Vous êtes sur la page 1sur 42

Relative Effects of IFRS Adoption and IFRS Convergence

on Financial Statement Comparability*

STEVE LIN, Florida International University

WILLIAM N. RICCARDI, University at Albany – SUNY

CHANGJIANG WANG, University of Cincinnati†

ABSTRACT
One of the primary objectives of both adoption of IFRS and convergence between IFRS and U.S.
GAAP is to increase financial statement comparability. Using a unique setting in Germany, we
compare the effectiveness of these two approaches in achieving this desired outcome. Our empiri-
cal tests show that both adoption and convergence lead to an increase in comparability after the
new enforcement regulation in 2005. However, difference-in-differences tests show that adoption
does not lead to a significant incremental increase in comparability beyond convergence. The find-
ings of this study should be of interest to regulators and standard setters as they assess alternative
methods of aligning domestic standards with IFRS.

Incidence comparée de l’adoption des IFRS et de la convergence


vers les IFRS sur la comparabilité des états financiers
RÉSUMÉ
L’un des principaux objectifs tant de l’adoption des Normes internationales d’information finan-
cière (IFRS) que de la convergence des principes comptables généralement reconnus des États-
Unis (PCGR des É.-U.) et des IFRS est d’accroître la comparabilité des états financiers. Dans
un cadre exclusif à l’Allemagne, les auteurs comparent l’efficacité de ces deux approches dans
l’obtention du résultat visé. Les tests empiriques auxquels ils procèdent montrent que l’adoption et
la convergence entraînent toutes deux une augmentation de la comparabilité après l’instauration
des nouvelles règles d’application en 2005. Toutefois, des tests d’écart dans les différences
révèlent que l’adoption n’entraîne pas d’augmentation marginale importante de la comparabilité
au-delà de la convergence. Les résultats de l’étude sont susceptibles d’intéresser les autorités de
réglementation et les normalisateurs appelés à évaluer les différentes méthodes possibles d’aligne-
ment des normes nationales sur les IFRS.

* Accepted by Ole-Kristian Hope. An earlier version of this paper was presented at the 2016 Contemporary Accounting
Research Conference, generously supported by the Chartered Professional Accountants of Canada. We are grateful for the
valuable comments from two anonymous reviewers, Ole-Kristian Hope, Patrick Hopkins (discussant), Gary Kabureck (dis-
cussant), Ervin Black, Agnes Cheng, Jennifer Ho, Andrew Leone, Richard Morris, Grace Pownall, Katherine Schipper,
Alan Teixeira (Deloitte, London), Wayne Thomas, Ram Venkataraman, Shiheng Wang, and workshop and conference par-
ticipants at the 2016 CAR conference, University of Cincinnati, Florida International University, University of Texas at
Arlington, University of Oklahoma, Hong Kong Polytechnic University, National Cheng Kung University, and National
Cheng Chi University. Earlier versions of this paper circulated under the title “Relative benefits of adoption of IFRS and
convergence between IFRS and U.S. GAAP: Evidence from Germany.”
† Corresponding author.

Contemporary Accounting Research Vol. 36 No. 2 (Summer 2019) pp. 588–628 © CAAA
doi:10.1111/1911-3846.12475
Relative Effects of IFRS Adoption and IFRS Convergence 589

1. Introduction
Using a unique setting in Germany as a natural experiment, this study examines the respective
and relative effects of mandatory adoption of IFRS from U.S. GAAP and convergence between
U.S. GAAP and IFRS on financial statement comparability (hereafter, “comparability”). This
study is motivated by the two approaches used by different countries as they aligned their domes-
tic accounting standards with IFRS over the last decade. Countries such as the member states of
the European Union (EU) directly switched to IFRS from their domestic accounting standards in
2005 (hereafter, “adoption”), while other countries, such as the United States, Japan, and China,
chose to gradually converge their national accounting standards with IFRS (hereafter, “conver-
gence”).1 So far, there has been no systematic evaluation and analysis about the relative costs and
benefits of adoption and convergence.2
Both adoption and convergence share the same goal of developing a single set of high-quality
global accounting standards. One important aspect of the intended benefits of both approaches is to
increase comparability of financial statements prepared by firms located in different countries.
Increased comparability has been stated as a desired objective of a shift toward globalized account-
ing standards. The FASB and IASB (hereafter, “the Boards”) both argue in their respective concep-
tual frameworks that comparability, one of the qualitative characteristics to “enhance the usefulness
of information that is relevant and faithfully represented,” is a desired outcome of adopting a uni-
form set of accounting standards (FASB 2010, 19; IASB 2010, 35–36). Thus, comparability should
improve financial reporting quality by allowing financial statement users “to identify and understand
similarities in, and differences among, items” (FASB 2010, 19; IASB 2010, 36).
Intuitively, one might expect that adoption of IFRS would eliminate any accounting differences
and significantly improve the comparability between firms that previously followed different account-
ing standards. However, there are several reasons why IFRS adoption may have a limited effect on
comparability. First, since IFRS are more principles-based, financial statements reported under IFRS
could be adversely affected by managerial discretion (Schipper 2003). Second, since the IASB does
not have authority to impose compliance with IFRS in individual countries, there are significant inter-
national variations related to IFRS implementation and practices (Nobes 2006; Henry et al. 2009;
Nobes 2011; Cascino and Gassen 2015). Finally, uniform accounting standards alone may not neces-
sarily result in comparable accounting outcomes due to differences in institutional and economic fac-
tors across countries (Ball 2006; Holthausen 2009; Cascino and Gassen 2015).
The convergence projects between the Boards aim to make their existing financial reporting
standards fully compatible as soon as is practicable and to coordinate their future work programs
to ensure that, once achieved, comparability is maintained (Norwalk Agreement, FASB and IASB
2002). However, many criticized that the process of convergence was slow and ineffective. While
certain projects were completed with various levels of success, there were also projects that were
discontinued or resulted in diverged standards. A summary of revisions to accounting standards
over the sample period is reported online in Appendix A.3 Using data during 2004–2006, Henry

1. The approach taken by individual countries may differ with respect to convergence. For example, convergence as
referenced in this study—between U.S. GAAP and IFRS—has resulted in changes to both sets of standards.
Conversely, China has chosen to adopt certain IFRS so far but has promised to fully converge Chinese GAAP to
IFRS in the near future. There has been convergence between IFRS and Japanese GAAP since 2005. Most Japanese
companies were permitted to use IFRS beginning in 2013.
2. The United States is the only country that has considered both adoption (Roadmap, SEC 2008) and convergence
(Norwalk Agreement, FASB and IASB 2002) as potential strategies to align U.S. GAAP with IFRS. No U.S. firms
have adopted IFRS thus far.
3. Please see supporting information: “Appendix A: Summary of revisions to accounting standards over sample
period” as an addition to the online article. As some examples, the share-based payment (IFRS 2 and SFAS 123)
and business combination (IFRS 3 and SFAS 141R) projects were successfully converged, while the income tax,
intangible assets, and asset impairment projects were discontinued. However, the lease accounting project encoun-
tered many obstacles that led to each Board developing its own model.

CAR Vol. 36 No. 2 (Summer 2019)


590 Contemporary Accounting Research

et al. (2009) find that early convergence projects reduced the accounting differences between
IFRS and U.S. GAAP, but significant differences remained.
Adoption and convergence share the same goal but are different in some aspects. Whereas
adoption of IFRS immediately eliminates any accounting differences, convergence only focuses
on certain major differences; some significant accounting differences have never been addressed
in the convergence projects and will continue to exist (e.g., use of LIFO for inventory valuation).
Therefore, convergence could have a less significant impact on accounting differences. Further,
convergence may result in changes to either or both sets of accounting standards, whereas adop-
tion replaces U.S. GAAP with IFRS.
Previous studies have examined the extent to which adoption affects comparability across
countries (Yip and Young 2012; Cascino and Gassen 2015).4 However, little is known about the
effect of convergence on comparability and whether the effects of adoption and convergence on
comparability might differ. An analysis on the relative benefits of adoption and convergence
would help regulators and standard setters as they assess alternative approaches of aligning
domestic standards with IFRS.
The unique setting of the German market allows us to analyze these issues.5 Prior to the
adoption of IFRS by the EU, the German government permitted firms to prepare their financial
statements in accordance with German GAAP, U.S. GAAP, or IFRS. After the mandatory adop-
tion of IFRS by the EU in 2005, all publicly listed German firms began to report under IFRS.6
Therefore, the German market provides a natural setting where there are firms that switched to
IFRS from U.S. GAAP (“German U.S. GAAP firms”) and other firms that applied IFRS continu-
ously (“German IFRS firms”) throughout our sample period. In addition, German regulators con-
currently introduced a new enforcement regulation not only to support the implementation of
IFRS but also to ensure full compliance, which renders the adoption of IFRS in Germany more
credible (Daske et al. 2013; Christensen et al. 2013).
Following prior research, this study employs various pair-based measures to operationalize
comparability. The changes in comparability for matched German U.S. GAAP–German IFRS
firm pairs are attributable to the combined effects of adoption, convergence, and enforcement,
while the changes in comparability for German IFRS–U.S. GAAP firm pairs are the result of con-
vergence and the same enforcement regulation. By employing a difference-in-differences regres-
sion model, we are able to examine whether adoption leads to a significant incremental increase
in comparability beyond convergence and enforcement. Our findings suggest that both adoption
and convergence led to an increase in comparability. These findings are robust after controlling
for the enforcement effect except for the changes in comparability due to convergence. Our
difference-in-differences regression results suggest that adoption does not lead to a significant
incremental increase in comparability beyond convergence and enforcement.
To complement the above empirical findings, we have performed three additional tests that
involve hand-collected financial statement data. First, based on the U.S. GAAP-to-IFRS reconcilia-
tion disclosure from a sample of German U.S. GAAP firms in 2005, we find that a switch to IFRS
from U.S. GAAP in Germany reduces approximately 5 and 3 percent of the accounting differences
in net income and shareholders’ equity, respectively, indicating that the effect of adoption on com-
parability is economically material. Second, we follow Cascino and Gassen (2015) to investigate
the extent to which the German U.S. GAAP firms included in our sample are in compliance with

4. These studies examine the impact of a switch to IFRS from domestic accounting standards on comparability,
whereas this study investigates the respective and relative effects of convergence between U.S. GAAP and IFRS
and mandatory adoption of IFRS from U.S. GAAP on comparability.
5. We are unable to use other countries’ data to investigate these questions due to a lack of the required institutional
setting and insufficient number of observations.
6. German firms were permitted to delay adoption of IFRS until no later than 2007 if they cross-listed outside of
Germany. After we impose the sample selection criteria, no remaining firms are cross-listed on U.S. exchanges.

CAR Vol. 36 No. 2 (Summer 2019)


Relative Effects of IFRS Adoption and IFRS Convergence 591

IFRS in the post-adoption period. Specifically, we investigate the average levels of both measure-
ment and disclosure compliance as required by IFRS in the areas that have been identified as key
accounting differences. We find that the average measurement and disclosure compliance rates are
high except for two disclosure items related to pension and post-employment benefits. Similarly,
we validate the convergence effect by comparing the average measurement and disclosure compli-
ance rates for those completed convergence projects up to 2010 between a sample of German IFRS
firms and matched U.S. firms. We find that the average compliance rates are high and generally
comparable between the two samples, except for a few items in segment reporting and business
combinations. Lack of compliance, therefore, is an unlikely alternative explanation for our findings.
This study offers several contributions to the literature. First, previous studies find that a
switch to IFRS from non-U.S. domestic accounting standards increases comparability (Barth
et al. 2012; Yip and Young 2012; Cascino and Gassen 2015). Our results show that comparability
also increases when firms switched to IFRS from U.S. GAAP, despite skepticism that such a shift
in the reporting regime may not necessarily lead to more comparable reporting. Second, this study
is the first to investigate the relative effects of adoption and convergence on comparability using
a unique setting in Germany. Our results suggest that both adoption and convergence improve
comparability after the new enforcement regulation, but adoption may not necessarily be more
beneficial than convergence in achieving greater comparability.7 We believe that our findings pro-
vide timely evidence as countries and accounting regulators consider alternative venues to align
their domestic accounting standards with IFRS.
Our study is not without caveats. First, the sample size is relatively small because there are a
limited number of German firms that reported under U.S. GAAP before adopting IFRS. Second,
Germany’s new enforcement regulation has an asymmetric impact on the two series of firm pairs
(both firms in the German U.S. GAAP–German IFRS pairs are influenced by the enforcement
effect, while only German firms in the German IFRS–U.S. GAAP pairs are affected), which can-
not be completely controlled by the difference-in-differences test. It is also noteworthy that,
although the setting in Germany may circuitously reflect a potential outcome if the United States
decides to switch to IFRS, the results of this study may not have direct implications for the
United States for the following reasons. This study compares the effect of a full adoption of IFRS
with partial (i.e., ongoing) convergence between U.S. GAAP and IFRS. The outcome could be
different if U.S. regulators continued toward achieving full convergence, such as the strategy of
convergence to IFRS once used by Canada. Additionally, the reporting and regulatory environ-
ment in Germany is substantially different from that in the United States. Finally, the decision
between adoption and convergence is influenced by a variety of factors, including control over
the standard-setting process, compliance costs, and political issues.

2. Institutional background
Adoption and convergence and their impacts on comparability
Of the 140 jurisdictions whose profiles have been posted by IASB, 116 jurisdictions (83 percent)
require IFRS for all or most domestic public companies.8 These jurisdictions, including the EU
member states, directly switched to IFRS from their domestic accounting standards over the last
decade. However, some of the remaining 24 jurisdictions, including the United States, Japan, and
China, have been converging their domestic accounting standards with IFRS.9 The goal of both

7. This is an especially important consideration given that there remains a possibility that financial reporting quality
could decline following a switch from U.S. GAAP to IFRS (Lin et al. 2012; Ahmed et al. 2013).
8. Please refer to http://www.ifrs.org/Use-around-the-world/Documents/Financial-Reporting-Standards-World-Economy-
June-2015.pdf.
9. These jurisdictions include Bermuda, Bolivia, Cayman Islands, China, Egypt, Guatemala, Guinea-Bissau, Honduras,
India, Indonesia, Japan, Macao, Madagascar, Nicaragua, Niger, Panama, Paraguay, Saudi Arabia, Suriname, Switzerland,
Thailand, the United States, Uzbekistan, and Vietnam.

CAR Vol. 36 No. 2 (Summer 2019)


592 Contemporary Accounting Research

adoption and convergence is to develop a single set of high-quality global accounting standards
that could be used for both domestic and cross-border financial reporting as a means to increase
comparability. The FASB (2014) states that working toward more comparable global accounting
standards is consistent with its core mission, arguing that “investors, companies, auditors, and
other participants in the U.S. financial reporting system will benefit from the increased compara-
bility that can result from closer alignment of standards used internationally.” Further, the FASB
believes that “more comparable standards have the potential to reduce costs for both users and
preparers of financial statements and make worldwide capital markets more efficient.”
In an attempt to achieve the goal of improved comparability, the U.S. SEC and FASB have
worked together with the IASB over the last decade to converge U.S. GAAP and IFRS through a
series of projects to remove the major differences between the two sets of standards.
A Roadmap (SEC 2008) was issued to provide the timeline for the potential adoption of
IFRS in the United States if the proposed seven milestones were achieved.10 However, no deci-
sion has been reached so far regarding whether and/or when U.S. firms may be permitted or
required to prepare their financial statements in accordance with IFRS (SEC 2012). The delay in
reaching a decision has been due, in part, to inconsistent views from the SEC on the potential
benefits of IFRS adoption and resistance from the stakeholders in the United States.11 Academics
have expressed their opinions on both sides of this debate. Bradshaw et al. (2010) propose that
continued convergence between the Boards is preferable to adoption of IFRS by U.S. firms in the
near future due to the expected costs of a complete change in the reporting system. In contrast,
Jamal et al. (2010) argue that it is unlikely to achieve comparability and consistency of financial
reporting on a global basis through convergence and instead propose that U.S. firms be allowed
to choose between U.S. GAAP and IFRS.
To the best of our knowledge, no study has examined the relative effects of adoption and
convergence on comparability. This is a timely and important question given the importance of
comparability to global capital markets and dichotomous approaches used by countries to align
their domestic accounting standards with IFRS. Evidence on this topic may also inform the
Boards as to the success of their efforts to achieve comparability through convergence, compared
to the outcome achieved through adoption. Germany provides an ideal setting to investigate this
issue.

