-
d Department of Accounting, College of Commerce and Economics,
Sultan Qaboos University,
examines the voluntary Internet nancial reporting practices of
local authorities. Six
variables associated with voluntary disclosure are examined:
political competition, size,
* Corresponding author. Tel.: +64 6 356 9099; fax: +64 6 350
5658.
E-mail address: [email protected] (F. Laswad).
Journal of Accounting and Public Policy 24 (2005) 101121
www.elsevier.com/locate/jaccpubpolP.O. Box 20, AL Khod PC 123,
Oman
Abstract
The reform of public sector (local and central government)
nancial reporting in
New Zealand in the early 1990s has aligned such reporting with
reporting practices in
the private sector (business enterprises). Literature examining
the behaviour of manag-
ers in the public (government) sector suggests that agency
relationships in the sector
motivate such managers to provide information to enable the
monitoring of their
actions. This literature identies a number of characteristics
and variables that proxy
for agency costs in the public sector. The recent development of
the Internet provides
an opportunity for examining voluntary disclosure in the public
sector and, in particu-
lar, in the local government environment. Some New Zealand local
government author-
ities elect to voluntarily provide nancial information on their
websites. This paperDeterminants of voluntary Internetnancial
reporting by localgovernment authorities
Fawzi Laswad a,*, Richard Fisher b, Peter Oyelere c,d
a School of Accountancy, Massey University, Private Bag 11 222,
Palmerston North, New Zealandb Department of Accountancy, Finance,
and Information Systems, University of Canterbury,
Private Bag 4800, Christ church, New Zealandc Centre of
Accounting Education and Research, Lincoln University, New
Zealand0278-4254/$ - see front matter 2005 Elsevier Inc. All rights
reserved.doi:10.1016/j.jaccpubpol.2004.12.006
-
leverage, municipal wealth, press visibility, and type of local
authority. Results indicate
that leverage, municipal wealth, press visibility, and type of
council are associated with
the Internet nancial reporting practices of local authorities in
New Zealand. Policy
implications and possible limitations of the study as well as
suggestions for future
research are discussed in the paper.
2005 Elsevier Inc. All rights reserved.
Keywords: Internet nancial reporting; Local governments;
Voluntary disclosure; New Zealand
1. Introduction
1 The Public Finance Act 1989 requires that the Crown, Crown
entities and agencies, and
102 F. Laswad et al. / Journal of Accounting and Public Policy
24 (2005) 101121departments prepare their accounts in accordance
with generally accepted accounting practice. An
amendment to the Local Government Act 1974, in 1989, also
requires that audited nancial
statements of local authorities be prepared in accordance with
generally accepted accounting
practice.2 The annual report provides historical nancial
information while the annual plan provides
forward-looking nancial information.3 The Act does not specify
the medium for the dissemination of nancial information. It
requires
only that local authorities send copies of their annual reports
and plans to the secretary of local
government, the auditor general, and the parliamentary library.
Local authorities are also required
to make their annual reports and plans available for public
inspection in their oces and librariesThis paper examines the
characteristics of local government (authorities)
that inuence the voluntary dissemination of nancial information
on the
Internet. Such examination enhances the understanding of
incentives to use
the Internet as a medium for discretionary disclosure in the
public sector.
Following reforms in the early 1990s, nancial reporting by
public sector
entities in New Zealand was aligned with private sector nancial
reporting.
Public sector entities prepare their nancial statements in
accordance with
the same set of nancial reporting standards that apply to
private sector enti-ties.1 The Local Government Act 1974 (sections
223D and 223E) requires local
authorities to prepare two nancial reports, an annual report,
and an annual
plan.2 Some local authorities voluntarily elect to use the
Internet to publish
nancial information.3
A number of studies examine the determinants of discretionary
nancial
reporting on the Internet by business enterprises (e.g.,
Ashbaugh et al., 1999;
Craven and Marston, 1999; Pirchegger and Wagenhofer, 1999;
Oyelere et al.,
2003). These studies primarily focus on managerial incentives in
the private sec-tor. The determinants of such discretionary
disclosure in the public sector have
not been examined.and to make copies of such reports available
to the public either free of charge or at a reasonable
charge.
-
and the political manager is the agent (Banker and Patton,
1987). In this sce-
nario, politicians are assumed to be self-interested, maximising
agents, wherebythe maximisation of their wealth depends on their
re-election, advancement,
and current and future income, both pecuniary and non-pecuniary
(Zimmer-man, 1977). Voters are also assumed to be self-interested
and to act in such
a way as to increase their wealth. Voters wealth is related to
the actions of theiragents. Accordingly, each voter has an
incentive to monitor the behaviour of
politicians (Zimmerman, 1977). Further, voters may align
themselves with
interest groups, which have greater power to aect the outcome of
elections
than individual voters (Baber, 1983).
Similar to other agency relationships, the interest group (the
principal) and
the politician (the agent in the public sector) share certain
benets and also cer-tain costs. Therefore, elected politicians
supply monitoring information to
show that they are honouring pre-election promises, and their
incentives to
do so increase as political competition increases (Baber,
1983).To investigate why managers of local authorities elect to
provide nancial
information on the Internet, this study compares the
characteristics of local
authorities that use the Internet as a disclosure medium with
those that choose
not to use such a medium. The results indicate that leverage,
municipal wealth,
press visibility, and council type are the primary determinants
in the decision to
report nancial information on the Internet.The paper is
structured as follows: The following section reviews the
litera-
ture on determinants of nancial reporting practices in the
public sector. Sec-
tion 3 discusses the characteristics and obligations of local
authorities in New
Zealand. Section 4 presents the research hypotheses, while
Section 5 discusses
the research design. Data analysis, and results and discussion
are provided in
Sections 6 and 7 respectively. The last section presents the
summary and
conclusion.
2. Literature review
The literature examining voluntary disclosure in the government
sector is
largely based on the application of agency theory in that
sector. The following
is a review of literature on agency relationships and
determinants of voluntary
disclosure in the public sector.
2.1. Agency relationships and the incentives for disclosure in
the public sector
The relationship between the political (public sector) manager
and the voter
can be described as an agency relationship whereby the voter is
the principal
F. Laswad et al. / Journal of Accounting and Public Policy 24
(2005) 101121 103The discussion above indicates that agency
relationships in the public
sector provide incentives to public sector managers to
voluntarily disclose
-
information that allows the monitoring of their actions. A
number of studies
examine the voluntary disclosure of managers in the public
sector. These
studies are reviewed in the following section.
