CEO equity incentives and the remediation of material weaknesses in internal control

• Author: Xiaohong Liu,Xuejiao Liu
• Date: 01 October 2017
• DOI: http://doi.org/10.1111/jbfa.12265
• Published date: 01 October 2017

DOI: 10.1111/jbfa.12265

CEO equity incentives and the remediation

of material weaknesses in internal control

Xuejiao Liu1Xiaohong Liu2

1Schoolof Business, University of International

Businessand Economics, Beijing, China

2Schoolof Business, The University of Hong

Kong,Pokfulam, Hong Kong

Correspondence

XuejiaoLiu, School of Business, University

ofInternational Business and Economics,

No.10 East Huixin Street, Beijing, China.

Email:xjluibe@outlook.com,

xuejiaoliu@uibe.edu.cn

JELClassification: G14, M41, M48, M52

Abstract

This study examines how CEO equity incentives affect the reme-

diation of material weaknesses (MWs) in internal control disclosed

pursuant to the Sarbanes-Oxley Act (SOX). We find that the sensi-

tivity of CEO equity portfolios to stock price (CEO price sensitiv-

ity, or delta) has a positive impact on firm promptness in remedy-

ing MWs, whereas the sensitivity of CEO equity portfolios to stock

return volatility (CEO volatility sensitivity, or vega) has a negative

impact on firm promptness in remedying MWs. In addition, we pro-

vide evidence that effective boards of directors mitigate the unde-

sirable, negative effect of CEO volatility sensitivity on remediation

of MWs. Our results shed light on the effects of equity compensa-

tion structures on internal control quality in the more transparent,

post-SOX environment.

KEYWORDS

CEO equity incentives, internal control weakness, remediation,

Sarbanes-Oxley Act

1INTRODUCTION

This study examineshow chief executive officer (CEO) equity incentives affect the remediation of material weaknesses

(MWs) in internal control that are disclosed pursuant to the Sarbanes-Oxley Act (SOX).1Healthy internal control is

essential to ensure the reliability of a firm’s financial reporting and the achievement of its operating objectives [Com-

mittee of Sponsoring Organizations of the Treadway Commission (COSO), 2011]. In response to a series of major

accounting scandals and corporate failures in the early 2000s, the US Congress passed SOX in 2002 to improve the

integrity of financial reporting and protect investors from accounting fraud (US House of Representatives, 2002). In

particular, Sections 302 and 404 of SOX (SOX 302 and SOX 404 hereafter) require that firms disclose any MW in

internal control over financial reporting in their periodic financial reports (SEC, 2002, 2003).2On the one hand, public

1An MW in internal control is ‘asignificant deficiency, or combination of significant deficiencies, that results in more than a remote likelihood that a material

misstatementof the annual or interim financial statements will not be prevented or detected’ [Public Company Accounting Oversight Board (PCAOB), 2004].

2Asin prior research (e.g., Doyle, Ge, & McVay,2007a; Goh, 2009), we view disclosure of an MW in internal control as mandatory, but disclosure of a significant

deficiencyin internal control as voluntary. According to the Office of the Chief Accountant (SEC, 2004), ‘aregistrant is obligated to identify and publicly disclose

all material weaknesses [in internal control]’; however,‘if management identifies a significant deficiency it is not obligated by virtue of that fact to publicly

disclosethe existence or nature of the significant deficiency.’

1338 c

2017 John Wiley & Sons Ltd wileyonlinelibrary.com/journal/jbfa JBus Fin Acc. 2017;44:1338–1369.

LIU ANDLIU 1339

disclosure of MWstriggers a significant decline in a firm’s stock price (e.g., DeFranco,Guan, & Lu, 2005; Gupta & Nayar,

2007; Hammersley,Myers, & Shakespeare, 2008), which can act as a disciplinary mechanism for managers whose com-

pensation is tied to stock price. On the other hand, as firms with internal control weaknesses tend to have higher

information and operating risk (e.g., Ashbaugh-Skaife, Collins, Kinney,& LaFond, 2009), managers whose compensa-

tion increases with firm risk may have an incentiveto delay the remediation of internal control weaknesses. Our main

objective is to examinehow these opposing incentives resulting from internal control disclosure and managers’ equity

compensation (stocks and stock options) impact management actions in the remediation of MWs.

