The impact of the voluntary and mandatory adoption of IFRS on
earnings quality: Evidence from Germany.
Author: Ernst van Zanten University of Groningen
The impact of the voluntary and mandatory adoption of IFRS on
earnings quality: Evidence from Germany.
Master thesis Accountancy
University of Groningen
Abstract
The demand for more severe financial reporting regulations emerged after some extensive financial scandals in the past. But to what extent does the on-going development of this regulatory environment actually improve financial reporting quality? This study contributes to this discussion by examining both the voluntary and mandatory adoption of IFRS in relation to earnings quality for German listed firms. In addition, I examine to what extent audit firm size affects this relation. To draw a conclusion about the level of earnings quality I apply three different measures of earnings quality including discretionary accruals, earnings smoothing and the value relevance of earnings. By using a sample consisting of more than 5.000 firm year observations between 1999- 2009, I infer that earnings quality did not significantly increase for both voluntary and mandatory IFRS adopters. Moreover, I find that the amount of earnings management activities increased after the adoption of IFRS. With respect to audit firm size, the results suggest that firms which are audited by a Big 3 audit firm are related to a lower decrease in earnings quality after the adoption of IFRS compared to non-Big 3 audit firms. Overall, I conclude that the specific regulatory change that came with the introduction of IFRS did not improve earnings quality.
Name: Ernst van Zanten
Student number: S3158128 Supervisor: Dr. C.A. Huijgen Second Assessor: Dr. T.A. Marra
Date: 23-6-2018
1
Introduction
There has been much discussion about the effect of regulations on earnings quality in recent years. Due to some extensive financial scandals in the past the demand for more transparent and severe regulations emerged. For this reason, the number of regulations in the field of financial reporting and corporate governance has grown in past years. Some examples of these regulations are the introduction and development of the Sarbanes-Oxley act (SOX) in the US and the mandatory adoption of the International Financial Reporting Standards (IFRS) for listed firms in Europe. These regulations should, among other things, improve the transparency and reliability of financial information and accordingly enhance earnings quality.
Earnings management is one of the aspects that influences earnings quality (Dechow et al., 2010). Moreover, earnings management is one of the main reasons for the disruption in the transparency and reliability of financial information. New regulations should contribute to reducing the use of earnings management and consequently enhance earnings quality. However, there is disagreement among researchers about the effect of these regulations on the amount of earnings management. For instance, Aubert and Grudnitski (2014) find statistical evidence that the introduction of SOX has contributed to reducing opportunistic earnings management behavior. On the other hand, Evans et al. (2015) offers evidence that suggests that the heavily regulated environment in the U.S. does not necessarily lead to a reduction in the total amount of earnings management.
To contribute to this discussion, I will investigate the effect of the voluntary and mandatory adoption of IFRS on earnings quality in Germany. In the 1990s several German companies needed capital injections from the international market. For this reason, the German legislator allowed German listed firms to voluntarily adopt IFRS in 1998 (Haller et al., 2009). Since 1 January 2005 the adoption of IFRS is mandatory for all listed firms in Europe. The adoption of IFRS should mainly lead to a more transparent and decision useful financial reporting mechanism (Weißenberger et al., 2004).
Another factor that influences earnings quality is the quality of audit services (Dechow et al., 2010). Audit quality is also highly debated recently as is expressed in reform of the EU statutory audit market in June 2016 (EU Memo/16/2244). In this research I will address this discussion by examining to what extent audit firm size affects the impact of the transition from German GAAP to IFRS. Therefore, I will make a distinction between firms which are audited by large audit firms and firms which are audited by smaller audit firms. According to DeAngelo (1981), Big 4 auditors achieve higher audit quality than non-Big 4 auditors due to the higher client independency. Moreover, Big 4 auditors would be better able to adapt to (international) regulatory changes (e.g. Street and Gray, 2002). However, Lawerence et al. (2011) argues that Big 4 and non-Big 4 auditors both need to follow the same
legislations and professional standards and one therefore should expect approximately the same audit quality.
Because larger audit firms are expected to provide higher audit quality and IFRS focusses on investor protection I expect a higher level of earnings quality at firms which adopted IFRS and are audited by a large audit firm. Based on this expectation I formulated the following research question:
To what extent does the voluntary and mandatory adoption of IFRS affect earnings quality of firms with larger versus smaller auditors?
To determine earnings quality I will examine three aspects of earnings quality. First, I will apply the modified Jones model to measures the amount of discretionary accruals. Second, I will determine whether companies engaged in earnings smoothing. Finally, the value relevance of earnings will be studied to determine earnings quality.
To my knowledge, this area of research has not been approached in this specific way. Many studies aim to find a decent explanation for the effect of regulations on earnings quality. This study will contribute to this research subject by examining the effect of the adoption of IFRS on earnings quality. This subject remains a relevant research direction as the financial regulatory environment frequently changes.
The German setting is well applicable for research in adopting IFRS since German firms voluntarily adopted IFRS before the mandatory adoption of IFRS in 2005. For this reason, a comparison can be made between firms which voluntary adopted IFRS and firms which initially applied German GAAP and were eventually forced to adopt IFRS.
Further, I will contribute to the ongoing discussion on audit quality by investigating the effect of audit firm size on earnings quality during this regulatory transition. The distinction between large audit firms and smaller audit firms is feasible in the German setting, since the audit engagements of listed firms are roughly evenly distributed among Big 3 and non-Big 3 audit firms in contrast to many other countries.
In the period before the mandatory adoption of IFRS, Van Tendeloo and Vanstraelen (2005)
conducted research in a similar area using data between 1999 and 2002. I will extend this research by focusing on earnings quality instead of solely on earnings management. Furthermore, I will provide more extensive evidence using a considerably larger sample. The number of companies which
voluntary adopted IFRS grew exponentially in the years before the mandatory adoption in 2005. The more even distributed use of IFRS and German GAAP offers the possibility to find a more well-founded explanation. Finally, I will extend the research of Van Tendeloo and Vanstraelen (2005) by investigating the effects of the mandatory adoption of IFRS in the period 2005 until 2009.
This paper will be structured as follows. In chapter 2, I will discuss the relevant literature and prior research in the field of earnings quality. Subsequently, the hypotheses will be developed based on this literature review. In chapter 3, the research methodology is explained. In chapter 4, I will discuss the research results. Finally, in chapter 5, the conclusions, limitations and future research directions will be addressed.
2
Literature review
I start this chapter by providing the main underlying definitions of earnings quality and earnings management. Subsequently, I will discuss prior research in the field of earnings quality in relation to regulatory change and audit firm size. Finally, the hypotheses will be formulated based on this literature review.
