Advances in Accounting, incorporating Advances in International Accounting 29 (2013) 108–123

Contents lists available at SciVerse ScienceDirect

Advances in Accounting, incorporating Advances inInternational Accounting

j ourna l homepage: www.e lsev ie r .com/ locate /ad iac

The impact of IFRS on accounting quality: Evidence from Greece

Panagiotis E. Dimitropoulos a,⁎, Dimitrios Asteriou b, Dimitrios Kousenidis c, Stergios Leventis d,1

a University of Peloponnese, Department of Sport Management, Orthias Artemidos and Plataion Str., P.C. 23100, Sparta, Greeceb Hellenic Open University, School of Social Sciences, Bouboulinas 57–59 Str., 26222 Patras, Greecec Aristotle University of Thessaloniki, Department of Economics, University Campus, P.C. 54124, Thessaloniki, Greeced International Hellenic University, School of Economics and Business Administration, 14th klm Thessaloniki-Moudania, 57101 Thessaloniki, Greece

⁎ Corresponding author. Tel.: +30 2731089669.E-mail addresses: dimitrop@uop.gr (P.E. Dimitropou

(D. Asteriou), dkous@econ.auth.gr (D. Kousenidis), s.lev1 Tel.: +30 2310 807 541.

0882-6110/$ – see front matter © 2013 Elsevier Ltd. Allhttp://dx.doi.org/10.1016/j.adiac.2013.03.004

a b s t r a c t

a r t i c l e i n f o

Keywords:

IFRSValue RelevanceAsymmetric TimelinessEarnings ManagementGreece

This paper examines the impact of IFRS adoption on the quality of accounting information within the Greekaccounting setting. Using a sample of 101 firms listed in the Athens Stock Exchange (ASE) for a period of eightyears (2001–2008) we find convincing evidence that the implementation of IFRS contributed to less earningsmanagement, more timely loss recognition and greater value relevance of accounting figures, compared tothe local accounting standards. Also, our findings document that audit quality further complements thebeneficial impact of IFRS since those companies that are audited by Big-5 audit firms exhibit higher levelsof accounting quality. Our findings are robust in regard to different model specifications and after controllingfor firm-specific effects like size, risk, profitability and growth opportunities.

© 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Effective 2005, all listed companies in the Athens Stock Exchange(ASE) are required to comply with International Financial ReportingStandards (hereafter IFRS). The research questions we address inthe present paper are: first, whether IFRS have mitigated earningsmanagement behavior in Greek listed firms; second, whether IFRS in-fluence the relevance of accounting information (earnings and bookvalues) compared to the Greek accounting standards and third, wetest whether the adoption of IFRS impacts on the level of reportingconservatism in financial statements.

The primary motivation for this study is to further our understand-ing of the influence of IFRS on the quality of reported earnings. Overall,IFRS are principle-based standards that are market-oriented andrequire extensive disclosure in comparison with prior standards,i.e. local generally accepted accounting principles (GAAP). The Interna-tional Accounting Standards Board (IASB) removed allowable account-ing alternatives (which existed in most countries under theirrespective GAAPs) and required accounting measurements that betterreflect a firm's economic position and performance. Based on this, intheory, the adoption of IFRS should significantly restrict the ability toengage in earnings management behavior (upwards or downwards)and increase the overall quality of disclosed information. However,

los), d.a.asteriou@eap.grentis@ihu.edu.gr (S. Leventis).

rights reserved.

Barth, Landsman, and Lang (2008) also note that the opposite may betrue. For example, limiting managerial discretion relating to accountingalternatives could eliminate a firm's ability to report accounting mea-surements that are more reflective of the firm's economic positionand performance. They also note that if enforcement of these standardsis lax, companies could still engage in earnings management.

In the current literature there are two opposing views regarding theinfluence of IFRS on accounting quality. On the one side, researchersargue that IFRS improve the reliability of financial reporting by limitingopportunisticmanagerial discretion (Ashbaugh & Pincus, 2001; Barth etal., 2008; Ewert & Wagenhofer, 2005). The contrary view is that theflexibility inherent in IFRS and lax enforcement might provide greateropportunity for firms to manage earnings (Ball, Robin, & Wu, 2003;Breeden, 1994; Burgsthaler, Hail, & Leuz, 2006; Cairns, 1999; Street &Gray, 2002). Recent evidence by Ahmed, Neel, and Wang (2010)suggests that IFRS adoption results in more earnings smoothing, moreaggressive reporting of accruals and a reduction in timeliness of lossrecognition relative to gain recognition. They attribute this finding totheminimal implementation guidance on IFRS, to managers' incentivesto manipulate earnings and to weak enforcement mechanisms.

Similar to conflicting views on the influence of IFRS on accountingquality, as noted above, research on the influence of IFRS on earningsmanagement has also provided mixed results. Van Tendeloo andVanstrelen (2005) for example, found that earnings management be-havior was not significantly different between German companiesthat adopted IFRS when compared to those that relied on GermanGAAP. However, these findings are not corroborated by Barth et al.(2008) who, using a sample of 21 countries, concluded that in thepost-adoption period firms applying IFRS displayed significantly less

109P.E. Dimitropoulos et al. / Advances in Accounting, incorporating Advances in International Accounting 29 (2013) 108–123

earnings management relative to the pre-adoption period. Also Chen,Tang, Jiang, and Lin (2010) document that since IFRS adoption thequality of accounting information has increased within the EU interms of less targeted earnings management, smaller magnitude of ab-solute discretionary accruals and higher accruals quality. Similar evi-dence, documented by Iatridis (2008, 2010) in the UK and furtherverified by Doukakis (2010) in Greece, suggests that IFRS has impactedon the variability and persistence of earnings and a reduction of incomesmoothing practices, which obviously results in higher reportingquality.

Part of the motivation for this study is to shed further light on thesecompeting views with regard to the efficiency of IFRS in reducingearnings management and improving earnings quality within anaccounting framework characterized by a long history of historical-accounting principles. According to Tsalavoutas and Evans (2010), theGreek accounting setting is heavily based on state regulation and oncertain rules associated with an increased monitoring cost. There is aclose link between taxation and accounting rules which, in turn,creates increased incentives for creative accounting and earnings man-agement practices. Ding, Hope, Jeanjean, and Stolowy (2007) point outthat Greece is the country with the highest divergence of issues be-tween its local GAAP and IFRS and that this divergence is closely relatedto the distinctive culture of this country. Hofstede (1980) argues thatGreek accountants prefer to reduce uncertainty by abiding by therules and Ballas, Skoutela, and Tzovas (2010) and Tsakumis (2007)also document that Greek accounting professionals are “characterizedby the acceptance of inequality and a preference for achievement andmaterial success, focusing on the group which they belong ratherthan on themselves”. Within this cultural framework the introductionof IFRS, which are principle-based standards, is expected to create abreach of this cultural status quo and thus the impact of this newregulation on the quality of published information remains an openempirical question.

In addition, Greece presents a unique economic environment whichdiffers significantly from other code-law countries. Greece is a bank-oriented capital market, a fact which aligns corporate reporting to-wards the protection of the creditors, while managers pay lessattention to minority shareholders (Ballas et al., 2010). Contrarily,IFRS are accounting standards which focus more on the needs of inves-tors rather than the creditors. Therefore, considering the countervailingcultural and economic environment of Greece, the adoption of IFRS bythe Greek government introduced several provisions which shake theestablished accounting practices of the last 25 years, such as increasedlevels of disclosure, the issues covered by accounting rules and thebroadened role of corporate boards in evaluating accounting values.Whether the adoption of IFRS improved accounting quality withinthis distinctive environment remains an issue which warrants furtherinvestigation.

Furthermore, the quality of accounting information is closely re-lated to the quality of the external audit. Tsalavoutas and Evans(2010) document that large multinational audit firms (Big-5) enjoygreater independence from their clients in Greece and have greaterexperience and resources for implementing IFRS, and of course facegreater reputation costs or litigation risks. These characteristics con-tribute towards higher quality audits. Caramanis and Lennox (2008)documented that within a weak legal enforcement environment(such as in Greece), the auditor's effort to ensure accounting qualityis motivated by reputation costs and since Big-5 audit firms facegreater litigation costs they also produce higher quality audits. Fur-thermore, a recent study by Tsalavoutas (2011) suggested that Big-5audit corporations act as monitoring mechanisms contributing to in-creased transparency, better quality financial statements and higherlevels of compliance with IFRS. Thus, our study adds to a growingbody of literature since it examines if and how the quality of reportedaccounting information is influenced in any way by the implementa-tion of IFRS, taking into consideration the quality of audit. We provide

additional evidence to the ongoing debate about international ac-counting standards and accounting quality.

Our sample comprises of 101 firms listed in the ASE covering aneight-year period from 2001 until 2008 where 25 had voluntarilyadopted IFRS prior to the initial enforcement in 2005 and theremaining 76 adopted IFRS when mandatorily enforced. We examinewhether accounting information during the post-IFRS period (2005–2008) exhibits less earnings management, more timely loss recognitionand higher value relevance, when compared to the relative amountsduring the pre-IFRS period (2001–2004). After controlling for firm-specific characteristics such as size, growth opportunities, risk andaudit quality we found that the adoption of IFRS contributed to lessearnings manipulation (lower magnitude of discretionary accruals),more timely loss recognition and greater value relevance of earningsand book values, compared to the local accounting standards. Also,our findings document that audit quality further complements the ben-eficial impact of IFRS since those companies that are audited by theBig-5 multinational audit firms exhibit higher levels of accountingquality compared to their non-Big-5 counterparts.

2. Greek accounting standard setting and IFRS

An extensive amount of research has examined the value rele-vance of accounting information within countries under differentaccounting regimes. Hope (2003) argues that the quality of financialaccounting information depends on both the quality of accountingstandards and the regulatory enforcement, since even the best ac-counting standards are inefficient if not enforced adequately.

The Greek accounting system dates back to 1835 when the first in-dependent Greek Republic was established after almost 400 years ofOttoman occupation. At that time the French commercial code wastranslated into Greek and adopted in order to serve as the basis ofGreek commercial law. The companies act law was introduced in1920 and modified in 1955 and it still exists now to a great extent.Accounting regulation in Greece was imposed by the governmentand specifically by the Minister of Economics and Finance. However,in 1988 the National Accounting Council was established in order tobe entrusted with the issuance of more detailed regulations. In con-trast to other countries, the Greek National Accounting Council hadlimited decision making authority and its function was clearly adviso-ry to the government. Before the council could express an opinion ithad to follow a process of consultation with the Economic Chamberof Greece which could be very time-consuming (Ballas, 1994).

In June 2003, the National Accounting Council was abolished and inits place the government established the Board of AccountingStandards which forms one division of the Committee of AccountingStandards and Control (along with the Board of Quality Control andthe Executive Committee). The Board of Accounting Standards isgoverned by a five-member board of directors and its main responsibil-ity is to provide consultative services on issues of accounting standard-ization regarding Greek accounting standards and the codification ofIFRS. Also, it has the authority to publish directives regarding the im-plementation of Greek GAAP and IFRS and also clarifications on sometechnical accounting issues. These advisory decisions need to be ap-proved by the Committee of Accounting Standards and Control beforethey are proposed to the Ministry of Economics for final approval(Ballas et al., 2010; Dimitropoulos & Asteriou, 2009).

Additionally, before 1991 the ASE was an unregulated stock marketand the only supervision performed was once again by the Ministry ofEconomics. As a result, until that period the accounting regulationspromoted financial disclosure for tax purposes only, since the stockmarket was unregulated and investment interest was quite limited.Furthermore, the overall setting was primarily used as an instrumentof government policy — most notably tax policy, but also wage andpricing policy (Ballas et al., 2010).

110 P.E. Dimitropoulos et al. / Advances in Accounting, incorporating Advances in International Accounting 29 (2013) 108–123

However, since 1991 important accounting and financial reformshave taken place on many levels of financial regulation. First, theGreek government established the Hellenic Capital Market Commissionwhich was entrusted with the governing of the stock market. In doingso the government wanted to establish a sound regulatory frameworkfor the evolvement of the stock exchange as an integral part of eco-nomic growth. Additionally, in 1992 the Code of Books and Recordswas enacted, determining the financial statements which firms mustreport and the specific details of accounts that must be disclosed inthe financial statements. Finally, in 1993 the audit profession in Greecewas liberalized by allowing foreign audit companies to operate withinthe country, something which was forbidden prior to 1993. This facthelped the evolution of the accounting profession and the stability ofthe capital market since those firms increased the creditworthiness ofthe stock market by providing accurate and high-quality audits. AsBallas and Fafaliou (2008) argue, the concentration of the audit marketin Greece shows a significant increase, especially after 2001 where theBig-5 multinational audit corporations are equally represented andcontrol almost 67% of statutory audits within the capital market. Thesame phenomenon exists in Germany (though the relevant concentra-tion ratio is less than the Greek context) and it is attributable to theregulatory reform that took place in Germany as it conformed to theIFRS, causing the market share of local audit firms to decrease, a factexpected to affect all European markets from the year 2005 onwards(Tsalavoutas & Evans, 2010).

