The determinants of Material Weaknesses in internal
control: US Empirical evidence 2005-2008
Edzard Kalff1 Student Number: 1459236 Master thesis MSc BA Finance
Faculty of Economics and Business, University of Groningen The Netherlands, February 2010
Abstract
This thesis explores the determinants of material weaknesses in reporting over internal control for 96 firms that disclosed a material weakness between January 2005 and December 2008. These firms are younger, have a lower credit rating and a lower quality of governance compared to firms not reporting a material weakness. Furthermore, material weakness firms have more business segments (operating and geographical segments), and are more involved in foreign transactions and restructurings. Their sales level appears to belong to the top 25% or they have declining sales. Firms with company wide material weaknesses are smaller, younger, have a lower credit rating, and have a higher probability of dealing with foreign transactions compared to firms reporting an account specific material weakness. The determinants of material weaknesses also significantly differ when looking at the underlying reason of the material weakness. For example, firms with complexity issues as reason for the material weakness have more SPE’s, more often report aggregate losses, are more involved in restructurings and are having a better corporate governance quality.
When the characteristics of a firm are equal to determinants of material weakness as presented in this study, the stakeholders should be aware of an increased possibility of having a material weakness.
JEL Classification: M41
Keywords: Internal control reporting, Sarbanes-Oxley Act, Material weakness
Supervisor: dr. L. Dam
2
Index
1. Introduction ... 3
2. Background and prior research ... 6
2.1 Background of Sarbanes-Oxley... 7
2.2 Fields of research... 9
2.3 Prior research on the determinants of material weaknesses ... 10
3. Hypotheses ... 12
4. Data and methodology ... 18
4.1 Material weakness classifications ... 19
4.1.1 Severity classification scheme ... 20
4.1.2 Reason classification scheme ... 21
5. Results ... 22
5.1 Multivariate analysis... 24
5.2 Severity and reason classification schemes statistics and regressions ... 28
6. Conclusions ... 36
7. References... .39
Appendix 1: Determinants of material weakness ... 44
Appendix 2: Corporate Governance Quotient® ... 46
3 1. Introduction
The objective of this study is to explore the determinants of material weaknesses in internal control over financial reporting. The empirical evidence on the determinants of material weaknesses is still limited. Therefore, this study contributes to the preliminary results of prior literature, and specifically Doyle et al. (2007), in providing evidence on the determinants of material weaknesses. This contribution is informative for managers and auditors, because they have to identify material weaknesses within the firms. The evidence is also of interest for regulators and users of financial statements. When firm characteristics are equal to the determinants of material weaknesses as been described in this study, all stakeholders should bear in mind that there is an increased possibility of material weakness presence.
According to the United States Securities and Exchange Commission2 a material weakness is a deficiency, or a combination of deficiencies, in internal control over financial reporting, such that there is a likelihood that a material misstatement of the company's annual or interim financial statements will not be prevented or detected on a timely basis. There are several reasons why material weaknesses are chosen as the proxy for internal control deficiencies. Material weaknesses are the most severe form of a significant deficiency and are expected to offer the greatest power for the statistical tests. Furthermore, disclosure of a material weakness is mandatory (Public Company Accounting Oversight Board report 2004/2007)3 and the consequences of having material weaknesses could be severe.
Sarbanes-Oxley Act (SOX) became effective in August 2002, in order to address the increasing concern of investors about the integrity of firms financial reports. This act was enabled to avoid future scandals involving worldwide respected companies (e.g. Enron and WorldCom) (Zhang et al.; 2007). Section 404 of SOX, creates an ongoing requirement of the firms’ management, and should cause companies to continue to monitor and strengthen their internal control over financial reporting. Internal control is a broad concept; it can be used to describe operational tasks, such as product quality assurance and plant maintenance (McMullen et al. 1996). In this study internal control is related to financial reporting, and its function is to detect and correct errors in the financial statements. This is also the first role of the Sarbanes-Oxley Act, and the second is to audit financial information that is reported publicly. The Sarbanes-Oxley Act is a legal obligation of all United States listed firms.
2 www.sec.gov
3
4 Furthermore, errors in the published financial information (e.g. annual reports) could lead to material weaknesses (DeFond and Jiambalvo; 1991).
Some examples that could result in a material weakness are an overstatement of the companies’ revenues in the public financial statements of $10 million dollar (on a total revenue of $80 million), or an incorrect tax report. An overstatement of incorrect tax report will lead to a material weakness when the impact is material4. Existence of a material weakness can have several consequences. Hammersley et al. (2009) found companies with a material weakness to pay higher audit fees. Secondly, they have a higher probability that the auditor resigns, and thirdly the possibility exists that their bond rating is downgraded. Especially a downgraded bond rating can have major implications. According to the literature bond rating downgrades reveal bad news about a firm’s financial condition, and increase the transaction costs (He et al. 2007). These phenomena increase the cost of capital and companies are eager to prevent this. When the material weakness is maintained the consequence are even more severe (e.g. delisting from stock exchange).
To illustrate the impact of section 404 of SOX on firms, a real-life example of a material weakness disclosure and its consequences is given.
Apparently the company described in the example above did not report their taxes correctly; this resulted in a material weakness.
4
Materiality concept defined by Financial Accounting Standard Board: The omission or misstatement of an item is material in a financial report, if, in light of surrounding circumstances, the magnitude of the item is such that it is probable that the judgment of a reasonable person relying upon the report would have been changed or influenced by the inclusion or correction of an item. www.fasb.org
5
Annual report Tenneco 2007, reported on www.sec.gov Example material weakness: Tenneco
In connection with our assessment of internal control over financial reporting under Section 404 of the Sarbanes-Oxley Act of 2002. We, Tenneco, identified a material weakness in our internal control over financial reporting relating to our accounting for income taxes as of December 31, 2006.
5 Previous research by Doyle et al. (2007) has been investigating the determinants of material weaknesses in internal control. Doyle et al. (2007) looked at all disclosures in financial statements between August 2002 and August 2005. Due to the limited time frame of this study, it is well possible that the sample was biased. Moreover, at that time management and auditors where not yet familiar with the process of implementing, evaluating and reporting on internal control. For this reason it is likely that Doyle’s study did not accurately reflect the relevant material weakness disclosures which makes it interesting to perform a study that further investigates the determinants of material weaknesses, particularly for a timeframe not directly after the implementation of Sarbanes-Oxley.