2.2. Incentives for voluntary disclosure in the public
sector
Relative to the private sector, studies examining voluntary
disclosure prac-
tices in the public sector are limited. The following is a brief
review of the var-
ious disclosure incentives examined in the literature.
Baber (1983) argues that the higher the political competition,
the higher the
incentive for the political manager to supply monitoring
information to the
principals. Several studies have empirically tested this
relationship (e.g., In-
gram, 1984; Baber, 1983; Baber and Sen, 1984; Evans and Patton,
1987; Gir-
oux, 1989). Most of these show a positive relationship between
politicalcompetition and disclosure.
Several studies have also examined the relationship between the
form of
local government and monitoring incentives (e.g., Zimmerman,
1977; Evans
and Patton, 1987; Ingram and DeJong, 1987; Giroux, 1989).
Generally, these
studies reveal a signicant relationship between the form of
local government
and disclosure choice.
A number of studies examine the association between size and
monitoring
behaviour (e.g., Evans and Patton, 1987; Ingram and DeJong,
1987; Ingram,1984; Baber, 1983; and Christiaens, 1999) but provide
inconclusive evidence.
Ingram (1984) and Evans and Patton (1987) report no statistical
association
while Baber (1983), Ingram and DeJong (1987), and Christiaens
(1999), show
a signicant relationship between size and monitoring
behaviour.
Zimmerman (1977) argues that politicians have incentives to
reduce the cost
of debt, thus increasing the resources available for other
programs that may
increase the politicians welfare. Such incentives motivate
public sector man-agers to provide information for the monitoring
of their actions. The resultsof studies that examine the
association between debt and disclosure in the pub-
lic sector are mixed. For example, Ingram and DeJong (1987) and
Evans and
Patton (1987) report a signicant relationship between debt and
disclosure,
while Baber (1983), Baber and Sen (1984) and Christiaens (1999)
nd that debt
is not associated with disclosure. Gore (2004) believes that a
failure to ade-
quately control for dierences in regulatory environments (e.g.,
dierences in
state-specic regulations) and, in some cases, focusing only on
larger municipal
authorities, may explain some of the inconsistencies in prior
studies. After con-trolling for these factors, Gore (2004) found,
in the unregulated state, a positive
association between disclosure levels and proxies for bond
market interaction.
Zimmerman (1977) also notes that the press and public media
inuence the
104 F. Laswad et al. / Journal of Accounting and Public Policy
24 (2005) 101121agency relationship between voters and politicians.
He points out that the
uncovering of political graft and scandals sells newspaper,
attracts television
-
forms, regional councils and territorial authorities.
Territorial authorities are
either district or city councils. Regional councils are
responsible for resourcemanagement, bio-security, catchment
control, harbour administration, regio-
nal civil defence, and regional land transport. City and
district councils are
responsible for community well-being and development,
environmental health
and safety (including building control, civil defence, and
environmental health
matters), infrastructure (roads and transport, sewerage,
water/stormwater),
recreation and culture, and resource management including land
use planningand development control.
Eligible voters are entitled to vote for councillors and, in the
case of territo-
rial authorities, a mayor. The council appoints a chief
executive ocer (CEO)
to whom heads of departments within the council are responsible.
Under the
Local Government Act 1974, the CEO (city manager) is charged
with the
responsibility for implementing decisions of the council,
providing advice to
council, and ensuring eective, ecient and economical management
of the
activities and planning of the local authority.Over the last
decade, local government has become increasingly independent
of central government. Central government has simultaneously
devolved cer-
tain responsibilities, and reduced the extent of nancial
assistance and subsi-viewers and radio listeners, and therefore
aects the advertising rate. The po-
sitive relationship between press coverage and disclosure is,
however, not sup-
ported in Ingrams (1984) study that nds a negative relationship
betweennewspaper circulation and disclosure quality of nancial
reports.
It is, perhaps, not surprising that prior studies have yielded
inconsistent re-
sults for certain variables. Relative to the private sector, the
theory of volun-tary disclosure in the public sector is at a
comparatively early stage and still
requires the development of a critical mass of research
literature. Further, un-
like the private sector, which is generally more homogeneous in
nature, the
diversity of public sector contexts, public sector entities
objectives, regulatoryenvironments, and available archival data,
has necessitated the use of dierent
metrics for similar concepts across prior studies. The present
study attempts to
contribute towards the development of a critical body of
voluntary disclosure
research in the public sector using well established constructs
drawn from priorpublic sector literature, and, where possible,
similar metrics. Further, the study
is limited to a single regulatory environment and includes local
authorities of
varying sizes, ranging from small to large.
3. Local government in New Zealand
There are 86 local authorities in New Zealand structured into
two principal
F. Laswad et al. / Journal of Accounting and Public Policy 24
(2005) 101121 105dies, to local government. The United Nations
Economic and Social
Commission for Asia and the Pacic (2001) estimates that 90% of
New Zealand
-
local government funding is now locally sourced. Corresponding
accountabil-
ity is achieved through the requirement of local authorities to
engage in a con-
sultative planning process, the publication and dissemination of
an annual plan
(forecast information) and an audited annual report, and the
requirement that
local authorities develop, implement, and maintain long-term
nancial strate-
gies together with investment and borrowing policies. Despite
the largelyautonomous nature of local government, the Minister of
Local Government
ing have been removed with the passing of the Local Government
Amendment
Act 1996 (No. 3).
106 F. Laswad et al. / Journal of Accounting and Public Policy
24 (2005) 101121A number of agency relationships can be identied in
the New Zealand local
authority setting. These include relationships between voters
(and interest
groups) and local political managers, creditors and local
political managers,
central government bureaucrats and local political managers, and
counciland appointed local political managers.
4. Hypotheses
This section develops the studys hypotheses relating to the
factors expectedto aect New Zealand local authorities discretionary
use of the Internet fornancial reporting. It draws heavily on the
literature and theory reviewed inthe preceding section. The study
does not attempt to model in detail all possible
agency relationships pertaining to local authority entities.4
Instead, it focuses
on those that we believe are most relevant to a discussion of
external nancial
reporting on the Internet.
