ARTICLE IN PRESS

Journal of Accounting and Economics 40 (2005) 7599

www.elsevier.com/locate/jae

Acquirers abnormal returns and the non-Big 4

auditor clientele effect$
Henock Louis
Smeal College of Business, Pennsylvania State University, University Park, Pennsylvania,
PA 16802-1912, USA
Available online 29 August 2005

Abstract

I analyze the effect of auditor choice on acquirers values around merger announcements
and the factors affecting the interaction between auditor size and the market reaction to
merger announcements. I nd that acquirers audited by non-Big 4 accounting rms
outperform those audited by Big 4 rms. This effect is more pronounced when the targets
are privately held and when the likelihood of the auditors playing a prominent advisory role
increases. While the largest auditing rms are usually assumed to offer superior services, the
study suggests that smaller rms have a comparative advantage in assisting their clients in
merger transactions.
r 2005 Elsevier B.V. All rights reserved.

JEL classification: G34; G38; M41

Keywords: Merger; Auditor size; Auditor clientele; Non-audit services; Private target

$
This paper benets from comments by Dorothy Alexander-Smith, Dan Givoly, Steven Huddart,
Thomas Lys (the editor), Chunlin Mao, James McKeown, Santhosh Ramalingegowda, Philip Reckers,
Dahlia Robinson, Andrew Sbaraglia, Dan Simunic (a referee), an anonymous referee, and workshop
participants at Penn State University.
Tel.: +1 814 865 4160; fax: +1 814 863 8393.
E-mail address: hul4@psu.edu.

0165-4101/$ - see front matter r 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.jacceco.2005.03.001
 ARTICLE IN PRESS
76 H. Louis / Journal of Accounting and Economics 40 (2005) 7599

1. Introduction

I analyze the effect of auditor choice on acquirers market values around merger
announcements and the factors affecting the interaction between auditor size and the
market reaction to merger announcements. Carpenter and Strawser (1971), in a
management survey, nd that clients of non-Big 4 audit rms tend to switch to Big 4
rms prior to issuing stocks with the intent of reducing their cost-of-capital.1
Johnson and Lys (1990) also nd that anticipation of acquisitions is associated with
a rms decision to switch to a larger auditor. However, while a substantial
proportion of non-Big 4 audit rms clients switch to more prestigious auditors prior
to merger bids, most acquirers audited by non-Big 4 rms do not switch auditors.
This suggests that, for most clients of non-Big 4 audit rms, the continued use of
non-Big 4 auditors is optimal.
The Big 4 audit rms certainly have more resources than the non-Big 4 rms and
arguably provide better quality audit.2 However, the non-Big 4 rms are likely to
have comparative advantages in some areas. Merger and acquisition is probably one
such area. Non-Big 4 audit rms personnel presumably have superior knowledge of
the local markets and usually have close and long-time connections with and the
trust of their local business communities. According to William E. Balhoff,
Chairman, Executive Committee of the AICPA Public Company Practice Section,
[t]he CPA serves as the trusted advisor of the small business owner.3 In a survey
of chief executive ofcers (CEOs) of privately owned companies, Addams and Davis
(1994) nd that privately owned rms particularly value personal relationships with
their auditors, the auditors responsiveness, their advice, and their understanding of
the companies businesses. These needs seem to characterize not only privately
owned companies but also small businesses in general. Peale (1994) observes that
[s]mall [audit] rms tout their service, and lower prices, as proof that small
businesses should stick with small accounting rms. The claim of the smaller audit
rms seems consistent with existing studies in the organizational ecology area (see
Boone et al., 2000; Carroll et al., 2002). Boone et al. (2000, p. 363), for instance,
argue: [T]he services of small audit rms are much more personalized than those of
large auditors. In fact, the small-rm auditor frequently becomes the condant of the
small business manager, providing personal advice and information on many issues
y . Consistent with Boone et al. (2000), Berton (1994) observes that small clients
1
Throughout the paper, I use Big 4 generically to designate Big 4, Big 5, Big 6, and Big 8 audit rms,
depending on the period.
2
The presumed superior quality audit is reected in the clients of Big 4 audit rms having higher
earnings response coefcients (Teoh and Wong, 1993), higher audit fees (Beatty, 1989; Craswell et al.,
1995), lower litigation rates (St. Pierre and Anderson, 1984; Palmrose, 1988; Lys and Watts, 1994), and
lower discretionary accruals (Becker et al., 1998). According to Charles D. Niemeier, acting chairman of
the new accounting oversight board and former chief accountant in the SECs enforcement division, Big 4
audit rms are more likely to detect accounting problems (Hilzenrath, 2003). See Dopuch and Simunic
(1980), DeAngelo (1981), Titman and Trueman (1986), and Johnson and Lys (1990) for more detail on the
relation between audit rm size and audit quality.
3
Oversight Hearing on Accounting and Investor Protection Issues Raised by Enron and Other Public
Companies. U.S. Senate Committee on Banking, Housing, and Urban Affairs, March 14, 2002.
 ARTICLE IN PRESS
H. Louis / Journal of Accounting and Economics 40 (2005) 7599 77

are irritated by the poor service they receive from the large audit rms; they feel
especially shortchanged in the personnel assigned to them. Major accounting rms
are shifting small clients among employees so often that the practice is known as
churn and burn. The largest accounting rms also tend to neglect small clients in
favor of more lucrative businesses with larger clients.
Based on the suggestion that non-Big 4 audit rms have superior knowledge of
local markets and better relation with their clients, I posit that non-Big 4 audit rms
are likely to provide superior acquisition advice to their clients.4 While the Big 4
audit rms likely have more expertise in merger and acquisition activities, the non-
Big 4 rms have relatively more experience with privately held companies. Therefore,
I posit that, if non-Big 4 audit rms have a comparative advantage in assisting their
clients in merger and acquisition transactions, this advantage is likely to be more
evident when the target is a private company. The services of the Big 4 audit rms are
presumably more valuable when a merger target is large. However, in a large
acquisition, the acquirer is more likely to rely on the advice of investment bankers
than on the advice of auditors.
Consistent with my conjectures, I nd that acquirers audited by the non-Big 4
accounting rms signicantly outperform those audited by the Big 4 rms at merger
announcements. The results are robust to controlling for the usual factors affecting
the market reaction to merger announcements and potential sample-selection biases.
I also nd that the non-Big 4 audit rms clients outperform the Big 4 audit rms
clients signicantly when the targets are private companies. Furthermore, I nd
some evidence that the superior performance of the non-Big 4 audit rms clients is
related to the likelihood that the acquirers use in-house advisers instead of outside
investment bankers.
Since private rms have more concentrated ownership than public rms, a client
of a non-Big 4 rm might benet more from the acquisition of a private company
because of the increased monitoring that results from the creation of new block-
holdings by the former shareholders of the private target. Alternatively, investors
may perceive audits by non-Big 4 rms to be of lower quality. Hence, by voluntarily
submitting itself to the heightened scrutiny inherent in the merger process, a client of
a non-Big 4 audit rm, might, among other things, reveal to investors that it is of a
better type than they previously thought and, thus, deserves a lower cost-of-capital.5
Under these explanations, the auditor size effect would be driven by stock-for-stock
acquisitions;6 however, I nd no evidence consistent with this hypothesis. In fact, the

4
For a survey of the role of accountants as acquisition advisers, see Gorman (1988).
5
Titman and Trueman (1986) and Datar et al. (1991) show that the choice of a low quality auditor is
associated with less favorable private information by managers/entrepreneurs. Titman and Trueman
(1986) suggest that investors infer the nature of managers private information through their auditor
choice.
6
The merger process also increases rm scrutiny because of factors such as additional mandated
disclosure requirements, the need for the merging partners to exercise due diligence, increased investor
interest in the merging partners, and investors incentives to engage in costly information acquisitions (cf.
Wyser-Pratte, 1982; Boesky, 1985; Larcker and Lys, 1987; Stoughton, 1988; Shavell, 1994; Easley and
OHara, 2002). Because stock-for-stock acquirers are usually subject to more intense scrutiny than cash
 ARTICLE IN PRESS
78 H. Louis / Journal of Accounting and Economics 40 (2005) 7599

auditor size effect is stronger in cash acquisitions. The auditor size effect might also
be related to the acquirers visibility or investor recognition.7 One measure of a
rms visibility is the level of its analyst following. Everything else equal, a rm that
has low analyst coverage is likely to have higher cost-of-capital because fewer
analysts are providing research and disseminating information about the rm.
Hence, a low analyst coverage rm will probably benet more from an increase in
media and investor attention around a merger announcement. However, I nd no
evidence that the auditor size effect is related to the acquirers analyst following.
Finally, clients of non-Big 4 audit rms might be frugal companies, saving money on
the audit, and paying smaller premiums. In that case, the auditor choice effect would
simply be an artifact of the premium effect. Again, I nd no evidence supporting this
hypothesis.
The remainder of the study is organized as follows. The sample selection process is
described in the next section. Univariate statistics are reported in Section 3. In
Section 4, I control for the usual determinants of the market response to merger
announcements and potential sample selection biases, and test alternative explana-
tions for the auditor size effect. The study concludes in Section 5.