This study explores a sample of firms reporting a material weakness in internal control between January 2005 and December 2008, according to the Sarbanes-Oxley Act. In order to investigate the determinants of the material weakness disclosures, ten variables were selected for this study. Most of these variables were chosen based on their explanatory power already shown in prior research (Doyle et al. 2007, Ge and McVay; 2005). The following variables are included: firm size, firm age, aggregate loss, credit rating, SPE’s, business segments (operating and geographical segments), currency adjustment, sales growth, restructuring charge, and corporate governance score, which is an indicator aimed at measuring the overall quality of governance. This sample consists of all Fortune 600 companies that reported a material weakness between 2005 and 2008, resulting in 96 unique firms reporting a weakness. The control group consists of the Fortune 600 companies that did not report a material weakness and where all data is available (442 companies). The disclosures have been reported by Compliance Week6, and by United States Securities and Exchange Commission7.
The main findings are that firms reporting a material weakness are younger, have a lower credit rating and a lower governance score. Moreover, they have more business segments (operating and geographical segments), are more involved in foreign transactions and restructurings. Finally, their sales level is part of the top 25% sales growth or they have declining sales (further explanation will follow).
After investigating the differences between firms that did and did not report a material weakness, this research looks further into the severity and underlying reason of the material
6 www.complianceweek.com. A commercial website that lists all companies that report a material weakness 7
6 weakness. So the material weaknesses are classified into categories, as presented in table 1. This classification scheme is based upon previous research of Doyle et al. (2007).
Table 1: Classification scheme
Underlying reason classification Severity classification
• Company wide material weaknesses • Complexity issues related material weakness • Staffing issues related material weakness • Account-specific material weaknesses
• General issues related material weakness The results show that firms with a company wide material weakness are smaller, younger, have a lower credit rating, and are less involved in foreign transactions compared to companies with account-specific material weaknesses. More significant differences were expected, but results could be limited due to the sample size after classification. Moreover, these findings are in line with previous research of Doyle et al. (2007). When looking at the sub-sample of underlying reason of the material weaknesses it can be seen that these firms differ significantly in aggregate loss, the number of SPE’s, restructuring charge and corporate governance score.
The relations between the five individual classes (company wide, account-specific, complexity, staffing, and general) and the firms not reporting a material weakness is in general significant as expected. These expectations are based upon previous studies. There is a possibility that the data could have lost power as a result of the classification of the material weaknesses. The results section of this paper will further explain these relations.
The paper is organized as follows. Section 2 provides a background on SOX and prior regulations. Furthermore, research related to the determinants of material weakness disclosures will be discussed in this section. The hypotheses are stated in section 3. In the fourth section the sample and methodology will be described. In Section 5 the results of the different statistical tests are presented, and in the sixth and final section I will present the conclusions of this study.
2. Background and prior research
7 2.1 Background of the Sarbanes-Oxley Act
Internal control over financial reporting has long been recognized as an important feature of the company. Through time different acts have been implemented to regulate the securities markets (Mahoney; 2001, Kinney et al; 1990). Nowadays the Sarbanes-Oxley Act is used for this purpose, which was introduced August 2002. The first act originated from 1933, known as The Securities Act of 1933. This act passed the congress in the wake of the market crash of 1929 and the ensuing great depression (Simon; 1989). During the 1920’s the banking sector became more competitive, and practices as “beating the gun” became more and more common. To beat the gun the bank had to violate the syndicate and take orders from customers before the securities had been released for sale. The purpose of this act is “to provide full and fair disclosure of the character securities sold in interstate and foreign commerce and through the mails, and to prevent fraud in the sale thereof”.
To enforce this purpose the following laws were implemented:
1. Registration requirements: registration statement approved by Securities and Exchange commission became mandatory
2. Waiting period: 20-days waiting period between filing date and first date for sale purpose
3. Civil liabilities: buyer was empowered to sue any person signing the registration statement for losses due to “omission of facts” or misleading statements (Blum; 1936, Simon; 1989)
In 1977 the Foreign Corrupt Practices Act (FCPA) is born as result of a political scandal known as the Watergate affaire. During the 1970’s the special prosecutor found that many corporations had made substantial, illegal contributions during the 1972 presidential election campaign. Follow-up investigations by SEC revealed that over 400 US companies (e.g. American Airlines, Colgate-Palmolive) made questionable or illegal payments to foreign government or politicians. This was possible because internal corporate accounting controls were ineffective or easily subverted. The congress introduces the FCPA, to prevent bribery. It consists of three basic provisions:
1. Accounting: make and keep records, books and accounts; to reasonable detail. 2. Criminalization of foreign bribery,
8 Although these provisions are consistent with the establishment and maintenance of an effective system of internal controls, there is no materiality standard and there is little authority for enforcement of the act in the absence of evidence of wrong doing. (Beneish et al.; 2005). The concept materiality states that if information presented in the financial statements is of such magnitude that it has no influence on the user’s judgement and decision-making, it can be left out. (de Bos, and de Kimpe; 1995) According to Beneish et al (2005) the FCPA did not have a standard for materiality. This resulted in confusion to the extent by which information can be left out. An example is given; an error of thousand dollars in finished goods is of less importance when your total value of finished goods is over one million dollars, then when the company has finished goods worth ten thousand dollar (de Bos, and de Kimpe; 1995). The question was: which amounts are allowed to be left out? The authority over internal accounting controls is expanded in 1991 by the Federal Deposit Insurance Corporation Improvement Act (FDICA). This act requires banks operating in the United States to file an annual report with regulators in which management attests the effectiveness of their control, and for these controls to be subjected to independent audit. (Doyle et al.; 2007, Beneish et al.; 2005).
In 1992 the Committee of Sponsoring Organization’s (COSO) developed an internal Control-Integrated Framework. COSO consists of five interrelated components: control environment, risk assessment, control activities, information and communication, and monitoring. This framework is still an often used structure to evaluate the internal controls. (Klamm and Watson; 2009)
The well-documented accounting scandals (e.g. Enron, WorldCom) at the start of the new millennium resulted in the implementation of the Sarbanes-Oxley Act of 2002 (SOX). SOX emphasizes internal control, which is defined as “a process, effected by an entity’s board, management, and other personnel, designed to provide reasonable assurance regarding the achievement of objectives”, according to the COSO framework. (Zhang et al; 2007) New provisions emphasizing the certification and testing of internal controls by management and independent auditors were introduced to all industries in the United States (Beneish et al.; 2005).
9 controls on a quarterly basis, and report an overall conclusion about their effectiveness. Specifically, the signing officers have to compile a list of all deficiencies in internal control, or fraud of employees who are involved in internal control activities; and report all significant changes that could have a negative impact on the internal controls (Zhang et al.; 2007, Hammersley et al.; 2008).