As Spathis and Georgakopoulou (2007) and Spathis, Doumpos,and Zopounidis (2003) argue, the aforementioned Greek accountingsetting is based on the General Greek Accounting Plan (GGAP)which emphasizes financial reporting conformity with tax rules, theprotection of stakeholders, and conservatism. The Greek accountingsystem has been highly conservative due to the fact that the settingof accounting standards remains the sole responsibility of the state(since the Board of Accounting Standards plays only an advisoryrole in the standardization process and the final decision and approv-al lie within the Ministry of Economics). This has, and in part due tothe large number of small and family-owned businesses, limited thegoal of financial reporting to provide only tax-relevant information(Kousenidis, Ladas, & Negakis, 2009). Greek firms obviously havestrong incentives to reduce taxes and, consequently, financial ac-counting information is less likely to reflect economic realities whenthe firm's goal is to minimize taxes. This assertion has been verifiedby Guenther and Young (2000) who argue that, in countries wherethere is increased conformity between financial accounting and taxaccounting rules, accounting information may differ from underlyingeconomic activities. Additionally, certain rules imposed by the Greekaccounting standards give managers greater flexibility than IFRS forexecuting income smoothing techniques such as the capitalizationof start-up costs and their amortization within a five year period.

The most significant change that has taken place in the Greek ac-counting setting is the adoption of International Financial ReportingStandards (IFRS) for all listed firms since January 2005 which is charac-terized as the second cornerstone of accounting legislation in Greeceafter the adoption of the Hellenic General Accounting Plan (HGAP) in1980. This new framework impacted on the stakeholder-oriented andtax-driven nature of the abovementioned accounting setting. Themost important implication of IFRS was the introduction of the fairvalue principle (replacing the long-established historical-cost principle)to asset valuation and liability recognition. The rationale behind the fairvalue principle is that the disclosed values of accounting quantitiesshould be close to their relativemarket values, thusfinancial accountinginformation is more likely to reflect true economic events. IFRS also re-duced managers' flexibility in valuing assets at the lowest amount pos-sible so as to minimize tax liabilities. Moreover, IFRS imposed changeson many technical accounting issues such as the presentation of finan-cial statements, segment reporting, intangible assets, depreciation, re-lated party disclosures etc. (see Spathis & Georgakopoulou, 2007 for

details), aimed at promoting a ‘true and fair’presentation offinancial in-formation to facilitate investors' rational investment decisions.

IFRS are considered to be high-quality standards and their imple-mentation in countries with accounting settings based onhistorical-cost values (such as code-law countries) could help to im-prove the quality of financial accounting information (Hitz, 2007;Penman, 2007). Whether this assertion is true within the Greek ac-counting setting remains an open empirical question.

3. Literature review and hypothesis development

As discussed in the previous section, the role of the IASB is to de-velop accounting standards that are internationally acceptable and ofa high quality. In order to achieve this goal, the IASB has issuedprinciple-based standards and removed allowable alternatives in ac-counting estimations, thus leaving little room for managerial discre-tion by requiring accounting measurements to better reflect acompany's economic position and performance. However, thereseems to be a long debate over whether IFRS are able to achieve theaforementioned goal and to what extent. Several studies in the fieldhave focused on different aspects of accounting quality, namelyvalue relevance, conditional conservatism and earnings management.

Referring to value relevance (or how efficiently publicly availableaccounting information is incorporated into stock returns), Ashbaughand Pincus (2001), Ewert and Wagenhofer (2005), Lang, Raedy, andYetman (2003); Lang, Raedy, and Wilson (2006), Leuz, Nanda, andWysocki (2003) and Bartov, Goldberg, and Kim (2005) provide evi-dence that reported earnings based on IFRS are more value-relevantthan domestic GAAP. Additionally, Djatej, Gao, Sarikas, and Senteney(2009) document that the implementation of IFRS increases the qualityof public information and that this increase is more pronounced forEast European countries. This suggests that IFRS reduce informationasymmetry by moving important performance-related informationfrom the private into the public domain. In the same vein, Beuselincket al. (2009), Devalle, Onali, and Magarini (2010), Aharony, Barniv,and Falk (2010) and Landsman, Maydew, and Thornock (2010) corrob-orate the above findings by providing evidence of an improvement instock price informativeness in the EU after the adoption of IFRS.Moreover, Dargenidou and McLeay (2010) support the above findingsby demonstrating that mandatory IFRS adoption results in accountingestimates that reflect the underlying economics of firms within finan-cial markets in a more timely and comparable manner. Also Barth etal. (2008) verify this hypothesis by examining the value relevance ofincome and the book value of equity between firms following IFRSand local GAAP. Their findings indicate a higher association of account-ing figures with stock returns for those firms that adopted IFRS.

Another way of assessing the quality of accounting information isto examine to what extent firms withhold the recognition of losseswith the intention to mislead stakeholders about the true economicperformance of the company, or to affect contractual outcomes thatare tied to accounting earnings. Referring to the issue of timely lossrecognition Ball, Kothari, and Robin (2000), Lang et al. (2003, 2006),Leuz et al. (2003), Ball and Shivakumar (2005, 2006) and Conover,Miller, and Szakmary (2008) suggest that one characteristic ofhigher-quality earnings is that large losses are recognized as theyoccur instead of being deferred to future periods. Barth et al. (2008)using a sample from 21 countries found that firms operating underIFRS exhibit more timely recognition of losses compared to the non-IFRS firms. Also, Guenther, Gegenfurtner, Kaserer, and Achleitner(2009) document that conditional conservatism increased after theIFRS implementation for both mandatory and voluntary adopters inGermany.

Moreover, a number of studies suggest that the quality of reportedfinancial statements is determined by the economic and institutionalfactors influencing manager and auditor incentives. Leuz et al. (2003)document that earnings management (measuring the magnitude of

111P.E. Dimitropoulos et al. / Advances in Accounting, incorporating Advances in International Accounting 29 (2013) 108–123

discretionary accruals) is more prevalent in code-law countries com-pared to common-law countries. This difference is attributed to the rel-ative costs of earnings management being less than the benefits. IFRSpromote an accounting environment within which firms have incen-tives to report investor-oriented information and thus engage signifi-cantly less in earnings management (compared to the local GAAP).Barth et al. (2008) verify this assertion by finding evidence of less earn-ings management for firms adopting IFRS compared to their non-IFRScounterparts. A characteristic of the aforementioned studies is thatthey use discretionary accruals and earnings benchmarks as measuresof earningsmanipulation. Therefore,we expect that Greek firms operat-ing during the post-IFRS period (2005–2008) will publish financialstatements which will present less earnings management comparedto the pre-IFRS era.

We also assume that firms engaging in earnings smoothing willpresent greater values of the ratio of standard deviation (SD) of in-come to the standard deviation (SD) of cash flows (Ahmed et al.,2010). Ahmed et al. (2010) declare that if managers use accruals tosmooth changes in cash flows when reporting income, the varianceof the change in net income should be less than the variance of thechange in cash flows. Accordingly, less positive values of this ratiowill indicate greater income smoothing. Thus, firms operating in thepost-IFRS era might present greater values of the ratio of SD of in-come to the SD of cash flows. Also Ding et al. (2007) document thatcountries with low quality accounting standards (such as Greece)are associated with greater opportunities for earnings managementand they suggest that the expansion of coverage of accounting issuesby IFRS can curb this behavior.

However, there are a series of research papers which provide exact-ly the opposite picture regarding the efficiency of IFRS on increasing ac-counting quality. Van Tendeloo and Vanstrelen (2005), find nodifferences in earnings management behavior between firms applyingIFRS and local GAAP in Germany. Also Daske (2006), and Hung andSubramanyam (2007) found no evidence of a reduction in the cost ofcapital and an increase in accounting value relevance in amounts pre-pared under IFRS as opposed to German GAAP. Additionally, Eccherand Healy (2003) and Callao, Jarne, and Laínez (2007) argue thataccounting figures based on IFRS are not more value-relevant whencompared to relative amounts prepared under Chinese and SpanishGAAP respectively. Finally, relevant research on the emerging capitalmarkets of Kazakhstan and Egypt reported by Tyrrall, Woodward, andRakhimbekova (2007) and Hassan, Romilly, Giorgioni, and Power(2009) respectively found evidence of a weak relation between IFRSadoption and accounting value relevance as well as a negative relation-ship between IFRS and firm value. Furthermore, Callao Gastón et al.(2010) examined the impact of mandatory IFRS adoption on financialreporting in the UK and Spain and they found a reduction in the rele-vance of financial reporting in both countries. Overall, there exist twogroups of studies with conflicting evidence regarding the ability ofIFRS to enhance accounting quality.

Barth et al. (2008) argue that the main reasons for the above-mentioned conflicting findings are the gradual transition from localGAAP to IFRS; the lack of infrastructure to enforce the application ofIFRS; differences in the effectiveness of controls regarding the econom-ic environment and corporate incentives; and, finally, differential timeperiods and data metrics. Our study tries to mitigate these problems byexamining the impact of IFRS on accounting quality within a period(2005–2008) of obligatory IFRS enforcement (not voluntary as inmany prior studies), while simultaneously controlling for voluntaryadopters; and also a period when the transition to IFRS from localGAAP has been made thoroughly, since Greek firms were obliged toreport their 2004 financial statements in both local GAAP and IFRS for-mats in order to give investors a visualization of the upcoming account-ing changes. Moreover, our study is focused on a single code-lawcountry. This fact has the advantage of removing the need to controlfor the potentially confounding effects of country-specific factors that

are not related to the accounting system. Also, as Hitz (2007) andPenman (2007) argue, the implementation of IFRS on countries withaccounting settings based on historical-cost values (like Greece)could help to improve the quality of financial accounting information.Finally, two recent studies conducted on the Greek capital marketreport that earnings value relevance has increased and earningsmanagement decreased after IFRS adoption; and that, overall, IFRS in-creased the reliability, transparency and comparability of the financialstatements (Ballas et al., 2010; Iatridis & Rouvolis, 2010). However, asyet no Greek study has addressed the issue of audit quality, which isa crucial element of quality financial reporting, through the efficientimplementation of IFRS (Hodgdon, Tondkar, Adhikari, & Harless,2009). Therefore we expect that, after controlling for the aforemen-tioned problems, the adoption of IFRS within the Greek accounting set-ting has helped to enhance the quality of accounting information. AsBarth et al. (2008) argue, the advantages of this procedure are firstthat we can determine the source of accounting quality differences be-tween the two sample periods before and after the implementation ofIFRS. Second, we can rule out predictions for some of our metrics basedon predictions from another metric, thus having a clearer picture of theimpact of IFRS on accounting quality. Enhanced accounting quality willbe translated as an increase in value relevance and accounting conser-vatism and a reduction of earnings management behavior during thepost-IFRS period. Thus, our first hypothesis states:

H1. Financial statements prepared under IFRS are of a higher qualitythan those prepared under the Greek GAAP.

Additionally, as previously discussed, the quality of accounting in-formation is heavily dependent on the quality of the external audit.On this issue, Teoh and Wong (1993), Gul, Sun, and Tsui (2003); Gul,Tsui, and Dhaliwal (2006), Ghosh and Moon (2005) and Park andPincus (2001) predict and find convincing evidence that EarningsResponse Coefficients (ERCs) are a positive function of perceived audi-tor quality. This practically means that firms audited by the Big-5 mul-tinational auditing corporations report earnings and book values ofcommon equity that have higher value relevance (higher associationwith stock returns). Moreover, multinational audit firms have more re-sources (knowledge and specialized personnel) and IFRS-related expe-rience and could provide greater assistance in the implementation andtransition to IFRS compared to the local audit firms. Also, recent evi-dence by Hodgdon et al. (2009) and Tsalavoutas (2011) indicate thatIFRS compliance is positively associated with audit quality (defined asbeing audited by Big-5 audit corporations). Their evidence reinforcesthe importance of multinational audit firms and quality audits as ameans of encouraging compliance with IFRS. A recent study byIatridis (2011) in the UK indicated that firms audited by the Big-5audit corporations are associated with high quality accounting disclo-sures and are less prone to earnings manipulation. Along the sameline, Tsalavoutas and Evans (2010) declared that IFRS have improvedthe quality of accounting information in the Greek capital marketsince they report a statistically significant association between auditquality and the transition to IFRS by Greek listed firms, suggestingthat Big-5 audit corporations were able to attract experienced em-ployees from their foreign operations to assist in the transition and im-plementation of IFRS. Thus, we expect that companies audited by Big-5audit firms will present a higher value relevance of earnings and bookvalues during the post-IFRS period compared to their non-Big-5audited counterparts.

