Constituent lobbying and its impact on the development of financial reporting

Description
This paper examines the impact of constituent lobbying activity on accounting regulators during the transformation
of the Fourth European Company Law Directive into German accounting law. Using detailed published commentaries
prepared by representative organisations on draft accounting legislation, we provide evidence concerning the pre-
ferences of the three primary German constituenciesÐpreparers, auditors, and academic experts. Initially, a model that
merely distinguishes between the three constituencies suggests that the industry lobby group representing preparers
exerts the greatest in¯uence on the decisions of the German legislature. However, when the empirical model is extended
to include all two-way interaction e€ects, the relative power of preparers is seen to be far lower, with the in¯uence
exerted by industry depending crucially on the support of at least one of the remaining lobby groups.

Constituent lobbying and its impact on the development of
®nancial reporting regulations: evidence from Germany
Stuart McLeay
a,
*, Dieter Ordelheide
b
, Steven Young
c
a
School of Accounting, Banking, and Economics, University of Wales, Bangor, UK
b
School of Business and Economics, J.W. Goethe-UniversitaÈt, Frankfurt am Main, Germany
c
International Centre for Research in Accounting, Lancaster University, UK
Abstract
This paper examines the impact of constituent lobbying activity on accounting regulators during the transformation
of the Fourth European Company Law Directive into German accounting law. Using detailed published commentaries
prepared by representative organisations on draft accounting legislation, we provide evidence concerning the pre-
ferences of the three primary German constituenciesÐpreparers, auditors, and academic experts. Initially, a model that
merely distinguishes between the three constituencies suggests that the industry lobby group representing preparers
exerts the greatest in¯uence on the decisions of the German legislature. However, when the empirical model is extended
to include all two-way interaction e?ects, the relative power of preparers is seen to be far lower, with the in¯uence
exerted by industry depending crucially on the support of at least one of the remaining lobby groups. #1999 Elsevier
Science Ltd. All rights reserved.
1. Introduction
The economic consequences of policy decisions
make it impossible for accounting regulators to
select non-controversial treatments on any of the
issues they face based solely on technical con-
siderations. Instead, the choice of appropriate
accounting regulations re¯ects, at least in part, a
social decision whereby regulatory bodies attempt
to manage con¯ict between competing con-
stituencies by selecting the most socially acceptable
solution (Horngren, 1972; Sunder, 1988; Ze?,
1978). The political nature of accounting rule-
development naturally raises questions concerning
both the responsiveness of policy-makers to con-
stituent pressure, and the distribution of power
among competing interest groups. One means by
which prior research has explored the source and
extent of political in¯uence in the context of
accounting regulation is through the analysis of
constituent lobbying activity. This study builds on
extant research by examining public lobbying
activity in Germany during the period when the
Fourth European Company Law Directive (here-
after Fourth Directive) was transformed into
German commercial law, the regulatory regime at
the time being fundamentally di?erent to that
previously analysed.
Lobbying research in the accounting literature
has focused almost exclusively on ®nancial
reporting regimes in English-speaking countries,
most notably that of the US (Puro, 1984, 1985;
Sutton, 1984; Tandy & Wilburn, 1992, 1996), and
0361-3682/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PI I : S0361- 3682( 99) 00028- 8
Accounting, Organizations and Society 25 (2000) 79±98
www.elsevier.com/locate/aos
* Corresponding author.
to a lesser extent those of the UK (Hope & Gray,
1982; Nobes, 1991; Sutton, 1984), Australia
(Klumpes, 1994; Walker & Robinson, 1994) and
New Zealand (Rahman, Ng & Tower, 1994).
While this body of work a?ords crucial insights
into the nature of the standard-setting process, the
institutional similarity of these regimes limits the
generalisability of reported ®ndings and provides
little evidence as to the e?ect of the organisational
form on the lobbying process. Examination of
alternative regulatory environments is, therefore,
warranted. In this context, the system operating in
Germany until recently provided one of the most
obvious examples of codi®ed accounting regula-
tion (Nobes & Parker, 1991, p. 12), where the leg-
islature assumed primary responsibility for the
development of ®nancial reporting regulations,
and where accounting rules were drafted by law-
yers and administrators in the Ministries of Justice
and Finance.
Implementation of the Fourth Directive repre-
sented a landmark in German accounting regula-
tion, codifying as it did certain aspects of ®nancial
reporting which were already generally accepted in
countries such as the US and UK.
1
On the other
hand, as a key element of the European Union's
harmonisation programme, the Fourth Directive
also contained many aspects of German law which
were then implemented elsewhere in Europe, i.e.
the requirements for companies to adopt (i)
de®ned methods of asset valuation and (ii) com-
pulsory balance sheet and pro®t and loss formats.
In the case of Germany, the provisions of the
Fourth Directive were enacted into national law
via the Accounting Directives Law (Bilanzrichtli-
nien-Gesetz), forming the legal basis for ®nancial
reporting for German companies. To obtain
agreement on the provisions of the Fourth Direc-
tive at a European level, the document contained a
large number of national options, a?ording mem-
ber states signi®cant ¯exibility over the transfor-
mation of many of its speci®c provisions. At a
national level, the process of choosing between
these various alternatives has provided a powerful
setting within which to explore the political nature
of accounting regulation.
2
Using commentaries published by representative
organisations during the transformation process,
we identify the preferences of the three primary
German accounting constituencies (industry,
auditors, and academic experts) in relation to 169
separate ®nancial reporting issues spanning
aspects of valuation, recognition, disclosure and
®nancial statement format. A striking feature of
the approach adopted in Germany when the
Fourth Directive was implemented is the manner
in which representations made to the legislature
re¯ected the collective view of the members of
professional and industrial associations.
The relative in¯uence of each of the lobby
groups involved is assessed by examining the
extent to which constituents' stated preferences
were ultimately re¯ected in the Accounting Direc-
tives Law. The approach used to identify lobby-
ists' preferences is similar to that employed in
extant research whereby lobbyists' publicly avail-
able comments are classi®ed as either `for' or
`against' the associated exposure draft (e.g. Brown,
1981; Francis, 1987; Puro, 1984). Traditionally,
such a coding procedure has proved problematic
due to the relatively broad nature of both com-
ment letter responses and the exposure drafts to
which they pertain, resulting in the need for con-
siderable judgement when reconciling the two
(Puro, 1984).
3
In contrast, because the German
accounting rules under consideration were similar
in structure to US tax code regulations, lobbyists'
proposals and the sections of commercial law to
1
Notable examples include the `true and fair view' require-
ment and the need for increased disclosure.
2
Since this study examines only the transformation of the
Fourth Directive into German national law, rather than the
development of the Directive at a European level, it is possible
that an additional dimension of the political process has been
omitted from the analysis. At a national level, the transforma-
tion recommendations proposed by any lobby group(s) must
have been developed within the framework of the Fourth
Directive. Consequently, the framework of the Directive repre-
sents a signi®cant external constraint on the regulatory process
(Diggle & Nobes, 1994; Hussein & Ketz, 1991).
3
For example, Puro (1984) discusses the coding problems
arising from cases in which respondents support one or more
parts of an exposure draft while opposing or remaining silent
on others.
80 S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98
which they refer were concise and highly speci®c in
nature. This results in (i) greater objectivity with
respect to the identi®cation of lobbyists' pre-
ferences and (ii) a more reliable measure of power
(i.e. the extent to which these preferences are now
re¯ected in commercial law).
A generalised linear model is used to assess how
the German legislature incorporated constituents'
preferences in the regulations it promulgated. In
addition to providing an appropriate means of
modeling the relative in¯uence of each lobby
group in isolation, this approach enables a formal
analysis of the way in which constituents' pre-
ferences interact to determine policy outcomes. To
the extent that the support of all sectors of the
®nancial community is often seen as an essential
element in the development of accounting regula-
tions (Horngren, 1972), the potential in¯uence of
any particular lobbyist is expected to be condi-
tional, at least in part, on the position adopted by
other lobby groups (Benveniste, 1972; Dyckman,
1988). To date, however, this issue remains rela-
tively unexplored in the empirical literature. Initial
®ndings for our main e?ects models suggest that,
conditional on a proposal being made, industry
exerted the greatest in¯uence on the decisions of
the legislature, while the odds of success asso-
ciated with an academic proposal were sig-
ni®cantly less than one. However, when the model
is extended to include all two-way interactions
between lobbyists, industry's power is seen to have
declined, with the outcome depending crucially on
agreement between industry and one or more of
the remaining lobby groups. Additional tests con-
®rm that, rather than any single group's pre-
ferences consistently dominating the legislature's
decisions, it was the existence of agreement among
participant groups that ultimately determined the
odds of success for a given proposal. As with the
US, therefore, the process of accounting-rule
development in Germany at the time of the
implementation of the Fourth Directive appears to
have been characterised by `power equivalency'
and the desire for political consensus.
In addition to empirically examining the politics
of accounting rule-development within a funda-
mentally di?erent and empirically more amenable
regime to that previously considered, this study
adds to the literature in three further respects.
First, several commentators (e.g. Puro, 1984; Sut-
ton, 1984) have suggested that the distribution of
power among competing constituencies may vary
as a function of the particular accounting issue
under consideration. Because this study examines
the politics of accounting rule-development across
a wide range of accounting issues in relation to a