4 The objective of this study is to examine whether managerial
incentives models are appropriate
for public sector settings. While this study is not aimed at
directly comparing the public and private
sectors, it could be argued that the managerial incentives in
both sectors are somewhat similar for
the following reasons: (1) in New Zealand, managers move between
the two sectors. Appointments
open to public sector managers are also open to private sector
managers; (2) as part of the reform
of the public sector in New Zealand, greater emphasises has been
placed on commercial and
nancial objectives and transparency. The alignment of reporting
standards is motivated by thesein the central government has the
power to appoint a review authority to re-
view a specic council where the Minister believes that there has
been a signif-
icant and identiable failure of governance (e.g., mismanagement
of resources
or deciencies in the management or decision-making processes of
the
authority).
The four main sources of funds for New Zealand councils are
property rates
(tax), grants from central government, income from user charges,
and loans.Nowadays, the only kind of subsidy from central
government is for road con-
struction and maintenance. Many of the restrictions on local
authority borrow-objectives; and (3) a number of studies have used
managerial incentives in the private sector as
models for explaining public sector behaviour.
-
4.1. Disclosure incentives: Political competition
F. Laswad et al. / Journal of Accounting and Public Policy 24
(2005) 101121 107The supply of potential oce holders and the
corresponding risk of defeat at
future elections, reduces the deviation between voters and
political agentsinterests (Zimmerman, 1977). High levels of
political competition increase
the long run cost to elected ocials of ignoring pre-election
promises, andmotivate incumbents to agree to bear greater
monitoring costs (Baber, 1983;
Evans and Patton, 1987). The Internet is a cost-eective
mechanism for the
instantaneous and simultaneous dissemination of information to
voters, and
therefore may be an important medium for discharging political
agents incre-mental monitoring obligations.
Further, Baber and Sen (1984) argue that political competition
makes . . .incumbents more vulnerable to claims by political rivals
of irresponsible man-
agement, and therefore increases incentives to engage in
self-regulationthatis, to demonstrate a commitment to ecient
managementby using accepted
reporting methods . . . (p. 94). There is evidence that the use
of the Internet fornancial reporting is fast becoming the norm in
the private sector internation-
ally.5 Indeed, the US Securities and Exchange Commission (SEC)
has stated
that as more investors have access to and use the Internet, the
Commission will
consider encouraging the use of the Internet as a prime
dissemination tool
(SEC, 2001). Accordingly, the voluntary use of the Web for
nancial reporting
by some local authorities in New Zealand may represent an
attempt to demon-strate self-regulation in the form of compliance
with best practice reportingmethods.
The rst hypothesis (in alternate form), then is
H1 There is a positive association between political competition
and the vol-
untary use of the Internet for nancial reporting.
4.2. Disclosure incentives: Size
Baber (1983) argues that the number and magnitude of wealth
transfers
administered by political agents may increase competition for
public oce.
Increasing political competition makes it more costly for
incumbent political
agents to ignore pre-election agreements with supporting voters
(or interest
groups), so political agents have an incentive to bear greater
monitoring costs
by supplying more information which demonstrates their execution
of pre-elec-tion promises (Baber, 1983).5 For example, see Lymer et
al. (1999).
-
Given the need for greater disclosure by large local
authorities, it is expected
that such entities will tend to adopt a variety of disclosure
methods. The Inter-
net is likely to be a particularly ecient means of satisfying
the political man-
108 F. Laswad et al. / Journal of Accounting and Public Policy
24 (2005) 101121agers voluntary disclosure incentives in large
local authorities, as the cost ofinformation production and
dissemination on the Internet is likely to be largely
independent of their size (Ettredge et al., 2002; Pirchegger and
Wagenhofer,1999). As a consequence, the benets of disclosure over
the Internet are likely
to be increasing with size (Debreceny et al., 2002; Pirchegger
and Wagenhofer,
1999).
The second hypothesis (in alternate form), then is
H2 There is a positive association between local authority size
and the volun-
tary use of Internet nancial reporting.
4.3. Disclosure incentives: Leverage
It has been argued that the use of debt to nance public
activities provides
an incentive for political managers to reduce the cost of debt
(Zimmerman,
1977). This may be achieved through the voluntary disclosure of
information
that facilitates monitoring by creditors (Zimmerman, 1977;
Baber, 1983; Baber
and Sen, 1984; Ingram, 1984; Ingram and DeJong, 1987;
Christiaens, 1999;Gore, 2004). Such incremental disclosure can
occur through traditional hard
copy nancial statements and other media, such as Internet
nancial reporting.
Therefore:
H3 There is a positive association between local authority
leverage and the
voluntary use of Internet nancial reporting.
4.4. Disclosure incentives: Municipal wealth
Christiaens (1999) argues that municipal wealth should be
positively asso-ciated with increased disclosure because it
provides a signal of management
quality, which may benet local politicians by increasing their
chances of re-
election and reducing interest costs. Christiaens, following
Ingram (1984), uses
own revenue per capita6 as a proxy for municipal wealth. Poorer
performinglocal authorities may avoid using voluntary disclosure
techniques, such as
6 Ingram (1984) denes own revenue per capita as general revenues
other than intergovern-
mental transfers. Such intergovernmental transfers are
negligible in New Zealand since the reforms
to the public sector discussed in Section 3.
-
Internet nancial reporting (IFR), preferring instead to, . . .
restrict access toaccounting information to more determined users
(Craven and Marston,
1999, p. 323). Hence:
F. Laswad et al. / Journal of Accounting and Public Policy 24
(2005) 101121 109H4 There is a positive association between local
authority municipal wealth
and the voluntary use of Internet nancial reporting.
4.5. Disclosure incentives: Press visibility
As discussed in Section 2, Zimmerman (1977) identied the role of
the press
in the agency relationship between voters and politicians.
Consistent with Zim-
merman (1977), Ingram (1984, p. 130) argues that a strong press,
. . . mightinduce more disclosures to satisfy the information
demands of the press or evenas a defensive mechanism for
politicians. Thus:
H5 There is a positive association between the frequency of
press reporting on
local government activities, and the voluntary use of Internet
nancial
reporting by local authorities.
