GOVERNANCE STRUCTURES AND CONTRACTING OUT MUNICIPAL AUDITING IN FINLAND AND NORWAY A ˚ GE JOHNSEN,PENTTI MEKLIN,LASSE OULASVIRTA AND JARMO VAKKURI* INTRODUCTION The purpose of this paper is to analyse governance structures of municipal audit in Finland and Norway. By governance structures (Williamson, 1985) we in this study mean how the municipalities organise the provision of the mandatory audit. Municipalities can organise the mandatory audit by pro- ducing the audit internally within the hierarchy (unified governance), by producing it in co-operation with or in mixed ownership with other organisations (relational contracting), or finally, by contracting the audit from suppliers in the market (market contracting). Williamson (1985) called for special attention to the underdeveloped state of the theory of bureaucracy, and he argued that the study of bureaucratic failure was very primitive compared to the study of market failure. Furthermore, Williamson argued that public finance was an area where transac- tion cost economics held out considerable promise but that problems of incentives and governance are extremely difficult in particular political contexts. Conse- quently, he called for tolerance for greater variance in public sector efficiency assessments compared to private sector efficiency assessments. In this paper, we analyse governance structures and explore the effects of contracting out municipal auditing. We have addressed the political costs in addition to production and transaction costs, which traditionally are the two main efficiency measures in transaction costs analysis (TCA). The background for our study is the ongoing public sector reforms in many countries, including Finland and Norway. Some public sector services now permit competition, and there has been an increased emphasis on output and performance measurement, including performance auditing (Pollitt and Bouckaert, 2000). Finland harmonised its municipal accounting * The first author is from Oslo University College, Norway and the University of Edinburgh, UK. The second, third and fourth authors are from the University of Tampere, Finland. This study was part of the project ‘Contracting out municipal audit’ at Agder Research, Kristiansand, financed by the Norwegian Research Council programme ‘Public Sector in Transition’; and the ‘Local authority economic management’ project in the ‘Finnish Local Government 2004’ programme financed by the Association of Finnish Local and Regional Authorities. Address for correspondence: A ˚ ge Johnsen, Faculty of Business, Public Administration and Social Work, Oslo University College, P.O. Box 4 St. Olavs plass, NO-0130 Oslo, Norway. e-mail: [email protected]Financial Accountability & Management, 20(4), November 2004, 0267-4424 #Blackwell Publishing Ltd. 2004, 9600 Garsington Road, Oxford OX4 2DQ , UK and 350 Main Street, Malden, MA 02148, USA. 445
34
Embed
Governance Structures and Contracting Out Municipal Auditing in Finland and Norway
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
GOVERNANCE STRUCTURES AND CONTRACTINGOUT MUNICIPAL AUDITING IN FINLAND AND
NORWAY
AGE JOHNSEN, PENTTI MEKLIN, LASSE OULASVIRTA AND JARMO VAKKURI*
INTRODUCTION
The purpose of this paper is to analyse governance structures of municipal
audit in Finland and Norway. By governance structures (Williamson, 1985)
we in this study mean how the municipalities organise the provision of the
mandatory audit. Municipalities can organise the mandatory audit by pro-
ducing the audit internally within the hierarchy (unified governance), by
producing it in co-operation with or in mixed ownership with other organisations
(relational contracting), or finally, by contracting the audit from suppliers in the
market (market contracting). Williamson (1985) called for special attention to the
underdeveloped state of the theory of bureaucracy, and he argued that the study of
bureaucratic failure was very primitive compared to the study of market failure.
Furthermore, Williamson argued that public finance was an area where transac-
tion cost economics held out considerable promise but that problems of incentives
and governance are extremely difficult in particular political contexts. Conse-
quently, he called for tolerance for greater variance in public sector efficiency
assessments compared to private sector efficiency assessments. In this paper, we
analyse governance structures and explore the effects of contracting out municipal
auditing. We have addressed the political costs in addition to production and
transaction costs, which traditionally are the two main efficiency measures in
transaction costs analysis (TCA).
The background for our study is the ongoing public sector reforms in
many countries, including Finland and Norway. Some public sector services
now permit competition, and there has been an increased emphasis on
output and performance measurement, including performance auditing
(Pollitt and Bouckaert, 2000). Finland harmonised its municipal accounting
*The first author is from Oslo University College, Norway and the University of Edinburgh, UK.The second, third and fourth authors are from the University of Tampere, Finland. This study waspart of the project ‘Contracting out municipal audit’ at Agder Research, Kristiansand, financed bythe Norwegian Research Council programme ‘Public Sector in Transition’; and the ‘Localauthority economic management’ project in the ‘Finnish Local Government 2004’ programmefinanced by the Association of Finnish Local and Regional Authorities.
Address for correspondence: Age Johnsen, Faculty of Business, Public Administration and SocialWork, Oslo University College, P.O. Box 4 St. Olavs plass, NO-0130 Oslo, Norway.e-mail: [email protected]
Financial Accountability & Management, 20(4), November 2004, 0267-4424
system with accrual accounting and deregulated the municipal audit market
in 1997 (Monsen and Nasi, 1996). In 1992, the Norwegian Parliament
decided to keep a distinct (idiosyncratic) municipal accounting system
(Bergevarn et al., 1995) and a regulated municipal audit market, and this
decision was upheld in the revision of the municipal accounting system in 2000.
Unlike the other Nordic countries, Norway did not allow municipalities to
contract out municipal audit during the period of study (1996–2000). How-
ever, there was an increasing political demand for change during that period
of time. In 2002, the Norwegian centre-right coalition government proposed
a change in the existing legislation in order to allow municipalities to voluntarily
contract out municipal auditing. The Norwegian Parliament approved the new
legislation in 2003, and municipalities will be able to voluntarily contract out
municipal audits from 1 July, 2004. The Labour Party, the Centre Party and
the Socialist Left Party voted against this proposal. The re-organisation of
Norwegian municipal auditors, the increased emphasis upon performance
auditing, and the greater investment in audit competence, are events that
occurred during the late 1990s (Johnsen et al., 2001). All of these events can be
understood as being actions that were in accordance with the growing political
demands for change and for an alignment with auditing practices in the other
Nordic countries.
The increased competition that will soon be allowed, due to deregulation
and contracting out, is interesting to study from a TCA perspective. Muni-
cipal audit is characterised by the agency problem of asymmetric informa-
tion (Watts and Zimmerman, 1983). In addition, auditing is strongly
regulated by the norms of the accounting profession (Hatherly and Brown,
1996). Consequently, there may be problems related to human asset speci-
ficity (specialised competence), which could facilitate opportunism. These
potential problems may be measured in terms of effects on productivity, i.e.
the cost of audit services to the municipalities under non-competitive versus
competitive conditions and in terms of transaction costs for different govern-
ance structures. In this paper, we have analysed the costs (audit fees and
transaction costs) and governance structures of municipal audit services in
Norway, where there has been little or no competition in the municipal audit
market, and the governance structures and costs of similar services in
Finland, which deregulated the municipal audit market in 1997. Our three
research questions are: What governance structures are used in municipal
auditing? How does governance structure, including contracting out, affect
audit fees? How does governance structure affect transaction costs?
There are several reasons why it is important to study governance struc-
tures and contracting out municipal auditing. First, there is the need for
studies of governance structures and contracting out municipal audit in
European contexts. Contracting out has been an important element in
public sector reforms during the 1980s and 1990s, and there is extensive
documentation of the effects of contracting out many municipal services
# Blackwell Publishing Ltd 2004
446 JOHNSEN, MEKLIN, OULASVIRTA AND VAKKURI
(Boyne, 1998). Still, to our knowledge, there are no empirical studies of
contracting out municipal audit in Europe.
Second, there is a need for more knowledge about ‘hybrid forms’ in
transaction cost analysis (TCA). By hybrid forms we mean those governance
structures that organisations use instead of unified governance (hierarchy) or
market contracting (Powell, 1990). These are commonly referred to as rela-
tional contracting, networks and alliances. However, although all three of
these governance structures are important, unified governance and market
contracting have received most attention in empirical TCA studies. In 1985,
Oliver Williamson wrote:
Bilateral structures [one of the relational governance structures between unified govern-ance and market contracting] have only recently received the attention they deserve,and their operation is least well understood (Williamson, 1985, p. 76).?
Unfortunately, the neglect of hybrid forms in TCA has persisted. According
to Williamson (2000), the new institutional economics (transaction cost
analysis) has hitherto dealt incompletely with mixed ownership, e.g. govern-
ance structures between ‘in-house’ municipal audit and market contracting:
The attributes of mixed ownership modes (alliances, joint ventures, franchising, and thelike) as well as the mechanisms for supporting credible contracting between autonomousfirms are incompletely worked out (Williamson, 2000, pp. 610–11).
Few have addressed hybrid forms and their potential impact upon differ-
ent services, including audit. For instance, in a relatively recent study on the
boundaries of the firm and why firms outsource internal auditing (Widener
and Selto, 1999), only unified governance and market contracting of profes-
sional accounting firms were studied. However, Widener and Selto (1999)
relied on an early formulation of Williamson’s transaction cost theory in
their study (Williamson, 1975). While Williamson in his earlier formulations
of TCA regarded transactions of the ‘middle kind’ (hybrid forms), as being
very difficult to organise and unstable, he was later persuaded to regard such
transactions as being much more common. Furthermore, he also argued that
many of the most interesting problems of organisation involve both asset
specificity and, one of the core issues of interest within agency theory,
information asymmetry (Williamson, 1985).
As an example of a service that might be relevant in relational governance
structures, Williamson (1985) mentioned architects. Architects are, like nurses,
doctors, engineers, lawyers, accountants and auditors, professionals that render
more or less specialised services on an occasional or recurrent basis to different
organisations. Some organisations hire their own professionals. Others make
arrangements for long-term relationships to meet their service needs. Finally,
there are organisations that contract professional services in the market when
needed. This means that studies of governance structures should take the diversity
of options into account and not be satisfied with a unified governance–market
# Blackwell Publishing Ltd 2004
GOVERNANCE STRUCTURES AND CONTRACTING OUT 447
contracting dichotomy or with a public-private provider dichotomy. Audit is a
professional service that involves ongoing relationships (transactions) between audi-
tors and auditees. These transactions are subject to asymmetric information and
asset specificity. Thus, audit bears on one of the main concerns of TCA, which is
the choice of governance structure, determining the borders of the organisation.
