This article was downloaded by: [Cornell University]

On: 04 December 2012, At: 15:27

Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office:
Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Accounting and Business Research

Publication details, including instructions for authors and subscription
information:
http://www.tandfonline.com/loi/rabr20

An Examination of Audit Delay: Further

Evidence from New Zealand
a b
Charles A. P. N. Carslaw & Steven E. Kaplan
a
University of Nevada, Reno
b
Arizona State University
Version of record first published: 28 Feb 2012.

To cite this article: Charles A. P. N. Carslaw & Steven E. Kaplan (1991): An Examination of Audit Delay:
Further Evidence from New Zealand, Accounting and Business Research, 22:85, 21-32

To link to this article: http://dx.doi.org/10.1080/00014788.1991.9729414

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching, and private study purposes. Any substantial
or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or
distribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the
contents will be complete or accurate or up to date. The accuracy of any instructions, formulae,
and drug doses should be independently verified with primary sources. The publisher shall not
be liable for any loss, actions, claims, proceedings, demand, or costs or damages whatsoever or
howsoever caused arising directly or indirectly in connection with or arising out of the use of this
material.
 Accounting and Business Research. Vol. 22. No. 85, pp. 21-32. 1991 21

An Examination of Audit Delay:

Further Evidence from New Zealand*
Charles A. P. N. Carslaw and Steven E. Kaplan
Abstract-An important qualitative attribute of financial statements is timeliness. The recognition that the length
of the audit may be the single most important determinant affecting the timing of earnings’ announcements has
motivated recent research on audit delay. The present study extends prior research by examining the multivariate
relationship between a set of explanatory variables and audit delay for a large sample of New Zealand public
companies. Further, the study includes two explanatory variables, company control (i.e. owner control versus
manager control) and debt proportion, which have not previously been considered. The results indicate that both
Downloaded by [Cornell University] at 15:27 04 December 2012

company size and sign of income significantly affect audit delay across the two years examined. Five other
explanatory variables significantly affected audit delay for one of the two years examined. The adjusted R2s of the
regression models, however, were relatively low. Additional analysis was also performed on each company control
subsample. These results revealed that the effect of company size and income sign may be conditional upon company
control. Implications from the results of the study for future research are discussed.

Introduction recent investigations of factors affecting ‘audit

delay’,’ i.e. the length of time from a company’s
An important qualitative attribute of financial state- financial year end to the date of the auditor’s
ments is timeliness, which requires that information report (Ashton, Willingham, and Elliott, 1387;
should be made available to financial statement Ashton, Graul, and Newton, 1989; and Newton
users as rapidly as possible. Both analytical and and Ashton, 1989). These studies of the United
empirical evidence suggest that decisions based .States and Canadian audit markets improved upon
upon financial statement information may be prior research (Courtis, 1976; Davies and Whittred,
affected by the timeliness of information release. 1980; Dyer and McHugh, 1975; Garsombke, 1981;
For example, an analytical model developed by Gilling, 1977; and Whittred, 1980) by examining
Feltham (1972) shows that timeliness affects a de- the multivariate relationship between audit delay
cision maker’s expected payoff. Empirical research and a set of explanatory variables. Multivariate
has also shown that timeliness affects security analysis is appropriate because many factors poten-
prices (Chambers and Penman, 1984; Givoly and tially contribute to or are associated with audit
Palmon, 1982; and Kross and Schroeder, 1984). delay. To date, the predictive ability of audit delay
Givoly and Palmon (1982, p. 491) contend that has been relatively low, and Ashton et al. (1989) call
the ‘single most important determinant of the for additional research focusing on different
timeliness of the earnings announcement is the countries, different time periods and additional
length of the audit’. Systematic differences in audit explanatory variables.
length are likely to be observed because of differ- The purpose of the present study is to provide
ences in the nature, timing and extent of evidence further evidence on the determinants of audit delay
necessary to support an audit opinion across audit and to extend prior research in three ways. First,
engagements. Additionally, the auditor may vary the multivariate relationship between audit delay
the time of commencement of the audit. The and a set of explanatory variables for the New
central role that the auditor plays in the timing Zealand market is considered. Previous studies
of earnings announcements has motivated several of the New Zealand market have not employed
a multivariate approach. Second, the explanatory
*The authors are associate professors at, respectively, the variables comprise both company and auditor
University of Nevada at Reno and Arizona State University.
Support for this research was provided to Steven Kaplan by
the Dean’s Council of 100, College of Business, Arizona State ‘The period between the company’s year end and the date of
University. Data for this study was collected while Steven the auditor’s report is referred to here as ‘audit delay’ to be con-
Kaplan was on sabbatical at the University of Canterbury, sistent with prior literature. Alternatively, this could be referred
New Zealand. The helpful comments of Jan Bebbington, to as ‘timeliness’. Audit delay, as defined, may be affected by
John Hasseldine, Ann Neale and the anonymous reviewers are both when the audit is started and the amount of time required
gratefully acknowledged. for conducting the audit.
 22 ACCOUNTING A N D BUSINESS RESEARCH

attributes, and include two explanatory variables Dyer and McHugh, 1975; and Whittred, 1980),
(i.e. owner controlled companies versus manager they are higher than found elsewhere. For example,
controlled companies, and gearing) which have not studies of United States companies have reported
been considered in prior research. Third, the study an average audit delay of 53 days (Garsombke,
includes a more recent time period, 1987 and 1988, 1981) and 63 days (Ashton et al., 1987). Ashton
than prior research and covers two time periods et af. (1989) report average audit delay was 55 days
in which the economic conditions in New Zealand for Canadian companies.
were very different. Whereas the economy was Several reasons may be offered to suggest why a
buoyant prior to the stock market crash of October, reduction in audit delay might be expected for a
1987, it lapsed into a deep recession afterwards. more recent sample of New Zealand companies.
Thus, it is possible to assess the sensitivity of the First, there was a slight decrease in audit delay
results across two significantly different years. between 1974 and 1976. Second, auditors have
become more sophisticated in their use of technol-
ogy which may be expected to make the audit more
Background efficient and progress faster. Finally, the audit
The New Zealand Companies Act, 1955, requires market has become more competitive. Audit firms
an audit for all public companies. The audit must were allowed to advertise beginning in 1986, and
audit tendering began in 1987. An increase in
Downloaded by [Cornell University] at 15:27 04 December 2012

