Do Investors Find Audits of Material Weakness Remediation Disclosures to be Valuable?

Transcription

1 Do Investors Find Audits of Material Weakness Remediation Disclosures to be Valuable? Sanaz Aghazadeh* Lehigh University College of Business and Economics 339 Rauch Business Center 621 Taylor Street Bethlehem, PA Tel Fax Marietta Peytcheva Lehigh University College of Business and Economics 339 Rauch Business Center 621 Taylor Street Bethlehem, PA Tel Fax *Corresponding author We would like to acknowledge research support from The Martindale Center for the Study of Private Enterprise at Lehigh University. We thank Heibatollah Sami, Paul Coram, and participants of the ABO 2013 meeting.

2 Do Investors Find Audits of Material Weakness Remediation Disclosures to be Valuable? ABSTRACT Year-end audits of material weakness remediation disclosures are required. However, there is little evidence on the value of these audits to investors. We take advantage of the context provided by AS 4, to examine the value of a voluntary audit of a firm s material weakness (MW) remediation disclosure at quarter-end, as evidenced by nonprofessional investors willingness to invest in the firm. In a between-participants experiment, we manipulate the presence of an audit of management s MW remediation disclosure and the type of the previously existing MW. Furthermore, we examine whether a voluntary audit increases the credibility of the MW remediation disclosure (Mercer 2004). We find that management s choice to obtain an independent audit on their MW remediation disclosure increases disclosure credibility which in turn increases investors willingness to invest in the firm, for both entity level control MWs and account-specific control MWs. Keywords: auditing; assurance; material weakness remediation; disclosure credibility 1

3 INTRODUCTION The purpose of this study is to examine the value of an audit for management s disclosure of material weakness (MW) remediation a mandatory internal control disclosure. The Sarbanes-Oxley Act of 2002 (SOX) requires all public companies to report on the effectiveness of their internal controls and to obtain an audit of this evaluation. During the audit, auditors must evaluate whether the internal controls are effective. A MW exists if the internal controls are not effective, and management must disclose this. Management must remediate the MW, and the timing of the remediation has implications for firm value. Management can elect to remediate the MW prior to the release of the following year s 10-K, and report on the remediation in the quarterly financials (10-Q). Management has the option to obtain an independent audit that provides assurance regarding the early remediation of the MW. Auditing Standard No. 4 (AS 4), which became effective in 2006, provides guidance on such voluntary audits of early MW remediation. Thus far, research has not examined the implications of this standard. From a theory perspective, the context of AS 4 allows for the examination of the value of a voluntary audit for an important type of management disclosures the disclosure of MW remediation. 1 Prior literature has mainly focused on the value of an audit for financial disclosures (e.g., Lennox and Pittman 2011; Marshall et al. 2014). Further, AS 4 allows for the examination of the value of a voluntary audit when the disclosure is more or less inherently plausible. Inherent plausibility refers to the likelihood that, ceteris paribus, the information contained in the disclosure deviates from expectations (Mercer 2004). We examine inherent plausibility by manipulating whether the MW is at the entity-level or account-specific. From a practice perspective, it is important to reconcile regulators assessment of the value of AS 4 with the 1 We recognize that enhancing material weakness disclosures will likely lend credibility to the financial statement disclosures as well. 2

4 investors assessments of its usefulness. 2 In addition, AS 4 allows us to determine the value of an audit at quarter-end and the potential value of an audit at year-end. Typically, MW remediation disclosures are always audited at year-end, so companies have no way to know the incremental value of the audit. Our study addresses the issue of whether the presence of an audit covering management s early remediation assertion influences investors judgments about the company s stock. We propose a theoretical model grounded in Mercer s (2004) credibility framework to examine the value of a MW remediation audit. We consider two factors outlined by Mercer: the presence of assurance (MW remediation was audited vs. not) and the inherent plausibility of the disclosure (remediated MW was related to the entity-level vs. account-specific). Prior research has not considered the interactive effects of these factors. However, prior research suggests that entity-level MWs are more difficult to remediate than account-specific MWs (Goh 2009; Hammersley et al. 2012; Bedard et al. 2012). Given this, it is important to consider the potential differential effects of an audit for these different types of MW. Our model predicts that the audit of the MW remediation and the type of MW will have a joint effect on disclosure credibility, which in turn will influence financial statement users willingness to invest in the firm s stock. We expect that the presence of assurance on management s MW remediation assertion will increase disclosure credibility. We also predict that we will observe an interactive effect of the presence of an audit and the inherent plausibility of the disclosure. Specifically, we predict that an audit will have a greater effect on the credibility of the MW remediation disclosure when management remediates an entity-level MW. 2 See the SEC background on AS 4 (SEC Release No , File No. PCAOB ), the public comments to the PCAOB proposed AS 4 (File No. PCAOB , Exhibit 2 (a) (C)), and the webcast for the November 18, 2004 PCAOB SAG Meeting related to AS 4 (available at the PCAOB website). 3