Municipal audit may become a highly suitable test-bed for studies of governance
structure in many public sector professional services. New knowledge about con-
tracting out municipal audit could be useful information for audit practitioners,
policy makers and for students of auditing and economic organisation theory.
The rest of this paper is outlined as follows. The next section explains how
we have applied TCA to develop hypotheses about municipal audit. The
third section is an outline of the research design and a description of the data
we have used. The fourth section documents the empirical results, including
a multiple regression analysis of how governance structures affected muni-
cipal audit fees per capita in Finland and Norway. The paper ends with a
discussion of the results and conclusions, and includes a summary of the
findings, limitations of the present study, implications for theory and prac-
tice, and suggestions for future research.
THEORY AND HYPOTHESES
TCA is an interdisciplinary new institutional theory in the intersection of law,
economics and organisation (Williamson, 1985). TCA is predominantly concerned
about the boundaries of the firm, or more precisely in our context, of the organisa-
tion. TCA maintains that any issue that either arises or can be recast as a problem
of contracting is usefully examined in terms of transaction cost and that all contract
schemes should be examined micro analytically and assessed in a comparative
manner. However, because of bounded rationality, contracts and governance
structures are assessed relative to one another in TCA and without the claim of
optimality that is common within neoclassical economics. Subsequently, any gov-
ernance structure that is in use may, as a starting point, be assumed to be efficient
relative to the local common knowledge concerning other relevant governance
structures.
Marsh (1998) wrote that TCA implications for local governance had only
recently been appreciated and held that TCA appeared to offer a weapon
against fragmentation. Fragmentation is the term that is often used to
describe the dysfunctional effects of excessive contracting out, privatisation,
use of networks etc. instead of the traditionally integrated and top-down
controlled public sector bureaucracy (the ‘hierarchy’ in TCA terms). Thus,
too much contracting out may result in relatively high transaction costs and
fragmentation, which may result in a hollowing out of the state and reduce
the steering capacity of public policy (Rhodes, 1997). However, Marsh
(1998) also identified several limitations in the assumptions underlying
# Blackwell Publishing Ltd 2004
448 JOHNSEN, MEKLIN, OULASVIRTA AND VAKKURI
TCA. Most notably, he argued that the decision about governance structures
is not restricted to minimising the sum of production and transaction costs
but also includes the optimisation of the net benefit, and that different
governance structures can produce different kinds of transactions. For
example, the nature of trust and the net benefit of municipal audit could
vary depending upon whether or not the audit is contracted in the
market, produced in relational contracting or provided internally.
TCA assumes bounded rationality, which means that there is uncertainty
and that planning is difficult. Human nature ‘as we know it’ is seen as being
prone to opportunism, which means that some individuals are not trust-
worthy all of the time. Because of bounded rationality and opportunism,
classical complete contracting is difficult and costly to agree upon and to
enforce. Consequently, TCA focuses upon ex-post governance structures
and neoclassical incomplete contracting and relational contracting that can
facilitate efficient exchange across organisational boundaries by minimising
total production and transaction costs. Transaction costs are the costs of
running an economic system and should be distinguished from production
costs. Transaction costs can be divided into ex-ante and ex-post types.
Ex-ante transaction costs are the costs of drafting, negotiating and safeguarding
an agreement. Transaction costs of ex-post type are costs related to mal-
adaptation from the contract, haggling costs to correct ex-post misalignment,
set-up and running costs associated with governance structures (courts and
other institutions) for the resolution of disputes, and bonding costs for effecting
secure commitments.
According to TCA, the decisions determining the appropriate governance
structure to minimise production and transaction costs will depend upon
asset specificity, uncertainty, and the frequency of the transactions. Asset
specificity is divided into four types; site, physical asset, human asset and
dedicated assets specificity. Each of the four types can be categorised into
non-specific, mixed or idiosyncratic investment characteristics for the trans-
actions in question. Audit can be classified as a product with so called mixed
investment characteristics. The auditors may have specific human assets and
certain audits, and most notably performance audit, may also have site
specificity. The financial audit of firms may be relatively non-specific and
recurrent, something that favours market contracting. It should be noted,
though, that professional audit firms might specialise in industries or
branches, enhancing asset specificity. Municipal financial audit, on the
other hand, requires competence in municipal accounting and public sector
regulation, and may include cameral accounting (Monsen and Nasi, 1996) in
addition to accrual accounting. Municipal performance audit may require
organisation and/or service specific tacit and formal knowledge that requires
the use of other professionals, for example, medics and engineers in addition
to accountants. Some performance audits may be recurrent and conducted
on an annual basis, but these audits will nevertheless not cover the same
# Blackwell Publishing Ltd 2004
GOVERNANCE STRUCTURES AND CONTRACTING OUT 449
services or issues regularly. These forms of human asset specificity, in add-
ition to the widespread asymmetric information that exists in professional
public sector services, strongly indicate that there are opportunities for
opportunism and substantial transaction costs related to municipal audit.
TCA distinguishes between non-strategic (structural/exogenous) uncer-
tainty and strategic (behavioural) uncertainty, where opportunism is of
special importance in TCA. For instance, a market situation with a large
number of bidders at the outset may be transformed into a bilateral supply
after a contract has been settled (the fundamental transformation problem).
It should be noted that there could be substantial interaction effects between
asset specificity, uncertainty and transaction frequency. Furthermore, TCA
distinguishes between occasional and frequent transactions.
Regarding transaction frequency, one can argue that financial audit is
recurrent and frequent because it is conducted annually or more often.
Performance audit is occasional. The same project, service or programme
is not regularly or routinely audited. However, the same organisation may
undergo performance audit regularly. Sometimes large organisations will
have annual performance audits.
When there are mixed or specific investment characteristics and recurrent
or occasional transactions, as with performance audits, then the transactions
may be efficiently organised either as a bi- or trilateral governance structure
or as a unified governance structure. The relatively non-specific financial
audit is likely to be contracted from the market, unless the financial audit is
bundled together with the performance audit. In addition, the entire ques-
tion about municipal audit governance structure is made even more complex
by the issue of plural forms (Bradach and Eccles, 1989). As some audits may
be non-specific and recurrent, especially financial audit, and other audits
may be specific and occasional, typically internal (or management) audit and
performance audit, several governance structures can be chosen. Some of
the municipal audit may be conducted within unified governance, by mixed
ownership or other relational contracts, by contracting professional account-
ing firms, or by combinations of all of these governance structures (plural
forms). Plural forms means, in our context, that a municipality may have
some audit in unified governance, participate in the inter-municipal owner-
ship of auditors, and contract some audits from the market, simultaneously.
Unified Governance
TCA assumes that organisations adapt to minimise the sum of production and
transaction costs. Unified governance is suitable when the transactions are rela-
tively recurrent and idiosyncratic. In this situation there may be few external
suppliers, and if the auditee is a large organisation then economy of scale can be
as well utilised as it is by an efficient external supplier. We can expect that
# Blackwell Publishing Ltd 2004
450 JOHNSEN, MEKLIN, OULASVIRTA AND VAKKURI
relatively large municipalities could utilise the economy of scale and provide all the
more or less specialised audit services internally (hypothesis 1).
H1: Large organisations have more unified governance of audit than
small organisations.
Let X denote the proportion of a governance structure within a population,
let UG be unified governance, let L be a large municipality and S a small
municipality, and let i be period and j be country. We can then test hypoth-
esis 1 by studying if:
XðUGLÞij > XðUGSÞij :
Relational Contracting
One of the possible benefits of both relational contracting and market
contracting is to establish incentives for efficiency through contestable mar-
kets (Baumol et al., 1982). Contestability means that the sole supplier does
not have a permanent hold on the market or contract and could be replaced
by other suppliers. However, Williamson (1985) argued that due to asset
specificity, contestability through easy entry and exit of firms in the market
might not be feasible. After the contract has been settled between a munici-
pality and an auditor, the initial bidders may exit from the market and the
market may effectively disappear. This is called the fundamental (market)
transformation (Williamson, 1985). Relational contracting could have major
advantages compared to both unified governance and market contracting of
audit. In relational contracting the audit committee is not dependent on a
sole internal provider as in unified governance, and relational contracting
would not be subject to the fundamental transformation problem to the same
degree as in market contracting. Therefore, the auditee may try to minimise
audit (production) and transaction costs by establishing a long-term trust
based relationship with one or a few auditors.
Auditing is a professional service. It is therefore difficult for the audit
committee to identify the auditor’s competence, monitor efficiency and
verify audit quality regardless of governance structures. For instance,
Roberts et al. (1990) found that school districts in Texas that changed
auditors in order to cut audit fees, might have received audit of lower quality
than before. (In fact, some of this effect may even have been intended as
some of the districts wanted to avoid negative information provided by some of the
previously used auditors.) In establishing relational contracting the municipalities
may seek co-operation with other similar organisations in order to establish mixed
ownership of a shared auditor. This action could facilitate economy of scale in
audit production costs as well as a high professional standard in the contracting
process. This action could also enhance the municipalities bargaining power
# Blackwell Publishing Ltd 2004
GOVERNANCE STRUCTURES AND CONTRACTING OUT 451
relative to the auditor as well as sharing fixed contracting costs between the
municipalities. For instance: when the Finnish municipal audit market was deregu-
lated in 1997, Audiator OY (limited), a nation-wide auditor formed by many of
the previous internal municipal auditors, became the market leader. The interest-
ing issue is that Audiator OY, which was established in 1993, is owned 100 per
cent by the Association of Finnish Local and Regional Authorities. Since all
Finnish municipalities are members of the Association of Finnish Local and
Regional Authorities, the auditor that is now the market leader is owned indirectly
by its auditees. In Norway, an increasing number of municipalities have merged
their former internal auditors into district audits. As a consequence, we expect to
find relatively extensive use of relational contracting in auditing.
H2: Relational contracting is used relatively more than unified govern-
ance and market contracting in municipal auditing.
