Accounting Research Journal

Transcription

1 Ownershp Influence and CSR Dsclosure n Chna Journal: Manuscrpt ID ARJ R Manuscrpt Type: Artcle Keywords: CSR, Foregn nvestors, SOEs, Insttutonal nvestors, Stakeholder salence theory, Stock exchange

2 Page of Ownershp Influence and CSR Dsclosure n Chna Abstract Purpose Ths study examnes the relatonshp between ownershp type and the lkelhood of publcaton of a corporate socal responsblty (CSR) report. Desgn/methodology/approach Drawng on stakeholder salence theory, the Probt model s employed for a sample of, Chnese lsted frms to study how dfferent types of owners nfluence frm CSR engagement. Fndngs The analyss reveals that the Chnese stock exchanges exert a postve nfluence on the lkelhood of a frm producng a CSR report, an effect whch s more sgnfcant n state owned enterprses (SOEs). Foregn nvestors lead to a greater lkelhood of publcaton of a CSR report, though ths effect s weaker n SOEs. In contrast, the holdngs of state and domestc nsttutonal nvestors are broadly neutral. Practcal mplcatons The study helps corporate managers to recognse how partcular types of shareholders wll value ther efforts regardng CSR actvtes and dsclosure, and also asssts polcy makers n mprovng the level of CSR dsclosure through the development of new polcy. Socal mplcatons Apposte CSR dsclosure enhances trust and facltates the shared values on whch to buld a more cohesve socety. Orgnalty/value The novelty of ths study s that t addresses the effect of nsttutonal nvestors on Chnese frm CSR engagement and thus provdes an mportant nsght for frms, nvestors, and other stakeholders nto the nterplay of portfolo nvestment and CSR. Key words: CSR, foregn nvestors, nsttutonal nvestors, SOEs, stakeholder salence theory, stock exchange Paper type: Research paper

3 Page of Ownershp Influence and CSR Dsclosure n Chna. INTRODUCTION Corporate socal responsblty (CSR) dsclosure may be nfluenced by the motves and values of an organsaton s shareholders. However, the precse effect of shareholders on CSR s far from clear (Neubaum & Zahra, 00), wth exstng studes focusng on the examnaton of the mpact of nsttutonal nvestors n a US or European settng (.e. Cox et al., 00; Johnson & Greenng, ; Neubaum & Zahra, 00; Dam & Scholtens, 0). Somewhat less attenton s focused on emergng markets regardng how varous types of shareholders relate to CSR dsclosure (Khan et al., 0). In a Chnese settng, exstng CSR studes have nvestgated the mpact of companes beng state-owned or otherwse on the lkelhood of CSR dsclosure (e.g. L et al., 0 & 0; Marqus & Qan, 0). Despte the greatly ncreased prevalence of nsttutonal nvestor ownershp n Chna, ts mpact s stll unexplored. Gven the very dfferent economc settng, t s uncertan whether the relatonshp between CSR dsclosure and nsttutonal ownershp documented for developed natons wll apply readly to Chnese companes. Ths paper thus seeks to fll the research gap by examnng the relatonshp between Chnese frm CSR engagement and ownershp structure, by determnng the mpact not only of state ownershp, but also the mpact of nsttutonal nvestors and foregn nvestors. Drawng on stakeholder salence theory, we argue that the more stakeholder attrbutes (power, legtmacy and urgency) a stakeholder embodes, the greater ts nfluence wll be on a frm s decson to ssue a CSR report. We fnd that the Chnese stock exchanges exert a postve nfluence on the lkelhood of a frm producng a CSR report, and such nfluence s more sgnfcant n SOEs. In contrast, the holdngs of state and domestc nsttutonal nvestors are broadly neutral. Foregn nvestors exert a postve mpact on the lkelhood of