Furthermore, firms audited by the Big-5 audit corporations willpresent more timely loss recognition since multinational audit firmsare more likely to follow a prudent policy towards loss-recognitionto avoid any future reversals on profitability that could be a seriousstrike to their reputation. Therefore, multinational audit corporationshave greater incentives for recognizing losses as they occur instead ofbeing deferred to future periods. Also, a series of studies has shown

Table 1Sample data selection procedure and sample description.

Panel A: Sample selection procedure

Firms listed in the Athens Stock Exchange (ASE)Less: Financial services firms

294(43)

Remaining non-financial firmsLess: Firms with incomplete accounting and stock price data

251(125)

Non-financial firms with full accounting and stock price dataLess: Non-December fiscal year end firms

126(25)

Firms included in the final sample 101

Panel B: Year and industry distribution

Sample distribution byyear

Sample distribution by industry sector

Year Observations Industries Observations

2001 101 Raw materials 402002 101 Constructions 1082003 101 Industrial products and services 1082004 101 Food and beverages 962005 101 Personal and household products 842006 101 Health corporations 402007 101 Trade corporations 722008 101 Media corporations 40

Tourism and leisure 48Telecommunications and technology 76Services 64Real estate 32

Total 808 Total 808

112 P.E. Dimitropoulos et al. / Advances in Accounting, incorporating Advances in International Accounting 29 (2013) 108–123

that audit quality (if the company is audited by one of the Big-5 auditcorporations) constitutes a constraint on earnings managementreporting (Bauwhede, Willekens, & Gaeremynck, 2003; Francis &Wang, 2003; Gore, Pope, & Singh, 2001; Iatridis, 2011). Street andGray (2001, 2002), Glaum and Street (2003) and Prather-Kinseyand Meek (2004) all find support for the hypothesis that companiesaudited by a large audit firm are positively associated with IFRS com-pliance. Caramanis and Lennox (2008) demonstrate that Big-5 auditfirms work more hours compared to non-Big-5 firms in the Greekcapital market, and companies audited by Big-5 audit firms report asmaller magnitude of discretionary accruals, thus engaging in lessearnings management in comparison to companies audited by non-Big-5 corporations. Additionally, Leventis and Caramanis (2005) pro-vide evidence that audit effort in Greece is correlated with audit firmsize. Also Frankel, Johnson, and Nelson (2002) document that discre-tionary accruals can be viewed as evidence of audit independencesince independent auditors require their clients to report unmanagedfinancial statements. Therefore, we assume that Big-5 audit firmsprovide better quality audits than local audit firms, meaning thatthey have greater abilities and incentives to enforce more transparentfinancial reporting after the adoption of high-quality accounting stan-dards like IFRS. Thus, based on the above discussion we formulate oursecond hypothesis as follows:

H2. Accounting quality will be higher in the post-IFRS period forfirms with enhanced audit quality.

4. Data and methodology

4.1. Data selection procedure

Our study uses a sample consisting of 101 companies, all listed onthe ASE, where 76 followed the mandatory adoption of IFRS in 2005and the other 25 were early voluntary adopters. All companies hadfull annual data of reported earnings and stock prices, sales, assets,total debt, property-plant and equipment, cash flows and common eq-uity during the period 2001–2008. We have excluded banks and finan-cial services firms due to the different operations and accountingpolicies. The classification of similar firms was made according to theindustry classification proposed by the ASE. Data were collected fromthe ASE database. The initial sample contained 126 companies withfull data for the period under investigation but we restricted thegroup to companies with a December fiscal year-end limiting thefinal number to 101, resulting in 808 firm-year observations. Annualstock prices exclusive of dividends at the end of each year and annualearnings per share were used. The earnings and price data were adjust-ed for stock splits, stock dividends and stock issues. In order to elimi-nate the effect of outliers, we deleted the top and bottom 1% of allvariables during the pre and post-IFRS periods. All variables in the sam-ple were deflated by the year's average total assets (measured as themean value of the assets at the beginning and at the end of the fiscalyear). Table 1 describes the data selection procedure and the synthesisof the final sample. The sample is quite dispersed with the majority ofthe observations situated within the construction and manufacturingsectors, followed by personal products and telecommunications.

4.2. Testing for value relevance

Our approach in this section included testing whether the relationbetween earnings per share and the book value of equity per sharewith stock prices varies between local accounting principles andIFRS. Our value relevance metric is based on the explanatory powerof a regression of stock prices on net income and the book value ofequity. In order to obtain a measure of stock price that is unaffectedby mean differences across industries (which would affect the explan-atory power of the comparisons) we applied the methodology

proposed by Barth et al. (2008) and regressed stock price P on industryfixed effects. The next action was to model the residuals from the pre-vious step P* as the dependent variable on book value of equity pershare (BVPS) and net income per share (NIPS) separately for the preand post-IFRS periods. In order to ensure that accounting informationhas been fully disclosed to the market we followed Lang et al. (2003,2006) and estimated P six months after the fiscal year end. In thenext model, and all the following, we pooled both mandatory andvoluntary IFRS adopters in order to control for incremental differencesbetween the two groups of firms, so as to extract more salient infer-ences regarding the impact of IFRS on accounting quality (Byard, Li, &Yu, 2011). The general form of the regression model is as follows:

P�it¼ α0þα1NIPSitþα2BVPSitþα3DVOLitþα4DVOLit�NIPSitþα5DVOLit

�BVPSitþβControlit þ γYeardummies

þδIndustrydummiesþ εit

ð1Þ

where:

P* is the residuals from a regression of stock prices on industryfixed effects,

NIPS is net income per share,BVPS is the book value of common equity per share,DVOL is a dummy receiving one (1) for firms that voluntarily

adopted IFRS prior to 2005 and zero (0) otherwise.

We also included some control variables capturing size, risk, growthand profitability and finally we took into consideration the year andindustry effects of any omitted variables. The industry dummies followthe ASE industry classification code and we classified firms in our sam-ple according to the sectors depicted in Table 1. Model 1 is run sepa-rately for the pre-IFRS (2001–2004) and post-IFRS (2005–2008)periods. In the estimation of all models we applied the White (1980)test in order to control for heteroscedasticity in the error terms. Ifhypothesis H1 is valid, net income and the book value of equity willbe of higher value relevance in the post-IFRS period, thus we expectan increase in R2 during the post-IFRS period compared to the

113P.E. Dimitropoulos et al. / Advances in Accounting, incorporating Advances in International Accounting 29 (2013) 108–123

pre-IFRS period. Also, voluntary IFRS adopters are expected to havebenefited from the early adoption of IFRS, thus we expect their ac-counting numbers to be more value relevant during the pre-IFRS peri-od (2001–2004) compared to the mandatory group.

Additionally, since the question in this research is whether thequality of accounting information has improved after the adoptionof IFRS, the two accounting standards (Greek GAAP and IFRS) areset up as competing non-nested models. For this reason we followedDechow (1994) and controlled for the change in R2 between the preand post-IFRS periods by employing the Vuong (1989) likelihoodratio test. The test examines the null hypothesis that the twosub-periods are equally capable of explaining the true data againstthe alternative that one period is closer.2 According to Dechow(1994), the Z-statistic has several advantages over other non-nestedtests since it assumes that under the null neither model is true. So,if we set as model A the post-IFRS period and model B the pre-IFRSperiod a positive and significant Z-statistic suggests that the residualsproduced by the regression during the pre-IFRS period are larger inmagnitude than those during the post-IFRS period. Consequently, apositive and significant Z-statistic indicates that IFRS adoption hascontributed to increased accounting quality.

Moreover, in order to control for hypothesis H2 we introduced adummy variable AUD which takes the value of one (1) if the firm isaudited by a Big-5 international audit firm in a given year and zero(0) otherwise. By this means we examine whether companiesaudited by Big-5 audit firms present a higher value relevance of earn-ings and book values during the post-IFRS period compared to theirnon-Big-5 audited counterparts. Therefore, model 1 is modified tothe following form:

P�it¼ α0þα1NIPSitþα2BVPSitþα3DVOLitþα4DVOLit�NIPSit

þα5DVOLit�BVPSitþa6AUDitþα7AUDit

�NIPSitþα8AUDit�BVPSit

þβControlit þ γYeardummies þ δIndustrydummiesþ εit ð2Þ

where:

AUD*BVPS and AUD*NIPS are the interaction terms between AUDdummy and book value per share and net income per sharerespectively, and

DVOL*NIPS and DVOL*BVPS are the interaction terms of NIPS andBVPS with the DVOL variable.

As in the previous model 1, model 2 is estimated separately for thepre and post-IFRS periods. H2 is accepted if the explanatory power ofmodel 2 in the post-IFRS period is higher compared to model 2 in thepre-IFRS period and also if coefficients a4 and a5 are significant andhave higher magnitude relative to coefficients a1 and a2 (again in thepost-IFRS period).

4.3. Testing for timely loss recognition

Our next test includes the examination of the link between timelyloss recognition and IFRS adoption. In order to determine whetherthe implementation of IFRS leads firms to disclose the recognition oflosses in the period under study we applied Basu's (1997) reverse re-gression model between earnings per share and stock returns. Forthis purpose we estimated 12-month stock returns over the fiscalyear in order to proxy for the news regarding a firm's performancethat is publicly available. Basu's (1997) introduces a dummy variable(D) in the reverse regression model based on the assumption thatearnings are expected to be more highly correlated with market

2 We are grateful to Jackson Caskey (UCLA) for programming this test on STATA andmaking it available for use (http://personal.anderson.ucla.edu/judson.caskey/data.html).

returns in periods of depression in market values than in periods ofprosperity in market values. This qualitative variable is given thevalue of one (1) when returns are negative and zero (0) if they are pos-itive. Under these assumptions the final form of the Basu's (1997)model is as following:

EYi;t¼ b0þb1�Ritþb2

�Ditþb3�Rit

�Ditþb4DVOLitþb5Dit�DVOLit

þb6Rit�DVOLitþb7Rit

�Dit�DVOLitþβControlit þ γYeardummies

þδIndustrydummiesþ eit ð3Þ

where:

EY is the earnings yield estimated as earnings per share deflatedby the stock price at the beginning of the fiscal year,

R is the buy and hold stock return over the fiscal year,D is a dichotomous variable receiving one (1) when R is nega-

tive and zero (0) otherwise and the other variables are de-fined as previously stated.

We have included control variables for size, risk, growth and prof-itability as well as year and industry dummies.

The coefficients on R (b1) and R*D (b3) capture the reaction of earn-ings to contemporaneous ‘good news’ and ‘bad news’ respectively.Contextually, a positive b3 is an indication of accounting conservatism.Since conservative reporting delays recognition of ‘good news’, thelagged effect may appear as a persistent shock. Therefore, the intercept,b0, is expected to have a positive sign. In other words, if conservatismaffects earnings downwardly the intercept is expected to have apositive sign. The coefficient on D, b2, is explained as a reversal of theprevious year's market information in the light of current value de-creases. When b2 is positive, over-provisioning is reversed as a prioryear adjustment and, when b2 is negative, deferred income recognitionis scaled down (Raonic, McLeay, & Asimakopoulos, 2004). Model 3 isrun separately for the pre and post-IFRS periods and if IFRS have in-creased the timely recognition of losses we expect a significant increasein the explanatory power of the model (which will be tested usingVuong's (1989) likelihood ratio test) and coefficient b3 to be positiveand have a higher magnitude relative to the pre-IFRS period. Also, coef-ficient b7 is expected to be positive and significant during the pre-IFRSperiod and to have a larger magnitude compared to coefficient b3 indi-cating that voluntary IFRS adopters are more conservative during thepre-IFRS period relative to firms following the Greek GAAP.

Moreover, we have extended the Basu (1997) model for examin-ing the effect of audit quality on timely loss recognition, by introduc-ing a dummy variable AUD into the model which takes the value ofone (1) if the firm is audited by a Big-5 audit firm and zero (0) other-wise. The new variable was then interacted with the other variablesin the standard Basu (1997) model as depicted in the following equa-tion:

EYit¼ b0þb1Ritþb2Ditþb3Rit�Ditþb4DVOLitþb5Dit

�DVOLitþb6Rit

�DVOLitþb7Rit�Dit

�DVOLitþb8AUDitþb9Dit�AUDit

þb10Rit�AUDitþb11Rit

�Dit�AUDitþβControlit þ γYeardummies

þδIndustrydummiesþ eit: ð4Þ

Model 4 is estimated again separately for the pre and post-IFRSperiods. Hypothesis H2 expects that firms audited by the Big-5 auditcorporations will present more timely loss recognition since multina-tional audit firms are more likely to follow a prudent policy towardsloss recognition considering that they have greater incentives. If H2is true we expect this to be represented by a significant positive coef-ficient on the R*D*AUD interaction term.