SOX section 404 not only requires management to provide an assessment of internal control structure and procedures, but also requires auditors to provide an opinion on management’s assessment. This act became effective on the 15th of November 2004 for accelerated filers. The most severe form of a deficiency should be reported as a material weakness. The Securities and Exchange Commission defines accelerated filers as companies that; have been a public company for at least twelve calendar months; have filed at least one annual report; and have public float (the aggregate market value of the voting and non-voting common equity held by non-affiliates of the company) of $75 million or more.
It is noteworthy that even before the implementation of Sarbanes-Oxley section 404; companies began to report material weaknesses in their controls in response to the introduction of section 302, while this was not yet mandatory (Doyle et al.; 2007).
The material weakness is often reported in item 9-A of the annual report (known as 10-K filing), and item 8 consists a report of the Independent Registered Public Accounting Firm where they state whether there exists a material weakness. This information is also reported in the firms’ quarterly report (8-K filings).
2.2 Related research
The last century different acts have been adopted to increase the regulatory authority in order to overcome and make sure that firms report their financial statements correctly. This resulted in studies that investigated the effectiveness of the SOX. Is SOX effective in reducing uncertainty about the quality of firms financial reporting? Different studies have been performed to attest this question, and diverse conclusions have been drawn. Jain et al. (2004) and Li et al. (2007) have found a positive reaction, indicating that SOX is having a beneficial effect on the quality of financial reporting. When looking at earnings informativeness, and CEO/CFO certification requirement following the introduction of SOX, no significant positive market response can be identified. (Beneish et al.; 2005)
10 indicated that firms audited by non-Big 4 auditors were six times as likely to restate as those audited by Big-Four firms, and Kinney et al. (2004) found a significant negative relation between the tax-service fees and restatements. While Hoitash et al. (2009) and Ockree and Martin (2009) find that an auditor change is positive significant to the reporting of material weaknesses. Concluding, all studies report significant results that indicate audit committee quality has a positive impact on the internal control reporting quality. (Krishnan; 2005)
A third field of research is dedicated to the costs to comply with the Sarbanes-Oxley Act. Krishnan et al. (2008) finds compliance costs are positively correlated with firm size, the presence of internal control material weaknesses, the cost of setting up new computer systems and establishing formal internal control policies, the involvement of large auditors, and the appointment of new CEO’s. They found that compliance costs are negatively associated with firms in regulated industries, and with firms that raised new financing. The mean and median of the compliance costs are respectively $2.2 and $1.1 million.
Another study shows that the mean (median) audit fees to comply too SOX in 2004 compared to the fees in 2003 are 86 (126) percent higher. Audit fees for 43 percent higher for clients with a material weakness disclosure compared to clients without such disclosure. (Raghunandan and Rama; 2006)
Finally this study addresses the stream of research that is specifically related to this study, the determinants of material weakness. This subject has not been discussed intensively. However the studies performed are extensive and produce significant results that raise questions for further research. Next to these four main objects of research related to the internal control reporting there are many more individual studies that look into this subject. Since the Sarbanes-Oxley Act is active since 2002, I will first discuss previous research that provides indirect evidence on internal control.
2.3 Prior research on the determinants of material weaknesses
11 reduced the earnings the companies reported previously. Furthermore, their study reveals that companies incurring restatements are smaller, less profitable, more highly leveraged, and are growing slower compared to other companies without restatements.
A second study that looks into the characteristics of overstatements is performed by DeFond and Jiambalvo (1991) and is based upon the paper of Kinney and McDaniel. Their goal was to further investigate the determinants of companies that overstated their earnings, and defined new characteristics and used a sample of 41 companies that reported an overstatement. Their main findings presented that companies with an overstatement are more likely to have diffuse ownership, lower growth in earnings and are less likely to have audit committees. An important study performed by McMullen et al. (1996) investigated if there was a relation between financial reporting problems and the existence of a management report on internal controls (MRIC) within the company.
The paper of McMullen et al. (1996) defines a financial reporting problem as follows;
1. A Securities and Exchange Commission enforcement action, related to an accountant treatment, against the company
2. A correction of a previously reported financial statement
Based on their results they suggest that mandating management reporting on internal control may benefit companies. The costs that have to be incurred to prepare a management report are lower then the costs to be incurred when a financial reporting problem is identified. Finally McMullen et al. (1996) suggest that a management report on internal control, as is currently done with the implementation of the Sarbanes-Oxley Act in 2002, should be mandated. There are only a limited number of studies that have investigated the characteristics of internal control reporting problems or material weaknesses. This is mainly due too the lack of public data on internal controls and because of the complexity of the internal control process. (Kinney; 2000).
12 Ge and McVay (2005) suggest this is because of the larger resources of these auditors, by which they can detect more material weaknesses.
After this initiative study of Ge and McVay (2005) a second study performed by Doyle et al. (2007) gave even more insight in the determinants of material weaknesses in internal control reporting. They investigated 779 firms that disclosed a material weakness in internal control between August 2002 and August 2005. Next to comparing the firms that reported a material weakness with the control group, they also explored how determinants vary among different types of material weakness disclosures.
They used two classification schemes; one based on the severity and the second classification is based on the stated reason for the internal control problem. Their main findings showed that material weaknesses are more likely for firms that are smaller, younger, financially weaker, more complex, growing rapidly, or undergoing restructuring. Doyle’s et al. (2007) paper uses an extensive sample but is limited due to its timeframe. Furthermore, their sample could be subject to under-identification because there is a possibility that firms were unable to detect material weaknesses due too the short period after the implementation of SOX that has been taken into account.
3. Hypotheses
The first possible determinant of material weakness that will be examined is firm size. Kinney and McDaniel (1989) started with investigating the differences between firms that made corrections to their previously reported quarterly earnings and firms who did not have to make any corrections. One of their main findings is that firms with corrections are smaller (firm size). However, Krishnan (2005) did find a negative but not significant relation and researchers Wright and Wright (1995), and Xu and Tang (2007) even found a positive relationship, suggesting that larger firms would have better internal control.
13 internal auditors or consulting fees, and enjoy economies of scale when developing and implementing internal control systems (Ge and McVay, 2005; Doyle et al.; 2007). Based on prior research I expect to find less material weaknesses in larger firms. Firm size will be measured by the natural logarithm of the firms’ market capitalization.
H1: Firms with a larger market capitalization are less likely to have internal control material weaknesses
A second determinant that is likely to affect the presence of a material weakness is firm age. In line with Doyle et al. (2007), expectations were that older firms report less material weakness. The rationale behind this expectation is that longer established firms have more experience with, and better knowledge of their own internal control systems. Younger firms tend to focus relatively more on operations than on internal processes (Feng et al. 2009). Firm age is measured by the number of years DataStream8 has information available about the share price of these firms (limited to 36 years). This variable is selected to measure the experience and knowledge of financial reporting as required for listed companies.