Let RC denote relational contracting and MC market contracting. We can
thus test hypothesis 2 by studying if:
XðRCÞij > XðUG; MCÞij :
The choice of governance structure is assumed to be a function of both
production and transaction costs. Consequently, the municipalities are
expected to choose that governance structure that minimises the sum of
production and transaction costs. If all costs related to governance structures
were randomly distributed, then we could expect a distribution of 1/3 for
each of the three governance structures. However, there is a large number of
small municipalities in both Finland and Norway and we expected that
relational contracting, due to economy of scale in production costs, would
be the dominant governance structure. We wanted to test whether or not
hypothesis 2 would be supported due to transaction costs. For that reason we
only tested hypothesis 2 in the group of large municipalities in both coun-
tries, assuming that most small municipalities due to economy of scale
reasoning would not use unified governance.
Market Contracting
We have specified two hypotheses regarding market contracting. Hypothesis
3 concerns the level of audit fees and hypothesis 4 concerns the volatility of
audit fees.
TCA makes the claim that if there are non-specific and occasional or
recurrent transactions, as is the case with yearly financial audits, then market
contracting is the best and most efficient governance structure.1 Contracting
out is understood as opening up for competition in the public sector, where
the provision of services is still seen as being a governmental responsibility.
One argument for contracting out audit is that competitive or financial
# Blackwell Publishing Ltd 2004
452 JOHNSEN, MEKLIN, OULASVIRTA AND VAKKURI
pressures force a concentration upon core competencies, increasing specia-
lisation and demands for new skills and competencies in auditing beyond
traditional financial and internal audits. A second common argument for
contracting out audit is the increased supply of high quality audit that can be
made available to the public sector by the big professional accounting firms
(Rittenberg and Covaleski, 1999; and Widener and Selto, 1999). According
to Domberger and Jensen (1997), two prominent scholars of contracting out,
expenditures, on average, can be reduced by 20% after contracting out.
However, some of Domberger et als’ methods and estimates have been
contested. Hodge (2000), on the other hand, based on a meta-analysis of
the available literature on contracting out for the period 1976–1994, esti-
mates the reduced costs to average from 6 to 12%. Furthermore, according
to Domberger and Jensen (1997) and Hodge (2000), one may expect the
quality of the services to be maintained or enhanced.
Research on the US municipal audit market indicated that increased
competition reduced audit fees (Sanders et al., 1995). However, Rubin
(1992) argued that private auditors in the US may charge higher audit fees
than municipal auditors. Rittenberg and Covaleski (1999) found that con-
tracting out internal audit in British government reduced costs by 25% on
average. Contracting out also led to more focused audit reports, and many of
the internal auditors began to co-operate with external suppliers in order to
enhance the quality of existing services. (This corroborates the underlying
critique and the implicit assumption of TCA, that contracting out will
promote homogenous transactions between governance structures.) We
expected that the municipalities would pragmatically choose that govern-
ance structure that minimised audit fees and transaction costs and optimised
the benefit of audit according to local circumstances. Consequently, we
expected a negative relationship between the use of market contracting
and an increase in audit fees per capita, after the deregulation in Finland.
Moreover, because all auditors are now facing competition, we expected the
overall level of audit fees per capita in Finland to decline in real terms. We
expected the Finnish audit fees to be lower and to have increased less than
the audit fees in Norway (hypothesis 3). This expectation was due to our
assumption that the already existing greater competition in the deregulated
Finnish market would reduce costs more than the increased threat of con-
tracting out would do in the Norwegian market.
H3: Increasing competition reduces audit fees.
We can test hypothesis 3 by studying if:
p1Finland > p2Finland and p2Finland < p2Norway:
In markets, where there are homogenous (non-specific) products and where
the identity of the supplier is irrelevant, price would be the major competition
# Blackwell Publishing Ltd 2004
GOVERNANCE STRUCTURES AND CONTRACTING OUT 453
factor. In relational contracting, costs would tend to be fixed and quantity rather
than price would more likely be adjustable (Williamson, 1985). This was indirectly
corroborated for the audit market by Hatherly and Brown (1996) who argued that
auditors (professional audit firms) compete on image and costs rather than directly
on audit quality unless quality is made much more visible than it is at the present.
Audit quality is commonly defined as the probability that the audited financial
statements contain material errors (Copley and Doucet, 1993). However, this
definition is less applicable to performance audit where decision relevance (Mays-
ton, 1985), volume of public funds involved, financial risk and political salience
(Keen, 1999) may be better criteria for audit quality. Furthermore, because there
is no general agreement on what performance audit is (Bowerman, 1996; and
Johnsen et al., 2001), performance audit is not a homogenous product. The
identity of the supplier may be important, and this may explain why there may
be relatively more relational contracting in municipal auditing. (This will be tested
in hypothesis 2.) Nevertheless, when the audit is contracted from the market, we
commonly expect price adjustments (hypothesis 4).
H4: Audit fees are more volatile in market contracting than in relational
contracting.
Thorne et al. (2001) concluded that audit fees on average were lower for
fixed-fee contracts than for cost-imbursement contracts. Moreover, they also
concluded that the probability for negotiating a fixed-fee contract increased
for governmental units with relatively high financial risk when they engaged
an independent auditor with a great deal of knowledge about the public
sector. We assumed that auditors engaged by municipalities should have a
great deal of knowledge about the public sector. Hence, fixed-fee contracts
should be common. However, the presence of strong incentives for cost
efficiency and lowballing in market contracting could (depending on compe-
tition) materialise in relatively low and volatile audit fees. If so, market
contracting of municipal audit would lead us to expect to see more adjust-
ments in prices (audit fees) relative to relational contracting (hypothesis 4).
RESEARCH DESIGN, SAMPLE AND DATA
Research Design
A characteristic of empirical research in transaction cost economics is that
the main variables of study, transaction frequency, asset specificity and
uncertainty, are difficult to measure and often require the use of proxies.
For instance, accounting data, even detailed accounting data, are often
poorly suited for the needs of TCA. Therefore, direct measurements of
transaction costs are rarely attempted. Comparative institutional analyses
are often used to study the attributes of the transactions and those are tested
# Blackwell Publishing Ltd 2004
454 JOHNSEN, MEKLIN, OULASVIRTA AND VAKKURI
against the TCA’s predictions regarding governance structures. This is done
instead of directly measuring transaction costs. Moreover, Williamson (1985)
called for more depth (fewer and more relevant data) compared to breadth
(more observations) in research design tradeoffs, for the near term.
Ideally, we wanted to compare the effects of contracting out in Finland by
collecting data before and after the contracting out ‘experiment’ started in
1997. Our hope was to be able to use Norwegian data as a control. However,
due to changes in the Finnish municipal accounting system in 1997, some
municipalities were unable to report audit fees prior to 1997. As a result, we
decided to use audit fees data for 1997 and 2000, in Finland, and for 1996
and 2000, in Norway.
Finland
Finland had a population of 5.1 million inhabitants in 2000. Finnish local
government had only one tier and that consisted of 452 municipalities.
According to the Local Government Act of 1995 §71, 72 and 73, Finnish
municipalities were required to have both performance auditing and exter-
nal financial auditors, but the municipalities were free to select the external
auditor. The auditors’ terms were fixed for the entire four-year period of
municipal council government. The municipal audit committees decided on
the performance audits that were subsequently produced either through the
system of unified governance (local government officials and/or members
of the audit committees) or with the assistance by the contracted external
auditors. The five largest auditors operating in the Finnish local government
market in the 1997–2000 period were Audiator Oy, Suomen Kuntatarkastus
Oy, SVH Julkistarkastus Oy, KPMG Kunta Oy, and Kuntatarkastajien Oy
Clavis. Moreover, there were three other smaller firms. Audiator had a
leading position in many rural areas but not in urban areas. The other
four auditors were private firms, and they were subsidiaries of large audit
firms (including KPMG and Deloitte & Touche). In 2001, a total of 13
authorised audit firms expressed their interest in operating within the muni-
cipal audit markets. This shows that there has been a keen interest in
entering the municipal audit market and that the municipal options for
selecting auditors have increased relative to the period of time that we
have studied.
Norway
Norway had 4.5 million inhabitants in 2000. Norway had two tiers of local
government with 435 municipalities and 18 counties. There were 82 muni-
cipal and 18 county auditors in 2000, but we only studied the municipal
auditors. The municipal audit was regulated by the Municipal Act of 1992
and was mandatorily performed by municipal or district auditors (§60.3).
# Blackwell Publishing Ltd 2004
GOVERNANCE STRUCTURES AND CONTRACTING OUT 455
All municipal and district auditors were independent by regulation of the
Municipal Act of 1992 and reported directly to the municipal councils’ audit
committees. Furthermore, all the Norwegian municipal auditors audited the
municipalities on a non-profit basis. The municipal audit committees were
responsible for conducting financial, management and performance audits
that were produced by mandated municipal auditors. (The municipal audi-
tors were also responsible for auditing some non-municipal operations, but
in this study we have only collected data on municipal audit fees.) The
municipal councils decided the audit budgets (fees) annually, but even
though contracting out was not allowed, the municipalities were free to use
internal auditors (unified governance) or district auditors, which were jointly
owned inter-municipal entities (mixed ownership). In 2000, the district
auditors were responsible for the audit of 2–18 municipalities each.
Data Collection and Measurements
Sample
The primary data on municipal audit fees were collected from annual
reports and by postal surveys in Finland and by postal surveys in Norway.
In Finland, we collected data from 47 (10.4%) of the 452 municipalities. The
47 municipalities comprise a sample that has been used for many years in
examining the politics, economy and management of local government in
Finland. It is meant to be a representative sample of the different municipal
characteristics, including type of municipality (rural/urban), number of
inhabitants, language used (Finnish or Swedish), and size (large/small).
Although the sample is relatively small compared to the Norwegian sample,
all the five largest Finnish auditors that operated in the municipal audit
market in 2000 were represented in the data from the 47 municipalities in
the sample.
In Norway, we sent the questionnaires to all of the 82 municipal auditors.
After one round of following up the non-respondents, 56 (68.3%) of the
municipal auditors responded with data for both 1996 and 2000. These
auditors audited 298 (68.5%) of the 435 Norwegian municipalities. Many
of the non-respondents came from the northern-most regions where the
municipalities were small and rural, but the data represent auditors and
municipalities from all the regions of Norway.
