Financial Crisis And Earnings Management: The European Evidence

This study examines the impact of the 2008–2009 financial crisis on the earnings management behavior of European-listed firms. It finds that earnings management has significantly decreased in the crisis years and provides evidence suggesting that national characteristics and market forces affect the propensity of income smoothing but not accruals quality.

Financial Crisis And Earnings
Management: The European Evidence☆
Andrei Filipa,⁎, Bernard Raffournierb
a ESSEC Business School, France
b GSEM, University of Geneva, Switzerland
Received 28 March 2013
Available online 28 October 2014
Abstract
We examine the impact of the 2008–2009 financial crisis on the earnings management b
European-listed firms. We find that earnings management has significantly decreased in th
This trend is confirmed in most of the 16 countries under review. We also report a link betw
of earnings management and the economic growth rate and provide evidence suggesting t
characteristics and market forces affect the propensity of income smoothing but not accrua
© 2014 University of Illinois. All rights reserved.
JEL Classification: G01 Financial Crises; M41 Accounting
Keywords: Financial crisis; Earnings management; Europe; IFRS; Investor protection; Accounting quality
1. Introduction
In recentyears,earningsmanagementhas received considerableattention from
academics, to the point that there is now an extensive body of research on the det
and consequences ofthe manipulation ofearnings.A common characteristic ofthese
studies is that they do not take into consideration the macroeconomic environmenfirm.In other words,generaleconomic conditions are held constantor supposed notto
influence the incentives for earnings management.Nevertheless,it can be assumed that
dramatic changes in the economic climate have an impacton the firm's propensity to
http://dx.doi.org/10.1016/j.intacc.2014.10.004
0020-7063/© 2014 University of Illinois. All rights reserved.
Available online at www.sciencedirect.com
ScienceDirect
The International Journal of Accounting 49 (2014) 455 – 478
☆ We acknowledge with great appreciation the CPA Ontario for supporting The Journal of International A
Symposium held at Brock University on July 5–7, 2012. We appreciate the helpful comments and sugges
during the Symposium and from conference participants at the 2012. AMIS conference (especially from
⁎ Corresponding author.
E-mail addresses: [e-mail] (A. Filip), [e-mail] (B. Raffournier).

manipulate earnings and/or the sign of these manipulations. The purpose of this pa
explore the influence of significant variations in the economic environment by comp
the earnings management practices of European companies during the 2008 –2009
crisis and in the years before.
According to the accounting literature, the motivations for earnings managemen
classified into two categories: those relating to the market, and those resulting from
relationships. Concerning the market influence, several studies provide evidence co
with the intuition that firms manage earnings upward to avoid reporting losses,earnings
declines, or negative earnings surprises (Ayers, Jiang, & Yeung, 2006; Burgstahler &
1997;Degeorge,Patel,& Zeckhauser,1999).Firms are also suspected ofmanipulating
earnings to facilitate the success of security issues. This hypothesis is supported by
studies showing that firms tend to inflate their earnings prior to seasoned equity off
(Rangan, 1998; Teoh, Welch, & Wong, 1998) or initial public offerings (Teoh, Wong,
1998). Earnings management can also be used as a tool to influence the execution
between the firm and its stakeholders. Empirical studies also provide evidence cons
the idea that managers manipulate earnings to increase their earnings-based comp
(Guidry,Leone,& Rock,1999;Holthausen,Larcker,& Sloan,1995),or to avoid debt
covenant violations (DeFond & Jiambalvo, 1994; Dichev & Skinner, 2002).
Prior studiesinvestigatehow firm's attributes(e.g., presenceof bonusplans,
earnings-based managementcompensation,or debtcovenants) or a particular event(bond
or equity issue) create incentives to manage earnings. Nevertheless, there is also e
macroeconomicconditionsdo affectearningsquality.Johnson (1999),for example,
documents that the value relevance of accounting earnings is sensitive to the busin
0
500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
4,500
5,000
2003 2004 2005 2006 2007 2008 2009 2010
Fig. 1. Evolution of the EuroSTOXX50 index.
456 A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

more precisely thatthe association between earnings and stock returns is higherduring
expansion periods than during contraction years.
Similarly,Jenkins,Kane,and Velury (2009) argue thatsince accounting information
reflects both the consequences of generaleconomic conditions and the effects of firm's
activities, its information content may vary across the business cycle. Contrary to J
(1999), they find that earnings are more value relevant during contraction years th
expansion periods.If macroeconomicconditionsaffectthe information contentof
accounting data, it can be expected that the predictive ability of accounting-based
prediction models varies across the business cycle. Consistent with this conjecture
studies reportthatthese models are sensitive to the occurrence ofa recession (Kane,
Richarson, & Graybeal, 1996; Richardson, Kane, & Lobingier, 1998).
These findings highlight the need for contextual earnings management studies,
research thatwould take into accountthe macroeconomic conditionsin which firms
operate. The 2008–2009 financial crisis and the preceding years provide a unique
for such analysis.From 2003 to 2007,the European financialmarketexperienced a
moderate but continuous growth, reflected by the EuroSTOXX50 index, whose valu
from 2000 to 4500 points during this period. Following the subprime crisis, the ind
to less than 2000 points in February 2009.Since then,it partially recovered without
exceeding the 3000-point threshold (see Fig. 1).
The financialcrisis had significantconsequences on the levelof economic activity.
Globally,the period from the early 1970s to 2008,witnessed an uninterrupted period of
economic growth, with continuously positive GDP growth rates. In 2009, the gross
product declined for the first time in 40 years. In Europe, the crisis was particularly
since the GDP growth rate fell to –4.30% vs. –2.05% for the world in general (see F
Source: World Bank
-6
-4
-2
0
2
4
6
8
1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009
World European Union
Fig. 2. Annual GDP growth rates.
457A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

To measure the impact of the financial crisis on accounting manipulations, we analy
leveland sign of earnings managementby European companies over the period 2006–
2009.The choice ofEurope has severaladvantages.First,Europe has been severely
affected by the 2008–2009 financial crisis. Accordingly, if macroeconomic condition
affect earnings manipulations, significant differences in the magnitude and sign of
management are expected between the expansion period (2006–2007) and the cris
(2008–2009).1 Second, most prior studies were conducted at a single-country level, w
makes the external validity of their findings questionable. Extending the analysis to
geographicalarea neutralizescountry-specific influencesand,consequently,provides
strongerevidence.Assigning countriesinto subgroupswill also make itpossible to
examine the impact of transnational factors such as the legal regime or the prevaili
of financing.
Finally,in order to isolate the impact of macroeconomic conditions,it is necessary to
hold other factors constant.In this regard,the European experimentalsetting is a good
choice inasmuch as since 2005,European-listed companies must comply with IFRS.Allfirms in the sample are thus subject to the same accounting regulation regardless o
nationality. As a result, country differences in the level of earnings management ca
interpreted as reflecting differences in the permissiveness of local accounting stand
Following prior research,earnings managementis measured with severalindicators:
two metrics of income smoothing and three accrual quality measures. Our data com
16 European countries and coverthe period 2006–2009,leading to a sample of8266firm-year observations. We provide evidence that earnings management has signi fi
decreased in the crisis years and that this trend is confirmed in most of the countrie
review.Further,we testand find a directlink between the magnitude ofearnings
management and the economic growth rate of the firm's country.Finally,we also report
that institutional and market characteristics are associated with income-smoothing
but not with accrual quality measures.
The main contribution of this study is to show that earnings management is sens
generaleconomicconditions,a factorthat was omittedby prior research.More
specifically,it documents thatincome manipulations are more likely in growth periods
than in crisis years,a finding thatshould be ofinterestto investors.This paperalso
contributes to the accounting literature by providing evidence that national charact
do not equally affect all types of earnings management. In particular, income smoo
shown as more closely reflecting institutionaland marketcharacteristics than accruals
manipulations,suggesting thatfuture research should notindiscriminately considerall
types of earnings management.
The restof the paper is organized as follows.First,we presenta brief review of the
literature and discuss factors thatmay affectearnings managementduring periods of
1 The financial crisis started in the US in 2007 but hit its peak in September and October 2008 with the
severalmajor financialinstitutions (Lehman Brothers,MerrillLynch,Fannie Mae,Freddie Mac,Washington
Mutual,Wachovia,Citigroup,and AIG).However,as shown by Figs.1 and 2,the consequences of the crisis
started to be felt in Europe mostly in 2008. We thus refer to 2006 and 2007 as pre-crisis years, and 2008
as crisis years.
458 A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

economic stress. Second, we present the research design and the sample. Third, w
the results and make additional analyses to explain country differences. We conclu
a summary of the findings and their implications for further research.
2. Earnings management in troubled periods: the conflicting views
There are a number of reasons to believe that earnings management should be
periods of economic stress. First, in such periods, most firms probably exhibit lowe
which should motivate managers to engage in income-increasing earnings manage
compensate forthe decrease ofoperationalperformance (Ahmad-Zaluki,Campbell,&
Goodacre, 2011). Managers of the most affected firms in particular may manipulat
upward to avoid a large decline of the firm's stock price that would negatively imp
compensation (Charitou, Lambertides, & Trigeorgis, 2007).
Another reason is the presence of debt covenants. Because such covenants are
based on earnings (Dichev & Skinner,2002),income-increasing manipulations should
reduce the probability of violations (DeFond & Jiambalvo,1994;Iatridis & Kadorinis,
2009; Saleh & Ahmed, 2005; Sweeney, 1994).
Consistentwith the intuition thatperiods of economic turbulence incite managers to
manipulate earnings upward,Ahmad-Zalukiet al. (2011) reportthatincome-increasing
earnings management by Malaysian firms engaged in initial public offerings was lim
IPOs that occurred during the Asian crisis period (1997–1998).
Adversely,in troubled periods,some firms may have incentives to manage earnings
downward, in particular those that must undertake debt restructuring due to debt
violation or failure to meet a debt repayment.For these firms,reporting losses may help
obtain concessions from lenders.Banks have the option to refuse these concessions and
require the firm's liquidation. However, in recession periods, the realizable value o
assets is probably low (Shleifer& Vishny,1992),which should incite them to accept
conditionsthatthey would ordinarily refuse.Consistentwith thisintuition,Asquith,
Gertner, and Scharfstein (1994) provide evidence that rather than exercising their
call the loan, banks generally prefer to restructure the debt by waiving covenants,
principal and interest, or reducing the interest rate.
Income-reducing earnings managementmightalso resultfrom agency relationships
with employees.DeAngelo,DeAngelo,and Skinner(1994),for example,note thatby
reporting losses, managers portray the firm as seriously troubled, which may be us
extractconcessionsfrom employeeswho otherwisewould doubtthe existenceand
persistence of the firm's difficulties. In support of this assertion, DeAngelo and DeA
(1991) show that reported earnings are lower during union renegotiations.
The search for political advantage can also motivate the reduction of earnings. I
periods, governments are likely to provide support to firms in financial distress (Pe
1976). Governmental support may take various forms. Ahmed, Godfrey, and Saleh
mention thatduring the Asian financialcrisis,the Malaysian governmentinstituted
mechanisms to facilitate the debt restructuring of companies. In 2008–2009, gove
of many countries provided banks with public funds to mitigate the credit crisis. Fir
also use the economic downturn asa pretextto obtain advantagesor oppose new
regulations.Because the probability of obtaining governmental aid increases as finan
459A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

performance worsens, firms have an incentive to deflate earnings. The literature pr
evidence of income-reducing earnings management to obtain advantages from the
(Jones, 1991; Lim & Matolcsy, 1999; Navissi, 1999). In an analysis of accounting cho
troubled companies,DeAngelo etal.(1994) reportthatseveraldistressed firms used theirfinancial difficulties to argue for import relief or antitrust clearance.
All previous arguments suggest that periods of economic downturn should be ass
with higher level of earnings management, although they do not agree with the sig
manipulations. Nevertheless, there are also reasons to believe that crises are less f
earnings management than expansion periods.
First, it can be argued that during crises, firms are subject to increased monitorin
auditors,creditors,and other stakeholders,which should resultin managers having less
discretion to manage earnings (Chia,Lapsley,& Lee, 2007).Anotherreason is that
litigation risk is probably higher during periods of economic decline, when equity m
experience sharp drops in stock prices. Managers should respond to this risk increa
limitation of earnings management. Contraction periods should thus be associated
earnings managementand,consequently,more conservative (i.e.,more timely) earnings
(Jenkins et al., 2009). The influence of litigation risk on conservatism is well docume
Huijgen and Lubberink (2005),for example,show thatfirms reportmore conservative
earnings in high legalliability regimes.Clients of Big 4 firms have also been shown as
reporting more conservatively in the post-Enron contextthan before (Krishnan,2007;
Willekens & Bauwhede, 2003).
Lower level of earnings management in recession periods may also result from a
demand forconservative earnings.Because ofthe transitory nature ofcrises,earnings
reported in such periods are less persistent, and thus less useful for predictions. Co
with this assertion, several studies show that the value relevance of earnings (i.e., t
of association between accounting data and market prices) varies across the busine
(Ball & Shivakumar,2005; Brown,He,& Teitel,2006).This increased uncertainty about
future outcomes should motivate marketforces to demand more conservative earnings in
crisis periods (Jenkins et al., 2009), which should dissuade companies to manipulate
earnings. One can also consider that in crisis periods, the market is more inclined to
poor performance (Ahmad-Zaluki et al., 2011). As a consequence, firms have less in
engage in earnings management activities.
Bertomeu and Magee (2011) modelthe dynamics between accounting standards,the
quality of financial reporting, and the state of the economy. Their analytical results
thatfinancialreporting quality should reach its maximum when the economy is good
(expansionarytimes),decreaseas the economicconditionsbecomeless favorable
(moderate times),and increase again ifthe economy becomes recessionary.Therefore,
the financial reporting quality is non-monotonic with the state of the economy. Alth
their model refers to the regulator point of view,it assumes that the regulator passes the
reporting quality levelthatis supported by a majority ofthe agents in the economy
(Wagenhofer, 2011).
On the basis of the existing literature,the effectthat economic crises should have on
earnings managementis unclear.Some argue thatin such periods,firms are induced to
manipulate their earnings,whereas others think thatthey have more incentives to report
unbiased net income. Even among the proponents of increased earnings managem
460 A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

opinion is divided, some of them expecting income-increasing manipulations, while
argue thatperiodsof economic stressshould be associated with downward income
adjustments. The empirical analysis should tell us which of these conflicting views
most likely.
3. Measuring earnings management
To testif firms exhibitunusualearnings managementbehaviorduring the financial
crisis,we employ standard techniques from the income smoothing and accruals qua
literature. This section presents the research design and the econometric models ufirsttwo metrics capture income smoothing,whereas the lastthree are accrualquality
measures.
Our firstattribute of income smoothing is adapted from Leuz,Nanda,and Wysocki
(2003)and focuses on insiders'reporting choices.Specifically,IS1 is defined as the
standard deviation of cash flow from operations divided by the standard deviation
income.In case of income smoothing,the variability ofearnings (as measured by the
standard deviation ofnet income)should be lessthan the variability ofcash flow.
Following Leuz etal. (2003),2 we thus interprethigh values ofIS1 as evidence that
managers exercise accounting discretion to smooth reported earnings.
The second indicator of income smoothing is also taken from Leuz etal. (2003).It
represents the Spearman correlation between variations in accruals and variations
flow from operations.Because accruals buffer cash flow shocks,the correlation between
changesin accrualsand changesin cash flow should be negative (Dechow,1994).
However,if accruals are manipulated to smooth income,the absolute value oftheir
correlation with cash flowsshould be particularly high.We define accrualsas the
difference between net income and cash flow from operations. As in prior research
Shivakumar, 2005; Barth, Landsman, & Lang, 2008; Lang, Raedy, & Wilson, 2006 ),
interpret a high negative correlation as evidence of earnings smoothing. For consis
presentation and interpretation,the Spearman coefficientwas multiplied by − 1 so that
higher IS2 scores reflect higher level of income smoothing.
Our first accrualquality metric isbased on the modified Jonesmodel.Although
differentmodels have been developed to detectearnings management,the Jones (1991)
model and the modified Jones model developed by Dechow, Sloan, and Sweeney (1
are the mostextensively used in identifying discretionary accruals.The only difference
between them is the inclusion of changes in accounts receivables in the modi fied J
model.In addition,followingKothari,Leone,and Wasley(2005),we includea
performance measure, i.e., return on assets, to control for the impact of firm perfo
on unexpected accruals. We require at least 20 observations in the same industry
2 Leuz et al. (2003) use the ratio of the standard deviation of net income to the standard deviation of c
from operations. For consistency in presenting and interpreting our results, we reverse the ratio so that
mean high levels of income smoothing.
461A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

as two-digit SIC code). The modified Jones model is estimated cross-sectionally, usinfirm-year observations in the same industry. Its equation is as follows:
ACCit ¼ α0 þ α11=Ait−1þ α2 ΔREVit–ΔRECitð Þ þ α3PPENit þ α4ROAit þ ζit ð1Þ
Where:
ACCjt accruals (change in non-cash currentassets minus change in currentliabilities
adjusted forthe currentportion oflong-term debt,minusdepreciation and
amortization expense) scaled by lagged total assets of firm j in year t;
A it − t lagged total assets of firm j in year t – 1;
ΔREV jt change in sales scaled by lagged total assets of firm j in year t;
ΔREC jt change in receivables from clients scaled by lagged total assets of firm j in y
PPENjt net value of property, plant, and equipment scaled by lagged total assets of
in year t;
ROAjt net income scaled by lagged total assets of firm j in year t.
Following priorresearch,the un-standardized residualfrom Eq.(1) is deemed the
discretionary accruals componentor abnormalaccruals.The principalidea behind the
model is to determine the extent of the measurement error that can be unintention
businessactivity)or intentional(due to earningsmanagement).The variance ofthis
measurementerror can be viewed as an inverse measure of accrualquality (Dechow &
Dichev,2002;Francis,LaFond,Olsson,& Schipper,2005;Rajgopal& Venkatachalam,
2011). Accordingly, the first accrual proxy, JONES1, represents the standard deviati
the residuals from the modified Jones model.We interpreta low standard deviation of
residuals as reflecting high accrual quality, and thus low level of earnings managem
Our second accrualquality metric was also used by Jones,Krishnan,and Melendrez
(2008). It is inspired by Larcker and Richardson (2004), who included in the Jones m
additional independent variables that are correlated with measures of unexpected a
The book-to-market ratio serves as a proxy for expected growth in the firm's operat
rationale behind this inclusion is that growing firms are expected to have large accr
necessarily due to opportunistic managerialbehavior.The operating cash flow is also
includedas a betterspecifiedmeasureof performance.The modelis estimated
cross-sectionally foreach industry,with more than 20 observations using the following
equation:
ACCit ¼ α0 þ α11=Ait−1þ α2 ΔREVit–ΔRECitð Þ þ α3PPENit þ α4BMit þ α5CFOit þ ζit ð2Þ
Where:
BM jt book value of equity divided by market value of firm j in year t;
CFOjt cash flow from operations scaled by lagged total assets of firm j in year t;
All other variables defined before.
Un-standardized residuals from Eq. (2) represent an alternative proxy for discreti
accruals, and JONES2 represents the standard deviation of these residuals. We inte
462 A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

low standard deviation of residuals as reflecting high accrual quality, and thus low
earnings management.
The last accrual quality metric is based on the cross-sectional Dechow and Dich
model, as modified by McNichols (2002) and Francis et al. (2005). Some recent stu
Chen, Hope, Li, & Wang, 2011; Core, Guay, & Verdi, 2008; Jones et al., 2008; Kim &
2010) apply this modelto assess accruals as a measure of earnings quality,arguing that
accruals are temporary adjustments made to better measure firm performance. Fr
(2005) argue that the Dechow and Dichev (2002) approach is superior to the modi
(1991)model'sidentification ofabnormalaccrualsbecause ofa more directlink to
information risk. The intuition behind this model is that accrual quality may be de fi
extent to which accruals map into cash flow realizations. We estimate the following
each industry with more than 20 observations as follows:
ACCit ¼ α0 þ α1CFOit−1þ α2CFOit þ α3CFOitþ1 þ α4 ΔREVit–ΔRECitð Þ
þ α5PPENit þ ζit ð3Þ
Where: All variables defined before.
The residuals from Eq. (3) represent the estimation errors in the current accrual
notassociated with operating cash flows and thatcannotbe explained by the change in
revenue orthe levelof property,plant,and equipment.Our metric DD represents the
standard deviation of these residuals. We interpret a low value of DD as evidence o
accrualquality.In all variations ofthese proxies,a low value indicates high accruals
quality and thus a low level of earnings management.
4. Sampling and data collection
At the time when the European Parliament issued regulation no. 1606/2002 ann
the mandatory IFRS adoption in 2005 for all public firms, the European Union (here
EU) was composed of15 memberstates.Because accounting standards influence the
earningsmanagementbehaviorof firms,our study focuseson these 15 “early” EU
members that adopted a common set of high quality standards (IFRS).Luxembourg was
excluded because ofan insufficientnumberof observations.Although they are not
members of the EU, Switzerland and Norway have been added to the sample, beca
Table 1
The sample.
Public firms from 16 countries 6901
– Non-IFRS firms 2643
– Banks and financial institutions 901
= Firms included in the sample 3357
Firm-year observations for 2006 to 2009 13,428
– Observations with unavailable accounting data 4060
– Observations with negative equity 948
– Observations from industries with less than 20 observations 154
= Final number of observations 8266
463A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

also issued regulations requiring the use of IFRS by all listed companies. This leads
sample of6901 firms coming from 16 countries with data available on Worldscope.
Table 1 describes the sampling and data collection process.
To avoid ambiguity,2643 firmsfollowing accounting standardsotherthan IFRS
(mainly US orlocalGAAP) during the period 2005 to 2010 were dropped from the
sample.3 Although mostEuropean banks and financialinstitutions apply IFRS,they are
also subject to specific reporting regulations. Moreover, the empirical models that c
earnings managementbehavior have been developed for commercialor industrialfirms
and, as such, are not appropriate for the financial industry. Banks and financial inst
were thus excluded from the sample.
Accounting data were collected for the four-year period 2006–2009. According to
Cazavan-Jeny, Jeanjean, and Weiss (2011), managers may use discretion allowed un
(which regulates first-time adoption of IFRS) to increase earnings management. Asfirst IFRS reporting period for most companies, this year was not included in the ana
assure comparability across our differentmetrics of financialreporting quality,we required
available financialstatements data for allour variables.Data were notavailable for 4060firm-year observations,and another 948 observations were dropped from the sample du
negative equity. Finally, we required at least 20 observations for each industry (two
code). Our final sample consists therefore of 8266 firm-year observations.
Our income-smoothing metrics (IS1 and IS2) are computed by pooling all observa
of a particular year or time period.Our measures of accrualquality (JONES1,JONES2,
and DD) are obtained in two steps.First,we compute discretionary accruals with each
modelby pooling allobservations from allcountries and years within the same 2-digit
industry SIC code. These discretionary accruals are then grouped by year or time pe
5. Results
Table 2 reports the values ofthe five earnings managementmetrics forthe pooled
sample,for each year,and for the pre- and post-crisis periods.Both measures of income
smoothing (IS1 and IS2) exhibit a similar time pattern: the lowest value is for the ye
for IS1 (0.750) and for the year 2009 for IS2 (0.576), whereas highest values are in
2007 (IS1 = 1.015 and IS2 = 0.685).These observations denote a strong decrease in
income smoothing in the period of the financial crisis.
The accrual quality metrics are reported in the last three columns of the table. Th
with the highestaccrualquality (i.e.,loweststandard deviation of residuals) is the year
2009 in all cases.The quality of accruals seems to increase in the last three years und
review, since there is a monotonous decreasing pattern of earnings management in
from 2007 to 2009. For instance, the value of JONES1 is 0.172 and 0.203 for the pre
years, which compares to only 0.124 and 0.113 for the crisis period.
According to Bertomeu and Magee (2011), financial reporting quality should reac
maximum when the economy is good,decrease as the economic conditions become less
3 Regulation no. 1606/2002 allowed entities following other high quality accounting standards (i.e., US G
to postpone the adoption of IFRS to 2007. Moreover,in some countries, the financial market is segmented and
only firms that are listed on the main segment must comply with IFRS.
464 A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

favorable,and increase again ifthe economy becomes recessionary.Our results seem
consistent with this conjecture.
This intuition is confirmed by the comparison of the average values of indicators
pre-crisis(2006–2007)and the crisisperiod (2008–2009).All earningsmanagement
metrics experience a sharp decrease in the crisis years. Indicators of income-smoo
and IS2 dropped from 0.981 to 0.810 and 0.678 to 0.600 respectively. Similarly, th
of accrual quality measures fell from 0.189 to 0.119 (JONES1), 0.176 to 0.115 (JON
and 0.169 to 0.111 (DD).
The significance of these differences was tested using a bootstrapping procedur
randomly chooses a sample of100 firms-yearobservations with replacementfrom a
period.We computed our attributes of earnings management and repeated the proc
10,000 times. We next performed an independent-samples t-test. All differences b
the two sub-periods are statistically significant at usual levels, which is consistent
conjecture that firms engage less in earnings management during crisis periods.
The impact of the financial crisis on the earnings management behavior of firms
influenced by the management's incentives to manipulate earnings upwards or do
Therefore we separate the sample in positive and negative discretionary accruals.
Table 2
Earnings management metrics by period.
Year N Income smoothing Accrual quality
IS1 IS2 JONES1 JONES2 DD
Pool 8266 0.894 0.625 0.156 0.147 0.143
2006 1897 0.942 0.668 0.172 0.166 0.159
2007 2041 1.015 0.685 0.203 0.186 0.177
2008 2146 0.750 0.628 0.124 0.118 0.116
2009 2182 0.946 0.576 0.113 0.112 0.106
2006–2007 3938 0.981 0.678 0.189 0.176 0.169
2008–2009 4328 0.810 0.600 0.119 0.115 0.111
Difference 0.171***
(21.36)
0.078***
(66.95)
0.070***
(86.36)
0.061***
(88.17)
0.057***
(90.28)
IS1 is the ratio of the standard deviation of CFOjt to the standard deviation of NIjt; IS2 is the Spearman correlation
between variations in accruals (defined as the difference between net income and cash flow from opera
lagged assets and variations in CFOjt, multiplied by minus one; JONES1 is the standard deviation of the residuals
the modified Jones model:ACCjt = α0 + α11/Ajt − 1+ α2(ΔREVjt − ΔRECjt) + α3PPENjt + α4ROAjt + εjt (1);
JONES2 is the standard deviation ofthe residuals from the modified Jones model:ACCjt = α0 + α11/Ajt − 1+
α2(ΔREVjt − ΔRECjt) + α3PPENjt + α4BMjt + α5CFOjt + εjt (2); DD is the standard deviation of the residuals from
the modified Dechow and Dichev model:ACCjt = α0 + α1CFOjt − 1+ α2CFOjt + α3CFOjt + 1+ α4(ΔREVjt −
ΔRECjt) + α5PPENjt + εjt (3).
Where ACCjt = accruals (change in non-cash current assets minus change in current liabilities adjusted fo
current portion of long-term debt minus depreciation and amortization expense) scaled by lagged totalfirm j in year t; Ajt − 1= lagged total assets of firm j in year t − 1; ΔREVjt = change in sales scaled by lagged total
assets of firm j in year t; ΔRECjt = change in receivables from clients scaled by lagged total assets of firm j in
t; PPENjt = netvalue of the property,plantand equipmentscaled by lagged totalassets of firm jin year t;
ROAjt = net income scaled by lagged total assets of firm j in year t; BMjt = book-to-market ratio of firm j in year
t; CFOjt = operating cash flow of firm j in year t scaled by lagged total assets; NIjt = net income of firm j in year t
scaled by lagged total assets.
*, **, *** indicate statistical significance at 0.10, 0.05, and 0.01 respectively; t-statistics into brackets.
465A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

observation is defined as positive (negative)discretionary accrualif the un-standardized
residual is positive (negative) on all three discretionary accruals models (Eqs. (1), (2
(3)). The results are presented in Table 3.
They confirm the previous findings that earnings management is significantly de
in the crisis years, as the difference between the pre-crisis (2006–2007) and the cri
(2008–2009) is positive and statistically significant, irrespective of the sign of discre
accruals.However,this decrease is more pronounced for firms with positive discretiona
accruals,i.e.,engaging in income-increasing earnings management.These firms exhibita
higher level of earnings management in the pre-crisis period compared to those wit
discretionary accruals, while the gap between the two sub-samples is narrowing du
crisis years. With the exception of IS2, the difference in differences is positive and h
significant,which suggeststhatfirms tend to manageearningsupwardsmorethan
downwards.Again,the gap between the two subsamples is narrowing during the crisis
Table 3
Positive vs. negative discretionary accruals.
Income smoothing Accrual quality
N IS1 IS2 JONES1 JONES2 DD
Positive discretionary accruals
2006–2007 1463 1.082 0.691 0.237 0.216 0.203
2008–2009 1279 0.650 0.645 0.117 0.112 0.108
Difference 0.432***
(25.14)
0.046***
(41.25)
0.120***
(124.73)
0.103***
(125.25)
0.095***
(126.49)
Negative discretionary accruals
2006–2007 1572 0.968 0.660 0.109 0.110 0.106
2008–2009 1817 0.816 0.593 0.085 0.085 0.085
Difference 0.152***
(53.36)
0.067***
(58.68)
0.025***
(58.17)
0.024***
(59.32)
0.021***
(51.66)
Difference in difference 0.280***
(49.52)
− 0.021***
(− 68.22)
0.095***
(101.19)
0.079***
(104.70)
0.074***
(103.33)
IS1 is the ratio of the standard deviation of CFOjt to the standard deviation of NIjt; IS2 is the Spearman correlation
between variations in accruals (defined as the difference between netincome and cash flow from operations)
scaled by lagged assets and variations in CFOjt, multiplied by minus one; JONES1 is the standard deviation of the
residuals from the modified Jones model: ACCjt = α0 + α11/Ajt − 1+ α2(ΔREVjt − ΔRECjt) + α3PPENjt + α4ROA-
jt +εjt(1); JONES2 is the standard deviation of the residuals from the modified Jones model: ACCjt = α0 + α11/Ajt − 1+
α2(ΔREVjt − ΔRECjt) + α3PPENjt + α4BMjt + α5CFOjt + εjt (2); DD is the standard deviation of the residuals from
the modified Dechow and Dichev model:ACCjt = α0 + α1CFOjt − 1+ α2CFOjt + α3CFOjt+1+ α4(ΔREVjt −
ΔRECjt) + α5PPENjt + εjt (3).
Where ACCjt = accruals (change in non-cash current assets minus change in current liabilities adjusted for
current portion of long-term debt minus depreciation and amortization expense) scaled by lagged total afirm j in year t; Ajt − 1= lagged total assets of firm j in year t − 1; ΔREVjt = change in sales scaled by lagged total
assets of firm j in year t; ΔRECjt = change in receivables from clients scaled by lagged total assets of firm j in y
t; PPENjt = netvalue of the property,plantand equipmentscaled by lagged totalassets of firm jin year t;
ROAjt = net income scaled by lagged total assets of firm j in year t; BMjt = book-to-market ratio of firm j in year
t; CFOjt = operating cash flow of firm j in year t scaled by lagged total assets; NIjt = net income of firm j in year t
scaled by lagged total assets.
*, **, *** indicate statistical significance at 0.10, 0.05, and 0.01 respectively; t-statistics into brackets.
466 A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

years. These conclusions hold for all our accrual quality metrics, whereas the resul
income-smoothing metrics are less consistent.This confirms thatthe two categories of
proxies capture different faces of earnings management.
These results suggest the existence of a link between the macroeconomic condi
which the firm operates and its level of earnings management. To more deeply inv
this association,we regressed ourfive yearly metrics ofearnings managementon the
growth rate of Gross Domestic Product (GDP). The results are presented in Table 4
For all accrual quality metrics (JONES1, JONES2, and DD) and the second income
attribute (IS2),the response coefficients are positive and statistically significant.The only
growth metric that is not associated with earnings management is IS1. Globally, th
are consistentwith the conjecture thateconomic crisesare lessfavorable to earnings
management than expansion periods.
It is interesting to note that IS2 and accrual quality metrics follow a similar patte
not surprising, given that IS2 is measuring the correlation between variations in ac
variations in cash flow from operations, whereas IS1 is not a direct measure of acc
Table 4
Results of regressions: earnings management and economic growth
Model : EMkt ¼ α0 þ α1GROWTHkt þ ζkt
Where: EMktearnings management metric of country k in year t; GROWTHktgrowth rate of GDP of country k in
year t (source: http://epp.eurostat.ec.europa.eu).
Income smoothing Accrual quality
EM kt IS1 IS2 JONES1 JONES2 DD
Intercept 1.118***
(19.292)
0.632***
(33.285)
0.136***
(15.967)
0.130***
(17.108)
0.126***
(18.248)
GROWTHkt − 0.010
(− 0.081)
0.329***
(2.739)
0.336***
(2.811)
0.305**
(2.519)
0.348***
(2.924)
N 64 64 64 64 64
F 0.007 7.504*** 7.903*** 6.346** 8.547***
R2 0.000 0.108 0.113 0.093 0.121
IS1 is the ratio of the standard deviation of CFOjt and the standard deviation of NIjt; IS2 is the Spearman correlation
between variations in accruals (defined as the difference between net income and cash flow from opera
lagged assets and variations in CFOjt, multiplied by minus one; JONES1 is the standard deviation of the residual f
the modified Jones model: ACCjt = α0 + α11 / Ajt − 1+ α2(ΔREVjt − ΔRECjt) + α3PPENjt + α4ROAjt + εjt (1);
JONES2 is the standard deviation of the residual from the modified Jones model: ACCjt = α0 + α11 / Ajt − 1+ α2(ΔREVjt −
ΔRECjt) + α3PPENjt + α4BMjt + α5CFOjt + εjt (2); DD is the standard deviation of the residualfrom the modified
Dechow and Dichev model:ACCjt = α0 + α1CFOjt − 1+ α2CFOjt + α3CFOjt + 1+ α4(ΔREVjt − ΔRECjt) + α5PPEN-
jt + εjt(3).
Where ACCjt — accruals (change in non-cash current assets minus change in current liabilities adjusted fo
current portion of long term debt minus depreciation and amortization expense) scaled by lagged totalfirm j in year t; Ajt − 1— lagged total assets of firm j in year t − 1; ΔREVjt — change in sales scaled by lagged
total assets of firm j in year t; ΔRECjt — change in receivables from clients scaled by lagged total assets of firm
in year t; PPENjt — net value of the property, plant and equipment scaled by lagged total assets of firm j in
ROAjt — net income scaled by lagged total assets of firm j in year t; BMjt — book-to-market ratio of firm j in year
t; CFOjt — operating cash flow of firm j in year t scaled by lagged total assets; NIjt — net income of firm j in year t
scaled by lagged total assets.
*, **, *** indicate statistical significance at 0.10, 0.05, and 0.01 respectively; t-statistics into brackets.
467A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

Table 5
The influence of the financial crisis on earnings management: country results.
Income smoothing Accrual quality
N IS1 IS2 JONES1 JONES2 DD
AUT 82 0.952 0.754 0.170 0.168 0.164
88 1.246 0.687 0.104 0.097 0.108
− 0.294*** 0.067*** 0.066*** 0.071*** 0.056***
(− 6.90) (7.12) (14.15) (14.83) (10.91)
BEL 122 0.930 0.690 0.157 0.155 0.157
139 1.000 0.686 0.127 0.122 0.143
− 0.070*** 0.004 0.030 0.033 0.015***
(− 2.91) (1.35) (0.23) (0.28) (3.32)
CHE 202 0.880 0.629 0.153 0.159 0.144
217 0.737 0.619 0.112 0.118 0.120
0.143*** 0.009* 0.040*** 0.041*** 0.024***
(12.58) (1.70) (10.00) (8.58) (5.46)
DEU 571 1.198 0.688 0.201 0.184 0.177
610 1.123 0.621 0.134 0.123 0.115
0.076*** 0.068*** 0.067*** 0.061*** 0.062***
(5.90) (4.91) (11.12) (12.23) (13.49)
DNK 116 0.869 0.518 0.130 0.134 0.150
124 0.763 0.487 0.105 0.108 0.091
0.106*** 0.031*** 0.025*** 0.025*** 0.059***
(13.16) (3.56) (6.73) (8.26) (19.73)
ESP 144 1.546 0.890 0.302 0.123 0.112
165 1.117 0.716 0.113 0.101 0.091
0.430*** 0.174*** 0.189*** 0.022*** 0.021***
(11.06) (20.40) (8.48) (86.63) (8.17)
FIN 162 1.189 0.707 0.198 0.205 0.212
167 0.970 0.624 0.118 0.120 0.124
0.219*** 0.084*** 0.079*** 0.084*** 0.088***
(8.92) (8.49) (7.98) (8.42) (8.30)
FRA 665 1.025 0.751 0.182 0.178 0.165
690 1.121 0.683 0.094 0.093 0.092
− 0.096*** 0.068*** 0.088*** 0.085*** 0.073***
(− 2.15) (4.87) (10.01) (8.92) (9.76)
GBR 820 0.884 0.593 0.147 0.144 0.142
943 0.680 0.506 0.134 0.125 0.123
0.204*** 0.088*** 0.013*** 0.018*** 0.019***
(3.88) (4.99) (5.05) (5.97) (5.95)
GRC 194 1.093 0.870 0.203 0.196 0.193
209 0.613 0.775 0.116 0.121 0.120
0.479*** 0.096*** 0.087*** 0.075*** 0.073***
(2.30) (11.54) (15.96) (12.31) (11.43)
IRL 40 1.073 0.445 0.061 0.059 0.061
50 1.043 0.341 0.126 0.131 0.114
0.030*** 0.104*** − 0.066*** − 0.072*** − 0.053***
(3.96) (5.44) (− 40.49) (− 39.11) (28.51)
ITA 253 1.271 0.803 0.202 0.201 0.195
309 1.031 0.690 0.094 0.091 0.085
0.240*** 0.113*** 0.108*** 0.111*** 0.110***
(11.43) (10.98) (14.95) (14.15) (14.49)
468 A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

Table 5 reportsthe valuesof earningsmanagementmetricsfor each country and
sub-period. Due to a low number of observations in several countries, the signi fica
the differences was tested using a bootstrapping procedure consisting in repeating
times the procedure with a sample of 10 firms-year observations.
In most countries, the findings are consistent with the previously reported trend
earnings managementmetrics are lower in the crisis years than in the preceding period
However, in Austria, Belgium, France, Norway, and Portugal, income smoothing as
Table 5 (continued)
Income smoothing Accrual quality
N IS1 IS2 JONES1 JONES2 DD
NLD 159 1.026 0.734 0.137 0.126 0.123
167 0.724 0.485 0.127 0.134 0.139
0.301*** 0.249*** 0.010** − 0.008*** − 0.016***
(18.46) (18.99) (2.20) (− 3.38) (− 4.65)
NOR 136 0.916 0.568 0.367 0.363 0.317
164 1.124 0.401 0.166 0.164 0.130
− 0.208* 0.168*** 0.201*** 0.199*** 0.187***
(− 1.87) (12.15) (19.32) (19.88) (22.30)
PRT 52 1.642 0.764 0.091 0.087 0.083
61 3.011 0.776 0.081 0.077 0.078
− 1.369*** − 0.011 0.010*** 0.010*** 0.005***
(− 15.18) (− 0.44) (7.73) (6.76) (2.63)
SWE 220 0.890 0.522 0.138 0.134 0.129
225 1.002 0.535 0.075 0.080 0.076
− 0.112 − 0.013 0.063*** 0.054*** 0.053***
(− 1.18) (− 0.93) (22.36) (19.24) (17.80)
Countries are named using ISO country codes.
For each country:
– 1st line: mean value 2006–2007
– 2nd line: mean value 2008–2009
– 3rd line: difference (*, **, *** indicate statistical significance at 0.10, 0.05 and 0.01, respectively)
– 4th line: t-statistics into brackets
IS1 is the ratio of the standard deviation of CFOjt to the standard deviation of NIjt; IS2 is the Spearman correlation
between variations in accruals (defined as the difference between netincome and cash flow from operations)
scaled by lagged assets and variations in CFOjt, multiplied by minus one; JONES1 is the standard deviation of
the residuals from the modified Jones model: ACCjt = α0 + α11 / Ajt = 1+ α2(ΔREVjt − ΔRECjt) + α3PPENjt +
α4ROAjt + εjt (1); JONES2 is the standard deviation of the residuals from the modified Jones model: ACCjt =
α0 + α11 / Ajt − 1+ α2(ΔREVjt − ΔRECjt) + α3PPENjt + α4BMjt + α5CFOjt + εjt (2); DD is the standard devia-
tion of the residuals from the modified Dechow and Dichev model:ACCjt = α0 + α1CFOjt − 1+ α2CFOjt + α3-
CFOjt + 1+α4(ΔREVjt − ΔRECjt) + α5PPENjt + εjt (3).
Where ACCjt = accruals (change in non-cash current assets minus change in current liabilities adjusted fo
current portion of long-term debt minus depreciation and amortization expense) scaled by lagged totalfirm j in year t; Ajt − 1= lagged total assets of firm j in year t − 1; ΔREVjt = change in sales scaled by lagged total
assets of firm j in year t; ΔRECjt = change in receivables from clients scaled by lagged total assets of firm j in
t; PPENjt = netvalue of the property,plantand equipmentscaled by lagged totalassets of firm jin year t;
ROAjt = net income scaled by lagged total assets of firm j in year t; BMjt = book-to-market ratio of firm j in year
t; CFOjt = operating cash flow of firm j in year t scaled by lagged total assets; NIjt = net income of firm j in year t
scaled by lagged total assets.
469A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

by IS1 increases through time.On the other hand,IS2 is positive and significantin 13
countriesbut not in Belgium,Portugal,and Sweden.With regard to accrualsquality
measures, all countries except Ireland and the Netherlands exhibit significant incre
results show that although there is a general trend toward less earnings manageme
periods, not all countries exhibit the same evolution.
6. Explaining country differences
Differencesreported in Table 5 suggestthatearningsmanagementis sensitive to
country characteristics.As noted by Ball,Robin,and Wu (2003),financialreporting
practices depend on the incentives of managers who are in charge of financial state
preparation, and these incentives are the product of market and political forces tha
at the country level. In this section, we investigate the impact of these forces on the
earnings management in each sub-period.
6.1. The influence of the legal and regulatory environment
A common assertion is that earnings management is more frequent in countries
low level of shareholder protection. This argument is based on the idea that strong
protection restricts insiders' profit appropriation and thus reduces their incentive to
in earnings management to conceal their profit diversion activities (Kinnunen & Kos
2003).On the contrary,when shareholderprotection is low,the likelihood ofprofit
diversionby insidersis high, which shouldconstitutean incentivefor earnings
management.
Several studies report evidence consistent with these predictions. Leuz et al. (20
particular show that earnings management is negatively related to shareholder pro
Similar results were found by Haw, Hu, Hwang, & Wu (2004), Burgstahler, Hail, and
(2006), and Nabar and Boonlert-U-Thai (2007). Consistent with prior studies, one of
proxy for shareholder protection is the anti-director rights index (ANTIDIR) develope
La Porta, Lopez-de-Silanes, and Shleifer (1998). This index, whose values range from
6, is a combined measure of six individual rights of shareholders, such as suing dire
calling special shareholder meetings.It measures minority shareholders' protection from
expropriation by corporate insiders ormajority shareholders.Because this index was
elaborated in the '90s, its values do not capture the potential effects of important r
changes thatoccurred in Europe since thattime (such as the enactmentof the Market
Abuse Directive, adopted by the EU in response to the US Fair Disclosure Regulation
orderto have an updated measure ofinvestorprotection,we also use the Strength of
InvestorProtection index (INVPRO)developed by the World Bank.This indicator,
calculated each year, is the average of three indexes measuring the extent of disclo
extent of director liability, and the ease of shareholder suits in each country.
Strong legal rules are a necessary condition to guarantee that the rights of share
are protected, but not a sufficient one. Legal rules may remain largely ineffective w
properenforcement(Burgstahleret al., 2006).Furthermore,a solid system oflegal
enforcementcan also substitute for weak rules,since active and well-functioning courts
can rescue investors abused by managers (La Porta et al., 1998).
470 A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

The influence of law enforcement is supported by several studies dealing with va
aspects ofaccounting quality (Daske,Hail, Leuz,& Verdi, 2008;DeFond,Hung,&
Trezevant,2007).With regard to earnings management,Bushman and Piotroski(2006)
document that firms in countries with strong public law enforcement slow the reco
of good news in reported earnings relative to firms in countries with weak public la
enforcement. We thus expect a negative association between law enforcement an
management measures.
La Porta et al. (1998) proposed several proxies for the quality of law enforcemen
al. (2003) obtained a global indicator by averaging three of them in a single measu
it (ENFORC) to test whether stronger law enforcement constrains earnings manage
order to have an updated estimate, we also use the Rule of Law index (RULLAW) d
by the World Bank. This indicator reflects confidence that people have in the rules
in particular the quality of contract enforcement, property rights, the police, and th
Law is notthe only source ofprotection forinvestors.Compliance with corporate
governance recommendationsmay also reduce the levelof expropriation by insiders
(Renders & Gaeremynck,2007).Durnev and Kim (2005)show thatfirms with better
governance are valued higher on stock markets, especially where legal investor pr
is weak, which leads them to the conclusion that firms adapt to poor legal environm
establish efficient governance practices.
Thereis a growing body ofresearch on theinfluenceof corporategovernance
characteristics on accounting quality. The proportion of outside directors in particu
found as being positively related to earnings timeliness (Beekes, Pope, & Young, 2
negatively associated with earnings management(Klein,2002;Peasnell,Pope,& Young,
2000).Variablesaimed atmeasuring the quality oflegalenforcement,such asthose
developed by La Porta et al. (1998), cannot capture the extent of corporate govern
because they don'ttake into accountextra-legalregulations.InstitutionalShareholder
Services provides a Corporate Governance Quotientfor a large sample ofinternational
companies. Aggarwal, Erel, Stulz, and Williamson (2007) used it to construct two in
corporate governance quality at the national level. One of them focuses on seven
governance characteristics that have received the most attention in the academic
use this index (GOVERN)as a proxy forthe quality ofnationalcorporate governance
systems.
6.2. The influence of market forces
It is traditionalto oppose bank and market-oriented financialsystems.In the former,
banks are the main providers ofcompany financing,whereas in the latter,companies
generally prefer raising funds from the market. Ball, Kothari, and Robin (2000); Ba
(2003) argue that in market-oriented countries, the large number of stockholders g
a high demand foraccounting quality,in particularfor more timely incorporation of
economicincomein accounting earnings.Inversely,in bank-oriented countries,the
demand for high-quality accounting data is lower, because information asymmetry
likely to be resolved through insidercommunications with management.This view is
supported by several studies that document the influence of the capital market str
the accounting behavior of companies.Burgstahler etal. (2006) in particular show that
471A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

earnings management is less prevalent in countries with large and highly develope
markets than in bank-oriented economies.
La Porta, Lopez-de-Silanes, Shleifer, and Vishny (1997) proposed three measures
importance of stock market in an economy: market capitalization of minorities/GNP
of listed domestic firms/population;and numberof IPOs/population.Leuz etal. (2003)
synthesized them in a single measure by taking the average rank on these ratios (M
These variables have been extensively used by prior studies to reflect the prevailin
mode in each country. We also use the conglomerate index of financial structure de
by Demirguc-Kunt and Levine (1999) as an alternative measure of the relative impo
banks and markets in the financing of companies. This index (MKIMP2) is a combina
measures of size, activity, and efficiency of capital markets. A high value on this ind
means that a country's financial system relies more on market-oriented financing t
bank-based financing.
7. The results
Table 6 shows the association between earnings management measures and inst
or market characteristics. Panel A reports the Spearman's ρ for the pre-crisis period
2007). Both measures of income smoothing (IS1 and IS2) are significantly correlate
the quality of law enforcement (ENFORC and RULLAW), and the sign of the associat
is as expected:countries with a high levelof law enforcementexhibitlowerincome
smoothing than less rigorous jurisdictions.
Similarly, income smoothing decreases as the quality of corporate governance (G
increases.By contrast,income smoothing seems notconstrained by the levelof investor
protection, except for IS2, which exhibits a negative association with INVPRO. This s
thatlaw does notdissuade managers to manipulate earnings ifnotcompleted by strict
enforcementprovisions and/or extra-legalregulations,such as corporate governance codes.
Consistentwith the intuition thatearnings managementis more frequentin bank-basedfinancing environments, there is a negative association between our income-smoot
(IS1 and IS2) and both measures of financial market importance.
A striking finding is the contrast between the results obtained with income-smoo
measures and those related to other earnings management proxies. Whereas the fo
partially consistent with expectations, there is no significant association between a
quality measures and institutional or market characteristics.
Panel B of Table 6 reports the same analyses for the crisis period (2008–2009). T
are largely consistentwith those of previous years.The main difference is the negative
association between investorprotection,as measured by ANTIDIR,and thesecond
income-smoothing metric IS2,which becomes marginally significant(atthe 5% level).
This suggests that law is more dissuasive for managers during troubled periods.
Nevertheless, it is worth noting that the association is no longer significant for th
measure ofincome smoothing.Anotherdifference is the association between ourfirst
metric of income smoothing (IS1) and the rule of law index (RULLAW) that becomes
significant,while there is stilla significantassociation with the othermeasure oflaw
enforcement (ENFORC).
472 A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

As expected, income smoothing is lower in market-based economies and countr
high law enforcementand high quality corporate governance mechanisms;and these
relations hold irrespective of the period considered (crisis or expansion years). By c
accruals manipulations seem totally independent from country characteristics. Thi
that institutional characteristics and market forces govern the income smoothing bfirms,butdo notaffectthe quality of accruals.Although this resultmustbe taken with
prudence, given the small number of observations, it is, to our knowledge, the first
of such differences between two forms of earnings management.
8. Conclusion
The purpose of this paper was to investigate the influence of macroeconomic co
on income manipulations through an analysis of the earnings managementbehavior of
European-listed companies during the 2008–2009 financial crisis and the years be
main finding isa significantdecrease in income smoothing and an improvementof
accruals quality in the crisis period.
This resulthas severalinterpretations.First,it is possible thatmanagers have less
incentive to manipulate earnings in crisis periods due to a higher market tolerance
performance. It can also be argued that litigation risk increases during crises, whic
dissuade insiders to engage in earnings management. Finally, the change in the be
companies may also respond to a higher demand for more timely earnings in troub
periods.
European countries are relatively homogeneous in economic and political terms
belong to a deeply integrated economic area characterized by the absence of inter
barriers.All EU countries also obey a common regulation intended to favorinternal
competition and movements of capital.Since 2005,all European-listed firms mustalso
comply with IFRS, which minimizes an important factor (accounting standards) ide
by the literature as affecting earnings management.
Despite these commonalties,the intensity of earnings management varies considerab
from one country to another. We show that national characteristics such as law en
corporate governance quality, and importance of financial markets partially explai
differences in income smoothing, which is consistent with a fast-growing body of re
relating accounting quality to institutional factors.
Nevertheless, we also show that country characteristics have no impact on mea
accrualquality.This finding suggests thatwhereas nationalpoliticaland marketforces
shape multi-periodic earnings management such as income smoothing, European
on equal terms with regard to mono-periodic manipulations of accruals. As a conse
furtherresearch dealing with internationaldata should nottreatall forms ofearnings
manipulations indifferently,butclearly distinguish between these two types of earnings
management.
There has been little research on the impact of macro-economic conditions on e
management. Our results can only be compared to those of Chia et al. (2007), who
thatservice-oriented Singapore companies engaged in less earnings managementduring
the 1997 Asian financial crisis, a result consistent with our findings. Contrary to its
predecessor,the 2008 financial crisis had an impact on most countries.It would thus be
473A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

Table 6
Correlations between earnings management measures and institutional/market characteristics.
Earnings management measures Investor protection Law enforcement Corpora
ANTIDIR INVPRO ENFORC RULLAW GOVERN
Panel A: 2006–2007 (N = 32)
Income smoothingIS1 Spearman's ρ − 0.063 − 0.266 − 0.611*** − 0.489*** − 0.424
Sig. (2-tailed) 0.732 0.141 0.000 0.005 0.016
IS2 Spearman's ρ − 0.231 − 0.552*** − 0.593*** − 0.623*** − 0.540
Sig. (2-tailed) 0.203 0.001 0.000 0.000 0.001
Accrual quality JONES1 Spearman's ρ − 0.059 − 0.240 0.089 0.008 − 0.022
Sig. (2-tailed) 0.749 0.186 0.629 0.965 0.905
JONES2 Spearman's ρ − 0.155 − 0.213 0.193 0.079 0.061
Sig. (2-tailed) 0.398 0.241 0.291 0.666 0.740
DD Spearman's ρ − 0.168 − 0.163 0.218 0.134 0.079
Sig. (2-tailed) 0.358 0.373 0.230 0.464 0.666
Panel B: 2008–2009 (N = 32)
Income smoothingIS1 Spearman's ρ 0.011 0.000 − 0.404** − 0.289 − 0.485
Sig. (2-tailed) 0.951 0.998 0.022 0.109 0.005
IS2 Spearman's ρ − 0.354** − 0.416** − 0.454*** − 0.550*** − 0.335*
Sig. (2-tailed) 0.047 0.018 0.009 0.001 0.061

Earnings management measures Investor protection Law enforcement Corpora
ANTIDIR INVPRO ENFORC RULLAW GOVER
Accrual quality JONES1 Spearman's ρ 0.065 0.067 0.119 0.097 0.184
Sig. (2-tailed) 0.724 0.714 0.515 0.598 0.315
JONES2 Spearman's ρ 0.004 0.057 0.200 0.143 0.241
Sig. (2-tailed) 0.982 0.755 0.273 0.434 0.184
DD Spearman's ρ − 0.090 0.004 0.174 0.068 0.181
Sig. (2-tailed) 0.624 0.981 0.341 0.713 0.323
IS1 is the ratio of the standard deviation of CFOjt and the standard deviation of NIjt; IS2 is the Spearman correlation b
between net income and cash flow from operations) scaled by lagged assets and variations in CFOjt, multiplied by minus on
from the modified Jones model: ACCjt = α0 + α11 / Ajt − 1+ α2(ΔREVjt − ΔRECjt) + α3PPENjt + α4ROAjt + εjt (1); JONES
the modified Jones model: ACCjt = α0 + α11 / Ajt − 1+ α2(ΔREVjt − ΔRECjt) + α3PPENjt + α4BMjt + α5CFOjt + εjt (2); D
the modified Dechow and Dichev model: ACCjt = α0 + α1CFOjt − 1+ α2CFOjt + α3CFOjt + 1+ α4(ΔREVjt − ΔRECjt) + α5P
Where ACCjt — accruals (change in non-cash current assets minus change in current liabilities adjuste
amortization expense) scaled by lagged total assets of firm j in year t; Ajt − 1— lagged total assets of firm j in year t − 1jt —firm j in year t; ΔRECjt — change in receivables from clients scaled by lagged total assets of firm j in yearjt — net valu
lagged total assets of firm j in year t; ROAjt — net income scaled by lagged total assets of firm j in year t; BMjt — book-to-mark
flow of firm j in year t scaled by lagged total assets; NIjt — net income of firm j in year t scaled by lagged total a
ANTIDIR = Index of anti-director rights (source: La Porta et al., 1998); INVPRO = Strength of investor
ENFORC = Average of three legal enforcement measures (source: Leuz et al., 2003); RULLAW = Rule
GOVERN = Index of corporate governance quality based on 7 attributes (source:Aggarwalet al.,2007);MKIMP1 = Me
capitalisation held by minorities/GNP; number of listed domestic firms/population; number of IPOs/po
stock market orientation of the country's financial system (source: Demurgic-Kunt and Levine, 1999)
*, **, *** indicate statistical significance at 0.10, 0.05, and 0.01 respectively.

usefulto reiterate the analysis in other geographicalareas to find whether the European
evidence is specific or not.
After measuring the impact of the financial crisis on earnings management, it wo
interesting to examine its effect on other aspects of accounting quality. In particula
decrease of income smoothing in crisis years been associated with an increase of th
relevance and timeliness of accounting figures? Generally speaking,the consequences of
macro-economic changes on the quality of accounting information are largely unex
Interesting developments could result from investigations in this domain,especially if the
economy enters in turbulenttimes characterized by an alternation ofgrowth and crisis
periods.
References
Aggarwal, R., Erel, I., Stulz, R., & Williamson, R. (2007). Do U.S. firms have the best corporate governanc
cross-country examination of the relation between corporate governance and shareholder wealth.Working
paper. European Corporate Governance Institute.
Ahmad-Zaluki, N. A., Campbell, K., & Goodacre, A. (2011). Earnings management in Malaysian IPOs: The
Asian crisis,ownership control,and post-IPO performance.The InternationalJournalof Accounting,46,
111–137.
Ahmed,K., Godfrey,J. M., & Saleh,N. M. (2008).Marketperceptions ofdiscretionary accruals by debt
renegotiating during economic downturn. The International Journal of Accounting, 43, 114–138.
Asquith, P., Gertner, R., & Scharfstein, D. (1994). Anatomy of financial distress: An examination of junk-b
issuers. The Quarterly Journal of Economics, 109, 625–658.
Ayers, B. C., Jiang, J., & Yeung, P. E. (2006). Discretionary accruals and earnings management: An analys
pseudo earnings targets. The Accounting Review, 81, 617–652.
Ball, R., Kothari,S. P., & Robin,A. (2000).The effectof internationalinstitutionalfactors on properties of
accounting earnings. Journal of Accounting and Economics, 29, 1–51.
Ball, R., Robin, A., & Wu, J. S. (2003). Incentives versus standards: Properties of accounting income in fou
Asian countries. Journal of Accounting and Economics, 36, 235–270.
Ball, R., & Shivakumar,L. (2005).Earningsquality in U.K.private firms:Comparative lossrecognition
timeliness. Journal of Accounting and Economics, 39, 83–128.
Barth, M. E., Landsman, W. R., & Lang, M. H. (2008). International accounting standards and accounting q
Journal of Accounting Research, 46, 467–498.
Beekes, W., Pope, P., & Young, S. (2004). The link between earnings timeliness, earnings conservatism a
composition: Evidence from the UK. Corporate Governance: An international review, 12, 47–59.
Bertomeu, J., & Magee, R. (2011). From low-quality reporting to financial crises: Politics of disclosure regu
along the economic cycle. Journal of Accounting and Economics, 52, 209–227.
Brown,W. D., He, H., & Teitel,K. (2006).Conditionalconservatism and the value relevance of accounting
earnings: An international study. European Accounting Review, 15,605–626.
Burgstahler,D., & Dichev,I. (1997).Earnings management to avoid earnings decreases and losses.Journal of
Accounting and Economics, 24, 99–126.
Burgstahler,D., Hail,L., & Leuz,C. (2006).The importance of reporting incentives: Earnings management in
European private and public firms. The Accounting Review, 81, 983–1016.
Bushman,R. M., & Piotroski,J. D. (2006).Financialreporting incentives forconservative accounting:The
influence of legal and political institutions. Journal of Accounting and Economics, 42, 107–148.
Capkun, V., Cazavan-Jeny, A., Jeanjean, T., & Weiss, L. A. (2011). Setting the bar: Earnings management
a change in accounting standards. Working paper (available on www.ssrn.com).
Charitou, A., Lambertides,N., & Trigeorgis, L.(2007). Earnings behaviour of financially distressed firms: The
role of institutional ownership. Abacus, 43, 271–296.
Chen,F., Hope,O. -K., Li, Q., & Wang,X. (2011).Financialreporting quality and investmentefficiency of
private firms in emerging markets. The Accounting Review, 86, 1255–1288.
476 A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

Chia, Y. M., Lapsley, I., & Lee, H. -W. (2007). Choice of auditors and earnings management during the Afinancial crisis. Managerial Auditing Journal, 22, 177–196.
Core,J., Guay,W., & Verdi,R. (2008).Is accruals quality a priced risk factor? Journalof Accounting and
Economics, 46, 2–22.
Daske, H., Hail, L., Leuz, C., & Verdi, R. (2008). Mandatory IFRS reporting around the world: Early eviden
the economic consequences. Journal of Accounting Research, 46, 1085–1142.
DeAngelo, H., & DeAngelo, L. (1991). Union negotiations and corporate policy. Journal of Financial Econo
30, 3–43.
DeAngelo,H., DeAngelo,L., & Skinner,D. J. (1994).Accounting choice in troubled companies.Journalof
Accounting and Economics, 17, 113–143.
Dechow,P. M. (1994).Accounting earnings and cash flows as measures offirm performance:The role of
accounting accruals. Journal of Accounting and Economics, 18, 3–42.
Dechow, P. M., & Dichev, I. (2002). The quality of accruals and earnings: The role of accrual estimation e
The Accounting Review, 77, 35–59.
Dechow, P. M., Sloan, R. G., & Sweeney, A. P. (1995). Detecting earnings management. The Accounting
70, 193–225.
DeFond,M., Hung, M., & Trezevant,R. (2007).Investorprotection and the information contentof
annualearnings announcements:Internationalevidence.Journal of Accounting and Economics,43,
37–67.
DeFond,M. L., & Jiambalvo,J. (1994).Debtcovenantviolation and manipulation ofaccruals.Journal of
Accounting and Economics, 17, 145–176.
Degeorge, F., Patel, J., & Zeckhauser, R. (1999). Earnings management to exceed thresholds. Journal of
72, 1–33.
Dichev,I. D., & Skinner,D. J. (2002).Large-sample evidence on the debtcovenanthypothesis.Journalof
Accounting Research, 40,1091–1123.
Demirguc-Kunt,A., & Levine, R. (1999).Bank-based and market-based financialsystems:Cross-country
comparisons. Policy Research Working Paper 2143. The World Bank.
Durnev, A., & Kim, E. H. (2005). To steal or not to steal: firm attributes, legal environment, and valuatio
Journal of Finance, 60, 1461–1493.
Francis,J., LaFond,R., Olsson,P., & Schipper,K. (2005).The market pricing of accruals quality.Journal of
Accounting and Economics, 39, 295–327.
Guidry, F., Leone, A. J., & Rock, S. (1999). Earnings-based bonus plans and earnings management by bu
managers. Journal of Accounting and Economics, 26, 113–142.
Haw, I. -M., Hu, B., Hwang, L. -S., & Wu, W. (2004). Ultimate ownership, income management, and legal
extra-legal institutions. Journal of Accounting Research, 42, 423–462.
Holthausen,R. W., Larcker,D. F., & Sloan,R. G. (1995).Annualbonus schemes and the manipulation of
earnings. Journal of Accounting and Economics, 19, 29–74.
Huijgen, C., & Lubberink, M. (2005). Earnings conservatism, litigation and contracting: The case of crossfirms. Journal of Business Finance and Accounting, 32, 1275–1309.
Iatridis, G., & Kadorinis, G. (2009). Earnings management and firm financial motives: A financial investig
UK listed firms. International Review of Financial Analysis, 18, 164–173.
Jenkins, D. S., Kane, G. D., & Velury, U. (2009). Earnings conservatism and value relevance across the b
cycle. Journal of Business Finance and Accounting, 36, 1041–1058.
Johnson,M. F. (1999).Business cycles and the relation between security returns and earnings.Review of
Accounting Studies, 4, 93–117.
Jones, J. J. (1991). Earnings management during import relief investigations. Journal of Accounting Resea
193–228.
Jones,K. L., Krishnan,G. V., & Melendrez,K. D. (2008).Do models of discretionary accruals detect actual
cases of fraudulent and restated earnings? An empirical analysis. Contemporary Accounting Researc
499–531.
Kane, G. D., Richarson, F. M., & Graybeal, P. (1996). Recession-induced stress and the prediction of corp
failure. Contemporary Accounting Research, 13, 631–650.
Kim, D., & Qi, Y. (2010).Accruals quality,stock returns,and macroeconomic conditions.The Accounting
Review, 85, 937–978.
477A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478

Kinnunen,J., & Koskela,M. (2003).Who is Miss World in cosmetic earnings management? A cross-national
comparison of small upward rounding of net income numbers among eighteen countries. Journal of In
Accounting Research, 2, 39–68.
Klein, A. (2002).Audit committee,board ofdirectorcharacteristics,and earnings management.Journal of
Accounting and Economics, 33, 375–400.
Kothari, S. P., Leone, A., & Wasley, C. (2005). Performance matched discretionary accrual measures. Jour
Accounting and Economics, 39, 163–197.
Krishnan, G. V. (2007). Did earnings conservatism increase for former Andersen clients? Journal of Accou
Auditing, and Finance, 22, 141–163.
La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. (1998). Law and finance. Journal of Political Economy, 106
1113–1155.
La Porta, R., Lopez-de-Silanes, F., Shleifer, A., & Vishny, R. (1997). Legal determinants of external finance
Journal of Finance, 52, 1131–1150.
Lang, M.,Raedy, J.S., & Wilson, W.(2006). Earnings management and cross listing: Are reconciled earnings
comparable to US earnings? Journal of Accounting and Economics, 42, 255–283.
Larcker, D. F., & Richardson, S. A. (2004). Fees paid to audit firms, accrual choices, and corporate govern
Journal of Accounting and Economics, 42, 626–658.
Leuz, C., Nanda, D., & Wysocki, P. D. (2003). Earnings management and investor protection: An internati
comparison. Journal of Financial Economics, 69, 505–527.
Lim, S., & Matolcsy, Z. (1999). Earnings management of firms subjected to product price controls. Accoun
and Finance, 39, 131–150.
McNichols, M. (2002). Discussion of the quality of accruals and earnings: the role of accrual estimation er
The Accounting Review, 77, 61–69.
Nabar,S., & Boonlert-U-Thai,K. K. (2007).Earnings management,investor protection,and nationalculture.
Journal of International Accounting Research, 6, 35–54.
Navissi,F. (1999).Earningsmanagementunderprice regulation.Contemporary Accounting Research,16,
281–304.
Peasnell, K. V., Pope, P. F., & Young, S. (2000). Accruals management to meet earnings targets: UK evide
and post-Cadbury. British Accounting Review, 32, 415–445.
Peltzman,S. (1976).Toward a more generaltheory ofregulation.The Journalof Law and Economics,19,
211–240.
Rajgopal, S., & Venkatachalam, M. (2011). Financial reporting quality and idiosyncratic return volatility. Jo
of Accounting and Economics, 51, 1–20.
Rangan, S. (1998). Earnings management and the performance of seasoned equity offerings. Journal of F
Economics, 50, 10–122.
Renders,A., & Gaeremynck,A. (2007).The impactof legaland voluntary investorprotection on the early
adoption of International Financial Reporting Standards (IFRS). De Economist, 155, 49–72.
Richardson, F. M., Kane, G. D., & Lobingier, P. (1998). The impact of recession on the prediction of corpor
failure. Journal of Business Finance and Accounting, 25, 167–186.
Saleh,N. M., & Ahmed,K. (2005).Earningsmanagementof distressed firmsduring debtrenegotiation.
Accounting and Business Research, 35, 69–86.
Shleifer, A., & Vishny, R. (1992). Liquidation values and debt capacity: A market equilibrium approach. Jo
of Finance, 47, 1343–1366.
Sweeney, A. P. (1994). Debt-covenant violations and managers' accounting responses. Journal of Accoun
Economics, 17, 281–308.
Teoh, S. H., Welch, I., & Wong, T. J. (1998). Earnings management and the underperformance of seasone
offerings. Journal of Financial Economics, 50, 63–99.
Teoh,S. H., Wong, T. J., & Rao, G. R. (1998). Are accruals during initial public offerings opportunistic? Revi
of Accounting Studies, 3, 175–208.
Wagenhofer,A. (2011).Towards a theory of accounting regulation:A discussion of the politics of disclosure
regulation along the economic cycle. Journal of Accounting and Economics, 52, 228–234.
Willekens,M., & Bauwhede,H. (2003).Auditorreporting conservatism asa defence mechanism against
increased post-Enron litigation risk. Working paper. Belgium: Catholic University of Leuven.
478 A. Filip, B. Raffournier / The International Journal of Accounting 49 (2014) 455–478