DOI: 10.1111/1475-679X.12099 Journal of Accounting Research Vol. 54 No. 1 March 2016 Printed in U.S.A.

Internal Audit Quality and Financial Reporting Quality: The Joint Importance of Independence and Competence L A W R E N C E J . A B B O T T ,∗ B R I A N D A U G H E R T Y ,∗ S U S A N P A R K E R ,† A N D G A R Y F . P E T E R S‡ Received 5 November 2013; accepted 24 September 2015

ABSTRACT

In light of the growing importance of internal audit functions (IAF) and the limited archival evidence on internal audit quality, we examine an interactive model of IAF quality (comprised of competence and independence) to better understand the determinants of IAF effectiveness as a financial reporting monitor. Our tests support the hypothesis that the joint presence of competence and independence is a necessary antecedent to effective IAF financial reporting monitoring. In sum, our results show that, the answer to “what is the effect of internal audit competence (independence) on financial reporting quality?” is “it depends on the independence (competence) of the internal auditor.” Our study extends the understanding of IAF quality determinants in the realm of financial reporting as it relates to ongoing discussions by researchers, standard setters, regulators, and practitioners.

∗ University of Wisconsin–Milwaukee; † Santa Clara University; ‡ University of Arkansas. Accepted by Philip Berger. We are grateful for assistance from Tim Seidel and helpful comments from Cory Cassell, Dana Hermanson, Chris Hines, Adi Masli, Bill Messier, Marcy Shepardson, David Wood, and workshop participants at Santa Clara University, University of Arkansas, Texas Tech University, the 2013 AAA Annual Meeting, and the 2012 Auditing Section Midyear Conference. 3 C , University of Chicago on behalf of the Accounting Research Center, 2015 Copyright 

4

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS

JEL codes: D83; G39; M12; M41; M42 Keywords: internal audit; financial reporting quality; auditor independence; auditor competence

1. Introduction In 2013, the NASDAQ Stock Market LLC (NASDAQ) proposed a rule change that would require all NASDAQ registrants to maintain an internal audit function (IAF) (NASDAQ [2013]).1 The New York Stock Exchange (NYSE) has required registrants to maintain an IAF since 2006. The rationale for these requirements is that an effective IAF provides the audit committee and other financial reporting stakeholders with critical information pertaining to a company’s risks (including financial reporting risks) and internal controls (i.e., Harrington [2004], NASDAQ [2013]). Similarly, corporate governance proponents consistently emphasize the IAF’s role in enhancing financial reporting quality (Coram, Ferguson, and Moroney [2008], Prawitt, Smith, and Wood [2009], Cohen, Krishnamoorthy, and Wright [2010]). Nonetheless, the IAF’s role in the financial reporting process is not yet fully understood and empirical evidence concerning the impact of IAF quality is minimal. We investigate the potential impact of IAF quality as a joint function of the IAF’s competence and independence. We base this view upon the theoretical work of DeAngelo [1981], who notes that external audit quality is a function of the external auditor’s ability (i.e., competence) to detect accounting misstatements and willingness (i.e., independence) to oblige proper accounting treatments. Similar to DeAngelo [1981], we assume that, within the internal audit function, competence and independence are important and distinct constructs that must interact to result in quality outcomes. Despite the intuitive appeal of IAF quality positively impacting financial reporting quality, prior empirical evidence is not as strong as the intuition would suggest. For example, Prawitt, Smith, and Wood [2009], using data from 2000 to 2005, document mixed evidence between an overall composite measure of IAF quality and financial reporting quality. Their single composite IAF quality measure adds several, equally weighted individual IAF characteristics of competence and independence. One potential explanation for prior mixed results revolves around how these IAF characteristics interact with each other in creating IAF quality. Rather than an additive relation between competence and independence whereby an increase in competence can compensate for decreased independence, IAF quality 1 The NASDAQ subsequently tabled the proposal as registrants expressed concerns over the costs and benefits of the proposal. An objective of the current study is to provide evidence concerning one aspect of the benefits.

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY

5

may be more appropriately described as an interactive, two-factor function of both competence and independence. Thus, IAF competence (independence) may be unlikely to impact financial reporting quality unless it is in the presence of IAF independence (competence). In this paper, we develop and test a two-factor model of IAF quality as a function of the IAF’s ability to prevent/detect financial misstatements (i.e., competence) and its inclination to report the misstatements to the audit committee and/or external auditor (i.e., independence). Our study uses survey evidence from 189 Chief Internal Auditors (CIAs) from Fortune 1000 companies during fiscal 2009.2 From this uniquely detailed and rich set of IAF data, we are able to create separate measures of IAF competence and independence. Our measure of IAF competence is based on the average hourly rate of budgeted IAF resources (Abbott, Parker, and Peters [2012]). We also identify three potential factors related to IAF independence. First, consistent with Abbott, Parker, and Peters [2010, 2012], we measure the audit committee’s IAF influence vis-`a-vis management’s IAF influence across multiple oversight dimensions.3 If upper management (rather than the audit committee) wields greater IAF influence, this may cause the IAF to fear reprisal should the IAF question management’s financial reporting decisions, and thus diminish the IAF’s objectivity or independence (e.g., Cohen, Krishnamoorthy, and Wright [2010], Norman, Rose, and Rose [2010]). In addition to audit committee IAF influence, a review of the IAF literature reveals two other potential threats to IAF independence: whether the IAF serves as a management training ground (MTG) and the sizeable presence of an IAF outside service provider (OSP). In particular, when the IAF is used as an MTG, internal auditors may be more reluctant to report financial reporting issues in an effort to ingratiate themselves to upper management (Messier et al. [2011], Christ et al. [2015]). The presence of an OSP may create job security concerns for the in-house IAF, as registrants may find the variable cost nature of OSPs attractive as a means of achieving cost savings and/or financial reporting flexibility (Abbott et al. [2007]). We regress our IAF-related variables and a set of control variables against common measures of financial reporting quality, abnormal accruals 2 As compared to other extant IAF research, our sample follows the implementation of the Public Company Accounting Oversight Board’s (PCAOB) Auditing Standard No. 5 on the effectiveness of internal control over financial reporting (ICFR), a period when substantial IAF resources were diverted to ICFR testing. 3 In contrast, prior research commonly utilizes a dichotomous influence variable based upon whether the IAF’s formal reporting line is to the audit committee (e.g., Prawitt, Smith, and Wood [2009]). However, Abbott, Parker, and Peters [2010] note that 96% of Chief Internal Auditors agreed with the statement “the Internal Auditor reports to the audit committee.” As such, the utilization of the formal reporting arrangement as the sole IAF independence characteristic leads to the possibility that the dichotomous identification of the reporting line could simply capture a ceremonial structure and may not be indicative of a significant degree of IAF independence.

6

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS

(Kothari, Leone, and Wasley [2005], Prawitt, Smith, and Wood [2009]), and whether the firm just meets or beats analyst forecasts (Koh, Matsumoto, and Rajgopal [2008], Prawitt, Smith, and Wood [2009], Mande and Son [2012]). Our abnormal accruals are further segregated into income-increasing and income-decreasing abnormal accruals. Consistent with IAF quality being a two-factor, interactive function of independence and competence, we document several statistically significant relations between financial reporting quality and our interacted competence and independence variables. We find that interactions between the IAF competence and (1) the relative degree of audit committee IAF oversight and (2) the lack of a substantial OSP presence curtail both income-increasing and income-decreasing abnormal accruals. We obtain similar results when we examine the firm’s proclivity to just meet/beat analysts’ forecasts. In contrast, when independence is proxied by whether the IAF is not used as an MTG, the interaction between the IAF independence factor and competence exhibits a statistically significant, mitigating impact on incomedecreasing abnormal accruals. This finding is consistent with lower likelihoods of IAF reporting of inappropriate income-decreasing opportunistic reserve behaviors when the IAF is used as an MTG. In this manner, internal auditors hoping to move into a non-IAF position may endeavor to ingratiate themselves to management, or demonstrate their ability to be a team player when they perceive a lower downside of doing so.4 Our paper contributes to the IAF literature in several ways. Our study is the first to establish IAF characteristics as separate, distinct constructs that act jointly in creating IAF quality. In doing so, this study contributes to our understanding of IAF quality and the determinants of the IAF as an effective internally based financial reporting monitor. Second, our results suggest that there are at least three factors that can impact IAF independence and that these factors have differential interactive effects with IAF competence in influencing financial reporting quality. In contrast, prior IAF literature generally uses a dichotomous, single-variable independence measure and implicitly ignores other potential independence determinants (e.g., Ahlawat and Lowe [2004]). Moreover, we find that our IAF independence characteristics are either not correlated or only weakly related to each other or to our IAF competence measure. This is consistent with (1) IAF independence being characterized as a multifaceted attribute with at least three determinants and (2) IAF independence and competence being separate and distinct constructs. While there is a very rich literature on threats to external auditor independence, there is a paucity of archival evidence on the potential determinants of internal auditor independence. With respect to internal 4 In terms of income-increasing abnormal accruals, internal auditors whose IAF serves as a management training ground may simply defer auditing more contentious, higher profile income-increasing abnormal accruals to the external auditor. Internal auditors may do so with the knowledge that external auditors are likely to expend greater audit effort on incomeincreasing accruals (Abbott, Parker, and Peters [2006]).

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY

7

auditor competence, we also provide descriptive evidence of the qualifications of the IAF staff and Chief Audit Executives (CAEs) in large firms, and the types of functions those staff carry out. A more detailed understanding of the factors that influence IAF independence and competence should be of interest to researchers, standard setters, regulators, and practitioners in their efforts to hone the IAF’s role as a financial reporting monitor. Our study also provides archival evidence on the impact of third-party IAF outsourcing on in-house IAF independence. Prior research has treated outsourcing as a threat to in-house IAF independence when the provider of such services was the external auditor (prior to the Sarbanes-Oxley Act of 2002 (SOX) when such service were allowed) (e.g., Ahlawat and Lowe [2004], Abbott et al. [2007], Prawitt, Sharp, and Wood [2012]). The evidence provided herein indicates that job security concerns can also be created by a large outsourcing presence unrelated to the external auditor. Furthermore, we find that over 65% of our sample IAFs have an outsourcing agreement in place and that close to 33% of our sample IAFs have a sizeable portion (i.e., in excess of 20%) of their budget allocated to OSPs. Given the pervasiveness of OSP-provided IAF services, we believe that future researchers should consider the nature and magnitude of OSPs in analyzing IAF independence. Finally, we provide evidence that our IAF competence proxy, the hourly in-house IAF rate, effectively summarizes several key IAF traits used in prior literature that individually or collectively represent IAF quality, such as tenure and certification (e.g., Prawitt, Smith, and Wood [2009], Lin et al. [2011]). Our competence variable has several attractive features for use in future research as it (1) is continuous, (2) allows for comparison between IAFs of varying size, (3) does not require researcher judgment regarding the weighting of IAF traits, and (4) does not require subjective judgment by IAF respondents when collected via survey instrument. Our study has several potential limitations. We draw survey data from a single year, and 2009 was a year of significant economic events that may have affected the generalizability of our study in unknown ways.5 In addition, discretionary accruals as a measure of management financial reporting discretion, though widely used, are likely to measure the underlying construct with error. Our test variables, similarly, are likely to contain measurement error. We have conducted a number of validity tests that are detailed in the sensitivity discussion, and support our IAF variables and results. Last, common to examinations involving internal attributes of an organization’s financial reporting environments, it is often difficult to disentangle the potential endogeneity of the roles that reside within an organization (such as the IAF, Audit Committee, and Executive Suite). Nonetheless, we believe our study offers insights into the operations of an increasingly important component of the internal control systems of large companies. 5 We address this issue by conducting our analysis over the years 2010–2011, assuming that our test variables would remain consistent over a short horizon. We find that our results continue to be significant over that time frame.

