X
تبلیغات
مدیریت اجرایی -
وبلاگ گروهی دانشجویان کارشناسی ارشد رشته مدیریت اجرایی دانشگاه علوم و تحقیقات ارومیه
 

International Journal of Accounting Information Systems 9 (2010) 1 – 20 Do early and voluntary filers of financial information in XBRL format signal superior corporate governance and operating performance? % Ronald F. Premuroso %, Somnath Bhattacharya Florida Atlantic University, 777 Glades Road, Boca Raton, FL 33431-0991, United States Received 27 February 2009; received in revised form 23 November 2009; accepted 3 January 2010 Abstract On March 16, 2005, the SEC issued Final Rule 33-8529 encouraging registrants to voluntarily file tagged financial statement information on the EDGAR reporting System using XBRL format. In this paper, we examine whether early and voluntary filers of financial information in XBRL format demonstrate superior corporate governance and operating performance relative to their non-adopting peers. We investigate performance, market, and structure-related firm variables. Our results suggest that corporate governance is significantly and positively associated with a firm's decision to be an early and voluntary filer of financial information in XBRL format. At the same time, firm performance factors including liquidity and firm size are also associated with the early and voluntary XBRL filing decision. Our findings should be particularly interesting for the SEC, as it considers the corporate governance and firm-performance related associations between certain registrants' early and voluntary response and its call for XBRL-based filings. c 2010 Elsevier Inc. All rights reserved. Keywords: Corporate governance; Signaling theory; Voluntary disclosure; XBRL We are grateful for helpful comments from participants at the American Accounting Association Concurrent Session in August 2006 in Washington, D.C., as well as the participants at the 15th Annual AI/ET Research Workshop in August 2006; participants at the 6th Annual XBRL Conference at Bryant University in June 2006; and attendees at workshops of members of the School of Accounting and the Finance Department in the College of Business at Florida Atlantic University. We are also grateful for the comments from two reviewers, the associate editor, and the editor of the International Journal of Accounting Information Systems. % Corresponding author. Tel.: +1 561 297 3636. E-mail addresses: premuros@fau.edu (R.F. Premuroso), sbhatt@fau.edu (S. Bhattacharya). 1467-0895/$ - see front matter c 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.accinf.2010.01.002 % 2 R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 1. Introduction The Securities and Exchange Commission (SEC, hereafter) issued its Final Rule 33-8529 regarding the XBRL Voluntary Financial Reporting Program on the EDGAR System on March 16, 2005. This Rule encouraged SEC registrants to voluntarily file certain mandated reports using the eXtensible Business Reporting Language (XBRL) format as exhibits to EDGAR filings required under the Securities Exchange Act of 1934. These XBRL submissions were to occur simultaneously with existing HTML1 or ASCII2-based plain-text formats of reporting to its electronic data gathering, analysis and retrieval (EDGAR) system (Debreceny et al., 2005). The SEC's motivation for the rule was to examine the feasibility, desirability and usefulness of using XBRL-tagged data to registrants, investors, the Commission and the marketplace.3 Debreceny et al. (2005) critically examined the implications and feasibility of the rule as part of a working party under the aegis of the Information Systems and Artificial Intelligence/Emerging Technologies section of the American Accounting Association (AAA, hereafter). The Debreceny et al. (2005) Committee emerged with strong support for the SEC initiative for the purposes of furthering greater transparency, stewardship and the smooth functioning of capital markets (Debreceny et al., 2005) The Committee saw XBRL as vital for the democratization of markets and recommended that the SEC not only consider adopting XBRL for Form 8-K filings but eventually mandate the XBRL format for all submissions made to the SEC. This strong endorsement of the proposed rule by the AAA Committee and recent voluntary financial statement filings by U.S. firms in XBRL allow us to examine the differential associations and implications related to firm corporate governance and operating performance for early and voluntary adopters of the SEC Rule. Given the strong corporate governance-based support for the SEC proposal by the AAA Committee, we examine whether early adopters of the SEC rule (those that have commenced financial statement reporting in XBRL format4) signal superior corporate governance and operating performance to stakeholders, investors and other market participants. The SEC has begun offering incentives to companies to participate in the XBRL filing program (Baldwin et al., 2006). Firms too have many incentives to voluntarily file financial statements in the XBRL format. Financial reporting in the XBRL format can streamline internal and external financial reporting for the firm. This, in turn, reduces the costs of compiling and reporting financial information. Adopting XBRL as a disclosure tool also levels the playing field for investors and reduces potential disparities between firms with regard to disclosure level and content.5 Given these incentives, early filers in XBRL format may expect to be rewarded by the marketplace in the form of greater and/or easier access to capital markets and future positive or improved corporate governance or credit agency rating evaluations. Voluntary XBRL-related financial statement filing initiatives, both within the U.S. and abroad, therefore suggest a potential association between corporate governance, firm performance and early and voluntary adoption of the XBRL technology. The fact that some early adopters to date are vendors of the XBRL technology does not necessarily undermine this expected link between HyperText Markup Language. American Standard Code for Information Interchange. 3See: http://www.sec.gov/rules/final/33-8529.htm#VI for SEC Final Rule 33-8529. Filing in XBRL format is not (yet) a requirement for all firms in any country as of the date of this manuscript. 4The number and types of firms filing in XBRL format is slowly growing since the Final Rule was passed. 5The Wall Street Journal reports that SEC Chairman Cox has been encouraging the adoption of the XBRL format for corporate filings without necessarily mandating such adoption (Scannell, 2006b). 2 1 R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 3 XBRL adoption and corporate governance. Given the SEC's interest in the technology based on its desire to effect greater corporate transparency (Scannell, 2006b), it is hardly surprising that XBRL-related vendors are among the earliest filers in the XBRL format. Additionally, the fact that these vendors are also among the leaders of the technology sector serves to not only demonstrate the efficacy of the technology itself but also the fact that these firms expect to gain from the greater transparency effected by filing in the XBRL format. While some may argue that these firms, given their size and reputation, do not need to signal their corporate governance superiority via XBRL adoption — the corollary also holds true that, on the margin, such adoptions may actually bolster their reputation for corporate transparency in the capital markets. Hence, our primary motivation is to determine whether early and voluntary filers of financial statements in the XBRL format signal superior corporate governance and operating performance vis-a-vis their non-voluntarily filing peers. The rest of this paper is structured as follows: Section 2 contains our literature review. We develop our hypotheses in section 3; Section 4 describes the research methodology, including the statistical analyses performed on the data; Section 5 reports the results of the various statistical analyses and Section 6 presents the conclusion, limitations and areas for future research. 2. Literature review 2.1. XBRL XBRL – eXtensible Business Reporting Language – is the product of XBRL International (www.xbrl.org) — a non-profit consortium of over 4506 global financial service, technology, stock exchange, government, and accounting organizations. XBRL is part of a family of XML languages becoming standard means for businesses to communicate over the Internet. A key attribute of XBRL is an explicit semantic and machine-readable format for the wide variety of information found in business reporting in general and in financial statements in particular (Debreceny and Gray 2001). XBRL tags are context-sensitive and their use in XBRL instance documents allows for the specific identification and retrieval of individual items of data from reports such as financial statements. XBRL “tags” firm data and “permits the automatic exchange and reliable extraction of financial information across all software formats and technologies, including the Internet” (Xbrl.org White Paper, 2002, 5). XBRL “tagged” data also include critical non-financial information which is typically difficult to capture, aggregate and disseminate using current incompatible information systems (Pinsker, 2004). The business case for adopting XBRL is also strong. Benefits to business include the ability to use commercial off-the-shelf software; the automated exchange of data within the firm (such as from parent to subsidiary; or within the firm's supply chain); inter-company exchanges; and to ease the transfer of information between a company and its regulators (for example for the transmission of FDIC-required banking reports). 