114 P.E. Dimitropoulos et al. / Advances in Accounting, incorporating Advances in International Accounting 29 (2013) 108–123

4.4. Testing for earnings management

Our first test on earnings management is based on the ratio of thestandard deviation of the change in net income (ΔΝΙ) to the standarddeviation of the change in cash flows (ΔCF). We performed this testfor both sub-groups of firms based on the voluntary or mandatoryadoption of IFRS. In order to capture any confounding effects by fac-tors that are not attributable to the financial reporting setting, wefollowed Barth et al. (2008) and Ahmed et al. (2010). Thus, we com-pared the ratio of the standard deviation of the change in net income(ΔΝΙ*) to the standard deviation of the change in cash flow (ΔCF*) re-siduals from the following two regression models, instead of compar-ing the SD of the change of NI and CF directly. The relative modelshave the following form:

ΔΝΙit¼ a0þa1SIZEitþa2GROWTHitþa3LEVitþa4AUDitþeit ð5Þ

ΔCFit¼ a0þa1SIZEitþa2GROWTHitþa3LEVitþa4AUDitþeit ð6Þ

where:

ΔΝΙ is the annual change in net income divided by lagged totalassets

ΔCF is the annual change in the operating cash flow divided bylagged total assets

SIZE is the natural logarithm of end year total assetsGROWTH is the percentage change in salesLEV is the ratio of end year total liabilities to end year total com-

mon equityAUD is a dummy receiving (1) if the firm is audited by PwC,

KPMG, Grant Thornton, E&Y or D&T, and (0) otherwise.

Ahmed et al. (2010) declare that if managers use accruals tosmooth the changes in cash flows when reporting income, the vari-ance of the change in net income should be less than the variance ofthe change in cash flows. So, less positive values of this ratio wouldindicate greater income smoothing and more earnings manipulation.Therefore, listed firms during the IFRS adoption period are expectedto have higher values of the ratio SD(ΔΝΙ*)/SD(ΔCF*).

In order to test for the differences in the earnings smoothing met-ric (SD(ΔΝΙ*)/SD(ΔCF*)), we followed Barth et al. (2008) and used at-test based on the empirical distribution of the differences. Specifi-cally we randomly selected (with replacement) firm observationsthat were assigned to each sub-group (voluntary and mandatoryadopters) during the two sample periods and calculated the differ-ence between the two types of firms in that metric. We then repeatedthis procedure 1000 times and obtained the empirical distribution ofthe difference.

Finally, our second measure of earnings management aims to ex-amine the impact of IFRS adoption and audit quality on the magnitudeof performance-matched discretionary accruals. For this reason we es-timated the cross-sectional Jones (1991) model, as modified byKothari, Leone, andWasley (2005), in order to extract the discretionaryor abnormal accruals following a performancematching approach. Thismodel estimates discretionary accruals as a function of changes insales, the levels of property, plant and equipment, and the level of re-turn on assets by estimating the following OLS equation:

ACCit=TAt�1¼ a0 þα 1=TAt�1ð Þ þ β ΔSalesit=TAt�1ð Þ þ γ PPEit=TAt�1ð ÞþgROAitþeit

ð7Þ

where:

ACCit is total accruals defined as the difference between net in-come and

CF is the divided by lagged total assets.

ΔSalesit is the change in net sales deflated by lagged total assets(Salesit − Salesit − 1)

PPEit is the level of property plant and equipment for each yeardeflated by lagged total assets

ROA is the end of year return on assets estimated as net incomeover total assets

TAit is firm's total assets at the end of the fiscal year.

As Kothari et al. (2005) argue, the inclusion of a constant term in theJones (1991) model provides an additional control for hetero-scedasticity not alleviated by deflating the variables with total assets.A constant term also mitigates problems arising from omitted size vari-ables and produces discretionary accrual measures that are more sym-metric, overcoming model misspecifications and making the power ofthe test comparisons more clear. Finally, the inclusion of a profitabilitymeasure (ROA) is designed to enhance the effectiveness of the perfor-mance matching methodology. The discretionary accruals from themodified Jones (1991) model were defined as the residuals from esti-mating Eq. (7):

DACit¼ ACCit=TAit�1�α 1=TAit�1ð Þ þ β ΔSALESit=TAit�1ð Þþγ PPEit=TAit�1ð Þ þ gROA:

ð8Þ

In order to estimate the performance-matched discretionaryaccruals we followed Kothari et al. (2005) and matched each firmyear observation with another from the same industry classificationand year, and with the closest return on assets (ROA). The final stepwas to define the Jones model performance-matched discretionaryaccruals for firm i in year t as the discretionary accruals from Eq. (8)minus the matched firm's Jones model discretionary accruals foryear t. As Kothari et al. (2005) argue, the performance matchingapproach based on ROA and using the Jones model produces a moresalient measure of discretionary accruals since the means and me-dians in performance-related sub-samples are closest to zero moreoften than the other measures.

The absolute value of performance-matched discretionary accruals(|DACC|) from Eq. (8) is our second measure of earnings management.The absolute value is used because earnings management can involveeither income increasing or income decreasing accruals to meet earn-ings targets (Bowen, Rajgopal, & Venkatachalam, 2003; Klein, 2002;Reynolds & Francis, 2000; Wang, 2006; Warfield, Wild, & Wild,1995). A higher value indicates a greater level of manipulation in thefinancial statements and thus a lower earnings quality. Also followingAhmed et al. (2010), we estimate model 9 using the signed value ofdiscretionary accruals so as to control for the consistency of the results.In order to test hypotheses H1 and H2 we introduced the absolute andsigned value of performance-matched discretionary accruals (DACC)respectively as the dependent variable in the following model:

DACCit DACCj jit¼ c0þc1IFRSitþc2AUDitþc3DVOLitþc4AUD�DVOLit

þc5AUD�IFRSitþβControlit þ γYeardummies

þδIndustrydummiesþ uit

ð9Þ

where:

IFRS is a dummy receiving (1) for the post-IFRS period (2005–2008)and (0) otherwise

AUD is a dummy receiving (1) if the firm is audited by PwC,KPMG, Grant Thornton, E&Y or D&T, and (0) otherwise

DVOL is a dummy receiving (1) for voluntary IFRS adopters and(0) otherwise

AUD*IFRS is the interaction term between AUD and IFRS.

If H1 is true we expect a negative coefficient on the IFRS and DVOLvariables indicating that the adoption of IFRS for both voluntary and

115P.E. Dimitropoulos et al. / Advances in Accounting, incorporating Advances in International Accounting 29 (2013) 108–123

mandatory adopters has resulted in less earnings management behav-ior. Additionally, if H2 is valid we expect to find a negative and statisti-cally significant coefficient on the interaction terms AUD*IFRS andDVOL*AUD suggesting that Big-5 audit firms provide better quality au-dits than local audit firms, meaning that they have greater abilities andincentives for the enforcement of more transparent financial reportingafter adopting the high-quality accounting standards of IFRS.

4.5. Control variables

In all models, we included several control variables suggested byprevious research, and also, because we wanted to capture differencesin earnings management, timely loss recognition and value relevanceincentives. First of all we controlled for firm size measured as the nat-ural logarithm of total assets (LnTA). Watts and Zimmerman (1990)and Van Tendeloo and Vanstrelen (2005) argue that larger firms aremore likely to prefer downward earnings management because the po-tential for government scrutiny increases as firms become larger andmore profitable. Therefore, we expect a negative relation between theSIZE variable and the absolute value of discretionary accruals. Beekes,Pope, and Young (2004) and Vafeas, Trigeorgis, and Georgiou (1998)argue that earnings should be more informative for smaller firmssince there is less media information available compared with largerfirms. On the other hand, in small markets larger firms drawmore pub-licity and are more closely followed, thus their earnings quality is likelyto be known with less uncertainty compared to their smaller counter-parts. Therefore, investors may place greater trust in larger firms withwhich they are more familiar, and thus be more responsive to their fi-nancial reports. Ultimately, the mitigating role of firm size is again anopen empirical question.

Additionally, we controlled for the impact of firm leverage (LEV).Van Tendeloo and Vanstrelen (2005) and Billings (1999) documentthat the debt-equity hypothesis predicts that highly leveraged firmsare more likely to engage in upward earnings management to avoiddebt covenant violations and are associated with lower earnings

Table 2Descriptive Statistics of Sample Variables.

Panel A: Pre and post IFRS mandatory adopters

Variables Full sample (voluntary and mandatory) Pre-IFRS (mandatory)

Mean Median Std dev Min Max Mean Median

R 0.074 −0.124 1.954 −0.891 35.66 −0.138 −0.195P 5.232 2.411 9.212 0.000 100.00 4.846 2.865NIPS 0.063 0.018 1.066 −4.111 25.36 0.130 0.019BVPS 6.321 1.414 65.220 0.036 18.96 7.13 0.90ACC −0.021 −0.132 0.318 −3.519 2.40 −0.031 −0.025DACC −0.000 −0.002 0.244 −2.536 2.34 0.000 −0.002CFO 0.092 0.030 0.544 −3.481 12.52 0.065 0.042SIZE 5.242 4.874 1.444 1.558 10.97 4.451 4.471LEV 8.590 1.650 7.890 −29.070 60.37 16.540 2.89GR 3.210 0.050 22.890 −1.000 50.96 0.162 −0.009AUD 0.456 0.000 0.442 0.000 1.00 0.352 0.000

Panel B: Pre and post IFRS voluntary adopters

Variables Pre-IFRS (voluntary)

Mean Median Std dev Min Max

R 0.128 0.102 0.325 −0.632 12.5P 6.545 3.214 7.774 2.365 39.5NIPS 0.235 0.122 1.547 −0.789 25.3BVPS 9.541 2.121 87.540 2.361 18.9ACC −0.035 −0.010 0.256 −2.694 1.2DACC 0.005 −0.004 0.144 −1.878 1.5CFO 0.159 0.102 0.359 1.236 12.5SIZE 6.547 4.772 2.141 2.878 10.9LEV 9.256 2.920 10.230 0.548 11.6GR 2.559 1.989 2.451 2.547 30.0AUD 0.614 0.000 0.455 0.000 1.0

response coefficients. Therefore we expect that leverage (as mea-sured by the ratio of total debt to common equity) will have a nega-tive impact on the quality of accounting information. Moreover, wecontrolled for the effect of growth opportunities (GROWTH) mea-sured as the percentage change in sales from year t-1 to year t.Kumar and Krishnan (2008) document that the value relevance ofearnings and cash flows increases with investment opportunities.Therefore, firms with a positive annual change in sales are expectedto have more value relevant accounting information and presentgreater timely loss recognition. Thus, the coefficient of the GROWTHvariable is expected to have a positive sign.

Finally, we included the level of operating cash flows (CFO) deflat-ed by lagged total assets as a performance measure. Dimitropoulosand Asteriou (2009) document that cash flows are equally value rel-evant with earnings in determining stock return movements withinthe Greek accounting setting. Consequently we expect a positive coef-ficient on the CFO variable during the pre and post-IFRS periods. AlsoDechow, Sloan, and Sweeney (1996) and Young (1999) declare thatthe matching principle results in a natural smoothing of accrualswhich in turn causes negative discretionary accruals to occur in pe-riods of extreme positive cash flows. We included the absolutevalue of CFO so as to control for this potential misspecification. Ifthe abovementioned assertion is true we expect a positive relationbetween |DACC| and |CFO|.

5. Empirical results

5.1. Descriptive statistics and correlations

Table 2 presents the descriptive statistics of the sample variablesfor the whole period of investigation (2001–2008) as well as duringthe pre-IFRS period (2001–2004) and the post-IFRS period(2005–2008) for voluntary and mandatory IFRS adopters. The sampleincludes 101 non-financial firms listed in the Athens Stock Exchangefor the period 2001–2008, 76 are mandatory IFRS adopters and 25

Post-IFRS (mandatory)

Std dev Min Max Mean Median Std dev Min Max

0.410 −0.804 2.02 0.247 −0.063 2.667 −0.891 35.667.260 0.150 62.50 5.135 1.851 9.609 0.000 100.001.442 −1.203 24.55 −0.035 0.038 0.484 −4.111 3.031

101.400 0.100 17.90 3.787 2.172 7.755 0.036 17.5600.343 −3.519 2.40 −0.034 −0.002 0.337 −2.884 1.4820.276 −2.536 2.34 −0.000 −0.002 0.249 −2.179 1.5040.487 −3.481 3.48 0.102 0.010 0.700 −0.840 10.2501.310 1.711 9.00 4.889 5.047 1.445 1.558 9.137

12.900 0.010 19.85 2.275 1.436 7.126 −29.070 60.3702.542 −1.000 41.50 5.69 0.06 36.130 −1.000 50.9600.478 0.000 1.00 0.381 0.000 0.486 0.000 1.000

Post-IFRS (voluntary)

Mean Median Std dev Min Max

5 0.130 0.066 2.545 −0.541 15.228 6.845 2.958 8.598 2.653 42.226 0.214 0.044 0.389 −1.259 20.446 10.22 2.414 74.550 3.226 15.2847 −0.039 −0.007 0.277 −2.549 1.1447 0.004 −0.002 0.149 −1.778 1.412 0.199 0.088 0.458 1.598 11.447 6.897 3.658 1.897 2.102 9.455 9.887 2.487 7.445 0.447 12.995 3.897 1.547 2.850 3.258 28.5400 0.620 0.000 0.444 0.000 1.000

116 P.E. Dimitropoulos et al. / Advances in Accounting, incorporating Advances in International Accounting 29 (2013) 108–123

are voluntary adopters. The mean stock return presents an increaseduring the post-IFRS period to 24.7% compared to the pre-IFRS periodwhich is negative (−13.8%) and the same stands for the mean closingstock price (5.135). The same is true for the voluntary adopters but thedifference is not as large as for the mandatory adopters. Also the ratifi-cation of IFRS impacted on the estimation of income and the book valueof equity since both variables present a significant decrease during thepost-IFRS period for the mandatory adopters. However, mandatoryadopters seem to report less accruals in the post-IFRS period(−0.034) compared to the pre-IFRS period (−0.031) suggesting thatthe adoption of IFRS mitigated the distance between net income andoperating cash flows leaving managers little room for discretion onthe reporting of total accruals. Furthermore, voluntary adopterspresented an even greater decrease in accruals after 2005. Finally, man-datory adopters during the post-IFRS period are less leveraged (2.27),have more growth opportunities (5.69) and generate more cash fromtheir operating activities (0.102) as a percentage of total assets, com-pared with the pre-IFRS period. Additionally, their size shows an in-crease to 4.88 in the post-IFRS period compared to the pre-IFRSperiod (4.45), however this result can be attributed to the revaluationof assets that took place in 2005.