be performed by a member of the New Zealand

Society of Accountants. For the period investi- competition may be expected to improve the lag
gated, the New Zealand Stock Exchange also between balance sheet date and audit report date.
required companies to file audited financial state- More comprehensive investigations of the
ments within 180 days of the company’s year end.2 determinants of audit delay have been performed
However, conversations with employees of the in the US by Ashton et af. (1987), and in Canada
New Zealand Stock exchange indicate that the 180 by Ashton et al. (1989). In the earlier study, the
day deadline is not actively enforced and that no authors examine a sample of 488 Peat Marwick
immediate financial penalties are imposed on com- Mitchell & Co. US clients for 1981. Multivariate
panies (or auditors) who fail to meet the deadline. analysis was used to evaluate the effects of fourteen
Audit delay for New Zealand public companies independent variables, including several variables
has been the topic of two previous studies (Courtis, not publicly available. Because the distribution
1976 and Gilling, 1977). For a sample of 1974 of audit delay was positively skewed the authors
companies, Courtis (1976) partitioned companies used the log of audit delay as the dependent
into quartiles by length of audit delay. He then variable. The adjusted R2 was approximately 26%,
investigated whether several explanatory variables with five variables significantly associated with the
systematically differed between companies included dependent variable. These five variables were: log
in the longest and shortest audit delay quartiles. of revenue, quality of internal controls, operational
He found that companies comprising the shortest complexity, relative mix of interim and final work,
audit delay quartile reported higher levels of in- and whether company ownership was public or
come. No differences for company size, company private.
age, number of shareholders, or annual report The Ashton et al. (1987) analysis of company
length were found between companies for the two ownership found that audit delay was significantly
audit delay quartiles. In the second study, Gilling shorter for public companies. That is, after con-
(1977) explored the effect of several additional trolling for other factors, public companies were
factors on audit delay for a sample of 1976 listed audited faster than private companies. The study
companies. Based upon descriptive statistics, also separately analysed public and private compa-
Gilling concluded that audit delay was shorter for nies to explore whether the explanatory variables
(a) companies with large auditors, (b) companies were differentially related to the two subsamples.
with overseas ownership, and (c) large companies. The results from the two subsamples were not
It should be noted that Gilling’s finding that audit similar. For example, whereas company size sig-
delay is affected by company size is at odds with nificantly affected audit delay for private compa-
Courtis. nies it was not associated with audit delay for
Courtis and Gilling reported relatively compar- public companies. Further, the adjusted R2 was
able levels of average audit delay-83 days for the much larger for public than private companies.
1974 sample and 80 days for the 1976 sample. These findings suggest that company ownership
Although these figures are similar to audit delay may directly influence audit delay, and that the
reported in Australia (Davies and Whittred, 1980; relationship of other explanatory variables may be
contingent upon the type of company ownership.
Ashton et al. (1989) investigated the multivariate
2The filing requirements for companies listed on the New
Zealand Stock Exchange were changed in 1989. Companies are relations between eight publicly available indepen-
now required to file audited financial statementswithin 90 days dent variables and audit delay over six years, 1977-
of the year-end. 1982, for 465 public Canadian companies. For each
 W I N T E R 1991 23

Table 1
Definitions of Explanatory Variables and Expected Effect on Audit Delay
Expected Relationship
Explanatory Variable Explanation with Audit Delay
Company Size ( A S T ) Total assets of company. Negative
Industry ( I N D ) Industry classification represented by a Negative
dummy variable: ‘financial’ companies
assigned a 1; otherwise a 0.
Income ( L O S S ) Sign of current year income represented by Positive
a dummy variable: companies with negative
net incomes assigned a 1; otherwise a 0.
Extraordinary Item Extraordinary item represented by a dummy Positive
(EXTR) variable: companies with an extraordinary
item assigned a 1; otherwise a 0.
Audit Opinion ( O P I N ) Type of audit opinion represented by a dummy Positive
variable: opinions other than standard were
assigned a 1; standard opinions a 0.
Downloaded by [Cornell University] at 15:27 04 December 2012

Auditor ( A U D ) Type of audit firm represented by a dummy Negative

variable: large international audit firms
were assigned a 1; otherwise a 0.
Company Year-End ( Y E ) Month of fiscal year-end represented by a Positive
dummy variable: March through June year-ends
assigned a 1; otherwise a 0.
Company Ownership Type of company ownership represented by a Negative
(0W N ) dummy variable: owner controlled companies
assigned a 1; manager controlled companies
assigned a 0.
Debt Proportion (DEBT) The proportion of debt to total assets. Positive