2. Sample selection

The study covers merger bids that were announced between January 1980 and
December 2002, inclusive. The sample is obtained from the Security Data
Companys (SDC) online database of domestic mergers and acquisitions.
A transaction is included in the sample if it satises the following criteria:

(a) the acquirer is a publicly traded company;

(b) the method of payment is reported on SDC;
(c) the targets total assets are reported on SDC;
(d) the acquirer has necessary Compustat data on equity book value, price, number
of shares outstanding, net income, total assets, total liabilities, and auditor code;
and
(e) the acquirer has the necessary data on CRSP to compute abnormal return.

(footnote continued)
acquirers, the auditor size effect should be stronger for stock-for-stock acquisitions under the heightened
scrutiny hypothesis. Cash mergers are also subjected to scrutiny because the acquirers generally nance the
mergers by borrowing the necessary funds. I assume, however, that an acquirer is likely to be subjected to
more scrutiny when it is trading in its own stock.
7
Merton (1987) provides a model of capital market equilibrium with incomplete markets in which
expected return depends not only on risk but also on a rms investor base. Under the investor
recognition theory, market participants invest in securities that they are aware of, imposing a higher cost-
of-capital on less visible rms. This theory is consistent with empirical evidence on what Arbel et al. (1983)
label neglected stocks. Therefore, clients of non-Big 4 audit rm clients may outperform at merger
announcements because they are less visible and, hence, benet more from the exposure generated by the
merger process.
 ARTICLE IN PRESS
H. Louis / Journal of Accounting and Economics 40 (2005) 7599 79

There are 3707 mergers that meet the selection criteria. The sample has 205
acquirers audited by non-Big 4 accounting rms and 3502 audited by Big 4 rms.

3. Univariate analysis

3.1. Descriptive statistics

Descriptive statistics are reported in Table 1. Compared to clients of Big 4 audit

rms, clients of non-Big 4 audit rms and their targets are very small. They have
higher book-to-market ratios, lower return-on-assets (ROA), lower earnings-to-price
ratios, lower analyst following, and higher pre-merger stock volatility. They are also
less leveraged, more likely to acquire private companies, less likely to use the services
of investment bankers, less likely to merge within their industry, and less likely to
account for their acquisitions by the pooling-of-interest method.

3.2. Acquirers abnormal returns

Table 2 presents statistics on the abnormal returns of the acquirers over the 3-day
period centered on the merger announcement date (t [
1,1]). The market model is
used to estimate return expectations for the individual rms. The market return is
proxied by the CRSP value-weighted return. The model is estimated over the period
from 60 to 259 days prior to the merger announcement date. I rst report results for
the full sample. I nd a signicantly positive market reaction to merger
announcements by non-Big 4 clients and a signicantly negative reaction to merger
announcements by Big 4 clients. Both the t-value and the Wilcoxon Z statistic for
testing for difference in the abnormal returns of the two groups of bidders are
statistically signicant.
I then condition the analysis on the public status of the targets. The mean and
median abnormal returns for acquisitions of publicly traded companies are
insignicantly positive for the clients of the non-Big 4 audit rms and signicantly
negative for the clients of the Big 4 audit rms. The differences in the mean and
median abnormal returns for the non-Big 4 audit rms clients and the Big 4 audit
rms clients are not signicantly different from zero. On the other hand, the mean
and median abnormal returns for acquisitions of privately held companies are
signicantly positive for both clients of non-Big 4 audit rms and clients of Big 4
audit rms. However, the mean (median) abnormal return for acquisitions of
privately held companies by clients of non-Big 4 auditors is signicantly higher than
the mean (median) abnormal return for acquisitions of privately held companies by
clients of Big 4 auditors. The results are consistent with the conjecture that smaller
audit rms have a comparative advantage in providing advice in acquisition and
valuation of private rms.
Next, I partition on whether the acquirer uses the services of investment bankers.
For acquisitions in which the acquirers use the services of investment bankers, the
mean and median abnormal returns are insignicantly different from zero for the
 ARTICLE IN PRESS
80 H. Louis / Journal of Accounting and Economics 40 (2005) 7599

Table 1
Descriptive statistics

Non-Big 4 client N 205 Big 4 client N 3502 T-statistics for Z-statistics from
tests of mean the Wilcoxon
differences two-sample test
Mean Median Mean Median

MV 1,072.000 44.832 8,560.000 857.021
10.47***
14.93***

ASSET 1,348.000 41.663 7,059.000 753.642
6.32***
14.54***
LOGASSET 4.082 3.730 6.643 6.625
16.63***
14.53***
TASSET 460.703 19.300 909.984 90.450
1.58
9.44***
LOGTASSET 3.115 2.081 4.561 2.091
9.67***
9.44***
RSIZE 13.138 2.543 64.188 6.655
7.40***
8.64***
LEV 0.458 0.465 0.511 0.523
3.18***
3.15***
CASH 0.156 0.085 0.164 0.082 0.67 0.33
C 0.205 0.000 0.291 0.000
2.94***
2.65***
BM 0.579 0.484 0.495 0.398 2.84*** 3.42***
ROA
0.038 0.038 0.022 0.049
3.09***
3.43***
EP
0.026 0.041 0.010 0.043
1.86*
1.74*
ACOV 1.293 0.000 4.783 0.000
10.62***
7.24***
VOL 0.049 0.039 0.029 0.025 5.89*** 9.87***
REGIND 0.122 0.000 0.118 0.000 0.17 0.17
PRIVATE 0.283 0.000 0.141 0.000 4.41*** 5.53***
NIB 0.356 0.000 0.712 1.000
8.87***
8.00***
INHOUSE 0.693 1.000 0.397 0.000 8.87*** 8.36***
STOCK 47.300 30.100 45.200 32.000 0.61 0.73
POOL 0.107 0.000 0.175 0.000
3.00***
2.50**
INDR 0.390 0.000 0.463 0.000
2.06**
2.02**
FRIENDLY 0.925 1.000 0.927 1.000 0.07 0.07

Notes: Big 4 refers to clients of Big 4 (Big 5, Big 6, or Big 8, depending on the period) auditing rms. MV is
the acquirers total market value of equity (in millions of dollars); ASSET is the acquirers total assets (in
millions of dollars); LOGASSET is the log of ASSET; TASSET is the targets total assets (in millions of
dollars); LOGTASSET is the log of TASSET; RSIZE, relative size, is the ratio of the acquirers total assets
to the targets total assets; LEV is the acquirers leverage dened as total liabilities divided by total assets;
CASH is the acquirers cash divided by total assets; BM is the acquirers book-to-market ratio; ROA is the
acquirers return-on-assets; EP is the acquirers earnings-to-price ratio; VOL is a measure of the acquirers
pre-merger stock volatility proxied by the standard deviation of the acquirers return over the period from
60 days to 259 days before the merger announcement; ACOV is analyst coverage, measured by the number
of analysts forecasting an acquirers annual earnings in the month immediately prior to the earnings
announcement; REGIND is a binary variable taking the value one if the acquirer is in a regulated industry
(SIC 6069, and 49) and zero otherwise; PRIVATE is a binary variable taking the value one is the target is
a private company and zero otherwise; NIB is the number of investment bankers identied by SDC;
INHOUSE is a binary variable taking the value one if no investment banker is identied by SDC and zero
otherwise; STOCK is the percentage of the transaction nanced with common stock; C is a binary variable
that is equal to one if the merger is nanced entirely with cash and zero otherwise; POOL is a binary
variable taking the value one if the merger is accounted for by the pooling-of-interest method and zero if it
is accounted for by the purchase method; and INDR, a proxy for the industry relatedness of the merging
rms, it is equal to one if the two merging partners are in the same three-digit SIC code and zero otherwise;
and FRIENDLY equals one if the targets attitude to the proposed merger is characterized as friendly by
SDC, and zero otherwise. ***, **, and * indicate that the differences between Big 4 and Non-Big 4 audit
rms clients are signicant at the 1, 5, and 10 percent levels in a two-tail test. For the tests of mean
differences, I report the Satterthwaites t values, assuming that the variances for Big 4 and Non-Big 4 audit
rms clients are unequal.
 ARTICLE IN PRESS
H. Louis / Journal of Accounting and Economics 40 (2005) 7599 81

Table 2
Acquirers percentage cumulative abnormal return around the merger announcement: clients of non-Big 4
versus clients of Big 4 audit rms