8

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS

2. Background and Prior Literature Prior IAF research investigating audit outcomes generally focuses on the external auditor’s reliance on the IAF for financial statement audit assistance (e.g., Messier et al. [2011], Prawitt, Sharp, and Wood [2011], Abbott, Parker, and Peters [2012], Bame-Aldred et al. [2012]). This stream of research focuses on the two most prevalent IAF characteristics that extant audit standards guide external auditors to consider: competence and independence.6 Competence generally refers to the auditor’s ability to perform tasks diligently and in accordance with professional standards (e.g., IAASB [2013]). The Institute of Internal Auditors (IIA) defines competence as “the ability of an individual to perform a job or task properly, being a set of defined knowledge, skills and behavior” (IIA [2013, p. 2]). Within an IAF setting, independence is defined as “the freedom from conditions that threaten the ability of the internal audit activity to carry out internal audit responsibilities in an unbiased manner” (IAASB [2013]). In other words, independence is often framed as objectivity or as the means to protect against bias, conflict of interest, or undue influence of others that would override professional judgments. Prawitt, Smith, and Wood [2009] is the first archival study to link IAF quality to financial reporting quality. The IAF quality measure used in Prawitt, Smith, and Wood [2009] is a single, additive composite, comprised of equally weighted metrics representing experience, certification, training, IAF reporting structure, time spent on financial activities, and relative IAF size. Using data from 2000 to 2005, the authors find that their composite measure of IAF quality is associated with mitigation of incomedecreasing accruals, but not income-increasing accruals. When they disaggregate their IAF characteristics, they find positive relationships between financial reporting quality and the IAF’s professional certifications (income-decreasing accruals) and IAF size relative to industry (incomeincreasing accruals).7 Prawitt, Smith, and Wood [2009] do not find significant associations between the IAF independence characteristic (whether the IAF reports to the audit committee) and financial reporting quality.8 While their composite measure of IAF quality includes facets of competence and independence, it is unclear when both of these characteristics are present for a given firm and whether their relationship is interactive or additive. 6 The AICPA’s Statement of Auditing Standards 65 (SAS 65) discusses indicators of IAF competence and independence; however, very little overlap exists among the factors that SAS 65 describes as competence-related versus independence-related (AICPA [1991]). 7 They explain the unexpected result on IAF size as the possibility that IAF size proxies for the difficulty of monitoring the firm, and, when monitoring is more difficult, managers may be able to find avenues to exercise more discretion to increase income. 8 During the sample period involved (2000–2005), Prawitt, Smith, and Wood [2009] find significant variation in this measure: 69% of sample IAFs reported to the audit committee and, as such, this was likely to be an appropriate and parsimonious proxy for IAF independence.

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY

9

Recent research points to other potential determinants of IAF independence and measures of how audit committees influence IAF independence, such as outsourcing presence, MTG usage, and perceived audit committee influence. (Quarles [1994], Abbott et al. [2007], Messier et al. [2011], Abbott, Parker, and Peters [2012]). In sum, prior research has provided only limited evidence on the impact of IAF quality. Furthermore, a number of IAF competence and independence measures have yet to be linked on a stand-alone (or interactive) basis to financial reporting quality. Finally, the IAF profession has also identified other mechanisms that can influence IAF independence that have yet to be investigated.

3. Hypothesis Development 3.1

IAF MECHANISMS FOR IMPACTING FINANCIAL REPORTING QUALITY

The focus of this paper is on the association between IAF quality and financial reporting quality. Therefore, an important antecedent to our research question is a description of the specific mechanisms by which an IAF can influence financial reporting quality. We posit that these opportunities arise within at least four activities: assisting with the financial statement audit, financial statement audit of subsidiaries, compliance auditing, and special consulting projects.9 Within the areas of financial statement audit assistance and audits of subsidiaries, the IAF performs specific audit procedures that allow it to potentially influence accrual decisions. These include review of the financial closing process, reviewing procedures for nonstandard journal entries and postclosing adjustments, and specific reviews of critical accruals such as accounts receivable valuation and inventory reserves (IIA [2005]). Compliance auditing can involve testing transactions or journal entries for compliance with the company’s financial reporting policy. Special consulting projects may also involve the IAF delving into accounting matters that require greater judgment on the part of the preparer, such as asset impairments, warranty reserves, collectability reserves, and/or inventory write-downs (PwC [2009]). It should be noted that, in any of the prior tasks, the IAF may encounter high-level accruals choices. More specifically, the IIA advises that, as a part of the quarterly financial reporting process, the IAF should engage the IAF should review the policies, procedures, and process for reporting and related disclosures. In particular, the IIA advises that, as a part of the quarterly financial reporting process, the IAF should engage “special or specifically targeted reviews of high-risk, complex, and problem areas; including material accounting estimates, reserve valuations, off-balance sheet activities, major subsidiaries, joint ventures, and special purpose entities” (IIA [2005, p. 236]). 9 We developed this list from a review of the prior literature, examination of IAF professional guidance, and discussion with CIAs. As discussed in our results section, respondents indicate a nontrivial budget allocation to these four activities.

10

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS

3.2

AUDIT COMMITTEE IAF OVERSIGHT IN THE INTERACTIVE IAF QUALITY MODEL

This study’s examination of IAF quality—and its potential impact on financial reporting quality—asserts that competence and independence are necessary, but not individually sufficient determinants of IAF quality. We investigate the potential impact of IAF quality as a joint function of the IAF’s competence and independence. While it is expected that a competent auditor is more likely to discover a financial reporting misstatement, the reporting of the discovered misstatement is contingent upon the auditor’s independence or objectivity. Similar to DeAngelo [1981], we assume that competence and independence are distinct constructs that must interact to ultimately impact financial reporting quality. Our first hypothesis pertains to the interaction of IAF competence and IAF independence, measured by audit committee or C-Suite influence. Both external and internal auditing standards assert that internal auditor independence is a direct function of the reporting relationship between the audit committee and the IAF (IIA [2002], AICPA [2013]). However, even though official reporting by the IAF to the audit committee is common, the extent and effects of active oversight by the audit committee and unofficial oversight by management varies greatly (e.g., Ernst & Young [2008], Mabry and Schwartz [2008], Hoffelder [2012]). Moreover, prior academicians highlight the inherent conflicts of IAFs serving both management and the audit committee (Hermanson [2002], Anderson [2003], Hermanson and Rittenberg [2003]). Prior research supports an association between greater audit committee oversight and greater independence for the IAF. For example, greater audit committee oversight of the IAF is associated with the greater shielding from possible management pressure (Quarles [1994], Carcello, Hermanson, and Raghunandan [2005]).10 If the IAF is not sufficiently shielded from possible management pressure, the CEO or CFO can reduce the likelihood that the issue will be reported to the proper channel. When the departure from reporting policy originates at the C-Suite level, the IAF’s independence is particularly important. A lack of independence of the IAF from the C-Suite may preclude the IAF from initiating a review of financial reporting choices made at the CFO/CEO level. This may prevent the discovery and reporting of financial reporting policy departures.11

10 For example, the internal auditors may be charged with reviewing the proper application of the firm’s policy for recording an inventory reserve. Divisions within the firm with separate profit targets may have incentives to either under- or overstate the reserve in order to maximize internal rewards. 11 This mechanism may operate in two ways. First, the CEO/CFO may effectively pressure the IAF to “filter” any IAF-generated reports that are forwarded to either the audit committee or external auditor. Second, to the extent that the CEO/CFO can influence IAF budgets or activities, the CEO/CFO may direct the IAF toward other activities that have an immediate, beneficial impact on current earnings such as the examination of vendor rebates or other

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY

11

Following our expectations, we posit that, as audit committee IAF influence (IAF independence) increases, the likelihood that a misstatement discovered by a competent IAF is either (1) properly reported to the audit committee and external auditor or (2) corrected by the party responsible for the original misstatement also increases. Both scenarios lead to greater financial reporting quality. This leads to our first hypothesis (stated in alternative form): H1: The interaction between IAF competence and audit committee IAF influence is positively associated with financial reporting quality.

3.3

THE IAF AS AN MTG IN THE INTERACTIVE IAF QUALITY MODEL

The second IAF independence determinant interacted with IAF competence is the organization’s use of the IAF as an MTG. Prior research suggests that the use of IAF as an MTG is prevalent among corporate entities (Stewart and Subramaniam [2010], Messier et al. [2011], Christ et al. [2015]). Many organizations see this as a means to attract and develop corporate talent by instituting both formal and informal rotations within the IAF (Ernst & Young [2008]). Despite these intended benefits, other IAF constituents see this as a potential threat to the independence of the IAF (Goodwin and Yeo [2001], Ahlawat and Lowe [2004], Christopher, Sarens, and Leung [2009]). Although concerns about both the costs and benefits of this practice remain, the use of the IAF as an MTG is common (Prawitt, Smith, and Wood [2009]). To date, few studies directly investigate the impact of MTGs on IAF independence (Stewart and Subramaniam [2010]). Messier et al. [2011] document a negative association between the use of the IAF as an MTG and the external auditor’s assessment of IAF independence. Despite the negative effect on perceived independence, MTGs were not associated with perceived differences in internal audit competence. This provides further evidence suggesting that independence and competence are separate and distinct constructs. When the IAF is an MTG, the IAF may be less likely to report the financial misstatement to the appropriate channel. If the misstatement reflects negatively on the division’s management, those managers may be less likely to offer that particular internal auditor a position within that division. Positive internal references may be less likely for an internal auditor seeking a position at a different division within the same company. In order for a misstatement to be corrected prior to the consolidation of results at the parent level, both IAF competence and independence must be present. As with our prior hypothesis, we consider the financial reporting quality

operational concerns, etc. (Abbott, Parker, and Peters [2010]). In doing so, the IAF is effectively precluded from collecting evidence that might serve to repudiate other compliance deviations involving financial reporting decisions.

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS

12

consequences of the joint importance of IAF independence and competency when independence is potentially conditioned by the explicit use of IAF as an MTG. This leads to our second hypothesis (stated in alternative form): H2:

The interaction between IAF competence and an IAF that is not used as an MTG is positively associated with financial reporting quality.

3.4

OUTSOURCED INTERNAL AUDIT ACTIVITIES IN THE INTERACTIVE IAF QUALITY MODEL

The final IAF independence factor interacted with IAF competence is the presence of a substantial OSP. Outsourced IAF activities represent ongoing sourcing strategies for many internal audit departments (Stewart and Subramaniam [2010]). Outsourcing all or a portion of the IAF provides management with the flexibility of adjusting IAF costs during the course of the year because outsourced IAF hours are essentially variable in nature. In contrast, in-house IAF hours are fixed in nature due to salary structures. Moreover, outsourced IAF activities can allow management to defer expense recognition into subsequent accounting periods by simply scheduling outsourced IAF activities into following years. This contracting flexibility may be especially attractive to management when the firm is faced with revenue shortfalls. Despite potential benefits of IAF outsourcing, there may be a cost in the form of decreased, in-house IAF independence. In particular, Abbott et al. [2007] and Quarles [1994] note that a substantial IAF outsourcing presence may undercut the in-house IAF’s willingness to confront management on issues due to concerns over job status. If an internal auditor feels “replaceable” as a consequence of outsourcing, he may be less willing to report a misstatement to the appropriate outlet for fear of losing his job within the firm (Quarles [1994], Abbott et al. [2007]). As before, increased independence, in this case the lack of a substantial IAF outsourcing presence, increases the likelihood that a misstatement discovered by a competent IAF is properly reported or corrected by the party responsible for the misstatement.12 As already stated, we consider the financial reporting quality consequences of the joint importance of independence and competency when in-house IAF independence is potentially influenced by the presence of outsourced IAF services. This leads to our third hypothesis (stated in the alternative): H3:

The interaction between IAF competence and the lack of a substantial IAF outsourcing presence (increased IAF independence) is positively associated with financial reporting quality.

12 Though we feel the preponderance of evidence points to a positive association between reporting quality and the lack of an OSP, we acknowledge that it is possible that the use of an OSP might instead serve to spur the IIA to greater effort and independence in order to prove their worth to the organization.

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY

13

4. Research Design 4.1

SAMPLE SELECTION

Our study utilizes a survey questionnaire (see appendix) sent to 909 nonbank members of the FORTUNE 1000 (in terms of total sales).13 Consistent with most prior internal audit research (e.g., Pelfrey and Peacock [1995], Scarbrough, Rama, and Raghunandan [1998], Raghunandan, Read, and Rama [2001], Carcello, Hermanson, and Raghunandan [2005], Abbott, Parker, and Peters [2010, 2012]), the survey targets Chief Internal Auditors (CIAs) or Chief Audit Executives (CAEs). The survey includes questions about the types of IAF services provided (whether in-house or outsourced), assistance to the external auditors, presence and activities of any OSPs, the reporting relationship between the IAF and the audit committee, and whether the IAF serves as an MTG. We asked recipients to provide responses based upon fiscal year 2009. The first survey mailing (sent October 2009) resulted in a total of 118 usable responses. A follow-up mailing (December 2009) produced an additional 99 usable responses, for a total of 227.14 However, we were unable to obtain complete Compustat data for 38 of these firms, bringing our total sample to 189. Table 1 provides a distribution of observations by twodigit focus industry membership (Hogan and Jeter [1999]). To test for potential nonresponse bias, we compare the characteristics of the early and late responders with each other and test for significant differences in size, leverage, return on assets, presence of a loss, and cash flow from operations. None of the differences are significant.15 We also compare our respondents to the industry makeup of the original population of nonbank Fortune 1000 firms. We conduct a test of differences of proportions comparing the industry sample representation to the population and find only the underrepresentation of the energy and manufacturing sectors to be significant.

4.2

CHARACTERISTICS OF THE IAF

Our survey responses provide information about the current nature of the IAF function in terms of qualification and activities. We document that the overwhelming majority (over 80%) of CAEs have a CPA certification 13 Banks are excluded since they do not possess inventory and have unique regulatory environments. We identified 909 nonbank firms within the Fortune 1000. 14 Our effective response rate of 20.7% (or 189 responses/909 total companies) compares favorably with the 12.7% rate obtained by Felix, Gramling, and Maletta [2001] in their internal audit study. However, the Felix et al. rate may have been depressed by the need for survey responses from both the internal and external auditors to constitute a complete sample pair. Our response rate is higher than that reported in previous studies of senior managers. For example, Graham and Harvey [2001] report a 9% response rate. 15 As with all surveys, there is a possibility of unknown response bias and of incorrect responses. While we tested our early and late respondents for differences on a number of dimensions, as previously discussed, it is possible that undetected bias is present.

