how do accounting standards and insiders' incentives affect …drthomaswu.com/uicfat/8.pdf ·...

Emerging Markets Review 16 (2013) 78–99

Contents lists available at SciVerse ScienceDirect

Emerging Markets Review

j ourna l homepage: www.e lsev ie r .com/ locate /emr

How do accounting standards and insiders'incentives affect earnings management?Evidence from China☆

Yuyang Zhang a, Konari Uchida b,⁎, Hua Bu c

a Graduate School of Economics, Kyushu University, 6-19-1, Hakozaki, Higashiku, Fukuoka 812-8581 Japanb Faculty of Economics, Kyushu University, 6-19-1, Hakozaki, Higashiku, Fukuoka 812-8581 Japanc School of Management, China University of Mining and Technology, University Road, 221116 Xuzhou, Jiangsu Province, China

a r t i c l e i n f o

☆ This paper was financially supported by the JSPS⁎ Corresponding author. Tel./fax: +81 92 642 246

E-mail address: [email protected] (K.

1566-0141/$ – see front matter © 2013 Elsevier B.V.http://dx.doi.org/10.1016/j.ememar.2013.04.002

a b s t r a c t

Article history:Received 23 November 2012Received in revised form 16 April 2013Accepted 30 April 2013Available online 9 May 2013

Chinese listed companies recently experienced two important institu-tional changes: split share structure reform (SSSR) and the mandatoryadoption of IFRS-convergent new accounting standards (NAS). We findthat the introduction of NAS significantly increased earnings manage-ment. Although we do not find evidence that SSSR directly decreasedearnings management of the average firm, the increase in earningsmanagement surrounding the introduction of NAS is negatively relatedto the reduction in non-tradable shares. These results suggest thataccounting standards are the more important factor associatedwith thelevel of earnings management. Insiders' incentives affect earningsmanagement given a specific set of accounting standards.

© 2013 Elsevier B.V. All rights reserved.

JEL classification:G34G38M10M40

Keywords:Split-share structure reformIFRSCorporate governanceEarnings managementChina

1. Introduction

Numerous studies have investigated determinants of firms' earnings management. Dechow et al. (2010)suggest that reported earnings are determined both by fundamental performance and measurement of thisperformance, of which the latter is considerably affected by accounting standards. In other words, accountingstandards will determine the degree of managerial discretion regarding revenue and loss recognition, and

KAKENHI Grant Number 23330107.3.Uchida)

All rights reserved.

http://dx.doi.org/10.1016/j.ememar.2013.04.002

mailto:[email protected]

http://dx.doi.org/10.1016/j.ememar.2013.04.002

http://www.sciencedirect.com/science/journal/15660141

79Y. Zhang et al. / Emerging Markets Review 16 (2013) 78–99

thereby, affect the level of corporate earnings management. The market globalization, which has beenaccelerated during the past few decades, generated huge demands among investors for internationalharmonization in the financial reporting process (Barth et al., 2012; He et al., 2012). Accordingly, theInternational Accounting Standards Committee (IASC) and its successor, the International AccountingStandards Board (IASB), have developed internationally acceptable accounting standards, also known asthe International Financial Reporting Standards (IFRS) (Ashbaugh and Pincus, 2001; Daske et al., 2008;Liu et al., 2011). Following the attempt to converge accounting information across countries in theEuropean Union, IFRS has been adopted by over 100 countries and regions. The introduction of IFRS offersresearchers an appropriate opportunity to investigate whether accounting standards affect the degree ofearnings management, as well as raises the question of whether IFRS improves earnings quality uniformlyaround the world. Barth et al. (2008, 2012), Cormier et al. (2009), and Liu et al. (2011) suggest earningsquality increases with the adoption of IFRS. In contrast, Callao et al. (2007), He et al. (2012), and Rudra andBhattacharjee (2012) find that earnings management increased after the introduction of IFRS in Spain, Chinaand India.

Another prominent stream in earnings management research is to investigate the relationshipbetween earnings management and corporate governance structures (Dechow et al., 1996; Klein, 2002;Park and Shin, 2004; Peasnell et al., 2005). These studies generally found a negative relationship betweenthe efficiency of governance devices and earnings management: for board and audit committees(Dechow et al., 1996; Gabrielsen et al., 2002; Klein, 2002; Park and Shin, 2004; Xie et al., 2003); forownership structure (Warfield et al., 1995; Yeo et al., 2002); and for audit quality (Francis and Krishnan,1999; Francis and Wang, 2008; Iatridis, 2012; Krishnan, 2003). Studies also address the effects ofcorporate governance on firms in emerging markets that typically have concentrated ownershipstructures andweak investor protection. In these countries, divergence exists between control rights andcash flow rights through a pyramidal structure, cross-shareholdings and dual-class shares. Thesecontrolling shareholders are effectively released from disciplinary governance mechanisms (Fan andWong, 2002; Johnson et al., 2000; La Porta et al., 1999; Liu and Lu, 2007; Shleifer and Vishny, 1997) andpursue private benefits of control at the expense of minority shareholders' wealth (Claessens et al., 2002;Denis and McConnell, 2003; Leuz et al., 2003; Lins, 2003). Under these circumstances, insiders manageearnings to conceal their expropriation of minority shareholders' wealth (Fan and Wong, 2002; Kim andYi, 2006; Leuz et al., 2003; Shleifer and Vishny, 1997). Indeed, previous studies suggest that ownershipconcentration and the divergence between control and cash flow rights are positively associated with thedegree of earnings management (Fan and Wong, 2002; Firth et al., 2007; Gopalan and Jayaraman, 2012;Haw et al., 2004; Kim and Yi, 2006; Liu and Lu, 2007).

This study investigates how accounting standards and insiders' incentives affect earnings managementby using recent Chinese data. The novelty of the data is that Chinese companies have experienced twonoticeable reforms that substantially changed accounting standards and insiders' incentives: split-sharestructure reform (hereafter denoted by SSSR) and the mandatory adoption of IFRS-convergent newaccounting standards (hereafter denoted by NAS). Segmented share structure, which represents thecoexistence of non-tradable shares (NTS) and tradable shares (TS), was a feature of Chinese corporategovernance before the SSSR. The split-share structure engendered divergence between cash flow andcontrol rights, which motivates controlling shareholders to extract private benefits at the expense ofminority shareholders' wealth (Claessens et al., 2000; Dyck and Zingales, 2004; Johnson et al., 2000; LaPorta et al., 2000).1 Previous studies also suggest that firms that issue more NTS manage earnings more(e.g., Firth et al., 2007). On April 29, 2005, however, the China Securities Regulatory Commission (CSRC)initiated the SSSR to require firms to convert NTS to TS through negotiations with TS holders. Given previousresearch findings, the substantial reduction in NTS should considerably decrease insiders' incentives ofminorityshareholder expropriation and earnings management. The ChineseMinistry of Finance also obliged listed firmsto introduce IFRS-convergent NAS in 2007, which considerably affects the degree of earnings management.Previous research argues that accounting standards per se have only a limited effect on earnings qualitycompared to insiders' incentives in countries whose economic and institutional environments are incompatible

1 NTS holders could not sell their shares on the public markets to realize capital gains. As a result, NTS holders have been moreinterested in expropriation of firm resources for their private benefits than in maximization of firm value.

80 Y. Zhang et al. / Emerging Markets Review 16 (2013) 78–99

with IFRS (e.g., East-Asian countries) (Ball et al., 2003; Bruggemann et al., 2010; Burgstahler et al., 2006;He et al.,2012; Holthausen, 2009; Leuz et al., 2003). The important institutional changes (the introduction of the SSSRand NAS) allow us to compare the impact of accounting standards on earnings management to that of insiders'incentives.

The findings of this study are summarized as follows. Although this study does not find evidence thatthe average Chinese firm decreased earnings management surrounding the SSSR, there is a positive andsignificant relationship between the reduction in earnings management and NTS surrounding the SSSR.However, the mandatory implementation of NAS significantly increased earnings management. Importantly,the increase in earnings management surrounding the introduction of NAS is negatively related to thereduction in NTS. These results suggest that accounting standards have stronger direct impacts on earningsmanagement than insiders' incentives, and that insiders' incentives significantly affect how firms increaseearnings management in response to the introduction of NAS.

This papermakes significant contributions to the literature. In such, it shows evidence that accountingstandards are the primary determinant of the degree of earnings management by investigating thechange-to-change relationship between earnings management, accounting standards and insiders'incentives. This study also shows that insiders' incentives can explain the variation of earnings managementwithin a specific set of accounting standards. Finally, the results of this study suggest that China's economicand institutional environments are incompatible with the adoption of IFRS; the standards provide insidersmore opportunities to engage in earnings management.

The rest of the paper is organized as follows. Section 2 presents the literature review. Section 3 developsthe hypotheses. Section 4 describes the methodology of this research. Section 5 presents this study's sampleselection and data. Results are reported in Section 6. Section 7 concludes this research.

2. Literature review

2.1. Accounting standards and earnings management

Accounting standards determine the degree of earnings quality. In general, accounting regulators(e.g., IASB) attempt to increase the compatibility and value relevance of earnings information bydeveloping an internationally acceptable set of financial reporting standards (Barth et al., 2008). IFRS, whichhas been adopted bynumerous countries during the past two decades, is typical of such accounting standards.The introduction of IFRS provides us with an appropriate opportunity to examine whether accountingstandards significantly affect the level of earningsmanagement. The intended objective of IFRS is to provide “ahigh degree of transparency and comparability of financial statements and hence an efficient functioning of thecommunity capital market and of the internal market” (European Communities, 2002: Art.1). To achieve thisgoal, IFRS embraces principle-based standards (Ashbaugh and Pincus, 2001). IFRS also requires accountingmeasurements that better reflect underlying performance and that provide investors with informativeaccounting information for their decision making (IASC, 1989). However, previous studies regarding theimpact of IFRS introduction on earnings quality have presented mixed results. Barth et al. (2008, 2012) showevidence that firms in countries adopting the international accounting standards (IAS or IFRS) managereported earnings less thanfirms in countries that apply domestic standards. Liu et al. (2011) indicate that theaccounting quality of Chinese listed companies has improved since the mandatory implementation ofIFRS-convergent NAS in 2007. Similarly, Cormier et al. (2009) find that the first-time adoption of IFRS byFrench firms was associated with increased quality of the firms' financial statements. Clarkson et al. (2011)provide evidence that difference in value relevance of earnings between Common and Code law countriesdisappears after the adoption of IFRS, suggesting that IFRS implementation enhances this comparability.Djatej et al. (2011) find that the quality of public information increases due to the implementation of IFRS inboth Eastern and Western European firms.

In contrast, He et al. (2012) report that the adoption of fair value accounting (a vital feature of China'sNAS) motivates firms to smooth reported earnings by selling available-for-sale security for capital gains.Rudra and Bhattacharjee (2012) provide Indian evidence that firms adopting IFRS are more likely toengage in earnings management (income smoothing) compared to non-adopting firms. Callao et al.(2007) find that IFRS introduction does not improve the quality of financial reporting in Spain.


2.2. Insiders' incentives and earnings management

Insiders have different reporting incentives, depending on local institutional environments (Ball et al.,2003; Burgstahler et al., 2006; Holthausen, 2009). The U.S. and other common-law countries typicallyhave dispersed corporate ownership structures. Firms in these countries rely heavily on publicinformation and thus likely have incentives to provide reliable earnings information. In contrast, privatechannels are more important in countries with a concentrated corporate ownership structure (e.g., China)to reduce information asymmetry (Burgstahler et al., 2006; Leuz, 2006). As a result, insiders in thesecountries have relatively weaker incentives to report informative earnings. Rather, they tend to extractprivate benefits at the expense of minority shareholders' wealth. These insiders will have an incentive toconceal their value-decreasing behaviors and to avoid potential outside interventions by engaging inearnings management (Fan and Wong, 2002; Gopalan and Jayaraman, 2012; Haw et al., 2004; Kim and Yi,2006; Leuz et al., 2003).

