Compliance with International Financial Reporting

Compliance with International Financial Reporting Standards and auditor choice:New evidence on

the importance of the statutory audit

Christopher Hodgdon a ,Rasoul H.Tondkar b,?,

Ajay Adhikari c ,David W.Harless b

a

The University of Vermont,United States b Virginia Commonwealth University,United States

c The American University,Unite

d States Abstract

We examine the impact of auditor choice on IFRS compliance under the assumption of strict exogeneity of auditor choice.Our results reveal that compliance is positively related to auditor choice after controlling for firm size,profitability,leverage,degree of international diversification,and whether a firm has a U.S.listing or was audited according to International Standards of Auditing.We also find that auditor choice is positively related to firm compliance when controlling for unmeasured,firm-specific effects.The results of our study reinforce the importance of developing institutional mechanisms (e.g.,enforcement,auditing,or corporate governance structures)to encourage compliance with IFRS.

?2008University of Illinois.All rights reserved.

Keywords:International Financial Reporting Standards;IFRS compliance;Auditor choice;IASB;V oluntary

disclosure

The International Journal of Accounting 44(2009)33–

55

?Corresponding author.Department of Accounting,Virginia Commonwealth University,Richmond,V A 23284,(804)828-7156.

E-mail addresses:chodgdon@https://www.360docs.net/doc/2f17940410.html, (C.Hodgdon),rhtondka@https://www.360docs.net/doc/2f17940410.html, (R.H.Tondkar),

aadhika@https://www.360docs.net/doc/2f17940410.html, (A.Adhikari),dwharles@https://www.360docs.net/doc/2f17940410.html, (D.W.Harless).

0020-7063/$-see front matter ?2008University of Illinois.All rights reserved.

doi:10.1016/j.intacc.2008.12.003

34 C.Hodgdon et al./The International Journal of Accounting44(2009)33–55

1.Introduction

In recent years,the international accounting standards movement has gained real momentum.The European Union's(EU)decision requiring all EU-listed companies to follow International Financial Reporting Standards(IFRS)1for their consolidated accounts, beginning with the2005reporting year,represents considerable progress towards the goal of international convergence of accounting standards.Despite this progress,significant challenges remain.Foremost among these challenges is ensuring consistent compliance with IFRS,also an important consideration in light of the recent elimination of the U.S.Securities and Exchange Commission's(SEC)requirement for a reconciliation to U.S.GAAP by firms that apply IFRS in their U.S.filings.2

Proponents of the mandated use of IFRS argue that its use will lead to comparable information and expanded financial-statement disclosure resulting in increased market efficiencies.However,absent effective enforcement,even the best accounting standards will be inconsequential(Hope2003).Recent studies(Street,Gray&Bryant,1999;Street& Bryant,2000;Street&Gray,2001;Glaum&Street,2003)document significant non-compliance with the disclosure requirements of IFRS in many areas.Not surprisingly,the IASB and capital-market regulators are increasingly turning their attention to compliance and enforcement issues related to IFRS.

The importance of developing institutional mechanisms(e.g.,enforcement,corporate governance structures,and auditing)to encourage compliance with IFRS has been widely discussed in the literature(Ball,Robin&Wu,2003;Brown&Tarca,2005).There is also a growing realization that while countries can adopt international accounting standards relatively easily,developing the institutional mechanisms to ensure successful implementa-tion and foster compliance is much more complex and time consuming.Developing local regulatory and enforcement mechanisms,instituting corporate-governance structures,and building capacity(systems,methodologies,application guidance,training,and education) all require a major commitment and investment both in terms of time and money.In this context,independent audit and audit quality become important first-line responses to encourage compliance with IFRS.Importing assurance by turning to international audit firms which offer a degree of audit quality that goes beyond what is required by existing local regulatory and enforcement mechanism is one way for foreign firms to gain credibility for their financial reporting.Moreover,the emergence of international audit networks and the capacity of large international firms to build audit capabilities quickly by transferring audit technology and key personnel to affiliates in countries with weak accounting institutional frameworks are playing a special role in the implementation of IFRS in such countries.Both the statutory audit and effective enforcement mechanisms are important in ensuring compliance with IFRS;the two are naturally compatible and mutually aggrandizing.

1To simplify the exposition,we use the term IFRS to refer to both International Accounting Standards(IAS) issued by the International Accounting Standards Committee(IASC)and International Financial Reporting Standards(IFRS)issued by the IASC's successor,the International Accounting Standards Board(IASB).

2After many years of debate,the SEC recently dropped its Form20-F reconciliation requirement that required listed https://www.360docs.net/doc/2f17940410.html,panies using IFRS to reconcile their financial results to U.S.GAAP.The SEC is also seeking comments as to whether https://www.360docs.net/doc/2f17940410.html,panies should be allowed to choose between U.S.GAAP and IFRS in preparing their financial statements.

In this study,we examine the determinants of IFRS disclosure compliance and the impact of auditor choice on IFRS compliance under the assumption of strict exogeneity of auditor choice.We ascertain firm compliance with the disclosure requirements of IFRS through an examination of the 1999and 2000annual reports of a sample of non-U.S.firms that claim to comply with IFRS.We measure compliance using both a weighted and unweighted disclosure score.We focus on these two years because several new IFRS came into effect during this period,including IAS 1Revised ,which requires firms that claim full compliance with IFRS to fully comply with all applicable standards.

Our results using ordinary least squares regression on a pooled cross-sectional dataset reveal that compliance improves between 1999and 2000,and is positively related to auditor choice,which we define as being audited by a Big-5+2auditor (the largest international accounting firms at the time,KPMG,Price Waterhouse,Coopers and Lybrand,Ernst &Young,Arthur Andersen,plus BDO and Grant Thornton).We control for firm size,profitability,leverage,degree of international diversification,and whether the firm has a U.S.listing or was audited according to International Standards of Auditing.Further,we obtain the same finding of a positive relationship between auditor choice and IFRS compliance when controlling for unmeasured,firm-specific effects by using first differencing with the two-year panel dataset.

Our study contributes to the existing literature on factors that explain compliance with the disclosure requirements of IFRS.Several recent studies report considerable variation in the level of compliance with IFRS disclosures and find that the level of compliance varies with certain corporate characteristics (e.g.,non regional listing,Big-5+2audit,country of domicile)(Street &Bryant,2000;Street &Gray,2001;Glaum &Street 2003).We extend this research by examining compliance over a two-year period using panel-data statistical methods to control for unknown firm-specific effects that may affect IFRS compliance.Our findings also provide support for the viewpoint that the extent of compliance with accounting standards is as important as the standards themselves,since even the highest quality standards may amount to little more than the lowest quality standards in the absence of full compliance (Sunder,1997).More specifically,our results reinforce the importance of international auditing firms and quality audits as a first-line response in encouraging compliance with IFRS.