5 We test our predictions with a 2x2 between-participants experiment in which MBA students proxy for nonprofessional investors. We manipulate the type of MW (entity-level vs. accountspecific) and whether or not management obtains an audit of the MW remediation disclosure (audited vs. not audited). In our experiment, participants first examine year-end audited information about a firm with a MW (either entity-level or account-specific), and then examine subsequent quarterly information which includes management s assertion that the previouslyexisting MW has been remediated. Participants in the audited condition also receive information indicating that the independent auditor audited management s MW remediation disclosure. Participants are then asked to judge the disclosure credibility and to report their willingness to invest in the firm. Overall, we find that management s choice to obtain an audit of the MW remediation disclosure increases disclosure credibility both for entity-level and for account-specific MWs, which in turn leads to an increased willingness to invest in the firm s stock. Our findings suggest that a voluntary audit of a MW remediation disclosure provides value for both entity-level and for account-specific MWs. The use of an experimental setting provides research advantages for our study. By using an experiment, we are able to hold constant the financial information communicated to participants. In this way, we can focus on the value provided by an audit. Moreover, using an experiment allows us to obtain a direct measure of user perceptions of disclosure credibility. These measures allow us to determine the reason for the investor s investment decision. Our study contributes to the literature on the value of an audit by examining the important area of internal control MWs and their remediation. Prior literature has focused on MWs remediation disclosures that have already been audited. It is not immediately apparent that 4

6 these audits provide value for several reasons. First, prior research suggests that investors may not always perceive an audit to add value (Hope et al. 2011; Houghton 1983; Coram et al. 2009). Second, regulations in other countries (besides the U.S. and Japan) suggest that management attestation on internal controls may be sufficient. Specifically, only the U.S. and Japan require an audit of internal controls, while other countries consider management s assessment of internal controls to be adequate. Finally, the audit of the MW remediation disclosure occurs at year-end, so an audit provided earlier in the year may not provide incremental value. We contribute to the literature by providing evidence on the value of an audit in addition to management attestation on the remediation of MW disclosures. Management disclosures about remediation of internal control MWs are unique in that they represent non-financial information that has a direct potential impact on financial information. While prior studies have examined the value of audits of financial information (Lennox and Pittman 2011) and non-financial information such as balanced scorecard disclosures (Coram et al. 2009), to our knowledge this is the first study to examine the value of an audit of internal control disclosures and to address the important issue of the credibility of management s assertion about remediation of internal control MWs. Our study addresses this issue by using Mercer s (2004) credibility framework in the context of a MW remediation. Our study also contributes to practice. Although companies have the option of obtaining an audit of their MW remediation disclosure, the benefits of purchasing an audit of these disclosures have not been assessed. Our results indicate that obtaining an audit of the MW remediation disclosure will enhance the disclosure credibility of the MW remediation disclosure and will increase investors willingness to invest in the firm. This is important because this result suggests that management s disclosure of the MW remediation is not sufficient and an 5

7 audit of the MW remediation disclosure provides value to investors. Importantly, investors assess value in an audit of MW remediation disclosures both when the MW is account-specific and when it is entity-level. Nevertheless, it is important to keep in mind that other considerations, such as costs, may be part of managers decisions of whether or not to obtain audits on their MW remediation disclosures. Likewise, risks and legal liability associated with providing such audits would be part of the auditors decision of whether to provide such audits. Our paper does not conclude that managers should obtain such audits, neither do we propose that auditors should provide such audits. We do, however, provide initial evidence on one important aspect of the issue considered by regulators, companies, and auditors with respect to AS4: whether investors would find such audits valuable. Our results suggest that, although early remediation of MW can be a positive signal to investors in itself, investors value the credibility-enhancing signal an audit provides; management s decision to obtain an audit enhances the credibility of the disclosure and increases the attractiveness of the company s stock. 3 Our findings are informative to audit committees and other company leaders in their decision whether to engage their auditors for this potentially costly service. Additionally, auditors will be interested to know how financial statement users assess the value of their service. U.S. regulators will also be interested in our results because our results will help them understand the value of an audit based on AS 4 and other standards. Our findings are of interest to regulators in other countries that may be considering switching to mandatory audits of internal controls over financial reporting. 3 Although the MW remediation will be audited at year-end, our results show that investors assign value to a credibility enhancement associated with the early MW remediation disclosure. 6

