-
The Impact of Mandatory versus Voluntary Auditor Switch on Stock
Liquidity in the Korean Market
Ferdinand A Gul Monash University, Sunway Campus, Malaysia
Woo-Jong Lee
The Hong Kong Polytechnic University
Abstract
Using Korean-listed firms subject to the auditor designation
rule, this paper shows that, 1)
on average, stock liquidity is negatively related to auditor
switches, and that 2) the negative
liquidity effect of the auditor switch is concentrated in firms
that switch from high-quality
auditors to low-quality auditors. Meanwhile, firms with
designated auditors exhibit even
higher stock liquidity when they have higher audit risk,
consistent with enhanced audit
quality as a result of auditor designation.
Data Availability: The data are available from the sources
identified in the paper.
JEL Classification Code: G14, M48
Keywords: Auditor switch; stock liquidity; auditor
designation
_________________ * We gratefully acknowledge comments from
Sunhwa Choi, Simon Fung, Lee-Seok Hwang, Yongtae Kim, Jay Junghun
Lee, Se-Yong Lee, Jeffrey Pittman, Bahlgeun Roh, B. Charlie Sohn,
Wilson Tong, and Cheong H. Yi. We also thank participants at Monash
University, Kuala Lumpur seminar workshop for their helpful
comments. We thank Cheong H. Yi for generously providing auditor
designation data. We acknowledge financial support for this
research from The Hong Kong Polytechnic University. All errors are
our own. ** Correspondence: Woo-Jong Lee, School of Accounting and
Finance, The Hong Kong Polytechnic University, Hung Hom, Kowloon,
Hong Kong; Tel: +852 2766 4539; Fax: +852 2330 9845; Email:
[email protected].
-
1
The Impact of Mandatory versus Voluntary Auditor Switch on Stock
Liquidity in the Korean Market
1. Introduction
This paper investigates the association between auditor switches
and stock liquidity.
Auditors play an important role in determining the quality of
financial reporting and are
responsible for its effect in the stock market (Pitman, 2004).
Chief Justice Burger in a
unanimous decision of the U.S. Supreme Court had this to say
about the responsibility of
auditors to the public (Wild et al. 2001, p. 129):
By certifying the public reports that collectively depict a
corporations financial status, the independent auditor assumes a
public responsibility transcending any employment relationship with
the client. The independent public accountant performing his
special function owes ultimate allegiance to the corporations
creditors and stock holders, as well as the investing
public..Public faith in the reliability of a corporations financial
statements depend upon the public perception of the outside auditor
as an independent professionif investors were to view the auditor
as an advocate for the corporate client, the value of the audit
function might itself be lost.
Such a role of auditors in independently verifying financial
statements could be more
important in emerging markets where the quality of financial
reporting is generally poorer
(Francis and Wang, 2008). In addition, there is a growing demand
for an investigation of the
audit quality in emerging market countries because investors are
increasingly looking to these
markets to diversify their investment portfolios (Levich, 2001).
At a general level, however,
the literature on emerging markets has relatively neglected the
role of auditors in those
markets. We aim to fill this gap by examining how stock
liquidity is related to firms decision
to switch their incumbent auditors.
-
2
In this paper, we focus on Korean-listed companies for the
several reasons. First, the
Korean audit market enables us to compare the effects of
voluntary auditor switches and
mandatory auditor switches on liquidity. The Korean regulatory
authority, the Securities and
Futures Commission (SFC), mandates firms that have a propensity
to manage earnings to
replace their incumbent auditors with designated auditors
(a.k.a., auditor designation rule).
This in turn enables researchers to examine differential impacts
of two types of auditor
switches. Second, the Korean stock market is characterized by
high market liquidity.
According to Lesmond et al. (2004) and Lesmond (2005), while
U.S. firms listed in NYSE
and AMEX have approximately 23.5 percent zero returns over an
annual trading period,
Korean-listed firms exhibit about 13 percent zero returns,
indicating that the Korean stock
market is more liquid than the U.S. market.1 Third, there is a
high frequency of auditor
changes in the Korean audit market. In our sample spanning from
1995 to 2006, about 20
percent of our sample switch auditors, which is relatively high
compared with U.S. firms.2
Collectively, both high liquidity and frequent auditor switches
provides us with a better
setting to investigate the relation between auditor switches and
stock liquidity.
1 This high liquidity of Korean stocks, at least in part, can be
attributed to the significant presence of foreign investors in the
Korean stock market. Unlike other emerging markets, the Korean
capital market is widely open to foreign investors that demand high
quality financial reporting and provide sufficient liquidity,
especially in developing markets (Doidge et al., 2009). For
instance, in 2006, foreign investors held 35.16 percent of the
total shares in the Korean Stock Exchange, while domestic
institutional investors held only 20.08 percent. The average
foreign ownership of the 20 largest listed companies (based on
market capitalization) was 52 percent as of 2006. 2 Highly frequent
auditor switches may be attributable to the discrepancy between
high demand and low supply for high quality audit services. First,
the market share of the Big 4 auditors in Korea is very low
compared with that in the U.S. According to Choi and Wong (2007),
the market share of the Big 4 auditors in Korea is 66.36 percent,
while that in the U.S. is 95.79 percent. Second, as Fan and Wong
(2005) indicate, where legal institutions are less developed, firms
with serious agency problems tend to employ the Big 4 auditors to
alleviate them. Further, the Korean capital market is largely open
to foreign investors that demand high quality financial reporting
(Doidge et al., 2009). Such high demand for high-quality audit
services, accompanied by the low supply of the Big 4 audit
services, creates a unique opportunity for these Big 4 auditors to
rebalance their audit client portfolios with ease.
-
3
In this paper, we first analyze the differences in stock
liquidity across firms that switch
from high-quality to low-quality auditors (and vice versa) and
firms that do not switch
auditors. In doing so, we employ three stock return-based
constructs of stock liquidity: the
information cost as measured by Lesmond et al. (1999), the
proportion of zero returns, and
the effective spread as measured by Roll (1984). We expect these
stock return-based
constructs to reflect credibility of financial reporting. Using
factor analysis, we identify a
liquidity factor, which captures the information contained in
our three individual liquidity
variables. We then compare the liquidity of switching firms and
non-switching firms. 3
Second, we examine whether or not the effect of auditor switches
on liquidity is also
pronounced in firms that are required to change auditors under
the auditor designation rule.
While auditor switches generally reflect the tendency of
managers for opinion-shopping, Kim
and Yi (2009) document that mandatory auditor switches by
auditor designation improves
accounting quality. We thus posit that investors reaction to
mandatory auditor switches will
be different from that to voluntary auditor switches.
The premise of this paper is that auditor quality is positively
associated with credibility
of corporate reporting, and hence stock liquidity. This premise
is consistent with the notion
that high-quality auditors produce reliable information about
firms that may efficiently
3 Another set of conventional liquidity measures may include
Amihuds (2002) price impact measure and the bid-ask spread. Because
the use of these measures drastically reduces the sample size, we
report the results of these tests only as supplementary evidence
under the additional tests section at the end of the paper.
Meanwhile, there is no reason to believe that the use of such
high-frequency measures bring about different results.
Specifically, Goyenko et al. (2009) have shown that low-frequency
measures (i.e., stock return-based measures) capture high-frequency
measures (i.e., transaction-based measures) well.
-
4
resolve contracting problems (e.g. Jensen and Meckling, 1976;
Watts and Zimmerman, 1986;
Francis et al., 1999; Pittman and Fortin, 2004). For example,
high-quality auditors can
enhance the credibility of financial statements by improving the
precision of the earnings
reports of firms (DeAngelo, 1981; Balvers et al., 1988,
Kanagaretnam et al., 2010), thereby
lowering the firms cost of capital (Fortin and Pittman, 2007).
Both theory and empirical
evidence are consistent with this link (Easley and OHara, 2004;
Francis et al., 2005; Leuz
and Verrecchia, 2007). Furthermore, the literature also
documents that earnings quality is
positively associated with stock liquidity. Lang, Lins, and
Maffett (2011) document greater
liquidity for firms with less evidence of earnings management.
Therefore, the quality of
auditors in assuring the quality of financial reporting is to be
reflected in stock liquidity.
However, both theory and empirical evidence on the economic
effects of an auditor
switch are somewhat incomplete and difficult to interpret.
Theories suggest that the direction
of the capital market reaction to auditor change cannot be
clearly anticipated. First, given that
auditor choice is an optimal firm decision, only auditor changes
that enhance firm value
should be initiated, suggesting that the market should react
positively to such changes. If a
change in auditor is fully anticipated by market participants,
the market price of the firms
stock will then reflect its value. Therefore, the market price
should be, at worst, zero in an
efficient market (Johnson and Lys, 1990). Second, if managers
change auditors solely to
maximize their own utility rather than to maximize shareholder
wealth, market reactions to
-
5
auditor changes will be negative. Furthermore, if a firm changes
its incumbent auditor to
signal its future prospects, only a change from a high-quality
auditor to a low-quality auditor
will reveal negative private information on firm prospects,
leading to a short-term negative
impact on the value of a firm in a market with weak
informational efficiency. The upshot of
these different arguments is that it is difficult to predict the
market reaction to an auditor
switch.
Furthermore, the findings of empirical studies based on U.S.
data remain inconclusive.
Earlier empirical evidence shows a negative or insignificant
market reaction to auditor
change (Fried and Schiff, 1981; Nichols and Smith, 1983; John
and Lys, 1990). By focusing
on a shorter event-window, later studies succeeded in finding a
negative market reaction (e.g.,
Whisenant et al., 2003; Beneish et al., 2005; Knechel et al.,
2007). On the other hand, other
studies document a positive or non-negative market reaction
(Sankaraguruswamy and
Whisenant, 2004; Whisenant, 2006).4 In sum, evidence on the
economic consequences of an
auditor switch remains insufficient.5
Our empirical tests use two sets of data to address the
implications of an auditor switch.