Non-Big 4 client Big 4 client T-statistics for Z-statistics from

tests of mean the Wilcoxon
differences two-sample test
Mean Median Mean Median

Full sample 2.52 0.24
0.59
0.46 2.65+++ 2.99+++

(0.032) (0.086) (0.000) (0.000)
N 205 N 205 N 3502 N 3502

Publicly traded 0.62 0.51
0.99
0.68 1.26 0.7

targets (0.627) (0.435) (0.000) (0.000)
N 147 N 147 N 3007 N 3007

Privately held 7.36 5.45 1.82 0.76 3.48+++ 3.38+++

targets (0.005) (0.000) (0.000) (0.000)
N 58 N 58 N 495 N 495

Investment
0.55 0.01
1.35
0.91 0.73 1.15

banker (0.572) (0.847) (0.000) (0.000)
N 63 N 63 N 2112 N 2112

No investment 3.89 0.36 0.56 0.01 2.04++ 1.92++

banker (0.018) (0.035) (0.009) (0.183)
N 142 N 142 N 1390 N 1390

Notes: Big 4 refers to clients of Big 4 (Big 5, Big 6, or Big 8, depending on the period) auditing rms.
The cumulative abnormal return of the bidders is estimated over the day of the merger announcement, the
preceding day, and the day after (t [
1, 1]). I use the market model to estimate return expectations for
the individual bidders. The market return is proxied by the CRSP value-weighted return. The model is
estimated over the period from 60 to 259 days before the bid announcement date.
The label investment banker indicates that SDC reports that the acquirer uses the services of investment
bankers and no investment banker indicates that SDC identies no investment bankers.
Two-tail p-values are reported in parentheses for the means and the medians, based respectively on the t-
test and the Wilcoxon signed rank test. +++ and ++ indicate that the differences in the abnormal returns
of Big 4 and Non-Big 4 audit rms clients are signicant at the 1 and 5 percent levels in a one-tail test. For
the tests of mean differences, I report the Satterthwaites t values, assuming that the variances for Big 4
and Non-Big 4 audit rms clients are unequal.

clients of the non-Big 4 auditors and signicantly negative for the clients of the Big 4
auditors. The differences in the mean and median abnormal returns for clients of
non-Big 4 auditors and clients of Big 4 auditors are not signicantly different from
zero. On the other hand, for acquisitions for which SDC identies no investment
bankers, the mean abnormal returns are signicantly positive for both clients of non-
Big 4 auditors and clients of Big 4 auditors. The median abnormal return is
signicantly positive for clients of non-Big 4 auditors and insignicantly different
from zero for clients of Big 4 auditors. However, the mean and median abnormal
returns of the non-Big 4 audit rms clients are signicantly higher than the mean
and median abnormal returns of the Big 4 audit rms clients. These results are
 ARTICLE IN PRESS
82 H. Louis / Journal of Accounting and Economics 40 (2005) 7599

consistent with the conjectures that clients of non-Big 4 audit rms are more likely to
outperform clients of Big 4 rms when the likelihood of the auditors playing a
prominent advisory role increases.

3.3. Simple correlation

Table 3 reports simple correlations among various relevant variables. There is a

signicantly negative correlation between the acquirers merger announcement
abnormal returns and acquisitions by clients of Big 4 audit rms. The abnormal
return is also positively correlated with the acquirers book-to-market ratios, private
acquisitions, and the likelihood that the acquirers use in-house advisers, and
negatively correlated with the targets sizes, the acquirers ROA, analyst coverage,
stock-for-stock acquisitions, pooling-of-interest acquisitions, and intra-industry
acquisitions. Acquisitions by clients of Big 4 audit rms are negatively correlated
with the acquirers book-to-market ratios, private acquisitions, the likelihood that
the acquirers use in-house advisers, pooling-of-interest acquisitions, and intra-
industry acquisitions, and positively correlated with analyst coverage and the
acquirers total market value of equity, total assets, and ROA.

4. Multivariate analysis

4.1. Controlling for potential correlated variables

Previous studies nd that in general cash acquirers outperform stock-for-stock

acquirers. This result is interpreted as announcements of a stock-for-stock
acquisition signaling that managers of the bidding rms have adverse private
information (cf. Leland and Pyle, 1977; Jensen and Ruback, 1983; DeAngelo et al.,
1984; Myers and Majluf, 1984; Travlos, 1987). Pandit (2003) nds that the method
used to account for a merger is also associated with the merger announcement
return, even after controlling for the method of payment. Lys and Vincent
(1995) and Pandit (2003) suggest that, on average, pooling transactions are bad
investment decisions because they are driven by accounting earnings instead of cash
ows and are subjected to several restrictions under Accounting Principles Board
Opinion No. 16.
Chang (1998) and Fuller et al. (2002) also nd a positive market reaction to
announcements of privately owned rm acquisitions. Consistent with this nding,
Moeller et al. (2005) report that few of the large merger announcement losses come
from acquisitions of privately held companies. Chang (1998) suggests that, because
of the concentrated ownership in privately owned companies, a stock-for-stock
acquisition of a private rm decreases the cost to monitor the acquirer and reduces
information asymmetry between management and shareholders. Servaes and Zenner
(1996) also nd that acquirers that use the advice of investment bankers experience
lower acquisition announcement returns.
Table 3
Pearson correlation (N 3707)

BIG4 MV ASSET LOG ASSET TASSET LOG TASSET RSIZE BM ROA ACOV PRIVATE NIB INHOUSE STOCK C POOL INDR FRIENDLY

CAR
0.079
0.024
0.016
0.113
0.036
0.145 0.026 0.071
0.078
0.045 0.130
0.121 0.119
0.118 0.095
0.089
0.034
0.000
(0.000) (0.147) (0.335) (0.000) (0.028) (0.000) (0.108) (0.000) (0.000) (0.006) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.037) (0.993)

BIG4 0.056 0.037 0.262 0.020 0.156 0.040
0.044 0.072 0.096
0.091 0.120
0.137
0.010 0.044
0.041
0.033
0.001
(0.001) (0.024) (0.000) (0.227) (0.000) (0.015) (0.007) (0.000) (0.000) (0.000) (0.000) (0.000) (0.530) (0.008) (0.012) (0.043) (0.943)

MV 0.560 0.431 0.233 0.197 0.220
0.163 0.083 0.230
0.088 0.064
0.044 0.020 0.055 0.0240
0.040 0.037
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.008) (0.231) (0.001) (0.145) (0.015) (0.025)

ASSET 0.375 0.448 0.245 0.134
0.042 0.020 0.054
0.071 0.072
0.069
0.024 0.043
0.009
0.055 0.026
(0.000) (0.000) (0.000) (0.000) (0.010) (0.216) (0.001) (0.000) (0.000) (0.000) (0.146) (0.009) (0.603) (0.001) (0.120)

LOG 0.252 0.656 0.237
0.071 0.252 0.304
0.330 0.304
0.292
0.156 0.167
0.011
0.050
0.034
ASSET (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.497) (0.002) (0.041)

TASSET 0.364
0.031
0.015 0.019 0.031
0.066 0.148
0.110 0.034
0.036 0.042 0.008
0.002
(0.000) (0.000) (0.376) (0.248) (0.058) (0.000) (0.000) (0.000) (0.038) (0.029) (0.011) (0.642) (0.177)

LOG 0.340 0.025 0.166 0.101
0.402 0.475
0.456
0.108 0.013 0.004 0.023 0.165
TASSET (0.000) (0.121) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.411) (0.813) (0.168) (0.000)

RSIZE
0.044 0.033 0.164
0.086
0.081 0.082
0.041 0.093
0.043
0.084 0.046
(0.008) (0.046) (0.000) (0.000) (0.000) (0.000) (0.013) (0.000) (0.008) (0.000) (0.005)
ARTICLE IN PRESS

BM
0.179
0.178 0.003
0.067 0.072
0.164 0.018
0.184
0.050
0.049
(0.000) (0.000) (0.869) (0.000) (0.000) (0.000) (0.277) (0.000) (0.002) (0.003)

ROA 0.113
0.097 0.042
0.042
0.138 0.105 0.068
0.037
0.040
(0.000) (0.000) (0.011) (0.010) (0.000) (0.000) (0.000) (0.023) (0.014)

ACOV
0.044 0.073
0.112 0.098
0.013 0.193
0.019 0.097

H. Louis / Journal of Accounting and Economics 40 (2005) 7599

0.007 (0.000) (0.000) (0.000) (0.426) (0.000) (0.258) (0.000)

PRIVATE
0.208 0.218 0.074
0.133 0.055
0.025 0.110

(0.000) (0.000) (0.000) (0.000) (0.001) (0.125) (0.000)

NIB
0.860 0.045 0.049 0.082 0.055
0.070

(0.000) (0.006) (0.003) (0.000) (0.001) (0.000)
83
Table 3 (continued ) 84

BIG4 MV ASSET LOG ASSET TASSET LOG TASSET RSIZE BM ROA ACOV PRIVATE NIB INHOUSE STOCK C POOL INDR FRIENDLY

INHOUSE
0.067
0.043
0.129 0.056 0.043

(0.000) (0.010) (0.000) (0.001) (0.007)

STOCK
0.625 0.501 0.131 0.167

(0.000) (0.000) (0.000) (0.000)

C
0.286
0.081
0.159

(0.000) (0.000) (0.000)

POOL 0.079 0.041

(0.000)
0.012

INDR 0.033
(0.037)