single event, we are able to develop a more com-
prehensive model of accounting regulation than in
prior work, where the focus has been largely
restricted to the analysis of single (or small subsets
of) accounting issues in isolation. We present
results which suggest that power and in¯uence are
indeed conditional on the nature of the ®nancial
reporting issue: in particular, any relative in¯uence
that industry exerted on the legislature appears to
have been restricted to disclosure issues.
Secondly, although the standard-setting process
in many English-speaking countries has a strong
private sector tradition, a degree of state interven-
tion in regulatory initiatives remains a distinct
possibility (Vieten, 1995; Wyatt, 1991).
4
Even in
Germany and France, which have seen the recent
creation of standard-setting bodies, ®nancial
responsibility for accounting regulation remains
within the state budget. The precise e?ect of these
di?ering regulatory structures on the rule-devel-
opment process remains ambiguous: for example,
while Dyckman (1988) raises concerns over the
susceptibility of public sector regulators to political
pressure from those antagonistic to business, Sut-
ton (1984) contends that the deliberations of pri-
vate sector institutions are more susceptible to
constituent in¯uence due to their lack of legal
mandate.
5
To the extent that theoretical arguments
4
The FASB operates under the continual threat of increased
Congressional involvement in the standard-setting process
(Kirk, 1988). In the UK, recent changes to the standard-setting
structure have resulted in the creation of a direct link between
the Financial Reporting Council and the Department of Trade
and Industry and while at present the UK government chooses
to maintain a relatively passive stance, this position cannot be
guaranteed in the future.
5
Further, Chatov (1985) argues that because bodies like the
FASB act as agents of the network creating them, they are
more likely to accede to the network's wishes, and in the case of
con¯ict are more likely to favour the preferences of the most
powerful group.
S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98 81
alone cannot unambiguously delineate the poten-
tial costs and bene®ts of public sector involvement
in the regulatory process, the issue is an empirical
question. Comparison of the ®ndings reported in
this paper with the evidence documented for the
US and similar systems highlights the existence of
important commonalities across di?erent institu-
tional settings with respect to the accounting rule-
development process.
Finally, the globalisation of business and the
expansion of the capital markets has led to
increasing pressure for accounting harmonisation
and, by implication, a greater understanding of
alternative ®nancial reporting regimes. However,
while many in the ®nancial reporting community
have resisted this pressure, choosing instead to
focus attention on US standard-setting activity on
the grounds that it is these standards which ulti-
mately form the basis of international harmonisa-
tion e?orts, the importance of alternative
accounting approaches can no longer be denied
(Wyatt & Yospe, 1993). Nowhere is this more
clearly demonstrated than in the SEC's decision to
allow three international accounting standards (or
parts thereof) to be accepted in cross border ®lings
without reconciliation to US GAAP,
6
even though
the results produced by the application of these
standards are substantially di?erent from those
produced under US GAAP (Bayless, Cochrane,
Harris, Leisenring, McLaughlin & Wirtz, 1996).
Evidence concerning both the process by which
accounting rules are developed and the way in which
these rules re¯ect constituents' preferences, repre-
sents an important step towards understanding the
product of alternative ®nancial reporting regimes.
The remainder of the paper is organised as fol-
lows. The next section reviews extant lobbying
research. We then present an overview of the
institutional characteristics of accounting regula-
tion in Germany at the time of the implementation
of the Fourth Directive, together with details of the
European accounting harmonisation programme.
The following sections discuss the data and sam-
ple, describe the estimation procedure, and present
the results. The paper concludes with a summary
and discussion.
2. Literature review
Using Downs' (1957) model of voting, Sutton
(1984) identi®es the conditions under which
rational lobbying will occur. Given a choice
between two alternative proposals, lobbying is
predicted in those circumstances where the di?er-
ential wealth e?ect associated with the two pro-
posals, discounted by the perceived probability of
in¯uencing the outcome, exceeds the cost of lob-
bying. Hence, the propensity for lobbying is
hypothesised to be increasing in both (i) the mag-
nitude of the perceived wealth e?ect and (ii) the
expectation of in¯uencing the ®nal decision. To
the extent that the potential economic con-
sequences of securing a favoured proposal are
thought to be greater in absolute terms for pre-
parers than for users, evidence reported by Tandy
and Wilburn (1992) for the US indicating that
preparers are more active lobbyists than users
serves to support Sutton's ®rst prediction. Similar
lobbying patterns have been documented by Sut-
ton for the UK and Walker and Robinson (1994)
for Australia. In addition, the signi®cant ®nancial
interest of audit ®rms in their clients' welfare helps
explain the relatively high lobbying propensity
observed for large US public accounting ®rms,
while the low levels of lobbying observed for US
accounting academics has been partly attributed
to the lack of a signi®cant wealth e?ect (Tandy &
Wilburn, 1996). Further, intra-industry studies
indicate that the likelihood of an individual ®rm
engaging in lobbying activity is also positively
associated with the magnitude of the perceived
wealth e?ects (e.g. Francis, 1987; Grin, 1983;
Kelly, 1982, 1985; Watts & Zimmerman, 1978).
Tandy and Wilburn (1996) also document direct
evidence in support of Sutton's second prediction.
Speci®cally, they ®nd that one of the main reasons
cited by US academics against active participation
in the FASB's due process procedures is their per-
ceived low probability of success. However, since
6
The International Accounting Standards (IASs) permitted
by the SEC in cross border ®lings are IAS 7: Cash Flow State-
ments, parts of IAS 21: The E?ects of Changes in Foreign
Exchange Rates and IAS 22: Business Combinations. Further,
IASs are now accepted by the London Stock Exchange for
cross-border listings.
82 S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98
prior research has focused almost exclusively on
the type of ®nancial reporting system prevalent in
English-speaking countries, little evidence exists to
support Sutton's additional assertion that these
predictions will hold irrespective of the institu-
tional setting within which accounting regulations
are developed.
While the bene®t term in Sutton's lobbying
equation is expected to vary between preparers
and users in general, it may also di?er across
alternative accounting issues for any preparer±
user combination (Sutton, 1984), suggesting that
constituent lobbying propensity may be a?ected
by the scope and nature of the accounting issue
under consideration. For example, Tandy and
Wilburn (1992) document greater lobbying activ-
ity by the industry and academic constituent
groups on substantive standards compared with
both industry standards and amendments. Simi-
larly, substantive standards elicited a greater mean
number of submissions from public accounting
®rms than for industry standards. Further, Puro
(1984) provides evidence that the extent of pre-
ference alignment between preparer and auditor
lobby groups may also di?er according to the
underlying nature of the accounting issue: pre-
parers and auditors are more likely to adopt simi-
lar lobbying positions in relation to proposals
designed to standardise accounting treatments,
whereas they tend to adopt competing positions in
relation to issues of disclosure. To the extent that
agreement among two or more groups results in a
multiplier e?ect, whereby the combined in¯uence
of the coalition is signi®cantly greater than the
in¯uence of a single group acting alone (Benve-
niste, 1972; Dyckman, 1988), Puro's ®ndings sug-
gest that constituent in¯uence may also vary as a
function of the nature of the ®nancial reporting
issue under debate. To date, however, this issue
remains largely unexplored in the empirical litera-
ture.
Several studies (e.g. Brown, 1981; Hussein &
Ketz, 1980; Newman, 1981; Puro, 1985) have used
lobbying data to investigate the impact of pressure
groups on the policy decisions taken by standard-
setting agencies following allegations in the Met-
calf report (US Congress, 1976) that the FASB
was unduly in¯uenced in its policy-making process
by certain sectors of society (primarily the large
audit ®rms and their clients). However, the evi-
dence reported in these studies suggests that the
FASB's policy decisions are not consistently
dominated by any single group or coalition of
groups. Instead, results suggest the in¯uence of
many centers of power. Indeed, Hussein and Ketz
(1991) conclude that the FASB's deliberations are
characterised by a system of `power equivalency',
whereby the structural and relational contracts
make it impossible for any single agent or coali-
tion of agents to dominate the process. Similar
conclusions are presented by Hope and Gray
(1982) in relation to the UK standard-setting pro-
cess. While this body of work provides important
evidence concerning the political nature of
accounting regulation, Hussein and Ketz (1991)
argue that power and in¯uence are determined, at
least in part, by the prevailing regulatory struc-
ture. Consequently, di?erent regulatory structures
may be associated with di?erent political out-
comes. A broader understanding of the politics of
accounting regulation therefore demands the ana-
lysis of alternative regulatory systems.
3. Institutional background
Traditionally, decisions on accounting rules
have been viewed in Germany not only as a tech-
nical matter on which a group of accounting
experts should be competent but also as an issue
of public policy having a relatively broad social
basis (Ordelheide & Pfa?, 1994, p. 82). The
resulting law-based system of ®nancial reporting is
primarily dependent on commercial law, with
strong connections to tax law,
7
and this is still the
case following the recent reform which established
a body to develop German accounting standards.
Until 1985, the Stock Corporation Law (Aktien-
gesetz) of 1965 represented the primary source of
7
The strong association between ®nancial reporting and tax
accounting is the result of the Mageblichkeitsprinzip which
requires that the statutory accounts form the authoritative
basis for taxation and the `reverse authoritativeness principle'
(umgekehrte Mageblichkeitsprinzip), which requires that tax
rules must be observed in commercial accounts in order to
bene®t from tax incentives.
S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98 83
accounting requirements for listed companies,
supplemented by provisions in the Commercial
Code (Handelsgesetzbuch) and income tax law.
8
Following the transformation of the Fourth (and
Seventh) Directive into German commercial law