8 Our paper is organized as follows. In the next section, we examine the prior literature related to MWs and the value of an audit, and present our hypotheses. The following section describes the methodology, and is followed by a presentation of our results. In the final section of the paper, we discuss the implications of our findings, consider our study s limitations, and review avenues for future research. BACKGROUND AND HYPOTHESES DEVELOPMENT In the following sections, we review literature related to internal control deficiencies and MWs. We also examine prior literature regarding the type of MWs. We rely on prior literature regarding the value of an audit and credibility theory to develop our research hypotheses and the proposed mediation model. Background on SOX and AS 4 The Sarbanes-Oxley Act of 2002 (SOX) requires all public companies to report on the effectiveness of their internal controls and to obtain an audit of this evaluation. Management must disclose any MWs and auditors must audit this disclosure. Due to the high costs associated with these audits, it is important to understand the benefits of the audits. However, it can be difficult to assess the value of these audits since the audits are required of all public companies. 4 We exploit the context provided by AS 4 to examine this important research question. AS 4 is a result of the Sarbanes-Oxley Act of 2002 (SOX). SOX required all public companies to report on the effectiveness of their internal controls and to obtain an audit of this evaluation. If management determines that the internal controls are not effective (i.e., a MW exists), management must disclose this. However, management has the option to remediate the MW. If management remediates the MW prior to the release of the following year s 10-K, management 4 Krishnan and Yu (2012) examined the value of such audits for small public companies, but findings related to small companies do not necessarily apply to large companies. 7

9 has the option to obtain an audit that determines whether or not the MW is in fact remediated at any point prior to year-end. AS 4 provides guidance on such voluntary audits of MW remediation disclosures. The purpose of the audit is to provide reasonable assurance that the MW has been remediated (i.e., no longer exists). Therefore, the audit is focused on determining whether the controls management has implemented to remediate the MW are designed and operating effectively. The auditor only provides an opinion on the remediation of the MW specified by management, rather than an opinion on the internal controls over financial reporting as a whole. At the conclusion of the audit, the auditor issues a report on the effectiveness of the MW audited. Our study examines how the remediation process is enhanced when a company obtains a voluntary audit of the remediation of MWs. While prior studies have examined the value of audits of financial information (Lennox and Pittman 2011) and non-financial information such as balanced scorecard disclosures (Coram et al. 2009), to our knowledge this is the first study to examine the value of an audit of internal control disclosures and to address the important issue of the credibility of management s assertion about remediation of internal control MWs. Furthermore, prior research has only focused on the disclosure of MWs at year-end, prior research has not been able to disentangle the effects of the MW remediation disclosure from the audit of this disclosure. Therefore, we are able to isolate the value of audits of MW remediation disclosures to investors and whether or not management attestation on MW remediation disclosures is of sufficient value. Below we discuss prior literature related to MWs and the remediation of these MWs. 8

10 Hypothesis Development: Effect of MWs on Firm Value Research on internal control deficiency disclosures examines the market reaction to internal control deficiency disclosures, finding that MW disclosures have a negative impact on stock prices. For example, Hammersley et al. (2008) found that MW disclosures resulted in more negative stock market reactions especially when the MW is at the entity-level. 5 Rose et al. (2010) experimentally examined entity-level versus account-specific MW disclosures, and found that including less detailed discussion of the entity-level MW increases investment risk for entitylevel MW disclosures and decreases investment risk for account-specific MWs. Together, these studies indicate that entity-level MW decreases firm value due to increased risk perceptions associated with entity-level MWs. Other studies consider the disclosure of MWs together with the remediation of MWs, finding that accruals quality increases for firms that remediate previous deficiencies in comparison to firms that do not remediate the deficiencies (Ashbaugh-Skaife et al. 2008). Studies also examine the effects of remediation of different types of MW. For example, Goh (2009) finds that companies spend more time remediating entity-level MWs than account-specific MWs. Furthermore, Hammersley et al. (2012) find that companies with entity-level MWs are less likely to remediate the MW. Bedard et al. (2012) take a more specific approach to defining the type of MW, finding that MWs related to information technology, segregation of duties, reconciliations, tax, revenues, and inventory are most difficult to remediate. Together, these studies suggest that entity-level MWs have a smaller likelihood of being remediated. 6 Our study builds on earlier research on financial statement user perceptions of MW disclosures and MW remediation 6 Although the market studies report that the remediation is informative, these studies are focused on companies that are required to get an audit. Therefore, these studies are unable to disentangle the signal provided by the remediation from the actual remediation. Furthermore, these studies do not focus on information regarding quarterly information, while our study does. 9

11 disclosures. We examine the effect of management s choice to obtain a voluntary audit of its remediation of different types of MW on financial statement users perceptions of the company. In the next section, we develop a theoretical model to explore the value of an audit. Mercer s Credibility Framework and Proposed Theoretical Model In this section, we develop our theoretical model for the joint effects of a voluntary audit of management s MW remediation disclosure and the type of the remediated MW on investors willingness to invest in the firm. Our proposed theoretical model is shown in Figure 1. [insert Figure 1 here.] Mercer s (2004) framework defines disclosure credibility as investors perceptions of the believability of a particular disclosure. Disclosure credibility relates to the specific assertion communicated by management in the disclosure, while management credibility is a more enduring characteristic of the source of the disclosure: the company s management. Management credibility is based on past disclosures and is related to management s reputation for making credible disclosures. Mercer s framework for disclosure credibility proposes that the credibility of a specific management disclosure is determined by the following factors: (1) the degree of external and internal assurance, (2) disclosure characteristics, (3) management s credibility, and (4) management s situational incentives at the time of the disclosure. Our study examines the joint effects of a voluntary audit of management s MW remediation disclosure and the type of the MW. We therefore focus on two components of Mercer s credibility framework, external assurance and disclosure characteristics (specifically inherent plausibility), and we hold constant management s credibility and situational incentives. In the next section, we examine the first link of our theoretical model, external assurance. Value of External Assurance 10

