The impact of IFRS on accounting quality: Evidence from Greece

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

Contents lists available at SciVerse ScienceDirect

Advances in Accounting, incorporating Advances in International Accounting journal homepage: www.elsevier.com/locate/adiac

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, Greece Hellenic Open University, School of Social Sciences, Bouboulinas 57–59 Str., 26222 Patras, Greece c Aristotle University of Thessaloniki, Department of Economics, University Campus, P.C. 54124, Thessaloniki, Greece d International Hellenic University, School of Economics and Business Administration, 14th klm Thessaloniki-Moudania, 57101 Thessaloniki, Greece b

a r t i c l e

i n f o

Keywords: IFRS Value Relevance Asymmetric Timeliness Earnings Management Greece

a b s t r a c t This paper examines the impact of IFRS adoption on the quality of accounting information within the Greek accounting setting. Using a sample of 101 firms listed in the Athens Stock Exchange (ASE) for a period of eight years (2001–2008) we find convincing evidence that the implementation of IFRS contributed to less earnings management, more timely loss recognition and greater value relevance of accounting figures, compared to the local accounting standards. Also, our findings document that audit quality further complements the beneficial impact of IFRS since those companies that are audited by Big-5 audit firms exhibit higher levels of accounting quality. Our findings are robust in regard to different model specifications and after controlling for 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 Reporting Standards (hereafter IFRS). The research questions we address in the present paper are: first, whether IFRS have mitigated earnings management behavior in Greek listed firms; second, whether IFRS influence the relevance of accounting information (earnings and book values) compared to the Greek accounting standards and third, we test whether the adoption of IFRS impacts on the level of reporting conservatism in financial statements. The primary motivation for this study is to further our understanding of the influence of IFRS on the quality of reported earnings. Overall, IFRS are principle-based standards that are market-oriented and require extensive disclosure in comparison with prior standards, i.e. local generally accepted accounting principles (GAAP). The International Accounting Standards Board (IASB) removed allowable accounting alternatives (which existed in most countries under their respective GAAPs) and required accounting measurements that better reflect a firm's economic position and performance. Based on this, in theory, the adoption of IFRS should significantly restrict the ability to engage in earnings management behavior (upwards or downwards) and increase the overall quality of disclosed information. However,

⁎ Corresponding author. Tel.: +30 2731089669. E-mail addresses: [email protected] (P.E. Dimitropoulos), [email protected] (D. Asteriou), [email protected] (D. Kousenidis), [email protected] (S. Leventis). 1 Tel.: +30 2310 807 541. 0882-6110/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.adiac.2013.03.004

Barth, Landsman, and Lang (2008) also note that the opposite may be true. For example, limiting managerial discretion relating to accounting alternatives could eliminate a firm's ability to report accounting measurements that are more reflective of the firm's economic position and performance. They also note that if enforcement of these standards is lax, companies could still engage in earnings management. In the current literature there are two opposing views regarding the influence of IFRS on accounting quality. On the one side, researchers argue that IFRS improve the reliability of financial reporting by limiting opportunistic managerial discretion (Ashbaugh & Pincus, 2001; Barth et al., 2008; Ewert & Wagenhofer, 2005). The contrary view is that the flexibility inherent in IFRS and lax enforcement might provide greater opportunity 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, more aggressive reporting of accruals and a reduction in timeliness of loss recognition relative to gain recognition. They attribute this finding to the minimal implementation guidance on IFRS, to managers' incentives to manipulate earnings and to weak enforcement mechanisms. Similar to conflicting views on the influence of IFRS on accounting quality, as noted above, research on the influence of IFRS on earnings management has also provided mixed results. Van Tendeloo and Vanstrelen (2005) for example, found that earnings management behavior was not significantly different between German companies that adopted IFRS when compared to those that relied on German GAAP. However, these findings are not corroborated by Barth et al. (2008) who, using a sample of 21 countries, concluded that in the post-adoption period firms applying IFRS displayed significantly less

P.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 the quality of accounting information has increased within the EU in terms of less targeted earnings management, smaller magnitude of absolute discretionary accruals and higher accruals quality. Similar evidence, documented by Iatridis (2008, 2010) in the UK and further verified by Doukakis (2010) in Greece, suggests that IFRS has impacted on the variability and persistence of earnings and a reduction of income smoothing practices, which obviously results in higher reporting quality. Part of the motivation for this study is to shed further light on these competing views with regard to the efficiency of IFRS in reducing earnings management and improving earnings quality within an accounting framework characterized by a long history of historicalaccounting principles. According to Tsalavoutas and Evans (2010), the Greek accounting setting is heavily based on state regulation and on certain rules associated with an increased monitoring cost. There is a close link between taxation and accounting rules which, in turn, creates increased incentives for creative accounting and earnings management practices. Ding, Hope, Jeanjean, and Stolowy (2007) point out that Greece is the country with the highest divergence of issues between its local GAAP and IFRS and that this divergence is closely related to the distinctive culture of this country. Hofstede (1980) argues that Greek accountants prefer to reduce uncertainty by abiding by the rules and Ballas, Skoutela, and Tzovas (2010) and Tsakumis (2007) also document that Greek accounting professionals are “characterized by the acceptance of inequality and a preference for achievement and material success, focusing on the group which they belong rather than on themselves”. Within this cultural framework the introduction of IFRS, which are principle-based standards, is expected to create a breach of this cultural status quo and thus the impact of this new regulation on the quality of published information remains an open empirical question. In addition, Greece presents a unique economic environment which differs significantly from other code-law countries. Greece is a bankoriented capital market, a fact which aligns corporate reporting towards the protection of the creditors, while managers pay less attention to minority shareholders (Ballas et al., 2010). Contrarily, IFRS are accounting standards which focus more on the needs of investors rather than the creditors. Therefore, considering the countervailing cultural and economic environment of Greece, the adoption of IFRS by the Greek government introduced several provisions which shake the established accounting practices of the last 25 years, such as increased levels of disclosure, the issues covered by accounting rules and the broadened role of corporate boards in evaluating accounting values. Whether the adoption of IFRS improved accounting quality within this distinctive environment remains an issue which warrants further investigation. Furthermore, the quality of accounting information is closely related to the quality of the external audit. Tsalavoutas and Evans (2010) document that large multinational audit firms (Big-5) enjoy greater independence from their clients in Greece and have greater experience and resources for implementing IFRS, and of course face greater reputation costs or litigation risks. These characteristics contribute 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 quality is motivated by reputation costs and since Big-5 audit firms face greater litigation costs they also produce higher quality audits. Furthermore, a recent study by Tsalavoutas (2011) suggested that Big-5 audit corporations act as monitoring mechanisms contributing to increased transparency, better quality financial statements and higher levels of compliance with IFRS. Thus, our study adds to a growing body of literature since it examines if and how the quality of reported accounting information is influenced in any way by the implementation of IFRS, taking into consideration the quality of audit. We provide

109

additional evidence to the ongoing debate about international accounting standards and accounting quality. Our sample comprises of 101 firms listed in the ASE covering an eight-year period from 2001 until 2008 where 25 had voluntarily adopted IFRS prior to the initial enforcement in 2005 and the remaining 76 adopted IFRS when mandatorily enforced. We examine whether accounting information during the post-IFRS period (2005– 2008) exhibits less earnings management, more timely loss recognition and higher value relevance, when compared to the relative amounts during the pre-IFRS period (2001–2004). After controlling for firmspecific characteristics such as size, growth opportunities, risk and audit quality we found that the adoption of IFRS contributed to less earnings manipulation (lower magnitude of discretionary accruals), more timely loss recognition and greater value relevance of earnings and book values, compared to the local accounting standards. Also, our findings document that audit quality further complements the beneficial impact of IFRS since those companies that are audited by the Big-5 multinational audit firms exhibit higher levels of accounting quality compared to their non-Big-5 counterparts.

2. Greek accounting standard setting and IFRS An extensive amount of research has examined the value relevance of accounting information within countries under different accounting regimes. Hope (2003) argues that the quality of financial accounting information depends on both the quality of accounting standards and the regulatory enforcement, since even the best accounting standards are inefficient if not enforced adequately. The Greek accounting system dates back to 1835 when the first independent Greek Republic was established after almost 400 years of Ottoman occupation. At that time the French commercial code was translated into Greek and adopted in order to serve as the basis of Greek commercial law. The companies act law was introduced in 1920 and modified in 1955 and it still exists now to a great extent. Accounting regulation in Greece was imposed by the government and specifically by the Minister of Economics and Finance. However, in 1988 the National Accounting Council was established in order to be entrusted with the issuance of more detailed regulations. In contrast to other countries, the Greek National Accounting Council had limited decision making authority and its function was clearly advisory to the government. Before the council could express an opinion it had to follow a process of consultation with the Economic Chamber of Greece which could be very time-consuming (Ballas, 1994). In June 2003, the National Accounting Council was abolished and in its place the government established the Board of Accounting Standards which forms one division of the Committee of Accounting Standards and Control (along with the Board of Quality Control and the Executive Committee). The Board of Accounting Standards is governed by a five-member board of directors and its main responsibility is to provide consultative services on issues of accounting standardization regarding Greek accounting standards and the codification of IFRS. Also, it has the authority to publish directives regarding the implementation of Greek GAAP and IFRS and also clarifications on some technical accounting issues. These advisory decisions need to be approved by the Committee of Accounting Standards and Control before they 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 market and the only supervision performed was once again by the Ministry of Economics. As a result, until that period the accounting regulations promoted financial disclosure for tax purposes only, since the stock market was unregulated and investment interest was quite limited. Furthermore, the overall setting was primarily used as an instrument of government policy — most notably tax policy, but also wage and pricing 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 reforms have taken place on many levels of financial regulation. First, the Greek government established the Hellenic Capital Market Commission which was entrusted with the governing of the stock market. In doing so the government wanted to establish a sound regulatory framework for the evolvement of the stock exchange as an integral part of economic growth. Additionally, in 1992 the Code of Books and Records was enacted, determining the financial statements which firms must report and the specific details of accounts that must be disclosed in the financial statements. Finally, in 1993 the audit profession in Greece was liberalized by allowing foreign audit companies to operate within the country, something which was forbidden prior to 1993. This fact helped the evolution of the accounting profession and the stability of the capital market since those firms increased the creditworthiness of the stock market by providing accurate and high-quality audits. As Ballas and Fafaliou (2008) argue, the concentration of the audit market in Greece shows a significant increase, especially after 2001 where the Big-5 multinational audit corporations are equally represented and control almost 67% of statutory audits within the capital market. The same phenomenon exists in Germany (though the relevant concentration ratio is less than the Greek context) and it is attributable to the regulatory reform that took place in Germany as it conformed to the IFRS, causing the market share of local audit firms to decrease, a fact expected 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 accounting setting is based on the General Greek Accounting Plan (GGAP) which emphasizes financial reporting conformity with tax rules, the protection of stakeholders, and conservatism. The Greek accounting system has been highly conservative due to the fact that the setting of accounting standards remains the sole responsibility of the state (since the Board of Accounting Standards plays only an advisory role in the standardization process and the final decision and approval lie within the Ministry of Economics). This has, and in part due to the large number of small and family-owned businesses, limited the goal of financial reporting to provide only tax-relevant information (Kousenidis, Ladas, & Negakis, 2009). Greek firms obviously have strong incentives to reduce taxes and, consequently, financial accounting information is less likely to reflect economic realities when the firm's goal is to minimize taxes. This assertion has been verified by Guenther and Young (2000) who argue that, in countries where there is increased conformity between financial accounting and tax accounting rules, accounting information may differ from underlying economic activities. Additionally, certain rules imposed by the Greek accounting standards give managers greater flexibility than IFRS for executing income smoothing techniques such as the capitalization of start-up costs and their amortization within a five year period. The most significant change that has taken place in the Greek accounting setting is the adoption of International Financial Reporting Standards (IFRS) for all listed firms since January 2005 which is characterized as the second cornerstone of accounting legislation in Greece after the adoption of the Hellenic General Accounting Plan (HGAP) in 1980. This new framework impacted on the stakeholder-oriented and tax-driven nature of the abovementioned accounting setting. The most important implication of IFRS was the introduction of the fair value principle (replacing the long-established historical-cost principle) to asset valuation and liability recognition. The rationale behind the fair value principle is that the disclosed values of accounting quantities should be close to their relative market values, thus financial accounting information is more likely to reflect true economic events. IFRS also reduced managers' flexibility in valuing assets at the lowest amount possible so as to minimize tax liabilities. Moreover, IFRS imposed changes on many technical accounting issues such as the presentation of financial statements, segment reporting, intangible assets, depreciation, related party disclosures etc. (see Spathis & Georgakopoulou, 2007 for

