AUDITING: A JOURNAL OF PRACTICE & THEORY American Accounting Association

Vol. 36, No. 2 DOI: 10.2308/ajpt-51577

May 2017
pp. 45–62

Board Independence and Internal Control Weakness:

Evidence from SOX 404 Disclosures
Yangyang Chen
The Hong Kong Polytechnic University

W. Robert Knechel
University of Florida

Vijaya Bhaskar Marisetty

RMIT University
BITS Pilani

Cameron Truong
Monash University

Madhu Veeraraghavan
T. A. Pai Management Institute
SUMMARY: In this paper, we investigate whether board independence has an impact on the likelihood that a
company reports weaknesses in internal controls. Using a sample of 11,226 firm-year observations spanning the
period 2004–2012, we establish several findings. First, we document a negative relation between board
independence and the disclosure of internal control weaknesses. We also document that the negative relation is
stronger for firms with unitary leadership (combined positions of CEO and chairman) than for firms with dual
leadership. Next, we show that board independence is associated with both fewer account-specific and company-
level weaknesses. Finally, we show that board independence is associated with timely remediation of internal control
weaknesses and that the implementation of Auditing Standard No. 5 in 2007 weakens the effect of board
independence on the disclosure of ICW.
Keywords: internal control weakness; board independence; unitary versus dual leadership; SOX 404.

JEL Classifications: G10; G18.

INTRODUCTION

T
he corporate governance structure of an organization can be quite complex. Since one of the purposes of
corporate governance is to manage various aspects of risk, different characteristics of the governance structure
may serve as either substitutes or complements in reducing risk (Knechel and Willekens 2006). Management is
responsible for designing and implementing a system of internal control within an organization. In this study, we
examine how the structure of a firm’s corporate governance influences how well management performs this task. The
disclosure of a material weakness in internal control over financial reporting (ICOFR) is an explicit acknowledgement by
management (and the auditor) that the internal control system may not adequately prevent material errors in the firm’s

We are especially grateful to Rani Hoitash (editor) and two anonymous referees for the insightful comments and suggestions that have significantly
improved the paper.
Editor’s note: Accepted by Rani Hoitash.
Submitted: June 2014
Accepted: August 2016
Published Online: September 2016
45
46 Chen, Knechel, Marisetty, Truong, and Veeraraghavan

financial report.1 While prior research has broadly examined the determinants of material weaknesses in ICOFR
(Krishnan 2005; Doyle, Ge, and McVay 2007; Ashbaugh-Skaife, Collins, and Kinney 2007), less attention has been paid
to whether corporate governance plays a significant role in a firm’s internal control system (Goh 2009). Given the
general belief that board independence can provide effective monitoring and control over a firm’s activities, this study
investigates whether two important aspects of board structure are associated with the quality of a firm’s internal control
over financial reporting.
More specifically, in this paper, we examine how board independence and a unitary CEO/chairman impact the disclosure
of internal control weaknesses (ICWs) within a company. Bhagat and Black (1999) observe that most large U.S. public
companies have boards with a majority of independent directors, which reflects the view that the board’s central task is to
monitor management. The Securities and Exchange Commission (SEC), New York Stock Exchange (NYSE), and National
Association of Securities Dealers Automated Quotations (NASDAQ) have long espoused independence as a desirable and
required board attribute. In addition, the NYSE and NASDAQ mandate that only independent directors should serve on the
audit and compensation committees.2 Prior research strongly supports the view that independent boards provide superior
oversight that can complement management’s fundamental responsibility for designing and implementing a system of internal
control. Hence, we hypothesize that board independence should lead to a reduced likelihood that a company will have
weaknesses in internal control over financial reporting. In addition, we investigate whether the effect of board independence on
the disclosure of ICWs is influenced by whether a firm has unitary leadership (i.e., the titles of CEO and chairman vested in one
individual).
Using a comprehensive sample of 11,226 firm-year observations spanning the period 2004–2012, we document that board
independence is negatively related to the disclosure of ICWs. We also find that the effect of board independence on ICWs is
stronger for firms with unitary leadership than for firms with dual leadership.3 In additional analysis, we show that board
independence has an effect on both account-specific ICWs and company-level ICWs. We also show that board independence is
associated with timely remediation of ICWs and that the implementation of Auditing Standard No. 5 in 2007 somewhat
weakens the effect of board independence on the disclosure of ICWs.
Our paper contributes to the growing literature on corporate governance and control, board independence, and internal
control reporting. A number of studies have examined the relation between corporate governance and material weaknesses
(Krishnan 2005; Zhang, J. Zhou, and N. Zhou 2007; Goh 2009; Johnstone, Li, and Rupley 2011). We contribute to this stream
of literature by documenting the effect of board independence on the disclosure of ICWs under the Sarbanes-Oxley Act Section
404 (SOX 404) using a large sample over an extended period of time. Further, with the exception of Johnstone et al. (2011),
none of the prior studies have examined how variations in the structure of top management—unitary versus dual CEO and
chairman—influence the relation between board monitoring and material weaknesses in internal control over financial
reporting. Johnstone et al. (2011) hypothesize that disclosure of material weaknesses is positively associated with turnover of
top management and members of the board of directors and audit committee. Their argument is that material weaknesses are
indicative of poor performance, which will lead to increased turnover in senior management and governance-related positions
within a firm. This observation supports the thesis developed in this paper. Since prior research in corporate finance shows that
unitary leadership is undesirable, our investigation of the unitary/dual question extends our understanding of how internal
control relates to a comprehensive system of corporate monitoring and control.
We also make at least two important contributions to the literature on internal control reporting in the auditing arena. First,
we observe that the change in auditing standards related to internal control reporting is associated with a change in the
relationship between board independence and the disclosure of control weaknesses. Specifically, we find that the association
between board independence and ICWs is weaker after the issue of Auditing Standard No. 5 (AS 5). This result could suggest a
general improvement in internal control over time, or it could suggest that boards receive less detailed information about
internal control after AS 5 since it was designed to make the evaluation of internal control more holistic, top down, and risk
based. Second, we examine whether board independence is also associated with timely remediation of ICWs. In particular, we
extend the analysis of Goh (2009) to a significantly longer post-implementation time period (i.e., 2004–2012 rather than the

1
Essentially, firms’ internal control system should prevent or detect the errors in their financial statements (PCAOB 2007). Firms disclosing material
weaknesses in internal control must indicate that their internal control system is not effective, which suggests that management is not completely
fulfilling its responsibilities.
2
Both the NYSE and NASDAQ have regulations requiring certain board members to be independent, usually defined as an absence of any material
financial relationship with a company. Examples of situations that would violate board independence rules include serving as an officer or employee,
providing significant services outside the role as a board member, or entering into significant transactions with the company. Specific rules can be
found on the NYSE and NASDAQ websites.
3
We follow Brickley, Coles, and Jarrell (1997) in defining unitary and dual leadership. Unitary leadership is the case when the CEO and the chairman of
the board titles are vested in one individual and dual leadership is the case where the two positions are held by different individuals.

Auditing: A Journal of Practice & Theory

Volume 36, Number 2, 2017
Board Independence and Internal Control Weakness: Evidence from SOX 404 Disclosures 47

window of initial implementation of SOX). Specifically, we show that a higher proportion of independent directors on the
board is associated with faster remediation of control weaknesses.
The remainder of the paper is organized as follows. The second section presents prior literature and develops the testable
hypotheses. The third section describes the data and variables. Empirical results are presented in the fourth section and further
analysis is presented in the fifth section. The sixth section presents a supplementary analysis and the seventh section concludes.