The German market, adoption of IFRS, and enforcement implementation


In March 1997, the German government opened the New Market (Neue Markt) in an effort to
attract international investors to small- and medium-sized, high-growth corporations (Leuz 2003).
Firms that opted to list securities in this market were required to follow either U.S. GAAP or
IFRS.12 The government later permitted all publicly listed companies to choose among German
GAAP, U.S. GAAP, or IFRS beginning in April 1998 (German Parliament, “Law to Facilitate the
Raising of Capital”). Although originally successful, the economic downturn in the early part of
the last decade led to the closure of the New Market in 2002. Many German firms, however,
chose to continue reporting under U.S. GAAP until they were mandated to switch to IFRS in

10. They include “improvements in accounting standards; the accountability and funding of the IASC Foundation; the
improvement in the ability to use interactive data (XBRL) for IFRS reporting; education and training relating to
IFRS; limited early use of IFRS where this would enhance comparability for U.S. investors; the anticipated timing
of future rulemaking by the Commission; and the implementation of mandatory use of IFRS by U.S. issuers.” (SEC
2008, 10).
11. While the former SEC Chairman, Christopher Cox, supported adoption of IFRS (SEC 2008), his successor, Mary
Schapiro, preferred not to make such a decision and instead supported the continuation of the convergence projects.
12. IAS were the standards developed by the International Accounting Standards Committee (IASC), the predecessor
of the IASB. Some of the IAS are still effective. The term IFRS as used in this study refers to both sets of
standards.

CAR Vol. 36 No. 2 (Summer 2019)


Relative Effects of IFRS Adoption and IFRS Convergence 593

2005. Similarly, there were also many German firms that chose to apply IFRS before the manda-
tory requirement.
In addition to the adoption of IFRS effective on January 1, 2005, there were concurrent
changes to enforcement of accounting standards. Effective from the fourth quarter of 2005, the
Financial Reporting Enforcement Panel (FREP) was created to oversee publicly listed firms’ com-
pliance with reporting standards. The FREP can recommend that the financial regulatory authority
in Germany, the Bundesanstalt für Finanzdienstleistungsaufsicht (“BaFin”), take action in cases
of noncompliance. Prior studies find that enforcement is an important institutional factor that ren-
ders the adoption of IFRS effective, rather than firms merely adopting these standards as a label
(Daske et al. 2013). Thus, while it is difficult to disentangle the effects of adoption and enforce-
ment, without these regulatory changes the effect of adoption would be potentially undermined.
Consistent with this notion, Christensen et al. (2013) find that improvements in market liquidity
around mandatory IFRS adoption are limited to firms in EU countries that also enacted concurrent
changes in enforcement.13 Likewise, Cascino and Gassen (2015) find that compliance with IFRS
is an important factor to warrant increased comparability after IFRS adoption.

3. Prior research and hypothesis development


Previous studies find that a switch from local accounting standards to IFRS increases comparabil-
ity among non-U.S. firms and between U.S. firms and non-U.S. firms. For instance, Yip and
Young (2012) find that the mandatory adoption of IFRS in the EU increases cross-country com-
parability, measured by the similarity of accounting functions, the degree of information transfer,
and the similarity of the information content of accounting numbers. Barth et al. (2012) find that
the comparability between non-U.S. and U.S. firms increased after non-U.S. firms adopted IFRS.
Brochet et al. (2013) provide evidence that adoption of IFRS in the United Kingdom also
improves comparability, measured by reduced private information benefits. In theory, mandatory
adoption of IFRS in Germany should have immediately eliminated any accounting differences if
German companies fully complied with IFRS. We predict that comparability in Germany increased
after all German firms reported in accordance with IFRS. We therefore state our first hypothesis in
the alternative form:

HYPOTHESIS 1. A switch to IFRS from U.S. GAAP increases the comparability between
German U.S. GAAP and IFRS firms after IFRS adoption in 2005.

Convergence projects between the Boards strive to not only reduce the major accounting
differences between the two sets of standards but also jointly issue new accounting standards.
Consistent with this notion, using the IFRS-to-U.S. GAAP reconciliation disclosure in cross-listed
firms’ Form 20-F, Henry et al. (2009) find that the accounting differences between IFRS and
U.S. GAAP during 2004–2006 significantly decreased after the convergence projects between the
Boards were launched. In addition, using the model of mapping earnings to returns, Barth et al.
(2012) find that comparability between non-U.S. firms applying IFRS and U.S. firms is strongest
in more recent years, which could be at least partially due to the ongoing convergence projects.
In the setting of Germany, the convergence projects should have increased the comparability
between German IFRS and matched U.S. firms, all else being equal. We therefore state our sec-
ond hypothesis in the alternative form:

HYPOTHESIS 2. Convergence increases the comparability between German IFRS firms and
matched U.S. firms.

13. Finland, the Netherlands, Norway, and the United Kingdom are among other European countries to impose new
enforcement regulations alongside IFRS adoption in 2005.

CAR Vol. 36 No. 2 (Summer 2019)


594 Contemporary Accounting Research

The above two hypotheses predict that both adoption and convergence could increase com-
parability, but it is unclear whether the effects of adoption and convergence on comparability
would differ. On one hand, adoption should immediately eliminate any accounting differences
between U.S. GAAP and IFRS. On the other hand, comparability may be compromised by the
nature of principles-based IFRS that are normally more flexible and offer more discretionary
choices than U.S. GAAP. Thus, there is a possibility that comparability may not significantly
increase after IFRS adoption. Cascino and Gassen (2015), for instance, find that comparability
only marginally increases after 29 countries switched to IFRS. Using sample firms from both
Germany and Italy, their results show that only high levels of compliance result in increased
comparability.
Convergence may increase comparability if the major accounting differences between U.S.
GAAP and IFRS are significantly reduced or eliminated by the converged accounting standards.
For example, SFAS 123 was modified to be consistent with IFRS 2 that the value of employee
stock options be expensed as of the grant date. Similarly, SFAS 141 and IFRS 3 were revised
word for word, which resulted in a high degree of convergence when accounting for business
combinations. However, the convergence projects only cover some major accounting differences
between IFRS and U.S. GAAP and do not attempt to eliminate all differences (Norwalk Agreement,
FASB and IASB 2002). Hence, any accounting differences that are not reconciled by convergence
will continue to exist. Second, convergence is a continuous process, which has taken a relatively
long time to progress, whereas the effect of adoption is immediate. Since it is not clear ex ante
whether adoption leads to a more significant increase in comparability than convergence, we state
our final hypothesis in the null form as follows:

HYPOTHESIS 3. Adoption does not lead to a significant incremental increase in comparability


beyond convergence.

4. Research design
Measuring the effects of adoption and convergence on comparability
Following previous studies, we use a matched sample research design in assessing changes in
comparability. We first identify German U.S. GAAP firms that reported under U.S. GAAP in
2002, 2003, and 2004 and adopted IFRS by mandate in 2005. We then match these firms with
German IFRS firms by industry (based on 2-digit SIC codes) and market capitalization at the end
of 2004. We then classify the sample period into the pre-adoption (2002–2004) and post-adoption
(2006–2010) periods, assess any changes in comparability between these paired firms from the
pre- to post-adoption period, and attribute the changes in comparability to adoption, convergence,
and enforcement. This is because ongoing convergence and the new enforcement regulation
affected both German U.S. GAAP and German IFRS firms.14
We follow the same procedure described above and create a second matched sample that
consists of the same German IFRS firms included in the first matched sample and U.S. GAAP
firms during the sample period. We then assess any changes in comparability between these
paired firms from the pre- to post-adoption period. Since both German IFRS and U.S. GAAP
firms were affected by the convergence projects and German IFRS firms were affected by the
new enforcement regulation, we attribute any changes in comparability of the second matched
sample to convergence and enforcement.
To examine whether adoption leads to a significant incremental increase in comparability
beyond convergence and enforcement, we pool these two matched samples and employ the

14. We do not include 2005 (the adoption year) in our analyses, to mitigate concerns that detected changes are con-
founded by a possible temporary effect during the transition to IFRS.

CAR Vol. 36 No. 2 (Summer 2019)


Relative Effects of IFRS Adoption and IFRS Convergence 595

following difference-in-differences regression model to separate the changes in comparability


driven by convergence and adoption, respectively, from other potential economic factors.15

COMPn ijt = α0 + α1 POST ijt + α2 ADOPT ijt + α3 POST ×ADOPT ijt


+ α4 MV_Ratioijt + α5 SALESGROW_Ratioijt
+ α6 MB_Ratioijt + α7 LEVERAGE_Ratioijt
+ INDUSTRY + εijt : ð1Þ

In equation (1), COMPnijt represents a comparability metric (n = 1, 2, or 3) for the matched


pair of firms i and j in period t. We measure the comparability metrics in such a way that a larger
(or less negative) value indicates a greater comparability. POST is an indicator variable equal to
one if t is after 2005. ADOPT is an indicator variable equal to one if the matched pair of firms
i and j consists of a German U.S. GAAP and a German IFRS firm, and zero if the matched pair
of firms consists of a German IFRS firm and a U.S. firm. The coefficient on POST captures the
effects of convergence and enforcement on comparability as it reflects the change in comparabil-
ity for the matched German IFRS firms and U.S. firms. Our variable of interest is the interaction
term of POST×ADOPT. A positive and significant coefficient on this term (α3) suggests that
adoption leads to an incremental increase in comparability beyond convergence and enforcement.
We also follow Yip and Young (2012) to include MV_Ratio, measured as the ratio of the relative
market values of the firms included in each matched pair to control for the possibility that change
in firm size over time between the matched firms impacts comparability.
However, the choice to adopt IFRS early raises a potential concern of self-selection bias.
To mitigate this concern, we follow prior studies (Leuz 2003; Christensen et al. 2015) to con-
trol for relative sales growth (SALESGROW_Ratio), market-to-book (MB_Ratio), and leverage
(LEVERAGE_Ratio) ratios of the matched firms, as they have been found to be associated with
accounting standard choice, as well as industry fixed effects.

Comparability metrics
Following prior literature, we use three different comparability measures. De Franco et al. (2011)
posit that accounting comparability can be described as the degree to which accounting functions
similarly translate economic events (proxied by stock return) into financial statement information
(proxied by reported earnings). We therefore estimate the following regression to model each
firm’s accounting function:

NI=Pit = α i + β i RET it + εit , ð2Þ

where NI/Pi,t is calculated as net income before extraordinary items per share divided by the stock
price at the beginning of the year for firm i in period t. RETit is the stock return of firm i in period
t, adjusted for dividends and stock splits. The coefficients (αi and βi) represent the estimated
accounting function of firm i. Following Yip and Young (2012), we estimate equation (2) at the
firm level using semiannual data separately in the pre- and post-adoption periods.
We construct the first comparability metric as follows. First, for each matched pair of
German U.S. GAAP firm (firm i) and German IFRS firm (firm j), we estimate equation (2) sepa-
rately at the firm level to obtain the coefficients representing firm i’s accounting function (αi and
βi) and the coefficients representing firm j’s accounting function (α j and β j). Second, for firm i in
each semiannual period t, we calculate the expected values of NI/P using its own accounting

15. As a robustness test, we also include interaction terms between control variables and POST. Our results are qualita-
tively similar.

CAR Vol. 36 No. 2 (Summer 2019)


596 Contemporary Accounting Research

function (αi and βi) and the corresponding matched firm j’s accounting function (α j and β j), yielding
two predicted NI/P (ENI/Piit and ENI/P jit). The absolute value of the difference between the two
predicted values of NI/P is computed as | ENI/Piit − ENI/P jit |. Third, we repeat this process for firm
j in each semiannual period and obtain two predicted NI/P (ENI/P jjt and ENI/Pijt). Again, the
absolute value of the difference in the two predicted values of NI/P is computed as | ENI/P jjt −
ENI/Pijt |. Fourth, we calculate the mean of the above two absolute values to obtain the first compa-
rability metric (COMP1). We then multiply COMP1 by −1 so that a larger (or less negative) value
indicates greater comparability. We follow the same steps to obtain COMP1 for the matched
German IFRS firms and U.S. firms.
We next consider the comparability measure employed by Barth et al. (2012), based on the
mapping of earnings levels and changes into stock returns.

RET it = δ i 0 + δ i 1 NI=Pit + δ i 2 ΔNI=Pit + δ i 3 LOSSit


+ δ i 4 ðLOSSit × NI=Pit Þ + δ i 5 ðLOSSit × ΔNI=Pit Þ + εit : ð3Þ

In equation (3), RETit is the stock return of firm i in period t, measured from nine months
before until three months after the fiscal year-end and adjusted for dividends and stock splits.
NI/P is net income per share scaled by the stock price at the beginning of the fiscal year, and
ΔNI/P is the change in annual net income per share scaled by the stock price at the beginning of
the fiscal year. We also include an indicator variable equal to one for firms with negative net
income in year t (LOSS) and allow the coefficients on NI/P and ΔNI/P to differ for loss firms
(Hayn 1995).
The procedure to obtain our second comparability metric (COMP2) from equation (3) is iden-
tical to the procedure described above for COMP1, with the following exception. Because only
annual data is available for the variables needed to construct this metric, the estimation of firm-
level regressions is impractical. We therefore follow prior studies (Lang et al. 2010; Barth et al.
2012) to estimate equation (3) at the industry level (based on 2-digit SIC codes) and separately
for the pre- and post-adoption periods for each group of firms.
The above two comparability measures rely on the fundamental association between earnings
and stock returns that might have changed following IFRS adoption. As an additional approach,
we follow Cascino and Gassen (2015) to construct a third comparability measure that is based on
the mapping of operating cash flows into accruals. This measure captures not only a central aspect
of the accounting recognition process (Dechow 1994; Ball and Shivakumar 2005) but also has
the advantage of capturing fundamental economic events via cash flows.

ACC it = δ i 0 + δ i 1 CFOit + εit : ð4Þ

In equation (4), ACC is total accruals, deflated by lagged total assets.16 CFO is net operating
cash flow, deflated by lagged total assets. The procedures to obtain our third comparability metric
(COMP3) from equation (4) are identical to the procedures to obtain COMP2.

5. Sample and data


Accounting standards data are obtained from Worldscope (WS). We identify 74 (206) German
U.S. GAAP (German IFRS firms).17 From the 74 German U.S. GAAP firms, we eliminate 8 firms
with insufficient data requirements and 3 firms that cannot be matched based on industry. We

16. Total accruals are calculated as the difference between income from continued operations and net operating
cash flow.
17. We follow Appendix A of Daske et al. (2013) in coding accounting standards based on WS “Accounting Standards
Followed” (Field 07536).