in 1985, the Accounting Directives Law became
the legal basis for ®nancial reporting in Germany.
The Accounting Directives Law amended the
Commercial Code, the latter containing the gen-
eral accounting and auditing rules applicable to all
companies, together with a special section relating
to stock corporations and limited liability compa-
nies.
While the Accounting Directives Law codi®ed
many important accounting principles and stan-
dards, the German legal provisions are not
exhaustive. For example, foreign currency trans-
lation and accounting for leasing and government
grants are not regulated explicitly in the Commer-
cial Code. In the absence of such regulations,
companies have resorted to tax law and tax court
rulings for authoritative and legally-binding inter-
pretations of the Commercial Code. Additionally,
certain non-authoritative legal interpretations are
combined with authoritative rulings to form the
`Correct Accounting Principles' (GrundsaÈtze ord-
nungsmaÈiger BuchfuÈhrung or GoB) explicitly
referred to in the Commercial Code.
9
This combi-
nation of legislative rules and legal interpretations
provides a distinctive hierarchical structure to
accounting regulation in Germany (Ordelheide,
1999), creating two primary mechanisms through
which parties have been able to shape ®nancial
reporting practice. First, agents have in¯uenced
commercial law directly by lobbying the legislative
body during the rule-development process. Such
participation by representative bodies in law-making
has constituted an important part of the due pro-
cess of accounting regulation in Germany, actively
encouraged by the legislature. Second, agents can
contribute to the market for legal interpretations.
Law commentaries and interpretative articles
which seek to clarify legal requirements have been
produced regularly by barristers, judges, auditors,
experts from business and the public sector, and
academics.
10
Until now, the complete system of
legal interpretations has been collected together
and published in law commentaries and ®nancial
accounting handbooks [e.g. Adler-DuÈring-
Schmaltz (Forster, 1997)].
Re¯ecting the codi®ed approach to regulation in
Germany, the accounting rule-development pro-
cess was coordinated until recently by the Ministry
of Justice. In contrast with the Anglo±American
approach, only a relatively minor role was ascribed
to the audit profession, for which the traditional
emphasis was concerned with clients' compliance
with the law rather than with the development of
binding accounting principles and procedures.
11
While accounting in Germany has been in¯uenced
by a wide range of other parties, including private
research institutes, academic accountants, pre-
parer groups organised by the Confederation of
German Industries, and the audit profession, the
®nancial reporting regulations examined in this
study were drafted by lawyers and administrators
in the Ministries of Finance and Justice.
8
No additional speci®c regulations existed for alternative
corporate forms such as limited liability companies or for non-
corporations such as partnerships and sole traders.
9
The GoB has been interpreted as meaning `those principles
which are not comprehensively codi®ed but which, by applica-
tion in speci®c cases, lead to a correct accounting treatment by
reference to the objectives of ®nancial statements. They can be
determined deductively by making full use of statute and case
law, accounting theory, pronouncements of the Institute of
Auditors, as well as accounting practice' (Brooks & Mertin,
1986).
10
For example, the Institute of Auditors (Institut der Wirt-
schaftspruÈfer) regularly publishes opinions on legislative pro-
nouncements, laws, and contemporary ®nancial accounting and
auditing issues. While these opinions are not legally binding,
the risk of an auditor being accused of professional malpractice
is higher if an auditor has certi®ed ®nancial statements that are
in con¯ict with such recommendations (Ordelheide, 1999).
11
Whereas in the US, performance of the audit function is
not a prerequisite for aliation to the AICPA, membership of
the Institute of Auditors is con®ned to practising auditors only.
This distinction re¯ects fundamental di?erences in the role
attributed to the profession in each country: in the US, the
audit profession assumes signi®cant responsibility for the
development of accounting standards; by contrast, the German
audit profession operates within a more tightly de®ned jur-
isdiction, where the emphasis rests ®rmly on performance of
the statutory audit function (Vieten, 1995).
84 S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98
3.1. European accounting harmonisation and the
Fourth Directive
An objective of the European Union is the con-
vergence not only of economic conditions but also
of member states' respective national laws to the
extent required for the common market to func-
tion in an orderly manner. To achieve this aim, a
programme of legal harmonisation has been
implemented involving the development of a series
of Company Law Directives. When the European
Commission has obtained agreement on a set of
proposals relating to the harmonisation of a par-
ticular topic, it places a Draft Directive before the
Council of Ministers. If the Directive is adopted,
governments of the member states have a speci®ed
period in which to enact the legislation and incor-
porate the Directives' provisions into their
national law. To obtain agreement at a European
level, it is usual for a Directive to contain a range
of national options, a?ording member states sig-
ni®cant ¯exibility when enacting its provisions
into national law. It is the choice among these
national options with respect to the Fourth
Directive that provides the background to our
analysis of accounting regulation in Germany.
Integration of the provisions of the Fourth
Directive into German law began with the pub-
lication of the legislature's transformation recom-
mendations. Comments on this document were
invited from all interested parties in Germany and
used by the Ministries of Justice and Finance to
produce a pre-draft law. Five subsequent draft
laws were published prior to enactment of the
Accounting Directives Law in December 1985 and
comments were received following each stage.
While 27 organisations were ocially invited to
comment during the drafting stage, only industry,
academia and the audit profession engaged in
formal lobbying on a material number of
accounting issues and on successive drafts of the
law. The restriction of lobbying activity to this
limited set of ®nancial reporting experts re¯ects, at
least in part, the highly complex nature of German
commercial law and the resulting barriers to entry
this creates. The absence of formal lobbying
activity by user associations may be partially
explained by the decision of the German banks
(the primary users of ®nancial reporting informa-
tion) to adopt a preparer position for lobbying pur-
poses. The greater involvement of preparers relative
to users in the lobbying process is also consistent
with both Sutton's (1984) theoretical model of lob-
bying activity and with Tandy and Wilburn's (1992)
study of constituent lobbying of the FASB.
Lobbying was conducted both publicly in the
form of published commentaries and submissions
to ocial hearings, and privately in the form of
unpublished letters and informal discussions.
Re¯ecting both the collectivist nature of German
society (Power, 1997) and cost eciencies asso-
ciated with organised lobbying activity (Sutton,
1984), all formal proposals were issued through
representative agencies. The preparer viewpoint
was represented by the Association for Finance
and Management (Gesellschaft fuÈr Finanzwis-
senschaft in der UnternehmensfuÈhrung) and the
Combined Associations of German Industry
(SpitzenverbaÈnde der deutschen Wirtschaft).
12
Pro-
posals from each association were developed by a
series of working parties, membership of which
was drawn from the corporate sector, and pub-
lished in the journal Der Betrieb.
13
The audit pro-
fession's viewpoint was jointly represented by the
Institute of Auditors (Institut der Wirtschafts-
pruÈfer) and the Chamber of Auditors (Wirtschaft-
spruÈferkammer). Proposals were drawn up by
working groups, issued as professional opinions,
and published at the outset in Der Betrieb and
later in Die WirtschaftspruÈfung.
14
Finally, the
12
The former is a private organisation of senior accounting
and ®nance managers from the large German corporations, while
the latter is an umbrella organisation for various business asso-
ciations, including representatives from the retail, banking, and
small ®rm sectors. For the purpose of enacting the Fourth Direc-
tive, the following groups were involved: the Federation of Ger-
man Industries; the Association of German Industry and
Commerce; the Federation of German Wholesale and Foreign
Trade; and the Federation of German Banks (Ordelheide, 1999).
13
Der Betrieb Vol. 31 (1978) pp. 1464±1466; Vol. 32 (1979)
pp. 1±8 and 1093±1097; Vol. 33 (1980) pp. 1±12; Vol. 34 (1981)
pp. 2448±2462; Vol. 37 (1984) pp. 1±12; Vol. 39 (1986) pp.
1985±1988 and 2553±2556.
14
Der Betrieb: Vol. 32 (1979) pp. 1237±1242 and 1296±1300;
Die WirtschaftspruÈfung: Vol. 33 (1980) pp. 501±523; Vol. 34
(1981) pp. 609±621; Vol. 37 (1984) pp. 125±140; Vol. 38 (1985)
pp. 349±353; Vol 39 (1986) pp. 537±553.
S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98 85
academic position was represented by the
Accounting Committee of the Association of
German Business Studies Academics (Kommission
Rechnungswesen im Verband der Hochschullehrer
fuÈr Betriebswirtschaft). Formed in 1977 speci®cally
to facilitate academic participation in the process
of transforming European Company Law Direc-
tives into German law, the Committee appointed
two working groups to debate the draft legislation.
The results of these debates were then submitted
to the Committee and, after subsequent modi®ca-
tion, published as ocial opinions in Die
Betriebswirtschaft.
15
4. Data and sample
Empirical tests are conducted using commen-
taries published by the representative bodies iden-
ti®ed in Section 3.1 during the draft law stages of
the transformation process.
16
For all accounting
issues examined, both industry associations made
identical proposals. To simplify the empirical
analysis and increase model parsimony, we aggre-
gate these two associations to form a single
industry group, denoted IND. Similarly, because
the Institute of Auditors and the Chamber of
Auditors acted in unison throughout the transfor-
mation process (Ordelheide, 1999), we aggregate
these two parties to form a single auditor group,
denoted AUD. Proposals made by the Accounting
Committee of the Association of German Business
Studies Academics are denoted ACA.
For each lobby group (IND, AUD and ACA),
all publicly available comments were surveyed to
identify each formal proposal. A total of 169
separate issues covering aspects of valuation,
recognition, disclosure and ®nancial statement
format were identi®ed on which at least one of the
three groups commented formally.
17
The nature of
these proposals was such that they clearly indi-
cated (i) the speci®c legal provision under discus-
sion and (ii) the lobbyist's proposed accounting
treatment.
Consistent with much of the extant literature on
accounting regulation, we adopt a pluralist model