Ball et al.(2003),Leuz (2003),and Francis,Khurana,and Pereira,(2005)argue that the global accounting debate focuses too much attention on standards per se and not enough attention on market and institutional forces that influence reporting incentives.Our results reinforce the importance of developing institutional mechanisms (e.g.,enforcement,auditing,or corporate-governance structures)to encourage compliance with IFRS.The results of our study,using two years of data,reveal a positive relationship between IFRS compliance and audit firm size,and provide evidence of the importance of the audit function.

Our study also informs the public policy debate on the implementation of IFRS by providing evidence of the importance of the statutory audit in ensuring that compliance is adequate.Prior studies find that audit quality is related to audit-firm size,and that higher-quality auditors are more likely to improve accounting controls and detect accounting irregularities due to greater knowledge and skills (DeAngelo,1981;Douthett,Duchac &Warren,2001;Moizer,1997;Dunn &Mayhew,2004).

The paper is organized as follows.In Section 2,we discuss the relationship between the statutory audit and IFRS compliance,and in Section 3we present the hypotheses and control 35

C.Hodgdon et al./The International Journal of Accounting 44(2009)33–55

36 C.Hodgdon et al./The International Journal of Accounting44(2009)33–55

variables.We discuss our data,sample selection,and descriptive statistics in Sections4and5.In Sections6and7,we present our research design and results of the study.Finally,in Section8, we provide a summary and conclusion of the study.

2.IFRS compliance and the statutory audit

The objective of the IASB is to develop“a single set of high quality,understandable and enforceable global accounting standards that require transparent and comparable information in general purpose financial statements”(IASB,2007).Zhang Wei Guo, chief accountant with the China Securities and Regulatory Commission,recently stated,“No matter how high the quality of an accounting standard,there will not be high quality and comparable accounting information without reasonable and consistent interpretations,

and strict enforcement”(KPMG,2006,41).Application of IFRS in China is particularly challenging since the IFRS principles-based approach requires more individual judgment than the traditional Chinese approach(KPMG,2006).

Compliance with IFRS is a contentious issue.Cairns(1999)documents nine categories of firm compliance with IFRS,ranging from full compliance to“unqualified description of differences.”Cairns(1998,1999)finds that some companies use a mixture of IFRS and domestic GAAP,while others use IFRS with stated exceptions.Street et al.(1999)report significant noncompliance in many areas.In a subsequent study,Street and Bryant(2000) report that the overall level of compliance for all sample firms was equal to or less than75%for many of the standards examined and that firm compliance is higher for firms with U.S.listings versus those without such a listing.Thus,while firms may be claiming full compliance with IFRS,significant deviations exist.

Concerned about the casual approach many companies were taking towards the application of IFRS,the IASB issued IAS1Revised“Presentation of Financial Statements”in1997(effective for all firms whose fiscal year begins on or after July1,1998).It states the following:

An enterprise whose financial statements comply with International Accounting Standards should disclose that fact.Financial statements should not be described as complying with International Accounting Standards unless they comply with all the requirements of each applicable Standard and each applicable Interpretation of the Standing Interpretations Committee.

With the passage of IAS1Revised,the emphasis of the IASB has shifted from encouraging firms to use IFRS to ensuring full compliance,with the result being a reduction in the number of firms that claim to comply fully with IFRS.Since IAS1Revised does not prohibit firms from claiming compliance“with exceptions,”many firms still use IFRS for accounting issues not covered by local requirements.

Consistent application of IFRS is a stated requirement by the Federation des Experts Comptables Europeens(FEE,2001).However,there is currently no international,pan-European securities regulator charged with enforcement of IFRS.In2001,the FEE conducted a study assessing the enforcement mechanisms within14EU countries,two non-EU countries,and three central and eastern European countries,and found that in nearly half of the countries surveyed there is no institutional-oversight system in place(FEE,

2001).In Switzerland,for instance,enforcement is carried out on a complaint basis,and there is “no intervention of public bodies in the process of preparation and audit of the financial statements ”(FEE,2001,70).The FEE study also noted that in some countries,such as Ireland and Spain,regulator-based enforcement systems are being created.In Germany,for example,the newly created Financial Reporting Enforcement Panel serves to enforce financial reporting requirements of listed companies (Cairns,2004).

In a subsequent discussion paper,the FEE called on all EU member states to review their arrangement for enforcement of IFRS.Rather than calling for a single,pan-European enforcement body similar to the U.S.SEC,the FEE called for the creation of an enforcement-coordination mechanism,dubbed the European Enforcement Coordination (EEC),which would coordinate the national enforcement bodies of member states.According to the FEE,arrangements for the enforcement of accounting standards at the national level,rather that at the European or global level,is consistent with the EU principle of “subsidiarity ”(FEE,2002,48).

Given the challenges of establishing or modifying the enforcement bodies in the EU and the additional challenge of creating the EEC,the FEE acknowledges that the statutory audit of financial information will continue to play a major role in the enforcement of IFRS.The statutory audit function is intended to ensure the proper application of accounting standards,if audit firms maintain high quality and consistent auditing standards.Thus,the FEE has called for the establishment of benchmarks for ethics standards and quality assurance among international auditing firms (FEE,2001).

For most firms,the breadth and volume of data required to comply with IFRS represents a data-collection challenge (Meall,2004).The importance of auditors is evidenced by the degree to which auditors are involved in working with their clients,assisting them in learning complex standards,and in setting up their accounting systems so that they may prepare IFRS-compliant financial statements (Brown &Tarca,2005).To address the important issue of IFRS compliance,the largest international audit firms are developing training programs and initiating audit controls specifically for IFRS audits (Street,2002).Absent a strong,pan-European enforcement body similar to the SEC,or the effectively coordinated enforcement bodies of member states,the statutory audit constitutes the single most important element in ensuring complete compliance with IFRS.Thus,at the present time,the importance of the auditor in ensuring adequate compliance with IFRS cannot be overstated.

Several studies examine the impact of audit firm choice on IFRS compliance.Examining the 1998annual reports of 279companies that claim to comply with IFRS,post IAS 1Revised ,Street and Gray (2001)report considerable variation in the level of compliance with IFRS disclosures and find that compliance is positively associated with the firm being audited by a Big 5+2accounting firm.