CAR Vol. 36 No. 2 (Summer 2019)


Relative Effects of IFRS Adoption and IFRS Convergence 597

then match the remaining 63 firms to German IFRS firms based on industry and firm size as of
the end of 2004. We eliminate 16 firm pairs whose closest match based on the firms’ market
value of equity is inadequate.18 Thus, the final sample consists of 47 pairs of German U.S. GAAP
and German IFRS firms. We form our second matched sample using the same criteria but match
the German IFRS firms with U.S. firms. A list of firms used in this study is reported in online
Appendix B.19 All financial statement data is extracted from WS, and market data is collected
from Datastream.
Panel A of Table 1 summarizes our sample selection process. Panels B and C of Table 1
present the industry distribution of sample firms for the semiannual and annual analyses,
respectively. A majority of the sample firms are in business services, industrial equip-
ment, electronic equipment, medical instruments and supplies, and management services
industries.
Table 2 presents descriptive statistics for the input variables used to compute comparability
metrics. To mitigate the effects of outliers, we winsorize all continuous variables used in our ana-
lyses at the top and bottom 1 percent of their distributions.20 Panel A presents the descriptive sta-
tistics for the input variables that are used to construct COMP1. Both net income (NI/P) and stock
return (RET) are measured for each semiannual period. In contrast, panel B shows the descriptive
statistics for the input variables that are used to construct COMP2 and COMP3. ΔNI/P is the
change in annual net income per share, and all other variables are as previously defined. We find
that all three groups of firms appear to have similar mean stock performance, profitability, and
firm size. Univariate tests indicate that German U.S. GAAP firms tend to have lower sales growth
and leverage than German IFRS firms, while U.S. firms tend to have lower leverage than German
IFRS firms.21
The estimated parameters of models (2), (3), and (4) in the pre- and post-adoption periods
are important for constructing the three comparability measures used in this study. Untabulated
results show that, across the three different groups of firms and in both the pre- and post-adoption
periods, the parameters of estimating all three comparability measures are stable and consistent
with the predicted signs.22

6. Empirical results
Comparability metrics and effects of adoption and convergence on comparability
Table 3 presents a univariate analysis of the comparability metrics and firm characteristic ratios
for the German U.S. GAAP and IFRS matched firm pairs. Panel A of Table 3 reports the descrip-
tive statistics for the variables related to the analysis of COMP1. The mean (median) of COMP1
increases from −0.303 (−0.093) in the pre-adoption period to −0.077 (−0.047) in the post-
adoption period, and the differences are statistically significant at the 1 percent level. Panel B

18. Specifically, we calculate a ratio of the matched pair firms’ relative market values and exclude those matched pairs
with this ratio less than 0.50 or greater than 2 (Barth et al. 2012).
19. Please see supporting information “Appendix B: The list of firms used in this study” as an addition to the online
article.
20. Since our sample is relatively small, as a robustness test we use the bootstrapping method to obtain our parametric
estimates. Specifically, we construct 500 samples and estimate our primary models 500 times. The results are con-
sistent with our main finding that adoption and convergence do not have different effects on the comparability mea-
sures from De Franco et al. (2011), Barth et al. (2012), and Cascino and Gassen (2015).
21. Comparing German U.S. GAAP with U.S. firms, untabulated results show that, on average, German U.S. GAAP
firms are generally smaller, rely more on debt, and have lower market-to-book ratio and sales growth than U.S. firms.
We have controlled for these factors in our regression analysis.
22. We find that all the regression parameters are not significantly different between the pre- and post-adoption periods,
except the intercepts for German U.S. GAAP and German IFRS firms when using COMP1. These results suggest
that the observed increases in comparability are due to greater predictability between matched firm pairs rather than
the changing relationship within firms in the pre- and post-adoption periods.

CAR Vol. 36 No. 2 (Summer 2019)


598 Contemporary Accounting Research

TABLE 1
Sample selection and industry composition

Panel A: German U.S. GAAP firms selection

Comparability metrics
1 2,3
COMP COMP
Data type Semiannual Annual
German firms that use U.S. GAAP in 2002, 2003, and 74 74
2004 and use IFRS in 2006, 2007, 2008, 2009, and 2010
Exclusions
Firms with missing price or earnings data 8 5
Firms that cannot be matched to German IFRS firms based on industry 3 2
Firms with inadequate match based on size 16 15
Firms in industries with less than two firms — 7
Total number of German U.S. GAAP firms in each sample 47 45
Panel B: Industry composition for German U.S. GAAP firms with COMP1

Industry 2-digit SIC Frequency Percent

Pharmaceuticals and chemicals 28 1 2.13


Steel, iron, and minerals 33 1 2.13
Industrial equipment 35 7 14.89
Electronic equipment 36 6 12.77
Motor vehicles and equipment 37 1 2.13
Medical instruments and supplies 38 5 10.64
Insurance 63 1 2.13
Holding and investment 67 1 2.13
Business services 73 19 40.41
Motion pictures 78 1 2.13
Management services 87 4 8.51
Total 47 100.00
Panel C: Industry composition for German U.S. GAAP firms with COMP2,3

Industry 2-digit SIC Frequency Percent

Industrial equipment 35 7 15.56


Electronic equipment 36 7 15.56
Medical instruments and supplies 38 7 15.56
Business services 73 20 44.43
Management services 87 4 8.89
Total 45 100.00
Notes: This table presents the sample selection procedure and industry composition for German firms that
switched from U.S. GAAP to IFRS in 2005 (i.e., German U.S. GAAP firms). Panel A shows the numbers of
German U.S. GAAP firms with semiannual basis comparability metric COMP1 and annual basis
comparability metric COMP2 and COMP3, respectively. Panels B and C display the industry compositions
for those firms, respectively.

reports the correlations (Pearson correlations are below while Spearman correlations are above
the diagonal) of COMP1 and firm characteristic ratios. The Pearson correlation matrix shows that
COMP1 is positively correlated with POST (0.295), indicating that comparability is higher in the
post-adoption period. COMP1 is negatively correlated with MV_Ratio and LEVERAGE_Ratio

CAR Vol. 36 No. 2 (Summer 2019)


TABLE 2
Descriptive statistics: Input variables of comparability metrics and firm characteristic variables

Panel A: Input variables for COMP1 and firm characteristic variables

German U.S. GAAP firms (n = 701) German IFRS firms (n = 694) U.S. firms (n = 709)

Mean Median SD Mean Median SD Mean Median SD t-statistic# t-statistic## t-statistic###

NI/P − 0.069 0.021 0.410 − 0.050 0.024 0.620 − 0.030 0.018 0.229 0.67 0.78 2.54
RET − 0.018 0.031 0.425 − 0.019 0.010 0.396 0.004 0.053 0.459 0.06 1.08 0.92
MV 484.969 50.520 1,537.490 522.217 55.190 1,588.900 485.512 68.610 1,408.470 0.44 0.46 0.01
SALESGROW 0.047 0.036 0.361 0.099 0.052 0.345 0.103 0.068 0.347 2.75 0.26 3.01
MB 1.998 1.405 2.555 1.601 1.242 1.368 2.421 1.551 3.486 3.62 5.78 2.60
LEVERAGE 0.429 0.402 0.214 0.539 0.567 0.225 0.402 0.396 0.212 9.39 11.78 2.40
2,3
Panel B: Input variables for COMP and firm characteristic variables

German U.S. GAAP firms (n = 347) German IFRS firms (n = 336) U.S. firms (n = 339)

# ## ###
Mean Median SD Mean Median SD Mean Median SD t-statistic t-statistic t-statistic
RET − 0.038 − 0.004 0.700 − 0.038 − 0.010 0.665 − 0.007 0.000 0.665 0.00 0.58 0.58
NI/P − 0.035 0.043 0.266 − 0.047 0.046 0.302 − 0.058 0.035 0.384 0.58 0.41 0.82
ΔNI/P 0.075 0.015 0.394 0.145 0.013 0.646 0.117 0.014 0.587 1.73 0.70 1.01
LOSS 0.352 0.000 0.478 0.333 0.000 0.472 0.372 0.000 0.484 0.50 1.04 0.55
ACC − 0.053 − 0.049 0.110 − 0.054 − 0.034 0.117 − 0.067 − 0.060 0.134 0.11 1.35 1.40
CFO 0.058 0.066 0.125 0.053 0.050 0.106 0.058 0.074 0.155 0.51 0.47 0.01
MV 216.650 46.110 555.914 171.631 41.650 342.486 299.914 56.010 1,140.710 127 1.98 1.22
SALESGROW 0.066 0.043 0.415 0.107 0.053 0.391 0.104 0.072 0.380 1.31 0.09 1.25
MB 2.005 1.392 3.312 1.668 1.253 1.621 2.928 1.522 9.012 1.68 2.52 1.79
LEVERAGE 0.408 0.384 0.204 0.520 0.547 0.223 0.394 0.359 0.219 6.87 7.39 0.83

Notes: Panels A and B report the descriptive statistics for the input variables used to derive COMP1, COMP2, and COMP3, as well as firm characteristic variables for all three groups of firms used in
this study: German U.S. GAAP firms, German IFRS firms, and U.S. firms. In panel A, NI/P is net income per share scaled by the stock price at the beginning of the semiannual period. RET is the
semiannual stock return. In panel B, RET is computed from nine months before until three months after fiscal year-end (i.e., annual stock return). NI/P is the annual net income per share scaled by the
stock price at the beginning of the fiscal period. ΔNI/P is the annual change in net income per share. LOSS is an indicator variable that is equal to one if NI/P is negative, and zero otherwise. ACC is
total accruals, calculated as net income less operating cash flows, deflated by lagged total assets. CFO is cash flows from continuing operations deflated by lagged total assets. In both panels, MV is the
Relative Effects of IFRS Adoption and IFRS Convergence

company’s total market value, in millions of euros, at the end of the fiscal period. SALESGROW is the percentage change in sales. MB is the market-to-book ratio, calculated as the market value of
equity divided by the book value of equity. LEVERAGE is total liabilities divided by total assets. #, ##, and ### indicate the corresponding t-statistic based on the difference in the means (non-
directional) between German U.S. GAAP firms and German IFRS firms, between U.S. firms and German IFRS firms, and between German U.S. GAAP firms and U.S. firms, respectively.
599

CAR Vol. 36 No. 2 (Summer 2019)


600 Contemporary Accounting Research

( − 0.326 and − 0.190, respectively), indicating that comparability is lower when German
U.S. GAAP and IFRS firms have greater difference in firm size and leverage. The Spearman corre-
lation matrix presents consistent results. Panel C reports the descriptive statistics for the variables
related to the analysis of COMP2 and COMP3. The mean (median) of COMP2 increases from
− 0.528 ( − 0.300) in the pre-adoption period to − 0.205 ( − 0.100) in the post-adoption period, and
the differences are statistically significant at the 1 percent level. Similarly, the mean (median) of
COMP3 increases from − 0.052 ( − 0.035) in the pre-adoption period to − 0.024 ( − 0.017) in the
post-adoption period. The differences are statistically significant at the 1 percent level. The Pearson
correlation matrix reported in panel D shows that COMP2 and COMP3 are positively correlated
with each other (0.332) and both are positively correlated with POST. The Spearman correlation
matrix shows consistent results. Finally, both panels A and C show that German U.S. GAAP firms
have higher market capitalization than their matched German IFRS firms, but the difference reduces
in the post-adoption period.
We next discuss the comparability between matched German IFRS and U.S. firms. Panel A
of Table 4 shows that the mean (median) of COMP1 increases from − 0.277 ( − 0.109) in the
pre-adoption period to − 0.088 ( − 0.052) in the post-adoption period, and the differences are sta-
tistically significant at the 1 percent level. The Pearson correlation matrix reported in panel B
shows that COMP1 is positively correlated with POST (0.296), indicating that comparability
increases in the post-adoption period. COMP1 is also negatively correlated with MV_Ratio and
LEVERAGE_Ratio ( − 0.144 and − 0.178, respectively). In panel C of Table 4, the mean
(median) of COMP2 increases from − 0.760 ( − 0.387) in the pre-adoption period to − 0.303
( − 0.176) in the post-adoption period, and these changes are also statistically significant at the
1 percent level. Similarly, the mean (median) of COMP3 increases from − 0.067 ( − 0.047) in
the pre-adoption period to − 0.053 ( − 0.046) in the post-adoption period, and the change is sta-
tistically significant at the 1 percent (10 percent) level. The Pearson correlation matrix reported
in panel D shows that both COMP2 and COMP3 are positively correlated (0.385), and that both
COMP2 and COMP3 are positively associated with POST (0.310 and 0.169, respectively), indi-
cating that comparability increases in the post-adoption period. Finally, U.S. firms have higher
market-to-book ratios than their matched German IFRS firms, but the difference reduces in the
post-adoption period.
To illustrate the changes in comparability over the sample period, we convert the univari-
ate results by year into the graphs presented in Figure 1 and highlight the following trends.
First, there is a sharp increase in COMP1 over the sample period for both the means and
medians of adoption pairs (ADOPT_MEAN and ADOPT_MED) and convergence pairs (CON-
V_MEAN and CONV_ MED). The increase in COMP1 appears to taper off after 2005. Second,
although a similar pattern is observed for COMP2, the effect fluctuates somewhat throughout
the post-adoption period. Finally, for the adoption pairs, there is a significant increase in
COMP3, which is sustained throughout the post-adoption period. On the other hand, while
there appears to be an increasing trend in COMP3 for the convergence pairs, it is a much less
drastic effect.

Regression results: Relative effects of adoption and convergence on comparability


Results described in the previous section suggest that both adoption and convergence increase
comparability after the new enforcement regulation in 2005. This section examines if adoption
leads to a significant incremental increase in comparability beyond convergence and enforcement.
Table 5 reports the regression results from estimating equation (1). Reported p-values are based
on robust standard errors clustered at the firm-pair level. The coefficient on POST is positive and
statistically significant at the 1 percent level when comparability is measured as either COMP1 or
COMP2 and at the 10 percent level when comparability is measured as COMP3. These results are
consistent with our earlier univariate results that convergence appears to have improved compara-
bility. The coefficient on POST × ADOPT, however, is insignificant when comparability is

CAR Vol. 36 No. 2 (Summer 2019)


TABLE 3
Descriptive statistics of comparability metrics and relative firm characteristic ratios for German U.S. GAAP firm and German IFRS firm matched pairs

Panel A: COMP1 and firm characteristic ratios for the German U.S. GAAP and IFRS firm matched pairs

Pre-adoption (n = 263) Post-adoption (n = 391)

Mean Median SD Mean Median SD t-statistic Wilcoxon z

COMP1 − 0.303 − 0.093 0.552 − 0.077 − 0.047 0.102 − 7.90 − 6.26


MV_Ratio 2.985 1.128 10.521 1.515 0.910 1.702 2.71 4.42
SALESGROW_Ratio 0.331 0.456 6.178 0.832 0.476 5.614 − 1.07 − 1.07
MB_Ratio 1.876 1.138 2.217 1.695 1.122 2.073 1.07 0.75
LEVERAGE_Ratio 0.888 0.816 1.275 0.756 0.736 0.925 1.53 1.31
Panel B: The correlations of COMP1 and firm characteristic ratios for the German U.S. GAAP and IFRS firm matched pairs

COMP1 POST MV_Ratio SALESGROW_Ratio MB_Ratio LEVERAGE_Ratio

COMP1 0.245 − 0.130 0.083 − 0.034 − 0.138


POST 0.295 − 0.173 0.042 − 0.030 − 0.051
MV_Ratio − 0.326 − 0.106 0.010 0.461 − 0.093
SALESGROW_Ratio 0.084 0.042 − 0.035 0.014 − 0.152
MB_Ratio − 0.031 − 0.042 0.111 0.049 0.060
LEVERAGE_Ratio − 0.190 − 0.060 0.002 − 0.058 0.105
2 3
Panel C: COMP , COMP , and firm characteristic ratios between German U.S. GAAP and IFRS firm matched pairs

Pre-adoption (n = 126) Post-adoption (n = 199)

Mean Median SD Mean Median SD t-statistic Wilcoxon z

COMP2 − 0.528 − 0.300 0.660 − 0.205 − 0.100 0.264 − 6.17 − 6.91


COMP3 0.051 0.022
Relative Effects of IFRS Adoption and IFRS Convergence

− 0.052 − 0.035 − 0.024 − 0.017 − 6.62 − 4.42


MV_Ratio 3.082 1.186 10.755 1.724 1.124 1.856 1.74 1.99
(The table is continued on the next page.)
601