Audit Fees
The dependent variable in this study is municipal audit fees per capita. Note
that the audit fees include internal (management) audit and performance
audit, when feasible, in addition to the traditional financial audit. Since
many of the Finnish municipalities did not procure external performance
# Blackwell Publishing Ltd 2004
456 JOHNSEN, MEKLIN, OULASVIRTA AND VAKKURI
audit, the costs for such audits are not included for all municipalities in our
data. The Finnish audit fees exclude a 22% value added tax (VAT). The
Norwegian data are without VAT costs, since municipalities in Norway were
exempted from paying VAT in the first place. The surveys distinguish
between financial audit fees and performance audit fees and also provide
total audit fees. Only a minority of the auditors were able to specify financial
and performance audit fees separately. As a consequence, we analysed the
total municipal audit fees but included a control variable for the inclusion of
performance audit in the Finnish data, see below.
Governance Structures
In this paper we have distinguished four governance structures, from no to
full market contracting. The first is internal municipal audit (unified govern-
ance). The second is audit provided in co-operation or mixed ownership with
other organisations (relational contracting). The third is audit procured from
private or public auditors in the market (market contracting). The fourth is
audit services provided in a combination of the former three governance
structures (plural forms). The last category, plural forms, did not apply to our
level of analysis as we lacked sufficiently detailed data to use this category,
but the entire Finnish market might be considered as belonging to this
category.
Performance Auditing
We assumed that performance auditing is related to human asset specificity
and we had reason to believe that the municipal auditors, during the 1990s,
emphasised performance audit more than they previously had done. (Per-
formance audit was made mandatory by the Local Government Acts, in both
countries, during the 1990s.) In order to measure changes (increase) at the
municipal level, we included a dummy variable with the value 0 if there was
no performance audit in either 1996/97 or 2000 or performance audit in
both 1996/97 and in 2000 (meaning no change) and the value 1 if there was
no performance audit in 1996/97 and some performance audit in 2000. Five
of the 47 Finish municipalities (10.6%) went from no performance audit in
1997 to some in 2000. However, our Finnish data were incomplete on this
issue. For instance, one Finnish municipality (Joroinen) had performance
audit in 1997 but not in 2000. This was coded as 0 because we had no reason
to assume that the municipality did not undertake any performance audit,
for instance, internally, in 2000. In 1996, 10.7% of the 298 municipalities in
Norway had no performance audit, but in 2000 all the Norwegian munici-
palities in the sample had at least some performance audit. In any case, the
trend of increased performance audit has required the focused training and
the recruitment of staff with specific competence, and we have identified this
# Blackwell Publishing Ltd 2004
GOVERNANCE STRUCTURES AND CONTRACTING OUT 457
as being an increase in human assets specificity in the municipal audit. Based
on the above data and on previous studies (Johnsen et al., 2001), we had no
reason to believe that these developments were significantly different at the
national level in the two countries that we have studied.
In addition to governance structure and the change in human asset
specificity, as independent variables, we also included a number of control
variables that were used in previous audit market studies.
Political Preferences
Banker et al. (1992) argued that political competition is a distinguishing
characteristic of the public sector, and that this variable is a highly relevant
one to include, as an explanatory variable for contracting out municipal
auditing. We used the ratio of left-wing representatives in the municipal
councils, as an indicator to measure political preferences, assuming that the
preferences of the majority or of the largest group are influential in the
political competition (Baber, 1983). In Finland, the political preference
variable included representatives from the Social Democratic Party of
Finland, the Left Alliance, the For Peace and Socialism – Communist Worker’s
Party, and the Communist Party of Finland. In Norway, this variable
included representatives from the Labour Party, the Socialist Left Party,
the Red Election Alliance, and the Norwegian Communist Party. We
expected that the higher the ratio of left-wing representatives in the muni-
cipal council, the more the council would rely upon unified governance or
mixed ownership (relational contracting) rather than market contracting.
Decisions on governance structures and contracting out require some time
in order to be implemented. Consequently, we measured political prefer-
ences on 1 January, 1997, in Finland, and after the 1995 local elections in
Norway, in order to conform to the regression analysis assumption that cause
should precede effect.
Auditor’s Size
We identified this variable as the auditors’ turnover in 1999, for Finland,
and as gross municipal audit fees, including estimated but not actually paid
property rents, IT support and other costs that the municipalities granted
the municipal auditors in 2000, for Norway. We expected that the large
auditors, due to their anticipated relatively high market power and compe-
tence/expertise, were positively related to audit fees per capita.
Municipal (Auditee) Size
Many studies have focused upon the relation between auditing fees and
auditee (organisational) size (Baber et al., 1987; Rubin, 1988; and Copley,
# Blackwell Publishing Ltd 2004
458 JOHNSEN, MEKLIN, OULASVIRTA AND VAKKURI
1989). All three studies found significant relationships between local govern-
ment audit fees and government size. In these studies, municipal size was
used as proxy for audit complexity. We expected that economy of scale and
the large municipalities’ relatively strong bargaining power would result in
lower audit fees per capita for large municipalities than for small munici-
palities. We used two indicators to measure municipal size: gross municipal
operating income in 1999, and the number of municipal inhabitants on
31 December, 1999, in Finland, and on 1 January, 2001, in Norway.
The smallest and the largest municipal population figures in our sample
were 1,161 and 209,667, in Finland, and 232 and 508,726, in Norway,
respectively.
Municipal (Audit) Financial Risk
During the 1990s there has been increasing attention on audit quality and on
auditors’ independence and financial responsibility. Auditor independence is
defined as the probability that the auditors will report a discovered breach
(Watts and Zimmerman, 1983), and as the absence of collusion between the
auditor and the manager of the client firm (Lee and Gu, 1998). There is an
increasing risk for the external auditors to be financially liable. We expected
that external auditors would raise their fees in order to compensate for the
anticipated financial risk involved in the audit. We used debt per capita
defined as long-term liabilities divided by municipal inhabitants (Rubin,
1988), to measure the financial risk in the municipalities. We expected
a positive relationship between debt per capita and audit fees per capita. More-
over, we expected this relationship to be stronger in Finland than in Norway.
Agency Costs and Municipal Dependence on Constituency Groups
Public officials may have incentives to demonstrate accountability to major
stakeholders. Ward et al. (1994) found some evidence indicating that audit
fees were positively related to agency variables that measured the extent of
taxpayer funding of services. We measured the agency cost variable with
municipal personal income taxes per capita as indicator. If the public
officials use audit to demonstrate or verify accountability, then we could
expect relatively high audit fees in highly taxed environments.
Municipal Urbanisation/Centrality
There were reorganisations in the municipal audit markets in the 1990s, in
Norway and Finland, resulting in bigger auditors in Norway and the
entrance of private audit firms into the market in Finland. We expected
that these new audit entities had their offices centrally located and that
auditing centrally located urban municipalities would be less expensive
# Blackwell Publishing Ltd 2004
GOVERNANCE STRUCTURES AND CONTRACTING OUT 459
than auditing more peripheral and rural municipalities. We included muni-
cipal urbanisation/centrality as an explanatory variable for relative audit
production costs that could lead to differences in audit fees. The degree of
municipal urbanisation/centrality was measured with an ordinal level scale.
The municipal urbanisation/centrality category indicator for Finland was
originally: 1¼ urban, 2¼ densely populated, 3¼ rural. We reversed this
scale to match the Norwegian indicator. For Norway we applied a four-
category centrality scale used by Statistics Norway. The categories were
originally: 0¼ least central, 1¼ less central, 2¼ quite central, and 3¼most
central. We merged the two lowest categories into the value 1.
Performance Audit Productions Costs
Because the Finnish data were incomplete regarding inclusiveness of all
relevant audit costs, we added a dummy variable for Finland that measured
the inclusion of performance audit in the total audit fees. Thirty percent of
the Finnish and 100 percent of the Norwegian municipalities had some
performance audit included in our audit fees data for 2000.
Analysis and Comparison
We assumed that both non-strategic (environmental) and strategic uncer-
tainty (opportunism), and transaction frequency were relatively equal in the
municipal services and in the municipal audit for Finland and Norway. On
the other hand, there are some issues that make direct comparisons of the
data between the two countries difficult. There are important institutional
differences regarding accounting and audit between many European coun-
tries (Maijoor et al., 2000). Due to the fact that the two countries have
dissimilar institutions and possibly different content in their audit services,
the municipal audit fees cannot be directly compared between the two
countries. There are several possible reasons for pre and post contracting
out differences in the content of the audit fees.
First, different traditions and institutions may have shaped the audit
markets in varying ways. Norwegian local government may have been
more under the control of central government, implying relatively extensive
use of municipal audit. Finland has undergone considerable restructuring
in the economy during the recession in the early 1990s. The recession in
Finland was linked to the collapse of its major trading partner the Soviet
Union in 1989, to the poorly designed deregulation of financial markets
in the 1980s, and other related unsuccessful policies and institutions. (See
Kiander and Virtanen, 2002.) Data from OECD Productivity Database
show that the annual average growth rate in the Finnish GDP was �0.8%
for the 1990–1995 period. In comparison, the annual average growth rate
in GDP in Norway was 3.9% in the same period. The accompanying
# Blackwell Publishing Ltd 2004
460 JOHNSEN, MEKLIN, OULASVIRTA AND VAKKURI
restructuring in the Finnish economy, including the public sector, during this
period may have resulted in better long term efficiency relative to Norway.
Finnish local government had only one tier while Norway had two. Finnish
municipalities and auditors were on average larger than Norwegian munici-
palities and auditors. Finnish government had harmonised the municipal
accounting system to accrual accounting. These factors and possibly others,
may have resulted in different economies of scale and transaction costs in
the municipal audit and subsequently led to different governance structures,
audit activities, audit costs and accounting practices between the two coun-
tries.
Second, the accounting (book keeping) practices may have been different
in the two countries. In Norway, the municipal audit fees covered financial,
management and performance audit. In Finland, the audit committees
were responsible for contracting the financial auditors and for producing
management and the annual performance audit. However, in many munici-
palities audit firms provided assistance in the performance audit process. It is
likely that some costs related to management and performance auditing, in
Finland, were not included in the Finnish audit fees data, while these costs
were encompassed by the Norwegian audit fees data. For those reasons,
we think that the Norwegian audit fees, which were the highest, measured
as municipal audit fees per capita, may have contained more costs related
to management and performance audit than was the case for the Finnish data.