In a study examining the extent to which companies listed on Germany's New Market comply with IFRS disclosure requirements in their 2000annual reports,Glaum and Street (2003)find that compliance ranges from 100%to 41.6%,with an average of 83.7%,and that compliance is positively related to the size of the audit firm.

In a study of the effects of IAS 14Revised ,Prather-Kinsey and Meek (2004)examine how IAS 14Revised affected the segment-disclosure practices of companies claiming to comply with IFRS.They further explore whether firms'segmental reporting practices are 37

C.Hodgdon et al./The International Journal of Accounting 44(2009)33–55

38 C.Hodgdon et al./The International Journal of Accounting44(2009)33–55

related to certain firm-specific characteristics and find that compliance with IAS14Revised is positively related to a firm being audited by a Big-5auditor.

In the latest empirical study of the factors that influence IFRS compliance,Ali,Ahmed, and Henry(2004)examine the level of compliance with the disclosure requirements of14 IFRS-based national accounting standards in South Asia.They find significant variation in compliance levels across countries,and that compliance is not significantly related to the quality of the firm's external auditors.

Thus,with regard to the impact of audit-firm size on IFRS compliance,the results of prior studies are somewhat mixed.Glaum and Street(2003),Street and Gray(2001),and Prather-Kinsey and Meek(2004)find a positive relationship while Ali et al.(2004)find no significant relationship.

Unlike these prior studies,we examine IFRS compliance over a two-year period,thereby making it possible to control for time-invariant,firm-specific factors.It would not be unreasonable to suggest that there may be other,firm-specific factors that affect compliance, but that are unknown or unmeasurable and thus have not been controlled for in prior studies. Using two-year panel data allows us to effectively control for those factors particular to each firm(i.e.,firm-specific effects)that may affect IFRS compliance.Two-year panel data also allows us to test the assumption of strict exogeneity of audit firm size.

3.Main hypotheses and control variables

3.1.Hypotheses

As previously mentioned,prior studies find that audit quality is not independent of audit firm size,and that higher-quality auditors are more likely to improve accounting controls and detect accounting irregularities due to greater knowledge and skills(DeAngelo,1981; Douthett et al.,2001;Moizer,1997;Dunn&Mayhew,2004).In general,prior research finds that the type of auditor may be associated with a firm's overall level of disclosure (Firth,1979;Inchausti,1997;Dumontier&Raffournier,1998).Additionally,most prior studies of the effects of firm-specific factors on IFRS compliance find IFRS compliance is higher for firms audited by large,international audit firms(Glaum&Street,2003;Street& Gray,2001;Prather-Kinsey&Meek,2004).

Prior studies do not,however,address the concern that the relationship between IFRS compliance and auditor choice may not be strictly exogenous,i.e.,there is concern that firms may select auditors as part of their overall disclosure strategy,violating the assumption about the direction of causality made in those studies(Dunn&Mayhew,2004). If the assumption of strict exogeneity does not hold,then it could be the case that a firm's increasing compliance precedes a change to a Big5+2auditor,and not vice-versa.3We, therefore,explicitly test for the assumption of strict exogeneity of auditor choice.We also 3The assumption of strict exogeneity holds if for each time period t,the expected value of the error term e

,

t conditional on the explanatory variables for all time periods,is zero.In the context of this paper,strict exogeneity implies that firms'choice of auditors is not affected by changes in IFRS compliance.For example,a positive error in time t(a higher level of compliance)would not affect auditor choice,either in the current or future time periods [for other examples of strict exogeneity,see Wooldridge(2003)page332].

examine the relationship between IFRS compliance and auditor size using information from the change in compliance and change in audit-firm size.Our primary hypothesis is as follows:

H1.Ceteris paribus ,compliance with the disclosure requirements of IFRS is positively related to being audited by a Big-5+2auditor.

One problem frequently encountered in international studies is how to control for variations in GAAP regimes and other firm and country-level factors.There are innumerable firm and country-level factors that could potentially impact IFRS compliance.Some factors that affect IFRS compliance may be time-invariant.We refer to such factors particular to each firm as “firm fixed effects.”In our study,variation in GAAP regimes is not an issue,since all of our sample firms claim compliance with IAS.To control for time-invariant factors that may affect IFRS compliance,we perform first differencing with two-year panel data.That is,we examine the change in IFRS compliance as a function of the change in the quality of the firm's external auditor,as well as the change in the control variables.We argue that increases in firm compliance with IFRS disclosure requirements should be associated with improved quality of the external auditor.Thus,our second hypothesis is as follows:

H2.Ceteris paribus ,the change in compliance with the disclosure requirements of IFRS is positively related to the change to being audited by a Big-5+2auditor.

3.2.Control variables

We develop parsimonious models by including only those explanatory variables found in prior studies to be statistically significant determinants of IFRS compliance.Factors not considered in our study (but considered in prior studies)include country of origin,industry,and the manner in which firms refer to the use of IFRS in their accounting policy footnotes.With respect to country of origin and industry,we do not examine these variables because they are time-invariant,so that the first-differences of these variables equal zero.With respect to the accounting policy variable,studies prior to the effective date of IAS 1Revised considered this variable because companies continued to refer to the use of IFRS in a variety of ways,a practice explicitly disallowed by IAS 1Revised .Since we study firms post-IAS 1Revised ,this variable is not necessary.The control variables examined in this study are as follows.

3.2.1.International Standards of Auditing

Generally speaking,audit quality should be higher when a large audit firm performs the audit,but quality may also depend on the type of auditing standards employed.An audit by a large,international audit firm does not automatically mean strict adherence to the highest standards of auditing.Prior studies indicate that IFRS compliance is higher when a firm is audited using International Standards of Auditing (ISA)(Street &Bryant,2000;Glaum &Street,2003).The evidence of the relationship is mixed,however,with Street and Gray (2001)finding no evidence of a positive association.Based upon the evidence from prior studies,we control for the association between IFRS compliance and whether an auditor 39

C.Hodgdon et al./The International Journal of Accounting 44(2009)33–55

40 C.Hodgdon et al./The International Journal of Accounting44(2009)33–55

employed ISA,with the expectation that compliance is greater when ISA is used.We determine whether or not ISA was employed in the audit through an examination of the firm's audit report.We define the variable ISA as equal to“one”if the firm's audit report states that ISA was employed in the audit and“zero”otherwise.