First, using a large sample of Korean-listed firms that exhibit
a high frequency of auditor
4 As for Korean evidences, since Korean firms started to
disclose auditor switch news after 2002, it was practically
impossible to examine stock market responses around auditor switch
news in Korea before 2002. Using 201 Korean firms in 2002 and 2003,
Sohn and Kim (2004) find that stock market negatively responds to
auditor switch. 5 It is also argued that the economic consequences
of an auditor switch vary with the type of the switch. For
instance, regulators expect switches from a low-quality auditor
(i.e., a non-Big 4 auditor) to a high-quality auditor (i.e., a Big
4 auditor) to enhance the credibility of the firms financial
statements, improve corporate transparency, and hence benefit
investors (e.g., Securities and Exchange Commission, 1988). This
therefore implies that the economic consequences of a change in
auditor are asymmetric to the type of auditor change (e.g.,
Eichenseher et al., 1989). Extant evidence fairly consistently
supports that investors interpret switches negatively when the
analysis more narrowly isolates those involving downgrades (i.e.,
from Big N to non-Big N).
-
6
switches, we analyze the effects of auditor switches on stock
liquidity. Second, by
categorizing our sample into firms subject to mandatory auditor
change (i.e., where the
auditor is designated by the SFC) and those that change their
auditor on a voluntary basis (i.e.,
where the auditor switch is initiated by the client firm), we
examine whether or not auditor
switches that are mandatory also have an effect on stock
liquidity. The SFC has assigned
statutory auditors to firms suspected of accounting fraud under
a prescribed set of conditions
since 1990. Given that this type of regulatory designation
forces firms that would not
otherwise switch auditors, we expect an auditor switch initiated
by the rule to have a non-
negative effect on stock liquidity. More details on the
institutional uniqueness of the auditor
designation rule in Korea are presented in Section 2.
In this work, we run cross-sectional firm-level regressions
using these two datasets and
find lower stock liquidity for firms that switch auditors than
those that do not. Further
analyses reveal that this effect is concentrated on BN firms
(i.e., firms that switch from Big 4
to non-Big 4 auditors).6 In one of our specifications that use
zero return frequency as a proxy
for stock liquidity, the percentage of days without trades is
higher in BN firms by 1 percent,
which represents about a 3 percent deterioration in liquidity
relative to firms that do not
switch auditors. Even when we aggregated all three liquidity
proxies into a single liquidity
factor for parsimony, the tenor of the results does not change.
6 We used NB, BN, BB, and NN to designate firms that switch from
non-Big 4 to Big 4 auditors, firms that switch from Big 4 to
non-Big 4 auditors, firms that switch from one Big 4 to another Big
4 auditor, and firms that switch from one non-Big 4 to another
non-Big 4 auditor, respectively. Likewise, our industry specialty
tests also used similar terms, such as NS, SN, SS and SS, where N
(S) denotes a non-specialist (specialist) auditor.
-
7
We also relax the assumption that a firms decision on switching
auditors is exogenous.
In reality, the auditor choice decision is likely to be
endogenous. We use both Heckman
(1979)s 2 stage regression approach and propensity score
matching to address the concern,
thereby ensuring that our findings are not due to the
endogeneity of the auditor switching
decision, but are more likely to represent a causal
relationship.
While the liquidity effects for firms switching auditors are
statistically and
economically significant, these effects are generally smaller
for firms whose auditor switches
are mandated by the regulatory body than firms voluntarily
switching auditors. This suggests
that investors are more concerned about voluntary auditor
switches than mandatory auditor
switches. Interestingly, we find that firms with designated
auditors do not exhibit lower
liquidity, consistent with the market perception that auditor
designation results in enhanced
audit quality. When a new auditor is assigned to a firm
suspected of earnings manipulation
under certain conditions, such as being subject to high default
risk, investors expect the
auditor designation to improve the credibility of the firms
financial statements, which is
consistent with the view that designated auditors improve audit
quality (Kim and Yi, 2009).
This study makes several important contributions to the
literature. First, while prior
studies have examined different dimensions of liquidity in
different markets (Bacidorea et al.,
2005; Frino et al., 2008; Kryzanowski et al., 2010), in this
study we focus on Korea, a rapidly
developing economy. In this way we extend the vast literature on
liquidity issues, in general,
-
8
and the links between financial reporting and liquidity, in
particular. Second, prior studies
regarding the link between the economic consequences of an
auditor switch and stock returns
are inconclusive and mixed. In this paper, we instead focus on
stock liquidity around the time
of the auditor switch as an alternative proxy for the capital
market effect. As the only
exception, Boone and Raman (2001) find that liquidity decreases
for firms whose auditors
resigned, but not for firms whose auditor is dismissed. While
they focus on auditors
resignation to adjust client portfolios, we discuss clients
incentives to switch auditors and
extend their findings. Third, the concern of Korean regulators
in auditor switches is that
auditor changes are motivated by management opportunism. Hence,
the findings should be of
interest to regulators and policy makers in Korea as well as in
any other countries considering
the adoption of similar regulation. Fourth, our study makes a
novel contribution in that it
examines the effect of a mandatory auditor switch on liquidity.
This is particularly interesting
because the economic consequences of the auditor switch depend
on who actually initiates
the decision to switch auditors. Finally, our paper adds to the
general body of knowledge
concerning the role played by high quality audits in reducing
information asymmetry in the
market (Autore et al., 2009; Pittman and Fortin, 2004).
The remainder of the paper is organized as follows. Section 2
develops our hypotheses
and reviews the literature. Section 3 specifies the research
design we use for hypothesis
testing and Section 4 presents the empirical results. Section 5
summarizes and concludes this
-
9
paper. In the Appendix, we provide additional details on the
construction of the three
liquidity proxies.
2. Literature review and hypothesis development
2.1. Auditor switch and stock liquidity
Rational managers may maximize their own utility at the expense
of the shareholders
rather than boosting the shareholders wealth by enhancing firm
value. However, it would be
still fair to assume that the objective of the auditor change
decision is the maximization of
firm value. Empirical evidence in this line of research is
largely inconclusive. Most
importantly, while some of the prior literature shows that
opportunistic managers switch
auditors to obtain more favorable audit opinions (Lennox, 2000),
other empirical evidence on
opinion shopping is not supportive of this finding because
post-switch opinions are no more
favorable than pre-switch opinions (e.g., Krishnan and Stephens,
1995). Numerous papers
that use stock returns to proxy for the capital market effect
document that investors react
negatively to auditor changes and that stock price drops vary
cross-sectionally with auditor or
firm characteristics (Smith, 1988; Wells and Loudder, 1997; Shu,
2000; Whisenant et al.,
2003; Beneish et al., 2005; Knechel et al., 2007). While the
prior studies noted above report a
negative response to auditor switch announcements, some studies
document a positive or
non-negative market reaction (Sankaraguruswamy and Whisenant,
2004; Whisenant, 2006),
thereby raising questions over the opportunistic auditor switch
assumption.
-
10
While tests based on stock returns conducted in prior studies
are aimed at reflecting
perceived changes in firm value surrounding the auditor switch,
our tests attempt to
determine whether or not the auditor switch implies the
willingness of investors to face risks
stemming from changes in the credibility of the firms financial
statements. An important
distinction between stock return and liquidity tests is that the
former reflects changes in the
expectations of the market as a whole, while the latter reflects
changes in the expectations of
individual investors. Although news of an auditor switch may be
neutral in that it does not
change the expectations of the market as a whole, it may greatly
alter the expectations of
individual investors. In an extreme case, we can envisage a
shift in portfolio positions, which
is captured by a change in liquidity without any price reaction.
If so, a return-based test may
be less sensitive to news of an auditor switch than a
liquidity-based test. In a similar vein, we
expect stock liquidity to be more sensitive to auditor switch
news than stock returns to the
extent that we assume informed and uninformed traders react
differently to such news.
Informed traders are likely to be indifferent to information
about the credibility of financial
statements because of their greater access to private
information about the firm, while
uninformed traders may indeed be affected by such an
announcement. Stock liquidity then
better captures the heterogeneity in the expectations of
informed and uninformed traders than
do stock returns. Hence, to better examine the market perception
of auditor changes, we
-
11
employ stock liquidity instead of stock returns and test the
following hypothesis in alternative
form:
H1A: The stock liquidity of firms that switch auditors is
significantly different from that of
firms that do not switch auditors, other things being equal.
The premise underlying this hypothesis is that an auditor switch
conveys new
information to capital market participants about the firms
future prospects. We expect that
switching from a high-quality auditor to a low-quality auditor
(from a low-quality auditor to a
high-quality auditor) makes corporate reporting less (more)
credible, increases (decreases)
information asymmetry, and hence decreases (increases) stock
liquidity.
2.2. Effect of auditor designation
In Korea, the regulatory authority designates external auditors
for firms deemed to have
strong incentives and/or great potential for fraudulent
accounting and orders these firms to
replace their incumbent auditors, appoint new auditors, and
retain them for a certain period.
This regulatory regime is known as auditor designation or
mandatory auditor assignment.
(Kim and Yi, 2009; Jeong et al., 2005).
-
12
The auditor designation rule in Korea is unique and different
from mandatory auditor
rotation in three ways. By designating auditors for problematic
firms, the auditor designation
rule aims to alleviate alleged problems arising from long-term
relationships between auditors
and clients while mandatory auditor rotation typically requires
auditor changes for all firms
after a certain period of an initial engagement. Furthermore,
the auditor designation rule
mandates retention of newly designated auditors for a certain
period to protect designated
auditors from early dismissal, and thereby mitigating a major
threat to auditor independence.