H2: Older firms are less likely to report material weaknesses
A third possible factor that is expected to determine the internal control is a firm’s financial health. Firms with a stronger financial position are able to spend more resources on internal control than firms that have to struggle to survive. Raghunandan and Rama (2006) investigated the costs to comply with Sarbanes-Oxley. Their main finding is an increase in compliance cost of 86% after the implementation of SOX. Another research performed by Gaeremynck et al. (2008) concluded that firms with a better financial health use auditing firms that supply higher audit quality, which results in higher client financial reporting quality. In line with previous research I expect firms with a stronger financial position to have fewer internal control material weaknesses.
The first measure of financial health that is used is aggregate loss. If the line item ‘profit before extraordinary items 2009’ and ‘profit before extraordinary items 2008’ sum to less than zero are accredited one, and zero otherwise. The measurement of this variable is based on the methodology used by Doyle et al. (2007). In this study the creation of a dummy variable was to compare the results, with those of Doyle et al. Further explanation will be given in the next section.
14 A second measure of financial health that is often used is the probability from the hazard default prediction model developed by Shumway (2001). Due to unavailability of this data, this study used the credit rating as measurement tool. Credit rating governs the degree of the issuer “event risk” in the portfolio. Different types of securities have different risk profiles, this paper looked at the long term credit rating developed by Standard and Poor’s9. The modifiers of this credit rating are also taken into account, because these effectively split single- and double-A ratings grades into three sub-categories by using + and – signs. (Saperstein, 2006)
H3a: Firms with an aggregate loss are more likely to report material weaknesses H3b: Firms with a higher credit rating are less likely to report a material weakness
Firm size, firm age and financial resources are expected to affect the firms’ abilities to report their internal control in a proper manner as described above. The complexity of the firms’ business transactions or business segments will increase the need for a well organized internal control system. This idea is supported by the complexity theory that was originally designed for understanding the cost of operating formal computer logarithms and later generalized by Bailey et al. (1981). According to Wu and Hahn (1989) it is important that the internal control should be integrated into the structure, and that the complexity of this structure should be determined in order to get a fair representation of the financial statements. In this study three different tools are used to assess the complexity of the firm.
1. Special purpose entity (SPE) this is a legally distinct entity with a limited life, created to carry out a narrowly defined pre-specified activity, typically to benefit only one “sponsor” company (Coallier; 2002. An SPE is attractive from a financial reporting point of view because it is excluded from the sponsor’s balance sheet and more importantly the sponsor’s income statement could also be managed. This results in providing the sponsor with more liquidity and lower capital costs. Next to these advantages, these financial arrangements will increase the complexity of the firm’s structure and increase the possibility of material weaknesses (Feng et al. 2006; Doyle et al.; 2007).
2. Business segments; The International Accounting Standards Committee10 defines a business segment as a distinguishable component of an enterprise engaged in providing an individual product or service or a group of related products or services and subject to risks
9 www.standardandpoors.com 10
15 and returns that are different from that of other segments. Business segments are divided in operating and geographical segments. A segment qualifies for an operation segment when:
1. It engages in business activities that may or may not earn revenues or incur expenses.
2. Its operating results are regularly reviewed by the enterprise's chief operating decision maker to make decisions about resources to be allocated to the segment and assess its performance
3. Discrete financial data are available.
A geographical segment is a distinguishable component of an enterprise engaged in providing products or services within a particular geographic or common bloc, and subject to risks and returns that are different from those operating in other areas or blocs.
Geographical segments are based either on the location of an enterprise's input factors or the location of its final customers. The number of business segments is the sum of the number of
operating and geographical segments, when the number of business segments within a firm increases this will affect the complexity of the firm. So the number of business segments is used as a measure for complexity. (Albrecht and Chipalkatti; 1998)
3. Currency adjustment; when a foreign currency adjustment is present, it is assumed that the firm deals with foreign transactions or subsidiaries. (Shank and Shamis,1979) This will lead to a more complex structure.
H4: Firms with a higher number of SPE’s or business segments or/and presence of a currency adjustment are more likely to report a material weakness.
A fifth factor that could be a possible determinant of internal control material weaknesses is sales growth. Firms that realize a high growth percentage may outgrow their internal control it has in place. Firms will hire extra personnel or establish new processes in order to update their internal control. Firms with a high sales growth (part of top 25% growth of sample) are expected to report more material weaknesses (Ge and McVay, 2005; Doyle et al.; 2007). On the other hand firms with declining sales will possibly reduce the number of controlling / reporting personnel or focus on opportunities to increase their sales, or reduce costs. This could also lead to higher number of reported internal control material weaknesses.
16 Another possible determinant is the presence of a restructuring charge during the past three years. Restructuring charges are the result of actions that are aiming for an improvement of their operating performance (Atiase et al., 2004). Often a restructuring is paired with reduced work force, increased product innovation, or curtail investment in excess production capacity (Baxter 1999). The reduced labour force can lead to a lack of adequately educated accounting staff. Also difficult accrual estimations and adjustments have to be made (tax, impairments, and restatements) (Doyle et al., 2007). Lack of staff and ascribed difficulties will lead to more internal control material weaknesses. The dummy variable “restructuring charges” will equal one if the firm restructured during the year 2007 or 2008, and zero otherwise.
H6: Firms that reported a restructuring charge the past two years are more likely to report a material weakness
The final possible determinant of internal control quality this paper reviews is corporate governance. The financial reporting system provides a mean by which providers of capital can monitor managers (Sloan, 2001). As representatives of owners, corporate boards of directors are responsible for supervising the financial reporting function. SOX and its related regulations provide an increased emphasis on corporate governance as well as sources of data through which to measure the linkage of governance to internal control quality (Hoitash et al.; 2009). Proceeding previous research expectations were that well governed firms report fewer material weaknesses. In this study corporate governance is measured by Corporate Governance Quotient® developed by the RiskMetrics Group. The CGQ is divided in eight categories: board, audit, charter/bylaws, state of incorporation, executive and director compensation, qualitative factors, ownership, and director education. These main categories are subdivided in 65 factors that are enclosed in appendix 2
H7: Firms with a higher score on corporate governance are less likely to report material weaknesses
17
11 www.standardandpoors.com 12
www.businessweek.com and www.reuters.com
Table 2: Measures of hypothesized determinants of material weaknesses and the predicted relation
Determinant Predicted direction Measure Firm size Firm age Financial health Aggregate loss Credit rating Complexity SPE’s Business segments Currency adjustment Other determinants Sales growth Restructuring charge Quality of governance - - + - + + + + + -
The natural logarithm of market capitalization
The natural logarithm of the years that DataStream has stock price data available
A dummy variable that is equal to one if the earnings before extraordinary items in years 2007 and 2008 sum to less than zero, and zero otherwise
Index number based upon the long-term credit rating of Standard and Poor’s; AAA – C corresponding to 22 – 111
The natural logarithm of the number of Special Purpose Entities as documented by the Securities and Exchange Commission in 2008
The natural logarithm of the number of business segments (sum of operating and geographical segments) reported by either Thomson Reuters and BusinessWeek12
Dummy variable that is equal to one if in the cash flow statement or elsewhere in the annual report an adjustment for foreign currencies is stated, and zero otherwise
Dummy variable that is equal to one if sales growth is part of top 4th quartile or sales is declining, and zero otherwise
A dummy variable that is equal to one If the firm performed a restructuring during the year 2007 or 2008, and zero otherwise Governance score measured by the Corporate Governance Quotient®, tool that consists of 63 corporate governance variables
18 4. Data and methodology
The material weaknesses observed in this study are reported between January 2005 and December 2008 by the Fortune 600 companies according to the list compiled in 2008. Compliance week is a website that collects all monthly, quarterly and annual reports on firms that have reported a material weakness. 13 This website is used to obtain the data of the companies that have reported a material weakness in the time frame. 96 Individual companies have reported a severe internal control deficiency, a material weakness.