as a means of operationalising the concept of
political power.
18
Within this framework, power is
assessed by distinguishing those lobbyists whose
proposals are ultimately adopted (i.e. successful)
from those whose proposals are rejected (i.e.
unsuccessful). Lobbyists with the highest propor-
tion of `successes' are then considered to exhibit
the greatest political in¯uence while those with the
lowest proportion are assumed to be the least
in¯uential. For the 169 issues on which at least
one of the three groups proposed a change to the
law, the proposals were compared with the legal
requirement contained in the ®nal text of the
Accounting Directives Law. For each constituent, a
recommendation corresponding to the accounting
treatment ultimately required by the Accounting
Directives Law is categorised as `successful'.
Alternatively, a recommendation which fails to
correspond with the treatment required by the
Accounting Directives Law is assumed to have
been `unsuccessful'. Finally, in the event that a
lobbyist makes no formal (i.e. publicly available)
proposal on a particular issue, an absence of lob-
bying activity is inferred.
The procedure described above provides a rela-
tively objective measure of the extent to which
lobbyists' proposals are ultimately realised by the
legislature in commercial law. Admittedly, as the
approach relies on ®nal outcomes to generate the
observable counts of `successes', it does not allow
for a detailed investigation of strategic lobbying
behaviour in a multi-period framework where
15
Die Betriebswirtschaft: Vol. 38 (1978) pp. 453±455; Vol. 39
(1979) pp. 1±70; Vol. 40 (1980) pp. 589±597; Vol. 43 (1983) pp.
5±15.
16
Consistent with prior work, this study focuses on obser-
vable lobbying behaviour. To the extent that private lobbying
was either more in¯uential than public lobbying, or di?ered in
respect to the nature of the preferences expressed, the results of
this study should be interpreted with caution.
17
Of the 169 points of accounting law identi®ed, 46 related
to issues of valuation, 46 to issues of recognition, 36 to issues of
®nancial statement format and 41 to issues of disclosure.
18
The pluralist model of power has been the subject of sig-
ni®cant criticism [see Lukes (1974) and Walker and Robinson
(1993) for a discussion]. However, in defence of the pluralist model,
Lukes (1974) acknowledges that in many cases it represents the
most useful method of operationalizing and analysing power.
86 S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98
there may be variation in the level of interest with
respect to individual points of law. Nevertheless,
the database provides a unique setting in which to
estimate lobbyist in¯uence and to assess the over-
all impact on such in¯uence of agreement and
disagreement between lobbyists. Indeed, an
important aspect of constituent lobbying is that
success depends not only on each interest group's
own actions but also on the degree of consensus
and con¯ict among the participant groups (Ben-
veniste, 1972). An explicit consideration of the
interaction between lobbyists' stated preferences
therefore seems appropriate. A limitation of our
data, however, is that when two or more parties
submit similar proposals, we are unable to distin-
guish between those cases which arise as a result of
an explicit coalition and those cases where it
represents two independently determined posi-
tions. Consequently, when modeling lobbyists'
interactions, we de®ne `agreement' simply as the
stylised fact that two equivalent proposals have
been made, while `disagreement' is de®ned as the
presence of a proposal and a counterproposal.
Further, an implication of the participation of
three lobby groups in the rule-development pro-
cess is the occurrence of situations where two of
the interest groups make the same proposal while
the third makes a counterproposal. We model
such situations as a set of two-way interactions
comprising one case of agreement (e.g. between
IND and AUD) and two separate cases of dis-
agreement (between IND and ACA and between
AUD and ACA). Finally, an additional character-
istic of the data is that in the case of disagreement,
one of the proposals always becomes law. That is,
con¯ict amongst lobby groups always leads to suc-
cess for one or more of the parties involved. On the
other hand, uncontested proposals made by one,
two or three parties may fail to become law. Here,
the lobbying process in Germany is modelled on the
basis of these stylised facts.
Table 1 presents the observed frequencies of
success, failure and non-participation, cross-clas-
si®ed by lobby group. The three lobby groups
made a total of 269 proposals, 148 of which were
successful. The draft law was amended by the
regulator with respect to 63% of the issues on
which comments were made (i.e. 106 issues out of
169). When only one of the lobbying groups made a
proposal, less than half (43 of 97) of the respective
regulations were revised. In contrast, the regulator
changed the legal text in 63 of the 72 issues which
attracted proposals from more than one lobby
group. In summary, the conditional probability that
a proposal would be successful if made by just one
lobbying group was only 0.443, but it was as high as
0.875 if made by more than one lobbying group.
A fuller understanding of these success prob-
abilities requires an approach which controls not
only for conditioning factors such as the di?er-
ences between the parties in the odds of making a
proposal in the ®rst place but also for the like-
lihood of agreement or disagreement over the
nature of the proposed amendment to the law. In
the latter case, these probabilities can be estimated
either for the general case (i.e. that there exists
agreement or disagreement, or both, over an issue)
or the speci®c (i.e. that two parties in particular
agree or disagree). In this way, an idea can be gained
not only of the odds that a proposal will succeed but
also of how such odds vary across the parties when
they agree or disagree. In this paper, we refer to the
odds of making a proposal as and to the odds that
it will succeed as !, whilst agreement is denoted by
and disagreement by '. A more detailed discussion
of model development in this context is given below,
followed by the empirical results.
5. Model speci®cation
The count in each cell of Table 1 follows a mul-
tinominal frequency distribution with unknown
probabilities (Francis, Green & Payne, 1993). A
statistical model to estimate the probability struc-
ture of this data may be written in the form of a
generalised linear model (Nelder & Wedderburn,
1972). For situations in which the response variate
is a vector of frequencies from a multi-way con-
tingency table and the explanatory variables are
categorical (i.e. no proposal, proposal accepted, or
proposal rejected), McCullagh and Nelder (1989)
demonstrate that the appropriate formulation is a
log-linear model with a log link function and a
Poisson distributed error. Under this scheme, the
count in a given cell may be expressed as a function
S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98 87
of the main e?ects (subscripted IND, AUD, and
ACA) and the associated interactions. As well as
this approach representing the most appropriate
modeling procedure given the nature of the data,
it a?ords the additional advantage of facilitating
an explicit analysis of the way in which individual
lobbyists' preferences interact in determining Ger-
man accounting law.
19
The standard approach in log-linear modeling
assumes that hierarchical models are to be ®tted,
such that the existence of a higher-order interac-
tion necessarily implies the inclusion of all lower
order terms marginal to it (Francis et al., 1993, p.
249). We therefore ®t a number of levels of the
model to account for the extent of agreement or
disagreement, the relative in¯uence of the indivi-
dual lobbying organisations, the odds of success
when such consensus and con¯ict exists, and
®nally, the e?ect on those odds of the mix of par-
ties involved.
Table 1
Constituent lobbying activity in relation to the transformation of the Fourth Directive into German commercial law, cross-classi®ed
by constituent group and lobbying outcome
a,b
Auditors Industry
No proposal Accepted Rejected Total
Academics=No proposal No proposal 0 22 22 44
Accepted 13 10 0 23
Rejected 11 3 5 19
Total 24 35 27 86
Academics=Accepted No proposal 8 5 6 19
Accepted 3 6 3 12
Rejected 0 1 0 1
Total 11 12 9 32
Academics=Rejected No proposal 21 10 0 31
Accepted 0 8 0 8
Rejected 2 8 2 12
Total 23 26 2 51
Total
Academics=Total No proposal 29 37 28 94
Accepted 16 24 3 43
Rejected 12 12 7 32
Total 58 73 38 169
a
The sample consists of 169 separate accounting issues on which at least one constituent group (i.e., industry, auditors, or aca-
demics) lobbied the German legislature. For each accounting issue, lobbyists' proposals were compared with the ultimate legal
requirement contained in the Accounting Directives Law. Proposals corresponding to the Accounting Directives Law requirement
were considered successful and coded `accepted', while proposals which did not correspond with the Accounting Directives Law
requirement were considered unsuccessful and coded `rejected'. In the event that a lobbyist made no formal (i.e., publicly available)
proposal on a particular issue, an absence of lobbying activity was inferred.
b
The industry constituency was represented by two organisations: The Association for Finance and Management and the Combined
Associations of German Industry. Since both organisations adopted the same lobbying position, for all 169 cases examined we aggregate
these two associations to form a single industry group. The auditor constituency was jointly represented by the Institute of Auditors and the
Chamber of Auditors on all 169 cases examined. These two parties are again aggregated to form a single auditor group. Finally, the aca-
demic constituency was represented by the Accounting Committee of the Association of German Business Studies Academics.
19
Since our interest is restricted to those issues on which at
least one of the three lobbyists made a proposal, the corner cell
IND=ACA=AUD=0 is treated as a structural zero (Francis
et al., 1993, p. 272).
88 S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98
5.1. Conditional independence
The lowest order log-linear model is that of
complete independence, where the expected value
is equal in each cross-classi®ed cell, except for the
structural zero in the corner cell where none of the
lobbyists makes a proposal. However, since this
simplistic model is unlikely to provide any mate-
rial insights into the structure of our data, we
begin by ®tting a model where the probabilities of
making a proposal are allowed to vary across the
constituent groups. We refer to this initial model
as the model of conditional independence, since
the estimates of expected acceptances and rejec-
tions are conditional on a proposal having been
made but are independent of the relative in¯uence
of the lobbyists and any possible interactions
between them. The model of conditional indepen-
dence therefore represents the null against which
each group's relative in¯uence, and the impact of
agreement and disagreement, may be assessed. For
the model of conditional independence, the log of
the observed count may be generalised as
log C
0
?
X
i