14

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS TABLE 1 Sample Selection Results

Focus Industry Construction Consumer product & food Energy Financial services Information & Communication Manufacturing Personal services & health care Professional, commercial services, education Real estate Retail & wholesale Transportation All other Totals

Related TwoDigit SIC Codes

No. of % of No. of Fortune % of Fortune 1000 Firms∗ Firms Sample 1000 Firms∗

15–17 20–33

9 43

4.8 22.8

17 180

1.9 19.8

10–14, 46, 49 60–64, 67 48, 73, 78, 79, 84

32 11 20

16.9 5.8 10.6

172 62 126

18.9 6.8 13.9

34–39 72, 80, 83

31 2

16.4 1.1

140 28

15.4 3.1

75, 76, 82, 87, 89

1

0.5

11

1.2

65, 70 50–59 40–42, 44, 45, 47 1, 2,7, 8, 99

1 35 3 1 189

0.5 18.5 1.6 0.5 100

6 149 11 6 908

0.7 16.4 1.2 0.7 100

∗ Banks are excluded from both the population and the sample as these firms face additional regulation unique to their industry and do not have significant inventory accounts (making the model unfit for regression purposes). The remaining Fortune 1000 firms (in terms of total assets) serve as our sampling population.

while, perhaps surprisingly, only 39% have a Certified Internal Auditor (CIA) certification. Responses also indicate that the CISA certification is relatively rare among CAEs (at approximately 10%). As a reflection of the changing and increasing importance of systems-based auditing, we find that the CISA certification is much more prevalent (16%) among staff. At the staff level, the CPA certification dominates, though less than half of staff (40%) have a CPA certification, while over a quarter of staff have a CIA certification (28%). The CPA-CIA certification “gap” is lower for staff (40% vs. 28%) than it is for CAEs (82% vs. 39%). The CPA-CIA certification gap indicates that the current generation of CAEs was most likely hired from external audit firms. Experience in financial statement auditing is strongly correlated (0.88) with the possession of a CPA. We believe this provides information about the composition of the IAF at both the staff and executive level. We also provide descriptive statistics about what activities IAFs are performing. We note that 100% of survey respondents allocated a portion of their annual IAF budget to Section 404–related work. Second, the mean percentage of in-house IAF budgets allocated to Section 404 assistance was 27.5%—the highest budget allocation percentage among IAF activities. Sample results indicate that over 65% of our sample IAFs have an outsourcing agreement in place and that, on average, close to 31.8% of outsourced IAF budgets are dedicated to Section 404 assistance. Thus, if an outsourced

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY

15

IAF is utilized, it is utilized heavily to provide Section 404 assistance. This result is consistent with our contention that a significant OSP presence could represent a threat to internal IAF independence since there is a large overlap in the services that are provided by in-house IAFs vis-`a-vis OSPs.

4.3

REGRESSION MODEL

To test the relationship between financial reporting quality and the interactive nature of IAF competence and independence, we adapt the following regression model from prior research: ABNACC = b 0 + b1 IACOMP + b2 ACIAFINF + b3 NONMTG + b 4 NONOSP20 + b5 IACOMP ∗ ACIAFINF + b 6 IACOMP ∗ NONMTG + b7 IACOMP ∗ NONOSP20 + b 8 ASSETS + b9 AGE + b10 LEVERAGE + b11 SEGNUM + b 12 CFO + b13 SALESGROW + b14 MTB + b15 CFOVOL + b 16 ROA + b17 LOSS + b18 MATWEAK + b 19−28 INDUSTRY + ε.

(1)

The variables used in the empirical models are summarized in table 2 and discussed below.

4.4

DEPENDENT VARIABLE

Following prior audit literature, we utilize abnormal accruals (ABNACC) as a proxy for financial reporting quality (Francis [2011]). To measure abnormal accruals, we use the performance-adjusted cross-sectional variation of the modified Jones model (Dechow, Sloan, and Sweeney [1996]) as reported by Kothari, Leone, and Wasley [2005].16 The Kothari et al. model includes both an intercept term and a measure of performance. Consistent with prior research, we first estimate the model for firms with information available on Compustat for 2009 (excluding financial institutions), and apply the results by two-digit SIC code to the calculation of abnormal accruals for the firms in our sample.17 Our estimate of abnormal accruals is the residual from the following regression: [TAit /Ait−1 ] = β0 + β1 [1/Ait−1 ] + β2 [(REVit − DARit )/Ait−1 ] +β3 [PPEit /Ait−1 ] + β4 [NIit /Ait−1 ] + εit . 16 Our sample size, though small compared to many discretionary accruals studies, is similar to other survey-based work in the area (Prawitt, Smith, and Wood [2009], Prawitt, Sharp, and Wood [2012]). 17 In separate untabulated tests, we also limited data for the estimation model to only Fortune 1000 firms. However, there were a total of 47 firms (almost a quarter of our sample) for which we could not estimate a reliable accrual estimation model. Thus, our tabulated ABNACC variable based upon the Compustat population represents a tradeoff between capturing additional business norms that drive transactional accruals within a given industry (population based) versus transaction accruals that might be limited to only a large-firm environment (large-firm subsample).

16

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS TABLE 2 Variable Definitions

Variable

Description

ABNACC

The Kothari, Leone, and Wasley (2005) version of the modified Jones model measure of abnormal accruals. Abnormal accruals is the error term of the equation below: [T Ait /Ait−1 ] = β0 + β1 [1/Ait−1 ] + β2 [(DR E Vit − DARit )/Ait−1 ] + β3 [P P E it /Ait−1 ] + β4 [N Iit /Ait−1 ] + εit , where total accruals for estimation portfolio firm i for year t. TA, or total accruals, are defined as income before extraordinary items (Compustat Data Item #18) minus operating cash flows (Data Item #308). Ait −1 is total assets (Data Item #6) at t−1 for firm i. REVit is the change in net revenues (Data Item #12) for estimation portfolio firm i for year t. ARit is the change in accounts receivable (Data Item #2) for estimation portfolio firm i for year t. PPEit is gross property, plant, and equipment (Data Item #7) for estimation portfolio firm i for year t. NIit is net income (Data Item #172) for estimation portfolio firm i for year t. Dichotomous dependent variable coded “1” in instances where firm met the forecast or exceeded it by the consensus, annual EPS forecast scaled by price at the beginning of the year by more than 0.0005; “0” else. Dichotomous variable calculated using a two-step process. First, we calculate unmanaged earnings by backing out discretionary abnormal accruals from reported earnings. We then identify firms that would have missed analysts’ forecasts, without management’s exercise of the accounting discretion. For these firms, JM/BEATaccr equals “1” and for all other firms JM/BEATaccr equals “0.” Average IAF resource expenditure per hour. Calculated by dividing total in-house IAF budget per survey question #4 by total in-house IA hours per question #1. Relative audit committee IAF influence vis-`a-vis management (CEO and CFO). Measured as the ratio of the total agreement points for the audit committee in the numerator divided by the total agreement points for the CFO and CEO in the denominator. The numerator is the sum of agreement points (per the 1–5 Likert-scale responses) on survey questions #12a, 12d, and 12g. The denominator is the sum of agreement points (per the 1–5 Likert-scale responses) on survey questions # 12a–12i. The 1–5 Likert-scale response values are recalibrated to 0–4 for purposes of computing this variable. Indicator variable coded “1” when the IAF does not serve as a management training ground per survey question #13 and “0” otherwise. Coded “1” (“0”) when the following ratio constructed from survey question #4 is less than (greater than) 0.20: (Budgeted OSP $)/(Budgeted OSP $ + Budgeted in-house IAF $). Total assets in millions (Compustat Data Item #6). Number of years the firm was listed on Compustat, truncated at 25 years The sum of long-term debt (Compustat Data Item #9) and current liabilities (Data Item #5) of a company divided by total assets (Data Item #6). Number of disclosed segments in which the company operates. Cash flows from operations (Compustat Data Item #308).

JM/BEAT∗

∗ JM/BEATaccr

IACOMP

ACIAFINF

NONMTG

NONOSP20

ASSETS AGE LEVERAGE

SEGNUM CFO

(Continued)

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY

17

T A B L E 2—Continued Variable SALESGROW MTB CFOVOL ROA LOSS MATWEAK

INDUSTRY

∗

Description One year sales growth (Compustat Data Item #12FY2009 – Item #12 FY2008 / Item #12 FY2008 ). A company’s market-to-book ratio (Compustat Data Item #24 ∗ Item #25 / Item #216). Standard deviation of operating cash flows for 2005–2009. Return on assets (Compustat Data Item #172 / Data Item # 6). Coded “1” if the firm experienced a loss in fiscal year 2008, “0” else. Indicator variable coded “1” for client firm disclosing a material weakness in internal controls over financial reporting during the prior two years and “0” otherwise. Coded “1” if firm’s two-digit SIC code is included in specific focus industry per table 1, “0” otherwise. Focus industry groupings per Hogan and Jeter (1999).

Variable used only in additional analysis section.

We then segregate our abnormal accruals as positive (income-increasing) and negative (income-decreasing) abnormal accruals. Our predicted relations are symmetrical for negative and positive accruals. We argue that, if the information used for internal decision-making is incorrectly biased in either direction, there may be career penalties for the internal auditor.

4.5

TEST VARIABLES

Our test variables are IACOMP, ACIAFINF, NONMTG, and NONOSP20, and their respective interaction terms. Consistent with Abbott, Parker, and Peters [2012], IACOMP is a proxy for IAF’s competence. IACOMP is defined as the average in-house IAF resource expenditure per budget hour. We divided total in-house IAF budget per survey question #4 by total in-house IA hours per question #1. On average, the resource commitment per hour of budgeted audit staff should capture and summarize several disparate dimensions of IAF competence such as educational level, professional experience, and certification of the IAF staff. When we compare IACOMP to the Prawitt, Smith, and Wood [2009] IAF quality measure, we find that most of the elements of the Prawitt, Smith, and Wood [2009] measure are related to competence (experience, certifications, training, IAF size (a measure of unexpected spending on the IAF)). We suggest these individual IAF traits are captured by an hourly, in-house IAF compensation rate.18

18 We believe our composite IACOMP measure offers other certain advantages, as it incorporates many different IAF elements (i.e., experience, certification, training) into one summary measure, is continuous in nature (allowing for easier comparison between different-sized IAFs), and does not require an implicit equal weighting of inputs. Due in part to the underlying relationship between budgets and hours, we include additional subsequent tests to ascertain whether our proxy is correlated with various dimensions of competence. See additional analysis discussions. Utilizing per hour amounts strengthens our ability to capture individual auditor traits that might otherwise be lost when utilizing a measure of competence based upon

18

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS

Following Abbott, Parker, and Peters [2012], we incorporate a separate continuous measure of audit committee influence into our tests of IAF independence. Prior internal audit research suggests that oversight roles or influence, though potentially prescribed by a charter, do not necessarily manifest themselves in practice as dichotomous characteristics (Kaplan and Schultz [2006], Ernst and Young [2008], Rose and Norman [2008]).19 Therefore, ACIAFINF is defined as the relative level of influence exerted over the IAF by the audit committee vis-`a-vis management. To measure ACIAFINF, we asked (survey question #12) CIAs to state their level of agreement concerning the amount of influence exhibited by the audit committee versus management (CEO and CFO) on reporting lines, termination rights, and budget determination.20 The level of agreement can range from strongly disagree (“1”) to strongly agree (“5”). We then construct a ratio of the total agreement points for the audit committee in the numerator (i.e., the sum of agreement points for survey questions #12a, 12d, and 12g) divided by the total agreement points for the audit committee, CFO, and CEO in the denominator (i.e., the sum of agreement points for survey questions #12a–12i).21 To further illustrate ACIAFINF, if the CIA answers “5” (i.e., strongly agrees) to all survey questions #12a–12i, then the audit committee is seen as an equal IAF oversight partner with management. Our ACIAFINF variable would receive a value of 0.33, denoting 33% relative oversight of the IAF, vis-`a-vis management. If, however, the CIA strongly disagrees with any CEO/CFO relationship, while strongly agreeing with all audit committee relationships, our ACIAFINF variable equals 1, denoting full IAF control by the audit committee.22

