Governance Structures and Abnormal Accruals Characteristics. Stacie O. Kelley University of Washington. Ping-Sheng Koh* UQ Business School

Transcription

1 Governance Structures and Abnormal Accruals Characteristics Stacie O. Kelley University of Washington Ping-Sheng Koh* UQ Business School Yen H. Tong University of Washington First Draft: 15 September 2004 Current Draft: 15 September 2004 Preliminary draft. Please do not quote without permission. Comments welcomed. *Corresponding author: UQ Business School The University of Queensland Brisbane QLD 4072 Australia Telephone: + (61 7) p.koh@business.uq.edu.au JEL Classification: M41, M43 and G34 Acknowledgements: This paper is based on work done when Ping-Sheng Koh was visiting the University of Washington, Seattle. We thank Allen Craswell for providing access to the Who Audits Australia? database. Able research assistance of Courtney Chugg is acknowledged. Kelley acknowledges financial support from the Deloitte Foundation, and Tong acknowledges financial support from Nanyang Technological University, Singapore.

2 Governance Structures and Abnormal Accruals Characteristics Abstract We examine the twin roles of accountability and value creation of corporate governance in the context of financial reporting. We investigate the accountability role by examining the association between governance structures and abnormal accruals. We differentiate governance mechanisms that have direct roles in the financial reporting process (audit related governance structures) from mechanisms that have indirect roles (non-audit related governance structures). We find that independence of non-audit related governance structures is negatively associated with absolute abnormal accruals whilst internal audit related governance structures have an incremental and negative association with income increasing abnormal accruals. Our evidence indicates that the governanceabnormal accruals association is not symmetrical between income increasing and income decreasing abnormal accruals. We examine the value creation role of governance mechanisms by investigating whether abnormal accruals predicted by governance structures are associated with future performance (future cash flows). We find that abnormal accruals predicted by audit related (non-audit related) governance structures are positively associated with one-year (two-year ahead) future cash flows. Our evidence is consistent with the independence of non-audit governance mechanisms and the internal audit function of audit committees being important attributes of governance in financial reporting. Also, the evidence is consistent with both audit and non-audit related governance mechanisms being value enhancing to firms.

3 1. Introduction The recent spate of corporate collapses (such as Enron and WorldCom in the U.S., HIH and Harris Scarfe in Australia) has drawn attention to corporate governance failures in preventing or forewarning these corporate events. In August 2002, the Australian Stock Exchange (ASX) formed the ASX Corporate Governance Council (the Council hereafter), consisting of 21 different constituencies, to develop recommendations which reflect international best practices in corporate governance for Australian firms. The Council issued the Principles of Good Corporate Governance and Best Practice Recommendations in March The Council views corporate governance as the system by which companies are directed and managed and good corporate governance structures are aimed at encouraging companies to create value and provide accountability and control systems commensurate with the risks involved (ASX, 2003, p.3). This view suggests corporate governance mechanisms play both accountability and value creation roles. In this study we use the context of financial reporting to examine the accountability and value creation roles of corporate governance structures. We examine the accountability role by testing whether corporate governance structures are associated with abnormal accruals. We also investigate the value creation role of corporate governance by testing the association of abnormal accruals as predicted by the corporate governance structures in place with future firm performance. We examine the accountability and value creation roles of governance structures in the financial reporting context because ensuring integrity in financial reporting is an integral part of international best practices (e.g., BRC in the U.S., Cadbury Report in the 2

4 UK, Toronto Guidelines in Canada, OECD principles for OECD countries and ASX principles in Australia). The Australian context provides a unique setting to examine these issues because Australia is considered to have a shareholder model of corporate governance and strong legal protection of minority shareholders, similar to the U.S. and the U.K. (La Porta et al., 1999; Kiel and Nicholson, 2003). However, the ownership structure in Australia is more concentrated than in these countries, which deviates from the Berle and Means depiction of widely held firms to which the shareholder model is founded (e.g., Cheffins, 2001; La Porta et al., 1999). 1 With high ownership concentration being the more common ownership structure around the world than the Berle and Mean corporations (La Porta et al., 1999), Australia shares a common ownership structure with many of these countries while having a well developed capital market and strong minority shareholder protection and other characteristics that typify the U.K./U.S. shareholder model more so than these other economies. With the convergence of corporate governance practices internationally, the Australian experience can provide valuable lessons to policy-makers of countries with concentrated ownership that are moving towards stronger protection of minority shareholders and strengthening their capital markets. 2 We differentiate between governance structures that have a direct role in ensuring financial reporting integrity (audit related governance structures) from those that have an indirect role (non-audit related governance structures). By direct role we mean that the 1 La Porta et al. (1999) report that, among large publicly traded firms, 100% and 80% of them are considered as widely held in the U.K. and U.S. respectively. In contrast, only 65% of large publicly traded firms in Australia are widely held. 2 While the governance practices in the U.S. and U.K. typify the shareholder model, closer examinations reveal the local practices in these countries are quite different on some fundamental aspects (Monks and Minow, 2001). Such differences in countries that have been classified as having an almost identical model suggests that country specific studies outside the U.S. and U.K. such as this paper can contribute to our understanding of the effects of governance practices as these practices continue to evolve. 3