i
; ?1?
where:
C
/
=a vector of observations representing the
observed counts, C, in Table 1, adjusted for the
mean e?ect;

i
=the log-odds that lobby group i makes a
proposal;
i=the industry (IND), auditor (AUD), or aca-
demic (ACA) lobby groups.
For each lobby group i, is de®ned as a two-
level factor, taking the value one if i made a pro-
posal and zero otherwise.
5.2. Relative in¯uence
This model evaluates the unconditional odds of
success for each of the lobbyists. In this case, the
log of the mean-adjusted observed count is
log C
0
?
X
i

i
?
X
i
!
i
; ?2?
where:
!
i
=the log-odds that a proposal made by lobby
group i will be accepted.
All other variables are as previously de®ned
and, for each lobby group i, !
i
is de®ned as a two-
level factor, taking the value one if i's proposal
was successful and zero if it was rejected.
5.3. Agreement and disagreement
Adding the main e?ects of agreement and dis-
agreement to model (2) generates the following
model:
log C
0
?
X
i

i
?
X
i
!
i
? ? ' ?3?
where:
=the overall log-odds that two parties will
make the same proposal, de®ned as a two-level
factor taking the value one if any two lobbyists
make the same proposal, and zero otherwise;
'=the overall log-odds that two proposals dif-
fer, de®ned as a two-level factor taking the value
one if any two lobbyists make di?erent proposals
on a given accounting issue, and zero otherwise.
While model (3) is expressed as a main e?ects
model, interaction terms can also be included to
account for those cases where two lobbyists are in
agreement while the third seeks a di?erent out-
come (i.e. '), and also for the log odds that an
identical proposal put forward by more than one
party will lead to a change in the law (i.e. !). At
the general level, the interaction between dis-
agreement and success (!') remains constant in
this study, since for each case involving con¯icting
proposals, one of the proposals always became
law.
5.4. Agreement and disagreement with lobby group i
The generalised models presented above can be
rewritten to allow for the relative odds of agree-
ment or disagreement with the ith lobby group, as
follows:
log C
0
?
X
i

i
?
X
i
!
i
?
X
i

i
?
X
i
'
i
?4:1?
S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98 89
where:

i
=a two-level factor taking the value one when a
proposal put forward by lobby group i is in agree-
ment with any other proposal, and zero otherwise;
'
i
=a two-level factor taking the value one when
a proposal put forward by lobby group i is in dis-
agreement with any other proposal, and zero
otherwise.
Model (4.1) may also be extended by including
interaction terms to account for the relative odds
of success when there is agreement and disagree-
ment. This extended model is written as
log C
0
?
X
i

i
?
X
i
!
i

i
?
X
i
!
i
'
i
?4:2?
where:
!
i

i
=the sum of the main e?ects that a propo-
sal made by lobbyist i will be successful and that a
similar proposal to lobbyist i's will be made by
another lobby group, plus the interaction e?ect on
the success of a proposal by lobbyist i when a
similar proposal has been made by another lob-
byist (i.e. !
i

i
= !
i
÷
i
÷ !
i

i
);
!
i
'
i
=the sum of main and interaction e?ects of
a successful counterproposal from lobbyist i (i.e.
!
i
'
i
= !
i
÷ '
i
÷ !
i
'
i
).
5.5. Agreement and disagreement between lobby
groups i and j
The ®nal model allows for each of the possible
two-way cases of agreement and disagreement
between each pair of lobbyists i and j, where i ,= j.
This model may be speci®ed as
log C
0
?
X
i

i
?
X
i
!
i
?
X
i
X
j

ij
?
X
i
X
j
'
ij
; ?5:1?
where:

ij
=a two-level factor taking the value one
when i and j are in agreement, and zero otherwise;
'
ij
=a two-level factor taking the value one when i
and j are in disagreement, and zero otherwise.
Finally, the relative odds of succeeding in chan-
ging the draft law when there is either agreement or
disagreement between i and j may now be added as
interactions. The odds of success for lobbyist i are
given by !
i

ij
in the case where the proposals put
forward by i and j are the same and by !
i
'
ij
where
they di?er, giving the following model:
log C
0
?
X
i

i
?
X
i
X
j
!
i

ij
?
X
i
X
j
!
i
'
ij
?5:2?
where:
!
i

ij
=the sum of the main e?ects that a propo-
sal made by lobbyist i will be successful and that a
similar proposal to lobbyist i's will be made by
lobby group j, plus the interaction e?ect on the
success of a proposal by lobbyist i when groups i
and j put forward the same proposal;
!
i
'
ij
=the sum of the main e?ects that a propo-
sal made by lobbyist i will be successful and that a
counter proposal to lobbyist i's will be made by
lobby group j, plus the interaction e?ect on the
success of a proposal by lobbyist i when groups i
and j make con¯icting proposals.
Since the relative odds of success when i and j
are in agreement are necessarily identical for the
two parties involved, the interaction terms !
i

ij
and !
j

ij
are equal. In contrast, since the relative
odds that lobby group i will win a contest with j
are the inverse of the relative odds that lobby
group j will win, !
i
'
ij
is equivalent to ÷ !
j
'
ij
.
A summary of the hierarchy of main e?ects
models and their associated factor structure is
presented in Table 2.
5.6. Goodness of ®t
The appropriate measure of goodness of ®t for a
log-linear model with a Poisson error is the
deviance (D). The deviance compares the max-
imised log-likelihood for the nth ®tted model (m
n
)
and fully saturated model and provides a general
test of the adequacy of the ®tted model.
20
20
A signi®cant value for the deviance indicates that a sig-
ni®cant term has been omitted from the model. For the Normal
model, the deviance is equal to the residual sum of squares,
while in the case of the Poisson model it has an interpretation
similar to the Pearson goodness of ®t statistic (Francis et al.,
1993, p. 275).
90 S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98
Accordingly, the relative goodness of ®t of two
nested models (m
1
and m
2
) is measured as the
di?erence between their associated deviances
(D
1
÷ D
2
) where m
1
is nested in m
2
. D
1
÷ D
2
is
asymptotically chi-squared distributed with
d
1
÷ d
2
, degrees of freedom. A signi®cant value for
D
1
÷ D
2
indicates that m
2
is a better model than
m
1
or equivalently, that the terms omitted from m
1
are signi®cant. The appropriate F-ratio compares
the mean change in deviance (D
1
÷ D
2
)=(d
1
÷ d
2
)
and the mean residual deviance (D
2
=d
2
), and is
distributed F ~ d
1
÷ d
2
( ); d
2
[ [.
6. Empirical results
Model (1) re¯ects the propensity for a particular
lobbyist to submit a proposal. The coecients
21
Table 2
The hierarchy of main e?ects models of the politics of accounting regulation in Germany
a
Conditional independence
b
log C
/
=
P
i

i
(1)
Relative in¯uence log C
/
=
P
i

i
÷
P
i
!
i
(2)
Agreement and disagreement log C
/
=
P
i

i
÷
P
i
!
i
÷ ÷ ' (3)
Agreement and disagreement with i log C
/
=
P
i

i
÷
P
i
!
i
÷
P
i

i
÷
P
i
'
i
(4.1)
Agreement and disagreement between i and j log C
/
=
P
i

i
÷
P
i
!
i
÷
P
i
P
j

ij
÷
P
i
P
j
'
ij
(5.1)
where:
C
/
=a vector of observations representing the observed counts in Table 1, less a constant term;

i
=the odds of a proposal being made by lobby group i;
!
i
=the odds of a proposal made by i succeeding;
i=industry (IND), auditors (AUD) or academics (ACA);
=the odds that, for a given issue, two lobby groups make the same proposal;
'=the odds that, for a given issue, two lobby groups make di?erent proposals;

i
=the odds of agreement with i;
'
i
=the odds of disagreement with i;

ij
=the odds that i and j make the same proposal, where i ,= j;
'
ij
=the odds that i and j make con¯icting proposals, where i ,= j;
a
The model is a log-linear model with a log link function and a Poisson distributed error. The dependent variable is a vector
representing the log of the observed counts in Table 1, while the explanatory variables are a series of n-level factors. All estimated
coecients are log-odds ratios (i.e., the exponent of the regression coecient is the relative odds of the count falling into one of the
possible categories). A regression coecient of zero therefore implies odds equal to one, while a positive (negative) coecient implies
odds greater (less) than one.
b
The model of conditional independence allows the probability of a proposal being made to vary across the three constituent
groups while holding constant their relative in¯uence on the decisions of the legislature. The model of relative in¯uence evaluates
the odds of success for each of the three lobbyists, conditional on a proposal being made. The model of agreement and dis-
agreement extends the model of relative in¯uence to account for the main e?ects of agreement and disagreement. Generalising
further, model (4.1) accounts for the relative odds of agreement or disagreement with lobby group i. This model is further
extended by adding the interaction terms !
i

i
and !
i
'
i
which, respectively, capture i's relative odds of success given agreement
or disagreement with either of the remaining constituent groups. Finally, model (5.1) accounts for all possible two-way cases of
agreement and disagreement between each pair of lobbyists i and j. This model is extended by adding the interactions !
i