2.1 Earnings quality and earnings management
There are many different definitions of earnings quality defined in prior literature. In this research I approach earnings quality from the definition of Dechow et al. (2010). She defines earnings quality as: “Higher quality earnings provide more information about the features of a firm’s financial
performance that are relevant to a specific decision made by a specific decision-maker”
Following the definition of Dechow et al. (2010), earnings quality refers to a broad perspective of financial reporting. Earnings management can be considered as a less comprehensive concept and is considered as one of the influencing factors on earnings quality. Many studies concerning earnings quality are conducted from the perspective of earnings management. This study will, besides other determinants, also use earnings management as a measure of earnings quality. In this research I will approach earnings management from the definition of Healy and Wahlen (1999). This commonly used definition of earnings management is formulated as:
“Earnings management occurs when managers use judgment in financial reporting and in structuring transactions to alter financial reports to either mislead some stakeholders about the underlying economic performance of the company or to influence contractual outcomes that depend on reported accounting numbers”.
2.2 The impact of regulations on earnings quality
International Financial Reporting StandardsThe development of IFRS started in 1971 when the International Accounting Standards (IAS) were published by the International Accounting Standards Board (IASB). As successor to the IAS, the first IFRS was initiated in 2003. Eventually, IFRS became mandatory for European listed firms since 1 January 2005. IFRS had two main objectives, enhancing the quality of financial reporting and improving the comparability of financial reporting between different countries (De George et al., 2016). With
respect to reporting quality, the IASB aims to increase the transparency of financial reports by developing a single set of high quality accounting standards (De George et al., 2016). Further, the increased comparability of financial reporting would enable more international capital market opportunities. Many German listed companies for this reason choose to voluntary adopt IFRS before the mandatory adoption in 2005 (Weißenberger et al., 2004).
German GAAP
Before the mandatory adoption of IFRS the German Generally Accepted Accounting Principles (GAAP) was the commonly applied accounting standard under listed firms in Germany. Due to the ownership structure in Germany these accounting principles were mainly focused on preserve equity, creditor protection and expedite the computation of taxable income (Van Tendeloo and Vanstraelen, 2005). Compared to the more international focused accounting standards, German GAAP was considered as a more conservative accounting standard.
Comparison between IFRS and German GAAP
The transition from German GAAP to IFRS demanded a substantial change from German companies in the field of financial reporting. Firstly, these two reporting standards are in essence based on different underlying objectives. IFRS focuses on investor protection in contrast to German GAAP which is mainly focused on creditor protection. This is reflected in the way the different standards cope with ownerships structures and financing decisions. Where the conservative German GAAP is mainly focused on debt financing, aims IFRS to stimulate public equity financing (Van Tendeloo and Vanstraelen, 2005). IFRS tries to stimulate this public equity market by providing more transparent financial information in order to enable investors to create a better view of the true economic value of a company. Consequently, IFRS limits the flexibility of reporting standards and hence is expected to decrease information asymmetry and earnings management activities (De George et al., 2016). Further, Weißenberger (2004) recognizes several financial objectives for switching from German GAAP to IFRS. These objectives aimed to increase investor diversification, to stimulate international investor community, to reduce of the cost of equity and to enhance credit ratings. Considering these assumptions, the switch from German GAAP to IFRS would predominantly lead to improvements in the financial reporting environment and therefore reduce earnings management.
Prior research on the effect of regulations on earnings quality
With respect to the general impact of financial reporting regulations, Ewert and Wagenhofer (2005) find that tighter accounting standards will increase earnings quality. However, they argue that tighter accounting standards will make accrual based earnings management less effective but will not directly reduce it. In addition, Evans et al. (2015) suggests that a more effective regulatory environment would discourage accrual based earnings management but, on the other hand, encourage real earnings
management. As a result, the overall effect of tighter regulations would not lead to a significant reduction of earnings management. Further, Ball et al. (2003) argues that high-quality accounting standards will not naturally lead to high-quality financial reporting. In addition, accounting standards alone are not sufficient to improve financial reporting (Jeanjean and Stolowy, 2008). Jeanjean and Stolowy (2008) argue that national institutional factors and management incentives are at least as important.
Much research has been conducted in the field of the specific impact of IFRS on financial reporting quality. Heemskerk and van der Tas (2006) for example, assume that on one hand the less flexible standards under IFRS will reduce earnings management but on the other hand earnings management will increase due to the higher level of subjectivity under IFRS. They found that the amount of accrual based earnings management and earnings smoothing increased after the adoption of IFRS. However, Barth et al. (2008) find, more or less, opposite results concerning the adoption of IFRS. By examining 21 different countries they found that the application of IFRS led to less earnings management, higher accounting information relevance and more timely loss recognition. Moreover, Daske et al. (2008) conducted a more in depth analysis on the difference between voluntary and mandatory adopters of IFRS. Their results show that voluntary adopters experienced capital market benefits due to positive liquidity and valuation effects. Mandatory adopters, on the other hand, did not experience such positive effects from adopting IFRS.
In addition to the large number of studies being conducted on the general impact of IFRS, there is also research available in the specific German setting. For instance, Tendeloo and Vanstraelen (2005) investigated the effect of voluntary adoption of IFRS in Germany. They find no evidence for a
difference in the amount of earnings management between companies which voluntarily adopted IFRS and companies which applied German GAAP. Further, they find that companies under IFRS apply more accrual based earnings management. Finally, they found, similar to Heemskerken and van der Tas (2006), that companies engage more in earnings smoothing under IFRS. Gassen and Selhorn (2006) also investigated the voluntary adoption of IFRS in the German setting, while they found that companies applying IFRS present higher earnings quality. Consequently, the adoption of IFRS would also lead to lower information asymmetry. Finally, Christensen et al. (2015) finds mainly similar results with their German orientated research. They conclude, consistent with the results of Daske et al. (2008), that voluntary adopters of IFRS are related to lower levels of earnings management, more timely loss recognition and increased value relevance. In contrast, they found virtually no evidence for these improvements with respect to mandatory IFRS adopters.
2.3 The role of auditors
Audit servicesAccording to the International Auditing and Assurance Standards Board (IAASB), an assurance engagement is intended to enhance the confidence of the user in certain information. This refers to the underlying concept that auditing of financial information can be seen as a monitoring activity and therefore provides an incentive for managers to act in best interest of owners. Consequently, this would lead to a lower level of information asymmetry and therefore reduce agency costs (Jensen and Meckling, 1976). To what extent the objective of an assurance engagement is achieved mainly depends on the quality of an audit service. A common used definition of audit quality is formulated by
DeAngelo (1981). She defines audit quality as:
“The market-assessed joint probability that a given auditor will both (a) discover a breach in the client’s accounting system, and (b) report the breach.”
Prior research recognizes many different factors influencing audit quality. Knechel et al. (2013) for example defined influencing indicators on audit quality. These indicators include professional skepticism, expertise, auditor-client negotiations, audit partner compensation and auditor tenure. Moreover, Knechel et al. (2013) emphasizes the impact of audit firm size on financial reporting quality which I discuss next.