details), aimed at promoting a ‘true and fair’ presentation of financial information to facilitate investors' rational investment decisions. IFRS are considered to be high-quality standards and their implementation in countries with accounting settings based on historical-cost values (such as code-law countries) could help to improve the quality of financial accounting information (Hitz, 2007; Penman, 2007). Whether this assertion is true within the Greek accounting 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 develop accounting standards that are internationally acceptable and of a high quality. In order to achieve this goal, the IASB has issued principle-based standards and removed allowable alternatives in accounting estimations, thus leaving little room for managerial discretion by requiring accounting measurements to better reflect a company's economic position and performance. However, there seems to be a long debate over whether IFRS are able to achieve the aforementioned goal and to what extent. Several studies in the field have focused on different aspects of accounting quality, namely value relevance, conditional conservatism and earnings management. Referring to value relevance (or how efficiently publicly available accounting information is incorporated into stock returns), Ashbaugh and Pincus (2001), Ewert and Wagenhofer (2005), Lang, Raedy, and Yetman (2003); Lang, Raedy, and Wilson (2006), Leuz, Nanda, and Wysocki (2003) and Bartov, Goldberg, and Kim (2005) provide evidence that reported earnings based on IFRS are more value-relevant than domestic GAAP. Additionally, Djatej, Gao, Sarikas, and Senteney (2009) document that the implementation of IFRS increases the quality of public information and that this increase is more pronounced for East European countries. This suggests that IFRS reduce information asymmetry by moving important performance-related information from the private into the public domain. In the same vein, Beuselinck et al. (2009), Devalle, Onali, and Magarini (2010), Aharony, Barniv, and Falk (2010) and Landsman, Maydew, and Thornock (2010) corroborate the above findings by providing evidence of an improvement in stock price informativeness in the EU after the adoption of IFRS. Moreover, Dargenidou and McLeay (2010) support the above findings by demonstrating that mandatory IFRS adoption results in accounting estimates that reflect the underlying economics of firms within financial markets in a more timely and comparable manner. Also Barth et al. (2008) verify this hypothesis by examining the value relevance of income and the book value of equity between firms following IFRS and local GAAP. Their findings indicate a higher association of accounting figures with stock returns for those firms that adopted IFRS. Another way of assessing the quality of accounting information is to examine to what extent firms withhold the recognition of losses with the intention to mislead stakeholders about the true economic performance of the company, or to affect contractual outcomes that are tied to accounting earnings. Referring to the issue of timely loss recognition 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 of higher-quality earnings is that large losses are recognized as they occur instead of being deferred to future periods. Barth et al. (2008) using a sample from 21 countries found that firms operating under IFRS exhibit more timely recognition of losses compared to the nonIFRS firms. Also, Guenther, Gegenfurtner, Kaserer, and Achleitner (2009) document that conditional conservatism increased after the IFRS implementation for both mandatory and voluntary adopters in Germany. Moreover, a number of studies suggest that the quality of reported financial statements is determined by the economic and institutional factors influencing manager and auditor incentives. Leuz et al. (2003) document that earnings management (measuring the magnitude of

P.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 compared to common-law countries. This difference is attributed to the relative costs of earnings management being less than the benefits. IFRS promote an accounting environment within which firms have incentives to report investor-oriented information and thus engage significantly less in earnings management (compared to the local GAAP). Barth et al. (2008) verify this assertion by finding evidence of less earnings management for firms adopting IFRS compared to their non-IFRS counterparts. A characteristic of the aforementioned studies is that they use discretionary accruals and earnings benchmarks as measures of earnings manipulation. Therefore, we expect that Greek firms operating during the post-IFRS period (2005–2008) will publish financial statements which will present less earnings management compared to the pre-IFRS era. We also assume that firms engaging in earnings smoothing will present greater values of the ratio of standard deviation (SD) of income to the standard deviation (SD) of cash flows (Ahmed et al., 2010). Ahmed et al. (2010) declare that if managers use accruals to smooth changes in cash flows when reporting income, the variance of the change in net income should be less than the variance of the change in cash flows. Accordingly, less positive values of this ratio will indicate greater income smoothing. Thus, firms operating in the post-IFRS era might present greater values of the ratio of SD of income to the SD of cash flows. Also Ding et al. (2007) document that countries with low quality accounting standards (such as Greece) are associated with greater opportunities for earnings management and they suggest that the expansion of coverage of accounting issues by IFRS can curb this behavior. However, there are a series of research papers which provide exactly the opposite picture regarding the efficiency of IFRS on increasing accounting quality. Van Tendeloo and Vanstrelen (2005), find no differences in earnings management behavior between firms applying IFRS and local GAAP in Germany. Also Daske (2006), and Hung and Subramanyam (2007) found no evidence of a reduction in the cost of capital and an increase in accounting value relevance in amounts prepared under IFRS as opposed to German GAAP. Additionally, Eccher and Healy (2003) and Callao, Jarne, and Laínez (2007) argue that accounting figures based on IFRS are not more value-relevant when compared to relative amounts prepared under Chinese and Spanish GAAP respectively. Finally, relevant research on the emerging capital markets of Kazakhstan and Egypt reported by Tyrrall, Woodward, and Rakhimbekova (2007) and Hassan, Romilly, Giorgioni, and Power (2009) respectively found evidence of a weak relation between IFRS adoption and accounting value relevance as well as a negative relationship between IFRS and firm value. Furthermore, Callao Gastón et al. (2010) examined the impact of mandatory IFRS adoption on financial reporting in the UK and Spain and they found a reduction in the relevance of financial reporting in both countries. Overall, there exist two groups of studies with conflicting evidence regarding the ability of IFRS to enhance accounting quality. Barth et al. (2008) argue that the main reasons for the abovementioned conflicting findings are the gradual transition from local GAAP to IFRS; the lack of infrastructure to enforce the application of IFRS; differences in the effectiveness of controls regarding the economic environment and corporate incentives; and, finally, differential time periods and data metrics. Our study tries to mitigate these problems by examining the impact of IFRS on accounting quality within a period (2005–2008) of obligatory IFRS enforcement (not voluntary as in many prior studies), while simultaneously controlling for voluntary adopters; and also a period when the transition to IFRS from local GAAP has been made thoroughly, since Greek firms were obliged to report their 2004 financial statements in both local GAAP and IFRS formats in order to give investors a visualization of the upcoming accounting changes. Moreover, our study is focused on a single code-law country. This fact has the advantage of removing the need to control for the potentially confounding effects of country-specific factors that

111

are not related to the accounting system. Also, as Hitz (2007) and Penman (2007) argue, the implementation of IFRS on countries with accounting 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 market report that earnings value relevance has increased and earnings management decreased after IFRS adoption; and that, overall, IFRS increased the reliability, transparency and comparability of the financial statements (Ballas et al., 2010; Iatridis & Rouvolis, 2010). However, as yet no Greek study has addressed the issue of audit quality, which is a crucial element of quality financial reporting, through the efficient implementation of IFRS (Hodgdon, Tondkar, Adhikari, & Harless, 2009). Therefore we expect that, after controlling for the aforementioned problems, the adoption of IFRS within the Greek accounting setting has helped to enhance the quality of accounting information. As Barth et al. (2008) argue, the advantages of this procedure are first that we can determine the source of accounting quality differences between the two sample periods before and after the implementation of IFRS. Second, we can rule out predictions for some of our metrics based on predictions from another metric, thus having a clearer picture of the impact of IFRS on accounting quality. Enhanced accounting quality will be translated as an increase in value relevance and accounting conservatism and a reduction of earnings management behavior during the post-IFRS period. Thus, our first hypothesis states: H1. Financial statements prepared under IFRS are of a higher quality than those prepared under the Greek GAAP. Additionally, as previously discussed, the quality of accounting information 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 and Pincus (2001) predict and find convincing evidence that Earnings Response Coefficients (ERCs) are a positive function of perceived auditor quality. This practically means that firms audited by the Big-5 multinational auditing corporations report earnings and book values of common equity that have higher value relevance (higher association with stock returns). Moreover, multinational audit firms have more resources (knowledge and specialized personnel) and IFRS-related experience and could provide greater assistance in the implementation and transition to IFRS compared to the local audit firms. Also, recent evidence by Hodgdon et al. (2009) and Tsalavoutas (2011) indicate that IFRS compliance is positively associated with audit quality (defined as being audited by Big-5 audit corporations). Their evidence reinforces the importance of multinational audit firms and quality audits as a means of encouraging compliance with IFRS. A recent study by Iatridis (2011) in the UK indicated that firms audited by the Big-5 audit corporations are associated with high quality accounting disclosures and are less prone to earnings manipulation. Along the same line, Tsalavoutas and Evans (2010) declared that IFRS have improved the quality of accounting information in the Greek capital market since they report a statistically significant association between audit quality and the transition to IFRS by Greek listed firms, suggesting that Big-5 audit corporations were able to attract experienced employees from their foreign operations to assist in the transition and implementation of IFRS. Thus, we expect that companies audited by Big-5 audit firms will present a higher value relevance of earnings and book values during the post-IFRS period compared to their non-Big-5 audited counterparts. Furthermore, firms audited by the Big-5 audit corporations will present more timely loss recognition since multinational audit firms are more likely to follow a prudent policy towards loss-recognition to avoid any future reversals on profitability that could be a serious strike to their reputation. Therefore, multinational audit corporations have greater incentives for recognizing losses as they occur instead of being deferred to future periods. Also, a series of studies has shown

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 audit corporations) constitutes a constraint on earnings management reporting (Bauwhede, Willekens, & Gaeremynck, 2003; Francis & Wang, 2003; Gore, Pope, & Singh, 2001; Iatridis, 2011). Street and Gray (2001, 2002), Glaum and Street (2003) and Prather-Kinsey and Meek (2004) all find support for the hypothesis that companies audited by a large audit firm are positively associated with IFRS compliance. Caramanis and Lennox (2008) demonstrate that Big-5 audit firms work more hours compared to non-Big-5 firms in the Greek capital market, and companies audited by Big-5 audit firms report a smaller magnitude of discretionary accruals, thus engaging in less earnings management in comparison to companies audited by nonBig-5 corporations. Additionally, Leventis and Caramanis (2005) provide evidence that audit effort in Greece is correlated with audit firm size. Also Frankel, Johnson, and Nelson (2002) document that discretionary accruals can be viewed as evidence of audit independence since independent auditors require their clients to report unmanaged financial statements. Therefore, we assume that Big-5 audit firms provide better quality audits than local audit firms, meaning that they have greater abilities and incentives to enforce more transparent financial reporting after the adoption of high-quality accounting standards like IFRS. Thus, based on the above discussion we formulate our second hypothesis as follows: H2. Accounting quality will be higher in the post-IFRS period for firms with enhanced audit quality.