Nevertheless, the gathering and classification of data was a time consuming process. First, for all 600 companies the annual reports for the period 2005 to 2008 had to be scanned for material weaknesses, these were extracted from the Securities and Exchange Commission database. Secondly, all firms that reported a material weakness had to be found in the database of Compliance week in order to determine the severity and underlying reason of the reported material weakness. Afterwards, a check was performed by locating all the information of each firm reported on Compliance week in order to indentify firms that indentified a material weakness, but could not be found on the Securities and Exchange Commission database.
When a firm reports several material weaknesses I do not include this firm multiple times in the sample, descriptive statistics and statistical analyses, but regard it simply as having a material weakness. Furthermore, the EDGAR database has been used, which is constructed and maintained by the US Securities and Exchange Commission. After identifying companies that reported a material weakness I obtained financial data about these companies, resulting in exclusion of seven companies due to lack of data. This set of companies is not as extended as in previous research, but compared to previous research, my time frame covers a longer period and does not include the period directly after the implementation of SOX compliance in to consideration. This would reduce the chance that companies are unaware of the legal obligation to comply to the Sarbanes-Oxley legislation. Next to this, both companies and auditors have more experience in detecting material weaknesses.
The control firms that have been selected are all the Fortune 600 companies that did not report a material weakness between January 2005 and December 2008. 491 Companies did not report a material weakness. After excluding the companies that reported a less severe
19 internal control deficiency and/or companies with a lack of data the final control sample consisted of 442 companies. The total sample therefore consists of a crossection of 538 companies. It is important to state that complying with section 404 of the SOX is a legal obligation. In table 3 the sample is presented it is noteworthy that 17.16% of all Fortune 600 companies identified a material weakness. Previous research of Doyle et al. (2007) presented that 16.36% of all companies reported a material weakness.
Table 3: Sample selection
Sample size (1) (2)
Material weakness firms (Fortune 600) 103 970
Excluded: as a result of a lack of data
Total companies that report one or more material weaknesses during timeframe
(7) 96
(191) 779
Control group (Fortune 600)
All firms
Firms that reported material weakness
600 (96)
5,935 (779)
Firms that reported less severe deficiency or exclusion due to lack of data (60) (109)
Final control Group 442 5047
This table presents the sample size of this study in row (1). The second row (2) presents the sample size of previous research performed by Doyle et al. (2007). First, the sample size of the firms with material weaknesses is presented and secondly the sample size of the firms without material weaknesses.
4.1 Material weakness classifications
A Material weakness is the most severe internal control deficiency, but exists in many firms. The classification of internal control material weaknesses exhibits two dimensions. First, the severity of the weakness differs from firm to firm. Second, the underlying reason; there are numerous underlying reasons for material weaknesses. These dimensions are based upon prior research from Doyle et al. (2007) and annual reports of firms that reported a material weakness. In the past many different classification schemes have been used.
The GAO (Report Pursuant to Section 704 of the Sarbanes-Oxley Act of 2002)14 classifies internal control weaknesses into four broad categories:
1. Improper revenue recognition (improper timing, fictitious revenue, and improper valuation)
14
20 2. Improper expense recognition (improper capitalization/deferral, overstating ending Inventory values, improper use of restructuring and other liability reserves, understating reserves for bad debts and loan losses, and failure to record asset impairments)
3. Improper accounting in connection with business combinations (improper asset valuation, improper use of merger reserves and inappropriate application of purchase/pooling methods)
4. Other areas of improper accounting (inadequate disclosures in management discussion, failure to disclose related party transactions, inappropriate accounting for non-monetary and roundtrip transactions, foreign payments in violation of the FCPA, improper use of Non-GAAP Financial Measures, and improper use of Off-Balance Sheet Arrangements).
Ge and McVay (2005) grouped nine different deficiency types (1) Account-Specific, (2) Training, (3) Period-End Reporting/Accounting Policies, (4) Revenue Recognition, (5) Segregation of Duties, (6) Account Reconciliation, (7) Subsidiary-Specific, (8) Senior Management, and (9)Technology Issues. This classification has also been used by Frank and Cheh (2006). Further research performed by Doyle et al (2007) used a more summarized classification scheme. Since this scheme was found to be most suitable, it will be used throughout this study. This research adopts the classification schemes developed by Doyle et al. (2007), the first one will look into the severity of the material weakness and the second scheme investigates the underlying reason of the internal control problem.
4.1.1 Severity classification scheme
To determine whether a material weakness is severe a classification scheme developed by the bond rating company moody’s has been used. According to this classification an internal control problem can be defined as; (Bizarro et al.; 2007)
21 2. Company level material weaknesses are included in Category B and include control weaknesses such as an ineffective control environment, weak overall financial reporting processes, tone at the top, delinquent filers and ineffective personnel. Category B weaknesses also include those companies that report several Category A weaknesses and those companies that have reported material control weaknesses for the second year running. This group would seem to contain younger, financially weaker and poorly governed firms that do not have the resources (time and money) to spend on an internal control system.
This classification scheme is mutually exclusive, internal control problems are either account specific or company wide and not both.
4.1.2 Reason classification scheme
A second classification scheme developed by Doyle et al. (2007) is based upon the stated reason for the material weakness. Three categories of material weakness have been created.