2.2. Benefits of XBRL According to Weber (2003), XBRL reduces the costs associated with obtaining and analyzing information from businesses by addressing and eliminating incompatible reporting formats. Hodge et al. (2004), for instance, concluded that using XBRL helps nonprofessional financial statement users acquire and integrate related financial statement and footnote information when 6 As of May 31, 2009. 4 R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 making investment decisions. XBRL also allows regulators to further the standardization and harmonization of international business reporting standards. For instance, as of January 2005, European stock exchanges required all registrants to prepare consolidated financial statements in accordance with International Financial Reporting Standards (IFRS). The use of IFRS is also expected to improve the transparency, comparability, and quality of financial reporting across the European Union (EU) and lead to a deepening of the EU's capital markets. Japanese firms are also embracing IFRS; however, Japanese firms must also reconcile differences in local financial statements to U.S. Generally Accepted Accounting Principles (GAAP) in filings with the SEC. Several South American countries are also either in the process of or have already adopted International Accounting Standards and/or IFRS to take advantage of the European capital markets (Catacora and Hannon 2005). The adoption and use of XBRL is expected to help avoid the extra effort and complications associated with such multiple reconciliations (Duangploy and Gray, 2005; Scannell, 2006b).7 XBRL International also claims that firms experience a reduction in their cost of capital due to improved, transparent and real-time financial reporting (via the Internet) and disclosure of data in the XBRL format (Xbrl.org White Paper, 2002). Based upon this, capital markets should reward firms that are early and voluntary filers of key financial statement information in the XBRL format. Prior studies8 (Botosan 1997; Trabelsi et al., 2004; Gibbons et al., 1990) suggest that firms seeking specific advantages in the capital markets would be expected to voluntarily opt for such XBRL tagged financial reporting disclosure (AAA, 2004). Furthermore, in a more recent context, the addition of the XBRL infrastructure by firms is expected to reinforce and make it easier to implement and comply with the various provisions (especially the relatively vague “disclosure” provisions of Section 401, Disclosures in Periodic Reports) of the Sarbanes-Oxley Act of 2002 (SOX). 2.3. Recent actions of various global regulatory agencies related to XBRL In the United States, the SEC has already established a voluntary program for registrants to furnish XBRL data in specific EDGAR filings under the Securities Exchange Act of 1934 and the Investment Company Act of 1940. Some U.S. public firms have already commenced reporting of basic financial statement information in XBRL format, and an increase in the number of firms filing their financial statements in XBRL format is expected over the next few years. The Financial Executives International (FEI, 2006) lists XBRL among its top ten (number 6) financial reporting challenges for 2009. The FDIC's XBRL experimental filing program for U.S. banks also started in October 2005. Internationally, there are several XBRL pilot projects in process. These include, but are not limited to, those of the UK Inland Revenue Authority (HM Revenues & Customs), the UK Almost 100 countries as of the date of this manuscript now require or allow the use of IFRS for the preparation of financial statements by listed companies, and other countries are moving to do the same. In addition, the SEC has issued a Concept Release (33-8831) dated August 7, 2009 evaluating the possibility of allowing U.S. issuers to prepare and file financial statements in accordance with IFRS, and issued a Proposed Rule (33-8818) dated July 2, 2009 accepting from foreign firms financial statements prepared in accordance with IFRS without a reconciliation to U.S. GAAP. 8Botosan (1997) studied the disclosure level of firms and their cost of capital and concluded that an increase by a firm in disclosure level reduces its cost of equity capital. Similarly, Trabelsi et al. (2004) adapted the Internet framework of Gibbons et al. (1990) to show that firms manage both their disclosure content and secondary features (including medium, data organization, timing and usability issues) in their financial reports. 7 R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 5 Companies House (filing service for the financial statements of small companies), EDGAR Online, the China Securities Regulatory Commission (CSRC), KOSDAQ (the Korean Stock Exchange), the Tokyo Stock Exchange, and the National Tax Agency of Japan. To date, members of the XBRL International include the following stock exchanges: Australia, China, Germany, Korea, London, New Zealand, and Tokyo.9 Sixteen countries have established their own XBRL organizations, including most recently, France. The Dutch government has launched the Dutch Taxonomy project, which is designed to reduce the administrative burden of Dutch business reporting to the government. The Banco de Espana has stated its desire to use XBRL as a monitoring vehicle to stop money laundering activities of local Spanish firms (CPA Technology Advisor, 2005). Lastly, all of the Big Four global accounting firms are also members of XBRL International. The actions of various global regulatory and other agencies mentioned above appear to signal their support for and involvement in XBRL adoption and may be viewed as part of a global effort to link improved corporate governance and transparency through the adoption of XBRL for financial reporting. Members of XBRL International also get direct assistance from XBRL International, which mitigates the start-up costs of reporting in XBRL format.10 2.4. XBRL and voluntary disclosure There are a number of reasons why companies may decide to voluntarily file their financial statements in XBRL format when such filing is not yet an SEC requirement. For example, the voluntary disclosure literature suggests that the motivation for firm voluntary disclosures include signaling good or improving firm performance (Verrecchia 1983, 2001; Dye 1995). Consistent with this, Lang and Lundholm (1993) and Miller (2002) find that better firm performance is associated with higher overall levels of disclosure. In the context of capital markets, the primary motivation for filing in XBRL may be related to lowering the firm's cost of capital, as the additional information reduces investor uncertainty about the quality of the firm and the expected returns from its securities (Gray et al., 1995). Prior research shows the firm disclosure practices may be related to the amount of leverage, especially present levels of debt (Jensen and Meckling 1976). One of the objectives of XBRL is the reduction of information asymmetry resulting from incompatible global reporting formats. In other words, XBRL theoretically should “level” the disclosure playing field, allowing any type of investor to evaluate financial statement information across a large number and differing sizes of firms. We can therefore posit that, among other goals, firms that voluntarily file financial information in XBRL format at this early stage do so in order to signal the increased transparency of their financial information. This, in turn, allows us to examine whether there is a relationship between the corporate governance ratings, firm performance, and a firm's decision to be early and voluntary filer of information in the XBRL format. China began with 50 companies voluntarily using XBRL in 2003; the program has grown in 2009 to more than 800 companies. In 2009, China became the first country to mandate XBRL reporting, and is now requiring interactive data filing for the full financial statements for all listed companies in quarterly, half-year, and annual reports under rules of both the CSRC and the Shanghai Stock Exchange. Japan has mandated public company reporting using XBRL for full financial statements of all listed companies beginning with quarterly reporting in the second quarter of 2010. And in Korea, a voluntary XBRL filing program was established in 2006, and beginning in October 2009, all publicly held companies were required to file financial statements using XBRL on the electronic filing system of the Korean Financial Supervisory Service (www.sec.gov/news/press/2009/2009-227.htm). 