Table 3 presents the Pearson correlation coefficients among thesample variables for the whole period of investigation (2001–2008)as well as during the pre-IFRS period (2001–2004) and the post-IFRSperiod (2005–2008) for voluntary and mandatory adopters. As wecan see, the stock price (P) is positively correlated with net income

Table 3Pearson correlation coefficients of sample variables.

Panel A: Full sample

Variables R P NIPS BVPS ACC

P 0.074NIPS 0.032 0.053BVPS 0.005 0.041 0.685ACC 0.021 0.018 0.021 0.005DACC 0.005 0.009 0.041 0.004 0.7CFO 0.007 0.211 0.077 0.008 −0.6SIZE −0.154 −0.110 0.063 0.039 −0.0LEV −0.055 −0.041 −0.018 −0.018 −0.0GR 0.102 0.079 −0.011 0.008 −0.2AUD 0.085 0.112 0.008 −0.041 −0.0

Variables R P NIPS BVPS ACC

Panel B: Pre-IFRS sampleR 1.000 0.220 0.212 0.054 −0.026P 0.167 1.000 0.185 0.044 −0.021NIPS −0.031 −0.020 1.000 0.189 −0.054BVPS −0.049 −0.019 0.774 1.000 −0.022ACC −0.069 −0.007 −0.000 −0.002 1.000DACC −0.033 0.019 −0.002 0.000 0.804CFO 0.099 0.042 0.129 −0.000 −0.685SIZE 0.073 0.223 0.051 0.066 −0.159LEV 0.007 −0.017 −0.015 −0.008 −0.021GR 0.053 0.011 −0.022 −0.016 −0.504AUD 0.019 0.097 −0.026 −0.044 −0.067

Panel C: Post-IFRS sampleR 1.000 0.204 0.285 0.044 −0.036P 0.066 1.000 0.154 0.085 −0.054NIPS 0.193 0.150 1.000 0.214 −0.007BVPS 0.017 0.711 0.120 1.000 −0.019ACC 0.023 0.022 0.167 0.023 1.000DACC 0.010 0.069 0.207 0.039 0.739CFO −0.010 0.154 0.049 0.122 −0.729SIZE −0.054 0.268 0.111 0.195 −0.033LEV −0.016 −0.059 −0.063 −0.064 −0.021GR 0.004 0.103 0.007 0.091 −0.138AUD 0.029 0.104 0.149 −0.010 0.045

Correlations in bold indicate statistical significance at less than 0.05.Coefficients below the diagonal are for mandatory IFRS adopters and above the diagonal fo

(NIPS, 0.15) and book value of equity (BVPS, 0.71) only in thepost-IFRS period for mandatory adopters, while voluntary adoptersreported value relevant accounting information prior to 2005. Thesefindings indicate that the adoption of IFRS resulted in an increasedvalue relevance of earnings and book value of equity. Moreover, thecorrelation coefficient between operating cash flows (CFO) and totalaccruals is more negative in the post-IFRS period (−0.729) comparedto the pre-IFRS period (−0.685) for mandatory adopters. Finally, re-garding the other control variables, large firms report higher operatingcash flows, have more growth opportunities and are characterized byincreased audit quality compared to their smaller counterparts.

5.2. Regression results on value relevance and IFRS

In this section we tested whether the relationship between earn-ings per share and book value of equity per share with stock pricesvaries between local accounting principles and IFRS. For this reasonmodels 1 and 2 were estimated separately for the pre and post-IFRSperiods. The empirical findings are presented in Table 4.

According to our discussion, it is expected that the value relevance ofnet income and the book value of equity will be higher in the post-IFRSperiod for mandatory adopters and this will be depicted by an increasein: (a) model 1's R2, and (b) the magnitude and significance of net in-come (NIPS) and book value of common equity (BVPS) coefficients dur-ing the post-IFRS periodwhen compared to the pre-IFRS period. Resultsin Table 4 verify this assertion since the coefficients on NIPS and BVPS

DACC CFO SIZE LEV GR

1470 −0.40822 0.008 0.00641 −0.019 −0.001 0.09908 0.005 0.188 −0.059 −0.01119 0.040 0.073 0.040 0.090 −0.009

DACC CFO SIZE LEV GR AUD

−0.065 0.105 0.089 −0.054 0.047 0.011−0.099 0.148 0.305 −0.121 0.088 0.074−0.062 0.212 0.055 −0.219 0.089 0.025−0.014 0.105 0.018 −0.055 0.088 −0.022

0.744 −0.569 −0.207 −0.033 −0.486 −0.1141.000 −0.585 −0.055 −0.018 −0.063 −0.008

−0.567 1.000 0.184 −0.263 0.401 0.130−0.016 0.123 1.000 0.005 0.305 0.214−0.025 0.002 0.017 1.000 0.005 0.141

0.021 0.324 0.205 0.011 1.000 0.084−0.009 0.097 0.295 0.113 0.059 1.000

−0.074 0.122 0.047 −0.089 0.023 0.010−0.096 0.195 0.332 −0.132 0.087 0.055−0.052 0.263 0.042 −0.208 0.091 0.033−0.043 0.112 0.012 −0.041 0.090 −0.015

0.722 −0.554 −0.184 −0.022 −0.498 −0.1551.000 −0.536 −0.085 −0.015 −0.017 −0.009

−0.253 1.000 0.201 −0.274 0.444 0.1370.101 0.208 1.000 0.004 0.322 0.244

−0.033 −0.015 0.028 1.000 0.005 0.185−0.001 0.219 0.136 −0.009 1.000 0.045

0.077 0.036 0.326 0.104 −0.033 1.000

r voluntary adopters.

3 Absence measures the difference between domestic accounting standards and IFRSas the extent to which the rules regarding specific accounting issues are missing in do-mestic standards while covered by IFRS.

Table 4Regression results on value relevance and IFRS controlling for voluntary IFRS adopters.

Model 1: P�it¼ α0þα1NIPSitþα2BVPSitþα3DVOLitþα4DVOLit�NIPSitþα5DVOLit�BVPSitþβControlit þ γYeardummies þ δIndustrydummiesþ εit

Model 2:P�it¼ α0þα1NIPSitþα2BVPSitþα3DVOLitþα4DVOLit�NIPSitþα5DVOLit�BVPSitþa6AUDitþα7AUDit

�NIPSitþα8AUDit�BVPSit

þβControlit þ γYeardummies þ δIndustrydummies þ εit

Variables Pre-IFRS Post-IFRS

Model 1 Model 2 Model 1 Model 2

Constant (a0) −12.10a

(−2.74)−14.34a

(−2.71)−2.61b

(−1.77)−2.08b

(−1.68)NIPS (a1) −0.28

(−0.24)1.05(0.98)

0.886b

(1.75)0.441b

(1.74)BVPS (a2) 0.014

(0.54)0.020(0.57)

1.005a

(2.52)0.944a

(2.54)DVOL (a3) 0.112c

(1.57)0.141b

(1.69)DVOL*NIPS (a4) 0.189b

(1.86)0.205a

(2.54)DVOL*BVPS (a5) 0.054b

(2.05)0.065b

(2.02)AUD (a6) −0.159

(−0.44)−1.418b

(−1.63)AUD*NIPS (a7) −1.987

(−1.21)1.547a

(2.73)AUD*BVPS (a8) 0.811c

(1.56)1.066a

(5.17)SIZE (β1) 1.144a

(3.47)1.201a

(3.57)0.543a

(2.62)0.414b

(1.84)LEV (β2) −0.0011

(−0.31)−0.0012(−0.39)

−0.010(−0.27)

0.018(0.46)

CFO (β3) 0.454(0.68)

0.312(0.66)

0.487(1.14)

0.454(1.03)

GR (β4) −0.111(−0.57)

−0.088(−0.66)

0.0044(0.43)

0.005(0.57)

R2-adjusted 6.7% 6.9% 67.1% 70.7%F-stat 3.23a 3.77a 12.11a 13.30a

Year and Industrydummies

Included Included Included Included

N 404 404 404 404Vuong's Z statistic Model 1 post-IFRS vs

model 1 pre-IFRS8.25a P (b0.001)

Model 2 post-IFRS vs model 2 pre-IFRS 10.44a P (b0.001)

t-statistics are in the parentheses.a Significant at 1% (two-tailed test).b Significant at 5% (two-tailed test).c Significant at 10% (two-tailed test).

117P.E. Dimitropoulos et al. / Advances in Accounting, incorporating Advances in International Accounting 29 (2013) 108–123

are positive and significant only in the post-IFRS period (0.886 and1.005 respectively). The relative coefficients for mandatory adoptersin the pre-IFRS period are negative and non-significant suggestingthat the adoption of IFRS had a significant positive impact on thevalue relevance of earnings and the book value of equity, thus verifyinghypothesis H1. Also the interaction terms of DVOL for voluntaryadopters with NIPS and BVPS are positive and significant suggestingthat voluntary adopters are characterized by enhanced value relevancerelative to the other firmswhich followedGreekGAAP. The negative co-efficients on the accounting variables during the pre-IFRS period for themandatory adopters could also be attributed to the financial distressthat was experienced after the ASE crash in 1999–2000, which mayhave impacted on the perceived quality of accounting information.Stated differently, investors may have lost confidence in the capitalmarket due to severe deficiencies during the crash period whichinfluenced the relevance of financial statements as a whole, resultingin a smaller association between returns, earnings and book values.However, the introduction of well-respected and high-quality account-ing standards, such as IFRS,may have reinstated investors' confidence infinancial reports and increased the relevance of accounting information.

Both models have been estimated including industry and yearfixed effects. The Vuong Z statistic is the likelihood ratio test devel-oped by Vuong (1989) for non-nested model selection. A significantpositive Z indicates that models 1 and 2 during the pre-IFRS period

should be rejected in favor of models 1 and 2 during the post-IFRSperiod.

Thus, hypothesis H1 is verified by the increase in R2 up to 67.1%during the post-IFRS period compared to the pre-IFRS period of6.7%. The Vuong Z statistic is significant at the one percent signifi-cance level indicating that the model during the post-IFRS period bet-ter explains the original data relative to the same model during thepre-IFRS period.

This finding verifies recent evidence provided by Ding et al. (2007)who document that the higher the “absence” between domestic ac-counting standards and IFRS the lower the value relevance of account-ing information to themarket.3 Greece was found to be the top countryin the absence score suggesting that the introduction of IFRS and theenrichment of coverage of the local accounting standards are crucialfor improving the transparency and relevance of accounting disclo-sures. Thus, we can conclude that the adoption of IFRS resulted in animprovement in the value relevance of accounting information com-pared to the local accounting standards. However, our results standin opposition to the results of Ahmed et al. (2010) in their international

118 P.E. Dimitropoulos et al. / Advances in Accounting, incorporating Advances in International Accounting 29 (2013) 108–123

study (including Greece). One explanation for the Ahmed et al. (2010)findings is that their results are more pronounced for firms from coun-tries with a strong rule of law. Greece is a country with a weak rule oflaw and, in contrast, strong law countries even prior to IFRS adoptionmay already have had high quality accounting standards and therewould therefore be less room for improvement after the adoption ofIFRS. Conversely, in Greece where there is high book-tax conformityand weak legal enforcement, there seems to be more room for im-provement after the mandating of IFRS since the quality of accountingis lower compared to other code-law countries.

