Asian Economic and Financial Review THE EFFECT OF IFRS, INFORMATION ASYMMETRY AND CORPORATE GOVERNANCE ON THE QUALITY OF ACCOUNTING INFORMATION

Transcription

1 Asian Economic and Financial Review journal homepage: hp://aessweb.com/journal-deail.php?id=5002 THE EFFECT OF IFRS, INFORMATION ASYMMETRY AND CORPORATE GOVERNANCE ON THE QUALITY OF ACCOUNTING INFORMATION Hui-Sung Kao Associae Professor, Deparmen of Accouning, Feng Chia Universiy, Taiwan Tzu-Han Wei Graduae Suden, Deparmen of Accouning, Feng Chia Universiy, Taiwan ABSTRACT This paper invesigaes he relaionship beween he informaion asymmery, he ownership srucure, he pledge of direcors-supervisor, respecively, and he qualiy of accouning informaion under differen accouning sandards. By considering A and B sock marke of China, which apply China GAAP and IFRS, we discuss wheher IFRS can reduce negaive effecs of he informaion asymmery, he ownership srucure, he pledge of direcors, and furhermore promoe he qualiy of accouning informaion effecively. The findings provide oher counries will use IFRS as a reference. Firs, we find ha IFRS improves he predicive value and imeliness, and i can influences represenaional faihfulness significanly. Second, he informaion asymmery degrades he qualiy of accouning informaion. Neverheless, IFRS can improve he informaion asymmery bu promoe he qualiy of accouning informaion is nonsignificanly. Third, sae ownership, manager ownership, blockholder and direcors-supervisor ownership would affec he qualiy of accouning informaion. IFRS would resrain negaive effec of sae ownership, manager ownership, blockholder and direcors-supervisor ownership and could enhance predicive value and imeliness. Finally, he pledge of direcors-supervisor would reduce he qualiy of accouning informaion. However, IFRS can confine negaive effecs of he pledge of direcors-supervisor bu can promoe he qualiy of accouning informaion significanly. As a resul, adoping IFRS could enhance he qualiy of accouning informaion significanly. Neverheless, IFRS need o reduce informaion asymmery and use corporae governance mechanism o promoe he qualiy of accouning informaion. The findings of his paper can provide IFRS s insiuion and regulaor promoing and using IFRS sysem as a reference. Keywords: IFRS, Informaion asymmery, Corporae governance, Accouning informaion qualiy. 226

2 1. INTRODUCTION More and more counries will adop Inernaional Financial Reporing Sandards (IFRS) in he near fuure. However, here is no consisen empirical finding regarding wheher he adopion of IFRS produces informaion qualiy superior o oher accouning sandards. Some scholars argue ha here is no difference in he informaion qualiy provided by differen accouning sandards (Leuz and Verrechia, 2000; Leuz, 2003; Leuz e al., 2003). They do no hink he selecion of accouning sandards maer much so here is no need for a uniform se of accouning sandards. However, oher scholars believe ha he adopion of IFRS leads o beer informaion qualiy (Ashbaugh and Olsson, 2002; Chen e al., 2006). Therefore, his paper inends o explore wheher he adopion of IFRS has posiive effecs as well as o ensure wheher he adopion of IFRS produces superior accouning qualiy o comparing oher accouning sandards. The capial marke in China had a lae sar, wih he esablishmen of Shanghai Sock Exchange in lae 1990 and Shenzhen Sock Exchange in early During he pas 20 years, he capial marke in China has been aracing global capial via reformaions of sae-own companies. In he early days, he Chinese governmen divided he invesmen arges ino A shares and B shares, in order o arac foreign capial and conrol is currency. A-share lised companies prepare financial accouns pursuan o Chinese Accouning Sandards (CAS); whereas B-share lised companies prepare accouns according o IFRS. Companies lised as A shares and B shares a he same ime mus provide he financial saemens of wo accouning sandards. Leuz e al. (2003) compare he earnings managemen and earning qualiy in counries of he coninenal law sysem and find ha invesors proecion and law reinforcemen are he key deerminans o reporing qualiy. The unique capial marke in China presens an opporuniy o invesigae he variances of accouning informaion qualiy under he China GAAP and IFRS. In addiion o a comparison of he variances in accouning informaion qualiy under he China GAAP and IFRS, his paper also examines he influence of informaion asymmery and corporae governance on accouning informaion qualiy. Assuming an efficien marke and raional behaviour, he convenional heoriss of financial accouning argues ha full disclosure of financial informaion can reduce informaion asymmery, i.e. he advanages of insides and disadvanages of ousiders. IFRS is principle-based and encourages he use of foonoes o enhance he ransparency of financial saemens. Foonoes can be divided ino he basis of financial numbers, addiional informaion and managemen informaion from he perspecive of he managemen. I is expeced ha he adopion of IFRS can miigae negaive effecs of informaion asymmery and improve he qualiy of accouning informaion. Meanwhile, his paper explores wheher he adopion of IFRS and he full disclosure of informaion will improve or deeriorae he equiy resrucure and he pledge on shares by direcors and supervisors. This paper samples he companies wih boh A shares and B shares lised on Shanghai Sock Exchange or Shenzhen Sock Exchange and examines heir accouning informaion qualiy under he China GAAP sysem and he IFRS sysem. The purpose is o compare wheher he IFRS sysem yields accouning informaion qualiy superior o he China GAAP sysem. In addiion o he 227