year a separate regression model was determined. current model. The model of audit delay also
The adjusted R2s from the six models ranged includes two additional explanatory variables: type
between 8.8% and 12.6%. Four of these variables of company ownership and debt proportion. The
(client industry, type of audit opinion, presence of hypothesised relationship between each of these
extraordinary items, loss for the year) were signifi- nine variables and audit delay, as well as the
cant for at least four of the six years. Three underlying rationale, is discussed below. Table 1
additional variables (log of total assets, fiscal year- summarises the proposed relationships between
end, and audit firm) had consistent signs across the each explanatory variable and audit delay.
six years but reached significance in three or fewer Company Size (AST). Total assets was used as
years. A dummy variable indicating a contingent a measure of company size. Total assets have
liability for the company was characterised by commonly been used to measure size in previous
weak significance and inconsistent signs. studies of audit delay (Ashton et al., 1989; Courtis,
1976; Davies and Whittred, 1980; Garsombke,
1981; Gilling, 1977; Newton and Ashton, 1989). It
Method is possible to offer reasons why company size could
The present study investigated factors affecting be either positively or negatively associated with
audit delay for New Zealand public companies for audit delay. Based upon the results of prior studies,
the years 1987 and 1988. Because of the need to however, a negative association between audit
obtain information from annual reports, the study delay and company size is expected. Several factors
was restricted to public companies. Audit delay is may account for this relationship. For example,
defined as the number of days between the date larger companies may have stronger internal con-
of the financial statements and the date of the trols, which in turn should reduce the propensity
auditor’s report. A model of audit delay was for financial statement errors to occur and enable
developed comprising nine explanatory variables. auditors to rely on controls more extensively and
The model is similar to one employed by Ashton to perform more interim work. Also, larger com-
et al. (1989), but differs in several key respects. The panies may be able to exert greater pressures on the
contingent liability variable was excluded because auditor to start and complete the audit in a timely
of the inconsistent signs reported by Ashton et al. fashion.
(1989). The remaining seven explanatory variables Industry (IND). The New Zealand Stock
used by Ashton et al. (1989) are included in the Exchange classifies companies by primary industry.
 24 A C C O U N T I N G A N D BUSINESS R E S E A R C H

This classification scheme was used to distinguish auditor-company conflict which would also tend to
between financial services companies and other increase audit delay.
types of companie~.~ Following Ashton et al. (1989), Auditor (AUD). Auditors were classified into
financial services companies were coded 1 and two groups-international accounting firms (e.g.
others were coded 0. Audit delay is expected to be members of what was at the time the Big Eight)’
shorter for financial services companies because and all other auditors. Large audit firms were
such companies typically have little or no inven- assigned a 1 and other firms were assigned a 0.
tory. Inventories are difficult to audit and represent Based on Gilling’s (1977) conclusion, audit delay
an area where material errors frequently occur. for companies with an international audit firm was
Thus, a lower percentage of inventory assets, rela- expected to be less than for audits from other firms.
tive to other types of assets, may lower audit delay International firms, because they are larger firms,
for financial services companies. might be able to audit more efficiently, and have
Sign of Income (LOSS).Companies reporting a greater flexibility in scheduling to complete audits
loss for the period were expected to have a longer on a timely basis.
audit delay, and were assigned a 1 . The remaining Company Year End ( Y E ) .The final explanatory
companies were assigned a 0. The expected role of variable from Ashton et al. (1989) is the company’s
a reporting loss on audit delay is suggested for year end. The two most common year ends for
several reasons. First, where a loss occurs, com- companies in New Zealand are March 3 1 and June
Downloaded by [Cornell University] at 15:27 04 December 2012

panies may wish to delay the bad news. A company 30, and the period between March and June is
with a loss may request the auditor to schedule considered the busy season for auditors in New
the start of the audit later than usual. Second, an Zealand.6 Year end was treated as a dummy vari-
auditor may proceed more cautiously during the able. Companies with a year end of March 31, June
audit process in response to a company loss if the 30 or between were assigned a 1 and others 0. A
auditor believes the company’s loss increases the positive association between audit delay and YE is
likelihood of financial failure or management fraud. expected. A large number of audits with the same
Extraordinary Item (EXTR).Companies report- year end date may cause scheduling problems for the
ing an extraordinary item were expected to have a auditor and increase the time to complete the audit.
longer audit delay, and were assigned 1. Compa- Company Ownership (0W N ) .Gilling (1977) and
nies without an extraordinary item were assigned Ashton et al. (1987) each contend that company
a 0. Extraordinary items by definition are indica- ownership may potentially affect audit delay. For
tive of unusual reporting items, so that additional this study, companies were classified as either owner
time may be needed for audit. Further, the auditor controlled or manager controlled, which reflects the
may have significant uncertainty as to whether a two dominant types of control among New Zealand
particular item is extraordinary or not, which, in public companies. A company was considered
turn, may lead to extended negotiations between owner controlled if 30% or more of the ordinary
the auditor and the company. stock was controlled by a single outside i n ~ e s t o r . ~
Audit Opinion (OPZN). Companies not receiving All remaining companies were categorised as man-
a standard (e.g. unqualified) audit opinion were ager controlled.* Thus, company ownership was
expected to have a longer audit delay.4 Standard treated as a dummy variable, with owner con-
audit opinions were assigned a 0, and all others trolled companies coded 1 and manager controlled
were assigned a 1. The rationale, in part, is similar companies coded 0. This definition of owner and
to the income variable. That is. comDanies receiv-
ing a qualification may view this as bad news and
’low down the audit process’ For a SAlthough Arthur Andersen did not have offices in New
Zealand, they were associated with Lawrence Anderson and
company might not respond in a fashion to Buddle (LAB). For purposes of this study, audits performed by
requests from the auditor. The receipt of a non- LAB were coded 0.
standard audit report might be symptomatic of 6Firth (1985) examined the determinants of audit fees among
New Zealand public companies. He found that company - _ year
_
end was not associated with audit fees.
’Financial companies are defined on the basis of their ’An outside investor is defined as one that is not part of
industry coding by the New Zealand Stock Exchange. Com- the management group or the Board of Directors. For purposes
panies coded 8 (financial and lending institutions), 12 (insurance of clarifying this definition, consider the following example. An
companies), or 13 (investment companies) were classified as external corporate investor owns at least 30% of the ordinary
financial companies based on the nature of their assets. stock of a company and appoints an individual to serve on the
For this study the term ‘qualified’ or ‘non-standard’ audit Board. Assume that the individual does not own at least 30% of
report includes any report which contains a significant modifi- the ordinary stock. Since the member on the Board does not
cation of the New Zealand standard auditor report. These own at least 30% of the ordinary stock, the company would be
include comments, or ‘tags’, in the audit report which emphasise classified as owner controlled.
material matters such as uncertainties or departures from 81nformation on company ownership was taken from the
normal accounting practice even though such comments may be disclosures in annual reports. Disclosures of the 20 largest
part of what is, technically, an unqualified audit opinion. The shareholders is required by the New Zealand Stock Exchange to
absence of the terms ‘subject to’ or ‘except for’ in the audit be included as part of the annual report for companies on the
report distinguish a ‘tag’ report from a ‘qualified’report. Exchange.
 WINTER 1991 25