Notes: CAR is the percentage cumulative abnormal return of the bidder computed over the day of the merger announcement, the preceding day, and the day after
(t [
1,1]); BIG4 is a binary variable that is equal to one for clients of Big 4 audit rms and zero for clients of non-Big 4 rms; MV is the acquirers total market
value of equity (in millions of dollars); ASSET is the acquirers total assets (in millions of dollars); LOGASSET is the log of the acquirers total assets; TASSET is
the targets total assets (in millions of dollars); LOGTASSET is the log of the targets total assets; RSIZE, relative size, is the ratio of the acquirers total assets to
the targets total assets; BM is the acquirers book-to-market ratio; ROA is the acquirers return-on-assets; ACOV is analyst coverage, measured by the number of
analysts forecasting an acquirers annual earnings in the month immediately prior to the earnings announcement; PRIVATE is a binary variable taking the value
one is the target is a private company and zero otherwise; NIB is the number of investment bankers identied by SDC; INHOUSE is a binary variable taking the
value one if no investment banker is identied by SDC and zero otherwise; STOCK is the percentage of the transaction nanced with common stock; C is a binary
ARTICLE IN PRESS

variable that is equal to one if the merger is nanced entirely with cash and zero otherwise; POOL is a binary variable taking the value one if the merger is
accounted for by the pooling-of-interest method and zero if it is accounted for by the purchase method; INDR, a proxy for the industry relatedness of the merging
rms, is equal to one if the two merging partners are in the same three-digit SIC code and zero otherwise; and FRIENDLY equals one if the targets attitude to the
proposed merger is characterized as friendly by SDC, and zero otherwise. Two-tail p-values are reported in parentheses.
H. Louis / Journal of Accounting and Economics 40 (2005) 7599
 ARTICLE IN PRESS
H. Louis / Journal of Accounting and Economics 40 (2005) 7599 85

Lang et al. (1989) and Servaes (1991) report a signicant correlation between
Tobins q and bidders abnormal returns. Rau and Vermaelen (1998) argue that high
book-to-market acquirers tend to be more cautious before engaging in major
transactions and their acquisitions are less likely to be motivated by hubris.
Accordingly, Louis (2004a) nds a positive association between book-to-market
ratios and acquirers abnormal returns at merger announcements.
The extant literature also documents a positive correlation between a rms
market value and the level of concentration in its lines of business.8 Scanlon et al.
(1989) and Louis (2004b) also nd that acquirers abnormal returns decrease in the
size of the targets. The size effect is explained by the fact that small targets are
usually more manageable.
To ensure that the auditor size effect documented in Table 2 is not biased due to
correlated omitted variables, I control for the determinants of acquirers
performance at merger announcements. I model the market reaction to merger
announcements as follows:

CARi b0 b1 BIG4i b2 STOCK i b3 C i b4 POOLi b5 PRIVATE i

b6 NIBi b7 INHOUSE i b8 BM i b9 ROAi b10 INDRi
b11 ACOV i b12 FRIENDLY i b13 LOGTASSET i b14 RSIZE i ei ,

where CAR is the cumulative abnormal return of the bidder computed over the 3
days centered on the merger announcement date (t [
1,1]); BIG4 is a binary
variable that is equal to one for clients of Big 4 audit rms and zero for clients of
non-Big 4 rms; STOCK is the percentage of the transaction nanced with common
stock; C is a binary variable that is equal to one if the merger is nanced entirely with
cash and zero otherwise; POOL is a binary variable taking the value one if the
merger is accounted for by the pooling-of-interest method and zero if it is accounted
for by the purchase method; PRIVATE is a binary variable taking the value one if
the target is a private company and zero otherwise; NIB is the number of investment
bankers identied by SDC; INHOUSE is a binary variable taking the value one if no
investment banker is identied by SDC and zero otherwise; BM is the acquirers
book-to-market ratio; ROA is the acquirers return-on-assets; INDR is a proxy for
the industry relatedness of the merging rms; it is equal to one if the two merging
partners are in the same three-digit SIC code and zero otherwise; ACOV is analyst
coverage, measured by the number of analysts forecasting the acquirers annual
earnings in the month immediately prior to the earnings announcement; FRIENDLY
equals one if the targets attitude to the proposed merger is characterized as friendly
by SDC, and zero otherwise; LOGTASSET is the log of the targets total assets; and
RSIZE, relative size, is the ratio of the acquirers total assets to the targets total
assets.

8
See Morck et al. (1990), Lang and Stulz (1994), Berger and Ofek (1995), John and Ofek (1995), and
Comment and Jarrell (1995).
 ARTICLE IN PRESS
86 H. Louis / Journal of Accounting and Economics 40 (2005) 7599

I control for the percentage of the acquisitions nanced with equity and the
number of investment bankers hired by the acquirer. I also add C and INHOUSE to
the model to control for potential discontinuities around zero. In other words, I want
to make sure that any effect that I document is not due to the failure to control for
the potential nonlinearity in the relation between the market reaction, and the
percentage of the transaction nanced with common stock and the number of
investment bankers advising the acquirer. In particular, controlling for INHOUSE is
consistent with Servaes and Zenner (1996) who study whether the use of investment
bankers advice is related to acquirers merger announcement returns. I also control
for the potential drivers of the auditor choice. Because the Big 4 clients are many
times as large as the non-Big 4 clients and because, ceteris paribus, an acquirers
return from a merger linearly decreases in the size of the acquirer,9 the bias
potentially has a very large effect on the difference in the abnormal returns of the
two groups of acquirers.10 In addition, there are potentially other factors, besides
size, that determine auditor choice. To the extent that these factors are also
correlated with the error term in regression (1), b1, which is the difference in the
abnormal returns of clients of non-Big 4 audit rms and clients of Big 4 audit rms,
will be biased. To ensure that the documented correlation between the markets
reaction to a merger announcement and the size of the audit rm is not driven by the
determinants of the auditor choice, I use Heckmans (1979) two-step procedure.
First, I estimate the following probit model:
BIG4i a0 a1 LOGASSET i a2 LEV i a3 CASH i
a4 VOLi a5 EPi a6 REGINDi ui , 2
where LOGASSET is the log of the acquirers total assets (in million); LEV is
the acquirers leverage dened as total liabilities divided by total assets; CASH is
the acquirers cash divided by total assets; EP is the acquirers earnings-to-price
ratio; VOL is a measure of the acquirers pre-merger stock volatility proxied
by the standard deviation of the acquirers return over the period from 60 to
259 days before the merger announcement; and REGIND is a binary variable taking
the value one if the acquirer is in a regulated industry (SIC 6069, and 49) and zero
otherwise.
I include leverage in the model because rms may choose prestigious auditors for
contracting reasons related to their debts. Highly leveraged rms may be required by
their lenders to have their nancial reports attested by certain audit rms.
Alternatively, the more leveraged rms may have chosen large audit rms to reduce
their borrowing costs. Chow (1982) and Ettredge et al. (1994) nd that the demand
for the attest service increases in leverage. Lys and Watts (1994) suggest that auditors
9
Assuming that an acquirers benet from a merger is $10. This represents a one percent return for an
acquirer that has a market value of $1000 and 10 percent for an acquirer that has a market value of $100.
10
As a robustness check, I add different combinations of target size, acquirer size, and the relative size of
the merging partners directly in the regression model. I obtain qualitatively the same results as those
reported in the paper. I actually nd that target size and not acquirer size is associated with an acquirers
abnormal return at the merger announcement. There is no signicant association between the abnormal
return and acquirer size after controlling for target size.
 ARTICLE IN PRESS
H. Louis / Journal of Accounting and Economics 40 (2005) 7599 87

whose clients have liquidity problems are more likely to be sued. Therefore,
these rms may tend to hire smaller audit rms because deep pocket auditors
may be reluctant to audit them. I use the ratio of cash to assets to proxy for
liquidity.11
Datar et al. (1991) show that the demand for high-quality audit increases in a
clients risk. Accordingly, I include the acquirers stock price volatility in the model:
volatile clients are likely to involve high audit risk. Such rms are likely to be
avoided by auditors that have the most at stake in terms of reputation and litigation
costs. I also control for earnings-to-price (EP) ratio because the values of the rms
that have high EP ratios tend to be less sensitive to earnings and to the accounting
report in general. Therefore, such rms have less incentive to hire high-quality
auditors. Finally, I include a regulated industry dummy. Because nancial reporting
in the nancial and utility sectors is regulated, audit quality should be less valuable in
these sectors. Hence, nancial and utility companies are less likely to pay a premium
to hire Big 4 auditors.
Then, I use the inverse Mills ratio from the probit model to control for the auditor
choice as follows:

CARi b0 b1 BIG8i b2 STOCK i b3 C i b4 POOLi

b5 PRIVATE i b6 NIBi b7 INHOUSE i b8 BM i
b9 ROAi b10 INDRi b11 ACOV i
b12 FRIENDLY i b13 LOGTASSET i
b14 RSIZE i b15 LAMBDAi ei , 3

where LAMBDA is the inverse Mills ratio.