12 In this section, we review related literature, and use Mercer s (2004, 2005) disclosure credibility framework to develop our hypothesis for the value of a company s choice to have an independent audit of its MW remediation disclosure. The International Framework for Assurance Engagements (IAASB 2010) suggests the purpose of an audit is to improve the credibility and quality of information for decision makers. Furthermore, the disclosure credibility framework posits that investors assess audited disclosures as more credible than unaudited disclosures (Mercer 2004). We examine prior literature regarding the value of an audit to develop our hypotheses. Prior research has investigated the value of auditing for enhancing required financial statement information. Watts and Zimmerman (1983) indicate that audits enhance the credibility of financial reports because they reduce information asymmetry. Studies show that a demand exists for audits, the stock market differentially prices independently audited information, and audits of financial information can reduce interest rates (Dopuch et al. 1997; Blackwell et al. 1998; Kim et al. 2011). Studies also examine the information conveyed by voluntary audits of financial statements. Lennox and Pittman (2011) compare the value of mandatory versus voluntary financial statement audits by examining the unique setting provided by a change in standards for private companies in the United Kingdom. Lennox and Pittman (2011) find that when a company voluntarily chooses to undergo a financial statement audit, it experiences a differential increase in credit rating compared to when the company was under a mandatory audit regime. Therefore, a company s choice to have audited financial statements sends out a signal that the company is low-risk. Furthermore, studies also examine the value of audits for public companies. Marshall et al. (2014) find that audits of financial statements reduce information asymmetry and result in more value relevant information. 11

13 Other studies examine the value of audits for voluntary disclosures, such as environmental disclosures. For example, Simnett et al. (2009) examine the manager s decision to obtain an audit of environmental disclosures, and found that managers who were interested in increasing the credibility of their reports and increasing their reputation were more likely to obtain an audit. Moroney et al. (2012) examine the impact of voluntary audits for environmental disclosures, finding that the quality of the environmental disclosures is enhanced when the disclosures are audited. Together, these studies suggest that audits enhance the value of both financial and voluntary environmental disclosures. We extend this research to examine the important area of internal control MW remediation a context in which disclosure about nonfinancial data can affect investors assessment of financial data. Consistent with this idea, we use disclosure credibility theory to predict how investors will respond to the audit of MW remediation disclosures. These studies suggest that management s choice to obtain a voluntary audit of its MW remediation disclosure should increase the credibility of the disclosure. The first link of our theoretical model reflects this hypothesis. We formally propose: H1: Management s choice to obtain an audit of the MW remediation disclosure increases disclosure credibility. While prior literature and theory supports our hypothesis, it is possible that the hypothesis would not be supported by empirical data. First, management s attestation regarding the MW remediation may be sufficient. Only the U.S. and Japan require companies to obtain audits of management s internal control disclosures. This suggests that investors may find management s attestation on the MW disclosure to be sufficiently credible, leading the audit to not add any additional value. Second, the MW remediation disclosure is audited at year-end. Given this 12

14 knowledge, as well as the fact that financial statements themselves are not audited at quarter-end, investors may not find an audit at quarter-end to be of value. Therefore, while we hypothesize management s choice to have an audit will affect disclosure credibility, we recognize that there is a contrasting possible result. Interactive Effects of External Assurance and Disclosure Characteristics Mercer s (2004) framework discusses the main effects of the determinants of disclosure credibility, but recognizes the likelihood that these factors may interact with one another. For example, management s incentives have been shown in prior studies to interact with management credibility and with disclosure characteristics (Hutton et al. 2003; Williams 1996). Management credibility has also been shown to interact with disclosure characteristics to affect disclosure credibility (Hirst et al. 1999). We extend these studies by examining the interaction of two components of the disclosure credibility framework: (1) the degree of external and internal assurance and (2) disclosure characteristics (specifically inherent plausibility). While prior research suggests that audits typically add value, other research suggests that this is not always the case, and that the value of an audit is contingent on other factors (Hope et al. 2011; Houghton 1983). Coram et al. (2009) examine the voluntary audit of balanced scorecard disclosures, finding that increases in stock prices resulted from the choice to have an audit in some contexts. Specifically, they found that obtaining an audit increased stock price estimates only when the balanced scorecard information being audited was consistent with management incentives due to investor perceptions of disclosure credibility. Similarly, Pflugrath et al. (2011) find that voluntary audits increase the credibility of environmental disclosures to a greater extent when the audit is provided by a professional accountant rather than a consultant specializing in sustainability due to investor perceptions of management credibility. These 13