IAF budgets deflated by firm size. Defining IACOMP as Budgets / Firm size yields inconclusive results, likely due in part to its lack of correlation with individual auditor competence traits that drive per hour costs. We thank the anonymous reviewer for bringing this to our attention. 19 Ernst and Young [2008] notes that audit committees take different approaches to the level of active oversight of the internal audit department, in which the internal auditor may perceive different levels of actual influence or authority exhibited by the corresponding overseer. 20 Similar to Abbott, Parker, and Peters [2010], we do not utilize answers to survey question #12j–12l (reviewing/approving the IAF’s annual risk assessment plan) when calculating the ACIAFINF variable. 21 Within upper management, the IAF can report to either or both the CEO and CFO. We combine the CEO and CFO responses since both share common risk preferences (compared to the Audit Committee) and are required by SOX Section 302 to certify the financial statements (SOX [2002]). 22 Survey responses to questions #12a–12i can range from 1 to 5, but are recalibrated to a scale of 0–4. By doing so, our ACIAFINF variable captures the intuition behind the relative IAF influence. For example, assume that the CIA respondent strongly agrees with the audit committee’s influence over IAF reporting, termination, and budgeting (i.e., responds with a “5” for survey questions 12a, 12d, and 12g). Also assume the CIA respondent also strongly disagrees with the CEO and CFO’s influence over IAF reporting, termination, and budgeting (i.e., responds with a “1” for survey questions 12b/12c, 12e/12f, and

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY

19

NONMTG represents our second facet of the IAF’s independence, namely, whether IAF is formally used as an MTG. NONMTG is coded “1” when the IAF is not part of a management training rotation, but instead is considered a separate career position within the company (survey question #13). The results of viewing the IAF as a distinct career position potentially induces greater attachment by the staff to the role of professional internal auditor, and thus a greater commitment to independence. In addition, those professionally attached to the IAF are less likely to feel pressure from audited managers in order to protect future placement opportunities. NONOSP20 represents the extent of the threat to the in-house IAF presented by the extent of outsourcing. Limited outsourcing may be related to the need for specialized expertise not possessed by the IAF, but larger levels of outsourcing may indicate a willingness on the part of a company to consider outsourcing core IAF functions. We identify outsourcing of more than 20% of the total budget as an appropriate level to indicate a possible threat to the IAF. The initial 20% threshold is based upon informal conversations with internal auditors concerning the level at which outsourcing would appear to threaten job security. We expect that, as the level of outsourcing increases, the in-house IAF may view itself as vulnerable to replacement, thus reducing the internal IAF’s independence. NONOSP20 is coded “1” if the level of outsourcing does not reach 20%, reflecting little or no threat to IAF independence. The extent of outsourcing is constructed from survey question #4:     Budgeted OSP $ / Budgeted OSP $ + Budgeted in-house IAF $ . Each of our stated hypotheses addresses the interactive nature of IAF competence and separate measures of independence. We test H1 by interacting IACOMP and ACIAFINF. We test H2 by the interacting IACOMP and NONMTG. We test H3 by interacting IACOMP and NONOSP20. In all three interactions, we expect a negative (positive) association with abnormal accruals (financial reporting quality).

4.6

CONTROL VARIABLES

We include control variables that may impact the level of abnormal accruals. We expect that ASSETS (log of company size) will magnify the size

12h/12i). Without recalibration to a 0–4 scale, the ACIAFINF value would equal 0.833 (e.g., (5+5+5)/(5+1+1+5+1+1+5+1+1)). With recalibration, the ACIAFINF value becomes 1 or ((4+4+4)/(4+0+0+4+0+0+4+0+0)). A second example further illustrates. In this case, assume that the CIA strongly disagrees with the audit committee’s IAF reporting, termination, and budgeting influence (i.e., responds with a “1” for survey questions 12a, 12d, and 12g) and also strongly agrees with the CFO’s reporting, termination, and budgeting influence (i.e., responds with a “5” for survey questions 12b, 12e, and 12h). Also assume that the CIA strongly disagreed with the CEO’s IAF reporting, termination, and budgeting influence (i.e., responds with a “1” for survey questions 12c, 12f, and 12i). In this case, ACIAFINF equals 0 (e.g., (0+0+0)/(0+4+0+0+4+0+0+4+0)).

20

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS

of accruals (Dechow and Dichev [2002]). For this reason, we predict a positive association between company size and positive accruals and a negative association between company size and negative accruals. AGE (number of years the company has been listed on Compustat, truncated at 25 years) is included because firms may experience different accrual patterns as they age (Prawitt, Smith, and Wood [2009]). We expect that LEVERAGE (total debt/total assets) will be associated with more income-increasing accruals to allow for nonviolation of debt covenants (Press and Weintrop [1990]) and income-decreasing accruals (DeAngelo, DeAngelo, and Skinner [1994]) to reduce earnings for contractual renegotiations. We include a variable that proxies for firm complexity using the number of operating segments a firm discloses in its 10K (SEGNUM). Firms with greater complexity may have greater financial reporting latitude due to the inherent complexity of their operations. We therefore expect that SEGNUM will be positively (negatively) associated with income-increasing (income-decreasing) abnormal accruals. CFO (operating cash flows), SALESGROW (sales growth from the prior year), and MTB (market to book) are included to control for growth, and CFOVOL (operating cash flow volatility) is included because it may impact the accrual calculation (Dechow, Sloan, and Sweeney [1996], Matsumoto [2002], Menon and Williams [2004]). Low performance provides an incentive for accruals management, so we include ROA (net income/assets) and LOSS (coded “1” if the firm experienced a loss in the preceding year, “0” else). In terms of income-increasing abnormal accruals, increases in ROA may impact the calculation of abnormal accruals and we expect a positive relation. In terms of incomedecreasing accruals, positive ROA provides incentives to “smooth” earnings and we expect a negative association between ROA and income-decreasing abnormal accruals. Firms with a net loss may have an incentive to magnify income-increasing abnormal accruals to avoid debt covenant violations and magnify income-decreasing abnormal accruals to “take a bath.” We therefore expect a positive (negative) association between LOSS and incomeincreasing (decreasing) abnormal accruals. We also include an indicator variable for the presence of material internal control weaknesses. Material weaknesses have been shown to be associated with an increase in abnormal accruals (Doyle, Ge, and McVay [2007]). Finally, consistent with prior accruals-based research, we include dummy variables for focus industry membership.

5. Results 5.1

DESCRIPTIVE STATISTICS

Table 3 presents descriptive statistics for the 189 respondents. Panel A provides information related to the IAF budget. The mean number of in-house (outsourced) IAF hours is 34,873 (4,195) and the mean inhouse (outsourced) budget is $2,056,236 ($484,312). The mean (median)

21

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY TABLE 3 Descriptive Statistics Panel A: Internal audit budget descriptive statistics Variable Name

Mean

Median

25th %

75th %

Total outsourced internal audit hours (n = 133)∗ Total outsourced internal audit budget Total in-house internal audit hours (n = 189) Total in-house internal audit budget ($) Hourly OSP IA resource expenditures∗∗ Hourly in-house IA (IACOMP) resource expenditures∗∗ Total combined IAF hours Total combined IAF budgets ($) % of OSP $ budget to total IAF $ budget IAFs with OSP budget > 20% of overall IAF budget (0,1) IAFs with OSP budget < 20% of overall budget (NONOSP20) IAF used as a management training ground (0,1) IAF not used as a management training ground (NONMTG)

4,195

1,800

0

5,000

$484,312

$450,000

$0

$900,000

34,873

20,000

8,000

37,750

$2,056,236

$1,500,000

$900,000

$2,500,000

$178.73/hr

$150/hr

$100/hr

$250/hr

$102.42/hr

$87.72/hr

$75/hr

$115/hr

40,983 $3,580,450

24,000 $2,225,000

8,460 $1,250,000

46,000 $5,250,000

0.177

0.072

0.00

0.332

0.338

0.00

0.00

1.00

0.662

1.00

0.00

1.00

0.8465

1.00

1.00

1.00

0.1535

0.00

0.00

0.00

∗ Only 133 of our 189 responding firms outsourced internal audit tasks and therefore our sample size for outsourced IAF data is 133. ∗∗ To obtain this figure, we calculated individual per hour IAF resource expenditures rates for each sample firm. We then averaged these 133 (189) rates to arrive at a mean (median) and percentiles figures for the OSP (in-house) IAF departments, respectively.

IACOMP is $102.42 (87.72). Approximately 85% of the respondents reported that IAF was an MTG. The mean ratio of OSP budget dollars to inhouse dollars is 17.7%, with a median of 7.2% and a range of 0 at the 25th percentile to 33.2% at the 75th percentile. Almost 34% of the 133 firms that outsourced had an OSP budget higher than 20%, indicating the pervasive nature of outsourcing portions of the IAF. Panel B of table 3 provides details concerning ACIAFINF. The mean (median) overall ACIAFINF measure is 0.41 (0.42), suggesting that on average the organizational oversight of the IAF is almost evenly split between management and the audit committee. As in Abbott, Parker, and Peters [2012], there is substantial, interfirm variation among the oversight roles. We find that, for the lower 25% of sample firms, the audit committee has relatively little influence over the IAF with a value of 0.26. The upper 25% of sample firms suggest that the IAF’s potential objectivity is relatively

22

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS T A B L E 3—Continued

Panel B: Audit committee and management oversight of internal audit function Degree of Agreement with survey statements #12a–12i IAF reports to the audit committee IAF reports to the CFO IAF reports to the CEO Audit committee authorized to terminate Chief Internal Auditor The CFO authorized to terminate Chief Internal Auditor The CEO authorized to terminate Chief Internal Auditor The audit committee determines Internal Audit’s annual budget The CFO determines Internal Audit’s annual budget The CEO determines Internal Audit’s annual budget ACIAFINF∗∗∗

Mean

Median

25th %

75th %

4.73

5.00

4.00

5.00

3.65 2.50 4.52

4.00 2.00 5.00

3.00 2.00 3.00

4.00 3.00 5.00

3.65

4.00

3.00

4.00

3.05

3.00

3.00

4.00

3.22

3.00

3.00

4.00

3.75

4.00

3.00

4.00

3.02

3.00

3.00

4.00

0.41

0.42

0.26

0.61

∗∗∗

ACIAFINF = The ratio of the total agreement points for the audit committee in the numerator divided by the total agreement points for the CFO and CEO in the denominator. The numerator is the sum of agreement points (per the 1–5 Likert-scale responses) on survey questions #12a, 12d, and 12g. The denominator is the sum of agreement points (per the 1–5 Likert-scale responses) on survey questions #12a– 12i. The 1–5 Likert-scale responses are recalibrated to 0–4 for purposes of computing this variable.

more protected from management influence, with the 75th percentile value of ACIAFINF equaling 0.61. The audit committee’s oversight is strongly manifested in reporting lines and termination authority. The respondents agree strongly with the statement that the IAF reports to the audit committee (mean of 4.73 and median of 5.00), and only slightly less strongly with the statement that the audit committee has termination authority over the CIA (mean of 4.52 and median of 5.00). However, in terms of reporting relationships, internal audit appears to be reporting to both the audit committee and the CFO or CEO. This could be indicative of a “solid line” or functional relationship with the audit committee and a “dotted line” or administrative relationship with the CFO or CEO. We observe lower agreement with the statement that the audit committee determines the IAF’s budget (mean of 3.22 and median of 3.00). Budgetary authority is greater for the CFO than the audit committee (mean of 3.75 and median of 4.00). The budgetary authority response represents the primary difference from results reported by Abbott, Parker, and Peters [2012] in a similar survey from an earlier time period (2005). We believe that the economic situation in our sample year, 2009, is a likely

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY

23

T A B L E 3—Continued Panel C: Control variables descriptive statistics Variable name

Mean

Median

25th %

75th %

ABNACC JM/BEAT∗∗∗∗ ASSETS (mil) AGE LEVERAGE SEGNUM CFO (mil) SALESGROW MTB CFOVOL ROA LOSS MATWEAK

0.0289 0.1957 $19,602 20.49 0.4729 4.3280 $1,780 −0.1177 2.5294 496.9 0.0321 0.2698 0.0423

−0.0139 0.0000 $5,318 25.00 0.4531 4.000 $566.9 −0.0947 1.6393 132.9 0.0365 0.0 0.0

−0.0788 0.0000 $2,613 15.00 0.3598 2.0000 $258.4 −0.2129 1.0786 66.85 0.0038 1.0 0.0

0.0269 1.000 $12,828 25.00 0.5644 6.0000 $1,126 −0.0105 2.7566 340.2 0.0698 0.0 0.0

All variable definitions can be found in table 2. ∗∗∗∗ Variable used in additional analysis section only. ABNACC is Kothari, Leone, and Wasley’s (2005) version of the modified Jones model measure of abnormal accruals. JM/BEAT is a dichotomous dependent variable coded “1” in instances where the firm met the consensus annual EPS forecast scaled by price at the beginning of the year or exceeded the forecast by more than 0.0005; “0” elses. ASSETS equals total assets in millions. AGE is the number of years the firm was listed on Compustat, truncated at 25 years. LEVERAGE equals the sum of long-term debt and current liabilities of a company divided by total assets. SEGNUM equals the number of disclosed segments in which the company operates. CFO equals cash flows from operations. SALESGROW equals one-year sales growth percentage. MTB equals market-to-book ratio. CFOVOL is the standard deviation of operating cash flows for 2005–2009. ROA equals return on assets. LOSS is coded “1” if the firm experienced a loss in fiscal year 2008, “0” else. MATWEAK is coded “1” for a client firm disclosing a material weakness in internal controls over financial reporting during the prior two years and “0” otherwise.

explanation for the difference. In tight economic conditions, it seems probable that all areas of the firm, including the IAF, would be scrutinized by the CFO for possible cost savings. Table 3, panel C, shows the control variable descriptive statistics. Our companies are quite large and “old,” as might be expected, with a mean (median) asset size of $19.6 billion ($5.3 billion) and a mean (median) age of 20.49 (25) years. Their leverage is fairly high (mean of 47%, median of 45%), and most have been public companies for at least 25 years, and many for much longer. Our sample firms have on average 4.32 operating segments. Operating cash flow is $1.78 billion (mean), $567 million (median). Notably, mean sales growth from 2008 to 2009 was −11.8%, and mean ROA was 3.2%, reflecting the 2009 economic environment. The mean market-to-book ratio is 2.6, while mean cash flow volatility is 496.9. Reflecting the economic downturn during our sample period, the incidence of a prior year loss (LOSS) is almost 27%. All these measures (except loss incidence) are lower than those reported in Prawitt, Smith, and Wood [2009], again likely reflective of the updated economic environment in our study. We also note a very low incidence of reported material weaknesses, with the mean of MATWEAK of 0.0405. This is consistent with a dramatic decrease in Section 404 weaknesses after the issuance of Audit Standard No. 5 (PCAOB [2007]). Table 3, panel C, also shows that the mean ABNACC is 0.0289, and the median is −0.0139.