ij
and
!
i
'
ij
which, respectively, capture i's relative odds of success in the event of agreement or disagreement between lobbyists i and j.
21
The
i
estimates frommodel (1), representing the log-odds that
lobby group i will make a proposal, were as follows:
IND
= 3:034
(t = 2:98);
AUD
= 2:314 (t = 7:62),
ACA
= 2:479 (t = 6:61). T-
tests for the signi®cance of individual parameter estimates are based
on the assumption of a symmetric likelihood function. Reported t-
statistics represent only a guide to the signi®cance of a particular
parameter with respect to a null of zero, rather than indicating the
relative likelihood, and must therefore be interpreted with caution.
S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98 91
from the model of conditional independence pro-
vide evidence of industry's high submission rate
and the low level of public involvement by the
German audit profession in the rule-development
process: industry submitted proposals on 111 of
the 169 issues examined (66% proposal rate)
compared with 75 proposals submitted by the
audit profession (44% proposal rate). Further,
German accounting academics are seen to display
a higher propensity for public lobbying than the
audit profession, in contrast to the ®ndings docu-
mented for the US and similar regimes. While this
result may partly re¯ect the German auditor's
traditional concern for the application of ®nancial
reporting rules, rather than for the development of
these rules (Vieten, 1995), the high level of public
lobbying activity observed for German accounting
academics may result from the evolution of the
legal perspective as the dominant research
approach to date in Germany, whereby account-
ing researchers have been actively involved in a
form of jurisprudence, by interpreting existing
laws where doubts have arisen.
22
Table 3 presents tests of the change in deviance
for the hierarchy of models discussed above, along
with the estimated coecients. The change in
deviance associated with ®tting model (2), which
provides evidence as to the relative power of the
three lobby groups, is not signi®cant at conven-
tional levels (p=0.241). However, it is notable that
the odds of success are signi®cantly greater than
one for industry (log-odds=0.652; odds=1.92 : 1;
t=3.27) and signi®cantly less than one for the
academic experts (log-odds=÷0.466; odds=1 :
1.59; t=÷2.07). These ®ndings re¯ect the fact that
industry succeeded in 73 out of 111 proposals
submitted (i.e. 63%) whereas the academics suc-
ceeded in only 32 of their 83 proposals (i.e. 38%).
The insigni®cant coecient estimate on the !
AUD
term indicates that the odds of success for the
audit profession were approximately equal to one.
Examination of the results documented in panel B
of Table 3 con®rms that, after accounting for the
general e?ects of agreement () and disagreement
('), industry's odds remain signi®cantly greater than
one (t = 3:31), while the odds of success for the
auditor constituency remain equal to one and those
for the academic constituency are again signi®cantly
less than one. In sum, the ®ndings for models (2)
and (3) suggest that German industry exerted
most in¯uence over the legislature during the
transformation of the Fourth Directive into com-
mercial law. In contrast, the in¯uence of the aca-
demic community appears to have been relatively
low, despite their active participation.
Extension of model (3) to include the general
interaction term !, which accounts for the log-
odds of success given agreement between any two
lobby groups, provides further insights into the
politics of accounting rule-development in Ger-
many. The positive and signi®cant coecient on
! (t = 2:65) indicates that the overall odds of
success are increased by more than 2:1 when any
two groups take similar positions with respect to a
particular ®nancial reporting issue, reinforcing the
view that the responsiveness of the German legis-
lature is a positive function of consensus among
lobby groups. The ®ndings for Germany that
agreement among constituent groups has been an
important factor in¯uencing the decision of the
legislature is consistent with the views expressed
with respect to the US by Dyckman (1988) and
Horngren (1972), both of whom argue that the
search for political consensus underlies many of
the ®nancial reporting solutions proposed by the
FASB.
Models (4.1) and (4.2) account for the relative
odds of agreement or disagreement with the ith
lobby group. Panel C of Table 3 reports a sig-
ni®cant and positive estimate of
AUD
(t = 3:14)
under the main e?ects model (4.1), suggesting a
high likelihood that lobbying by auditors would
have been accompanied by consensus. This e?ect
alone, however, does not change the measures of
relative in¯uence, with industry continuing to
appear more successful than the other groups.
Model (4.2) extends model (4.1) by including
interaction terms which account for the relative
odds of success when there is agreement or
22
In serving the market for interpretations, some academics
will derive support from audit ®rms or direct from industry. In
this respect, they are not necessarily disinterested individuals
who serve merely to interpret the law. However, participation
in the process of redrafting the law is through the auspices of
an academic body which acts on behalf of the German aca-
demic profession as a whole, not individual academics.
92 S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98
Table 3
The relative in¯uence of constituent lobby groups in Germany (t-statistics in parentheses)
a
Panel A: Relative in¯uence
Relative in¯uence of lobby group i
!
IND
!
AUD
!
ACA
Deviance ÁDeviance
b
F
c
p-value
Model (2) 0.652 0.295 ÷0.466 71.747 17.234
(1, 2)
1.521 0.241
(3.27) (1.27) (÷2.07)
Panel B: Agreement and disagreement
d
Relative in¯uence of lobby group i Agreement () and disagreement (')
!
IND
!
AUD
!
ACA
' Deviance ÁDeviance F
c
p-value
Model (3) 0.665 0.279 ÷0.518 ÷1.380 ÷1.708 53.515 18.232
(2, 3)
1.817 0.185
(3.31) (1.19) (÷2.25) (÷2.54) (÷3.55)
Panel C: Agreement and disagreement with each lobby group
Relative in¯uence of lobby group i Agreement () and disagreement (') with i
i=IND i = AUD i=ACA
!
IND
!
AUD
!
ACA

i
'
i

i
'
i

i
'
i
Deviance ÁDeviance F
c
p-value
Model (4.1): Main e?ects 0.593 0.238 ÷0.411 0.613 1.129 1.243 ÷0.675 0.393 0.391 50.991 20.756
(2, 4.1)
0.882 0.534
(2.92) (0.99) (÷1.78) (1.57) (1.75) (3.14) (÷1.75) (1.05) (0.82)
Model (4.2): Main e?ects ÷0.117 ÷0.054 ÷0.632 ÷0.199 0.795 1.032 ÷0.745 0.206 0.394 39.943 10.72
(4.1, 4.2)
0.323 0.905
(÷0.34) (÷0.13) (÷1.63) (÷0.38) (1.02) (1.98) (÷1.38) (0.46) (0.67)
:Interaction e?ects !
i

i
!
i
'
i
!
i

i
!
I
'
i
!
I

i
!
i
'
i
1.242 0.505 0.292 0.038 0.390 ÷0.052
(2.55) (0.75) (0.53) (0.50) (0.75) (÷0.07)
Panel D: Agreement and disagreement between lobby groups i and j
Relative in¯uence of lobby group i Agreement () and disagreement (') between i and j
i, j=IND, AUD i, j=AUD, ACA i, j=ACA, IND
!
IND
!
AUD
!
ACA

i
'
i

i
'
i

i
'
i
Deviance ÁDeviance F
c
p-value
Model (5.1): Main e?ects 0.548 0.276 ÷0.398 2.710 1.510 1.971 0.325 1.394 2.686 25.562 46.185
(2, 5.1)
3.915 0.019
(4.29) (1.09) (÷1.64) (4.45) (2.25) (4.29) (0.66) (2.27) (4.35)
Model (5.2): Main e?ects 0.103 0.240 ÷0.753 2.480 1.920 1.749 0.586 ÷0.010 2.461 12.899 12.663
(5.1, 5.2)
1.145 0.426
(0.35) (0.62) (÷1.99) (3.16) (2.87) (2.95) (0.01) (0.01) (3.85)
: Interaction e?ects !
i

ij
!
i
'
ij
!
i

ij
!
i
'
ij
!
I

ij
!
i
'
ij
0.166 ÷1.353 0.150 ÷1.392 2.095 0.353
(0.25) (÷1.53) (0.21) (÷1.14) (2.17) (0.55)
a
Reported coecients are from a log-linear model with a log link function and a Poisson distributed error. The dependent variable is the log of the counts in the cells of Table 1, adjusted for a mean e?ect.
Coecient estimates are log-odds ratios: a regression coecient of zero implies odds equal to one, while a positive (negative) coecient implies odds greater (less) than one. Subscripts IND, AUD and ACA indicate the
industry, auditor and academic lobby groups, respectively. Models are described in Table 2.
b
The subscripts in parentheses indicate the nested models for which the change in deviance is reported. The deviance for the model of conditional independence (model 1) is 88.981. Note that the
i
coecients in
model 1 are also included in all subsequent estimations, but for parsimony are not recorded.
c
The F-statistic represents a test of the relative goodness of ®t of two nested models. The relative goodness of ®t is measured as the di?erence in the associated deviances (ÁDeviance) of two models, x and y, where
x is nested in y. The change in deviance is asymptotically F-distributed. A signi®cant F-statistic indicates that y is a better ®tting model than x, or equivalently, that the terms omitted from x are signi®cant.
d
The model of agreement and disagreement may be modi®ed to include interaction terms measuring the general odds of success when there is either consensus (!) or con¯ict (!'). When these interaction terms
are included instead of the separate !
i
e?ects for the individual lobby groups, the deviance falls from 88.981 for the model of conditional independence to 63.751 (F = 1:345, p = 0:292). The relevant coecient estimates
are ! = 0:842 (t = 2:65) and !' = ÷0:145 (t = ÷0:75), again emphasising the positive and signi®cant impact of consensus within the lobbying process.
S
.
M
c
L
e
a
y
e
t
a
l
.
/
A
c
c
o
u
n
t
i
n
g
,
O
r
g
a
n
i
z
a
t
i
o
n
s
a
n
d
S
o
c
i
e
t
y
2
5
(
2
0
0
0
)
7
9
±
9
8
9
3
disagreement with lobbyist i. Consistent with the
view that agreement among lobby groups repre-
sents an important element in the rule-develop-
ment process, the relative power of industry is seen
to have fallen when the empirical model is exten-
ded to include interactions with each lobby group,
suggesting the existence of important multiplier
e?ects in relation to industry's perceived in¯uence
on the decisions of the legislature. Speci®cally, the
relative odds of industry's preferred accounting
treatment being incorporated into commercial law
now declines from approximately 2:1 (t = 3:27) in
the relative power model to just less than 1 : 1
(t = ÷0:34) in Model (4.2) in those cases where the
only proposal to have been made is from industry.
However, the positive and signi®cant coecient
for the !
IND