CAR Vol. 36 No. 2 (Summer 2019)


602

TABLE 3 (continued)

Panel C: COMP2, COMP3, and firm characteristic ratios between German U.S. GAAP and IFRS firm matched pairs

Pre-adoption (n = 126) Post-adoption (n = 199)

Mean Median SD Mean Median SD t-statistic Wilcoxon z

SALESGROW_Ratio 0.835 0.385 7.281 1.848 0.494 8.617 − 1.09 − 0.93


MB_Ratio 1.845 1.074 2.285 1.672 1.145 2.003 0.72

CAR Vol. 36 No. 2 (Summer 2019)


− 0.16
LEVERAGE_Ratio 0.775 0.746 1.664 0.760 0.681 1.167 0.09 0.62
Panel D: The correlations of COMP2, COMP3, and firm characteristic ratios between German U.S. GAAP and IFRS firm matched pairs

COMP2 COMP3 POST MV_Ratio SALESGROW_Ratio MB_Ratio LEVERAGE_Ratio


Contemporary Accounting Research

COMP2 0.407 0.384 − 0.049 0.064 − 0.045 0.023


COMP3 0.332 0.245 − 0.067 − 0.008 − 0.102 0.029
POST 0.325 0.346 − 0.110 0.052 0.009 − 0.034
MV_Ratio − 0.125 − 0.056 − 0.096 0.014 0.431 − 0.077
SALESGROW_Ratio 0.013 0.014 0.061 − 0.025 0.058 − 0.140
MB_Ratio − 0.076 − 0.062 − 0.040 0.106 0.018 0.023
LEVERAGE_Ratio − 0.063 − 0.124 − 0.005 − 0.001 − 0.004 0.164
Notes: This table presents the descriptive statistics and correlations (Pearson correlation is below while Spearman correlation is above the diagonal) of the comparability
metrics and firm characteristic ratios for the German U.S. GAAP and German IFRS firm matched pairs in the pre-adoption (2002–2004) and post-adoption (2006–2010)
periods. Correlation coefficients in bold are significant at the 0.01 level. The coefficients in italics are significant at the 0.10 level (two-tailed). The last two columns
of panels A and C report the statistical differences in the means (t-test) and medians (Wilcoxon z-test) of these variables across the pre- and post-adoption periods.
Panels A and B are based on the semiannual sample. Panels C and D are based on the annual sample. COMP1, COMP2, and COMP3 are the comparability metrics.
All the comparability measures are multiplied by − 1 so that a larger (or less negative) metric value indicates greater comparability. MV_Ratio,
SALESGROW__Ratio, MB_Ratio, and LEVERAGE_Ratio are the relative market value, sales growth, market-to-book, and leverage ratios of the matched German
U.S. GAAP firm and German IFRS firm.
TABLE 4
Descriptive statistics of comparability metrics and relative firm characteristic ratios for U.S. firm and German IFRS firm matched pairs

Panel A: COMP1 and firm characteristic ratios between U.S. firm and German IFRS firm matched pairs

Pre-adoption (n = 242) Post-adoption (n = 411)

Mean Median SD Mean Median SD t-statistic Wilcoxon z

COMP1 − 0.277 − 0.109 0.466 − 0.088 − 0.052 0.102 − 7.90 − 7.65


MV_Ratio 2.100 1.383 2.622 2.297 1.140 4.178 − 0.66 3.74
SALESGROW_Ratio 0.388 0.208 8.363 0.923 0.474 7.627 − 0.83 − 0.93
MB_Ratio 2.862 1.421 3.981 1.948 1.170 2.990 3.33 3.68
LEVERAGE_Ratio 0.995 0.712 1.208 0.949 0.729 0.855 0.56 0.02
Panel B: The correlations of COMP1 and firm characteristic ratios between U.S. firm and German IFRS firm matched pairs

COMP1 POST MV_Ratio SALESGROW_Ratio MB_Ratio LEVERAGE_Ratio

COMP1 0.300 − 0.083 0.135 − 0.025 − 0.179


POST 0.296 − 0.147 0.036 − 0.144 − 0.001
MV_Ratio − 0.144 0.026 − 0.024 0.334 0.000
SALESGROW_Ratio 0.047 0.033 − 0.146 − 0.070 − 0.085
MB_Ratio 0.009 − 0.129 0.073 − 0.079 0.053
LEVERAGE_Ratio − 0.178 − 0.022 0.010 0.061 0.055
2 3
Panel C: COMP , COMP , and firm characteristic ratios between U.S. firm and German IFRS firm matched pairs

Pre-adoption (n = 122) Post-adoption (n = 193)

Mean Median SD Mean Median SD t-statistic Wilcoxon z

COMP2 − 0.760 − 0.387 0.965 − 0.303 − 0.176 0.421 − 5.77 − 7.64


COMP3 0.049 0.033
Relative Effects of IFRS Adoption and IFRS Convergence

− 0.067 − 0.047 − 0.053 − 0.046 − 3.04 − 1.97


MV_Ratio 2.098 1.395 2.394 2.704 1.453 4.933 − 1.27 1.30
SALESGROW_Ratio 0.539 0.094 7.097 2.296 0.487 11.715 − 1.49 − 1.35
603

CAR Vol. 36 No. 2 (Summer 2019)


(The table is continued on the next page.)
604

TABLE 4 (continued)

Panel C: COMP2, COMP3, and firm characteristic ratios between U.S. firm and German IFRS firm matched pairs

Pre-adoption (n = 122) Post-adoption (n = 193)

Mean Median SD Mean Median SD t-statistic Wilcoxon z

MB_Ratio 2.690 1.417 3.501 1.930 1.195 2.857 2.10 2.32

CAR Vol. 36 No. 2 (Summer 2019)


LEVERAGE_Ratio 0.914 0.670 1.812 0.948 0.686 0.920 − 0.22 − 0.05
Panel D: The correlations of COMP2, COMP3, and firm characteristic ratios between U.S. firm and German IFRS firm matched pairs

COMP2 COMP3 POST MV_Ratio SALESGROW_Ratio MB_Ratio LEVERAGE_Ratio


Contemporary Accounting Research

COMP2 0.226 0.431 − 0.071 0.130 − 0.063 − 0.074


COMP3 0.385 0.111 − 0.043 − 0.061 − 0.021 − 0.073
POST 0.310 0.169 − 0.073 0.076 − 0.131 0.003
MV_Ratio − 0.142 − 0.086 0.071 − 0.061 0.355 0.068
SALESGROW_Ratio 0.052 − 0.120 0.084 − 0.095 − 0.047 − 0.092
MB_Ratio − 0.034 0.017 − 0.118 0.037 0.138 0.075
LEVERAGE_Ratio − 0.101 − 0.075 0.012 0.033 0.106 0.051
Notes: This table presents the descriptive statistics and correlations (Pearson correlation is below while Spearman correlation is above the diagonal) of the comparability
metrics and firm characteristic ratios for the U.S. firm and German IFRS firm matched pairs in the pre-adoption (2002–2004) and post-adoption (2006–2010) periods.
Correlation coefficients in bold are significant at the 0.01 level. The coefficients in italics are significant at the 0.10 level (two-tailed). The last two columns of panels A
and C report the statistical differences in the means (t-test) and medians (Wilcoxon z-test) of these variables across the pre- and post-adoption periods. Panels A and B are
based on the semiannual sample. Panels C and D are based on the annual sample. COMP1, COMP2, and COMP3 are the comparability metrics. All the comparability
measures are multiplied by − 1 so that a larger (or less negative) metric value indicates greater comparability. MV_Ratio, SALESGROW_Ratio, MB_Ratio, and
LEVERAGE_Ratio are the relative market value, sales growth, market-to-book, and leverage ratios of the matched U.S. firm and German IFRS firm.
Relative Effects of IFRS Adoption and IFRS Convergence 605

Figure 1 Changes in comparability metrics over sample period

Notes: ADOPT_MEAN and CONV_MEAN indicate the mean increased comparability caused by adoption
and convergence, respectively. ADOPT_MED and CONV_MED indicate the median increased comparability
caused by adoption and convergence, respectively. The plotted points on each graph are the corresponding
annual predicted values.

measured as either COMP1 or COMP2, but it is significant at the 5 percent level (coeffi-
cient = 0.227, p = 0.047) when comparability is measured as COMP3. In sum, these results sug-
gest that adoption does not necessarily lead to a significant incremental increase in comparability
beyond convergence and enforcement.

Robustness tests
We have performed a variety of additional tests to ensure that our results are robust to alternative
research design choices.

Potential effect of the recent financial crisis


To further investigate whether our main findings are affected by the recent financial crisis, we
compare the results using sample periods of 2002–2008 and 2002–2010 and find that results
using both test periods are qualitatively consistent. Together with our finding that the estimated
parameters of our comparability regression models appear to be reasonably stable in both the pre-
and post-adoption periods, the recent financial crisis does not appear to have any significant
adverse effect on our inferences.

CAR Vol. 36 No. 2 (Summer 2019)


606 Contemporary Accounting Research

TABLE 5
Incremental effects of adoption on comparability beyond convergence and enforcement

COMP1 COMP2 COMP3

Estimate p-value Estimate p-value Estimate p-value

Intercept − 0.621 0.003 − 1.125 < 0.0001 − 0.107 < 0.0001


ADOPT − 0.015 0.731 0.221 0.074 0.015 0.017
POST 0.197 0.003 0.453 < 0.0001 0.014 0.063
POST×ADOPT 0.009 0.859 − 0.149 0.227 0.013 0.047
MV_Ratio − 0.013 < 0.0001 − 0.010 0.034 0.000 0.814
SALESGROW_Ratio 0.002 0.059 0.002 0.132 0.000 0.107
MB_Ratio 0.012 0.211 − 0.001 0.934 0.000 0.391
LEVERAGE_Ratio − 0.024 0.291 − 0.027 0.162 − 0.001 0.315
INDUSTRY Yes Yes Yes
Model 11.7336 24.7129 33.9325
Adjusted R2 0.3102 0.2133 0.2953
Number of observations 1307 640 640
Notes: This table reports the difference-in-differences results derived from estimating equation (1), which is
used to assess the incremental effect of IFRS adoption from U.S. GAAP on comparability beyond
convergence between U.S. GAAP and IFRS and the new enforcement regulation. COMPnijt denotes one of
the three comparability metrics (COMP1, COMP2, or COMP3) for the matched pair with firm i and firm j in
period t (the subscripts are omitted in the table). COMP1, COMP2, and COMP3 are defined in Tables 3 and
4. A larger (or less negative) metric value indicates greater comparability. POST is an indicator variable
equal to one if year t is after 2005. ADOPT is an indicator variable equal to one for the matched pairs
consisting of a German U.S. GAAP firm and a German IFRS firm, and zero for the matched pairs consisting
of a U.S. firm and a German IFRS firm. MV_Ratio, SALESGROW_Ratio, MB_Ratio, and LEVERAGE_Ratio
are the relative market value, sales growth, market-to-book, and leverage ratios of the matched pairs.
INDUSTRY represents dummy variables to control for industry fixed effects. Reported p-values (two-tailed)
are based on the t-statistics using the White (1980) heteroskedasticity-adjusted robust variance estimates,
further clustered at firm-pair level.
Controlling for the impact of enforcement
Germany passed a new financial reporting enforcement regulation in 2005 to ensure full compliance
with IFRS, which is a potentially confounding event with IFRS adoption. Strong enforcement could
significantly limit managerial discretion and increase financial statement comparability. Christensen
et al. (2013) argue that both enforcement and IFRS adoption are bundled and cannot be easily sepa-
rated in countries like Germany.
The difference-in-differences design employed in this study partially mitigates the above con-
cern because the effects of both adoption and convergence are affected by the enforcement regu-
lation. To further shed light on whether enforcement affects comparability and whether the
enforcement effect dominates our finding, we conduct the following tests. First, we measure the
enforcement effect by examining the change in comparability of the pair-matched German IFRS
firms (i.e., one German IFRS firm is matched with another German IFRS firm) in the pre- and
post-adoption periods.23 Conceptually, since all German IFRS firms complied with the same

23. We create the matched sample for the same German IFRS firms used in the adoption and convergence samples. We
match each of these German IFRS firms with another German IFRS firm in the same industry (based on 2-digit SIC
codes) and that has the most similar size (measured as market value of equity) at the end of 2004. We exclude those
matched pairs with MV_Ratio less than 0.50 or greater than 2. We further impose the requirement of not having a
mutual match. In other words, if one German IFRS firm i is matched with one German IFRS firm j, we don’t allow
firm j to also be matched with firm i.

CAR Vol. 36 No. 2 (Summer 2019)


Relative Effects of IFRS Adoption and IFRS Convergence 607

accounting standards, any increased comparability between the pair-matched German IFRS firms
over the sample period can be attributed to enforcement.
We compare the observed adoption effect with any changes in comparability driven by
enforcement. The model used to investigate this issue is as follows:

COMPn ijt = α0 + α1 POST t + α2 ADOPT_ENF ijt + α3 POST t × ADOPT_ENF ijt


+ α4 MV_Ratioijt + α5 SALESGROW_Ratioijt + α6 MB_Ratioijt
+ α7 LEVERAGE_Ratioijt + INDUSTRY + εijt , ð5Þ

where COMPnijt denotes one of the three comparability metrics for the matched pair containing
German IFRS firms i and j in period t; ADOPT_ENF is an indicator variable equal to one for the
matched pairs consisting of a German U.S. GAAP firm and a German IFRS firm, and zero for the

TABLE 6
Incremental effects of adoption on comparability beyond enforcement

COMP1 COMP2 COMP3

Estimate p-value Estimate p-value Estimate p-value

Intercept − 0.748 0.001 − 1.005 0.019 − 0.092 < 0.0001


ADOPT_ENF 0.150 0.187 − 0.059 0.513 − 0.015 0.006
POST 0.286 0.000 0.143 0.197 0.010 0.055
POST × ADOPT_ENF − 0.096 0.363 0.176 0.095 0.018 0.045
MV_Ratio − 0.077 0.008 0.002 0.666 0.000 0.099
SALESGROW_Ratio 0.006 0.031 − 0.019 0.390 0.000 0.805
MB_Ratio 0.016 0.181 − 0.002 0.683 0.000 0.849
LEVERAGE_Ratio − 0.002 0.894 0.000 0.853 0.000 0.850
INDUSTRY Yes Yes Yes
Model 6.47 13.3711 63.689
Adjusted R2 0.3973 0.1693 0.3929
Number of observations 1171 573 573
Notes: This table reports the difference-in-differences results from estimating equation (5), which is
used to assess the incremental effect of IFRS adoption from U.S. GAAP on comparability beyond what
may be caused by enforcement. The enforcement effect is derived from changes in comparability
between matched pairs of German IFRS firms. COMPnijt denotes one of the three comparability metrics
(COMP1, COMP2, or COMP3) for the matched pair with firm i and firm j in period t (subscripts are omitted
in the table). COMP1, COMP2 and COMP3 are defined in Tables 3 and 4. A larger (or less negative) metric
value indicates greater comparability. POST is an indicator variable equal to one if year t is after 2005.
ADOPT_ENF is an indicator variable equal to one for the matched pairs consisting of a German U.S. GAAP
firm and a German IFRS firm, and zero for the matched pairs consisting of a German IFRS firm and another
German IFRS firm. MV_Ratio, SALESGROW_Ratio, MB_Ratio, and LEVERAGE_Ratio are the relative
market value, sales growth, market-to-book, and leverage ratios of the matched pairs. INDUSTRY represents
dummy variables to control for industry fixed effects. Reported p-values (two-tailed) are based on the
t-statistics using the White (1980) heteroskedasticity-adjusted robust variance estimates, further clustered at
firm-pair level.

matched pairs consisting of a German IFRS firm and another German IFRS firm. Other variables
are as previously defined.
Table 6 shows that the coefficients on POST are positive and significant for two out of three
comparability measures, suggesting that enforcement increases comparability. More importantly,

CAR Vol. 36 No. 2 (Summer 2019)


608 Contemporary Accounting Research

we find that POST × ADOPT_ENF is statistically significant at conventional levels when using
COMP2 and COMP3 (coefficient = 0.175 and 0.018, p = 0.095 and 0.45, respectively). In other
words, we find evidence that increased comparability caused by adoption of IFRS is greater than
the increased comparability that might be attributable to the new enforcement regulation in
Germany. This is the first study to compare the effects of IFRS adoption and concurrent changes
in enforcement within a single country setting. We also follow the same procedures to compare
the observed convergence effect with any changes in comparability driven by enforcement.
Untabulated results show that convergence does not provide a significant incremental increase in
comparability beyond enforcement, which could be caused by the fact that the new enforcement
regulation has an asymmetric effect on the two series of firm pairs. That is, both firms in the
German IFRS–German IFRS pairs are influenced by the new enforcement regulation, while only
German firms in the German IFRS–U.S. GAAP pairs are affected. Finally, untabulated results
also show that adoption does not lead to a significant incremental increase in comparability
beyond convergence and enforcement.