24

5.2

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS

UNIVARIATE RESULTS

Table 4 provides the results of examining the sample, based upon whether the respondent firms do not use the IAF as an MTG (NONMTG) or do not have substantial OSP (NONOSP20). In panel A, firms not utilizing the IAF as an MTG (n = 28) are larger and have higher cash flows than firms that use the IAF as an MTG (n = 161). This is generally consistent with larger IAFs being able to provide an internal audit “career” (Messier et al. [2011]). Panel B of table 4 provides univariate differences based upon firms with less than 20% of their budgets attributed to outsourcing (n = 125) and those with more than 20% (n = 64). We note that firms without a substantial OSP presence have a greater audit committee IAF influence and a less severe deceleration in sales. The former result suggests that audit committees may exert a preference for in-house IAF. The latter result indicates that outsourcing may be an attractive means of cutting costs. We find no difference in the lack of an MTG between the two subsamples. Table 4, panel C, provides univariate comparisons based upon whether firms just meet or beat their analyst’s EPS forecast (n = 37) and those that do not (n = 152). We find a low level of just meeting or beating analyst forecasts (less than 20%), consistent with the Sarbanes-Oxley Act curtailing (but not eliminating) earnings management (Koh, Matsumoto, and Rajgopal [2008]). Consistent with prior literature, firms that just meet or beat analysts’ forecasts have more leverage, are more complex, have more operating cash flow, and have a greater market-to-book ratio. We find a marginally smaller IACOMP value for firms that just meet or beat analysts’ forecasts. Consistent with panels A and B, there are no univariate differences in IACOMP, NONMTG, or NONOSP20. Overall, the univariate results are consistent with (1) our three IAF independence measures capturing different independence mechanisms and (2) the inability of competence and independence to significantly influence financial reporting quality on a stand-alone basis.

5.3

MULTIVARIATE RESULTS

Tables 5 and 6 present our main results.23 Table 5 summarizes the results from regressing positive (income-increasing) accruals on our IAFrelated variables and controls variables. In table 5, our predicted signs for our interaction variables are all negative, as we predict lower incomeincreasing accruals when the IAF exhibits both greater competence and independence.24 With respect to the interactive term between IACOMP and ACIAFINF (coefficient = −0.00017, t-stat. = −2.302), our results suggest that the effect of competence depends on the amount of oversight provided by the audit committee and vice versa (in support of H1). In other 23 In the interest of brevity, we do not report the coefficient estimates on our INDUSTRY dummy variables. 24 We are careful to not overinterpret our stand-alone variables, as they are not necessarily reflective of main effects due to the continuous nature of some of our test variables (Jaccard and Turrisi [2003]).

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY

25

TABLE 4 Univariate Comparisons Panel A: Univariate comparison of IAF’s management training grounds Variable Name

Firms with NONMTG = 1

Firms with NONMTG = 0

IACOMP ACIAFINF NONOSP20 ASSETS (mil) AGE LEVERAGE SEGNUM CFO (mil) SALESGROW MTB CFOVOL ROA LOSS MATWEAK Observations

$107.59/hr 0.439 0.6786 $24,145 19.97 0.4907 4.4643 $1,929 −0.1332 2.41 448.2 0.0298 0.2500 0.0357 28

$101.52/hr 0.405 0.6584 $18,812 20.62 0.4698 4.2919 $1,754 −0.1150 2.54 505.4 0.0325 0.2733 0.0435 161

Difference

Mann-Whitney Statistic

$6.07/hr 0.034 0.0202 $5,333 −0.65 0.0209 0.1724 $175 0.0182 −0.1300 −57.2 −0.0027 −0.0233 −0.0078

1.0051 1.2323 0.9974 1.6736∗ 0.9929 0.6788 0.7505 1.5029∗ 0.7471 0.8515 0.4793 0.5555 0.6099 0.3311

Panel B: Univariate comparison of IAFs lacking/with a significant (>20%) OSP presence Variable Name

Firms with NONOSP20 = 1

Firms with NONOSP20 = 0

IACOMP ACIAFINF NONMTG ASSETS (mil) AGE LEVERAGE SEGNUM CFO (mil) SALESGROW MTB CFOVOL ROA LOSS MATWEAK Observations

$104.89/hr 0.441 0.168 $20,038 20.53 0.4780 4.2480 $1,775 −0.0988 2.47 515.1 0.0293 0.2720 0.0400 125

$97.59/hr 0.349 0.157 $18,750 20.41 0.4629 4.4844 $1,789 −0.1546 2.61 487.6 0.0335 0.2656 0.0468 64

Difference

Mann-Whitney Statistic

$7.30/hr 0.092 0.011 $1,288 0.12 0.0151 −0.2364 −$14 −0.0558 −0.14 27.5 −0.0042 0.0640 −0.0680

1.1011 1.6993∗ 0.0694 0.8828 0.3324 0.7247 0.8791 1.0801 3.4804∗∗ 0.3779 0.1155 0.9013 0.6528 0.4999

Panel C: Univariate comparison of firms that just meet/beat analyst forecast to other firms

Variable Name

Firms that Just Meet/Beat Analysts’ Forecasts

All Other Firms

Difference

Mann-Whitney Statistic

IACOMP ACIAFINF NONMTG NONOSP20 ASSETS (mil) AGE LEVERAGE

$99.87/hr 0.3606 0.1622 0.6486 $18,087 19.57 0.4994

$103.04/hr 0.4220 0.1513 0.6652 $19,971 20.65 0.4664

−$3.17/hr −0.0614 0.0109 −0.0166 −$1,904 −1.0800 0.0330

1.4011 1.7804∗ 1.0501 0.7822 1.8736∗ 0.4424 1.9942∗ (Continued)

26

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS T A B L E 4—Continued

Panel C: Univariate comparison of firms that just meet/beat analyst forecast to other firms

Variable Name SEGNUM CFO (mil) SALESGROW MTB CFOVOL ROA LOSS MATWEAK Observations

Firms that Just Meet/Beat Analysts’ Forecasts

All Other Firms

4.8214 $2,010 −0.1103 2.82 481.4 0.0339 0.2432 0.0541 37

4.2422 $1,724 −0.1195 2.46 500.3 0.0317 0.2763 0.0395 152

Difference

Mann-Whitney Statistic

0.5792 $286 −0.0092 0.36 −18.9 0.0022 −0.0233 0.0146

1.7305∗ 1.8544∗ 1.1008 2.0011∗∗ 0.3021 0.4933 0.6224 0.7754

Significance levels (one-tailed if in predicted direction): ∗ ,∗∗ , ∗∗∗ = p-value < 0.10, 0.05, 0.01, respectively. IACOMP equals average IAF resource expenditure per hour. ACIAFINF is the relative audit committee IAF influence vis-`a-vis management (CEO and CFO). NONMTG is an indicator variable coded “1” when the IAF does not serve as a management training ground per survey question #13 and 0 otherwise. NONOSP20 coded “1” (“0”) when the budgeted outsourced IAF services are less than 20% of IAF total budget. ASSETS equals total assets in millions. AGE is the number of years the firm was listed on Compustat, truncated at 25 years. LEVERAGE equals sum of long-term debt and current liabilities of a company divided by total assets. SEGNUM equals number of disclosed segments in which a company operates. CFO equals cash flows from operations. SALESGROW equals one-year sales growth percentage. MTB equals market-to-book ratio. CFOVOL is the standard deviation of operating cash flows for 2005–2009. ROA equals return on assets. LOSS is coded “1” if the firm experienced a loss in fiscal year 2008, “0” else. MATWEAK is coded “1” for client firm disclosing a material weakness in internal controls over financial reporting during the prior two years and “0” otherwise.

words, the impact of the IAF on financial reporting quality is jointly dependent on the level of competence and independence of the IAF. The reduction in the size of positive abnormal accruals for a certain level of IAF competence (audit committee oversight of the IAF) is conditional upon audit committee oversight of the IAF (IAF competence). Economically, this suggests that firms overinvesting in competence (independence), but not in independence (competence), are in danger of establishing an IAF that is not effective in strengthening the firms’ financial reporting process. In other words, the presence of both competence and independence is a necessary antecedent for effective financial reporting monitoring by the IAF. The significantly negative coefficient (coefficient −0.00013, t-stat −3.248) for the interaction term between IACOMP and NONOSP20 suggests that the abnormal accrual reduction effect of IACOMP is greater for firms that do not have a substantial OSP presence (in support of H3). In other words, greater amounts of investment in the in-house IAF activities (both in terms of budgets and hours provided by the in-house IAF) yield greater reductions in positive abnormal accruals (i.e., increase in financial reporting quality). We do not find a significant interaction effect between IACOMP and NONMTG. Several of the control variables are significant. ASSETS, LEVERAGE, CFO, MTB, CFOVOL, ROA, and LOSS are all significant and in the predicted direction. Table 6 summarizes the results from regressing negative (incomedecreasing) accruals on our IAF-related variables and controls variables.

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY

27

TABLE 5 OLS Regression Results: Positive Abnormal Accruals Positive ABNACC = b0 + b1 IACOMP + b2 ACIAFINF + b3 NONMTG + b4 NONOSP20 + b 5 IACOMP ∗ ACIAFINF + b6 IACOMP ∗ NONMTG + b 7 IACOMP ∗ NONOSP20 + b8 ASSETS + b9 AGE + b10 LEVERAGE + b 11 SEGNUM + b12 CFO + b13 SALESGROW + b14 MTB + b15 CFOVOL + b 16 ROA + b17 LOSS + b18 MATWEAK + b19−28 INDUSTRY + ε. Independent Variable Intercept IACOMP ACIAFINF NONMTG NONOSP20 IACOMP ∗ ACIAFINF IACOMP ∗ NONMTG IACOMP ∗ NONOSP20 ASSETS AGE LEVERAGE SEGNUM CFO SALESGROW MTB CFOVOL ROA LOSS MATWEAK INDUSTRY Observations Adjusted R2

Expected Sign ? ? ? ? − − − + − + + − + + + − + +

Coefficient Estimate

t-Statistic

0.0467 −0.00004 −0.0369 −0.0187 −0.0101 −0.00017 −0.00005 −0.00013 0.0003 −0.0004 0.0412 0.0052 −0.0001 0.0069 0.0048 0.0001 −0.0101 0.0108 0.0128 Included 81 0.2399

1.1099 −1.1921 −1.4201∗ −0.8645 −0.6779 −2.3022∗∗ −1.2199 −3.2481∗∗∗ 2.5778∗∗∗ −0.3101 1.7844∗ 2.9941∗∗∗ −1.4245∗ 0.7274 2.9998∗∗∗ 2.2217∗∗ 0.3741 0.9577 0.5299

Significance levels (one-tailed if in predicted direction): ∗ ,∗∗ , ∗∗∗ = p-value < 0.10, 0.05, 0.01, respectively. ABNACC is the Kothari, Leone, and Wasley (2005) version of the modified Jones model measure of abnormal accruals. IACOMP equals average IAF resource expenditure per hour. ACIAFINF is the relative audit committee IAF influence vis-`a-vis management (CEO and CFO). NONMTG is an indicator variable coded “1” when the IAF does not serve as a management training ground per survey question #13, and 0 otherwise. NONOSP20 is coded “1” (“0”) when the budgeted outsourced IAF services are less than 20% of the IAF total budget. ASSETS equals total assets in millions. AGE is the number of years the firm was listed on Compustat, truncated at 25 years. LEVERAGE equals the sum of long-term debt and current liabilities of a company divided by total assets. SEGNUM equals the number of disclosed segments a company operates in. CFO equals cash flows from operations. SALESGROW equals one-year sales growth percentage. MTB equals the market-to-book ratio. CFOVOL is the standard deviation of operating cash flows for 2005–2009. ROA equals return on assets. LOSS is coded “1” if the firm experienced a loss in fiscal year 2008, “0” otherwise. MATWEAK is coded “1” for a client firm disclosing a material weakness in internal controls over financial reporting during the prior two years and “0” otherwise.