Moreover, H2 expects that the value relevance of net income andthe book value of equity will be higher in the post-IFRS period forfirms with increased audit quality because multinational audit firmshave more resources and experience of the issues of international ac-counting standards, and could provide greater assistance for the im-plementation and transition to IFRS compared to local audit firms.In order to test this assumption model 1 and model 2 are estimatedseparately for the pre and post-IFRS periods. H2 will be accepted ifthe explanatory power of model 2 in the post-IFRS period is highercompared to model 2 in the pre-IFRS period and also if coefficientsa7 and a8 are significant and have a higher magnitude relative to co-efficients a1 and a2 (again in the post-IFRS period). The findings inTable 4 again verify H2 since the explanatory power of model 2 inthe post-IFRS period presents a huge increase to 70.7% compared tothe pre-IFRS period where the relative R2 is only 6.9% and theVuong statistic is also significant at the one percent level.

Furthermore, the coefficients of the interaction variables NIPS*AUDand BVPS*AUD (from model 2 during the post-IFRS period) are bothpositive and statistically significant (1.547 and 1.066 respectively)and their relative magnitudes are higher compared to the coefficientsof the NIPS and BVPS variables (0.441 and 0.944 respectively). Regard-ing the control variables, the only significant coefficient was found inthe SIZE variable which is positive and statistically significant at theone percent significance level indicating that larger firms reportaccounting figures of increased value relevance. This result could be at-tributed to the fact that larger firms in Greece drawmore publicity thansmaller firms, and thus their earnings quality is likely to be disclosedwith less uncertainty. As a result, investors place greater trust in largerfirms and are more responsive to their financial reports.

Overall we can argue that firms characterized by increased auditquality present a higher value relevance of earnings and book valueof equity compared to firms of a lower audit quality and this positiveimpact seems to be further complemented by the implementation ofIFRS. Thus, our findings corroborate the results of Ewert andWagenhofer (2005), Lang et al. (2003, 2006), Leuz et al. (2003) andBarth et al. (2008) who argue that high-quality accounting standards(such as IFRS) result in accounting figures that have a higher valuerelevance.

5.3. Regression results on timely loss recognition and IFRS

Our next step was to examine the impact of IFRS adoption ontimely loss recognition. For this purpose we applied the classic Basu(1997) reverse regression model between earnings per share andstock returns. Model 3 is estimated separately for the pre andpost-IFRS periods and if IFRS increase the timely recognition of losseswe expect an increase in the explanatory power of the model and co-efficient b3 to be positive and has a higher magnitude relative to thepre-IFRS period. The empirical findings are presented in Table 5.

Regarding mandatory IFRS adopters, coefficient b3(D*R) of model3 is positive and statistically significant only in the post-IFRS periodwhile the relative coefficient for voluntary adopters (b7) is positiveand significant during the pre-IFRS period. Also, the R2 of model 3presents a small increase during the post-IFRS period up to 7.2% com-pared to the pre-IFRS R2 of 6.1%. The Vuong Z statistic in this table issignificant at the one percent significance level suggesting that the

model during the post-IFRS period better explains the original datarelative to the same model during the pre-IFRS period. In addition,voluntary IFRS adopters exhibited significant conservative reportingcompared to firms following the Greek GAAP since coefficient b7 ispositive and significant. Therefore, we can argue that the adoptionof IFRS resulted in an increase in conservatism within the Greek listedfirms. This finding is consistent with Barth et al. (2008) who foundthat firms operating under IFRS exhibit more timely recognition oflosses compared to the non-IFRS firms.

In addition, hypothesis H2 expects that firms audited by Big-5audit corporations will present more timely loss recognition. If thisis true we expect a significant positive coefficient on the R*D*AUD(b11) interaction term. The results of model 4 in Table 5 verify theaforementioned hypothesis since the coefficient b11 is positive andstatistically significant and presents a higher magnitude in thepost-IFRS period (0.410) relative to the pre-IFRS period (0.248). Inaddition, the R2 of model 4 in the post-IFRS period presents an in-crease to 21.1% relative to the pre-IFRS period where the respectiveR2 is only 7.4%. Moreover, the Vuong Z statistic is significant at theone percent significance level suggesting that model 4 during thepost-IFRS period better explains the original data relative to thesame model during the pre-IFRS period. Regarding the control vari-ables, SIZE was found to have a positive and significant coefficientonly in the post-IFRS period suggesting that larger firms reportmore timely losses after the adoption of IFRS. Furthermore, the LEVvariable has the expected sign but is marginally significant only dur-ing the post-IFRS era. Finally, the GR variable was found to be nega-tive and marginally significant only during the pre-IFRS period, andthe CFO coefficient was positive and statistically significant as weexpected in both sub-periods indicating that firms generating in-creased cash flows are more timely loss reporters.

Consequently, if higher accounting quality is measured by the rec-ognition of large losses as they occur instead of being deferred to fu-ture periods (Ball & Shivakumar, 2005, 2006) our results suggest thatIFRS are characterized by increased quality since they have proven tocontribute significantly to the more timely recognition of losses with-in Greek listed firms.

5.4. Empirical findings on earnings management and IFRS

Our last section presents the empirical findings regarding the im-pact of IFRS adoption on earnings management behavior. In order totest this association we applied two different metrics of earnings ma-nipulation. Our first test, which controls for earnings smoothing, isthe ratio of the standard deviation of the change in net income(ΔΝΙ*) to the standard deviation of the change in cash flow (ΔCF*) re-siduals during the pre and post-IFRS periods (Ahmed et al., 2010). Therelevant results are presented in Table 6.

The differences between the ratios were estimated by applying at-test based on the distribution of the differences. ΔΝΙ* and ΔCF* arethe change of net income and change of cash flow residuals fromthe following four regression models, instead of comparing thesemeasures directly into the specific tests.

The empirical findings indicate the ratio of standard deviation ofthe change in net income (ΔΝΙ*) to the standard deviation of thechange in cash flow (ΔCF*) residuals, as well as the difference ofthis ratio between firms audited by Big-5 audit and non-Big-5 auditcorporations during the pre and post-IFRS adoption periods. This dif-ference was estimated using a t-test based on the distribution of thedifferences (Barth et al., 2008). Evidence suggests that the specificratio is higher during the post-IFRS adoption period for mandatoryadopters (difference 0.0007 and significant at 1%), while the highestdifference is demonstrated by firms audited by the Big-5 (significantat 1%). Overall, the evidence in Table 6 suggests that Greek listedfirms which adopted IFRS after its inauguration in 2005, report lesssmoothed financial statements after the implementation of IFRS,

Table 5Regression results on timely loss recognition and IFRS controlling for voluntary IFRS adopters.

Model 3: EYi;t¼ b0þb1�Ritþb2

�Ditþb3�Rit

�Ditþb4DVOLitþb5Dit�DVOLitþb6Rit

�DVOLitþb7Rit�Dit

�DVOLitþβControlit þ γYeardummies þ δIndustrydummiesþ eit

Model 4:EYit¼ b0þb1Ritþb2Ditþb3Rit

�Ditþb4DVOLitþb5Dit�DVOLitþb6Rit

�DVOLitþb7Rit�Dit�DVOLitþb8AUDitþb9Dit�AUDitþb10Rit

�AUDitþb11Rit�Dit

�AUDit

þβControlit þ γYeardummiesþ δIndustrydummies þ eit

Variables Pre-IFRS Post-IFRS

Model 3 Model 4 Model 3 Model 4

Constant (b0) −0.981(−0.77)

−1.221(−1.11)

−0.528(−1.20)

0.0080(0.33)

R (b1) −0.141(−0.45)

−0.187(−0.60)

0.030a

(2.57)−0.068a

(−3.18)D (b2) −0.044

(−0.21)−0.009(−0.40)

0.0038(0.25)

−0.051(−0.98)

D*R (b3) 0.214(1.58)

0.221b

(1.69)0.251b

(1.90)0.338b

(2.12)DVOL (b4) 0.157b

(1.74)0.184b

(1.90)DVOL*D (b5) −0.027

(−0.84)−0.031(−0.91)

DVOL*R (b6) −0.094(−1.12)

−0.105(−1.36)

DVOL*D*R (b7) 0.187b

(2.31)0.210b

(2.27)AUD (b8) 0.051

(0.19)0.042(0.90)

AUD*D (b9) −0.198(−0.67)

0.204b

(1.71)AUD*R (b10) 0.094

(0.88)0.192a

(6.50)AUD*D*R (b11) 0.248b

(2.12)0.410b

(2.17)SIZE (β1) 0.041

(0.67)0.044(0.91)

0.059b

(1.93)0.027b

(1.66)LEV (β2) −0.001

(−0.28)0.0008(0.69)

−0.009(−0.96)

−0.010(−1.27)

CFO (β3) 0.487a

(2.66)0.518a

(2.71)0.131b

(2.57)0.191b

(2.22)GR (β4) −0.044

(−1.21)−0.044(−1.17)

−0.0040(−0.19)

−0.0017(−0.27)

R2-adjusted 6.1% 7.4% 7.2% 21.1%F-stat 2.88a 2.77a 3.24a 7.03a

Year and Industry dummies Included Included Included IncludedN 404 404 404 404Vuong's Z statistic Model 3 post-IFRS vs. Model 3 pre-IFRS 3.87a P (b0.001)

Model 4 post-IFRS vs. Model 4 pre-IFRS 4.27a P (b0.001)

t-statistics are in the parentheses.a Significant at 1% (two-tailed test).b Significant at 5% (two-tailed test).

119P.E. Dimitropoulos et al. / Advances in Accounting, incorporating Advances in International Accounting 29 (2013) 108–123

verifying previous arguments made by Barth et al. (2008) and Chen etal. (2010) that IFRS adoption leads to less earnings manipulation.

Our second test on earnings management is based on the estima-tion of model 9 which includes the absolute value of performance-matched discretionary accruals as the dependent variable (seeKothari et al., 2005 for details). The relevant results are presented inTable 7.

The coefficient of the IFRS dichotomous variable is negative and sta-tistically significant (−0.181) and the relative coefficient for the volun-tary adopters (DVOL) is also negative and significant (−0.145)indicating that the adoption of IFRS resulted in a decrease in themagnitude of performance-matched discretionary accruals. Thus, thisfinding corroborates the evidence by Aubert and Grudnitski (2011)who argue that mandatory IFRS adoption in Greece improved accrualquality. In addition, the coefficient of the interaction term IFRS*AUDis also negative and statistically significant at the one percent signifi-cance level (−0.125) verifying H2 and suggesting that Big-5 auditfirms provide better quality audits than local audit firms. Thus, beingaudited by a Big-5 audit firm further mitigates any incentives for ma-nipulating accounting income. Overall, the evidence from the two

metrics of earnings management tends to suggest that firms whichadopted IFRS after its inauguration in 2005 report better quality finan-cial statements, verifying previous arguments made by Barth et al.(2008) and Chen et al. (2010).

Finally, regarding the control variables, SIZE is significant with anegative coefficient similar to Watts and Zimmerman (1990) andVan Tendeloo and Vanstrelen (2005). Also the coefficient of the abso-lute value of operating cash flows (|CFO|) is positive and statisticallysignificant (0.166) as predicted, suggesting that the matching princi-ple results in a natural smoothing of accruals which in turn causesnegative discretionary accruals to occur in periods of extreme posi-tive cash flows, verifying Dechow et al. (1996) and Young (1999).

Overall the empirical results indicate that the adoption of IFRS hasimproved the quality of accounting information by increasing thevalue relevance of earnings and book value of equity, improving thetimely recognition of losses and by decreasing discretionary manipula-tion of earnings. Consequently, our findings allow us to accept bothhypotheses H1 and H2 and claim that accounting figures preparedunder IFRS are of a higher quality than those prepared under theGreek GAAP and that audit quality contributes to enhanced accounting

Table 6Univariate evidence on earnings smoothing.

Ratio of SD(ΔΝΙ*)/SD(ΔCF*) between the standard deviation of the change in net income and cash flows residuals.

Model 5: ΔΝΙit = a0 + a1SIZEit + a2GROWTHit + a3LEVit + a4AUDit + eit

Model 6: ΔCFit = a0 + a1SIZEit + a2GROWTHit + a3LEVit + a4AUDit + eit

Sample Ratio Voluntary adoption Mandatory adoption

Pre-IFRS Post-IFRS Difference Pre-IFRS Post-IFRS Difference

Full sample SD(ΔΝΙ*)/SD(ΔCF*) 0.0019 0.0020 0.0001 0.0021 0.0028 0.0007a

N 100 100 100 304 304 304Big-5 SD(ΔΝΙ*)/SD(ΔCF*) 0.0021 0.0019 0.0002 0.0027 0.0035 0.0008a

N 40 40 40 116 116 116Non Big-5 SD(ΔΝΙ*)/SD(ΔCF*) 0.0030 0.0029 0.0001 0.0024 0.0027 0.0003a

N 60 60 60 188 188 188

a Significant at 1% significance level.

Table 7Regression results on earnings management and IFRS controlling for voluntary IFRSadopters.