3.2.2.Firm size

Prior research suggests that larger firms disclose more information than smaller firms (Atiase,1980,1985).In international accounting studies,larger firms have been found to have a more diverse shareholder base,lower information processing and production costs,and lower competitive-disadvantage costs(Chow&Wong-Boren,1987;Gray&Roberts,1989; Meek,Roberts&Gray,1995;Zarzeski,1996).Studies of the relationship between IFRS compliance and firm size are mixed,with Prather-Kinsey and Meek(2004)and Ali et al. (2004)finding that firm size is positively related to IFRS compliance,while Street and Gray (2001)and Glaum and Street(2003)find no significant relationship.Since the findings of prior studies are mixed,we have no prediction regarding the impact of size on IAS compliance. Consistent with Meek et al.(1995),we measure firm size using year-end revenue,converted to U.S.dollars at year-end exchange rates,so that all amounts are stated on a consistent basis. Revenue-recognition rules across GAAP regimes are not as diverse as asset-valuation and recognition rules,and thus the use of revenue as a proxy for firm size is appropriate.

3.2.3.U.S.listing

Adhikari and Tondkar(1992)report significant variability in the amount of disclosure required as part of the listing and filing requirements of stock exchanges around the world. Of a total of35stock exchanges examined,the New York Stock Exchange(NYSE)had the most stringent disclosure requirements.Other studies find that listing status has a statistically significant association with the level of disclosure provided by the firm(Meek et al.,1995;Gray,Meek&Roberts,1995).Street and Bryant(2000)find that the extent of compliance with IFRS is greater for firms with a U.S.listing.Firms that use IFRS and have a U.S.listing may be providing more information to the market through the filing of their Form20-F in the U.S.We define the variable USLIST as equal to“one”if the firm has a United States listing and“zero”otherwise.

3.2.

4.Profitability

Several prior studies have examined the effect of firm profitability on the level of disclosure.The results of these studies are mixed.For example,Wallace,Naser and Mora (1994)and Wallace and Naser(1995)find a significant association,while Meek et al. (1995)and Dumontier and Raffournier(1998)provide no evidence of an association.Street and Bryant(2000),Street and Gray(2001),and Glaum and Street(2003)find no association between firm profit and IFRS disclosure compliance.Consequently,we offer no prediction regarding the impact of profitability on IFRS compliance.In this study, profitability is defined as the ratio of net income before tax to total shareholders'equity.

3.2.5.International diversification

We further test for the effects of international diversification on IFRS compliance.Firms with international operations may face additional regulatory hurdles with regard to

disclosure compliance,and thus these firms may have a greater incentive to provide timely and complete disclosures.Street and Gray (2001)and Glaum and Street (2003)find no association between their measure of multinationality and IFRS compliance.We include in our models a control variable for international diversification equal to the number of geographic segments in which a firm operates.

3.2.6.Leverage

Prior research suggests that leverage is a market force that is significantly associated with the level of annual report disclosures.Sengupta (1998)finds that firms with high disclosure quality ratings enjoy a lower cost of issuing debt,consistent with the argument that disclosures reduce the perception of default risk.Zarzeski (1996)finds that lower debt ratios are associated with a higher level of annual-report disclosures.She argues that this negative relationship exists because in many countries (e.g.,Germany and Switzerland),bank financing is the primary source of capital,and firm disclosures are more private than public.The argument does not,however,consider the markets for public debt where public debtors may demand more public financial disclosures.Although Ali et al.(2004)find that leverage is unrelated to the level of IFRS disclosure compliance,no other prior study of IFRS compliance has examined the effects of leverage.We define leverage as the firm's year-end total liabilities divided by total assets.We make no prediction of the directional effects of firm leverage on IFRS disclosure compliance.

4.Data

4.1.Instrument to measure compliance with IFRS:The disclosure index

We use a disclosure index that focuses on items disclosed in the financial statements and footnotes and is specifically targeted to those IFRSs where previous research has indicated noncompliance (Hodgdon et al.2008).This disclosure index includes the disclosure items examined in the Street and Gray (2001)study plus disclosure items that were not in effect at the time of their study.Table 1provides a summary of the IFRS disclosures included in the disclosure index with citations of the relevant literature that justify their inclusion.If a firm indicates early adoption of an IFRS in its annual report,the expanded disclosure requirements are considered mandatory.Consistent with Street and Bryant (2000),for those IFRS that have been revised but are not yet effective for the years 1999and/or 2000,the number of applicable mandatory disclosures is based upon the IFRS as originally issued.4

Our disclosure index includes 209disclosure items,each coded as disclosed,not disclosed,or not applicable,for a sample of 101firms.The disclosure is deemed not relevant for the firm if it appears that disclosure of the information is not mandatory.In contrast,if it is evident that an item of disclosure is applicable,based upon a review of the firm's entire annual report (e.g.,there is mention in the annual report of an asset under capital leases,but no disclosure of the amounts)then the disclosure is coded as not disclosed.As opposed to a simple disclosed/not disclosed procedure,we review the entire annual report and make judgments as to the applicability of certain required disclosures.4The complete IFRS disclosure checklist is available from the authors upon request.

41

C.Hodgdon et al./The International Journal of Accounting 44(2009)33–55

This approach is consistent with Cooke (1989),who developed a scoring scheme to capture levels of disclosure for a sample of Swedish firms.One researcher was primarily responsible for data collection.The collection of two years of data served as a natural check on the accuracy of the coding,e.g.,when geographic disclosures for one firm were present for 1999but not for 2000,the researcher went back and double checked both years for accuracy.

Based on these 209disclosure items,we compute an unweighted and weighted compliance score for firm j in year t .Our unweighted IFRS compliance score for each firm in each year is a relative measure,defined as the number of mandatory disclosures actually provided by a firm in year t divided by the number of mandatory disclosures that should have been provided by the firm in year t ,as follows:

PR X C jt =

X 209i =1d ijt r ijt =X 209i =1r ijt e1T

where,

PR _C jt is the unweighted IFRS compliance score for firm j in year t ;

d ijt indicates th

e disclosure o

f item i ,by firm j ,in year t ,i =1...,209;j =1…,101,and

t =1999,2000.d ijt =1if item i is disclosed by firm j in year t ,and d ijt =0if the item is either not disclosed or not relevant.Table 1

Components of IAS disclosure index a

IAS

Disclosures Inclusion supported by:IAS 12,income taxes

All required disclosures Street and Bryant (2000)IAS 14,segment reporting

All required line of business and geographic disclosures Street and Bryant (2000);Cairns (1999);UNCTAD (1998)IAS 16,property,plant,and

equipment

All required disclosures for revalued property,plant,and equipment Street et al.(1999)IAS 17,leases