Table 1 Sample data selection procedure and sample description. Panel A: Sample selection procedure Firms listed in the Athens Stock Exchange (ASE) Less: Financial services firms Remaining non-financial firms Less: Firms with incomplete accounting and stock price data Non-financial firms with full accounting and stock price data Less: Non-December fiscal year end firms Firms included in the final sample

294 (43) 251 (125) 126 (25) 101

Panel B: Year and industry distribution Sample distribution by year

Sample distribution by industry sector

Year

Observations

Industries

Observations

2001 2002 2003 2004 2005 2006 2007 2008

101 101 101 101 101 101 101 101

Total

808

Raw materials Constructions Industrial products and services Food and beverages Personal and household products Health corporations Trade corporations Media corporations Tourism and leisure Telecommunications and technology Services Real estate Total

40 108 108 96 84 40 72 40 48 76 64 32 808

4. Data and methodology 4.1. Data selection procedure Our study uses a sample consisting of 101 companies, all listed on the ASE, where 76 followed the mandatory adoption of IFRS in 2005 and the other 25 were early voluntary adopters. All companies had full annual data of reported earnings and stock prices, sales, assets, total debt, property-plant and equipment, cash flows and common equity during the period 2001–2008. We have excluded banks and financial services firms due to the different operations and accounting policies. The classification of similar firms was made according to the industry classification proposed by the ASE. Data were collected from the ASE database. The initial sample contained 126 companies with full data for the period under investigation but we restricted the group to companies with a December fiscal year-end limiting the final number to 101, resulting in 808 firm-year observations. Annual stock prices exclusive of dividends at the end of each year and annual earnings per share were used. The earnings and price data were adjusted for stock splits, stock dividends and stock issues. In order to eliminate the effect of outliers, we deleted the top and bottom 1% of all variables during the pre and post-IFRS periods. All variables in the sample were deflated by the year's average total assets (measured as the mean value of the assets at the beginning and at the end of the fiscal year). Table 1 describes the data selection procedure and the synthesis of the final sample. The sample is quite dispersed with the majority of the observations situated within the construction and manufacturing sectors, followed by personal products and telecommunications. 4.2. Testing for value relevance Our approach in this section included testing whether the relation between earnings per share and the book value of equity per share with stock prices varies between local accounting principles and IFRS. Our value relevance metric is based on the explanatory power of a regression of stock prices on net income and the book value of equity. In order to obtain a measure of stock price that is unaffected by mean differences across industries (which would affect the explanatory power of the comparisons) we applied the methodology

proposed by Barth et al. (2008) and regressed stock price P on industry fixed effects. The next action was to model the residuals from the previous step P* as the dependent variable on book value of equity per share (BVPS) and net income per share (NIPS) separately for the pre and post-IFRS periods. In order to ensure that accounting information has been fully disclosed to the market we followed Lang et al. (2003, 2006) and estimated P six months after the fiscal year end. In the next model, and all the following, we pooled both mandatory and voluntary IFRS adopters in order to control for incremental differences between the two groups of firms, so as to extract more salient inferences regarding the impact of IFRS on accounting quality (Byard, Li, & Yu, 2011). The general form of the regression model is as follows: P
it ¼ α0 þα1 NIPSit þα2 BVPSit þα3 DVOLit þα4 DVOLit
NIPSit þα DVOL
BVPS þβControl þ γYear dummies 5

it

it

it

ð1Þ

þδIndustry dummies þ εit where: P* NIPS BVPS DVOL

is the residuals from a regression of stock prices on industry fixed effects, is net income per share, is the book value of common equity per share, 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, growth and profitability and finally we took into consideration the year and industry effects of any omitted variables. The industry dummies follow the ASE industry classification code and we classified firms in our sample according to the sectors depicted in Table 1. Model 1 is run separately 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. If hypothesis H1 is valid, net income and the book value of equity will be of higher value relevance in the post-IFRS period, thus we expect an increase in R2 during the post-IFRS period compared to the

P.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 have benefited from the early adoption of IFRS, thus we expect their accounting numbers to be more value relevant during the pre-IFRS period (2001–2004) compared to the mandatory group. Additionally, since the question in this research is whether the quality of accounting information has improved after the adoption of IFRS, the two accounting standards (Greek GAAP and IFRS) are set up as competing non-nested models. For this reason we followed Dechow (1994) and controlled for the change in R 2 between the pre and post-IFRS periods by employing the Vuong (1989) likelihood ratio test. The test examines the null hypothesis that the two sub-periods are equally capable of explaining the true data against the alternative that one period is closer. 2 According to Dechow (1994), the Z-statistic has several advantages over other non-nested tests 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-IFRS period a positive and significant Z-statistic suggests that the residuals produced by the regression during the pre-IFRS period are larger in magnitude than those during the post-IFRS period. Consequently, a positive and significant Z-statistic indicates that IFRS adoption has contributed to increased accounting quality. Moreover, in order to control for hypothesis H2 we introduced a dummy variable AUD which takes the value of one (1) if the firm is audited by a Big-5 international audit firm in a given year and zero (0) otherwise. By this means we examine whether companies audited by Big-5 audit firms present a higher value relevance of earnings and book values during the post-IFRS period compared to their non-Big-5 audited counterparts. Therefore, model 1 is modified to the following form: P
it ¼ α0 þα1 NIPSit þα2 BVPSit þα3 DVOLit þα4 DVOLit
NIPSit þα5 DVOLit
BVPSit þa6 AUDit þα7 AUDit
NIPSit þα8 AUDit
BVPSit þβControlit þ γYear dummies þ δIndustry dummies þ εit

ð2Þ

where: AUD*BVPS and AUD*NIPS are the interaction terms between AUD dummy and book value per share and net income per share respectively, and DVOL*NIPS and DVOL*BVPS are the interaction terms of NIPS and BVPS with the DVOL variable. As in the previous model 1, model 2 is estimated separately for the pre and post-IFRS periods. H2 is accepted if the explanatory power of model 2 in the post-IFRS period is higher compared to model 2 in the pre-IFRS period and also if coefficients a4 and a5 are significant and have higher magnitude relative to coefficients a1 and a2 (again in the post-IFRS period). 4.3. Testing for timely loss recognition Our next test includes the examination of the link between timely loss recognition and IFRS adoption. In order to determine whether the implementation of IFRS leads firms to disclose the recognition of losses in the period under study we applied Basu's (1997) reverse regression model between earnings per share and stock returns. For this purpose we estimated 12-month stock returns over the fiscal year in order to proxy for the news regarding a firm's performance that is publicly available. Basu's (1997) introduces a dummy variable (D) in the reverse regression model based on the assumption that earnings are expected to be more highly correlated with market 2 We are grateful to Jackson Caskey (UCLA) for programming this test on STATA and making it available for use (http://personal.anderson.ucla.edu/judson.caskey/data. html).

113

returns in periods of depression in market values than in periods of prosperity in market values. This qualitative variable is given the value of one (1) when returns are negative and zero (0) if they are positive. Under these assumptions the final form of the Basu's (1997) model is as following: EYi;t ¼ b0 þb1
Rit þb2
Dit þb3
Rit
Dit þb4 DVOLit þb5 Dit
DVOLit þb R
DVOL þb R
D
DVOL þβControl þ γYear dummies 6

it

it

7 it

it

it

it

þδIndustry dummies þ eit

ð3Þ

where: EY R D

is the earnings yield estimated as earnings per share deflated by the stock price at the beginning of the fiscal year, is the buy and hold stock return over the fiscal year, is a dichotomous variable receiving one (1) when R is negative and zero (0) otherwise and the other variables are defined as previously stated.

We have included control variables for size, risk, growth and profitability as well as year and industry dummies. The coefficients on R (b1) and R*D (b3) capture the reaction of earnings 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’, the lagged effect may appear as a persistent shock. Therefore, the intercept, b0, is expected to have a positive sign. In other words, if conservatism affects earnings downwardly the intercept is expected to have a positive sign. The coefficient on D, b2, is explained as a reversal of the previous year's market information in the light of current value decreases. When b2 is positive, over-provisioning is reversed as a prior year adjustment and, when b2 is negative, deferred income recognition is scaled down (Raonic, McLeay, & Asimakopoulos, 2004). Model 3 is run separately for the pre and post-IFRS periods and if IFRS have increased the timely recognition of losses we expect a significant increase in the explanatory power of the model (which will be tested using Vuong's (1989) likelihood ratio test) and coefficient b3 to be positive and have a higher magnitude relative to the pre-IFRS period. Also, coefficient b7 is expected to be positive and significant during the pre-IFRS period and to have a larger magnitude compared to coefficient b3 indicating that voluntary IFRS adopters are more conservative during the pre-IFRS period relative to firms following the Greek GAAP. Moreover, we have extended the Basu (1997) model for examining the effect of audit quality on timely loss recognition, by introducing a dummy variable AUD into the model which takes the value of one (1) if the firm is audited by a Big-5 audit firm and zero (0) otherwise. The new variable was then interacted with the other variables in the standard Basu (1997) model as depicted in the following equation: EYit ¼ b0 þb1 Rit þb2 Dit þb3 Rit
Dit þb4 DVOLit þb5 Dit
DVOLit þb R
DVOL þb R
D
DVOL þb AUD þb D
AUD 6 it

it

7 it

it

it

8

it

9

it

it

þb10 Rit
AUDit þb11 Rit
Dit
AUDit þβControlit þ γYear dummies þδIndustry dummies þ eit :