Big 4 versus non-Big 4
Does audit firm size impact earnings quality? This is a frequently asked question in previous literature. Most studies divide the concept of audit firm size in two segments; Big 4 and non-Big 4. Many studies claim that Big 4 audit firms provide higher audit quality. However, there exists no unambiguous inference that larger audit firms are related to higher earnings quality.
First, according to the study of DeAngelo (1981) larger audit firms are related to higher quality audits. The main reason for this would be that larger sized audit firms behave less opportunistic because of the lower client dependency. Further, the large capital requirements to obtain audit technology would counteract smaller sized audit firms. In line with these results, Lennox (1999) argues that large auditors provide more accurate audit reports since they are able to deploy more resources in their audit engagements. Moreover, large auditors are more susceptible to reputation damage and therefore expected to deliver more accurate audit reports. In addition, Balsam et al. (2003) suggests that Big 4 firms are positively related to earnings quality due to the fact that they possess more industry specific knowledge.
In contrast, Lawrence et al. (2011) claims that the expected audit quality under Big 4 versus non-Big 4 auditors should be almost the same. They argue that both Big 4 and non-Big 4 firms are held to the same regulations and standards and therefore expected to deliver the same audit quality on average. Their results indicate that audit quality may not merely depend on auditor characteristics but mostly on client characteristics. Boone et al. (2010) conducted similar research by comparing the audit quality of Big 4 audit firms and second tier audit firms. According to their results, there does not exist a
significant difference in the amount of abnormal accruals between these two groups. Consequently, they find no evidence for divergent financial reporting qualities between Big 4 and second tier audit firms.
More related to this particular research topic is the way different sized audit firms deal with complex and changeable regulatory environments. Following the reputational argument of Lennox (1999), Dumontier and Raffournier (1998) expect large audit firms to force their clients to apply with tighter regulations in order to protect their own reputation. Moreover, large auditors would be better able to adapt to international regulations because they are more internationally orientated and hence possess more experience and knowledge in this field. Various studies confirm these assumptions including that of Street and Gray (2002). They investigated the influencing factors on compliance with international accounting standards and found a significant positive relation between companies which were audited by large audit firms and the level of compliance with IAS disclosure requirements. In addition, Glaum and Street (2002) investigated this relationship for the specific German regulatory setting and found similar results. They found a positive relationship between Big 5 audit firms and the level of international compliance with both US GAAP and IAS for German listed firms.
2.4 Hypothesis development
In this last section I will discuss the hypothesis development. Based on the literature review in previous sections I formulated four different hypotheses. These hypotheses are expected to answer to the following research question:
To what extent does the voluntary and mandatory adoption of IFRS affect earnings quality under large versus small audit firms?
The main underlying subject of this research is to investigate the influence of more severe regulations on earnings quality. As the literature review indicates, there does not exist overall agreement among researchers about the impact of tighter regulations and more specific the adoption of IFRS on earnings quality. To contribute to this research I will investigate this relationship by using the regulatory change in the German market. As mentioned before, many German companies voluntarily adopted IFRS
before the mandatory adoption in 2005. This specific and not very common setting does allow me to draw a decent comparison between voluntary and mandatory adopters. In line with the results of Daske (2008), I expect that firms which voluntarily adopted IFRS show a higher earnings quality compared to firms which applied German GAAP. Further, I expect that the overall level of earnings quality increased after the mandatory adoption of IFRS compared to the pre-IFRS period. For this reason, I formulated the following hypotheses:
H1: German companies which voluntarily adopted IFRS show a higher earnings quality compared to companies which applied German GAAP.
H2: Earnings quality of German companies increased after the mandatory adoption of IFRS compared to the pre-IFRS period.
The second part of the research question relates to the connection between auditor firm size and regulatory changes. Based on the results of DeAngelo (1981), I expect that companies which are audited by large audit firms show higher earnings quality compared to companies which are audited by smaller audit firms. Moreover, several studies find a positive relationship between large audit firms and compliance with international audit standards (e.g. Street and Gray, 2002). Further, Dumontier and Raffournier (1998) assume that larger audit firms are better able to adapt to international
regulatory changes. Based on these results, I expect that the positive effect of IFRS on earnings quality will be reinforced at companies which are audited by larger audit firms. To operationalize these assumptions I developed the following hypotheses:
H3: The voluntary adoption of IFRS has more impact on earnings quality for German companies which are audited by Big 3 audit firms compared to companies which are audited by non-Big 3 audit firms.
H4: The mandatory adoption of IFRS has more impact on earnings quality for German companies which are audited by Big3 audit firms compared to companies which are audited by non-Big 3 audit firms.
A perhaps striking aspects of these hypotheses is that it separates Big 3 audit firms from other audit firms instead of Big 4 or even Big 5 audit firms which many studies do. The reason for this is that Deloitte is not part of the traditional Big 4 in German because it only represents 5% of the total market share, measured by the number of clients, compared to the other three firms which together represent over 50% of the total market share divided by nearly equal shares.
3
Research methodology
In the first section of this chapter I will discuss the data sample that will be used to conduct this research. Thereafter, in section 2, I will explain the employed measurement techniques and the corresponding regression models.
3.1 Sample
In order to conduct this research I will use an existing dataset of German listed firms and their auditors between 1999 and 2009. In total this dataset contains 5013 firm year observations. Companies which operate in the financial services industry (SIC code 6000-6800) are not included in this data set as they are subject to specific financial regulatory requirements.
For each observation several variables are available including the appointed audit firm, 4 digit SIC code, applied reporting standard and the necessary accounting variables. Data between 1999 and 2004 will be used to test the first and third hypotheses which addresses the voluntary adopters of IFRS. To test the second and fourth hypotheses I will use firm year observations between 1999 and 2009. In order to test the impact of mandatory adopters on earnings management, I will eliminate firms which voluntary adopted IFRS from this sample.
Further, I will make a distinction between two types of audit firms namely Big 3 and non-Big 3. The distribution between these two types of audit firms is almost even, representing both about 50% of the observations measured by the number of clients. Further, I will consider Arthur Andersen as a Big 3 office for the years 1999 until the collapse of the company in 2002 because of their substantial market share in this period. Thereby, the audit services of Arthur Andersen Germany were ultimately acquired by EY.
The distribution of the applied accounting standard varies over time among the observations. As figure 1 illustrates, over 80% of the companies applied German GAAP in the first year of observation. The number of German GAAP appliers shrank to just over 40% in the last year of observation. The reason for this is that a large number of companies choose to voluntary adopt IFRS as the mandatory adoption came closer. I will consider companies which applied German GAAP in 2004 as mandatory IFRS adopters for the remaining part of the sample. Further, I will include companies which applied US GAAP in the sample as IFRS adopters because both standards are based on the same underlying goals such as providing transparent and reliable information for investors (Weißenberger et al., 2004).