1. Staffing: when a material weakness is a result of a lack of adequate personnel, resources or when there is an inadequate segregation of duties the weakness will be addressed to staffing
2. Complexity: weaknesses as result of complex accounting standards will be addressed to complexity, often this are internal control problems related to hedging and derivatives.
3. General: weaknesses that are related to period-end reporting problems or revenue recognition issues will be subscribed to the general category
Example
“We did not maintain effective controls over the financial reporting process due to an insufficient complement of personnel with an appropriate level of accounting knowledge”
Core-Mark holding - December 2007
Example
“We concluded that our foreign exchange forward contracts did no qualify for cash flow hedge accounting since…” Western Union - 31 December 2006
Example
22 This classification scheme is not mutually exclusive. An identified material weakness can be assigned to 1, 2, or 3 of the categories. For example a company having difficulties with period-end reporting as a result of a staffing problem so severe that it leads to a material weakness will be assigned to the general and staffing category. I expect the determinants of the material weaknesses to differ across these categories. Staffing issues will be present at companies that are financially weaker, younger. Complexity problems are expected at older larger and diversified companies, while general issues are expected at firms with a bad governance rating. (Doyle et al.; 2007)
5. Results
The descriptive statistics of the material weakness firms and control firms (Fortune 600; excluding firms that did report a material weakness) are presented in table 4. The t-test and Wilcoxon rank-sum test are performed to measure the differences in mean and median between the material weakness firms and the control group for the selected variables. As mentioned above for some variables the natural logarithm has been taken in the regression analysis. However the descriptive statistics are based upon the raw data (no logarithms). This applies to the variables market capitalization, firm age, SPE’s, and number of business segments.
The first variable firm size, measured by the log of market capitalization is significant lower for firms that reported a material weakness. The table shows that the mean of firms with material weakness is 9.8 B US dollars and the control group has a mean market capitalization of 21.8B US dollars, which results in a significant difference. The mean and median market capitalization presented is much higher then documented in previous research of Doyle et al. ($180 M and $224 M). This is a result of selecting only the 600 largest companies. These results will be further examined during the multivariate tests.
23 compared to the control group. These findings are in line with the expectations and previous studies.
The variables SPE’s and business segments do not differ significantly between both groups when the natural logarithm has been taken, indicating that these variables do not determine the presence of a material weakness. I expected companies which reported a material weakness to have more business segments and SPE’s, as was found by prior research of Doyle et al. (2007). To measure the foreign transactions, the reporting of a currency adjustment is measured. Here also seems to be the preliminary proof that the probability of having a material weakness is higher for firms ’which reported a material weakness. This is in line with the expectations and previous studies.
The sales growth variable is measured by a dummy variable, that is equal to one if growth is part of top 25% of sales growth or when sales is declining, otherwise zero. This variable is significant higher for firms reporting a material weakness indicating that firms reporting a material weakness more often present extreme growth numbers (top 25%) or declining sales. This finding is in line with the expectations. According to the literature firms with high growth rates may outgrow their internal controls, and may need time to establish new procedures or hiring new personnel. (Doyle et al. 2007) Firms with declining sales may lack resources and will reduce the number of control staff or focus on opportunities that increase their sales. Sales growth and the earlier mentioned aggregate loss look at first sight slightly the same, but when looking at the correlation matrix there in no strong correlation between these possible determinants. This is confirmed in the results section. Restructuring is also significant different, indicating firms that perform a restructuring are reporting a material weakness more often. This finding is support by prior research; restructurings come with layoffs and difficult accounting estimates which leads to internal control deficiencies. (Baxter; 1999)
24 Table 5 presents a Pearson correlation matrix and shows that there are some variables correlating with each other. Examples given; Aggregate loss is correlating negatively with the variables governance score, market capitalization, and credit rating. While Credit rating, firm age and governance score are correlating positively with each other. The findings described above are based upon the t-test, Wilcoxon rank-sum test, and Pearson correlation matrix and will be further investigated during the multivariate analysis.
5.1 Multivariate analysis
The probability of disclosing a material weakness in internal control is modelled using a logistic regression, also known as logit regression. This methodology is used because the dependent variable, material weakness, is a dummy variable that can take values of 0 and 1. (Brooks; 2008) The following equation is estimated:
+ + + + + + + + + + + = ε β β β β β β β β β β β Governance 10 ing Restructur 9 growth Sales 8 adj. Currency 7 Segments 6 SPE´s 5 rating Credit 4 loss Aggregate 3 age Firm 2 size Firm 1 0 ) weakness material ( f prob (1)
Here f is the logistic function. The estimation results can be found in table 6. The first column presents the predicted direction of the variables based upon earlier research. The second column presents the coefficients and significance. The direction of logit regression estimates are all in the same direction as predicted, though not al coefficients are significant. Looking at the significances it can be seen that firm age, credit rating sales growth, restructuring, and corporate governance are highly significant. The variables firm age and currency adjustment are also significant, but on a lower level. SPE’s and market capitalization appear not to be a determinant of material weakness when these two groups were compared.
25 Table 4: Descriptive statistics of material weakness firms and Fortune 600 (excluding material weakness firms)
Material weakness firms (96) Fortune 600, no material weakness firms
Variable Mean Median MIN MAX Std. dev. Predicted
Direction
Mean Median MIN MAX Std. dev
Market capitalization 9,8 B 3,7 B 13,78 M 163,2 B 21,7 B < 21,8BA 7,5 BA 7,9 M 344,5 B 39,3 B Firm age 21,2 19 1 36 12,33 < 23,98A 24B 1 36 11,61 Aggregate loss 0.43 1 0 1 0,49 > 0,19A 0A 0 1 0,39 Credit rating 9,06 8 1 21 4,32 < 14,23A 15A 1 21 3,47 SPE’s 40,46 18 0 902 96,94 > 30,49 7 0 1336 99,15 Business segments 5.84 5 2 15 2,95 > 3,54 3 0 19 2,02 Currency adj. 0,69 1 0 1 0,46 > 0,53A 1B 0 1 0,49 Sales growth 0,70 1 0 1 0,46 > 0,25A 1A 0 1 0,44 Restructuring charge 0,82 1 0 1 0,38 > 0,42A 0A 0 1 0,49 Corporate governance 0,54 0,55 0,04 0,99 0,23 < 0,70A 0,63A 0,04 0,96 0,28
26 Table 5: Pearson correlation matrix
Material weakness
Market capitalization
Firm age Aggregate
27
Table 6: The determinants of material weaknesses (Fortune 600 for the years 2005-2008 )
Dependent variable = MW
Independent variable Predicted
sign Logit estimate
Intercept -4.24 (-1.51) Market capitalization - -0.11 (-0.37) Firm age - -0.729C (-1.36) Aggregate loss + 0.21 (0.50) Credit rating - -0.36A (-6.24) SPE’s + 0.103 (0.38) Business segments + 5.36A (6.09) Currency adjustment + 0.52C (1.40) Sales growth + 2.00A (5.52) Restructuring charge + 2.50A (5.54) Corporate governance - -2.14A (-2.89)
Number of material weaknesses 96
N
Likelihood ratio Χ2
538 282.37
p-value (0.000)
This table contains a logit regression. Material weakness is a dummy variable, which is equal to one if a material weakness is present, and zero otherwise. All determinants are in the predicted direction. The hypothesized determinants are explained in more detail in table 2. The estimated equation:
+ + + + + + + + + + + = ε β β β β β β β β β β β Governance 10 ing Restructur 9 growth Sales 8 adj. Currency 7 Segments 6 SPE´s 5 rating Credit 4 loss Aggregate 3 age Firm 2 size Firm 1 0 ) weakness material ( f prob A, B, and C
28 5.2 Severity and reason classification schemes statistics and regressions
As mentioned earlier a classification to severity has been made, resulting in the groups’ account-specific material weakness and company-wide material weakness disclosures. Company-wide weaknesses can be seen as the more severe internal control disclosures. The reason classification consists of General, Complexity, and Staffing issues. These three groups are possible reasons for disclosures that have been identified by previous research and are used in this study. In Table 7 an overview is given of the number of material weakness in each class, and as can be seen these frequencies are in line with those of Doyle et al (2007. The material weaknesses reported in the underlying reason classification scheme do not add up to the total of all material weaknesses, because this scheme is not mutually exclusive. Indicating that firm’s can have a material weakness that is for example related to complexity and staffing.