Kobelsky, Lim, and Jha (2013) and Balsam, Jiang, and Lu (2014) are the first to assess the role of executive com-

pensation in the context of firm internal control. They report an inverse relation between managers’ price sensitivity

(the sensitivity of equity compensation to stock price, also known as delta) and the existence/disclosure of MWs by

comparing firms that disclosed MWs with those that did not disclose in the first two yearsfollowing enactment of SOX

404. However,because SOX is an unexpected regulatory change, internal control statuses disclosed in the short period

following SOX enactment largely reflect managers’ pre-SOXinternal control decisions. 3Thus, the inferences drawnon

the post-SOXequity incentives in these studies are likely based on a sample comprising mainly pre-SOX internal control

decisions (which are revealed due to the SOXdisclosure regulation).

In addition, in analyzing equity incentives,these studies do not consider the risk-seeking incentive provided by man-

agers’ volatility sensitivity,that is, the sensitivity of equity compensation to stock return volatility, or vega. Since delta

and vega are positively correlated (see Table3) and are inherent in stock- and option-based compensation, omission of

vega may lead to incomplete and biased inferences.4

Our research objective and design differ from those of the previously mentioned studies in two keyrespects. First,

we exploitthe shift in disclosure regime (brought about by SOX), which caused significant stock price declines for firms

reporting MWs in internal control, to investigate how managers respond to losses in their stock price-based wealth

through changing/remedying ineffective internal controls. By analyzing managers’ remediation behavior,we investi-

gate the tradeoff between equity benefits and costs in moving awayfrom ineffective internal controls under the more

transparent, post-SOX disclosure regime.

More specifically, firms’ initial disclosures of MWs under SOX present a quasi-experimental setting where we

attempt to identify a causal relation between managers’ equity incentives and their remediation behavior in a short

window of one year following MW disclosure. In this setting, remediation incentives derive from market reaction to

unanticipated MW disclosure and managers’ pre-existing (i.e., pre-disclosure) equity holdings. Therefore, the equity

incentives are exogenousto the determination of managers’ subsequent remediation behavior. This quasi-experiment

helps us to examine the effects of equity incentiveson internal control decisions.5

The second keydifference between our research and both Kobelsky et al. (2013) and Balsam et al. (2014) is that we

distinguish between the different roles of price and volatility sensitivities derived from equity compensation, whereas

they consider only price sensitivity.We thus mitigate the omitted correlated variable problem in their studies and pro-

vide a more comprehensive perspective on the effects of managerial equity incentiveson firm internal control.

Our main analyses focus on CEO incentives because CEOs are the ultimate leaders, planners, and coordinators of

internal control, especially for those control activities that are at the firm level or cut across different units of a firm,

such as prevention of management override, fraud detection controls, and compliance with corporate policies and

procedures (Moody’s, 2006; COSO, 2011).

3Sincefirms need time to adjust their internalcontrol positions (i.e., to fix MWs) following the external shock of SOX, most of the MWs disclosed in the short

period following enactment of SOXreflect managers’ pre-SOX internal control decisions rather than their optimal decisions in the post-SOX period. Consistent

withthis view, we find that the proportion of firms disclosing MWs in SOX 404 reports decreased from 15.8 percent in 2004 to 3.9 percent in 2009 and leveled

offafter 2009, suggesting a transition period in which firms adapted to the post-SOX environment.

4Consistent with this concern, Armstrong, Larcker,Ormazabal, and Taylor (2013, p. 327) find that ‘jointly considering the incentive effects of both portfolio

deltaand portfolio vega substantially alters inferences reported in [the] prior literature [on earnings management].’

5Ourresults cannot be inferred from those of Kobelsky et al. (2013) and Balsam et al. (2014). They find that CEO price sensitivity is not significantly associated

withthe existence/disclosure of MWs when chief financial officer (CFO) price sensitivity is separatelyincluded in the regression. In contrast, we findthatCEO

price sensitivity has a significantly positive impact on the remediation of MWseven after controlling for CFO price sensitivity (see Section 5). We also note

that, in an additional analysis, Balsam et al. (2014, table 7) examinethe relation between MW remediation and both the level and change in price sensitivity.

Theresults in their remediation analysis differ from our results.