All required disclosures for lessees Street and Bryant (2000)IAS 19,retirement benefits

All required disclosures for defined benefit pensions Street et al.(1999);Street and Bryant (2000);Cairns (1999)IAS 23,borrowing costs Accounting policy for borrowing cost

and required disclosures for capitalized

borrowing cost where applicable

Street and Bryant (2000)IAS 32,financial instruments Information about extent and nature of

financial instruments;information about

interest rate risk;information about

credit rate risk

Street &Gray (2001)IAS 33,earnings per share All required disclosures

Street and Bryant (2000)IAS 35,discontinued operations All required disclosures

Not examined in prior literature IAS 36,impairment of assets All required disclosures

Not examined in prior literature IAS 37,provisions,contingent liabilities and contingent assets All required disclosures

Not examined in prior literature IAS 38,intangible assets

All required disclosures Not examined in prior literature a Only IFRS found by prior studies to be problematic in terms of disclosure compliance are considered in this study,as well as newly issued IFRS applicable for firms in 1999and 2000.IFRS not considered in this study include the following:IFRS 1–8;IAS 1–11,18,20,21,24–31,34,39–41.

42 C.Hodgdon et al./The International Journal of Accounting 44(2009)33–55

r ijt indicates whether disclosure item i is required for firm j in year t .Thus r ijt =1if

disclosure for item i is required for firm j in year t ,and r ijt =0if the disclosure is not required.

The unweighted,dichotomous approach is consistent with Street and Bryant (2000),and Street and Gray (2001).

4.2.Alternative compliance measure:The Saidin index

In computing unweighted disclosure scores,we assume equal weighting of all disclosure items.We also use an alternative disclosure index,the “Saidin ”index,which weights each disclosure item by the percentage of firms in the sample that do not comply with the item (Spetz &Baker,1999).5The Saidin index assumes disclosures that few firms provide are more costly and may have greater informational value to financial analysts.Less common disclosures therefore,receive a greater weighting in the Saidin index,with the weight approaching one for disclosures that no firm provides.Conversely,a disclosure that is less costly or onerous to provide and is provided by all firms would receive a weight closer to zero.

In order to create the Saidin index for firm j in year t ,we consider the same 209disclosures used in calculating the unweighted compliance score,each one indexed as i =1,…,209.For each disclosure,a weight a i ,is calculated:

a i =1àP 2t =1P

101j =1d ijt r ijt

P 2t =1P

101j =1

r ijt e2Twhere d ijt and r ijt are as previously defined.In the circumstance where every item is required for every firm (r ijt =1for all i ,j ,and t ),then the denominator of the ratio (to be subtracted from one)would simply be the product of the number of firms and the number of time periods.If all firms provide this required disclosure,then its weight equals zero (i.e.,1?1/1),and if none provide this required disclosure,then its weight equals 1(i.e.,1?0/1).The weights are used to compute the Saidin score for firm j in year t as follows:

Saidin jt =X

209i =1a i d ijt r ijt =X 209i =1a i r ijt e3T

5

Spetz and Baker (1999)devised the Saidin index in order to measure the effect of Health Maintenance Organizations on the availability of medical technologies in hospitals.The Saidin index is a weighted sum of indicators for various technologies and services,with each weight being the percentage of hospitals in the sample that did not possess the technology or service.Rare technologies (such as MRI)received higher weights while technologies that were commonplace (such as operating rooms)received lower weights.A hospital's Saidin index score increases with the addition of new technologies but increases relatively more with the addition of rare technologies than with the addition of technologies that are more commonplace.43

C.Hodgdon et al./The International Journal of Accounting 44(2009)33–55

44 C.Hodgdon et al./The International Journal of Accounting44(2009)33–55

The Saidin score for firm j in year t is the sum of the weighted disclosures provided by the firm divided by the sum of the weighted disclosures that should have been provided.

Certain reservations must be made with regard to the application and interpretation of the results using the Saidin index.Spetz and Baker's(1999)application of the Saidin index to health-maintenance organizations assume that less common medical technologies are more important or valuable and thus indicative of the quality of available medical care.Our application of the Saidin index to financial disclosures is only valid to the extent that rare disclosures are more valuable to the users of financial information than more commonplace disclosures,a key assumption that may be difficult to substantiate,given that ours is the first study to apply this technique to financial-disclosure data.

5.Sample selection and descriptive statistics

Following Hodgdon et al.(2008),the sample of interest in this study consists of firms that claim to comply with IFRS for the two years1999and2000.In1999,the IASB posted on its website a list of831firms that claimed to use IFRS as their primary basis of financial reporting.After IAS1Revised,many firms stopped using IFRS.In May,2000,the IASB revised its list of firms that use IFRS.The subsequent list contained183firms.

Using these two lists,and checking for firm addresses and websites in the Worldscope Database and International Companies Database of Lexis-Nexis,we compiled a list containing 633firms that claim,or claimed at some point in the past,to apply IFRS.We contacted these firms and requested a copy of their annual reports for the fiscal years1999and2000.

Of the377companies that responded to the request for their1999and2000annual reports,only208claim to comply with IFRS for both1999and2000.Of these208 companies,107companies are excluded since they belong to specialized industries(e.g., banking,insurance,oil and gas,etc.),or for other reasons(e.g.,not included in IBES database,annual reports not available in English,etc.),resulting in101firms in our sample. Table2gives details of the final sample determination.Table3provides details of the country of origin of the sample firms.

Table2

IAS companies—sample determination

The IASB's1999list of companies that refer to the use of IAS831 Companies whose addresses were not available from either Worldscope or Lexis/Nexis International

(198) Companies Database

Mailed letter requesting copy of1999and2000annual reports633 Companies not responding to mailing(256) Companies responding and eligible for inclusion in sample377 Companies that did not apply IAS for1999and2000(169) Companies that applied IAS for1999and2000208 Companies not included in the IBES database(30) Financial industry,natural resource,utility companies(46) Companies whose fiscal-year-end is other than December31(12) Companies whose1999or2000annual report was unavailable(18) Size=zero(Company declared no revenues)(1) Final sample101

Descriptive statistics for the independent and dependent variables are reported in Table 4.The mean and standard deviation of PR_C (the unweighted compliance score)for 1999and 2000are 58%and 64%,and 21%and 19%,respectively,and are consistent with Street and Gray (2001)who find a mean of 72%and standard deviation of 19%,using a similar unweighted compliance score.The SAIDIN (weighted compliance score)also indicates that there is variability in the level of disclosure provided by IFRS companies.There are no prior studies that utilize a weighted disclosure score with which to benchmark this finding.A very low value for PR_C is noted for one company for both years (5%for 1999and 4%for 2000).Further investigation revealed that the company noted minimal compliance with IFRS in its accounting policy footnotes for both years.Very high compliance is observed for one company (96%),and is also consistent with Street and Gray (2001)who note high compliance by some sample https://www.360docs.net/doc/2f17940410.html,parison-of-means tests for 1999and 2000provide evidence that the mean of our unweighted compliance score has increased between 1999and 2000(p b .05).