RELATED LITERATURE AND HYPOTHESES DEVELOPMENT

Board Independence and the Incidence of Internal Control Weakness

Research on internal control has intensified since the introduction of SOX 404 in 2002. Numerous attributes of a firm have
been shown to be associated with reported ICWs. For example, Ashbaugh-Skaife et al. (2007) find that firms disclosing ICWs
have more complex operations, recent organizational changes, greater accounting risk, more auditor resignations, and fewer
resources available for internal control. Doyle et al. (2007) show that ICWs are associated with firm size, firm age, financial
health, financial reporting complexity, rapid growth, restructuring charges, and corporate governance. They also document that
the determinants of ICWs vary depending on the type of ICW reported.
Internal control weaknesses have also been shown to be associated with indications of declining performance within a firm
(DeFond and Jiambalvo 1991). For example, Krishnan (2005) shows that firms with internal control problems also are more
likely to report financial losses.4 Ashbaugh-Skaife, Collins, Kinney, and LaFond (2008) document that firms reporting an ICW
have lower-quality accruals relative to firms that do not disclose an ICW. Feng, Li, and McVay (2009) suggest that firms with
weak internal control release less accurate management guidance. Li et al. (2011) find that firms with ICWs have less qualified
CFOs and higher CFO turnover. Kim, Song, and Zhang (2011) report that the loan spread is 28 basis points higher for ICW
firms than for non-ICW firms. Finally, Zhang et al. (2007) shows that firms that disclose internal control deficiencies have less
financial expertise on the audit committee. All of these results can be symptoms or indications of poor corporate and managerial
performance.
On the other hand, prior research also shows that board independence plays a positive role in firm performance and reduces
the likelihood of corporate fraud. For instance, Weisbach (1988) attributes the relation between higher firm value and board
independence to better monitoring. Beasley (1996) finds that the inclusion of an outside director reduces the probability of
financial statement fraud. Uzun, Szewczyk, and Varma (2004) demonstrate that the likelihood of corporate fraud decreases with
an increase in board independence. More generally, strong boards are negatively associated with earnings management,
restatements, and fraud, and are positively associated with audit effort and earnings quality (Dechow, Sloan, and Sweeney
1996; Beasley, Carcello, Hermanson, and Lapides 2000; Carcello, Hermanson, Neal, and Riley 2002; Klein 2002; Bédard,
Chtourou, and Courteau 2004). In short, high-quality corporate governance tends to offset many potential corporate problems.
The question addressed in this paper is whether better corporate governance (board independence) counterbalances potential
problems in internal control over financial reporting.
More specifically, Doyle et al. (2007) hypothesize that corporate governance plays a significant role in the quality of a
firm’s internal control. They measure corporate governance using the governance score developed by Brown and Caylor
(2006), a composite measure of 51 factors covering eight governance categories (audit, board of directors, charter/by laws,
director education, executive and director compensation, ownership, progressive practices, and state of incorporation). Doyle et
al. (2007) do not find a significant relation between material weakness disclosures and the corporate governance score. Our
study builds on Doyle et al. (2007), but the principal difference is that we focus on one critical dimension of corporate
governance, i.e., board independence. Our use of a single measure follows the strand of literature that considers board
characteristics as key determinants of corporate governance. For instance, Hermalin and Weisbach (1998, 2003) and Bhagat
and Black (2002) advance board independence as an important determinant of good corporate governance, while Brickley et al.
(1997) suggest that separating the chairman and CEO titles will reduce agency costs and improve performance. Bhagat and
Bolton (2008, 258) argue that:
Corporate boards have the power to make, or at least ratify, all important decisions including decisions about
investment policy, management compensation policy, and board governance itself. It is plausible that board members
with appropriate stock ownership will have the incentive to provide effective monitoring and oversight of important
corporate decisions noted above; hence board independence or ownership can be a good proxy for overall good
governance.

4
Krishnan (2005) also finds that prior to SOX, firms with more effective audit committees report fewer internal control problems in their 8-Ks when
reporting an auditor change.

Auditing: A Journal of Practice & Theory

Volume 36, Number 2, 2017
48 Chen, Knechel, Marisetty, Truong, and Veeraraghavan

Adding to these arguments, Larcker, Richardson, and Tuna (2007) note that results from prior studies using multiple
measures of corporate governance are mixed and inconclusive because these measures exhibit only a modest level of reliability
and construct validity. Bhagat and Bolton (2008) argue that on both economic and econometric grounds it is possible for a
single characteristic to be an effective measure of corporate governance, and they conclude that board independence is a good,
standalone measure of corporate governance. A number of recent studies offer strong support for this view. For example, Goh
(2009) documents that firms with more independent boards remediated material weaknesses more quickly around the time of
SOX, while Johnstone et al. (2011) document a positive association between disclosure of ICWs and subsequent turnover in
corporate boards, audit committees, and top management. Further, Johnstone et al. (2011) find that the remediation of ICWs is
associated with improvements in the operational effectiveness of boards and audit committees. Overall, these results indicate
that poor internal control can lead to changes in an organization, which lead to improved performance.

Hypotheses
Extant research has examined some aspects of the association between some board characteristics and the disclosure and
remediation of ICWs. For example, Goh (2009) examines whether the effectiveness of the board of directors is associated with
firms’ timeliness in the remediation of an ICW at the time that ICOFR reporting was first implemented. He documents that
firms with more independent boards, large audit committees, and audit committees with greater nonaccounting financial
expertise are more likely to remediate ICWs in a timely manner. U. Hoitash, R. Hoitash, and Bedard (2009) also focus on board
characteristics and find that higher-quality corporate governance is associated with more effective internal controls. In
particular, they find that a lower likelihood of disclosing a material weakness is associated with relatively more audit committee
members having accounting and supervisory experience.
Our study focuses on the direct relationship between board independence and the disclosure of internal control weaknesses.
Our main argument is that independent directors require a transparent information environment to effectively monitor and
advise management, which can best be achieved through the use of an effective internal reporting and control system that
provides the information that will be included in financial reports. This implies that there could be a negative relationship
between board independence and the disclosure of ICWs, as boards demand better information for monitoring purposes. In a
related paper, Johnstone et al. (2011) find that remediation of weaknesses in internal control occurs in conjunction with
improvements in boards, audit committees, and top management. This result suggests that an effective board, i.e., one that is a
highly independent monitor of management, will have a stronger effect on the quality of the internal control, as they demand
increased transparency from the internal reporting and control systems. This leads to our first hypothesis:
H1: Board independence is negatively associated with the disclosure of ICWs.
Our second hypothesis recognizes that not all boards are the same, even when they are independent, and considers the role
of unitary versus dual leadership styles on the association between board independence and the disclosure of ICW. It is often
argued that the CEO and chairman positions should be held by different individuals in a strong governance environment. For
instance, Jensen (1993, 866) states:
The function of the chairman is to run board meetings and oversee the process of hiring, firing, evaluating, and
compensating the CEO. Clearly the CEO cannot perform this function apart from his or her personal interest. Without
the direction of an independent leader, it is much more difficult for the board to perform its critical function. Therefore,
for the board to be effective, it is important to separate the CEO and chairman positions. The independent chairman
should, at a minimum, be given the rights to initiate board appointments, board committee assignments, and ( joint
with the CEO) the setting of the board’s agenda.
Conyon and Peck (1998) state that unitary leadership (where the CEO and the chairman of the board titles are vested in one
individual) can provide the CEO with a wider power base and locus of control, but they go on to argue that the two roles should
be separated because an independent board chair will facilitate objective assessment of the performance of the CEO and top
management. Finkelstein and Hambrick (1996) also observe that unitary leadership is a clear indicator of the CEO’s power
over the board.
A number of papers have empirically examined the use of dual titles in corporate governance with somewhat mixed results.
Consistent with the traditional view that separating the CEO and chairman positions is desirable, Goyal and Park (2002) find
that the sensitivity of CEO turnover to performance is lower when titles are combined, which is consistent with the view that a
unitary structure increases the CEO’s power over the board when the individual also serves as chairman. In a similar vein,
Adams, Hermalin, and Weisbach (2010) find that an individual in a unitary leadership position holds greater influence over
corporate decision making. However, they also note that although combining the titles would mean that an individual has more
influence over the firm, it does not necessarily mean that separating the titles will lead to improved performance. In the end,

Auditing: A Journal of Practice & Theory

Volume 36, Number 2, 2017
Board Independence and Internal Control Weakness: Evidence from SOX 404 Disclosures 49

they conclude that policymakers should be wary of calls for prohibiting the CEO serving as the chairman of the board. This
view is supported by Brickley et al. (1997), who find that the costs of separation are larger than the benefits of separation for
most firms, and that combining the titles of CEO and chairman can be efficient and consistent with shareholder interests.
Thus, the question as to whether a board can be more effective if the chairman and CEO positions are held by different
individuals is unsettled. The connection between board structure and the quality of internal control is even less obvious. If a unitary
structure results in higher agency costs, then management may have less inclination to foster a strong internal control system because
such an environment may inhibit its decision latitude and increase accountability. On the other hand, if a unitary structure increases
operating efficiency and effectiveness, then management may encourage a strong internal control system in order to better manage
internal operations and decision making. Further, the indirect link between ICOFR and the board structure implies there may be
intervening factors that will influence the nature of their association. As a result, we frame our second hypothesis in a null form:
H2: Board leadership style does not have any impact on the relation between board independence and the disclosure of
ICWs.