5 intended purpose of these structures is financial reporting, while the governance structures that play an indirect role have other purposes such as general oversight, strategic management and remuneration. We find that both audit and non-audit related governance structures are effective in mitigating discretionary financial reporting behavior. Specifically, we find that governance structures associated with board independence and internal audit are negatively associated with absolute and positive abnormal current accruals respectively. Our evidence also suggests that the association between governance structures and abnormal accruals is not symmetrical between income increasing and income decreasing abnormal accruals, as the association appears to be stronger among firms with income increasing abnormal accruals. In addition, we find that the predicted components of abnormal current accruals related to audit (non-audit) related governance structures are positively associated with one-year (two-year) ahead future cash flows. This suggests audit related governance structures are value enhancing over the shorter term whilst nonaudit related board structures are value enhancing over a longer term. Our paper makes several contributions to the literature. This is the first known study to distinguish between governance structures that have a direct role in the financial reporting process from those that have an indirect role, and find evidence suggesting differential associations with abnormal accruals and future cash flows exist. This suggests future studies should differentiate the roles of various governance mechanisms to suit the context of their studies to enhance our understanding of the effects of the relevant governance structures. We are also among the first to examine both the accountability and value creation roles of corporate governance mechanisms to provide 4

6 more coherent evidence on the benefits of these mechanisms. 3 Prior studies have implicitly assumed individual governance mechanisms exist and operate in isolation (e.g., Klein, 2002 and Xie et al., 2003). We relax this assumption and assume related governance mechanisms operate in tandem to achieve their collective objectives. This assumption better reflects how governance mechanisms work in practice. This study represents a significant extension to the only known Australian study on the association between governance structures and accruals management (Davidson et al., 2004). Specifically, we extend Davidson et al. (2004) to a multiple-year study that differentiates governance structures that have a direct and an indirect role in the financial reporting process, and examine the value creation role of these structures. We also extend Park and Shin (2004), which uses a similar institutional setting (Canada) as Australia, by examining a more comprehensive set of governance structures during a more recent time period and specifically examining the value creation role of governance structures. 4 The rest of the paper is organized as follows. Section 2 presents our hypothesis development. In Section 3, we detail the research design while in Section 4 we discuss our results. Finally, Section 5 summarizes the paper. 3 Bowen et al. (2003) is the only known study that is closest, in spirit, to our study in this respect but conducted in the U.S. 4 The Canadian and Australian institutional environments are similar in the sense that both countries follow a shareholder model of corporate governance, have well developed capital markets and strong legal protections for minority shareholders and have similar level of ownership concentration (e.g., La Porta et al., 1999). However, there remain some differences. For example, among large publicly traded firms, out of 35-40% of firms that are not widely held, Canadian firms are more likely to be family owned whilst Australian firms are more likely to be controlled by widely held corporations (La Porta et al., 1999). Among medium-sized publicly traded firms, much lower percentage of Australian firms is widely held than Canadian firms (30% vs. 60%). For the non-widely held medium-sized firms, both countries have similar percentage that are family owned (70-75%) while the remaining Australian firms are corporation owned, the Canadian firms are state owned (La Porta et al., 1999). Hence, while Canadian and Australian institutional environment are comparable they are not identical. We do not exploit these differences as they are beyond the scope of this study. 5

7 2. Hypothesis Development In this study, we divide governance structures into governance mechanisms that have a direct role in the financial reporting process (audit related governance structures) and mechanisms that pertain to the general oversight and running of firms (non-audit related board structures). This distinction enables us to more clearly examine whether mechanisms set up to ensure the integrity of the financial reporting process achieve their intended objectives. While governance mechanisms related to general oversight and operation of the business are likely to have both value adding and accountability roles, these mechanisms are more likely to have a second order effect on ensuring financial statement integrity. 2.1 Audit Related Governance Structures When laying out the general principle on safeguarding financial reporting integrity, the Council recommends that a company should have in place a structure that can independently verify and safeguard the integrity of the company s financial reporting (ASX, 2003). The Council further suggests that such a structure would involve both internal mechanisms (audit committee) as well as external mechanisms (external auditors). The primary role of the audit committee is to oversee and review the financial reporting process of a company and to act as an intermediary between external auditors, managers and the board of directors to ensure open information flow and discussions (Klein, 2002; BRC, 1999). The effectiveness of the audit committee in discharging its responsibilities is dependent upon on its composition and frequency of 6