Comparability between German U.S. GAAP firms and U.S. firms after IFRS adoption
Both German U.S. GAAP and U.S. firms prepared their financial statements in the pre-adoption
period using U.S. GAAP. However, in the post-adoption period, German U.S. GAAP firms pre-
pared their financial statements under IFRS. To the extent that firms using the same accounting
standards are more comparable than firms that use different accounting standards, we expect to
observe a decrease in accounting comparability between German U.S. GAAP firms and U.S. firms
in the post-adoption period relative to the pre-adoption period. We investigate this issue using the
following model:

COMPn ijt = α0 + α1 POST t + α2 REVERSEijt + α3 POST t × REVERSE ijt


+ α4 MV_Ratioijt + α5 SALESGROW_Ratioijt + α6 MB_Ratioijt
+ α7 LEVERAGE_Ratioijt + INDUSTRY + εijt , ð6Þ

where REVERSE is an indicator variable equal to one for the matched pairs consisting of a
German U.S. GAAP firm and a U.S. firm, and zero for the matched pairs consisting of a Ger-
man IFRS firm and a U.S. firm. The other variables are as previously defined. Consistent with
our prediction, Table 7 shows that POST × REVERSE is negative and significant at conven-
tional levels using the comparability measures of COMP2 and COMP3 (coefficient = − 0.243
and − 0.011, p = 0.046 and 0.05, respectively), indicating that accounting comparability
between German U.S. GAAP firms and U.S. firms decreases significantly in the post-adoption
period.

Using all German IFRS firms as the basis to measure the convergence effect
This study uses matched firm pairs of German U.S. GAAP and German IFRS firms to investi-
gate the adoption effect. The same German IFRS firms used for investigating the adoption
effect are used to match U.S. firms to investigate the convergence effect. To maximize the
sample size and potentially provide a better estimate of the increased comparability driven by
convergence, we repeat our analyses using all available German IFRS firms and their matched
U.S. firms to investigate whether our previous findings continue to hold. We find that conver-
gence leads to an improvement in comparability over our sample period for two out of three
comparability measures. Our difference-in-differences tests, however, show that in this speci-
fication adoption leads to a significant increase in comparability beyond convergence and
enforcement for two out of three comparability measures. However, we note that since

CAR Vol. 36 No. 2 (Summer 2019)


Relative Effects of IFRS Adoption and IFRS Convergence 609

TABLE 7
Comparability between German U.S. GAAP firms and U.S. firms after IFRS adoption

COMP1 COMP2 COMP3

Estimate p-value Estimate p-value Estimate p-value

Intercept − 0.541 0.003 − 1.444 < 0.0001 − 0.119 < 0.0001


REVERSE 0.059 0.338 0.126 0.147 0.013 0.006
POST 0.185 0.002 0.477 0.000 0.014 0.059
POST×REVERSE − 0.064 0.301 − 0.243 0.046 − 0.011 0.050
MV_Ratio − 0.007 0.052 − 0.022 0.066 − 0.001 0.025
SALESGROW_Ratio 0.000 0.967 0.004 0.219 0.000 0.335
MB_Ratio 0.013 0.144 0.010 0.011 0.000 0.771
LEVERAGE_Ratio − 0.001 0.950 0.002 0.873 0.001 0.514
INDUSTRY Yes Yes Yes
Model 6.706 13.5905 39.904
Adjusted R2 0.2462 0.2041 0.3003
Number of observations 1,313 642 642
Notes: This table reports the difference-in-differences results from estimating equation (6), which is
used to assess whether the departure from using U.S. GAAP has reduced the comparability between
German U.S. GAAP firms and U.S. firms, with the control of effects from convergence between IFRS
and U.S. GAAP on comparability. COMPnijt denotes one of the three comparability metrics (COMP1,
COMP2, or COMP3) for the matched pair with firm i and firm j in period t (the subscripts are omitted in the
table). COMP1, COMP2, and COMP3 are defined in Tables 3 and 4. A larger (or less negative) metric value
indicates greater comparability. POST is an indicator variable equal to one if year t is after 2005. REVERSE
is an indicator variable equal to one for the matched pairs consisting of a German U.S. GAAP firm and a
U.S. firm, and zero for the matched pairs consisting of a German IFRS firm and a U.S. firm. MV_Ratio,
SALESGROW_Ratio, MB_Ratio, and LEVERAGE Ratio are the relative market value, sales growth, market-
to-book, and leverage ratios of the matched pairs. INDUSTRY represents dummy variables to control for
industry fixed effects. Reported p-values (two-tailed) are based on the t-statistics using the White (1980)
heteroskedasticity-adjusted robust variance estimates, further clustered at firm-pair level.

this result is not based on matched firm pairs, the adoption effect may not be comparable with
the convergence effect given the differences in firm size, industrial sector, and other firm
characteristics.

Trade-off between treatment effect and population generalization


Cram et al. (2009) argue that the matching approach that is widely used in accounting literature
reflects a trade-off between identifying the treatment effects and generalizing the results to the full
population. As a result, the matching approach could reduce the power of tests or bias empirical
results due to systematic differences in the subsamples from the full population. We address this
concern in two ways. First, we include pair fixed effects as additional control variables. Second,
we use the pairwise difference as the dependent variable and regress it directly on POST. Untabu-
lated results are robust to both of these adjustments.

Adoption and convergence compliance


Our finding that adoption and convergence do not have different effects on comparability may
have several possible alternative explanations. In this section, we describe our analyses of these

CAR Vol. 36 No. 2 (Summer 2019)


610 Contemporary Accounting Research

alternatives. First, it is possible that companies attempted to align their accounting practices such
that the effect of adoption would result in limited accounting changes. For example, German
U.S. GAAP firms might make accounting choices that are closer to IFRS before 2005 or make
accounting choices that are closer to U.S. GAAP after 2005. We believe that the first scenario
would be rare because U.S. GAAP generally have fewer accounting choices than IFRS.24 Inven-
tory methods are one of the few exceptions as only U.S. GAAP admits LIFO. We examine the
financial statements of 15 German U.S. GAAP firms and find that no companies switched from
LIFO to FIFO prior to IFRS adoption.25 The second scenario could occur if the accounting
choices under IFRS and U.S. GAAP are the same and firms continue to use the same accounting
practices before and after IFRS adoption. We find that it is a common practice that firms continue
to follow the same reporting practices under U.S. GAAP if they are also permitted by IFRS.
Although it is possible that the above trend may have resulted in a somewhat more limited effect
of adoption, it is difficult to attribute such choices to an intent to minimize the effect of adoption
of IFRS.
Second, adoption and convergence may not provide different effects on comparability if
the accounting differences between IFRS and U.S. GAAP are small. To investigate this possi-
bility, we perform two additional tests using the same sample of 15 German U.S. GAAP
firms. First, using the U.S. GAAP-to-IFRS reconciliation disclosure, we identify the key items
causing the accounting differences and their magnitude. Panel A of Table 8 shows that a man-
datory switch from U.S. GAAP to IFRS reduced the accounting differences in net income and
shareholders’ equity by 5 and 3 percent, respectively, which are economically material. Third,
German U.S. GAAP firms may not have been in full compliance with IFRS after adoption.
To address this possibility, we investigate whether German U.S. GAAP firms complied with
the measurement and disclosure requirements following IFRS adoption for the following key
accounting differences: share-based payments, pension and post-employment benefits, intangible
assets, impairments, and business combinations. Panel B of Table 8 shows that the average mea-
surement and disclosure compliance rates are high except for two disclosure items related to pen-
sions and post-employment benefits. Thus, we can reasonably conclude that neither immaterial
accounting differences nor lack of compliance are likely the alternative explanations for our
findings.
We also validate the results for the convergence effect by comparing the average mea-
surement and disclosure compliance rates for a sample of German IFRS firms and matched
U.S. firms. This test is based on the converged accounting standards up to 2010, including
share-based payments, discontinued operations, borrowing costs, segment reporting, and busi-
ness combinations.26 Panel C of Table 8 shows that, in general, the average measurement and
disclosure compliance rates are high and comparable for both groups of firms except for a
few items related to segment reporting. In summary, we find consistent evidence that compli-
ance is high for both German IFRS firms and matched U.S. firms following the converged
accounting standards.

24. We investigate whether the first scenario could apply to any accounting items causing major accounting differences,
including share-based payments, pension and post-employee benefits, intangible assets, impairments, and business
combinations. In all cases, IFRS include the same accounting choices as U.S. GAAP, plus at least one additional
option, which excludes the possibility of the first scenario.
25. These 15 German U.S. GAAP firms switched to IFRS in 2005 and provided financial statements in English for
2004 and 2006 on their company websites.
26. Our sample for this analysis includes 15 German IFRS firms with financial statements available in English on their
websites for the period 2005–2010 and their matched U.S. firms.

CAR Vol. 36 No. 2 (Summer 2019)


TABLE 8
Adoption and convergence effects

Panel A: Accounting differences between U.S. GAAP and IFRS

Accounting Difference Difference


Company differences in NI NI (U.S. GAAP) Percentage in EQ EQ (U.S. GAAP) Percentage

Syzygy AG 1 40 2,895 1.38 0 55,737 0.00


Dialog Semiconductor PLC 1, 2, 3 533 5,743 9.28 12,908 121,135 10.66
SinnerSchrader AG 1 8 531 1.51 0 8,054 0.00
Jubii Europe NV 0 45,476 0.00 0 146,198 0.00
Pfeiffer Vacuum Technology AG 1, 4, 5, 6 159 22,748 0.70 633 112,631 0.56
Intershop Communications AG 1, 7 145 8,921 1.63 0 2,655 0.00
Realtech AG 0 1,760 0.00 0 42,442 0.00
Continental AG 7, 8 61,000 1,096,400 5.56 94,900 2,842,300 3.34
Leoni AG 8, 9, 10, 11, 12 5,551 33,225 16.71 6,437 371,340 1.73
GfK AG 1, 2, 5, 7, 8, 13, 14 10,367 63,502 16.33 8,036 264,786 3.03
Evotec AG 9, 10 6,391 84,203 7.59 8,498 102,010 8.33
CeoTronics AG 1, 3 51 1,001 5.09 63 10,756 0.59
artnet AG 1 8 518 1.54 0 1,056 0.00
Jungheinrich AG 5, 7, 8, 14 3,760 45,568 8.25 8,057 391,772 2.06
CLAAS KGaA 1, 5, 6, 7, 8 380 21,497 1.77 62,730 311,641 20.13
Average (1,000 euros or %) 5,893 95,599 5.16 13,484 318,968 3.36
Accounting Total number of Accounting Total number of
difference Issues occurrences difference Issues occurrences

1 Stock-based compensation 9 9 Impairment of fixed assets 2


2 Goodwill 2 10 Impairment of intangible assets (including goodwill) 2
3 Intangible assets 2 11 Sales from development contract 1
4 Valuation 1 12 Depreciation 1
5 Income tax 3 13 Provision 1
6 Minority interest 2 14 Consolidation 2
Relative Effects of IFRS Adoption and IFRS Convergence

7 Research and development 4 15 Leases 1


8 Pension 5
611

CAR Vol. 36 No. 2 (Summer 2019)


(The table is continued on the next page.)
TABLE 8 (continued)
612

Panel B: Adoption compliance test

Applicable Percentage
Topic/subtopic Criteria tested to compliant

Stock-based compensation 7 firms


Measurement Does the entity state that equity investments are valued as the fair value of goods/services received? N/A N/A
compliance
Disclosure compliance Does the entity provide a general description of stock-based compensation arrangements? 7 100
Does the entity provide a description of determination of “fair value of goods or services received”? N/A N/A

CAR Vol. 36 No. 2 (Summer 2019)


Does the entity disclose the effect on profit/loss for the current period? 7 100
Pensions and post-employment benefits 9 firms
Measurement Does the entity state that post-employment benefits are recognized at net present value of future 7 78
compliance obligations?
Contemporary Accounting Research

Disclosure compliance Does the entity provide a general description of post-employment benefits? 9 100
Does the entity disclose methods used to determine actuarial gains/loss? 9 100
Does the entity provide a reconciliation between actual and book? 5 56
Does the entity provide reconciliation between the beginning and end-of-the-period value of the 6 67
pension obligation?
Asset impairments 13 firms
Measurement Does the entity disclose determination of the recoverable amount? 13 100
compliance Does the entity disclose use of an annual impairment test for goodwill? 12 92
Disclosure compliance Is the basis for determining “value in use” disclosed? 13 100
Is the basis for determining “fair value less cost to sell” disclosed? 13 100

Intangible assets 15 firms


Measurement The entity does NOT capitalize research, start-up, or advertising costs 15 100
compliance Does the entity directly state that it generally expenses R&D? 14 93
Disclosure compliance Is the useful life and/or amortization rate disclosed? 15 100
Is the amortization method disclosed? 15 100
Does the entity provide a reconciliation of the carrying amount of intangibles from the beginning to 15 100
the end of the period?
(The table is continued on the next page.)
TABLE 8 (continued)

Business combinations 6 firms


Measurement Does the entity disclose use of the acquisition (purchase) method? 6 100
compliance Does the entity disclose use of the recognition principle or the measurement principle in valuing 6
acquired assets and assumed liabilities? 100
Does the entity separately disclose any capitalized goodwill (recognized gain on bargain purchases)? 5 83
Disclosure compliance Does the entity disclose detailed information about the acquired companies and related accounts? 6 100
Panel C: Convergence compliance test

U.S. GAAP IFRS

Applicable Percentage Applicable Percentage


Topic/subtopic Criteria tested (based on IFRS) to compliant to compliant

Stock-based compensation 15 firms 6 firms


Measurement Does the entity state that equity investments are valued as the fair value of N/A N/A N/A N/A
compliance goods/services received?
Disclosure compliance Does the entity provide a general description of stock-based compensation 15 100 6 100
arrangements?
Does the entity provide a description of determination of “fair value of 15 100 5 83
goods or services received”?
Does the entity disclose the effect on profit/loss for the current period? 14 93 5 83

Discontinued operations 4 firms 1 firm


Measurement None
compliance
Disclosure compliance Does the entity report net profit/loss from discontinued operations on the 4 100 1 100
face of the income statement?
Does the entity disclose the net profit/loss from discontinued operations 4 100 1 100
separately from continuing operations?
Does the entity disclose the adjustments to discontinued operation in prior N/A N/A N/A N/A
Relative Effects of IFRS Adoption and IFRS Convergence

year separately?
Does the entity reclassify any included discontinued operations in income N/A N/A N/A N/A
from continuing operations when they are ceased?
613

CAR Vol. 36 No. 2 (Summer 2019)


(The table is continued on the next page.)
TABLE 8 (continued)
614

Panel C: Convergence compliance test

U.S. GAAP IFRS

Applicable Percentage Applicable Percentage


Topic/subtopic Criteria tested (based on IFRS) to compliant to compliant

Borrowing costs 0 firms 2 firms


Measurement Does the entity only consider the costs directly attributable to the acquisition, 0 0 2 100
compliance construction, and production of a qualifying asset?