Lastly, and perhaps most importantly, contracting out in Finland may
have ‘pruned’ the content of the audit, relative to Norway, in such a way that
the audit content and perhaps the quality of the audit are incommensurable
between the two countries. We assumed that there were different market
regulations and accounting institutions within each of these two countries
and that there were probably significant differences in the pre and post
contracting out audit content in Finland compared to the audit in Norway.
The data on municipal audit fees, for these reasons, are not directly com-
mensurable between Finland and Norway. As a consequence, we performed
separate analyses for each country and compared our findings.
RESULTS
Univariate and Bivariate Analysis
The descriptive statistics of mean, standard deviation and correlation coeffi-
cients are given for Finland and Norway, in Tables 1 and 2, respectively.
Eighty-five percent of the Finnish and 98 percent of the Norwegian munici-
palities in our sample used auditors with mixed ownership. (In fact, in 2000,
only 9 (2%) of all the 435 municipalities in Norway had unified governance.
Our sample included 7 of the 9 municipalities with unified governance.)
Thus, the answer to our first research question is that the municipalities in
both countries predominantly used auditors with mixed ownership (rela-
tional contracting) as their governance structure. The correlation coefficients
between governance structure and gross municipal operating income was
0.33 (p< 0.05) in Finland, and �0.55 (p< 0.01) in Norway. These figures
indicate that the small municipalities in Finland mostly used market con-
tracting and the small municipalities in Norway mostly used relational
contracting. The negative correlation of �0.13, in Finland, between the
ratio of left-wing representatives and governance structures corroborated
our expectation that municipalities with left-wing councils had relatively
less market-based governance structures. The correlation had the same
sign, but was hardly noticeable (�0.00), in Norway. However, since Norway
had no contracting out, it seems reasonable to assume that there were
no specific political preferences regarding the choice between internal or
district auditors in Norway. The correlation coefficients between audit fees per
capita and municipal inhabitants were �0.39 (p< 0.01) in Finland, and �0.20
(p< 0.01) in Norway, indicating economy of scale in the municipal audit.
Further analysis showed that there were decreasing audit fees per capita until
ca. 16,000 inhabitants, in Finland, and ca. 10,000 inhabitants, in Norway.
Multiple Regression Analysis of Audit Fees Per Capita
The regression analyses in Tables 3 and 4 address our second research
question on the relationship between governance structure and audit fees.
Because gross municipal operating income and municipal inhabitants were
highly correlated (see Tables 1 and 2), we only included municipal inhabi-
tants as the independent variable in the regression equation, in order to
avoid co-linearity.
The governance variable with three values was transformed into a dummy
variable for each country. For Finland, the governance variable had the
value 0 for mixed ownership and 1 for market contracting. For Norway, the
variable had the value 0 for unified governance and 1 for mixed ownership.
The resulting general regression equation is given by:
ðMunicipal audit fees per capitaÞ ¼ b0 þ b1ðgovernance structureÞþ b2aðchange in performance auditingÞ þ b3ðleftist political preferencesÞþ b4ðauditor sizeÞ þ b5ðmunicipal sizeÞ þ b6ðdebt per capitaÞþ b7ðmunicipal income tax per capitaÞ þ b8ðmunicipal centralityÞþ b2bðperformance auditing ½only for Finland�Þ þ e:
To test for the assumption of normally distributed data in regression
analysis, we run one-sample Kolmogorov-Smirnoff tests of normality. We
found all the size related variables, auditor turnover, municipal inhabitants
# Blackwell Publishing Ltd 2004
464 JOHNSEN, MEKLIN, OULASVIRTA AND VAKKURI
Table
3
RegressionofMunicipalAuditFeesa
PerCapitain
Finland(N
¼47)
ModelI(basicmodel)
ModelII
(som
elogtransformed
independentvariables)
ModelIII(som
elogtransformed
independentvariablesandlog
transformed
dependentvariable)
Independent
Variables
Regression
Coefficients
Standardised
Regression(Beta)
Coefficients
Regression
Coefficients
Standardised
Regression(Beta)
Coefficients
Regression
Coefficients
Standardised
Regression(Beta)
Coefficients
Constant
14.07
25.19*
4.99***
Governance
structure
b�6.30
�0.80
�2.15
�0.27
0.23
0.16
Changein
perform
ance
audit
�2.56*
�0.31
�1.11
�0.13
0.06
0.04
Left-w
ingrepresentatives
7.71*
0.38
5.50
0.27
�0.55
�0.15
Auditorturnovera
�0.00
�0.51
––
––
(Log)Auditorturnovera
––
�0.00
�0.05
0.00
0.14
Operatingincomea
�0.00
�0.04
––
––
(Log)Operatingincomea
––
�0.00***
�0.62
�0.00***
�1.04
Debtpercapitaa
�0.00
�0.15
�0.00
�0.07
0.00
0.01
Taxpercapitaa
�0.00
�0.28
�0.00
�0.22
0.00
0.05
Centrality
�0.50
�0.16
0.93
0.30
0.13
0.24
Perform
ance
audit
1.77*
0.29
1.34
0.22
�0.00
�0.01
AdjustedR2
0.25
0.39
0.75
F-value
2.67**
4.28***
16.52***
Notes:
aFIM
.bDummyvariable
(Finland):0¼mixedownership
and1¼marketcontracting.
Significance
levels:*¼p<0.10,**
¼p<0.05,***¼p<0.01.
# Blackwell Publishing Ltd 2004
GOVERNANCE STRUCTURES AND CONTRACTING OUT 465
Table
4
RegressionofMunicipalAuditFeesa
PerCapitain
Norw
ay(N
¼298)
ModelI(basicmodel)
ModelII
(som
elogtransformed
independentvariables)
ModelIII(som
elogtransformed
independentvariablesandlog
transformed
dependentvariable)
IndependentVariables
RegressionCoefficients
Standardised
Regression(Beta)
Coefficients
Regression
Coefficients
Standardised
Regression(Beta)
Coefficients
Regression
Coefficients
Standardised
Regression(Beta)
Coefficients
Constant
41.74
452.74***
19.22***
Governance
structure
b�16.07
�0.04
�67.70***
�0.17
�2.53***
�0.18
Changein
perform
ance
audit
20.89**
0.11
24.07***
0.12
0.62***
0.09
Left-w
ingrepresentatives
�24.59
�0.02
36.70**
0.09
�0.16
�0.01
Auditorturnovera
0.00***
0.18
––
––
(Log)Auditorturnovera
––
0.01***
0.17
0.00***
0.12
Operatingincomea
�0.00***
�0.23
––
––
(Log)Operatingincomea
––
�0.04***
�0.69
�0.00***
�0.78
Debtpercapitaa
0.00***
0.30
0.00***
0.22
0.00***
0.13
Taxpercapitaa
0.00***
0.40
0.00***
0.40
0.00***
0.31
Centrality
�24.95***
�0.35
�8.76***
�0.12
�0.15
�0.06
Perform
ance
audit
––
––
––
AdjustedR2
0.40
0.66
0.66
F-value
25.71***
73.59***
72.68***
Notes:
aNOK.
bDummyvariable
(Norw
ay):0¼unifiedgovernance
and1¼mixedownership.
Significance
levels:*¼p<0.10,**
¼p<0.05,***¼p<0.01.
# Blackwell Publishing Ltd 2004
466 JOHNSEN, MEKLIN, OULASVIRTA AND VAKKURI
and gross municipal operating income, as well as the dependent variable
audit fees per capita, to be significantly non-normally distributed (p< 0.05).
We also used models with natural log transformed variables in the analyses,
but the auditor turnover variable remained significantly non-normally dis-
tributed. In order to test for the assumption of no co-linearity between
the independent variables, we examined the tolerance and variance inflator
(VIF) indexes. Only governance structure (dummy variable) and auditor turnover,
in Finland, (where Audiator OY was very large and had a market share of 85%, in
our sample) were the independent variables in our models that exceeded the
critical values of being below 0.10 or above 10, respectively.
Tables 3 and 4 present the results from ordinary least square (OLS)
multiple regression analyses with municipal audit fees per capita as the
dependent variable for Finland and Norway, respectively. The main hypo-
thesis in our study, hypothesis 3, that competition (less use of unified govern-
ance) is related to relatively low audit fees per capita, was corroborated in
both Finland and Norway, but the relationship was significant (p< 0.01) only
for two models with the Norwegian data. Change in performance audit as
proxy for increased asset specificity was positive, as expected, in one model
for Finland, and positive and significant (p< 0.05) for all three models for
Norway. The ratios of left-wing representatives in the municipal councils
were both negatively and positively related to audit fees per capita in Finland
and Norway, resulting in an indeterminate conclusion on this issue. There
was an indeterminate relationship in Finland, but a positive relationship
(p< 0.01) in Norway, between auditors’ turnover and audit fees per capita.
The coefficients for the relationship between audit fees per capita and the
municipalities’ bargaining power indicator (log of) municipal income, were
negative in Finland and Norway, and significantly so (p< 0.01) in five
models. Our expectation of a positive relationship between debt per capita
and audit fees per capita was corroborated in one model, for Finland, and
significantly (p< 0.01), in all three models for Norway. However, Nordic
municipalities have a relatively high credit rating. As a consequence, exter-
nal auditors may not consider their own financial risk to be a matter of
concern in their appraisals of the financial situation of their municipal
clients, and the issue may not be a major decision variable in any of the
Nordic municipal audit markets. Our expectation that agency costs (audit
fees per capita) were positively related to municipal personal income taxes
per capita, was supported in one model for Finland, and was significantly
supported (p< 0.01) for all three models in Norway. We expected a negative
relationship between centrally located municipalities and audit fees per
capita, and this expectation was supported in one model in Finland, and it
was significantly supported (p< 0.01) in all models for Norway. In Finland,
the inclusion of performance audit in the data was positively related to audit
fees per capita in two models and the relationship was significant (p< 0.10)
in one of the three models.
# Blackwell Publishing Ltd 2004
GOVERNANCE STRUCTURES AND CONTRACTING OUT 467
Multivariate Analysis of Governance Structure and Audit Fees Per Capita
The above analyses documented that governance structure and investment
in asset specificity do seem to have an effect on audit fees. Furthermore, the
auditors’ and the auditees’ relative bargaining strengths and agency and
production costs may also matter. We may now turn our attention to how
the actual governance structures conform to production and transaction
costs reasoning, in greater detail, keeping in mind that transaction costs
are unobservable variables in our research design.