Ball et al. (2003) investigate earnings quality (timely loss recognition) in four East Asian countrieswhose accounting standards are viewed as high quality (following the standards of common-lawcountries that are IFRS-convergent). Interestingly, they find that earnings quality in these countries is nobetter than that in code-law countries. This result suggests that insiders' incentives have a strong impacton the level of earnings management. For the sample of 9 East Asian and 13 Western European countries,Haw et al. (2004) find that income management is positively associated with the divergence betweencontrolling shareholders' voting rights and cash flow rights. Gopalan and Jayaraman (2012) examineearnings management of firms from 22 countries. They show evidence that insider-controlled firms withgreater divergence between control rights and cash flow rights engage more in earnings management.

3. Hypotheses development

3.1. Accounting standards and earnings management

China issued its IFRS-convergent NAS on February 15, 2006 (He et al., 2012; Liu et al., 2011). Chineselisted firms have been required to follow NAS since January 1, 2007. Given that IFRS is designed to enhancethe transparency of accounting information, its introduction should be accompanied by improved earningsquality in China. Indeed, several previous studies suggest that IFRS introduction increases earnings quality(Barth et al., 2008, 2012; Cormier et al., 2009). Liu et al. (2011) support this idea for Chinese companies andwe present the following hypothesis:

H1a. Earnings management has decreased following the implementation of China's NAS.

However, the literature presents mixed evidence regarding the effectiveness of IFRS on earnings quality.There are several reasons why IFRS does not contribute to improved earnings quality in China. Firstly, theflexibility inherent in principles-based standards is likely to provide managers more opportunities to engagein earnings management (Breeden, 1994). For example, traditional Chinese accounting standards haveallowed firms to measure trading securities at lower historical costs or market values. However fair-valueaccounting in China's NAS requires firms to reflect the change in those securities' values in current earnings.He et al. (2012) suggest this considerable change will give managers the opportunity to manipulate earningsby selling available-for-sale securities.

Secondly, several researchers point out that Chinese economic and institutional conditions areincompatible with the IFRS-convergent NAS. For instance, Peng and Bewley (2010) argue that the recentadoption of fair value-oriented NAS has not been accompanied by the necessary changes in China's capitaland other market infrastructures. This incompatibility potentially induces more opportunistic earningsmanagement that uses the inherent flexibility in IFRS (He et al., 2012; Holthausen, 2009). IFRS is drafted incountries with developed equity markets in which control rights of firms are widely dispersed (Nobes,1998; Sunder, 2009; Tyrrall et al., 2007). Besides, the IFRS is based mainly on the practices of common-lawcountries (Liu et al., 2011). However, China has followed a central planned economy (Ding and Su, 2008)and adopts a code-law system in which political interventions play essential roles. He et al. (2012) arguethat the goal of fair value accounting in IFRS—namely, supplying relevant information to investors—can be


achieved only when the fair value of assets is obtained through arm's-length negotiations between thefirm and its counterparts. However, this is unavailable in China because firms and their counterparts areusually related parties. Rather, the adoption of fair value accounting is more likely to provide insiderswith opportunities to engage in earnings management. Additionally, Callao et al. (2007) and Rudra andBhattacharjee (2012) show evidence that financial reporting quality has worsened after the applicationof IFRS in Spain and India, respectively. As for China, He et al. (2012) show that earnings managementincreases after the introduction of IFRS.

Thus, we propose an alternative hypothesis regarding the impact of the adoption of IFRS-convergentNAS on earnings management in China:

H1b. The level of earnings management has increased following the implementation of NAS.

3.2. Insiders' incentives and earnings management

In China, the existence of NTS generates divergence between cash flow and control rights, andtherefore, insiders have an incentive to extract private benefits. In the 1990s, the Chinese Governmentlaunched its privatization program to list some SOEs on the Shanghai and Shenzhen stock exchanges.However, only a minority part of shares of these companies (TS) can be traded on the secondary market.The majority of the shares (NTS) has been retained mainly by the founders (usually central or localgovernments) and is not tradable in the secondary market. The segmented share structure was originallydesigned for different levels of governments to maintain their control over SOEs, especially in key industries(e.g., agricultural, chemical and petroleum). The split-share structure engenders severe divergence betweencash flow and control rights, since holders of NTS neither benefit from capital gains nor bear losses from stockprice reductions. Prior research suggests that this divergence gives insiders an incentive to extract privatebenefits at the expense of minority shareholders' wealth (Claessens et al., 2000; Dyck and Zingales, 2004;Johnson et al., 2000; La Porta et al., 2000). Those insiders are likely to engage in earnings management toconceal evidence of their value-destroying behaviors. Indeed, Fan and Wong (2002) show that controldivergence weakens informativeness of accounting earnings. Firth et al. (2007) find that the proportion of NTSover total shares is positively related to earnings management. Notably, CSRC initiated the SSSR on April 29,2005 to convert NTS to TS. NTS holders are required to negotiate with TS holders regarding compensation NTSholders pay to TS holders in exchange for obtaining tradability of their shares. The SSSR, which substantiallydecreased NTS, is likely to reduce insiders' incentives to extract private benefits and manage earnings. Thesediscussions give rise to the following hypothesis:

H2. The level of earnings management has decreased following completion of the SSSR.

4. Methodology

4.1. Measurement of earnings management

To capture the magnitude of earnings management, we adopt the value of discretionary accruals (DA),which is measured by the difference between total accruals (TA) and non-discretionary accruals (NDA),divided by total assets at the beginning of the year. Among various measures of earnings management, DAis one of the most commonly-used measures in the research (Dechow et al., 2010). DA is suitable for ourresearch because it computes the level of earningsmanagementwithout any a priori assumptions concerningthe goal of earnings management.2 Different from timely loss recognition, DA does not use stock returns,which will be highly affected by expropriation problems, as a measure of firms' true performance.

We follow previous studies in using a cross-sectional Modified Jones Model to estimate earningsmanagement (Dechow et al., 1995; Kothari et al., 2005).3 Bartov et al. (2001) suggest the cross-sectionalModified JonesModel is superior tomodels that use time-series estimation. Dechow et al. (2010) suggest that

2 The usage of earnings persistence requires the assumption that majority shareholders have an incentive to report less persistentearnings (earnings that are not related to previous earnings) to conceal expropriation problems. If we use earnings smoothness, wehave to assume that less volatile earnings can hide expropriation problems. It is extremely difficult to justify those assumptions.

3 Industry classification is based on the “Classification Guidance of Chinese Listed Companies,” issued by the CSRC in 1998.


the cross-sectional Modified Jones Model increases the power of the Jones Model (1991) by reflecting creditsales manipulation.

In order to estimate NDA, we estimate the following Eq. (1) for firm-years in the same industry beforethe institutional changes (SSSR and introduction of NAS). We delete data after the introduction of the SSSRand NAS for the estimation, since reported earnings after the institutional changes potentially showabnormal patterns. This method allows NDA to incorporate the whole effects of the institutional changeson earnings management.

4 Somadopt tWarfiel

TAi;t

Assetsi;t−1¼ β0

1Assetsi;t−1

!þ β1

ΔSalesi;t−ΔARi;t

Assetsi;t−1

!þ β2

PPEi;tAssetsi;t−1

!þ β3 ROAi;t−1

� �þ εi;t ð1Þ

subscripts i and t represent firm and year, respectively. Δ denotes the change from the previous
whereyear. Assets is total assets; Sales is annual sales; AR refers to net accounting receivable, PPE represents totalfixed assets; and ROA is return on assets, which is included in the model, as suggested by Kothari et al.(2005).
Following Dechow et al. (1995), we calculate TA as

TAi;t ¼ ΔCAi;t−ΔCashi;t� �

− ΔCLi;t−ΔSTLi;t−ΔTPi;t

� �−Depreciationi;t ð2Þ

CA is current assets, Cash is cash and cash equivalents, CL is current liabilities, STL is short-term loan
whereincluded in current liabilities, TP is taxes payable, and Depreciation is depreciation expenses.
We compute DA as TA minus NDA, which is calculated by using the estimated coefficients for Eq. (1).Since we focus on the magnitude of DA rather than the direction of DA, the following analysis employs theabsolute values of discretionary accruals (ADA).4 A definition of variables is presented in Table 1.

4.2. Research design

To test our hypotheses, we investigate the change of earnings management (ADA) surrounding theSSSR (H2) and surrounding the implementation of NAS (H1a and H1b). Throughout the followinganalyses, we adopt a year before 2005 as benchmark year since ADA after the initiation of SSSR potentiallyincludes abnormal patterns. To test H1a and H1b, we implement statistical tests on whether the mean andmedian ADA are different between year 2003 and year 2007 + j (j ≥ 0). Similarly for H2, we conductstatistical tests on whether the mean andmedian ADA are different between year tb and year ta. Year tb (ta)denotes a year before (after) the firm's SSSR initiation (completion).

H2 also implies that the change in ADA is positively associated with the change in the number of NTS,since the SSSR decreases insiders' incentives for earnings management through decreasing NTS. Toaddress this issue, we execute regression analyses of the change in ADA (C_ADA) before and after the SSSR.We employ the change in the proportion of NTS over total outstanding shares (C_NTSR) as the keyindependent variable.

C ADAi;tbeta ¼ β0 þ β1C NTSRi;tbeta þ β2NTSRi;tbþ∑

k¼3βkControlk;ita–2 þ εi: ð3Þ

The subscript i denotes the firm and tb~ta for change variables (C_ADA and C_NTSR) denotes the periodduring year tb to year ta. For other variables, the subscript t denotes the data year. We include theproportion of NTS over outstanding shares (NTSR) to control for the potential effects of the initial level ofdivergence between cash flow and control rights.

Although H1a and H1b simply predict that ADA decreased and increased surrounding the introductionof NAS due to the change in managerial discretion on reported earnings, insiders' responses to the changeare likely to depend on their incentive of earnings management. For instance, managers of firms with high

e previous studies also combine the effect of income-increasing and income-decreasing earnings management, and therefore,he absolute values of discretionary accruals as a proxy for earnings management (Chung et al., 2002; Firth et al., 2007;d et al., 1995).

Table 1Definition of variables.

ADA The absolute value of discretionary accruals. Discretionary accruals (DA) are measured by the differencebetween total accruals (TA) and non-discretionary accruals (NDA). The cross-sectional Modified JonesModel is adopted for estimation.

D_NAS A dummy variable that takes a value of one if a firm-year is after the implementation of IFRS-convergent newaccounting standards, and zero otherwise

NTSR The percentage of non-tradable shares over total outstanding sharesCONCENTRATION The percentage ownership by the largest shareholderD_SOE A dummy variable that takes a value of one for state-owned enterprises, and zero for othersBOARDSIZE The natural logarithm of the number of board membersBINDEPENDENCE The proportion of independent directors over total board membersPDIREC_SOWN The proportion of board members who own the firm's shares over all board membersD_DUAL A dummy variable that takes a value of one if the CEO also serves as the chairperson of the boardSUPERSIZE The natural logarithm of the number of supervisory board members.PSUPER_SOWN The proportion of supervisory board members who own shares of the company over total supervisorsSIZE The natural logarithm of total assetsROA Return on assetsLEV Liabilities divided by equityQ The total book value of liabilities and market value of equity divided by the book value of total assets


NTSR (or small reduction in NTSR during the SSSR) have a strong incentive to manage earnings and takeadvantage of the accounting standard change (if NAS increases their discretion on reported earnings). Totest this idea, we estimate the following regression of C_ADA. The inclusion of C_NTSR is important alsobecause we adopt as the dependent variable the change in ADA during year 2003 to year 2007 + j, whichincludes the period of SSSR.