4.6. Disclosure incentives: Council type
While the incentives for individual voters to monitor political
behaviour ap-
pear limited, coalitions of voters (such as political parties,
industrial or occupa-
tional coalitions, etc.) have greater incentives to do so due to
their potential
ability to inuence the outcome of elections via collective
voting (Baber,
1983; Ingram, 1984). Local political managers of councils
subject to the scru-
tiny of cohesive coalitions will have corresponding incentives
to facilitate the
supply of information useful in the monitoring of their actions.
As urbanisa-tion facilitates the formation of coalitions (Ingram,
1984), it is probable that
local political managers in councils with proportionately large
populations in
metropolitan areas have greater incentives to voluntarily
provide monitoring
information than managers in councils with relatively large
rural populations.
Coupled with the fact that urbanisation has been associated with
Internet ac-
cess and usage,7 it is also probable that voluntary IFR
disclosures will be high-
est among city councils and regional councils relative to
district councils. Both
city and regional councils have higher proportions of rate
payers in major
7 For example, see Falling Through the Net: Dening the Digital
DivideA Report on theTelecommunications and Information Technology
Gap in America (July 1999) and subsequent
updates produced by the National Telecommunications and
Information Administration (NTIA).
-
metropolitan areas relative to district councils. Consequently,
the sixth hypoth-
esis is as follows:
H6 District councils engage in less voluntary use of Internet
nancial reporting
than do regional or city councils.
5. Research design
The 86 local authorities in New Zealand comprise 12 regional, 15
city,
and 59 district councils, respectively. The website addresses of
local authori-
ties were identied through the Local Government New Zealand
website
, Local Government Online , andLocal Government Website Index .
Local authorities
not listed on these three websites were contacted by telephone
to obtain their
website addresses. Table 1 shows that 61 out of 86 local
authorities maintain
websites.
Local authorities use the Internet to provide a wide range of
information. Of
particular relevance to this study, only 30 (about 49%) of local
authorities with
Council type Total
110 F. Laswad et al. / Journal of Accounting and Public Policy
24 (2005) 101121Regional City District No. of local
authorities
%
Financial highlights only 1 3 4 13.3
Annual reports only 1 1 2 6.7
Annual plan only 1 6 7 14 46.7
Combinations of annual reports,
plans and nancial highlights
3 4 3 10 33.3Table 1
Local authorities maintenance of websites
Type Regional
council
City council District
council
Total
No. % No. % No. % No. %
With website 11 92 15 100.0 35 59 61 71
Without website 1 8 0 0.0 24 41 25 29
Total 12 100.0 15 100.0 59 100.0 86 100.0
Table 2
The nature of published nancial information on the Internet (n =
30)Total 6 10 14 30 100
-
Table 3
Research variables
Variable Denition
Political competition
Ratio of candidates to positionNumber of candidates
Positions available
Size
Total assets Average total assets
Total revenue Average total revenue
Leverage
Long-term liabilities: total assetsLong-term liabilities
Total assets
Long-term liabilities
Total public equity
Municipal wealth
Own revenue per capitaAverage total revenue
Population
Press visibility
News item count Number of news items in the print press
in which the local authority appeared during 2000,
as measured by a count search on Newzindex
Council type
F. Laswad et al. / Journal of Accounting and Public Policy 24
(2005) 101121 111websites provide nancial information on those
sites. Table 2 provides a break-
down of the type of nancial information provided on
websites.
Table 3 presents the denitions of the research variables used in
this study.
Financial data for local authorities were collected from the
nancial informa-
tion on the websites of local authorities (where available) and
from hard copies
of their annual reports and plans. The nancial accounting data
cover the
three-year period 19971999. Two proxies each were used to
alternatively rep-
resent the size and leverage variables. Size was represented by
total assets andtotal revenue, while leverage was represented by
total long-term liabilities to
total assets and total long-term liabilities to total public
equity.
Consistent with Christiaens (1999) and Ingram (1984), municipal
wealth was
proxied by own revenue per capita. The ratio of candidates to
positions in thelast local government election is employed as a
proxy for political competition.
This proxy measures the intensity of competition in local
government elections.
Data for this measure was obtained from the Local Authority
Election Statis-
tics publication (Department of Internal Aairs, 1999). Press
visibility is mea-sured by the number of news items appearing in
the print press in 2000
obtained from the New Zealand electronic database Newsindex. In
prior re-
search, newspaper circulation has been used as a surrogate for
strength of press
(Ingram, 1984). However, Ingram (1984) concluded that this
measure was not
Form of local authority Type of council: District, city, or
regional
-
Table 4
Descriptive statistics and univariate sample tests
Variable Statistic All local
authorities
(n = 86)
Local authorities with websites
N-IFRAsa
(n = 31)
IFRAsa
(n = 30)
t-
Statistic
Panel A: Descriptive statistics and sample t-test of continuous
variables
Political competition
Ratio of candidates
to position
Mean 2.340 2.280 2.550 1.497*
Median 2.300 2.300 2.400
Std Deviation .676 .647 .739
Minimum 1.000 1.000 2.000
Maximum 4.000 4.000 4.000
Size
Total assets Mean 462,976,620 428,723,543 685,017,202 1.500*
Median 313,585,333 358,378,167 351,140,547
Std Deviation 583,818,905 404,488,828 843,789,091
Minimum 23,847,016 41,350,667 79,517,333
Maximum 3,514,577,000 2,093,541,667 3,514,577,000
Total revenue Mean 59,040,543 49,381,664 102,011,514 1.658*
Median 31,512,110 33,865,500 42,209,450
Std Deviation 107,957,370 59,803,279 163,249,508
Minimum 3,271,316 5,229,119 5,593,333
Maximum 828,545,333 268,047,333 828,545,333
Leverage
Long-term liabilities:
total assets
Mean .067 .053 .075 1.193
Median .036 .041 .058
Std Deviation .122 .058 .083
Minimum .000 .000 .000
Maximum .967 .282 .436
Long-term liabilities:
total public equity
Mean .080 .059 .103 1.207
Median .038 .038 .065
Std Deviation .162 .082 .182
Minimum .000 .000 .000
Maximum 1.008 .418 1.008
Municipal wealth
Own revenue
per capita
Mean 902 872 938 .424
Median 849 818 878
Std Deviation 525 610 589
Minimum 87 103 87
Maximum 3,432 3,432 2,629
Press visibility
News item count Mean 17.480 7.480 38.030 2.443***
Median 3.000 3.000 6.000
Std Deviation 43.657 12.119 68.532
Minimum 0.000 0.000 0.000
Maximum 276.000 48.000 276.000
112 F. Laswad et al. / Journal of Accounting and Public Policy
24 (2005) 101121
-
authorities in New Zealand using both univariate and
multivariate approaches.