4 Page of ssung a CSR report, though ths relatonshp s moderated by the presence of state ownershp. Ths paper makes several contrbutons. Frst, t adds to the emergng lterature on CSR practces by provdng emprcal evdence on the CSR-shareholder relatonshp n Chna, whle much of the exstng lterature s lmted to examnng the relaton between CSR and fnancal performance or corporate governance. Second, whle pror studes tend to focus exclusvely on state ownershp, ths s the frst study of the effect of dfferent nsttutonal nvestors on CSR dsclosure and corporate governance n Chna. We account for nvestment horzon heterogenety across nsttutonal nvestors due to dfferences n ther scale, role and socetal poston. Understandng the mpact of shareholder type on CSR dsclosure should offer an mportant nsght for frms, nvestors, and other stakeholders nto the nterplay of portfolo nvestment and CSR (Dam & Scholtens, 0). Our results may be used to help promote the CSR agenda and ncrease the level of CSR dsclosure n Chnese lsted frms. Fnally, our study extends the transton economy CSR lterature by examnng ownershp type. The remander of our paper s organsed as follows. Secton summarses the salent lterature on CSR n Chna. Secton dscusses the use of stakeholder salence theory as the theoretcal framework for our study, and then proceeds to formulate the hypotheses. Secton outlnes our research desgn. The results are then presented n Secton, and dscussed further n Secton. Secton sets out our conclusons and the mplcatons and lmtatons of the research.. CSR DEVELOPMENT IN CHINA It was not untl 00 that CSR was explctly recognsed n Company Law n Chna. In that year, the Shenzhen Stock Exchange (SZSE) ssued the Gude on the Socal Responsblty of

5 Page of Lsted Companes, followed by a Crcular publshed by the Shangha Stock Exchange (SSE) n 00. In the same year, the State-owned Assets Supervson and Admnstraton Commsson ssued the CSR Gudelne for Central Government Controlled SOEs requrng them to publsh CSR reports wthn three years of the ssued gudelne. Meanwhle, the Mnstry of Commerce released the Gude on Implementaton of CSR n 00 to promote socal responsblty n foregn-funded enterprses. Despte government endorsement of CSR, mplementaton of the gudelnes was far from effectve due to ther lack of clarty (Hu & Karbhar, 0). Exstng CSR studes for Chna are manly concerned wth examnng the level and determnants of the CSR dsclosure decson, ncludng the nfluence of poltcal connecton (Marqus & Qan, 0) and fnancal performance (L et al., 0), and the coverage of the CSR report. Few studes drectly examne the relatonshp between ownershp type and CSR dsclosure exclusvely for state ownershp. Lau et al. (0) and L et al. (0) fnd that CSR dsclosure s postvely assocated wth state ownershp, whle no such relatonshp s found by L et al. (0) and Marqus and Qan (0). These studes nether examne the effect of foregn or nsttutonal nvestors on Chnese frm CSR engagement, nor do they address the mpact of CSR ntatves by government agences (such as the two stock exchanges) on CSR dsclosure, the sole excepton beng Marqus and Qan (0). Our study, however, dffers from Marqus and Qan n two ways. Frst, we examne the effects of a varety of ownershp types on frm CSR dsclosure by employng stakeholder salence theory n contrast to the poltcal legtmacy perspectve employed by the authors. Second, we study how the effect of CSR ntatves by the two Chnese stock exchanges on the CSR reportng decson dffers between SOEs and non-soes.

6 Page of STAKEHOLDER SALIENCE THEORY AND HYPOTHESES DEVELOPMENT Freeman (, p.) defnes stakeholders as any group or ndvdual who can affect or s affected by the achevement of the organsaton s objectves, but fals to address the prortsaton of stakeholder clams. To dentfy stakeholder relevance, Mtchell et al. () propose the stakeholder salence theory. They defne salence as the degree to whch managers gve prorty to competng stakeholder clams (p.), whch s postvely related to the stakeholder s power to nfluence the frm, the legtmacy of the stakeholder s relatonshp wth the frm, and the urgency of the stakeholder s clam on the frm. The more attrbutes managers assess as strong, the hgher the salence of a partcular stakeholder. The stakeholder salence model s dynamc because the attrbutes can change for any partcular entty or stakeholder-manager relatonshp and are not objectve as they are based on management percepton. Further, the stakeholders may or may not be aware that they possess a partcular attrbute or may not be wllng or wsh to act on that attrbute. Drawng on stakeholder salence theory, we argue that the more powerful and vocal the shareholders, the greater ther nfluence on a frm s decson to produce a CSR report... State ownershp Nearly 0% of Chnese lsted frms are state-owned enterprses (SOEs) (L et al., 0), gvng state owners mportant votng rghts on key corporate decsons (Chen et al., 00) and provdng them wth greater power, legtmacy and nfluence over frm objectves, ncludng frm support for CSR ntatves. The nherent poltcal te enables SOEs to enjoy fnancal and regulatory favour (Wu et al., 0), but also pressures them to pursue poltcal and strategc goals n addton to proft maxmsng. Exstng studes on the effect of state holdngs on the CSR reportng decson provde lttle emprcal evdence n a developed naton settng (Van der Zee, 0; Dam & Scholtens,