Finally, by requiring the regulatory authority to choose
incoming auditors, the auditor
designation rule enhances auditor independence, and thereby
improving audit quality. To our
knowledge, no country other than Korea has ever introduced or
considered an auditor
designation policy. We take advantage of this unique
institutional feature to examine whether
differences in stock liquidity are less pronounced in firms with
designated auditors.
Empirical evidence suggests that the appointment of a designated
auditor may result in
higher audit quality. Jeong et al. (2005) find that designated
auditors charge significantly
higher audit fees than others. More recently, Kim and Yi (2009)
find that the auditor
designation rule is effective in deterring managers from
engaging in income-increasing
earnings management. They also find that firms subject to
mandatory auditor changes
effected through the auditor designation regime report
significantly lower discretionary
-
13
accruals than firms that make voluntary auditor changes,
suggesting that the appointment of a
designated auditor enhances audit quality.
To the extent that the auditor designation rule is effective in
improving auditor
independence and thus audit quality, designated auditors are
likely to be less lenient towards
managerial opportunism than are non-designated auditors. As a
consequence, firms with
designated auditors should exhibit higher stock liquidity, or at
least they should not exhibit
lower stock liquidity. We expect the liquidity effects of
mandatory auditor changes (i.e.,
auditor designation) to be less pronounced than those of
voluntary auditor changes and thus
test the following hypothesis.
H2: The stock liquidity effects of auditor switches are less
pronounced when firms are
mandated to switch auditors.
3. Research design and sample composition
3.1. Research design
To test our hypotheses, we specify the following regression
model linking stock
liquidity with auditor switch and other control variables.
Liquidityit = 0 + 1 SWITCHit + 2 SIZEit + 3 LEVit + 4 ATURNit +
5 ROAit + 6 SGROWit + 7 QUICKit + 8 RETVARit + Year + Industry +
it, (1)
where for firm i and year t:
-
14
Dependent Variables Liquidity stock liquidity proxies; LOT,
ZERO, ROLL, and FACTOR LOT = The natural logarithm of a yearly
estimate of the transaction costs implied
by the trading behavior of investors developed by Lesmond et al.
(1999). The estimation period spans month -5 through month +7 so
that the firms annual reports are publicly available for a few
months, which should enable markets to impound the auditor switch
effect.
ZERO = The proportion of zero daily returns out of the maximum
potential number of trading days in a given year. The measurement
period spans month -5 through month +7 relative to the firms fiscal
year-end.
ROLL = Rolls (1984) effective spread based on the bid-ask
bounce-induced negative serial correlation in returns. The
measurement period spans month -5 through month +7 relative to the
firms fiscal year-end.
FACTOR = a liquidity factor based on factor analysis using LOT,
ZERO, and ROLL
Independent Variables Switch = an indicator variable which is
one if a firm switches its auditor and zero
otherwise BN = an indicator variable which is one if a firm
switches from a Big 4 auditor to
a non-Big 4 auditor and zero otherwise NB = an indicator
variable which is one if a firm switches from a non-Big 4
auditor to a Big 4 auditor and zero otherwise BB = an indicator
variable which is one if a firm switches from one Big 4 auditor
to another Big 4 auditor and zero otherwise NN = an indicator
variable which is one if a firm switches from one non-Big 4
auditor to another non-Big 4 auditor and zero otherwise Control
Variables
SIZE = the natural logarithm of the book value of total assets
LEV = financial leverage measure by the ratio of total debt to
total assets ATURN = asset turnover measured as sales deflated by
assets ROA = net income deflated by total assets SGROW = sales
growth measured as the difference in sales deflated by lagged sales
QUICK = current assets deflated by current liabilities RETVAR =
return variability measured as the standard deviation of daily
returns
To conduct the empirical tests, we create a binary indicator
variable, Switch, that takes
the value of one for firms that switch auditors. This variable
should capture the average
liquidity effect around the time of the auditor switch as the
key variable of interest. We
introduce four separate indicator variables for firm-year
observations from firms that switch
-
15
auditors. We distinguish BN, NB, BB, and NN switchers,
respectively, depending on the types
of the former and successor auditors.
For dependent variables, we use four proxies LOT, ZERO, ROLL,
and FACTOR for
stock liquidity.7 We construct the first three proxies using
daily individual stock returns and
aggregate market returns. Assuming that daily returns are
normally distributed, we estimate
all three measures for each firm and year over the period from
month -5 through month +7
relative to the firms fiscal year-end.8 The first dependent
variable, LOT, is a yearly estimate
of the total round trip transaction costs implied by the trading
behavior of investors developed
by Lesmond et al. (1999). The second dependent variable, ZERO,
is the proportion of zero
daily returns out of the maximum potential number of trading
days in a given year. Lesmond
et al. (1999) argue that zero-return days occur when the
transaction costs of trading outweigh
the value of new information not yet contained in prices and it
is better not to trade. The zero-
return metric commonly serves as a proxy for illiquidity (e.g.,
Bekaert et al., 2007). The third
measure is ROLL, which is designed to capture the bid-ask
bounce-induced negative serial
autocorrelation in returns to estimate the effective spread. If
the serial autocovariance is
positive, we force it to be negative and use the Roll estimate
as if a negative serial
autocovariance were being estimated (Harris, 1989). Harris
(1990) explains that positive
7 In Table A1, we also report the results based on the price
impact measure and the bid-ask spread. We avoid using these two
measures in the main analyses because of limited data availability.
8 Korean listed firms have not been required to disclose auditor
switch news before 2002. Without the exact date of auditors being
replaced, we are unable to examine short-term market responses
around auditor switch dates. Admitting potential confounding
effects can exist in our long-horizon test, we also use two
different estimation periods for our liquidity proxies (month -6 to
month +6, and month -4 to month +8). Our results hold robust to the
changes in estimation periods.
-
16
autocovariance can result from closing prices that cluster at
the ask, violating Rolls
assumption of trade independence. Due to the strengths and
weaknesses of each liquidity
measure and proxy, we employ all three measures to construct
FACTOR using factor analysis
and use it as a separate dependent variable in the analyses.9
Note that higher values of all of
our liquidity proxies indicate lower stock liquidity by
definition. In common with prior
studies, we control for firm size, leverage, asset turnover,
return-on-assets ratio, sales growth,
the quick ratio, and return variability in the regressions
(Chordia et al., 2000; Leuz and
Verrecchia, 2000).
3.2. Sample composition
The initial sample for our study consists of all firms listed on
the Korea Stock Exchange
(KSE) and the Korea Securities Dealers Automated Quotation
(KOSDAQ). We retrieve
financial data from the KIS-Value database developed by the
Korea Information Service
(KIS), the top credit rating agency affiliated with Moodys. This
database has been used in
several recent studies on accounting and finance (Kim and Yi,
2009). We obtain monthly
stock returns data from the 2006 Korea Securities Research
Institute database (KSRI 2006).
The sample period spans from 1995 to 2006.10
9 Factor analysis is initially run on three measures of stock
liquidity and the factor obtained is used to test the hypotheses.
Factor analysis is a data reduction technique that simplifies
complex and diverse relationships that exist among a set of
variables by generating new variables, the factors of which extract
the main sources of variation among the original variables (Dillon
and Goldstein, 1984). Specifically, we combine three liquidity
proxies (i.e., trading costs based on Lesmond et al. (1999), zero
returns, and effective spreads based on Roll, 1984) by factor
analysis with one oblique rotation. The liquidity factor is the
factor score from the first (and only) factor with an eigenvalue
greater than one. 10 We start from 1995 because this is the first
year for which stock return data are available in the KSRI
database.
-
17
We begin the sample collection procedure with all observations
for which we have the
necessary data to compute the variables used in the firm-year
regressions described above.
We exclude financial companies from our sample because the
financial reporting
environment for financial institutions differs from that for
companies in other industries. The
final sample includes 9,241 firm-year observations after we
delete observations with missing
financial data. For our auditor designation sample, we obtain a
list of 583 firm-year
observations for 291 distinct firms whose auditors were
designated in the 19952002 period
from the Korea Financial Supervisory Service. 11 For the
analysis testing our second
hypothesis, the data consist of 5,093 firm-year observations
including 583 auditor-designated
observations.
4. Empirical results
4.1. Sample description
Table 1 summarizes our variable definitions and Table 2
describes the summary
statistics. Panel A of Table 2 provides a breakdown of the
number of observations and shows
the composition of the samples used for our analyses. The full
sample consists of 7,445 non-
switching firm-year observations and 1,796 switching firm-year
observations. The second
dataset, the auditor designation sample, spans the period from
1995 to 2002 and consists of
583 firm-year observations in which mandatory auditor switches
take place. The number of 11 We thank Professor Cheong H. Yi for
providing the list of firms subject to auditor designation during
the 1995-2002 period. We dont extend the sample period because SFC
practically stopped designating auditors for listed companies.
Since 2003, only a small number of firms subject to delisting have
been mandated to change auditors.
-
18
observations in our non-designated sample during the same period
is 4,510, yielding 5,093
firm-year observations in total.12
Panel B of Table 2 shows that the mean of SWITCH is 19.4% for
the final sample,
indicating that about one-fifth of our firm-year observations
are switchers, a higher
proportion than those observed in U.S. samples. Among these
switching firms, about half
(8.5%) change from one Big 4 auditor to another Big 4 auditor,
suggesting that caution is
required when drawing inferences based solely on SWITCH as a
proxy for audit quality. The
means of LOT and ZERO are -4.835 and 0.089, respectively. The
8.9% of trading days with
zero returns equates to 23 days in each year without price
changes. The ROLL measure is
2.3% on average, which is a fairly low effective spread cost. As
previously discussed, the
Korean stock market has better liquidity than other emerging
markets, a feature we expect to
provide us with reasonable test power in the following analyses.
We winsorize all the
variables other than those with natural lower and upper bounds
at the 1st and 99th percentiles.