15 studies suggest that an audit of voluntary disclosures may provide more value in certain contexts (i.e., for private firms, when voluntary disclosures are consistent with management incentives and when a professional accountant is used to audit voluntary disclosures). Given the importance of internal control to the integrity of financial information, it is therefore especially important to examine the value of an audit in the context of internal control MW remediation. We hypothesize that the influence of the voluntary audit on disclosure credibility will be dependent on the inherent plausibility (a disclosure characteristic) of the disclosure. We operationalize inherent plausibility by examining entity-level and account-specific MWs. Mercer (2004) provides the example of a reported earnings growth being more credible if it comes from a company with consistent previous earnings growth than if it comes from a company with consistent previous earnings decline. Specifically, the type of the MW could imply different levels of inherent plausibility. We outline several reasons for this relation in the following paragraph. Investors likely view entity-level MWs as leading to a higher likelihood of errors because they apply to the entire business, while account-specific MWs may be viewed as less prone to errors because they are contained to a single account or accounting cycle (Asare and Wright 2012, Asare and Wright 2011; Rose et al. 2010; Hammersley et al. 2008; Doyle et al. 2007). Further, it is more difficult to audit entity-level MWs leading to users associating greater verification risk with entity-level MWs (Hoag and Hollingsworth 2011; Hoitash et al. 2008; Moody s 2004; Credit Suisse First Boston 2005; Leech 2004; Nicolaisen 2004; Hammersley et al. 2008; Asare and Wright 2011; Asare and Wright 2012). Moreover, management override can be a bigger problem for entity-level MWs (Hoitash et al. 2008). As discussed earlier, extant research on MW has shown that entity-level MWs are also more difficult to remediate (Goh 14

16 2009; Hammersley et al. 2012; Bedard et al. 2012). Given the aforementioned characteristics of entity-level MWs, investors may not expect that the entity-level MW is in fact remediated. Thus, am entity-level MW is low in inherent plausibility. Therefore, when MW disclosure is related to an entity-level control, the investor is likely to view the audit as adding more credibility than when MW disclosure is related to an account-specific control because the entity-level MW s remediation is less inherently plausible. We therefore propose the following interaction hypothesis. H2: The presence of a voluntary audit will have a greater effect on disclosure credibility when the remediated MW is at the entity-level than when the MW is account-specific. Management s MW remediation is a disclosure of good news. Prior research has shown that bad news disclosures are always informative, but the informativeness of good news disclosures is dependent on disclosure credibility (Hutton et al. 2003). Studies indicate investors are sensitive to disclosure credibility (Hutton et al. 2003; Williams 1996, Mercer 2004) and that disclosure credibility is an important predictor of investors willingness to invest in a company (Jennings 1987; Healy and Palepu 2001; Mercer 2004). We therefore expect that greater credibility of the good news MW remediation disclosure will increase investors willingness to invest in the company s stock. Our formal hypothesis is stated below. H3: Greater credibility of the MW remediation disclosure increases investors willingness to invest in the company s stock. 15

17 METHOD Design Overview In our experiment, participants first examined year-end audited information about a firm with a MW in its internal controls, and then examined subsequent quarterly information which included management s assertion that the previously-existing MW had been remediated. Participants then reported their willingness to invest in the stock of the firm. The experiment manipulated between participants the type of the MW (entity-level vs. account-specific) and whether management s assertion in their quarterly report that the MW had been remediated was audited by the firm s independent auditor or not (audited vs. not audited). Participants and Task Participants in the experiment were MBA students at a private university. Prior research shows that MBA students are reasonable surrogates for nonprofessional investors (e.g., Elliott et al. 2007; Rose et al. 2010). The experiment was administered both in class sessions and online. 7 Participation in the study was voluntary. Upon completion of the study, each participant received a $5 amazon.com gift card and was entered into a lottery to win one of five $50 gift cards. Of the total of 96 participants who provided responses to the study, 95 participants provided usable responses to all dependent variables, and are used in all analyses. Of the 95 participants, 54 participated online and 41 participated in a classroom. An indicator variable for online participation was not significant for the analyses (all p >.73), so we performed our analyses on the full sample of 95 participants. Participants mean (median) level of experience was 10 (8) years; 72 percent of the participants were male. There were no significant demographic 7 The response rate for studies distributed online was 22% (54 usable responses returned from 246 invitations). 16