CAR Vol. 36 No. 2 (Summer 2019)


Disclosure compliance Does the entity disclose the amount of borrowing costs capitalized during the 0 0 2 100
period?
Does the entity disclose the capitalization rate used? 0 0 2 100

Segmental reporting 12 firms 13 firms


Contemporary Accounting Research

Measurement Does the entity disclose the management decision approach? 4 33 13 100
compliance
Disclosure compliance Does the entity disclose how operating segments are identified? 12 100 13 100
Does the entity disclose the judgments made by management in applying the 4 33 1 8
aggregation criteria?
Does the entity disclose the profit or loss for each reportable segment? 12 100 13 100
Does the entity disclose the total assets and liabilities for each reportable 12 100 12 92
segment?
Does the entity provide information about the measurement of segment profit 3 25 5 38
or loss, segment assets, and segment liabilities?
Does the entity provide a reconciliation of the totals of segment revenues, 6 50 9 69
profit or loss, assets, and liabilities?
Does the entity provide information about each product and service or 11 92 13 100
groups of products and services?
Does the entity provide analyses of revenues and certain non-current assets 10 83 13 100
by geographical areas?
Does the entity provide information about transactions with major 6 50 6 46
customers?
(The table is continued on the next page.)
TABLE 8 (continued)

Panel C: Convergence compliance test

U.S. GAAP IFRS

Applicable Percentage Applicable Percentage


Topic/subtopic Criteria tested (based on IFRS) to compliant to compliant

Business combinations 3 firms 2 firms


Measurement Does the entity disclose use of the acquisition (purchase) method? 3 100 0 0
compliance Does the entity disclose use of the recognition principle or the measurement 3 100 2
principle in valuing acquired assets and assumed liabilities? 100
Does the entity separately disclose any capitalized goodwill (recognized gain 3 100 1 50
on bargain purchases)?
Disclosure compliance Does the entity disclose detailed information about the acquired companies 3 100 2 100
and related accounts?
Does the entity report any follow-up information related to adjustments of N/A N/A 1 50
past business combinations?
Notes: This table summarizes our findings from three validation tests. In panel A, we examine the impact of the accounting differences between U.S. GAAP and IFRS
on firms’ net income and equity. The data was hand-collected from the U.S. GAAP-to-IFRS reconciliation disclosure as reported by German U.S. GAAP firms in
2005. In panel B, using the disclosure information from 2006 financial statements, we examine the compliance rates of individual measurement and disclosure
requirements of five IFRS that are related to the major accounting differences between IFRS and U.S. GAAP in 2005. Each question or criterion listed is based on the
IASB’s listed measurement and disclosure requirements for each individual accounting standard. Similarly, in panel C, we examine the compliance rates of individual
measurement and disclosure requirements of those converged accounting standards (i.e., converged U.S. GAAP for U.S. firms and converged IFRS for German IFRS
firms). Each question or criterion listed is based on the IASB’s listed measurement and disclosure requirements for each individual accounting standard. In both panels
B and C, the compliance rate is calculated by number of firms that have complied with a specific measurement or disclosure requirement divided by the total number
of firms to which each individual accounting standard is applicable. N/A indicates that a specific measurement or disclosure requirement does not apply to the firms.
Relative Effects of IFRS Adoption and IFRS Convergence
615

CAR Vol. 36 No. 2 (Summer 2019)


616 Contemporary Accounting Research

7. Conclusion
This study is motivated by the prevalent use of either the adoption or convergence approach when
countries aligned their domestic accounting standards with IFRS over the last decade. We identify
a unique setting in Germany to facilitate our investigation on whether adoption and convergence,
respectively, increase comparability. More importantly, we examine whether the effects of adop-
tion and convergence on comparability differ.
Using various comparability metrics, we find that both adoption and convergence increase
comparability after the new enforcement regulation in 2005. The results are robust after control-
ling for the enforcement effect except for the changes in comparability due to convergence.
Employing a difference-in-differences design, our results generally suggest that the effects of
adoption and convergence on comparability are not statistically different. Our findings are robust
to a number of alternative empirical specifications and are complemented by an investigation of
the compliance rates among our sample firms. The findings of this study should be of interest to
regulators and standard setters as they continue to debate alternative approaches to achieving
accounting standard globalization.
As a caveat, our results should not be interpreted as directly applicable to the potential adop-
tion of IFRS by the United States. It is unclear whether these results would be observed should
U.S. firms be permitted to adopt IFRS or if U.S. GAAP are fully converged to IFRS. Moreover,
increased financial statement comparability is only one outcome of accounting standard globaliza-
tion, and the decision between adoption and convergence within the United States is ultimately
driven by a number of factors. Future research may seek to evaluate whether differences between
the effects of adoption and convergence have any implication for other qualitative characteristics
of accounting information, such as value relevance. As many countries embrace a global frame-
work for accounting standards, evidence will certainly be useful in evaluating the success of such
regulatory efforts, as well as the corresponding benefits to capital markets.

References
Ahmed, A. S., M. Neel, and D. Wang. 2013. Does mandatory adoption of IFRS improve accounting quality?
Preliminary evidence. Contemporary Accounting Research 30 (4): 1344–72.
Ball, R. 2006. International financial reporting standards (IFRS): Pros and cons for investors. Accounting
and Business Research 36 (1): 5–27.
Ball, R., and L. Shivakumar. 2005. Earnings quality in U.K. private firms. Journal of Accounting and Eco-
nomics 39 (1): 83–128.
Barth, M. E., W. R. Landsman, M. H. Lang, and C. Williams. 2012. Are IFRS-based and U.S. GAAP-based
amounts comparable? Journal of Accounting and Economics 54 (1): 68–93.
Bradshaw, M., C. Callahan, J. Ciesielski, E. A. Gordon, L. Hodder, P. E. Hopkins, M. Kohlbeck, R. Laux,
S. McVay, T. Stober, P. Stocken, and T. L. Yohn. 2010. Response to the SEC’s proposed rule—
Roadmap for the potential use of financial statements prepared in accordance with International Finan-
cial Reporting Standards (IFRS) by U.S. issuers. Accounting Horizons 24 (1): 117–28.
Brochet, F., A. Jagolinzer, and E. J. Riedl. 2013. Mandatory IFRS adoption and financial statement compara-
bility. Contemporary Accounting Research 30 (4): 1373–400.
Cascino, S., and J. Gassen. 2015. What drives the comparability effect of mandatory IFRS adoption? Review
of Accounting Studies 20 (1): 1–41.
Christensen, H., L. Hail, and C. Leuz. 2013. Mandatory IFRS reporting and changes in enforcement. Journal
of Accounting and Economics 56 (2): 147–77.
Christensen, H., E. Lee, M. Walker, and C. Zheng. 2015. Incentives or standards: What determines account-
ing quality changes around IFRS adoption? European Accounting Review 24 (1): 31–61.
Cram, D., V. Karan, and I. Stuart. 2009. Three threats to validity of choice-based and matched sample stud-
ies in accounting research. Contemporary Accounting Research 26 (2): 477–516.

CAR Vol. 36 No. 2 (Summer 2019)


Relative Effects of IFRS Adoption and IFRS Convergence 617

Daske, H., L. Hail, C. Leuz, and R. Verdi. 2013. Adopting a label: Heterogeneity in the economic conse-
quences around IAS/IFRS adoption. Journal of Accounting Research 51 (3): 495–547.
Dechow, P. M. 1994. Accounting earnings and cash flows as measures of firm performance: The role of
accounting accruals. Journal of Accounting and Economics 18 (1): 3–42.
De Franco, G., S. P. Kothari, and R. S. Verdi. 2011. The benefits of financial statement comparability. Jour-
nal of Accounting Research 49 (4): 895–931.
FASB. 2010. Chapter 1: The objective of general purpose financial reporting, and Chapter 3: Qualitative
characteristics of useful financial information. In Statement of Financial Accounting Concepts No. 8:
Conceptual Framework for Financial Reporting. Norwalk, CT: FASB.
FASB. 2014. Comparability in international accounting standards—An overview. Available at http://www.fasb
.org/jsp/FASB/Page/SectionPage&cid=1176156245663, accessed September 2016.
FASB and IASB. 2002. Memorandum of Understanding (“The Norwalk Agreement”). Norwalk, CT: Finan-
cial Accounting Standards Board.
Hayn, C. 1995. The information content of losses. Journal of Accounting and Economics 20 (2): 125–53.
Henry, E., S. Lin, and Y. W. Yang. 2009. The European-US GAAP gap: IFRS to US GAAP Form 20-F rec-
onciliations. Accounting Horizons 23 (2): 121–50.
Holthausen, R. W. 2009. Accounting standards, financial reporting outcomes, and enforcement. Journal of
Accounting Research 47 (2): 447–58.
IASB. 2010. The Conceptual Framework for Financial Reporting. IASB: London.
Jamal, K., S. Penman, R. Bloomfield, T. E. Christensen, R. Colson, J. A. O. Moehrle, G. Previts, T. Stober,
S. Sunder, and R. L. Watts. 2010. A framework for financial reporting standards: Issues and a suggested
model. Accounting Horizons 24 (3): 471–85.
Lang, M. H., M. G. Maffett, and E. L. Owens. 2010. Earnings comovement and accounting comparability:
The effects of mandatory IFRS adoption. Working paper, University of Rochester.
Leuz, C. 2003. IAS versus U.S. GAAP: Information asymmetry-based evidence from Germany’s New
Market. Journal of Accounting Research 41 (3): 445–72.
Lin, S., W. N. Riccardi, and C. Wang. 2012. Does accounting quality change following a switch from
U.S. GAAP to IFRS? Evidence from Germany. Journal of Accounting and Public Policy 31 (6): 641–57.
Nobes, C. 2006. Survival of international accounting differences under IFRS: Towards a new research
agenda. Accounting and Business Research 36 (3): 233–45.
Nobes, C. 2011. International variations of IFRS adoption and practice. London: Certified Accountants Educa-
tional Trust.
Schipper, K. 2003. The introduction of international accounting standards in Europe: Implications for inter-
national convergence. European Accounting Review 14 (1): 101–26.
SEC. 2008. Roadmap for the potential use of financial statements prepared in accordance with International
Financial Reporting Standards by U.S. issuers. Release No. 33-8982; 34-58960.
SEC. 2012. Work plan for the consideration of incorporating International Financial Reporting Standards
into the financial reporting system for U.S. issuers. Final Staff Report (July 13). Washington, DC:
Government Printing Office.
White, H. 1980. A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroske-
dasticity. Econometrica 48 (4): 817–38.
Yip, R., and R. Young. 2012. Does mandatory IFRS adoption improve information comparability? The
Accounting Review 87 (5): 1767–89.

SUPPORTING INFORMATION
Additional Supporting Information may be found in the online version of this article:
Appendix A: Summary of revisions to accounting standards over sample period
Appendix B: The list of firms used in this study

CAR Vol. 36 No. 2 (Summer 2019)


618 Contemporary Accounting Research

Discussion of “Relative Effects of IFRS Adoption and IFRS


Convergence on Financial Statement Comparability”*

PATRICK E. HOPKINS, Indiana University†

1. Introduction
Lin, Riccardi, and Wang (2019) (hereafter LRW) exploit the 2005 mandated adoption of IFRS in
Germany as a setting for a quasi-experimental, cross-sectional, archival investigation of the rela-
tive effects on financial statement comparability of (i) a “single-shot” adoption of IFRS in
Germany, where some firms were previously using U.S. GAAP, versus (ii) the incremental con-
vergence between GAAP and IFRS that was coincidentally occurring in the United States. To
measure adoption effects, LRW identify 47 publicly traded German firms that switched from
U.S. GAAP to IFRS in 2005 (“German U.S. GAAP firms”) and match them (i.e., based on
2004’s 2-digit SIC code) with German firms that applied IFRS throughout the 2002–2010 sample
period (“German IFRS firms”). To measure convergence effects, LRW take the previously identi-
fied German IFRS firms and match them to publicly traded U.S. domestic firms that applied
U.S. GAAP throughout the 2002–2010 sample period (i.e., “U.S. GAAP firms”).
For each of these two sets of pairwise matched firms, LRW compute three different pair-
wise proxies for financial reporting comparability: COMP1, an earnings on returns regression-
type proxy popularized by DeFranco et al. (2011); COMP2, a returns on earnings regression-type
proxy proposed by Barth et al. (2012); and COMP3, an accruals on cash flows regression-type
proxy used by Cascino and Gassen (2015). For each of these comparability measures, LRW pool
2002–2010 data for the 47 firms, and use the now ubiquitous difference-in-differences design to
measure the effects of a sudden, structural shift from U.S. GAAP to IFRS (i.e., the COMP measures
for the pairs of German U.S. GAAP firms and German IFRS firms) versus the incremental conver-
gence between U.S. GAAP and IFRS (i.e., the COMP measures for the pairs of German IFRS firms
and U.S. GAAP firms). LRW interpret their results as suggesting that both adoption and conver-
gence are associated with an increasing trend in each of the COMP measures during the 2002
through 2010 time period; however, these slopes (mostly) are not statistically different between
the pre-2005 and the post-2005 time period, suggesting that adoption and convergence are associ-
ated with similar trends in comparability (i.e., there are no statistically reliable difference in the
differences).
This discussion is based on the remarks I made during the 2016 Contemporary Accounting
Research Conference at the University of Waterloo, and proceeds as follows: First, given the
IASB’s sponsorship of part of the 2016 conference, I describe where the enhancing qualitative
characteristic, comparability, fits into the overall IASB Conceptual Framework.1 Next, I provide a
high-level discussion of the paper’s research design, with an eye toward evaluating construct
validity across comparability-related research studies. Finally, I make suggestions for improving
comparability-related inferences in future research studies.

* Accepted by Michael Welker. This discussion is based on my remarks at the 2016 Contemporary Accounting Research
conference, generously supported by the Chartered Professional Accountants of Canada. I thank Michael Welker for the
opportunity to discuss the manuscript, and Spencer Anderson, Leslie Hodder, and Amanda Zawadski for their extremely
valuable input. I also thank the Kelley School of Business and SungKyunKwan University for financial support.
† Corresponding author.
1. For the remainder of this discussion, all references to the “Conceptual Framework” relate to the IASB’s version.