The correlations reported in Panel C of Table 2 show that all
three liquidity variables are
positively associated with the auditor switch indicator
(SWITCH), and similarly, return
variability (RETVAR) is positively correlated with the switch
indicator. The three liquidity
proxies are also significantly and positively related with one
another, suggesting the effective
internal validity of these measures. Firm size (SIZE), financial
leverage (LEV), asset turnover
12 The availability of audit fee data used to construct
fee-based industry specialization variables restricts our final
dataset to 4,850 observations since audit fee data is publicly
available since 2001.
-
19
(ATURN), the return-on-assets ratio (ROA), sales growth (SGROW),
the quick ratio (QUICK),
and return variability (RETVAR) are significantly related to
liquidity measures, thereby
confirming that we need to control for these variables in
examining the incremental liquidity
effect of an auditor switch.
4.2. Auditor switch and stock liquidity
4.2.1. Multivariate analyses
We now test the association between auditor switches and stock
liquidity for the full
sample. We estimate Eq. (1) using our full sample of 9,241
firm-year observations covering
1,677 unique firms. In Table 3, we report the regression
coefficients and their t-values
computed using standard errors adjusted for clustering at the
firm level for our pooled OLS
regressions to control for serial correlation problems.
The results reported in Panel A of Table 3 show that the
coefficient for SWITCH is
significantly positive in all cases in which a stock liquidity
measure is used as the dependent
variable. These findings are consistent with the view that
auditor switches have negative
economic consequences and suggest that auditor switching firms
are less liquid than their
non-switching counterparts.
We then run regressions using separate indicators of the auditor
switch types BN, NB,
BB, and NN in Panel B of Table 3. We find that the negative
stock liquidity effect of an
auditor switch is pronounced only in BN firms. In our
specification, the percentage of days
-
20
without trades increases by 1% for BN firms, which represents a
deterioration in liquidity of
about 3% (= 0.009/0.274) relative to the liquidity of firms that
do not switch auditors. The
coefficients of both LOT and ROLL are also significantly
positive (t = 3.94 and 2.18,
respectively). When we aggregate all three liquidity proxies
into a single liquidity factor for
parsimony, we also find a significantly lower liquidity only for
BN firms (coefficient = 0.196,
t = 3.69).
Turning to the control variables, we observe several significant
relations. Using the
liquidity factor as a dependent variable, we find that
consistent with the prior market
microstructure literature (e.g., Chordia et al., 2000; Leuz and
Verrecchia, 2000), larger firms
and firms with higher profitability, higher sales growth, a
higher quick ratio, and higher
return variability exhibit higher stock liquidity.
4.2.1. Accounting for endogenous auditor switching decisions
The reverse causality problem may not be a big concern in Table
3 because it is unlikely
that firms facing the prospect of lower stock liquidity will
tend to replace high-quality
auditors with low-quality auditors. Rather, as Fan and Wong
(2005) argue, firms with a
higher level of information risk will hire high-quality auditors
to mitigate investors concerns
about information risk. In this regard, the potential reverse
causality issue, if any, is unlikely
to explain our results because it works against finding
significant results in our analyses.
-
21
On the other hand, because auditor switching decisions are
unlikely to be exogenous,
this may raise some concerns over our previous findings on the
liquidity effect of an auditor
switch. To address the concern, we use matched sample techniques
that have been developed
in the literature (e.g., Clatworthy et al., 2009). Specifically,
we estimate the probit model of
the probability of switching auditor using a set of firm
characteristics as explanatory variables,
as described below:
SWITCHit = 0 + 1 Xit-1 + Year + Industry + it, (2)
where SWITCH is an indicator of whether firm i switches auditor
in year t, X is the vector of
firm characteristics including KSE (an indicator of whether a
firm is listed on the Korean
Stock Exchange), Size (the natural logarithm of total assets),
ROA (the return-on-assets ratio),
LOSS (an indicator of a firm reporting negative net income), the
interaction of ROA and
LOSS, and NSWITCH (the frequency of auditor switches during the
sample period).13 We
take the predicted value from Equation (2) to construct the
propensity score of auditor
switches for each firm i in year t. Next, using this propensity
score, we match auditor-
switching firm-years with non-switching firm-years that have the
closest propensity score to
the matched switching firm for the same year.14 The resultant
sample thus consists solely of
firms that have a similar probability of switching observations.
By doing so, we control for
13 At the first stage, all the explanatory variables other than
SIZE are statistically significant at the 5% significance level.
The model also has reasonable fit as its log likelihood is
-3,676.49. The coefficients are estimated as:
SWITCH=-1.30-0.51*KSE-0.01*SIZE+1.44*ROA-2.37*ROA*LOSS+0.11*LOSS+0.47*NSWITCH.
14 The mean value of the differences in propensity scores between
the two samples is 0.003, with a standard deviation of 0.013,
indicating our propensity matching exercise is reasonably
successful.
-
22
firm characteristics that could possibly drive auditor switch
decisions. In such a sample, the
coefficient on our SWITCH indicator variable can be reliably
interpreted as a causal effect of
the auditor switch on stock liquidity.15
Before analyzing the matched sample, we first use observations
for firms that switch
auditors only once during our sample periods shown in columns
(1) and (2) of Table 4. This
sample consists of 2,819 observations for 542 firms. We choose
one-time switchers only
because we expect they exhibit a clearer difference in stock
liquidity after changing auditors
than firms that switch auditors repetitively. By doing so, we
can compare stock liquidity of
pre-switching periods with that of post-periods for such firms.
In column (1), the coefficient
of SWITCH is positive but insignificant. But in column (2), the
coefficient of BN remains
significantly positive, suggesting that firms experience lower
liquidity after they switch high
quality auditors to low quality auditors.16 This reconfirms our
previous results that stock
liquidity is related to the quality of auditors.
15 Propensity score matching has several advantages over
instrumental variable (IV) methods as a solution of the endogeneity
problem (Core, 2010). First, it does not require the identification
of reliable IVs which is very challenging in many empirical
settings. For example, Clatworthy et al. (2009) and Francis,
Lennox, and Wang (2010) find that the inferences from Heckmans
two-stage estimation are sensitive to the choice of IVs.
Furthermore, Larcker and Rusticus (2010) show that if IVs are not
appropriate, ordinary least squares estimates can be better than
two-stage or three-stage least squares (2SLS or 3SLS) estimates.
Second, it is robust to misspecification of the functional form
because it is not subject to parametric assumptions of IV methods
(Armstrong et al., 2010). However, propensity score matching does
not address endogeneity directly as do IV methods. We thus discuss
the robustness of our results using Heckmans two-stage estimation
in footnote 20. 16 Examining changes in liquidity around the event
of auditor switches is known as an effective way in controlling for
potential biases from omitted variables. However, the event date
cannot be identified in our setting. Even though auditor switches
can occur at any time during the fiscal year, in Korea, firms
switching auditors have not been required to announce the news at
the event before 2002. As a result, it is practically impossible to
figure out when the auditor switch news is released to the public.
Therefore, it may not be appropriate to investigate the impact of
auditor switches on the change in liquidity variables over time. It
is noted that matching samples can be an alternative in such a case
(Mayhew and Mihov, 2004).
-
23
In columns (3) and (4), we match all 1,796 switching year
observations to the same
number of non-switching year observations, yielding 3,592
firm-year observations in total.
The matched non-switching observations should have the closest
propensity score in the same
year of the switching observations. Specifically, a switching
firm-year observation is matched
with a non-switching firm-year observation having the most
similar likelihood of switching
an auditor in the same year, which means the benchmark group in
columns (3) and (4)
consists of non-switching observations in the same year of
switching observations. For
completeness, in columns (5) and (6), we choose the switching
year observations of 542 one-
time switchers that are used in columns (1) and (2) and match
them to the same number of
observations for firms that have never changed auditors,
yielding a total of 1,084
observations. The matching also requires the closest propensity
score in the same year.
The results in columns (3) to (6) corroborate our previous
findings that auditor switches
have negative implications for stock liquidity. The findings
suggest that, even after we
control for firm characteristics that might affect auditor
switch decisions, the negative
liquidity effect of auditor switch holds robust. Especially, the
within-firm approach in
columns (1) and (2) confirms that the results in Table 3 are not
due to the potential bias of
auditor switching decisions, but are instead likely to reflect a
causal relationship.17
17 Despite several advantages of the propensity matching
technique, its major limitation is that matching is based only on
observable variables and therefore it cannot control for
selectivity based on unobservables. We thus conduct Heckmans (1979)
2SLS test for robustness check and find similar results. With the
inclusion of the inverse Mills ratio estimated from the first stage
probit model in the second stage OLS, the coefficient of SWITCH is
positive but becomes insignificant (coeff. = 0.102, t = 0.98),
while the coefficient of BN remains significantly positive (coeff.
= 0.238, t = 2.10). However, we are also cautious in interpreting
this result because of the difficulty of identifying an exogenous
instrument variable in the first stage.
-
24
4.3. The liquidity effects of a mandatory auditor switch:
evidence from an auditor designation
sample
4.3.1. The effect of a mandatory auditor switch on stock
liquidity
In this section, we compare the liquidity effects of mandatory
auditor changes (i.e.,
auditor designation) with those of voluntary auditor changes to
investigate further whether
the effect of mandatory auditor change differs systematically
from that of voluntary auditor
change. In so doing, we construct a subsample in which auditors
are switched on either a
mandatory or a voluntary basis from 1995 to 2002. We use 5,093
firm-year observations
consisting of 583 firm-year observations for designated auditor
switches and 4,510 firm-year
observations for non-designated auditor switches.