DATA AND VARIABLES

We obtain the ICW data for this study from Audit Analytics. Under SOX 404, accelerated filers are required to file both a
management report and an auditor’s attestation on internal controls over financial reporting beginning with annual reports filed
after November 15, 2004.5 Audit Analytics extracts the ICW data primarily from the firm’s 10-K, 10-K/A, 20-F, and 40-F
forms, which cover all SEC registrants that have disclosed their assessments of internal controls over financial reporting in
electronic filings. We obtain board information from RiskMetrics. The database collects information about the board of
directors such as name, age, gender, and affiliation for S&P 1500 firms. The database classifies each director into three
categories: employee, independent, and linked. We define a director as independent if he/she is classified by RiskMetrics as
independent. Further, we obtain financial information from Compustat, stock return information from CRSP, and M&A
information from SDC Platinum. Since the ICW data start from 2004, we restrict our sample period to 2004–2012. Our final
sample consists of 2,048 unique firms with 11,226 firm-year observations.
We measure the disclosure of ICWs using an indicator variable (INCICW), which is equal to 1 if the firm discloses ICWs
in the current year, and 0 otherwise. Board independence (INDEP) is measured as the fraction of independent directors on the
board. The selection of control variables follows Ge and McVay (2005), Ashbaugh-Skaife et al. (2007), Doyle et al. (2007),
and Ogneva, Subramanyam, and Raghunandan (2007). Specifically, board size (LDIR) is measured as the natural log of the
number of directors on the board. It is included as prior literature (e.g., Yermack 1996) shows that the size of the board is
related to the effectiveness of board monitoring. We also include two variables to control for the complexity of the firm’s
business operations. Business segments (LSEG) is measured as the natural log of the number of the firm’s business segments.
Foreign currency translation (FRGN) is an indicator variable equal to 1 if the firm has foreign currency translation in the current
year, and 0 otherwise. Since organizational change within the firm might also affect the disclosure of ICWs, we include M&A
activity (M&A), an indicator variable equal to 1 if the firm was involved in mergers or acquisitions over the three-year period
including the current year, and 0 otherwise, and restructure (RESTR), an indicator variable equal to 1 if the firm is involved in
restructuring over the three-year period including the current year, and 0 otherwise.
Further, we control for a firm’s accounting and measurement risk using top sales growth (QSGRW), defined as an indicator
variable equal to 1 if the firm’s industry-adjusted sales growth falls into the top quintile, and 0 otherwise, and inventory (INV),
calculated as the ratio of inventories over the book value of assets. Last, we control for the firm’s resource constraints in building
an effective internal control system using three variables. Firm size (LMVE) is measured by the natural log of the market value of
equity. Operating loss (LOSS) is defined as an indicator variable equal to 1 if the firm’s operating earnings are negative, and 0
otherwise. Reverse Z-score (RDAZ) is the decreasing decile rank of the firm’s Altman Z-score. We further include firm age
(LAGE), defined as the natural log of the number of years the firm exists in the CRSP database, since Doyle et al. (2007) suggest
that older firms tend to have better internal control systems. Detailed variable definitions are available in Appendix A.
Panel A of Table 1 presents the industry distribution of ICWs. The panel shows that the ‘‘Other’’ group has the highest
percentage of ICWs, while hotels and personal services have the lowest. We also report that 8.7 percent of observations in the

5
Non-accelerated filers are firms with market capitalization less than $75 million. These firms are not required to comply with Section 404 reporting
provisions until their fiscal year ending on or after July 15, 2007. The Dodd-Frank Act exempts the non-accelerated filers from the requirements to file
auditor attestation reports on internal control over financial reporting. More specifically, SOX Section 404(b), which requires the firms’ auditor to attest
and to report on management’s assessment, applies only to accelerated and large accelerated filers. This exemption significantly reduces the costs of
being public companies. For accelerated filers, Section 404 became effective for fiscal year-end after November 15, 2004. Accelerated filers include
companies that have an aggregate market value of at least $75 million as of the end of their most recently completed second quarter.

Auditing: A Journal of Practice & Theory

Volume 36, Number 2, 2017
50 Chen, Knechel, Marisetty, Truong, and Veeraraghavan

TABLE 1
Descriptive Statistics

Panel A: Industry Distribution of Internal Control Weakness

Internal Control Weakness
No Yes
Industry Obs. Percent Obs. Percent Total
Agriculture 21 91.30 2 8.70 23
Mining 450 97.83 10 2.17 460
Construction 173 98.86 2 1.14 175
Food and Tobacco 320 97.26 9 2.74 329
Textile and Apparel 124 95.38 6 4.62 130
Lumber, Furniture, Paper, and Print 401 96.39 15 3.61 416
Chemicals 731 96.57 26 3.43 757
Petroleum, Rubber, and Plastic 190 97.44 5 2.56 195
Leather, Stone, and Glass 109 99.09 1 0.91 110
Primary and Fabricated Metals 302 94.38 18 5.63 320
Machinery 1,461 96.37 55 3.63 1,516
Transport Equipment 269 96.42 10 3.58 279
Instruments and Miscellaneous Manufacturing 715 96.49 26 3.51 741
Transport, Communications, and Utilities 1,139 96.69 39 3.31 1,178
Wholesale Trade 364 96.81 12 3.19 376
Retail Trade 869 95.60 40 4.40 909
Finance, Insurance, and Real Estate 2,001 97.42 53 2.58 2,054
Hotels and Personal Services 20 100.00 0 0.00 20
Services 1,549 94.74 86 5.26 1,635
Other 18 90.00 2 10.00 20
Total 11,226 96.42 417 3.58 11,643

Panel B: Summary Statistics

Mean S.D. 25 Percent Median 75 Percent
INCICW 0.036 0.186 0.000 0.000 0.000
INDEP 0.762 0.125 0.667 0.778 0.875
NUMDIR 9.379 2.428 8.000 9.000 11.000
LDIR 2.206 0.254 2.079 2.197 2.398
NUMSEG 5.618 5.334 1.000 3.000 9.000
LSEG 1.245 1.015 0.000 1.099 2.197
FRGN 0.639 0.480 0.000 1.000 1.000
M&A 0.449 0.497 0.000 0.000 1.000
RESTR 0.492 0.500 0.000 0.000 1.000
SGRW 0.070 0.172 0.008 0.071 0.152
INV 0.095 0.119 0.004 0.051 0.141
MVE ($M) 8,178 19,157 821 2,062 6,098
LMVE 7.782 1.482 6.711 7.631 8.716
LOSS 0.127 0.332 0.000 0.000 0.000
AZ 1.771 1.251 0.787 1.720 2.589
RDAZ 5.500 2.873 3.000 5.000 8.000
AGE 26.738 19.195 13.000 20.500 37.000
LAGE 3.028 0.752 2.565 3.020 3.611
Obs. 11,226
This table presents the industry-wide distribution of internal control weaknesses identified under SOX 404 and the summary statistics of the variables used
in the analysis.
Variable definitions are presented in Appendix A.