Table 2
Descriptive Statistics for the Dependent and Explanatory Variables by Year
1987 1988
Standard Standard
Mean Deviation Percentage * Mean Deviation Percentage *
Audit Delay 87.7 35.3 95.5 46.5
AST ($000) $259,154 $986,798 $282,632 $1,183,508
IND 23% 24%
LOSS 23 % 46%
EXTR 59% 72%
OPIN 14% 23%
AUD 88% 89%
YE 70% 73%
0 WN 51 Yo 44%
DEBT 0.46 0.24 0.47 0.26
Sample 245 206
Downloaded by [Cornell University] at 15:27 04 December 2012

*Percentage of companies for which dummy variable coded as 1.

manager controlled companies has previously been Results

used by Carslaw (1988) to categorise New Zealand
public companies and by Kamin and Ronen (1978) The results are presented in two sections. In the
to categorise US public companies. A negative first section, descriptive statistics related to the
association between OWN and audit delay is sample companies are reported. This is followed by
expected. Large external investors, having an acute a section detailing the multivariate test results
need for timely information, may be expected to related to the proposed model of audit delay.
pressure the company and auditor to start and to Descriptive Statistics
complete the audit as rapidly as practicable.
Debt Proportion (DEBT). Debt proportion The annual report of the New Zealand Stock
has not previously been used as an explanatory Exchange was used to identify companies of inter-
variable for audit delay. It has, however, been used est. The report contained 263 and 239 companies
in studies of audit pricing and auditor selection for 1987 and 1988, respectively. Annual reports
(Francis, 1984; Francis and Stokes, 1986; Francis were obtained from the University of Canterbury
and Wilson, 1988; Palmrose, 1984; Simunic and Accounting Library and the New Zealand Stock
Stein, 1987). The proportion of debt to total assets Exchange. As shown in Table 2, complete inform-
is expected to be positively related to audit delay. ation was collected on 245 firms for 1987 and 206
Two reasons may be offered for this expectation. for 1988. Thus, the sample for 1987 represents over
First, the relative proportion of debt to total assets 93% of the listed companies and the sample for
may be indicative of the financial health of the 1988 represents over 86% of the listed companies.
company. A high proportion of debt to total assets Companies were excluded from the sample if com-
will increase a company’s likelihood of failure and plete information was not available or no report
may raise in the auditor’s mind additional concerns was available.’O
that the financial statements may be less reliable Descriptive statistics for sample companies for
than normal. For example, poor financial health the two years are reported in Table 2. As shown,
may raise the possibility of management fraud or for 1987 and 1988 mean audit delay was approxi-
unintentional errors related to reductions in staff mately 88 and 95 days, respectively. These levels of
and/or inadequate staff training. Such concerns mean audit delay represent an increase from those
would tend to increase the length of the audit previously reported for New Zealand public com-
engagement. Second, the audit of debt may be panies. Thus, contrary to our speculation, mean
more time consuming than the audit of equity. audit delay has not declined over time. Perhaps
Generally the audit of debt is more involved and the mean audit delay has increased because overall
complicated than the audit of equity.’
”For 1987 and 1988, respectively, four and six annual reports
did not contain complete information sets. In all cases the
yThe number of debt holders will affect the amount of time auditor’s report was not dated. Information was not available
needed to audit in this area. For example, if debt is owed to only on the remaining companies excluded from the analysis.
one source, the audit may be relatively simple and require little Conversations with the employees of the New Zealand Stock
time. On the other hand, to the extent that debt is owed to Exchange indicate that the non-filers typically represent
many, the audit is expected to require more time. companies that have gone into receivership or liquidation.
 26 ACCOUNTING A N D BUSINESS RESEARCH

Table 3
Descriptive Statistics for the Dependent and Explanatory Variables for
Owner Controlled/Manager Controlled Subsamples
Panel A: 1987
Owner Controlled Subsample Manager Controlled Subsample
Standard Standard
Mean Deviation Percentage * Mean Deviation Percentage *

Audit Delay 82.6 32.83 93.00 37.18

AST ($000) $195,609 $364,403 $325,345 $1,360,602
IND 20 % 26 %
LOSS 24% 23%
EXTR 62% 56%
OPIN 12% 18%
AUD 92% 84%
YE 70% 70%
DEBT 0.49 0.23 0.43 0.25
Downloaded by [Cornell University] at 15:27 04 December 2012

Sample 125 120

Panel B: 1988
Audit Delay 89.5 49.2 100.93 43.72
AST ($000) $215,341 $456,003 $336,963 $1,538,494
IND 22% 25%
LOSS 40% 49 %
EXTR 75% 75%
OPIN 22% 25%
AUD 91% 88%
YE 74% 72%
DEBT 0.46 0.26 0.50 0.28
Sample 92 114
*Percentage of companies for which dummy variable coded as 1.