3 measuremen of relevance, his paper also invesigaes he reliance of accouning informaion qualiy. To sum up, his paper ses ou o explore: (1) how he IFRS adopion affecs accouning informaion qualiy; (2) how informaion asymmery affecs accouning informaion qualiy under he IFRS sysem; (3) how he shareholder srucure affecs accouning informaion qualiy under he IFRS sysem; (4) how he pledge of shares by direcors and direcors affecs accouning informaion qualiy under he IFRS sysem. This empirical sudy finds ha: (1) The IFRS adopion enhances predicabiliy and imeliness, bu i does no have significan effecs on faihful presenaions; (2) Informaion asymmery reduces accouning informaion qualiy. The IFRS adopion miigaes such negaive effecs of informaion asymmery bu i has limied effecs on he improvemen of accouning informaion qualiy; (3) The ownership by he Sae, senior managers, blockholders, direcors and supervisors does lower he qualiy of accouning informaion. The IFRS adopion can undermine such negaive effecs, bu i only enhances predicabiliy and imeliness; finally, (4) he pledge of shares by direcors and supervisors has adverse effecs on accouning informaion qualiy. The IFRS adopion can reduce such negaive effecs bu i does no have significan effecs on he improvemen of accouning informaion qualiy. The remainder of his paper is organized as follows. Secion 2 provides a review of he lieraure. Secion 3 explains he research design including he daa sources, sample selecion, variable definiion and empirical model. The empirical resuls and analysis are presened in Secion 4. The addiional analyses are demonsraed in Secion5. Finally, he conclusions and suggesions are presened in Secion LITERATURE REVIEW 2.1. IFRS, China GAAP and Chinese Capial Marke In 1904, he firs meeing of Inernaional Federaion of Accounans (IFAC) was held in S. Louis and his convenion marks he beginning of he IFRS esablishmen. In 1972, he 10h IFAC meeing was convened in Sydney, wih he proposal o esablish he Inernaional Accouning Sandards Board (IASB). In June 1973, he IASB was founded in London, and his even was a waershed for he IFRS developmen. Since incepion, he IASB has been dedicaed o he synchronizaion of accouning principles by seing up accouning sandards and enhancing he comparabiliy of inernaional sandards. In 1993, he IASB proposed 40 core principles. In May 2000, Inernaional Organisaion of Securiies Commissions recified and recommended 30 principles of inernaional accouning sandards, marking a milesone for he synchronizaion of global financial reporing (Pan, 2004). IFRS is considered a "principles based" se of sandards; herefore, i akes a while o evaluae he appropriae accouning principles. The process requires exensive and professional judgemen. Meanwhile, as accouning reamens are deermined on he basis of he naure of ransacions, hey can more faihfully presen he ransacions. Also, IFRS pays aenion o noebooks as addiional disclosure o enhance he ransparency and comparabiliy of financial saemens. One of he major reforms IFRS aims for is o inroduce he concep of fair 228

4 values (in conras wih hisorical values), so ha financial saemens can beer presen he economic ruh and rue value of companies. Before 1979, China implemened a cenrally planned economy, wih is accouning sysem following he Sovie Union. I was in essence a highly regulaed and rigid sysem. The role of accounans was o assis governmen planning and conrol he producion quoas of sae-owned enerprises. Accounans had o deermine how much informaion should be provided in order o reflec he conflic beween invesors and managers. Informaion disclosed depended on social coss and social benefis. In 1985, he Wesern accouning pracices were inroduced, wih he esablishmen of Join Venure (JV) Accouning Sandards. In 1992, he Chinese governmen published Pilo Company Accouning Sysem and Company Reporing Principles o se up a concepual framework. In 1993, Deloie ToucheTohmasus Inernaional was commission o draf No. 30 Accouning Principle, as he firs sep o link he China GAAP o IFRS. In 1998, he pracice of lower of cos or marke price was inroduced, and i requires he provision of allowance for bad debs, invenory losses, shor-erm and long-erm invesmen losses. Meanwhile, he accouning reforms in 1998 were accompanied wih he release of seven accouning principles, e.g. he principle for he preparaion of cash flow saemens, disclosure of major evens afer he compilaion of he balance shee, deb resrucuring and liquidaion principles, income principles, invesmen principles, conracual principles, accouning policies and changes of accouning esimaes. In 2001, he lower of cos or marke price pracice was exended o non-cash asses, e.g. fixed asses, inangible asses, impairmen ess of consrucion-in-progress and enrused loans, in order o properly link wih IFRS and miigae he concerns of he inernaional communiy over he opaqueness of marke informaion and overesimaes of earnings in he Chinese marke. In 2005, China Accouning Sandards Commiee iniiaed anoher round of reforms, by modifying he exising 17 accouning principles and adding 21 new ones. In 2006, he Chinese governmen sared o bring is accouning principles closer o IFRS and published new accouning and audiing principles. In 2007, he new sysem ook effec wih lised companies. The new China GAAP is known for (1) is synchronizaion wih IFRS and he consideraion of local characerisics; (2) quaniaive requiremens for fair values (including he impac of ime value on paymen exensions). Alhough i is in sync wih IFRS, he new China GAAP has is own pariculariies. The Chinese pracice is mainained in he reversal of asse impairmens, he disclosure of affiliaed pary ransacions and he accouning reamens of cerain governmen subsidies (Deloie, 2007). To avoid he impacs of foreign capial inroducion on he fledgling sock markes in Shanghai and Shenzhen, he Chinese governmen sipulaes ha A-share lised companies should prepare financial saemens pursuan o China GAAP whils B-share lised companies should prepare wo ses of financial saemens pursuan o China GAAP and IFRS. Originally, A shares were raded by Chinese naionals; whereas B shares were raded by foreign invesors. However, he resricions were relaxed by allowing Chinese naionals o rade B shares in 2001 and qualified foreign 229