Table 4also reports descriptive statistics for the first-difference (change)of our variables between 1999and 2000.The average change in IFRS compliance between 1999and 2000is five percentage points.Three firms changed auditors to a Big-5+2audit firm,and no firms changed from a Big-5+2firm to a non-Big-5+2firm,reflected in the average change in AUDIT of .03.Seven firms changed to being audited according to ISA,while two firms discontinued being audited according to ISA,resulting in a mean change of .05.Prior studies of IFRS compliance have not examined first differences,so we cannot provide a benchmark with which to compare these results.The number of firms that changed to being audited by a Big-5+2audit firm or changed to being audited according to ISA is limited,but is nonetheless incorporated in the standard errors and resulting tests of statistical significance.

Table 5reports the correlation matrix for the variables in our regression analysis.There is a noticeable correlation (.98)between LNPR_C (the log-transformed,unweighted disclosure score)and LNSAIDIN (the log-transformed weighted disclosure score).This is to be expected since LNPR_C and LNSAIDIN are similar measures.Since both variables Table 3

IFRS companies by country

Country

Frequency Austria

3China

3Czech

1Denmark

5Finland

3Germany

34Hong Kong

3Hungary

5Italy

3Netherlands

1South Africa

1Sweden

2Switzerland

37Total 101

45C.Hodgdon et al./The International Journal of Accounting 44(2009)33–55

Table4

Descriptive statistics1999,2000and change in variables1999–2000a

Variable(n=101)19992000Change in variables

1999–2000Comparison-of-means (2000versus1999)t-statistics

PR_C

Mean.58.64.05 1.99* Std.dev.21.19.08

Min.05.04?.18

Max.95.96.40

SAIDIN

Mean.45.50.04 1.53

Std.dev.21.19.08

Min.02.01?.11

Max.88.92.33

AUDIT

Mean.88.91.03.68

Std.dev.32.28.17

Min000

Max111

SIZE??

Mean3,7704,280505.40

Std.dev8,3409,4201,440

Min 2.910.3?804

Max51,70057,7008,700

USLIST

Mean.15.20.05.90

Std.dev.36.40.21

Min000

Max111

PROFIT

Mean.15.10?.04?1

Std.dev.16.39.35

Min?.67?3.03?2.9

Max.54.92.78

ISA

Mean.68.73.05.77

Std.dev.46.44.29

Min00?1

Max111

INTDIV

Mean 3.77 3.78.01.03

Std.dev 1.78 1.78.55

Min11?3

Max993

46 C.Hodgdon et al./The International Journal of Accounting44(2009)33–55

are considered in separate models,there is no concern for multicollinearity.Correlations for other variables are within conventional thresholds.6

6.Research design

To test for functional misspecification,we use several specification tests.We use Ramsey's (1969)Regression Specification Error Test (RESET)as a general test for functional form misspecification.7We find evidence of a highly significant,nonlinear relationship between auditor choice and IFRS compliance,and use the Davidson and MacKinnon (1981)test to determine whether the dependent variable (IFRS disclosure compliance)should appear in our models as log transformed (i.e.,the natural log of IFRS compliance)or in levels.We also use the Davidson and MacKinnon test to determine whether the variable SIZE (firm size)should appear in our models as log transformed or in levels.8

6

Kennedy (1998),page 187,considers correlations greater than .80as the threshold for determining if a correlation is high and a source of potential problems.We also calculate variance inflation factors (VIFs)for the regressors in the first-differences models,and find mean VIFs of 1.32using both PR_C (unweighted compliance)and Saidin (weighted compliance)as dependent variables.VIFs provide a measure of each regressor's independence from the other regressors in the model.The presence of multicollinearity is usually indicated when the largest VIF is greater than 10or the mean VIF is larger than one.None of our VIFs are greater than two,and our mean VIF is sufficiently close to 1to suggest that multicollinearity is not a problem in our models.7RESET adds polynomials of the OLS fitted values from the estimated model to detect general kinds of functional form misspecification (Wooldridge,2003,293).

8The results of these tests suggest that SIZE should appear in our models as log transformed (p =.45),but that IFRS compliance may remain in levels or log transformed (i.e.,the tests indicate no preference,p =.29),in order to achieve proper functional-form specification.Accordingly,we log transform SIZE and our dependent variable,since we believe that the log-transformed variables have the advantage of a more straightforward interpretation of the coefficients.Table 4(continued )

Variable (n =101)

19992000Change in variables 1999–2000Comparison-of-means (2000versus 1999)t -statistics LEV

Mean

.48.50.01.60Std.dev

.21.20.09Min

.02.01?.25Max

.84.92.49a PR_C is the unweighted percentage compliance score for firm j in period t ;SAIDIN is the weighted percentage compliance score for firm j in period t ;AUDIT is a dummy variable with one indicating the firm is audited by a Big-5+2(includes BDO and Grant Thornton)firm,and zero otherwise;SIZE is firm size,as measured by year-end revenue,converted to U.S.dollars at year-end exchange rates for firm j in period t ;USLIST is a U.S.listing dummy variable equal to one if the company has a U.S.listing,and zero otherwise;PROFIT is firm profitability,measured as the ratio of net income before tax to total shareholders'equity;ISA is a dummy variable with one indicating the firm was audited according to International Standards of Auditing (ISAs),zero otherwise;INTDIV is the international diversification variable equal to the number of geographic segments in which a firm operates;and LEV is firm leverage,defined as the firm's year-end total liabilities divided by total assets.

*p b .05.

