The Development of Accounting Quality of IAS and IFRS Over Time:

The Case of Germany1

April 2008, Revised July 2008.

Mari Paananen2
Business School
University of Hertfordshire,
College Lane, Hatfield, AL10 9AB
United Kingdom
m.paananen@herts.ac.uk

Henghsiu Lin
University of Portland
Pamplin School of Business Administration
5000 North Willamette Blvd.
Portland, OR 97203
lin@up.edu

1
We thank the anonymous reviewers of the Journal of International Accounting Research, Dr. Eva
Jermakowicz, the participants at the AAA 2008 Semi-annual International Accounting Section meeting in
San Diego, and the participants at workshops at Göteborg University, Keele University and Warwick
University for helpful comments.
2
Corresponding author

Electronic copy available at: http://ssrn.com/abstract=1066604

 Abstract
We examine the characteristics of accounting amounts using a sample of German

companies reporting under IAS during 2000-2002 (IAS period), and IFRS during 2003-

2004 (IFRSvoluntary period) and 2005-2006 (IFRSmandatory period). We find a decrease in

accounting quality after the mandatory EU adoption in 2005. Our findings on earnings

smoothing and timely loss recognition corroborates largely our findings related to value

relevance of accounting information. Our results indicate that accounting quality has not

improved but worsened over time. Further analysis shows that this development is less

likely be driven by new adopters of IFRS but is driven by the changes of the standards.

Contrary to the intention with the adoption of the European adoption of IFRS, this makes

it harder for investors to base their decisions on the IFRS financial reporting.

Electronic copy available at: http://ssrn.com/abstract=1066604

I. Introduction

The main objective of this study is to examine and compare the quality of

accounting numbers under International Accounting Standards (IAS) during 2000-2002

with those under International Financial Reporting Standards (IFRS) during 2003-20063.

Accordingly, we compare the characteristics of accounting amounts using a sample of

German companies reporting under IAS during 2000-2002 (IAS period), and IFRS during

2003-2004 (IFRSvoluntary period) and 2005-2006 (IFRSmandatory period). Specifically, we

investigate whether there is a change in accounting quality during these three time

periods. We limit our investigation to German companies to hold constant certain

institutional factors such as stock listing requirements, accounting disclosure

requirements, market microstructures and regulatory environments that may confound the

results, thereby strengthening the reliability of our findings.

Our inquiry is mainly motivated by the major revisions of IASs and the

development of new IFRSs since the implementation of the new structure with IASB

taking over the standard setting responsibilities from IASC in April 2001. The IASB has

focused on developing a set of high quality standards to promote global accounting

harmonization. This has led to significant changes in standards; only 31 of the 41 IASs

remained in effect as of January 2005. In addition, by this time, IASB has issued eight

new IFRSs (IASB). Many of these revisions and new additions of accounting standards

reflect IASB’s preference for fair value measurement of assets and liabilities (Alexander

and Jermakowicz 2006; Hung and Subramanyam 2007; Schipper 2005; Whittington

2005). Considering the developments in the international standards, we predict that these

3
Beginning in 2005, all listed companies in the European Union (EU) are required to prepare their
consolidated financial statements in accordance with IFRS. Prior to that, using IAS/IFRS was voluntary in
many European countries.

3
changes are likely to affect the quality of accounting amounts as a result of IASB’s

increased orientation towards fair value accounting.

The European Union’s (EU) adoption has made IFRS the most widely accepted

financial accounting model in the world. It is very important that current and potential

investors as well as the standard setters understand the implication of IFRS on accounting

variables. Hence, we also examine the effect of the EU’s mandatory adoption of IFRS in

2005 and 2006 on the quality of accounting. Specifically, we examine whether the

mandatory adoption of IFRS in the EU may have changed the structure of the set of

companies reporting under IFRS and how this may have impacted the overall accounting

quality.

Prior research has compared properties of accounting numbers using samples of

German companies (e.g., Hung and Subramanyam 2007; Bartov and Kim 2005; Barth et

al. 2008, Barth et al. 2006). Our study differs from prior research on quality of IAS and

IFRS accounting measures in that we examine the change in quality of accounting caused

by the revisions made to IASs and the development of new IFRSs. This study compares

the characteristics of accounting amounts using a sample of German companies reporting

under IAS during 2000-2002, and IFRS during 2003-2004 and 2005-2006. Specifically,

we investigate whether there is a change in accounting quality during these three time

periods as IASB revises existing IAS and issues new IFRS to formulate a set of high

quality international accounting standards for global financial reporting purpose.

Contrary to our expectations, our results suggest a decrease in accounting quality after the

mandatory EU adoption of IFRS. We find that earnings and book value of equity are

becoming less value relevant during the IFRS periods compared to the IAS period. Our

4
findings on earnings smoothing and timely loss recognition largely corroborates our

findings with respect to the value relevance of accounting information. Our results

indicate that accounting quality has not improved but worsened over time. Further

analysis shows that this is less likely to be driven by new adopters of IFRS in 2005 and

more by the change in international accounting standards. When using a matched sample

we cannot find any clear indication of either an improved or a worsened quality of

financial reporting. We also analyzed the voluntary and the mandatory adopters in the

period 2005 to 2006. We found only weak indications that the decrease in quality was

caused by the mandatory adopters. Furthermore, we also investigated whether our results

were driven by a new dominating industry group, Financials. We find nothing supporting

that this is the case and our interpretation is that the decrease in accounting quality is

driven by the revisions of IASs and the addition of new of IFRSs around the time of the

European mandatory adoption. Contrary to the intention with the European adoption of

IFRS, this might make it harder for investors to base their decisions on the accounting

information.

Our primary contribution is that we exclusively examine the impact of

international standards over time on accounting quality and value relevance of accounting

measures as these standards go through revisions and new standards are issued. No

study, to our knowledge, has empirically examined this issue. Our second contribution is

that we include more recent data and investigate the effects of the mandatory EU

adoption of IFRS since 2005 on accounting quality.

The remainder of our paper is organized as follows. The next section describes

the development of international accounting standards over time. Section III briefly

5
discusses prior research and develops hypotheses. Section IV discusses the research

design and how we examine earnings smoothing, timely loss recognition and value

relevance over time. Section V describes the sample, and the sample selection criteria.

Section VI presents our findings while Section VII concludes.

II. The Development of International Accounting Standards over Time

During the period of our investigation, a number of revisions to International

Accounting Standards took place. These changes are summarized in Table 1, which

outlines the revisions of existing IASs and the issuance of new IFRSs in the

chronological order these changes went into effect. Of these changes, we consider three

to have had a major impact on companies’ financial reporting and thereby the possibility

of a major impact on the value relevance on book value of equity and earnings. Theses

changes are related to IAS 36, Impairment of assets, IAS 38, Intangible assets, and IFRS

3, Business combinations. IAS 36 requires a review of assets including intangible assets

with an indefinite useful life for impairment, and measurement of recoverable amounts

on an annual basis. Any impairment loss is recorded as an expense in the income

statement. IAS 38 requires the recognition of an intangible asset when it is probable that

future benefits of an intangible asset will benefit the company, and the cost of the

intangible asset can be measured reliably. In addition, intangible assets should also be

assessed for impairment in accordance with IAS 36. IFRS 3 allows only the purchase

method for business combinations. All identifiable assets and liabilities are valued at fair

value. Goodwill is not amortized, but subject to an impairment test annually. Negative

goodwill is recognized immediately in the income statement. Since the changes in these

6
accounting standards are all fair-value oriented, we expect the informativeness of IFRS

earnings and book values to increase compared to that of IAS earnings and book values.

<Table 1 about here>

II. Prior Research and Hypotheses Development

As noted earlier, some recent studies compare IAS accounting measures to those

under other GAAPs. Hung and Subramanyam (2007) compare the financial statement

effects of using IAS to those using German GAAP for a sample of German companies

that elected to adopt IAS by examining these companies’ restatements of prior years

accounting numbers in the adoption year. They find that the adjustments between the

two reporting systems are value relevant for book values of equity, but not for earnings.

But they do not find any difference in value relevance of book value of equity and

earnings under IAS and German GAAP. They also find that total assets and book value

of equity are significantly higher under IAS and that there is a higher variability in book

value of equity and earnings under IAS. Finally, they find that IAS adopters exhibit

larger loss provisions. Bartov et al. (2005) also examine and compare the value relevance

of earnings based US GAAP, IAS and German GAAP. They, on the other hand, find that

IAS earnings are more value relevant than those based on German GAAP. The

difference in the results of these two studies may be found in that Bartov et al. (2005)

exclude loss-firm observations in their estimations while these are included in the Hung

Subramanyam (2007) study.