Table 7; Frequency table.
(1) (2) # of material Material Weaknesses Classification as % of all material weakness Doyle’s # of Material Weakness Classification as % of all material weakness All material weaknesses 96 100% 779 100% Severity - Account specific - Company wide 69 27 71,8% 28,2% 491 286 63% 37% Underlying reason - Complexity - Staffing - General 36 23 68 37,5% 24% 70,8% 251 347 519 32,2% 44.5% 66,2%
This table presents the number of material weaknesses for this study and the study performed by Doyle et al. (2007). There are two classifications schemes. First, severity classification contains company wide and account specific material weakness. The second classification is based upon the underlying reason of the material weakness and contains complexity, staffing and general material weaknesses. The first row (1) presents the classification scheme for this study. The second row (2) indentifies the classification scheme of previous research performed by Doyle et al. (2007)
29 When comparing the firms that reported a company wide material weakness with those who reported an account-specific material weakness it can be seen that the firms reporting company-wide disclosures are smaller, younger, have a lower credit rating, and have a higher probability of dealing with foreign transactions. For the variables aggregate loss, SPE’s, business segments, sales, restructuring, and governance no significant differences have been found. The firms reporting account-specific weaknesses are older, more mature, have a better credit rating and report less foreign transactions. Surprisingly the currency adjustment is not in the predicted direction. I expected the account-specific disclosures to deal more with foreign transactions, because these firms are larger. These findings are for a large part in line with the expectations and prior research. Doyle et al. (2007) found significant support that account specific material weaknesses report less aggregate losses. This study did not find significant evidence for this relation. In the sample an interesting significant result has been found, indicating that the credit-rating is higher for firms which reported an account-specific material weakness, indicating that firms with a lower credit rating more often report company wide material weakness.
Table 8: descriptive statistics of internal control disclosure by severity classification scheme
Account specific Company-wide
Variable Mean Median Predicted Direction Mean Median
Market capitalization 12.4B 4.9B > 2.9B B 1.3B A Firm age 22.67 21 > 17.48A 18B Aggregate loss 0.45 0 < 0.37 0 Credit rating 9.78 9 > 7.18A 7A SPE’s 45.14 19 > 28.33 17 Business segments 5.88 5 > 5.74 5 Currency adj. 0.63 1 > 0.85A 1B Sales growth 0.68 1 > 0.74 1 Restructuring charge 0.82 1 > 0.81 1 Corporate governance 0.55 0.57 > 0.52 0.50
This table presents the descriptive statistics of the severity classification scheme. The hypothesized determinants are described in more detail in table 2. The t-test is used to measure the difference in means, while the medians are compared using the Wilcoxon rank-sum test. The company-wide material weakness (MW) group consists of 26 companies and the account-specific MW group of 70 companies. The variables market capitalization, firm age, SPE’s, and business segments are presented in the raw data form. For testing their significances the natural logarithm of these variables has been taken. Sales growth is a dummy variable; 1 indicates a sales growth that belongs to top 25% of sample or when the sales are declining.
30 Table 9 presents the classification scheme by underlying reason of the material weakness disclosure. This scheme consists of 36 observations that are related to complexity, 23 to staffing problems, and finally 68 observations where the material weakness was a result of general issues. It is important to stress again these classifications are not mutually exclusive. In the descriptive statistics table the mean and medians of the selected variables between the groups are compared by using the t-test and the Wilcoxon rank-sum test. The number of significant differences between the identified variable between the groups is lower then expected. The preliminary result that can be seen from these descriptive statistics is that the reason behind the material weakness disclosures can only be partly determined by the selected variables.
When looking at the complexity issues group it can bee seen that these companies report significantly smaller aggregate losses. In line with the expectations and prior research the complexity group also has more SPE’s, when comparing with the general and staffing issues groups. Another significant finding is related to the restructuring variable. Firms facing more complexity score significantly higher, indicating that this group performs more restructurings then the staffing and general group. A final finding indicates that complexity firms’ score higher on the corporate governance measure developed by the RiskMetrics Group when comparing the mean and median with the general issues group. Firms with material weakness disclosures related to staffing issues have significantly less SPE’s then the complexity group, but more then the general group. Next to this their restructuring variable is significantly the lowest of all groups, indicating that less restructurings are performed within this group.