5 insiuional invesors (QFIIs) o rade A shares in This paper samples he A-share and B- share lised companies on Shanghai Sock Exchange and Shenzhen Sock Exchange Financial Accouning Sandards and Accouning Informaion Qualiy Leuz and Verrechia (2000) sugges ha mos companies in Germany adop IFRS because i improves he qualiy and ransparency of financial reporing. Dumonier and Raffourier (1998) indicae ha he early adopion of IFRS by Swizerland, as a resul of poliical coss and exernal marke pressures, mees wih inernaional requiremens and reduces capial incenives. El-Gazzar e al. (1999) sample Swiss companies and find ha he companies under he IFRS sysem enjoy more inernaional rades and lower leverage (measured by deb/equiy raio). Sree and Gray (2002) and Ball e al. (2003) argue ha axaions limis he adopion and he validiy of IFRS. Barh e al. (2006) examine companies in differen regions and compares he adopion of IFRS and US GAAP. Their resul shows ha he US GAAP produces beer accouning qualiy. However, oher scholars argue ha he managemen is given leeway o make judgmens under he principle-based IFRS. Therefore, hey know how o ruhfully reflec operaing performances wih appropriae accouning reamens and provide beer accouning qualiy. The Chinese regulaory framework is no robus enough; herefore, he Chinese governmen inends o synchronize wih IFRS. However, here are inconsisencies beween China GAAP and IFRS in erms of accouning policies and reamens of cerain economic aciviies. To arac foreign capial, he Chinese governmen sipulaes he adopion of boh accouning sysems. Bao and Chow (1999) examine wheher B-share lised companies on he IFRS sysem disclose more accouning informaion han he companies on he China GAAP sysem. The resuls show ha he prices esimaed wih he IFRS model are more relevan o he acual prices of B shares. Therefore, IFRS has exra explanaory power compared o China GAAP. Eccher and Healy (2000) look ino he wo accouning sysems adoped in China and explore which sysem can provide more useful informaion of economic aciviies. Their research delves ino he following issues: (1) Assuming informaion is useful for invesors o forecas fuure cash flows, which accouning sandard is more relevan o he predicions of fuure cash flows; (2) Assuming share prices are linked wih company performances during he informaion period, how he wo accouning sysems are relevan o reurns o share prices. Their sudy suggess ha China GAAP and IFRS have he same explanaory power over fuure cash flows. However, China GAAP is more relevan han IFRS in erms of reurns o share prices. Heibaollah and Haiyan (2004) compare A-share lised companies under he China GAAP sysem and B-share lised companies under he IFRS sysem and invesigae he variances in he relevance of accouning informaion beween hese wo groups. The resuls sugges ha he relevance of accouning informaion of B-share lised companies is beer han ha of A-share lised companies. Chen e al. (2006) examine he informaion qualiy of differen accouning sandards by sampling companies wih boh A-shares and B-shares lised in Shenzhen or Shanghai in 2001~2003. The sudy finds ha here is a sronger relevance of earnings o share price reurns 230

6 among B-share lised companies under he IFRS sysem. Meanwhile, he B-share marke responds o unexpeced earnings informaion calculaed by IFRS and herefore, IFRS yields beer informaion qualiy. Wu (2006) examines he companies wih boh A-shares and B-shares lised in Shenzhen or Shanghai and explores he relevance of accouning informaion in he China GAAP- IFRS dual sysem in 1994~2005. The resuls sugges ha earnings informaion disclosed under he IFRS sysem is more relevan compared o ha under he China GAAP sysem. Barov e al. (2002) also indicae ha earnings repored under he IFRS are more relevan. According o No. 1 of Saemen of Financial Accouning Sandards, relevance consiss of predicabiliy, feedback and imeliness. Therefore, his paper anicipaes IFRS produces accouning qualiy of beer relevance. Meanwhile, his paper also expecs IFRS o make more faihful presenaions because i is principle-based and accouning reamens are based on he naure of ransacion. Therefore, his paper esablishes H1: H1:The IFRS adopion yields beer accouning informaion qualiy. H1-1:The IFRS adopion yields beer predicabiliy value. H1-2:The IFRS adopion yields beer feedback value. H1-3:The IFRS adopion yields beer imeliness. H1-4:The IFRS adopion yields beer faihful presenaion IFRS, Informaion Asymmery and Accouning Informaion Qualiy Informaion asymmery persiss in capial markes. To proec he informaion disadvanaged, here are many accouning sandards and requiremens in place. For example, he laws sipulae professional conrol over he qualiy of accouning informaion. The Financial Supervisory Commission in Taiwan demands public companies o disclose cerain informaion. There are also accouning principles released by IASB and FASB. Leuz (2003) suggess ha he main purpose of he securiies marke and accouning regulaors is o miigae informaion asymmery. La Pora e al. (2000) indicae ha legal srucures are an imporan facor ha influences financial markes, capial marke srucures, dividend policies and privae ownership of companies. Ball e al. (2000) emphasize ha financial reporing qualiy is subjec o he influence of juridical sysems. They analyze seven counries of varying monioring mechanisms and he resuls sugges ha here are soluions o differing degrees of informaion asymmery. A comparison beween he counries of unwrien laws and cusomary laws finds ha he recogniion of losses is on he realime basis in he counries of cusomary laws. Leuz e al. (2003) compare he earnings managemen and earnings qualiy in 31 counies. Their sudy finds ha he key deerminan o reporing qualiy is he laws o proec invesors and he abiliy o reinforce he laws. Ball e al. (2003) also second his view. Chen e al. (2006) examine he companies wih boh A shares and B shares lised, i.e. he companies reporing under he China GAAP sysem and he IFRS sysem a he same ime. Their sudy indicaes ha IFRS yields beer informaion qualiy. Sami and Zhou (2004) also suppor ha B-share accouning informaion is more relevan han A-share accouning informaion. Leuz and 231

7 Verrechia (2000) believe ha IFRS resuls in beer ransparency of financial reporing and reduce informaion asymmery, uncerainies and esimaion risks. I can furher lower capial cos and enhance marke liquidiy. Trueman and Timan (1988) and Dye (1988) indicae ha he presumpion of earnings managemen is informaion asymmery. Richardson (2000) analyzes he variances beween analyss forecass and he spread of share ransacions as an indicaor o informaion asymmery. The empirical resuls show ha here is a posiive correlaion beween informaion asymmery and earnings managemen. The wors he informaion asymmery, he more likely earnings managemen is. Lieraure suggess ha he level of ransparency affecs he probabiliy of financial repor users spoing he earnings manipulaions by he managemen (Hirs and Hopkins, 1998; Maines and McDaniel, 2000). Lobo and Zhou (2001) and Hunon e al. (2006) hink ha he ransparency of financial reporing and qualiy of informaion disclosure can lessen he willingness and inenion o earnings managemen. To sum up, he beer he ransparency of financial reporing and he qualiy of informaion disclosure (i.e. low degrees of informaion asymmery), he lower he probabiliy of earnings managemen and he beer he qualiy of accouning informaion. China inends o learn from he exensive experience of he IFRS sysem in is adjusmen of he China GAAP sysem. Many of is accouning principles are synchronized wih IFRS. However, here are sill many variances (Pan, 2004; Deloie, 2006). As IFRS is esablished on he basis of accouning principles in developed counries, i is now he adoped or soon-o-be-adoped sandards in mos counries around he world. To China, who is sill developing is accouning sandards, he IFRS adopion can enhance he ransparency of financial reporing, lower he level of informaion asymmery and improve he qualiy of accouning informaion. Therefore, his paper esablishes H2: H2 : The IFRS adopion can undermine he negaive effecs of informaion asymmery on accouning informaion qualiy. H2-1:The IFRS adopion can undermine he negaive effecs of informaion asymmery on predicabiliy value. H2-2:The IFRS adopion can undermine he negaive effecs of informaion asymmery on feedback value. H2-3:The IFRS adopion can undermine he negaive effecs of informaion asymmery on imeliness. H2-4:The IFRS adopion can undermine he negaive effecs of informaion asymmery on faihful presenaion IFRS, Corporae Governance and Accouning Informaion Qualiy China was originally a planned economy. There has been a reformaion o privaize saeowned enerprises since he 1980s. However, he Chinese governmen sill has 1/3 sake of lised companies. Ye and Ma (1999) argue ha sae shareholders play a passive role in corporae governance due o he double-whamming agency problems associaed wih he sae ownership and 232