Auditing: A Journal of Practice & Theory

Volume 36, Number 2, 2017
Board Independence and Internal Control Weakness: Evidence from SOX 404 Disclosures 51

TABLE 2
Correlation Matrix

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)
(1) INCICW 1.000

(2) INDEP 0.061 1.000

(0.000)
(3) LDIR 0.050 0.137 1.000
(0.000) (0.000)
(4) LSEG 0.034 0.005 0.012 1.000
(0.000) (0.606) (0.191)
(5) FRGN 0.017 0.090 0.002 0.102 1.000
(0.068) (0.000) (0.804) (0.000)
(6) M&A 0.016 0.040 0.038 0.227 0.108 1.000
(0.094) (0.000) (0.000) (0.000) (0.000)
(7) RESTR 0.027 0.164 0.070 0.108 0.353 0.083 1.000
(0.004) (0.000) (0.000) (0.000) (0.000) (0.000)
(8) QSGRW 0.002 0.039 0.102 0.020 0.008 0.060 0.091 1.000
(0.843) (0.000) (0.000) (0.034) (0.411) (0.000) (0.000)
(9) INV 0.020 0.029 0.075 0.051 0.101 0.057 0.054 0.005 1.000
(0.037) (0.002) (0.000) (0.000) (0.000) (0.000) (0.000) (0.569)
(10) LMVE 0.099 0.150 0.470 0.031 0.162 0.115 0.068 0.012 0.125 1.000
(0.000) (0.000) (0.000) (0.001) (0.000) (0.000) (0.000) (0.210) (0.000)
(11) LOSS 0.070 0.021 0.061 0.003 0.014 0.021 0.121 0.073 0.046 0.262 1.000
(0.000) (0.022) (0.000) (0.761) (0.133) (0.027) (0.000) (0.000) (0.000) (0.000)
(12) RDAZ 0.026 0.035 0.176 0.113 0.226 0.011 0.063 0.001 0.441 0.037 0.220 1.000
(0.006) (0.000) (0.000) (0.000) (0.000) (0.257) (0.000) (0.888) (0.000) (0.000) (0.000)
(13) LAGE 0.035 0.181 0.310 0.130 0.059 0.069 0.100 0.168 0.047 0.267 0.024 0.033 1.000
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.009) (0.000)
This table presents the correlation matrix of the variables used in the analysis. Numbers in parentheses are the significance level of the correlation
coefficient.
Variable definitions are presented in Appendix A.

agriculture industry report ICWs. Industries with a high proportion of ICWs also include textiles and apparel, retail trade, and
primary and fabricated metals.
Panel B of Table 1 presents the summary statistics of the variables used in our analysis. The panel shows that 3.6 percent of
the observations in our sample report an ICW.6 The average board has 9.379 directors, and board independence is 76.2 percent
over our sample period. On average, each firm has 5.618 business segments, and 63.9 percent of the firms have foreign
currency translation. Of the total sample, 44.9 percent and 49.2 percent of firms are involved in M&A and restructuring in the
three-year period including the current year, respectively. The sample firms have an average sales growth rate of 7 percent, and
average inventory of 9.5 percent. 12.7 percent of them report an operating loss. Finally, the average market capitalization is
$8,178 million, and the average firm age is 26.738 years.
The correlations are reported in Table 2. The first column shows that the indicator variable representing the disclosure of an
ICW is negatively correlated with the proportion of independent directors (p , 0.01), suggesting that higher board
independence is associated with a lower probability of ICWs. This result is consistent with our first hypothesis. Further, the
indicator variable is positively correlated with business segments, M&A activity, restructurings, top growth rate, inventory,
operating losses, and reverse Altman Z-score, while negatively correlated with board size, firm size, and firm age.

6
Our 3.6 percent ICWs is comparable to other studies in the literature. For example, Floyd (2016) finds 4.13 percent ICWs in a sample of 12,952 firm-
year observations.

Auditing: A Journal of Practice & Theory

Volume 36, Number 2, 2017
52 Chen, Knechel, Marisetty, Truong, and Veeraraghavan

RESULTS

Univariate Analysis
In this section, we report the results of the univariate analysis. We divide the sample into deciles based on the fraction of
independent directors and examine how the disclosure of ICWs changes across the deciles. Decile 1 includes firms with the
lowest fraction of independent directors on the board, while Decile 10 includes those with the highest fraction. The results of
the univariate analysis are presented in Panel A of Table 3. The table shows that 6.95 percent of observations in Decile 1 report
an ICW. When we move to Decile 2, the fraction drops to 4.81 percent. The fraction drops further when we move from decile
to decile. While the pattern is not strictly monotonic, we can clearly observe the downward trend as we move from a low level
of independence to a high level of independence. A comparison between Deciles 1 and 10 reveals that the probability of
reporting ICWs is about 75 percent lower for firms with the highest level of board independence than for firms with the lowest
level of board independence. The difference is statistically significant at the 1 percent level. Overall, the univariate analysis
shows a negative relation between board independence and ICWs, which is consistent with our first hypothesis.

Regression Analysis
In this section, we estimate the following regression model for the relation between board independence and the disclosure
of an ICW:
DUMICWi;t ¼ a þ b  FIDIRi;t þ c1  LDIRi;t þ c2  LSEGi;t þ c3  FRGNi;t þ c4  M&Ai;t þ c5  RESTRi;t þ c6
 QSGRWi;t þ c7  INVi;t þ c8  LMVEi;t þ c9  LOSSi;t þ c10  RDAZi;t þ c11  LAGEi;t þ Indj;t þ ei;t
ð1Þ
where i denotes firm, t denotes year, j denotes industry, Ind is industry fixed-effects—defined at the two-digit SIC code level,
and e is the error term. We predict a negative sign for b, which indicates a negative association between board independence
and the disclosure of ICWs. The regression results are presented in Panel B of Table 3.
Column (1) shows that board independence is negatively and significantly related to ICWs after controlling for industry
fixed-effects, suggesting that higher board independence leads to a lower probability of ICWs. In Column (2), we add the
natural log of the number of directors (a measure of board size) as a control variable, and in Column (3) we add a number of
firm characteristics as controls. The coefficient for board independence retains its statistical significance in both the models. The
coefficient of board independence in Column (3) shows that a one-standard-deviation increase in board independence results in
a 0.88 percent drop in the probability of ICWs. Since the unconditional probability of ICWs is 3.6 percent, this constitutes a
24.31 percent drop in the probability. Therefore, the effect of board independence on the probability of ICWs is economically
significant. These results support H1.
The results for control variables are consistent with prior literature (e.g., Ogneva et al. 2007). We find that the disclosure of
ICWs is positively and significantly associated with business segments, foreign currency translation, and inventory, suggesting
that firms with more complex operating environments are more likely to report an ICW. Further, the coefficient for firm size is
negative and significant. These findings indicate that small firms have a higher probability of reporting an ICW. The disclosure
of ICWs is also positively associated with the reverse Z-score, showing that firms with a higher probability of bankruptcy are
more likely to report an ICW. The coefficients for restructuring and operating losses are both positive and significant. Finally,
the coefficients for board size, M&A activity, top sales growth, and firm age are mostly insignificant.

The Role of Unitary and Dual Leadership Styles

In this section, we analyze the role of unitary leadership on the relation between board independence and the incidence of
internal control problems. We follow Brickley et al. (1997) in defining unitary and dual leadership, i.e., unitary leadership is the
case when the CEO and the chairman of the board titles are vested in one individual and dual leadership is the case where the
two positions are held by different individuals. We create CEO unitary (CEOUNI), an indicator variable equal to 1 if the firm
has unitary leadership, and 0 if the firm has dual leadership. Approximately 62 percent of our sample has a unitary board. The
results for unitary and dual leadership styles are presented in Table 4.
In Column (1) of Table 4, we include CEO unitary in the regression specification in Equation (1). We find that the
coefficient of board independence remains negative and statistically significant, while the coefficient of CEO unitary is
statistically insignificant. The results suggest that unitary board leadership does not have a direct impact on the disclosure of
ICWs after controlling for board independence. In Column (2), we interact CEO unitary with board independence. We find that
the coefficient of the interaction term is negative and statistically significant—and the coefficient for the independence variable
when there is a dual board is insignificant (INDEP)—suggesting that the effect of board independence on the disclosure of

Auditing: A Journal of Practice & Theory

Volume 36, Number 2, 2017
Board Independence and Internal Control Weakness: Evidence from SOX 404 Disclosures 53

TABLE 3
Board Independence and Internal Control Weakness

Panel A: Univariate Analysis

Internal Control Weakness
No Yes
INDEP
Decile Obs. Percentage Obs. Percentage Total
1 1,084 93.05% 81 6.95% 1,165
2 1,108 95.19% 56 4.81% 1,164
3 1,122 96.39% 52 4.47% 1,164
4 1,122 96.39% 42 3.61% 1,164
5 1,128 96.82% 37 3.18% 1,165
6 1,134 97.42% 30 2.58% 1,164
7 1,130 97.08% 34 2.92% 1,164
8 1,130 97.08% 34 2.92% 1,164
9 1,124 96.56% 30 2.58% 1,164
10 1,144 98.20% 21 1.80% 1,165
Total 11,226 96.42% 417 3.58% 11,643

Panel B: Regression Analysis

Dependent Variable:
INCICW INCICW INCICW
(1) (2) (3)
INDEP 0.085 0.077 0.070
(5.723)*** (5.234)*** (5.260)***
LDIR 0.031 0.005
(3.749)*** (0.593)
LSEG 0.006
(2.417)**
FRGN 0.011
(2.473)**
M&A 0.004
(1.146)
RESTR 0.008
(2.027)**
QSGRW 0.003
(0.630)
INV 0.052
(2.328)**
LMVE 0.011
(6.276)***
LOSS 0.010
(1.994)**
RDAZ 0.004
(3.998)***
LAGE 0.002
(0.676)
Industry FE Included Included Included
Obs. 11,226 11,226 11,226
R2 0.031 0.036 0.082
*, **, *** Denote significance at the 10 percent, 5 percent, and 1 percent levels, respectively.
This table presents results of the univariate and multivariate analysis on the relation between board independence and internal control weakness. Industry
fixed-effects (FE) are based on two-digit SIC codes. The regressions are performed by Probit model, and t-statistics in parentheses are calculated from the
Huber-White sandwich heteroscedastic consistent errors, which are also corrected for correlation across observations for a given firm.
Variable definitions are specified in Appendix A.