Table 6 indicates that our results for income-decreasing accruals are also generally consistent with our hypotheses. In support of H1, H2, and H3, the coefficient estimates for all three interactive competence and independence terms (IACOMP∗ACIAFINF, IACOMP∗NONOSP20, IACOMP∗NONMGT) are significantly positive (indicating associations with smaller negative accruals, that is, negative accruals that are closer to zero). The same control variables are significant as in table 5, except that the cash flow variables lose significance and SALESGROW gains significance. We interpret this as suggesting that income-decreasing (negative) abnormal

28

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS TABLE 6 OLS Regression Results: Negative Abnormal Accruals

Negative ABNACC = b0 + b1 IACOMP + b2 ACIAFINF + b3 NONMTG + b4 NONOSP20 + b5 IACOMP ∗ ACIAFINF + b6 IACOMP ∗ NONMTG + b 7 IACOMP ∗ NONOSP20 + b8 ASSETS + b9 AGE + b10 LEVERAGE + b 11 SEGNUM + b12 CFO + b13 SALESGROW + b14 MTB + b15 CFOVOL + b 16 ROA + b17 LOSS + b18 MATWEAK + b19−28 INDUSTRY + ε. Independent Variable Intercept IACOMP ACIAFINF NONMTG NONOSP20 IACOMP ∗ ACIAFINF IACOMP ∗ NONMTG IACOMP ∗ NONOSP20 ASSETS AGE LEVERAGE SEGNUM CFO SALESGROW MTB CFOVOL ROA LOSS MATWEAK INDUSTRY Observations Adjusted R 2

Expected Sign ? ? ? ? + + + − + − − − + + − − − −

Coefficient Estimate

t-Statistic

−0.3110 0.00054 0.0104 0.0171 0.0027 0.00018 0.00013 0.00006 −0.0034 0.0012 −0.0877 0.0029 0.0001 0.0549 0.0008 −0.0002 −0.4176 −0.0728 −0.0504 Included 108 0.201

−1.0772 1.6005∗ 1.2002 0.8895 0.8874 2.9444∗∗∗ 2.6222∗∗ 2.9004∗∗∗ −0.4747 0.1111 −1.4123∗ 0.3133 0.5998 1.2987∗ 2.8956∗∗∗ −0.0765 −3.1767∗∗∗ −2.5649∗∗∗ −0.2772

Significance levels (one-tailed if in predicted direction): ∗ ,∗∗ , ∗∗∗ = p-value < 0.10, 0.05, 0.01, respectively. All variable definitions can be found in table 2. ABNACC is the Kothari, Leone, and Wasley (2005) version of the modified Jones model measure of abnormal accruals. IACOMP equals average IAF resource expenditure per hour. ACIAFINF is the relative audit committee IAF influence vis-`a-vis management (CEO and CFO). NONMTG is an indicator variable coded “1” when the IAF does not serve as a management training ground per survey question #13, and 0 otherwise. NONOSP20 is coded “1” (“0”) when the budgeted outsourced IAF services are less than 20% of the IAF total budget. ASSETS equals total assets in millions. AGE is the number of years the firm was listed on Compustat, truncated at 25 years. LEVERAGE equals the sum of long-term debt and current liabilities of a company divided by total assets. SEGNUM equals the number of disclosed segments in which the company operates. CFO equals cash flows from operations. SALESGROW equals one-year sales growth percentage. MTB equals market-to-book ratio. CFOVOL is the standard deviation of operating cash flows for 2005–2009. ROA equals return on assets. LOSS is coded “1” if the firm experienced a loss in fiscal year 2008, “0” otherwise. MATWEAK is coded “1” for a client firm disclosing a material weakness in internal controls over financial reporting during the prior two years and “0” otherwise.

accruals are less likely to be associated with cash flow considerations, and also that they are less (closer to zero) when sales are growing. This is consistent with fewer incentives for “big bath” behavior when growth is present. We note that, in order for IAF quality to impact financial reporting quality, it is necessary for the IAF to be involved in the monitoring of the financial reporting process. Per survey question #3 in the appendix, such opportunities include financial statement audits of subsidiaries and/or assisting the external auditor with the financial statement audit. We observed

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY

29

that all respondents had nonzero budget allocations to these activities, indicating that all sample IAF departments had the opportunity to monitor the financial reporting process.25 In summary, the evidence of tables 5 and 6 is consistent with the characterization of a two-factor IAF quality function, whereby competence and independence must combine with each other to promote the IAF as an effective, internally based financial reporting monitor.26

5.4

ADDITIONAL ANALYSIS: OTHER FINANCIAL REPORTING QUALITY PROXIES

While accruals-based measures of earnings management are the most pervasive proxy for financial reporting quality, extant research also uses the ability to just meet/beat analysts’ forecasts and the use of positive accruals to just meet/beat analysts’ forecasts as proxies for financial reporting quality (Koh, Matsumoto, and Rajgopal [2008], Prawitt, Smith, and Wood [2009], Mande and Son [2012]). We also incorporate these additional financial reporting quality proxies into our analysis by utilizing logistic regression where our dependent variable is a dichotomous variable as to whether the firm just meets or beats analysts’ expectations (JM/BEAT) or uses positive abnormal accruals to just meet or beat analysts’ expectations (JM/BEATaccr ). To compute which firms just meet/beat analysts’ expectations, we divide how close the firm was to consensus forecast by price at the beginning of the year and then examine a bin that is within 0.0005 (Prawitt, Smith, and Wood [2009]). This results in 37 firms that meet or just beat analysts’ forecasts (JM/BEAT = 1) and 152 other firms (JM/BEAT = 0). Our set of independent variables is consistent with those found in the prior section. Table 7, column A, provides the results for JM/BEAT as the dependent variable. The logistic regressions were conducted in a manner consistent with Norton, Wang, and Ai [2004] and Ai and Norton [2003]. Consistent with the main results, coefficient estimates for our interactive terms of IACOMP∗ACIAFINF and IACOMP∗NONOSP20 are statistically significant. Our coefficient estimate for IACOMP∗NONMTG is in the predicted direction, but is not statistically significant. This provides additional evidence in support of H1 and H3. In general, the joint presence of audit committee IAF oversight and IAF competence appears to reduce the risk of firms artificially just meeting or beating the consensus EPS forecast. Similarly, when competence increases when there is not a substantial

25 Our survey respondents reported spending 15.4% of in-house audit hours on assisting with the financial audit, 10.3% on audits of subsidiaries, 6.8% on compliance auditing, and 9.8% on special projects. For comparison, the highest single percentage of time was devoted to SOX-related controls work (27.15%). 26 Breusch-Pagan tests failed to detect the presence of heteroskedasticity.

30

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS TABLE 7 Logistic Regression Results: Just Meet/Beat Analysts’ Forecasts

JM /BEAT = b0 + b1 IACOMP + b2 ACIAFINF + b3 NONMTG + b4 NONOSP20 + b 5 IACOMP ∗ ACIAFINF + b6 IACOMP ∗ NONMTG + b 7 IACOMP ∗ NONOSP20 + b8 ASSETS + b9 AGE + b10 LEVERAGE + b 11 SEGNUM + b12 CFO + b13 SALESGROW + b14 MTB + b15 CFOVOL + b 16 ROA + b17 LOSS + b18 MATWEAK + b19−28 INDUSTRY + ε. Column A Independent Variable Intercept IACOMP ACIAFINF NONMTG NONOSP20 IACOMP ∗ ACIAFINF IACOMP ∗ NONMTG IACOMP ∗ NONOSP20 ASSETS AGE LEVERAGE SEGNUM CFO SALESGROW MTB CFOVOL ROA LOSS MATWEAK INDUSTRY Observations Pseudo-R 2

Expected Sign ? ? ? ? − − − + − + + − + + − + − −

Column B

Coefficient Estimate

Chi-Square Statistic

Coefficient Estimate

Chi-Square Statistic

−0.3110 0.0054 0.0104 0.0171 −0.0027 −0.0098 −0.0013 −0.0048 −0.0034 0.0049 1.3944 0.3249 0.0001 0.3551 0.0498 −0.0002 2.8783 −0.4445 −0.0513 Included 189 0.201

1.0772 1.0504 2.4691 0.6738 0.8874 4.4823∗∗ 1.6788 5.9004∗∗ 1.2626 0.9203 3.9487∗∗ 2.8777∗ 0.0075 1.0013 4.7145∗∗ 0.0765 5.1055∗∗ 1.3011 0.7381

−0.4971 0.0054 0.0089 0.0124 0.0022 −0.0117 −0.0005 −0.0054 −0.0301 0.0012 1.6553 0.0829 −0.0002 0.2817 0.0502 −0.0002 2.6533 −0.6235 −0.0719 Included 189 0.137

1.0772 1.6603 1.8550 1.7776 0.6911 4.4187∗∗ 2.5558 5.1234∗∗ 1.7231 0.1111 3.7349∗ 2.7792∗ 0.0978 1.1113 2.8277∗ 0.0765 3.6783∗ 0.9645 0.5334

Significance levels (one-tailed if in predicted direction): ∗ ,∗∗ , ∗∗∗ = p-value < .10, .05, .01, respectively. Column A present results when the dependent variable is defined as JM/BEAT is a dichotomous dependent variable coded “1” in instances where the firm met the consensus annual EPS forecast scaled by price at the beginning of the year or exceeded the forecast by more than 0.0005; “0” else. Column B presents results when the dependent variable is defined as JM/BEATaccr , coded “1” if the firm would have missed analysts’ forecasts, without abnormal positive accruals, “0” else. All other variables are defined as previously described.

outsourcing presence, there is a reduction in the likelihood of a firm just meeting or beating the consensus EPS forecast.27 Table 7, column B, provides the results for JM/BEATaccr as the dependent variable. Following Koh, Matsumoto, and Rajgopal [2008], we calculate unmanaged earnings by backing out discretionary abnormal accruals from reported earnings. We then identify firms that would have missed analysts’ forecasts, without management’s exercise of the accounting discretion and code these firms as JM/BEATaccr = 1. We find 25 firms that belong to 27 In untabulated results, we isolate those firms that appear to have utilized abnormal accruals to just meet/beat forecasts, and find similar results, though at lower levels of significance.

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY

31

this category (JM/BEATaccr = 1) and the remaining 164 (JM/BEATaccr = 0) firms do not. We obtain evidence similar to that provided in column A. The coefficient estimates for the interactive terms of IACOMP∗ACIAFINF and IACOMP∗NONOSP20 suggests that the disciplining interactive impact of competence and independence are acute for this form of financial reporting quality.

5.5

ADDITIONAL ANALYSIS: DETERMINANTS OF IAF COMPETENCE

We assert that IACOMP is a parsimonious measure that aggregates many individual, IAF competence-related characteristics. To test this assertion, we regress IACOMP against several IAF characteristics that may be associated with competence. Individual IAF characteristics include measures of IAF certification, CIA certification, and tenure, which relate to survey questions #6–8. We further augment our model to include other noncompetence variables that have been shown to impact overall IAF budgets (Carcello, Hermanson, and Raghunandan [2005] and Anderson et al. [2012]). These additional variables include firm size, leverage, percentage of inventory to total assets, and operating cash flows. Finally, as a means of discerning if any of our independence-related measures are potentially determinants of IACOMP, we include ACIAFINF, NONOSP20, and NONMTG as regressors.28 The untabulated results show that professional certification, experience, percentage of assets in inventory, and firm size are positively related to IACOMP. The significant, positive association between IACOMP and the percentage of total assets comprised of inventory is likely attributable to the more complex accounting systems needed to accumulate and analyze product costs. Similarly, firm size may proxy for both the relative amount of firm resources available to devote to the IAF, as well as firm complexity. Only one of the independence measures, ACIAFINF, is marginally associated with IACOMP. As the relative degree of audit committee-IAF oversight increases, per-hour IAF compensation increases as well. This result is consistent, although at a much lower economic and statistical significance, with Carcello, Hermanson, and Raghunandan [2005] and Anderson et al. [2012]). These results provide evidence that our measures of competence and independence capture separate and distinct constructs.

5.6

SENSITIVITY ANALYSIS

To address potential endogeneity, we execute a two-stage least squares analysis using an instrumental variable approach (Prawitt, Smith, and Wood [2009]). Given that our IAF independence variables are not highly

28 Carcello, Hermanson, and Raghunandan [2005] also find that the presence of an OSP negatively impacts overall the IAF budget. Instead of using a dichotomous OSP “presence” variable per Carcello, Hermanson, and Raghunandan [2005], we use NONOSP20, which is substantively similar to the variable used in Carcello, Hermanson, and Raghunandan [2005]. We note that neither NONOSP20 nor a dichotomous OSP presence variable is significantly associated with IACOMP.