audit firm growth has not kept pace with the overall a loss, an extraordinary item, or receiving a non-
growth in the size and number of companies listed standard audit opinion were significantly higher
on the New Zealand Stock Exchange. Or perhaps in 1988. The sample contained a lower proportion
the operating and financing environment that com- of owner controlled companies for 1988, but not
panies face has become increasingly sophisticated significantly so.
and complicated. The increase in mean audit delay Table 3 presents descriptive statistics for the
in 1988 from the prior year may be attributable, two subsamples of 0 WN. Two differences may be
in part, to the difficult economic conditions that observed between the owner controlled subsample
companies and auditors faced in 1988. and the manager controlled subsample. First,
Table 2 also indicates the presence of significant for both years audit delay for owner controlled
variability for audit delay and AST. For example, companies was approximately ten days less than
the standard deviation in audit delay was over 35 manager controlled companies. Secondly, owner
days in 1987 and over 46 days in 1988. The results controlled companies tended to be smaller com-
from the Kolmogorov D test statistic indicated panies, on average, than management controlled
that the distributions for each of these two vari- companies. In other respects, the two subsamples
ables, audit delay and AST, were not normal. were roughly equivalent, with no significant
For this reason the log of each variable is used in differences among the variables.
the regression models. Thus, the dependent
measure used in the regression models is the log Multivariate Results
of audit delay and the log of AST is used as an To explore the multivariate relationship between
explanatory variable. This approach was also used audit delay and the nine explanatory variables,
by Ashton et al. (1989) in response to non-normal a multiple regression was performed. The depen-
distributions. dent variable was the natural log of audit delay
A comparison across the two years reveals (1nAUDLA Y ) . Also, because total assets were not
three significant differences. Chi-square test results normally distributed, the natural log of total assets
indicated that the frequency of companies having (1nAST) was used as the explanatory variable for
 W I N T E R 1991 27

Table 4
Multiple Regressions of Ln Audit Delay on the Explanatory Variables
for the Overall Sample
1987 1988
Coeficient t -value Coeficient t -value

lnAST -0.062 -3.79*** - 0.061 -2.81***

IND 0.061 0.96 -0.171 -2.06**
LOSS 0.219 3.33*** 0.207 2.71***
EXTR 0.160 2.99*** 0.080 0.97
OPIN 0.139 1.85* 0.056 0.65
AUD 0.012 0.16 -0.060 -0.52
YE 0.066 1.14 0.1 10 1.38
0 WN -0.126 -2.37** -0.114 - 1.59
DEBT 0.059 0.52 0.43 1 3. IS***
Intercept 5.64 5.41
Sample 245 206
Downloaded by [Cornell University] at 15:27 04 December 2012

Adjusted R2 17.0 14.3

F-Ratio 6.55 4.8 1
***significant at 0.01.
**significant at 0.05.
*significant at 0.10.

company size. The remaining eight explanatory with the 1987 results, audit delay was negatively
variables are as described in Table 1. associated with InAST and positively associated
Table 4 presents the multiple regression results with LOSS. In addition, two explanatory variables
by year for the entire sample. For 1987, five of that were not significant in 1987 were significant in
the nine explanatory variables were significant. 1988. As shown in Table 4, financial services com-
The coefficients for these five variables were all in panies and companies with low debt proportions
the predicted direction. Audit delay was positively experienced a shorter audit delay during 1988. For
associated with the LOSS, EXTR and OPIN the five variables that did not reach significance in
explanatory variables. Thus, an increase in audit 1988, the coefficients were in the predicted direction.
delay was observed in the presence of a loss, an The adjusted RZ for the 1988 overall sample is
extraordinary item, or a qualified audit opinion. 0.143. Although a decline from 1987, this level still
Additionally, audit delay was negatively associated exceeds the levels reported by Ashton et al. (1989).
with lnAST and 0 WN." That is, on average, audit A comparison of the results from the two years
delay increased for companies with smaller total indicates that four variables were significant in
assets and manager controlled companies. Con- one but not both of the periods. Additional analy-
cerning the four explanatory variables that did not sis was conducted on two of these variables, EXTR
reach significance, the coefficients for two of the and OPIN, to determine whether the inconsistency
four were in the predicted direction. The co- may be attributed to qualitative changes over
efficients for IND and AUD were not in the time. Regarding EXTR, the proportion of income
predicted directions. The adjusted R2 for the 1987 increasing versus income decreasing items for each
overall sample is 0.170 which, although relatively year were compared." The Chi-square test results
low, is an improvement over the levels reported by indicate that year and type of extraordinary item
Ashton et al. (1989) for Canadian companies. are not significantly associated. Thus, the differ-
Considering the 1988 results, four of the nine ences in results for EXTR for the two years do not
explanatory variables were significant. Consistent appear to be related to income effects.
An analysis was also conducted on the non-
standard audit opinions for 1987 and 1988. As
"To test the sensitivity of the OWN results, additional regres-
sions for 1987 and 1988 were conducted in which a company shown in Table 5, these opinions were assigned
was considered owner controlled if 20% or more of the ordinary to one of four categories (e.g. tag emphasis, un-
stock was controlled by a single investor. Twenty percent was qualified except for departure from statement of
also used because it is generally used as a benchmark indicating standard accounting procedure or accounting
associated company status. This definition only changed the policy departure, qualified subject to going concern
classification for seven companies in 1987 and nine companies
in 1988. The results for OWN are similar to the results reported
in the paper. We use the 30% definition to be consistent with I2If a company had multiple extraordinary items, the net
prior literature. effect was determined and used in the analysis.
 28 ACCOUNTING A N D BUSINESS RESEARCH

Table 5
Classification of Non-Standard Audit Opinions by Year
Qualifiedfor Qualified
departure from subject to
standard or going concern or
Tag emphasis accounting policy asset valuation Disclaimer Total
1987 17 7 8 3 35
1988 6 11 25 5 47
- - - - -
23 18 33 8 82

or asset valuation, and disclaimer of opinion). A audit delay for each year in the total sample were
chi-square test was also employed to test for an significant in only one of the subsamples, but not
association between year and category of audit both. For 1988, DEBT was significant for both
opinion. The results indicate a significant associ- subsamples. These results suggest that company
ation (Chi-square = 13.95, p < 0.01). A review of ownership may moderate the relationship between
Downloaded by [Cornell University] at 15:27 04 December 2012