C ADAi;2003 e 2007þj ¼ β0 þ β1C NTSRi;2003 e 2007þj þ β2NTSRi;2003

þ∑k¼3

βkControlk;i;2006 þ εi:ð4Þ

In addition to share non-tradability, ownership concentration will also engender expropriationproblems (Fan and Wong, 2002; Firth et al., 2007; Liu and Lu, 2007; Lo et al., 2010; Shleifer and Vishny,1997). A few insiders who own the majority of a company's shares will be released from disciplinarymechanisms (Fan and Wong, 2002; Johnson et al., 2000; La Porta et al., 1999; Liu and Lu, 2007; Shleiferand Vishny, 1997). Consequently, concentrated ownership structures would give insiders more power toexpropriate minority shareholders' wealth and distort financial reports (entrenchment effect). Fan and Wong(2002) and Firth et al. (2007)find that ownership concentration significantlyweakens earnings quality. It is alsolikely that SSSR reduced earnings management through decreasing ownership concentration. To test this idea,we adopt the percentage of shares owned by the largest shareholder as a measure of ownership concentration(CONCENTRATION) and estimate Eqs. (3) and (4) that replace NTSR and C_NTSR by CONCENTRATION and thechange in CONCENTRATION (C_CONCENTRATION).

We also include the following variables to account for various factors that potentially affect earningsmanagement. One unique feature in Chinese corporate governance is the existence of listed companiesthat are controlled by governments (state-owned enterprises; namely SOEs). SOEs principally pursuesocial and political objectives (e.g., employment, tax revenue, social welfare and so on) rather thanmaximizing shareholders' value (Liu and Lu, 2007). As a result, SOEs are likely to engender expropriationproblems and engage in earnings management. We define SOEs as firms whose largest shareholder is thecentral government or local governments. A dummy variable that takes a value of one for SOEs and zerofor non-SOEs is adopted in this research (D_SOE).

The corporate board is an important governance device that is expected to monitor insiders' behaviors.Numerous studies have investigated whether board characteristics (e.g., board size and independence)affect earnings quality (Ahmed et al., 2006; Beasley, 1996; Dimitropoulos and Asteriou, 2010; Firth et al.,2007; Klein, 2002; Liu and Lu, 2007; Lo et al., 2010; Park and Shin, 2004; Vafeas, 2000; Wang et al., 2007;Xie et al., 2003). We follow these previous studies and adopt the natural logarithm of the number of boardmembers as a proxy for board size (BOARDSIZE). It is also a common idea that independent boards make it


difficult for insiders to expropriate minority shareholders' wealth and to manage earnings (Klein, 2002;Liu and Lu, 2007; Lo et al., 2010; Xie et al., 2003). Therefore, we employ the percentage of independentdirectors over total board members (BINDEPENDENCE) to test this idea.

Board monitoring will become more effective as directors' personal wealth is more sensitive to firmvalue. High board ownership is likely to prevent insiders from conducting value-decreasing behaviors thatincrease their private benefits. Accordingly, earnings management is likely to decrease with boardownership. It is noteworthy that directors in Chinese listed firms typically hold no shares or only aninsignificant portion of shares. The variation in directors' ownership is so limited that we employ thepercentage of directors who own shares (PDIREC_SOWN). When board leadership and CEO power arevested in one person, the CEO is able to exert significant power and board members will find it difficult tooversee managerial behaviors (Jensen, 1993; Raheja, 2005). Dechow et al. (1996) and Forker (1992) findthat firms that manipulate earnings are more likely to have a CEO who simultaneously serves as chairmanof the board. Following these, we adopt a dummy variable that takes a value of one when the CEO alsoserves as chairman of the board, and zero otherwise (D_DUAL).

Chinese firms adopt a two-tier monitoring system that consists of a board of directors and a board ofsupervisors. The Guideline for Chinese Listed Companies (2006) requires companies to have supervisors whoare independent of a board of directors andmanagers.5 Chinese Company Law empowers a supervisory boardtomonitor a firm's accounting system and to request necessary changes in accounting procedures.6We adoptthe natural logarithm of the number of supervisory board members (SUPERSIZE) and the percentage ofsupervisors who own shares over all supervisors (PSUPER_SOWN).

It is likely that large firms engage less in earnings management due to close attention from investorsand regulators (Holland and Jackson, 2004). Indeed, previous studies find that disclosure quality increaseswith firm size (Atiase, 1985; Freeman, 1987). We include the natural logarithm of total assets (SIZE) as aproxy for firm size. Following Butler et al. (2004) and Firth et al. (2007), we include ROA to control for theeffect of profitability on earnings management. Dhaliwal et al. (1991) suggest that firms with highleverage (high default risk) tend to exploit discretionary accounting methods to conceal poor performance(Watts and Zimmerman, 1990). Therefore, we include leverage (LEV), which is liabilities divided by equity.Tobin's Q (Q), which ismeasured by the book value of liabilities and themarket value of equity divided by thebook value of total assets, is included to control for the effect of a firm's growth opportunity on earningsmanagement. Growing firms have an incentive to manage earnings to meet earnings requirements for newsecurity issues or to decrease costs of capital (Beaver et al., 1970; Minton and Schrand, 1999).

5. Sample selection and data

Our sample firms consist of Chinese companies listed on the Shanghai and Shenzhen stock exchangesfrom 2001 through 2010. Corporate financial data are obtained from the CSMAR (China Stock Market andAccounting Research) database operated by GTA Finance & Education Group. We also collect governancedata from the CCER (China Center for Economic Research) database provided by the Sinofin InformationService Company of China, LTD. Financial companies are excluded from our sample due to their regulatoryenvironments and the different financial statement format (Firth et al., 2007; Peasnell et al., 2005; Vafeas,2000). We delete firms for which necessary data are not available. As a result of these procedures, 11,939firm-year observations (involving 1316 firms) are adopted as our final sample. In such, we winsorize theADA at the 1st and 99th percentiles.

Panel A of Table 2 shows summary statistics for the whole sample. Slightly more than 40% of thefirm-years are in the NAS introduction year (2007) or later. NTS accounts for a substantial part of totaloutstanding shares (mean NTSR is 45.9%). The average largest shareholder holds 38.8% of total outstandingshares (CONCENTRATION). In contrast, the mean percentage ownership by the second- to fifth-largestshareholders is less than 15% (the median is 11.8%) (not reported). These figures suggest that Chineseownership structures are highly concentrated so that the largest shareholder can exert dominant controlover listed companies. Panel A also indicates that the government or governments control about 70% of

5 This guideline prohibits the manager, directors and financial officers of the company from being members of a supervisory board.6 The Code of Corporate Governance for Listed Companies allows the board of supervisors to report directly to regulatory

authorities if it finds any violations of laws, regulations, accounting standards or firm charters.

Table 2Descriptive statistics.

Variable Mean Median S.D. 1st quartile 3rd quartile

Panel A: Whole sample (N = 11,939)D_NAS 0.421 0.000 0.494 0.000 1.000NTSR 0.459 0.516 0.225 0.329 0.632CONCENTRATION 0.388 0.365 0.163 0.259 0.515D_SOE 0.697 1.000 0.460 0.000 1.000Number of board members 9.621 9.000 2.303 9.000 11.000BINDEPENDENCE 0.321 0.333 0.109 0.333 0.364PDIREC_SOWN 0.164 0.091 0.206 0.000 0.286D_DUAL 0.132 0.000 0.339 0.000 0.000Number of supervisory board members 4.213 4.000 1.455 3.000 5.000PSUPER_SOWN 0.189 0.000 0.270 0.000 0.333SIZE 21.385 21.291 1.096 20.659 22.024ROA 0.021 0.029 0.080 0.009 0.054LEV 1.336 1.007 1.634 0.549 1.704Q 2.357 1.834 1.669 1.349 2.710

Panel B: Pre-SSSR observations (N = 5353)NTSR 0.601 0.619 0.113 0.538 0.688CONCENTRATION 0.426 0.411 0.168 0.289 0.571

Panel C: during-SSSR observations (N = 1533)NTSR 0.527 0.540 0.128 0.441 0.619CONCENTRATION 0.368 0.344 0.151 0.246 0.490

Panel D: Post-SSSR observations (N = 5053)NTSR 0.289 0.301 0.222 0.025 0.469t-statistics (Z-statistics) 89.525*** 68.558***CONCENTRATION 0.354 0.334 0.152 0.229 0.470t-statistics (Z-statistics) 23.107*** 21.750***

Notes:Panel A indicates descriptive statistics of variables for the whole sample. Panels B to D show summary statistics of NTSR andCONCENTRATION separately for pre-, during and post-SSSR periods. Panel D also indicates t-statistics (Z-statistics) for the nullhypothesis that the mean (median) is identical between the pre- and post-SSSR periods. *, ** and *** indicate significance at the 10%,5%, and 1% levels, respectively. D_NAS is a dummy variable that takes a value of one if a firm-year is after the implementation ofIFRS-convergent new accounting standards, and zero otherwise. NTSR is the percentage of non-tradable shares over totaloutstanding shares. CONCENTRATION is the percentage ownership by the largest shareholder. D_SOE is a dummy variable that takesa value of one for state-owned enterprises, and zero for others. BOARDSIZE is the natural logarithm of the number of board members.BINDEPENDENCE is the proportion of independent directors over total board members. PDIREC_SOWN is the proportion of boardmembers who own the firm's shares over all board members. D_DUAL is a dummy variable that takes a value of one if the CEO alsoserves as the chairperson of the board. SUPERSIZE is the natural logarithm of the number of supervisory board members.PSUPER_SOWN is the proportion of supervisory board members who own shares of the company over total supervisors. SIZE is thenatural logarithm of total assets. ROA is the return on assets. LEV is the liabilities divided by equity. Q is the total book value ofliabilities and market value of equity divided by the book value of total assets.


listed companies (D_SOE). The average (median) board has 9.6 (9.0) members. Independent directorsaccount for about one-third of total board members (BINDEPENDENCE). This figure suggests that Chinesecompanies adopt the minimum level of independent directors required by the CSRC.7

As mentioned, the SSSR considerably decreases the NTS of Chinese companies. Panels B to D of Table 2show descriptive statistics of NTSR and CONCENTRATION separately for pre-, during and post-SSSRobservations. The mean NTSR is about 60.1% for the pre-SSSR firm-years and this becomes 28.9% for thepost-SSSR observations. The mean and median differences in this variable are statistically and economicallysignificant between the pre- and post-SSSR periods. It is noteworthy that the firm-years after the SSSR have amuch higher standard deviation of NTSR than the firm-years before the SSSR do. This fact suggests that widevariations exist in NTSR in those periods. CONCENTRATION also shows a statistically and economically

7 The Guideline for Establishing Independent Directors System for Listed Companies,which is issued by CSRC in 2001, requires Chineselisted companies to gradually establish an “Independent Directors System” and to make qualified independent directors account for atleast one-third of board members.


significant reduction; the mean CONCENTRATION is approximately 43% for the pre-reform firm-years andthis percentage decreases to about 35% for post-reform observations.