IND
interaction term indicates that
the relative odds of an industry-submitted propo-
sal being recognised in commercial law were more
than three times greater (t = 2:55) when industry's
preferences were aligned with those of one of the
remaining lobbying organisations (see Fig. 1). We
interpret these ®ndings as evidence that industry's
in¯uence on the decisions of the legislative body
was crucially dependent on the support of at least
one of the remaining lobby groups.
Several factors may help explain these observed
multiplier e?ects. First, Watts and Zimmerman
(1986) argue that industry's lobbying position is
primarily motivated on opportunistic grounds. This
self-interest motive can serve to reduce the cred-
ibility of industry-submitted proposals, thereby
reducing the likelihood of their acceptance. How-
ever, to the extent that industry's position appears
more credible when supported by a group of
acknowledged ®nancial reporting experts (in this
case, either the audit profession or accounting
academics), the likelihood of acceptance increases.
Secondly, it is widely acknowledged that industry
represents a powerful lobby group in Germany.
Consequently, for cases in which their position
was matched by one or more ®nancial reporting
constituencies, a body of opinion may have been
created whose political in¯uence was simply too
great for the legislature to ignore.
Models (5.1) and (5.2) account for agreement and
disagreement between industry and auditors, indus-
try and academics, or auditors and academics.
Model (5.1) estimates the main e?ects for agreement
and disagreement between lobbyists i and j and
leads to a signi®cant fall in the deviance (p=0.019)
with respect to the model of relative in¯uence, as
documented in panel D of Table 3, suggesting that
identi®cation of the speci®c parties involved is an
important explanatory factor. The ®ndings con-
®rm the inference drawn previously that the audi-
tors were likely to adopt the position of another
group, either industry (t = 4:45) or academia
(t = 4:29), while the latter two were more likely to
be in con¯ict with each other (t = 4:35). Model
(5.2) extends model (5.1) to account for the odds
of success in cases of agreement and disagreement
between groups i and j. Results indicate that
industry's relative power declined substantially
when such interactions were accounted for, again
suggesting that the ability of German industry to
in¯uence the policy decisions of the legislature
may have been largely conditional on the support
of at least one of the remaining constituent
groups. In particular, model (5.2) suggests the
existence of a signi®cant multiplier e?ect when
there is agreement between industry and academia
Fig. 1. The e?ect of agreement or disagreement between con-
stituent lobby groups in Germany on the likelihood of success
of a proposal to amend the draft accounting law.
Note: the log odds ratios are given in Panel C of Table 3: as
!
i
estimates in the case of one lobbyist alone, !
i