18 differences across experimental conditions, and none of the demographic variables were significant as covariates in any tests of hypotheses (all p >.18). 8 Participants provided informed consent to participate in the study, and then proceeded to work on the experimental task. Participants examined excerpts from the firm s 10-K report and the subsequent 10-Q report, and made decisions about investing in the firm s stock. Next, they assessed the credibility of the firm s management, responded to manipulation checks, and provided demographic information. Experimental Case The experimental case provided background information on a biopharmaceutical firm with successful commercial products and licensing agreements. 9 The firm had received an unqualified opinion on its year-end financial statements as indicated in the firm s 10-K. Participants reviewed excerpts from the firm s financial statements included in the 10-K and then read management s assessment of the effectiveness of the firm s internal controls. Specifically, the firm s management had concluded that internal controls over information technology did not operate effectively as of year-end, and this constituted a MW. Specifically, the MW resulted from a failure to maintain adequate password and access controls in the ERP system. Management described the details of the MW which were associated with a different type of MW (entity-level or account-specific). Following management s assessment, participants were presented with the independent auditors report on internal controls over financial reporting. In the report, the auditors described the MW, concluded that controls were not effective as of year- 8 Years of experience had a p-value of.182 and gender had a p-value of.193 when individually included as covariates in the analysis. The p-values for all other demographic variables included as covariates were greater than We provided positive financial information because negative information is always credible (Hutton et al. 2003). 17

19 end, and issued an adverse opinion on internal controls. There was no restatement associated with the MW. The subsequent quarterly report (for the first quarter of the following fiscal year) contained unaudited financial statement data and an update regarding the effectiveness of internal controls. Participants read an excerpt from management s disclosure which stated that during the first quarter management completed corrective actions to remediate the previously existing MW. Management asserted that the previously reported MW related to IT controls had been remediated as of the end of the quarter. This concluded the case for the not audited condition. Participants in the audited condition examined an auditors report related to the remediation of the MW, in addition to management s report. We describe the details in the following section Independent Variables Audit of the MW Remediation We manipulated whether management s MW remediation disclosure made in the quarterly report was audited (audited condition) or not (not audited condition) by an independent auditor. In the not audited condition, participants read an excerpt from management s report stating that the MW had been remediated, and that IT controls testing indicated the corrective actions had been implemented, and the controls were operating effectively. Participants in the audited condition were presented with an independent auditor s report regarding the remediation disclosure, in addition to management s report. In their report, the independent auditors concluded that the MW no longer existed as of the quarter s end because the IT controls testing indicated that corrective actions had been implemented and the controls were operating effectively. The wording in the auditor s report was based on the illustrative auditor s reports presented by the PCAOB in AS 4 (PCAOB 2006). 18

20 MW Type We manipulated the type of the MW in internal controls using the manipulation in Rose et al. (2010), resulting in two conditions: entity-level or account-specific MW. In the entity-level MW condition, the 10-K report stated that the MW resulted from a failure to maintain adequate password and access controls over the entire ERP system. In the subsequent 10-Q report, management stated that remediation changes were made to the entire ERP system. In the account-specific MW condition, the 10-K report stated that the MWs resulted from a failure to maintain adequate password and access controls over the portion of the ERP system that ensures proper recording of sales in the appropriate reporting period. In the subsequent 10-Q report, management stated that remediation changes were made to that same portion of the ERP system. Dependent Measures Disclosure Credibility Participants assessment of the credibility of the remediation disclosure (disclosure credibility) was measured with the question How likely is it that the reported MW was remediated as of 3/31/2012 by management? Participants responses were assessed on a 7-point Likert scale with anchors not at all likely and extremely likely. Willingness to Invest To ensure participants responses to disclosure credibility have an impact on future decisions, we also examine whether disclosure credibility affects participants willingness to invest. Participants willingness to invest was measured with the three questions used in Elliott et al. (2012). Participants responses for the first two questions were measured on 101-point scales with anchors 0 and 100. The first question was, How attractive is an investment in IMMUNTALITY Co.'s stock? (0 = very unattractive, 100 = very attractive). The second question 19

21 was, How likely are you to invest in IMMUNTALITY Co. s stock? (0 = very unlikely, 100 = very likely). The third question instructed participants to assume that they have $10,000 to invest in the biopharmaceutical industry, and asked them to indicate how much of the $10,000 they would invest in the company s stock. RESULTS Manipulation Checks The effectiveness of the manipulation of presence vs. absence of an audit of the MW remediation was tested with participants ratings of the statement The specific assertion made by IMMUNTALITY management that the previously reported MW was remediated as of 3/31/2012, was audited by IMMUNTALITY s independent auditor (1 = strongly disagree, 7 = strongly agree). Participants in the MW remediation-audited condition reported higher agreement (mean of 5.77) than participants in the MW remediation not audited condition (mean of 3.20), and the difference was significant (t = 8.01, p < 0.01, two-tailed), indicating success for this experimental manipulation. The effectiveness of the MW type manipulation was tested with participants ratings of the statements The material weakness reported in the 12/31/2011 financial statements of IMMUNTALITY Co. related to a portion of the Enterprise Resource Planning system and The material weakness reported in the 12/31/2011 financial statements of IMMUNTALITY Co. related to the entire Enterprise Resource Planning system (1 = strongly disagree, 7 = strongly agree). Participants in the account-specific MW condition agreed more that the MW related to a portion of the ERP system (mean of 5.09) than participants in the entity-level MW condition (mean of 3.94), and the difference was significant (t = 3.00, p < 0.01). Participants in the entitylevel MW condition agreed more that the MW related to the entire ERP system (mean of 4.23) 20