Auditing: A Journal of Practice & Theory

Volume 36, Number 2, 2017
54 Chen, Knechel, Marisetty, Truong, and Veeraraghavan

TABLE 4
The Role of Unitary and Dual Leadership Styles
Dependent Variable:
INCICW INCICW
(1) (2)
INDEP 0.076 0.033
(5.699)*** (1.438)
INDEP  CEOUNI 0.068
(2.565)**
CEOUNI 0.005 0.004
(1.442) (0.769)
LDIR 0.000 0.001
(0.032) (0.090)
LSEG 0.004 0.004
(1.570) (1.620)
FRGN 0.014 0.014
(3.312)*** (3.297)***
M&A 0.006 0.006
(1.622) (1.653)*
RESTR 0.009 0.009
(2.328)** (2.331)**
QSGRW 0.002 0.002
(0.449) (0.436)
INV 0.043 0.042
(1.835)* (1.802)*
LMVE 0.012 0.012
(6.982)*** (6.859)***
LOSS 0.009 0.009
(1.810)* (1.825)*
RDAZ 0.003 0.003
(3.782)*** (3.702)***
LAGE 0.002 0.002
(0.686) (0.591)
Industry FE Included Included
Obs. 11,226 11,226
R2 0.081 0.084
*, **, *** Denote significance at the 10 percent, 5 percent, and 1 percent levels, respectively.
This table presents the results of the regression analysis on the relation between board independence and internal control weakness for unitary and dual
leadership styles. Industry fixed-effects (FE) are based on two-digit SIC codes. The regressions are performed by Probit model, and t-statistics in
parentheses are calculated from the Huber-White sandwich heteroscedastic consistent errors, which are also corrected for correlation across observations
for a given firm.
Variable definitions are specified in Appendix A.

ICWs is stronger for firms with unitary board leadership than firms with dual board leadership. The findings are consistent with
H2 that board independence plays a more important role for firms with unitary leadership where board monitoring is more
critical.

ADDITIONAL ANALYSIS

Types of Internal Control Weakness

As a next step in our analysis, we examine the association between board independence and the disclosure of specific types
of ICWs. Our initial analysis is based on the 21 types of weaknesses identified in the Audit Analytics database. We classify the
21 types of ICWs into two broad categories identified by Doyle et al. (2007); namely, account-specific (or transaction-level)

Auditing: A Journal of Practice & Theory

Volume 36, Number 2, 2017
Board Independence and Internal Control Weakness: Evidence from SOX 404 Disclosures 55

material weaknesses and company-level material weaknesses.7 Doyle et al. (2007) define the classification scheme based on the
severity of the ICW, and employ the logic proposed by Moody’s Investors Service (Doss and Jonas 2004) to determine whether
the ICW is severe. They state that auditors can more easily adjust to account-specific weaknesses than company-level
weaknesses.
To analyze the effect of board independence on the two categories of ICWs, we estimate our regression specification in
Equation (1) but with the dependent variables indicating the disclosure of account-specific and company-level ICWs,
respectively (i.e., the indicator variable for account-specific ICWs is equal to 1 if the firm has account-specific ICWs in a given
year, and 0 otherwise, and the indicator variable for company-level ICWs is equal to 1 if the firm has company-level ICWs in a
given year, and 0 otherwise). We present the results in Columns (1) and (2) of Table 5. The columns show that a higher fraction
of independent directors on the board is associated with lower incidence of both categories of ICWs.8 Further, the coefficient of
board independence is higher in account-specific ICW regression than in company-level ICW regression. The difference
between the two coefficients is statistically significant at the 10 percent level, suggesting that board independence has a large
effect on account-specific ICWs, which are easier to fix.

Timely Remediation of Internal Control Weaknesses

In addition to the disclosure of ICWs, we examine whether board independence is also associated with timely remediation
of ICWs. Goh (2009) finds that the effectiveness of the board of directors and audit committee is associated with firms’
timeliness in the remediation of material weaknesses in internal controls under Section 302 of SOX. Bedard, R. Hoitash, U.
Hoitash, and Westermann (2012) find that different types of ICWs vary in their remediation rates. They further find that
remediation of specific types of ICWs is differentially associated with availability of corporate resources and strength of
corporate governance.
We extend the analysis to ICWs under Section 404 of SOX over a longer time period (i.e., 2004–2012 instead of 2003–
2004 as used in Goh [2009]). Specifically, we identify the first year of each ICW as the initial year of the ICW event. For
example, if a firm reports an ICW in 2006 and 2007 but does not report an ICW in 2005 and 2008, then we identify this as an
ICW event and regard 2006 as the first year of the event. Then, we define remediation (REM) as an indicator variable equal to 1
if ICW is remediated in the next year, and 0 if ICW is not remediated in the next year. On average, 76.5 percent of the identified
ICWs are remediated in the subsequent year. We also define the length of remediation (LENGTHREM) as the number of years
for an ICW to be remediated. The average time to remediation is 1.3 years. Our sample size for this test is small as our focus is
on remediation of ICWs and the main condition for being included in this test is that a firm must have disclosed ICWs before.
Hence, we perform this test with only 315 observations.
We then estimate two models based on these ICW events (i.e., one observation for each event) with the dependent
variables being remediation and length of remediation, respectively, and independent variables being board independence
and all other control variables in Equation (1). The regression results are presented in Columns (3) and (4) of Table 5.
The regression in Column (3) is conducted using a Probit model and shows that the coefficient of board independence is
positive and statistically significant at the 10 percent level, suggesting that a higher proportion of independent directors on
the board is associated with a higher likelihood of remediation of an ICW in the following year. This is consistent with
Bedard et al. (2012). The regression in Column (4) is conducted by ordered Logit model and shows a negative and
significant coefficient for board independence as well. Consistent with Column (3), the results in Column (4) suggest that
a higher proportion of independent directors on the board is associated with a shorter time for the ICWs to be remediated.
Therefore, our findings are consistent with Goh (2009) and Bedard et al. (2012) with regard to the role of board
monitoring in the remediation of ICWs.

7
The account-specific ICWs in Audit Analytics types are Type 1 (Accounting documentation, policy, or procedures), Type 4 (Inadequate disclosure
controls), Type 6 (Ineffective, non-existent, or understaffed audit committee), Type 8 (Insufficient or non-existent internal audit function), Type 9
(Journal entry control issues), Type 11 (Material or numerous auditor adjustments), Type 12 (Non-routine transaction control issues), Type 13
(Remediation of internal weakness identified), Type 14 (Restatement or non-reliance of company filings), Type 15 (Restatement of previous 404
disclosures), Type 16 (SAB 108 adjustments noted), Type 17 (Scope (disclaimer of opinion) or other limitations), Type 18 (SEC or other regulatory
investigations or inquiries), and Type 21 (Untimely or inadequate account reconciliations). The company-level ICWs are Type 2 (Accounting
personnel resources, competency/training), Type 3 (Ethical or compliance issues with personnel), Type 5 (Ineffective or understaffed audit committee),
Type 7 (Information technology, software, security and access issue), Type 10 (Management/board/audit committee investigation (s)), Type 19
(Segregations of duties/design of controls (personnel)), and Type 20 (Senior management competency, tone, reliability issues). Details of the 21 types
of ICWs are available in the Audit Analytics Data Manual.
8
Note, when we rerun the analysis using a reduced sample that excludes company-level (account-specific) ICWs from the analysis of account-specific
(company-level) ICWs, we get qualitatively similar results.