32

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS

correlated with each other and do not appear to significantly influence IAF competence, our primary focus is predicting IACOMP in a first-stage regression. Similar to Prawitt, Smith, and Wood [2009], our objective is to utilize instrumental variables that are correlated with IACOMP, but not correlated with our measures of financial reporting quality. Our two instrumental variables include an average IACOMP score by industry and the amount of inventory relative to assets.29 We then use the predicted IACOMP value in a second-stage regression using the same set of dependent variables found in tables 5–7. We obtain virtually identical results to those reported in tables 5–7. We also considered numerous categories of potentially omitted correlated variables. With respect to auditor-related variables, prior research has indicated that auditor tenure (Davis, Soo, and Trompeter [2009]), auditor industry specialization (Reichelt and Wang [2010]) and the magnitude of non-audit fees (Frankel, Johnson, and Nelson [2002]) may impact financial reporting quality. Our results remained unchanged when we included: (1) a continuous variable defined as total auditor tenure (capped at 25); (2) a dichotomous variable coded “1” when the external auditor audited the largest percentage of client assets per the two-digit focus industry per table 1, and “0” otherwise and (3) a continuous variable defined as the ratio of non-audit service fees to audit fees. We utilized various alternative control variables for corporate governance and/or audit committee effectiveness, including the Gompers, Ishii, and Metrick [2003] corporate governance index, and the type and percentage of financial experts present on the audit committee (Dhaliwal, Naiker, and Navissi [2010]).30 We also examined certain CFO characteristics (age, tenure, board membership, stockholdings, and gender) given their potential impact on financial reporting decisions (Ge, Matsumoto, and Zhang [2011]). Inclusion of the above variables did not substantively alter the results of tables 5–7. We examined whether IACOMP might be capturing the effect of high cost of living areas by including “high-cost” dummy variables.31 We included a variable for the percentage of IAF budget allocated to financial statement auditing per Lin et al. [2011]. We examined the potential effects of litigious industries and labor intensiveness (Frankel, Johnson, and Nelson [2002], Matsumoto [2002]). We included a dichotomous variable coded “1” in instances where the primary SIC code indicated a litigious industry 29 The industry-wide average of IACOMP is likely to be associated with an individual firm’s IACOMP because internal auditors have the opportunity to transfer their skills from one company to the other and would do so as a means of increasing or maintaining their current salary. Similar to Prawitt, Smith, and Wood [2009], the industry-wide IACOMP value is unlikely to be correlated with our dependent variables since we do not find that financial reporting quality is highly concentrated in a few industries. 30 Experts are current or former certified public accountants, CFOs, vice presidents of finance, controllers, and those with other major accounting positions. 31 The high-cost cities include New York, Boston, Los Angeles, Chicago, and San Francisco.

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY

33

per Frankel, Johnson, and Nelson [2002].32 We also included a continuous variable calculated as 1 minus the ratio of gross property, plant, and equipment (PPE) to total assets. The results of tables 5–7 remained qualitatively similar. The time period of our study was a recession year, and, thus, our results may not be generalizable to other years. We assumed that our test variables would remain relatively constant over a short horizon, and tested our model for 2010 and 2011. We also included abnormal returns and a dummy variable for negative abnormal return to our regression as a control for accelerated recognition of losses in a recession year (Ball and Shivakumar [2006]). Our results remained qualitatively similar in both these tests. Last, we attempted to separate the relative influence of the CEO and CFO over the IAF. In our first test, we created a dichotomous variable (CFODOMAIN) coded “1” in instances where the survey respondent indicated a strong level of agreement with survey questions 12b/12e/12h, while simultaneously indicating strong disagreement with survey questions 12c/12f/12i. We then interacted IACOMP∗ACIAINF∗CFODOMAIN. Inclusion of these variables in our regression did not substantively alter the results of tables 5–7. In a second set of analyses, we redefined CFODOMAIN as “1” in instances where the sum of survey responses to 12b/12e/12h was greater than survey responses to 12c/12f/12i (that is, in observations where the CIA noted that the CFO had relatively greater IAF influence). CFODOMAIN occurred in slightly more than 80% of our observations— indicating that the CFO has relatively greater influence over the IAF vis-`a-vis the CEO in the overwhelming majority of instances. Inclusion of this redefined CFODOMAIN variable in our regressions did not impact our overall results and the coefficient estimates on CFODOMAIN on a stand-alone or interactive basis were statistically insignificant.33

6. Conclusion We examine whether IAF quality—and its ability to foster higher quality financial reporting—can be reasonably characterized as a joint function of both competence and independence. Our tests rely on the argument that the effectiveness of the IAF rests upon the complementary roles of competence and independence. For example, even though a competent IAF may 32 High-litigation industries are industries with SICs of 2833–2836, 3570–3577, 3600–3674, 5200–5961, and 7370–7374 (Frankel, Johnson, and Nelson [2002]). 33 There were zero observations where the CIA indicated a strong level of disagreement with the following survey questions: 12a/12d/12g. That is, survey respondents all stated that the audit committee had some sort of influence over the IAF. In slightly less than 3% of our observations, the CIA indicated that the audit committee had sole province over the IAF (that is, strongly agreed with survey questions 12a/12d/12g, while simultaneously strongly disagreeing with survey questions 12b/12c/12e/12f/12h/12i). These results indicate that it is extremely rare for the audit committee to have sole control over the IAF and that there are virtually no instances where the audit committee has no influence over the IAF.

34

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS

be more likely to identify a misstatement, an IAF that lacks independence may be less likely to report it to the appropriate channel or demand that the party originally responsible for the misstatement to correct it. Despite a large amount of practitioner guidance on the importance of independence, little is known about how independence interacts with competence in fostering the IAF’s role as an effective financial reporting monitor. In sum, our results show that, if one wishes to ask “what is the effect of IAF competence (independence) on financial reporting quality?,” the answer is “it depends on the independence (competence) of the internal auditor.” Our study is also motivated by the relative scarcity of archival evidence on factors that potentially impact IAF competence, IAF independence, and their joint influence on the IAF quality. Our proxy for competence is captured by the organization’s investment in average hourly rates of budgeted IAF resources. We consider three specific artifacts of independence: the relative degree of audit committee oversight vis-`a-vis management oversight, the preponderance of internal audit outsourcing, and the use of internal audit as an MTG. We regress several financial reporting quality proxies on our set of competence and independence variables, as well as the interaction of the competence and independence variables. We find that, consistent with a two-factor IAF quality function, we document statistically significant associations between our interaction terms and financial reporting quality. The overall results provide evidence consistent with the hypothesis that the combined presence of both competence and independence is a necessary antecedent to effective IAF financial reporting monitoring. With respect to the IAF characteristics, we find results consistent with independence being enhanced by relatively greater degrees of audit committee oversight of the IAF, versus management oversight. We utilize a measure of oversight influence that extends beyond formal reporting lines. This enhanced independence interacts with IAF competence as a means of curtailing financial reporting discretions in both income-increasing and incomedecreasing environments. We also document a similar set of relationships when we interact IAF competence and the relative lack of IAF outsourcing. Finally, we find lower occurrences of income-decreasing abnormal accruals as IAFs jointly reflect both greater competence and are not used as an MTG. Our findings extend prior research and deepen our understanding of the determinants of IAF quality. We believe that more research is needed in this area to enrich our understanding of how and why IAFs impact financial reporting quality. Our study points to the need to better understand how an IAF can benefit a firm, but also how researchers can potentially demonstrate and quantify the value of a competent and independent IAF. APPENDIX

Dear Sir or Madam: We are conducting a brief survey to obtain data about internal audit functions, audit risk factors, and the relationship between internal audit and the audit committee. Your insight is vital to this project, so please take a few

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY

35

minutes to complete the survey and return it in the self-addressed envelope enclosed. Please be assured that all responses will be strictly confidential, and no company or individual will be specifically identified. If you would like a copy of our findings, please include an e-mail and/or mailing address. Thank you for your time and cooperation on this survey. 1. During fiscal 2009, approximately how many hours were devoted to internal audit services by: OUTSIDE SERVICE PROVIDERS (OSPs) hours INTERNAL PROVIDERS

hours

2. Of the total internal audit hours provided by outside service provider(s) per question 1, please give the approximate percentage distribution of activities (note the percentages should add to 100%): Activity Description Financial statement audits of subsidiaries Special consulting projects Compliance auditing Risk management activities Assisting external auditor in financial statement audit

% of hours

Activity Description EDP auditing

% of hours

Internal control evaluation Section 404-related work Fraud audits Other (explain)

3. Of the total internal audit hours provided by the in-house internal audit department per question 1, give the approximate percentage distribution of activities (note the percentages should add to 100%): Activity Description Financial statement audits of subsidiaries Special consulting projects Compliance auditing Risk Management Activities Assisting external auditor in financial statement audit

% of hours

Activity Description EDP Auditing

% of hours

Internal control evaluation Section 404-related work Fraud audits Self-assessment programs/ Other (explain)

4. Please indicate the total fiscal year 2009 internal audit-related expenditures for: OUTSIDE SERVICE PROVIDERS $ IN-HOUSE INTERNAL AUDIT $ 5. Please indicate the total Section 404-related fees paid to your external auditor in 2009. $

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS

36

6. Please indicate the percentage of your staff who have the following certifications: CPA

CIA

CISA

% of staff with at least 1 certification

7. With respect to you personally, please indicate how many years of work experience you have as: an internal auditor a financial statement auditor 8. With respect to you personally, which certification(s) do you possess? (check all that apply): CPA

CIA

CISA

Other

9. Which of the following best describes how controls over financial reporting are monitored at your entity? a. Separate Evaluations b. Controls self-assessments c. Other (explain) 10. Which of the following best describes how your entity conducts its assessments of the effectiveness of controls over financial reporting? a. Ongoing testing throughout the year b. Point-in-time near year-end c. Other (explain). 11. What are the key components of your organization’s internal audit risk assessment and weighting of each (not considered, low, medium, high).

Factor Extent of routine versus nonroutine transactions Dollar materiality Ethical tone set by management Changes in upper management Complexity of the operation Quality of application-level controls Physical controls over assets Quality of manual controls Complexity of accounting guidance Financial condition of the entity Program level changes since last audit Asset liquidity Controls over period-end reporting Complexity of controls Budget variances of auditable entities Changes in nonmanagement personnel Downsizing Foreign Corrupt Practices Act compliance Currency fluctuation exposure Political risk Environmental regulation Exposure to financially distressed third parties (i.e., customers, vendors)

Not applicable

Low

Medium

High

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY

Factor

Not applicable

Low

Medium

37 High

Quality of management forecasts Revenue fluctuation Other (please explain) Other (please explain) Other (please explain)

12. Please indicate your level of agreement with the following statements: (1= strongly disagree; 2 = disagree; 3 = neutral; 4 = agree; 5 = strongly agree)

Level of agreement (circle one number)

Statement a. Internal audit reports to the Audit Committee b. Internal audit reports to the Chief Financial Officer (CFO) c. Internal audit reports to the Chief Executive Officer (CEO) d. The Audit Committee has authorization to terminate the Chief Internal Auditor e. The CFO has authorization to terminate the Chief Internal Auditor f. The CEO has authorization to terminate the Chief Internal Auditor g. The Audit Committee determines Internal Audit’s annual budget h. The CFO determines Internal Audit’s annual budget i. The CEO determines Internal Audit’s annual budget j. The Audit Committee reviews and approves Internal Audit’s annual risk assessment plan k. The CFO reviews and approves Internal Audit’s annual risk assessment plan l. The CEO reviews and approves Internal Audit’s annual risk assessment plan

12345 12345 12345 12345 12345 12345 12345 12345 12345 12345 12345 12345

13. Does internal audit serve as a training ground for future management No positions? Yes REFERENCES ABBOTT, L. J.; S. PARKER; AND G. F. PETERS. “Earnings Management, Litigation Risk, and Asymmetric Audit Fee Responses.” Auditing: A Journal of Practice and Theory 25 (2006): 85–98. ABBOTT, L. J.; S. PARKER; AND G. F. PETERS. “Serving Two Masters: The Association Between Audit Committee Internal Audit Oversight and Internal Audit Activities.” Accounting Horizons 24 (2010): 1–24. ABBOTT, L. J.; S. PARKER; AND G. F. PETERS. “Audit Fee Reductions from Internal AuditProvided Assistance: The Incremental Impact of Internal Audit Characteristics.” Contemporary Accounting Research 29 (2012): 94–118.