Local authorities are categorised into those providing nancial
reports on theInternet (IFRAs) (n = 30) and those that maintain a
website but choose not to
use it to disseminate nancial information (N-IFRAs) (n =
31).
Univariate independent sample t-tests are carried out on the
independent
variables for the two main groupings of local authorities to
test for possible dif-ferences in the means of the selected
variables between IFRAs and N-IFRAs.an eective proxy. Descriptive
statistics for all the present studys independentvariables are
presented in Table 4.
A local authoritys engagement in Internet nancial reporting is
measured asa dichotomous variable, which reects whether the local
authority does or does
not provide nancial information on the web.
6. Data analysis
The aim of this study is to identify the determinants of IFR
among local
Table 4 (continued)
Pearson Chi-square statistic
Council type N-IFRA IFRA Total Value df Sig. (2-tailed)
Panel B: Chi-square test of categorical independent variable
Regional 5 6 11 3.142 2 .208
City 5 10 15
District 21 14 35
Total 31 30 61
***,* Denote signicance at the 1%, 5%, and 10% (one-tailed)
levels, respectively.a IFRAs = Internet nancial reporting local
authorities; N-IFRAs = Non-Internet nancial
reporting local authorities.
F. Laswad et al. / Journal of Accounting and Public Policy 24
(2005) 101121 113Table 4 presents the results of these tests.
Logit analysis is used to determine the association of the
independent vari-
ables with the dichotomous (IFRA vs. N-IFRA) Internet nancial
reporting
practice dependent variable. Logit allows the predication of a
discrete outcome
(e.g., IFRA vs. N-IFRA) from a set of independent variables,
which may be
continuous, discrete, dichotomous, or a mixture. Logit can be
used to ascertain
which variables predict an outcome and how these variables aect
an outcome(Tabachnick and Fidell, 1996).
Expressed in its full form with respect to this study, the logit
equation is
IFRi a b1Political competitioni b2Sizei b3Leveragei b4Municipal
wealthi b5Press visibilityi b6Council typei li 1
-
similar relationship is apparent for the alternative measure of
leverage, totallong-term liabilities to total assets.On average,
IFRAs generated more wealth (own revenue per capita) than
N-IFRAs. Their mean of $938 compares with $872 for N-IFRAs.
IFRAs are
also more visible in the print press, with a mean press
visibility count of 38
news items as compared to only 7 news items for N-IFRAs. The
level of
publicity experienced, however, varies signicantly among IFRAs,
ranging
between a minimum of no news items and a maximum of 276 news
items.
N-IFRAs minima was nil, with a maxima of 48.
7.2. Univariate data analysiswhere, for the, ith authority, IFR
is the IFR practice: 0 for N-IFRAs and 1 for
IFRAs, a is constant of the equation, Political competition is
the ratio of candi-dates to council positions, Size is either total
assets or total revenue, Leverage is
either long-term liabilities: total assets or long-term
liabilities: total public equity,
Municipal wealth is own revenue per capita, Press visibility is
the news items
count, Council type is 1 if district council and 0 otherwise and
l is the error term.As predictor variables in logit models do not
have to be normally distrib-
uted, linearly related, or of equal variance within each group
(Tabachnick
and Fidell, 1996, p. 575; Hair et al., 1995, p. 130), none of
the independent vari-
ables are transformed for the logit analysis. Table 5 contains
the logit results.
7. Results and discussion
7.1. Descriptive statistics
Table 4 reveals that the intensity of political competition is
higher in IFRAs
relative to N-IFRAs. The mean ratio of candidates competing for
council posi-
tions is 2.55 (median = 2.40) for IFRAs as compared to 2.28
(median 2.30) for
N-IFRAs.
Table 4 also shows that IFRAs are larger than N-IFRAs. The mean
total
assets of IFRAs, at $685.017 million, is considerably greater
than that of N-IFRAs ($428.724 million). The largest IFRA had total
assets of $3514.577 mil-
lion, as compared to $2093.542 million for the largest N-IFRA.
This trend is
repeated across the alternative measure of size, total revenue.
This preliminary
result is similar to those reported by studies of IFR in the
private sector (Ashb-
augh et al., 1999; Oyelere et al., 2003).
IFRAs appear more highly leveraged, with mean long-term
liabilities to
total public equity ratio of 0.10:1 as compared to 0.06:1 for
N-IFRAs. A
114 F. Laswad et al. / Journal of Accounting and Public Policy
24 (2005) 101121The results of univariate tests of dierences
between IFRAs and N-IFRAs
are presented in Panels A and B of Table 4. The results indicate
statistically sig-
-
Table 5
Multivariate regression results
Logit analysis
IFRi = a + b1Political competitioni + b2Sizei + b3Leveragei +
b4Municipal wealthi+ b5Press visibilityi + b6Council typei + li
Variable Expected sign Model coecients (Std. error)
Model Aa Model Ba Model Ca Model Da
Constant +/ .169 .209 .462 .516(1.537) (1.544) (1.513)
(1.513)
F. Laswad et al. / Journal of Accounting and Public Policy 24
(2005) 101121 115nicant dierences between the two groups of local
authorities on the bases of
political competition, size, and press visibility.
Panel A of Table 4 reveals that political competition is more
intense in
IFRAs than N-IFRAs. The ratio of candidates to council positions
in IFRAs
is statistically signicantly higher at the 10% level.
Similarly, IFRAs are signicantly larger than N-IFRAs (at the 10%
level
using either proxy for size). This result is consistent with the
ndings of previ-
ous studies on the characteristics of local authorities (and
their variants), con-rming size as a determining variable at the
univariate level (Baber, 1983;
Evans and Patton, 1987; Christiaens, 1999).