3.2 Measuring earnings quality
To draw solid conclusions about the level of earnings quality I will measure three aspects of earnings quality. These measures aim to determine the level of discretionary accruals, the degree of earnings smoothing and the value relevance of earnings.
3.2.1 Measuring discretionary accruals
First, I will apply an accrual based measure to determine the level of earnings management. By using the modified Jones model (Dechow et al., 1995) I will estimate the discretionary accruals.
Discretionary accruals are subject to managerial judgement and decision making and therefore are assumed to be a reflection of earnings management. In order to determine the discretionary accruals, I will first estimate the non-discretionary accruals. These accruals are assumed to be influenced by firm specific economic circumstances (Jones, 1991). In order to account for this effect, the modified Jones model computes firm-specific parameters. Further, Dechow et al. (1995) argues that revenues can be influenced by management decisions and hence need to be adjusted for in the discretionary part. This argumentation leads to the following model:
= + ∆ ∆ + + (1)
where:
= Non-discretionary accruals in year t.
= Difference between revenues in year t and year t-1. = Difference between net receivables year t and year t-1.
= Property, plant and equipment year t.
= Total assets t-1.
The firm specific parameters can be computed using the following equation:
= + ∆ ∆ + + (2)
where:
= Total accruals in year t.
The total accruals will be calculated by extracting the cash flow from operations from net income. Finally, the discretionary part of the total accruals will be calculated by extracting the non-discretionary accruals from the total accruals.
I will estimate the model for each different industry separately in order to account for industry specific factors. A sufficient number of observations by industry will be required to obtain reliable firm specific coefficients. In the case of the two digit SIC code industry classification a sufficient number of observations will not be obtained. For this reason, I will apply the industry classification introduced by Barth (1998). In table 1 the distribution among the different industry categories is provided.
Categories 6, 12 and 13 are not included in this table because none of the observations are part of these industry categories. These categories consist of extractive industries, financial institutions and real estate companies respectively.
Regression models discretionary accruals
As addressed in the literature review, earnings quality can be influenced by many different factors, just as the level of discretionary accruals. For this reason I will include several control variables in my research design. These control variables will account for size, financing structure and growth potential. One of the underlying motives for earnings management activities are lending contracts (Healy and Wahlen, 1999). Many of these lending contracts include debt covenants that should prevent
opportunistic management behavior (Jensen and Meckling, 1976). However, in some cases managers apply earnings management in order to avoid debt covenants violations. For this reason, I expect that a large amount of debt will have a positive effect on the level of earnings management. I will control for this effect by using variable FIN_LEV which is measured by total debt divided by total assets. The second control variable will account for the possible effects of size. According to Healy and Wahlen (1999), larger companies will apply earnings management to prevent regulatory interventions. The control variable SIZE will be used to account for this effect. This variable is measured by the natural logarithm of total assets and is expected to have a positive effect on the level of discretionary accruals.
Another aspect that influences the level of earnings management and consequently earnings quality are firms with a high growth potential (Skinner and Sloan, 2002). Companies with many growth
opportunities are expected to engage more in earnings management. By taking into account the book-to-market ratio I will control for the effects of GROWTH. A low book-book-to-market ratio will indicate a high growth potential firm and therefore expected to have a positive effect on the level of discretionary accruals.
Further, I will include Big 3 as an control variable in the regression models of hypotheses 1 and 2. As mentioned in the literature review, audit firm size is found to be a influencing factor on audit quality (e.g. DeAngelo, 1981). In line with these results, I expect that Big 3 audit firms have a negative effect on the absolute level of discretionary accruals. I will control for this effect by using a dummy variable which equals 1 if a company is audited by a Big 3 audit firm and 0 otherwise.
This research design leads to the following regression models concerning the absolute discretionary accruals (ADA): Hypothesis 1. (3) = + _ + _ + + + 3 Hypothesis 2. (4) = + _ + _ + + + 3 Hypothesis 3. (5) = + _ + 3 + _ ∗ 3 + _ + + Hypothesis 4. (6) = + _ + 3 + _ ∗ 3 + _ + +
As the models indicate, I will use the absolute value of the discretionary accruals. The reason for this is that the financial situation and financial motives of a specific company are unknown. As I am not interested in the direction of discretionary accruals but in the deviation of discretionary relative to zero, the absolute value of the discretionary accruals will be used. Further, the variable PRE_IFRS will be equal to 1 for voluntary IFRS adopters and 0 otherwise. Consequently, the variable POST_IFRS will indicate 1 after the mandatory adoption of IFRS and 0 before the mandatory adoption in 2005. Finally, I will use an interacting variable for hypotheses 3 and 4 to measure the effect of IFRS in relation to a Big 3 audit firm. Finally, year dummies will be included to account for year related effects.
3.2.2 Measuring earnings smoothing
The second aspect I will measure to determine earnings quality is earnings smoothing. Earnings smoothing is one of the patterns a manager can use to manage earnings (Scott, 2015). This is a commonly used pattern as both managers and shareholders prefer smooth earnings paths (De Jong et al., 2013).
Albrecht and Richardson (1990) developed a model to determine if companies engage in artificial smoothing. In the case of artificial smoothing managers shift cost and revenues between periods to establish a smooth earnings path. By dividing the coefficient of variation of net income (CVNI) by the
coefficient of variation in revenues (CVR) the relative coefficient of variation (CVIR) can be
calculated. Albrecht and Richardson (1990) argue that if the outcome of the model is less than 1 a company can be considered as a “smoother”. This model can be expressed in the following equation:
= ∆∆ (7)
= ∆∆ (8)
= (9)
where:
∆ = Net income in year t – Net income in year t-1 ∆ = Revenues in year t – Revenues in year t-1
In order to determine if a company can be considered as a smoother a largest possible time-series of observations is called for. For this reason, I only included companies for which a maximum number of observations is available in the dataset. In other words, only companies for which 6 observations are available before 2005 and 5 observations after 2005 are included in this model.
Further, companies which applied IFRS in 2004 are considered as voluntary adopters. In order to make a comparison between Big 3 and non-Big 3 audit firms, the engaging audit firm in 2004 will be considered as leading for the sample from 1999 until 2004. For the sample from 1999 until 2009 the engaging audit firm in 2009 will be considered as leading. This gives the following sample distribution concerning this measure as shown in table 2.
Regression models earnings smoothing
The smoothing coefficient will, similar to the discretionary accruals measure, be included as the dependent variable in the regression analyses. In the regression models, two control variables will be included to account for the effects of both operating and financial leverage.