8 meeting, among others (e.g., ASX, 2003; BRC, 1999). Empirical evidence is generally consistent with an independent audit committee and active audit committee being associated with lower levels of abnormal accruals in the U.S. (Klein, 2002; Xie et al., 2003) but no such evidence is found in Australia (Davidson et al. 2004). As for the external mechanisms, external auditors, their effectiveness in ensuring the integrity of financial reporting is likely to depend on the auditors quality and auditors independence. Existing studies have argued and found evidence consistent with higher quality auditors (proxied by Big 5) being associated with lower levels of abnormal accruals (Becker et al., 1998; Francis et al., 1999), and lower incidence of accounting errors (DeFond and Jiambalvo, 1991) relative to lower quality auditors. Similarly, firms that have higher levels of abnormal accruals are found to be more likely to receive a qualified audit opinion (Francis and Krishnan, 1999). The independence of external auditors is affected by the economic bond between the external auditors and their audit client (DeAngelo, 1981). Increasing auditors reliance on non-audit services fees is argued to compromise auditors independence (e.g., Frankel et al., 2002; Levitt, 2000). The maintained hypothesis of this paper is that effective audit related governance mechanisms (both internal and external) are important to ensure financial reporting integrity. Hence, our first hypothesis is: H1: Effective audit related governance structures are associated with lower levels of absolute abnormal accruals. If the impact of effective audit related governance structures is symmetrical between income increasing and income decreasing abnormal accruals then it follows that: 7

9 H1a: Effective audit related governance structures are associated with lower levels of income increasing abnormal accruals. H1b: Effective audit related governance structures are associated with lower levels of income decreasing abnormal accruals. We argue that audit related governance mechanisms work together to ensure financial reporting integrity. We identify five characteristics to capture the essence of audit related governance mechanisms, including: 1) audit committee composition; 2) number of audit committee meetings during the financial year; 3) external auditors identity; 4) external auditors audit opinion; and 5) external auditors reliance on nonaudit fees. We discuss these characteristics further in the research design section. 2.2 Non-Audit Related Board Governance Structures While companies are encouraged to establish audit related governance structures to ensure the integrity of the companies financial reporting, the ultimate fiduciary duties remain with the board of directors. That is, the board of directors has the duty to monitor management to protect shareholders interest and ensure financial reporting integrity. Given the board can delegate the monitoring of financial reporting processes to the audit committee (Kiel and Nicholson, 2003), the non-audit related board structures are likely to play a secondary role in monitoring financial reporting processes. Prior research has argued that the effectiveness of boards in monitoring management and financial reporting is largely dependent on their independence and activity (e.g., Fama, 1980; Fama and Jensen, 1983; BRC, 1999). Empirical evidence is generally consistent with this argument as independent boards are associated with a lower 8

10 likelihood of financial statement fraud (Beasley, 1996) and SEC enforcement for GAAP violations (Dechow et al., 1996), and lower levels of abnormal accruals (Peasnell et al., 2000; Klein, 2002; Xie et al., 2003; Davidson et al., 2004). Consistent with prior literature, effective non-audit related board governance structures are expected to contribute to financial reporting integrity. As such, our second hypothesis is: H2: Effective non-audit related board governance structures are associated with lower levels of absolute abnormal accruals. Similar to our first hypothesis, if the impact of effective non-audit related governance structures is symmetrical between income increasing and income decreasing abnormal accruals then it follows that: H2a: Effective non-audit related board governance structures are associated with lower levels of income increasing abnormal accruals. H2b: Effective non-audit related board governance structures are associated with lower levels of income decreasing abnormal accruals. Drawing from best practice recommendations and prior studies, we capture the effectiveness of non-audit related board structures via several aspects, including 1) the separation of CEO-Chair roles; 2) majority non-executive directors on the board; 3) majority independent directors on the board; 5 4) number of board meetings during the financial year; 5) majority non-executive directors sitting on the remuneration committee; 5 We define an independent director as a non-executive director without any related party transactions with the firm. While it is impossible to establish a perfect measure of independence, our operationalization is consistent with two of the three aspects the Council s definition, namely independent of management (i.e., non-executive director) and free of business relationships (i.e., no related party transactions). 9

11 and 6) majority of board members sitting on two or more other boards. We discuss these characteristics further in the research design section. 2.3 Corporate Governance and Value Creation Good corporate governance is intended to add value to firms in addition to its monitoring role (ASX, 2003; Hilmer, 1998). None of the existing studies on the association between abnormal accruals and corporate governance examine the value creation role of corporate governance. 6 We examine the value creation role of corporate governance by investigating whether abnormal accruals predicted by corporate governance structures are associated with future performance. Consistent with prior literature, the preceding section predicts effective governance structures to be associated with lower level of abnormal accruals. However, the directional association between abnormal accruals and effective corporate governance mechanisms by itself does not inform readers whether these governance mechanisms are adding value to firms. That is, the negative association between effective governance structures and abnormal accruals can be value enhancing or reducing. Therefore, whether corporate governance s role in financial reporting is value adding or not remains an empirical question. If the association between abnormal accruals and corporate governance enhances firm value, then following Core et al. (1999), Hanlon et al. (2003) and Bowen et al. (2003), we expect abnormal accruals predicted by corporate governance mechanisms to be positively associated with future performance. Conversely, if the predicted abnormal 6 One exception is Bowen et al. (2003) that study the association between accounting discretion and corporate governance, and how the accounting discretion predicted by corporate governance is associated with future performance. 10