Model 9:DACCj jit¼ c0þc1IFRSitþc2AUDitþc3DVOLitþc4AUD � DVOLit

þc5AUD � IFRSitþβControlit þ γYeardummiesþδIndustrydummiesþ uit

Variables Model 9(signed DACC)

Model 9(absolute DACC)

Constant (c0) 0.310a

(2.70)0.355a

(3.05)IFRS (c1) −0.191b

(−2.18)−0.181b

(−1.99)AUD (c2) 0.088a

(3.01)0.084a

(2.77)DVOL (c3) −0.122

(−1.54)−0.145c

(−1.62)DVOL*AUD (c4) −0.074

(−1.19)−0.087(−1.25)

IFRS*AUD (c5) −0.144a

(−2.87)−0.125a

(−2.75)|CFO| (β1) 0.174a

(9.21)0.166a

(8.88)SIZE(β2) −0.018b

(−1.95)−0.021b

(−2.05)LEV(β3) −0.0010

(−0.55)−0.0009(−0.22)

GR (β4) −0.0004(−0.11)

−0.0003(−0.27)

Adjusted R2 22.7% 19.7%F-statistic 14.08a 13.08a

Year and Industry dummies Included IncludedN 808 808

120 P.E. Dimitropoulos et al. / Advances in Accounting, incorporating Advances in International Accounting 29 (2013) 108–123

quality during the post-IFRS period. These results corroborate the argu-ments of several studies such as Ashbaugh and Pincus (2001), Ewertand Wagenhofer (2005), Bartov et al. (2005), Hitz (2007), Penman(2007) and Barth et al. (2008).

5.5. Sensitivity analysis

In order to check the robustness of our results, we performed sev-eral sensitivity tests related to the specification of the empiricalmodels and the research design.4 First, we re-estimated model 9 in-cluding additional interaction variables between IFRS and the controlvariables (CFO, SIZE, LEV, GR) in order to control for any effects ofIFRS that are not picked up by the sole control variables. Resultsremained qualitatively unchanged compared to those reported inTable 7. Also, we replaced discretionary accruals as the dependentvariable on model 9 with total accruals in order to capture any prob-lems from possible accrual misspecification arising from the estima-tion of the Jones (1991) model. The results remained unchangedafter this modification.

Furthermore, we controlled for possible misspecification in the esti-mation of the performance-matched discretionary accruals. FollowingKothari et al. (2005) we re-estimated the Jones (1991) model as fol-lows: (1) without a constant, (2) without matching firms based onROA but estimating DACC including ROA as an additional regressor,and 3) by implementing the Jones model cross-sectionally using allfirms instead of just firms within the same industry. In addition, were-estimated models 1 and 2 using the actual stock price and the annu-al buy-and-hold return as the dependent variables, and the level ofearnings and the book value of equity deflated by the beginning ofthe year market capitalization as the explanatory variables. The resultsare qualitatively similar to the results presented in Table 4. Also, were-estimated models 3 and 4 during the pre and post-IFRS periods byreplacing the dependent variable (pre-tax earnings deflated by laggedtotal assets) with earnings per share deflated by the opening shareprice. Again the results did not change after this modification. Addi-tionally, we controlled for the level of family ownership and cases ofcross-listed firms following the discussion in Section 2 suggestingthat the large number of family-owned firms in Greece has encouragedthe limitation of financial reporting to provide tax-relevant informationonly. For this purpose we included a dummy coded one (1) forcross-listed firms and zero (0) otherwise and took into account thepercentage of stocks owned by the founding family members of eachcompany (following Wang, 2006), before re-estimating all models.The results proved to be unaffected by these variables since thecross-listing dummy was insignificant and the ownership variable

4 Untabulated results can be obtained from the corresponding author.

was significant only on the earnings management regression,suggesting that family-owned firms engage in less financial statementmanipulation after the adoption of IFRS.

Furthermore, Johnson (1999) finds convincing evidence that thevalue relevance of earnings fluctuates over time as a function oftime-varying expected returns. She argues that interest rates shouldbe included in models whenever inferences about changes in the in-formation relevance of earnings due to non-interest rate factors(e.g. change in accounting methods) are hypothesized. For this reasonmodel 1 was re-estimated for the pre and post-IFRS periods includingthe term-spread of interest rates as an additional explanatory vari-able. Results remained unchanged relative to those reported inTable 4. In addition, we re-estimated model 1 combining the variablesduring the pre and post-IFRS adoption periods by interacting NIPSand BVPS with a dummy IFRS receiving one (1) for the period

t-statistics are in the parentheses.a Significant at 1% significance level (two-tailed test).b Significant at 5% significance level (two-tailed test).c Significant at 10% significance level (two-tailed test).

121P.E. Dimitropoulos et al. / Advances in Accounting, incorporating Advances in International Accounting 29 (2013) 108–123

2005–2008 and zero (0) otherwise. The results were qualitatively thesame compared with those in Table 4. Moreover, we re-estimated allempirical models excluding the control variables and the year andindustry dummies so as to check whether these variables confoundthe explanatory power of the models but the results remainedunchanged.

In addition, we controlled for a ‘learning curve’ effect. Based on thisrationale, as the firm continues applying IFRS concepts and practices itbecomes more familiar with proper IFRS application. Cohen andLeventis (2013) found that time has a beneficial effect on compliancewith accounting standards in Greece. This is another explanation ofwhy empirical findings may point towards IFRS not contributing tothe improvement of accounting quality. For this reason we dividedthe post-adoption period into two sub-periods, namely 2005–2006and 2007–2008, and re-estimated models 9, 2 and 4. The resultssuggested that there is some learning curve effect since the resultsfor value relevance and earnings management were more significantduring the 2007–2008 sub-period. Contrarily however, accountingconservatism does not present any significant differences betweenthe two sub-periods.

Finally, we controlled for possible cross-sectional dependence andheterogeneity in our data by applying the methodology proposed byKousenidis et al. (2009). The rationale behind this modification isthat the level of conservatism and value relevance may not be directlyrelated to the members of the data panel. To be more specific, 1, 2, 3,and 4 were re-estimated for the pre- and post-IFRS periods by includ-ing the cross-sectional means of the dependent and independent var-iables namely P, NIPS and BVPS for model 1 and EBT and R for model2. Results remained unchanged after this specification.

6. Conclusions

The issue of international accounting standards and their impact onaccounting quality have been a part of the international research agen-da for a number of years. Many studies have documented that theadoption of high-quality accounting standards results in an increasein the quality of accounting information. For instance, researchersargue that IFRS improves the reliability of financial reporting by limit-ing opportunistic managerial discretion (Ashbaugh & Pincus, 2001;Barth et al., 2008; Ewert & Wagenhofer, 2005). The contrary view isthat the flexibility inherent in IFRS and lax enforcement might providegreater opportunities for firms to manage earnings (Burgsthaler et al.,2006; Ball et al., 2003; Street & Gray, 2002; Cairns, 1999; Breeden,1994 among others). While two conflicting views exist regarding theinfluence of IFRS, empirical research on the impact of IFRS on earningsmanagement has also provided mixed results. Van Tendeloo andVanstrelen (2005), for example, found that earnings management be-havior was not significantly different between companies in Germanythat adopted IFRS and those that relied on German GAAP. Conversely,Barth et al. (2008) concluded that firms applying IFRS displayed signif-icantly less earnings management, meaning that they report more reli-able and transparent financial statements.

Part of the motivation for this study was to shed further light on thecompeting views regarding the efficiency of IFRS in reducing earningsmanipulation and improving earnings quality within an accountingframework characterized by a long history of historical-accountingprinciples. We examined whether accounting information during thepost-IFRS period (2005–2008) demonstrates less earnings manage-ment, more timely loss recognition and higher value relevance, whencompared to the relative amounts reported during the pre-IFRS period(2001–2004). After controlling for firm-specific characteristics such assize, growth opportunities, risk and audit quality we found that the in-troduction of IFRS contributed to less earnings management, moretimely loss recognition, and greater value relevance of earnings andbook value of equity compared to the local accounting standards. Fur-thermore, our findings document that a firm's audit quality further

complements the beneficial impact of IFRS since companies auditedby a Big-5 audit firm exhibit higher levels of accounting quality com-pared to their non-Big-5 counterparts.

Our findings contribute to the ongoing debate over whetherhigh-quality accounting standards are effective and sufficient incode-law countries and countries with a historical-cost accountingbase. Moreover, our results could prove useful to investors sincethey indicate that financial accounting information prepared underIFRS and audited by the Big-5 audit firms is more reliable and has en-hanced value relevance.

However, the results of this study should be treated with cautionsince they are subject to several limitations. Although we have con-trolled for various incentives of reporting discretion, we acknowledgethat there may be other incentives which remained uncontrolled byour empirical models. Thus, there may be an additional source ofbias that affects our inferences regarding the estimation of model 9.Finally, the small size of our sample and the focus on a single countrysetting could be another source of bias which does not allow us toinfer any generalization of the results onto firms operating outsidethe Greek borders.

Future research, however, could help resolve this problem by ex-amining the impact of IFRS adoption on accounting quality betweendifferent EU countries which share common accounting characteris-tics, thus providing more salient inferences to the current debatethrough cross-country evidence. Furthermore, a recent paper byCahan, Emanuel, and Sun (2009) declares that even if the accountingsystem (IFRS) can generate high-quality accounting information in acountry with a weak institutional infrastructure, the informationwill still be less important for market participants. Consequently, itwould be interesting to examine the impact of IFRS adoption on ac-counting quality under any legislative improvements taking place inthe European business setting regarding the issues of investor protec-tion and institutional transparency.

References

Aharony, J., Barniv, R., & Falk, H. (2010). The impact of mandatory IFRS adoption onequity valuation of accounting numbers for security investors in the EU. The EuropeanAccounting Review, 19, 535–578.

Ahmed, A. S., Neel, M., & Wang, D. (2010). Does mandatory adoption of IFRS improveaccounting quality? Preliminary evidence. Working paper, Texas A&M University.

Ashbaugh, H., & Pincus, M. (2001). Domestic accounting standards, internationalaccounting standards and the predictability of earnings. Journal of AccountingResearch, 39, 417–434.

Aubert, F., & Grudnitski, G. (2011). The impact and importance of mandatory adoptionof International Financial Reporting Standards in Europe. Journal of InternationalFinancial Management and Accounting, 22, 1–26.

Ball, R., Kothari, S. P., & Robin, A. (2000). The effect of international institutional factors onproperties of accounting earnings. Journal of Accounting and Economics, 29, 1–51.

Ball, R., Robin, A., &Wu, J. S. (2003). Incentives versus standards: Properties of accountingincome in four East Asian countries. Journal of Accounting and Economics, 36,235–270.

Ball, R., & Shivakumar, L. (2005). Earnings quality in UK private firms: Comparative lossrecognition timeliness. Journal of Accounting and Economics, 39, 83–128.

Ball, R., & Shivakumar, L. (2006). The role of accruals in asymmetrically timely gain andloss recognition. Journal of Accounting Research, 44, 207–242.

Ballas, A. A. (1994). Accounting in Greece. The European Accounting Review, 3, 107–121.Ballas, A. A., & Fafaliou, I. (2008). Market shares and concentration in the EU auditing

industry: The effects of Andersen's demise. International Advances in EconomicResearch, 14, 485–497.

Ballas, A. A., Skoutela, D., & Tzovas, C. A. (2010). The relevance of IFRS to an emergingmarket: Evidence from Greece. Managerial Finance, 36, 931–948.

Barth, M. E., Landsman, W. R., & Lang, M. H. (2008). International accounting standardsand accounting quality. Journal of Accounting Research, 43, 467–498.

Bartov, E., Goldberg, S., & Kim, M. (2005). Comparative value relevance among German,U.S. and International Accounting Standards: A German stock market perspective.Journal of Accounting, Auditing and Finance, 20, 95–119.

Basu, S. (1997). The conservatism principle and the asymmetric timeliness of earnings.Journal of Accounting and Economics, 24, 3–37.

Bauwhede, H. V., Willekens, M., & Gaeremynck, A. (2003). Audit firm size, publicownership and firms' discretionary accruals management. The International Journalof Accounting, 38, 1–22.

Beekes, W., Pope, P., & Young, S. (2004). The link between earnings timeliness, earningsconservatism and board composition: Evidence from the UK. Corporate Governance:An International Review, 12, 47–59.

122 P.E. Dimitropoulos et al. / Advances in Accounting, incorporating Advances in International Accounting 29 (2013) 108–123

Beuselinck, C., Joos, P., Khurana, I. K., & Van der Meulen, S. (2009). Mandatory IFRSreporting and stock price informativeness. Working paper, Tilburg University.

Billings, B. K. (1999). Revisiting the relation between the default risk of debt and theearnings response coefficient. The Accounting Review, 74, 509–522.

Bowen, R. M., Rajgopal, S., & Venkatachalam, M. (2003). Accounting discretion, corpo-rate governance and firm performance. Working paper, University of Washingtonand Duke University.

Breeden, R. (1994). Foreign companies and US markets in a time of economic transfor-mation. Fordham International Law Journal, 17, S77–S96.