ð4Þ

Model 4 is estimated again separately for the pre and post-IFRS periods. Hypothesis H2 expects that firms audited by the Big-5 audit corporations will present more timely loss recognition since multinational audit firms are more likely to follow a prudent policy towards loss recognition considering that they have greater incentives. If H2 is true we expect this to be represented by a significant positive coefficient 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

ΔSalesit

Our first test on earnings management is based on the ratio of the standard deviation of the change in net income (ΔΝΙ) to the standard deviation of the change in cash flows (ΔCF). We performed this test for both sub-groups of firms based on the voluntary or mandatory adoption of IFRS. In order to capture any confounding effects by factors that are not attributable to the financial reporting setting, we followed Barth et al. (2008) and Ahmed et al. (2010). Thus, we compared the ratio of the standard deviation of the change in net income (ΔΝΙ*) to the standard deviation of the change in cash flow (ΔCF*) residuals from the following two regression models, instead of comparing the SD of the change of NI and CF directly. The relative models have the following form:

PPEit ROA TAit

is the change in net sales deflated by lagged total assets (Salesit − Salesit − 1) is the level of property plant and equipment for each year deflated by lagged total assets is the end of year return on assets estimated as net income over total assets is firm's total assets at the end of the fiscal year.

where:

As Kothari et al. (2005) argue, the inclusion of a constant term in the Jones (1991) model provides an additional control for heteroscedasticity not alleviated by deflating the variables with total assets. A constant term also mitigates problems arising from omitted size variables and produces discretionary accrual measures that are more symmetric, overcoming model misspecifications and making the power of the test comparisons more clear. Finally, the inclusion of a profitability measure (ROA) is designed to enhance the effectiveness of the performance matching methodology. The discretionary accruals from the modified Jones (1991) model were defined as the residuals from estimating Eq. (7):

ΔΝΙ

^ ðΔSALES =TA Þ ^ ð1=TAit1 Þ þ β DACit ¼ ACCit =TAit1 α it it1

ΔΝΙit ¼ a0 þa1 SIZEit þa2 GROWTHit þa3 LEVit þa4 AUDit þeit

ð5Þ

ΔCFit ¼ a0 þa1 SIZEit þa2 GROWTHit þa3 LEVit þa4 AUDit þeit

ð6Þ

is the annual change in net income divided by lagged total assets ΔCF is the annual change in the operating cash flow divided by lagged total assets SIZE is the natural logarithm of end year total assets GROWTH is the percentage change in sales LEV is the ratio of end year total liabilities to end year total common equity AUD 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 to smooth the changes in cash flows when reporting income, the variance of the change in net income should be less than the variance of the change in cash flows. So, less positive values of this ratio would indicate greater income smoothing and more earnings manipulation. Therefore, listed firms during the IFRS adoption period are expected to have higher values of the ratio SD(ΔΝΙ*)/SD(ΔCF*). In order to test for the differences in the earnings smoothing metric (SD(ΔΝΙ*)/SD(ΔCF*)), we followed Barth et al. (2008) and used a t-test based on the empirical distribution of the differences. Specifically we randomly selected (with replacement) firm observations that were assigned to each sub-group (voluntary and mandatory adopters) during the two sample periods and calculated the difference between the two types of firms in that metric. We then repeated this procedure 1000 times and obtained the empirical distribution of the difference. Finally, our second measure of earnings management aims to examine the impact of IFRS adoption and audit quality on the magnitude of performance-matched discretionary accruals. For this reason we estimated the cross-sectional Jones (1991) model, as modified by Kothari, Leone, and Wasley (2005), in order to extract the discretionary or abnormal accruals following a performance matching approach. This model estimates discretionary accruals as a function of changes in sales, the levels of property, plant and equipment, and the level of return on assets by estimating the following OLS equation: ACCit =TAt1 ¼ a0 þ αð1=TAt1 Þ þ βðΔSalesit =TAt1 Þ þ γðPPEit =TAt1 Þð7Þ þgROAit þeit where: ACCit CF

is total accruals defined as the difference between net income and is the divided by lagged total assets.

ð8Þ

^ ðPPEit =TAit1 Þ þ g^ ROA: þγ In order to estimate the performance-matched discretionary accruals we followed Kothari et al. (2005) and matched each firm year observation with another from the same industry classification and year, and with the closest return on assets (ROA). The final step was to define the Jones model performance-matched discretionary accruals for firm i in year t as the discretionary accruals from Eq. (8) minus the matched firm's Jones model discretionary accruals for year t. As Kothari et al. (2005) argue, the performance matching approach based on ROA and using the Jones model produces a more salient measure of discretionary accruals since the means and medians in performance-related sub-samples are closest to zero more often 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 involve either income increasing or income decreasing accruals to meet earnings 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 the financial statements and thus a lower earnings quality. Also following Ahmed et al. (2010), we estimate model 9 using the signed value of discretionary accruals so as to control for the consistency of the results. In order to test hypotheses H1 and H2 we introduced the absolute and signed value of performance-matched discretionary accruals (DACC) respectively as the dependent variable in the following model: DACCit jDACCjit ¼ c0 þc1 IFRSit þc2 AUDit þc3 DVOLit þc4 AUD
DVOLit þc AUD
IFRS þβControl þ γYear dummies 5

it

it

ð9Þ

þδIndustry dummies þ uit 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 DVOL variables indicating that the adoption of IFRS for both voluntary and

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

115

response coefficients. Therefore we expect that leverage (as measured by the ratio of total debt to common equity) will have a negative impact on the quality of accounting information. Moreover, we controlled for the effect of growth opportunities (GROWTH) measured as the percentage change in sales from year t-1 to year t. Kumar and Krishnan (2008) document that the value relevance of earnings and cash flows increases with investment opportunities. Therefore, firms with a positive annual change in sales are expected to have more value relevant accounting information and present greater timely loss recognition. Thus, the coefficient of the GROWTH variable is expected to have a positive sign. Finally, we included the level of operating cash flows (CFO) deflated by lagged total assets as a performance measure. Dimitropoulos and Asteriou (2009) document that cash flows are equally value relevant with earnings in determining stock return movements within the Greek accounting setting. Consequently we expect a positive coefficient on the CFO variable during the pre and post-IFRS periods. Also Dechow, Sloan, and Sweeney (1996) and Young (1999) declare that the matching principle results in a natural smoothing of accruals which in turn causes negative discretionary accruals to occur in periods of extreme positive cash flows. We included the absolute value of CFO so as to control for this potential misspecification. If the abovementioned assertion is true we expect a positive relation between |DACC| and |CFO|.

mandatory adopters has resulted in less earnings management behavior. Additionally, if H2 is valid we expect to find a negative and statistically significant coefficient on the interaction terms AUD*IFRS and DVOL*AUD suggesting that Big-5 audit firms provide better quality audits than local audit firms, meaning that they have greater abilities and incentives for the enforcement of more transparent financial reporting after adopting the high-quality accounting standards of IFRS. 4.5. Control variables In all models, we included several control variables suggested by previous research, and also, because we wanted to capture differences in earnings management, timely loss recognition and value relevance incentives. First of all we controlled for firm size measured as the natural logarithm of total assets (LnTA). Watts and Zimmerman (1990) and Van Tendeloo and Vanstrelen (2005) argue that larger firms are more likely to prefer downward earnings management because the potential for government scrutiny increases as firms become larger and more profitable. Therefore, we expect a negative relation between the SIZE 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 firms since there is less media information available compared with larger firms. On the other hand, in small markets larger firms draw more publicity and are more closely followed, thus their earnings quality is likely to be known with less uncertainty compared to their smaller counterparts. Therefore, investors may place greater trust in larger firms with which they are more familiar, and thus be more responsive to their financial reports. Ultimately, the mitigating role of firm size is again an open empirical question. Additionally, we controlled for the impact of firm leverage (LEV). Van Tendeloo and Vanstrelen (2005) and Billings (1999) document that the debt-equity hypothesis predicts that highly leveraged firms are more likely to engage in upward earnings management to avoid debt covenant violations and are associated with lower earnings

5. Empirical results 5.1. Descriptive statistics and correlations Table 2 presents the descriptive statistics of the sample variables for the whole period of investigation (2001–2008) as well as during the pre-IFRS period (2001–2004) and the post-IFRS period (2005–2008) for voluntary and mandatory IFRS adopters. The sample includes 101 non-financial firms listed in the Athens Stock Exchange for the period 2001–2008, 76 are mandatory IFRS adopters and 25

Table 2 Descriptive Statistics of Sample Variables. Panel A: Pre and post IFRS mandatory adopters Variables

R P NIPS BVPS ACC DACC CFO SIZE LEV GR AUD

Full sample (voluntary and mandatory)

Pre-IFRS (mandatory)

Post-IFRS (mandatory)

Mean

Median

Std dev

Min

Max

Mean

Median

Std dev

Min

Max

Mean

Median

Std dev

Min

Max

0.074 5.232 0.063 6.321 −0.021 −0.000 0.092 5.242 8.590 3.210 0.456

−0.124 2.411 0.018 1.414 −0.132 −0.002 0.030 4.874 1.650 0.050 0.000

1.954 9.212 1.066 65.220 0.318 0.244 0.544 1.444 7.890 22.890 0.442

−0.891 0.000 −4.111 0.036 −3.519 −2.536 −3.481 1.558 −29.070 −1.000 0.000

35.66 100.00 25.36 18.96 2.40 2.34 12.52 10.97 60.37 50.96 1.00

−0.138 4.846 0.130 7.13 −0.031 0.000 0.065 4.451 16.540 0.162 0.352

−0.195 2.865 0.019 0.90 −0.025 −0.002 0.042 4.471 2.89 −0.009 0.000

0.410 7.260 1.442 101.400 0.343 0.276 0.487 1.310 12.900 2.542 0.478

−0.804 0.150 −1.203 0.100 −3.519 −2.536 −3.481 1.711 0.010 −1.000 0.000

2.02 62.50 24.55 17.90 2.40 2.34 3.48 9.00 19.85 41.50 1.00

0.247 5.135 −0.035 3.787 −0.034 −0.000 0.102 4.889 2.275 5.69 0.381

−0.063 1.851 0.038 2.172 −0.002 −0.002 0.010 5.047 1.436 0.06 0.000

2.667 9.609 0.484 7.755 0.337 0.249 0.700 1.445 7.126 36.130 0.486

−0.891 0.000 −4.111 0.036 −2.884 −2.179 −0.840 1.558 −29.070 −1.000 0.000

35.66 100.00 3.031 17.560 1.482 1.504 10.250 9.137 60.370 50.960 1.000

Panel B: Pre and post IFRS voluntary adopters Variables

R P NIPS BVPS ACC DACC CFO SIZE LEV GR AUD

Pre-IFRS (voluntary)

Post-IFRS (voluntary)