Table 5 shows the results for audit fees per capita by governance struc-
tures and municipal size in Finland and Norway. All of the 1996 and 1997
audit fees in the analyses have been inflated to 2000 currency, using the
consumer price index (CPI). Table 5 shows that no Finnish municipality in
our sample (apparently) used unified governance2 and obviously no Nor-
wegian municipality used market contracting. The Norwegian data showed
that no municipalities with less than 15,000 inhabitants used unified govern-
ance. These findings corroborate our first hypothesis. Hypothesis 1 states
that large organisations have more unified governance of audit than small
organisations. However, it may be the case that some Finnish municipalities
had unified governance of some audit but did not include or report the costs
for this as audit fees. Interestingly, most of the Finnish (85%) and even more
of the Norwegian municipalities (98%) chose to use an auditor with mixed
ownership, despite the fact that relational contracting was relatively costly,
measured as mean audit fees per capita, for most of the municipalities in
both Finland and Norway. Relational contracting was the dominant govern-
ance structure in the group of large municipalities in both Finland and
Norway. This finding supports hypothesis 2.
Table 5 presents the data for testing hypothesis 3, that increasing compe-
tition reduces audit fees. It shows that the mean audit fee per capita was
apparently much lower in Finland than in Norway. The audit fees per capita
were FIM 6.30 in Finland, and NOK 100.30 in Norway, corroborating our
notion that there might be substantial differences between the two countries
in the content of the audit. The findings support the notions that competi-
tion and contracting out may have effects on the content of the audit and on
audit fees. (In order to compare the audit fees directly, the currencies must
be converted. See the notes to Table 5.) Contracting out in Finland might
have resulted in relatively low audit fees due to high efficiency and to
lowballing. At the same time, audit fees in Norway may initially have been
relatively high, due to the lack of competition in the Norwegian municipal
audit market. However, and as stated above, audit fees should not be
compared directly between the two countries. Table 5 also shows that the
mean municipal audit fees per capita, in Finland, rose by 7.4% from 1997 to
2000. This was an annual increase of 2.5%. The audit fees per capita, in
Norway, rose by 6.5% from 1996 to 2000. Audit fees per capita, in Norway,
# Blackwell Publishing Ltd 2004
468 JOHNSEN, MEKLIN, OULASVIRTA AND VAKKURI
Table
5
MeanMunicipalAuditFeesPer
Capitain
Finland(N
¼47)andNorw
ay(N
¼298)byGovernance
Structure
andMunicipalSize
Finland(FIM
)Norway
(NOK)
Governance
1997
2000
1996
2000
Structure
Smalla
Largeb
Total
Smalla
Largeb
Total
Smalla
Largeb
Total
Smalla
Largeb
Total
Unified
––
––
––
–50.28
50.28
–53.43
53.43
governance
––
––
––
–(10.96)
(10.96)
–(10.18)
(10.18)
––
––
––
–[37.52–70.74]
[37.52–70.74]
–[40.92–70.99]
[40.92–70.99]
n¼0
n¼0
N¼0
n¼0
n¼0
N¼0
n¼7
N¼7
n¼7
N¼7
Relational
6.45
5.51
6.19
7.19
5.70
6.78
101.05
50.51
94.45
109.14
50.40
101.47
contracting
(3.13)
(2.68)
(3.01)
(2.79)
(2.46)
(2.76)
(56.72)
(12.67)
(55.74)
(62.29)
(13.68)
(61.55)
[2.26–13.34]
[1.46–10.65]
[1.46–13.34]
[2.38–14.38]
[3.02–10.91]
[2.38–14.38]
[34.23–457.59]
[21.40–81.47]
[21.40–457.59]
[37.79–560.56]
[18.29–84.36]
[18.29–560.56]
n¼29
n¼11
N¼40
n¼29
n¼11
N¼40
n¼253
n¼38
N¼291
n¼253
n¼38
N¼291
Market
5.25
2.01
4.32
4.77
1.81
3.92
––
––
––
contracting
(2.89)
(0.57)
(2.85)
(0.95)
(0.44)
(1.65)
––
––
––
[2.14–9.54]
[1.60–2.41]
[1.60–9.54]
[3.20–5.61]
[1.50–2.12]
[1.50–5.61]
––
––
––
n¼5
n¼2
N¼7
n¼5
n¼2
N¼7
n¼0
n¼0
N¼0
n¼0
n¼0
N¼0
Total
6.28
4.97
5.92
6.83
5.10
6.36
101.05
50.47
93.41
109.14
50.87
100.34
(3.08)
(2.79)
(3.03)
(2.74)
(2.68)
(2.80)
(56.72)
(12.30)
(55.51)
(62.29)
(13.14)
(61.27)
[2.14–13.34]
[1.46–10.65]
[1.46–13.34]
[2.38–14.38]
[1.50–10.91]
[1.50–14.38]
[34.23–457.59]
[21.40–81.47]
[21.40–457.59]
[37.79–560.56]
[18.29–84.36]
[18.29–560.56]
n¼34
n¼13
N¼47
n¼34
n¼13
N¼47
n¼253
n¼45
N¼298
n¼253
n¼45
N¼298
Notes:
aSmallmunicipalities<15,000inhabitants.
bLargemunicipalities�15,000inhabitants.
FIM
andNOK
in2000valuespercapita.The1997auditfeeswere
inflatedto
2000valuesbyusingtheFinnishconsumerprice
index(C
PI).TheCPIwas101.8
in1997and108.0
in2000with1995asthebase
year.The1996auditfeeswere
inflatedto
2000valuesbyusingtheNorw
egianconsumerprice
index(C
PI).TheCPIwas95.3
in1996and106.7
inDecember2000with1998asthebase
year.
Standard
deviationsin
parenthesis()
andminim
um
andmaxim
um
valuesin
brackets
[].
Thecu
rrency
exch
angerates31December,2000,were:1NOK¼0.7202FIM
¼0.07641GBP¼0.1211EURO.
# Blackwell Publishing Ltd 2004
GOVERNANCE STRUCTURES AND CONTRACTING OUT 469
rose only by 1.6% annually, from 1996 to 2000. On the basis of these
statistics, we can see that the annual increase in audit fees per capita was
higher in Finland, which we assumed had become relatively more competi-
tive, than in Norway. However, this finding must be seen in relation to the
high initial audit fees per capita, in Norway, the decrease in population in
some Finnish municipalities, and to the possible lowballing in Finland when
the market was deregulated in 1997. Thus, hypothesis 3, which claims that
more competitive markets have relatively lower audit fees, was corroborated,
but the notion that the most competitive market also experienced the lowest
increase in audit fees, was not supported. Note, however, that if the audit
fees had been measured in terms of the costs of external audit per audit day
purchased, the conclusion is different. These costs did not rise in Finland,
from 1997 to 2000, when the costs are adjusted for inflation. This finding
would support our hypothesis.
Hypothesis 4 predicted relatively volatile audit fees in market contracting.
In Finland, the audit fees in relational contracting rose by 9.5% while the
audit fees in market contracting fell by �9.3%. There are, however, too few
observations of municipalities with market contracting to be able to test
hypothesis 4 in any rigorous manner. Consequently, based on our evidence,
there was at best only partial corroboration for hypothesis 4.
There may be several reasonable explanations for the many substantial
changes in audit fees, in the Finnish market, during the period of our study,
which was the first contract period for auditors after the 1997 accounting
reform. There may have been uncertainties regarding the new accounting
and market regulations and inexperience in negotiating new and incomplete
contracts. There may well have been renegotiations concerning the content
of the audit and the level of audit fees, in many municipalities, during the
first four-year contract period. In addition, many Finnish municipalities
experienced a decline in population during this period of time, which may
partially explain the relatively large increase in audit fees per capita in
relational contracting. Moreover, in Finland, many municipalities owned
Audiator OY indirectly, through the Association of Finnish Local and
Regional Authorities. This may well have resulted in relatively weak muni-
cipal bargaining power, relative to this auditor, since Audiator OY was so
very large. In comparison, findings for the large municipality group with
mixed ownership, in Norway, show that audit fees per capita fell, from 1996
to 2000. Finally, for market contracting, there may well have been low audit
fees that were due to lowballing.
DISCUSSION AND CONCLUSIONS
In this study we have documented three main findings: First, widespread use
of relational contracting in the municipal audit market in both Finland and
# Blackwell Publishing Ltd 2004
470 JOHNSEN, MEKLIN, OULASVIRTA AND VAKKURI
Norway; second, low municipal audit fees (possibly partly due to lowballing)
and then a growth in audit fees in Finland; and third, relatively high (initial)
municipal audit fees in Norway.
A striking feature of the data was the large proportion of the municipal
audit that was governed through mixed ownership (relational contracting)
despite the fact that the audit fees per capita were seemingly relatively high
for this governance structure in both Finland and Norway. On the other
hand, the audit fees did not reflect transaction costs. If the majority of the
municipalities chose the most efficient governance structure, then there must
have been the anticipation of relatively high transaction costs, low audit
quality, or high political costs associated with using municipal auditors in
unified governance or contracted from the market. Low decision relevance
may have stemmed from a paucity of performance audit relative to economic
and organisational needs or called for by the government or by the munici-
palities. Such considerations have been major arguments for the establish-
ment of district audits in Norway. In Norway, 98% of the municipalities
chose district audits rather than unified governance. To a large extent,
Finnish municipalities have continued to use the auditors they had before
the 1997 reform (Audiator OY). However, one criterion for selecting market
contracting, for small Finnish municipalities, was to acquire general eco-
nomic and financial expertise, which means management audit and consult-
ing services, in addition to the traditional municipal audit. The majority of
the Finnish municipalities chose relational contracting, even though market
contracting is seemingly cost efficient. This finding strengthens the case for
asset specificity and economy of scale, as important factors for deciding upon
the governance of municipal audit. The finding of extensive use of relational
contracting supports Williamson’s (2000) call for an enhanced attention to
hybrid forms (mixed ownership, relational contracting) as a relevant and
important governance structure in studies of institutional economics and
organisation. A tentative conclusion might be that municipalities chose
relational contracting because the sum of audit fees, transaction costs and
political costs were minimised and the net benefit of audit were optimised in
this governance structure.