8 he company managemen, as well as wih he public-service funcion and he sae ownership. A high concenraion of ownership (by he Chinese governmen) is prone o agency problems (DeFond e al., 2000; Huang, 2007), and conflics of ineres beween conrolling shareholders and reail invesors. Conrolling shareholders decide on he reporing of accouning informaion in he conex of heir own benefis. As a resul, accouning informaion does no ruhfully reflec he ruh of economic ransacions (Fan and Wong, 2002). Wang e al. (2007) explore he relaionship beween sae ownership (i.e. he governmen s inervenion) and accouning informaion qualiy. They find ha companies of high sae-ownership end o exercise poliical pressures o influence audiing qualiy and his reduces accouning informaion qualiy. Liu and Tu (2003) sugges ha when he Chinese governmen is he larges shareholder of a lised company, he likelihood of frauds in financial reporing is high. This paper anicipaes ha he IFRS adopion will enhance he ransparency and informaion disclosure and hence, miigae he negaive effecs of sae ownership and improve accouning informaion qualiy. Therefore, his paper esablishes H3: H3:The IFRS adopion can undermine he negaive effecs of Sae-ownership on accouning informaion qualiy. H3-1:The IFRS adopion can undermine he negaive effecs of Sae-ownership on predicabiliy value. H3-2:The IFRS adopion can undermine he negaive effecs of Sae-ownership on feedback value. H3-3:The IFRS adopion can undermine he negaive effecs of Sae-ownership on imeliness. H3-4:The IFRS adopion can undermine he negaive effecs of Sae-ownership on faihful presenaion. According o Jensen and Ruback (1983), a low percenage of holdings by he managemen preven he deviaion of he managemen from he goal of maximizing he company s ineress, as he managemen wans o secure heir jobs. However, as heir job securiy is proeced wih a high percenage of shareholdings, he managemen may walk away from he bes ineres of he shareholders. Yermack (1997) suggess ha a high percenage of holdings by he managemen mean heir personal wealh is subjec o he share prices. In such insances, hey are likely o manage earnings by manipulaing financial reporing in order o enhance share prices for heir personal gains. Namely, he managemen s ownership is posiively correlaed earnings managemen, and negaively correlaed wih accouning qualiy. Aboody and Kasznik (2000) believe ha before he share subscripion scheme for employees, he managemen will manipulae earnings downward o reduce heir subscripion coss. Therefore, here is a posiive correlaion beween he managemen s holdings and earnings managemen. Yeo e al. (2002) argue ha when he managemen s sake is below 25%, he managemen is less likely o manipulae earnings wih discreionary accruals. However, if he managemen s sake is above 25%, he managemen increases heir manipulaion of discreionary accruals. Hence, his paper anicipaes ha he IFRS adopion can enhance he qualiy of accouning informaion and enhance he ransparency of 233

9 financial reporing. I can miigae he behavior of earnings managemen or manipulaion by he managemen. Therefore, his paper esablishes H4: H4:The IFRS adopion can undermine he negaive effecs of ownership by senior managers on accouning informaion qualiy. H4-1:The IFRS adopion can undermine he negaive effecs of ownership by senior managers on predicabiliy value. H4-2:The IFRS adopion can undermine he negaive effecs of ownership by senior managers on feedback value. H4-3:The IFRS adopion can undermine he negaive effecs of ownership by senior managers on imeliness. H4-4:The IFRS adopion can undermine he negaive effecs of ownership by senior managers on faihful presenaion. According o Convergence-of-ineres Hypohesis, he companies whose direcors and supervisors have a high sake may aemp o window-dress earnings in order o lower capial coss or o mee oher purposes. As a resul, heir accouning qualiy will be affeced (Oswald and Jahera, 1991). Meanwhile, Enrenchmen Hypohesis suggess ha in he companies whose direcors and supervisors have a high sake bu do no have o deal wih couner-balance mechanisms, he direcors and supervisors are likely o embezzle company asses. They may aemp o cover up heir sealing wih earnings managemen. As a resul, he qualiy of accouning informaion will deeriorae (Pound, 1988). Kuo (1996) argues ha he higher he ownership by direcors and supervisors, he wors he earnings managemen and he earnings qualiy. Beasley (1996) samples he companies reporing frauds in financial saemens and finds ha here is a posiive correlaion beween he likelihood of unscrupulous financial reporing and he ownership by direcors and supervisors. In oher words, he higher he ownership by direcors, he more likely he frauds in financial reporing and he lower he qualiy of accouning informaion. This paper anicipaes ha he IFRS adopion will enhance he ransparency and disclosure of a company and diminish he negaive effecs of he ownership by major shareholders, direcors and supervisors. As a resul, accouning informaion qualiy will improve. Therefore, his paper esablishes H5: H5:The IFRS adopion can undermine he negaive effecs of ownership by blockholders (major shareholders), direcors and supervisors on accouning informaion qualiy. H5-1:The IFRS adopion can undermine he negaive effecs of ownership by blockholders (major shareholders), direcors and supervisors on predicabiliy value. H5-2:The IFRS adopion can undermine he negaive effecs of ownership by blockholders (major shareholders), direcors and supervisors on feedback value. H5-3:The IFRS adopion can undermine he negaive effecs of ownership by blockholders (major shareholders), direcors and supervisors on imeliness. H5-4:The IFRS adopion can undermine he negaive effecs of ownership by blockholders (major shareholders), direcors and supervisors on faihful presenaion. 234