6. Empirical results

6.1. ADA change before and after the SSSR

Firstly, we investigate the change in ADA before and after the SSSR to test H2. In order to examine thepure effects of the SSSR, we limit our sample to 1936 observations that completed their SSSR in 2005.

Table 3Analysis of earnings management surrounding the SSSR and NAS.

Year Mean Median 1st quartile 3rd quartile N t-Statistic

Panel A: ADA surrounding the SSSR for firms that completed their SSSR in 2005−4 0.141 0.108 0.051 0.190 119−3 0.114 0.081 0.036 0.147 133−2 0.128 0.088 0.045 0.191 147−1 0.121 0.096 0.036 0.169 1690 0.108 0.088 0.039 0.150 220 −1.074+1 0.110 0.089 0.039 0.154 232 −0.851+2 0.201 0.171 0.092 0.284 228 5.896***+3 0.152 0.123 0.063 0.212 231 2.931***+4 0.150 0.115 0.066 0.203 228 2.119**+5 0.160 0.136 0.073 0.230 229 3.340***Total 0.140 0.110 0.052 0.198 1936

Panel B: ADA surrounding the NAS for firms that completed their SSSR in 2005 or 20062001 0.117 0.085 0.036 0.156 8152002 0.109 0.084 0.037 0.146 8872003 0.112 0.084 0.043 0.147 9542004 0.105 0.079 0.035 0.141 10132005 0.101 0.076 0.034 0.146 10992006 0.107 0.080 0.038 0.147 11282007 0.175 0.150 0.080 0.241 1098 11.749***2008 0.136 0.110 0.050 0.187 1120 4.626***2009 0.140 0.111 0.055 0.192 1101 4.962***2010 0.152 0.126 0.060 0.216 1108 7.136***Total 0.126 0.097 0.045 0.174 10,323

Panel C: ADA surrounding the NAS for the entire sample2001 0.117 0.085 0.036 0.156 9822002 0.110 0.084 0.036 0.146 10572003 0.114 0.086 0.043 0.149 11312004 0.108 0.080 0.036 0.145 11952005 0.106 0.080 0.036 0.151 12792006 0.113 0.083 0.039 0.156 12712007 0.178 0.152 0.080 0.243 1221 13.120***2008 0.143 0.113 0.054 0.197 1263 6.342***2009 0.147 0.112 0.056 0.196 1257 6.448***2010 0.160 0.128 0.061 0.221 1283 9.046***Total 0.130 0.099 0.045 0.178 11,939

Notes:This table indicates earnings management (ADA) surrounding the SSSR (Panel A) and NAS (Panels B and C). In order to measure thepure effect of the SSSR, Panel A focuses only on firms that completed their SSSR in 2005. In order to avoid the SSSR contaminating theresults of NAS, Panel B employs as sample firms those that completed the SSSR in 2005 or 2006. We include the whole sample inPanel C. In Panel A, Year 0 refers to the years in the process of the SSSR. Year−1 (+1) indicates one year before (after) the initiation(completion) of the SSSR. The t-statistic in Panel A is for the null hypothesis that the mean ADA is identical to that of Year −1. InPanels B and C, the t-statistic is for the null hypothesis that the mean ADA is identical to that of Year 2003. *, ** and *** indicatesignificance at the 10%, 5% and 1% levels, respectively. ADA is the absolute value of discretionary accruals estimated by thecross-sectional Modified Jones Model.

Table 4Correlation matrix.

C_ADA C_NTSR NTSR C_CONCENTRATION CONCENTRATION D_SOE BOARDSIZE

Panel A: Correlation matrix among variables adopted in regression of C_ADA (−1, +1) (for firms that completed their SSSR in 2005)C_ADA 1.000C_NTSR 0.148 1.000NTSR −0.116 −0.265 1.000C_CONCENTRATION 0.090 0.041 −0.013 1.000CONCENTRATION −0.063 0.320 0.460 −0.385 1.000D_SOE 0.027 0.251 −0.109 −0.125 0.404 1.000BOARDSIZE 0.001 −0.045 0.082 −0.068 0.023 0.150 1.000BINDEPENDENCE 0.037 −0.043 0.124 0.098 −0.054 −0.151 −0.103PDIREC_SOWN 0.277 −0.154 −0.216 −0.045 −0.157 −0.060 −0.039D_DUAL −0.075 −0.132 0.028 0.069 −0.142 −0.361 −0.006SUPERSIZE −0.092 −0.002 0.051 −0.067 0.199 0.339 0.349PSUPER_SOWN 0.252 −0.100 −0.183 −0.092 −0.041 0.025 −0.019SIZE 0.074 0.263 −0.095 0.112 0.271 0.412 0.185ROA 0.039 0.085 0.090 0.067 0.094 0.029 0.054LEV 0.023 −0.096 −0.006 −0.052 −0.100 −0.050 0.109Q −0.094 0.004 0.092 0.115 −0.095 −0.140 −0.129

Panel B: Correlation matrix among variables adopted in the regression of C_ADA (2003, 2008) (for the whole sample)C_ADA 1.000C_NTSR 0.116 1.000NTSR 0.057 −0.286 1.000C_CONCENTRATION 0.075 0.340 −0.122 1.000CONCENTRATION 0.046 0.091 0.477 −0.478 1.000D_SOE 0.001 0.039 0.120 −0.134 0.337 1.000BOARDSIZE 0.034 −0.034 0.069 0.010 −0.022 0.159 1.000BINDEPENDENCE 0.044 −0.046 −0.016 0.008 −0.054 −0.061 −0.274PDIREC_SOWN −0.052 −0.028 −0.172 −0.108 0.027 0.094 −0.037D_DUAL −0.024 −0.018 −0.004 0.038 −0.079 −0.111 −0.061SUPERSIZE 0.039 −0.020 0.047 −0.015 0.070 0.206 0.323PSUPER_SOWN −0.037 −0.051 −0.110 −0.095 0.041 0.087 0.000SIZE 0.000 0.038 −0.037 −0.061 0.246 0.272 0.261ROA −0.063 −0.017 0.064 0.024 0.121 0.093 0.079LEV 0.037 0.024 −0.049 0.049 −0.067 −0.036 0.025Q 0.025 0.093 0.077 0.145 −0.029 −0.083 −0.071

BINDEPENDENCE PDIREC_SOWN D_DUAL SUPERSIZE PSUPER_SOWN SIZE ROA LEV Q

Notes:Panel A presents correlation coefficients among variables used in the regression of C_ADA (−1, +1). C_Variable (tb, ta) indicates thechange in Variable during Year tb to Year ta. Year −1 (+1) refers to the year before (after) the initiation (completion) of the SSSR.Panel A adopts the same interval (−1, +1) for C_NTSR and C_CONCENTRATION, while we use the Year−1 value for other variables.In accordance with the following regression analysis, Panel A limits our attention to firms that completed the SSSR in 2005. Panel Bindicates correlation coefficients among variables used in the regression of C_ADA (2003, 2008). The same interval (2003, 2008) isemployed for C_NTSR and C_CONCENTRATION, while we adopt the year 2003 value for NTSR and CONCENTRATION and the year2006 value for other variables. This panel presents figures for the whole sample. ADA is the absolute value of discretionary accrualsestimated by the cross-sectional Modified Jones Model. NTSR is the percentage of non-tradable shares over total outstanding shares.CONCENTRATION is the percentage ownership by the largest shareholder. D_SOE is a dummy variable that takes a value of one forstate-owned enterprises, and zero for others. BOARDSIZE is the natural logarithm of the number of board members.BINDEPENDENCE is the proportion of independent directors over total board members. PDIREC_SOWN is the proportion of boardmembers who own the firm's shares over all board members. D_DUAL is a dummy variable that takes a value of one if the CEO alsoserves as the chairperson of the board. SUPERSIZE is the natural logarithm of the number of supervisory board members.PSUPER_SOWN is the proportion of supervisory board members who own shares of the company over total supervisors. SIZE is thenatural logarithm of total assets. ROA is the return on assets. LEV is the liabilities divided by equity. Q is the total book value ofliabilities and market value of equity divided by the book value of total assets.


Panel A of Table 3 presents the results. We define Year 0 as the year in the process of SSSR (yearsbetween Year −1 and Year +1; this is year 2005 for the sample of this analysis). The t-statistics arepresented for the null hypothesis that the mean ADA for the year is identical to that of Year −1. Wealso employ different years (Year −2 to −4) for the benchmark year and obtain the qualitatively sameresult (not reported). The result indicates that observations of Year +1 do not show a significantdecline in the level of earnings management compared to those of Year −1. This result does not

Table 4 (continued)

BINDEPENDENCE PDIREC_SOWN D_DUAL SUPERSIZE PSUPER_SOWN SIZE ROA LEV Q

Panel A: Correlation matrix among variables adopted in regression of C_ADA (−1, +1) (for firms that completed their SSSR in 2005)

1.000−0.120 1.000

0.047 0.069 1.000−0.119 −0.076 −0.137 1.000−0.131 0.757 −0.023 0.001 1.000−0.100 0.146 −0.200 0.264 0.216 1.000−0.016 −0.035 −0.002 0.002 −0.084 0.106 1.000

0.003 −0.050 0.038 0.033 0.025 0.135 −0.018 1.0000.104 −0.159 0.072 −0.167 −0.175 −0.364 0.384 −0.286 1.000

Panel B: Correlation matrix among variables adopted in the regression of C_ADA (2003, 2008) (for the whole sample)

1.000−0.083 1.000

0.011 −0.027 1.000−0.065 0.010 −0.079 1.000−0.030 0.618 −0.071 −0.019 1.000−0.012 0.206 −0.085 0.190 0.168 1.000

0.001 0.059 0.002 0.049 0.033 0.284 1.0000.003 −0.026 0.019 −0.017 0.019 0.117 −0.147 1.000

−0.014 −0.109 0.044 −0.042 −0.077 −0.315 −0.005 −0.180 1.000


support H2 that Chinese firms decrease earnings management due to the decreased NTS through theimplementation of the SSSR. Panel A also suggests that ADA significantly increases in Year +2 andlater. We should not consider this result to test H2, since the introduction of IFRS-convergent NAS islikely to contaminate the result. Shen et al. (2009) argue that earnings management increases due tothe SSSR, since controlling shareholders are obsessed with huge capital gains. Our results are alsoinconsistent with their finding.


6.2. ADA change before and after the introduction of NAS

Panels B and C of Table 3 show ADA before and after the mandatory adoption of IFRS-convergent NAS(Year 2007). Most listed companies in China initiated their SSSR in 2005 or 2006. In order to avoid theSSSR contaminating our result, Panel B employs as sample firms those that completed the SSSR in 2005 or2006. In Panel C, we present results for the entire sample as a robustness check. Panels B and C presentt-statistics for the null hypothesis that the mean ADA for the year is identical to that for year 2003. We alsoadopt different years (2001, 2002 and 2004) as benchmark years and find the qualitatively same result(not reported). Both panels suggest that ADA is much higher for year 2007 and later compared to year2003 (H1b is supported). For example, the ADA for year 2007 is approximately 1.5 times that of year 2003.All the mean differences are statistically significant at the one percent level. A potential criticism to theresult is that the ADA change does not accurately capture the effect of NAS introduction, since it includesthe period of SSSR. Given that SSSR should decrease ADA, however, the inclusion of the SSSR period shouldbias the result downward. In addition, our former analysis finds an insignificant change in ADA surroundingthe SSSR. Taken all together, the present result serves as strong evidence that Chinese companies increasedearnings management in response to the introduction of NAS. This result supports the previous researchfindings that IFRS decreases earnings quality in several countries including China (Callao et al., 2007; He et al.,2012; Rudra and Bhattacharjee, 2012), as well as the argument that Chinese economic and institutionalconditions are incompatible with the IFRS-convergent NAS (He et al., 2012; Peng and Bewley, 2010).