1340 LIU ANDLIU

We posit that CEO equity holdings provide two opposite incentives for the remediation of MWs in internal con-

trol following their initial disclosure, depending on the relation of the equity holdings to stock price and stock return

volatility.First, the value of stock and stock options decreases as stock price slumps in response to disclosure of MWs

(e.g., Gupta & Nayar,2007; Hammersley et al., 2008). This relation gives CEOs an incentive to promptly remedy MWs

to prevent further losses in their stock price-based wealth. We thus hypothesizethat, ceteris paribus, CEO price sensi-

tivity has a positive impact on firm promptness in remedying MWs in internal control. We measure price sensitivity as

the dollar change in the value of the CEO’s stock and option portfolios resulting from a 1 percent change in stock price

(Core & Guay,2002).

Second, CEOs could forgo some of their stock risk/volatility-basedwealth by moving from weak internal controls to

stronger internal controls, because the value of stock options is positivity related to firm risk (Black & Scholes, 1973),

and firms tend to have higher information risk and excessiveoperating risk under a weak control and monitoring sys-

tem (e.g., Ashbaugh-Skaife et al., 2009). Therefore, CEOs with option holdings have an incentive to delay rectifying

internal control deficiencies. We thus hypothesize that, ceteris paribus, CEO volatility sensitivity has a negative impact

on firm promptness in remedying MWs in internal control. Wemeasure volatility sensitivity as the partial derivative of

CEO equity holdings with respect to stock return volatility,namely, the dollar change in the value of the CEO’s option

portfolio resulting from a 1 percent change in stock return volatility (Core & Guay,2002).6

Wethen empirically test these hypotheses using a sample of firms that disclosed MWs in internal control for the first

time at their fiscal year-end, between November 2003 and August 2006, under SOX 302 or SOX 404.7Our empirical

measure of firm remediation promptness is an indicator variable that takes the value of one (zero) if a firm remedies

(fails to remedy) the MWs, as attested by auditor opinions, in the fiscal yearfollowing initial disclosure of MWs.

Our empirical results are consistent with these hypotheses. Specifically,we find that, ceteris paribus, as the sensitiv-

ity of CEO equity portfolios to stock prices (delta) increases, firms are more likelyto promptly remedy MWs in internal

control. In contrast, as the sensitivity of CEO equity portfolios to stock return volatility (vega) increases, firms are less

likely to promptly remedy MWs in internal control. In terms of economic significance, our results suggest that a one

standard deviation increase in CEO delta from its mean increases the likelihood of timely remediation of MWs by12

percent, whereas a similar increase in CEO vega reduces the likelihood of timely remediation of MWs by 9 percent.

Finally, weprovide evidence that effective boards of directors mitigate the undesirable, negative effects of CEO vega

on the remediation of MWs. This moderating effect from corporate governanceprovides additional evidence consis-

tent with our argument that our results are likely to be causal instead of correlational (e.g., Jha & Chen, 2015). Our

results are also robust to the use of a two-stage instrumental variable estimation approach that further accounts for

the endogenous nature of CEO equity incentives (Armstrong & Vashishtha,2012).

We provide three main contributions. First, our study sheds light on the CEO tradeoff between equity benefits

and costs in moving away from ineffective firm internal controls under the more transparent, post-SOX disclosure

regime. Our results imply that when the price-driven losses managers suffer from public revelation of internal con-

trol weaknesses exceed any risk-driven benefits they enjoy under a weak control system, these personal losses not

only impel managers to remedy existing weaknesses, they also act as a credible threat to deter managers from choos-

ing ineffective internal controls in the future. This evidence suggests that the disclosure regulation in SOX provides

a channel for a compensation structure that creates high delta, rather than one that creates high vega, to facilitate

market-based discipline on managers, therebystrengthening firm internal control. Our results are consistent with evi-

dence of a decrease in the (positive) relation between executivedelta and earnings management from the pre- to the

6Sinceprice and volatility sensitivities provide opposite incentives for MW remediation, the net effect of a CEO’s equity portfolio on that CEO’s remediation

actiondepends on which of the two opposite incentives arising from the equity portfolio is dominant.

7Asexplained subsequently (in Section 3), our sample (of firms that disclosed MWs for the first time under SOX) begins November 15, 2003, one year before

thefirst batch of SOX 404 audit opinions under Auditing Standard No. 2 (or AS2; see PCAOB, 2004) became available, so that we can check all disclosed MWs

with subsequent audit opinions in firm annual reports to confirm their remediation statuses. Our sample ends August 27, 2006, one yearprior to the earliest

adoption of Auditing Standard No. 5 (or AS5; see PCAOB, 2007), which provides a different set of guidelines for auditors to expressan opinion on internal

controleffectiveness compared with AS2, to avoid potential confounding effects of the disparities between AS2 and AS5.