Political competition + .274 .312 .479 .507(.637) (.640) (.612)
(.614)
Size + .000 .000 .000 .000
(.000) (.000) (.000) (.000)
Leverage + 1.276 1.278 2.399 1.856
(4.775) (2.843) (4.628) (2.733)
Municipal wealth + .001 .001 .001 .001
(.001) (.001) (.001) (.001)
Press visibility + .039* .038* .032 .031
(.023) (.023) (.023) (.022)
Council type 1.016 1.011 .985 1.007(.810) (.806) (.842)
(.843)
Log likelihood 70.444 70.291 71.438 71.169
Nagelkerke R2 .233 .236 .214 .219
Chi-square statistic 11.330* 11.484* 10.337 10.606
Degrees of freedom 6 6 6 6
Number of observations 60 60 60 60
Correctly predicted
N-IFR 75.9 72.4 72.4 72.4
IFR 56.7 56.7 56.7 56.7
Overall 66.1 64.4 64.4 64.4
* Denote signicance at the 1%, 5%, and 10% levels,
respectively.a Proxies for size and leverage, respectively, are as
follows: Model A-total assets and long-term
liabilities: total assets; Model B-total assets and long-term
liabilities: total public equity; Model C-
total revenue and long-term liabilities: total assets; and Model
D-total revenue and long-term
liabilities: total public equity.
-
IFRAs are not statistically signicantly more highly nancially
leveraged
than N-IFRAs. However, at the 1% level, IFRAs are more visible
in the press
116 F. Laswad et al. / Journal of Accounting and Public Policy
24 (2005) 101121than N-IFRAs. This high level of positive
association between IFR practice
and press visibility may not be surprising, given that both are
media for com-
municating information, nancial or otherwise, to the
stakeholders of the local
authorities. Ingram (1984) discusses the potential inuence of
the press onnancial disclosure levels of governments. Panel B of
Table 4 indicates no asso-
ciation between council type and the practice of Internet
nancial reporting.
To summarise the ndings of univariate tests carried out in this
study, on
average IFRAs are signicantly larger, have greater political
competition for
their councillor positions, and are more politically visible in
the press than
N-IFRAs.
7.3. Multivariate regression analysis
Table 5 presents the results of the multivariate analysis
(corresponding to Eq.
(1)) where the dependent variable is a dichotomous variable
indicating whether
a local authority does or does not provide Internet nancial
information.
Since two alternative proxies are used to measure size and
leverage, the re-
sults of the estimation of four combinations (models) are
reported. Total assets
is used as a proxy for size in Models A and B, while total
revenue is used in-
stead in Models C and D. With respect to leverage, total
long-term liabilitiesto total assets is used in Models A and C,
while long term liability to total pub-
lic equity is used in Models B and D.
The results of Cooks Distance test revealed that one outlier had
a dispropor-tionately large inuence on the models.8 Consequently,
the relevant case was ex-
cluded from subsequent analysis, leaving a total of 60 local
authorities. The
overall predictive ability of the four logit models ranged from
64.4% to 66.1%.
Of the four models reported in Table 5, two (models A and B) are
signicant
as a whole, based on Chi-square statistics. In both of these
models, only pressvisibility is signicant in the predicted
direction, providing weak support for H5.
Press visibility is statistically signicantly related to local
authorities IFR atthe 10% level in both Models A and B. Local
authorities that are highly visible
in the press media are more likely to engage in IFR. This is as
predicted, and
indicates that the press media inuences the agency relationship
between voters
and local authority managers. Local authorities that are more
press-visible are
more likely to proactively seek and use other channels of
communication such
as the Internet. This result is consistent with Ingrams (1984)
hypothesis withrespect to press strength and voluntary hard copy
nancial disclosure.8 Removing problematic outliers assists in
removing distortion from statistical tests and leads to a
better understanding of the phenomena being examined (Hair et
al., 1995, pp. 5758).
-
At rst sight, the results reported in Table 5 appear to provide
weak support
for only one of the studys hypotheses. However, an alternative
interpretationof the results is possible by virtue of the fact that
the studys sample consists ofthe entire population of New Zealand
local authorities with websites at the
time of the study. As long as there is no intention to make
inferences from
the results beyond New Zealand local authorities, the studys
sample may, infact, be regarded as a 100% sample. When a sample
represents the entire pop-
ulation, tests of signicance are generally less meaningful
because any non-zero
regression coecient must, with certainty, be non-zero for the
population as a
whole.9 Each regression coecient is regarded as an actual
population value
(which is a constant) rather than as an estimator (with
associated sampling dis-
tribution) of a population value (Kish, 1965).
F. Laswad et al. / Journal of Accounting and Public Policy 24
(2005) 101121 117Using this alternative interpretation of the
results, it is apparent that, in
addition to press visibility, three further variables (leverage,
municipal wealth,and council type) are non-zero and in the
predicted direction across all
models.10
Among New Zealand local authorities, leverage (H3) appears to be
posi-
tively associated with IFR practices. Table 5 reveals such a
relationship across
all four models. This result suggests that political managers of
New Zealand
municipalities, which are subject to relatively high debt
burdens, perceive
IFR as a potential means of facilitating monitoring by
creditors. The result
is also consistent with the ndings in several previous public
sector voluntarydisclosure studies conducted in the US (e.g.,
Ingram and DeJong, 1987; Evans
and Patton, 1987; Gore, 2004).
The results in Table 5 suggest that municipal wealth (H4), as
proxied by
own revenue per capita, is directly associated with the practice
of IFR forNew Zealand local authorities. This nding is consistent
with that of Ingram
(1984) and suggests that this variable is perceived by
management as an impor-
tant indicator of their quality in the New Zealand public sector
environment.
9 It is acknowledged that tests of signicance could be
meaningful in these circumstances if
measures of dependent and explanatory variables are regarded as
being random draws from
distributions of possible values. This could be seen to be the
case, for example, for a measure such
as stock returns based on closing price minus opening price. One
more or less transaction on the
relevant day could have resulted in a dierent closing stock
price. This is consistent with the
superpopulation concept, under which a nite population may be
regarded as being drawn from a
larger universe (Cassel et al., 1977). However, we do not
believe that this view is particularly
applicable to this study, as its measurements of variables, such
as size, are likely to be highly
correlated from day to day and even from year to year. (We are
grateful to one of the manuscriptsreviewers for this
observation.)10 Interpretation of the results reported in Table 5
in the proposed manner must be tempered by
the possibility that, whilst the coecients corresponding to the
three additional variables arereliably dierent from zero, they may
be close to zero (i.e., have shallow slopes). This may explainthe
prevailing lack of signicance in the models generally.