In contrast, Liu et al. (2011) find that earnings quality in Chinese listed companies has improved withrestricted earnings management since 2007. A potential reason for the difference is that Liu et al. (2011)employ income smoothing for a measure of earnings management. Given that relatively long-term dataare needed to accurately measure income smoothing, we argue that our method (use of discretionaryaccruals that take one value for every single firm-year) presents a reasonable analysis to trace the effect of aspecific event on earnings management. The result provides evidence that accounting standards have asignificant impact on the degree of earningsmanagement. Together with the results of Panel A, we argue thataccounting standards have stronger direct impacts on earnings management than insiders' incentives do.

In addition, the result suggests that IFRS-convergent NAS is incompatible with Chinese economic andinstitutional environments. As mentioned, IFRS adopts principles-based standards, which provide muchdiscretion on earnings management. This characteristic is likely to damage earnings quality, combined withChinese corporate insiders' incentives for extracting private benefits.

6.3. Regression results of the change in ADA

As mentioned, H2 implies a positive relationship between the changes in earnings management andNTSR surrounding the SSSR. To test this idea, we execute regression analyses of Eq. (3) by using companiesthat completed their SSSR in 2005. We adopt as the dependent variable the change in ADA from Year tb toYear +1 (C_ADA (tb, +1)) and the most important explanatory variable is C_NTSR during the same period(C_NTSR (tb, +1)). We also include C_CONCENTRATION and other variables to examine their effects onthe change in earnings management surrounding the SSSR. Table 4 exhibits the correlation matrix amongvariables. We find that the correlations are relatively low, except for that of NTSR and CONCENTRATION.We include NTSR and CONCENTRATION (as well as C_NTSR and C_CONCENTRATION) separately in theindependent variables to avoid potential multicollinearity problems.

Table 5 presents the regression results. C_NTSR has a positive and significant coefficient in theregressions of C_ADA (−3, +1) and C_ADA (−1, +1). The regression of C_ADA (−4, +1) also engendersa marginally significant coefficient on C_NTSR. These results suggest that firms showing a reduction inearnings management also show a reduction in NTS. This result is consistent with H2, although Panel A ofTable 3 does not show evidence that SSSR significantly decreases earnings management of the averagefirm. To a certain degree, insiders' incentives have an explanatory power of earnings management.Table 5 does not engender a significant coefficient on C_CONCENTRATION except in the regression ofC_ADA (−3, +1). We interpret that the reduction in NTS represents the decreased incentives forearnings management better than the reduction in CONCENTRATION.

Although there is a positive and significant relationship between the changes in NTSR and ADAsurrounding the SSSR, Table 3 suggests that this effect is relatively marginal and eliminated by the

Table 5Regression results of the change in earnings management surrounding the SSSR.

Interval Year (−4, +1) Year (−4, +1) Year (−3, +1) Year (−3, +1) Year (−2, +1) Year (−2, +1) Year (−1, +1) Year (−1, +1)

Dependent variable C_ADA (−4, +1) C_ADA (−4, +1) C_ADA (−3, +1) C_ADA (−3, +1) C_ADA (−2, +1) C_ADA (−2, +1) C_ADA (−1, +1) C_ADA (−1, +1)

Coef t-Stat Coef t-Stat Coef t-Stat Coef t-Stat Coef t-Stat Coef t-Stat Coef t-Stat Coef t-Stat

C_NTSR 0.202* 1.81 0.242** 2.26 0.136 1.33 0.215*** 2.68NTSR −0.073 −0.65 0.102 0.74 0.131 1.39 −0.004 −0.04C_CONCENTRATION 0.118 0.67 0.356*** 2.70 0.117 0.67 0.211 1.29CONCENTRATION 0.009 0.07 0.125 1.18 0.158* 1.92 0.010 0.12D_SOE 0.007 0.19 0.018 0.48 0.009 0.35 0.006 0.20 0.006 0.21 −0.004 −0.14 0.004 0.17 0.020 0.73BOARDSIZE 0.042 0.76 0.040 0.68 0.019 0.32 0.037 0.61 0.043 0.75 0.055 0.92 0.039 0.74 0.033 0.64BINDEPENDENCE 0.136 0.51 0.093 0.38 0.052 0.27 0.033 0.18 0.094 0.35 0.120 0.43 0.240 1.20 0.199 1.04PDIREC_SOWN −0.031 −0.42 −0.029 −0.39 0.141* 1.86 0.144** 2.07 −0.051 −0.47 −0.054 −0.5 0.143* 1.91 0.120 1.58D_DUAL −0.014 −0.28 −0.017 −0.32 −0.022 −0.41 −0.032 −0.63 0.006 0.15 0.004 0.10 −0.033 −0.89 −0.037 −1.01SUPERSIZE 0.019 0.43 0.006 0.12 −0.010 −0.25 −0.021 −0.55 −0.038 −1.09 −0.044 −1.24 −0.039 −1.03 −0.048 −1.26PSUPER_SOWN 0.062 1.24 0.047 0.89 −0.033 −0.70 −0.042 −0.97 0.080 1.17 0.077 1.15 0.047 0.98 0.053 1.03SIZE −0.007 −0.37 −0.002 −0.09 −0.015 −0.94 −0.015 −0.93 −0.002 −0.15 −0.004 −0.25 −0.012 −0.82 −0.010 −0.61ROA −0.032 −0.09 −0.061 −0.17 0.001 0.01 −0.047 −0.24 0.090 0.37 0.079 0.34 0.258 1.20 0.284 1.31LEV −0.016 −0.81 −0.019 −0.84 −0.001 −0.05 −0.001 −0.10 0.001 0.06 0.003 0.21 0.006 0.43 0.004 0.24Q −0.032 −1.61 −0.026 −1.30 −0.0002 −0.01 0.005 0.38 −0.030 −1.53 −0.027 −1.45 −0.020 −1.22 −0.023 −1.49Constant 0.091 0.21 −0.068 −0.15 0.228 0.65 0.214 0.61 −0.074 −0.23 −0.069 −0.22 0.145 0.47 0.115 0.35R2 0.082 0.049 0.074 0.086 0.066 0.070 0.151 0.130N 119 119 133 133 147 147 169 169

Notes:This table presents regression results of the change in earnings management (C_ADA) surrounding the SSSR. In order to obtain the pure effect of the SSSR, we focus on firms that completed theirSSSR in 2005. We choose Year (tb, +1) as the interval to compute C_ADA, C_NTSR and C_CONCENTRATION (tb = −4, −3, −2 or −1). We use the Year tb value for NTSR and CONCENTRATION,while we use the Year−1 value for other governance factors. Year 0 refers to the years in the process of the SSSR. Year −1 (+1) indicates one year before (after) the initiation (completion) of theSSSR. *, ** and *** indicate significance at the 10%, 5% and 1% levels, respectively. ADA is the absolute value of discretionary accruals estimated by the cross-sectional Modified Jones Model. NTSR isthe percentage of non-tradable shares over total outstanding shares. CONCENTRATION is the percentage ownership by the largest shareholder. D_SOE is a dummy variable that takes a value of onefor state-owned enterprises, and zero for others. BOARDSIZE is the natural logarithm of the number of board members. BINDEPENDENCE is the proportion of independent directors over total boardmembers. PDIREC_SOWN is the proportion of board members who own the firm's shares over all board members. D_DUAL is a dummy variable that takes a value of one if the CEO also serves as thechairperson of the board. SUPERSIZE is the natural logarithm of the number of supervisory board members. PSUPER_SOWN is the proportion of supervisory board members who own shares of thecompany over total supervisors. SIZE is the natural logarithm of total assets. ROA is the return on assets. LEV is the liabilities divided by equity. Q is the total book value of liabilities and market valueof equity divided by the book value of total assets.

91Y.Zhang

etal./

Emerging

Markets

Review16

(2013)78

–99

Table 6Regression results of the change in earnings management surrounding the NAS.

Interval (2003, 2007) (2003, 2007) (2003, 2008) (2003, 2008) (2003, 2009) (2003, 2009) (2003, 2010) (2003, 2010)

Dependent variable C_ADA(2003, 2007)

C_ADA(2003, 2007)

C_ADA(2003, 2008)

C_ADA(2003, 2008)

C_ADA(2003, 2009)

C_ADA(2003, 2009)

C_ADA(2003, 2010)

C_ADA(2003, 2010)

Coef t-Stat Coef t-Stat Coef t-Stat Coef t-Stat Coef t-Stat Coef t-Stat Coef t-Stat Coef t-Stat

Panel A: Regression results of C_ADA for firms that completed their SSSR in 2005 or 2006C_NTSR 0.129** 2.53 0.079** 1.99 0.089*** 3.23 0.080*** 2.74NTSR 0.059 1.21 0.124*** 2.75 0.163*** 3.02 0.129** 2.27C_CONCENTRATION −0.040 −0.55 0.054 0.80 0.148** 2.33 0.046 0.75CONCENTRATION 0.035 0.89 0.094** 2.44 0.114*** 2.87 0.079* 1.84D_SOE 0.012 0.95 0.009 0.74 −0.008 −0.68 −0.011 −0.89 0.003 0.21 −0.002 −0.14 0.004 0.35 −0.001 −0.09BOARDSIZE 0.026 1.04 0.022 0.89 0.001 0.02 0.009 0.35 0.026 1.06 0.038 1.52 0.028 0.99 0.038 1.34BINDEPENDENCE 0.126 1.27 0.110 1.12 0.220** 2.12 0.222** 2.14 0.176* 1.69 0.179* 1.72 0.016 0.14 0.018 0.15PDIREC_SOWN 0.004 0.12 −0.007 −0.20 −0.006 −0.19 −0.010 −0.31 0.003 0.08 0.002 0.08 0.029 0.78 0.029 0.81D_DUAL −0.014 −0.93 −0.014 −0.89 −0.016 −1.02 −0.016 −1.01 −0.018 −1.11 −0.018 −1.14 0.012 0.71 0.011 0.63SUPERSIZE 0.018 0.99 0.018 1.00 0.027 1.53 0.025 1.44 0.031* 1.82 0.026 1.55 0.027 1.43 0.025 1.32PSUPER_SOWN 0.006 0.28 0.003 0.14 0.004 0.18 −0.0002 −0.01 0.010 0.49 0.009 0.43 0.010 0.42 0.007 0.27SIZE 0.005 0.81 0.007 1.20 0.005 0.79 0.003 0.42 −0.003 −0.52 −0.006 −1.07 0.004 0.62 0.002 0.36ROA 0.111 1.07 0.153 1.49 −0.041 −0.41 −0.039 −0.39 −0.145 −1.26 −0.153 −1.37 −0.168* −1.67 −0.165* −1.66LEV 0.010** 2.48 0.011*** 2.66 0.005 0.99 0.005 1.06 0.005 1.17 0.004 1.07 0.014** 2.47 0.014** 2.43Q 0.004 0.85 0.006 1.19 0.001 0.22 0.001 0.30 −0.008 −1.25 −0.009 −1.41 −0.001 −0.21 −0.001 −0.22Constant −0.219 −1.52 −0.263* −1.92 −0.259* −1.89 −0.204 −1.52 −0.124 −0.96 −0.045 −0.35 −0.217 −1.48 −0.188 −1.32R2 0.034 0.028 0.024 0.020 0.037 0.032 0.041 0.033N 924 924 944 944 928 928 930 930