Table 5 indicates that (a) the proportion of audit audit delay and other explanatory variables. The
reports with a tag emphasis was much greater in contingent nature of the results may be highlighted
1987 than 1988, and (b) the proportion of audit by observing the pattern of results regarding the
reports qualified subject to going concern or asset lnAST and LOSS variables. For the entire sample,
valuation issues was much greater in 1988 than these two variables were significant in both 1987
1987. Thus, the differences in results for OPZN for and 1988. However, they were differentially signifi-
the two years may in part be related to changes in cant across the two subsamples. LnAST was only
the mix of non-standard opinions from one year to significant for owner controlled companies, whereas
the next. This is discussed further in the next LOSS was only significant for management con-
section. trolled companies. We discuss this finding below.
To provide further evidence on the role of Table 7 presents the correlations among explan-
0 WN on audit delay, separate multiple regressions, atory variables to investigate the possible effects
presented in Table 6, were performed on each of multicollinearity. As shown, the levels of cor-
0 WN subsample. With one exception, the four relations are relatively low, with no correlation
variables (excluding 0 W N ) significantly affecting exceeding 0.30. The correlations between lnAST

Table 6
Multiple Regression of Ln Audit Delay on the Explanatory Variables for
Owner Controlled/Manager Controlled Subsamples
1987 1988
Owner Controlled Manager Controlled Owner Controlled Manager Controlled
Regression Regression Regression Regression
CoefJicient t -value Coeficient t -value Coeflcient t -value Coeficient t -value
lnAST -0.100 -3.99*** -0.032 -1.54 -0.079 -2.13** -0.042 -1.60
IND 0.080 0.78 0.052 0.77 -0.386 -2.51** 0.001 0.01
LOSS 0.133 1.34 0.304 3.58*** 0.185 1.48 0.237 2.52**
EXTR 0.281 3.35*** 0.043 0.65 0.226 1.62 -0.067 -0.67
OPIN 0.004 0.03 0.209 2.30** 0.035 0.26 0.141 1.35
AUD -0.103 -0.70 -0.003 -0.38 -0.226 -1.08 -0.001 -0.01
YE 0.061 0.69 0.058 0.79 0.121 1.08 0.132 1.37
DEBT 0.121 0.68 0.017 0.11 0.666 2.83*** 0.278 1.71*
Intercept 6.01 5.35 5.55 5.05
Sample 125 120 92 114
Adjusted R2 16.9 16.8 19.2 10.9
F-Ratio 4.16 4.01 3.71 2.73
***significant at 0.01.
**significant at 0.05.
*significant at 0.10.
 W I N T E R 1991 29

Table 7
Correlations Between Explanatory Variables
Panel A: 1987
lnAST IND LOSS EXTR OPIN AUD YE OWN DEBT
InAST -

IND 0.20 -
LOSS 0.27 -0.03 -
EXTR -0.00 -0.12 0.04 -
OPIN 0.04 0.01 -0.21 0.06 -
AUD 0.04 -0.01 0.03 -0.02 0.02 -

YE 0.09 0.05 -0.13 -0.06 0.11 -0.04 -

0 WN 0.09 -0.06 0.02 0.07 0.08 0.12 0.01 -

DEBT 0.30 0.09 0.02 -0.05 -0.06 -0.06 -0.09 0.11 -

Panel B: 1988
lnAST -

IND 0.14 -
Downloaded by [Cornell University] at 15:27 04 December 2012

LOSS -0.23 0.08 -

EXTR 0.22 -0.08 0.07 -
OPIN 0.09 -0.03 -0.24 0.02 -
AUD 0.09 0.08 -0.06 0.01 -0.04 -

YE 0.09 0.09 -0.05 0.04 -0.08 0.07 -

0 WN 0.11 -0.04 -0.09 0.04 0.04 0.06 0.02 -

DEBT 0.27 0.12 0.06 0.17 -0.08 -0.02 -0.03 -0.06 -

and DEBT were 0.30 in 1987 and 0.27 in 1988. finding concerning company size on audit delay for
For each year two additional regressions were New Zealand public companies. The differences in
performed, first dropping lnAST and then drop- the company size results between the two studies
ping DEBT. No changes in the signs, relative sizes may be explained, in part, by the current study’s
or significance levels of the coefficients of the larger sample size and stronger statistical tests.I3
remaining variables were observed. This suggests Secondly, the existence of a loss also contributed
that multicollinearity does not pose a problem in significantly to audit delay. This result is also
interpreting the results. consistent with audit delay research from other
countries. Several underlying factors may explain
this relationship. For example, the company may
Discussion wish to defer announcement of a loss and so is less
This study examined potential explanatory vari- cooperative in supplying information to the auditor,
ables affecting audit delay for New Zealand public or the auditor agrees to complete the audit later
companies. The audit delay model used in this than usual (either by starting the audit late or by
study was an expanded version of a model used by conducting the audit slowly). Additionally, there
Ashton er al. (1989). Two explanatory variables, may be problems arising due to disputes with a
company ownership and debt proportion, that were company which is attempting to enhance results,
excluded by Ashton et al. (1989) are part of the or an auditor may proceed more slowly because of
current model. Additionally, the association of ex- heightened concerns over the possibility of fraud
planatory variables and audit delay was investigated or failure.
for two separate time periods, 1987 and 1988. This study specifically tested whether audit delay
For the entire set of sample companies, only two is affected by two explanatory variables previously
of the nine variables significantly affected audit overlooked in the literature. DEBT was significant
delay across both years. First, for both years audit in 1988 but not for 1987. The differences between
delay was inversely related to company size. This the two years may be due in part to the impact
seems to confirm previous studies conducted in of the general economy, which we discuss below.
countries other than New Zealand that efficiencies With respect to 0 WN, a significant effect was found
and stronger control systems are more likely to be for 1987. The results were in the predicted direction
found in larger companies which reflects on their
ease of audit. It may also relate to better servicing
l 3 Courtis reduced his sample size by only examining differ-
of larger clients by firms to ensure client satis- ences between two quartiles and employed univariate non-
faction. The significant company size effect found parametric statistics to determine whether differences existed
here, however, is at odds with Courtis’ (1976) between quartiles.
 30 ACCOUNTING A N D BUSINESS RESEARCH