CAR Vol. 36 No. 2 (Summer 2019)


Discussion of “Relative Effects of IFRS Adoption and IFRS Convergence” 619

2. Comparability in the Conceptual Framework


The Conceptual Framework is a nonbinding, aspirational statement that describes the objectives,
characteristics, and structural details that should be embodied in general purpose financial statements
prepared for external users. As an aspirational statement, the Conceptual Framework is primarily
intended to guide standard-setting-related choices made by accounting standard setters. In addition,
when accounting standards yield unambiguous accounting treatments for transactions and events, the
Conceptual Framework is part of the third level of the accounting hierarchy and is supposed to guide
the applications of accounting standards by preparers, the evaluation by auditors of those applica-
tions, and the interpretation of financial statements by users (IASB 2010).2 Thus, the evaluation of
actual financial reporting outcomes in the context of the Conceptual Framework is a worthwhile
research goal and should be informative to the IASB and other standard-setting bodies.
As noted in the Conceptual Framework, “the objective of general purpose financial reporting
is to provide financial information about the reporting entity that is useful to existing and poten-
tial investors, lenders and other creditors in making decisions about providing resources to the
entity” (IASB 2010, OB2). The IASB decided that useful financial reporting information should
have two “fundamental qualitative characteristics”: relevance and faithful representation. Rele-
vance is defined as the ability of financial information to make a difference in the decisions of
financial statement users (IASB 2010, QC6) and includes the ability of financial information to
assist in financial prediction and in feedback or reflection. Faithful representation is a state that is
achieved when financial information is complete, neutral, and free from error (IASB 2010,
QC12). Conceptually, the two fundamental qualitative characteristics are independent; that is, in
theory, a relevant piece of financial information could be depicted in an incomplete, biased, and
error-prone manner, while the converse could also be true. Decision-useful financial information
will maximize the joint product of relevance and faithful representation, and does not favor one
fundamental qualitative characteristic over the other (IASB 2010, QC17).
Comparability, verifiability, timeliness, and understandability are “enhancing qualitative char-
acteristics” and are only considered after relevance and faithful representation are jointly maxi-
mized (IASB 2010, QC19). None of the enhancing qualitative characteristics takes priority over
the others, and the emphasis of one or the other can be contextual (e.g., upon adoption of a new
accounting standard, comparability may be initially diminished but then enhanced in later periods).
The focus of LRW is comparability, defined by the IASB as “the qualitative characteristic
that enables users to identify and understand similarities in, and differences among, items” (IASB
2010, QC21). Comparability is different from uniformity because comparability means “like
things must look alike and different things must look different” (IASB 2010, QC23). The IASB
also notes that while relevant economic phenomena can be faithfully represented in many ways,
allowing “alternative accounting methods for the same economic phenomenon diminishes compa-
rability” (IASB 2010, QC25). Finally, the IASB notes that noncomparable financial reporting can
be partially remedied through appropriate disclosure (IASB 2010, QC34).
The compensatory relation between noncomparable reporting and disclosure can be illus-
trated via lessee accounting under International Accounting Standard (IAS) 17, which requires
capitalization and balance sheet recognition of implicit assets and liabilities for financing leases
but, for operating leases, requires off-balance sheet treatment and expense recognition for the
lease payments.3 This accounting results in noncomparable balance sheets and income statements
for lessees because it takes similar obligations for future lease payments, and treats one as an on-
balance sheet liability (i.e., with periodic interest charges run through the performance statement)
and the other as an unrecognized contingency that generates performance-statement recognition

2. In contrast, the FASB’s Conceptual Framework is not part of the GAAP hierarchy.
3. IFRS 16 will supersede IAS 17, is effective for annual reporting periods beginning after January 1, 2019, and gen-
erally requires lessees to apply financing-lease treatment to most leases (i.e., except for leases of less than 12 months
and leases for low-value assets). This treatment should result in increased comparability for lease obligations.

CAR Vol. 36 No. 2 (Summer 2019)


620 Contemporary Accounting Research

of the lease payments. However, IAS 17 also requires companies to disclose minimum future
lease payments for operating leases, and this information can be used to compute implicit lease
obligations by financial statement users. Bratten et al. (2013) investigate a similar capitalize-ver-
sus-expense, lessee accounting regime in the United States, and find that investors equally
impound into cost of capital firms’ off-balance sheet “as-if” computed operating lease obligations
and firms’ recognized capital lease obligations. Their findings suggest that, under certain condi-
tions (e.g., when disclosures are salient, are not based on management estimates, and allow basic
techniques for imputing as-if recognized amounts), disclosure can mitigate the deleterious effects
of noncomparable reporting.
This discussion of the Conceptual Framework has three important implications for the inter-
pretation of LRW. First, comparability is one of four second-order characteristics that can influ-
ence the usefulness of financial accounting information after the joint and primary characteristics
of relevance and faithful representation are maximized. If, for example, an accounting standard is
issued that changes from a single, less relevant accounting treatment (i.e., accounting is bad, but
comparable) to a single, more relevant accounting treatment (i.e., accounting is better, and still
comparable), then the decision usefulness of accounting information will increase without any of
that improvement in decision usefulness coming from changes in comparability. Second, the deci-
sion usefulness of financial reporting can be improved without improving comparability if the
IASB issues standards that provide enhanced disclosures that allow financial statement users to
“work around” the incomparability. Third, all six of the fundamental and enhancing qualitative
characteristics influence the decision usefulness of accounting information. Therefore, researchers
interested in investigating individual qualitative characteristics must be careful that their operatio-
nalized proxies capture the qualitative characteristic of interest, to the exclusion of the other five
(i.e., construct validity is sufficiently high). I focus on construct validity in the next section.

3. Thoughts on construct validity


Figure 1 provides a stylized representation of LRW’s research design.4 The top two boxes (i.e., linked
by arrow 1) represent the unobservable constructs that the study intends to investigate, and capture an
unambiguously causal relationship between the constructs. These boxes should be based on theory,
which is then used to develop the study’s hypotheses. The bottom three boxes represent the things the
researchers actually did in the study. The left two boxes on the bottom of Figure 1 (i.e., linked by
arrow 4) represent the primary operationalization that tests the theoretical constructs in the top row.
The bottom right-most box is extremely important in archival settings because typical archival
research designs do not allow for randomization of subjects (e.g., firms) to quasi-experimental treat-
ment. Arrows 2 and 3 capture elements of construct validity and external validity, while arrows 4 and
5 capture elements of internal validity and statistical conclusion validity.
Because they have implications for studies beyond LRW’s, I focus my research design-
related remarks on construct validity.5 A necessary condition for an observed proxy to possess
high levels of construct validity is for that proxy to covary with the construct of interest and to be
orthogonal with all other potential constructs. In Figure 1, the correspondence represented by
arrow 2 reflects the paper’s claim that splitting the world into pre-2005 and post-2005 categories
is a sufficiently good proxy to capture the nature of change in accounting principles (i.e., mandated
versus converged) in Germany. In addition, implicit in this operationalization is a second claim: that
the pre–post-2005 split is sufficiently diagnostic that it captures only the causal agent in mandated

4. Figure 1 illustrates the Predictive Validity Framework proposed by Runkel and McGrath (1972, 160), and intro-
duced to the accounting literature by Bob Libby in his discussion of an article by Bob Ashton (Libby 1976, 19) and
in his seminal monograph (Libby 1981, 11).
5. The issues in LRW related to internal validity and statistical conclusion validity are fairly standard (e.g., self-selection)
and should be obvious to most readers.

CAR Vol. 36 No. 2 (Summer 2019)


Discussion of “Relative Effects of IFRS Adoption and IFRS Convergence” 621

Figure 1 LRW’s research design

Concept A Concept B

Nature of change in 1 Differential changes in


accounting principles financial statement
(i.e., mandated versus comparability
converged)

2 3

Operational Operational Moderating


definition A definition B variables
4
Difference in COMP for 5 Pairwise controls for
Pre/post-2005 in matched pairs of ratio of firms’ MV,
Germany versus U.S. G-GAAP & G-IFRS Sales growth, M/B,
versus leverage & industry
G-IFRS & U.S. Firms (self-selection)

versus converged changes in accounting principles, to the exclusion of all other potential influential
factors. This is a rather bold assumption given the complexity of economies, cultures, and societies.
By using a fixed date as the operational treatment in the study, LRW run the risk of spurious
correlation with other constructs (i.e., it has low construct validity). For example, using other events
occurring in 2005, the results of the study are sufficient for us to conclude that the association
between earnings and returns was influenced by Angela Merkel’s election to the German chancel-
lorship, Hurricane Katrina’s devastation of the Gulf coast of the United States, and the death of
Pope John Paul II. While this conclusion might seem absurd, these events do all share the same
statistical association with the year 2005 as the mandated German switch to IFRS.
Close inspection of the trends illustrated in LRW’s Figure 1 indicate that something defi-
nitely happened during the 2003–2006 time period. However, this pattern does not make sense
when one considers the nature of sudden change versus gradual convergence. If we focus on the
pattern of COMP1 means (i.e., the upper left graph), we notice that the average level of compara-
bility for the adopted and converged pairs of firms were identical in 2006. Does this suggest that
the mandated adoption of IFRS in Germany (or the election of Angela Merkel as German chan-
cellor) in 2005 caused U.S. GAAP to become converged with IFRS? Does this make sense given
the different underlying processes of adoption versus convergence? This pattern causes me to
infer that something else is going on during this time period, and whatever it is appears to be
highly correlated with the year 2005 (and everything that happened in that year). I believe we
would learn much more about standard-setting institutions if we took the time and effort to
exactly define what is embodied in those institutions. With respect to LRW, this means instead of
relying on a coarse, fixed-date proxy for the treatment, we would learn much more if we
attempted to design proxies that capture the complex dimensions of the actual treatment (i.e., in
this case, the key features of mandated shifts in standards versus convergence).
Next, I consider the construct validity of the relation represented by arrow 3 in Figure 1;
specifically, do the comparability proxies (i.e., COMP1, COMP2, and COMP3) capture the IFRS’s
concept of comparability to the exclusion of other qualitative characteristics and to the exclusion
of other correlated, omitted factors? Because of space limitations, I will constrain my remarks to
COMP1 because it is based on the most common comparability proxy in the extant literature;
however, the spirit of these remarks also applies to COMP2 and COMP3.
COMP1 is derived from an output-based measure of interfirm earnings–return similarity
proposed in De Franco et al. (2011). De Franco et al.’s (2011) proxy is based on the following

CAR Vol. 36 No. 2 (Summer 2019)


622 Contemporary Accounting Research

theoretical relation between financial statements and the economic events and transactions
affecting firm i:

Financial Statementsi = fi ðEconomic Eventsi Þ: ð1Þ

Who can argue with equation (1)? The function, f (•), translates the real underlying economic
events (i.e., including transactions) of the firm into financial statements through a set of accounting
processes and procedures. If the accounting between two firms is comparable, then, given a set of
identical economic events, one should observe identical financial statements. Of course, the trick
here is in the definitions of financial statements and economic events. Consistent with De Franco
et al. (2011), LRW proxy for the entire package of financial statements and related note disclosures
using NIi ,t /Pi ,t −1 (i.e., the inverse of the price-earnings ratio) and the entire set of economic events
and transactions that affect the firm using (Pi,t −Pi,t −1)/Pi,t −1 (i.e., contemporaneous returns),
resulting in the following relation:

NI i, t =Pi, t −1 = α + β½ðPi, t – Pi, t − 1 Þ=Pi, t − 1  + ε: ð2Þ

To construct COMP1, LRW include two estimations for the α^ and β^ parameters from run-
ning, separately for each firm in the pair, regressions based on equation (2): pre-adoption
(2002–2004) and post-adoption (2006–2010). The parameters are used to generate the expected
level of NI/P (based on the actual returns) in each reporting period for each of the firms in the
pair. For each pair of firms in each period, these expected levels of NI/P are differenced, the abso-
lute value for each is computed, and then are averaged. Because LRW use semiannual earnings
data, in the pre (post) period, each pair of firms generates 6 (10) within-sample COMP1 observa-
tions based on the identical α^ and β^ parameters, for a total of 16 observations across the entire
panel.6
The pairwise nature of COMP1 is the basis for using it as a proxy for comparability
(i.e., comparability implicitly requires at least two firms). However, from a construct validity
perspective, the more important attribute is that COMP1 is based on the earnings-returns
relation summarized in equation (2). Regardless of direction (i.e., Returns = f(Earnings) or
Earnings = f(Returns)), this relation has likely been the subject of more empirical estimation
and investigation than any other set of accounting and market outputs. For example, varia-
tions on this relation are the basis for Ball and Brown (1968), the vast literature on earnings
response coefficients (e.g., Collins and Kothari 1989), common proxies for accounting con-
servatism (e.g., Basu 1997), and, since 2011, proxies for financial reporting comparability.
The fact that this relation has been used in so many settings is either testament to its
versatility or an indication of its overuse. In the context of comparability, we would benefit
from a bit more thought on the economic and financial reporting characteristics captured by
the earnings-returns relation.
One aspect of the earnings-returns relation that is ignored by LRW, De Franco et al.
(2011), and others is the low proportion of contemporaneous returns-based economic infor-
mation that is captured by earnings. For example, De Franco et al. (2011) report a median
R2 = 7 percent and Basu (1997) reports 10 percent R2 when a negative-return indicator vari-
able is interacted with returns. Other studies have suggested that the low explanatory power

6. Although further in the weeds than I’d like to go with this discussion, I do have the following specification con-
cerns with LRW’s operationalization of COMP1: (i) differences in volatility of pre and post returns will mechani-
cally generate differences in pre and post COMP1 values, (ii) all computed E/P values (and related differences) are
incorporated into the COMP1 proxy as if they are measured without error, and (iii) all computed E/P values (and
related differences) are incorporated into the COMP1 proxy as if they are independent.

CAR Vol. 36 No. 2 (Summer 2019)


Discussion of “Relative Effects of IFRS Adoption and IFRS Convergence” 623

is an artifact of the low timeliness of when economic information makes its way through
accrual accounting systems into earnings. For example, Ryan and Zarowin (2003) find that a
regression of contemporaneous earnings on returns has a 5 percent R2, but that a regression
also including three lagged periods of returns triples the R2 to 16 percent. This result suggests
that the timeliness (i.e., another enhancing qualitative characteristic) of accounting treatments
can affect the earnings-returns relation. Further, hypothetically, if different standard setters
(e.g., IASB and FASB) issue standards that result in different accounting treatments, but that
both improve timeliness, then the earnings-returns relation would change and the pairwise
correlation between industry-matched firms in the two regimes would improve; however,
comparability would be low.
The potential issues with COMP1’s construct validity extend beyond the equivocal out-
comes of improved comparability and improved timeliness. As their primary goal, the FASB
and the IASB are supposed to promulgate standards that increase relevance and faithful repre-
sentation. Holding comparability constant, would one expect more relevant information to
have an impact on the earnings-returns relation and the co-movement of that relation between
firms? What about enhancements in the understandability (i.e., another enhancing qualitative char-
acteristic) of reported information? The single most important question that must be answered when
considering LRW’s research design is why, given the two primary qualitative characteristics and
the four equally important enhancing characteristics, does a pairwise statistical proxy based on the
earnings-returns relation best exemplify comparability in isolation when one would expect other
qualitative characteristics to also affect the earnings-returns relation? Ultimately, a satisfactory
answer to this question is necessary for accounting researchers and standard setters to consider the
results of LRW in making meaningful inferences about financial reporting comparability. In the
next section, I provide suggestions for how accounting research can provide useful insights related
to financial reporting comparability.

4. Suggestions for comparability-themed research


During the last half century, archival accounting research has placed tremendous pressure on
earnings—a highly aggregated, ill-defined summary statistic—to reveal subtle attributes about
financial reporting and capital markets institutions.7 LRW continue this practice by using pairwise
correlations between earnings and returns to study an important, but second-order, attribute of
financial reporting: comparability. While I admire LRW’s attempt, I am not convinced the manu-
script provides us with any better understanding of the effects on comparability of adoption versus
convergence. If we wish to make real progress in understanding financial reporting comparability,
then we should think hard about how comparability effects would be manifest in the complete
financial reporting package.
Ultimately, comparability is an attribute of the total set of information included in general
purpose financial statements, including both the content and its form. The Conceptual Framework
includes at least four areas that determine the content and form of general purpose financial
statements:

Recognition
Does a particular transaction, event or condition suggest an item satisfies the definition of a finan-
cial statement element (e.g., liability versus disclosed contingency)?

7. Earnings (i.e., net income) is not defined in the Conceptual Framework. The IASB states that “profit or loss, total
OCI and total comprehensive income are not elements of financial statements. They are subtotals or totals derived
by summing items of income or expense” (IASB 2013, 2:13). Although the FASB’s Conceptual Framework
includes comprehensive income as an element, it does not define earnings. Thus, under both conceptual frame-
works, earnings is the aggregation and netting of those things identified as revenues, expenses, gains, and losses.

CAR Vol. 36 No. 2 (Summer 2019)


624 Contemporary Accounting Research

Measurement
Conditional on recognizing a financial statement element, how should firms measure the financial
statement element (cost accumulation versus fair value)?