The results of the probit model are reported in Table 4. The data t the
model quite well. The model classies 81 percent of the observations correctly,
the Nagelkerkes (1991) pseudo-R2 is 0.22, and chi-squares testing whether
the coefcient estimates are jointly zero are signicant with p-values below 0.000.
Acquirer size, cash, volatility, and earnings-to-price ratio are signicantly associated
with the auditor choice. Consistent with the results reported in Table 1,
I nd a signicant correlation between leverage and auditor choice when acquirer
size is not included in the model. However, after controlling for size, leverage is not
signicant.
In Table 5, I report the conditional association between audit rm size and
acquirers abnormal returns at merger announcements. There is some weak evidence
of sample selection biases. The coefcient on LAMBDA, the inverse Mills ratio, is
negatively correlated with the merger announcement return. However, the market
reactions to the merger announcements are signicantly lower for the Big 4 clients
than for the non-Big 4 clients, even after controlling for Lambda and the other
determinants of merger announcement abnormal returns.

11
I use cash to proxy for liquidity instead of a broader working capital measure to maximize the sample
size of the non-Big 4 clients.
 ARTICLE IN PRESS
88 H. Louis / Journal of Accounting and Economics 40 (2005) 7599

Table 4
Probit regression: auditor choice model (N 3707)

BIG4i a0 a1 LOGASSET i a2 LEV i a3 CASH i a4 VOLi a5 EPi a6 REGINDi ui

Variable Coefcient Standard Wald w2 p-value

estimate error

INTERCEPT 0.291 0.202 2.070 0.150

LOGASSET 0.258 0.024 118.040 0.000
LEV 0.033 0.196 0.030 0.866
CASH 1.035 0.234 19.600 0.000
VOL
7.750 2.266 11.700 0.001
EP
0.359 0.197 3.320 0.069
REGIND
0.246 0.122 4.100 0.043
w2 p-value

Likelihood ratio test 304.282 0.000

Score test 360.493 0.000
Wald test 235.626 0.000
Percent concordant (discordant) 81.0 (18.2)
Max-rescaled R2 0.227

Notes: BIG4 is a binary variable that is equal to one for clients of Big 4 audit rms and zero for clients of
non-Big 4 rms; LOGASSET is the log of the acquirers total assets; LEV is the acquirers leverage dened
as total liabilities divided by total assets; CASH is the acquirers cash divided by total assets; EP is the
acquirers earnings-to-price ratio; VOL is a measure of the acquirers pre-merger stock volatility proxied
by the standard deviation of the acquirers return over the period from 60 to 259 days before the merger
announcement; and REGIND is a binary variable taking the value one if the acquirer is in a regulated
industry (SIC 6069, and 49) and zero otherwise.

4.2. Explaining the audit firm size effect

There are many potential explanations for the auditor size effect. Under the
heightened scrutiny hypothesis, acquirers audited by the non-Big 4 accounting rms
outperform at merger announcements because the non-Big 4 rms are presumed to
provide lower quality audit and, therefore, their clients stocks are priced at a
discount. By voluntarily submitting itself to greater scrutiny through the merger
process, a client of a non-Big 4 audit rm reveals to investors that it is of a better
type than they previously thought and, therefore, deserves a lower cost-of-capital.
The neglected rm hypothesis suggests that non-Big 4 audit rm clients are priced at
a discount because they have a lower investor-base due to their low visibility.
Therefore, the exposure generated by the merger process benets them by increasing
their visibility. Finally, according to the superior clientele-advising hypothesis, non-
Big 4 audit rms direct their services to entrepreneurial clients and their partners
and staff have close connections with and the trust of their local business
communities. This creates an environment where the non-Big 4 audit rms are
likely to provide superior advice in acquisition and valuation of private rms. The
 ARTICLE IN PRESS
H. Louis / Journal of Accounting and Economics 40 (2005) 7599 89

Table 5
Determinants of the audit size effect (N 3707)

CARi b0 b1 BIG4i b2 STOCK i b3 C i b4 POOLi

b5 PRIVATE i b6 NIBi b7 INHOUSE i
b8 BM i b9 ROAi b10 INDRi b11 ACOV i
b12 FRIENDLY i b13 LOGTASSET i b14 RSIZE i
b15 BIG4i  STOCK i b16 BIG4i  ACOV i
b17 BIG4i  PRIVATE i b18 BIG4i
 PRIVATE i  STOCK i b19 BIG4i  INHOUSE i
b20 BIG4i  INHOUSE i  STOCK i
b21 LAMBDAi ei ,

BIG4i a0 a1 LOGASSET i a2 LEV i a3 CASH i a4 EPi a5 VOLi a6 REGINDi ui

BIG4 0 if BIG4  o0 and BIG4 1 if BIG4  40

Coefcient Variable Expected sign Model 1 Model 2

b0 Intercept 7
0.422***
0.422***

(
2.91) (
2.92)
b1 BIG4

5.263+++
2.871
(
2.69) (
1.02)
b2 STOCK
0.019+++
0.033+++
(
4.05) (
2.45)
b3 C + 1.064+++ 0.955+++
(2.51) (2.25)
b4 POOL

0.267
0.123
(
0.58) (
0.26)
b5 PRIVATE + 2.329+++ 5.142+++
(5.13) (3.61)
b6 NIB

0.571+
0.589+
(
1.33) (
1.38)
b7 INHOUSE + 0.002 1.287
(0.000) (0.84)
b8 BM + 0.919+++ 1.006+++
(2.63) (2.89)
b9 ROA

2.914+++
2.794+++
(
3.64) (
3.50)
b10 INDR +
0.125
0.135
(
0.42) (
0.46)
b11 ACOV + 0.006
0.031
(0.31) (
0.21)
b12 FRIENDLY +
0.046 0.225
(
0.08) (0.39)
b13 LOGTASSET

0.282+++
0.284+++
(
3.01) (
2.99)
 ARTICLE IN PRESS
90 H. Louis / Journal of Accounting and Economics 40 (2005) 7599

Table 5 (continued )

Coefcient Variable Expected sign Model 1 Model 2

b14 RSIZE + 0.000 0.000

(0.67) (0.62)
b15 BIG4*STOCK


0.001
(
0.04)
b16 BIG4*ACOV
_ 0.034
(0.23)
b17 BIG4*PRIVATE
_
5.015+++
(
3.21)
b18 BIG4*PRIVATE*STOCK + _ 0.035+++
(3.62)
b19 BIG4*INHOUSE
_
2.350+
(
1.55)
b20 BIG4*INHOUSE*STOCK + _ 0.024+++
(3.53)
b21 LAMBDA 7
1.826*
1.448
(
1.81) (
1.29)
Adjusted R2 0.055 0.064

Notes: CAR is the percentage cumulative abnormal return of the bidder computed over the day of the
merger announcement, the preceding day, and the day after (t [
1,1]); BIG4 is a binary variable that is
equal to one for clients of Big 4 audit rms and zero for clients of non-Big 4 rms; STOCK is the
percentage of the transaction nanced with common stock; C is a binary variable that is equal to one if the
merger is nanced entirely with cash and zero otherwise; POOL is a binary variable taking the value one if
the merger is accounted for by the pooling-of-interest method and zero if it is accounted for by the
purchase method; PRIVATE is a binary variable taking the value one is the target is a private company
and zero otherwise; NIB is the number of investment bankers identied by SDC; INHOUSE is a binary
variable taking the value one if no investment banker is identied by SDC and zero otherwise; BM is the
acquirers book-to-market ratio; ROA is the acquirers return-on-assets; INDR, a proxy for the industry
relatedness of the merging rms, is equal to one if the two merging partners are in the same three-digit SIC
code and zero otherwise; ACOV is analyst coverage, measured by the number of analysts forecasting the
acquirers annual earnings in the month immediately prior to the earnings announcement; FRIENDLY
equals one if the targets attitude to the proposed merger is characterized as friendly by SDC, and zero
otherwise; LOGTASSET is the log of the targets total assets; LOGASSET is the log of the acquirers total
assets; RSIZE, relative size, is the ratio of the acquirers total assets to the targets total assets; and
LAMBDA is the inverse Mills ratio from the probit model. I de-mean all the independent variables. The
variables in the probit model are dened in Tables 4. T-statistics are reported in parentheses. *** (+++)
and * (+) indicate signicance at the 1 and 10 percent levels in a two-tail (one-tail) test.

services of Big 4 audit rms are likely to be more valuable when the target is large;
however, in a large acquisition, the acquirer is more likely to rely on the advice of
investment banks than on the advice of audit rms. To test these alternative
hypotheses, I extend regression (3) as follows:
CARi b0 b1 BIG4i b2 STOCK i b3 C i b4 POOLi
b5 PRIVATE i b6 NIBi b7 INHOUSE i
b8 BM i b9 ROAi b10 INDRi b11 ACOV i
 ARTICLE IN PRESS
H. Louis / Journal of Accounting and Economics 40 (2005) 7599 91

b12 FRIENDLY i b13 LOGTASSET i b14 RSIZE i

b15 BIG4i  STOCK i b16 BIG4i  ACOV i
b17 BIG4i  PRIVATE i b18 BIG4i  PRIVATE i
 STOCK i b19 BIG4i  INHOUSE i b20 BIG4i
 INHOUSE i  STOCK i b21 LAMBDAi ei . 4