Jermakowicz et al. (2007) examine German companies’ adoption of IFRS and US

GAAP over the period 1995 to 2004. Specifically, they investigate the usefulness,

proxied as value relevance, before and after the adoption of these GAAPs and the

7
perceived benefits and costs related to the process of implementing IFRS among the

DAX-30 companies.4 They find a significant increase in the value relevance of earnings

after the adoption of these GAAPs. They also find that the key challenges related to the

adoption of IFRS are the complexity of IFRS, the costs involved, and the lack of

implementation guidance. The challenges related to the adoption of IFRS documented by

Jermakowicz et al. (2007) and Soderstrom and Sun (2007) may explain the findings of

Christensen et al. (2007). Christensen et al. (2007) investigate the change in earnings

management and timely loss recognition among German firms that voluntarily adopt

IFRS and those who wait until the adoption of IFRS is mandatory. They find that

companies that voluntarily adopt are less prone to earnings management and recognize

losses more timely compared to those that resist and wait until the adoption of IFRS

becomes mandatory. They interpret their findings as a sign of how certain companies (i.e.

insider oriented companies) have less incentive to adopt IFRS since they will not benefit

and the challenges involved are considerable.

Finally, Barth et al. (2008) also study IAS adopters from a number of countries,

whereof Germany is one of the countries with greatest representation in the sample.

They find that firms that adopt IAS are less prone to engage in earnings smoothing and

recognize losses more timely.

There are also other recent studies on the effect of German and other GAAPs’ on

accounting quality and cost of capital. Leuz and Verrecchia (2000) investigate the bid-

ask spreads, trading volume, and stock return volatility as proxies for the information

asymmetry part of cost of capital. Comparing the above proxies for German companies

4
DAX-30 (Deutscher Aktien IndeX 30 (formerly Deutscher Aktien-Index-30) is a Blue Chip stock market
index consisting of the 30 major German companies trading on the Frankfurt Stock Exchange.

8
which switch from German GAAP to either IAS or US GAAP, as they predict, they find

that the bid-asked spread decreases, and the trading volume increases, however they find

no reduction in stock return volatility. Daske (2006) builds on Leuz and Verrecchia’s

(2000) study using data from 1993 and 2002. He, on the other hand, does not find any

sign of a lower cost of capital for companies that switch to IAS or US GAAP. On the

contrary, Daske (2006) finds an increase in cost of capital for these companies. Finally,

Platikanova and Nobes (2006) compare the information asymmetry component of the

bid-ask spread among companies before and after EU’s adoption of IFRS in 2005. They

find a larger volatility in the information asymmetry for UK and German companies.

Contrary to expectations, they also find that companies from countries where earnings

management is more common exhibit a lower information asymmetry component

compared to other groups of countries. They interpret this result as income smoothing

reduces information asymmetry.

Overall, the results of these studies do not provide clear evidence on how the

recent development in the global accounting standards impacts the quality of the

accounting amounts. For instance, Barth et al. (2008) and Jermakowicz et al. (2007)

cover a period including both IAS and IFRS data, which makes it difficult to interpret

their results regarding the impact on accounting quality as the international accounting

standards go through changes over time. In addition, the fact that Bartov et al. (2005)

exclude loss-firm observations and obtain a result different from Hung and Subramanyam

(2005) suggests that certain characteristics of the companies reporting under international

accounting standards may drive the results. This notion is supported by the findings of

both Jermakowicz et al. (2007) and Christensen et al. (2007). Jermakowicz et al. (2007)

9
results suggests that the value relevance of earnings increases after companies adopt

IFRS or US GAAP, a notion that make sense considering the sample used in the study

(DAX-30 companies), a set of companies that are most likely to be able to cope with the

complexity of implementing these GAAPs. They also find that, in spite of these

companies’ ability to cope with an adoption to a more complex GAAP, they still find the

adoption of IFRS to be a major challenge due to its complexity, high cost, and the lack of

implementation guidance. Christensen et al.’s (2007) results suggests that companies that

have an incentive to implement a more challenging GAAP are more likely to maintain a

higher accounting quality (proxied as earnings management and timely loss recognition)

compared to those who do not.

We assume that the recent developments in the international accounting standards

have led to changes in the quality of financial reporting over time. Therefore, the

question remains whether the accounting quality is higher as a result of the IASB’s

initiatives and actions. As the IASB reduces the allowable alternative accounting

methods and choices and provides a more consistent approach to accounting

measurement for the goal of developing a single set of high quality international

accounting standards, we predict that these changes in recent years improve the quality of

accounting as evidenced by higher value relevance of earnings and book value of equity,

less earnings smoothing, and more timely recognition of losses.

With respect to value relevance, we expect to see higher association between

stock prices and earnings and book value of equity for firms with higher quality of

reported accounting numbers. Moreover, we expect that firms with less earnings

smoothing will exhibit more variability in change in net income, a higher ratio of the

10
variability of change in net income to variability of change in cash flow, a less negative

correlation between accruals and cash flows and less frequency of reporting small

positive earnings as the accounting quality improves. With respect to timely loss

recognition, we predict that firms with higher accounting quality show a larger frequency

of large losses.

III. Research Design

We follow Barth et al. (2008) and Lang et al. 2005) when testing our predictions

of higher quality of accounting as the IASB revises IAS and issues new IFRS in the

recent years, we divide our study period (2000-2006) to three time periods: the IAS

period ranging from 2000-2002; the IFRSvoluntary period ranging from 2003-2004 and the

IFRSmandatory period ranging from 2005-2006. Following prior research, we

operationalize quality of accounting using earnings smoothing, timely recognition of

losses measures, and value relevance.

Four measures of earnings smoothing are used in this study. They are the

variability of the change in net income, the ratio of the variability of the change in net

income to the variability of the change in operating cash flows, the correlation between

accruals and cash flows and the frequency of small positive net income (Lang et al. 2005;

Barth et al. 2006, 2008). Our first earnings smoothing metric is the variability of the

change in net income scaled by total assets, ∆NI (Barth et al. 2006, 2008; Lang et al.

2006). To control for other economic factors that affect earnings variability unrelated to

the financial reporting system, we regress ∆NI on a number of control variables identified

in prior literature (Ashbaugh 2001; Pagano et al., 2002; Lang et al., 2003; Tarca 2004;

11
Lang et al. 2006; Barth et al., 2006, 2008), and the variances of the residuals of the

regression is our measure of the earnings variability. The ∆NI is estimated as follows:

∆NIit = β 0 + β 1LEV 1it + β 2Growth it + β 3Eissueit + β 4 Dissueit + β 5Turnit + β 6 Sizeit +

(1),
β 7CFOit + β 8 AUDit + β 9 NUMEX it + β10 XLISTit + β11 FFit + εit.

where:

LEV = the total liabilities divided by shareholders’ equity;

GROWTH = the percentage of change in sales;

Eissue = the percentage change in common shareholders’ equity;

Dissue = the percentage change in total liabilities;

Turn = sales divided by total assets;

Size = the natural log of total assets;

CFO = the cash flow from operating activities scaled by total assets

AUD= a dummy variable taking the value one if the firm’s auditor is PwC, KPMG,

Arthur Andersen, E&Y, or D&T and zero otherwise;

NUMEX = the number of stock exchanges on which a firm’s stock is listed;

XLIST = a dummy variable taking the value one if the firm is listed on a U.S. stock

exchange (the U.S. not being the primary exchange) and zero otherwise.

FF = the average number of shares traded the last day of the month during the fiscal year

divided by number of common shares outstanding at the fiscal year end.

We estimate equation (1) pooling observations in each of the three time periods examined

and compare the variances of the residuals of the regression for each time period using a

two-tailed variance ratio F-test.

Our second measure of earnings smoothing is the ratio of the variability of the

change in net income, ∆NI, to the variability of the change in operating cash flows,

12
∆CFO (Barth et al., 2006; Lang et al., 2006). ∆CFO is the change in cash flows scaled by

total assets. As with ∆NI, to control for other economic factors that affect cash flows

variability unrelated to financial reporting system, we regress ∆CFO on a number of

control variables similar to equation (1), but with ∆CFO as the dependent variable.