Presentation
Conditional on recognizing a financial statement element, how should firms present the element
(e.g., in earnings or comprehensive income)?

Disclosure
Given noncomparable recognition, measurement or presentation, can standard setters provide other
information that mitigates the noncomparability (e.g., supplemental disclosure for operating leases)?
The first three of these concepts—recognition, measurement, and presentation—are attri-
butes of the primary financial statements and are the source of what one might define as financial
statement comparability. The fourth concept—disclosure—is outside the three primary financial
statements and is a way for standard setters to mitigate problems with noncomparable primary finan-
cial statements.
Given that comparability is a multidimensional attribute of financial reporting, then an indi-
vidual interested in understanding comparability would be well-served to consider the exact
source of comparability. In the context of LRW, what are the adopted and/or converged standards
during 2002 through 2010 that caused changes in recognition, presentation, or measurement? Did
these changes affect the timing of when specific components of earnings were recognized? Did
these changes affect the way earnings were measured? Was the salience of earnings and its com-
ponents affected by changes in presentation? How would those changes be manifest in COMP?
Alternatively, if these attributes remained static, did changes in accounting disclosure result in
changes in the earnings-returns relation and in COMP?
If we extend beyond the proxies included in LRW, our understanding of the determinants
of comparability will be enhanced if we explicitly make states of, and changes in, recogni-
tion, measurement, presentation, and disclosure the subjects of our investigations. By isolat-
ing and separately investigating each of these features of accounting standards, we can
greatly increase the power of our tests and provide inferences that have a higher likelihood of
providing useful information to accounting researchers and standard setters. Can we leverage
newer text-scraping technologies to identify and catalog the sources of changes in compara-
bility (e.g., specific accounting standards) and the reporting outcomes of those changes? Can
we identify specific financial statement line items (e.g., revenues) that should be affected? We
can do much more in attempting to isolate comparability to the exclusion of other financial
reporting attributes (e.g., timeliness), and these efforts would do well to rely less on the
earnings-returns relation.
Finally, the comparability-related studies can be greatly improved if researchers attempt to explic-
itly investigate the attributes that separate accounting comparability from uniformity: “Like things
must look alike and different things must look different” (QC23). Implicitly, by operationalizing
comparability-related proxies based on the pricing of earnings across pairs of companies, extant
research has primarily focused on the characteristics of earnings that cause like things to look alike.
Alternatively, one could also argue that these studies investigate settings in which accounting is more
uniform. Accounting standard setters and researchers would benefit from studies that explicitly con-
sider cases in which “different things are supposed to look different” (e.g., Anderson 2017; Yip and
Young 2012).

5. Conclusion
LRW exploit the 2005 mandated adoption of IFRS in Germany to investigate the relative effects
on financial statement comparability of all-at-once adoption versus the incremental convergence

CAR Vol. 36 No. 2 (Summer 2019)


Discussion of “Relative Effects of IFRS Adoption and IFRS Convergence” 625

between GAAP and IFRS that was coincidentally occurring in the United States. In this discus-
sion, I identify comparability as one of four enhancing qualitative characteristics that are second-
ary to the fundamental qualitative characteristics, relevance, and representational faithfulness.
Next, I discuss the construct validity of LRW’s treatment and outcome proxies and describe how
the former is equivocal to many other events that occurred in 2005 and that the latter likely does
not have sufficiently high discriminant validity to meaningfully differentiate it from other qualita-
tive characteristics, like relevance, timeliness, or understandability. Thus, any inferences about
comparability are tenuous, at best. I conclude this discussion with suggestions for ways to
improve the empirical study of comparability.

References
Anderson, S. E. 2017. The effect of financial statement incomparability and notes labeling on investment
decisions. Working paper, University of Illinois – Urbana-Champaign.
Ball, R., and P. Brown. 1968. An empirical evaluation of accounting income numbers. Journal of Account-
ing Research 6 (2): 159–78.
Barth, M. E., W. R. Landsman, M. H. Lang, and C. Williams. 2012. Are IFRS-based and U.S. GAAP-based
amounts comparable? Journal of Accounting and Economics 54 (1): 68–93.
Basu, S. 1997. The conservatism principle and the asymmetric timeliness of earnings. Journal of Accounting
and Economics 24 (1): 3–37, https://doi.org/10.1016/S0165-4101(97)00014-1
Bratten, B., P. Choudhary, and K. Schipper. 2013. Evidence that market participants assess recognized and
disclosed items similarly when reliability is not an issue. The Accounting Review 88 (4): 1179–210,
https://doi.org/10.2308/accr-50421
Cascino, S., and J. Gassen. 2015. What drives the comparability effect of mandatory IFRS adoption? Review
of Accounting Studies 20 (1): 1–41.
Collins, D. W., and S. P. Kothari. 1989. An analysis of intertemporal and cross-sectional determinants of
earnings response coefficients. Journal of Accounting and Economics 11 (2–3): 143–81, https://doi.org
/10.1016/0165-4101(89)90004-9
De Franco, G., S. P. Kothari, and R. S. Verdi. 2011. The benefits of financial statement comparability. Jour-
nal of Accounting Research 49 (4): 895–931, https://doi.org/10.1111/j.1475-679X.2011.00415.x
IASB. 2010. Conceptual Framework for Financial Reporting. Available at http://eifrs.ifrs.org/, retrieved June 2017.
IASB. 2013. Discussion Paper DP/2013/1: A Review of the Conceptual Framework for Financial Reporting.
http://archive.ifrs.org/Current-Projects/IASB-Projects/Conceptual-Framework/Discussion-Paper-July-2013
/Documents/Discussion-Paper-Conceptual-Framework-July-2013.pdf
Libby, R. 1976. Discussion of cognitive changes induced by accounting changes: Experimental evidence on the
functional fixation hypothesis. Journal of Accounting Research 14 (3): 18–24, https://doi.org/10.2307/2490444
Libby, R. 1981. Accounting and human information processing: Theory and applications. Englewood Cliffs,
NJ: Prentice-Hall.
Lin, S., W. N. Riccardi, and C. Wang. 2019. Relative effects of IFRS adoption and IFRS convergence on financial
statement comparability. Contemporary Accounting Research, this issue.
Runkel, P. J., and J. E. McGrath. 1972. Research on human behavior: A systematic guide to method.
New York: Holt, Rinehart and Winston.
Ryan, S. G., and P. A. Zarowin. 2003. Why has the contemporaneous linear returns-earnings relation
declined? The Accounting Review 78 (2): 523–53, https://doi.org/10.2308/accr.2003.78.2.523
Yip, R. W. Y., and D. Young. 2012. Does mandatory IFRS adoption improve information comparability?
The Accounting Review 87 (5): 1767–89, https://doi.org/10.2308/accr-50192

CAR Vol. 36 No. 2 (Summer 2019)


626 Contemporary Accounting Research

Discussion of “Relative Effects of IFRS Adoption and IFRS


Convergence on Financial Statement Comparability”*

GARY KABURECK, International Accounting Standards Board (IASB)†

1. Introduction
From my point of view as a standard setter, I would summarize what the paper (Lin, Riccardi,
and Wang 2019) does as making a contribution to understanding the relative utility of adoption
versus convergence. My comments start with some context from the Conceptual Framework. I
then make some observations on the paper, look at its implications and its relationship to planned
IASB activities, and, finally, make some suggestions for future research.

2. The Conceptual Framework


Patrick Hopkins’s discussion (Hopkins 2019) covers a lot of the Conceptual Framework background.
Comparability is an enhancing characteristic that helps us see similarities and differences more
clearly. It is often confused with consistency, but consistency is essentially a tool to help us achieve
comparability. An auditor’s report, for example, does not say that the financial statements are “com-
parable;” it says that they are “consistent.” Doing the same things in the same way is consistency.
Comparability is a goal set in order to help financial analysts, for example, draw meaningful conclu-
sions between companies within a year, between companies within an industry, and so on.
At the same time, the Conceptual Framework makes the point that comparability does not
equal uniformity. Effectively it means standards and financial reports have to be “close enough”
to draw comparisons and that any differences that exist do not affect the analysts’ conclusions.

3. Implications of the paper


The paper’s findings comport with what I would have intuitively expected, so it provides confir-
matory value. It tells us that both methods—adoption and convergence—improve the information
content and enhance the comparability of affected financial statements. It supports the assumption
that sometimes convergence, when there is enough of it, is sufficient on its own and that full
adoption is not always required. When you have a high degree of convergence in standards, the
additional costs of full adoption might exceed the benefits.
You have to be careful when you are talking about convergence. You can have wholesale
convergence, with only a few things not converged, or you can have one or two things converged
and a hundred things not converged. So researchers studying convergence need to understand the
degree of convergence and how it changes over the study period.
I think the paper’s findings support the IASB’s approach of having principles-based account-
ing standards. They show that IFRS reporting is comparable with U.S. GAAP reporting, even
though U.S. GAAP is usually described as rules-based. Of course, U.S. GAAP has principles and
IFRS has some rules, but, relatively, IFRS is more principles-based and U.S. GAAP is more
rules-based. The paper’s conclusion that the two regimes produce similar outcomes suggests that
they both work, but also confirms that IFRS’s principles-based methodology is good enough to
have tremendous value and utility.

* Accepted by Ole-Kristian Hope. This discussion is based on the version of the paper presented at the 2016 IASB Research
Forum, held in Waterloo, Ontario, Canada, in October 2016, and on Patrick Hopkins’s discussion of the paper at the same event.
† Corresponding author.

CAR Vol. 36 No. 2 (Summer 2019)


Discussion of “Relative Effects of IFRS Adoption and IFRS Convergence” 627

The paper’s findings also suggest that the IASB and the U.S. FASB did the right thing in
working on the convergence projects. We did not get all the projects done, but we got a lot of
them done—some almost completely converged, some heavily converged, some a little bit con-
verged, and some not at all converged.
The paper is based on a sample of 47 German firms and the study period is some time ago,
but if its findings were confirmed globally, the SEC would probably conclude that there really is
not much need to discuss full adoption of IFRS in the United States because the two sets of stan-
dards are already close.
I think the paper also shows that investors are better off whether there is full adoption or
heavy convergence. When I meet with investors as part of the IASB’s outreach activities, they
tend to have very specific questions about differences in the two sets of standards and they spend
a lot of time trying to understand the differences. But what they really want to know is that, when
they compare IFRS financial statements and U.S. GAAP financial statements for companies in
the same industry, are they close enough?

4. Relationship to planned IASB activities


What will follow the era of convergence with the United States? Hans Hoogervorst, the IASB’s
chairman, has said more than once recently that IFRS adoption in the United States is not going to
happen. And there are no more joint projects with the FASB coming up. But we are going to have
continued cooperation. We are working on many of the same projects, and there are extensive staff
discussions—on the disclosure initiative, for example, and a little bit on the Conceptual Framework.
We might come to different conclusions, but we want to be aware when that is happening.
The IASB’s primary outreach in the United States will be to investors. We will certainly be
at academic conferences in the United States, and I as a former preparer will certainly be talking
to preparers there, but our targeted outreach will be aimed at investors. There are about 1,000
mutual funds in the United States that almost exclusively hold international investments.
U.S. investors hold about $7 trillion of foreign securities, and most of the investees are IFRS-
reporting companies or at least in countries that have adopted IFRS. At the same time U.S. corpo-
rations and so on have about $7 trillion in foreign direct investments abroad. So the United States
has a very big stake in IFRS; even if the United States has not adopted it, U.S. investors care an
awful lot about it, and we are spending time with them. We also have an Investor Centre on our
website, and this is another way of engaging with investors.
It should be understood that where convergence has been achieved, there is no force—no
law, no regulation, no enforcement capability—that guarantees that it will continue. It really
depends on the professionals in the organizations concerned and on continuing pressure from
markets. Markets expect that where we are converged, we should not diverge in future and we
should get closer together if we can.
In that context, it is of interest that right now the IASB and the FASB are separately working
on goodwill and impairment. This is not a joint project, but the two boards have been in contact
on it, including speaking together at two public sessions. Goodwill is fully converged. Impair-
ment has a lot of similarities between IFRS and U.S. GAAP, but it is not converged. There has
always been pressure on this issue—whether goodwill should be impaired or amortized or some
combination of the two. And at the moment there is definitely pressure on whether we should
change the impairment models to make them easier to use and reflect impairments in a more
timely way. So this is a good project to watch to see what happens to convergence.
The IASB finished an agenda consultation in 2016 and several themes came out of it for the
next five years (which was the time horizon for this consultation); two of them are relevant to
comparability.
One is that the IASB should focus more on implementation and similar support efforts to
ensure more consistent worldwide application of IFRS. That can only enhance comparability.

CAR Vol. 36 No. 2 (Summer 2019)


628 Contemporary Accounting Research

Another focus is on promoting better communication between preparers and investors. That
takes us to the Disclosure Initiative and principles of disclosure project, a project on primary
financial statements, and some other projects. Better communication with investors, so that they
read things the way they are intended to be read, improves comparability through enhanced
understanding.

5. Suggestions for future research


I like the paper, but it acknowledges some limitations—for example, that it only looks at 47
German companies. Another limitation to my mind is the time that has elapsed since the study
period, which is 2002–2010. There have been a lot more converged standards since then, such as
the revenue recognition standard, which are starting to go into operation. And so I think the study
should be updated.
I also think that it would be useful to look at more countries. Japan is far and away the most
obvious one, although it might be necessary—to build on Patrick Hopkins’s comments—to hire a
research assistant who speaks or reads Japanese. I suggest Japan first of all because the paradigm
is different; IFRS is voluntary, so companies are adopting it because they want to—they see a
business need for it or a business goal that can be attained by doing it. Also, over the last five or
six years, Japan has gone from having about 60 companies using U.S. GAAP to about 20, and
that trend is continuing. At the same time, about 200 companies have moved from Japanese
GAAP to IFRS. So it would be possible to compare a changeover from U.S. GAAP to IFRS and
a changeover from local GAAP to IFRS. India, China, and Taiwan would also be good markets
to look at.
On the paper’s conclusions and the absence of a marked difference in comparability between
adoption and convergence, I think it is necessary to take the next step and look at the costs of full
adoption and at the likely diminishing returns from further improvements in comparability. Also,
is the remaining gap a tolerable one? How close is “close enough”? These questions suggest the
need for some sort of materiality test.
The paper has just a few sentences on the possible effects of management behavior, some-
thing of particular interest to me, having spent most of my career as a preparer. As the paper
notes, one possibility is that “companies attempted to align their accounting practices such that
the effect of adoption would result in limited accounting changes.” On the other hand, manage-
ment might instead take the view that you should never let a big change go to waste, take the
opportunity to unload some inconvenient debits, and blame it on the accounting change. Both
possibilities are worth researching.
As Patrick Hopkins pointed out, there are problems in measuring comparability. I appreciate
that net income and stock return are convenient measures to use, but there is also information in
the notes to the accounts, in the audit report (Europe now has a greatly expanded audit report,
which has not yet reached the United States), in management commentary, and so on. So the
question is: How do you include all these sources of information in measures of comparability? I
do not know the answer, but somehow there have to be measures of comparability that are not
restricted to the information in the two primary financial statements.

References
Hopkins, P. E. 2019. Discussion of “Relative effects of IFRS adoption and IFRS convergence on financial
statement comparability.” Contemporary Accounting Research, this issue.
Lin, S., W. N. Riccardi, and C. Wang. 2019. Relative effects of IFRS adoption and IFRS convergence on
financial statement comparability. Contemporary Accounting Research, this issue.

CAR Vol. 36 No. 2 (Summer 2019)


Copyright of Contemporary Accounting Research is the property of Canadian Academic
Accounting Association and its content may not be copied or emailed to multiple sites or
posted to a listserv without the copyright holder's express written permission. However, users
may print, download, or email articles for individual use.