Table 5 reports the subsample results for the liquidity effects
of auditor designation. For
simplicity, we report only the results based on the liquidity
factor (FACTOR), although the
results obtained using the three different individual measures
are qualitatively similar to those
reported for the composite factor. Column (1) shows the result
of Eq. (1) using the subsample
and reports a positive coefficient of SWITCH. We include the
auditor designation dummy
(DSG) and define an interaction term, DSG*SWITCH, that captures
any incremental liquidity
effect for firms subject to the auditor designation rule. As
shown in column (3), the
Note that it is essential to identify exogenous independent
variables from the first stage choice model that can be validly
excluded from the second stage regression to successfully implement
the Heckman (1979) procedure (Francis et al., 2010). We thus
consider NSWITCH as an exogenous variable to capture firms tendency
to frequently switch auditors. We also expect NSWITCH not to be
related to the error terms in Equation (1) as it is counted over
the sample period and there is no reason to expect it has an impact
on stock liquidity in a particular year.
-
25
coefficient of DSG*SWITCH is significantly negative, which
indicates that the negative
liquidity effect of auditor switching is less pronounced if the
switch is driven by the auditor
designation rule. As the rule aims to improve auditor
independence and thus audit quality,
switching to an incoming designated auditor is perceived to have
positive economic
consequences in the capital market. The F-test result confirms
that the negative impact of an
auditor switch is no longer observed when the switch is
mandatory. As the auditor is
designated when a firm is suspected of fraudulent accounting,
this finding is consistent with
the view that designated auditors improve audit quality (Kim and
Yi, 2009).
Columns (3) and (4) show that the negative liquidity effects of
auditor-switching firms
are concentrated in firms that switch from a Big 4 auditor to a
non-Big 4 auditor and that such
effects are less pronounced when the switch is initiated by the
rule. The positive liquidity
effects of auditor designation, which we report in Columns (2)
and (4), are mostly driven by
firms switching from a non-Big 4 auditor to a Big 4 auditor.
4.3.2. The effect of audit risk on the relation between a
mandated auditor switch and stock
liquidity
If the non-negative liquidity impact of a mandated auditor
switch is attributed to the
fact that investors are well aware of the regulatory
intervention objective of improving audit
quality, this impact is likely to be greater when the firm is
subject to higher audit risk. We
-
26
thus further examine whether the liquidity effects of designated
auditor switches vary with
the level of audit risk. As shown in Panel A of Table 6, Article
10 of the Financial
Supervisory Service (FSC) Regulation under the Act on External
Audit stipulates that listed
firms will be subject to an auditor designation requirement if
they meet one of the 10
specified conditions. Following Kim and Yi (2009), we consider
that the first four conditions
describe instances of lower client risk, while the last six
describe circumstances of higher
client risk. Among the 583 auditor designation sample
observations, 184 (399) belong to the
low-risk (high-risk) group.18 Panel B shows that the liquidity
effect is pronounced only in
NB firms subject to higher risk. Difference tests between the
two risk groups show that
liquidity effects can be observed in firms switching to Big 4
auditors, regardless of the type
of auditor formerly engaged.
The results imply that the positive liquidity effect of a
mandatory auditor switch is
related to the potential impact of auditor designation on the
audit risk of client firms. The
finding that there is a positive liquidity effect of auditor
switches in firms with higher audit
risk suggests that investors expect designated auditors to play
a more significant role in such
firms. Accordingly, the results reported in Table 6 corroborate
our previous findings about
the implications of mandatory auditor switches.
4.4. Additional tests 18 We follow Kim and Yi (2009) in
classifying risk. Of the ten criteria for auditor designation, four
are regarded as indicating low client risk and the other six are
regarded as indicating high client risk. See Table 6 for
details.
-
27
4.4.1. An alternative proxy for audit quality: industry
specialization
Prior studies have employed auditor brand name or auditor size
(i.e., Big 4) as a proxy
for audit quality. However, audit firm brand name is only one
possible indication of audit
quality (e.g., Knechel et al., 2007). To mitigate the concern
over using an empirical proxy for
audit quality, recent studies have hypothesized that auditor
industry specialization also
contributes to audit quality. Knechel et al. (2007) find that
firms switching between Big 4
auditors experience significant positive abnormal returns when
the successor auditor is an
industry specialist and experience significant negative abnormal
returns when the successor
auditor is not a specialist. Balsam et al. (2003) report that
client firms of specialist auditors
have lower discretionary accruals and higher earnings response
coefficients than client firms
of non-specialists. On the whole, the industry specialist
indicator is perceived as an effective
proxy for audit quality. We thus corroborate our analyses by
focusing on industry
specialization. Given that the market perceives audit quality
differences based on industry
specialization (Knechel et al., 2007), this test will confirm
that differences in the credibility
of financial statements are relevant to stock liquidity.
We identify auditor expertise by considering each auditors
industry market share in a
specific year. To identify industry specialists, we manually
collect audit fee information
disclosed by Korean listed firms. We code an indicator variable
(Specialist) equal to one if
the firm is audited by an industry specialist that reports the
greatest market share in terms of
-
28
the audit fee in each two digit industry. We also categorize
switchers as SN, NS, SS, or NN
based on the industry specialization of their successor
auditors. In so doing, we examine
whether a switch from an industry specialist auditor to a
non-specialist auditor has a
significant impact on stock liquidity. If the auditor switch has
a discernible effect on stock
liquidity, we expect the firm switching from a specialist (a
non-specialist) auditor to a non-
specialist (specialist) auditor to exhibit a lower (higher)
level of stock liquidity than a firm
that does not switch its auditor. This test is carried out using
4,850 firm-year observations
during the 20012005 period, a timeframe for which audit fee data
are publicly available.
Table 7 confirms that the negative liquidity effect of an
auditor switch remains constant
in the subset of data for the limited sample period from 2001 to
2005. Similar to the previous
result showing a concentrated liquidity effect among BN firms,
the liquidity effect of auditor
switches is most pronounced among SN firms (i.e., firms
switching from an industry
specialist auditor to a non-specialist auditor). NN firms also
exhibit lower stock liquidity,
suggesting that market participants penalize firms that switch
to non-specialist auditors
regardless of the type of auditor formerly engaged. Given that
the market perceives audit
quality differences based on industry specialization (Knechel et
al., 2007), these results
suggest that employing industry specialists to improve the
credibility of financial statements
-
29
is relevant to stock liquidity. In this regard, our findings can
be regarded as robust regardless
of the measure of audit quality employed.19
4.4.2 Other measures of stock liquidity
Following prior studies, we also use the price impact measure of
Amihud (2002) and the
bid-ask spread as additional measures of liquidity. However,
trading volume and transaction
data, both of which are essential to construct these two
measures, are available only for KSE
(Korean Stock Exchange) firms and not for KOSDAQ (Korean
Securities Dealers Automated
Quotation) firms. For example, the bid-ask spread can be
calculated for only 5,089 firm-year
observations from the KSE, which represents about 55% of our
total sample. To avoid such a
substantial loss of data, we have reported our main results
without relying on these two
measures. However, when we repeat our main analyses based on
these two measures for
completeness, the tenor of our main results does not change.
These results, which are
reported in Table A1, give us more confidence in our
conclusions.
5. Concluding remarks
19 We also use alternative measures of industry specialization
such as auditors industry share based on auditees sales or assets
(Balsam et al., 2003; Knechel et al., 2007). Although we find
directionally consistent results, the coefficient on SN is not
statistically significant (t=0.45 when we use asset-based industry
leadership and t=0.21 when we use sales-based industry leadership).
Rather, the coefficient of NN becomes positive and marginally
significant. (t=1.86 when we use asset-based industry leadership
and t=1.88 when we use sales-based industry leadership). It is
possible that an auditor can increase the number of clients (thus
the summation of sales or assets of the clients) by proposing a
discounted price (Ettredge and Greenberg, 1990; Pearson and
Trompeter, 1994). Alternatively, another auditor can charge higher
audit fees and at the same time increase the market share in an
industry. The latter case is possible only when the auditor
provides higher-quality audit service enough to justify the higher
price than the other auditors charge. In contrast, it is hard to
say that the auditors in the former case are the industry expertise
auditor. Thus, with caution, we argue that the industry specialist
measure based on audit fee is a better proxy for the auditor
industry specialist. Consistent with this argument, Craswell et al.
(1995) document the existence of significant industry specialist
auditor fee premium.
-
30
In this paper, we examine the effect of auditor switches on
stock liquidity. We find that
other things being equal, firms that switch auditors are more
likely to exhibit lower stock
liquidity than firms that do not switch auditors, thereby
supporting the agency risk
explanation of auditor switches. We also find that the negative
liquidity effect is driven by
firms switching from a high-quality auditor to a low-quality
auditor.
We further examine whether auditor switches mandated by a
regulatory body under the
auditor designation rule in Korea are also associated with stock
liquidity. Consistent with the
notion that auditor designation improves audit quality, we find
a non-negative association
between stock liquidity and a designated auditor switch.
Furthermore, we document an even
positive association between them when firms with newly
designated auditors have higher
audit risk. Overall, our results are consistent with the notion
that the auditor designation rule
in Korea enhances audit quality and thus the credibility of
financial reporting.
Our study is subject to several limitations. First, as many
previous studies point out,
auditor switches often coincide with other substantial changes
in the firm. It is possible that
our results reflect the joint effects of these changes and hence
cannot be attributed solely to
the auditor switch. Second, if the stock market is efficient
enough to anticipate the
announcement of the auditor switch, this is likely to reduce the
power of our tests. Finally,
the results are based solely on data on Korean firms and
therefore cannot be easily
generalized to other countries.
-
31
Despite these limitations, the study makes meaningful
contributions to the literature. In
particular, we provide evidence regarding the economic
consequences of an auditor switch in
terms of the significant capital market costs incurred around
the time of the announcement. In
contrast, the results of prior studies based on stock returns
have been inconclusive and mixed.