7 Page of ). Research n a Chnese corporate settng has proved nconclusve. Some studes fnd that CSR dsclosure s postvely related to state ownershp (Lau et al., 0; L et al., 0). However, L et al. (0) and Marqus and Qan (0) fnd no such relatonshp. The government, as the paramount shareholder, provdes crtcal resources and legtmacy to Chnese SOEs whch are mportant to ther contnued vablty and success. Stakeholder salence theory suggests that large state holdngs can weld nfluence over frm managers decsons. Therefore, SOE managers attend to the demands of powerful state owners and prortse ther socal nterests more than other frms. We thus hypothesze: Hypothess : SOEs are more lkely to ssue CSR reports than non-soes. L et al. (0) argue that SOEs and non-soes have dfferent prortes n stakeholder management. Non-SOEs focus on proft maxmsaton whle SOEs tend to have wder strategc and poltcal objectves (Wu et al., 0), causng them to engage n and report upon socally responsble actvtes as a result of pressure from ther government owners. In addton, the CSR Gude and Crcular ssued by the two stock exchanges place a demand of legtmacy and urgency on SOEs dsclosng CSR nformaton. The combned pressure from the state owner and the two stock exchanges strengthens the mperatve of ther CSR demands to SOE executves. Consstent wth stakeholder salence theory, when two or more stakeholders jon forces to communcate ther respectve postons on CSR ssues, they are lkely to ncrease ther mpact. We argue that SOE managers are more lkely to respond to CSR dsclosure demands from state owners and the two stock exchanges n order to mantan ther legtmacy. We thus hypothesze: Hypothess : Where requred to publsh CSR reports by the two stock exchanges, SOEs are more lkely to ssue CSR reports than non-soes.

8 Page of Insttutonal nvestors Insttutonal nvestors have come under ncreasng pressure to account for socal performance n ther portfolos (Cox et al., 00) n the hope of dfferentatng ther servces and to demonstrate ther relablty and responsblty (Segel & Vtalano, 00). However, nsttutonal nvestors dffer n ther orentaton towards the frm s socal nvestments. Blar () fnds that nsttutonal owner nvestment horzons nfluence ther behavour and ncentves, arsng from ther dfferent mandates and dvergent stakeholder expectatons. Stakeholder salence theory suggests that the CSR demands of short-term nsttutonal nvestors may not be vewed as paramount compared to those of long-term nsttutonal nvestors due to ther lmted nfluence on frm management (Neubaum & Zahra, 00). We segment nsttutonal ownershp by nvestment horzon to examne whether frm decsons to ssue a CSR report are related to short-term nvestments, ncludng domestc mutual funds and nsurance companes, or longer-term nvestments, the Natonal Socal Securty Fund (NSSF), and the qualfed foregn nsttutonal nvestors (QFIIs). Short-term nsttutonal owners Mutual funds n Chna are subject to the scrutny of regulators, nvestors and the publc, are requred to make quarterly portfolo dsclosures, and must follow mandated nvestment styles (Natonal People s Congress, 00), leadng to an extremely hgh turnover rate. Such shorttermsm s nconsstent wth CSR objectves whch are nherently long-term (Graves & Waddock, ). It was not untl 00 that Chnese nsurers were allowed to nvest up to % of ther total assets n A-shares (stock exchange lsted) drectly and an addtonal % n shares ndrectly through mutual funds (Aggarwal et al., 0). To seek hgher nvestment yelds, they engaged n sgnfcant short term stock tradng (Huang et al., 0). Pror studes evdence mxed results on the effect of short-term nsttutonal ownershp on CSR performance. Neubaum and Zahra (00) fnd such owners are less lkely to focus on