The first control variable refers to financial leverage (FIN_LEV) and is similar to the control variable that will be used in the regression analyses concerning discretionary accruals. This variable will be calculated by dividing total debt by total assets. On one hand, a high financial leverage ratio is expected to have a positive relation on the smoothing coefficient. The reason for this is that a relative high amount of debt will increase the fixed costs due to higher interest costs. On the other hand, according to Healy and Wahlen (1999), managers are expected to apply earnings management to avoid debt covenants violations. Following this argument, financial leverage is expected to be negatively related to the smoothing coefficient. Based on this argumentation I cannot make an unambiguous assumption about the expected direction of this variable.
Second, the degree of operating leverage (OP_LEV) refers to the ratio between fixed costs and total costs. In general, firms with a high operating leverage face more income fluctuations. Therefore, I expect that operating leverage positively effects the smoothing coefficient. Because the fixed costs are not specified in the data set a proxy variable will be used to control for this effect. This variable is calculated by dividing fixed assets by total assets.
This leads to the following regression models:
Hypothesis 1. (10) = + _ + _ + _ Hypothesis 2. (11) = + _ + _ + _ Hypothesis 3. (12) = + _ .+ 3 + _ .∗ 3 + _ + _ Hypothesis 4. (13) = + _ .+ 3 + _ .∗ 3 + _ + _
3.2.3 Measuring the value relevance of earnings
The final method I will apply to determine earnings quality is the value relevance of earnings. The earnings response coefficient (ERC) expresses the relationship between shareholder returns and unexpected earnings and can be used to determine the value relevance of earnings. According to Scott (2015) earnings quality is, among other things, reflected in the ERC. More specific, he argues the higher the ERC the higher the earnings quality is because investors would be better able to estimate future cash flows. In addition, according to Nwaeze (2011), earnings uncertainty will lead to higher information asymmetry which will affect the perception across investors and consequently lowers the ERC.
As the ERC can be an appropriate proxy of estimating earnings quality, the measurement of this variable also brings some challenges. One of the main challenges concerns the determination of the expected earnings. The level of expected earnings depends on the persistency of earnings through time. When earnings are mostly persistent the expected earnings are relatively easy to estimate based on the results of previous years. In contrast, when earnings are less persistent the expected earnings are harder to estimate.
One way to address this challenge is introduced by Easton and Harris (1991). By using a regression model they aim to identify the correlation between the shareholder returns and the amount of unexpected earnings. By both examining the level of net income and the change in net income they found a strong correlation between shareholder returns and current earnings. In line with Easton and Harris (1991), I will both use the level of net income as well as the change in net income in order to determine the ERC. The following two equations will be used to calculate the ERC:
∶ = + + (14)
And:
ℎ ∶ = ∅ + ∅ ∆ + (15)
where:
= Fiscal year return in year t.
= Net income in year t
= Market value of equity in year t -1
I will use interacting variables in order to analyze the relationship between the ERC and applied accounting standard. The interacting variables in regression model 3 and 4 also include the engaged audit firm. Below the different regression models are presented based on Easton and Harris (1991).
Hypothesis 1. = + + ∗ + ∆ + ∆ ∗ (16) Hypothesis 2. = + + ∗ + ∆ + ∆ ∗ (17) Hypothesis 3. = + + ∗ ∗ 3 + ∆ + ∆ ∗ ∗ 3 (18) Hypothesis 4. = + + ∗ ∗ 3 + ∆ + ∆ ∗ ∗ 3 (19)
4
Results
In this chapter I will discuss the results. First, I will go into the results with respect to the modified Jones model. Thereafter, in section 2, the results concerning the earnings smoothing coefficient will be explained. Subsequently, in section 3, the results based on the earnings response coefficients will be presented. Finally, the results with respect to the additional tests will be discussed in section 4.
4.1 Discretionary accruals results
In table 3 the descriptive statistics concerning the accrual based measure are presented. DA indicates the discretionary accruals measured by the modified Jones model. The average value of this variable is slightly negative (-0,031) which indicates that firms apply earnings decreasing measures on average. The ADA variable indicates the absolute value of the discretionary accruals. The outliers within the variable DA are eliminated by using the winsorizing technique. The 99th and 1st percentile of the
values within this variable are adjusted by using this technique.
The dummy variable Big3 indicates 1 if a company is audited by a Big 3 audit firm and 0 if the company is audited by a non-Big 3 audit firm. As mentioned before, the distribution between Big 3 and non-Big 3 audit engagements is mainly even (0,518).
The second dummy variable indicates which financial reporting standard is applied. This variable equals 1 if a company applies IFRS and 0 if a company applies German GAAP. The mean of this variable indicates that 67,4% (0,674) of the total sample applied IFRS. The interacting variable Big 3 * Standard equals 1 if a company applies IFRS and in the same time is audited by a Big 3 audit firm. If German GAAP is applied or the company is audited by a non-Big 3 audit firm this variable equals 0.
In table 4 the descriptive statistics with respect to the absolute discretionary accruals are presented by individual hypothesis.
I applied an independent samples T-test in order to determine if specific means significantly differ
from each other. The descriptive statistics of H1 indicate that the companies which voluntary applied IFRS are related to significant higher discretionary accruals (0,108) compared to companies which applied German GAAP (0,092). Further, the average level of discretionary accruals does not significantly differ in the post-IFRS period (0,098) compared to the pre-IFRS period (0,092). No significant difference is found between the level of discretionary accruals under companies which voluntary adopted IFRS and are audited by a Big 3 auditor (0,098) and companies which are audited by non-Big 3 audit firms (0,098). Similar to the voluntary adopters of IFRS, the average level of discretionary accruals under mandatory adopters is not significantly lower for companies which are audited by a Big-3 audit firm (0,089) compared to a non-Big 3 audit firm (0,096).
In table 5 the Pearson correlation matrix is presented. This test is conducted to determine if variables are positively or negatively correlated and if variables are possibly subject to the influence of multicollinearity. According to this test, variable Big 3 is significantly negatively correlated (-0,043**) with the absolute discretionary accruals. This correlation is in line with the expectation that Big 3 audit firms are related to higher earnings quality and therefore a lower level of discretionary accruals. The same applies to Pre_IFRS * Big 3 (-0,039**) which is also expected to be negatively associated with the level of discretionary accruals. A similar negative correlation is found for the interacting variable Post_IFRS * Big 3 (-0,052). Further, the variable SIZE is significantly negatively correlated with the absolute discretionary accruals (-0,201**), which is contradictory to the expected negative correlation. FIN_LEV (-0,010), shows a negative correlation which also is in contrast to the expectation.
The strongest measured correlation is detected at the interaction variables. Obviously, these variables are highly correlated with the underlying dummy variables Pre-/Post-IFRS and Big 3. Based on the results of this test no unusual highly correlated variables are detected in this data. In addition, the VIF scores are calculated to test for multicollinearity. These results indicate no issues concerning
In table 6 the results of the regression analyses are presented. The four columns each represent a regression model regarding the four hypotheses. Both the coefficients and the P-values of the different variables are included in this table. In the second column the expected relation is included based on the literature review. The R-squared values indicate the explanatory power of the regression models.