Auditing: A Journal of Practice & Theory

Volume 36, Number 2, 2017
56 Chen, Knechel, Marisetty, Truong, and Veeraraghavan

TABLE 5
Additional Analysis
Accounting Standard
Types of ICW ICW Remediation (AS) Subperiods
Account- Company-
Specific Level Dummy Length AS No. 2 AS No. 5
Dependent Variable:
LENGTH
INCICW INCICW REM REM INCICW INCICW
(1) (2) (3) (4) (5) (6)
INDEP 0.070 0.032 0.134 0.501 0.074 0.014
(5.275)*** (4.181)*** (1.809)* (1.733)* (2.356)** (2.103)**
LDIR 0.004 0.011 0.140 1.172 0.008 0.003
(0.514) (2.410)** (0.877) (1.247) (0.388) (0.571)
LSEG 0.006 0.003 0.135 0.748 0.012 0.001
(2.381)** (1.831)* (3.048)*** (2.932)*** (2.110)** (0.603)
FRGN 0.011 0.004 0.130 0.701 0.045 0.002
(2.458)** (1.464) (1.531) (1.269) (4.256)*** (0.584)
M&A 0.004 0.001 0.049 0.262 0.003 0.001
(1.099) (0.316) (0.723) (0.707) (0.291) (0.448)
RESTR 0.007 0.004 0.042 0.221 0.017 0.007
(2.005)** (1.812)* (0.576) (0.541) (1.556) (2.512)**
QSGRW 0.003 0.003 0.001 0.062 0.002 0.002
(0.658) (1.248) (0.014) (0.154) (0.157) (0.638)
INV 0.052 0.016 0.043 0.339 0.100 0.001
(2.302)** (1.312) (0.099) (0.139) (1.633) (0.059)
LMVE 0.010 0.005 0.014 0.097 0.027 0.008
(6.265)*** (5.108)*** (0.460) (0.541) (6.012)*** (5.776)***
LOSS 0.010 0.003 0.055 0.239 0.040 0.006
(2.030)** (0.876) (0.735) (0.572) (2.477)** (1.603)
RDAZ 0.004 0.001 0.027 0.154 0.011 0.000
(3.975)*** (2.678)*** (1.538) (1.689)* (4.286)*** (0.030)
LAGE 0.002 0.003 0.008 0.066 0.007 0.002
(0.712) (1.924)* (0.162) (0.254) (0.952) (0.964)
Industry FE Included Included Included Included Included Included
Obs. 11,226 11,226 315 315 3,675 7,551
R2 0.086 0.098 0.133 0.156 0.114 0.080
*, **, *** Denote significance at the 10 percent, 5 percent, and 1 percent levels, respectively.
This table presents the regression results of the additional analyses on the relation between board independence and internal control weakness. Industry
fixed-effects (FE) are based on two-digit SIC codes. The regression in Column (4) is performed by ordered Logit model and all other regressions are
performed by Probit model. t-statistics in parentheses are calculated from the Huber-White sandwich heteroscedastic consistent errors, which are also
corrected for correlation across observations for a given firm.
Variable definitions are specified in Appendix A.

The Effect of Auditing Standard Change

In this section, we examine the effect of a change in auditing standards related to internal control reporting on the relation
between board independence and the disclosure of control weaknesses. Auditing Standard No. 5 (AS 5) superseded Auditing
Standard No. 2 (AS 2) effective on November 15, 2007. AS 5 was considered to be less detailed and more process and risk
based than AS 2. As a result, we expect the association between board independence and ICWs to be stronger for firms under
AS 2.9 To test the effect of auditing standard change, we divide our sample period into an AS 2 subperiod (2004–2006) and an

9
Also, given the passage of time and the learning curve of large companies in identifying and remediating ICWs, many pre-existing ICWs would have
been addressed by the time AS 5 became effective.

Auditing: A Journal of Practice & Theory

Volume 36, Number 2, 2017
Board Independence and Internal Control Weakness: Evidence from SOX 404 Disclosures 57

AS 5 subperiod (2007–2012) and run the regression in Equation (1) on the two subperiods. Results are presented in Columns
(5) and (6) of Table 5 and show that the coefficient of board independence is negative and statistically significant for both
subperiods. However, the coefficient is larger in magnitude for the AS 2 subperiod than for the AS 5 subperiod. The difference
between the two coefficients is statistically significant at the 5 percent level and is consistent with our expectation that the
association between board independence on ICWs is weaker after the issue of AS 5.10 We speculate that this result could be due
to the nature of the auditor’s testing of internal control under the two regimes. AS 5 was intended to make the testing of internal
control more top down and holistic, and less detailed. This approach may produce less auditor insight into the internal control
system to share with the Board, reducing their potential impact on the quality of internal control overall.

SUPPLEMENTAL ANALYSIS
The significant negative association between board independence and the probability of reporting an ICW could be due to
the fact that independent directors are better monitors of the firm. However, it is possible that both board independence and
ICWs are correlated with variables omitted from the regression. It is also possible that the disclosure of an ICW can have a
significant impact on the firm’s governance structure, including the independence of the board. To mitigate the omitted variable
concern, we include firm fixed-effects in the regression, which controls for time-invariant omitted firm characteristics.11 The
regression results are presented in Column (1) of Table 6 and show that the negative coefficient of board independence remains
even after controlling for firm fixed-effects.
To address the potential reverse causality problem, we first employ the subsample regressions approach. In this approach,
the first subsample includes observations in year 2004 only. Since 2004 is the first year that firms had to evaluate their internal
control, board characteristics are less likely to change in response to the outcome of the evaluation. The second subsample
includes observations without a previously identified ICW. For example, if a firm reports an ICW in 2006 for the first time, then
we drop all observations after 2006 and only keep those from 2004 to 2006. The purpose of using the second subsample is
similar to that of the first one, except that the sample size in this case is much larger. We use the lagged value of board
independence as the independent variable to address the reverse causality problem further. Since firms do not report a previous
ICW, board independence is less likely to have changed as a result of the disclosed ICW. This experimental setting ensures that
the causality goes from board independence to ICW but not vice versa. The regression results for the two subsamples are
presented in Columns (2) and (3) of Table 6. We show that the negative relation between an ICW and the lagged value of board
independence holds in both subsamples, suggesting that the relation is not driven by reverse causality.
Further, we conduct a formal test to address potential endogeneity problems. We adopt an approach following Wooldridge
(2002) where we first regress board independence against determinants of an ICW. In the second stage, the predicted value of
board independence from the first-stage regression is employed as the instrumental variable.12 We follow Linck, Netter, and
Yang (2008) and employ a number of determinants of an ICW in the first stage. Detailed definitions of these variables are
presented in Appendix A. The regression results are presented in Table 7. Panel A presents the prediction model and shows that
the coefficients of the independent variables are largely consistent with prior literature (e.g., Linck et al. 2008). For example,
large firms and firms with high leverage have a higher proportion of independent directors on the board. Further, the under-
identification test indicates that the instruments are relevant for the first-stage regression (p , 0.01) and the Sargan’s J test of
over-identification does not reject the null hypothesis that these instruments are jointly uncorrelated with the error term from the
second-stage regression (p . 0.10), supporting the validity of these instrumental variables. We report the second-stage
Wooldridge (2002) regression in Panel B of Table 7. The panel shows that the coefficient of board independence remains
negative and statistically significant, suggesting that our findings are robust to the endogeneity concerns.

CONCLUSION
In this paper, we argue that the fiduciary responsibility and monitoring role of independent directors should encourage
better accounting practices and internal control mechanisms by management. Hence, board independence should be negatively
related to the propensity for firms to report weaknesses in internal control over financial reporting (ICOFR). We also argue that
the effect of board independence on the disclosure of ICWs depends on the board leadership structure, specifically whether the
company has a unitary or dual CEO/chairman.

10
The disclosure of ICWs has declined over our sample period. To control for this decline, we add year fixed-effects to the model for the different
auditing standard subperiods. The results are qualitatively the same as our primary analysis.
11
Since 96.42 percent of the observations in our sample do not report ICWs, firms without ICWs throughout our sample period are dropped automatically
after including firm fixed-effects into the Probit model. This reduces our sample to a large extent. Therefore, we use OLS in this test, which does not
suffer from this problem.
12
We cannot employ the more traditional two-stage least squares technique, as our dependent variable in the second stage (INCICW) is a binary variable.