92Y.Zhang

etal./

Emerging

Markets

Review16

(2013)78

–99

Panel B: Regression results of C_ADA for the whole sampleC_NTSR 0.150*** 3.30 0.131*** 3.44 0.121*** 4.68 0.104*** 4.03NTSR 0.070 1.51 0.137*** 3.17 0.188*** 3.57 0.120** 2.27C_CONCENTRATION 0.018 0.26 0.159** 2.46 0.199*** 3.44 0.071 1.26CONCENTRATION 0.051 1.37 0.123*** 3.34 0.127*** 3.35 0.075* 1.87D_SOE 0.011 0.96 0.010 0.82 −0.007 −0.61 −0.008 −0.70 −0.002 −0.20 −0.005 −0.40 −0.001 −0.11 −0.006 −0.45BOARDSIZE 0.021 0.84 0.018 0.75 0.030 1.22 0.038 1.54 0.040 1.60 0.052** 2.05 0.035 1.30 0.045* 1.66BINDEPENDENCE 0.094 0.99 0.081 0.85 0.204** 2.02 0.204** 2.00 0.167 1.52 0.167 1.51 0.051 0.44 0.047 0.41PDIREC_SOWN −0.003 −0.10 −0.016 −0.50 −0.014 −0.42 −0.015 −0.45 −0.001 −0.03 −0.001 −0.02 0.041 1.16 0.041 1.20D_DUAL −0.015 −0.98 −0.014 −0.89 −0.010 −0.60 −0.009 −0.59 −0.007 −0.45 −0.008 −0.48 0.021 1.24 0.020 1.17SUPERSIZE 0.015 0.87 0.014 0.84 0.016 0.98 0.014 0.85 0.031* 1.91 0.028* 1.69 0.030* 1.67 0.028 1.57PSUPER_SOWN 0.001 0.05 0.001 0.06 −0.002 −0.11 −0.006 −0.31 0.014 0.68 0.013 0.63 0.0004 0.02 −0.002 −0.07SIZE 0.003 0.48 0.004 0.71 0.001 0.22 −0.002 −0.26 −0.006 −0.99 −0.010* −1.75 0.002 0.28 0.0003 0.05ROA 0.100 1.17 0.108 1.27 −0.148 −1.63 −0.169* −1.84 −0.183* −1.82 −0.205** −2.05 −0.169** −2.36 −0.208*** −2.93LEV 0.005 1.36 0.005 1.42 0.003 0.69 0.003 0.73 0.001 0.35 0.002 0.39 0.007* 1.69 0.007* 1.65Q 0.008* 1.70 0.011** 2.22 0.002 0.37 0.002 0.37 −0.006 −1.07 −0.006 −0.99 0.002 0.48 0.004 0.70Constant −0.141 −1.05 −0.167 −1.29 −0.209 −1.59 −0.148 −1.15 −0.094 −0.72 0.014 0.11 −0.174 −1.26 −0.162 −1.18R2 0.032 0.020 0.035 0.029 0.051 0.041 0.043 0.028N 1018 1018 1055 1055 1046 1046 1060 1060

Notes:This table presents the regression results of the change in earnings management (C_ADA) surrounding the NAS. To avoid SSSR contaminating the result, we focus on firms that completed their SSSRin 2005 or 2006 in Panel A, while we adopt the whole sample in Panel B. We choose (2003, 2007 + j) as the interval to compute C_ADA, C_NTSR and C_CONCENTRATION (j = 0, 1, 2, or 3). We usethe year 2003 value for NTSR and CONCENTRATION, while we adopt the year 2006 value for other governance factors. *, ** and *** indicate significance at the 10%, 5% and 1% levels, respectively.ADA is the absolute value of discretionary accruals estimated by the cross-sectional Modified Jones Model. NTSR is the percentage of non-tradable shares over total outstanding shares.CONCENTRATION is the percentage ownership by the largest shareholder. D_SOE is a dummy variable that takes a value of one for state-owned enterprises, and zero for others. BOARDSIZE is thenatural logarithm of the number of board members. BINDEPENDENCE is the proportion of independent directors over total board members. PDIREC_SOWN is the proportion of board members whoown the firm's shares over all board members. D_DUAL is a dummy variable that takes a value of one if the CEO also serves as the chairperson of the board. SUPERSIZE is the natural logarithm of thenumber of supervisory board members. PSUPER_SOWN is the proportion of supervisory board members who own shares of the company over total supervisors. SIZE is the natural logarithm of totalassets. ROA is the return on assets. LEV is the liabilities divided by equity. Q is the total book value of liabilities and market value of equity divided by the book value of total assets.

93Y.Zhang

etal./

Emerging

Markets

Review16

(2013)78

–99


introduction of NAS, which increases insiders' discretion on earnings management. Nevertheless, thereduction in NTS potentially affects the change in earnings management before and after the introductionof NAS through its effects on insiders' incentives. Specifically, insiders who still have strong incentives forextracting private benefits will aggressively utilize the increased discretion on earnings management. Totest this idea, we implement regression analyses of Eq. (4).

Table 6 presents the regression results. Again, Panel A employs only firms that completed their SSSR in2005 or 2006, while Panel B replicates the analysis for the whole sample as a robustness check. In bothpanels, we choose 2003 as the benchmark year to compute C_ADA, C_NTSR and C_CONCENTRATION. Wealso adopt different years (2001, 2002 and 2004) as a benchmark year and find the qualitatively sameresult (not reported).

C_NTSR has a positive and significant coefficient in all models. This result suggests that firms thatdecreased NTS less during the SSSR tended to increase their earnings management more after theintroduction of NAS. As mentioned, the reduced NTS through SSSR does not reduce the earningsmanagement of the average Chinese firm (Table 3). However, insiders' reduced incentive to extract privatebenefits affects how they increase earnings management in response to the introduction of NAS. In otherwords, insiders' incentives affect the level of earnings management, given a specific set of accountingstandards. Another possible story is that the reduction in NTS represents insiders' incentives to extractprivate benefits. Insiders who strongly desire to expropriate minority shareholders' wealth did notsubstantially decrease NTS under the SSSR and increased earnings management considerably. In eithercase, insiders' incentives affect how they take advantage of increased discretion on earnings management.This argument is reinforced by the finding that NTSR, at year 2003, has a positive and significant coefficientin many specifications. Overall, our change-to-change regression results support the conventional wisdomthat the divergence between control and cash flow rights increases earnings management (Fan andWong,2002; Firth et al., 2007; Gopalan and Jayaraman, 2012; Haw et al., 2004). The results also provide apotential explanation for why Chinese firms increased earnings management after the introduction ofNAS. Chinese corporate insiders have a strong incentive to extract private benefits due to the existence ofNTS. Those insiders are likely to view the flexibility in financial reporting (principle-based standards) as agood opportunity for earnings management.

As with the former result (Table 5), the coefficient of C_CONCENTRATION is insignificant in manymodels. LEVhas a positive and significant coefficient in Table 6, suggesting that high-leveraged companies havestrong incentives for earningsmanagement and thus utilize increased discretion on earningsmanagement.Wefind no other variables to significantly affect the change in earnings management.

6.4. Regression results of the level of ADA

In this section, we conduct a regression analysis of the level of ADA as an additional analysis. We introducea dummy variable that takes a value of one for year 2007 and later, and zero otherwise (D_NAS), to test H1aand H1b. Another key independent variable is NTSR (for H2). We also include the interaction term of D_NASand NTSR to test the idea that the increase in earnings management due to the adoption of NAS is affected byNTSR. Table 7 shows the regression results. We employ the firm fixed-effects model (Panels C and D ofTable 7) for estimation, as well as pooled OLS (Panels A and B). We compute t-statistics in the OLS estimationby using firm-clustering standard errors. The Hausman test also suggests that we adopt the fixed-effectsmodel rather than the random-effects model. Panels A and C present the regression results of ADA for firmsthat completed their SSSR in 2005 or 2006, while Panels B and D show the results for the whole sample.

Consistent with the former results (H1b), D_NAS has a positive and significant coefficient in mostestimations, which suggests thatfirms tend tomanage earningsmore in the years after the implementation ofIFRS-convergent NAS. The introduction of NAS is incompatible with Chinese corporate environments andprovides managers with more opportunities to engage in earnings management.

As for insiders' incentives, NTSR has a positive and significant coefficient. This result, which is consistentwith the findings of Firth et al. (2007), provides support for the suggestion that NTS engenders expropriationproblems and gives insiders incentives for earnings management to conceal the evidence. Importantly, themodels carry a positive and significant coefficient on the interaction term of D_NAS and NTSR. As with theformer finding, this result provides evidence that the increase in earnings management due to theintroduction of NAS is positively affected byNTSR. This result supports the idea that insiders' incentives affect

Table 7Regression results of the level of earnings management.

Coef t-Stat Coef t-Stat Coef t-Stat Coef t-Stat

Panel A: OLS results of ADA for firms that completed their SSSR in 2005 or 2006D_NAS 0.051*** 13.26 0.031*** 3.88 0.030*** 9.34 0.008 1.12NTSR 0.073*** 9.17 0.044*** 3.39D_NAS × NTSR 0.040*** 2.57CONCENTRATION 0.021** 1.96 −0.001 −0.07D_NAS × CONCENTRATION 0.059*** 3.13D_SOE −0.012*** −3.73 −0.012*** −3.66 −0.012*** −3.38 −0.012*** −3.34BOARDSIZE 0.002 0.27 0.002 0.26 0.005 0.79 0.005 0.76BINDEPENDENCE −0.008 −0.63 −0.010 −0.83 −0.016 −1.25 −0.017 −1.34PDIREC_SOWN −0.038*** −4.70 −0.040*** −4.85 −0.043*** −5.25 −0.041*** −5.02D_DUAL 0.003 0.71 0.003 0.71 0.003 0.75 0.003 0.83SUPERSIZE −0.0004 −0.09 −0.0002 −0.04 −0.001 −0.17 −0.001 −0.16PSUPER_SOWN 0.004 0.61 0.004 0.63 0.003 0.57 0.004 0.63SIZE 0.010*** 5.78 0.010*** 5.63 0.010*** 5.50 0.010*** 5.25ROA −0.022 −0.88 −0.020 −0.83 −0.016 −0.66 −0.017 −0.70LEV 0.004*** 3.30 0.004*** 3.30 0.004*** 3.25 0.004*** 3.15Q 0.008*** 5.73 0.008*** 5.83 0.008*** 5.91 0.008*** 5.94_cons −0.162*** −3.88 −0.141*** −3.26 −0.125*** −3.03 −0.107*** −2.57R2 0.068 0.068 0.057 0.058N 10,323 10,323 10,323 10,323