10The costs of rearranging internal and external reporting guidelines and reporting may not be a trivial matter, as many firms have developed tailored management reports over a long period of time and at high cost. 9 6 R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 3. Hypotheses development As one of the objectives of XBRL is the reduction in information asymmetry resulting from incompatible global reporting formats, we hypothesize that early filers of financial information in XBRL format already evidence strong qualities (compared to their peers) in financial statement transparency and related disclosure. They are therefore considered to demonstrate stronger corporate governance characteristics vis-a-vis their non-voluntarily filing peers. Following Lang and Lundholm (1993), Wallace et al. (1994), Camfferman and Cooke (2002), and Alsaeed (2005), we consider firm characteristics as proxies for the degree of variation related to firm voluntary disclosure to form our independent variables and classify them into three categories: (1) performance-related variables (liquidity, net profit margin and return on equity); (2) marketrelated variables (auditor type and firm type); and (3) structure-related variables (leverage ratio and firm size). We use a corporate governance variable for each sample firm — initially as an independent variable in our logistic regression model and then as a dependent variable in our multiple regression model. 3.1. Performance-related variables 3.1.1. Liquidity Liquidity can be defined as the ability of a firm to pay its short-term liabilities. Prior literature offers many measures which proxy for liquidity as well as conflicting evidence regarding the relationship between liquidity and the extent of firm disclosure. For instance, Belkaoui and Kahl (1978) found no relationship between liquidity and firm disclosure; Wallace et al. (1994) found a significantly negative relationship and Camfferman and Cooke (2002) found significantly positive relationship between the liquidity of Dutch firms and the extent of disclosure and a negative relationship for U.K. firms. Cooke (1989), however, concluded that the more liquid the financial condition of the firm, the greater the incentive for it to disclose and signal its strength to the market. We believe that early and voluntary filers in XBRL format are likely to be more liquid than their non-voluntarily filing peers, and use this membership as one of the ways to signal their strong financial condition and their ability to meet current obligations to short-term lenders or suppliers. Therefore, we posit that: H1: Firms with higher liquidity are more likely to be early and voluntary filers of financial information in XBRL format. Operationalization: Current Ratio is defined as current assets divided by current liabilities. 3.1.2. Net profit margin and return on equity Singhvi and Desai (1971) found that firms with higher profitability supply additional information to both placate investors and increase management compensation. However, Lang and Lundholm (1993) found that the relationship between the extent of firm disclosure and profitability performance is unclear. In fact, Camfferman and Cooke (2002) found a significant negative relationship between the net profit margin ratio and the extent of disclosure by U.K. firms. Section 409 of SOX – Real Time Issuer Disclosures (which became effective on August 23, 2004) – dramatically expanded the types of events requiring real-time disclosure (on Form 8- K) for publicly-traded firms. Therefore, it may be argued that larger firms with high levels of profitability have greater exposure to public scrutiny and may respond to such scrutiny with increased measures of financial disclosure. Early and voluntary filers of financial information in XBRL format with strong profitability ratios will therefore have greater incentive to disclose their R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 7 financial condition to the marketplace and to signal this condition to existing and potential shareholders than peers that are less profitable.11 Therefore, we posit that: H2: Firms with higher net profit margin are more likely to be early and voluntary filers of financial information in XBRL format; and H3: Firms with higher returns on equity are more likely to be early and voluntary filers of financial information in XBRL format. Operationalizations: Net Profit Margin is defined as net income before extraordinary items divided by net sales. Return on Equity is defined as net income before extraordinary items divided by the previous year's (the last audited financial year-end) book value of equity. 3.2. Market-related variable 3.2.1. Auditor type The world's auditing firms today are divided into the Big 4 and the non-Big 4. The Big 4 operate globally while the non-Big 4 operate primarily in their countries of origin. Prior research on the relationship between audit type and disclosure standards of firms is unclear. For instance, Camfferman and Cooke (2002) found a significantly positive relationship between disclosures of U.K. firms and audit type. On the other hand, Wallace and Naser (1995) found a significantly negative relationship between audit type and disclosure levels among Hong Kong firms. Wallace et al. (1994) also found an association between the content of annual reports and the size of the audit firm in a study of Spanish public companies. It is important to note that all of the Big 4 audit firms are presently members of XBRL International. It is therefore realistic to expect that these firms encourage their clients to report in the XBRL format in order to ensure compliance with international regulations regarding financial reporting — especially regulations related to the timely reporting of material events. Furthermore, PricewaterhouseCoopers (PwC, hereafter), one of the Big 4 audit firms, is one of the major proponents of the XBRL initiative; thus, PwC in particular would be expected to push for the early adoption of XBRL amongst its global audit clients. The remaining members of the Big 4 audit firms also would be expected to follow suit — particularly in light of both their own membership in XBRL International and in light of PwC's efforts with its clients. Furthermore, since Big 4 audit clients tend to be larger global public firms they may also be expected to lean toward increased disclosures to fulfill their own public responsibilities. Finally, US public firms also have the added pressure of compliance with SOX and may determine that reporting in the XBRL format (especially with SEC encouragement) helps them achieve overall compliance with SOX. Therefore, given these XBRL-favorable conditions, it is realistic to expect that firms engaging Big 4 audit firms are more likely to be early and voluntary filers of financial information in XBRL format. In keeping with these expectations, we posit that: H4: Firms engaging Big 4 audit firms are more likely to be early and voluntary filers of financial information in XBRL format. Murphy and Zimmerman (1993) advanced the cover-up hypothesis which states that managers' discretionary choices conceal poor performance by their superiors and themselves; however, looking at the profile of the majority of the early firm filers in Appendix A, the cover-up hypothesis does not appear to apply to early and voluntary filers of information in XBRL format. 11 8 R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 Operationalization: Big 4: 1 if the Company has a Big 4 audit firm as its auditor, 0 otherwise. 3.3. Structure-related variables Consistent with prior studies such as Lang and Lundholm (1993), Wallace et al. (1994), Camfferman and Cooke (2002), and Alsaeed (2005), we include firm leverage ratio and firm size as structure-related variables. 3.3.1. Leverage ratio If a firm raises debt in the securities market, it has more requirements to satisfy the needs of related creditors than firms with little or no public debt. Jensen and Meckling (1976) suggest that firms with high debt can alleviate the associated monitoring costs for creditors by achieving more comprehensive levels of disclosure. This view is supported in studies by Belkaoui and Kahl (1978) and Malone et al. (1993). Highly leveraged firms may tend to disclose more financial information to reassure their creditors regarding their future debt requirements than less leveraged firms, especially if the debt is raised in public security markets. Agency theory also suggests that highly leveraged firms disclose more information voluntarily in order to satisfy the needs of their creditors. Such firms' debt holders may also demand collateralized arrangements such as sinking fund requirements, minimum deposit requirements, and other types of reassurances. Therefore, highly leveraged firms are more likely to be early and voluntary filers of financial information in XBRL format in order to reduce monitoring costs, to satisfy the needs of creditors, and to signal to the public security markets their confidence regarding their abilities to service their debt and covenant requirements. H5: Highly leveraged firms are more likely to be early and voluntary filers of financial information in XBRL format. Operationalization: Leverage is defined as the ratio of long-term debt to total stockholders' equity. 3.3.2. Firm size Many past studies such as Cooke (1989) and Botosan (1997) have found a positive and significant relationship between firm size and firm disclosures. Camfferman and Cooke (2002, 10) write, “size has been found to be a very significant variable in most studies with a positive association between size and the extent of disclosure.” Larger firms may tend to disclose more financial information for many reasons, including: (1) revealing more information may positively impact a firm's future cost of obtaining new funds (Botosan 1997); (2) larger firms are in the public spotlight more than smaller firms, and they respond to this pressure by increasing disclosures; and (3) larger firms have more financial resources than smaller firms to prepare, collect, analyze, and present financial data in new formats (like XBRL) than smaller firms. Hence, we posit that: H6: Larger firms are more likely to be early and voluntary filers of financial information in XBRL format. Operationalization: Total Assets: firm size is defined as total assets. 