∆CFOit = β 0 + β 1LEV 1it + β 2Growth it + β 3Eissueit + β 4 Dissueit + β 5Turnit + β 6 Sizeit +

(2),
β 7CFOit + β 8 AUDit + β 9 NUMEX it + β10 XLISTit + β11 FFit + εit.

The variability of the change in cash flows is the variance of residuals from equation (2).

Then the second measure of earnings smoothing is the ratio of the variability of ∆NI to

the variability of ∆CFO. The rationale for using this ratio is that it is plausible that the

variability of net income is affected by the firm-specific volatility of cash flows and by

using the ratio we control for this (Barth et al. 2006, 2008; Lang et al. 2006). We measure

the difference between these ratios across time periods using ranksum test of the

permuted dataset.

The third measure of earnings smoothing is the spearman correlation between

accruals and cash flows. As with the previous tests, to control for economic factors

unrelated to earnings smoothing, we run separate regressions of accruals and cash flows

on the control variables as included in Equations 1 and 2, except CFO.

CFOit = β 0 + β 1LEV 1it + β 2Growth it + β 3 Eissueit + β 4 Dissueit + β 5Turnit + β 6 Sizeit +

(3)
β 7 AUDit + β 8 NUMEX it + β 9 XLISTit + β10 FFit + εit.

ACCit = β 0 + β 1LEV 1it + β 2Growth it + β 3 Eissueit + β 4 Dissueit + β 5Turnit + β 6 Sizeit +

(4)
β 7 AUDit + β 8 NUMEX it + β 9 XLISTit + β10 FFit + εit.

We then compare and test the correlation of the residuals from equations (3) and (4)

between the three time periods based on Cramer’s (1987) squared correlation test.

13
 To test managing towards positive earnings, we run the following model

including two periods (IAS vs. IFRSvoluntary and IFRSvoluntary vs. IFRSmandatory) at a time to

examine if firms in one period are more likely to manage towards positive earnings

(Barth et al., 2006; Lang et al., 2006) than the other. The coefficient on the small positive

net income is our measure of managing towards positive earnings.

IAS (0,1)it = β 0 + β 1SPOit + β 2 LEVit + β 3Growth it + β 4 Eissueit + β 5 Dissueit +

β 6Turnit + β 7 Sizeit + β 8 CFOit + β 9 AUDit + β10 NUMEX it + β11 XLISTit + (5)
β12 FFit + εit.

and

IFRS (0,1)it = β 0 + β 1SPOit + β 2 LEVit + β 3Growth it + β 4 Eissueit + β 5 Dissueit +

β 6Turnit + β 7 Sizeit + β 8 CFOit + β 9 AUDit + β10 NUMEX it + β11 XLISTit + (6)
β12 FFit + εit.

IAS(0,1) in the first estimation is equal to 1 for the IAS period and zero for the

IFRSvoluntary period and in the second estimation, the IAS(0,1) is equal to 1 for the

IFRSvoluntary and zero for the IFRSmandatory period. SPO is a binary variable equal to 1 if

net income scaled by total assets is between 0 and 0.01 (Barth et al., 2006; Lang et al.,

2006; Lang et al., 2003). A positive coefficient on SPO in the estimation covering the

IAS (IFRSvoluntary) period and the IFRSvoluntary (IFRSmandatory) period indicates that firms in

the IAS (IFRSvoluntary) period manage earnings toward small positive amounts more

frequently than firms in the IFRSvoluntary(IFRSmandatory) period.

For the measure of timely recognition of losses, we also estimate an equation

similar to equation (5) and (6), but replacing SPO with LNEG.

IAS (0,1)it = β 0 + β 1LNEGit + β 2 LEVit + β 3Growth it + β 4 Eissueit + β 5 Dissueit +

β 6Turnit + β 7 Sizeit + β 8 CFOit + β 9 AUDit + β10 NUMEX it + β11 XLISTit + (7)
β12 FFit + εit.

14
and

IFRS (0,1)it = β 0 + β 1LNEGit + β 2 LEVit + β 3Growth it + β 4 Eissueit + β 5 Dissueit +

β 6Turnit + β 7 Sizeit + β 8 CFOit + β 9 AUDit + β10 NUMEX it + β11 XLISTit + (8)
β12 FFit + εit.

LNEG is a binary variable taking on the value of 1 for observations with annual earnings

scaled by total assets less than negative 0.2, and zero otherwise. The coefficient on

LNEG is our measure of timely loss recognition. A positive coefficient on LNEG

indicates that firms in the IAS (IFRSvoluntary) period recognize large losses more

frequently than those in IFRSvoluntary (IFRSmandatory).

In addition, we also use Basu’s (1997) reverse regressions of earnings on a

dummy variable for bad news (negative returns), annual return, and an interaction

variable of return and the dummy variable for bad news. We expect that more timely loss

recognition will result in a larger coefficient on bad news earnings. We compare the

magnitude of the interaction coefficient across the three periods to evaluate the timeliness

of bad news reflected in earnings. A larger coefficient indicates more timely loss

recognition.

The test of relative value relevance is based on a valuation framework provided

by Ohlson (1995) where a firm’s share price is a function of both earnings and book

value of equity.

Pit = β 0 + β 1 Eit + β 2 BVE + ε it (9),

where Pit is the market price per share three months after fiscal year end in year t of

companyi, and Eit, and BVEit are earnings before extraordinary items, and book value of

stockholders’ equity per share, respectively, and εit is the other value-relevant information

15
of company i in year t. The regression model’s R2 indicates the strength of the

association between the respective accounting variable(s) and stock price.

We also use a reverse regression with earnings as the dependent variable and

returns as the independent variable. We expect that if losses are recognized in a timely

manner, the association between returns and earnings is stronger, mirroring that more

information about earnings reaches the equity market on in the period the loss occurs and

not later (Basu 1997). Hence, we predict an increase in the association between earnings

and return over the three time periods under investigation.

IV. Sample Selection and Descriptive Statistics

The initial sample consists of all industrial German listed companies found in the

Datastream database in the years 2000 – 2002, 2003 – 2004 and 2005-2006, indicating

IAS/IFRS as their primary accounting standards.5 German companies often are traded on

more than one domestic stock exchange; we choose the common stock issue of the

highest market value as the company’s primary stock issue, and use the stock price of that

issue when a company has multiple issues of common stocks. For the value relevance

test, the sample selection process yields a German IAS sample of 187 firm-year

observations for 107 companies, a German IFRSvoluntary sample of 204 firm-year

observations and companies, and a German IFRSmandatory sample of 448 firm-year

observations and companies. Table 2 outlines our sample selection procedures for the

value relevance tests.

In addition, to mitigate the effect of different firms in each period on the

regression estimations, we match the sample firms in one period with the same firms in

5
We include all firms where the applied accounting standards are either international accounting standards
or IFRS.

16
the next period, and re-run all equations with the matched sample. This procedure yields

159 observations (92 companies) under the IAS period, 92 observations (92 companies)

for the IFRSvoluntary period, and 90 observations (90 companies) under the IFRSmandatory

period.

<Table 2 about here>

As shown in Table 3 below, the Electronic industry is the largest industry cluster

in the IAS and the IFRSvoluntary periods, and Financials is the largest in the IFRSmandatory

period. The Financials industry is the second largest industry during the IAS period, and

the Financials and Machinery and Equipment firms are both the second largest industries

during the IFRSvoluntary period and Financials during the IFRSmandatory period. One of the

most notable change in industry specialization is an increase in number of Financial firms

in the IFRSmandatory period. It seems that most of the Financials firms were not using the

IAS/IFRS until the mandatory adoption of IFRS in 2005. Another change is the steady

relative decrease of firms within the Recreation industry.

<Table 3 about here>

Table 4 presents the descriptive statistics of earnings smoothing, timely loss

recognition metrics, and value relevance, followed by the control variables.6 The earnings

per share (EPS) increases significantly across the periods, which is plausible considering

the economic upturn between 2003 and 2006. There is a significant increase in book

value of shareholders’ equity per share (BVPS) between the IFRSvoluntary period and the

IFRSmandatory period. The change in net income (∆NI) increases significantly from the IAS

period to the IFRSvoluntary periods. The ∆NI then decreases between the IFRSvoluntary period

and the IFRSmandatory period. There is no significant difference in change in cash flows
6
Variables in all our analyzes are winsorized at the 1st and 99th percentiles to control for outliers.