Stock-for-stock acquirers are usually subject to more intense scrutiny than cash
acquirers. If the heightened scrutiny hypothesis is supported by the data, I expect the
auditor size effect to increase in the proportion of equity used in the transaction.
Hence, under the heightened scrutiny hypothesis, I expect clients of the largest
auditors to fare signicantly worse than those of the smaller auditors in stock-for-
stock mergers. Therefore, I expect b15 to be negative. Similarly, if the neglected rm
hypothesis holds, I expect b16, the coefcient on the interaction between the auditor
size effect and analyst coverage, to be negative. Because non-Big 4 auditors have
relatively more experience with privately owned companies, I conjecture that if they
provide more valuable advice to their clients, the effect should be more manifest
when the clients are targeting privately owned rms. Therefore, under the superior
clientele-advising hypothesis, the clients of the largest audit rms should fare
signicantly worse than the clients of the smaller auditors in acquisitions of private
rms. Hence, I expect b17 to be negative. Because stock-for-stock mergers are more
complex than cash acquisitions, the largest audit rms should have a relative
advantage in advising their clients in stock-for-stock transactions. Therefore,
I expect b18 to be positive. That is, the negative effect of the interaction between
BIG4 and PRIVATE dampens as the share of the transaction nanced with stock
increases. Finally, I conjecture that auditors are likely to play a more prominent
advisory role in the merger process as their clients rely less on the services of
investment bankers. Therefore, under the superior clientele-advising hypothesis,
I expect the clients of the largest audit rms to fare signicantly worse than the
clients of the smaller auditors as the acquirers reliance on the services of investment
bankers decreases. Hence, I expect b19 to be negative. However, because the largest
audit rms presumably have a relative advantage in advising clients in stock-for-
stock acquisitions, the effect of the interaction between BIG4 and INHOUSE should
dampen as the proportion of the transaction nanced with stock increases.
Therefore, I expect b20 to be positive.
A signicantly negative coefcient on the interaction between BIG4 and
PRIVATE is also consistent with Changs (1998) heightened monitoring hypothesis.
Chang suggests that acquirers experience signicantly higher abnormal returns in
acquisitions of privately owned rms because such acquisitions result in greater
monitoring of the acquirers by the former shareholders of the acquired rms.
Evidence that the superior merger announcement performance of non-Big 4
auditors clients is driven by private acquisitions could be construed as evidence
supporting the heightened monitoring hypothesis. The benets of the heightened
monitoring that result from private acquisitions would be relatively greater for
clients of non-Big 4 auditors because these rms have lower monitoring to begin
 ARTICLE IN PRESS
92 H. Louis / Journal of Accounting and Economics 40 (2005) 7599

with. However, a signicantly negative coefcient on the interaction between BIG4

and PRIVATE is also consistent with the superior advising hypothesis. The
heightened monitoring and the superior advising hypotheses are distinguishable
because the effect of the interaction between BIG4 and PRIVATE should increase in
the proportion of the merger nanced with stock under the heightened monitoring
hypothesis and decrease in the proportion of the merger nanced with stock under
the superior advising hypothesis. In other words, b18 is expected to be negative if the
heightened monitoring hypothesis dominates and positive if the superior advising
hypothesis dominates.
The results are reported in the last column of Table 5. I nd no support for either
the heightened scrutiny or the neglected rm hypotheses. There is no evidence that
the auditor size effect is increasing in the proportion of equity used to nance the
acquisitions or in analyst following. The results also lend no support to the
heightened monitoring hypothesis. First, contrary to the prediction of the heightened
monitoring hypothesis, b15, the coefcient on the interaction between BIG4 and
STOCK is insignicantly different from zero. More importantly, the coefcient on
the interaction between BIG4, PRIVATE, and STOCK, b18, is signicantly positive,
which is inconsistent with the predictions of the heightened monitoring hypothesis.
Consistent with the superior advising hypothesis, I nd strong evidence that the
auditor size effect is particularly driven by acquisitions of privately owned
companies. The coefcient on the interaction between BIG4 and PRIVATE, b17, is

5.015, with a t-value of
3.21. This implies that, at the announcement of cash-
based acquisitions of private companies, a Big 4 auditing rms average client
experiences an incremental loss of about 5 percent relative to the performance of a
non-Big 4 rms average client. Moreover, consistent with the superior advising
hypothesis, the effect of the interaction between BIG4 and PRIVATE signicantly
decreases in the percentage of the transaction nanced with stock; b18 is 0.035 with a
t-value of 3.62. The average market value of the Big 4 (non-Big 4) accounting rms
clients that engage in acquisitions of private companies is about $1958 (80) million.
Hence, the results suggest that, in cash-based mergers, the average acquirer of a
private target experiences an incremental decrease in shareholder wealth of
approximately $98 million if audited by a Big 4 accounting rm. Similarly, relative
to the Big 4 audit rms clients, the non-Big 4 audit rms average client acquiring a
private company in a cash-based merger gains approximately an additional $4
million. However, for each incremental 1-percent of a transaction nanced with
stock, the loss of a Big 4 audit rms average client decreases by about $685,000
(totaling $68.5 million in a pure stock-for-stock transaction) and the gain of a non-
Big 4 audit rms average client decreases by about $28,000 (totaling of $2.8 million
in a pure stock-for-stock transaction).
I also nd some evidence that the superior performance of the non-Big 4 audit
rms clients is related to the likelihood that the acquirers use in-house advisers,
which is consistent with the superior clientele-advising hypothesis. The coefcient on
the interaction between BIG4 and INHOUSE, b19, is 2.35, with a t-value of 1.55.
Consistent with the superior advising hypothesis, the effect of the interaction
between BIG4 and INHOUSE signicantly decreases in the percentage of the
 ARTICLE IN PRESS
H. Louis / Journal of Accounting and Economics 40 (2005) 7599 93

transaction nanced with stock; b20 is 0.024 with a t-value of 3.53. The average size
of the acquirers audited by Big 4 (non-Big 4) accounting rms for which no
investment banker is identied on SDC is about $7202 (250) million. Hence, the
results suggest that, in cash-based mergers, those acquirers, on average, have
destroyed approximately an additional $169 million in shareholders wealth if they
are audited by Big 4 accounting rms. Similarly, relative to the Big 4 audit rms
clients, the non-Big 4 rms average client gains approximately an additional $6
million. However, for each incremental 1-percent of a transaction nanced with
stock, the loss of a Big 4 audit rms average client decreases by about $1.73 million
(for a total of $173 million in a pure stock-for-stock transaction) and the gain of a
non-Big 4 audit rms average client decreases by about $60,000 (for a total of $6
million in a pure stock-for-stock transaction).

5. Sensitivity analyses

5.1. Is the auditor choice effect an artifact of the premium effect?

Another plausible alternative explanation for the difference in the abnormal

returns of the clients of Big 4 and non-Big 4 auditors is that clients of non-Big 4
auditors are just frugal companies, saving money on the audit (hence using a smaller
audit rm), and paying smaller premiums. In that case, the auditor choice effect
would simply be an artifact of the premium effect. To test this hypothesis, I use a
restricted sample for which the merger premium over book value of equity is
reported on SDC.12 I use the premium over book value because it is available on
SDC for many acquisitions of publicly traded companies and some acquisitions of
privately held companies whereas the premium over market value is observable only
for acquisitions of publicly traded companies.
I rst model the merger premium as a function of the auditor choice, the method
of payment, the method used to account for the transaction, the public status of the
target, the likelihood that the acquirer uses in-house advisers, the acquirers book-to-
market ratio, the acquirers return-on-assets, the industry relatedness of the merging
partners, analyst coverage, the targets attitude toward the proposed merger, the
targets size, the relative size of the merging rms, and the inverse Mills ratio from
regression (2). Non-tabulated results provide no evidence that the merger premium is
associated with the auditor choice variable; the t-value for the coefcient on the
auditor choice variable is
0.08. I further restrict the sample to publicly traded
targets and re-estimate the model using the premium over market value 4 weeks prior
12
I do not use the restricted sample as the primary sample for the study because of the severe attrition
that results from requiring data on the merger premium. The attrition is particularly severe for the clients
of the non-Big 4 audit rms. For these rms, the sample is reduced almost in half from 205 to 106. For the
clients of the Big 4 rms, the sample is reduced from 3502 to 2378. I use only non-negative book premiums.
However, adding the negative premiums does not qualitatively change the results.
 ARTICLE IN PRESS
94 H. Louis / Journal of Accounting and Economics 40 (2005) 7599

to the merger announcement. I obtain qualitatively the same results; the t-value for
the coefcient on the auditor choice variable is
0.72.
I then model the merger announcement return as a function of the auditor choice,
the merger premium over book value, the method of payment, the method used to
account for the transaction, the public status of the target, the likelihood that the
acquirer uses in-house advisers, the acquirers book-to-market ratio, the acquirers
return-on-assets, the industry relatedness of the merging partners, analyst coverage,
the targets attitude toward the proposed merger, the targets size, the relative size of
the merging rms, and the inverse Mills ratio from regression (2). Non-tabulated
results indicate that the auditor size effect is robust to controlling for the merger
premium. The coefcient on the Big 4 indicator variable is
8.088 percent with a t-
value of 3.04.