for 1988 and approached statistical significance are generally consistent with the pattern of results
(p c 0.12). These results suggest that further re- observed by Ashton er al. (1989). In their study the
search examining company ownership appears majority of the explanatory variables examined
warranted and may contribute towards an im- had a significant effect on some but not all of the
proved understanding of audit delay. Two direc- years investigated.
tions for further research may be offered. First, The results based on the entire sample also pro-
attempts should be made to determine whether the vide clues about the role of the general economy
0 WN results generalise to other countries. The New on audit delay. For example, DEBT was found
Zealand Stock Exchange differs from the stock to be significant only for 1988 but not for 1987.
exchanges in other countries, especially the United One might speculate that this pattern of results
States, in many ways. For example, the New for DEBT may reflect the impact of the general
Zealand Stock Exchange comprises a relatively economy. A high debt proportion may not be an
small number of listed companies that are approxi- important signal of poor financial health when the
mately evenly divided between owner controlled economy is buoyant. That is, when the economy is
and manager controlled companies. Also of import- healthy the ability to repay debt may appear
ance are differences across countries in regulatory strong. Thus, DEBT was not significant in 1987.
requirements and institutional arrangements, which However, this ratio may be much more sympto-
could affect the observed relationships. Thus, it is matic of poor financial health when the economy
Downloaded by [Cornell University] at 15:27 04 December 2012

inappropriate to generalise the results from this is in a recession. In this case, the ability to generate
study to other countries. Secondly, operationalis- earnings to repay the debt may be much more
ation of the OWN variable may be refined. For in doubt. Consequently, in 1988 we observe a
example, further research could partition owner significant effect for DEBT.
controlled firms between those with associated ZND was significant in 1988 but not in 1987.
company status (e.g. 20%-50% ownership by a Concerning these results, we speculate that in a
single investor) and subsidiary status (e.g. over healthy economy the audit implications stemming
50% ownership by an in~estor).’~ from financial and non-financial resources may be
The results also showed that the AUD was not similar. However, this may not be the case during
significant in either year. This result is consistent times of a severe economic downturn. In a poor
with Garsombke’s (1981) analysis of United States economy valuation issues may be particularly
companies and generally consistent with Ashton problematic for non-financial resources relative
et al.’s (1989) analysis of Canadian companies, to financial resources. That is, the increased
who found AUD significant in only one of six uncertainties stemming from an economic down-
years. The results reported here, however, are in turn may interfere with the auditor’s ability to
contrast to the conclusion reached by Gilling determine an objective valuation for non-financial
(1977), based on an analysis of New Zealand public assets to a greater extent than for financial
companies. This difference could be due either to assets.
differences in methodology or to changes in the A comparison of the two years also produces
audit market. Since Gilling’s study, the concen- some counterintuitive results. First, EXTR was
tration of Big Eight auditors among New Zealand significant in 1987 but not in 1988. Given the
public companies has increased substantially. decline in economic conditions, one might have
Thus, the small number of observations of com- expected EXTR to have a stronger impact on audit
panies using non-Big Eight auditors diminishes the delay for 1988 than 1987. Perhaps the results reflect
ability to detect audit firm differences. the change in the frequency of EXTR. As the
Overall seven of the nine variables reached frequency of extraordinary items increased to over
significance (p c 0.10) for at least one of the two 70% in 1988, the audit in this area may have
years. Variables may have been significant in one become more routine. Thus, the presence of an
year but not both because of differences in the extraordinary item did not affect audit delay.
samples across the two years. Approximately The results for OPZN, significant in 1987 but
20% of New Zealand public companies went into not in 1988, were also counterintuitive. This is
receivership in the eighteen months following the particularly the case given the analysis reported in
1987 stock market collapse. To the extent that Table 5. As shown, more serious qualifications
going into receivership is not a random event, were more common in 1988, yet OPIN did not
differences in the samples may, in part, explain the affect audit delay. Again, perhaps, the audit of
lack of consistent finding for the two years. These companies receiving non-clean opinions may have
results, however, still provide partial support for become more routine as their frequency increased.
the proposed audit delay model. Further, the results Or, perhaps, companies anticipating that non-stan-
dard opinions would be more common, placed less
I4It should be noted that the proposed refinement of the importance on receiving a standard opinion in
ownership variable would use 20% as the cut-off point between 1988. Thus, companies might have been more
owner controlled and manager controlled firms. accepting of a non-standard audit opinion.
 W I N T E R 1991 31

When partitioned by the nature of company evidence presented here and in Ashton et al. (1987)
ownership, the results indicated that AST signifi- suggests that ownership moderates the relationship
cantly affected only the owner controlled com- between other explanatory variables and audit
panies, whereas LOSS significantly affected only delay. It is also apparent that large differences in
manager controlled companies. That is, although mean audit delay exist across countries. Research
the coefficients were in the predicted directions, has yet to provide an adequate explanation for
AST did not significantly affect manager controlled such differences among countries. A fruitful av-
companies nor did LOSS significantly affect owner enue of research would be a study containing
controlled companies. We believe the results related samples from multiple countries.
to the LOSS variable are particularly interesting
and offer a possible explanation. Consider that References
audit delay was not influenced by LOSS when
Ashton, R. H., P. R. G r a d and J. D. Newton, ‘Audit Delay
the company was owner controlled. An owner and the Timeliness of Corporate Reporting’, Contemporary
controlled company has, as defined, a highly con- Accounting Research, Spring, 1989, pp. 657-673.
centrated external ownership. For example, the Ashton, R. H., J. J. Willingham and R. K. Elliott, ‘An
financial report may be part of the financial state- Empirical Analysis of Audit Delay’, Journal of Accounting
ments of a larger reporting entity. Such a highly Research, Autumn 1987, pp. 275-292.
Downloaded by [Cornell University] at 15:27 04 December 2012