10 Pledge on shares can access capial bu mainain conrol; hence, i is a common pracice for direcors and supervisors o borrow money by using company shares as collaeral. This leverage by credi expansion enhances he correlaion beween he personal wealh of major shareholders, direcors and supervisors and he company s share prices (Shen and Huang, 2001). Once share prices drop, conrolling shareholders will face he pressure of providing addiional collaerals. Therefore, hey have incenives o manage earnings. Kao (2002) invesigaes accouning informaion in he conex of agency problems associaed wih pledge on shares owned by direcors and supervisors and suggess ha his pracice worsens he agency problems beween conrolling shareholders and exernal shareholders. I deepens he influence of direcors and supervisors on he managemen and undermines he predicabiliy of earnings of he curren period o fuure performances. Chang e al. (2007) sample lised companies in Taiwan and find ha earnings managemen becomes more pronounced along wih an increase of pledge on shares owned by direcors and supervisors. However, his paper anicipaes ha IFRS can enhance accouning informaion qualiy because i enhances he ransparency of financial reporing and miigaes he negaive effecs of pledge on shares owned by supervisors and direcors. Hence, his paper esablishes H6: H6:The IFRS adopion can undermine he negaive effecs of pledge of shares owned by direcors and supervisors on accouning informaion qualiy. H6-1:The IFRS adopion can undermine he negaive effecs of pledge of shares owned by direcors and supervisors on predicabiliy value. H6-2: The IFRS adopion can undermine he negaive effecs of pledge of shares owned by direcors and supervisors on feedback value. H6-3:The IFRS adopion can undermine he negaive effecs of pledge of shares owned by direcors and supervisors on imeliness. H6-4:The IFRS adopion can undermine he negaive effecs of pledge of shares owned by direcors and supervisors on faihful presenaion. 3. RESEARCH DESIGN 3.1. Daa Sources and Sample Selecion This paper examine he 42 companies issuing boh A shares and B shares on Shenzhen Sock Exchange and 44 companies issuing boh shares on Shanghai Sock Exchange. As each of hese company provides observaions of A shares and B shares, his paper samples a oal of 172 company daa in 1999~2009. The purpose is o examine he relaionship beween accouning informaion qualiy and informaion asymmery, as well as wih ownership srucure, pledge on shares by direcors and supervisors under differen accouning sysems. A oal of 1,892 observaion values are sampled. However, he research period is 2002~2009 due o he resricion of daa acquisiions for he purpose of faihful presenaion. In his regard, a oal of 1,376 observaion values are esablished. The empirical daa is sourced from China Finance and economic daabase of TEJ, a daabase offered by Taiwan Economic Journal. 235

11 3.2. Empirical Model This paper aims o explore he effecs of informaion asymmery, ownership srucure and pledge on shares by direcors and supervisors on he qualiy of accouning informaion under differen financial reporing sandards. According o lieraure, he facors ha influence of accouning informaion qualiy include company sizes (Aiase, 1985; Chaney and Jeer, 1992), deb raios (Dhaliwal e al., 1982), book/marke muliple value (Felham and Ohlson, 1995), company losses (Hsu, 1994), sales growh (Myers, 1977), payou raio (Kasanen e al., 1996), audiing by Big 4 (Becker e al., 1998; Francis and Krishan, 1999; Francis e al., 1999), ownership by insiuional invesors (Rajgopal e al., 1999) and reurn on equiy (Freeman e al., 1982). This paper aemps o ensure he robusness of model se-up and selecs conrol variables ha influence he qualiy of accouning informaion wih sepwise regression. On he basis of he empirical resuls of sepwise regression, his paper incorporaes a oal of eigh variables as conrol variables. They are company sizes, deb raios, book/marke muliples, company losses, payou raios, audiing by Big 4, ownership by insiuional invesors and reurn on equiy Financial Accouning Sandards and Accouning Informaion Qualiy Firs Hypohesis demonsraed he accouning informaion qualiy beween China GAAP and IFRS in disinc financial accouning sandards. The empirical model is as follows: AIQ 0 1IFRS 2SIZE 3LEVi, 4ROEi, 5BM 6 LOSS 7DIVP 8B4NB4 9INSTi, (1) IFRS, Informaion Asymmery and Accouning Informaion Qualiy The second hypohesis demonsraed if using IFRS could reduce negaive effec of informaion asymmery o furhermore enhance he accouning informaion qualiy. The empirical model is as follows: AIQ 0 1IFRS 2IAi, 3IFRS IAi, 4SIZE 5LEVi, 6 ROE (2) 7BM 8LOSSi, 9DIVP 10B NB4 4 INST IFRS, Corporae Governance and Accouning Informaion Qualiy This paper examines ownership srucures and pledge on shares by direcors and supervisors as wo corporae governance facors. Ownership srucure issue is divided ino hree componens, i.e. sae ownership, ownership by senior managers and ownership by major shareholders, direcors and supervisors. This is o validae H3, H4 and H5 by examining wheher he IFRS adopion can 236

12 diminish he negaive effecs of sae ownership, ownership by senior managers and ownership by major shareholders, direcors and supervisors, so as o improve he qualiy of accouning informaion. H6 assumes he IFRS adopion can miigae he negaive effecs of pledge on shares by direcors and supervisors and hence enhance he qualiy of accouning informaion. The empirical model is as follows: A. Ownership srucure AIQ 0 1IFRS 2GOVi, 3MANAGER 4B & Di, 5IFRS GOV 6 IFRS MANAGER 7IFRS B & Di, 8SIZE 9 LEV 10ROEi, 11BM 12LOSSi, 13DIVPi, 14B NB4 (3) B. The pledge of direcors and supervisors 4 INST 15 AIQ 0 1IFRS 2PLEDGEi, 3IFRS PLEDGEi, 4SIZE 5 LEV 4 INST (4) 6ROEi, 7BM 8LOSSi, 9DIVP 10B NB4 The all variables of above models idenify as nex secion: Variable Measuremens Dependen variables: Accouning Informaion Qualiy (AIQ) This paper evaluaes he qualiy of accouning informaion (AIQ) in he conex of FASB srucure by examining predicabiliy, feedback, imeliness and faihful presenaion (as dependen variables) and summarizes in able 1. The measuremen of dependen variables is based on he valuaion of financial informaion qualiy developed by Liao (2004). Table-1. The definiion of accouning informaion qualiy Variables Definiion Measuremens Predicabiliyvalue (PV) Feedback value( FV) Operaing ne income / Ne income ΔP%/ΔEPS% Timeliness(TM) β 1+ R 2 ( Toalize he One-off ne incomes= gains (losses) from disposal of invesmens + gains (losses) from disposal of fixed asses + gains (losses) from foreign currency conversions + oher gains (losses) + gains (losses) of disconinued operaions + exraordinary gains (losses) + accumulaed effecs of accouning principle changes. Ne incomes= Recurring ne incomes + one-off ne incomes. ΔP%= Percenage of changes o share prices from April 1 of he year o March 31 of he year +1. ΔEPS%= Percenage of changes o basic EPS of years and -1 (excluding exraordinary iems). EARN = α 0 +α 1 NEG +β1ret +β 2 NEG *RET +ε EARN : Ne incomes before exraordinary iems and disconinued 237