Panel B: OLS results of ADA for the whole sampleD_NAS 0.054*** 14.65 0.023*** 2.88 0.030*** 10.03 −0.0001 −0.02NTSR 0.090*** 11.42 0.045*** 3.48D_NAS × NTSR 0.063*** 4.13CONCENTRATION 0.035*** 3.42 0.004 0.38D_NAS × CONCENTRATION 0.083*** 4.59D_SOE −0.015*** −4.78 −0.015*** −4.68 −0.015*** −4.62 −0.015*** −4.50BOARDSIZE −0.002 −0.27 −0.002 −0.27 0.002 0.39 0.002 0.37BINDEPENDENCE 0.013 1.09 0.009 0.78 0.006 0.54 0.005 0.44PDIREC_SOWN −0.036*** −4.55 −0.037*** −4.74 −0.042*** −5.32 −0.039*** −4.95D_DUAL 0.003 0.83 0.003 0.80 0.003 0.88 0.003 0.95SUPERSIZE −0.002 −0.52 −0.002 −0.42 −0.003 −0.54 −0.002 −0.50PSUPER_SOWN 0.002 0.37 0.002 0.35 0.002 0.40 0.003 0.45SIZE 0.011*** 6.35 0.010*** 6.12 0.009*** 5.59 0.009*** 5.21ROA −0.088*** −4.63 −0.085*** −4.53 −0.084*** −4.46 −0.083*** −4.43LEV 0.002** 2.51 0.002** 2.49 0.003*** 2.69 0.002** 2.49Q 0.010*** 8.43 0.011*** 8.50 0.011*** 8.88 0.011*** 8.91_cons −0.168*** −4.38 −0.135*** −3.40 −0.115*** −3.00 −0.089** −2.31R2 0.082 0.084 0.067 0.070N 11,939 11,939 11,939 11,939

Panel C: Fixed-effects model results of ADA for firms that completed their SSSR in 2005 or 2006D_NAS 0.042*** 10.27 0.031*** 3.18 0.024*** 6.55 0.012* 1.72NTSR 0.073*** 8.54 0.055*** 3.31D_NAS × NTSR 0.021 1.18CONCENTRATION 0.043** 2.03 0.033 1.62D_NAS × CONCENTRATION 0.033* 1.74D_SOE −0.002 −0.42 −0.002 −0.42 −0.0003 −0.06 −0.001 −0.13BOARDSIZE 0.006 0.67 0.006 0.66 0.010 1.16 0.010 1.11BINDEPENDENCE −0.022 −1.60 −0.024* −1.73 −0.026* −1.88 −0.026* −1.86PDIREC_SOWN −0.033*** −2.69 −0.033*** −2.69 −0.033*** −2.72 −0.032*** −2.58D_DUAL −0.001 −0.13 −0.001 −0.12 −0.0002 −0.05 −0.00003 −0.01SUPERSIZE −0.001 −0.17 −0.001 −0.14 −0.001 −0.16 −0.001 −0.14PSUPER_SOWN 0.002 0.30 0.002 0.31 0.004 0.51 0.004 0.53SIZE 0.027*** 6.37 0.027*** 6.27 0.026*** 6.14 0.024*** 5.85ROA 0.041* 1.75 0.042* 1.78 0.043* 1.81 0.043* 1.83LEV 0.001 1.02 0.001 1.00 0.001 0.88 0.001 0.84Q 0.010*** 7.30 0.010*** 7.31 0.010*** 7.11 0.010*** 7.14_cons −0.523*** −5.65 −0.505*** −5.37 −0.481*** −5.21 −0.453*** −4.89R2 within 0.069 0.070 0.060 0.061

(continued on next page)


Table 7 (continued)

Coef t-Stat Coef t-Stat Coef t-Stat Coef t-Stat

R2 between 0.031 0.032 0.019 0.020R2 overall 0.048 0.049 0.036 0.038N 10,323 10,323 10,323 10,323

Panel D: Fixed-effects model results of ADA for the whole sampleD_NAS 0.049*** 12.66 0.031*** 3.27 0.030*** 8.72 0.014** 2.17NTSR 0.089*** 10.45 0.059*** 3.65D_NAS × NTSR 0.035** 2.01CONCENTRATION 0.078*** 3.99 0.063*** 3.36D_NAS × CONCENTRATION 0.044** 2.42D_SOE −0.010** −2.05 −0.010** −2.05 −0.010* −1.93 −0.010* −1.95BOARDSIZE 0.003 0.42 0.003 0.38 0.008 1.06 0.008 0.98BINDEPENDENCE −0.005 −0.39 −0.008 −0.61 −0.005 −0.39 −0.005 −0.38PDIREC_SOWN −0.035*** −3.02 −0.035*** −3.03 −0.037*** −3.13 −0.035*** −2.98D_DUAL 0.001 0.26 0.001 0.26 0.001 0.33 0.002 0.34SUPERSIZE −0.009 −1.25 −0.009 −1.19 −0.010 −1.30 −0.010 −1.26PSUPER_SOWN −0.007 −0.86 −0.006 −0.86 −0.004 −0.58 −0.004 −0.56SIZE 0.023*** 6.04 0.023*** 5.90 0.021*** 5.43 0.019*** 5.01ROA −0.009 −0.49 −0.009 −0.46 −0.006 −0.32 −0.005 −0.29LEV −0.0002 −0.15 −0.0002 −0.18 −0.0002 −0.24 −0.0003 −0.30Q 0.010*** 8.73 0.010*** 8.73 0.010*** 8.63 0.010*** 8.66_cons −0.430*** −5.08 −0.401*** −4.68 −0.368*** −4.36 −0.330*** −3.90R2 within 0.076 0.077 0.066 0.067R2 between 0.047 0.050 0.021 0.026R2 overall 0.064 0.066 0.046 0.049N 11,939 11,939 11,939 11,939

Notes:This table presents the regression results of the level of earnings management (ADA). We use pooled OLS (in Panels A and B) and thefirm-fixed effects model (in Panels C and D) for estimation. The t-statistics in pooled OLS regressions are computed by usingfirm-clustering standard errors. Panels A and C present the regression results of ADA for firms that completed their SSSR in 2005 or2006. Panels B and D show the regression results of ADA for the whole sample as a robustness check. *, ** and *** indicate significance atthe 10%, 5% and1% levels, respectively. ADA is the absolute value of discretionary accruals estimated by the cross-sectionalModified JonesModel. D_NAS is a dummy variable that takes a value of one if a firm-year is after the implementation of IFRS-convergent new accountingstandards, and zero otherwise. NTSR is the percentage of non-tradable shares over total outstanding shares. CONCENTRATION is thepercentage ownership by the largest shareholder. D_SOE is a dummy variable that takes a value of one for state-owned enterprises, andzero for others. BOARDSIZE is the natural logarithmof the number of boardmembers. BINDEPENDENCE is the proportion of independentdirectors over total board members. PDIREC_SOWN is the proportion of board members who own the firm's shares over all boardmembers. D_DUAL is a dummy variable that takes a value of one if the CEO also serves as the chairperson of the board. SUPERSIZE is thenatural logarithm of the number of supervisory board members. PSUPER_SOWN is the proportion of supervisory board members whoown shares of the company over total supervisors. SIZE is the natural logarithm of total assets. ROA is the return on assets. LEV is theliabilities divided by equity. Q is the total book value of liabilities and market value of equity divided by the book value of total assets.

Panel C: Fixed-effects model results of ADA for firms that completed their SSSR in 2005 or 2006


how firms increase earnings management in response to increased discretion on reported earnings. Wereplicate the analyses by replacing NTSR by CONCENTRATION. Although the former analyses find aninsignificant relationship between changes in ownership concentration and earnings management afterthe introduction of NAS, Table 7 engenders a positive and significant coefficient on the interaction term ofD_NAS and CONCENTRATION. As mentioned, Fan and Wong (2002) and Firth et al. (2007) find thatownership concentration significantly weakens earnings quality. The result is also consistent with theidea that insiders' incentives affect how firms increase earnings management in response to the increasedopportunity of earnings management (introduction of NAS).

With respect to the control variables, the coefficient of D_SOE is negative and statistically significant.Firms controlled by central or local governments have less incentive to manage reported earnings. This resultsupports the finding of Lo et al. (2010). Board ownership (represented by PDIREC_SOWN) is expected to alignthe interests of insiders with those of minority shareholders, thereby mitigating earnings management.Consistent with this idea, Table 7 engenders a negative coefficient on PDIREC_SOWN. However, we do notfind a significant coefficient on other governance variables (BOARDSIZE, BINDEPENDENCE and D_DUAL).Firm size (SIZE) has a positive impact on earnings management, which is consistent with the argument of


Watts and Zimmerman (1978) that large companies are more likely to manage (decrease) earnings in orderto reduce the possibility of social scrutiny and political intervention. Consistent with the former result, wefind that firms with high leverage (LEV) tend to manage earnings more (in OLS regression results), whichsupports thefinding of Dhaliwal et al. (1991). Additionally, growing companieswith higher Tobin's Qmanageearnings more. This result is consistent with the findings of Beaver et al. (1970) and Minton and Schrand(1999).

7. Conclusions

This paper investigates the impacts of accounting standards and insiders' incentives on earningsmanagement by using recent Chinese data. The data offer us an appropriate research setting to address theissue, since recent Chinese listed companies have experienced two noticeable institutional changes thataffect those factors. The SSSR, which was introduced in 2005, substantially decreased NTS, which inducesinsiders' expropriation problems. The mandatory implementation of IFRS-convergent NAS changes thedegree of managerial discretion on reported earnings due to the inherent flexibility in principle-basedstandards.

Although we do not find evidence that the SSSR decreased earnings management, on average, there is apositive and significant relationship between the reductions in earnings management and NTSR surroundingthe SSSR. However, the mandatory implementation of IFRS-convergent NAS in China significantly increasesearnings management. These results suggest that accounting standards have a stronger direct impact onearnings management than insiders' incentives do. The result is also consistent with the argument that theeconomic and institutional environments in China are incompatiblewith the adoption of IFRS, which expandsinsiders' opportunity to engage in earningsmanagement. Notably, the reduction in NTSR is negatively relatedto the increase in earnings management surrounding the introduction of NAS. This result suggests thatinsiders' incentives affect how firms increase earnings management in response to increased discretion onreported earnings. In brief, we argue that insiders' incentives affect the level of earnings management given aspecific set of accounting standards.

Numerous studies show evidence that corporate governance structures are associated with earningsquality. This study provides an important contribution to the literature by showing that accountingstandards are the primary determinant of the degree of earnings management and corporate governancestructures (or insiders' incentives) are subordinate. It would be also noteworthy that insiders' incentiveshave explanatory power on the variation of earnings management within a specific set of accountingstandards. Our results also present an important policy implication: SSSR aims at eliminating NTS, which isan important source of expropriation problems; however, the forced reduction of NTS does notsubstantially improve earnings quality, although there is a statistically significant linkage between thereductions in NTS and earnings management. Besides, the mandatory adoption of IFRS-convergent newaccounting standards provides more opportunities to engage in earnings management in China, whereinsiders have strong incentive to extract private benefits and the institutional settings are incompatible withthe new standards. Regulators should focus on the unintended effects of new regulations, with attention totheir incompatibility with local environments.

More detailed analyses on accounting techniques that have been used after the introduction of NASremain an important part of future studies. Future research also needs to take the characteristics ofaccounting standards into consideration to examine the relationship between earnings management andcorporate governance.

References

Ahmed, K., Hossain, M., Adams, M., 2006. The effects of board composition and board size on the informativeness of annualaccounting earnings. Corporate Governance: An International Review 14, 418–431.

Ashbaugh, H., Pincus, M., 2001. Domestic accounting standards, international accounting standards, and the predictability ofearnings. Journal of Accounting Research 39, 417–434.

Atiase, R., 1985. Predisclosure information, firm capitalization, and security price behavior around earnings announcements. Journalof Accounting Research 23, 21–36.

Ball, R., Robin, A., Wu, J.S., 2003. Incentives versus standards: properties of accounting income in four East Asian countries. Journal ofAccounting and Economics 36, 235–270.

http://refhub.elsevier.com/S1566-0141(13)00034-4/rf0005









Barth, M., Landsman, W., Lang, M., 2008. International accounting standards and accounting quality. Journal of Accounting Research46, 467–498.

Barth, M., Landsman, W., Lang, M., Williams, C., 2012. Are IFRS-based and US GAAP-based accounting amounts comparable? Journalof Accounting and Economics 54, 68–93.