12 accruals are not, or are negatively, associated with future performance, then the effect of governance structures on abnormal accruals is not value enhancing. Consistent with the value creation role of corporate governance, our third hypothesis is: H3: Abnormal accruals predicted by governance structures are positively associated with future performance. 3. Research Design 3.1 Sample selection Our initial sample is drawn from the intersection between the Connect4 and Aspect Fin Analysis database from This leads to an initial sample of 2,259 firm-year observations. We delete observations (a) without GICS industry codes (298), (b) in the finance-related industries 8 (228), (c) with change in financial year end (58), (d) insufficient financial data (155), (e) that are not incorporated in Australia (41), and (f) within industries that have less than eight observations per year (546). Our final sample comprises 933 firm-year observations from , consisting of 337 distinct firms. Financial information is drawn from the Aspect Fin Analysis database whereas governance structure information is hand collected from firms annual reports. Auditor, audit opinion, audit and non-audit fee information is obtained from the Who Audits Australia? database. 9 7 The Connect4 database contains the annual reports of the Top 500 listed Australian companies from The Aspect Fin Analysis database contains machine-readable financial information for about 1200 listed Australian companies over the last 12 years. We use the intersection between these two databases, as the governance variables are hand collected from the annual reports. 8 Firms in the GICS codes (banks), (diversified financials) and (insurance) are considered finance-related. 9 We thank Allen Craswell for providing these data. 11

13 3.2 Governance Structure Variables To extract the underlying common constructs, we perform factor analysis separately on the audit related and non-audit related governance structures. This approach allows the interaction between various governance structures to be captured without introducing a multicollinearity problem into subsequent regression analysis (Dechow et al., 1996). This represents an improvement to Klein (2002) and Xie et al. (2003) where they investigate various governance variables one at a time, which implicitly assumes the effects of various governance mechanisms are independent of each other. Audit related governance structure We identify five characteristics that are consistent with prior literature and international best practices to capture the effectiveness of audit related governance mechanisms, including the number of audit committee meetings during the financial year (ACMEET), the majority of non-executive directors sitting on the audit committee (MAC), being audited by a big 5 auditor (KPMG, Arthur Anderson, Deloitte and Touche, Ernst and Young and PricewaterhouseCoopers) (BIG5), the audit opinion, AUDOP, where zero if a clean opinion and one otherwise; and independence as measured by an indicator variable equal to one if the ratio of non-audit fees to total fees is equal to or greater than 50%, zero otherwise (FEE). Independence has two components independence in fact and independence in appearance. The ratio of non-audit to total fees has been used to proxy for both aspects of independence, even though it is arguably a better proxy for independence in appearance, and has been shown to be associated with 12

14 financial reporting characteristics such as abnormal accruals and conservatism (e.g. Ashbaugh et al., 2003; Chung and Kallapur, 2003; Kelley, Shores and Tong, 2004; and Ruddock, Taylor and Taylor, 2004). The first two variables, ACMEET and MAC, are internal audit mechanisms and are positively correlated (see Table 1), while the other three are external audit mechanisms and their relation is mixed. BIG5 is positively correlated with FEE and negatively correlated with AUDOP, as are AUDOP and the FEE variable. Given the audit committee is expected to have close interactions with external auditors in discharging their duties, internal and external audit mechanisms are likely to act in tandem. To capture this interaction characteristic, we perform a factor analysis on these five characteristics to obtain summary measures of the effectiveness of audit related governance structures. 10 The factor analysis process identifies two factors that capture the five characteristics. We call the first factor internal audit factor (IAFAC) and it is positively associated with the number of audit committee meetings (ACMEET) and majority of NED sitting on the audit committee (MAC). This is consistent with the effectiveness of the audit committee being dependent on its independence and activity (e.g., ASX, 2003; BRC, 1999). We call the second factor external audit factor (EAFAC). It is positively associated with Big 5 auditors (BIG5), the independence variable (FEE), and negatively associated with audit opinion (AUDOP). This factor is consistent with effective external audit mechanisms where Big 5 auditors provide higher 10 We do not separately perform factor analysis on internal and external audit related governance structures, as they do not function in isolation of each other in practice. Performing a single factor analysis on all audit related variables enables such interactions to be captured in the factor scores. 13