38

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS

ABBOTT, L. J.; S. PARKER; G. F. PETERS; AND D. V. RAMA. “Corporate Governance, Audit Quality, and the Sarbanes-Oxley Act: Evidence from Internal Audit Outsourcing.” The Accounting Review 82 (2007): 803–35. AHLAWAT, S. S., AND D. J. LOWE. “An Examination of Internal Auditor Independence: InHouse Versus Outsourcing.” Auditing: A Journal of Practice and Theory 23 (2004): 147–58. AI, C., AND E. NORTON. “Interaction Terms in Logit and Probit Models.” Economic Letters 80 (2003): 123–29. AMERICAN INSTITUTE OF CERTIFIED PUBLIC ACCOUNTANTS (AICPA). AU Section 322: SAS No. 65 The Auditor’s Consideration of the Internal Audit Function in an Audit of Financial Statements. New York, NY: AICPA, 1991. AMERICAN INSTITUTE OF CERTIFIED PUBLIC ACCOUNTANTS (AICPA). Proposed Statement on Auditing Standards: Using the Work of Internal Auditors. New York, NY: AICPA, 2013. ANDERSON, U. “Assurance and Consulting Services” in Research Opportunities in Internal Auditing, edited by A.D. Bailey, A. A. Gramling, and S. Ramamoorti. Altamonte Springs, FL: Institute of Internal Auditors Research Foundation, 2003. ANDERSON, U.; M. CHRIST; K. JOHNSTONE; AND L. RITTENBERG. “A Post-SOX Examination of Factors Associated with the Size of the Internal Audit Functions.” Accounting Horizons 26 (2012): 167–91. BALL, R., AND L. SHIVAKUMAR. “The Role of Accruals in Asymmetric Timely Gain and Loss Recognition.” Journal of Accounting Research 44 (2006): 207–42. BAME-ALDRED, C.; D. BRANDON; B. MESSIER; L. RITTENBERG; AND C. M. STEFANIAK. “A Summary of Research on External Auditor Reliance on the Internal Audit Function.” Auditing: A Journal of Practice and Theory 32 (2012): 251–86. CARCELLO, J.; D. HERMANSON; AND K. RAGHUNANDAN. “Factors Associated with U.S. Public Companies’ Investment in Internal Auditing.” Accounting Horizons 19 (2005): 69–84. CHRIST, M. H.; A. MASLI; N. Y. SHARP; AND D. A. WOOD. “Rotational Internal Audit Programs and Financial Reporting Quality: Do Compensating Controls Help?” Accounting, Organizations and Society 44 (2015): 37–59. CHRISTOPHER, J.; G. SARENS; AND P. LEUNG. “A Critical Analysis of the Independence of the Internal Audit Function: Evidence from Australia.” Accounting, Auditing, & Accountability Journal 22 (2009): 200–20. COHEN, J.; G. KRISHNAMOORTHY; AND A. WRIGHT. “Corporate Governance in the Post-Sarbanes Oxley Era: Auditors’ Experiences.” Contemporary Accounting Research 27 (2010): 751–86. CORAM, P.; C. FERGUSON; AND R. MORONEY. “The Internal Audit, Alternative Internal Audit Structures, and the Level of Misappropriation of Assets Fraud.” Accounting and Finance 48 (2008): 543–59. DAVIS, L.; B. SOO; AND G. TROMPETER. “Auditor Tenure and the Ability to Meet or Beat Earnings Forecasts.” Contemporary Accounting Research 26 (2009): 517–48. DEANGELO, H.; L. DEANGELO; AND D. J. SKINNER. “Accounting Choice in Troubled Companies.” Journal of Accounting and Economics 17 (1994): 113–43. DEANGELO, L. “Auditor Size and Audit Quality.” Journal of Accounting and Economics 3 (1981): 189–99. DECHOW, P. M., AND I. D. DICHEV. “The Quality of Accruals and Earnings: The Role of Accrual Estimation Errors.” The Accounting Review 77 (2002): 35–59. DECHOW, P. M.; R. SLOAN; AND A. SWEENEY. “Causes and Consequences of Earnings Manipulation: An Analysis of Firms Subject to Enforcement Actions by the SEC.” Contemporary Accounting Research 13 (1996): 1–36. DHALIWAL, D. S.; V. NAIKER; AND F. NAVISSI. “The Association Between Accruals Quality and the Characteristics of Accounting Experts and Mix of Expertise on Audit Committees.” Contemporary Accounting Research 27 (2010): 787–827. DOYLE, J. T.; W. GE; AND S. MCVAY. “Accruals Quality and Internal Control over Financial Reporting.” The Accounting Review 82 (2007): 1141–70. ERNST & YOUNG (EY). Global Internal Audit Survey, 2008. Available at www.ey.com.

INTERNAL AUDIT QUALITY AND FINANCIAL REPORTING QUALITY

39

FELIX, W.; A. GRAMLING; AND M. MALETTA. “The Contribution of Internal Audit as a Determinant of External Audit Fees and Factors Influencing this Contribution.” Journal of Accounting Research 39 (2001): 513–34. FRANCIS, J. “A Framework for Understanding and Researching Audit Quality.” Auditing: A Journal of Practice and Theory 30 (2011): 125–52. FRANKEL, R.; M. JOHNSON; AND K. NELSON. “The Relation Between Auditors’ Fees for Nonaudit Services and Earnings Management.” The Accounting Review 77 (2002): 71–105. GE, W.; D. MATSUMOTO; AND J. ZHANG. “Do CFOs Have Style? An Empirical Investigation of the Effect of Individual CFOs on Accounting Practices.” Contemporary Accounting Research 28 (2011): 1141–79. GOMPERS, P.; J. ISHII; AND A. METRICK. “Corporate Governance and Equity Prices.” The Quarterly Journal of Economics 118 (2003): 107–156. GOODWIN, J., AND T. Y. YEO. “Two Factors Affecting Internal Audit Independence and Independence: Evidence from Singapore.” International Journal of Auditing 5 (2001): 107–25. GRAHAM, J. R., AND C. R. HARVEY. “The Theory and Practice of Corporate Finance: Evidence from the Field.” Journal of Financial Economics 60 (2001): 187–243. HARRINGTON, C. “Internal Audit’s New Role.” Journal of Accountancy (September) (2004): 25–40. HERMANSON, D. “The Growing Stature of Internal Auditing.” Internal Auditing 17 (2002): 43– 44. HERMANSON, D., AND L. RITTENBERG. “Internal Audit and Organizational Governance.” In Research Opportunities in Internal Auditing. Altamonte Springs, FL: Institute of Internal Auditors Research Foundation, 2003. HOFFELDER, K. “New Guidance Could Add Teeth to Internal Audit.” CFO.com, 2012. Available at http://www3.cfo.com/article/2012/10/auditing˙eisneramper-institute-of-internalauditors-chief-audit-executive-standards/. HOGAN, C., AND D. JETER. “Industry Specialization by Auditors.” Auditing: A Journal of Practice and Theory 18 (1999): 1–17. INSTITUTE OF INTERNAL AUDITORS (IIA). The IIA’s Recommendations to the Conference Committee on H.R. 3703. Altamonte Springs, FL: IIA, 2002. INSTITUTE OF INTERNAL AUDITORS (IIA). Professional Practices Framework.” Altamonte Springs, FL: IIA, 2005. INSTITUTE OF INTERNAL AUDITORS (IIA). Global Internal Audit Competency Framework. Altamonte Springs, FL: IIA, 2013. INTERNATIONAL AUDITING AND ASSURANCE STANDARDS BOARD (IAASB). ISA 610 (Revised March 2013), Using the Work of Internal Auditors. International Federation of Accountants (IFAC), New York. Available at http://www.ifac.org/sites/default/files/publications/files/ ISA-610-(Revised-2013).pdf. JACCARD, J. J., AND R. TURRISI. Interaction Effects in Multiple Regression, Second edition, Sage Publications, Newbury Park, CA: 2003. KAPLAN, S. E., AND J. J. SCHULTZ JR. The Role of Internal Audit in Sensitive Communications. The Institute of Internal Audit Research Foundation. Altamonte Springs, FL: IIA, 2006. KOH, K.; D. A. MATSUMOTO; AND S. RAJGOPAL. “Meeting or Beating Analyst Expectations in the Post-Scandals World: Changes in Stock Market Rewards and Managerial Actions.” Contemporary Accounting Research 25 (2008): 1067–98. KOTHARI, S. P.; A. J. LEONE; AND C. E. WASLEY. “Performance Matched Discretionary Accrual Measures.” Journal of Accounting and Economics 39 (2005): 163–97. LIN, S.; M. PIZZINI; M. VARGUS; AND I. R. BARDHEN. “The Role of the Internal Audit Function in the Disclosure of Material Weakness.” The Accounting Review 86 (2011): 287–323. MABRY, B., AND L. SCHWARTZ. Working with Auditors and Management. Ernst & Young LLP, 2008. MANDE, V., AND M. SON. “CEO Centrality and Meeting or Beating Analysts’ Earnings Forecasts.” Journal of Business Finance and Accounting 39 (2012): 82–112. MATSUMOTO, D. “Management’s Incentives to Avoid Negative Earnings Surprises.” The Accounting Review 77 (2002): 483–514. MENON, K., AND D. WILLIAMS. “Former Audit Partners and Abnormal Accruals.” The Accounting Review 79 (2004): 1095–118.

40

L. J. ABBOTT, B. DAUGHERTY, S. PARKER, AND G. F. PETERS

MESSIER, W. F.; J. K. REYNOLDS; C. A. SIMON; AND D. A. WOOD. “The Effect of Using the Internal Audit Function as a Management Training Ground on the External Auditor’s Reliance Decision.” The Accounting Review 86 (2011): 2131–154. NASDAQ STOCK MARKET LLC (NASDAQ). “Notice of Filing of Proposed Rule Change to Require that Listed Companies Have an Internal Audit Function.” Release No. 34-69030; File No. SR-NASDAQ-2013-032, 2013. Available at http://www.sec.gov/rules/sro/nasdaq/ 2013/34-69030.pdf. NORMAN, C. S.; A. M. ROSE; AND J. M. ROSE. “Internal Audit Reporting Lines, Fraud Risk Decomposition, and Assessments of Fraud Risk.” Accounting, Organizations, and Society 35 (2010): 546–57. NORTON, E.; H. WANG; AND C. AI. “Computing Interaction Effects and Standard Errors in Logit and Probit Models.” The Stata Journal 4 (2004): 154–67. PELFREY, S., AND E. PEACOCK. “A Current Status Report on Outsourcing.” Internal Auditing Fall (1995): 26–32. PRAWITT, D. F.; N. Y. SHARP; AND D. A. WOOD. “Reconciling Archival and Experimental Research: Does Internal Audit Contribution Affect the External Audit Fee?” Behavioral Research in Accounting 23 (2011): 187–206. PRAWITT, D. F.; N. Y. SHARP; AND D. A. WOOD. “Internal Audit Outsourcing and the risk of Misleading or Fraudulent Financial Reporting: Did Sarbanes-Oxley Get It Wrong?” Contemporary Accounting Research 29 (2012): 1109–36. PRAWITT, D. F.; J. L. SMITH; AND D. A. WOOD. “Internal Audit Characteristics and Earnings Management.” The Accounting Review 84 (2009): 1255–80. PRESS, E. G., AND J. B. WEINTROP. “Accounting-Based Constraints in Public and Private Debt Agreements.” Journal of Accounting and Economics 12 (1990): 65–95. PRICEWATERHOUSECOOPERS LLP (PWC). Business Upheaval: Internal Audit Weighs Its Role Amid the Recession and Evolving Enterprise Risks. New York, NY: PWC, 2009. PUBLIC COMPANY ACCOUNTING OVERSIGHT BOARD (PCAOB). Auditing Standard 5: An Audit of Internal Control over Financial Reporting that Is Integrated with an Audit of Financial Statements. Washington, D.C.: PCAOB, 2007. QUARLES, N. R. “An Examination of Promotion Opportunities and Evaluation Criteria as Mechanisms for Affecting Internal Auditor Commitment, Job Satisfaction and Turnover Intentions.” Journal of Managerial Issues 6 (1994): 176–94. RAGHUNANDAN, K.; W. READ; AND D. V. RAMA. “Audit Committee Composition, ‘Grey Directors,’ and Interaction with Internal Auditing.” Accounting Horizons 15 (2001): 105–18. REICHELT, K., AND D. WANG. “National and Office-Specific Measures of Auditor Industry Expertise and Effects on Audit Quality.” Journal of Accounting Research 48 (2010): 647–86. ROSE, J., AND C. S. NORMAN. Internal Audit Reporting Lines: Fraud Risk Decomposition, and Assessments of Fraud Risk. Altamonte Springs, FL: The Institute of Internal Auditors Research Foundation, 2008. SARBANES-OXLEY ACT OF 2002 (SOX). “Corporate and Auditing and Accountability, Responsibility, and Transparency Act of 2002.” U.S. Public Law 107-204. 107th Cong., 2d sess., 30 July 2002. SCARBROUGH, D.; D. V. RAMA; AND K. RAGHUNANDAN. “Audit Committee Composition and Interaction with Internal Auditing: Canadian Evidence.” Accounting Horizons 12 (1998): 51– 62. STEWART, J., AND N. SUBRAMANIAM. “Internal Audit Independence and Independence: Emerging Research Opportunities.” Managerial Auditing Journal 25 (2010): 328–60.