17
from operations (∆CFO) between the IAS and the IFRSvoluntary periods, while there is a

significant decrease between the the IFRSvoluntary and the IFRSmandatory periods. There is a

significant increase in accruals (ACC), measured as net income minus cash flow from

operations scaled by total assets, between the IFRSvoluntary and the IFRSmandatory periods,

which could possibly be related to a buildup of accruals caused by the abolishment of

amortization of acquired goodwill. There is no significant difference in the prevalence of

reporting of small positive earnings (SPOS) across the three periods. There is a

significant decrease in the reporting of large negative earnings (Lneg) across the IAS and

the IFRSvoluntary periods. This could be a sign of an increase in income smoothing

behavior; however, this development could also be driven by the economic situation

during the IAS period and the ongoing upturn during the two following period. The

descriptive statistics on the control variables suggests that there is a decrease in growth

between the IAS and the IFRSvoluntary periods but an increase in growth between the

IFRSvoluntary and the IFRSmandatory periods (measured as change in sales). The latter

increase could be related to the new companies adopting IFRS in 2005. There is no

statistically significant difference in change in common stock over the three periods.

There is no significant difference in changes in total liabilities (Dissue) between the IAS

and the IFRSvoluntary periods but there is significant increase between the IFRSvoluntary and

the IFRSmandatory periods. The significant increase in the IFRSmandatory period could be

driven by the fact that many more companies now must adopt IFRS and this result in the

inclusion of companies that are less capital market oriented and more reliant on debt in

the IFRSmandatory sample. However, it should be noted that there is no significant

difference in leverage (measured as total liabilities to total shareholders’ equity) across

18
the three periods. There is a statistically significant increase in the asset turnover rate

between the IAS and the IFRSvoluntary periods. Although, this is development is reversed

between the IFRSvoluntary and the IFRSmandatory periods. The latter is most likely a

consequence of the inclusion of a large number of new IFRS adopters in the sample

group. The size of the sample companies (measured as the natural log of total assets)

decreases significantly between the IAS period and the IFRSvoluntary period. This could be

driven by a lot of write downs during the IAS period due to the economic situation at that

time, which is corroborated by the fact that in this period there was a significantly higher

prevalence of reporting of large negative earnings. Finally, the cash flow from operations

increased significantly between the IAS period and the IFRSvoluntary period, this may also

be explained by the improved economic conditions in the following periods. However,

the cash flow from operations decreased significantly between the IFRSvoluntary and the

IFRSmandatory periods, which could be related to the inclusion of a large number of new

IFRS adopters in the sample group. Finally, there is a significant decrease in the free float

between the IAS period and the IFRSvoluntary period.

<Table 4 about here>

V. Results

Earnings Smoothing

As reported in Table 5, Panel A, the results of the tests of earnings smoothing are

contrary to our expectations in some instances. The variability in the change in net

income, ∆NI*, does increase significantly between the IAS and the IFRSvoluntary periods,

however, there is a significant decrease between the IFRSvoluntary and the IFRSmandatory

periods, suggesting an increase in income smoothing behavior. We also control for the

19
firm-specific volatility in cash flow from operations by using the ratio of income

variability and cash flow from operations variability. As predicted, the variability

increases significantly between the IAS and the IFRSvoluntary periods but decrease sharply

and end up below the IAS level in the IFRSmandatory period, however, in this test the

difference is statistically significant on the 5% level. Once again, this suggests an

increase in income smoothing behavior. The correlation between the residuals from the

regression on accruals (ACC) and cash flow from operations CFO shows an increase in

the magnitude of the negative correlation indicating a significant increase in earnings

management across the three periods. It should be noted that the correlation between

ACC and CFO is positive in the IAS period. We interpret this as a sign that companies

are growing and the growth is driving both an increase in CFO and a buildup of accruals.

This notion is also confirmed by the measure of growth that was significantly larger in

this period compared to the later two. Finally, there is no significant difference between

the small positive income variable (SPOS) across the three periods.

Timely Loss Recognition Tests

Table 5, Panel B shows a significant decrease in the reporting of large negative

earnings. As previously pointed out, the change in reporting of large negative earnings

may be the result of an improvement of the economic conditions since the IAS period.

However, most of our tests support the notion that the quality of accounting has

decreased among German companies reporting under IAS and IFRS over time.

Furthermore, the coefficient of the interaction variable of return and bad news shows a

significant increase between the IAS period and the IFRSvoluntary period, as predicted.

However, there is a significant decrease between the IFRSvoluntary period and the

20
IFRSmandatory period indicating less timely loss recognition. This also may have something

to do with the number of new adopters in the IFRSmandatory period.

Value Relevance Test

We measure value relevance in terms of the ability of accounting measures to

explain stock prices. As shown in Table 5, Panel C, the overall R2 of the regression

models for each of the time period examined was 0.44 in the IAS period, 0.09 in the

IFRSvoluntary period, and 0.19 in the IFRSmandatory period. Cramer’s (1987) test indicates a

significant difference in R2 between all three periods. Contrary to our expectations, the

R2s of both the IFRS periods are lower than the R2 of the IAS period, indicating a lower

usefulness of financial reporting under IFRS compared to IAS.

We also measure the R2 of a reverse regression where earnings is dependent

variable and returns as the independent variable (Basu 1997). As predicted, there seem to

be an increase in the association between earnings and returns between the IAS and the

IFRSvoluntary periods (from 0.19 to 0.28). However, this development is once again

reversed between the IFRSvoluntary and the IFRSmandatory periods (from 0.28 to 0.16).

However, only the increase between the the IAS and the IFRSvoluntary periods is

statistically significant. Turning to the bad news observations only, once again, we find

an expected increase in the association between the IAS and the IFRSvoluntary period.

However, contrary to expectations, this turns into a significant decrease from 0.35 to 0.11

between the IFRSvoluntary period and the IFRSmandatory period.

<Table 5 about here>

21
 In summary, our tests show consistent evidence that the quality of accounting

increased between the IAS period and the IFRSvoluntary but that this development reverses

between the IFRSvoluntary period and the IFRSmandatory period.

Sensitivity Analysis

Our findings in the main analysis may be a result of a structural change in the type

of companies that report under IFRS since it became mandatory for most public

companies in Germany. When IFRS was a choice, certain types of firms may have had

more incentives to opt to report under IFRS than others. Germany has a large and liquid

capital market, which means that large and well established companies have relatively

less incentive to switch to IAS/IFRS to attract foreign investors since they can raise

capital reporting under German GAAP. However, small information technology and less

well established companies may not have the same access to the German capital market,

and therefore, have reasons to switch to IAS/IFRS to be able to raise capital. Also, as

shown in Table 4, the distribution of companies in different industries changed between

the IFRSvoluntary periods when the Electronics industry dominated and the IFRSmandatory

period when the financial sector became dominating.

In order to examine whether a self-selection bias in our pre-2005 sample drives

the results we also rerun all tests using a sub-sample consisting of companies with firm-

year observations in both the IAS and the IFRS period. This sample has 159 observations

(92 companies) for the IAS period and 92 observations (92 companies) for the

IFRSvoluntary period and 90 observations (90 companies) for the IFRSmandatory period.

As shown in Table 6 Panel A, the results for tests of earnings smoothing and

timely loss recognition to some extent support the findings in the analysis using the

22
whole sample, although the reduction in accounting quality seems to be considerably

lesser. The variability in the change in net income, ∆NI*, increases significantly between

the IAS and the IFRSvoluntary periods but, as in the case with the whole sample, decreases

significantly compared to the IFRSmandatory period. When controlling for the volatility in

cash flow from operating activities, by using the ratio of net income variability and cash

flow from operations variability, we find the same pattern, a significant increase between

the IAS and the IFRSvoluntary periods followed by a significant between the IFRSvoluntary

and the IFRSmandatory periods. The correlation between accruals and cash flow from

operations shows a significant decrease between the IAS and the IFRSvoluntary periods.

The correlation is positive in both of these periods, although it becomes negative in the

IFRSmandatory , however, the change is not statistically significant. We find no significant

difference in the frequency of reporting small positive earnings across the two periods.