3.4. Corporate governance measures Many types of summary measures are available in the literature as proxy measures for a firm's corporate governance. Brown and Caylor (2006a, b), using a unique dataset provided by Institutional Shareholder Services (ISS), created Gov_Score, a broad summary measure of both internal and external firm corporate governance, and found that Gov_Score was related to both R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 9 firm operating performance and firm valuation.12 We used Gov_Score as the main proxy for corporate governance in this study, as Gov_score reflects recent changes in the corporate governance environment since SOX.13 Using Standard and Poor's 2002 Transparency and Disclosure ratings as a proxy for corporate governance, Premuroso and Bhattacharya (2009) found, using financial performance and structure-related variables, an association between firms' corporate governance ratings and firms' decisions to form a technology committee.14 We use Gov_score as our proxy for corporate governance due to the close proximity of the preparation of these governance ratings to the audited annual financial statement dates that our subject firms file their financial information in XBRL format. 4. Research methodology and statistical design 4.1. Sample selection and data collection A list of all voluntary filers of their latest annual audited financial information in XBRL interactive format as of May 31, 2009 was obtained from the U.S. Securities and Exchange Commission Interactive Financial Report Viewer website.15 Of the 30 firms voluntarily filing financial information with the SEC at such date, six were foreign-domiciled firms, two U.S. firms did not have classified balance sheets, one U.S. firm was not a separately publicly-traded stock, and one U.S. firm was from the banking industry.16 Financial data required for the variables used in the statistical analyses for the 20 remaining firms was obtained using Compustat or obtained directly from the firm's latest audited (in most cases, the calendar year ended 2006) annual report on Form 10-K. A matched pair (matched by 4-digit SIC code and latest annual revenues closest to December 31, 2006) for each of the 20 firms that were early and voluntary filers in XBRL format was obtained using Compustat. 4.2. Statistical analyses 4.2.1. Binary logistic regression Binary logistic regression is an alternative to traditional regression analysis and simultaneously predicts the probability of an event (Hair et al., 1998, Mertler and Vannatta 2005). Logistic Gov_Score is freely available for download at the following website: http://www.robinson.gsu.edu/accountancy/ gov_score.html. The details of the 51 ISS corporate governance provisions included in the determination of Gov_Score can be found in Brown and Caylor, 2006a, Table 1. The provisions are coded 1 (acceptable) or 0 (unacceptable) for each firm depending upon whether or not ISS considers the firm's governance to be minimally acceptable in accordance with the ISS Corporate Governance Best Practices User Guide and Glossary. 13Gov_Score (calculated as of February 1, 2005, the latest date available) is used as a proxy for corporate governance for both XBRL filers and the matched pair firms; the higher the Gov_Score, the stronger the relative level of corporate governance. The governance ratings for each firm are constructed by adding the 51 governance factors classified in a binary manner (Brown and Caylor, 2006a). 14We do not use the S&P Transparency and Disclosure survey as a proxy for corporate governance in this manuscript, as this survey and governance rating has not been updated for firms since 2002, and the SEC (post-SOX), the stock exchanges and publicly-listed firms themselves have made wholesale changes to governance rules and practices since 2002. 15Available at: http://216.241.101.197/viewer. This is the SEC’s Interactive Financial Report Viewer website, where each annual report (10-K) financial statement filing by any firm in XBRL format is called an annual report; in the EDGAR archives where the XBRL financial statements are actually filed, many of the annual reports are filed later than the formal filing of the traditional Form 10-K on EDGAR, and as such, these filings in XBRL format are filed in EDGAR under a Form 8-K. 16See Appendix A for a list of the 20 U.S. firms that are early and voluntary filers of audited annual financial information in XBRL format for their latest financial year ending before and within twelve months of May 31, 2009. 12 10 R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 regression is used in studies where the dependent variable is not a quantitative or continuous variable (George and Mallery 2000) and tests the ability of a model or group of independent variables to predict group membership as defined by a categorical variable. Logistic regression also provides several distinct advantages over multiple regression because the independent variables do not have to be normally distributed, linearly related or have equal variances within each group. This makes logistic regression more flexible than other parametric techniques (Tabachnick and Fidell 1996). We used a binary logistic regression with a categorical dependent variable (1 if the firm was an early and voluntary filer and 0 for the matched pair firms that are not early and voluntary filers of financial statements in XBRL format). Our aim was to predict membership or non-membership in the XBRL filer/non-filer groups using our independent variables and the corporate governance ratings for each firm. In order to run our binary logistic regressions, we combined our 40 firms (the 20 early and voluntary filers and the 20 matched pair firms, matched by annual revenues and 4-digit SIC code) into one combined data set. The logistic regression results also helped us identify the independent variables that were potentially (on an exploratory basis) the best predictors of early and voluntary filers of financial statements in XBRL format. 4.2.2. Multiple regression analysis Following the logistic regression analysis, we used multiple regression analysis to answer the main research question: Is there an association between the corporate governance rating (the dependent variable in our multiple regression analysis) and operating performance (using the remaining independent variables which are firm characteristics and proxies for the degree of variation related to firm voluntary disclosure decisions) for the 20 firms that have voluntarily filed financial statements in XBRL format? Tests for compliance with fundamental underlying assumptions for multiple regression analyses were conducted in order to ensure that the data were conducive to such analysis. For instance, the relationships between the independent variables as well as the relationships between the dependent and independent variables were analyzed via correlation coefficients for every potential pair of variables in the study. Multicollinearity tests were developed using variance inflation factors (VIF) to test for the presence of multicollinearity in each of the independent variables. The results of the tests for multicollinearity depended on the numeric values of the VIFs for all independent variables. Pearson correlation analysis with a confidence level of alpha = 0.10 were used. In the presence of multicollinearity, we considered removing the independent variable with the least predictive ability in the analysis. It was important to address multicollinearity in order to ensure that each of the independent variables adequately and correctly explained its role in bringing about changes in the dependent variable. 