Given that regulators interest in auditor switches is due to the
concern that auditor changes
are motivated by management opportunism, our findings are likely
to be of interest to
regulators and policymakers. Our study is also novel in that it
examines the liquidity effect of
mandatory auditor switches required by regulation. It implies
that investors are concerned not
only with changes in audit quality, but also with who initiates
the change. Overall, our paper
contributes to the ongoing debate on the economic costs and
benefits of voluntary versus
mandatory auditor changes and the literature on stock
liquidity.
-
32
Appendix Table A1. Regression results based on other liquidity
proxies This table reports the regression results for the relation
between auditor switch and stock liquidity using PRCIMPT and
SPREAD. Panel A (B) regresses the stock liquidity measures on the
auditor switch indicator (each auditor switch type). FACTOR5
represents the scores for single factors extracted from all five
liquidity variables (LOT, ZERO, ROLL, PRCIMPT, and SPREAD)
employing factor analysis. We include industry- and year-fixed
effects based on the two-digit SIC code, but we do not report the
coefficients. The table reports the OLS coefficient estimates and
t-statistics based on robust standard errors that are clustered by
firm. See Table 1 for variable definitions. (Panel A) Results based
on auditor switch indicator
Dependent var. = (1)
PRCIMPT (2)
SPREAD (3)
FACTOR5
Parameter Coef. t-stat. Coef. Coef. Coef. t-stat.Intercept
12.186 9.95 -3.064 -12.39 6.190 10.93
SWITCH 0.117 2.21 0.012 0.70 0.051 1.36SIZE -0.527 -12.55 -0.101
-11.98 -0.238 -12.10LEV -0.141 -0.46 0.139 2.07 0.025 0.17
ATURN -0.179 -1.73 -0.049 -2.07 -0.008 -0.16ROA 1.271 2.81
-0.473 -4.72 -0.399 -1.96
SGROW -0.199 -2.13 -0.013 -0.51 -0.040 -0.74QUICK -0.049 -1.22
0.011 0.95 0.015 0.62
RETVAR -44.528 -11.38 3.106 2.61 -8.767 -3.67Year Yes Yes
Yes
Industry Yes Yes Yes Adj.R2 0.334 0.167 0.296 #Obs. 4,852 5,089
4,147
(Panel B) Results based on auditor switch type
Dependent var. = (1)
PRCIMPT (2)
SPREAD (3)
FACTOR5
Parameter Coef. t-stat. Coef. t-stat. Coef. t-stat.Intercept
12.309 10.08 -3.064 -12.28 6.197 10.92
BN 0.230 2.14 0.082 2.23 0.182 2.42NB 0.062 0.60 -0.013 -0.39
0.010 0.13BB 0.199 2.71 0.005 0.20 0.035 0.65NN -0.224 -1.51 -0.030
-0.73 -0.021 -0.21
SIZE -0.531 -12.66 -0.101 -11.84 -0.238 -12.08LEV -0.152 -0.49
0.136 2.03 0.021 0.14
ATURN -0.179 -1.72 -0.050 -2.09 -0.010 -0.19ROA 1.286 2.85
-0.469 -4.67 -0.378 -1.87
SGROW -0.199 -2.13 -0.013 -0.50 -0.040 -0.75QUICK -0.051 -1.28
0.011 0.93 0.014 0.56
RETVAR -44.492 -11.34 3.112 2.61 -8.773 -3.67Year Yes Yes
Yes
Industry Yes Yes Yes Adj.R2 0.335 0.167 0.296 #Obs. 4,852 5,089
4,147
-
33
References Autore, D.M., Billingsley, R.S., Schneller, M.I.,
2009. Information uncertainty and
auditor reputation. Journal of Banking and Finance 33 (12),
183-192. Armstrong, C.S., Jagolinzer, A.D., Larcker, D.F., 2010.
Chief executive officer equity
incentives and accounting irregularities. Journal of Accounting
Research 48: 225-271.
Bacidore, J.M., Battalio, R., Galpin, N., Jenning, R., 2005.
Sources of liquidity for NYSE-listed non-US stocks. Journal of
Banking and Finance 29: 3075-3098.
Balsam, S., Krishnan, J., Yang, J., 2003. Auditor industry
specialization and the earnings response coefficient. Auditing: A
Journal of Practice & Theory 22: 71-97.
Balvers, R.J., McDonald, B., Miller, R.E., 1988. Underpricing of
new issues and the choice of auditor as a signal of investment
banker reputation. The Accounting Review 63: 605-622.
Bekaert, G., Harvey, R.H., Lundblad, C., 2007. Liquidity and
Expected Returns: Lessons from Emerging Markets. Review of
Financial Studies 20: 1783-1831.
Beneish, M.D., Hopkins, P.E., Jensen, I.P., Martin, R.D., 2005.
Do auditor resignations reduce uncertainty about the quality of
firms financial reporting? Journal of Accounting & Public
Policy 24: 357-390.
Boone, J.P., Raman, K.K., 2001. Auditor resignations versus
dismissals: an examination of the differential effects on market
liquidity and trading activity. Advances in Accounting 18:
47-75.
Choi, J-H., Wong, T.J., 2007. Auditors governance functions and
legal environments: an international investigation. Contemporary
Accounting Research 24: 13-46.
Chordia, T., Roll, R., Subrahmanyam, A., 2000. Co-movements in
bid-ask spreads and market depth. Financial Analysts Journal 56:
23-27.
Clatworthy, M., Makepeace, G., Peel, M., 2009. Selection bias
and the Big Four premium: new evidence using Heckman and matching
models. Accounting and Business Research 39: 139-166.
Core, J., 2010. Discussion of chief executive officer equity
incentives and accounting irregularities. Journal of Accounting
Research 48: 271-287.
Craswell, A., Francis, J., Taylor, S., 1995. Auditor Brand Name
Reputations and Industry Specializations. Journal of Accounting and
Economics 20: 297-322.
DeAngelo, L., 1981. Auditor size and audit quality. Journal of
Accounting and Economics 3: 183-199.
Dillon, W.R., Goldstein, M., 1984. Multivariate analysis:
methods and applications. John Wiley & Sons: New York, US.
Doidge, C., Karolyi, G., Stulz, R., 2009. Has New York become
less competitive in global markets? Evaluating foreign listing
choices over time. Journal of Financial Economics 91: 253-277.
Easley, D., OHara, M., 2004. Information and the cost of
capital. The Journal of Finance 59: 1553-1583.
Eichenseher, J.W., Hagigi, M., Shields, D., 1989. Market
reaction to auditor changes by OTC companies. Auditing: A Journal
of Practice & Theory 9: 29-40.
-
34
Ettredge, M., Greenberg, R., 1990. Determinants of Fee Cutting
on Initial Audit Engagements. Journal of Accounting Research 28:
198-210.
Fan, J.P.H., Wong, T.J., 2005. Do external auditors perform a
corporate governance role in emerging markets? Evidence from East
Asia. Journal of Accounting Research 43: 1-38.
Fortin, S., Pittman, J.A., 2007. The role of auditor choice in
debt pricing in private firms. Contemporary Accounting Research 24:
859-896.
Francis, J.R., Khurana, I.K., Pereira, R., 2003. The role of
accounting and auditing in corporate governance and the development
of financial markets around the world. Asia-Pacific Journal of
Accounting and Economics 10: 1-30.
Francis, J.R., Maydew, E.L., Sparks, H.C., 1999. The role of big
6 auditors in the credible reporting of accruals. Auditing: A
Journal of Practice & Theory 18: 17-34.
Francis, J., Lafond, R., Olsson, P., Schipper, K., 2005. The
market pricing of accruals quality. Journal of Accounting and
Economics 39: 295-327.
Francis, J.R., Lennox, C.S., Wang, Z., 2010. Selection models in
accounting research. Unpublished Paper.
Fried, D., Schiff, A., 1981. CPA switches and associated market
reaction. The Accounting Review 56: 326-341.
Frino, A., Gerace, D., Lepone, A., 2008. Liquidity in auction
and specialist market structures: Evidence from the Italian bourse.
Journal of Banking and Finance 32: 2581-2588.
Goyenko, R.Y., Holden, C.W., Trzcinka, C.A., 2009. Do liquidity
measures measure liquidity? Journal of Financial Economics 92:
153-181.
Harris, L., 1989. A day-end transaction price anomaly. Journal
of Financial and Quantitative Analysis 24: 29-45.
Harris, L., 1990. Statistical properties of the Roll serial
covariance bid-ask spread estimator. Journal of Finance 45:
568-579.
Heckman, J., 1979. Sample selection bias as a specification
error. Econometrica 47: 153-161.
Jensen, M.C., Meckling, W.H., 1976. Theory of the firm:
managerial behavior, agency costs, and ownership structure. Journal
of Financial Economics 3: 305-360.
Jeong, S.W., Jung, K., Lee, S-J., 2005. The effect of mandatory
auditor assignment and non-audit service on audit fees: evidence
from Korea. The International Journal of Accounting 40:
233-248.
Johnson, W.B., Lys, T., 1990. The market for audit services.
Journal of Accounting and Economics 12: 281-308.
Kanagaretnam , K., Lim, C. Y., Lobo, G.J., 2010. Auditor
reputation and earnings management: International evidence from the
banking industry. Journal of Banking and Finance 34: 2318-2327.
Kim, J-B., Yi. C., 2009. Does auditor designation by the
regulatory authority improve audit quality? Evidence from Korea.
Journal of Accounting & Public Policy 28: 207-230.
-
35
Knechel, W.R., Naiker, V., Pacheko, G., 2007. Does auditor
industry specialization matter? Evidence from market reaction to
auditor switches. Auditing: A Journal of Practice & Theory 26:
19-45.
Krishnan, J., Stephens, R.G., 1995. Evidence on opinion shopping
from audit opinion conservatism. Journal of Accounting & Public
Policy 14: 179-201.