Mean

Median

Std dev

Min

Max

Mean

Median

Std dev

Min

Max

0.128 6.545 0.235 9.541 −0.035 0.005 0.159 6.547 9.256 2.559 0.614

0.102 3.214 0.122 2.121 −0.010 −0.004 0.102 4.772 2.920 1.989 0.000

0.325 7.774 1.547 87.540 0.256 0.144 0.359 2.141 10.230 2.451 0.455

−0.632 2.365 −0.789 2.361 −2.694 −1.878 1.236 2.878 0.548 2.547 0.000

12.55 39.58 25.36 18.96 1.247 1.547 12.52 10.97 11.65 30.05 1.000

0.130 6.845 0.214 10.22 −0.039 0.004 0.199 6.897 9.887 3.897 0.620

0.066 2.958 0.044 2.414 −0.007 −0.002 0.088 3.658 2.487 1.547 0.000

2.545 8.598 0.389 74.550 0.277 0.149 0.458 1.897 7.445 2.850 0.444

−0.541 2.653 −1.259 3.226 −2.549 −1.778 1.598 2.102 0.447 3.258 0.000

15.22 42.22 20.44 15.28 1.14 1.41 11.44 9.45 12.99 28.54 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 increase during the post-IFRS period to 24.7% compared to the pre-IFRS period which is negative (− 13.8%) and the same stands for the mean closing stock price (5.135). The same is true for the voluntary adopters but the difference is not as large as for the mandatory adopters. Also the ratification of IFRS impacted on the estimation of income and the book value of equity since both variables present a significant decrease during the post-IFRS period for the mandatory adopters. However, mandatory adopters seem to report less accruals in the post-IFRS period (−0.034) compared to the pre-IFRS period (−0.031) suggesting that the adoption of IFRS mitigated the distance between net income and operating cash flows leaving managers little room for discretion on the reporting of total accruals. Furthermore, voluntary adopters presented an even greater decrease in accruals after 2005. Finally, mandatory adopters during the post-IFRS period are less leveraged (2.27), have more growth opportunities (5.69) and generate more cash from their operating activities (0.102) as a percentage of total assets, compared with the pre-IFRS period. Additionally, their size shows an increase to 4.88 in the post-IFRS period compared to the pre-IFRS period (4.45), however this result can be attributed to the revaluation of assets that took place in 2005. Table 3 presents the Pearson correlation coefficients among the sample variables for the whole period of investigation (2001–2008) as well as during the pre-IFRS period (2001–2004) and the post-IFRS period (2005–2008) for voluntary and mandatory adopters. As we can see, the stock price (P) is positively correlated with net income

(NIPS, 0.15) and book value of equity (BVPS, 0.71) only in the post-IFRS period for mandatory adopters, while voluntary adopters reported value relevant accounting information prior to 2005. These findings indicate that the adoption of IFRS resulted in an increased value relevance of earnings and book value of equity. Moreover, the correlation coefficient between operating cash flows (CFO) and total accruals is more negative in the post-IFRS period (−0.729) compared to the pre-IFRS period (−0.685) for mandatory adopters. Finally, regarding the other control variables, large firms report higher operating cash flows, have more growth opportunities and are characterized by increased 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 earnings per share and book value of equity per share with stock prices varies between local accounting principles and IFRS. For this reason models 1 and 2 were estimated separately for the pre and post-IFRS periods. The empirical findings are presented in Table 4. According to our discussion, it is expected that the value relevance of net income and the book value of equity will be higher in the post-IFRS period for mandatory adopters and this will be depicted by an increase in: (a) model 1's R2, and (b) the magnitude and significance of net income (NIPS) and book value of common equity (BVPS) coefficients during the post-IFRS period when compared to the pre-IFRS period. Results in Table 4 verify this assertion since the coefficients on NIPS and BVPS

Table 3 Pearson correlation coefficients of sample variables. Panel A: Full sample Variables

R

P

NIPS

BVPS

ACC

DACC

CFO

SIZE

LEV

GR

P NIPS BVPS ACC DACC CFO SIZE LEV GR AUD

0.074 0.032 0.005 0.021 0.005 0.007 −0.154 −0.055 0.102 0.085

0.053 0.041 0.018 0.009 0.211 −0.110 −0.041 0.079 0.112

0.685 0.021 0.041 0.077 0.063 −0.018 −0.011 0.008

0.005 0.004 0.008 0.039 −0.018 0.008 −0.041

0.714 −0.670 −0.022 −0.041 −0.208 −0.019

−0.408 0.008 −0.019 0.005 0.040

0.006 −0.001 0.188 0.073

0.099 −0.059 0.040

−0.011 0.090

−0.009

Variables

R

P

NIPS

BVPS

ACC

DACC

CFO

SIZE

LEV

GR

AUD

Panel B: Pre-IFRS sample R 1.000 P 0.167 NIPS −0.031 BVPS −0.049 ACC −0.069 DACC −0.033 CFO 0.099 SIZE 0.073 LEV 0.007 GR 0.053 AUD 0.019

0.220 1.000 −0.020 −0.019 −0.007 0.019 0.042 0.223 −0.017 0.011 0.097

0.212 0.185 1.000 0.774 −0.000 −0.002 0.129 0.051 −0.015 −0.022 −0.026

0.054 0.044 0.189 1.000 −0.002 0.000 −0.000 0.066 −0.008 −0.016 −0.044

−0.026 −0.021 −0.054 −0.022 1.000 0.804 −0.685 −0.159 −0.021 −0.504 −0.067

−0.065 −0.099 −0.062 −0.014 0.744 1.000 −0.567 −0.016 −0.025 0.021 −0.009

0.105 0.148 0.212 0.105 −0.569 −0.585 1.000 0.123 0.002 0.324 0.097

0.089 0.305 0.055 0.018 −0.207 −0.055 0.184 1.000 0.017 0.205 0.295

−0.054 −0.121 −0.219 −0.055 −0.033 −0.018 −0.263 0.005 1.000 0.011 0.113

0.047 0.088 0.089 0.088 −0.486 −0.063 0.401 0.305 0.005 1.000 0.059

0.011 0.074 0.025 −0.022 −0.114 −0.008 0.130 0.214 0.141 0.084 1.000

Panel C: Post-IFRS sample R 1.000 P 0.066 NIPS 0.193 BVPS 0.017 ACC 0.023 DACC 0.010 CFO −0.010 SIZE −0.054 LEV −0.016 GR 0.004 AUD 0.029

0.204 1.000 0.150 0.711 0.022 0.069 0.154 0.268 −0.059 0.103 0.104

0.285 0.154 1.000 0.120 0.167 0.207 0.049 0.111 −0.063 0.007 0.149

0.044 0.085 0.214 1.000 0.023 0.039 0.122 0.195 −0.064 0.091 −0.010

−0.036 −0.054 −0.007 −0.019 1.000 0.739 −0.729 −0.033 −0.021 −0.138 0.045

−0.074 −0.096 −0.052 −0.043 0.722 1.000 −0.253 0.101 −0.033 −0.001 0.077

0.122 0.195 0.263 0.112 −0.554 −0.536 1.000 0.208 −0.015 0.219 0.036

0.047 0.332 0.042 0.012 −0.184 −0.085 0.201 1.000 0.028 0.136 0.326

−0.089 −0.132 −0.208 −0.041 −0.022 −0.015 −0.274 0.004 1.000 −0.009 0.104

0.023 0.087 0.091 0.090 −0.498 −0.017 0.444 0.322 0.005 1.000 −0.033

0.010 0.055 0.033 −0.015 −0.155 −0.009 0.137 0.244 0.185 0.045 1.000

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

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

117

Table 4 Regression results on value relevance and IFRS controlling for voluntary IFRS adopters. Model 1: P
it ¼ α0 þα1 NIPSit þα2 BVPSit þα3 DVOLit þα4 DVOLit
NIPSit þα5 DVOLit
BVPSit þβControlit þ γYear dummies þ δIndustry dummies þ εit Model 2:

P
it ¼ α0 þα1 NIPSit þα2 BVPSit þα3 DVOLit þα4 DVOLit
NIPSit þα5 DVOLit
BVPSit þa6 AUDit þα7 AUDit
NIPSit þα8 AUDit
BVPSit þβControlit þ γYear dummies þ δIndustry dummies þ εit

Variables

Pre-IFRS Model 1

Model 2

Model 1

Model 2

Constant (a0)

−12.10a (−2.74) −0.28 (−0.24) 0.014 (0.54) 0.112c (1.57) 0.189b (1.86) 0.054b (2.05)

−14.34a (−2.71) 1.05 (0.98) 0.020 (0.57) 0.141b (1.69) 0.205a (2.54) 0.065b (2.02) −0.159 (−0.44) −1.987 (−1.21) 0.811c (1.56) 1.201a (3.57) −0.0012 (−0.39) 0.312 (0.66) −0.088 (−0.66) 6.9% 3.77a Included

−2.61b (−1.77) 0.886b (1.75) 1.005a (2.52)

−2.08b (−1.68) 0.441b (1.74) 0.944a (2.54)

0.543a (2.62) −0.010 (−0.27) 0.487 (1.14) 0.0044 (0.43) 67.1% 12.11a Included

−1.418b (−1.63) 1.547a (2.73) 1.066a (5.17) 0.414b (1.84) 0.018 (0.46) 0.454 (1.03) 0.005 (0.57) 70.7% 13.30a Included

404 8.25a

404 P (b0.001)

10.44a

P (b0.001)

NIPS (a1) BVPS (a2) DVOL (a3) DVOL*NIPS (a4) DVOL*BVPS (a5) AUD (a6) AUD*NIPS (a7) AUD*BVPS (a8) SIZE (β1) LEV (β2) CFO (β3) GR (β4) R2-adjusted F-stat Year and Industry dummies N Vuong's Z statistic

1.144a (3.47) −0.0011 (−0.31) 0.454 (0.68) −0.111 (−0.57) 6.7% 3.23a Included

Post-IFRS

404 404 Model 1 post-IFRS vs model 1 pre-IFRS Model 2 post-IFRS vs model 2 pre-IFRS

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).

are positive and significant only in the post-IFRS period (0.886 and 1.005 respectively). The relative coefficients for mandatory adopters in the pre-IFRS period are negative and non-significant suggesting that the adoption of IFRS had a significant positive impact on the value relevance of earnings and the book value of equity, thus verifying hypothesis H1. Also the interaction terms of DVOL for voluntary adopters with NIPS and BVPS are positive and significant suggesting that voluntary adopters are characterized by enhanced value relevance relative to the other firms which followed Greek GAAP. The negative coefficients on the accounting variables during the pre-IFRS period for the mandatory adopters could also be attributed to the financial distress that was experienced after the ASE crash in 1999–2000, which may have impacted on the perceived quality of accounting information. Stated differently, investors may have lost confidence in the capital market due to severe deficiencies during the crash period which influenced the relevance of financial statements as a whole, resulting in a smaller association between returns, earnings and book values. However, the introduction of well-respected and high-quality accounting standards, such as IFRS, may have reinstated investors' confidence in financial reports and increased the relevance of accounting information. Both models have been estimated including industry and year fixed effects. The Vuong Z statistic is the likelihood ratio test developed by Vuong (1989) for non-nested model selection. A significant positive 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-IFRS period. Thus, hypothesis H1 is verified by the increase in R 2 up to 67.1% during the post-IFRS period compared to the pre-IFRS period of 6.7%. The Vuong Z statistic is significant at the one percent significance level indicating that the model during the post-IFRS period better explains the original data relative to the same model during the pre-IFRS period. This finding verifies recent evidence provided by Ding et al. (2007) who document that the higher the “absence” between domestic accounting standards and IFRS the lower the value relevance of accounting information to the market. 3 Greece was found to be the top country in the absence score suggesting that the introduction of IFRS and the enrichment of coverage of the local accounting standards are crucial for improving the transparency and relevance of accounting disclosures. Thus, we can conclude that the adoption of IFRS resulted in an improvement in the value relevance of accounting information compared to the local accounting standards. However, our results stand in opposition to the results of Ahmed et al. (2010) in their international