9 Page of frm socal performance, though Johnson and Greenng () and Cox et al. (00) fnd no evdence of ths. To date, no studes have examned how nvestment by mutual funds and nsurance companes affects frm CSR dsclosure n Chna. Gven ther transent focus, stakeholder salence theory suggests that the CSR expectatons of short-term nsttutonal nvestors wll not receve as much management attenton as those of long-term nsttutons. Because the former are solely focused on a fnancal nvestment relatonshp wth an nvestee, CSR actvty may be merely seen as a costly actvty. We therefore hypothesze: Hypothess : The degree of short-term nsttutonal stock ownershp (e.g. mutual funds and nsurance companes) s negatvely assocated wth the probablty of frms ssung CSR reports. Long-term nsttutonal nvestors Stakeholder salence theory suggests that a long-term nvestment horzon enables long-term nsttutonal nvestors to establsh effectve and sustanable relatonshps wth nvestee frms, thus ncreasng ther salence to frm managers and ganng manager support for CSR dsclosure (Neubaum & Zahra, 00). Long-term nsttutonal owners are more lkely to nvest n frms wth good socal performance due to ts potentally favourable mpact on longrun rsk and return (Graves & Waddock, ). The Chnese publc penson fund, the NSSF, establshed n 000, began nvestng n the stock markets n 00. Snce then the NSSF has become the largest nsttutonal nvestor n Chna (Lecke & Pan, 00). The NSSF supports nvestment n long-term actvtes, consstent wth a commtment to CSR. Drawng on stakeholder salence theory, the sze and endurng focus of long-term nsttutonal holdngs can change frm management prortes and decsons, lendng such nvestors more authorty. Despte the lack of emprcal research on the NSSF and CSR n Chna, the broad lterature consensus suggests that such penson funds

10 Page of postvely promote frm CSR dsclosure (Johnson & Greenng, ; Neubaum & Zahra, 00). Thus, we state the followng hypothess: Hypothess : The degree of NSSF ownershp s postvely assocated wth the lkelhood of a frm ssung a CSR report... QFIIs QFIIs may empower frms to promote CSR dsclosure for the followng reasons. Frst, QFIIs are subject to a strct approval process to vet new entrants (Huang & Zhu, 0), confrmng that they mantan long-term nvestment phlosophes (CSRC, 0). Stakeholder salence theory suggests that executves wll pay more attenton to the preferences of long-term nsttutonal owners. Second, despte QFII holdngs accountng for only % of Chnese stocks, ther nfluence s mportant as they are more nclned to exercse ther voce rather than ext (Huang & Zhu, 0), consstent wth stakeholder salence theory whch argues that vocal stakeholders are lkely to have a greater nfluence on frm strategy. In addton, the ncreasng sgnfcance of foregn fnancng of Chnese stocks hghlghts the mportance of QFII demands n the mnds of nvestee frm managers, makng a CSR dsclosure request more compellng. We therefore state the followng hypothess: Hypothess : The degree of QFII ownershp s postvely assocated wth the lkelhood of a frm ssung a CSR report. Wu et al. (0) fnd that n weaker legal envronments, nsttutonal nvestors may have fewer channels to effectvely montor frm actvtes, and ths may affect the process of stakeholder prortsaton. Gven the unque nsttutonal features and weak legal envronment n Chna (Fan et al., 0), we propose that the nfluence of foregn ownershp on CSR dsclosure s moderated by the presence of state ownershp as the latter reduces relance on

11 Page of external captal market fundng, and may weaken foregn ownershp pressure to engage n CSR actvtes. Senor managers must balance competng expectatons, determnng whch stakeholders are salent and therefore should receve management attenton (Neubaum & Zahra, 00). We argue that foregn ownershp demand for CSR dsclosure may be moderated by state co-owners who enjoy greater power, legtmacy and urgency over lsted frms. Thus, we propose the followng hypothess: Hypothess : There s a weaker assocaton between the degree of QFII ownershp and the lkelhood of a frm ssung a CSR reports n SOEs than s the case n non- SOEs.. RESEARCH DESIGN Our sample conssts of all publcly traded Chnese frms lsted on the SSE and SZSE for the year 0, the frst and only year that CSR ratngs were dsclosed for all Chnese frms by the Chnese Academy of Socal Scences Corporate Socal Responsblty Research Centre (CASS-CSR). We obtan the CSR data from the Whtepaper on Chnese frms CSR by the CASS-CSR (0), the fnancal and ownershp data from the Chna Stock Market and Accountng Research database, and the lst of frms requred to ssue CSR reports from the Shangha and Shenzhen Stock Exchanges, respectvely. After removng frms wth mssng data, fnancal frms, and those n specal treatment, our fnal sample conssts of, frms, of whch publshed CSR reports. Table reports the number of frms ssung CSR reports by ndustry class where more than half of the CSR reports ssued were by manufacturng frms. The mean percentage of publshed CSR reports across all ndustres s %, wth the hghest percentage of % for the mnng ndustry. [Insert Table here]