Finally, as shown in Table 6 Panel B, we find no significant change in reporting of large

negative earnings, indicative for less timely loss recognition. However, contrary to the

previous measure, the coefficient of the interaction variable of annual return and negative

return is significantly larger for the IFRSvoluntary period compared to the IAS period. There

is no significant difference between the IFRSvoluntary and the IFRSmandatory periods.

Furthermore, as shown in Table 6 Panel C, the value relevance measures are also

contradicting. The overall R2 of the regression models for time periods examined was

0.47 in the IAS period and 0.15 in the IFRSvoluntary period. Cramer’s (1987) test indicates

a significant difference in R2 between the two periods on the 1% level. This is partially

reversed between the IFRSvoluntary and the IFRSmandatory periods (from 0.15 to 0.38,

significant on the 5% level). The analysis of the earnings on returns regressions shows a

23
significant increase between IAS period and the IFRSvoluntary periods and a decrease

between the IFRSvoluntary and the IFRSmandatory periods (though this is not significant).

When we analyze the earnings on returns regressions using bad news observations only

we find an incremental increase in the value relevance across all three periods.

<Table 6 about here>

We conclude that our measures of earnings management, timely loss recognition,

and value relevance using a matched sample are neither providing evidence of an

increase nor a decrease in accounting quality between the IAS and the IFRS period.

At the time of the mandatory adoption of IFRS, there is a shift in the structure of

the sample, the largest industry group was Electronics up to the end of 2004 after the

mandatory adoption in 2005 the largest industry group is now the Financial. We,

therefore, rerun our tests excluding financial observations in order to investigate whether

the results of our main analysis is driven by companies in the financial sector. As shown

in Table 7, there is no qualitative difference in our results from these tests compared to

the main analysis using the full sample.

<Table 7 about here>

In order to analyze whether our results are driven by the new adopters in 2005, we

also split the sample used in the IFRSmandatory period into a mandatory and a voluntary

adopter group. Where the mandatory adopters are those who waited to adopt IFRS until it

was made mandatory in 2005 and those who adopted IFRS before are classified as

voluntary adopters. In addition, when doing this we also excluded all adopters of U.S.

GAAP who waited to adopt IFRS until it was made mandatory. The reason for this is that

we do not consider these companies as resisting the use of international accounting

24
standards, and therefore, they are different from those companies who resist switching

from German GAAP to IFRS. As shown in Table 8, we find that there is slightly less

variability in residuals of change of net income and the variability of the ratio of the

variability of change in residuals of net income and cash flow from operations; however,

the differences are not statistically significant. There seem to be more large negative

earnings reported by the voluntary adopters but once again, the difference is not

statistically significant. The only significant difference between the two groups is found

in the value relevance tests, the resisters are consistently showing significantly lower

value relevance in all such tests.

<Table 8 about here>

In order to establish that the any decrease in the quality of financial reporting is

driven by the excluded set of U.S, GAAP adopters; we also compared these observations

to the voluntary adopters. As expected, we did not find any evidence of this. On the

contrary, the U.S. GAAP adopters show a significantly higher variability in change in

residuals of net income (also when controlling for variability in change in residuals of

cash flow from operations), a positive correlation between accruals and cash flow, a

significantly lower frequency in reporting small positive incomes. However, the value

relevance among these companies was lower than that of the voluntary adopters.

<Table 9 about here>

VI. Conclusion

This study compares the characteristics of accounting amounts using a sample of

German companies reporting under IAS during 2000-2002, and IFRS during 2003-2004

and 2005-2006. Specifically, we investigate whether there is a change in accounting

25
quality during these three time periods as IASB revises existing IAS and issues new IFRS

to formulate a set of high quality international accounting standards for global financial

reporting purpose. Following prior research, we operationalize accounting quality with

earnings smoothing, timely loss recognition, and value relevance metrics. Contrary to

our expectations, our results suggest a decrease in accounting quality over the last years.

We find that earnings and book value of equity are becoming less value relevant during

the IFRSmandatory period compared to both the IAS and the IFRSvoluntary period. The

findings on earnings smoothing and timely loss recognition corroborate largely our

findings with respect to the value relevance of accounting information. Our results

consistently indicate that accounting quality has worsened over time. When using a

matched sample we cannot find any clear indication of either an improved or a worsened

quality of financial reporting. Further analysis of the 2005 to 2006 period provides some

weak indications that this might have been partly driven by new adopters of IFRS in

2005, however, only the difference in value relevance is statistically significant. We also

investigated whether our results were driven by a new dominating industry group,

Financials. We find nothing suggesting that this is the case. In sum, it seems that the

decrease in accounting quality is mainly driven by changes in accounting standards, not

the new adopters in 2005. The implication of this is that the last revisions of IASs and the

addition of new IFRSs have caused a decrease the quality of financial reporting in

Germany and future research needs to establish which standards drive this development.

26
 REFERENCES

Alexander, D. and E.K. Jermakowicz. 2006. “A True and Fair View of the
Principles/Rules Debate.” Abacus 42: 132-164.

Ashbaugh, H. 2001. ”Domestic Accounting Standards, International Accounting

Standards, and the Predictability of Earnings. Journal of Accounting Research 39:
417- 434

Barth, M.E., W.H. Beaver, and W.R. Landsman. 1998. “Relative Valuation Roles of
Equity Book Value and Net Income as a Function of Financial Health.” Journal of
Accounting and Economics 25: 1-34.

_____, W.R. Landsman, and M. Lang. 2005. “International Accounting Standards and
Accounting Quality.” Journal of Accounting Research 46: 467-498.

_____, W.R. Landsman, M. Lang, and C. Williams. 2006. “Accounting Quality:

International Accounting Standards and US GAAP.” Working paper.

Bartov, E. and M. Kim. 2005. “Comparative value relevance among German, U.S. and
international accounting standards: a German stock market perspective.” Journal
of Accounting, Auditing & Finance 20: 95-119.

Basu, S., 1997. “The Conservatism Principle and The Asymmetric Timeliness of
Earnings.” Journal of Accounting and Economics (24): 3-37.

Christensen, H.B., E. Lee, and M. Walker. 2007. “Incentives or Standards: What

Determine Accounting Quality Changes Around IFRS Adoption?” Working
Paper Manchester Business School.

Collins, D.W., E.L. Maydew, and I.S. Weiss. 1997. “Changes in the Value-Relevance of
Earnings and Book Values over the Past Forty Years.” Journal of Accounting and
Economics (24): 39-67.

Cramer, J.S. 1987. “Mean and Variance of R2 in Small and Moderate Samples.” Journal
of Econometrics. 35: 253-266.

Daske, H. 2006. “Economic Benefits of Adopting IFRS or US-GAAP – Have the

Expected Cost of Equity Capital Really Decreased?” Journal of Business Finance
& Accounting. 33: 329-373.

Easton, P. 1985. “Accounting Earnings and Security Valuation: Empirical Evidence of

Fundamental Links.” Journal of Accounting Research. 23 (Supplement): 54-77.

Hayn, C. 1995. “The information content of losses.” Journal of Accounting and

Economics 20: 125-153.

27
Hung, M., and K.R. Subramanyam. 2007. “Financial statement effects of adopting
international accounting standards: the case of Germany.” Review of Accounting
Studies. 12: 623-657.
Jermakowicz, E.K., J. Prather-Kinsey, and I. Wulf. 2007. “The Value Relevance of
Accounting Income Reported by DAX-30 German Companies.” Journal of
International Financial Management and Accounting. 18: 151-191.

Lang, M., J. Raedy, and M. Yetman. 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.

Lang, M; J. Raedy, and W. Wilson. 2006. “Earnings Management and Cross-Listing: Are
Reconciled Earnings Comparable to U.S. Earnings?” Journal of Accounting &
Economics. 42: 255-283.

Leuz, C., and R. E. Verrecchia. 2000. The Economic Consequences of Increased

Disclosure. Journal of Accounting Research. 38 (Supplement): 91-124.

Leuz, C., D. Nanda, and P. Wysocki. 2003. “Earnings Management and Investor
Protection: An International Comparison.” Journal of Financial Economics 69:
505-527.

Ohlson, J.A. 1995. Earnings, book values, and dividends in security valuation.
Contemporary Accounting Research 11: 661-687.

Pagano, M., A. Röell, and J. Zehner. 2002. ”The Geography of Equity Listings: Why do
Companies List Abroad?” Journal of Finance. 57: 2651-2694.