22 than participants in the account-specific MW condition (mean of 3.45), and the difference was also significant (t = 1.96, p = 0.05), indicating success of the MW type manipulation. ANOVA of Disclosure Credibility on Manipulated Variables Descriptive statistics for disclosure credibility are included in Panel A of Table 1. Participants assess the credibility of the MW remediation disclosure in the presence of an audit to be higher, both for an account-specific MW (mean = 5.48) and for an entity-level MW (mean = 5.04), as compared to the absence of an audit (mean = 4.46 and mean = 4.09, respectively). ANOVA results are shown in Panel B of Table 1. The main effect of a MW remediation audit on disclosure credibility is significant (F = 9.79, p = 0.002). This is consistent with H1. However, the interaction (F = 0.01, p = 0.916) is not significant, providing no support for H2. [insert Table 1 here] ANOVA of Willingness to Invest on Manipulated Variables Descriptive statistics for the three willingness-to-invest questions are presented in Panel A of Table 2. The three questions measured (1) participants rating of the attractiveness of the stock, (2) the likelihood that participants would invest in the stock, and (3) the amount participants would invest in the stock of the company if they had $10,000 to invest in the biopharmaceuticals industry. Following Elliott et al. (2012), we conducted a principal component factor analysis of the three individual questions. The three questions loaded on a single factor, which we label willingness to invest. The factor has an eigenvalue of 2.45, which explains 82 percent of the variance in participants responses. As in Elliott et al. (2012), we use the willingness-to-invest factor score to test our hypotheses. We present the ANOVA model for willingness to invest in Panel B of Table 2. The presence of an audit of the MW remediation disclosure significantly increases participants willingness to 21

23 invest in the company (F = 4.23, p = 0.043, two-sided). This indicates the presence of a direct effect of a MW remediation audit on willingness to invest (ignoring any mediation effects). The interaction term is not significant (F = 0.02, p = 0.965). [insert Table 2 here.] Model Results We test our theoretical model using path analysis, in which each variable is simultaneously regressed on all preceding variables in the model. Results from the path analysis are shown in Figure 2. The presence of a voluntary audit of management s remediation disclosure increases disclosure credibility (path coefficient of 1.02, p = 0.01), supporting H1. However, the interaction is not significant (path coefficient of -0.07, p = 0.46), indicating lack of support for H2. Finally, we observe that disclosure credibility significantly increases the willingness to invest in the company (path coefficient of 0.23, p < 0.01). 10,11 This indicates that participants judgments of disclosure credibility affected their investment decisions, supporting H3. [insert Figure 2 here.] Additional Analysis Mercer s (2004) disclosure credibility framework lists existing management credibility as one of the factors that influence current disclosure credibility. In the context of Mercer s (2004) disclosure credibility framework, management s prior behavior and existing reputation affect the credibility of the present disclosure. In our study, we hold constant management s prior behavior 10 The observed model fits the data well (chi-sq. = 0.96; CFI = 1.00; TLI = 1.05; RMSEA < 0.01; SRMR = 0.03). 11 We also conduct a mediation analysis using the bootstrapping technique described in Preacher and Hayes (2008). Using the Preacher and Hayes (2008) macro we find that the presence of an audit of the MW remediation significantly increases participants perception of disclosure credibility (path coefficient of 0.97, p < 0.01), and that disclosure credibility increases participants willingness to invest (path coefficient of 0.23, p < 0.01), while the path from audit of the MW remediation to willingness to invest is not significant (path coefficient of 0.17, p = 0.26). The mediation effect is significant (one-tailed p = 0.01), and we observe that the 5% bootstrap confidence interval for the indirect effect excludes zero. 22