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

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 countries with a strong rule of law. Greece is a country with a weak rule of law and, in contrast, strong law countries even prior to IFRS adoption may already have had high quality accounting standards and there would therefore be less room for improvement after the adoption of IFRS. Conversely, in Greece where there is high book-tax conformity and weak legal enforcement, there seems to be more room for improvement after the mandating of IFRS since the quality of accounting is lower compared to other code-law countries. Moreover, H2 expects that the value relevance of net income and the book value of equity will be higher in the post-IFRS period for firms with increased audit quality because multinational audit firms have more resources and experience of the issues of international accounting standards, and could provide greater assistance for the implementation and transition to IFRS compared to local audit firms. In order to test this assumption model 1 and model 2 are estimated separately for the pre and post-IFRS periods. H2 will be accepted if the explanatory power of model 2 in the post-IFRS period is higher compared to model 2 in the pre-IFRS period and also if coefficients a7 and a8 are significant and have a higher magnitude relative to coefficients a1 and a2 (again in the post-IFRS period). The findings in Table 4 again verify H2 since the explanatory power of model 2 in the post-IFRS period presents a huge increase to 70.7% compared to the pre-IFRS period where the relative R 2 is only 6.9% and the Vuong statistic is also significant at the one percent level. Furthermore, the coefficients of the interaction variables NIPS*AUD and BVPS*AUD (from model 2 during the post-IFRS period) are both positive and statistically significant (1.547 and 1.066 respectively) and their relative magnitudes are higher compared to the coefficients of the NIPS and BVPS variables (0.441 and 0.944 respectively). Regarding the control variables, the only significant coefficient was found in the SIZE variable which is positive and statistically significant at the one percent significance level indicating that larger firms report accounting figures of increased value relevance. This result could be attributed to the fact that larger firms in Greece draw more publicity than smaller firms, and thus their earnings quality is likely to be disclosed with less uncertainty. As a result, investors place greater trust in larger firms and are more responsive to their financial reports. Overall we can argue that firms characterized by increased audit quality present a higher value relevance of earnings and book value of equity compared to firms of a lower audit quality and this positive impact seems to be further complemented by the implementation of IFRS. Thus, our findings corroborate the results of Ewert and Wagenhofer (2005), Lang et al. (2003, 2006), Leuz et al. (2003) and Barth et al. (2008) who argue that high-quality accounting standards (such as IFRS) result in accounting figures that have a higher value relevance. 5.3. Regression results on timely loss recognition and IFRS Our next step was to examine the impact of IFRS adoption on timely loss recognition. For this purpose we applied the classic Basu (1997) reverse regression model between earnings per share and stock returns. Model 3 is estimated separately for the pre and post-IFRS periods and if IFRS increase the timely recognition of losses we expect an increase in the explanatory power of the model and coefficient b3 to be positive and has a higher magnitude relative to the pre-IFRS period. The empirical findings are presented in Table 5. Regarding mandatory IFRS adopters, coefficient b3(D*R) of model 3 is positive and statistically significant only in the post-IFRS period while the relative coefficient for voluntary adopters (b7) is positive and significant during the pre-IFRS period. Also, the R 2 of model 3 presents a small increase during the post-IFRS period up to 7.2% compared to the pre-IFRS R 2 of 6.1%. The Vuong Z statistic in this table is significant at the one percent significance level suggesting that the

model during the post-IFRS period better explains the original data relative to the same model during the pre-IFRS period. In addition, voluntary IFRS adopters exhibited significant conservative reporting compared to firms following the Greek GAAP since coefficient b7 is positive and significant. Therefore, we can argue that the adoption of IFRS resulted in an increase in conservatism within the Greek listed firms. This finding is consistent with Barth et al. (2008) who found that firms operating under IFRS exhibit more timely recognition of losses compared to the non-IFRS firms. In addition, hypothesis H2 expects that firms audited by Big-5 audit corporations will present more timely loss recognition. If this is true we expect a significant positive coefficient on the R*D*AUD (b11) interaction term. The results of model 4 in Table 5 verify the aforementioned hypothesis since the coefficient b11 is positive and statistically significant and presents a higher magnitude in the post-IFRS period (0.410) relative to the pre-IFRS period (0.248). In addition, the R 2 of model 4 in the post-IFRS period presents an increase to 21.1% relative to the pre-IFRS period where the respective R 2 is only 7.4%. Moreover, the Vuong Z statistic is significant at the one percent significance level suggesting that model 4 during the post-IFRS period better explains the original data relative to the same model during the pre-IFRS period. Regarding the control variables, SIZE was found to have a positive and significant coefficient only in the post-IFRS period suggesting that larger firms report more timely losses after the adoption of IFRS. Furthermore, the LEV variable has the expected sign but is marginally significant only during the post-IFRS era. Finally, the GR variable was found to be negative and marginally significant only during the pre-IFRS period, and the CFO coefficient was positive and statistically significant as we expected in both sub-periods indicating that firms generating increased cash flows are more timely loss reporters. Consequently, if higher accounting quality is measured by the recognition of large losses as they occur instead of being deferred to future periods (Ball & Shivakumar, 2005, 2006) our results suggest that IFRS are characterized by increased quality since they have proven to contribute significantly to the more timely recognition of losses within Greek listed firms. 5.4. Empirical findings on earnings management and IFRS Our last section presents the empirical findings regarding the impact of IFRS adoption on earnings management behavior. In order to test this association we applied two different metrics of earnings manipulation. Our first test, which controls for earnings smoothing, is the ratio of the standard deviation of the change in net income (ΔΝΙ*) to the standard deviation of the change in cash flow (ΔCF*) residuals during the pre and post-IFRS periods (Ahmed et al., 2010). The relevant results are presented in Table 6. The differences between the ratios were estimated by applying a t-test based on the distribution of the differences. ΔΝΙ* and ΔCF* are the change of net income and change of cash flow residuals from the following four regression models, instead of comparing these measures directly into the specific tests. The empirical findings indicate the ratio of standard deviation of the change in net income (ΔΝΙ*) to the standard deviation of the change in cash flow (ΔCF*) residuals, as well as the difference of this ratio between firms audited by Big-5 audit and non-Big-5 audit corporations during the pre and post-IFRS adoption periods. This difference was estimated using a t-test based on the distribution of the differences (Barth et al., 2008). Evidence suggests that the specific ratio is higher during the post-IFRS adoption period for mandatory adopters (difference 0.0007 and significant at 1%), while the highest difference is demonstrated by firms audited by the Big-5 (significant at 1%). Overall, the evidence in Table 6 suggests that Greek listed firms which adopted IFRS after its inauguration in 2005, report less smoothed financial statements after the implementation of IFRS,

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

119

Table 5 Regression results on timely loss recognition and IFRS controlling for voluntary IFRS adopters. Model 3: EYi;t ¼ b0 þb1
Rit þb2
Dit þb3
Rit
Dit þb4 DVOLit þb5 Dit
DVOLit þb6 Rit
DVOLit þb7 Rit
Dit
DVOLit þβControlit þ γYear dummies þ δIndustry dummies þ eit Model 4:

EYit ¼ b0 þb1 Rit þb2 Dit þb3 Rit
Dit þb4 DVOLit þb5 Dit
DVOLit þb6 Rit
DVOLit þb7 Rit
Dit
DVOLit þb8 AUDit þb9 Dit
AUDit þb10 Rit
AUDit þb11 Rit
Dit
AUDit þβControlit þ γYear dummies þ δIndustry dummies þ eit

Variables

Pre-IFRS Model 3

Constant (b0) R (b1) D (b2) D*R (b3) DVOL (b4) DVOL*D (b5) DVOL*R (b6) DVOL*D*R (b7) AUD (b8) AUD*D (b9) AUD*R (b10) AUD*D*R (b11) SIZE (β1) LEV (β2) CFO (β3) GR (β4) R2-adjusted F-stat Year and Industry dummies N Vuong's Z statistic

Post-IFRS Model 4

Model 3

Model 4

−1.221 (−1.11) −0.187 (−0.60) −0.009 (−0.40) 0.221b (1.69) 0.184b (1.90) −0.031 (−0.91) −0.105 (−1.36) 0.210b (2.27) 0.051 (0.19) −0.198 (−0.67) 0.094 (0.88) 0.248b (2.12) 0.041 0.044 (0.67) (0.91) −0.001 0.0008 (−0.28) (0.69) 0.487a 0.518a (2.66) (2.71) −0.044 −0.044 (−1.21) (−1.17) 6.1% 7.4% 2.88a 2.77a Included Included 404 404 Model 3 post-IFRS vs. Model 3 pre-IFRS Model 4 post-IFRS vs. Model 4 pre-IFRS

−0.528 (−1.20) 0.030a (2.57) 0.0038 (0.25) 0.251b (1.90)

0.0080 (0.33) −0.068a (−3.18) −0.051 (−0.98) 0.338b (2.12)

0.059b (1.93) −0.009 (−0.96) 0.131b (2.57) −0.0040 (−0.19) 7.2% 3.24a Included 404 3.87a 4.27a

0.042 (0.90) 0.204b (1.71) 0.192a (6.50) 0.410b (2.17) 0.027b (1.66) −0.010 (−1.27) 0.191b (2.22) −0.0017 (−0.27) 21.1% 7.03a Included 404 P (b0.001) P (b0.001)

−0.981 (−0.77) −0.141 (−0.45) −0.044 (−0.21) 0.214 (1.58) 0.157b (1.74) −0.027 (−0.84) −0.094 (−1.12) 0.187b (2.31)