Kryzanowski, L., Lazrak, S., Rakita, I., 2010. Behavior of
liquidity and returns around Canadian seasoned equity offerings.
Journal of Banking and Finance 34: 2954-2967.
Lambert, R., Leuz, C., Verrecchia, R.E., 2007. Accounting
information, disclosure, and the cost of capital. Journal of
Accounting Research 45: 385-420.
Lang, M.H., Lins, K.V., Maffett, M.G., 2011. Transparency,
liquidity, and valuation: International evidence on when
transparency matters most. Journal of Accounting Research,
forthcoming.
Larcker, D.F., Rusticus, T.O., 2010. On the use of instrumental
variables in accounting research. Journal of Accounting and
Economics 49: 186-205.
Lennox, C.S., 2000. Do companies successfully engage in
opinion-shopping? Evidence from the U.K. Journal of Accounting
& Economics 29: 321-337.
Lesmond, D., 2005. Liquidity of emerging markets. Journal of
Financial Economics 77: 411-452.
Lesmond, D., Ogden, J., Trzcinka, C., 1999. A new estimate of
transaction costs. Review of Financial Studies 12: 1113-1141.
Lesmond, D., Schill, M., Zhou, C., 2004. The illusory nature of
momentum profits. Journal of Financial Economics 43: 4421-1142.
Leuz, C., Verrecchia, R.E., 2000. The economic consequences of
increased disclosure. Journal of Accounting Research 38:
91-124.
Levich, R., 2001. The importance of emerging capital markets.
Brookings-Wharton Papers on Financial Services: 145.
Mayhew, S., Mihov, V., 2004. How do exchanges select stocks for
option listing? Journal of Finance 58, 447-471.
Nichols, D.R., Smith, D.B., 1983. Auditor credibility and
auditor changes. Journal of Accounting Research 21: 534-544.
Pearson, T., Trompeter, G., 1994. Competition in the Market for
Audit Services: The Effect of Supplier Concentration on Audit Fees.
Contemporary Accounting Research 11: 115-135.
Pittman, J.A., Fortin, S., 2004. Auditor choice and the cost of
debt capital for newly public firms. Journal of Accounting and
Economics 37: 113-136.
Roll, R., 1984. A simple implicit measure of the effective
bid-ask spread in an efficient market. Journal of Finance 39:
1127-1139.
Sankaraguruswamy, S., Whisenant, J.S., 2004. An empirical
analysis of voluntarily supplied client-auditor realignment
reasons. Auditing: A Journal of Practice & Theory 23:
107-121.
Securities and Exchange Commission (SEC), 1988. Disclosure
Amendments to Regulation S-K, Form 8-K and Schedule 14a Regarding
Changes in Accountants and Potential Opinion Shopping Situations.
Financial Reporting Release NO. 31. Washington, D.C.: Government
Printing Office.
-
36
Sengupta, P., 1998. Corporate disclosure and the cost of debt.
The Accounting Review 73: 459-474.
Shu, S.Z., 2000. Auditor resignations: clientele effects and
legal liability. Journal of Accounting & Economics 29:
173-205.
Smith, D.B., 1988. An investigation of Securities and Exchange
Commission regulation of auditor change disclosures: the case of
accounting series release no. 165. Journal of Accounting Research
26: 134-145.
Sohn, S., Kim, Y.H., 2004. Information Content Study of the
Auditor Switch. Korean Management Review 33: 685-702 (in
Korean).
Watts, R., Zimmerman, J.L., 1986. Positive accounting theory.
Englewood Cliffs, N.J.: Prentice-Hall, Inc.
Wells, D.W., Loudder, M.L., 1997. The market effects of auditor
resignations. Auditing: A Journal of Practice & Theory 16:
138-144.
Whisenant, J.S., 2006. An analysis of the information content of
auditor change announcements. The Grant Thornton Commentary Series
Investor Perceptions of Audit Firm Parity 1 (1).
Whisenant, J.S., Sankaraguruswamy, S., Raghunandan, K., 2003.
Market reactions to disclosure of reportable events. Auditing: A
Journal of Practice & Theory 22: 181-194.
Wild, J.J, Bernstein, L.A., Subramanyam, K.R., 2001. Financial
Statement Analysis, McGraw Hill Book Company.
-
37
Table 1 Variable definitions
Variable Description Dependent variables LOT The natural
logarithm of a yearly estimate of the transaction costs
implied by the trading behavior of investors developed by
Lesmond et al. (1999). The estimation period spans month -5 through
month +7 so that the firms annual reports are publicly available
for a few months, which should enable markets to impound the
auditor switch effect.
ZERO The proportion of zero daily returns out of the maximum
potential number of trading days in a given year. The measurement
period spans month -5 through month +7 relative to the firms fiscal
year-end.
ROLL Rolls (1984) effective spread based on the bid-ask
bounce-induced negative serial correlation in returns. The
measurement period spans month -5 through month +7 relative to the
firms fiscal year-end.
PRCIMPT Amihuds (2002) illiquidity measure which captures the
price impact of trades. We measure illiquidity as the median daily
price impact over the year and compute the price impact as the
daily absolute price change in percentage terms divided by trading
volume (measured in million KRW). To avoid the misclassification of
days with no or low trading activity, we omit zero-return days from
yearly median calculations. We take the natural logarithm of the
value for expositional purposes. The measurement period runs from
month -5 to month +7 relative to the firms fiscal year-end.
SPREAD The bid-ask spread. We obtain the closing bid and ask
prices for each day from the intraday transaction data of the
Korean Stock Exchange and compute the daily quoted spread as the
difference between the two prices divided by the midpoint. We then
compute the median daily spread over the year. The measurement
period runs from month -5 to month +7 relative to the firms fiscal
year-end.
Auditor switch variables SWITCH 1 if a firm changes its auditor
in the period; 0 otherwise BN 1 if a firm changes from a Big 4
auditor to a non-Big 4 auditor; 0
otherwise NB 1 if a firm changes from a non-Big 4 auditor to a
Big 4 auditor; 0
otherwise BB 1 if a firm changes from one Big 4 auditor to
another Big 4
auditor; 0 otherwiseNN 1 if a firm changes from one non-Big 4 to
another non-Big 4
auditor; 0 otherwise Industry specialization variablesSN 1 if a
firm changes its auditor from an industry specialist to a non-
specialist; 0 otherwiseNS 1 if a firm changes its auditor from a
non-specialist to a specialist;
0 otherwise SS 1 if a firm changes its auditor from one
specialist to another
specialist; 0 otherwiseNN 1 if a firm changes its auditor from
one non-specialist to another
-
38
non-specialist; 0 otherwise
Other variables SIZE The natural logarithm of total assetsLEV
The leverage ratio, calculated as the sum of short-term and
long-
term debt divided by total assetsATURN The asset turnover ratio,
calculated as sales divided by total assetsROA The return-on-assets
ratio, calculated as net income divided by
total assets SGROW Sales growth, calculated as the change in
sales divided by sales in
the previous yearQUICK The current ratio, calculated as current
assets divided by current
liabilities RETVAR Return variability, calculated as the
standard deviation of daily
returns in the year. The measurement period spans from month -5
to month +7 relative to the firms fiscal year-end.
DSG 1 if a firms auditor is designated by regulation; 0
otherwise FO Percentage of foreign ownership
-
39
Table 2 Descriptive statistics Panel A presents the number of
observations for each auditor switch type. Panel B presents the
summary statistics for the full sample. Panel C reports the Pearson
correlation matrix for the key variables. The full sample comprises
a maximum of 9,421 firm-year observations between 1995 and 2006
with non-missing financial data. See Table 1 for variable
definitions. (Panel A) Number of observations by auditor switch
type
Type Full sample Auditor designation sample (1995 to 2006) (1995
to 2002)
Designated Others Non-switch 7,445 286 3,839
Switch 1,796 297 671 BN 353 75 124 NB 347 57 149 BB 786 117 262
NN 310 48 136
Total 9,241 583 4,510 (Panel B) Full sample (N = 9,241, 1995 to
2006)
Variable Mean Std Dev Min. Q1 Median Q3 Max.SWITCH 0.194 0.396
0.000 0.000 0.000 0.000 1.000
BN 0.038 0.192 0.000 0.000 0.000 0.000 1.000 NB 0.038 0.190
0.000 0.000 0.000 0.000 1.000 BB 0.085 0.279 0.000 0.000 0.000
0.000 1.000NN 0.034 0.180 0.000 0.000 0.000 0.000 1.000 LOT -4.836
0.551 -6.123 -5.199 -4.852 -4.516 -3.175
ZERO 0.089 0.053 0.016 0.052 0.076 0.111 0.310 ROLL 0.023 0.014
0.000 0.013 0.020 0.030 0.113 SIZE 25.648 1.458 23.035 24.593
25.410 26.423 30.057 LEV 0.405 0.240 0.003 0.230 0.396 0.556
1.300
ATURN 0.951 0.514 0.140 0.620 0.855 1.162 3.249 ROA -0.015 0.168
-0.993 -0.013 0.019 0.055 0.216
SGROW 0.089 0.297 -0.655 -0.049 0.065 0.189 1.476 QUICK 1.950
1.914 0.235 0.930 1.361 2.116 12.022
RETVAR 0.039 0.014 0.015 0.028 0.038 0.049 0.074
-
40
(Panel C) Pearson correlation matrix
LOT ZERO ROLL SIZE LEV ATURN ROA SGROW QUICK RETVAR SWITCH 0.062
0.024 0.070 -0.007 0.047 -0.011 -0.011 0.005 -0.020 0.079
0.00 0.02 0.00 0.52 0.00 0.31 0.28 0.61 0.06 0.00 LOT 0.754
0.126 -0.194 0.145 -0.008 -0.146 -0.040 -0.064 0.114
0.00 0.00 0.00 0.00 0.42 0.00 0.00 0.00 0.00 ZERO 0.183 -0.020
0.076 0.022 0.022 0.016 -0.080 -0.372
0.00 0.05 0.00 0.04 0.03 0.12 0.00 0.00 ROLL -0.203 0.151 -0.075
-0.227 -0.039 0.000 0.650
0.00 0.00 0.00 0.00 0.00 0.99 0.00 SIZE 0.137 -0.063 0.208 0.063
-0.270 -0.294
0.00 0.00 0.00 0.00 0.00 0.00 LEV 0.064 -0.388 -0.023 -0.513
0.171
0.00 0.00 0.03 0.00 0.00 ATURN 0.105 0.150 -0.154 -0.078
0.00 0.00 0.00 0.00 ROA 0.227 0.070 -0.302
0.00 0.00 0.00 SGROW -0.054 -0.055
0.00 0.00 QUICK 0.029
0.01
-
41
Table 3 Regression results for the relation between auditor
switch and stock liquidity This table reports the regression
results for the relation between auditor switch and stock
liquidity. Panel A (B) regresses the stock liquidity measures on
the auditor switch indicator (each auditor switch type). FACTOR
summarizes our liquidity proxies and represents the scores of a
single factor extracted from liquidity variables (1) through (3)
employing factor analysis. We include industry- and year-fixed
effects based on the two-digit SIC code, but we do not report the
coefficients. The table reports the OLS coefficient estimates and
t-statistics based on robust standard errors that are clustered by
firm. See Table 1 for variable definitions. (Panel A) Results based
on auditor switch indicator
Dependent var. = (1)
LOT
(2)ZERO
(3)ROLL
(4) FACTOR
Parameter Coef. t-stat. Coef. t-stat. Coef. t-stat. Coef.