i
when there is
agreement with another party (i.e. consensus) and !
i
'
i
when
there is disagreement (i.e. con¯ict). Agreement with another
party increases the odds of success considerably for all lobby-
ists. On the other hand, when there is disagreement between
lobbyists, only industry tends to be more successful.
94 S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98
(t = 2:17). Thus, while the academic community
appears to have exerted relatively little in¯uence
over the decisions of the legislature when con-
sidered in isolation, German academics are seen to
have derived signi®cant in¯uence through their
interaction with industry.
6.1. Partitioning by accounting issue
Extant research by Puro (1984) and Sutton
(1984) suggests that the distribution of power
among alternative lobby groups may vary as a
function of the nature of the ®nancial reporting
issue under consideration. To explore this possibi-
lity further, we partitioned the 169 separate points
of accounting law into two groups, one of which
comprised all points of law relating to issues of
valuation and recognition (N=92) while the other
comprised all points of law relating to issues of
disclosure and ®nancial statement format
(N = 77). While acknowledging the inherently
arbitrary nature of this partitioning scheme, the
approach at least partly captures Sutton's char-
acterisation of preparers as more concerned with
regulations which a?ect the measurement of
earnings and the valuation of assets and liabil-
ities.
23
Models (1)±(5) were re-estimated for each sample
partition. Table 4 presents the coecient estimates
and associated deviances for models (2), (5.1) and
(5.2). The focus of this table is con®ned to the !
i
terms as the remaining coecient estimates were
Table 4
The relative success of industry, auditors and academics, partitioned by type of accounting issue (t-statistics in parentheses)
a
Model Relative in¯uence Agreement and disagreement between i and j
(Model 2) (Main e?ects model 5.1) (Interaction e?ects model 5.2)
Valuation and
recognition
Format and
disclosure
Valuation and
recognition
Format and
disclosure
Valuation and
recognition
Format and
disclosure
!
IND
0.251 1.308 0.073 1.206 0.030 0.262
(1.00) (3.68) (0.26) (3.34) (0.07) (0.62)
!
AUD
0.550 0.000 0.576 0.122 0.486 ÷0.022
(1.70) (0.00) (1.61) (0.31) (0.86) (0.04)
!
ACA
÷0.133 ÷0.898 ÷0.234 ÷0.72 ÷0.859 ÷0.730
(÷0.45) (÷2.51) (÷0.72) (÷1.86) (1.51) (1.41)
Deviance 46.922 50.431 16.034 17.461 9.283 1.674
ÁDeviance
b
4.191
(1, 2)
23.447
(1, 2)
30.888
(2, 5.1)
32.970
(2, 5.1)
6.751
(5.1, 5.2)
15.787
(5.1, 5.2)
F
c
0.566 2.945 4.173 4.091 0.848 11.005
p-value 0.644 0.059 0.014 0.015 0.571 0.003
a
The initial sample of 169 separate accounting issues has been partitioned into those dealing with valuation and recognition mat-
ters (N = 92) and those dealing with format and disclosure matters (N = 77). Separate generalised linear models were then estimated
for each partition. For parsimony, only the relative in¯uence coecients (!
i
) are reported for all models. The coecients are from a
log-linear model with a log link function and a Poisson distributed error. The dependent variable is the log of the counts from either
the valuation and recognition contingency table or format and disclosure contingency table. Coecient estimates are log-odds ratios: a
regression coecient of zero implies odds equal to one, while a positive (negative) coecient implies odds greater (less) than one.
Subscripts IND, AUD and ACA indicate the industry, auditor and academic lobby groups, respectively.
b
The subscripts in parentheses indicate the nested models for which the change in deviance is reported. The deviance for the
model of conditional independence (Model 1) is 88.981. Note that the
i
coecients in model (1) are also included in all subsequent
estimations, but for parsimony are not recorded.
c
The F-statistic represents a test of the relative goodness of ®t of two nested models. The relative goodness of ®t is measured as the
di?erence in the associated deviances (ÁD) of two models, x and y, where x is nested in y. This di?erence is asymptotically F-dis-
tributed. A signi®cant F-statistic indicates that y is a better ®tting model than x, or equivalently, that the terms omitted from y are
signi®cant.
23
Partitioning at a ®ner level of accounting issue was not
possible due to an insucient number of observations in each
partition to allow reliable estimation.
S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98 95
quantitatively similar to those reported in Table 3.
The relative in¯uence attributable to the industry
constituency appears to be restricted to issues of
disclosure and ®nancial statement formats: in the
case of model (2), industry's relative log-odds of
success are 1.308 (t = 3:68) for format and dis-
closure issues, compared with only 0.251
(t = 1:00) for issues of valuation and recognition.
However, consistent with the ®ndings presented in
Table 3, all !
i
coecient estimates in both sample
partitions become insigni®cant when the e?ects of
agreement or disagreement between lobbyists are
accounted for, again emphasising the importance
of political consensus with respect to the account-
ing rule-development process. The di?erential
impact of agreement and disagreement on lobby-
ing success across the various ®nancial reporting
issues may be inferred from examination of the
reductions in the deviance associated with the
estimation of model (5.2) for the two sub-samples:
the change in deviance is signi®cant for format
and disclosure issues (p = 0:003) but not for
valuation and recognition issues (p = 0:571).
7. Summary and conclusions
This paper presents evidence on the impact of
constituent lobbying activity on the policy deci-
sions of the German legislature during the trans-
formation of the Fourth Directive into German
commercial law. Three main lobby groups are
identi®ed: industry, auditors and academics. With
the exception of the levels of lobbying activity
observed for the academic constituency, the active
groups display a remarkable similarity to those
documented for the US and similar regulatory
regimes. The high level of participation by Ger-
man accounting academics, relative to the levels
observed in the US, can be partially explained by
the active market for legal interpretations in Ger-
many. Together, these results support Sutton's
(1984) model of lobbying activity and provide
prima facie evidence in support of Sutton's addi-
tional conjecture that these patterns will hold
irrespective of the institutional setting.
Using publicly available proposals, we identify
the preferences of each lobby group with respect
to 169 separate accounting issues. These preferences
are then reconciled with the eventual accounting
treatment required by the Accounting Directives
Law as a means of empirically operationalizing the
concepts of power and in¯uence. Several important
®ndings are documented. First, analysis of the main
e?ects models suggests that the industry lobby
group exerted the greatest level of relative power
over the legislature, while the proposals issued by
the academic community display the least like-
lihood of success. However, when the empirical
model is extended to include interaction terms
which account for agreement and disagreement
between the lobbyists, industry's relative power is
seen to have declined, suggesting that the in¯uence
of preparers on the decision of the legislature
depended crucially on the support of at least one
of the remaining lobby groups. While the aca-
demic community on its own appears to have
exerted relatively little in¯uence over the decisions
of the legislative body, academic experts acting in
unison through a formal lobbying group appear to
have derived signi®cant in¯uence through their
contribution to a process in which agreement
between lobbyists was a key feature of accounting
lawmaking. Finally, additional tests indicate that
industry's relative power was largely restricted to
format and disclosure issues, again depending
crucially on the support of either the academic or
auditor groups.
Interestingly, German law was amended in May
1998 to mandate a private sector institution to
develop accounting standards applicable to group
accounts and to advise the Ministry of Justice on
changes to accounting law. The Deutscher Stan-
dardisierungsrat has since been founded and, once
published by the Ministry of Justice, the DSR's
standards should have the standing of legally-
recognised accounting principles.
24
The DSR has
been modelled on the FASB, and is sta?ed by
independent experts: three from industry, two
auditors, one ®nancial analyst and one academic.
24
New legislation has also been passed to allow German
listed companies the option to use internationally accepted
accounting rules in their group accounts as an interim measure
until 2004. In turn, the new DSR has been given the task of
adapting German accounting principles to international norms
by that date.
96 S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98
Although the DSR will instigate due process for the
development of its standards, ®nal decisions will
require a two-thirds majority of its board. There-
fore, in conclusion, we can say that the consensus
between the parties which we have demonstrated in
this paper is now being institutionalised within the
new framework.
Acknowledgements
The authors gratefully acknowledge the helpful
suggestions made by Brian Francis, Pelham
Gore, Chris Nobes, Ken Peasnell, and workshop
participants at Frankfurt University, Lancaster
University, the University of Wales, Bangor, the
1997 British Accounting Association Conference,
the 1997 European Accounting Association Con-
ference and the 1997 Financial Accounting and
Auditing Research Conference. Research assis-
tance was provided by Marc WaÈ hrisch. Financial
assistance was provided by the European Union
Human Capital and Mobility Programme, Con-
tract ERBCHRXCT.
References
Bayless, R., Cochrane, J., Harris, T., Leisenring, J., McLaugh-
lin, J., & Wirtz, J. P. (1996). International access to US
capital markets: an AAA forum on accounting policy.
Accounting Horizons, 10, 75±94.
Benveniste, G. (1972). The politics of expertise. London: Croom
Helm.
Brooks, J. P., & Mertin, D. (1986). Neues Deutsches Bilanz-
recht/New German accounting legislation. DuÈ sseldorf: IDW-
Verlag GmbH.
Brown, P. (1981). A descriptive analysis of select input bases of
the Financial Accounting Standards Board. Journal of
Accounting Research, 19, 232±246.
Chatov, R. (1985). The possible new shape of accounting reg-
ulation in the United States. Journal of Accounting and Public
Policy, 4, 161±174.
Diggle, G., & Nobes, C. (1994). European rule-making in
accounting: the seventh directive as a case study. Accounting
and Business Research, 24, 319±334.
Downs, A. (1957). An economic theory of democracy. New
York: Harper & Row.
Dyckman, T. R. (1988). Credibility and the formulation of
accounting standards under the Financial Accounting Stan-
dards Board. Journal of Accounting Literature, 7, 1±30.
Forster, K.H. (1997). Adler-During-Schmaltz, Rechnungslegung
und PruÈfung der Unternehmens. Stuttgart: Poeschel.
Francis, J. R. (1987). Lobbying against proposed accounting
standards. Journal of Accounting and Public Policy, 6, 35±57.
Francis, B., Green, M., & Payne, C. (1993). GLIM4. Oxford:
Oxford University Press.
Grin, P. A. (1983). Management's preferences for FASB State-
ment no. 52: predictive ability results. Abacus, 19, 130±138.
Hope, T., & Gray, R. (1982). Power and policy making: the
development of an R&D standard. Journal of Business,
Finance, and Accounting, 9, 531±558.
Horngren, C. T. (1972). Accounting principles: private or pub-
lic sector? Journal of Accountancy, 134, 37±41.
Hussein, M., & Ketz, J. (1980). Ruling elites of the FASB: a
study of the big eight. Journal of Accounting, Auditing and
Finance, 3, 354±367.
Hussein, M., & Ketz, J. (1991). Accounting standards-setting in
the US: an analysis of power and social exchange. Journal of
Accounting and Public Policy, 10, 59±81.
Kelly, L. (1982). Corporate lobbying and changes in ®nancing
or operating activities in reaction to FAS No. 8. Journal of
Accounting and Public Policy, 2, 153±173.
Kelly, L. (1985). Corporate management lobbying on FAS No.
8: some further evidence. Journal of Accounting Research, 23,
619±632.
Kirk, D. J. (1988). Looking back on fourteen years at the
FASB: the education of a standard setter. Accounting Hor-
izons, 2, 8±17.
Klumpes, P. (1994). The politics of accounting rule-making:
developing an Australian pension accounting standard.
Abacus, 30, 159±189.
Lukes, S. (1974). Power: A radical view. London: Macmillan.
McCullagh, P., & Nelder, J. A. (1989). Generalised linear mod-
els (2nd ed.). London: Chapman and Hall.
Nelder, J., & Wedderburn, R. (1972). Generalised linear mod-
els. Journal of the Royal Statistical Society, 135, 370±384.
Newman, D. P. (1981). An investigation of the distribution of
power in the APB and FASB. Journal of Accounting
Research, 19, 247±262.
Nobes, C. (1991). Cycles in UK standard setting. Accounting
and Business Research, 21, 265±274.
Nobes, C., & Parker, R. (1991). Comparative international
accounting. Hemel Hempstead: Prentice Hall.
Ordelheide, D. (1999). Germany. In S. J. McLeay, Accounting
regulation in Europe. London: MacMillan.
Ordelheide, D., & Pfa?, D. (1994). European ®nancial report-
ing: Germany. London: Routledge.
Power, M. (1997). Academics in the accounting policy process:
England and Germany. In J. Flower & C. Lefebvre, Compara-
tive studies in accounting regulation in Europe. Leuven: Acco.
Puro, M. (1984). Audit ®rm lobbying before the Financial
Accounting Standards Board: an empirical study. Journal of
Accounting Research, 22, 624±646.
Puro, M. (1985). Do large accounting ®rms collude in the
standards-setting process? Journal of Accounting, Auditing,
and Finance, 8, 165±177.
Rahman, A. R., Ng, L. W., & Tower, G. D. (1994). Policy
choice and standard setting in New Zealand: an exploratory
study. Abacus, 30, 98±117.
S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98 97
Sunder, S. (1988). Political economy of accounting standards.
Journal of Accounting Literature, 7, 31±45.
Sutton, T. G. (1984). Lobbying of accounting standard-setting
bodies in the UK and the USA: a Downsian analysis.
Accounting, Organisations and Society, 9, 81±95.
Tandy, P., & Wilburn, N. (1992). Constituent participation in
standards-setting: the FASB's ®rst 100 statements. Account-
ing Horizons, 6, 47±58.
Tandy, P., & Wilburn, N. (1996). The academic community's
participation in standard setting: submission of comment let-
ters on SFAS Nos. 1±117. Accounting Horizons, 10, 92±111.
US Congress. Senate Subcommittee on Reports, Accounting
and Management of the Committee on Government Opera-
tions (1976). The accounting establishment: a sta? study
(Metcalf Sta? Report). 94th Congress, 2nd Session.
Vieten, H. (1995). Auditing in Britain and Germany compared:
professions, knowledge, and the state. European Accounting
Review, 4, 485±511.
Walker, R. G., & Robinson, S. P. (1993). A critical assessment
of the literature on political activity and accounting regula-
tion. Research in Accounting Regulation, 7, 3±40.
Walker, R. G., & Robinson, S. P. (1994). Related party trans-
actions: a case study of inter-organisational con¯ict and the
`development' of disclosure rules. Abacus, 30, 18±43.
Watts, R., & Zimmerman, J. (1978). Towards a positive theory
of the determination of accounting standards. The Account-
ing Review, 53, 112±134.
Watts, R., & Zimmerman, J. (1986). Positive accounting theory.
New Jersey: Prentice Hall.
Wyatt, A. (1991). Accounting standards setting at a crossroads.
Accounting Horizons, 5, 110±114.
Wyatt, A., & Yospe, J. F. (1993). Wake-up call to American
business: international accounting standards are on the way.
Journal of Accountancy, 176, 80±85.
Ze?, S. (1978). The rise of economic consequences. Journal of
Accountancy, 146, 56±63.
98 S. McLeay et al. / Accounting, Organizations and Society 25 (2000) 79±98

doc_798779441.pdf