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

verifying previous arguments made by Barth et al. (2008) and Chen et al. (2010) that IFRS adoption leads to less earnings manipulation. Our second test on earnings management is based on the estimation of model 9 which includes the absolute value of performancematched discretionary accruals as the dependent variable (see Kothari et al., 2005 for details). The relevant results are presented in Table 7. The coefficient of the IFRS dichotomous variable is negative and statistically significant (−0.181) and the relative coefficient for the voluntary adopters (DVOL) is also negative and significant (−0.145) indicating that the adoption of IFRS resulted in a decrease in the magnitude of performance-matched discretionary accruals. Thus, this finding corroborates the evidence by Aubert and Grudnitski (2011) who argue that mandatory IFRS adoption in Greece improved accrual quality. In addition, the coefficient of the interaction term IFRS*AUD is also negative and statistically significant at the one percent significance level (−0.125) verifying H2 and suggesting that Big-5 audit firms provide better quality audits than local audit firms. Thus, being audited by a Big-5 audit firm further mitigates any incentives for manipulating accounting income. Overall, the evidence from the two

metrics of earnings management tends to suggest that firms which adopted IFRS after its inauguration in 2005 report better quality financial statements, verifying previous arguments made by Barth et al. (2008) and Chen et al. (2010). Finally, regarding the control variables, SIZE is significant with a negative coefficient similar to Watts and Zimmerman (1990) and Van Tendeloo and Vanstrelen (2005). Also the coefficient of the absolute value of operating cash flows (|CFO|) is positive and statistically significant (0.166) as predicted, suggesting that the matching principle results in a natural smoothing of accruals which in turn causes negative discretionary accruals to occur in periods of extreme positive cash flows, verifying Dechow et al. (1996) and Young (1999). Overall the empirical results indicate that the adoption of IFRS has improved the quality of accounting information by increasing the value relevance of earnings and book value of equity, improving the timely recognition of losses and by decreasing discretionary manipulation of earnings. Consequently, our findings allow us to accept both hypotheses H1 and H2 and claim that accounting figures prepared under IFRS are of a higher quality than those prepared under the Greek GAAP and that audit quality contributes to enhanced accounting

120

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

Table 6 Univariate 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

Full sample

SD(ΔΝΙ*)/SD(ΔCF*) N SD(ΔΝΙ*)/SD(ΔCF*) N SD(ΔΝΙ*)/SD(ΔCF*) N

Big-5 Non Big-5 a

Voluntary adoption

Mandatory adoption

Pre-IFRS

Post-IFRS

Difference

Pre-IFRS

Post-IFRS

Difference

0.0019 100 0.0021 40 0.0030 60

0.0020 100 0.0019 40 0.0029 60

0.0001 100 0.0002 40 0.0001 60

0.0021 304 0.0027 116 0.0024 188

0.0028 304 0.0035 116 0.0027 188

0.0007a 304 0.0008a 116 0.0003a 188

Significant at 1% significance level.

quality during the post-IFRS period. These results corroborate the arguments of several studies such as Ashbaugh and Pincus (2001), Ewert and 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 several sensitivity tests related to the specification of the empirical models and the research design. 4 First, we re-estimated model 9 including additional interaction variables between IFRS and the control variables (CFO, SIZE, LEV, GR) in order to control for any effects of IFRS that are not picked up by the sole control variables. Results remained qualitatively unchanged compared to those reported in Table 7. Also, we replaced discretionary accruals as the dependent variable on model 9 with total accruals in order to capture any problems from possible accrual misspecification arising from the estimation of the Jones (1991) model. The results remained unchanged after this modification. Furthermore, we controlled for possible misspecification in the estimation of the performance-matched discretionary accruals. Following Kothari et al. (2005) we re-estimated the Jones (1991) model as follows: (1) without a constant, (2) without matching firms based on ROA but estimating DACC including ROA as an additional regressor, and 3) by implementing the Jones model cross-sectionally using all firms instead of just firms within the same industry. In addition, we re-estimated models 1 and 2 using the actual stock price and the annual buy-and-hold return as the dependent variables, and the level of earnings and the book value of equity deflated by the beginning of the year market capitalization as the explanatory variables. The results are qualitatively similar to the results presented in Table 4. Also, we re-estimated models 3 and 4 during the pre and post-IFRS periods by replacing the dependent variable (pre-tax earnings deflated by lagged total assets) with earnings per share deflated by the opening share price. Again the results did not change after this modification. Additionally, we controlled for the level of family ownership and cases of cross-listed firms following the discussion in Section 2 suggesting that the large number of family-owned firms in Greece has encouraged the limitation of financial reporting to provide tax-relevant information only. For this purpose we included a dummy coded one (1) for cross-listed firms and zero (0) otherwise and took into account the percentage of stocks owned by the founding family members of each company (following Wang, 2006), before re-estimating all models. The results proved to be unaffected by these variables since the cross-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 statement manipulation after the adoption of IFRS. Furthermore, Johnson (1999) finds convincing evidence that the value relevance of earnings fluctuates over time as a function of time-varying expected returns. She argues that interest rates should be included in models whenever inferences about changes in the information relevance of earnings due to non-interest rate factors (e.g. change in accounting methods) are hypothesized. For this reason model 1 was re-estimated for the pre and post-IFRS periods including the term-spread of interest rates as an additional explanatory variable. Results remained unchanged relative to those reported in Table 4. In addition, we re-estimated model 1 combining the variables during the pre and post-IFRS adoption periods by interacting NIPS and BVPS with a dummy IFRS receiving one (1) for the period

Table 7 Regression results on earnings management and IFRS controlling for voluntary IFRS adopters. Model 9:

jDACCjit ¼ c0 þc1 IFRSit þc2 AUDit þc3 DVOLit þc4 AUD
DVOLit þc5 AUD
IFRSit þβControlit þ γYear dummies þδIndustry dummies þ uit

Variables

Model 9 (signed DACC)

Model 9 (absolute DACC)

Constant (c0)

0.310a (2.70) −0.191b (−2.18) 0.088a (3.01) −0.122 (−1.54) −0.074 (−1.19) −0.144a (−2.87) 0.174a (9.21) −0.018b (−1.95) −0.0010 (−0.55) −0.0004 (−0.11) 22.7% 14.08a Included 808

0.355a (3.05) −0.181b (−1.99) 0.084a (2.77) −0.145c (−1.62) −0.087 (−1.25) −0.125a (−2.75) 0.166a (8.88) −0.021b (−2.05) −0.0009 (−0.22) −0.0003 (−0.27) 19.7% 13.08a Included 808

IFRS (c1) AUD (c2) DVOL (c3) DVOL*AUD (c4) IFRS*AUD (c5) |CFO| (β1) SIZE(β2) LEV(β3) GR (β4) Adjusted R2 F-statistic Year and Industry dummies N

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).

P.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 the same compared with those in Table 4. Moreover, we re-estimated all empirical models excluding the control variables and the year and industry dummies so as to check whether these variables confound the explanatory power of the models but the results remained unchanged. In addition, we controlled for a ‘learning curve’ effect. Based on this rationale, as the firm continues applying IFRS concepts and practices it becomes more familiar with proper IFRS application. Cohen and Leventis (2013) found that time has a beneficial effect on compliance with accounting standards in Greece. This is another explanation of why empirical findings may point towards IFRS not contributing to the improvement of accounting quality. For this reason we divided the post-adoption period into two sub-periods, namely 2005–2006 and 2007–2008, and re-estimated models 9, 2 and 4. The results suggested that there is some learning curve effect since the results for value relevance and earnings management were more significant during the 2007–2008 sub-period. Contrarily however, accounting conservatism does not present any significant differences between the two sub-periods. Finally, we controlled for possible cross-sectional dependence and heterogeneity in our data by applying the methodology proposed by Kousenidis et al. (2009). The rationale behind this modification is that the level of conservatism and value relevance may not be directly related 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 including the cross-sectional means of the dependent and independent variables namely P, NIPS and BVPS for model 1 and EBT and R for model 2. Results remained unchanged after this specification. 6. Conclusions The issue of international accounting standards and their impact on accounting quality have been a part of the international research agenda for a number of years. Many studies have documented that the adoption of high-quality accounting standards results in an increase in the quality of accounting information. For instance, researchers argue that IFRS improves the reliability of financial reporting by limiting opportunistic managerial discretion (Ashbaugh & Pincus, 2001; Barth et al., 2008; Ewert & Wagenhofer, 2005). The contrary view is that the flexibility inherent in IFRS and lax enforcement might provide greater 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 the influence of IFRS, empirical research on the impact of IFRS on earnings management has also provided mixed results. Van Tendeloo and Vanstrelen (2005), for example, found that earnings management behavior was not significantly different between companies in Germany that adopted IFRS and those that relied on German GAAP. Conversely, Barth et al. (2008) concluded that firms applying IFRS displayed significantly less earnings management, meaning that they report more reliable and transparent financial statements. Part of the motivation for this study was to shed further light on the competing views regarding the efficiency of IFRS in reducing earnings manipulation and improving earnings quality within an accounting framework characterized by a long history of historical-accounting principles. We examined whether accounting information during the post-IFRS period (2005–2008) demonstrates less earnings management, more timely loss recognition and higher value relevance, when compared to the relative amounts reported during the pre-IFRS period (2001–2004). After controlling for firm-specific characteristics such as size, growth opportunities, risk and audit quality we found that the introduction of IFRS contributed to less earnings management, more timely loss recognition, and greater value relevance of earnings and book value of equity compared to the local accounting standards. Furthermore, our findings document that a firm's audit quality further