Relational contracting may have several economic advantages. Engaging
in long-term relational contracting instead of market contracting could avoid
recurrent start-up costs, which are high in the audit market (DeAngelo,
1981). Relational contracting of audit in mixed ownership could also facil-
itate economy of scale that only large municipalities with unified governance
or large accounting firms could otherwise have achieved. Mixed ownership
allows auditors to audit more than one municipality and to become larger
than would be the case in unified governance. These auditors can specialise
and invest in idiosyncratic human capital, a requirement to be able to
competently undertake public sector specific activities, like performance
auditing.
# Blackwell Publishing Ltd 2004
GOVERNANCE STRUCTURES AND CONTRACTING OUT 471
Relational contracting may have the lowest transaction costs for most
municipalities. Even though there is extensive uncertainty and asset specifi-
city in auditing, the high level of commitment in relational contracting (De
Ruyter and Wetzels, 1999) may save the audit clients the costs of continu-
ously monitoring the auditor (DeAngelo, 1981). Simultaneously, the govern-
ance structure of mixed ownership may facilitate sharing economic risk and
monitoring costs by the municipalities. The mixed client and ownership
status may also give municipalities an enhanced bargaining position to
counter opportunism (and reduce audit fees) relative to the auditors. More-
over, by relying on relational contracting the municipalities can potentially
avoid ex post contracting problems, in the circumstance that the initial
contract provides the auditor with a first mover advantage that prohibits
effective competition for subsequent contracts. Potential contracting out
would always pose the threat of competition and facilitate incentives for
efficiency. Finally, and perhaps most importantly, the long-term relationship
may develop trust in a way that substantially reduces opportunism and
related transaction costs, limiting the contingencies of incomplete contracts.
The majority of the municipalities may have chosen relational contracting in
order to avoid being dependent upon internal production or (large) external
accounting firms.
In addition to production costs and transaction costs, there may also be
political costs that have influenced the dominant choice of mixed ownership.
We have noted that Norwegian municipal councils were apparently more
‘left-wing’ than Finnish municipal councils, and that Norway did not allow
contracting out municipal audit while Finland did. Contracting out munici-
pal audit can become a contestable political question and generate substan-
tial political costs. As such, the extensive use of mixed ownership could
reduce the political costs relative to contracting out, as well as serve eco-
nomic purposes. Since mixed ownership is not pure market contracting, it
would probably not generate as much political resistance, from unions and
the Left, as full market contracting would have done.
Our study has essentially been an analysis of (a) market(s) in transition.
The research design has major limitations with regard to internal and
external validity claims. Results must be regarded as preliminary. Based on
the available evidence, we have no reason to put forward any claim that the
widespread use of relational contracting (instead of market contracting) in
municipal audit will continue. It may be the case that municipal audit has a
product life cycle (PLC) like many ordinary products. If so, then there may
very well be new developments where and when municipal audit technology
becomes relatively more common. Over time, a fully developed market
supply of municipal audit may evolve. An increase in market supply may
reduce the incentives to provide municipal audit in unified governance or in
relational contracting and this may be especially true for small municipali-
ties. If municipal audit conforms to PLC reasoning, then the seemingly
# Blackwell Publishing Ltd 2004
472 JOHNSEN, MEKLIN, OULASVIRTA AND VAKKURI
dominant governance structure of relational contracting/mixed ownership
may be a transitional state. When more municipalities get experience with
the diversity of available governance structures, they may begin to use
several governance structures simultaneously. Municipal audit markets
could evolve into more plural forms, subject to production costs, transaction
costs and the hazards of fragmentation.
Municipal audit fees in Finland had a larger annual increase than the
audit fees in Norway in the late 1990s. Price increases in the Finnish market
may have many explanations. Auditors may have competed for market
shares in the initial rounds, after the 1997 reform, by lowballing and reduced
audit fees in general. It is possible that audit fees in Finland rose because
pricing strategies within private audit firms included attempts at making
profits in the longer run. Finnish municipalities may have had problems
with incomplete contracts. In particular, too little audit time may have been
bought initially, due to inexperience with competitive tendering. More audit
time may subsequently have been bought at higher prices in later years.
Increased competition and contracting out municipal audit could lead to
higher audit fees, in the long run, in the market than would be the case with
a regulated market. This potentiality would depend upon a slew of factors,
including accounting regulation, uncertainty, initial market efficiency, bar-
gaining power, entry barriers and market rivalry. However, since we did not
have data for Finland prior to deregulation and since the long-run outcome
in Finnish and Norwegian markets are still uncertain, we can only speculate
about this matter at this time.
More inclusive book keeping of performance audit costs might explain
some of the relatively high audit costs in Norway as compared to Finland.
Despite this, we could still make a case for questioning the relatively high
(initial) costs in Norway. We might claim that inefficiency results whenever
non-optimal governance structures in the municipal audit market are cho-
sen. We might further claim that inefficiency will result when there is an
absence of sufficient competition in the municipal audit market. Finally, we
might claim that inefficiency will result when there are high monitoring costs
associated with keeping an idiosyncratic municipal accounting system and a
two tier local government structure with many small municipalities. Unfor-
tunately, at present, we have no evidence to firmly support any of these
claims.
Much of the preceding analysis of transaction costs has assumed rational
choice. But we might as easily have assumed that Norway and Finland have
organised the municipal audit market in an inefficient manner. That could
well lead to the understanding that if relatively more municipal audit were
contracted out, then the overall municipal audit market performance would
improve, at least in Norway. The Norwegian government could alternatively
or additionally harmonise the municipal accounting system to accrual
accounting, which could lower the overall municipal monitoring costs and
# Blackwell Publishing Ltd 2004
GOVERNANCE STRUCTURES AND CONTRACTING OUT 473
reduce the entry barriers in the municipal audit market for private auditors.
Moreover, the Norwegian government could re-organise the local govern-
ment structure resulting in larger municipalities with relatively more bar-
gaining power. However, these considerations are tentative because we only
had a small number of cases of unified governance and market contracting
available for our comparative analysis.
There are at least three viable research strategies for pursuing studies of
municipal audit and transaction costs further. A first strategy is to study the
development of municipal audit fees as the municipal auditors re-organise
and merge under the threat of possible market competition. Norway, during
the 1990s and early 2000s, may be an opportune case. A second strategy is to
study audit and transaction costs when municipal accounting systems are
harmonised with accrual accounting and generally accepted accounting
practice. Sweden’s reform in 1987, and Finland’s reform in 1997, may be
candidate cases. A third strategy is to pursue further studies of deregulated
municipal audit markets with longitudinal panel data studies or with case
studies of municipal audit contracting. Common for all three research
strategies is the need to develop agreed upon definitions and readily avail-
able and comparable statistics on municipal audit fees. In conclusion:
Although the available data in this study to some extent has supported
many of the TCA and contracting out propositions, how contracting out
affects municipal audit in the long run is still an open question.
ACKNOWLEDGEMENTS
Thanks to Juha-Pekka Martikainen for assistance in data collection for Finland.
Thanks to Liv Bente Hannevik Friestad, Ole Kristian Rogndokken and the
Norwegian Municipal Auditor Association for their help and comments in the
data collection for Norway. The authors acknowledge the comments received
when earlier versions of this paper were presented at: the 16th Nordic Conference
on Business Studies, Uppsala University, Sweden, August 16–18, 2001; the Nor-
wegian Association for Municipal Auditors Management Meeting, Gardermoen,
24 October 2001; the 10th Nordic Conference in Municipal Research, Turku,
23–25 November 2001; the Norwegian Research Council Workshop on Public
Sector in Transition, Høsbjør, 26–27 November 2001; the 11th Annual Confer-
ence in Management, Norwegian School of Economics and Business Administra-
tion, Bergen, 10–11 January 2002; the European Institute for Advanced Studies in
Management (EIASM) International Conference on Accounting, Auditing and
Management in Public Sector Reforms, University College Dublin, September 5–
7, 2002; and at the Nordic Workshop on Transaction Cost Economics in Business
Administration, Norwegian School of Economics and Business Administration,
Bergen, 20–21 June 2003. Lastly, the authors acknowledge the comments from
Lawrence Young and two anonymous referees.
# Blackwell Publishing Ltd 2004
474 JOHNSEN, MEKLIN, OULASVIRTA AND VAKKURI
NOTES
1 One could possibly argue that classical market contracting of audit does not exist and that allaudits are either internal or relational. When a municipality buys audit from an accountingfirm, the contract is a relation for a specific period and the municipality still has to assist inthe process. Moreover, in order to secure buyer competence, some supplier competence maybe maintained by keeping some audit internally, at least in large organisations. Contractingout of mandatory audit could thus be classified as (bi-lateral) relational contracting and onlya few audits, and mostly non-mandatory management services (consulting), are strictlyspeaking market contracting. Even these services could be termed relational contractingbecause one of the motives for lowballing could be to position the auditor in order to obtainconsulting services from the audit client on a regular basis (Firth, 1997 and 2002). Lowballingis the pricing strategy where audit firms quote audit fees that are below their start-up costswith new clients. It is a practice that is based upon the belief that future fees will exceedfuture marginal audit costs (Dopuch and King, 1996). However, in this study we havecategorised the contracting out of mandatory audit as market contracting, but we acknow-ledge that this is an operational definition and that the conceptual definition relative to TCAwarrants more clarification.
2 Unified governance is not allowed in municipal financial auditing in Finland. The LocalGovernment Act requires that municipalities contract out the mandatory financial audit.Furthermore, the certified public sector financial auditors may not work for the auditeeorganisation. This issue has been understood as a matter of auditor independence.
REFERENCES
Baber, W.R. (1983), ‘Toward Understanding the Role of Auditing in the Public Sector’, Journalof Accounting and Economics, Vol. 5, No. 3, pp. 213–27.
——— E.H. Brooks and W.E. Ricks (1987), ‘An Empirical Investigation of the Market for AuditServices in the Public Sector’, Journal of Accounting Research, Vol. 25, No. 2, pp. 293–305.
Banker, R.D., W.W. Cooper and G. Potter (1992), ‘A Perspective on Research in GovernmentalAccounting’, The Accounting Review, Vol. 67, pp. 496–510.