Platikanova, P. and C. Nobes. 2006. ”Was the Introduction of IFRS in Europe Value-
Relevant?” Working Paper University Pompeu Fabra and University of Reading.

Schipper, K. 2005. “The Introduction of International Accounting Standards in Europe:

Implications for International Convergence”. European Accounting Review
14:101-126.
Soderstrom, N. S. and K. J. Sun. 2007. “IFRS Adoption and Accounting Quality: A
Review.” European Accounting Review 16:675-702.
Tarca, A. 2004. “International Convergence of Accounting Practices: Choosing Between
IAS and US GAAP.” Journal of International Financial Management and
Accounting. 15: 60-91.

Whittington, G. 2005. “The Adoption of International Accounting Standards in the

European Union.” European Accounting Review 14:127-153.

28
Table 1: Summary of Revisions of IAS and IFRS over Time
Panel A: Revisions During the IAS period

Standard Issued/ In effect Focus Revision Made

revised
Year
IAS 12 2000 2001 Income The standard was amended to include guidance on
Taxes accounting for tax consequences of dividends and
other distributions made by the reporting parent
company.
IAS 19 2000 2001 Employee Revised to include and regulate more the
benefits retirement benefit costs only.
IAS 40 2000 2001 Investment Was an attempt to impose fair value measurement
property of investment property, ended up allowing it as an
alternative to historical cost accounting.
Panel B: Revisions During the IFRS voluntary period

Standard Issued/ In effect Focus Revision Made

revised
Year
IAS 10 2003 2004 Events after Regulates the reporting of events after the
the reporting reporting period.
period
IAS 17 2003 2004 Leases Initial direct and incremental costs by lessors in
negotiating leases must be recognized over the
lease term.
IAS 32 2003 2005* Presentation Additional guidance on measurement of the
of financial components of compound instruments on initial
instruments recognition and to group all guidance on financial
instruments in one standard
IAS 33 2003 2005* Earnings per Clarifying the standard and eliminate alternatives
share allowed by the standard.
IAS 41 2001 2003 Agriculture Regulates accounting for agricultural assets. These
types of assets were previously not covered by
other IASs.
IAS 36 2004 2004 Impairment Requires measurement of recoverable amount of
of assets intangible assets with an indefinite useful life on an
annual basis (including goodwill and intangible
assets not yet available for use).
IAS 38 2004 2004 Intangible The assumption that all assets’ useful life is finite
assets is abolished. In addition, intangible assets with
infinite useful life should not be amortized.
IFRS 1 2003 2004 First-time Sets out the procedures for first-time Adoption.
adoption of
IFRS
IFRS 3 2004 2004 Business Prohibits the use of the pooling of interests method
combinations for business combinations. Goodwill is initially
defined as the net fair value of acquired assets and
liabilities. Goodwill etc. with infinite lives are mot
amortized. Also see IAS 36 and IAS 38.
IAS 1 2003 2005* Presentation Mostly transferring the policies application related
of Financial to changes in accounting estimates and errors to
Statements IAS 8 while the presentation issues is transferred to
IAS 1.

29
Panel B: Revisions During the IFRS voluntary period cont.

Standard Issued/ In effect Focus Revision Made

revised
Year
IAS 2 2003 2005* Inventories The main change is the prohibition of LIFO as a
cost formula.
IAS 8 2003 2005* Accounting Please refer to revisions described for IAS 1.
policies,
changes in
accounting
estimates and
errors
IAS 16 2004 2005* Property, Costs of dismantlement, removal or restoration are
plant, and included in capitalized amounts.
equipment
Fair value revaluation only if this is reliably
measurable.

Depreciation must start when the asset is available

for use and continues regardless if the asset is idle
or not.
IAS 19 2004 2006* Employee Revised to permit recognition of actuarial gain and
benefits losses in equity and to require additional
disclosure.
IAS 21 2003 2005* Changes in Removal of a limited option to capitalize exchange
foreign rate differences resulting from severe devaluation
exchange or depreciation of a currency against which there is
rates no means of hedging.
IAS 24 2003 2005* Related party Requires disclosure of compensation to key
disclosures management employees and expands the definition
of “related party” by adding joint ventures, etc.
IAS 27 2003 2005* Consolidated Minority interests are now presented within the
and separate equity as a separate line item.
financial
statements
IAS 28 2003 2005* Investments Investors must not only consider the carrying
in associates amount the investment but also other long-term
interests in the associate when recognizing its share
of losses of the associate.
IAS 31 2003 2005* Interests in Investors must disclose the method used to
joint recognize its interest in jointly controlled entities
ventures (proportional consolidation or the equity method).
IAS 39 2004 2005* Financial Added fair value accounting for a hedge of the
instruments: interest rate exposure of a portfolio of financial
recognition assets and liabilities.
and
measurement
IAS 40 2004 2005* Investment Defining the concept “investment property” and
property impose a consistent use of the fair value or the cost
model.
IFRS 2 2004 2005* Share-based Require recognition of all share-based payment
payment transactions using a fair value measurement basis.

30
Panel B: Revisions During the IFRS voluntary period cont.

Standard Issued/ In effect Focus Revision Made

revised
Year
IFRS 4 2004 2005* Insurance Prohibits catastrophe and equalization reserves.
contracts Requires testing of the adequacy of recognized
insurance liabilities and impairment tests of
reinsurance assets.
IFRS 5 2004 2005*Non-current Prescribes accounting for assets held for sale and
assets held the presentation and disclosure of discontinued
for sale and operations.
discontinued
operations
Panel C: Revisions During the IFRS mandatory period

Standard Issued/ In effect Focus Revision Made

revised
Year
IAS 39 2005 2006 Financial Restricted the use of the “fair value option” to
instruments: eliminate accounting or economic mismatches.
recognition
and
measurement
IFRS 6 2005 2006 Exploration Regulates the financial reporting of mineral
for and resources until IASB has completed a
evaluation of comprehensive project of this.
mineral
resources
IFRS 7 2006 2007 Financial Require disclosure of information on the
instruments significance of financial instruments for a
disclosure company’s financial position and profitability.
IFRS 8 2006 2009* Operating Extends and change the scope of segment
Segments reporting.
*
Earlier application is encouraged.

31
Table 2: Sample Selection Process.
IAS IFRSVOLUNTARY IFRSMANDATORY
Firm- Firm- Firm-
Firms Years Firms Years Firms Years

From Datastream 290 570 327 327 571 571

Excluded observations
due to missing data -183 -383 -123 -123 -123 -123
Total sample 107 187 204 204 448 448

32
Table 3: Analysis of Industry Specialization

IAS IFRSVOLUNTARY IFRSMANDATORY

Aerospace 0 0% 0 0% 1 0%
Apparel 3 2% 4 2% 7 2%
Automotive 4 3% 4 2% 6 1%
Beverages 0 0% 1 0% 6 1%
Chemicals 4 4% 5 2% 12 3%
Construction 4 4% 9 4% 20 4%
Diversified 2 2% 2 1% 6 1%
Drugs, cosmetics and health care 5 5% 10 5% 21 5%
Electrical 3 3% 5 2% 11 2%
Electronics 14 13% 31 15% 55 12%
Financials 12 11% 22 11% 65 15%
Food 1 1% 1 0% 6 1%
Machinery and equipment 10 9% 23 11% 44 10%
Metal producers 1 1% 1 0% 1 0%
Metal product manufacturers 1 1% 3 1% 5 1%
Oil, gas, coal and related services 0 0% 2 1% 2 0%
Paper 0 0% 2 1% 5 1%
Printing and publishing 0 0% 2 1% 3 1%
Recreation 6 6% 10 5% 16 4%
Retailers 2 2% 7 3% 13 3%
Textiles 2 2% 2 1% 4 1%
Transportation 2 2% 2 1% 3 1%
Utilities 4 4% 6 3% 14 3%
Miscellaneous 27 25% 50 25% 122 27%
107 204 451