5.2. Alternative event windows and size-adjusted return

Schwert (1996) nds signicant run-ups in targets prices as far back as 21 trading
days before merger announcements. Rumors about impending mergers by Big 4 and
non-Big 4 clients might reach the market at different speeds. To control for the
potential effects of leakage, I measure the abnormal return over the period from 21
days before to 1 day after the announcement. I also use two additional event
windows: [
3 +1] and [
5 +5]. The results of the study are qualitatively the same
when I use these alternative event windows. To ensure that the auditor size effect is
not an artifact of a potential size effect, I also compute size-adjusted returns using
the NYSE/AMEX/NASDAQ size-deciles portfolios and obtain qualitatively the
same results. I then compare the market reactions for the two groups of acquirers by
quintiles of the acquirers total assets. The evidence indicates that the auditor effect is
distinct from the size effect. The abnormal returns are lower for the Big 4 acquirers in
all the quintiles.

5.3. Controlling for potential contaminations

The merger announcements might be systematically associated with events that

would bias our ndings. I therefore remove from the sample merger announcements
that could be contaminated by other concurrent events. I combine the merger sample
with samples of stock repurchases reported on SDC, distributions to shareholders
(including stock splits and stock dividends) reported on CRSP, and earnings
announcements reported on Compustat. I then search the newswires (via LexisNexis
Academic Universe) for news about the non-Big 4 auditors clients and nd two
additional events: An announcement of a claim dismissal in a lawsuit settlement by
Elron Electronics Industries and the announcement of a new contract by CGI Group
Inc. In total, I identify 15 (187) merger announcements by clients of non-Big 4 (Big 4)
auditors that could be contaminated by other concurrent events. Deleting these
observations does not qualitatively change the results of the paper.
Firms that increase in size tend to switch from non-Big 4 to Big 4 auditors.
Thus, the favorable reaction to the acquisition announcements by the clients of the
 ARTICLE IN PRESS
H. Louis / Journal of Accounting and Economics 40 (2005) 7599 95

non-Big 4 auditors might be due to the market anticipating some the non-Big 4 rms
clients to switch to Big 4 auditors. To test this explanation, I delete from the sample
those rms that switch to Big 4 auditors in the year after the merger announcement. I
obtain qualitatively the same results as those reported in the paper.

5.4. Comparing clients of Big 4, national, and regional audit firms

I follow the extant literature and group the audit rms into Big 4 and non-Big 4
rms. However, some of the non-Big 4 rms have national operations while others
have only regional and local operations.13 As a sensitivity analysis, I examine the
extent to which the results vary across clients of national audit rms and clients of
regional rms. I classify the largest non-Big 4 audit rms, Main Hurdman,
Laventhol & Horwath, BDO Seidman, Grant Thornton, and McGladrey & Pullen as
national rms and the other non-Big 4 audit rms as regional rms.14 Note that the
labels national and regional are not intended to be fully descriptive as the
distinctions between national and regional and between regional and local
are not always clear and some national rms also have international operations. I
use the labels only for exposition purposes.
Un-tabulated results show no signicant difference in the abnormal returns of the
clients of the national audit rms and the clients of the regional audit rms. In
addition, clients of the Big 4 rms under-perform the clients of both the national and
the regional audit rms. Overall, the evidence suggests that the national audit rms
resemble the regional rms more than the Big 4 rms in terms of their ability and
opportunity to provide acquisition advice to their clients. There are at least four
reasons for this nding. First, the national rms are, by far, more similar to the
regional rms than to the Big 4 rms. The General Accounting Ofce (GAO), for
instance, in its July 2003 report to the Senate Committee on Banking, Housing, and
Urban Affairs and the House Committee on Financial Services on consolidation and
competition in the audit market estimates that [b]y any measure, the large public
company audit market is a tight oligopoly (p. 16). Based on data published in the
Public Accounting Report, the GAO estimates that the Big 4 rms currently audit
over 99 percent of all public company sales.15 Second, even though the national rms
13
For instance, the Panel on Audit Effectiveness (PAE), in its August 31, 2000 report, classies BDO
Seidman, Grant Thornton, and McGladrey & Pullen as a separate group. Excluding the largest audit
rms, BDO Seidman, Grant Thornton, and McGladrey & Pullen lead the market in audit revenue, staff,
and ofces. Two other rms that no longer exist, Main Hurdman and Laventhol & Horwath, could also
have been considered national. Prior to its takeover by KPMG Peat Marwick in 1987, Main Hurdman was
the largest Non-Big 4 rm. It was followed by Laventhol & Horwath that went bankrupt in 1990.
14
Prior to 1988, Compustat did not provide the names of the smaller audit rms. I use various volumes
of Who Audits America to identify the smaller audit rms from 1979 to 1987. According to the GAO
report, in 2002, Baird, Kurtz, and Dobson (BKD) and Crowe, Chizek and Co. had more total revenue
than McGladrey & Pullen although they had lower revenue from audit and attest services and fewer
ofces. As a robustness check, I also classify Baird, Kurtz, and Dobson (BKD) and Crowe, Chizek and Co
as national rms and obtain qualitatively the same results.
15
According to the GAO report, Deloitte & Touche had $5900 million in revenue in 2002 whereas Grant
Thornton, the largest among the national rms had only $400 million. The average Deloitte & Touches
 ARTICLE IN PRESS
96 H. Louis / Journal of Accounting and Economics 40 (2005) 7599

have national and international operations, unlike the Big 4 audit rms, they are
essentially associations of local rms that likely benet from the close connection to
the local communities that characterizes non-Big 4 audit rms in general.16 Third,
according to the Panel on Auditing Effectiveness (PAE), unlike the Big 4 audit rms,
the national rms have directed their services primarily to entrepreneurial
clients.17 This likely gives them more familiarity with the operations and merger
activities of private companies. Finally, the national rms generally recruit relatively
more experienced professionals than of the Big 4 rms. This is likely to make a
substantial difference in client advising in complex transactions such as mergers.18

6. Summary

I analyze the effect of auditor size on acquirers abnormal returns around merger
announcements and the factors affecting the interaction between auditor size and the
market reaction to merger announcements. I nd strong evidence that acquirers
audited by non-Big 4 accounting rms signicantly outperform those audited by Big
4 rms at merger announcements. I test various potential explanations for the
auditor size effect.
First, it is generally assumed that smaller accounting rms provide lower quality
audit. Titman and Trueman (1986) and Datar et al. (1991) also suggest that investors
infer the nature of managers private information through the managers auditor
choice. Hence, by voluntarily submitting itself to greater scrutiny through the merger
process, a client of a smaller audit rm might reveal to investors that it is of a better

(footnote continued)
ofce had $72,840 million in revenue whereas the average Grant Thornton ofce had $7843 million. The
report can be found at http://www.gao.gov/new.items/d03864.pdf. It includes an extensive discussion on
the structure of the audit market.
16
The GAO, for instance, notes in its report: The various national practices of any given Big 4 rm are
separate and independent legal entities, but they often share common resources, support systems, audit
procedures, and quality and internal control structures. Market participants said that the afliates of
smaller rms, in contrast, tended to have lower degrees of commonality (p. 48). The difference in the
nature of the audit rms is also evident when we compare the rms relative numbers of ofces and their
relative revenues. According to the GAO report, in 2002, the total revenues per ofce were $47,308 million,
$6581 million, and $10,327 million for the Big 4, the national, and the regional rms, respectively. Thus,
on average, the national rms have even smaller ofces than the regional rms.
17
See page 184 of the PAE report at http://www.pobauditpanel.org/download.html. The panel was
appointed by the Public Oversight Board at the request of the chairman of the Securities and Exchange
Commission. The report provides a good description of the audit market.
18
For instance, discussing the reasons why rms are shifting from Big 4 audit rms to national rms
following the SarbanesOxley Act, Leland Graul, director at BDO Seidman, indicates: Better access to
top accounting professionals also is a factor. Given the complex new accounting rules, many companies
want ready access to the partner in charge of the audit Krantz (2004, p. B1). Graul also argues that the
national rms have relatively more experienced professionals because the largest rms rely heavily on
recent college graduates. Edward Nusbaum, CEO of Grant Thornton, also explains the defection from the
Big 4 by the need of mid-size rms for more personalized attention from their audit partners, which, he
suggests, his rm can better provide than the largest audit rms that specialize in service to large
companies (Yoon, 2004, p. 15).
 ARTICLE IN PRESS
H. Louis / Journal of Accounting and Economics 40 (2005) 7599 97

type than they previously thought and, therefore, deserves a lower cost-of-capital.
Second, clients of non-Big 4 audit rms might outperform at merger announcements
because they are less visible and, hence, neglected. Thus, they benet more from the
exposure generated by the merger process. Third, clients of non-Big 4 audit rms
may outperform those of Big 4 rms because non-Big 4 audit rms have a
comparative advantage in advising their clients. I postulate that their superiority
should be more evident when the clients are targeting privately held rms and less
evident when the clients rely on investment banks advice. Alternatively, since private
rms have more concentrated ownership than public rms, a client of a non-Big 4
audit rm might benet more from the acquisition of a private company because of
the increased monitoring that results from the creation of new block-holdings by the
former shareholders of the private target.
I nd no support for the heightened scrutiny hypothesis, the neglected rm
hypothesis, and the heightened monitoring hypothesis. The results, however, support
the suggestion that non-Big 4 audit rms provide superior acquisition advice to their
clients. They indicate that the clients of the non-Big 4 rms outperform the clients of
the Big 4 rms signicantly more at merger announcements if the targets are
privately held companies. There is also some indication that the auditor size effect is
more pronounced when the likelihood of the auditor playing a prominent advisory
role in the merger process increases. These effects are signicantly stronger in cash
acquisitions as opposed to stock swaps and are robust to controlling for the usual
factors affecting the market reaction to merger announcements and potential
sample-selection biases. While Big 4 auditing rms are usually assumed to offer
better services than the smaller ones, the study suggests that smaller audit rms have
a comparative advantage in serving their clientele.