24 and existing reputation across all conditions, and the only differences between conditions relate to the present disclosure made by management. Importantly, in a different study, Mercer (2005) examines the effects of the credibility of a disclosure on management s subsequent reporting credibility with investors. Specifically, Mercer (2005) finds that more credible disclosures increase management s subsequent, future credibility, at least in the short term. Since in our study we hold constant management s reputation for credibility prior to the disclosure, we can test whether the credibility of the current disclosure affects investors assessment of management s credibility in a way consistent with Mercer (2005). We measured participants assessment of management credibility with the Mercer (2005; Barton and Mercer 2005) six-item management credibility scale. 12 In a regression analysis (untabulated), we find that disclosure credibility significantly increases management s credibility (coefficient of 0.27, p < 0.01). DISCUSSION AND CONCLUSIONS In this paper, we assess the value of an audit of management s MW remediation disclosures by exploiting the AS 4 context. We investigate whether management s attestation on MW remediation disclosures is sufficient or whether obtaining an audit provides value. Specifically, we examine how a company s choice to obtain an audit of its MW remediation assertion influences disclosure credibility and investors willingness to invest in company stock. We further examine whether an audit brings greater value for entity-level MWs than for accountspecific ones. Our results show that a voluntary audit of management s disclosure that a previously existing MW has been remediated increases the credibility of management s disclosure and, ultimately, increases investors willingness to invest in the firm s stock. This effect is observed for both 12 The scale measures investors perceptions of two primary dimensions of management credibility: competence and trustworthiness (Mercer 2005). 23

25 entity-level and account-specific MW disclosures. Additional analysis shows that, consistent with Mercer (2005), the credibility of the current management disclosure impacts management s future reporting credibility with investors. Our study makes several contributions. To our knowledge this is the first study to examine the value of an audit of internal control MW remediation disclosures and to address the important issue of the credibility of management s assertion about remediation of internal control MWs. Prior studies have examined the value of audits of financial information (Lennox and Pittman 2011; Marshall et al. 2014) and non-financial information such as balanced scorecard disclosures (Coram et al. 2009). Management disclosures about remediation of internal control MWs are unique in that they represent non-financial information that has a direct potential impact on financial information. In addition, the context of AS 4 allows us to isolate the value of an audit at quarter-end and the potential value of an audit at year-end for MW remediation disclosures. Typically, MW remediation disclosures are always audited at year-end, so companies have no way of knowing the incremental value of the audit. Our study addresses the issue of whether the presence of an audit covering management s early remediation assertion influences investors judgments about the company s stock. Relatedly, the U.S. and Japan are currently the only countries that require internal control audits. The results of this study can be considered by regulators of other countries who may be considering requiring audits of internal controls. Our findings should be interpreted with caution, since they are subject to several limitations in addition to the general limitations of experimental research. While our study provides evidence on how a voluntary audit affects the investing decisions of nonprofessional investors, future research should examine whether our results would hold with analysts and other 24

26 professional investors. Future research can also examine whether the disclosure credibility effects observed in our study would increase management s reporting credibility with investors over the long run. Another fruitful avenue for future research would be to examine how the voluntary audit interacts with other factors from Mercer s (2004) credibility framework. For example, established strong management credibility may subdue the effect of an audit of a MW remediation disclosure. Similarly, management s situational incentives or the amount of supporting information provided by management in their remediation disclosure may interact with the credibility-increasing effect of a voluntary audit. Beyond its contributions to audit research, our study should also be informative to accounting practitioners and audit firms. Audit firms may be interested in the credibility-enhancing effect of their assurance and the value the marketplace assigns to their audits. Regulators, such as the PCAOB, may be interested in our findings to increase their understanding of the relevance of AS 4, which is one of the newer auditing standards. Regulators outside the U.S. will benefit from the results of the study by having evidence that management attestation on internal controls is not adequate and that an audit provides value for these disclosures. Finally, in determining whether to engage the auditors for a voluntary audit of a MW remediation, company leadership can consider the potential value of such an audit regarding the credibility of management s remediation assertions and, ultimately, on the attractiveness of the company s stock to the public. While the results seem promising with respect to the value of a voluntary audit of management s MW remediation assertions, future research is needed to examine the cost-benefit effect of a decision for such an engagement. Our study contributes to the literature on the value of an audit by examining the important area of internal control MWs and their remediation. Prior literature has focused on MWs 25

27 remediation disclosures that have already been audited. It is not immediately apparent that these audits provide value for several reasons. First, prior research suggests that investors may not always perceive an audit to add value (Hope et al. 2011; Houghton 1983; Coram et al. 2009). Second, regulations in other countries (besides the U.S. and Japan) suggest that management attestation on internal controls may be sufficient. Specifically, only the U.S. and Japan require an audit of internal controls, while other countries consider management s assessment of internal controls to be adequate. Finally, the audit of the MW remediation disclosure occurs at year-end, so an audit provided earlier in the year may not provide incremental value. We contribute to the literature by providing evidence on the value of an audit in addition to management attestation on the remediation of MW disclosures. Management disclosures about remediation of internal control MWs are unique in that they represent non-financial information that has a direct potential impact on financial information. While prior studies have examined the value of audits of financial information (Lennox and Pittman 2011) and non-financial information such as balanced scorecard disclosures (Coram et al. 2009), to our knowledge this is the first study to examine the value of an audit of internal control disclosures and to address the important issue of the credibility of management s assertion about remediation of internal control MWs. 26