15 quality audits, have a greater capacity to deliver non-audit services, and due to more selective client choice, are more likely to have clients with clean audit opinions. 11 Non-audit related board governance structures We use six variables to capture the effectiveness of non-audit related board governance based on international best practices and prior literature, including CEO- Chair duality (DUAL), majority non-executive directors on the board (MNED), majority independent directors on the board (MIND), 12 number of board meetings during the financial year (BMEET), majority non-executive directors sitting on the remuneration committee (MRC), majority of board members sitting on two or more other boards (MMULT). Descriptive statistics in Table 1 show that boards meet an average of 11 times per year, only 10% of the observations include a dual CEO and Chairman, and nearly 89% have boards with a majority of non-executive directors. The non-audit related board governance variables are included to capture similar constructs, which is reinforced by the correlations in Table 1. Not surprisingly, having a CEO as the chair of the board (DUAL) is negatively and significantly correlated with MNED, MINDN and MRC, but only marginally negatively associated with MMULT (p-value < 10%). The relation between the number of board meetings (BMEET) and the other variables is somewhat mixed. 11 An alternative view would be that this factor (EAFAC) is associated with ineffective external audit mechanisms where the greater reliance on non-audit fees by Big 5 jeopardize their objectivity which lead to less qualified audit opinion. 12 We define an independent director as a non-executive director without any related party transactions with the firm. While it is impossible to establish a perfect measure for directors independence, our operationalization is consistent with two of the three aspects of the Council s definition on independence, namely independent of management (i.e., non-executive director) and free of business relationships (i.e., no related party transactions). 14

16 The factor analysis process identifies two separate factors. We refer the first as board independence (BIFAC). It is positively associated with majority non-executive directors on the board (MNED), majority independent directors on the board (MIND), majority non-executive directors sitting on the remuneration committee (MRC), and majority of board members sitting on two or more other boards (MMULT). We call the second factor board activity (BAFAC) and it is positively associated with number of board meetings during the financial year (BMEET), and negatively associated with CEO- Chair duality (DUAL). Both factors are consistent with the effectiveness of the board being dependent on its independence and activity. 3.3 Characteristics of Accountability Abnormal Accruals Measure We examine abnormal current accruals because previous research suggests that managerial discretion over accounting choices and estimates manifests itself more in current accruals than in long-term accruals such as depreciation (Teoh, Welch and Wong 1998a, 1998b; Beneish 1998). 13 Consistent with Kothari, Leone, and Wasley (2005), we use a cross-sectional modified-jones model with lagged return-on-assets (ROA) to estimate abnormal current accruals: 1 CACC = α + α ( REV REC ) + α ( ROA ) + ε. (1) it, 1 2 it, it, 3 it, 1 it, TAit, 1 13 As sensitivity tests, we also examine abnormal total accruals. Untabled results are qualitatively similar to results using abnormal current accruals discussed in Section 4. 15

17 CACC i,t is current accruals measured as net profit before depreciation and amortization minus operating cash flows, REC it, is change in accounts receivables measured as accounts receivables in year t minus accounts receivables in year t-1. REV it, is change in revenue, calculated by taking revenue in year t minus revenue in year t-1. CACC i,t, REC it, and REV it, are deflated by beginning of the year total assets ( TAit, 1). ROAit, 1 is lagged return on assets, measured as net profit before extraordinary items divided by beginning of the year total assets. Equation (1) is estimated cross-sectionally by year using six-digit GICS codes with at least eight observations. Firm-year specific abnormal current accruals (ACA i,t ) are the residuals (ε i,t ) from the estimation of the equation. 14 Abnormal Accruals and Governance Structures We examine the association between abnormal accruals and the governance factors by estimating the following regression: ACA = b 0 + j = 4 j= 1 b YEAR + b 5 BAFAC + b 6 BIFAC + b 7 EAFAC + b 8 IAFAC j + b 9 DIRSH + b 10 LEV + b 11 SIZE + b 12 CFO + b 13 EXPER + b 14 BTM + b 15 XLIST + b 16 CEOCH + b 17 SUB + ε (2) 14 As in Kothari et al. (2005), the estimation of the modified-jones model includes changes in account receivables in the estimation stage. This is different from the original modified-jones model which only includes changes in receivables during the test stage (Dechow, Sloan, and Sweeney, 1995). Kothari et al. (2005) argue that the original modified-jones approach is likely to generate large estimated discretionary accruals whenever a firm experiences extreme growth during the test stage compared to the estimation stage. Including the change in receivables during the estimation stage (which assumes that all changes in credit sales arise from earnings management) mitigates this problem. 16

18 ACA is either absolute or signed abnormal current accruals. When ACA is absolute abnormal current accruals, we expect the coefficients on BAFAC, BIFAC, EAFAC and IAFAC to be significantly negative, consistent with H1. For signed ACA, we partition the sample into income increasing (positive) and income decreasing (negative) abnormal current accruals. Consistent with H1a and H2a (H1b and H2b), we expect the coefficients on BAFAC, BIFAC, EAFAC and IAFAC to be negative (positive) when ACA is income increasing (decreasing). In equation (2), we control for year effects by including year dummies. We also control for other contracting and economic factors that have been shown by prior studies to be associated with abnormal accruals. Prior literature has shown that accruals management is associated with insider ownership (Warfield et al., 1995), concentrated ownership (Peasnell et al., 1999), leverage (Press and Weintrop, 1990; DeFond and Jiambalvo, 1994), firm size (Han and Wang, 1998), operating cash flows (Dechow, 1994), extreme earnings performance (Dechow et al., 1995), growth opportunities (Klein, 2002) and changes in CEOs (Wells, 2002). Further, cross-listing firms are bonding themselves to more stringent scrutiny, enforcement and/or disclosure requirements (e.g., Coffee, 2002; Lang, Lines and Miller, 2003). This suggests cross-listing is likely to be associated with accruals management and other firm characteristics including governance structures. We measure insider ownership (DIRSH) as directors shareholdings to total shares outstanding; concentrated ownership (SUB) as total substantial shareholders ownership to total shares outstanding; 15 total leverage (LEV) as the ratio of gross borrowing to total assets; firm size (SIZE) as the log of market value of equity; cash flow 15 Substantial shareholders are shareholders with 5% or more of a firm s outstanding shares. 17