References

Addams, L., Davis, B., 1994. Privately held companies report reasons for selecting and switching auditors.
The CPA Journal 64, 3841.
Arbel, A., Carvell, S., Strebel, P., 1983. Giraffes, institutions and neglected rms. Financial Analysts
Journal 39, 28.
Beatty, R., 1989. Auditor reputation and the pricing of initial public offerings. The Accounting Review 64,
693709.
Becker, C., et al., 1998. The effect of audit quality on earnings management. Contemporary Accounting
Research 15, 124.
Berger, P., Ofek, E., 1995. Diversications effect on rm value. Journal of Financial Economics 37, 3965.
Berton, L., 1994. Booked up: Big accounting rms, striving to cut costs, irritate small clientsMany
companies say service has slipped badly, hurt in part by staff turnoverI got tired of being ignored.
Wall Street Journal (April 21), A1.
Boesky, I., 1985. Merger Mania: ArbitrageWall Streets Best-Kept Money-Making Secret. Holt,
Rinehart & Winston, New York.
Boone, C., Brocheler, V., Carroll, G., 2000. Custom service: application and tests of resource-partitioning
theory among Dutch auditing rms for 18961992. Organization Studies 21, 355381.
Carpenter, C., Strawser, R., 1971. Displacement of auditors when clients go public. Journal of
Accountancy 131, 5558.
 ARTICLE IN PRESS
98 H. Louis / Journal of Accounting and Economics 40 (2005) 7599

Carroll, G., Dobrev, S., Swaminathan, A., 2002. Theorie der Ressourcenteilung in der Organisationso-
kologie. Kolner Zeitschrift fur Soziologie und Sozialpsychologie 42, 381413.
Chang, S., 1998. Takeovers of privately held targets, method of payment, and bidder returns. Journal of
Finance 53, 773784.
Chow, C., 1982. The demand for external auditing: size, debt and ownership inuences. The Accounting
Review 57, 272291.
Comment, R., Jarrell, G., 1995. Corporate focus and stock returns. Journal of Financial Economics 37,
6787.
Craswell, A., Francis, J., Taylor, S., 1995. Auditor brand name reputation and industry specialization.
Journal of Accounting and Economics 20, 297322.
Datar, S., Feltham, G., Hughes, J., 1991. The role of audits and audit quality in valuing new issues.
Journal of Accounting and Economics 14, 349.
DeAngelo, L., 1981. Auditor size and audit quality. Journal of Accounting and Economics 3, 183199.
DeAngelo, H., DeAngelo, L., Rice, E., 1984. Going private: Minority freezeouts and stockholder wealth.
Journal of Law and Economics 27, 367401.
Dopuch, N., Simunic, D., 1980. The nature of competition in the auditing profession: a descriptive and
normative view. In: Buckley, J., Weston, F. (Eds.), Regulation and the Accounting Profession.
Lifetime Learning Publication, Belmont, CA.
Easley, D., OHara, M., 2002. Information and the cost of capital. Working Paper, Cornell University.
Ettredge, M., Simon, D., Smith, D., Stone, M., 1994. Why do companies purchase timely quarterly
reviews? Journal of Accounting and Economics 19, 131155.
Fuller, K., Netter, J., Stegemoller, M., 2002. What do returns to acquiring rms tell us? Evidence from
rms that make many acquisitions. Journal of Finance 57, 17631793.
Gorman, J., 1988. M&As: the CPA as acquisition adviser. Journal of Accountancy 166, 3640.
Heckman, J., 1979. Sample selection bias as specication error. Econometrica 47, 153161.
Hilzenrath, D., 2003. Big rms avoiding SEC ire: enforcement record lopsided. The Washington Post
(January 17), E01.
Jensen, M., Ruback, R., 1983. The market for corporate control: the scientic evidence. Journal of
Financial Economics 11, 550.
John, K., Ofek, E., 1995. Asset sales and increase in focus. Journal of Financial Economics 37, 105126.
Johnson, W., Lys, T., 1990. The market for audit services: evidence from voluntary auditor changes.
Journal of Accounting and Economics 36, 281308.
Krantz, M., 2004. More rms ee Big Four accountants. USA Today 1 (September 27), 1B.
Lang, L., Stulz, R., 1994. Tobins q, corporate diversication and rm performance. Journal of Political
Economy 102, 12481280.
Lang, L., Stulz, R., Walking, R., 1989. Managerial performance, Tobins q and the gain from successful
tender offers. Journal of Financial Economics 24, 137154.
Larcker, D., Lys, T., 1987. An empirical analysis of the incentives to engage in costly information
acquisition: the case of risk arbitrage. Journal of Financial Economics 16, 111126.
Leland, H., Pyle, D., 1977. Information asymmetries, nancial structure and nancial intermediation.
Journal of Finance 32, 371387.
Louis, H., 2004a. Earnings management and the market performance of acquiring rms. Journal of
Financial Economics 74, 121148.
Louis, H., 2004b. The cost of using bank mergers as defensive mechanisms against takeover threats.
Journal of Business 77, 295310.
Lys, T., Vincent, L., 1995. An analysis of value destruction in AT&Ts acquisition of NCR. Journal of
Financial Economics 39, 353378.
Lys, T., Watts, R., 1994. Lawsuits against auditors. Journal of Accounting Research 32, 6593.
Merton, R., 1987. Presidential address: a simple model of capital market equilibrium with incomplete
information. Journal of Finance 42, 483510.
Moeller, S., Schlingemann, F., Stulz, R., 2005. Wealth destruction on a massive scale? A study of
acquiring-rm returns in the recent merger wave. Journal of Finance 60, 757782.
 ARTICLE IN PRESS
H. Louis / Journal of Accounting and Economics 40 (2005) 7599 99

Morck, R., Shleifer, A., Vishny, R., 1990. Do managerial objectives drive bad acquisitions? The Journal of
Finance 45, 3148.
Myers, S., Majluf, N., 1984. Corporate nancing and investment decisions when rms have information
that investors do not have. Journal of Financial Economics 12, 187222.
Nagelkerke, N., 1991. A note on general denition of the coefcient of determination. Biometrika 78,
691692.
Palmrose, Z., 1988. An analysis of auditor litigation and audit service quality. The Accounting Review 63,
5573.
Pandit, S., 2003. Accounting choice and operating performance in stock-for-stock mergers. Working
Paper, University of Rochester.
Peale, C., 1994. Small businesses like service, price at small CPA rms. Canadian Business Conditions
(June 6), 1.
Rau, P., Vermaelen, T., 1998. Glamour, value and the post-acquisition performance of acquiring rms.
Journal of Financial Economics 49, 223253.
Scanlon, K., Trifts, J., Pettway, R., 1989. Impact of relative size and industrial relatedness on returns to
shareholders of acquiring rms. The Journal of Financial Research 12, 103112.
Schwert, G., 1996. Markup pricing in mergers and acquisitions. Journal of Financial Economics 41,
153192.
Servaes, H., 1991. Tobins Q and gains from takeovers. The Journal of Finance 54, 547580.
Servaes, H., Zenner, M., 1996. The role of investment banks in acquisitions. The Review of Financial
Studies 9, 787815.
Shavell, S., 1994. Acquisition and disclosure of information prior to sale. The Rand Journal of Economics
25, 2036.
St. Pierre, K., Anderson, J., 1984. An analysis of the factors associated with lawsuits against public
accountants. The Accounting Review 59, 242263.
Stoughton, N., 1988. The information content of corporate merger and acquisition offers. Journal of
Financial and Quantitative Analysis 23, 175197.
Teoh, S., Wong, T., 1993. Perceived auditor quality and the earnings response coefcient. The Accounting
Review 68, 346367.
Titman, S., Trueman, B., 1986. Information quality and the valuation of new issues. Journal of
Accounting and Economics 8, 159172.
Travlos, N., 1987. Corporate takeover bids, method of payment, and bidding rms stock returns. The
Journal of Finance 42, 943963.
Who Audits America, 719th ed. Data Financial Press, Menlo Park, CA.
Wyser-Pratte, G., 1982. Risk Arbitrage II, Monograph Series in Finance and Economics, vols. 34. New
York University, New York.
Yoon, L., 2004. Bolting from the big four. CFO Magazine (April 1), 15.