Bartov, E., Gul, F.A., Tsui, J.S.L., 2001. Discretionary-accruals models and audit qualifications. Journal of Accounting and Economics 30,421–452.

Beasley, M., 1996. An empirical analysis between the board of director composition and financial statement fraud. The AccountingReview 71, 443–466.

Beaver, W., Kettler, P., Scholes, M., 1970. The association between market determined and accounting determined risk measures. TheAccounting Review 45, 654–682.

Breeden, R., 1994. Foreign companies and U.S. markets in a time of economic transformation. Fordham International Law Journal 17,S77–S96.

Bruggemann, U., Hitz, J., Sellhorn, T., 2010. Intended and unintended consequences of mandatory IFRS adoption: review of extantevidence and suggestions for future research. Working paper Lancaster University.

Burgstahler, D.C., Hail, L., Leuz, C., 2006. The importance of reporting incentives: earnings management in European private andpublic firms. The Accounting Review 81, 983–1016.

Butler, M., Leone, A.J., Willenborg, M., 2004. An empirical analysis of auditor reporting and its association with abnormal accruals.Journal of Accounting and Economics 37, 139–165.

Callao, S., Jarne, J.I., Lainez, J.A., 2007. Adoption of IFRS in Spain: effect on the comparability and relevance of financial reporting.Journal of International Accounting, Auditing and Taxation 16, 148–178.

Chung, R., Firth, M., Kim, J.B., 2002. Institutional monitoring and opportunistic earnings management. Journal of Corporate Finance 8,29–48.

Claessens, S., Djankov, S., Lang, L.H.P., 2000. The separation of ownership and control in East Asian corporations. Journal of FinancialEconomics 58, 81–112.

Claessens, S., Djankov, S., Fan, J.P.H., Lang, L.H.P., 2002. Disentangling the incentive and entrenchment effects of large shareholdings.Journal of Finance 57, 2741–2771.

Clarkson, P., Hanna, J.D., Richardson, G.D., Thompson, R., 2011. The impact of IFRS adoption on the value relevance of book value andearnings. Journal of Contemporary Accounting and Economics 7, 1–17.

Cormier, D., Demaria, S., Lapointe-Antunes, P., Teller, R., 2009. First-time adoption of IFRS, managerial incentives, and value-relevance:some French evidence. Journal of International Accounting Research 8, 1–22.

Daske, H., Hail, L., Leuz, C., Verdi, R., 2008. Mandatory IFRS reporting around the world: early evidence on the economic consequences.Journal of Accounting Research 46, 1085–1142.

Dechow, P.M., Sloan, R.G., Sweeney, A.P., 1995. Detecting earnings management. The Accounting Review 70, 113–144.Dechow, P.M., Sloan, R.G., Sweeney, A.P., 1996. Causes and consequences of earnings manipulation: an analysis of firms subject to

enforcement actions by the SEC. Contemporary Accounting Research 13, 1–36.Dechow, P.M., Ge, W.L., Schrand, C., 2010. Understanding earnings quality: a review of the proxies, their determinants and their

consequences. Journal of Accounting and Economics 50, 344–401.Denis, D.K., McConnell, J.J., 2003. International corporate governance. Journal of Financial and Quantitative Analysis 38, 1–36.Dhaliwal, D., Lee, K., Fargher, N., 1991. The association between unexpected earnings and abnormal security returns in the presence

of financial leverage. Contemporary Accounting Research 8, 20–41.Dimitropoulos, P.E., Asteriou, D., 2010. The effect of board composition on the informativeness and quality of annual earnings:

empirical evidence from Greece. Research in International Business and Finance 24, 190–205.Ding, Y., Su, X., 2008. Implementation of IFRS in a regulated market. Journal of Accounting and Public Policy 27, 474–479.Djatej, A., Gao, G., Sarikas, R.H.S., Senteney, D.L., 2011. Eastern and Western European firms' public and private information quality:

the comparative impact of degree of implementation of IFRS. Emerging Markets Review 12, 111–129.Dyck, A., Zingales, L., 2004. Private benefits of control: an international comparison. Journal of Finance 59, 537–599.European Communities, 2002. Regulation (EC) no. 1606/2002 of the European Parliament and of the Council of 19 July 2002 on the

application of international accounting standards. Official Journal of the European Communities L 243, 1–4.Fan, J.P.H., Wong, T.J., 2002. Corporate ownership structure and the informativeness of accounting earnings in East Asia. Journal of

Accounting and Economics 33, 401–425.Firth, M., Fung, P.M.Y., Rui, O.M., 2007. Ownership, two-tier board structure, and the informativeness of earnings—evidence from

China. Journal of Accounting and Public Policy 26, 463–496.Forker, J.J., 1992. Corporate governance and disclosure quality. Accounting and Business Research 22, 111–124.Francis, J., Krishnan, J., 1999. Accounting accruals and auditor reporting conservatism. Contemporary Accounting Research 16,

135–166.Francis, J., Wang, D., 2008. The joint effect of investor protection and Big 4 audits on earnings quality around the world. Contemporary

Accounting Research 25, 157–191.Freeman, R., 1987. The association between accounting earnings and security returns for large and small firms. Journal of Accounting

and Economics 9, 195–228.Gabrielsen, G., Gramlich, J.D., Plenborg, T., 2002. Managerial ownership, information content of earnings, and discretionary accruals

in a non-US setting. Journal of Business Finance and Accounting 29, 967–988.Gopalan, R., Jayaraman, S., 2012. Private control benefits and earnings management: evidence from insider controlled firms. Journal

of Accounting Research 50, 117–157.Haw, I., Hu, B., Hwang, L., Wu, W., 2004. Ultimate ownership, income management, and legal and extra-legal institutions. Journal of

Accounting Research 42, 423–462.He, X.J., Wong, T.J., Young, D.Q., 2012. Challenges for implementation of fair value accounting in emerging markets: evidence from

China. Contemporary Accounting Research 29, 538–562.Holland, K., Jackson, R.H.G., 2004. Earnings management and deferred tax. Accounting and Business Research 34, 101–123.Holthausen, R., 2009. Accounting standards, financial reporting outcomes, and enforcement. Journal of Accounting Research 47,

447–458.








































































Iatridis, G.E., 2012. Audit quality in common-law and code-law emerging markets: evidence on earnings conservatism, agency costsand cost of equity. Emerging Markets Review 13, 101–117.

International Accounting Standards Committee (IASC), 1989. Framework for the preparation and presentation of financial statements.IASC, London, UK.

Jensen, M.C., 1993. The modern industrial revolution, exit, and the failure of internal control systems. Journal of Finance 48, 830–881.Johnson, S., La Porta, R., Lopez-De-Salinas, F., Shleifer, A., 2000. Tunneling. American Economic Review Papers and Proceedings 90,

22–27.Kim, J.B., Yi, C.H., 2006. Ownership structure, business group affiliation, listing status and earnings management: evidence from

Korea. Contemporary Accounting Research 23, 427–464.Klein, A., 2002. Audit committee, board of director characteristics, and earnings management. Journal of Accounting and Economics

33, 375–400.Kothari, S.P., Leone, A.J., Wasley, C.E., 2005. Performance matched discretionary accrual measures. Journal of Accounting and

Economics 39, 163–197.Krishnan, G., 2003. Audit quality and the pricing of discretionary accruals. Auditing: A Journal of Practice and Theory 22, 109–126.La Porta, R., Lopez-de-Silanes, F., Shleifer, A., 1999. Corporate ownership around the world. Journal of Finance 54, 471–518.La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R.W., 2000. Agency problems and dividend policies around the world. Journal of

Finance 55, 1–33.Leuz, C., 2006. Cross listing bonding and firms reporting incentives: a discussion of Lang, Ready and Wilson (2006). Journal of

Accounting and Economics 42, 285–299.Leuz, C., Nanda, D., Wysocki, P.D., 2003. Earnings management and investor protection: an international comparison. Journal of

Financial Economics 69, 505–527.Lins, K.V., 2003. Equity ownership and firm value in emerging markets. Journal of Financial and Quantitative Analysis 38, 159–184.Liu, Q., Lu, J., 2007. Corporate governance and earnings management in the Chinese listed companies: a tunneling perspective.

Journal of Corporate Finance 13, 881–906.Liu, C.H., Yao, L.J., Hu, N., Liu, L., 2011. The impact of IFRS on accounting quality in a regulated market: an empirical study of China.

Journal of Accounting, Auditing and Finance 26, 659–676.Lo, A.W.Y., Wong, R.M.K., Firth, M., 2010. Can corporate governance deter management from manipulating earnings? Evidence from

related-party sales transactions in China. Journal of Corporate Finance 16, 225–235.Minton, B., Schrand, C., 1999. The impact of cash flow volatility on discretionary investment and the costs of debt and equity

financing. Journal of Financial Economics 54, 423–460.Nobes, C., 1998. Towards a general model of the reasons for international differences in financial reporting. Abacus 34, 162–187.Park, Y.W., Shin, H.H., 2004. Board composition and earnings management in Canada. Journal of Corporate Finance 10, 431–457.Peasnell, K., Pope, P., Young, S., 2005. Board monitoring and earnings management: do outside directors influence abnormal accruals.

Journal of Business Finance and Accounting 32, 1311–1346.Peng, S.L., Bewley, K., 2010. Adaptability of fair value accounting in China: assessment of an emerging economy converging with IFRS.

Working paper York University.Raheja, C., 2005. Determinants of board size and composition: a theory of corporate boards. Journal of Financial and Quantitative

Analysis 40, 283–306.Rudra, T., Bhattacharjee, D., 2012. Does IFRS influence earnings management? Evidence from India. Journal of Management Research

4, 1–13.Shen, H.H., Huang, Z.K., Wu, L.S., 2009. The effects of Split-share Structure Reform on earnings management. China Journal of

Accounting Research 8, 40–48 (in Chinese).Shleifer, A., Vishny, R.W., 1997. A survey of corporate governance. Journal of Finance 52, 737–775.Sunder, S., 2009. IFRS and the accounting consensus. Accounting Horizons 23, 101–111.Tyrrall, D., Woodward, D., Rakhimbekova, A., 2007. The relevance of international financial reporting standards to a developing

country: evidence from Kazakhstan. The International Journal of Accounting 42, 82–110.Vafeas, N., 2000. Board structure and the informativeness of earnings. Journal of Accounting and Public Policy 19, 139–160.Wang, H.C., Liu, T.L., Lu, C., 2007. Corporate governance and earnings quality: evidence from China. China Soft Science 11, 122–128

(in Chinese).Warfield, T.D., Wild, J.J., Wild, K.L., 1995. Managerial ownership, accounting choices, and informativeness of earnings. Journal of

Accounting and Economics 20, 61–91.Watts, R.L., Zimmerman, J.L., 1978. Towards a positive theory of the determination of accounting standards. The Accounting Review

53, 112–134.Watts, R.L., Zimmerman, J.L., 1990. Positive accounting theory: a ten year perspective. The Accounting Review 56, 131–156.Xie, B., Davidson, W.N., DaDalt, P.J., 2003. Earnings management and corporate governance: the role of the board and the audit

committee. Journal of Corporate Finance 9, 295–316.Yeo, G.H.H., Tan, P.M.S., Ho, K.W., Chen, S.S., 2002. Corporate ownership structure and the informativeness of earnings. Journal of

Business Finance and Accounting 29, 1023–1046.



























































how do accounting standards and insiders' incentives affect …drthomaswu.com/uicfat/8.pdf ·...

Documents