The first model aims to find the effect of the voluntary adoption of IFRS on the level of discretionary accruals. The results of the first regression model indicate a significantly positive coefficient (0,025) at the one percent significance level for the dummy variable Pre_IFRS. This indicates that the voluntary adoption of IFRS has a significant positive effect on the level of discretionary accruals which is contradictory to the expected negative relation.
The second regression model compares the level of discretionary accruals between the pre-IFRS period and post-IFRS period. The results of this model indicate a negative coefficient (-0,004) for variable Post_IFRS. This implies that the mandatory adoption of IFRS negatively affects the level of discretionary accruals which is in line with the expected assumption. However, according to the P-value this relation is not significant at a 10% level.
The third regression model addresses the effect of Big 3 audit firms on the level of discretionary accruals for voluntary IFRS adopters. According to the results, the interacting variable Pre_IFRS * Big 3 has a significant negative effect on the level of discretionary accruals (-0,018) which corresponds to the expected relation.
The effect of Big 3 audit firms on level of discretionary accruals in the post-IFRS period is examined in the fourth regression model. This model finds, in line with the third regression model, a negative coefficient for the interacting variable Post_IFRS * Big 3 (-0,011). This indicates that Big 3 audit firms also have a negative effect on the level of discretionary accruals in the Post-IRS period. These results are, however, not significant.
Varying results are found with respect to the control variables. First, the control variable FIN_LEV indicates a small insignificant negative effect on the level of discretionary accruals around zero which does not clearly answer to the initial positive assumed relation. The regression results of control variable SIZE indicate a significant negative relation with the absolute discretionary accruals. These results are contradictory to the positive expected relation. Further, the control variable GROWTH indicates a negative effect on the level of discretionary accruals. Finally, the R-squared values vary between 0,054 and 0,039 which indicate that between 4 to 5 percent of the variance is explained by the variables in these regression models.
In summary, for the first and third regression model significant results are found. This implies that companies which voluntary adopted IFRS are related to a significant higher level of discretionary accruals and therefore are expected to engage more in earnings management activities. This
relationship is in line with the results of Tendeloo and Vanstraelen (2005) which also found a positive relation between accruals based earnings management and voluntary IFRS adopters. However, the
regression results indicate that voluntary adopters of IFRS which are audited by a Big 3 audit firm are related to a significant lower level of discretionary accruals compared to firms which are audited by a non-Big 3 audit firm. With respect to the remaining two regression models no significant results are found. This implies that the level of discretionary accruals did not significantly decline after the mandatory adoption of IFRS. Finally, the level of discretionary accruals under Big 3 audit firms for mandatory adopters is not significantly lower.
4.2 Earnings smoothing results
The model of Albrecht and Richardson (1990) is used to determine whether a company engaged in earnings smoothing activities. The values of CVIR with a Z-score above +/- 3 are eliminated to address
for the potential effects of outliers. Further, in line with Albrecht and Richardson (1990), the absolute value of the earnings smoothing coefficient is being used. In table 7 the descriptive statistics of the separate hypotheses are presented.
As the mean value of hypothesis 1 indicate, companies which apply German GAAP are associated
with a more volatile earnings path compared to voluntary IFRS adopters (9,69 vs 6,94). However, according to the independent samples T-test these two values do not significantly differ from each other. Second, the results indicate that there has been an increase in the amount of smoothing behavior in the Post-IFRS period (4,78) compared to the pre-IFRS period (9,69). According to the results of the independent samples T-test these values significantly differ from each other. Further, the results indicate that there does not exist a significant difference in the average smoothing coefficient between companies which are audited by Big 3 (6,60) versus non-Big 3 (9,20) audit firms and voluntary adopted IFRS. Similar insignificant results are found for audit firms in the post-IFRS period.
In table 8 the Pearson correlation matrix with respect to the smoothing coefficient is presented. First, the results indicate that the dummy variable IFRS is slightly negatively correlated with the smoothing coefficient which is in line with values presented in the descriptive statistics. The remaining variables do not show any significant correlations with the smoothing coefficient. As expected, the interacting variable is highly correlated with the underlying dummy variables IFRS and Big 3. Finally, the control variables indicate small significant correlations with the dummy variables. The VIF scores are
analyzed to test for potential problems of multicollinearity. No unusual high values are found which indicates no problems concerning multicollinearity.
In table 9 the outcomes of the regression models are presented in a similar way as the accrual based regression results. According to the results of the first regression model, companies which voluntary adopted IFRS are related to a lower earnings smoothing coefficient (-3,177). This would imply that these companies engage more in earnings smoothing activities. However, no significant P-value for this relation is found. The results of the second regression model indicate that companies engaged more in earnings smoothing activities after the mandatory adoption of IFRS (-5,089). These results are, in line with the T-test results, significant at a 5 percent level. The third regression model shows contradictory results with respect to the expected relation. As the results indicate, voluntary IFRS adopters which are audited by a Big 3 audit firm are negatively related to the earnings smoothing coefficient. A mainly similar negative relation is found for the fourth regression model, which means that mandatory adopters are related to a slightly lower smoothing coefficient when they are audited by a Big 3 audit firm. However, no significant results are found for these regression models.
To summarize, only for the second regression model significant results are found. This means that earnings smoothing behavior after the mandatory adoption of IFRS increased. This result is in line with previous research where increased earnings smoothing behavior is documented after the mandatory adoption of IFRS (e.g. Heemskerk and van der Tas, 2006. Tendeloo and Vanstraelen, 2005). Further, no significant evidence is found to infer that voluntary IFRS adopters are related to less earnings smoothing behavior. Finally, no significant differences in the smoothing coefficient are found between companies which are audited by Big 3 audit firms compared to non-Big 3 audit firms for both voluntary and mandatory IFRS adopters.
4.3 Value relevance of earnings results
The third and last measure to determine earnings quality is the value relevance of earnings. By using the model of Easton and Harris (1991) the earnings response coefficient is estimated based on both the level and the change in net income.
Table 10 includes the descriptive statistics of the variables which are used to apply the model of Easton and Harris (1991). First, the potential effects of outliers are addressed in line with the method of Easton and Harris (1991). This implies that for the level of net income and the changes in net income, scaled by market value of equity, only values within a range of +1,5 and -1,5 are included in the sample. Following this method, 202 observations are eliminated from the initial dataset. Second, the return figures are winsorized for values within 1st and 99th percentile of the sample.
In table 11 the Pearson correlation matrix is presented based on the level and change in net income. As the correlation coefficients indicate, all the independent variables are positively correlated with the variable return. Further, the VIF values do not indicate any problems with respect to multicollinearity.