Auditing: A Journal of Practice & Theory

Volume 36, Number 2, 2017
58 Chen, Knechel, Marisetty, Truong, and Veeraraghavan

TABLE 6
Robustness Checks
Firm Fixed-Effects 2004 Only No ICW Before
Dependent Variable:
INCICW INCICW INCICW
(1) (2) (3)
INDEP 0.082 0.155 0.061
(2.465)** (2.146)** (5.556)***
LDIR 0.042 0.013 0.000
(1.638) (0.280) (0.055)
LSEG 0.000 0.009 0.004
(0.075) (0.574) (2.389)**
FRGN 0.005 0.059 0.008
(0.476) (2.321)** (2.128)**
M&A 0.006 0.027 0.004
(1.027) (1.204) (1.412)
RESTR 0.008 0.004 0.008
(1.083) (0.177) (2.390)**
QSGRW 0.007 0.034 0.002
(1.141) (1.339) (0.598)
INV 0.038 0.552 0.063
(0.375) (2.887)*** (2.970)***
LMVE 0.010 0.028 0.009
(1.444) (2.789)*** (6.703)***
LOSS 0.007 0.112 0.006
(0.656) (2.692)*** (1.398)
RDAZ 0.007 0.028 0.003
(2.578)** (3.829)*** (3.826)***
LAGE 0.113 0.011 0.002
(5.271)*** (0.648) (1.041)
Industry FE Included Included Included
Obs. 11,226 711 9,920
R2 0.321 0.191 0.098
*, **, *** Denote significance at the 10 percent, 5 percent, and 1 percent levels, respectively.
This table presents the results for robustness checks. Industry fixed-effects (FE) are based on two-digit SIC codes. The regressions are performed by OLS
in Column (1) and Probit model in Columns (2) and (3). t-statistics in parentheses are calculated from the Huber-White sandwich heteroscedastic
consistent errors, which are also corrected for correlation across observations for a given firm.
Variable definitions are specified in Appendix A.

Using a comprehensive sample of U.S. firms spanning the period 2004–2012, we establish several key findings. First, we
show that board independence is negatively associated with the disclosure of ICWs. Second, we observe that the negative
relation between board independence and ICWs is strongest in a company that has unitary leadership. In an additional analysis,
we find that board independence is associated with timely remediation of ICWs and that the implementation of Auditing
Standard No. 5 in 2007 weakens the effect of board independence on the disclosure of ICWs. Our findings are robust to omitted
variable and reverse causality problems.
Our paper makes several contributions to the literature. First, we show a significant negative association between board
independence and the disclosure of an ICW. Second, we document that the effect of board independence is conditional on the
structure of firm leadership (dual versus unitary). Taken together, these findings suggest that the effectiveness of internal
control over financial reporting should be evaluated in the context of the overall corporate governance environment. In the
broadest sense, effective control in an organization is achieved through many avenues including ICOFR and corporate
governance, but also through internal and external auditing, direct involvement by dominant shareholders and, even, the
influence of other external stakeholders (e.g., creditors). We also add to the literature on internal control reporting by showing
that board structure influences different types and the remediation of internal control weaknesses. Future research should
examine more of the links across sources of internal control (i.e., governance, internal auditing) since the overall performance

Auditing: A Journal of Practice & Theory

Volume 36, Number 2, 2017
Board Independence and Internal Control Weakness: Evidence from SOX 404 Disclosures 59

TABLE 7
Instrumental Variable Approach

Panel A: Prediction Model

Dependent Variable:
INDEP
LMVE 0.004
(2.048)**
DEBT 0.053
(3.465)***
LSEG 0.001
(0.244)
LAGE 0.047
(3.141)***
LAGE2 0.011
(3.966)***
MB 0.000
(0.026)
R&D 0.026
(0.657)
STDRET 0.049
(2.160)**
CEOOWN 0.018
(0.263)
DIROWN 0.438
(13.528)***
FCF 0.026
(0.670)
INDROA 0.062
(1.841)*
CEOAGE 0.002
(0.433)
CEODUAL 0.010
(2.427)**
Industry FE Included
Obs. 8,321
R2 0.204

Panel B: Second-Stage Wooldridge (2002) Regression

Dependent Variable:
INCICW
Instrumented INDEP 2.334
(4.394)***
LDIR 0.121
(1.027)
LSEG 0.056
(1.925)*
FRGN 0.165
(2.673)***
M&A 0.045
(0.825)
RESTR 0.142
(2.357)**
QSGRW 0.009
(0.138)
(continued on next page)

Auditing: A Journal of Practice & Theory

Volume 36, Number 2, 2017
60 Chen, Knechel, Marisetty, Truong, and Veeraraghavan

TABLE 7 (continued)
Dependent Variable:
INCICW
INV 0.302
(1.281)
LMVE 0.154
(6.229)***
LOSS 0.133
(1.776)*
RDAZ 0.042
(3.787)***
LAGE 0.026
(0.626)
Industry FE Included
Obs. 8,321
R2 —
*, **, *** Denote significance at the 10 percent, 5 percent, and 1 percent levels, respectively.
This table presents the results for the instrumental variable test. In the first stage, we regress board independence on a number of determinants. In the
second stage, we use the predicted value of board independence as an instrument in a standard instrumental variable estimation. Industry fixed-effects (FE)
are based on two-digit SIC codes. t-statistics in parentheses are calculated from the Huber-White sandwich heteroscedastic consistent errors, which are also
corrected for correlation across observations for a given firm.
Variable definitions are specified in Appendix A.

of management and the company is likely to be affected by the joint impact of these various control mechanisms. Nevertheless,
the current findings should be of interest to regulators who oversee governance and financial reporting in the financial markets,
investors and shareholders who are represented by the board, and auditors who incorporate assessments of corporate
governance in their assessment of risk within the audit process.

REFERENCES
Adams, R. B., B. E. Hermalin, and M. S. Weisbach. 2010. The role of boards of directors in corporate governance: A conceptual
framework and survey. Journal of Economic Literature 48: 58–107.
Ashbaugh-Skaife, H., D. W. Collins, and W. R. Kinney. 2007. The discovery and reporting of internal control deficiencies prior to SOX-
mandated audits. Journal of Accounting & Economics 44: 166–192.
Ashbaugh-Skaife, H., D. W. Collins, W. R. Kinney, and R. LaFond. 2008. The effect of SOX internal control deficiencies and their
remediation on accrual quality. The Accounting Review 83: 217–250.
Beasley, M. 1996. An empirical analysis of the relation between the board of director composition and financial statement fraud. The
Accounting Review 71: 443–465.
Beasley, M., J. Carcello, D. Hermanson, and P. Lapides. 2000. Fraudulent financial reporting: Consideration of industry traits and
corporate governance mechanisms. Accounting Horizons 14: 441–454.
Bédard, J., S. Chtourou, and L. Courteau. 2004. The effect of audit committee expertise, independence, and activity on aggressive
earnings management. Auditing: A Journal of Practice & Theory 23: 13–25.
Bedard, J. C., R. Hoitash, U. Hoitash, and K. Westermann. 2012. Material weakness remediation and earnings quality: A detailed
examination by type of control deficiency. Auditing: A Journal of Practice & Theory 31: 57–78.
Bhagat, S., and B. Black. 1999. The uncertain relationship between board composition and firm performance. Business Lawyer 54: 921–
963.
Bhagat, S., and B. Black. 2002. The non-correlation between board independence and long-term firm performance. Journal of
Corporation Law 27: 231–274.
Bhagat, S., and B. Bolton. 2008. Corporate governance and firm performance. Journal of Corporate Finance 14: 257–273.
Brickley, J. A., J. I. Coles, and G. J. Jarrell. 1997. Leadership structure: Separating the CEO and chairman of the board. Journal of
Corporate Finance 3: 189–220.
Brown, L. D., and M. L. Caylor. 2006. Corporate governance and firm valuation. Journal of Accounting and Public Policy 25: 409–434.
Carcello, J., D. Hermanson, T. Neal, and R. A. Riley. 2002. Board characteristics and audit fees. Contemporary Accounting Research 19:
365–384.
Conyon, M. J., and S. I. Peck. 1998. Board control, remuneration committees, and top management compensation. The Academy of
Management Journal 41: 146–157.