-
(40%) engage in IFR than regional (55%) or city (67%) councils,
respectively.
The Internet is, at least at its initial stage, an urban
phenomenon. This perhapsexplains its relatively greater use by city
councils and, to a lesser extent, regio-
nal councils, for nancial reporting purposes. Ingram (1984) also
found urban-ization to have a signicant inuence on disclosure
practices.
As is apparent from Table 5, the expected positive association
between polit-
ical competition (H1) and New Zealand local authorities IFR
practices was notfound. This nding is surprising given previous
ndings in the governmental
sector nancial disclosure literature (Baber, 1983; Baber and
Sen, 1984; In-
gram, 1984). It appears that IFR is not seen by political
managers of New Zea-
land local authorities as a critical means of demonstrating
self-regulation in the
form of compliance with best practice reporting methods, even in
the presenceof intense political competition. In fact, the greater
the intensity of political
competition, the lesser the tendency of managers of local
authorities to use
the Internet to voluntarily report nancial information.
Also surprising was the fact that size (H2) (across all four
models) was found
not to be a predictor of IFR among New Zealand local
authorities. It appears
that, unlike the ndings of earlier studies in both the public
sector (Baber,
1983; Evans and Patton, 1987; Christiaens, 1999) and private
sector nancial
disclosure literature (Courtis, 1979; McNally et al., 1982;
Cooke, 1989a,b,1991, 1992; Hossain et al., 1995; Wallace et al.,
1994; Wallace and Naser,
1995; Inchausti, 1997; Owusu-Ansah, 1998; Ashbaugh et al., 1999;
Debreceny
et al., 2002; Ettredge et al., 2002; Oyelere et al., 2003), size
does not appear to
be an important explanatory variable for IFR practices of New
Zealand local
authorities. Larger New Zealand local authorities may not be
taking full
advantage of the potential economies of scale benets that may be
available
from the use of the Internet for nancial information
dissemination.
In summary, support was found for four out of the six alternate
hypothesesstated in this study based on the results of multivariate
analysis. The predicted
positive associations between leverage (H3), municipal wealth
(H4), and press
visibility (H5), on the one hand, and IFR practices on the
other, were supported.
Also, the predicted negative relationship between council type
(H6) and IFR
practices was also borne out by the studys results. However, the
hypothesespostulating a positive relationship between IFR
practices, on the one hand,
and political competition (H1) and size (H2) on the other, were
not supported.
8. Summary and conclusionCouncil type (H6) is a predictor of IFR
among New Zealand local author-
ities. As evident in Panel B of Table 4, a lower proportion of
district councils
118 F. Laswad et al. / Journal of Accounting and Public Policy
24 (2005) 101121The reform of public sector nancial reporting in
New Zealand, where it is
aligned with practices in the private sector, and the
development of the Internet
-
and other theories highlighting the incentives for voluntary
disclosures in the
public sector.
Results are largely consistent with expectations. The results of
logit analysis
indicates that whether or not local authorities in New Zealand
use IFR can be
predicted based on their level of nancial leverage, municipal
wealth, press vis-
ibility and council type. Both local authority size and level of
political compe-
tition were not found to be useful predictors of IFR. Local
authorities that are
more highly leveraged or that create relatively more municipal
wealth than
other authorities, are more likely to engage in IFR. New Zealand
local author-
ities that are more visible in the press are also more likely to
use the Internet toprovide nancial information, a conrmation of the
role of a strong press in
ensuring greater nancial accountability and transparency. Also,
it is apparent
that district councils engage in less voluntary nancial
disclosure through the
Internet than city and regional councils, possibly due to the
urban nature
of the Internet and relatively lower level of access. This eect
is likely to dissi-
pate signicantly in the near future.
Future research opportunities include the development of a
disclosure index
for disclosure on the Internet that has not been included in
other reporting out-lets, such as hard copy nancial reports. The
inclusion of other nancial and
non-nancial variables, such as management compensation and
qualication,
may assist the development of improved predictive models of
voluntary disclo-
sure in the public sector. Future research may control for other
variables such
as the strength of the information technology departments, where
the greater
the expenditure the local authority allocates to such
departments, the more
likely it is to disclose on the Internet. Future studies may
also consider exam-
ining the timeliness of disclosure on the Internet and how it
compares with thetimeliness of reporting in other media. This study
is based on New Zealand
practices; studies in other countries and international
comparisons of determi-
nants of IFR would be useful in the development of a
comprehensive predictive
disclosure model in the public sector, or at least a segment of
the public sector,
in an electronic environment such as the Internet.
Acknowledgments
We are grateful for the valuable comments we have received from
the Jour-
nals reviewers on an earlier draft of the paper. We acknowledge
the researchas an information medium provide a unique opportunity
for the development
and renement of disclosure behaviour models in the public
sector. This study
examined the possible determinants of discretionary Internet
nancial disclo-
sure practices by local authorities in New Zealand in the
context of agency
F. Laswad et al. / Journal of Accounting and Public Policy 24
(2005) 101121 119assistance provided by Shireen Zaharudin and
Fareeha Shareef, and the com-
ments and suggestions received from participants at the 2002
Annual Meeting
-
Debreceny, R., Gray, G., Rahman, A., 2002. The determinants of
Internet nancial reporting.Journal of Accounting and Public Policy
21 (4/5), 371394.
Department of Internal Aairs, 1999. Local Authority Election
Statistics 1998. Department of
Internal Aairs, Wellington, New Zealand.
Ettredge, M., Richardson, V.J., Scholz, S., 2002. Dissemination
of information for investors at
corporate websites. Journal of Accounting and Public Policy 21
(4/5), 357369.
Evans III, J.H., Patton, J.M., 1987. Signalling and monitoring
in public-sector accounting. Journal
of Accounting Research 25 (Suppl.), 130158.
Giroux, G., 1989. Political interests and governmental
accounting disclosure. Journal of
Accounting and Public Policy 8 (3), 199217.of the American
Accounting Association, the 9th IAAER World Congress of
Accounting Educators, 2002 and the 2001 Annual Conference of the
Account-
ing Association of Australia and New Zealand. We also beneted
from useful
comments by Jesse Hughes, Ken Smith, Stephen Owusu-Ansah and the
late
R.S. Olusegun Wallace. Financial support was provided from the
Summer
Scholarship and Research Assistantship Grants of Lincoln
Universitys Com-merce Division.