Carslaw, C. A. P. N., ‘Anomalies in Income Numbers: Evidence

concentrated ownership will be able to exert strong of Goal Oriented Behavior’, Accounting Review, April 1988,
influence on management, which, in turn, can pp. 321-327.
influence the auditor to provide timely service, Chambers, A. E. and S. H. Penman, ‘Timeliness of Reporting
regardless of the sign of income. In contrast, and the Stock Price Reaction to Earnings Announcements’,
Journal of Accounting Research, Spring 1984, pp. 21-47.
for manager controlled companies where strong Courtis, J. K., ‘Relationships Between Timeliness in Corporate
external ownership does not exist, audit delay was Reporting and Corporate Attributes’, Accounting and
a function of the sign of income. Business Research, Winter 1976, pp. 45-56.
Perhaps the results also shed light on why LOSS Davies, B. and G. P. Whittred, ‘The Association Between
affects audit delay. If a loss increased audit risk and Selected Corporate Attributes and Timeliness in Corporate
Reporting: Further Analysis’, Abacus, June 1980, pp. 48-60.
required additional audit work, then LOSS should Dyer, J. D. and A. J. McHugh, ‘The Timeliness of the Aus-
be significant across both samples. These results tralian Annual Report’, Journal of Accounting Research,
tentatively suggest that a reporting loss increases Autumn 1975, pp. 204-219.
audit delay because of management’s desire to Feltham, G. A. Information Evaluation, Studies in Accounting
Research no. 5, Sarasota, Fla.: American Accounting Associ-
delay the reporting of bad news. Separate analysis ation, 1972.
of OWN samples also showed that company size Firth, M., ‘An Analysis of Audit Fees and Their Determinants
did not significantly affect audit delay for manager in New Zealand’, Auditing: A Journal of Practice & Theory,
controlled companies. Apparently, company size Spring, 1985, pp. 23-37.
is associated with other offsetting factors that Francis, J. R.,‘The Effect of Audit Firm Size on Audit Prices:
A Study of the Australian Market’, Journal of Accounting and
adversely affect audit delay for manager controlled Economics, August 1984, pp. 133-151.
companies but not owner controlled companies. Francis, J. R. and D. J. Stokes, ‘Audit Prices, Product Differ-
In evaluating the results presented here (as well entiation, and Scale Economies: Further Evidence from the
as audit delay research in general) an important Australian Market’, Journal of Accounting Research, Autumn
1986, pp. 383-393.
limitation should be noted. The proposed relation- Francis, J. R. and E. R. Wilson, ‘Auditor Changes: A Joint Test
ship between an independent variable and audit of Theories Relating to Agency Costs and Auditor Differ-
delay is typically couched in terms of either a entiation’, Accounting Review, October 1988, pp. 663-682.
presumed effect on the timing of the start of the Garsombke, H. P., ‘The Timeliness of Corporate Financial
audit or the speed of conducting the audit. With Disclosure’, in J. K. Courtis (ed.), Communication via Annual
Reports, AFM Exploratory Series No. 11, University of New
the exception of Ashton et al. (1987), audit delay England, Armidale, N.S.W., 1981, pp. 204-218.
research has not corroborated the association Gilling, D. M., ‘Timeliness of Corporate Reporting: Some
between independent variables and actual audit Further Comment’, Accounting and Business Research,
events. Ashton et al. (1987) initiated work on this Winter 1977, pp. 34-36.
Givoly, D. and D. Palmon, ‘Timeliness of Annual Earning
issue by treating the relative amount of interim Announcements: Some Empirical Evidence’, Accounting
audit work performed on a company as an inde- Review, July 1982, pp. 486-508.
pendent variable. Further research should attempt Kamin, J. Y. and J. Ronen, ‘The Smoothing of Income
to incorporate additional audit information such Numbers: Some Empirical Evidence on Systematic
as when the audit started, and the number and Differences Among Management Controlled and Owner-
Controlled Firms’, Accounting, Organizations and Society,
kind of personnel assigned to the audit. 1978, pp. 141-157.
In closing, we may say that the results provide Kross, W. and D. A. Schroeder, ‘An Empirical Investigation of
further evidence on the determinants of audit the Effect of Quarterly Earnings Announcement Timing on
delay. The ability to predict audit delay remains Stock Returns’, Journal of Accounting Research, Spring 1984,
pp. 153-176.
relatively low. Thus, current models of audit delay Newton, J. D. and R. H. Ashton, ‘The Association Between
are incomplete. Several additional directions for Audit Technology and Audit Delay’, Auditing: A Journal of
further research may be offered. For example, the Practice & Theory, Supplement 1989, pp. 22-37.
 32 ACCOUNTING A N D BUSINESS RESEARCH

Palmrose, Z., ‘The Demand for Quality-DifferentiatedAudit Issues, Canadian Certified General Accountants’ Research
Services in an Agency-Cost Setting’, Auditing Research Foundation, 1987.
Symposium, University of Illinois, 1984, pp. 229-252. Whittred, G. P., ‘Audit Qualification and the Timeliness of
Simunic, D. and M. Stein, Product Differentiation in Corporate Annual Reports’, Accounting Review, October
Auditing: A Study of Auditor Effects in the Market for New 1980, pp. 563-511.
Downloaded by [Cornell University] at 15:27 04 December 2012