Auditing: A Journal of Practice & Theory

Volume 36, Number 2, 2017
Board Independence and Internal Control Weakness: Evidence from SOX 404 Disclosures 61

Dechow, P. M., R. G. Sloan, and A. P. Sweeney. 1996. Causes and consequences of earnings manipulation: An analysis of firms subject
to enforcement actions by the SEC. Contemporary Accounting Research 13: 1–36.
DeFond, M. L., and J. Jiambalvo. 1991. Incidence and circumstances of accounting errors. The Accounting Review 66: 643–655.
Doss, M., and G. Jonas. 2004. Section 404 Reports on Internal Control: Impact on Ratings Will Depend on Nature of Material
Weaknesses Reported. Available at: http://www.alacrastore.com/moodys-credit-research/Section-404-Reports-on-Internal-Control-
Impact-on-Ratings-Will-Depend-on-Nature-of-Material-Weaknesses-Reported-PBC_89482
Doyle, J., W. Ge, and S. McVay. 2007. Accruals quality and internal control over financial reporting. The Accounting Review 82: 1141–
1170.
Feng, M., C. Li, and S. McVay. 2009. Internal control and management guidance. Journal of Accounting & Economics 48: 190–209.
Finkelstein, S., and D. Hambrick. 1996. Strategic Leadership: Top Executives and Their Effects on Organizations. St. Paul, MN: West
Publishing Company.
Floyd, W. 2016. Financial Reporting Quality and Voluntary Disclosure: Evidence from Internal Control Weaknesses. Working paper,
Stanford University.
Ge, W., and S. McVay. 2005. The disclosure of material weakness in internal control after the Sarbanes-Oxley Act. Accounting Horizons
19: 137–158.
Goh, B. W. 2009. Audit committees, boards of directors, and remediation of material weaknesses in internal control. Contemporary
Accounting Research 26: 549–579.
Goyal, V. K., and C. W. Park. 2002. Board leadership structure and CEO turnover. Journal of Corporate Finance 8: 49–66.
Hermalin, B. E., and M. S. Weisbach. 1998. Endogenously chosen boards of directors and their monitoring of the CEO. The American
Economic Review 88: 96–118.
Hermalin, B. E., and M. S. Weisbach. 2003. Boards of directors as an endogenously determined institution: A survey of the economic
literature. Economic Policy Review 9: 7–26.
Hoitash, U., R. Hoitash, and J. C. Bedard. 2009. Corporate governance and internal control over financial reporting: A comparison of
regulatory regimes. The Accounting Review 84: 839–867.
Jensen, M. 1993. The modern industrial revolution, exit, and the failure of internal control systems. The Journal of Finance 48: 831–880.
Johnstone, K., C. Li, and K. H. Rupley. 2011. Changes in corporate governance associated with the internal control material weaknesses
and their subsequent remediation. Contemporary Accounting Research 28: 331–383.
Kim, J.-B., B. Y. Song, and L. Zhang. 2011. Internal control weakness and bank loan contracting: Evidence from SOX Section 404
disclosures. The Accounting Review 86: 1157–1188.
Klein, A. 2002. Audit committee, board of director characteristics, and earnings management. Journal of Accounting & Economics 33:
375–400.
Knechel, W. R., and M. Willekens. 2006. The role of risk management and governance in determining audit demand. Journal of Business
Finance & Accounting 33: 1344–1367.
Krishnan, J. 2005. Audit committee quality and internal control: An empirical analysis. The Accounting Review 80: 649–675.
Larcker, D. F., S. A. Richardson, and I. Tuna. 2007. Corporate governance, accounting outcomes, and organizational performance. The
Accounting Review 82: 963–1008.
Linck, J., J. Netter, and T. Yang. 2008. The determinants of board structure. Journal of Financial Economics 87: 308–328.
Ogneva, M., K. R. Subramanyam, and K. Raghunandan. 2007. Internal control weaknesses and cost of equity: Evidence from SOX
Section 404 disclosures. The Accounting Review 82: 1255–1297.
Public Company Accounting Oversight Board (PCAOB). 2007. Annual Report. Available at: https://pcaobus.org/About/Administration/
Documents/Annual%20Reports/2007.pdf
Uzun, H., S. H. Szewczyk, and R. Varma. 2004. Board composition and corporate fraud. Financial Analysts Journal 60: 33–43.
Weisbach, M. S. 1988. Outside directors and CEO turnover. Journal of Financial Economics 20: 432–460.
Wooldridge, J. M. 2002. Econometric Analysis of Cross Section and Panel Data. Cambridge, MA: MIT Press.
Yermack, D. 1996. Higher market valuation of companies with a small board of directors. Journal of Financial Economics 40: 185–211.
Zhang, Y., J. Zhou, and N. Zhou. 2007. Audit committee quality, auditor independence, and internal control weaknesses. Journal of
Accounting and Public Policy 26: 300–327.

Auditing: A Journal of Practice & Theory

Volume 36, Number 2, 2017
62 Chen, Knechel, Marisetty, Truong, and Veeraraghavan

APPENDIX A
Variables, Sources, and Definitions
Variable Source Definition
AGE CRSP Number of years the firm exists in the CRSP database.
AZ Compustat 3.3 times pretax income (OIBDP) plus sales (SALE) plus 1.4 times retained earnings (RE) plus 1.2 times
net working capital, then divided by total assets (AT). Net working capital is defined as current assets
(ACT) minus current liabilities (LCT).
CEOUNI RiskMetrics Indicator variable equal to 1 if the CEO is also the chairman of the board, and 0 otherwise.
CEOOWN RiskMetrics Number of shares owned by the CEO divided by number of shares outstanding.
DEBT Compustat Long-term debt (DLTT) divided by total assets (AT).
DIROWN RiskMetrics Number of shares owned by all the board directors divided by number of shares outstanding.
FCF Compustat Income before extraordinary items (IB) minus total tax (TXT) minus change in deferred tax (TXDB) minus
interest expense (XINT) minus common dividend (DVC) minus preferred dividend (DVP), then divided
by total assets (AT).
FRGN Compustat Indicator variable equal to 1 if the firm has a non-zero foreign currency translation (CICURR), and 0
otherwise
INCICW Audit Analytics Indicator variable equal to 1 if the firm has internal control material weaknesses (ICW), and 0 otherwise.
INDEP RiskMetrics Fraction of independent directors on the board.
INDROA Compustat The difference between the firm’s ROA and the average ROA over the two-digit SIC industry. ROA is
defined as operating income after depreciation (OIADP) divided by total assets (AT).
INV Compustat Inventories (INVT) divided by total assets (AT).
LDIR RiskMetrics Natural logarithm of the variable NUMDIR.
LMVE Compustat Natural logarithm of the variable MVE.
LOSS Compustat Indicator variable equal to 1 if earnings before extraordinary items (IB) is negative, and 0 otherwise.
LSEG Compustat Natural logarithm of the variable NUMSEG.
LAGE CRSP Natural logarithm of the variable AGE.
M&A SDC Indicator variable equal to 1 if the firm was involved in mergers or acquisitions over the three-year period
including the current year, and 0 otherwise.
MB Compustat MVE divided by book equity (CEQ).
MVE Compustat Market value of equity, defined as number of shares outstanding (CSHPRI) times stock price (PRCC_F).
NUMDIR RiskMetrics Number of directors on the board.
NUMSEG Compustat Number of business segments in the firm.
QSGRW Compustat Indicator variable equal to 1 if industry-adjusted SGRW falls into the top quintile, and 0 otherwise.
R&D Compustat R&D expenses (XRD) divided by total assets (AT).
RDAZ Compustat Decreasing decile rank of the variable AZ.
RESTR Compustat Indicator variable equal to 1 if the firm is involved in restructuring over the three-year period including the
current year (at least one of restructuring costs after tax [RCA], restructuring costs diluted EPS effect
[RCD], restructuring costs basic EPS effects [RCEPS], and reserves for credit losses [RC]) is not equal
to 0), and 0 otherwise.
SGRW Compustat Annual growth rate of sales revenue (SALE).

Auditing: A Journal of Practice & Theory

Volume 36, Number 2, 2017
Copyright of Auditing: A Journal of Practice & Theory is the property of American
Accounting Association and its content may not be copied or emailed to multiple sites or
posted to a listserv without the copyright holder's express written permission. However, users
may print, download, or email articles for individual use.