5. Results 5.1. Descriptive statistics Table 1A shows the descriptive statistics for the 20 firms voluntarily filing in XBRL format and the 20 matched pair firms. Table 1B shows the comparison of the sample means (t-test) and medians (Wilcoxon Signed Rank Test) for each variable between the XBRL filers and the matched pair firms. There was a wide variation in some of the variables, as indicated by the minimum, maximum values and standard deviations in Table 1A. Total assets showed considerable dispersion within the two groups; consequently, this variable was converted to a natural log amount in all subsequent statistical analyses. R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 Table 1A Descriptive statistics: voluntary XBRL financial statement filers and matched pair firms Variable Voluntary XBRL filers N = 20 Minimum Gov_score Gompers G-index Current Ratio Net Profit Margin Return on Equity Big 4 Leverage Total Assets (in millions) 22.000 4.000 0.275 − 0.771 − 0.154 0.000 − 3.231 19.000 Maximum 43.000 12.000 4.444 0.400 1.807 1.000 0.799 114837.000 S.D. 4.487 2.049 1.047 0.230 0.391 0.308 0.854 37606.532 Matched pair firms N = 20 Minimum 19.000 6.000 0.556 0.005 −3.000 0.000 −1.000 9.000 Maximum 40.000 16.000 2.914 0.308 0.465 1.000 1.721 70556.000 S.D. 11 5.748 2.511 0.694 0.083 0.723 0.224 0.525 19275.330 A number of other trends also appear in the descriptive statistics for the 20 XBRL filing firms, including (in terms of mean values): - - - - relatively strong liquidity, as evidenced by a current ratio of 1.7 to 1; return on equity is also strong at almost 26%; the majority use Big 4 audit firms for their annual audits; and the firms are modestly leveraged at a debt-to-equity ratio of 17.7%. The mean of the Gov_score governance variable measure for the XBRL filing firms is 34.35, which is higher than the mean for the matched pair firms (32.10) and significantly different using a standard t-test (p-value = 0.097). A significant difference in both the means and medians between the XBRL filers and the matched pair firms is also found for the total assets variable (ttest p-value 0.019, and Wilcoxon-test p-value 0.044, respectively). Not surprisingly, larger firms with higher corporate governance ratings appear to be early and voluntary filers of their financial information in XBRL format — these firms, presumably, have the most to gain from being early and voluntary filers of their audited financial information in XBRL format. We concluded at this Table 1B Comparison of sample means and medians: XBRL filers and matched pair firms Variable XBRL filers N = 20 Mean Gov_score Gompers G-index Current Ratio Net Profit Margin Return on Equity Big 4 Leverage Total Assets (in millions) 34.350 9.100 1.732 0.070 0.257 0.900 0.176 36,155.150 Median 35.000 9.000 1.574 0.089 0.183 1.000 0.330 26,532.500 Matched pair firms N = 20 Mean 32.100 8.900 1.485 0.123 0.039 0.950 0.322 18,878.900 Median 33.000 9.000 1.317 0.103 0.202 1.000 0.265 16,659.500 T-test p-value 0.097 0.691 0.232 0.300 0.207 0.330 0.542 0.019 Wilcoxon-test p-value 0.111 0.723 0.117 0.709 0.588 0.317 0.811 0.044 p-values are two-tailed. Gov_score is a summary binary governance measure of 51 firm-specific provisions using data from ISS. Gompers G-index is a summary binary governance measure of 24 firm-specific provisions using data from IRCC. Current Ratio is current assets divided by current liabilities. Net Profit Margin is net income before extraordinary items divided by net sales. Return on Equity is net income before extraordinary items divided by last year's book value of equity. Big 4 is 1 if the firm's latest annual financial statements are audited by a member of the Big 4 auditing firms, 0 otherwise. Leverage is total long-term debt divided by total shareholders' equity. Total Assets is total consolidated assets. (All financial data is from the latest audited annual financial statements for each firm). 12 R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 stage that our matched pair methodology (matching firms that were early filers in the XBRL format with their non-adopting peers using firms in the same 4-digit SIC code and annual revenues) was sufficient for us to proceed to the next stage of our statistical analyses. 5.2. Logistic regression results Binary logistic regression was used to determine which of our variables significantly predict that a firm will file financial statements in XBRL format. After combining our 20 XBRL filers and the 20 matched pair firms into one data set (dependent variable = 1 if firm is an XBRL filer, 0 otherwise), a preliminary multiple regression was conducted to identify potential outliers.17 For outliers, Mahalanobis (1936) distance is evaluated as a chi-square statistic with degrees of freedom equal to the number of variables in the analysis (Tabachnick and Fidell 1996). The chisquare critical value (df = 7, p b .001) of 24.322 was then compared to the obtained value for the Mahalanobis' (1936) distances for each of the 40 firms.18 We checked for multicollinearity (using Pearson Correlation Coefficients) at this stage. Tolerance for all variables was greater than 0.1. According to Mertler and Vannatta (2005), if the tolerance statistic for any individual independent variable is less than 0.1, the regression should be repeated by omitting the violating independent variable. The variance inflation factors (VIFs) for each individual independent variable was greater than 1.5, averaged 3.0 for all of the variables, and none individually exceeded 6.0. While there are generally no universal rules about VIF values, Myers (1990) suggests that an individual VIF value of 10 is cause for concern. We concluded at this stage that our logistic regression model was adequate and robust to potential outliers or multicollinearity. We then performed the binary logistic regression model on the combined 40 firms – the 20 firms voluntarily reporting in XBRL format and the 20 peer non-adopting firms, with all of the independent variables entered simultaneously into the model. The results of the binary logistic regression model are shown in Table 2. The regression results indicate that the overall model, including two of the predictor variables (current ratio p-value = .031 and total assets p-value = .035) was statistically reliable in distinguishing between firms that are early and voluntary filers of financial statements in XBRL format and their non-adopting peers (− 2 Log Likelihood-index of model of fit = 40.490 and Chi-Square-goodness of fit (7) = 14.962, p = .036). The Cox and Snell R2 (.312) and Nagelkerke R2 (.416) measures (essentially two different estimates of R2, indicating the proportion of variability in the dependent variable that can be accounted for by all of the independent variables in the model) were also acceptable. The binary logistic regression model applied to the combined set of 40 firms classified 82.5% of the firms correctly, either as early and voluntary filers in XBRL format or as non-XBRL adopting firms. We then performed our binary logistic regression model using various methods of entry for the independent variables to further examine the variables most important in predicting early filers in We calculate the Mahalanobis (1936) distance to identify outliers. The Mahalanobis (1936) distance is defined as the distance of each individual firm's data case from the centroids of the remaining firm’s data cases, where the centroids is the point created by the means of all the variables (Tabachnick and Fidell 1996, Mertler and Vannatta 2005, p. 29). 18The accepted criterion for outliers is a value for Mahalanobis (1936) distance which is significant at p b .001 (Mertler and Vannatta, 2005). Two outliers were identified in this process (1 from each of the XBRL filer and non-adopting peer groups-each resulting from extreme negative amounts in at least two of the independent variables included in the regression model). All subsequent analyses were performed with and without these two outliers, and our overall results remained unchanged. Therefore, for parsimony, we present only statistical results in this section related to combined data set of the 40 firms. 17 R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 Table 2 Binary logistic regression results-using Gov_score as corporate governance proxy Variable Constant Gov_score Current Ratio Net Profit Margin Return on Equity Big 4 Leverage Total assets B − 26.721 0.018 1.757 − 9.075 0.894 8.858 − 0.927 1.046 Std. Error 11.090 0.099 0.816 7.032 2.393 6.105 1.294 0.496 Wald 5.806 0.033 4.636 1.666 0.139 2.105 0.514 4.449 df 1 1 1 1 1 1 1 1 13 p-value 0.016 0.856 0.031 0.197 0.709 0.147 0.474 0.035 Model: Prob (XBRL Filer = 1) = F(Z1 + Z2Gov_score + Z3Current Ratio + Z4Net Profit Margin + Z5Return on Equity + Z6Big 4 + Z7Leverage + Z8Total Assets). p-values are two-tailed. Gov_score is a summary binary governance measure of 51 firm-specific provisions using data from ISS. Current Ratio is current assets divided by current liabilities. Net Profit Margin is net income before extraordinary items divided by net sales. Return on Equity is net income before extraordinary items divided by last year's book value of equity. Big 4 is 1 if the firm's latest financial statements are audited by a member of the Big 4 auditing firms, 0 otherwise. Leverage is total long-term debt divided by total shareholders' equity. Total Assets is natural log of total assets. XBRL. Forward (LR, or logs ratio) logistic regression was utilized, whereby each of the independent variables (IVs) was entered one at a time and the likelihood ratio was used to fixate on the independent variable(s) required to predict voluntarily filers in XBRL. Other methods used included forward (Wald — whereby the IVs are entered one at a time, and the Wald statistic is used to determine variable selection); backward (LR — whereby all of the IVs are entered at once, then the IVs are removed from the model one at a time and the Likelihood-Ratio is used to determine variable removal); and backward (Wald — all of the IVs are entered at once, then the IVs are removed one at a time and the Wald statistic is used to determine variable removal). An important overall outcome of these binary logistic regression methods was that removing Gov_score reduced the overall effectiveness of the models significantly (p b .002). Removing any of the other seven independent variables (individually) did not have the same significant impact on the model. This suggested that the Gov_score was driving the classification (and perhaps the decision process) of XBRL firms. We interpret this to suggest that firms that are early filers in XBRL signal their confidence in the quality of their corporate governance standings relative to their non-adopting peers. While our results support our hypotheses, it is important to note that the population of filers of annual audited financial information in XBRL format as of May 31, 2009 is limited to just 20 firms. This makes it difficult to draw any definitive conclusions from the logistic regression results alone. However, in order to further investigate the potential impact of the Gov_score rating variable on a firm's decision to voluntarily file financial statements in XBRL format, we developed a model with Gov_score (the corporate governance rating for the 20 early and voluntary filers in XBRL format) as the dependent variable in a multiple regression model in order to formally determine the Table 3A Regression model summary: Gov_score as dependent variable R 0.713 R2 0.508 Adjusted R2 0.333 Std. Error of the estimate 3.665 Predictors: (Constant), Current Ratio, Net Profit Margin, Big 4, Leverage, Total Assets. 