13 Faihfulpresenaion (RF) percenile of β 1 and R 2 ) R 2 operaions of he year, deflaed by he equiy marke value a he beginning of he year. NEG : dummy variable, 1 if RET <0, 0 if oherwise. RET : annualized reurns on he company s shares during year (average of he 12 monhs reurns). ε : residual iem. R 2 : coefficien of deerminaion. CFO = α 0 + α 1 EARN -1 + DA + ε CFO : operaing cash flows per share of he year. EARN -1 : Earnings/No. of shares ousanding for he year -1. DA : Discreionary accruals are measured by Modified Jones Model (Dechow e al., 1995). R 2 : coefficien of deerminaion Independen variables: IFRS, Informaion Asymmery and Corporae Governance The independen variables of empirical model are as follows. IFRS: The adopion of IFRS is defined as 1; he non-adopion of IFRS is defined as 0. IA: Informaion asymmery is measured wih (he lowes ask quoe he highes bid quoe)/(he lowes ask quoe + he highes ask quoe)/2 as of April 30, i.e. he deadline for he publicaion of annual repors. The highes bid quoe is based on he closing prices during he five days before April 30; whereas he lowes ask quoe is based on he closing prices during he five days afer April 30. GOV: Sae ownership is measured wih (No. of shares held by he governmen + No. of shares held by sae-owned insiuional invesors)/ (No. of shares ousanding). MANAGER: Ownership by senior managers is measured wih (No. of shares held by senior managers)/ (No. of shares ousanding). B&D: Ownership by major shareholders, direcors and supervisors is measured wih (No. of shares held by major shareholders, direcors and supervisors)/ (No. of shares ousanding). PLEDGE: The percenage of pledged shares is measured by (No. of pledged shares by major shareholders, direcors and supervisors)/ (No. of shares ousanding) Conrol Variables The conrol variables of empirical model are as follows. SIZE: Company size is measured wih he naural logarihm of oal asses a he end of he period. LEV: Deb raio is measured wih oal liabiliies a he end of he period divided by oal asses a he end of he period. ROE: Reurn on equiy is measured wih ne income divided wih shareholders equiy. BM: Book/marke muliple is measured wih book value of shareholders equiy a he end of he period divided by he marke value of one a he end of he period. LOSS :If he ne income during he period is smaller han 0, he dummy variable for company losses is defined as 1; oherwise, i is defined as 0. DIVP: Payou raio is measured wih he naural logarihm of (cash dividends + sock dividends)/eps. B4NB4: The dummy variable for audiing by Big 4 is defined as 1; oherwise i is defined as 0. INST: Ownership by insiuional invesors is calculaed as he percenage of holdings by domesic 238

14 founding insiuional invesors plus he holdings by insiuional invesors in he second marke plus he holdings by foreign founding insiuional invesors. 4. EMPIRICAL RESULT 4.1. Descripive Saisics and Analysis Table 2 summarizes he descripive saisics of differen variables. The mean of Informaion Asymmery (IA) of sampled companies is 0.016, and he median is This suggess ha he minimum ask prices are higher han he maximum bid prices, i.e. a sligh discrepancy beween buyers and sellers in erms of equiy valuaions. In erms of corporae governance, he mean of he percenage of governmen-owned shares (GOC) is 30.6%and he median is 31.5%. Despie effors since he 1980s o privaize sae-owned enerprises, he Chinese governmen sill holds a high sake in lised companies. The mean and he median of he shareholding percenage by senior managers (MANAGER) is 0.1% and 0%, a very low level. The mean and he median of he shareholding percenage by major shareholders, direcors and supervisors (B&D) is 8% and 2.4%, respecively, no a high level eiher. These numbers sugges ha he Chinese governmen has a large sake in domesic lised companies. Therefore, i is no a common phenomenon in China for direcors and supervisors o borrow money by seing up a pledge on heir shares, as evidenced by he mean of 0.8% and he median of 0% in erms of he pledged shares owned by insiders (PLEDGE) Pearson Correlaion Coefficiens Table 3 summarizes he Pearson correlaion coefficiens. Single variable analysis suggess ha here is a significan and posiive correlaion wih IFRS, predicabiliy (PV) and imeliness (TM). These numbers indicae ha he IFRS adopion can enhance predicabiliy and imeliness. There is a significan and negaive correlaion wih informaion asymmery (IA), feedback (FV), imeliness (TM) and faihful presenaion (RF). The wors he informaion asymmery, he lower he qualiy of accouning informaion is. Sae ownership (GOV) is significanly and negaively correlaed wih predicabiliy (PV) and faihful presenaion (RF). This implies ha sae ownership reduces he predicabiliy and faihful presenaion of accouning informaion. However, here is a significan and posiive correlaion wih imeliness (TM). Ownership by senior managers (MANAGER) is significanly and negaively correlaed wih feedback (FV) and imeliness (TM). The higher he ownership by senior managers, he wors he qualiy of accouning informaion is. Ownership by major shareholders, direcors and supervisors (B&D) is significanly and negaively correlaed wih imeliness (TM) and faihful presenaion. A high percenage of holdings by major shareholders, direcors and supervisors reduce he imeliness and faihful presenaion of accouning informaion. Pledge on shares (PLEDGE) is significanly and negaively correlaed wih predicabiliy (PV), imeliness (TM) and faihful presenaion (RF). This means he higher he percenage of pledged shares by direcors and supervisors, he wors he qualiy of accouning informaion. 239