In table 12 the results of the regression models are included. The results of the first regression model indicate that companies which voluntary applied IFRS are related to a lower ERC based on the level of net income (-0,116). However, this relation is not significant. In contrast, based on the net income changes, the ERC indicates an opposite relation. According to this variable, companies which voluntary applied IFRS are significantly related to a higher ERC (0,483). Accordingly, this would imply that investors derive more value from net income changes for companies which voluntary adopted IFRS.
The second regression model compares the ERC’s between the Pre- and Post-IFRS period. The results of the first interacting variable show a significantly positive relation (0,178) between companies in the post-IFRS period and the ERC based on the level of net income. A same relation is found based on the changes in net income (0,047), these results are, however, not significant. These results indicate that investors derive more value from the level of net income after the mandatory adoption of IFRS. In the third and fourth regression model, the differences in ERC between Big 3 and non-Big 3 audited firms are being investigated. According to the results of the third regression model, companies which voluntary adopted IFRS and are audited by a Big 3 audit firm are significantly related to a lower ERC
(-0,314) based on the level of net income. An opposite relation is, however, found for the ERC based on changes in net income (0,349). On one hand this would mean that investors derive less explanatory power from the level of net income. On the other hand, the explanatory power has increased based on net income changes for these companies. A similar conclusion can be derived from the results of the fourth regression model. These results also indicate a contradicting relationships based on the level and change in net income. However, significant results for the positive coefficient of the change in net income are not found.
To summarize, voluntary adopters of IFRS are significantly related to a higher ERC based on net income changes and are therefore expected to deliver a higher earnings quality. Further, based on net income levels, a higher ERC is documented in the Post-IFRS period which would indicate an increase in earnings quality. These results are partially in line with Christensen (2015) which found a higher value relevance for voluntary IFRS adopters. However, his research did not document significant results for mandatory IFRS adopters. Finally, the results indicate that investor derive less (more) explanatory power from the level (change) of net income when a company is audited by a Big 3 audit firm.
4.4 Additional tests
In order to test the robustness of the results presented in the previous section, three additional tests are applied by using alternative measurement techniques.
Defond and Park Model
An alternative way to estimate the level of discretionary accruals is introduced by Defond & Park (2001). In contrast to the modified Jones model, this model does not apply firms specific parameters but estimates the discretionary working capital accruals based on previous year revenues. This model is formulated as follows:
= − [ ∗ ] (20)
where:
= Discretionary working capital accruals scaled by lagged total assets. = (Current Assets – Cash and Short-term Investments) –
(Current Liabilities – Short-term Debt) = Revenues in year t
In line with the modified Jones model the outliers within the 99th and 1st percentile of the sample are
winsorized. Further, similar to the modified Jones model, the absolute values of the discretionary working capital accruals are used. The summarized descriptive statistics and T-test results are included in the appendix (table 16). The descriptive statistics and T-test results show mostly similar results compared to the modified Jones model. According to the Defond and Park model, the level of discretionary working capital accruals is higher under voluntary (H1) and mandatory (H2) IFRS adopters. Further, both H3 and H4 indicate that companies which are audited by Big 3 audit firms are not related to a different level of discretionary working capital accruals compared to non-Big 3 audit firms. The condensed regression results of this model are presented in table 13.
The regression results are in line with the results based on the modified Jones model. This implies that for regression model 1 significant results are found which demonstrates that the level of discretionary accruals has increased for voluntary IFRS adopters. For the mandatory adoption of IFRS (H2) and the engagement of a Big 3 audit firm (H3 and H4) no significant results are found. A possible explanation for the different results compared to the modified Jones model is that no industry specific effects are taken into account by using the Defond & Park model. The same control variables as the modified Jones model are used in the regressions. These variables are not included in this table as they show
mainly similar results. Finally, I correlated the results based on the modified Jones Model with the results of the Defond and Park model. This test shows a correlation coefficient of 0,551 which clarifies the consistency of the results but also indicates the added value of the Defond and Park model in explaining the relation with respect to accrual based earnings management.
Earnings smoothing
In order to test the robustness of the earnings smoothing results, I apply an additional smoothing measure introduced by Burgstahler et al. (2006). They measure earnings smoothing by dividing the standard deviation of net income by the standard deviation of cash flow from operations. As the outcome is being multiplied by -1 higher values indicate more earnings smoothing behavior. This leads to the following equation:
= −1 ∗ (21)
In line with Leuz et al. (2003), I scaled both net income and cash flow from operations by lagged total assets in order to account for size related effects. Ultimately, I eliminated outlying values consistent with the smoothing measure of Albrecht and Richardson (1990). In table 14 the condensed regression results are presented:
As the results indicate, no significant relations are found with respect to the applied reporting standard and the effect of Big 3 audit firms. Consequently, these results indicate that the amount of earnings smoothing behavior did not significantly decline during both the voluntary and the mandatory adoption of IFRS. Further, in line with the results based on the model of Albrecht and Richardson (1990), I found no evidence to infer that Big 3 audit firms are related to a lower amount of earnings
smoothing compared to non-Big 3 audit firms. The slightly different results compared to the initial earnings smoothing measure can be due to the fact that this model does not use the coefficient of variation and hence do not take into account the average level of net income and cash flows from operations. The difference between the two measures of earnings smoothing is also reflected in the correlation coefficient as this shows a relatively low value (0,126).
Earnings conservatism
Earnings conservatism will be analyzed, in order to test the robustness of the ERC results. Earnings conservatism refers to the way in which companies respond to both positive and negative returns. One way to determine earnings conservatism is introduced by Basu (1997). This model can be formulated in the following way:
= + + _ + ∗ _ (22)
where:
= Net Income before extraordinary items in year t.
= Market value of equity in year t-1
= Fiscal year return
_ = Dummy variable which equals 1 if the annual returns are negative and 0 if earnings are positive.
Further, dummy variables are added for applied reporting standard and engaged audit firm to test the different regression models. In line with the previous applied models I winsorized the 99th and 1st
percentile of earnings, returns and market value of equity to account for the possible effect of outliers. In table 15 the condensed regression results of this model are presented. The complete regression results are included in the appendix (table 17).
The results of the interacting variable NEG_RET * RET indicate a significant positive coefficient for all four regression models. Following Basu (1997), this indicates that the companies in these samples can be considered as conservative on average. The results of the first regression model indicate a slightly positive coefficient (0,088) which implies that voluntary adopters of IFRS can be considered as more conservative than German GAAP appliers. This result, however, is not significant. The second regression model shows a opposite relation for mandatory IFRS adopters (-0,220). This suggests that companies became less conservative after the mandatory adoption of IFRS. Significant evidence for this relation is not found. Finally, the third and fourth regression model both show a negative coefficient. This would indicate that companies which are audited by Big 3 audit firms are less conservative for both voluntary and mandatory adopters of IFRS. Only for the third regression models significant evidence is found. This implies that voluntary adopters which are audited by a Big 3 audit firm are significantly less conservative.
