money’’ contributions of $2.8 million were
made by the insurance industry during this time
period (Common Cause, 2000). For this time per-
iod, the insurance industry’s soft money contribu-
tions were more than that of any other industry
(Common Cause, 2000). Table 2 summarizes the
corporate political activities of the insurance indus-
try. In total, we documented over $30 million of
corporate political activity expenditures by this
industry.
Although soft money contributions cannot be
directly linked to individual legislators, PAC con-
tributions are made to individual congressional
campaigns. We examined the relationship between
insurance PAC contributions and legislator mem-
bership on either the House Ways and Means
Committee or the Senate Finance Committee by
developing and testing an empirical model that
controlled for other factors that could in?uence
10
The insurance industry also expressed concerns over pro-
posed changes that a?ected individual taxpayers. This was
because the provisions, if adopted, could have business reper-
cussions. These issues included privatization of social security
(Brosto?, 1997a), increased deductibility of health insurance
premiums for self-employed taxpayers (Kerley, 1997a), changes
in estate taxes (Kerley 1997a; Brosto?, 1997c; King, 1997), and
decreases in the capital gains tax rate (Kerley, 1997a). While
not explored further in our study, this further demonstrates the
extent to which tax-related issues a?ect corporate business and
tax-related political activities.
9
The industry was able to obtain clarifying language in the
provision that limited this provision to non-employees. Further
the tax rule was written such that it does not apply to coverage
of former employees, o?cers and directors (Brosto?, 1997b).
Thus, interest expense for general borrowing will not be dis-
allowed as a result of unborrowed policy cash values, except for
the case of mortgage-owned life insurance.
576 R.W. Roberts, D.D. Bobek / Accounting, Organizations and Society 29 (2004) 565–590
insurance PAC giving. The regression results for the
House of Representatives are shown in Table 3
and the results for the Senate are shown in Table 4.
Each table provides descriptive statistics for each
variable in the model and ordinary least squares
regression results.
11
Panel A of each table presents
descriptive statistics as well as the t-test results of
the di?erences in the mean PAC contribution for
each level of the dichotomous independent vari-
ables. The reported means indicate the dollar
impact of each of the independent variables. Panel
B of each table presents the multivariate OLS
regression results for the PAC contribution model.
We also present descriptive data regarding legisla-
tors’ political party a?liation but this variable is
not included in the regressions because it is highly
correlated with the legislators’ voting records. The
results provide strong evidence that the insur-
ance industry allocated its PAC contributions to
Table 2
Insurance industry summary of corporate political activities
1/1/97–6/30/97 1997 Tax Act lobbying expenditures
a
Insurance Companies (n=50) $ 15,189,740
Insurance Industry Associations (n=16) 6,362,646
Total Lobbying Expenditures by Insurance
Industry related to 1997 Tax Act, for ?rst
Six months of 1997 $ 20,512,993
PAC contributions to individual house and senate members by insurance companies and
insurance industry associations who lobbied for 1997 Tax Act: 1998 election cycle
b
House Members $ 5,005,851
Senate Members $ 1,942,667
Total PAC Contributions $ 6,948,518
‘‘Soft money’’ contributions to political parties by insurance industry
c
To the Democratic Party $ 902,161
To the Republican Party $ 1,942,667
Total ‘‘soft money’’ contributions $ 2,841,860
a
Lobbying reports are ?led with the Secretary of the Senate and the Clerk of the House of Representatives as required by the
Lobbying Disclosure Act. These reports include information about the speci?c bills that were being lobbied. We only included
amounts from lobbying reports that included the 1997 Tax Act as one of the lobbying targets. Note: in some cases the reported
amount was ‘‘ 70 seats
$1/seat on aircrafts with