15 4.3 Empirical Resuls and Analysis IFRS and Accouning Informaion Qualiy Table 4 shows how he IFRS adopion influences he qualiy of accouning informaion. According o Table 3, adjused R 2 in respecive models is 18.3%, 0.7%, 0.7% and 3.5%, respecively. Predicabiliy (PV) repors he sronges explanaory power on independen variables. In erms of model f he F values are (P-value=0), (P-value=0.017), (Pvalue=0.014) and (P-value=0), respecively, indicaing significan levels of model fi. Table-2. Descripive Saisics (N=1,892) Variables a Mean Median Min Max Sd. deviaion IFRS IA GOV MANAGER B&D PLEDGE PV FV TM RF SIZE LEV ROE BM LOSS DIVP B4NB INST a Definiion of variables: IFRS: use dummy variable. The adopion of IFRS is defined as 1; he non-adopion of IFRS is defined as 0. IA:Informaion Asymmery measured wih (he lowes ask quoe he highes bid quoe)/(he lowes ask quoe + he highes ask quoe)/2 as of April 30. GOV: Sae ownership is measured wih (No. of shares held by he governmen + No. of shares held by sae-owned insiuional invesors)/ (No. of shares ousanding). MANAGER: Ownership by senior managers is measured wih (No. of shares held by senior managers)/ (No. of shares ousanding). B&D: Ownership by major shareholders, direcors and supervisors is measured wih (No. of shares held by major shareholders, direcors and supervisors)/ (No. of shares ousanding). PLEDGE: The percenage of pledged shares is measured by (No. of pledged shares by major shareholders, direcors and supervisors)/ (No. of shares ousanding). PV: Predicabiliy value measure wih Operaing ne income / Ne income. FV: Feedback value measure wih ΔP%/ΔEPS%. TM: Timeliness measure wih β 1 + R 2 (Toalize he percenile of β 1 and R 2 ). RF: Faihful presenaion measure wih R 2 (measured mehod as able 1). SIZE: ake he naural logarihm of oal asses a he end of he period. LEV: Toal liabiliies a he end of he period divided by oal asses a he end of he period. ROE: Ne income divided wih shareholders equiy. BM: measure wih book value of shareholders equiy a he end of he period divided by he marke value of one a he end of he period. LOSS: If he ne income during he period is smaller han 0, he dummy variable for company losses is defined as 1; oherwise, i is defined as 0. DIVP: 240

16 ake he naural logarihm of (cash dividends + sock dividends)/eps. B4NB4: The dummy variable for audiing by Big 4 is defined as 1; oherwise i is defined as 0. INST: Ownership by insiuional invesors is calculaed as he percenage of holdings by domesic founding insiuional invesors plus he holdings by insiuional invesors in he second marke plus he holdings by foreign founding insiuional invesors. Table-3. Pearson correlaion coefficiens (N=1,892) b a Definiion of variables: IFRS: use dummy variable. The adopion of IFRS is defined as 1; he non-adopion of IFRS is defined as 0. IA: Informaion Asymmery measured wih (he lowes ask quoe he highes bid quoe)/(he lowes ask quoe + he highes ask quoe)/2 as of April 30. GOV: Sae ownership is measured wih (No. of shares held by he governmen + No. of shares held by sae-owned insiuional invesors)/ (No. of shares ousanding). MANAGER: Ownership by senior managers is measured wih (No. of shares held by senior managers)/ (No. of shares ousanding). B&D: Ownership by major shareholders, direcors and supervisors is measured wih (No. of shares held by major shareholders, direcors and supervisors)/ (No. of shares ousanding). PLEDGE: The percenage of pledged shares is measured by (No. of pledged shares by major shareholders, direcors and supervisors)/ (No. of shares ousanding). PV: Predicabiliy value measure wih Operaing ne income / Ne income. FV: Feedback value measure wih ΔP%/ΔEPS%. TM: Timeliness measure wih β 1 + R 2 (Toalize he percenile of β 1 and R 2 ). RF: Faihful presenaion measure wih R 2 (measured mehod as able 1). SIZE: ake he naural logarihm of oal asses a he end of he period. LEV: Deb raio a he end of he period. ROE: Ne income divided wih shareholders equiy. BM: measure wih book value of shareholders equiy a he end of he period divided by he marke value of one a he end of he period. LOSS: If he ne income during he period is smaller han 0, he dummy variable for company losses is defined as 1; oherwise, i is defined as 0. DIVP: ake he naural logarihm of (cash dividends + sock dividends)/eps. B4NB4: The dummy variable for audiing by Big 4 is defined as 1; oherwise 0. INST: Ownership by insiuional invesors is calculaed as he percenage of holdings 241

17 by domesic founding insiuional invesors plus he holdings by insiuional invesors in he second marke plus he holdings by foreign founding insiuional invesors.b *** (**) (*) Indicaes significance a 1% (5%) (10%). Table 4 suggess ha predicabiliy (PV) and IFRS are significanly and posiively correlaed. Compared o he China GAAP sysem, he IFRS sysem can improve he predicabiliy of accouning informaion. Therefore, H1-1 is suppored. However, here is a negaive bu no significan correlaion beween feedback (FV) and IFRS. This is probably because Chinese invesors are accusomed o he use of he financial repors in compliance wih China GAAP for heir invesmen decisions. Eccher and Healy (2000) poin ou ha he China GAAP sysem is more relevan o share prices compared o he IFRS sysem. Therefore, H1-2 is no suppored. However, he resul is inline wih Eccher and Healy (2000). IFRS and imeliness are significanly and posiively correlaed. Compared o he China GAAP sysem, he IFRS sysem can beer improve he imeliness of accouning informaion. Therefore, H1-3 is suppored. Finally, IFRS and faihful presenaion (RF) is posiively bu no significanly correlaed. This shows ha he IFRS adopion improves he faihful presenaion of accouning informaion bu he benefi is limied. Therefore, H1-4 is no suppored. Perhaps axaions resriced he effeciveness of he IFRS adopion (Ball e al., 2003). In erms of conrol variables, company sizes (SIZE) is significanly and posiively correlaed wih predicabiliy (PV) and faihful presenaion (RF). The larger company have beer predicabiliy and faihful presenaion of is accouning informaion. There is a negaive correlaion wih deb raios (LEV). The higher he leverage, he more moives he managemen has o manipulae he selecion of accouning mehods o enhance earnings. As a resul, he accouning informaion qualiy deerioraes. There is a significan and negaive correlaion beween reurn on equiy (ROE) and predicabiliy (PV). Tha mean he lower ROE, he beer predicabiliy. There are also a negaive correlaion wih book/marke muliples (BM). The lower book/marke muliple, he more conservaive he accouning policies have he beer accouning informaion qualiy. There is a significan and posiive correlaion beween company losses (LOSS) and predicabiliy (PV). Namely, when he ne income during he period urns ou o be negaive, i can serve as a predicaor o he company s fuure. Payou raios (DIVP) are significanly and posiively correlaed wih predicabiliy (PV) and faihful presenaion (RF). The low payou raio could mean he reenion of earnings o fund capial. As a resul, he accouning informaion provides beer predicabiliy and faihful presenaion. There is a posiive correlaion wih he audiing by Big 4 (B4NB4) and he ownership by insiuional invesors (INST). This is inline wih he expecaions of his paper: boh facors can enhance he qualiy of accouning informaion. 242