**In $1,000,000s.

47C.Hodgdon et al./The International Journal of Accounting 44(2009)33–55

Table5

Correlation matrix a

LNPR_C LNSAIDIN AUDIT LNSIZE USLIST PROFIT ISA INTDIV LEV D2000 LNPR_C 1.0000

LNSAIDIN0.9859*** 1.0000

AUDIT0.3510***0.3445*** 1.0000

LNSIZE0.3950***0.3952***0.3468*** 1.0000

USLIST?0.0021?0.00040.16130.3793 1.0000

PROFIT0.04340.04820.12060.3091***0.1158 1.0000

ISA0.16190.14970.13790.1209?0.00540.0870 1.0000

INTDIV0.2572**0.2547*0.15820.5058***0.10990.09730.0482 1.0000

LEV0.2641**0.2468*0.20590.4680***0.0383?0.07310.10390.3474*** 1.0000

D20000.13200.10240.04870.04460.0640?0.07060.05440.00280.0425 1.0000 ***,**,*Significant at the.001,.01,and.05levels,respectively.

a LNPR_C is the natural log of the unweighted compliance score for firm j in period t;LNSAIDIN is the natural log of the weighted compliance score for firm j in period t;AUDIT is a dummy variable with one indicating the firm is audited by a Big5+2(includes BDO and Grant Thornton)firm,and zero otherwise;LNSIZE is the natural log of firm size,as measured by year-end revenue,converted to U.S.dollars at year-end exchange rates for firm j in period t;USLIST is a U.S.listing dummy variable equal to one if the company has a U.S.listing,and zero otherwise;PROFIT is firm profitability,measured as the ratio of net income before tax to total shareholders'equity;ISA is a dummy variable with one indicating the firm was audited according to International Standards of Auditing(ISAs),zero otherwise;INTDIV is the international diversification variable equal to the number of geographic segments in which a firm operates;LEV is firm leverage,defined as the firm's year-end total liabilities divided by total assets;and D2000is a dummy variable equal to one for2000and zero otherwise.48

C. Hodgdon et al. / The International Journal of Accounting 44 (2009) 33

–55

The multivariate models for both the unweighted (4)and weighted (5)IFRS compliance scores are as follows:

LNPR X C j ;t =a j +b 1AUDIT j ;t +b 2LNSIZE j ;t +b 3USLIST j ;t +b 4PROFIT j ;t

+b 5ISA j ;t +b 6INTDIV j ;t +b 7LEV j ;t +b 8D2000j ;t +e j ;t

e4TLNSAIDIN j ;t =a j +b 1AUDIT j ;t +b 2LNSIZE j ;t +b 3USLIST j ;t +b 4PROFIT j ;t

+b 5ISA j ;t +b 6INTDIV j ;t +b 7LEV j ;t +b 8D2000j ;t +e j ;t

e5T

where:

LNPR_C j ,t the natural log of the unweighted compliance score for firm j in period t ;LNSAIDIN j ,t the natural log of the Saidin weighted disclosure score for firm j in period t ;AUDIT j ,t dummy variable with one indicating the firm is audited by a Big 5+2(KPMG,

Price Waterhouse,Coopers and Lybrand,Ernst &Young,Arthur Andersen,plus BDO and Grant Thornton)firm,and zero otherwise;

LNSIZE j ,t the natural log of firm size,as measured by year-end revenue,converted to U.S.

dollars at year-end exchange rates for firm j in period t ;

USLIST j ,t the U.S.listing dummy variable equal to one if the company has a U.S.listing,

and zero otherwise;

PROFIT j ,t firm profitability,measured as the ratio of net income before tax to total

shareholders'equity;

ISA j ,t dummy variable with one indicating the firm was audited according to

International Standards of Auditing (ISAs),zero otherwise;

INTDIV j ,t the international diversification variable equal to the number of geographic

segments in which a firm operates;

LEV j ,t firm leverage,defined as the firm's year-end total liabilities divided by total assets;D2000j ,t the dummy variable equal to one for 2000and zero otherwise;and

αj and e j ,t the fixed effect for each firm (i.e.,firm indicator variables)and the error term,

respectively.

Eqs.(4)and (5)are identical except that Eq.(4)uses the unweighted compliance score whereas Eq.(5)uses the SAIDIN weighted compliance score.The αj in Eqs.(4)and (5)are the firm fixed effects representing all time-invariant factors specific to firm j .In order to estimate the equations,we could use either the first-difference transformation or the within-transformation to eliminate the fixed effects and then apply Ordinary Least Squares to estimate the slope parameters.Note that in this circumstance,where there are two observations per firm,applying either the first-difference or within-transformations produces identical estimates and standard errors for the slopes.To be more specific,although the two methods produce identical regression coefficients and standard errors when t =2,fixed-effects estimation (the within-transformation)allows us to compute a test of the joint significance of the firm indicator variables (both reported in Tables 6and 7).

As previously stated,we examine the impact of auditor choice on IFRS compliance under the assumption of strict exogeneity of auditor choice.It is unclear,a priori ,whether 49

C.Hodgdon et al./The International Journal of Accounting 44(2009)33–55

auditors are responsible for greater disclosure,or if firms select auditors as part of their overall disclosure strategy,and thus the direction of causality may be in question (Dunn and Mayhew,2004).This latter possibility would represent a violation of the strict exogeneity assumption,in which case our estimates would be biased.

We test for the assumption of strict exogeneity of the AUDIT variable using a test suggested by Wooldridge (2002).The test is performed on Eqs.(4)and (5).Using Eq.(4)as an illustration,after first-differencing,the dependent variable is (LNPR_C j ,2000?LNPR_C j ,1999)and the AUDIT variable is (AUDIT j ,2000?AUDIT j ,1999).Under strict exogeneity,if we estimate the first-differenced Eq.(4)with the additional regressor AUDIT j ,2000(results not tabulated),then the additional regressor should not be significant,i.e.,the fact that a firm is audited by a Big-5+2auditor in year 2000should not be a statistically significant determinant of the change in IFRS compliance from 1999to 2000.If,however,a change in IFRS compliance systematically precedes the change in auditor,then the additional regressor should be significant.Thus,assuming strict exogeneity,we test whether the null hypothesis that the

Table 6

Estimation results Eq.(4):unweighted score a

Variables

Pooled OLS First differences AUDIT (Big 5+2auditor)

.3906??.4687??(.1482)(.1736)LNSIZE (firm size)

.0859???.0766(.0252)(.0681)USLIST (U.S.listing dummy)

?.2309?.0912(.1628)(.0637)PROFIT (firm profitability)

?.1278??.0699(.0650)(.0570)ISA (Int'l Standards of Auditing)

.0903.0314(.0976)(.0511)INTDIV (int'l diversification)

.0164?.0710(.0217)(.0585)LEV (firm leverage)