33
Table 4: Descriptive Statistics
IAS Firms N=187 IFRSV Firms N=204 IFRSM Firms N=448
Mean Median Standard Mean Median Standard Mean Median Standard
deviation deviation deviation
Test Variables
EPS1 -0.162 0.022 0.444 -0.034*** 0.052*** 0.300 0.019** 0.058** 0.226
2 **
BVPS 0.818 0.583 0.699 0.737 0.670 0.511 0.642 0.548*** 0.481
Return3 -36.298 -39.190 37.876 27.005*** 7.715*** 61.345 24.681 17.535*** 49.371
∆NI4 -0.054 -0.005 0.202 0.255*** 0.013*** 0.223 0.011 0.007*** 0.136
5 **
∆CFO 0.019 0.010 0.121 0.017 0.009 0.124 -0.006 -0.003*** 0.110
ACC6 -0.108 -0.051 0.197 -0.089 -0.060 0.190 -0.020*** -0.017*** 0.124
SPOS7 0.123 0.000 0.329 0.118 0.000 0.323 0.107 0.000 0.310
Lneg8 0.150 0.000 0.358 0.083** 0.000** 0.277 0.051 0.000 0.221
Control Variables
LEV9 0.579 0.614 0.252 0.578 0.608 0.263 0.557 0.575 0.246
Growth10 0.294 0.056 0.714 0.092*** 0.047* 0.350 0.184** 0.090*** 0.508
Eissue11 0.053 0.000 0.227 0.082 0.000 0.266 0.119 0.000 0.420
Dissue12 0.421 0.037 1.682 0.275 0.012 1.503 0.333 0.059*** 1.585
Turn13 1.093 0.929 0.829 1.191 1.118** 0.778 1.061 1.012** 0.716
Size14 13.354 12.384 2.721 12.467*** 11.758*** 2.634 12.479 12.053 2.358
CFO15 0.044 0.057 0.117 0.074** 0.075*** 0.123 0.046 0.055*** 0.104
NUMEX16 2.337 2.000 1.668 2.098 2.000 1.365 1.920 2.000* 1.211
AUD17 0.658 1.000 0.476 0.598 1.000 0.492 0.596 1.000 0.491
XLIST18 0.016 0.000 0.126 0.010 0.000 0.098 0.009 0.000 0.094
FF19 91.584 15.846 165,240.000 60.201** 10.541** 134,479.500 46.717 11.719 111,953.000
1
EPS is earnings per share at year end of the fiscal year deflated by the share price 6 months after the preceding fiscal year end.
2
BVPS is the book value of shareholders’ equity per share at the end of the fiscal year deflated by the share price 6 months after the preceding fiscal year end.
3
Return is the annual return of company i at time t.
4
∆NI is the change in annual earnings scaled by total assets.
5
∆CF is the change in cash flow from operating activities scaled by total assets.
6
ACC is earnings less cash flow from operating activities scaled by total assets.
7
SPOS is a dummy variable taking on the value 1 for observations for which the annual earnings scaled by total assets is between 0 and 0.01, and 0 otherwise.
8
Lneg is a dummy variable taking on the value 1 for observations for which the annual earnings scaled by total assets is less than -0.20, and 0 otherwise.
9
LEV is total liabilities divided by shareholders’ equity.
10
Growth is the percentage change in sales.
11
Eissue is the percentage change in common shareholders’ stock.

34
12
Dissue is the percentage change in total liabilities.
13
Turn is sales divided by total assets.
14
Size is the natural log of total assets.
15
CFO is the cash flow from operating activities scaled by total assets.
16
NUMEX is the number of stock exchange listings.
17
AUD is an indicator taking on the value of 1 if the firm’s auditor is PwC, KPMG, Arthur Andersen, E&Y, or D&T, and 0 otherwise.
18
XLIST is an indicator taking on the value of 1 if the firm is listed on any U.S. stock exchange and 0 otherwise.
19
FF is the free float measured as the average number of shares traded the last day of the month during the fiscal year divided by number of common shares
outstanding at the fiscal year end, divided by 1,000.
Asterisks indicate that there is significantly different from the previous time period using a two-tailed t-test: * p < 0.1, ** p < 0.05, and *** p < 0.01.

35
Table 5: Analysis of Accounting Quality
Panel A: Earnings IAS IFRSV IFRSM
Management N=187 N=204 N=448

Variability of ∆NI*1 0.038 0.048*** 0.018***

Variability of ∆NI* over
∆CFO*2,3 3.950 4.697** 1.840**
Correlation of ACC* and CFO*4 0.083 -0.048** -0.049
Small positive NI5 -0.273 0.125
Panel B: Timely Loss
Recognition

Large negative NI6 0.715### 0.565##

Basu Regression of Return *
Dum Coefficient7 0.003### 0.008***### 0.004***###
Panel C: Association of Stock
Prices and Returns with
Accounting Data4

Price8 0.442 0.088*** 0.191**

Return regression9:
Pooling good news and bad news
observations 0.194 0.283* 0.155
Basu Good News 0.026 0.022 0.013
Basu Bad News 0.129 0.345** 0.109**
*, **, ***
Significantly different between each category at the 0.01, 0.05, and the 0.10 respectively (two-tailed)
# ## ###
, , Significantly different from zero at the 0.01, 0.05, and the 0.10 respectively (two-tailed)
1
∆NI* is the variance of residuals from a regression of the ∆NI on the control variables.
2
∆CFO* is the variance of residuals from a regression of the ∆NI on the control variables.
3
Variability of ∆NI* over ∆CFO* is the ratio of ∆NI* divided by ∆CFO*.
4
Correlation of ACC* and CFO* is the partial Spearman correlation between the residuals from the ACC
and CFO regression.
5
SPOS is a dummy variable taking on the value 1 for observations for which the annual earnings scaled by
total assets is between 0 and 0.01, and 0 otherwise.
6
Lneg is a dummy variable taking on the value 1 for observations for which the annual earnings scaled by
total assets is less than -0.20, and 0 otherwise.
7
The regression is EPS = β 0 + β1 R + β 2 DUM + β 3 R * DUM + ε where EPS is annual earnings per share
deflated by share price at the beginning of the period, R is annual return, and DUM takes on the value 1 if
the return is negative and 0 otherwise.
8
The regression is P = β 0 + β 1 EPS + β 2 BVPS + ε where P is price as of three months after the fiscal year –
end, EPS is earnings per share and BVPS is the book value of shareholders’ equity per share. All variables
are scaled by share price six months after the preceding year-end.
9
The Basu good and bad news regression is EPS = β 0 + β1 R + ε , where EPS is earnings per share deflated
by price at the beginning of the year and R is the annual return. Good news observations are those for
which R is positive and bad news are those for which return is negative. We winsorize all continuous
variables at the 1% level to control for outliers.

36
Table 6: Analysis of Accounting Quality Using a Sub-Sample of voluntary IFRS adopters.
Panel A: Earnings IAS IFRSV IFRSM
Management N=159 N=92 N=90

Variability of ∆NI*1 0.037 0.053** 0.014***

Variability of ∆NI* over
∆CFO*2,3 4.125 8.152*** 1.924*
Correlation of ACC* and CFO*4 0.292 0.063*** -0.120
Small positive NI5 -0.075 0.120
Panel B: Timely Loss
Recognition

Large negative NI6 0.259 0.633

Basu Regression of Return *
Dum Coefficient7 0.003## 0.009###** 0.008###
Panel C: Association of Stock
Prices and Returns with
Accounting Data4

Price8 0.473 0.149*** 0.376**

Return regression9:
Pooling good news and bad news
observations 0.187 0.391** 0.252
Basu Good News -0.030 0.157** 0.013
Basu Bad News 0.124 0.347* 0.452
*, **, ***
Significantly different between each category at the 0.01, 0.05, and the 0.10 respectively (two-tailed)
# ## ###
, , Significantly different from zero at the 0.01, 0.05, and the 0.10 respectively (two-tailed)
1
∆NI* is the variance of residuals from a regression of the ∆NI on the control variables.
2
∆CFO* is the variance of residuals from a regression of the ∆NI on the control variables.
3
Variability of ∆NI* over ∆CFO* is the ratio of ∆NI* divided by ∆CFO*.
4
Correlation of ACC* and CFO* is the partial Spearman correlation between the residuals from the ACC
and CFO regression.
5
SPOS is a dummy variable taking on the value 1 for observations for which the annual earnings scaled by
total assets is between 0 and 0.01, and 0 otherwise.
6
Lneg is a dummy variable taking on the value 1 for observations for which the annual earnings scaled by
total assets is less than -0.20, and 0 otherwise.
7
The regression is EPS = β 0 + β1 R + β 2 DUM + β 3 R * DUM + ε where EPS is annual earnings per share
deflated by share price at the beginning of the period, R is annual return, and DUM takes on the value 1 if
the return is negative and 0 otherwise.
8
The regression is P = β 0 + β 1 EPS + β 2 BVPS + ε where P is price as of three months after the fiscal year –
end, EPS is earnings per share and BVPS is the book value of shareholders’ equity per share. All variables
are scaled by share price six months after the preceding year-end.
9
The Basu good and bad news regression is EPS = β 0 + β1 R + ε , where EPS is earnings per share deflated
by price at the beginning of the year and R is the annual return. Good news observations are those for
which R is positive and bad news are those for which return is negative. We winsorize all continuous
variables at the 1% level to control for outliers.