References
Ashbaugh, H., Johnstone, K., Wareld, T., 1999. Corporate
reporting on the Internet. Accounting
Horizons 13 (3), 241257.
Baber, W.R., 1983. Towards understanding the role of Auditing in
the public sector. Journal of
Accounting and Economics 5 (3), 213227.
Baber, W.R., Sen, P.K., 1984. The role of generally accepted
reporting methods in the public
sector: An empirical test. Journal of Accounting and Public
Policy 3 (2), 91106.
Banker, R.D., Patton, J.M., 1987. Analytical agency theory and
municipal accounting: An
introduction and an application. Research in Governmental and
Nonprot Accounting 3 (Part
B), 2950.
Cassel, C., Sarndal, C., Wretman, J.H., 1977. Foundations of
Inference in Survey Sampling. Wiley,
New York.
Christiaens, J., 1999. Financial accounting reform in Flemish
municipalities: An empirical
investigation. Financial Accountability and Management 15 (1),
2140.
Cooke, T., 1989a. Voluntary corporate disclosure by Swedish
companies. Journal of International
Financial Management and Accounting 1 (2), 171195.
Cooke, T., 1989b. Disclosure in the corporate annual reports of
Swedish companies. Accounting
and Business Research 19 (74), 113124.
Cooke, T., 1991. An assessment of voluntary disclosure in the
annual reports of Japanese
corporations. The International Journal of Accounting 26 (3),
174189.
Cooke, T., 1992. The impact of size, stock market listing and
industry type on disclosure in the
annual reports of Japanese listed corporations. Accounting and
Business Research 22 (87), 229
237.
Courtis, J.K., 1979. Annual report disclosure in New Zealand:
Analysis of selected corporate
attributes. Research Study No. 8, University of New England,
Armidale.
Craven, B., Marston, C., 1999. Financial reporting on the
Internet by leading UK companies. The
European Accounting Review 8 (2), 321333.
120 F. Laswad et al. / Journal of Accounting and Public Policy
24 (2005) 101121Gore, A.K., 2004. The eects of GAAP regulation and
bond market interaction on local
government disclosure. Journal of Accounting and Public Policy
23, 2352.
-
Hair, J., Anderson, R., Tatham, R., Black, W., 1995.
Multivariate Analysis with Readings, fourth
ed. Prentice-Hall, Inc., New Jersey.
Hossain, M., Perera, M.H.B., Rahman, A.R., 1995. Voluntary
disclosure in annual reports of New
F. Laswad et al. / Journal of Accounting and Public Policy 24
(2005) 101121 121Zealand companies. Journal of International
Financial Management and Accounting 6 (1),
6985.
Inchausti, B.G., 1997. The inuence of company characteristics
and accounting regulation on
information disclosed by Spanish rms. The European Accounting
Review 6 (1), 4568.
Ingram, R.W., 1984. Economic incentives and the choice of state
government accounting practices.
Journal of Accounting Research 22 (1), 126144.
Ingram, R.W., DeJong, D.V., 1987. The eect of regulation on
Local Government Disclosure
Practices. Journal of Accounting and Public Policy 6 (4),
245270.
Kish, L., 1965. Survey Sampling. John Wiley & Sons, Inc.,
New York.
Lymer, A., Debreceny, R., Gray, G., Rahman, A., 1999. Business
Reporting on the Internet.
International Accounting Standards Committee.
McNally, G.M., Lee, H.E., Hasseldine, R., 1982. Corporate
nancial reporting in New Zealand:
An analysis of user preferences, corporate characteristics and
disclosure practices for
discretionary information. Accounting and Business Research 13
(49), 1120.
National Telecommunications and Information Administration
(NTIA), 1999. Falling through the
net: Dening the digital divide. A Report on the
Telecommunications and Information
Technology Gap in America. Retrieved December 2003. Available
from: .
Owusu-Ansah, S., 1998. The impact of corporate attributes on the
extent of mandatory disclosure
and reporting by listed companies in Zimbabwe. The International
Journal of Accounting 33
(5), 605631.
Oyelere, P., Laswad, F., Fisher, R., 2003. Determinants of
Internet nancial reporting by New
Zealand companies. Journal of International Financial Management
and Accounting 14 (1),
2651.
Pirchegger, B., Wagenhofer, A., 1999. Financial information on
the Internet: A survey of the
homepages of Austrian companies. The European Accounting Review
8 (2), 383395.
Securities and Exchange Commission (SEC), 2001. Special Study:
Regulation Fair Disclosure
Revisited. [Retrieved October 2003]. Available from: .
Tabachnick, B.G., Fidell, L.S., 1996. Using Multivariate
Statistics, third ed. HarperCollins College
Publishers, New York.
United Nations Economic and Social Commission for Asia and the
Pacic, 2001. Country Paper:
New Zealand. [Retrieved January 2001]. Available from: .
Wallace, R.S.O., Naser, K., 1995. Firm specic determinants of
comprehensiveness of mandatory
disclosure in the corporate annual reports of rms on the stock
exchange of Hong Kong.
Journal of Accounting and Public Policy 14 (4), 311368.
Wallace, R.S.O., Naser, K., Mora, A., 1994. The relationship
between comprehensiveness of
corporate annual reports and rm characteristics in Spain.
Accounting and Business Research
25 (97), 4153.
Zimmerman, J.L., 1977. The municipal accounting maze: An
analysis of political incentives.
Journal of Accounting Research 15 (Suppl.), 107144.
Determinants of voluntary Internet financial reporting by local
government authoritiesIntroductionLiterature reviewAgency
relationships and the incentives for disclosure in the public
sectorIncentives for voluntary disclosure in the public sector
Local government in New ZealandHypothesesDisclosure incentives:
Political competitionDisclosure incentives: SizeDisclosure
incentives: LeverageDisclosure incentives: Municipal
wealthDisclosure incentives: Press visibilityDisclosure incentives:
Council type
Research designData analysisResults and discussionDescriptive
statisticsUnivariate data analysisMultivariate regression
analysis
Summary and conclusionAcknowledgmentsReferences