18 Table-4. Regressions of Accouning Informaion Qualiy on IFRS and conrols (N=1,892) b Variables a Predic sign PV FV TM RF Inercep ** *** (0.034) (0.164) (0.000) (0.432) IFRS *** *** (0.000) (0.198) (0.005) (0.313) SIZE? 0.038*** *** (0.000) (0.164) (0.631) (0.002) LEV * ** (0.122) (0.454) (0.093) (0.012) ROE? ** (0.024) (0.945) (0.394) (0.621) BM ** (0.025) (0.205) (0.206) (0.128) LOSS? 0.078*** (0.000) (0.402) (0.280) (0.701) DIVP? *** *** (0.000) (0.467) (0.270) (0.000) B4NB *** 1.053*** (0.000) (0.003) (0.334) (0.490) INST *** *** 0.181*** (0.000) (0.215) (0.004) (0.000) Adjused R F-saisic Prob(F-saisic) a Definiion of variables: PV: Predicabiliy value measure wih Operaing ne income / Ne income. FV: Feedback value measure wih ΔP%/ΔEPS%. TM: Timeliness measure wih β 1 + R 2 (Toalize he percenile of β 1 and R 2 ). RF: Faihful presenaion measure wih R 2 (measured mehod as able 1). IFRS: use dummy variable. The adopion of IFRS is defined as 1; he non-adopion of IFRS is defined as 0. SIZE: ake he naural logarihm of oal asses a he end of he period. LEV: Deb raio a he end of he period. ROE: Ne income divided wih shareholders equiy. BM: measure wih book value of shareholders equiy a he end of he period divided by he marke value of one a he end of he period. LOSS: If he ne income during he period is smaller han 0, he dummy variable for company losses is defined as 1; oherwise, i is defined as 0. DIVP: ake he naural logarihm of (cash dividends + sock dividends)/eps. B4NB4: The dummy variable for audiing by Big 4 is defined as 1; oherwise 0. INST: Ownership by insiuional invesors is calculaed as he percenage of holdings by domesic founding insiuional invesors plus he holdings by insiuional invesors in he second marke plus he holdings by foreign founding insiuional invesors. b The P-value are repored in parenheses below coefficien esimaes. *** (**) (*) Indicaes significance a 1% (5%) (10%). 243

19 IFRS, Informaion Asymmery and Accouning Informaion Qualiy Table 5 indicaes he effecs of informaion asymmery on accouning informaion qualiy under he IFRS sysem. The adjused R 2 in respecive models is 18.7%, 0.9%, 0.7% and 5.1%, respecively. Predicabiliy (PV) repors he sronges explanaory power on independen variables. In erms of model f he F values are (P-value=0), (P-value=0.012), (Pvalue=0.009) and (P-value=0), respecively, indicaing significan levels of model fi. According o Table 5, informaion asymmery and accouning informaion qualiy are negaively correlaed. Feedback (FV) and imeliness (TM) boh reach he 5% significan level. Faihful presenaion (RF) reaches he 1% significan level. This shows ha he wors he informaion asymmery, he poorer he accouning informaion qualiy. As a resul, feedback, imelines and faihful presenaion are all dampened. The produc of IFRS and informaion asymmery (IFRS*IA) repors a posiive bu no significan correlaion. Whils he IFRS adopion can diminish he negaive effecs of informaion asymmery and improve accouning informaion qualiy, he benefis are no saisfacory. Therefore, H2 is no suppored. The resricions on daily upicks and downicks on Shenzhen Sock Exchange also lesson he improvemen of informaion asymmery and accouning informaion qualiy under he IFRS sysem. This may be why H2 is no suppored. This conclusion is consisen wih (Chang, 2006). The resuls of conrol variables are he same as Table IFRS, Ownership Srucure and Accouning Informaion Qualiy According o Table 6, adjused R 2 in respecive models is 19.3%, 1.4%, 3% and 7.8%, respecively. Predicabiliy (PV) repors he sronges explanaory power on independen variables. In erms of model f he F values are (P-value=0), (P-value=0.001), (Pvalue=0) and (P-value=0), respecively, indicaing significan levels of model fi. Table 6 shows he effecs of sae ownership, ownership by senior managers, and ownership by major shareholders, direcors and supervisors on accouning informaion qualiy under he IFRS sysem. Sae ownership (GOV) is significanly and negaively correlaed wih predicabiliy (PV) and faihful presenaion (RF). This shows ha he governmen shareholders ake a posiive aiude in corporae governance and monioring. A high sae ownership means poor predicabiliy and faihful presenaion of accouning informaion. There is a posiive correlaion wih he produc of IFRS and sae ownership (IFRS*GOV). Meanwhile, feedback (FV) and imeliness (TM) reach he 5% and 10% Significance levels, respecively. The IFRS adopion can indeed lessens he negaive effecs of sae ownership and improve he feedback and imeliness of accouning informaion. Therefore, H3-2 and H3-3 are suppored. However, he posiive benefis of he IFRS adopion on predicabiliy and faihful presenaion are no significan in he conex of sae ownership. Therefore, H3-1 and H3-4 are no suppored. 244