14 R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 Table 3B Regression ANOVA summary table: Gov_score as dependent variable Model 1 Regression Residual Total Sum of Squares 194.511 188.039 382.550 df 5 14 19 Mean Square 38.902 13.431 F 2.896 Sig. 0.053 Predictors: (Constant), Current Ratio, Net Profit Margin, Big 4, Leverage, Total Assets. potential relationship between the independent variables that proxy for both firm performance and control variables, and the XBRL firms' corporate governance ratings. 5.3. Multiple regression results We first tested the data for outliers (again via Mahalanobis' (1936) distance in a preliminary regression procedure for all variables in our model) and found no outliers. Our tests for the presence of multicollinearity, using Pearson Correlation Coefficients (PCC), found multicollinearity concerns between the variables “leverage” and “return on equity” (coefficient = −.866, p = .000), and similar statistical multicollinearity concerns for these same variables individually in both tolerance (both less than .10) and VIF (greater than 10) factors.19 The remaining PCCs were insignificant or less than 0.4. Because of the aforementioned multicollinearity concerns, we decided to run our multiple regression models in two different ways — one excluding only the leverage variable, and again excluding only the return on equity variable. For parsimony, we present the results of the model excluding return on equity below (the results are essentially the same for the model excluding leverage compared to the model excluding return on equity). Tables 3A, 3B and 3C show the results of our multiple regression models. Table 3A shows the multiple regression model summary; the five independent variables (entered all at one time into the model) explain (adjusted R-square) 33.3% of the variation in the dependant variable, Gov_score for the 20 firms that are early and voluntary filers of financial information in XBRL format. Table 3B shows the ANOVA table — the F-ratio (F = 2.896) and the level of significance (.053) indicate that the multiple regression significantly predicts firm's that file voluntarily in XBRL format. Overall, Table 3C shows that the regression results are mixed regarding the association of the various variables with the Gov_score rating for XBRL adopting firms. Discussions of the multiple regression coefficient results follow. 5.4. Current ratio and net profit margin The results related to these performance-related variables differed. The coefficient for current ratio was positive and insignificantly (p = .480) related to Gov_score. This insignificant result is somewhat surprising, as past studies such as Alsaeed (2005), Naser and Nusuerbeh (2003) and Wallace et al. (1994) have found that larger firms with strong current ratios are more closely watched by investors, suggesting that larger firms would disclose more financial information than smaller firms. However, Camfferman and Cooke (2002) also found a negative and insignificant We are unable to transform the variables Leverage or Return on Equity using common transformation techniques, as both variables are ratio’s; for example, a square root transformation of the variables was not possible, as some of the 20 firms reported negative leverage or negative return on equity ratios, which were not transformable. 19 R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 Table 3C Regression model results: Gov_score as dependent variable Variable Unstandardized coefficients B Constant Current ratio Net profit margin Big 4 Leverage Total Assets (in millions) − 5.340 1.011 − 6.632 − 6.263 1.919 1.902 Std. Error 17.610 1.393 8.953 5.107 1.806 0.844 Standardized coefficients Beta 0.236 − 0.340 − 0.430 0.365 1.090 t −0.303 0.726 −0.741 −1.226 1.063 2.253 15 p-value 0.766 0.480 0.471 0.240 0.306 0.041 Model: Gov_score = Z1 + Z2Current Ratio + Z3Net Profit Margin + Z4Big 4 + Z5Leverage + Z6Total Assets. p-values are two-tailed. Gov_score is a summary binary governance measure of 51 firm-specific provisions using data from ISS. Current Ratio is current assets divided by current liabilities. Net Profit Margin is net income before extraordinary items divided by net sales. Big 4 is 1 if the firm's latest financial statements are audited by a member of the Big 4 auditing firms, 0 otherwise. Leverage is total long-term debt divided by total shareholders' equity. Total Assets is natural log of total assets. relationship between the extent of disclosure and current ratio among U.K. firms. It is important to note that in our binary logistic regression analysis, the current ratio was a statistically significant factor in predicting which firms will voluntarily report in XBRL format when compared (and included in a matched pair model) with non-adopting peers. The coefficient was negative and insignificant for net profit margin (p = .480) in both the logistic and multiple regression models, similar to the findings of Alsaeed (2005) and Wallace et al. (1994) for this variable. 5.5. Auditor type The Big 4 audit firm variable was found to be insignificant and negative (p = .240) in the multiple regression model. This finding was also somewhat surprising, as Camfferman and Cooke (2002) found a significantly positive association between level of disclosure and the use of Big 6 auditors among U.K. firms. However, this negative and insignificant coefficient was contradicted in the logistic regression analysis, where the Big 4 variable was positive and marginally insignificant (p = .147). 5.6. Leverage and size Leverage was found to be insignificant and positive in the regression — contrary to the findings of Camfferman and Cooke (2002) for UK firms, which found a negative coefficient, related to disclosure. The finding that size, measured by total assets, is positively and significantly associated with corporate governance (p = .041) is consistent with the research of Cooke (1989) and Lang and Lundholm (1993). 5.7. Robustness tests We use the Gompers G-index (Gompers et al., 2003, first constructed in the early 1990s)20 as a separate proxy for corporate governance and a robustness test for the firms included in our The Gompers G-index for each firm is determined as of 2006, which is very close to the annual audited financial statement reporting date for most of the firm’s included in our statistical analyses. The Gompers G-index has been obtained from the IRRC dataset in the Wharton Research Data Services (WRDS) database. 20 16 R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 Table 4 Binary logistic regression results-using Gompers G-index as the corporate governance proxy Variable Constant G-index Current Ratio Net Profit Margin Return on Equity Big 4 Leverage Total Assets B − 30.304 0.175 1.833 − 9.594 0.257 26.085 − 1.078 1.165 Std. Error 11.693 0.173 0.829 7.105 4.249 27870.749 1.331 0.465 Wald 6.716 1.027 4.894 1.823 0.004 0.000 0.656 6.292 df 1 1 1 1 1 1 1 1 p-value 0.010 0.311 0.027 0.177 0.952 0.999 0.418 0.012 Model: Prob (XBRL Filer = 1) = F(Z1 + Z2G-index + Z3Current Ratio + Z4Net Profit Margin + Z5Return on Equity + Z6Big 4 + Z7Leverage + Z8Total Assets). p-values are two-tailed. Gompers G-index is a summary binary governance measure of 24 firm-specific provisions using data from IRCC. Current Ratio is current assets divided by current liabilities. Net Profit Margin is net income before extraordinary items divided by net sales. Return on Equity is net income before extraordinary items divided by last year's book value of equity. Big 4 is 1 if the firm's latest financial statements are audited by a member of the Big 4 auditing firms, 0 otherwise. Leverage is total long-term debt divided by total shareholders' equity. Total Assets is natural log of total assets. statistical analyses. The Gompers G-index uses 24 distinct corporate-governance provisions derived from publications of the IRRC and constructs a governance index as a proxy for the balance of power between shareholders and managers; for every firm, Gompers et al. (2003) add one point for every provision that reduces shareholder rights. The higher the Gompers G-index for a firm, the weaker the related shareholder rights, and therefore the weaker the corporate governance rating of the firm The Gompers G-index