37
Table 7: Analysis of Accounting Quality Excluding Firm-Year Observations from the Financial Industry

Panel A: Earnings IAS IFRSV IFRSM

Management N=196 N=194 N=386

Variability of ∆NI*1 0.042 0.053 0.018***

Variability of ∆NI* over
∆CFO*2,3 4.074 4.989* 1.814***
Correlation of ACC* and CFO*4 0.049 0.002** -0.031***
Small positive NI5 -0.080 0.089
Timely Loss Recognition

Large negative NI6 0.738### 0.620###

Basu Regression of Return *
Dum Coefficient3 0.003## 0.009###*** 0.001#***
Association of Stock Prices and
Returns with Accounting Data4

Price 0.435 0.084*** 0.196**

Return regression:
Pooling good news and bad news
observations 0.186 0.124 0.096
Basu Good News 0.020 0.006 0.012
Basu Bad News 0.128 0.343** 0.133**
*, **, ***
Significantly different between each category at the 0.01, 0.05, and the 0.10 respectively (two-tailed)
# ## ###
, , Significantly different from zero at the 0.01, 0.05, and the 0.10 respectively (two-tailed)
1
∆NI* is the variance of residuals from a regression of the ∆NI on the control variables.
2
∆CFO* is the variance of residuals from a regression of the ∆NI on the control variables.
3
Variability of ∆NI* over ∆CFO* is the ratio of ∆NI* divided by ∆CFO*.
4
Correlation of ACC* and CFO* is the partial Spearman correlation between the residuals from the ACC
and CFO regression.
5
SPOS is a dummy variable taking on the value 1 for observations for which the annual earnings scaled by
total assets is between 0 and 0.01, and 0 otherwise.
6
Lneg is a dummy variable taking on the value 1 for observations for which the annual earnings scaled by
total assets is less than -0.20, and 0 otherwise.
7
The regression is EPS = β 0 + β1 R + β 2 DUM + β 3 R * DUM + ε where EPS is annual earnings per share
deflated by share price at the beginning of the period, R is annual return, and DUM takes on the value 1 if
the return is negative and 0 otherwise.
8
The regression is P = β 0 + β 1 EPS + β 2 BVPS + ε where P is price as of three months after the fiscal year –
end, EPS is earnings per share and BVPS is the book value of shareholders’ equity per share. All variables
are scaled by share price six months after the preceding year-end.
9
The Basu good and bad news regression is EPS = β 0 + β1 R + ε , where EPS is earnings per share deflated
by price at the beginning of the year and R is the annual return. Good news observations are those for
which R is positive and bad news are those for which return is negative. We winsorize all continuous
variables at the 1% level to control for outliers.

38
Table 8: Analysis of Accounting Quality Using Observations from the 2005-2006 Period, Comparing
Voluntary Adopters of IFRS and Mandatory Adopters of IFRS.
Panel A: Earnings IFRSV IFRSM
Management N=224 N=137

Variability of ∆NI*1 0.014 0.013

Variability of ∆NI* over
∆CFO*2,3 1.626 1.619
Correlation of ACC* and CFO*4 -0.245 0.049
Small positive NI5 -0.307
Timely Loss Recognition

Large negative NI6 0.432

Basu Regression of Return *
Dum Coefficient3 0.005### -0.001
Association of Stock Prices and
Returns with Accounting Data4

Price 0.402 0.007***

Return regression:
Pooling good news and bad news 0.058 0.100
observations
Basu Good News -0.005 0.100*
Basu Bad News 0.361 -0.029***
*, **, ***
Significantly different between each category at the 0.01, 0.05, and the 0.10 respectively (two-tailed)
# ## ###
, , Significantly different from zero at the 0.01, 0.05, and the 0.10 respectively (two-tailed)
1
∆NI* is the variance of residuals from a regression of the ∆NI on the control variables.
2
∆CFO* is the variance of residuals from a regression of the ∆NI on the control variables.
3
Variability of ∆NI* over ∆CFO* is the ratio of ∆NI* divided by ∆CFO*.
4
Correlation of ACC* and CFO* is the partial Spearman correlation between the residuals from the ACC
and CFO regression.
5
SPOS is a dummy variable taking on the value 1 for observations for which the annual earnings scaled by
total assets is between 0 and 0.01, and 0 otherwise.
6
Lneg is a dummy variable taking on the value 1 for observations for which the annual earnings scaled by
total assets is less than -0.20, and 0 otherwise.
7
The regression is EPS = β 0 + β1 R + β 2 DUM + β 3 R * DUM + ε where EPS is annual earnings per share
deflated by share price at the beginning of the period, R is annual return, and DUM takes on the value 1 if
the return is negative and 0 otherwise.
8
The regression is P = β 0 + β 1 EPS + β 2 BVPS + ε where P is price as of three months after the fiscal year –
end, EPS is earnings per share and BVPS is the book value of shareholders’ equity per share. All variables
are scaled by share price six months after the preceding year-end.
9
The Basu good and bad news regression is EPS = β 0 + β1 R + ε , where EPS is earnings per share deflated
by price at the beginning of the year and R is the annual return. Good news observations are those for
which R is positive and bad news are those for which return is negative. We winsorize all continuous
variables at the 1% level to control for outliers.

39
Table 9: Analysis of Accounting Quality Using Observations from the 2005-2006 Period, Comparing
Voluntary Adopters of IFRS and Adopters of IFRS Who Previously Reported Under US GAAP.
Panel A: Earnings IFRSV Pre-US
Management N=224 GAAP
N=87

Variability of ∆NI*1 0.014 0.035***

Variability of ∆NI* over
∆CFO*2,3 1.626 2.318***
Correlation of ACC* and CFO*4 -0.245 0.143
Small positive NI5 -1.782#
Timely Loss Recognition

Large negative NI6 0.265

Basu Regression of Return *
Dum Coefficient3 0.005### 0.006##***
Association of Stock Prices and
Returns with Accounting Data4

Price 0.402 0.230**

Return regression:
Pooling good news and bad news 0.058 0.139
observations
Basu Good News -0.005 -0.013
Basu Bad News 0.361 0.064**
*, **, ***
Significantly different between each category at the 0.01, 0.05, and the 0.10 respectively (two-tailed)
# ## ###
, , Significantly different from zero at the 0.01, 0.05, and the 0.10 respectively (two-tailed)
1
∆NI* is the variance of residuals from a regression of the ∆NI on the control variables.
2
∆CFO* is the variance of residuals from a regression of the ∆NI on the control variables.
3
Variability of ∆NI* over ∆CFO* is the ratio of ∆NI* divided by ∆CFO*.
4
Correlation of ACC* and CFO* is the partial Spearman correlation between the residuals from the ACC
and CFO regression.
5
SPOS is a dummy variable taking on the value 1 for observations for which the annual earnings scaled by
total assets is between 0 and 0.01, and 0 otherwise.
6
Lneg is a dummy variable taking on the value 1 for observations for which the annual earnings scaled by
total assets is less than -0.20, and 0 otherwise.
7
The regression is EPS = β 0 + β1 R + β 2 DUM + β 3 R * DUM + ε where EPS is annual earnings per share
deflated by share price at the beginning of the period, R is annual return, and DUM takes on the value 1 if
the return is negative and 0 otherwise.
8
The regression is P = β 0 + β 1 EPS + β 2 BVPS + ε where P is price as of three months after the fiscal year –
end, EPS is earnings per share and BVPS is the book value of shareholders’ equity per share. All variables
are scaled by share price six months after the preceding year-end.
9
The Basu good and bad news regression is EPS = β 0 + β1 R + ε , where EPS is earnings per share deflated
by price at the beginning of the year and R is the annual return. Good news observations are those for
which R is positive and bad news are those for which return is negative. We winsorize all continuous
variables at the 1% level to control for outliers.

40