5
Conclusion and Discussion
The main goal of this research is to evaluate the effect of stricter regulations on earnings quality by examining the change from national GAAP to IFRS in the German setting. More specific, I
investigated the change in earnings quality for voluntary and mandatory IFRS adopters. Moreover, I examined if there exists a positive relation between Big 3 auditors and earnings quality during this regulatory reform. In order to determine earnings quality I applied three different models. These models aim to determine the level of discretionary accruals, earnings smoothing and the value
relevance of earnings. By using these three determinants, I intended to find an answer on the following research question:
To what extent does the voluntary and mandatory adoption of IFRS affect earnings quality of firms with larger versus smaller auditors?
Based on the literature review, voluntary adopters of IFRS are expected to be positively related to earnings quality (e.g. Daske, 2008). In line with the results of Weißberger (2004), a similar relation is expected for mandatory IFRS adopters. For both voluntary and mandatory IFRS adopters an even higher improvement in earnings quality is expected at companies which are audited by a larger audit firm (Glaum and Street, 2002).
In this study, voluntary IFRS adopters were found to engage more in accrual based earnings
management compared to German GAAP appliers. However, evidence indicates that this increase of accrual based earnings management is smaller when firms are audited by a Big 3 audit firm. For mandatory IFRS adopters a significant change with respect to accrual based earnings management is not found. A possible explanation for this would be that audit firms became more experienced with respect to the application of IFRS and were therefore better able to detect accrual based earnings management.
Moreover, mandatory IFRS adopters are related to significantly more earnings smoothing behavior since the implementation in 2005. A difference in smoothing behavior for voluntary IFRS adopters is not found. The results also indicate that there does not exist a significant difference in earnings smoothing behavior between companies which adopted IFRS and are audited by a Big 3 audit firm compared to a non-Big 3 audit firm.
Another finding is that investors derive less (more) value from the level (change) of net income for companies which voluntary adopted IFRS and are audited by a Big 3 audit firm. No unambiguous evidence is found for such a change in the earnings response coefficient for mandatory adopters of IFRS.
Three alternative measures are used to test the robustness of these findings. Mainly similar results are found for accrual based earnings management based on the Defond & Park model. Further, based on the Burgstahler model no significant differences are found in smoothing behavior after the adoption of IFRS nor in relation to a Big 3 audit firm. Finally, the results concerning earnings conservatism indicate that firms which voluntary adopted IFRS and are audited by a Big 3 audit firm are significantly less conservative.
Based on the results of this study I conclude that earnings quality has not improved by both the voluntary and mandatory adoption of IFRS. Moreover, I infer that earnings management activities including discretionary accruals and smoothing behavior slightly increased after the adoption of IFRS. For this reason, I reject hypotheses 1 and 2. Further, some evidence is found to infer that voluntary adopters which are audited by a larger audit firm are related to higher earnings quality compared smaller audit firms. As this evidence is less convincing for mandatory adopters, I cannot draw an unambiguous conclusion based on these results. I therefore accept hypothesis 3 and reject hypothesis 4.
The results of this research are inevitably subject to some limitations and therefore need to be interpreted with certain caution. First, despite the use of different measurement techniques, an all-encompassing measure to determine earnings quality does not exist. For this reason, the measured earnings quality might differ from the actual perceived earnings quality. Second, there are some limitations with respect to the used data set. The results of this research are based on data between 1999 and 2009 which therefore may give a different view about the current level of earnings quality since IFRS remains into a process of further development. Further, the data set only contains information about the German market which means that the results of this research may not be generalizable for other European countries.
Finally, this research provides some possibilities for future research. First, in line with the limitations of this study, future research might focus on different countries and using more recent data. Further, only three of the various earnings quality measures were tested. For this reason, future research may use other earnings quality measures. This may, for example, imply the impact on the level of real earnings management in the mandatory setting as Van Tendeloo and Vanstraelen (2005) already documented varying results for voluntary IFRS adopters. However, it needs to be taken into account that real earnings management is to lesser extent affected by financial accounting regulations.
6
References
Albrecht, W., & Richardson, F. (1990). Income smoothing by economy sector. Journal of Business Finance & Accounting, 17(5), 713-730.
Aubert, F., & Grudnitski, G. (2014). The impact of SOX on opportunistic management behavior. International Review of Financial Analysis, 32, 188-198.
Ball, R., Robin, A., & Wu, J. (2003). Incentives versus standards: Properties of accounting income in four east Asian countries. Journal of Accounting and Economics, 36(1), 235-270.
Balsam, S., Krishnan, J., & Yang, J. (2003). Auditor industry specialization and earnings quality. Auditing: A Journal of Practice & Theory, 22(2), 71-97.
Barth, M., Beaver, W., & Landsman, W. (1998). Relative valuation roles of equity book value and net income as a function of financial health. Journal of Accounting and Economics, 25(1), 1-34.
Barth, M., Landsman, W., & Lang, M. (2008). International accounting standards and accounting quality. Journal of Accounting Research, 46(3), 467-498.
Basu, S. (1997). The conservatism principle and the asymmetric timeliness of earnings. Journal of Accounting and Economics, 24(1), 3-37.
Boone, J. P., Khurana, I. K., & Raman, K. K. (2009). Is audit quality different for Big 4 and mid-tier auditors? Working Paper, University of Texas.
Burgstahler, D., Hail, L., & Leuz, C. (2006). The importance of reporting incentives: Earnings management in European private and public firms. The Accounting Review, 81(5), 983-1016.
Christensen, H., Lee, E., Walker, M., & Zeng, C. (2015). Incentives or standards: What determines accounting quality changes around IFRS adoption? European Accounting Review, 24(1), 31-61. Daske, H., Hail, L., Leuz, C., & Verdi, R. (2008). Mandatory IFRS reporting around the world: Early evidence on the economic consequences. Journal of Accounting Research, 46(5), 1085-1142.
DeAngelo, L. (1981). Auditor size and audit quality. Journal of Accounting and Economics, 3(3), 183-199.
Dechow, P., Sloan, R., & Sweeney, A. (1995). Detecting earnings management. The Accounting Review, 70(2), 193-225.
Dechow, P., Ge, W., & Schrand, C. (2010). Understanding earnings quality: A review of the proxies, their determinants and their consequences. Journal of Accounting and Economics, 50(2-3), 344-401.
DeFond, M., & Park, C. (2001). The reversal of abnormal accruals and the market valuation of earnings surprises. The Accounting Review, 76(3), 375-404.
De George, E., Li, X., & Shivakumar, L. (2016). A review of adoption literature. Review of Accounting Studies, 21(3), 898-1004.
De Jong, A., Mertens, G., Van der Poel, M., & Van Dijk, R. (2014). How does earnings management influence investor’s perceptions of firm value? survey evidence from financial analysts. Review of Accounting Studies, 19(2), 606-627.