has become the de facto measure of corporate governance used in most academic research papers over the past decade. We first conducted our robustness test by using the individual Gompers G-index governance scores for the 40 firms as an independent variable in the binary logistic regression analysis, replacing Gov_score as the proxy for corporate governance.21 The binary logistic regression results in Table 4 were similar to Gov_score in Table 2 using the Gompers G-index as a proxy for corporate governance (constant p-value = 0.010, current ratio pvalue = .027, and total assets p-value = .012). We also found this model statistically reliable in predicting the early filers in XBRL format (− 2 Log Likelihood-index of model of fit = 52.679, Chi-Square-goodness of fit (7) = 13.517, p = .060). The Cox and Snell R2 (.299) and Nagelkerke R2 (.399) measures were also acceptable. The binary logistic regression model using G-index classified 84.2% of the firms correctly either as early and voluntary filers in XBRL format or nonadopting firms. Next, we use Gompers G-index as a replacement proxy for corporate governance for Gov_score for the 20 firms in the multiple regression analysis (results not shown for parsimony). Again, as in Table 3C, we find using the same five independent variables that the results using the Gompers Gindex were similar to those shown in Tables 3A, 3B, and 3C, including the coefficients, direction of the coefficients, and statistical significance of both the model and the variables themselves. As was previously mentioned, the number of firms filing voluntarily in XBRL format at the date of the preparation of this manuscript is 20 firms; our matched pair design is therefore also limited to 20 firms. To further test our multiple regression model, we calculate the industry median (by four- We first performed the same tests and checks for outliers and multicollinearity in this model using G-index; our revised model was sufficiently robust to outliers or threats related to multicollinearity. 21 R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 17 digit SIC code22 for calendar year 2006 using Compustat data) for each financial-statement based ratio variable (current ratio, net profit margin ratio, return on equity ratio, and the leverage ratio). We then convert the financial statement ratio variables for the 20 XBRL filing firms in our model to dichotomous variables (1 if the XBRL filing firm's financial ratio exceeds the four-digit SIC code median ratio for calendar 2006, 0 otherwise), and rerun the regression model. Our results (not shown in the interest of parsimony) are again similar to the results described above. 6. Conclusions, limitations and future research In this paper we investigate the association between firms' decisions to be early and voluntary filers of financial information in XBRL format and the corporate governance and operating performance characteristics of such early filers. We use a number of firm-specific financial variables and conduct statistical analyses to examine the corporate governance implications for firms making voluntary decisions to file their financial statements in the XBRL format. Our investigation was motivated by the basic question of why certain firms voluntarily and proactively adopt a technology that promises to bring greater transparency to their financial disclosures. Is it purely coincidence that some of the clientele of the largest accounting firms internationally are early and voluntary filers in XBRL format? Does the early adoption of a technology that yields greater transparency and thus signals greater affinity for corporate governance help firms differentiate themselves from their peers? Our findings suggest that superior corporate governance is indeed associated with a firm's decision to be an early filer of financial information in XBRL format. We also found that firm performance factors including liquidity and firm size are also associated with the early and voluntary XBRL filing decision. Our study and its findings echo a number of issues raised by Debreceny et al. (2005). For instance, Debreceny et al. (2005) call for research on the adoption of information systems innovations and their role in serving as catalysts in bringing together disparate parties in an information supply chain. We find that the adoption by firms of information systems innovations, namely the adoption of XBRL, appears to have a signaling agenda for such adopting firms. Debreceny et al. (2005) also call for an investigation of the economic justification for XBRL adoptions. The results of our study suggest that early and voluntary XBRL adoptions are indeed indicators of superior corporate transparency and related corporate governance that are expected to benefit adopting firms in the long run. Debreceny et al. (2005) also call for research examining whether XBRL adopting firms realize net positive benefits from such adoption. The results of our study suggest that they do. The AAA Working Party also seeks an exploration of XBRL representations of more qualitative variables of corporate transparency such as those that occur in MD&A disclosures. The capital markets' accelerating drive to incorporate such qualitative factors can also be seen in the growing trend among SEC registrants to form the likes of compliance committees (Lublin, 2006) along with the SEC Chairman Mr. Christopher Cox's self-professed interest in effecting greater transparency in the area of executive compensation (Scannell, 2006a). Our research into XBRL and its capabilities lead us to believe that XBRL's semantic clarity and the ability to form unique instance documents make it a natural tool for use by corporate boards and sub-committees on issues such as compliance and the ethical use of information technology (AAA, 2004). Furthermore, the SEC's requirement that proxy statements carry a new column detailing the top The composition of our 20 early and voluntary filers of audited annual financial information in XBRL format is as follows by 4-digit SIC code: 0-1999 = 1 firm; 2000-2999 = 8 firms; 3000-3999 = 4 firms; 4000-4999 = 2 firms; 7000- 7999 = 4 firms; and from 8000-8999 = 1 firm. 22 18 R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 five executives' compensation packages and carrying it over to the compensation for members of the boards of directors can also be effected through the use of XBRL instance documents. These efforts will, in fact, be consistent with the user-oriented frameworks of Penman (2003), Schipper and Vincent (2003) and the SEC's own goal-oriented principles as laid out in Wallman (1997). The current stage of development of XBRL also offers researchers significant opportunities as XBRL International grows in size, stature, and more firms globally start to report their financial results in the XBRL format. Potential research issues related to XBRL include: (1) investigating the impact on US and foreign firms' cost of capital (compared to peer non-members) as a result of the decision to file financial reports in the XBRL format. As reported by Debreceny et al. (2005), there are no empirical studies to date that have examined the effect of XBRL on corporate cost of capital issues. This should be of particular interest for emerging economies like China, where reportedly 600 firms have already filed in the XBRL format with the Shanghai and Shenzhen stock exchanges; (2) the effect of XBRL adoption on firms' internal reporting processes. Debreceny et al. (2005) propose that XBRL-based internal reporting may provide a consistent and relatively smooth bridge for firms to undertake XBRL-tagged external reporting; (3) examining the linkage between credit rating services or the changes in firm corporate governance ratings with the decision and timing of those decisions by firms to voluntarily file financial statements in XBRL format; (4) the incorporation of other potential independent variables that may help predict early adoption of and reporting in XBRL format; (5) extending the current study by country, especially to countries like Japan and Ireland that are openly advocating early XBRL adoption by local firms, and (6) extending the study to private members of XBRL to determine which independent variables influence private firms to report early in the XBRL format. As is true of early and exploratory studies, the limitations of our study include the very small sample size of firms filing voluntarily in XBRL format. We strongly urge a replication of our study in the not-too-distant future when a larger number of publicly listed firms file financial statements in XBRL format. However, it is important to note that a replication based on the current research design would require such replication to be performed at a time when the number of firms voluntarily filing in XBRL format is significantly larger than that tapped here but before such time that XBRL reporting becomes mandatory for public firms in the U.S. and/or elsewhere. Of course, this caveat does not imply that future research cannot or must not use other research designs. As may be noted in our own attempts at triangulation, we believe the issue of linkage between voluntary reporting in XBRL format and corporate governance to be one that will only benefit from the employment of other research designs and time frames. This will allow for the examination of whether