31
Table 9: descriptive statistics of internal control disclosure by reason classification scheme
Complex Staffing General
Variables 36 observations 23 observations 68 observations
Mean Median Mean Median Mean Median
Market capitalization $ 4.3B 13.0B 3.4B 6.7B 7.0B 3.3B Firm age 20.9 20.5 20.69 19 21.39 20.5 Aggregate loss 0.33 G3 0 0.39 0 0.48 C3 0 Credit rating 8.80 7.5 8.95 8 9.07 8 SPE’s 72.44 S3, G1 32 G1 33.39 C3, G3 23 G3 23.72 C1, S3 12.5 C1, S3 Business segments 6.58 6 6.21 5 5.73 5 Currency adjustment 0.72 1 0.73 1 0.69 1 Sales growth 0.07 0.06 0.07 0.08 0.09 0.09 Restructuring charge 0.94 S1, G1 1 S1 0.60 C1, G1 1 C1 0.79 G1, S1 1 Corporate governance 0.59 G2 0.60 G2 0.55 0.54 0.49 C2 0.51 C2 This table presents the descriptive statistics of the underlying reason classification scheme. The hypothesized determinants are described in more detail in table 2. As mentioned before, this classification scheme is not mutually exclusive. The mean of the groups; complex, staffing, and general are compared with each other. Again the variables market capitalization, firm age, SPE, and business segments are presented in their unlogged form, while the test have been performed with their logged equivalent
G1, G2, G3 When significantly different from General group at 1%, 5%, 10% level, tested with t-test for the mean
and Wilcoxon rank-sum test for the median.
S1, S2, S3 When significantly different from staffing group at 1%, 5%, 10% level, tested with t-test for the mean
and Wilcoxon rank-sum test for the median.
C1, C2, C3 When significantly different from staffing group at 1%, 5%, 10% level, tested with t-test for the mean
and Wilcoxon rank-sum test for the median.
32
Table 10: The determinants of material weaknesses. Classification by severity: account-specific and company wide material weaknesses (Fortune 600 for the years 2005-2008)
Dependent variable = MW account-specific
Dependent variable= MW company-wide Independent variable Logit estimate Logit estimate
Intercept -7.28B (-2.38) -1.83 (-0.23) Market capitalization 0.37 (1.16) -0.85 (-0.88) Firm age -0.89 (-1.45) 2.17C (1.66) Aggregate loss 0.49 (1.08) -1.03 (-1.00) Credit rating -0.32A (-5.37) -1.04A (-3.45) SPE’s 0.07 (0.26) 0.83 (0.89) Business segments 5.14A (5.64) 10.58A (3.26) Currency adjustment 0.14 (0.38) 8.62A (3.34) Sales growth 1.85A (4.89) 3.97A (3.35) Restructuring charge 2.49A (5.22) 2.99B (2.45) Corporate governance -2.11A (-2.66) -5.06 A (-2.76)
Number of material weaknesses 70 26
N Likelihood ratio Χ2 512 205.21 468 166.48 p-value 0.00 0.00
Account-specific MW and company-wide MW are the two severity classes developed by Moody’s. The test performed is a logit estimate regression. The dependent variable is the classified material weakness. A, B, and C correspond to significance levels of 1%, 5%, and 10%. The estimated equation where the material weakness is account-specific (AC) or company wide (CW):
33 First the regression outcomes for the account-specific material weakness disclosures will be discussed. The dependent variable MW account-specific is equal to 1 if a firm reported a material weakness that is related to a specific account, or part of the organization. The variables credit rating, business segments, sales growth, and corporate governance score are all significant in the predicted direction (table 2). According to the regressions, firms that report an account-specific related material weakness, have a lower credit rating, more business segments, restructure more, perform worse on their governance score and are having extreme sales growth, or declining sales compared to the Fortune 600. These findings were expected and supported by prior research of Doyle et al. (2007). But Doyle also found significant results for market capitalization, number of SPE’s, and currency adjustment.
The second column contains a logit regression with company wide material weaknesses as dependent variable. The dependent variable is equal to one if the company reported a material weakness related to the whole company (e.g. weak overall financial reporting processes). The outcomes of this logit regression look in many ways the same as the outcomes for account-specific related material weaknesses. The firms here are also having a lower credit rating, more business segments, restructure more, perform worse on their governance score and are having extreme sales growth, or declining sales compared to the Fortune 600. An interesting finding is that firms reported company-wide material weaknesses seem to be older than the control group. This was not expected and supported by prior studies. A possible explanation is that these older firms have more experience in detecting material weaknesses compared to younger firms, and therefore they report material weaknesses more often. The results of Doyle et al. (2007) only show that firms with company wide weaknesses are younger, report more aggregate losses and have more segments and SPE’s.
Table 11 presents the regression output for the second classification scheme where the underlying reason of the reported material weakness is investigated. The first column contains complexity as dependent variable. Complexity is equal to one if the firm disclosed a material weakness as a result of complexity issues, and zero if it did not disclose a material weakness. The second and third columns have staffing and general as dependent variable.
34 related to the complexity of the firm has more SPE’s and business segments, has to deal with foreign transactions and restructurings, its credit rating is lower and more often are its earning part of the top 25% sales growth or are the sales declining.
The second column contains a regression with staffing issues as dependent variable. The outcomes of this regression are for a part the same as those for complexity issues.
The positive significant predictors are the number of business segments, currency adjustment, sales growth, and restructuring charge. Negative significant predictors are the credit rating and governance structure. Overall, firms with material weakness disclosures as result of staffing problems have more business segments, have to deal with foreign transactions and restructurings, and have extreme sales growth or declining sales. Next to this their credit rating is lower and so is their governance score.
35
Table 11: The determinants of material weaknesses. Classification by underlying reason: complexity, staffing, and general issues (Fortune 600 for the years 2005-2008)
Dependent variable = MW complexity Dependent variable= MW Staffing Dependent variable= MW General Independent variable Logit estimate Logit estimate Logit estimate
Intercept -10.37 (-1.74) -3.94 (-0.94) -4.02 (-1.25) Market capitalization 0.05 (0.08) 0.07 (0.61) 0.06 (0.18) Firm age -0.47 (-0.28) 0.50 (0.15) 0.98C (1.55) Aggregate loss -1.17 (-1.40) 0.13 (0.19) 0.52 (1.08) Credit rating -0.51A (-4.09) -0.40A (-4.23) -0.35A (-5.37) SPE’s 1.58B (2.56) 0.20 (0.44) -0.32 (-0.97) Business segments 8.30A (1.86) 4.46A (3.51) 5.86A (5.54) Currency adjustment 1.61B (2.11) 1.74A (2.34) 0.85C (1.89) Sales growth 2.81A (4.01) 2.06A (3.46) 2.11A (4.95) Restructuring charge 4.24A (2.56) 0.90C (1.45) 2.63A (4.90) Corporate governance 0.29 (0.20) -1.91C (-1.70) -3.68A (-4.09) # of material weaknesses 36 23 68 N Likelihood ratio Χ2 478 181.63 465 86.93 510 231.20 p-value 0.00 0.00 0.00
Material Weakness (complex, staffing, and general) is a dummy variable. A logit estimate regression is performed. A, B, and C
correspond to the significance levels of 1%, 5%, and 10%. The estimated equation in which the material weakness is caused by complexity (C), staffing (S) or general (G) issues:
