Corporate accruals quality during the 2008–2010 Global Financial Crisis

Accepted Manuscript Title: Corporate Accruals Quality During the 2008-2010 Global Financial Crisis1 Author: Neal Arthur Qingliang Tang Zhiwei (Stanley) Lin PII: DOI: Reference:

S1061-9518(15)00025-7 http://dx.doi.org/doi:10.1016/j.intaccaudtax.2015.10.004 ACCAUD 205

To appear in:

Journal of International Accounting, Auditing and Taxation

Received date: Revised date: Accepted date:

24-4-2015 2-10-2015 3-10-2015

Please cite this article as: Arthur, N., Tang, Q., and Lin, Z. S.,Corporate Accruals Quality During the 2008-2010 Global Financial Crisis1 , Journal of the Chinese Institute of Chemical Engineers (2015), http://dx.doi.org/10.1016/j.intaccaudtax.2015.10.004 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Corporate Accruals Quality During the 2008-2010 Global Financial Crisis1

ip t

Neal Arthur

School of Business, The University of Sydney, Australia

cr

Email: [email protected] Qingliang Tang

us

School of Business, Western Sydney University Email: [email protected]

an

Zhiwei (Stanley) Lin

[Corresponding author]

M

College of Economics, Shenzhen University Nanshan District, Shenzhen, 518060, China

Ac ce p

te

d

Email: [email protected]

1

Zhiwei (Stanley) Lin acknowledges the support of grants from the National Natural Science Foundation of China

(No.71372215 and No. 71302195) and financial support from Shenzhen University (No. 836-00008108). Qingliang Tang acknowledges financial support from the School of Business, Western Sydney University. 1

Page 1 of 36

Corporate Accruals Quality

cr

ip t

During the 2008-2010 Global Financial Crisis

us

ABSTRACT

This study investigates the earnings quality (measured by accruals quality) of European firms during

an

the 2008–2010 financial crisis. Prior literature suggests that the quality of earnings would be low during the crisis. However, we found that the accounting standards-setting bodies and regulators of the capital market showed great concern regarding financial reporting policy, so we expected that there would be strong

M

incentives for firms to improve earnings quality during an economic recession. In the present study, we compare the earnings quality of sample firms in 14 European countries during the 2005–2007 period and

d

during the financial crisis period (2008–2010) and find that the sample firms tended to present higher-quality

te

financial reports during the financial crisis than prior to it. This finding suggests that reduced investor

Ac ce p

confidence and market liquidity engendered by the financial crisis motivated management to strategically enhance earnings quality in an attempt to increase investor confidence and reduce the negative impact of the economic recession. Our results are robust after taking into account other firm-level and country-level factors that may affect earnings management incentives and behavior. Key words: Accruals strategy and incentives; Financial crisis; Earnings management; Accruals quality JEL Classification: M41

2

Page 2 of 36

1. INTRODUCTION The global financial crisis, which began in September 2008, severely shook investor confidence, affected corporate earnings and stock market performance, and caused the failure of many firms. There has been

ip t

passionate debate regarding the role of accounting in the crisis. This debate concerns whether the accounting system, which used fair value measurement, decreased or increased investor confidence and thus

cr

either mitigated or exacerbated the crisis (Lin et al., 2013). This debate is based on the assumption that investors are concerned about the quality of financial reporting and use financial reporting information to

strategy to manage investor confidence during a financial crisis?

us

make decisions. If this is true, another question arises: would management strategically adjust their accruals

an

The extant literature indicates that the financial crisis was a crisis in investor confidence; investors were panicking and eager to sell shares (Okonjo-Iweala et al., 1999; Statman, 1999). User confidence was crucial,

M

and management was aware that financial reporting would affect users’ perceptions of a firm’s financial condition and operating performance. Would these factors give strong incentives for firms to manage

d

earnings to enhance investor confidence? This question is important, because an understanding of corporate

te

accruals strategy, behavior, and incentives during the crisis will help identify robust accounting practices and systems that better accommodate the needs of users and firms that struggle to survive periodically occurring

Ac ce p

financial downturns. This debate motivates the current study. Our evidence is likely to be relevant for accounting standard–setting bodies around the world(for example, the IASB and the FASB), which are considering the impact of accounting on the financial crisis so that accounting standards, such as fair value standards particularly, might be revised appropriately. Firms tend to adjust their operating strategy when facing pressure to survive a crisis, and financial reporting policy is integral to a firm’s overall business strategy. Our research question is: in response to the economic recession, did firms adjust their financial reporting policy to boost accounting performance? On the one hand, accounting performance was likely to be poor. Consequently, firms had a stronger incentive to use less conservative accounting methods and engage in more aggressive earnings management, so that they could show a more positive picture to investors. It is also possible that managers of firms that suffered a loss would be willing to “take a bath” to report a bigger loss. If this argument is true, we would observe lower 1

Page 3 of 36

earnings quality during the financial crisis. Investor confidence was lower during the financial crisis, causing illiquidity in the share market (Hameed et al., 2010). Investor confidence was so crucial that governments around the world spent hundreds of

ip t

billions of dollars to stabilize markets and restore investor confidence. Similarly, firms spared no effort in their attempts to maintain investor confidence. The literature provides evidence that financial reports affect

cr

investor decision-making (Ball and Brown, 1968; Beaver, 1968). Given that the quality of financial reporting is associated with investor confidence and the priority of management during the financial crisis was to

us

increase confidence, we expect that managers might adopt a more transparent financial reporting strategy to restore investor confidence and improve liquidity. Thus, we would observe higher quality financial reporting

an

during the financial crisis.

The present research looks for evidence that would be useful for an understanding of firms’ financial

M

reporting policy, strategy, and quality. Following previous studies, we use the magnitude of discretionary accruals (measured using the modified Jones model and the performance-controlled modified Jones model)

d

as a proxy for the degree of transparency (quality) of financial reporting. Our sample includes the financial

te

statements of listed European (EU) firms from 2005 to 2010. Our results show that EU firms, on average, engaged in less earnings management and presented higher quality accounting information during the

Ac ce p

financial crisis period than during the pre-crisis period. The evidence suggests that management had a strong incentive to increase investor confidence during the financial crisis by providing reliable financial statements that would reduce information asymmetry and improve market liquidity so as to mitigate the negative impact of the economic recession. Our results remain robust when we control for factors that affect firms’ earnings management incentives (including firm size, leverage, and profitability, etc.). This study makes a number of contributions to the literature. First, most of the previous research on financial reporting during the financial crisis explored the impact of fair value measures on the valuation of firm equity (e.g., Lin et al., 2013). The debate on fair value accounting has appeared to ignore the role of managerial incentives. In contrast, the current study examines the financial reporting strategy of management in maintaining investor confidence. Our findings complement prior studies by showing that managerial incentives play an important role in mitigating the negative effects of the crisis. Second, we found 2

Page 4 of 36

limited published research on the impact of the financial crisis on earnings quality and most were conducted in a domestic setting.2Our investigation in an international context is more comprehensive. The purpose of our multinational research design is to reflect the global nature of the crisis. Thus, our findings contrast with

ip t

those using a national setting and provide more insight into the relationship between economic conditions and financial reporting. Third, most previous research on earnings management incentives was conducted

cr

during normal economic conditions and under the assumption of “business as usual.”3However, during an economic downturn, the assumption of “business as usual” does not hold. We therefore examine financial

us

reporting strategy in a unique business environment and find that management adjusted earnings strategy in response to the changed economic environment. This issue was not considered in previous studies, and the

an

results are potentially useful to predict managerial behavior during future periodic economic recessions. The remainder of the paper is organized as follows. Section 2 develops the hypotheses, and section 3

M

discusses the research methodology. Section 4 presents the empirical results, and the final section offers some conclusions.

te

2.1. Literature Review

d

2. LITERATURE REVIEW and HYPOTHESIS DEVELOPMENT

Prior literature suggests that earnings management is intended to alter the perception of users by

Ac ce p

overstating (or understating) a firm’s earnings (Healy and Wahlen, 1999). Earnings management reduces the quality of financial reporting and increases information asymmetry and opacity. Managers have various incentives to engage in earnings management (Dechow et al., 2010), including reducing political costs and the probability of breaching debt contracts, increasing managerial compensation (Watts and Zimmerman, 1990),and boosting share price. If managers believe that investors are unable to detect opportunistic behavior(Abarbanell and Bernard, 1992; Ball and Bartov, 1996; Bernard and Thomas, 1990, 1989; Sloan, 1996), they will take advantage of the inherent subjectivity in accounting assumptions and the flexibility of standards to achieve earnings goals (Ahmed et al., 1999; Holthausen and Verrecchia, 1990; Healy and Palepu, 1993). 2

See next session on literature review for studies on financial crisis and earnings management.

3

Most of the research on earnings management was done in the context of normal economic conditions, rather than during a

period of financial depression (e.g.,Yoon and Miller, 2002; Teoh and Wong, 1998a; Leuz et al., 2003). 3

Page 5 of 36

Prior studies have documented evidence that a financial crisis can impact companies’ financial reporting policies. For example, Chia et al. (2007) report a lower level of total accruals of Singapore companies subsequent to the 1997 Asian financial crisis. Vichitsarawong et al. (2010) find that conservatism and

ip t

timeliness of earnings reporting in Hong Kong, Malaysia, Singapore, and Thailand during the 1997 Asian financial crisis period are low, but improved in the post-crisis period due to corporate governance reforms.

cr

They also find that financial turmoil negatively impacted the value relevance of accounting information such as book values and earnings figures of Thai companies during the Asian financial crisis (Graham et al 2000).

us

Davis-Friday et al (2006) provide similar findings on the value relevance of earnings and book value in four Asian countries, Indonesia, South Korea, Malaysia and Thailand, in the period surrounding the Asian financial

an

crisis. The results suggest the level of corporate-governance mechanisms as well as accounting systems (i.e., IAS or tax-based) have an impact on the extent of changes in the value relevance of book values (but not

M

earnings) resulting from the crisis.

For the recent 2008-2010 Global Financial Crisis, prior studies continue to present empirical evidence on

d

the effect of financial crisis on corporate financial reporting strategy. The results are mixed. For instance,

te

Habib et al. (2012) examines financial reporting by financially distressed firms in New Zealand and finds no effect of the global financial crisis on discretionary accruals. Rusmin et al. (2013) provide evidence of

Ac ce p

increased income smoothing by management of Asian transport firms during the global financial crisis. However, Filip and Raffournier (2014) examine the impact of the global financial crisis on the earnings management behaviour of European-listed firms and find that earnings management significantly decreased in the crisis years. Lin, Jiang, Tang and He, et al (2014) examine the role accounting plays in the financial crisis and specifically investigate the impact of the quality of financial reporting (as measured via earnings quality) on liquidity (measured by the bid-ask spread) in the equity market during the global financial crisis in the United Kingdom. They find that market liquidity was much lower during the crisis than prior to the crisis which was expected. However, firms with high-quality financial reporting suffered fewer negative effects as a result of the financial crisis. The findings support the notion that high-quality earnings information can reduce information asymmetry and hence enhance investor confidence during a financial crisis. In sum, a common characteristic of these studies, which is distinct from other prior literature, is that 4

Page 6 of 36

they take into account the macroeconomic environment of the firms. In other prior studies, general economic conditions are often held constant, so as not to influence incentives for earnings management. In contrast, the above mentioned research emphasizes that dramatic changes in the economic climate have an

ip t

impact on the firm's propensity to adjust earnings reporting policy, as well as its impact on the usefulness of accounting information. However, many questions remain unanswered. For instance, why and how did

cr

managers strategically modify their financial reporting policy during the crisis? In other words, it is not clear what the motivations are behind the phenomenon of a reduced earnings management. We attempt to

us

explore this unresolved issue using a different research design to address these questions. 2.2 Accruals quality during financial crisis period versus non–financial crisis period

an

The global financial crisis that began in 2008 was an unusual event with a complex impact on financial reporting practices. On the one hand, it can be argued that the quality of financial reporting would be low

M

during a period of economic recession. This is because a firm’s financial performance generally would be poor during a crisis. Managers tend to mask true firm performance to minimize the negative market reaction to

d

unexpected changes in earnings (Ball and Brown, 1968). In other words, since during a crisis firms experience

te

enormous financial difficulties and pressure, they have incentives to manage earnings to boost their reported performance and survive the economic downturn. In addition, it has been claimed that the accounting

Ac ce p

system was opaque during the recent crisis due to the use of fair value accounting, a method that might increase information asymmetry (Lin et al., 2013, Laux and Leuz, 2009). On the other hand, some factors during a financial crisis that are unlikely to exist under normal economic conditions may lead to enhanced quality of financial reporting. First, for many firms, low earnings performance was largely expected by investors; this would reduce managers’ incentives to artificially inflate profits (Filip and Raffournier 2014). Second, the relationship between equity-based management 4

compensation and earnings management may weaken during a crisis period , reducing the incentive for earnings manipulation. Third, one of the major problems during the 2008 financial crisis was market illiquidity, caused by the lack of confidence of investor, who sold shares in a panic (Lin et al., 2013). The lack of

4

See Lucian and Holger (2009) for a discussion of the role of executive compensation as a possible cause of the 2008 global

financial crisis. 5

Page 7 of 36

confidence was associated, at least in part, with information asymmetry between informed managers and uninformed investors (Lin et al., 2013), which significantly increased uncertainty regarding firms’ ongoing situations. Information asymmetry exists at all times but is greater during a financial crisis. If firms’ financial

ip t

reporting is opaque, this further decreases investor confidence. Managers who are concerned about investor confidence have powerful incentives to provide reliable financial information. This is consistent with the

cr

literature showing that more credible earnings reports reduce information asymmetry and enhance investor confidence and, thus, market performance (e.g., Teoh and Wong, 1993). In other words, when investors are

us

sensitive to earnings quality, management uses a less aggressive financial reporting strategy and gives clear signals to the market of actual economic performance. It is apparent that the sensitivity of confidence of

an

investors to earnings quality should be higher during a financial crisis period.

Fourth, auditor incentives during a financial crisis should result in less earnings management. Auditors

M

are likely to perceive an increased probability of client bankruptcy and hence litigation risk. Thus, auditors will be more conservative, which may lead to a greater tendency to issue qualified audit reports (Xu et al.,

d

2011), which would, in turn, curb earnings management (Pinnuck, 2012). Finally, accounting

te

standards–setting bodies, such as the IASB and the FASB, and capital market regulators, such as the U.S. Securities and Exchange Commission, undertook a series of actions during the crisis to improve financial 5

Ac ce p

reporting quality in hopes of restoring investor confidence. This in turn led to considerable pressure on managers and established strong incentives to enhance the credibility of financial reports, because lower earnings quality and greater uncertainty may result in additional monitoring by external stakeholders, especially financial market regulators (Chia et al., 2007). In sum, the global financial crisis was a unique event. During the crisis, the government, the general public, investors, the accounting profession and financial market rule makers showed unusually great concern and high expectations for corporate financial reporting. Such high expectations are not generally seen during normal business and economic periods. As a result, we would expect to observe an increased level of accruals 5

For example, the IASB modified its accounting standards for fair value accounting, and the Securities and Exchange

Commission conducted a study on mark-to-market accounting at the beginning of the global financial crisis in late 2008 (Lin et al., 2013). These actions sent a clear message that these bodies were seriously concerned about the impact of financial reporting on investor confidence. 6

Page 8 of 36

quality during the crisis period. This discussion leads to our first hypothesis: H1: Accruals quality was higher during the global financial crisis period than during the pre–financial crisis

ip t

period.

2.3 Access to capital markets and earnings management during the financial crisis

cr

Firms with growth opportunities are expected to raise capital to fund future investments. Evidence from the period before the crisis suggests that managers have greater incentives to manage earnings if they plan

us

to raise capital (e.g., Teoh et al., 1998a). However, it was more difficult to raise funds during the financial crisis than prior to the crisis, as lending was at significantly lower levels (Ivashina and Scharfstein, 2010).

an

Furthermore, in this economic environment, many new investments would be inherently more risky. Together, this suggests that lenders were more conservative and diligent in evaluating loan applications

M

during the crisis period than prior to it.

Prior evidence also indicates that the probability of auditor litigation increases as clients report more

d

income-increasing accruals (Henininger, 2001). Auditors of growing firms seeking to expand through raising

te

capital during the financial crisis would thus be concerned, as lenders rely on financial statements in making

Ac ce p

their lending decisions and because of the potential litigation costs arising from earnings management. Our second research hypothesis is as follows: H2: Firms with high funding needs engaged in less earnings management during the financial crisis.

3. RESEARCH DESIGN AND SAMPLE SELECTION 3.1 Empirical models

We use the following model to test H1 (refer to Table 1 for variable definitions): AQi,t = B0 + 1 CRISISi,t + 2ISSUEi,t + 3BIG4i,t + 4GROWTHi,t +5SIZE,I,t + 6LEVi,t +β7LOSSi,t +FixedEffects + i,t (1) We use the accruals quality proxy AQi,t as the dependent variable, which is measured as the absolute value of estimated discretionary accruals (this estimate is described in the next section). The first independent

7

Page 9 of 36

variable of interest is CRISIS, which equals 1 if the observation is from the period 2008 to 2010 inclusive and 0 otherwise. H1 predicts that 1 is negative. To test H2, we extend equation (1) to equation (1a): AQi ,t   0  1CRISIS i ,t   2 ISSUEi ,t   3CRISIS i ,t ×ISSUEi ,t   4 BIG 4i ,t   5GROWTH i ,t

ip t

  6 SIZE   7 LEVi ,t  8 LOSS i ,t  FixedEffects   i ,t

(1a)

cr

In equation (1) and (1a), we create a dummy variable (ISSUEi,t), which equals 1 if the proceeds from equity or debt issuance are larger than 5% of total assets and 0 otherwise (Teoh et al., 1998a,b).6 The control variables

us

(BIG4, GROWTH, SIZE, LEV, and LOSS) are explained and defined below in section 3.2.3. We also control for both country and (two-digit) industry fixed effects. Based on H2, we would expect the coefficient of

an

interaction term ( CRISIS×ISSUE ),  3 to be negative. According to extant literature, we expect the coefficient of ISSUE (  2 ) to be positive.

M

Prior evidence suggests that when performance is extremely poor, management “takes a big bath” (e.g., Jordan and Clark, 2011; Yoon and Miller, 2002). An economic crisis may thus prompt managers to take a bath,

d

implying that during a financial crisis there will be more negative accruals than during a non-crisis period. We

te

used the following model to determine whether firms’ financial reporting strategy during the financial crisis was different for positive and negative discretionary accruals.

Ac ce p

DA+i,t (DA-i,t) =1 CRISISi,t + 2ISSUEi,t + 3BIG4i,t + 4GROWTHi,t 5LEVi,t + 6LOSSi,t +FixedEffects + i,t(2) DAi+,t ( DAi,t ) is the estimated amount of positive (negative) discretionary accruals. We estimated model 2

separately for firm-year observations with positive and negative values for discretionary accruals. 3.2 Measurement of accruals quality

As previously discussed, accounting discretion allows managers to use earnings management strategies (e.g., to move earnings closer to targets or to avoid losses). Prior studies used discretionary accruals to 7

measure the extent of earnings management (Jones, 1991; Leuz et al., 2003; Dechow et al., 2010). Because earnings management can use income-increasing or income-decreasing accruals to meet earnings targets, in the current study we use the magnitude of the absolute amount of discretionary accruals as the proxy for 6

All financial data are measured in a common currency (US dollar).

7

Earnings management is the alteration of companies’ reported accounting numbers by insiders to either mislead

stakeholders or to influence contractual outcomes (Healy and Wahlen, 1999). 8

Page 10 of 36

8

earnings management (Chen et al., 2010; Reynolds and Francis, 2000; Wang, 2006). A higher magnitude of absolute discretionary accruals would therefore indicate a greater level of earnings management (and lower accruals quality).

ip t

Discretionary accruals are defined as total accruals minus estimated normal accruals. We use two widely used discretionary accruals models to estimate normal accruals: (1) the modified Jones model and (2) the

cr

modified Jones model with control for performance (Dechow et al., 1995, 1996; Kothari et al., 2005). 9

Both models are explained in the following two sections. 3.2.1 The modified Jones model

us

Discretionary accruals are estimated as the residuals from cross-sectional discretionary accruals models.

an

Following Dechow et al. (1995) and DeFond and Subramanyam (1998), we estimated normal accruals using the modified Jones model as follows:

M

TAi ,t   1 (1 Assetsi ,t -1 )   2 ( REVi ,t  RECi , t )   3 PPEi ,t   i ,t

TAi ,t is total accruals scaled by lagged total assets for firm i in year

t

(3)

, where total accruals is the difference

d

between income before extraordinary items and operating cash flows; Assetsi ,t -1 is the year-end total assets

te

for company i in year t-1; REVi ,t is the change in sales from year t-1 to year accounts receivable from year t-1 to year

t

t

; RECi ,t is the change in

; and PPEi ,t is gross property, plant, and equipment. REVi ,t ,

Ac ce p

RECi ,t , and PPEi ,t are scaled by Assetsi ,t -1 . We estimate the coefficients of model (3) from cross-sectional

industry regressions by two-digit Standard Industrial Classification (SIC) groups for each year. We require a minimum of 20 observations for each two-digit SIC group for each year. The residual from model (3) is the estimate of discretionary accruals, and the absolute value of the residual is referred to as A Q A Q _ M JM

_ M JM

. A higher

therefore indicates more earnings management (and lower earnings quality).

3.2.2 The modified Jones model with control for performance Kothari et al. (2005) argue that discretionary accruals estimated from the modified Jones model are

8

For example, managers may use income-increasing accruals to meet earnings thresholds, while they may use

income-decreasing accruals to avoid political costs from regulations. 9

Bartov et al. (2000) found that cross-sectional models were better than time-series models in detecting earnings

management. 9

Page 11 of 36

correlated with growth in operating performance and subject to measurement error. To avoid this problem, we add ROA i,t(model 4) and ROAi , t 1 (model 5) to the modified Jones model (model 3) to estimate normal 10

accruals.

TAi ,t  1 (1 Assetsi ,t -1 )   2 ( REVi ,t  RECi ,t )   3 PPEi ,t   4 ROAi ,t   i , t

ip t

(4)

TAi ,t  1 (1 Assetsi ,t -1 )   2 ( REVi , t  RECi ,t )   3 PPE i ,t   4 ROAi ,t 1   i , t

(5)

ROAi ,t ( ROAi , t 1 ) is income before extraordinary items for firm

in year

(t-1) divided by

t

cr

i

Assetsi ,t -1 ( Assetsi ,t -2 ). The definitions of the other variables in models 4 and 5 are the same as in model 3. The

residual (i.e., discretionary accruals) is referred to as

A Q _ P JM

(

A Q _ P JM 1

). Both

A Q _ P JM

and

when earnings management is taking place (i.e., lower accruals quality is present).

an

A Q _ P JM 1 increase

us

residual from model 4 (and model 5) is the estimate of discretionary accruals, and the absolute value of the

3.2.3 Control variables

M

In the present study we examine accruals practices during the 2008-2010 global financial crisis, which was global in scope. Thus, we believe that an international research setting is more appropriate than a

d

national setting. To strengthen our interpretation of the results of analysis of data from across broad capital

te

markets, we include in our model firm-level and country-level variables that we expect to be related to earnings management incentives. This is particularly important because our sample firms exhibit

Ac ce p

considerable diversity with respect to the national institutional systems with which they interact and with respect to the cultural background of their employees and other stakeholders. Prior research suggests that large auditors are perceived as more independent (DeAngelo, 1981a,b), more experienced, having greater industry expertise (Balsam et al., 2003; Krishnan, 2003), and making larger investments in professional education and training (Dopuch and Simunic, 1982). The “Big Four” auditors are known to be more conservative as a means of protecting their brand names, because they have more to lose 11

than non–Big Four auditors in the event of a loss of reputation (DeAngelo, 1981a,b).

Given this, it is

expected that high-quality auditors (i.e., Big Four auditors) will discourage aggressive reporting of accruals 10

In additional tests (not tabulated), we estimate normal accruals using the performance-matched modified Jones models(i.e.,

models 4 and 5);the results are qualitatively the same (see section 4.2.3, “Robustness tests”). 11

We use the term “Big Four” throughout the paper to refer to Deloitte Touche Tohmatsu, Ernst & Young, KPMG, and

PricewaterhouseCoopers. 10

Page 12 of 36

and persuade management to report economic losses in a timely fashion during a financial crisis. We include the Big Four as a control variable in our models, as a dummy variable that takes on the value of 1 for firm-years in which the company is audited by a Big Four auditor.

ip t

We use GROWTH i ,t , measured as the annual growth rate in sales, to control for sales growth, because the market severely penalizes growth firms that experience unexpected adverse earnings (Skinner and Sloan,

cr

2002). With larger firms, the level of market monitoring is greater and managers have less incentive to

us

manage earnings. We include the natural logarithm of annual sales ( SIZEi ,t ) as a proxy for market monitoring. Highly leveraged firms receive more monitoring from debt holders and therefore have less incentive to manipulate earnings (DeFond et al., 1994; Watts et al., 1990). We therefore include leverage ( LEVi ,t ) to

an

control for this effect. Managers of companies experiencing negative earnings may reduce earnings even further through the use of discretionary write-downs (Healy, 1985; Healy et al., 1999). We therefore

M

include a dummy variable ( LOSSi ,t ) to indicate whether a firm suffered a loss in year

t

.

We also control for country-level effects. For example, some countries may have introduced a more

d

effective enforcement system or more corporate governance regulations. Therefore, we include country

te

dummies in our model. In addition, we use two-digit industry dummies to control for other unobserved

Ac ce p

variations in the contracting environment of a given company. Differences in institutions and culture

The extant literature shows that financial reporting also varies with certain country-level factors, including the underlying legal protection of investors (La Porta et al., 1998; Leuz et al., 2003) and the level of law enforcement (Hope, 2003). Specifically, Leuz et al. (2003) report that earnings management is less frequent in economies with stronger outside investor rights and better legal enforcement. We control for these factors by including indicator variables as proxies for national levels of investor protection (H_ANTISD) and disclosure requirements (H_DISCLOSE). The indicator variable H_ANTISD is 1 for observations from a country with a value for the anti–self-dealing index above the median level for the index (Djankov et al., 2008) and 0 otherwise, while H_DISCLOSE equals 1 for observations from a country with a value that is above the

11

Page 13 of 36

12

median for the disclosure requirements index (La Porta et al., 2006) and 0 otherwise.

National culture has also been found to influence financial reporting systems and practices. Hofstede (1981) initially proposed four dimensions that could be used to group countries by cultural clusters and

ip t

developed four related measures, referred to as “uncertainty avoidance,” “individualism,” “power distance,” and “masculinity.”“Long-term versus short-term orientation,” and “indulgence” was recently identified as the 13

cr

fifth and sixth cultural dimension .The cultural dimensions developed by Hofstede (1981) were later argued to be related to a number of “value dimensions” in the accounting culture (Gray, 1988). Importantly, Gray

us

(1988) notes that only individualism and uncertainty avoidance are fully linked to all four accounting values and could therefore act as proxies for them. Similarly, Hope (2003) notes that the individualism and

an

uncertainty avoidance dimensions are likely to have the most straightforward implications for managers’ accounting choices. Gray’s model (1988) suggests that, where individualism is the dominant culture,

M

accountants will have the most flexibility in terms of self-governance (professionalism) and measurement. Managers within an individualistic society should also be predisposed to report the most optimistic numbers

d

allowed by institutions. Thus, earnings management in such an individualistic society is probably prevalent,

te

which might lead to a positive association between the magnitude of earnings management and individualism. In addition, under Gray’s model, countries that are highly uncertainty avoidant require more

Ac ce p

uniformity from accountants, with many rules and little self-governance in the selection of options for the accounting profession (statutory control) and what is presented in financial reports (uniformity). In uncertainty-avoidant countries, the rules would support a conservative approach to earnings management 12

Observations were classified based on the company’s country of residence.

13

Research by Michael Bond and colleagues among students in 23 countries led him in 1991 to adding a fifth dimension

called Long- versus Short-Term Orientation. In 2010, research by Michael Minkov allowed to extend the number of country scores for this dimension to 93, using recent World Values Survey data from representative samples of national populations. Long- term oriented societies foster pragmatic virtues oriented towards future rewards, in particular saving, persistence, and adapting to changing circumstances. Short-term oriented societies foster virtues related to the past and present such as national pride, respect for tradition, preservation of "face", and fulfilling social obligations. The sixth dimension, Indulgence versus Restraint is based on Minkov's World Values Survey data analysis for 93 countries. Indulgence stands for a society that allows relatively free gratification of basic and natural human drives related to enjoying life and having fun. Restraint stands for a society that suppresses gratification of needs and regulates it by means of strict social norms(Hofstede et al. 2010). 12

Page 14 of 36

(conservatism) but encourage secrecy. Because of the cultural interest in uniformity, these societies would provide fewer opportunities for earnings management. Subsequently, the link between national culture and accounting choices was empirically tested, and a

ip t

significant association was documented (Nabar and Boonlert-U-Thai, 2007; Han et al., 2010; Doupnik, 2008; Guan et al., 2006). Thus, we control for the effect of cultural differences in our models. We use the index

cr

values reported by Hofstede (2008) for the level of individualism (H_INDIV) and for uncertainty avoidance (H_UA). H_INDIV equals 1 if the observation was from a country with an index value for individualistic culture

uncertainty avoidance above the median and 0 otherwise.14

an

3.3 Sample selection

us

above the median and 0 otherwise. H_UA equals 1 if observations were made in a country with a value for

Our initial sample consists of all listed firms in 14 European countries.15 Our sample period starts in

M

2005 and ends in 2010. The EU and its member states formally adopted IFRS for the preparation of financial statements on January 1, 2005. To prevent this shift in accounting standards from influencing our analyses of

d

earnings quality, we begin our sample period in 2005. We end our sample period in 2010 when global

te

economic conditions were gradually stabilized.

We exclude firms with missing data and firms with negative owners’ equity. Consistent with previous

Ac ce p

studies (e.g., Leuz et al., 2003), financial institutions are excluded from our sample. As noted in section 3.2.1, when calculating discretionary accruals, we exclude industries with fewer than 20 firms. The final sample therefore includes 17,198 firm-year observations, of which 7,913 observations are from the non–financial crisis period (2005–2007) and 9,285 observations are during the financial crisis period (2008–2010). Table 2 16

reports the country distribution of the 17,198 firm-year observations. 14

The models include country fixed effects. We transformed the anti–self-dealing, disclosure, and culture variables to avoid

multicollinearity problems. As part of our additional tests (not tabulated), we substituted the actual values for these control variables; the results were qualitatively unchanged. 15

There were 27 member states in the EU as of December 31, 2010. Following Chen et al. (2010), we excluded the 12 member

states that joined the EU during the period 2004–2007, as these member states have emerging markets and the financial data for most of these countries are not available in the Worldscope database. Luxemburg was also excluded, as no cultural or institutional data were available for this country. 16

The sample was different in the pre- and crisis period. We used an unbalanced sample to avoid survivorship bias in the tests. 13

Page 15 of 36

We selected the EU as our research setting because it was significantly affected by the financial crisis. In addition, our sample includes multiple nations to reflect the global nature of the crisis. Such a research design allows us to examine the impact of the financial crisis beyond national borders, and we draw conclusions

ip t

relating to the overall impact of the financial crisis. Thus, we are able to overcome the limitations of previous research that focused on individual countries. EMPIRICAL RESULTS

cr

4.

4.1 Summary statistics and correlation analysis

us

Table 3 reports the descriptive statistics. All variables (except for the dummy variables) were winsorized at the 1st and 99th percentiles to mitigate the effects of outliers. Panel A of Table 3 presents descriptive

an

statistics for the full sample. Panel B reports that, consistent with H1, the mean values for AQ_MJM, AQ_PJM, and AQ_PJM1 (proxies for accruals quality) were all lower during the financial crisis period than during the

M

pre–financial crisis period (significant at the 1% level).

Table 4 reports the Pearson correlation statistics. We find a significant (at the 1% level) negative

d

correlation between CRISIS and the absolute value of discretionary accruals, which is consistent with the

te

notion that managers increase earnings quality to maintain investor confidence and improve market liquidity during a financial crisis. Further, we find that firms that had large securities issues had significantly higher

Ac ce p

levels of discretionary accruals than other firms. This result is consistent with extant literature that firms manage earnings when they issue capital.

With respect to the control variables, we find, consistent with prior literature, statistically significant correlations of accruals quality with the variables BIG4 (negative) and GROWTH (positive). The correlations between CRISIS and the control variables are all relatively low (below 0.2), suggesting that multicollinearity is 17

not present in the models.

17

Variance inflation factors (VIFs) were calculated for each regression model. For all models, VIF values were less than 5 for the

variables of interest. 14

Page 16 of 36

4.2 Results 4.2.1 Earnings management during the financial crisis We use model 1 to test H1. Table 5 reports the ordinary least squares regression results. We use the

ip t

absolute value of performance-controlled discretionary accruals as a proxy for accruals quality. The higher the value of AQ_PJM1, the higher the estimated amount of earnings management. H1 predicts that the

cr

coefficient of CRISIS should be negative. As reported in Table 5, we find a significant negative coefficient for CRISIS (t=-7.51; P<0.01), which supports the notion that firms reduced earnings management during the

us

financial crisis to maintain investor confidence and improve share liquidity (H1).

We use model 2 to determine whether there was any change in the level of positive and (separately)

an

negative earnings management during the financial crisis. For this test, we separate the sample into two subsamples based on the discretionary accruals (positive or negative). Column 1 in Table 6 presents the

M

regression results from model 1 for firm-years with positive discretionary accruals, while Column 2 in Table 6 presents the regression results for firm-years with negative discretionary accruals. We found that firms

d

engaged in less positive and negative earnings management during the financial crisis relative to the

te

pre-financial crisis years. This is also consistent with H1.

Ac ce p

4.3.2 Cross-sectional differences in earnings management in the financial crisis We next add interaction terms to the equations to examine the effects of interactions between international economic conditions and firm-specific factors on earnings management. Specifically, we examine the interactions of CRISIS with GROWTH, ISSUE, LEV, and BIG4. H2 predicts that firms with greater financing needs engaged in less earnings management during the financial crisis, which suggests that the coefficient of Crisis*ISSUE will be negative. We report the related regression results in Column 1 of Table 7. The coefficient for Crisis*ISSUE is -0.005, which is significant at the 10% level, supporting the predictions of H2. To determine whether institutions and culture affected firms’ earnings management incentives during the financial crisis, we add another set of interaction terms. The related interaction terms are labeled Crisis*H_ANTISD and Crisis*H_DISCLOSE. Columns 2 to 5 of Table 7 report on the effects of culture on firms’ 15

Page 17 of 36

earnings management incentives during the financial crisis. The regression results show that institutions and culture did not have significant effects on firms’ earnings management behavior during the financial crisis. We add masculinity and power distance as control variables and the main results of the paper remain

ip t

qualitatively the same. We then separate the full sample into two subsamples according to discretionary accruals

cr

(positive/negative) and run cross-sectional regressions to determine whether firm characteristics, institutions, and/or culture affected earnings management during the financial crisis. The results, reported in Tables 8 and

us

9, are consistent with the preceding analyses. 4.2.3 Robustness tests

an

We run several tests for robustness on our data. First, we calculate earnings quality using the modified Jones model and the performance-controlled Jones model with current-period ROA and use the two earnings

M

quality proxies (calculated using model 3 and model 4, described in section 3.2) as dependent variables. Second, we include an intercept in the discretionary accruals model (model 5) as an additional control for

d

heteroskedasticity (Kothari et al., 2005). Third, we estimate discretionary accruals models separately for each

te

country-year grouping and used the residuals from these regressions as a proxy for earnings management.

Ac ce p

The results were qualitatively the same following these robustness tests. Taken as a whole, our findings make incremental contributions beyond extant literature on economic crisis and financial reporting practice. First, although previous papers provide evidence on the Asian financial crisis (e.g. Vichitsarawong et al. 2010), we extend this line of research to the global financial crisis, which is much large in scope and deep in its impact on business. Second, Lin, Jiang, Tang and He, et al (2014) find the impact of the quality of financial reporting (as measured via earnings quality) on liquidity (measured by the bid-ask spread) in the equity market during the global financial crisis in the United Kingdom. Lin, Kang, Morris and Tang, (2014) find fair value accounting increases the information asymmetry during the global financial crisis.

But their focus is on the consequence of quality of accounting information. In addition, the

overwhelming majority of companies used historically based earnings measurement rather than fair value

16

Page 18 of 36

based earnings reporting18. Filip and Raffournier (2014) find that earnings management significantly decreased in the crisis years. Our paper provides consistent results. Although the research setting is similar, our paper uses a different

ip t

research design and is guided by different theories. First, model specification and sample size are different. For example, their study covers the period from 2006-2009 with total observations of 8266, while our

cr

sample size is 17,198, covering a longer time span from 2005 to 2010. Their major results are based on univariate tests while we adopt regression models to test our hypotheses. Second, they focus on country level

us

factors, such as national legal system and market force (macroeconomic variables) on earnings management, while we include many firm level variables (i.e. micro-mechanism) such as firm size, leverage, and profitability,

an

etc. Third, we control for more related variables so our results seem to be more robust. For example, their model did not control for national culture which is particularly relevant for an international study on earnings

M

management (Han et al 2010). Fourth and most importantly, our paper not only corroborates their results, but also offers alternative interpretations. They indicate decreased earnings management during the crisis

d

period is possibly due to a higher market tolerance for poor performance or increased litigation risk. However,

te

they did not consider the effect of availability of corporate financial resource and its need and capability of raising capital. This issue is important because during a period of financial crisis, it is extremely difficult to

Ac ce p

raise capital. So the confidence of investors and banks is vital for firms attempting to raise capital, which motivated managers to enhance financial reporting quality in order to maintain investors’ confidence. Therefore, we specifically test this conjecture and the theory is confirmed by our evidence that the improvement of earnings quality is more pronounced in firms that raised capital in troubled periods. We show that financial needs have strengthened the effect of financial crisis on earnings management. Thus, financial reporting quality appears to be used as a strategic tool to improve corporate financing capability. Overall, our study not only confirms the findings of Filip and Raffournier (2014), but provides additional insightful and more convincing evidence.

18

For example, some of the US financial institutions only used fair value to account for about 45% assets and 15% liabilities

before the global financial crisis (SEC report 2009). 17

Page 19 of 36

5.

CONCLUSIONS This paper addresses the issue of whether the quality of financial reporting, as measured using accruals

quality, increased or decreased during the financial crisis. This is an important issue because there has been

ip t

intense debate about the role of the accounting system during the financial crisis. Prior literature suggests that the quality of financial reporting would be low during an economic recession. This is partly because firms

cr

use fair value accounting, which would increase information asymmetry and thus reduce market liquidity (Lin et al., 2013). In addition, it can be argued that there would be stronger incentives for firms to manage

us

earnings to boost their reported performance during a bad economic period.

We use a sample of firms listed on the EU stock exchanges during the pre–financial crisis period

an

(2005-2007) and the financial crisis period (2008-2010) and use accruals quality as a proxy for financial reporting quality. Contrary to predictions from previous studies, our evidence shows that EU firms tended to

M

present higher-quality financial reports during the financial crisis. This evidence suggests that the financial crisis motivated managers to strategically enhance accruals quality (by reducing earnings management) in a

d

bid to improve investor confidence and reduce the negative impacts of the economic recession. We also

te

observe that large firms, high-growth firms, and particularly firms with financing needs tended to report higher-quality earnings. Our results hold after controlling for firm-level variables and variations in

Ac ce p

country-level institutions and cultures.

There are a number of limitations to this study. First, we examine only the 2008 financial crisis. We believe that documentation of the crisis is useful for the understanding of the nature of, and managerial reaction to, this financial recession. Nevertheless, we emphasize that caution should be exercised prior to extending these findings to other financial crises because of differences in factors such as institutional frameworks and cultural norms. Second, because this is an international study, we cannot reach any conclusions regarding individual countries. However, it should be recognized that different countries were affected by the crisis to different degrees. Thus, we do not claim that the results hold for all EU countries. Third, 60% of the sample firms are based in 3 countries (United Kingdom, France, and Germany), and the remaining 11 nations account for 40% of the sample firms. Thus, our results were heavily influenced by the top 3 countries, which have more developed financial market systems. Recognizing this limitation, we add a 18

Page 20 of 36

country fixed effect control (as well as industry effect) to our models and our main findings are unaltered. Despite this, we encourage future research to modify our model specifications to overcome these limitations. Finally, we only consider the difference in earnings quality between a pre-crisis period versus the financial

ip t

crisis period. Future research may examine the difference between the crisis period and a post-crisis period,

Ac ce p

te

d

M

an

us

cr

which might shed additional insight.

19

Page 21 of 36

Table 1 Definitions of Variables Variable

Definition Accruals quality.

AQ_MJM

Absolute value of discretionary accruals. Discretionary accruals are estimated using the cross-sectional modified

ip t

AQ

Jones model. AQ_PJM

Absolute value of discretionary accruals. Discretionary accruals are estimated using the cross-sectional modified

AQ_PJM1

cr

Jones model with ROA.

Absolute value of discretionary accruals. Discretionary accruals are estimated using the cross-sectional modified

CRISIS

us

Jones model with the last-year ROA.

Indicator variable; equals 1 for observations in the crisis period (2008-2010) and 0 for observations in the pre-crisis period (2005-2007).

Indicator variable; equals 1 if the auditor is Big Four and 0 otherwise.

GROWTH

Yearly sales growth rate.

SIZE

Log (market value).

LEV

Leverage, which equals total liabilities divided by total assets.

LOSS

Indicator variable; equals 1 if net income is negative in a given year and 0 otherwise.

ISSUE

Indicator variable; equals 1 if total issuance of equity and debt is larger than 5% of year-end total assets and 0

an

BIG4

M

Market value is defined as stock price times the number of shares outstanding.

otherwise.

Indicator variable; equals 1 if the anti–self-dealing index (Djankov et al., 2008) of the country is higher than the

d

H_ANTISD H_DISCLOSE

te

median in 14 EU countries and 0 otherwise.

Indicator variable; equals 1 if the disclosure requirement index (La Porta et al., 2006) of the country is higher than the median in 14 EU countries and 0 otherwise. Indicator variable; equals 1 if the individualism value (Hofstede, 2008) of the country is higher than the median

Ac ce p

H_INDIV

of the 14 sample EU countries and 0 otherwise.

H_UA

Indicator variable; equals 1 if the uncertainty avoidance value (Hofstede, 2008) of the country is higher than the median of 14 EU countries and 0 otherwise.

20

Page 22 of 36

Table 2 Country Distribution of Firm-Year Observations

Non-Crisis (05-07)

Crisis (08-10)

319 445 2,719 564 525 585 2,671 4,754 1,134 202 940 564 227 1,549 17,198

1.85 2.59 15.81 3.28 3.05 3.4 15.53 27.64 6.59 1.17 5.47 3.28 1.32 9.01 100

149 203 1,318 260 249 281 1,270 2,136 539 93 416 265 109 625 7,913

170 242 1,401 304 276 304 1,401 2,618 595 109 524 299 118 924 9,285

us

cr

ip t

Percent

an

Austria Belgium Germany Denmark Spain Finland France United Kingdom Greece Ireland Italy Netherlands Portugal Sweden

Frequency

M

Country

Ac ce p

te

d

Country Code AUT BEL DEU DNK ESP FIN FRA GBR GRC IRL ITA NLD PRT SWE Total

21

Page 23 of 36

Table 3 Descriptive Statistics Panel A Mean

Median

Standard Deviation

17,198

0.071

0.045

0.082

AQ_PJM

17,198

0.065

0.043

0.071

AQ_PJM1

17,198

0.07

0.044

0.079

DA_MJM

17,198

0

0

0.105

DA_PJM

17,198

0

-0.003

0.093

DA_PJM1

17,198

0

0

0.103

BIG4

17,198

0.63

1

0.483

CRISIS

17,198

0.54

1

GROWTH

17,198

0.167

0.072

LEV

17,198

0.201

0.179

Maximum

0.001

0.439

0.001

0.385

0.001

0.421

-0.36

0.391

-0.279

0.354

-0.344

0.379

0

1

0.498

0

1

0.514

-0.635

3.426

0.17

0

0.65

an

us

AQ_MJM

Minimum

ip t

N

cr

Variable

17,198

18.818

18.527

2.188

14.666

24.549

17,198

0.295

0

0.456

0

1

ISSUE

17,198

0.241

0

0.428

0

1

H_ANTISD

17,198

0.449

0

0.497

0

1

H_DISCLOSE

17,198

0.621

1

0.485

0

1

H_INDIV

17,198

0.668

1

0.471

0

1

H_UA

17,198

0.364

0.481

0

1

d

te

Panel B

M

SIZE LOSS

0

2005-2007

2008-2010

Difference

AQ_MJM

.076

.067

.009***

Ac ce p

Variable AQ_PJM

.071

.060

.011***

AQ_PJM1

.075

.066

.009***

All variables (except for dummy variables) are winsorized at the 1st and 99th percentiles to mitigate the effects of outliers.

*, **, *** indicate significance at the 10%, 5%, and 1% levels, respectively (two-tailed).

22

Page 24 of 36

ip t cr

Pearson Correlation Statistics BIG4

CRISIS

GROWTH

DA_PJM1

0.023***

BIG4

-0.121***

-0.028***

CRISIS

-0.051***

-0.000

-0.010

GROWTH

0.165***

0.069***

-0.016**

-0.093***

LEV

-0.123***

0.012

0.087***

0.034***

-0.061***

SIZE

-0.216***

0.006

0.416***

-0.119***

-0.012

LOSS

0.167***

-0.198***

-0.118***

0.116***

-0.003

ISSUE

0.213***

0.042***

-0.124***

-0.098***

H_ANTISD

0.006

-0.031***

0.062***

0.012

H_DISCLOSE

0.022***

-0.039***

0.060***

0.027***

H_INDIV

0.015*

-0.037***

0.088***

0.027***

H_UA

-0.098***

0.038***

-0.055***

-0.013*

LEV

SIZE

LOSS

ISSUE

H_ANTISD

H_DISCLOSE

0.170***

-0.366***

0.132***

-0.073***

-0.188***

0.196***

0.043***

-0.059***

0.012

0.056***

-0.059***

0.070***

-0.171***

-0.062***

0.055***

-0.059***

0.402***

0.064***

-0.161***

-0.028***

0.032***

-0.075***

0.384***

0.851***

0.215***

0.087***

-0.065***

-0.013*

-0.291***

-0.069***

ed

0.014*

-0.080***

H_INDIV

M an

DA_PJM1

ce pt

AQ_PJM1

us

Table 4

-0.032***

Ac

*, **, *** indicate significance at the 10%, 5%, and 1% levels, respectively (two-tailed).

23

Page 25 of 36

Table 5 Regression Results: Earnings Management Measured Using the Absolute Value of

-0.009*** 0.024*** -0.006*** 0.019*** -0.005*** -0.024*** 0.014*** 0.161*** Y Y 17,198 0.135 0.132

(-7.51) (14.29) (-4.01) (10.02) (-16.18) (-6.20) (9.11) (12.42)

us

cr

T-Stat

an

CRISIS ISSUE BIG4 GROWTH SIZE LEV LOSS INTERCEPT COUNTRY FE INDUSTRY FE N R2 adj. R2

AQ_PJM1

ip t

Abnormal Accruals

M

*, **, *** denote significance at the 10%, 5%, and 1% levels, respectively (two-tailed). Table 6

d

Regression Results: Positive and Negative Discretionary Accruals

-0.009*** 0.030*** -0.007*** 0.023*** -0.007*** -0.014** -0.015*** 0.216*** Y Y

Ac ce p

CRISIS ISSUE BIG4 GROWTH SIZE LEV LOSS INTERCEPT COUNTRY FE INDUSTRY FE

te

(1) DA_PJM1>=0

N R2 adj. R2

T-Stat

(2) DA_PJM1<0

(-5.58) (13.11) (-3.33) (9.74) (-14.59) (-2.49) (-7.01) (10.11)

8,554 0.141 0.135

0.009*** -0.016*** 0.004** -0.012*** 0.003*** 0.039*** -0.037*** -0.106*** Y Y

T-Stat (5.87) (-7.30) (2.34) (-5.11) (8.24) (7.85) (-19.31) (-7.29)

8,644 0.188 0.183

The dependent variable is signed discretionary accruals (DA_PJM1).

*, **, *** denote significance at the 10%, 5%, and 1% levels, respectively (two-tailed).

24

Page 26 of 36

ip t cr

AQ_PJM1

(1)

(2) ***

(3)

***

-0.008

-0.009

***

-0.008

M an

CRISIS

us

Table 7 The Effects of Firm Characteristics, Institutions, and Culture on Absolute Earnings Management During the Financial Crisis

(-6.42)

(-5.53)

(-4.56)

(4)

(5)

-0.009

***

(-4.73)

***

-0.010

(-6.29)

Firm Characteristics

Crisis*ISSUE

-0.005

*

ed

(-1.67)

Institutions and Culture

Crisis*H_INDIV Crisis*H_UA BIG4 GROWTH SIZE

Ac

Crisis*H_DISCLOSE

-0.000 (-0.05)

ce pt

Crisis*H_ANTISD

***

-0.006

(-4.00) 0.018

***

(10.02) ***

-0.001 (-0.37) 0.001 (0.25) 0.002 (1.05) ***

-0.006

(-4.01) 0.019

***

(10.02) ***

***

-0.006

(-4.01) 0.019

***

(10.02) ***

-0.006

***

(-4.01) ***

0.019

(10.02) ***

-0.005

-0.005

-0.005

-0.005

(-16.14)

(-16.18)

(-16.18)

(-16.15)

***

-0.006

(-4.02) 0.019

***

(10.03) ***

-0.005

(-16.19)

25

Page 27 of 36

ip t (-6.20) LOSS

0.014

ISSUE

0.027

(-6.20)

***

0.014

(9.11)

***

-0.024

(-6.21)

0.014

M an

(9.13)

***

cr

***

-0.024

-0.024

us

***

LEV

***

0.024

(11.28)

***

(14.30)

H_ANTISD

***

(9.10)

0.024

***

(14.31)

-0.024

***

(-6.20) ***

0.014

(9.12) ***

***

-0.024

(-6.21) 0.014

***

(9.09) ***

0.024

0.024

(14.29)

(14.30)

-0.003

(-0.55)

H_DISCLOSE

-0.003

H_UA

INDUSTRY FE N R

2

adj. R

2

ce pt

COUNTRY FE

0.160

Ac

INTERCEPT

***

ed

(-0.49)

H_INDIV

0.003 (0.52) *

-0.010

(-1.94) 0.161

***

0.161

***

***

0.162

0.170

***

(12.42)

(12.42)

(12.41)

(12.42)

(13.20)

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

17198

17198

17198

17198

17198

0.135

0.135

0.135

0.135

0.135

0.132

0.132

0.132

0.132

0.132

*, **, *** denote significance at the 10%, 5%, and 1% levels, respectively (two-tailed).

26

Page 28 of 36

ip t cr us

Table 8

The Effects of Firm Characteristics, Institutions and Culture on Positive Earnings Management in the Financial Crisis CRISIS

(1)

(2)

(3)

M an

DA_PJM1>=0

***

-0.006

-0.010

(-3.87)

***

(-4.52)

***

-0.010

(-3.89)

(4)

(5)

-0.010

***

(-3.73)

***

-0.011

(-5.11)

Firm Characteristics **

-0.010

(-2.35)

ed

Crisis*ISSUE

Crisis*H_DISCLOSE Crisis*H_INDIV Crisis*H_UA BIG4

Ac

Crisis*H_ANTISD

ce pt

Institutions and Culture

0.002 (0.47) 0.002 (0.56) (1.46)

***

-0.007 0.023

(0.76)

0.005

(-3.32)

GROWTH

0.002

***

(9.73)

-0.007

***

(-3.33) ***

0.023

(9.73)

***

-0.007

(-3.33) 0.023

***

(9.74)

-0.007

***

(-3.33) 0.023

***

(9.74)

***

-0.007

(-3.33) 0.023

***

(9.74)

27

Page 29 of 36

ip t ***

LEV

-0.014

-0.014

(-2.48)

(-2.49)

(-14.60)

(-14.58) **

***

LOSS

-0.015

-0.015

(-7.00) ISSUE

0.035

(-14.57)

-0.014

**

(-2.49)

M an

**

***

-0.007

***

(-7.00)

***

***

(10.61) H_ANTISD

cr

-0.007

-0.007

-0.007

us

***

SIZE

***

-0.015

(-7.00)

***

0.030

0.030

(13.10)

(13.11)

***

(-14.57) **

***

-0.007

(-14.62) **

-0.014

-0.014

(-2.49)

(-2.48)

-0.015

***

(-6.99) 0.030

***

(13.11)

***

-0.015

(-7.08) 0.030

***

(13.13)

0.007

(0.84)

COUNTRY FE INDUSTRY FE N R

2

adj. R

2

0.213

***

(-0.34) 0.005 (0.59) -0.000 (-0.05) ***

***

0.216

***

0.215

***

0.216

0.216

(10.16)

(10.09)

(10.13)

(10.13)

(10.05)

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

8554

8554

8554

8554

8554

0.142

0.141

0.141

0.141

0.141

0.136

0.135

0.135

0.135

0.135

Ac

INTERCEPT

ce pt

H_INDIV H_UA

-0.003

ed

H_DISCLOSE

*, **, *** denote significance at the 10%, 5%, and 1% levels, respectively (two-tailed).

28

Page 30 of 36

ip t cr

(1)

CRISIS

0.009

(2) ***

0.008

(3)

***

***

0.008

M an

DA_PJM1<0

us

Table 9 The Effects of Firm Characteristics, Institutions, and Culture on Negative Earnings Management During the Financial Crisis

(5.85)

(4.11)

(3.31)

(4) 0.009

(5) ***

(3.66)

0.009

***

(4.45)

Firm Characteristics Crisis*ISSUE

-0.000

ed

(-0.11)

Institutions and Culture

Crisis*H_INDIV Crisis*H_UA BIG4 GROWTH SIZE

Ac

Crisis*H_DISCLOSE

0.002 (0.56)

ce pt

Crisis*H_ANTISD

0.004

**

(2.34) ***

-0.012

(-5.11) 0.003

***

(8.24)

0.002 (0.69) -0.000 (-0.07) 0.000 (0.05) 0.004

**

(2.34) ***

-0.012

(-5.12) 0.003

***

(8.25)

0.004

**

(2.34) -0.012

***

(-5.12) ***

0.003

(8.26)

**

0.004

(2.34) -0.012

***

(-5.11) 0.003

***

(8.23)

0.004

**

(2.33) ***

-0.012

(-5.11) 0.003

***

(8.24)

29

Page 31 of 36

ip t cr 0.039

***

0.039

(7.85)

(7.85)

***

***

-0.037

(-19.30) ISSUE

***

0.039

(7.87)

***

-0.037

-0.037

(-19.29)

(-19.30)

M an

LOSS

***

us

LEV

***

***

-0.016

-0.016

(-5.36)

(-7.32)

H_ANTISD

-0.016

***

(-7.32)

0.039

***

(7.86)

***

***

0.039

(7.84) ***

-0.037

-0.037

(-19.32)

(-19.32)

***

-0.016

(-7.28)

-0.016

***

(-7.30)

-0.005

(-0.63)

0.002

H_INDIV

INDUSTRY FE N R

2

adj. R

2

***

-0.106

0.000 (0.01) 0.005 (0.91) ***

-0.106

-0.105

***

***

-0.106

-0.111

***

(-7.29)

(-7.28)

(-7.27)

(-7.27)

(-7.71)

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

8644

8644

8644

8644

8644

0.188

0.188

0.188

0.188

0.188

0.183

0.183

0.183

0.183

0.183

Ac

COUNTRY FE

ce pt

H_UA INTERCEPT

(0.26)

ed

H_DISCLOSE

*, **, *** denote significance at the 10%, 5%, and 1% levels, respectively (two-tailed)

30

Page 32 of 36

Ac ce p

te

d

M

an

us

cr

ip t

REFERENCES Abarbanell, J. S., and V. L. Bernard. 1992. Tests of Analysts' Overreaction/Underreaction to Earnings Information as an Explanation for Anomalous Stock Price Behavior. The Journal of Finance.47, 1181-1207. Ahmed, A. S., C. Takeda, and S. Thomas. 1999. Bank loan loss provisions: a reexamination of capital management, earnings management and signaling effects. Journal of Accounting and Economics.28, 1-25. AlNajjar, F. and A. Rainhi-Belkaoui, Growth Opportunities and Earnings Management, Managerial Finance, 27, 72-81. Ball, R., and E. Bartov. 1996. How naive is the stock market's use of earnings information? Journal of Accounting and Economics. 21, 319-337. Ball, R., and P. Brown. 1968. An Empirical Evaluation of Accounting Income Numbers. Journal of Accounting Research.6, 159-178. Balsam, S., J. Krishnan, and J. S. Yang. 2003. Auditor Industry Specialization and Earnings Quality. Auditing.22, 71-97. Barth, M. E., W. R. Landsman, and M. H. Lang. 2008. International Accounting Standards and Accounting Quality. Journal of Accounting Research.46, 467-498. Bartov, E., F. A. Gul, and J. S. L. J. S. L. Tsui. 2000. Discretionary-accruals models and audit qualifications. Journal of Accounting and Economics.30, 421-452. Beaver, W. H. 1968. The Information Content of Annual Earnings Announcements. Journal of Accounting Research, 6, 67-92. Bernard, V. L., and J. K. Thomas. 1989. Post-Earnings-Announcement Drift: Delayed Price Response or Risk Premium? Journal of Accounting Research.27, 1-36.———. 1990. Evidence that stock prices do not fully reflect the implications of current earnings for future earnings. Journal of Accounting and Economics.13, 305-340. Bebchuk, Lucian A. and Spamann, Holger, "Regulating Bankers' Pay" (2009). Harvard Law School John M. Olin Center for Law, Economics and Business Discussion Paper Series. Paper 634. http://lsr.nellco.org/harvard_olin/634. Chen, H., Q. Tang, Y. Jiang, and Z. Lin. 2010. International Financial Reporting Standards and Accounting Quality: Evidence from the European Union. Journal of International Financial Management & Accounting.21, 220-278. Chen, K.Y., R.J. Elder and S. Hung, The Investment Opportunity Set and Earnings Management: Evidence from the Role of Controlling Shareholders, Corporate Governence: An International Review, 18, 193-211. Chia, Y.M., I. Lapsley and H. Lee, 2007, Choice of Auditors and Earnings Manangement During the Financial Crisis. Managerial Auditing Journal.22, 177-196. DeAngelo, L. E. 1981a. Auditor Independence, 'Low Balling', and Disclosure Regulation. Journal of Accounting and Economics.3, 113-127. ———. 1981b. Auditor Size and Audit Quality. Journal of Accounting and Economics. 3, 183-199. Dechow, P., W. Ge, and C. Schrand. 2010. Understanding earnings quality: A review of the proxies, their determinants and their consequences. Journal of Accounting and Economics.50, 344-401. Dechow, P. M., R. G. Sloan, and A. P. Sweeney. 1995. Detecting Earnings Management. The Accounting Review.70, 193-225. 31

Page 33 of 36

Ac ce p

te

d

M

an

us

cr

ip t

———. 1996. Causes and Consequences of Earnings Manipulation: An Analysis of Firms Subject to Enforcement Actions by the SEC. Contemporary Accounting Research.13, 1-36. DeFond, M. L., and J. Jiambalvo. 1994. Debt covenant violation and manipulation of accruals. Journal of Accounting and Economics.17, 145-176. DeFond, M. L., and K. R. Subramanyam. 1998. Auditor changes and discretionary accruals. Journal of Accounting and Economics.25, 35-67. Djankov, S., R. La Porta, F. Lopez-de-Silanes, and A. Shleifer. 2008. The law and economics of self-dealing. Journal of Financial Economics.88, 430-465. Dopuch, N., and D. Simunic. 1982. Competition in Auditing: An Assessment. Symposium on Auditing Research IV, Urbana: University of Illinois. Doupnik, T. 2008 Influence of Culture on Earnings Management: A Note. Abacus 44, 317-340 Durnev, A. R. T., and E. H. Kim. 2005. To Steal or Not to Steal: Firm Attributes, Legal Environment, and Valuation. The Journal of Finance.60, 1461-1493. Fama, E. F., and K. R. French. 1995. Size and Book-to-Market Factors in Earnings and Returns. The Journal of Finance.50, 131-155. ———. 1996. Multifactor Explanations of Asset Pricing Anomalies. The Journal of Finance. 51, 55-84. Gray, S. 1988, "Towards a Theory of Cultural Influence on the Development of Accounting Systems Internationally, Abacus, 24(1), 1-15. Habib, A., B.U. Bhuiyan and A. Islam. 2012. Financial Distress, Earnings Management and Market Pricing of Accruals During the Global Financial Crisis, Managerial Finance. 39, 1-38. Hameed, A., W. Kang and S. Viswanathan, 2010, "Stock Market Declines and Liquidity", The Journal of Finance. 65, 257-293. Han, S., T. Kang, S. Salter, and Y. K. Yoo. 2010. A cross-country study on the effects of national culture on earnings management. Journal of International Business Studies.41, 123-141. Healy, P. M. 1985. THE EFFECT OF BONUS SCHEMES ON ACCOUNTING DECISIONS. Journal of Accounting & Economics.7, 85-107. Healy, P. M., A. P. Hutton, and K. G. Palepu. 1999. Stock Performance and Intermediation Changes Surrounding Sustained Increases in Disclosure. Contemporary Accounting Research.16, 485-520. Healy, P. M., and K. G. Palepu. 1993. The Effect of Firms' Financial Disclosure Strategies on Stock Prices. Accounting Horizons.7, 1-11. Healy, P. M., and J. M. Wahlen. 1999. A Review of the Earnings Management Literature and Its Implications for Standard Setting. Accounting Horizons.13, 365-383. Heninger, W.G., 2001, The Association bewtween auditor litigation and abnormal accruals, The Accounting Review, 76, 111-126. Research and Values Survey Modules 2008. Available at Hofstede, G. http://www.geerthofstede.nl/research--vsm.aspx (accessed on 23 January 2012). Hofstede G., Hofstede G. J. and Minkov M. 2010. Cultures and organizations: software of the mind: intercultural cooperation and its importance for survival, 3rd ed., New York: McGraw Hill Holthausen, R. W., and R. E. Verrecchia. 1990. The Effect of Informedness and Consensus on Price and Volume Behavior. The Accounting Review.65, 191-208. Hope, O.-K. 2003. Disclosure Practices, Enforcement of Accounting Standards, and Analysts' Forecast Accuracy: An International Study. Journal of Accounting Research.41, 235-272. 32

Page 34 of 36

Ac ce p

te

d

M

an

us

cr

ip t

Ivashina, V. and D. Scharfstein. 2010. Bank Lending During the Financial Crisis of 2008, Journal of Financial Economics. 97, 319-338. Jones, J. J. 1991. Earnings Management During Import Relief Investigations. Journal of Accounting Research.29, 193-228. Jones, K. L., G. V. Krishnan, and K. D. Melendrez. 2008. Do Models of Discretionary Accruals Detect Actual Cases of Fraudulent and Restated Earnings? An Empirical Analysis*. Contemporary Accounting Research.25, 499-531. Jordon, C.E. and S.J. Clark. 2011. Big Bath Earnings Management: The Case of Goodwill Impairment Under SFAS No. 142. Journal of Applied Business Research. 20, 63-69. Kothari, S. P., A. J. Leone, and C. E. Wasley. 2005. Performance matched discretionary accrual measures. Journal of Accounting and Economics.39, 163-197 Krishnan, G. V. 2003. Audit Quality and the Pricing of Discretionary Accruals. Auditing.22, 109-126. La Porta, R., F. Lopez-De-Silanes, and A. Shleifer. 2006. What Works in Securities Laws? Journal of Finance.61, 1-32. La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R. Vishny. 1998. Law and finance. Journal of Political Economy.106, 1113-1155. Laux, C. and C Leuz, 2009, Did Fair Value Accouunting Contribute to the Financial Crisis?, NBER Working Paper 15515. Leuz, C., D. Nanda, and P. D. Wysocki. 2003. Earnings management and investor protection: an international comparison. Journal of Financial Economics. 69, 505-527. Lin, J. W., and M. I. Hwang. 2010. Audit Quality, Corporate Governance, and Earnings Management: A Meta-Analysis. International Journal of Auditing.14, 57-77. Lin, L, R. Morris, H. Kang and Q Tang. 2013. Information Asymmetry of Fair Value Accounting during the Global Financial Crisis. Journal of Contemporary Accounting and Economics (A). Volume 9 Issue 2, December, 221-236. Nabar, S and K. K. Boonlert-U-Thai (2007) Earnings Management, Investor Protection, and National Culture. Journal of International Accounting Research: Fall, Vol. 6, No. 2, pp. 35-54. Okonjo-Iweala, N., V. Kwakwa, A. Beckwith and Z. Ahmed. 1999. Impact of Asia’s Financial Crisis on Cambodia and the Lao PDR. Finance & Development.36, 48-51. Pinnuk, M. 2012. A Review of the Role of Financial Reporting in the Global Financial Crisis. Australian Accounting Review. 60, 1-14. Reynolds, J. K., and J. R. Francis. 2000. Does size matter? The influence of large clients on office-level auditor reporting decisions. Journal of Accounting and Economics.30, 375-400. Rusmin, R., G. Scully and G. Tower. 2013. Income Smoothing Behaviour by Asian Transportation Firms. Managerial Auditing Journal. 28, 23-44. Securities and Exchange Commission (SEC), US. 2009. Report and Recommendations Pursuant to Section 133 of the Emergency Economic Stabilization Act of 2008: Study on Mark-To-Market Accounting. Skinner, D., and R. Sloan. 2002. Earnings Surprises, Growth Expectations, and Stock Returns or Don't Let an Earnings Torpedo Sink Your Portfolio. Review of Accounting Studies.7, 289-312. Sloan, R. G. 1996. Do Stock Prices Fully Reflect Information in Accruals and Cash Flows about Future Earnings? The Accounting Review.71, 289-315. Smith, C. W. and R. L. Watts, The Investment Opportunity Set and Corporate Financing, Dividend 33

Page 35 of 36

Ac ce p

te

d

M

an

us

cr

ip t

and Compensation Policies, Journal of Financial Economics, 32, 263-292. Statman, M. 1999. Behaviorial Finance: Past Battles and Future Engagements. Financial Analysts Journal. 55, 18-27. Teoh, S. H., and T. J. Wong. 1993. Perceived Auditor Quality and the Earnings Response Coefficient.The Accounting Review.68, 346-366. Teoh, S. H., I. Welch, and T. J. Wong. 1998a. Earnings Management and the Long-Run Market Performance of Initial Public Offerings. The Journal of Finance.53, 1935-1974. Teoh, S. H., I. Welch, and T. J. Wong. 1998b. Earnings management and the underperformance of seasoned equity offerings. Journal of Financial Economics. 50, 63-99. The Chinese Culture Connection. 1987. Chinese Values and the Search for Culture-Free Dimensions of Culture. Journal of Cross-Cultural Psychology, 18, 143-164. Wang, D. 2006. Founding Family Ownership and Earnings Quality. Journal of Accounting Research.44, 619-656. Watts, R. L., and J. L. Zimmerman. 1990. Positive Accounting Theory: A Ten Year Perspective. The Accounting Review.65, 131-156. Xu, Y., E. Carson, N. Fargher and L. Jiang. 2013. Responses by Australian Auditors to the Global Financial Crisis. Accounting and Finance, 53, 301-338. Yoon, S.S. and G.A. Miller. 2002. Cash from Operations and Earnings Management in Korea. The International Journal of Accounting, 37, 395-412.

34

Page 36 of 36