12 Page of The dependent varable CSR report CSR report s a dchotomous varable that equals f the frm ssued a CSR report n 0, and 0 otherwse, accordng to the CASS-CSR data... Shareholder varables Shareholder varables nclude state control, nsttutonal ownershp, and foregn ownershp. State control s a dummy varable takng the value of f the frm s ultmately controlled by the state, and 0 otherwse. To ensure ths varable reflects the nature of the ultmate controller of a frm, we measure control consstent wth La Porta et al. (). Insttutonal ownershp has three types: () mutual fund ownershp the proporton of shares owned by Chnese mutual funds; () nsurance ownershp the proporton of shares owned by Chnese nsurance companes; and () NSSF ownershp the proporton of shares owned by the NSSF. Fnally, foregn ownershp s the proporton of shares owned by QFIIs... Control varables We nclude control varables to account for ndustry and frm characterstcs. We compute the logarthm of total assets to gauge frm sze and employ the return on assets (ROA) to measure frm fnancal performance, both of whch have a postve effect on CSR dsclosure (Orltzky, 00; L et al., 0). Frm leverage, measured as total labltes to total assets, s used to gauge the nfluence of credtor power on CSR dsclosure (Roberts, ). Furthermore, we employ the Zeno ndex (Z-ndex) to gauge ownershp concentraton, consstent wth Lu et al. (0), and calculated as the number of shares owned by the largest shareholder deflated by the number of shares owned by the second largest shareholder. Sxteen ndustry dummes are created as Dam and Scholtens (0) fnd ndustry has a sgnfcant nfluence on CSR. We use CSRC ndustry codes, wth dversfed ndustres as the reference category. We use a

13 Page of stock exchange dummy to dstngush the frms lsted n the SSE from those n the SZSE, wth lstng on the SZSE as the reference category. Fnally, we nclude a requred dscloser dummy varable to equal f a frm s requred by ether the SSE or the SZSE to ssue a CSR report, and 0 otherwse. As the dependent varable s a bnary varable, we employ the Probt model to explan the lkelhood of a frm ssung a CSR report. The specfcaton for the baselne model s gven n equaton (). CSRRP = β + β SIZE + β ROA + β LEVERAGE + β OWNCON + β SSE + β SE + β SOE + β MF + β INSUR + β NSSF + β FOREIGN + ε ( ) 0 where CSRRP s the dchotomous varable measurng the presence of a CSR report for frm, SIZE s the logarthm of total assets, ROA s return on assets, LEVERAGE s measured as total labltes deflated by total assets, OWNCON measures ownershp concentraton, SSE s the Shangha Stock Exchange dummy, SE s a dummy varable for stock exchange requred dscloser, SOE s the state owned enterprse dummy, MF s the percentage of shares owned by mutual funds, INSUR s the percentage of shares owned by nsurance companes, NSSF s the percentage of shares owned by the NSSF, FOREIGN s the percentage of shares owned by QFIIs, β 0 s the ntercept, and ε s the model error. The specfcaton for the full model ncludng the nteracton terms s gven n equaton (). + β SE + β SOE + β MF + β INSUR + β NSSF + β FOREIGN + β SOE SE + β SOE FOREIGN + ε where the nteracton term SOE SE tests hypothess and the nteracton term SOE FOREIGN tests hypothess. CSRRP = β + β SIZE + β ROA + β LEVERAGE + β OWNCON + β SSE 0 ( )

14 Page of RESULTS.. Descrptve analyss Table reports the descrptve statstcs and correlaton matrx for our model varables. In total, % of frms ssued a CSR report and 0% are controlled by the state. Mutual fund nvestors are the bggest nsttutonal nvestors, on average holdng.% of the shares outstandng... The CSR models [Insert Table here] Table presents the results of seven specfcatons of the model explanng the lkelhood of a frm ssung a CSR report. Model presents results for the control varables. Frm sze and fnancal performance are postvely related to the lkelhood of a frm publshng a CSR report, consstent wth the exstng lterature (Waddock & Graves, ). Large frms are more lkely to ssue CSR reports because of ther greater publc vsblty. Proftable frms enjoy more organsatonal slack, thus are more lkely to nvest n CSR actvtes. Hgher ownershp concentraton reduces the lkelhood of a frm publshng a CSR report, consstent wth the exstng lterature (Roberts, ). Frms requred by the two stock exchanges to publsh CSR reports are more lkely to provde such a report. However, none of the ndustry dummes are sgnfcant, suggestng that ndustry membershp does not mpact upon the lkelhood of a frm ssung a CSR report. [Insert Table here] Models to test the hypotheses. Hypothess states that SOEs are more lkely to ssue CSR reports than non-soes. In Model, the coeffcent of state control s postve but nsgnfcant and therefore hypothess s not supported. In Model, the nteracton term