19 from operations (CFO) as cash flow from operations divided by total assets. Extreme performance (EXPER) is measured as one if return on assets is in the top or bottom 10% of the distribution and zero otherwise; growth opportunities as book-to-market value of equity (BTM); cross-listing (XLIST) takes the value of one if a firm is cross-listed in another exchange and zero otherwise; and CEO change (CEOCH) equals one if there is a CEO change during the financial year and zero otherwise. 3.4 Value Creation Our future performance metric to capture value creation is future cash flows from operations. We focus on future cash flows as it is less susceptible to accruals management than earnings-based performance measures such as return on assets. Unlike market-based performance metrics such as stock returns, cash flow-based performance metrics do not incorporate market expectations on accruals management. As such, cash flow performance is a more appropriate performance measure in examining our research question. We include one-year, two-year and cumulative two-year ahead cash flow from operations (deflated by total assets) as our future performance measures. 16 Cumulative two-year ahead cash flows are the sum of the one-year and two-year ahead cash flows from operations. We estimate the following regression to test H3: j = 4 j = 11 F UTCFO = b + b YEAR + b IND + b P_ BODACA + b P _ AUDACA 0 j j j = 1 j = 5 + b CFO + b SIZE + b EV + ε (3) R 16 We stop at two-year ahead as we lose one full year of observations for each year beyond the one-year ahead future performance. Using a reduced sample with three-year ahead future cash flows from operations, we obtain results qualitatively similar to tests using two-year ahead cash flows from operations. 18

20 FUTCFO is either one-year ahead, two-year ahead or cumulative two-year ahead cash flow from operations deflated by beginning of period total assets. In equation (3), we also control for factors that are likely to be associated with future cash flow from operations. These factors include current year cash flow from operations (CFO), current year firm size (SIZE) and current year sales revenues (REV). We also include year and industry dummies to control for year and industry specific effects on cash flow from operations. P_BODACA is the predicted component of abnormal current accruals associated with non-audit related governance factors BIFAC and BAFAC. Similarly, P_AUDACA is the predicted component of abnormal current accruals associated with audit related governance factors EAFAC and IAFAC. 17 If the coefficient on P_BODACA (P_AUDACA) is significant and positive, we interpret this as being consistent with non-audit related (audit related) governance structures adding value through their influence on financial reporting process, consistent with H3. 4. Analysis of Results 4.1 Characteristics of Accountability Abnormal Accruals and Governance Structures The results for estimating equation (2) are reported in Table 4. We first include the audit and non-audited related governance factors separately and then jointly in the 17 P_BODACA is obtained by first estimating equation (2) using BIFAC and BAFAC and all other control variables with signed abnormal current accruals as the dependent variable. P_BODACA is then calculated using the estimated coefficients of BIFAC and BAFAC multiply by the actual values of BIFAC and BAFAC for each firm-year observation. P_AUDACA is similarly obtained. 19

21 equation. We report results using absolute, income increasing and income decreasing abnormal accruals in Panel A, B and C of Table 4, respectively. The results in the first column of Panel A show that BIFAC is negatively and significantly associated with absolute abnormal current accruals (p-value = 0.03). 18 This suggests that effective non-audit related governance structures linked to board independence are successful in curbing discretionary reporting behavior, consistent with H2. This result persists when examining non-audit and audit related governance factors together (BIFAC = , p-value = 0.02), suggesting board independence has incremental effects on absolute abnormal accruals over other governance structures. As shown in Panel A, none of the audit related corporate governance factors are significant in explaining absolute abnormal current accruals (i.e., there is no support for H1). Panel B reports the results when we examine the partitioned sample with income increasing abnormal current accruals. We find that both the board independence factor (BIFAC) and the internal audit factor (IAFAC) are negatively related to income increasing abnormal current accruals (p-value = 0.04, and 0.02 respectively). However, when we consider the non-audit and audit related governance factors jointly, only IAFAC is significantly negative. This suggests that effective board independence and internal audit functions individually help in mitigating income increasing discretionary reporting behavior, lending support to hypotheses H1a and H2a. But, when considered together, only IAFAC has incremental effects as the significance level of BIFAC drops to eight percent. For the partitioned sample with income decreasing abnormal current accruals, we find that none of the factors are significant when audit and non-audit related governance 18 We label results with p-values less than or equal to 0.05 as statistically significant. 20