Burgsthaler, D., Hail, L., & Leuz, C. (2006). The importance of reporting incentives:Earnings management in European private and public firms. The AccountingReview, 81, 983–1016.

Byard, D., Li, Y., & Yu, Y. (2011). The effect of mandatory IFRS adoption on financialanalysts' information environment. Journal of Accounting Research, 49, 69–96.

Cahan, S. F., Emanuel, D., & Sun, J. (2009). The effect of earnings quality andcountry-level institutions on the value relevance of earnings. Review of Quantita-tive Finance and Accounting, 33, 371–391.

Cairns, D. (1999). Degrees of compliance. Accountancy International, 68–69.Callao Gastón, S., Ferrer García, C., Jarne, J. I., & Laínez, J. A. (2010). IFRS adoption in

Spain and the United Kingdom: Effects on accounting numbers and relevance. Ad-vances in Accounting, Incorporating Advances in International Accounting, 26,304–313.

Callao, S., Jarne, J. I., & Laínez, J. A. (2007). Adoption of IFRS in Spain: Effect on the com-parability and relevance of financial reporting. Journal of International Accounting,Auditing and Taxation, 16, 148–178.

Caramanis, C., & Lennox, C. (2008). Audit effort and earnings management. Journal ofAccounting and Economics, 45, 116–138.

Chen, H., Tang, Q., Jiang, Y., & Lin, Z. (2010). The role of international financial reportingstandards in accounting quality: Evidence from the European Union. Journal ofInternational Financial Management and Accounting, 21, 220–278.

Cohen, S., & Leventis, S. (2013). An empirical investigation of audit pricing in the publicsector: The case of Greek LGOs. Financial Accountability and Management, 29(1),74–98.

Conover, M., Miller, R., & Szakmary, A. (2008). The timeliness of accounting disclosuresin international security markets. International Review of Financial Analysis, 17(5),849–869.

Dargenidou, C., & McLeay, S. (2010). The impact of introducing estimates of the futureon international comparability in earnings expectations. The European AccountingReview, 19, 511–534.

Daske, H. (2006). Economic benefits of adopting IFRS or US-GAAP— Have the expectedcosts of equity capital really decreased? Journal of Business Finance and Accounting,33, 329–373.

Dechow, P. M. (1994). Accounting earnings and cash flows as measures of firmperformance: The role of accounting accruals. Journal of Accounting and Economics,18, 3–42.

Dechow, P., Sloan, R., & Sweeney, A. (1996). Causes and consequences of earningsmanipulation: An analysis of firms subject to enforcement actions by the SEC.Contemporary Accounting Research, 13, 1–36.

Devalle, A., Onali, E., & Magarini, R. (2010). Assessing the value relevance of accountingdata after the introduction of IGRS in Europe. Journal of International FinancialManagement and Accounting, 21, 85–119.

Dimitropoulos, P., & Asteriou, D. (2009). The impact of accruals and cash flows on thereturns-earnings relation: Evidence from Greece. International Journal of Account-ing, Auditing and Performance Evaluation, 5, 384–407.

Ding, Y., Hope, O. -K., Jeanjean, T., & Stolowy, H. (2007). Differences between domesticaccounting standards and IAS: Measurement, determinants and implications.Journal of Accounting and Public Policy, 26, 1–38.

Djatej, A., Gao, G., Sarikas, R. H. S., & Senteney, D. L. (2009). An investigation of the com-parative impact of degree of implementation of IFRS upon the public and privateinformation quality of East and West European firms. Advances in Accounting,Incorporating Advances in International Accounting, 25, 208–215.

Doukakis, L. C. (2010). The persistence of earnings and earnings components after theadoption of IFRS. Managerial Finance, 36, 969–980.

Eccher, E., & Healy, P. (2003). The role of International Accounting Standards in transi-tional economies: A study of the People's Republic of China. Working paper, Massa-chusetts Institute of Technology.

Ewert, R., & Wagenhofer, A. (2005). Economic effects of tightening accounting stan-dards to restrict earnings management. The Accounting Review, 43, 1101–1124.

Francis, J., & Wang, D. (2003). Are Big 4 audit firms conservative only in the UnitedStates? Working paper, University of Missouri–Columbia.

Frankel, R., Johnson, M., & Nelson, K. (2002). The relation between auditors' feesfor non-audit services and earnings management. The Accounting Review, 77, 71–105.

Ghosh, A., & Moon, D. (2005). Auditor tenure and perceptions of audit quality. The Ac-counting Review, 80, 585–612.

Glaum, M., & Street, D. (2003). Compliance with the disclosure requirements ofGermany's new markets: IAS versus US GAAP. Journal of International FinancialManagement and Accounting, 14, 64–100.

Gore, P., Pope, P. F., & Singh, A. K. (2001). Non-audit services, auditor independenceand earnings management. Working paper, Lancaster University.

Guenther, N., Gegenfurtner, B., Kaserer, C., & Achleitner, A. K. (2009). International fi-nancial reporting standards and earnings quality: The myth of voluntary vs. man-datory adoption. Working paper no 2009–09, Centre for Entrepreneurial andFinancial Studies, Technical University of Munich.

Guenther, D. A., & Young, D. (2000). The association between financial accounting andmeasures of real economic activity: A multinational study. Journal of Accountingand Economics, 29, 53–72.

Gul, F. A., Sun, S. Y. J., & Tsui, J. S. L. (2003). Audit quality, earnings and the Shanghaistock market reaction. Journal of Accounting, Auditing and Finance, 18, 411–427.

Gul, F. A., Tsui, J. S. L., & Dhaliwal, D. S. (2006). Non-audit services, auditor quality andthe value relevance of earnings. Accounting and Finance, 46, 1–21.

Hassan, O. A. G., Romilly, P., Giorgioni, G., & Power, D. (2009). The value relevance ofdisclosure: Evidence from the emerging capital market of Egypt. The InternationalJournal of Accounting, 44, 79–102.

Hitz, J. M. (2007). The decision usefulness of fair value accounting: A theoretical per-spective. The European Accounting Review, 16, 323–362.

Hodgdon, C., Tondkar, R. H., Adhikari, A., & Harless, D. W. (2009). Compliance with in-ternational financial reporting standards and auditor choice: New evidence on theimportance of the statutory audit. The International Journal of Accounting, 44,33–55.

Hofstede, G. (1980). Culture's consequences: International differences in work-relatedvalues. Newbury, CA: Sage Publications.

Hope, O. K. (2003). Disclosure practises, enforcement of accounting standards andanalysts' forecast accuracy: An international study. Journal of Accounting Research,41, 235–272.

Hung, M., & Subramanyam, K. R. (2007). Financial statement effects of adopting Inter-national Accounting Standards: The case of Germany. Review of Accounting Studies,12, 623–657.

Iatridis, G. (2008). Accounting disclosures and firms' financial attributes: Evidencefrom the UK stock market. International Review of Financial Analysis, 17, 219–241.

Iatridis, G. (2010). International Financial Reporting Standards and the quality of finan-cial statement information. International Review of Financial Analysis, 19, 193–204.

Iatridis, G. (2011). Accounting disclosures, accounting quality and conditional and un-conditional conservatism. International Review of Financial Analysis, 20, 88–102.

Iatridis, G., & Rouvolis, S. (2010). The post-adoption effects of the implementation ofInternational Financial Reporting Standards in Greece. Journal of International Ac-counting, Auditing and Taxation, 19, 55–65.

Johnson, M. F. (1999). Business cycles and the relation between security returns andearnings. Review of Accounting Studies, 4, 93–117.

Jones, J. (1991). Earnings management during import relief investigations. Journal ofAccounting Research, 29, 193–228.

Klein, A. (2002). Audit committee, board of director characteristics and earnings man-agement. Journal of Accounting and Economics, 33, 375–400.

Kothari, S. P., Leone, A. J., & Wasley, C. E. (2005). Performance matched discretionaryaccrual measures. Journal of Accounting and Economics, 39, 163–197.

Kousenidis, D. V., Ladas, A. C., & Negakis, C. I. (2009). Value relevance of conservativeand non-conservative accounting information. The International Journal of Account-ing, 44, 219–238.

Kumar, K. R., & Krishnan, G. V. (2008). The value relevance of cash flows and accruals:The role of investment opportunities. The Accounting Review, 83, 997–1040.

Landsman, W. R., Maydew, E. L., & Thornock, J. R. (2010). The information content ofannual earnings announcements and mandatory adoption of IFRS. Journal of Ac-counting and Economics, 53, 34–54.

Lang, M., Raedy, J., & Wilson, W. (2006). Earnings management and cross listing: Arereconciled earnings comparable to US earnings? Journal of Accounting and Econom-ics, 42, 255–283.

Lang, M., Raedy, J., & Yetman, M. (2003). How representative are firms that are crosslisted in the United States? An analysis of accounting quality. Journal of AccountingResearch, 41, 363–386.

Leuz, C., Nanda, D., &Wysocki, P. D. (2003). Earningsmanagement and investor protection:An international comparison. Journal of Financial Economics, 69, 505–527.

Leventis, S., & Caramanis, C. (2005). Determinants of audit time as a proxy of auditquality. Managerial Auditing Journal, 20, 460–478.

Park, C., & Pincus, M. (2001). Internal versus external equity funding sources and earningsresponse coefficients. Review of Quantitative Finance and Accounting, 16, 33–52.

Penman, S. (2007). Financial reporting quality: Is fair value a plus or minus? Accountingand Business Research, 37, 33–44.

Prather-Kinsey, J., & Meek, G. K. (2004). The effect of revised IAS 14 on segmentreporting by IAS companies. The European Accounting Review, 13, 213–234.

Raonic, I., McLeay, S., & Asimakopoulos, I. (2004). The timeliness of income recognitionby interlisted European companies: An analysis of institutional and market com-plexity. Journal of Business Finance and Accounting, 31, 115–148.

Reynolds, J. K., & Francis, J. R. (2000). Does size matter? The influence of large clients onoffice-level auditor reporting decisions. Journal of Accounting and Economics, 30,375–400.

Spathis, C., Doumpos, M., & Zopounidis, C. (2003). Using client performance measuresto identify pre-engagement factors associated with qualified audit reports inGreece. The International Journal of Accounting, 38, 267–284.

Spathis, C., & Georgakopoulou, E. (2007). The adoption of IFRS in South Eastern Europe:The case of Greece. International Journal of Financial Services Management, 2, 50–63.

Street, D. L., & Gray, S. J. (2001). Observance of international accounting standards:Factors explaining non-compliance. ACCA research report no 74, Association ofChartered Certified Accountants.

Street, D. L., & Gray, S. J. (2002). Factors influencing the extent of corporate compliancewith International Accounting Standards: Summary of a research monograph.Journal of International Accounting, Auditing and Taxation, 11, 51–76.

Teoh, S., & Wong, T. (1993). Perceived auditor quality and the earnings response coef-ficient. The Accounting Review, 68, 346–366.

Tsakumis, G. T. (2007). The influence of culture on accountants' application of financialreporting rules. Abacus, 43, 27–48.

Tsalavoutas, I. (2011). Transition to IFRS and compliance with mandatory disclosurerequirements: What is the signal? Advances in Accounting, incorporating Advancesin International Accounting, 27, 390–405.

123P.E. Dimitropoulos et al. / Advances in Accounting, incorporating Advances in International Accounting 29 (2013) 108–123

Tsalavoutas, I., & Evans, L. (2010). Transition to IFRS in Greece: Financial statement ef-fects and auditor size. Managerial Auditing Journal, 25, 814–842.

Tyrrall, D., Woodward, D., & Rakhimbekova, A. (2007). The relevance of InternationalFinancial Reporting Standards to a developing country: Evidence from Kazakhstan.The International Journal of Accounting, 42, 82–110.

Vafeas, N., Trigeorgis, L., & Georgiou, X. (1998). The usefulness of earnings in explainingstock returns in an emerging market: The case of Cyprus. The European AccountingReview, 7, 105–124.

Van Tendeloo, B., & Vanstrelen, A. (2005). Earnings management under German GAAPversus IFRS. The European Accounting Review, 14, 155–180.

Vuong, Q. H. (1989). Likelihood ratio tests for model selection and non-nested hypoth-eses. Econometrica, 57, 307–333.

Wang, D. (2006). Founding family ownership and earnings quality. Journal of Account-ing Research, 44, 619–656.

Warfield, T., Wild, J., & Wild, K. (1995). Managerial ownership, accounting choices, andinformativeness of earnings. Journal of Accounting and Economics, 20, 61–91.

Watts, R. L., & Zimmerman, J. L. (1990). Positive accounting theory: A ten year perspec-tive. The Accounting Review, 65, 131–156.

White, H. (1980). A heteroscedasticity consistent covariance matrix estimator and adirect test for heteroscedasticity. Econometrica, 48, 817–838.

Young, S. (1999). Systematic measurement error in the estimation of discretionaryaccruals: An evaluation of alternative modelling procedures. Journal of BusinessFinance and Accounting, 26, 833–862.