Baumol, W.J, J. Panzar and R. Willig (1982), Contestable Markets and the Theory of Industry Structure(New York, N.Y.: Harcourt Brace Jovanovich).
Bergevarn, L.E., F. Mellemvik and O. Olson (1995), ‘Institutionalization of Municipal Account-ing – a Comparative Study between Sweden and Norway’, Scandinavian Journal of Manage-ment, Vol. 11, No. 1, pp. 25–41.
Bowerman, M. (1996), ‘The Rise and Fall of Value for Money Auditing’, in I. Lapsley andF. Mitchell (eds.), Accounting and Performance Measurement. Issues in the Private and Public Sectors(London: Paul Chapman Publishing), pp. 193–212.
Boyne, G.A. (1998), ‘Competitive Tendering in Local Government: A Review of Theory andEvidence’, Public Administration, Vol. 76, No. 4, pp. 695–712.
Bradach, J.L. and R.G. Eccles (1989), ‘Price, Authority, and Trust: From Ideal Types to PluralForms’, Annual Review of Sociology, Vol. 15, pp. 97–118.
Copley, P.A. (1989), ‘Municipal Audit Fees: A Survey and Analysis’, Government Finance Review,Vol. 5, No. 5, p. 21.
——— and M.S. Doucet (1993), ‘The Impact of Competition on the Quality of GovernmentalAudits’, Auditing: A Journal of Practice and Theory, Vol. 12, No. 1, pp. 88–98.
DeAngelo, L.E. (1981), ‘Auditor Size and Audit Quality’, Journal of Accounting and Economics,Vol. 3, No. 4, pp. 183–99.
De Ruyter, K. and M. Wetzels (1999), ‘Commitment in Auditor–Client Relationships: Ante-cendents and Consequences’, Accounting, Organizations and Society, Vol. 24, No. 1, pp. 57–75.
Domberger, S. and P. Jensen (1997), ‘Contracting Out by the Public Sector: Theory, Evidence,Prospects’, Oxford Review of Economic Policy, Vol. 13, No. 4, pp. 67–78.
Dopuch, N. and R.R. King (1996), ‘The Effects of Lowballing on Audit Quality: An Experi-mental Market Study’, Journal of Accounting, Auditing and Finance, Vol. 11, No. 1, pp. 45–68.
# Blackwell Publishing Ltd 2004
GOVERNANCE STRUCTURES AND CONTRACTING OUT 475
Firth, M. (1997), ‘The Provision of Non-audit Services and the Pricing of Audit Fees’, Journal ofBusiness Finance & Accounting, Vol. 24, Nos. 3/4, pp. 511–25.
——— (2002), ‘Auditor-provided Consultancy Services and their Associations with Audit Feesand Audit Opinions’, Journal of Business Finance & Accounting, Vol. 29, Nos. 5/6, pp. 661–93.
Hatherly, D. and T. Brown (1996), ‘The Measurement of Audit Quality’, in I. Lapsley andF. Mitchell (eds.), Accounting and Performance Measurement. Issues in the Private and Public Sectors(London: Paul Chapman Publishing), pp. 77–86.
Hodge, G.A. (2000), Privatization: An International Review of Performance (Boulder, Colorado:Westview Press).
Johnsen, A., P. Meklin, L. Oulasvirta and J. Vakkuri (2001), ‘Performance Auditing in LocalGovernment: An Exploratory Study of Perceived Efficiency of Municipal Value for MoneyAuditing in Finland and Norway’, The European Accounting Review, Vol. 10, No. 3, pp. 583–99.
Keen, J. (1999), ‘On the Nature of Audit Judgements: The Case of Value for Money Studies’,Public Administration, Vol. 77, No. 3, pp. 509–25.
Kiander, J. and S. Virtanen (eds.) (2002), ‘Economic Crisis in the 1990s. The ResearchProgramme on the Economic Crisis of the 1990s in Finland – Final Report’, Helsinki:Government Institute for Economic Research (VATT), Publications 27:7.
Lee, C.W.J. and G. Zhaoyang (1998), ‘Low Balling, Legal Liability and Auditor Independence’,The Accounting Review, Vol. 73, No. 4, pp. 533–55.
Maijoor, S., R. Meuwissen and L. Quadackers (2000), ‘The Effects of National Institutions onAudit Research: Evidence from Europe and North America’, The European Accounting Review,Vol. 9, No. 4, pp. 569–87.
Marsh, A. (1998), ‘Local Governance: The Relevance of Transaction Cost Economics’, LocalGovernment Studies, Vol. 24, No. 1, pp. 1–18.
Mayston, D.J. (1985), ‘Non-Profit Performance Indicators in the Public Sector’, FinancialAccountability & Management, Vol. 1, No. 1, pp. 51–74.
Monsen, N. and S. Nasi (1996), ‘Local Government Accounting in Finland and Norway:A Historical Note on Cammeralism’, in J.L. Chan, R.H. Jones and K. Luder (eds.), Researchin Governmental and Nonprofit Accounting, Vol. 9, pp. 259–74.
Pollitt, C. and G. Bouckaert (2000), Public Management Reform. A Comparative Analysis (Oxford:Oxford University Press).
Powell, W.W. (1990), ‘Neither Market nor Hierarchy: Network Forms of Organization’, Researchin Organizational Behavior, Vol. 12, pp. 295–336.
Rhodes, R.A.W. (1997), Understanding Governance. Policy Networks, Governance, Reflexivity and Account-ability (Buckingham: Open University Press).
Rittenberg, L.E. and M. Covaleski (1999), ‘Outsourcing the Internal Audit Function: TheBritish Government Experience with Market Testing’, International Journal of Auditing,Vol. 3, pp. 225–35.
Roberts, R.W., G.W. Glezen and T.W. Jones (1990), ‘Determinants of Auditor Change in thePublic Sector’, Journal of Accounting Research, Vol. 28, No. 1, pp. 220–28.
Rubin, M.A. (1988), ‘Municipal Audit Fee Determinants’, The Accounting Review, Vol. 63, No. 2,pp. 219–36.
——— (1992), ‘Municipal Selection of a State or External Auditor for Financial StatementAudits’, Journal of Accounting and Public Policy, Vol. 11, pp. 155–78.
Sanders, G., A. Allen and L. Korte (1995), ‘Municipal Audit Fees: Has Increased Competitionmade a Difference?’, Auditing: A Journal of Practice and Theory, Vol. 14, No. 1 (Spring),pp. 105–14.
Thorne, J., S.A. Holmes, A.S. McGowan, C.A. Strand and R.H. Strawser (2001), ‘The Relationbetween Audit Pricing and Audit Contract Type: A Public Sector Analysis’, Journal ofAccounting and Public Policy, Vol. 20, pp. 189–215.
Ward, D. D., R.J. Elder and S.C. Kattelus (1994), ‘Further Evidence on the Determinants ofMunicipal Audit Fees’, The Accounting Review, Vol. 69, No. 2, pp. 399–411.
Watts, R.L. and J.L. Zimmerman (1983), ‘Agency Problems, Auditing, and the Theory of theFirm: Some Evidence’, Journal of Law and Economics, Vol. 26, pp. 613–33.
Widener, S.K. and F.H. Selto (1999), ‘Management Control Systems and Boundaries of theFirm: Why do Firms Outsource Internal Auditing Activities?’, Journal of Management Account-ing Research, Vol. 11, pp. 45–73.
# Blackwell Publishing Ltd 2004
476 JOHNSEN, MEKLIN, OULASVIRTA AND VAKKURI
Williamson, O.E. (1975), Markets and Hierarchies: Analysis and Antitrust Implications (New York: TheFree Press).
——— (1985), The Economic Institutions of Capitalism. Firms, Markets, Relational Contracting (NewYork: The Free Press).
——— (2000), ‘The New Institutional Economics: Taking Stock, Looking Ahead’, Journal ofEconomic Literature, Vol. 38, pp. 595–613.
# Blackwell Publishing Ltd 2004
GOVERNANCE STRUCTURES AND CONTRACTING OUT 477
Journal of Business Finance & AccountingSeptember/October 2004 Vol. 31 Nos 7 & 8
ContentsCorporate Data and Future Cash Flows
by ALI AL-ATTAR AND SIMON HUSSAIN
Net Present Value-Consistent Investment Criteria Based on Accruals: AGeneralisation of the Residual Income-Identity
by THOMAS PFEIFFER
Can Market Incompleteness Resolve Asset Pricing Puzzles?
by MARK C. FREEMAN
Public Predisclosure Information, Firm Size, Analyst Following, and MarketReactions to Earnings Announcements
by THEODORE E. CHRISTENSEN, TONI Q. SMITH AND PAMELA S. STUERKE
The Recognition and Timing of Deferred Tax Liabilities
by A. GAEREMYNCK AND L. VAN DE GUCHT
The Role of the Options Market in the Dissemination of Private Information
by TIMOTHY CAIRNEY AND JUDITH SWISHER
The Conditional Price of Basis Risk: An Investigation Using Foreign ExchangeInstruments
by JOELLE MIFFRE
Making a Market with Spreads and Depths
by KEE H. CHUNG AND XIN ZHAO
Firm Characteristics as Cross-sectional Determinants of Adverse Selection
by SHANTARAM P. HEGDE AND JOHN B. MCDERMOTT
The Influence of Monetary Conditions on the Response of Interest Rate Futuresto M1 Releases: 1976–1998
by THOMAS MANN AND RICHARD DOWEN
Effects of Bank Funds Management Activities on the Disintermediation of BankDeposits
by DAVID E. ALLEN AND JERRY T. PARWADA
Competition and Information Production in Market Maker Models
by ALAN D. MORRISON
Debt Overhang, Costly Expandability and Reversibility, and Optimal FinancialStructure
by JYH-BANG JOU AND TAN LEE
Subscriptions 2004 volume
UK/Europe The Americas Rest of the WorldInstitutional Premium Rate £639.00 $1175.00 £725.00Personal Rate £88.00 $175.00 £108.00
New orders and sample copy requests should be addressed to the Journals Marketing Manager atBlackwell Publishing Ltd., 9600 Garsington Road, Oxford, OX4 2DQ, UK. Editorial communicationsto JBFA, Manchester Business School, Booth Street West, Manchester M15 6PB, UK.