22 structures are considered separately or jointly. Therefore, there is no empirical support for hypotheses H1b and H2b. In summary, the results reported in Table 4 suggest that governance structures associated with greater board independence and effective internal audit functions help curb discretionary reporting behavior. However, these associations are primarily driven by firms with income increasing abnormal accruals and so the relation is not symmetrical between income increasing and income decreasing abnormal accruals. 4.2 Value Creation Table 5 reports the results of our analysis of value creation using future firm performance based on cash flows from operation. We consider both the predicted components of abnormal current accruals that are associated with non-audit (P_BODACA) and audit (P_AUDACA) related governance structures separately and jointly in Table 5. The first, second and third column of Table 5 details the results from using one-year, two-year and cumulative two-year ahead cash flows from operation, respectively. For one-year ahead cash flows from operations, the results show a positive and significant coefficient on P_AUDACA (p-value = 0.03) and an insignificant coefficient on P_BODACA (p-value = 0.25). This suggests that the abnormal current accruals associated with audit related governance structures are value enhancing when we consider one-year ahead cash flows from operations. This result persists even when we consider P_BODACA and P_AUDACA jointly. In contrast to the results for one-year ahead cash flows from operations, when we look at two-year ahead cash flows, the 21

23 coefficient on P_BODACA is positive and significant (p-value = 0.02) while the coefficient on P_AUDACA is insignificant (p-value = 0.21). When we examine P_BODACA and P_AUDACA jointly, only the former is positive and significant (pvalue = 0.01). Thus, for two-year ahead cash flows, the positive impact on value creation is derived mainly from effective non-audit related board governance structures. The results using cumulative two-year ahead cash flows from operation are qualitatively similar to those using two-year ahead cash flows from operation and whether we consider P_BODACA and P_AUDACA separately or jointly. In summary, we find that both audit and non-audit related governance structures are value enhancing in that they are positively associated with future firm performance as measured by future cash flows from operations. This is consistent with H3. However, audit related governance structures have a shorter term effect compared to non-audit related governance structures. This could be because non-audit related governance factors have longer term operational effects than audit related governance structures as the former is about the general and strategic oversight of the firms while the latter is only focused on the financial reporting process. 5. Summary and Conclusion We examine the accountability and value creation roles of corporate governance in the financial reporting context of Australian firms. Despite its similarity in capital market structure, corporate governance and financial reporting models with the U.S. and U.K., Australian firms have more concentrated ownership structures, similar to that observed in many other countries, as well as different governance practices from the U.S. 22

24 and U.K. (e.g., Kiel and Nicholson, 2003; La Porta, et al, 1999). The Australian context represents an interesting setting to examine our research issues associated with corporate governance and financial reporting because Australia has institutional characteristics that blend the shareholder model of corporate governance, strong financial reporting and legal protection of minority shareholders, as well as a more concentrated ownership structure together. Hence, the lessons from Australia can be instructive to policy-makers of economies with concentrated ownership that are moving towards strengthening their capital markets and protection of minority shareholders. Our results show that the board independence aspect of non-audit related board governance structures is negatively associated with absolute abnormal accruals while the internal audit aspect of audit related governance structures has an incremental and negative association with income increasing abnormal accruals. This is consistent with prior literature s interpretation that effective governance structures lead to lower levels of abnormal accruals. Our evidence also shows that the association between governance structures and abnormal accruals is not symmetrical between income increasing and income decreasing abnormal accruals. We also find evidence consistent with the value creation role of governance structures in the financial reporting context using future cash flows as our performance measure. Interestingly we find that audit related governance structures are value enhancing in the shorter term (one-year ahead performance) whilst longer term value creation is derived from non-audit related board governance structures (two-year ahead performance). This is consistent with non-audit related board structures being involved 23

25 in more general oversight and strategic direction of firms operations, which likely to generate value to firms in the longer term. Our evidence shows that in Australia board independence and internal audit functions are important governance attributes of the financial reporting process. More importantly both audit and non-audit related governance structures create value to firms. Therefore, in the context of financial reporting, Australian firms governance structures, on average, appear to fulfill their roles of accountability and value creation. Our study raises several future research opportunities. For example, do corporate governance structures in other countries also serve the accountability and value creation roles? Why is the governance-abnormal accruals association asymmetrical between income increasing and income decreasing abnormal accruals? In addition, this paper only examines one aspect of accountability as proxied by abnormal accruals. Another aspect of accountability is conservatism. Better understanding of these issues in different institutional settings can aid us in assessing the effectiveness of different governance practices under different institutional environment. This in turn can inform interested parties involved in the deliberation of governance best practice recommendations for their specific institutional environment. 24