t-stat.Intercept -1.765 -9.43 0.275 15.44 0.056 17.53 4.792
13.62
SWITCH 0.041 2.83 0.003 2.23 0.001 2.35 0.067 2.61SIZE -0.127
-19.08 -0.007 -10.56 -0.002 -15.71 -0.186 -14.93LEV 0.133 2.46
-0.011 -1.95 0.007 8.64 0.004 0.04
ATURN -0.003 -0.21 0.003 1.63 -0.001 -4.69 0.026 0.83ROA -0.618
-9.69 -0.030 -5.15 -0.009 -5.68 -0.880 -7.57
SGROW -0.061 -3.18 -0.002 -1.01 -0.001 -2.42 -0.075 -2.13QUICK
-0.015 -3.31 -0.001 -2.63 0.000 1.26 -0.026 -3.10
RETVAR -3.661 -5.37 -1.037 -16.11 0.188 13.23 -14.551 -11.64Year
Yes Yes Yes Yes
Industry Yes Yes Yes Yes Adj.R2 0.161 0.305 0.334 0.242 #Obs.
9,241 9,241 9,241 9,241
(Panel B) Results based on auditor switch type
Dependent var. = (1)
LOT
(2)ZERO
(3)ROLL
(4) FACTOR
Parameter Coef. t-stat. Coef. t-stat. Coef. t-stat. Coef.
t-stat.Intercept -1.780 -9.48 0.274 15.40 0.056 17.38 4.771
13.54
BN 0.115 3.94 0.009 3.15 0.001 2.18 0.196 3.69NB 0.011 0.40
0.002 0.74 0.000 0.72 0.027 0.58BB 0.026 1.31 0.002 1.15 0.001 1.19
0.046 1.27NN 0.024 0.73 0.000 -0.14 0.001 1.25 0.017 0.28
SIZE -0.126 -18.95 -0.007 -10.52 -0.002 -15.55 -0.185 -14.85LEV
0.132 2.44 -0.011 -1.98 0.007 8.62 0.001 0.01
ATURN -0.003 -0.19 0.003 1.65 -0.001 -4.69 0.026 0.85ROA -0.617
-9.68 -0.030 -5.15 -0.009 -5.66 -0.880 -7.57
SGROW -0.060 -3.16 -0.002 -0.99 -0.001 -2.41 -0.074 -2.12QUICK
-0.015 -3.31 -0.001 -2.63 0.000 1.27 -0.026 -3.10
RETVAR -3.654 -5.37 -1.035 -16.10 0.188 13.20 -14.528 -11.63Year
Yes Yes Yes Yes
Industry Yes Yes Yes Yes adj.R2 0.218 0.305 0.334 0.242 #Obs.
9,241 9,241 9,241 9,241
-
42
Table 4 Regression results after controlling for the likelihood
of switching auditors This table controls for the likelihood of
switching auditors in the regression results for the relation
between auditor switch and stock liquidity. In columns (1) and (2),
we use firm-year observations for firms that switch auditors only
once during our sample period. In columns (3) to (6), we first
construct the propensity score as a predicted probability of
auditor switch. Next, using this propensity score, we match
auditor-switching observations with non-switching observations in
the same year in columns (3) and (4). In columns (5) and (6), we
match switching year observations for firms that switch auditors
only once during the sample period with the same year observations
for firms that have never changed auditors. We use FACTOR and the
scores for single factors extracted from the three liquidity
variables - LOT, ZERO, and ROLL - as dependent variables. We
include industry- and year-fixed effects based on the two-digit SIC
code, but we do not report the coefficients. The table reports the
OLS coefficient estimates and t-statistics based on robust standard
errors that are clustered by firm. See Table 1 for variable
definitions.
Dep. Var = FACTOR
Pre-Post analysis: One-time switchers only
Switching and non-switching pairs matched by the propensity
score
Switching and non-switching pairs using one-time switchers and
never-
switching firms (matched by the propensity score)
(1) (2) (3) (4) (5) (6)Parameter Coef. t-stat. Coef. t-stat.
Coef. t-stat. Coef. t-stat. Coef. t-stat. Coef. t-stat. Intercept
4.284 7.48 4.200 7.34 4.794 8.32 4.757 8.18 5.395 6.24 5.209
6.05
SWITCH 0.039 0.97 0.082 2.29 0.121 1.81 BN 0.275 2.95 0.220 3.68
0.370 3.35 NB -0.070 -0.88 0.028 0.54 -0.005 -0.06 BB -0.027 -0.48
0.071 1.55 0.059 0.71 NN 0.009 0.10 0.017 0.25 0.096 0.90
SIZE -0.169 -7.81 -0.166 -7.68 -0.188 -9.47 -0.187 -9.31 -0.212
-7.08 -0.205 -6.90 LEV 0.135 0.72 0.130 0.69 0.000 0.00 -0.008
-0.05 0.201 0.90 0.194 0.87
ATURN 0.002 0.04 0.006 0.12 -0.006 -0.12 -0.005 -0.11 0.002 0.03
0.007 0.12 ROA -0.829 -3.97 -0.839 -4.06 -0.662 -3.67 -0.664 -3.69
-0.686 -2.53 -0.701 -2.59
SGROW -0.080 -1.64 -0.079 -1.63 -0.096 -1.30 -0.093 -1.27 -0.157
-1.89 -0.156 -1.91 QUICK -0.039 -3.05 -0.037 -2.97 0.001 0.09 0.002
0.11 -0.024 -1.58 -0.021 -1.39
RETVAR -14.253 -7.71 -14.155 -7.71 -14.933 -6.67 -14.870 -6.65
-20.308 -6.35 -20.067 -6.27 Year Yes Yes Yes Yes Yes Yes
Industry Yes Yes Yes Yes Yes Yes Adj.R2 0.258 0.260 0.223 0.225
0.251 0.258 #Obs. 2,819 2,819 3,592 3,592 1,084 1,084
-
43
Table 5 The effect of auditor designation on the relation
between auditor switch and stock liquidity This table reports the
regression results for the auditor designation effect on the
relation between auditor switch and stock liquidity. The sample
comprises 5,093 firm-year observations between 1995 and 2002. The
auditor designation rule is applied in 671 of these observations.
DSG is an indicator that equals 1 if a firm is audited by a
designated auditor and 0 otherwise. We use FACTOR and the scores
for single factors extracted from the three liquidity variables -
LOT, ZERO, and ROLL - as the dependent variables. We include
industry- and year-fixed effects based on the two-digit SIC code,
but we do not report the coefficients. The table reports the OLS
coefficient estimates and t-statistics based on robust standard
errors that are clustered by firm. See Table 1 for variable
definitions.
Depvar.=FACTOR (1) (2) (3) (4) Parameter Coef. t-stat. Coef.
t-stat. Coef. t-stat. Coef. t-stat.Intercept 2.524 4.59 2.512 4.58
2.453 4.49 2.439 4.46
SWITCH 0.085 2.65 0.092 2.28 BN 0.141 2.19 0.169 2.05NB 0.107
1.82 0.141 1.97BB 0.058 1.23 0.045 0.85NN 0.057 0.72 0.068 0.84
DSG 0.288 4.31 0.349 4.75DSG*SWITCH -0.170 -2.50
DSG*BN -0.327 -2.38DSG*NB -0.355 -2.56DSG*BB -0.203 -1.70DSG*NN
-0.292 -1.54
SIZE -0.091 -4.99 -0.090 -4.98 -0.087 -4.79 -0.086 -4.75LEV
-0.232 -1.69 -0.235 -1.70 -0.317 -2.32 -0.335 -2.44
ATURN 0.068 1.65 0.067 1.64 0.072 1.76 0.072 1.76ROA -0.987
-6.07 -0.987 -6.07 -0.981 -6.04 -0.970 -6.00
SGROW -0.130 -2.86 -0.129 -2.83 -0.127 -2.84 -0.126 -2.81QUICK
-0.003 -0.21 -0.003 -0.22 -0.006 -0.46 -0.007 -0.48
RET