the preliminary hints of linkage found here actually grow stronger or are lost as a result of other corporate exigencies. Another limitation is our use of the Gov_Score governance ratings as a measure of corporate governance. Prior studies including Spero (1979) have provided overwhelming evidence supporting the premise that there are no significant differences between weighted and unweighted disclosure indices. Nevertheless, it is possible that our use of the Gov_score governance ratings may not have captured the depth and importance of corporate governance ratings adequately. Lastly, some of the commonly used firm-specific independent variables (such as the leverage ratio and current ratio) may not be the best measures of the underlying metrics (for example, determining the exact ratio by firm of interest-bearing debt to assets or using the quick ratio may have changed some of our results). Finally, another limitation of our study is endogeneity, a problem that plagues all corporate governance related studies (Brown and Caylor, 2006a, Larker and Richardson 2004). Despite the limitations noted above, our study suggests that firms voluntarily filing their financial statements in XBRL format potentially harbor greater corporate governance and R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 19 transparency aspirations and view such membership as signaling mechanisms for conveying their governance-related superiority to stakeholders. These early filers appear to view the XBRL reporting technology as yet another means of differentiating themselves from their non-XBRL filing peers. In doing so, they hope to signal their superior corporate governance and financial reporting transparency underpinnings — traits expected to make them potentially larger beneficiaries of an increasingly discriminating global capital market. Appendix A. Early and voluntary filers of annual audited financial Information in XBRL format: May 31, 2009 3M Company Anadarko Petroleum Corporation Automatic Data Processing, Inc. Altria Group, Inc. Bowne & Co. Inc. Bristol-Myers Squibb Company Comcast Corporation The Dow Chemical Company EDGAR Online Inc. Microsoft Corporation PepsiCo, Inc. Pfizer, Inc. R.R. Donnelley & Sons Company Radyne Corporation Adobe Corporation Englobal Corporation Lockheed Martin Corporation United Technologies Corporation Xerox Corporation XM Satellite Radio, Inc. References American Accounting Association (AAA). Comments on SEC Proposed Rule-33-8496. XBRL Voluntary Financial Reporting Program on the Edgar System; 2004. [Prepared by the Information Systems and Artificial Intelligence/ Emerging Technologies Sections of the American Accounting Association]; [October 31]. Alsaeed K. The association between firm-specific characteristics and disclosure: the case of Saudi Arabia. J Am Acad Bus 2005;7(1):310–21. Baldwin AA, Brown CE, Trinkle BS. XBRL: an impacts framework and research challenge. J Emerg Technol Account 2006:97–116. Belkaoui A, Kahl A. Corporate financial disclosure in Canada. Research Monograph of the Canadian Certified General Accountants Association, Vancouver; 1978. Botosan CA. Disclosure level and the cost of equity capital. Account Rev 1997;20(3):421–39. Brown LD, Caylor ML. Corporate governance and firm operating performance. Rev Quant Financ Account 2006a. Brown LD, Caylor ML. Corporate governance and firm valuation. J Account Public Policy 2006b;25(4):409–34. Camfferman K, Cooke T. An analysis of disclosure in annual reports of U.K. and Dutch companies. J Int Account 2002;1:28. Catacora F, Hannon N. XBRL and IFRS in Latin America. Strateg Financ 2005:59–60 [February]. Cooke TE. Voluntary corporate disclosure by Swedish companies'. J Int Financ Manag Account 1989;2:113–24 [Summer]. CPA Technology Advisor. Software Vendors Unveil A Variety of XBRL Solutions, Vol. 15(4). 2005. p. 3. [August], [Anonymous author]. Debreceny R, Gray G. The production and use of semantically rich accounting reports on the Internet: XML and XBRL. Int J Account Inf Syst 2001;2(1):47–74. Debreceny R, Chandra A, Cheh JJ, Guithues-Amrhein D, Hannon NJ, Hutchison PD, et al. Financial reporting in XBRL on the SEC's EDGAR System: a critique and evaluation. J Inf Syst 2005;19(2):191–210. Duangploy O, Gay D. International harmonization impact compared: illustration of United States and Japan Financial Statement Ratio Analysis. Journal of American Academy of Business 2005:225–30 [March]. Dye R. Disclosure of nonproprietary information. J Acc Res 1995;23(1):123–45. FEI. Lists top ten financial reporting challenges for 2005. IOMA's Rep Financ Anal Plan Rep 2006;5(3):5–6. George D, Mallery P. SPSS for Windows Step-by-Step: A Simple Guide and Reference. 2nd Edition. Boston: Allyn and Bacon; 2000. Gibbons M, Richardson A, Waterhouse J. The management of corporate financial disclosures: opportunism, realism, policies, and procedures. J Acc Res 1990;28(1):121–43. 20 R.F. Premuroso, S. Bhattacharya / International Journal of Accounting Information Systems 9 (2010) 1–20 Gompers P, Ishii J, Metrick A. Corporate governance and equity prices. Q J Econ 2003:107–55 [February]. Gray SJ, Meek GK, Roberts CB. International capital market pressures and voluntary annual report disclosures by U.S. U.K. multinationals. J Int Financ Manag Account 1995;6(1):43–68. Hair Jr JF, Anderson RE, Tatham RL, Black WC. Multivariate Data Analysis. 5th Edition. New Jersey: Prentice-Hall; 1998. Hodge FD, Kennedy JJ, Maines LA. Does search-facilitating technology improve the transparency of financial reporting? Account Rev 2004;79(3):687–703. Jensen M, Meckling W. Theory of the firm: managerial behavior, agency costs and ownership structure. J Financ Econ 1976;3:305–60. Lang M, Lundholm R. Cross-sectional determinants of analyst ratings of corporate disclosure. J Acc Res 1993:306–60 [Autumn]. Larker D, Richardson S. Fees paid to audit firms, accrual choices, and corporate governance. J Acc Res 2004;42:625– [June]. Lublin JS 2006. Compliance Panels Slowly Take Hold. The Wall Street Journal (1/9/06): B4. Mahalanobis P. On the generalized distance in statistics. Proc Natl Inst Sci India 1936;2:49–55 [Calcutta]. Malone D, Fries C, Jones T. An empirical investigation of the extent of corporate financial disclosure in the oil and gas industry. J Account Audit Financ 1993;8:249–73 [Summer]. Mertler CA, Vannatta RA. Advanced and Multivariate Statistical Methods. 3rd Edition. Pyrczak Publishing; 2005. Miller GS. Earnings performance and discretionary disclosure. J Acc Res 2002;40:173–204. Murphy KJ, Zimmerman JL. Financial performance surrounding CEO turnover. J Account Econ 1993;16(1/2/3):273–315 Myers R. Classical and Modern Regression with Applications. 2nd Edition. Boston, MA: Duxbury; 1990. Naser K, Nusuerbeh R. Quality of financial reporting: evidence from listed Saudi non-financial companies. Int J Account 2003;38:41–69. Penman SH. The quality of financial statements: perspectives from the recent stock market bubble. Account Horiz 2003;17(Supplement):77–96. Pinsker R. A theoretical framework for examining the corporate adoption decision involving XBRL as a real-time business reporting technology. A Working Paper presented to the 2004 Information Systems Section Midyear Meeting; 2004. Sarbanes-Oxley Act of 2002. Section 409. Premuroso RF, Bhattacharya S. Is there a relationship between firm performance, corporate governance, and a firm's decision to form a technology committee? Corporate Governance: An International Review 2005;15(6):1260–76. Scannell K. 2006a. SEC to Propose Overhaul of Rules on Executive Pay. The Wall Street Journal (1/10/06): A1. Scannell K. 2006b. Cox Puts More Tech Into SEC. The Wall Street Journal (3/22/06): C1. Schipper K, Vincent L. Earnings quality. Account Horiz 2003;17(Supplement):97–110. Singhvi S, Desai HB. An empirical analysis of the quality of corporate financial disclosure. Account Rev 1971;46:621–32 Spero L. The Extent of Voluntary Disclosure of Financial Information in Three European Capital Markets: An Exploratory Study. Unpublished Doctoral Dissertation. Cambridge, MA: Harvard University; 1979. Tabachnick BG, Fidell LS. Using Multivariate Statistics. 3rd Edition. New York: HarperCollins; 1996. Trabelsi SR, Labelle R, Laurin C. The management of financial disclosure on corporate websites: a conceptual model. Account Perspect 2004;13(2). Verrecchia RE. Discretionary disclosure. J Account Econ 1983;5:179–94. Verrecchia RE. Essays on disclosure. J Account Econ 2001;32:97–180. Wallace RSO, Naser K. Firm specific determinants of the comprehensiveness of mandatory disclosure in the corporate annual reports of firms listed on the stock exchange of Hong Kong. J Account Public Policy 1995;14:311–68. Wallace RSO, Naser K, Mora A. The relationship between the comprehensiveness of corporate annual reports and firm characteristics in Spain. Account Bus Res 1994;25(97):41–53. Wallman SMH. The future of accounting and financial reporting, Part IV: Access Accounting. Account Horiz 1997;11(2): Weber RA. XML, XBRL, and the Future of Business and Business Reporting, in Trust and Data Assurances in Capital Markets: The Role of Technology Solutions. In: Roohani ST, editor. Smithfield, RI: Bryant College; 2003. p. 3–6. Xbrl.org. XRBL: Understanding the XML Standard for Business Reporting and Finance; 2002 [White paper, www.xbrl.org].
|+| نوشته شده توسط امیر عباسی در یکشنبه هجدهم دی 1390  |
 
 
بالا