26 References Ashbaugh, H., R. LaFond, and B.W. Mayhew Do nonaudit services compromise auditor independence? Further evidence. The Accounting Review 78 (3): Australian Stock Exchange (ASX) Corporate Governance Council Principles of Good Corporate Governance and Best Practice Recommendations. Australian Stock Exchange: Sydney. Beasley, M An empirical analysis of the relation between the board of director composition and financial statement fraud. The Accounting Review 71 (4): Becker, C.L., M.L. DeFond, J. Jiambalvo, and K.R. Subramanyam The effect of audit quality on earnings management. Contemporary Accounting Research 15 (1):1-24. Beneish, M.D Discussion of "Are accruals during initial public offerings Opportunistic". Review of Accounting Studies 3: Blue Ribbon Committee (BRC) Report and Recommendations of the Blue Ribbon Committee on Improving the Effectiveness of Corporate Audit Committees. NYSE and NASD: New York. Bosch Committee Corporate Practices and Conduct. 3 rd Edition, FT Pitman Publishing: Melbourne. Bowen, R.M., S. Rajgopal, and M. Venkatachalam Accounting discretion, corporate governance and firm performance. Working Paper, University of Washington and Duke University. Cadbury Committee Report of the Committee on the Financial Aspects of Corporate Governance. Gee and Company Ltd: London. Cheffins, B.R Comparative corporate governance and the Australian experience: A research agenda. Working Paper, University of Cambridge. Chung, H. and S. Kallapur Client Importance, Non-Audit Services, and Abnormal Accruals. The Accounting Review 78 (October): Coffee, J.C Racing towards the top?: The impact of cross-listing and stock market competition on international corporate governance. Working Paper, Columbia Law School. 25

27 Core, J.E., R.W. Holthausen, and D.F. Larcker Corporate governance, chief executive officer compensation, and firm performance. Journal of Financial Economics 51: Davidson, R., J. Goodwin, and P. Kent Internal governance structures and earnings management. Accounting and Finance forthcoming. DeAngelo, H Auditor size and audit quality. Journal of Accounting and Economics 3 (3): Dechow, P.M., R.G. Sloan, and A.P. Sweeney Causes and consequences of earnings manipulations: An analysis of firms subject to enforcement actions by the SEC. Contemporary Accounting Research 13 (Spring):1-36. Dechow, P.M., R.G. Sloan, and A.P. Sweeney Detecting earnings management. The Accounting Review 70 (2): DeFond, M., and J. Jiambalvo Debt covenant violation and manipulation of accruals. Journal of Accounting and Economics 17: DeFond, M.L., and J. Jiambalvo Incidence and circumstances of accounting errors. The Accounting Review 66 (3): Fama, E.F Agency problems and the theory of the firm. Journal of Political Economy 88 (21): Fama, E.F., and M.C. Jensen Separation of ownership and control. Journal of Law and Economics 26 (June): Francis, J.R., and J. Krishnan Accounting accruals and auditor reporting conservatism. Contemporary Accounting Research 16 (1): Francis, J.R., E.L. Maydew, and H.C. Sparks The role of Big 6 auditors in the credible reporting of accruals. Auditing: A Journal of Practice & Theory 18 (2): Frankel, R., M.F. Johnson, and K.K. Nelson The relation between auditors' fees for nonaudit services and earnings management. The Accounting Review 77 (Supplement): Han, J.C.Y., and S. Wang Political costs and earnings management of oil companies during the 1990 Persian Gulf Crisis. The Accounting Review 73 (1): Hanlon, M., S. Rajgopal, and T. Shevlin Are executive stock options associated with future earnings? Journal of Accounting and Economics 36:

28 Hilmer, F.G Strictly Boardroom: Improving Governance to Enhance Company Performance. Information Australia: Melbourne. Kelley, S.O., D. Shores, Y.H. Tong Independence in Appearance, Earnings Conservatism, and Prediction of Future Cash Flows. Working Paper, University of Washington. Kiel, G., and G. Nicholson Boards that Works: Anew Guide for Directors. McGraw-Hill: Sydney. Klein, A Audit committee, board of directors' characteristics, and earnings management. Journal of Accounting and Economics 33: Kothari, S.P., A.J. Leone, and C. Wasley Performance matched discretionary accruals measures. Journal of Accounting and Economics 39 (1): forthcoming. La Porta, R., F. Lopez-de-Silanes, and A. Shleifer Corporate ownership around the world. Journal of Finance 57 (2): Lang, M., K.V. Lins, and D.P. Miller ADRs, analysts, and accuracy: Does cross listing in the United States improve a firm's information environment and increase market value? Journal of Accounting Research 41 (2): Larcker, D.F., and S.A. Richardson Corporate governance, fees for non-audit services and accruals choices. Journal of Accounting Research 42 (3): Levitt, A Renewing the covenant with investors. Remarks delivered at the NYU Center for Law and Business, New York, May 10. Monks, R., and N. Minow Corporate Governance. 2 nd edition. Blackwell Publishers Ltd: Malden, MA. Organization for Economic Corporation and Development (OECD) OECD Principles of Corporate Governance. OECD: Paris. Park, Y.W., and H-H. Shin Board composition and earnings management in Canada. Journal of Corporate Finance 10 (3): Peasnell, K.V., P.F. Pope, and S. Young Accrual management to meet earnings targets: UK evidence pre- and post-cadbury. British Accounting Review 32: Press, E.G., and J.B. Weintrop Accounting-based constraints in public and private debt agreements: Their association with leverage and impact on accounting choice. Journal of Accounting and Economics 12 (1-3):