121

complements the beneficial impact of IFRS since companies audited by a Big-5 audit firm exhibit higher levels of accounting quality compared to their non-Big-5 counterparts. Our findings contribute to the ongoing debate over whether high-quality accounting standards are effective and sufficient in code-law countries and countries with a historical-cost accounting base. Moreover, our results could prove useful to investors since they indicate that financial accounting information prepared under IFRS and audited by the Big-5 audit firms is more reliable and has enhanced value relevance. However, the results of this study should be treated with caution since they are subject to several limitations. Although we have controlled for various incentives of reporting discretion, we acknowledge that there may be other incentives which remained uncontrolled by our empirical models. Thus, there may be an additional source of bias that affects our inferences regarding the estimation of model 9. Finally, the small size of our sample and the focus on a single country setting could be another source of bias which does not allow us to infer any generalization of the results onto firms operating outside the Greek borders. Future research, however, could help resolve this problem by examining the impact of IFRS adoption on accounting quality between different EU countries which share common accounting characteristics, thus providing more salient inferences to the current debate through cross-country evidence. Furthermore, a recent paper by Cahan, Emanuel, and Sun (2009) declares that even if the accounting system (IFRS) can generate high-quality accounting information in a country with a weak institutional infrastructure, the information will still be less important for market participants. Consequently, it would be interesting to examine the impact of IFRS adoption on accounting quality under any legislative improvements taking place in the European business setting regarding the issues of investor protection and institutional transparency. References Aharony, J., Barniv, R., & Falk, H. (2010). The impact of mandatory IFRS adoption on equity valuation of accounting numbers for security investors in the EU. The European Accounting Review, 19, 535–578. Ahmed, A. S., Neel, M., & Wang, D. (2010). Does mandatory adoption of IFRS improve accounting quality? Preliminary evidence. Working paper, Texas A&M University. Ashbaugh, H., & Pincus, M. (2001). Domestic accounting standards, international accounting standards and the predictability of earnings. Journal of Accounting Research, 39, 417–434. Aubert, F., & Grudnitski, G. (2011). The impact and importance of mandatory adoption of International Financial Reporting Standards in Europe. Journal of International Financial Management and Accounting, 22, 1–26. Ball, R., Kothari, S. P., & Robin, A. (2000). The effect of international institutional factors on properties of accounting earnings. Journal of Accounting and Economics, 29, 1–51. Ball, R., Robin, A., & Wu, J. S. (2003). Incentives versus standards: Properties of accounting income 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 loss recognition timeliness. Journal of Accounting and Economics, 39, 83–128. Ball, R., & Shivakumar, L. (2006). The role of accruals in asymmetrically timely gain and loss 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 Economic Research, 14, 485–497. Ballas, A. A., Skoutela, D., & Tzovas, C. A. (2010). The relevance of IFRS to an emerging market: Evidence from Greece. Managerial Finance, 36, 931–948. Barth, M. E., Landsman, W. R., & Lang, M. H. (2008). International accounting standards and 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, public ownership and firms' discretionary accruals management. The International Journal of Accounting, 38, 1–22. Beekes, W., Pope, P., & Young, S. (2004). The link between earnings timeliness, earnings conservatism 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 IFRS reporting and stock price informativeness. Working paper, Tilburg University. Billings, B. K. (1999). Revisiting the relation between the default risk of debt and the earnings response coefficient. The Accounting Review, 74, 509–522. Bowen, R. M., Rajgopal, S., & Venkatachalam, M. (2003). Accounting discretion, corporate governance and firm performance. Working paper, University of Washington and Duke University. Breeden, R. (1994). Foreign companies and US markets in a time of economic transformation. 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 Accounting Review, 81, 983–1016. Byard, D., Li, Y., & Yu, Y. (2011). The effect of mandatory IFRS adoption on financial analysts' information environment. Journal of Accounting Research, 49, 69–96. Cahan, S. F., Emanuel, D., & Sun, J. (2009). The effect of earnings quality and country-level institutions on the value relevance of earnings. Review of Quantitative 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. Advances 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 comparability 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 of Accounting and Economics, 45, 116–138. Chen, H., Tang, Q., Jiang, Y., & Lin, Z. (2010). The role of international financial reporting standards in accounting quality: Evidence from the European Union. Journal of International Financial Management and Accounting, 21, 220–278. Cohen, S., & Leventis, S. (2013). An empirical investigation of audit pricing in the public sector: The case of Greek LGOs. Financial Accountability and Management, 29(1), 74–98. Conover, M., Miller, R., & Szakmary, A. (2008). The timeliness of accounting disclosures in international security markets. International Review of Financial Analysis, 17(5), 849–869. Dargenidou, C., & McLeay, S. (2010). The impact of introducing estimates of the future on international comparability in earnings expectations. The European Accounting Review, 19, 511–534. Daske, H. (2006). Economic benefits of adopting IFRS or US-GAAP — Have the expected costs 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 firm performance: The role of accounting accruals. Journal of Accounting and Economics, 18, 3–42. Dechow, P., Sloan, R., & Sweeney, A. (1996). Causes and consequences of earnings manipulation: 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 accounting data after the introduction of IGRS in Europe. Journal of International Financial Management and Accounting, 21, 85–119. Dimitropoulos, P., & Asteriou, D. (2009). The impact of accruals and cash flows on the returns-earnings relation: Evidence from Greece. International Journal of Accounting, Auditing and Performance Evaluation, 5, 384–407. Ding, Y., Hope, O. -K., Jeanjean, T., & Stolowy, H. (2007). Differences between domestic accounting 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 comparative impact of degree of implementation of IFRS upon the public and private information 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 the adoption of IFRS. Managerial Finance, 36, 969–980. Eccher, E., & Healy, P. (2003). The role of International Accounting Standards in transitional economies: A study of the People's Republic of China. Working paper, Massachusetts Institute of Technology. Ewert, R., & Wagenhofer, A. (2005). Economic effects of tightening accounting standards to restrict earnings management. The Accounting Review, 43, 1101–1124. Francis, J., & Wang, D. (2003). Are Big 4 audit firms conservative only in the United States? Working paper, University of Missouri–Columbia. Frankel, R., Johnson, M., & Nelson, K. (2002). The relation between auditors' fees for 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 Accounting Review, 80, 585–612. Glaum, M., & Street, D. (2003). Compliance with the disclosure requirements of Germany's new markets: IAS versus US GAAP. Journal of International Financial Management and Accounting, 14, 64–100. Gore, P., Pope, P. F., & Singh, A. K. (2001). Non-audit services, auditor independence and earnings management. Working paper, Lancaster University. Guenther, N., Gegenfurtner, B., Kaserer, C., & Achleitner, A. K. (2009). International financial reporting standards and earnings quality: The myth of voluntary vs. mandatory adoption. Working paper no 2009–09, Centre for Entrepreneurial and Financial Studies, Technical University of Munich. Guenther, D. A., & Young, D. (2000). The association between financial accounting and measures of real economic activity: A multinational study. Journal of Accounting and Economics, 29, 53–72.

Gul, F. A., Sun, S. Y. J., & Tsui, J. S. L. (2003). Audit quality, earnings and the Shanghai stock 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 and the value relevance of earnings. Accounting and Finance, 46, 1–21. Hassan, O. A. G., Romilly, P., Giorgioni, G., & Power, D. (2009). The value relevance of disclosure: Evidence from the emerging capital market of Egypt. The International Journal of Accounting, 44, 79–102. Hitz, J. M. (2007). The decision usefulness of fair value accounting: A theoretical perspective. The European Accounting Review, 16, 323–362. Hodgdon, C., Tondkar, R. H., Adhikari, A., & Harless, D. W. (2009). Compliance with international financial reporting standards and auditor choice: New evidence on the importance of the statutory audit. The International Journal of Accounting, 44, 33–55. Hofstede, G. (1980). Culture's consequences: International differences in work-related values. Newbury, CA: Sage Publications. Hope, O. K. (2003). Disclosure practises, enforcement of accounting standards and analysts' forecast accuracy: An international study. Journal of Accounting Research, 41, 235–272. Hung, M., & Subramanyam, K. R. (2007). Financial statement effects of adopting International Accounting Standards: The case of Germany. Review of Accounting Studies, 12, 623–657. Iatridis, G. (2008). Accounting disclosures and firms' financial attributes: Evidence from the UK stock market. International Review of Financial Analysis, 17, 219–241. Iatridis, G. (2010). International Financial Reporting Standards and the quality of financial statement information. International Review of Financial Analysis, 19, 193–204. Iatridis, G. (2011). Accounting disclosures, accounting quality and conditional and unconditional conservatism. International Review of Financial Analysis, 20, 88–102. Iatridis, G., & Rouvolis, S. (2010). The post-adoption effects of the implementation of International Financial Reporting Standards in Greece. Journal of International Accounting, Auditing and Taxation, 19, 55–65. Johnson, M. F. (1999). Business cycles and the relation between security returns and earnings. Review of Accounting Studies, 4, 93–117. Jones, J. (1991). Earnings management during import relief investigations. Journal of Accounting Research, 29, 193–228. Klein, A. (2002). Audit committee, board of director characteristics and earnings management. Journal of Accounting and Economics, 33, 375–400. Kothari, S. P., Leone, A. J., & Wasley, C. E. (2005). Performance matched discretionary accrual measures. Journal of Accounting and Economics, 39, 163–197. Kousenidis, D. V., Ladas, A. C., & Negakis, C. I. (2009). Value relevance of conservative and non-conservative accounting information. The International Journal of Accounting, 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 of annual earnings announcements and mandatory adoption of IFRS. Journal of Accounting and Economics, 53, 34–54. Lang, M., Raedy, J., & Wilson, W. (2006). Earnings management and cross listing: Are reconciled earnings comparable to US earnings? Journal of Accounting and Economics, 42, 255–283. Lang, M., Raedy, J., & Yetman, M. (2003). How representative are firms that are cross listed in the United States? An analysis of accounting quality. Journal of Accounting Research, 41, 363–386. Leuz, C., Nanda, D., & Wysocki, P. D. (2003). Earnings management 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 audit quality. Managerial Auditing Journal, 20, 460–478. Park, C., & Pincus, M. (2001). Internal versus external equity funding sources and earnings response coefficients. Review of Quantitative Finance and Accounting, 16, 33–52. Penman, S. (2007). Financial reporting quality: Is fair value a plus or minus? Accounting and Business Research, 37, 33–44. Prather-Kinsey, J., & Meek, G. K. (2004). The effect of revised IAS 14 on segment reporting by IAS companies. The European Accounting Review, 13, 213–234. Raonic, I., McLeay, S., & Asimakopoulos, I. (2004). The timeliness of income recognition by interlisted European companies: An analysis of institutional and market complexity. Journal of Business Finance and Accounting, 31, 115–148. Reynolds, J. K., & Francis, J. R. (2000). Does size matter? The influence of large clients on office-level auditor reporting decisions. Journal of Accounting and Economics, 30, 375–400. Spathis, C., Doumpos, M., & Zopounidis, C. (2003). Using client performance measures to identify pre-engagement factors associated with qualified audit reports in Greece. 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 of Chartered Certified Accountants. Street, D. L., & Gray, S. J. (2002). Factors influencing the extent of corporate compliance with 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 coefficient. The Accounting Review, 68, 346–366. Tsakumis, G. T. (2007). The influence of culture on accountants' application of financial reporting rules. Abacus, 43, 27–48. Tsalavoutas, I. (2011). Transition to IFRS and compliance with mandatory disclosure requirements: What is the signal? Advances in Accounting, incorporating Advances in International Accounting, 27, 390–405.

P.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 effects and auditor size. Managerial Auditing Journal, 25, 814–842. Tyrrall, D., Woodward, D., & Rakhimbekova, A. (2007). The relevance of International Financial 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 explaining stock returns in an emerging market: The case of Cyprus. The European Accounting Review, 7, 105–124. Van Tendeloo, B., & Vanstrelen, A. (2005). Earnings management under German GAAP versus IFRS. The European Accounting Review, 14, 155–180. Vuong, Q. H. (1989). Likelihood ratio tests for model selection and non-nested hypotheses. Econometrica, 57, 307–333.

123

Wang, D. (2006). Founding family ownership and earnings quality. Journal of Accounting Research, 44, 619–656. Warfield, T., Wild, J., & Wild, K. (1995). Managerial ownership, accounting choices, and informativeness of earnings. Journal of Accounting and Economics, 20, 61–91. Watts, R. L., & Zimmerman, J. L. (1990). Positive accounting theory: A ten year perspective. The Accounting Review, 65, 131–156. White, H. (1980). A heteroscedasticity consistent covariance matrix estimator and a direct test for heteroscedasticity. Econometrica, 48, 817–838. Young, S. (1999). Systematic measurement error in the estimation of discretionary accruals: An evaluation of alternative modelling procedures. Journal of Business Finance and Accounting, 26, 833–862.