.0451.2466(.2391)(.2092)D2000(year 2000dummy)

.1056???.1008???(.0204)(.0260)Number of observations

202101b R 2=.27R 2=.21

F statistic F =9.08???F =3.39??c

LNPR àC j ;t =a j +b 1AUDIT j ;t +b 2LNSIZE j ;t +b 3USLIST j ;t +b 4PROFIT j ;t +b 5ISA +b 6INTDIV j ;t

+b 7LEV +b 8D2000j ;t +e j ;t

a

Standard errors beneath the coefficients.The standard errors for pooled OLS are adjusted for within-firm correlation of the error terms.

b There are 202/2=101observations for the first-differences estimation.

c Test of the significance of firm fixe

d effects,i.e.,that all αj =α,F =16.48***.?Significant at th

e 0.05level.??Significant at the 0.01level.???Significant at the 0.001level.50 C.Hodgdon et al./The International Journal o

f Accountin

g 44(2009)33–55

coefficient on the additional regressor AUDIT j ,2000is equal to zero.We find no evidence to reject the assumption of strict exogeneity (p =.33).9

7.Empirical results

7.1.Results for the unweighted IFRS compliance score

Table 6presents a summary of the results of Eq.(4),which tests H1and H2using the natural log of our unweighted IFRS compliance score (LNPR_C).Both pooled OLS and OLS on the first-differences are used to test our hypotheses.The pooled OLS multiple regressions examine the effects of the independent variables on IFRS compliance in each firm's 1999and 2000annual reports.The first-differences model controls for the presence of the firm fixed effects (different intercepts for each firm)by first differencing the data so that the dependent variable becomes change in LNPR_C (i.e.,LNPR_C j,t ?LNPR_C j ,t ?1),Table 7

Estimation results Eq.(5):Saidin weighted score a

Variables

Pooled OLS First differences AUDIT (Big 5+2auditor)

.5313??.6459?(.1974)(.2746)LNSIZE (firm size)

.1242????.0621(.0349)(.0706)USLIST (U.S.listing dummy)

?.3207?.1021(.2403)(.1013)PROFIT (firm profitability)

?.1816??.0196(.0866)(.0777)ISA (Int'l Standards of Auditing)

.1106?.0934(.1337)(.0801)INTDIV (int'l diversification)

.0221?.0679(.0297)(.0941)LEV (firm leverage)

?.0076.4830(.3408)(.2834)D2000(year 2000dummy)

.1062???.1353???(.0289)(.0324)Number of observations

202101b R 2=.25R 2=.24

F statistic F =8.62???F =3.79??c

LNSAIDIN j ;t =a j +b 1AUDIT j ;t +b 2LNSIZE j ;t +b 3USLIST j ;t +b 4PROFIT j ;t +b 5ISA +b 6INTDIV j ;t

+b 7LEV +b 8D2000j ;t +e j ;t

a

Standard errors beneath the coefficients.The standard errors for pooled OLS are adjusted for within-firm correlation of the error terms.

b There are 202/2=101observations for the first-differences estimation.

c Test of the significance of firm fixe

d effects,i.e.,that all αj =α,F =19.89***.?Significant at th

e 0.05level.??Significant at the 0.01level.???Significant at the 0.001level.9See Wooldridge (2002,285)for a more detailed description o

f this test.

51C.Hodgdon et al./The International Journal of Accounting 44(2009)33–55

52 C.Hodgdon et al./The International Journal of Accounting44(2009)33–55

which is regressed on the changes in the independent variables.In contrast to the first-differences where the intercept is allowed to vary across firms,the pooled OLS regression constrains the intercept to be the same for all observations.

Table6reports the results of pooled OLS regression and OLS using first-differences.The R-squared statistic for Eq.(4)using pooled OLS is0.27,consistent with prior https://www.360docs.net/doc/2f17940410.html,ing pooled OLS,the coefficient for Big-5+2auditor(AUDIT)is significant(p b.01)and in the expected direction.Under the assumption of strict exogeneity of the AUDIT variable,this finding provides evidence that being audited by a Big-5+2auditor causes an increase in IFRS compliance,and supports H1.The coefficient on the AUDIT variable using pooled OLS suggests that being audited by a Big-5+2auditor(versus being audited by a non-Big-5+2 auditor)results in a48%increase in IFRS compliance.

The coefficient for LNSIZE(firm size)is highly significant(p b.001)and in the expected direction,indicating that larger firms are better able to comply with IFRS disclosure requirements.This finding is consistent with prior studies that find larger firms disclose more information than smaller firms(Atiase,1980,1985).The coefficient for firm profitability (PROFIT)is significant(p b.05)and negative,implying that IFRS compliance decreases as a firm's ratio of net income to shareholders'equity increases.As discussed earlier,because the results of prior studies are mixed with regard to the association between profitability and disclosure,we make no prediction regarding the impact of profitability on IFRS compliance.

Results from using first-differences estimation of Eq.(4)provide evidence in support of H2,that the change to being audited by a Big-5+2auditor results in greater IFRS compliance.OLS using first-differences assumes that an examination of the change in IFRS compliance as a function of the change in the independent variables will provide estimates that are superior to those obtained using pooled OLS,which are biased if such fixed effects exist and are correlated with the other regressors in the model.One indication of the appropriateness of a model specification that controls for fixed effects is a test of the joint significance of the firm indicator variables.Another indication is a comparison of the coefficients estimated using pooled OLS with those obtained using first differences,with large differences suggesting that OLS is biased.

We see from Table6that there are two indications that OLS estimates are biased.First,as shown in Table6(note c),the indicator variables for each firm are jointly significant.10 Second,many of the coefficient estimates are quite different between the pooled OLS and first-differences estimation,most noticeably for USLIST,PROFIT,ISA,INTDIV,and LEV.The coefficient estimates for the AUDIT variable,however,are very similar(.39versus.46).After controlling for firm fixed effects,the coefficient for AUDIT continues to be very significant, providing strong support for H2.The coefficient on the AUDIT variable using first-differences suggests that a change to being audited by a Big-5+2auditor would result in a60%increase in IFRS compliance.The R-squared statistic for the first-differences model is0.21,notably large given that we are regressing change in LNPR_C on the change in the regressors.Thus,the results of Eq.(4)suggest a positive and significant relationship between LNPR_C and AUDIT,and provide strong evidence in support of H1and H2.

10It should be noted that OLS with the full set of indicator variables for each firm is equivalent to the OLS on the first-differences of the variables when there are only two observations(two years)per firm.