15 Page of between state control and stock exchange s postve and sgnfcant at the % level, supportng hypothess, whch states that where requred to publsh CSR reports by the two stock exchanges, SOEs are more lkely to ssue CSR reports than non-soes. Hypothess states that the degree of short-term nsttutonal stock ownershp s negatvely assocated wth the probablty of frms ssung CSR reports. In Model, the coeffcent for mutual fund ownershp s postve but nsgnfcant, and the coeffcent for nsurance ownershp s negatve and sgnfcant at the % level, offerng weak support for hypothess. Hypothess states that the extent of NSSF ownershp s postvely assocated wth the presence of a CSR report. The coeffcent of NSSF s postve but nsgnfcant, and thus hypothess s not supported. Hypothess states that the degree of QFII (foregn) ownershp s postvely assocated wth the lkelhood of a frm ssung a CSR report. In Model, the coeffcent for foregn ownershp s postve and sgnfcant at the % level, offerng weak support for hypothess. Hypothess states that there s a weaker assocaton between the degree of QFII ownershp and the presence of a CSR report n SOEs than s the case n non-soes. In Model, the coeffcent of the nteracton term between foregn ownershp and state control s negatve and sgnfcant at the % level, thereby supportng hypothess. Model ncludes all shareholder varables and the results are consstent wth those of the reduced form models. Two tests are conducted to check the robustness of the results. Frst, for all models, we replace the Z-ndex wth the ownershp percentage of the largest shareholder to measure ownershp concentraton. Second, n Models and, we replace shares by foregn nvestors wth a bnary varable measurng whether or not a frm has foregn nvestors. The unreported results suggest that the fndngs regardng ownershp concentraton and foregn nvestment are robust.

16 Page of DISCUSSION.. The effect of the state and the stock exchanges on CSR dsclosure Our results show that frms under state control are not more lkely to publsh CSR reports than prvately controlled frms, consstent wth Marqus and Qan (0) and Dam and Scholtens (0). However, SOEs are more lkely to ssue CSR reports than non-soes where requred to publsh CSR reports by the two stock exchanges. These fndngs can be explaned by stakeholder salence theory. Frst, although the Chnese state s clearly a powerful and legtmate stakeholder of the frms n whch t has controllng shares, and promotng CSR s one of ts socal objectves, ts demand for managers to publsh CSR reports lacks urgency. Mtchell et al. () argue that for stakeholders to acheve salence to frm managers, they must have ether power to enforce ther wll or ther clam should be perceved as urgent. Our results contrast wth the study by Lau et al. (0) who report a postve relatonshp, though ths may be attrbuted to ther use of the percentage of shares owned by the state as the measure of state control. Second, the stock exchanges strengthen the salence of the CSR demand of state owners. Mtchell et al. () argue that domnant stakeholders (e.g. state owners) wth power and legtmacy may become salent f ther legtmate stake becomes urgent, and therefore earn the manager s mmedate prorty. When the specfc requrement from the stock exchanges for a CSR report s absent, managers appear not to treat the state owner s demand for CSR reports as ther top prorty. However, when a frm s requred by the stock exchanges to publsh a CSR report, the state owner s demand suddenly becomes urgent; thus, the managers of these frms are more lkely to comply than s the case for prvately controlled frms.

17 Page of The effect of nsttutonal nvestors on CSR dsclosure Our results show that short-term nsttutonal nvestors have no mpact on the lkelhood of frm CSR dsclosure, consstent wth fndngs for developed countres such as n Johnson and Greenng (), Neubaum and Zahra (00) and Cox et al. (00). The short-termsm of Chnese mutual funds and nsurance companes dscussed n the extant lterature s nconsstent wth the objectves of CSR, and thus funds earn less salence n ther relatonshp wth frm managers. We fnd that NSSF ownershp has no drect mpact on the lkelhood of CSR dsclosure. Ths s somewhat surprsng as numerous studes for other countres document a postve relaton (Johnson & Greenng, ; Cox et al. 00; Oh et al. 0). Our results may be explaned by the potental corporate governance shortcomngs of the NSSF tself. Accordng to Lecke and Pan (00), the NSSF has no clearly defned objectves and purposes, no external ndependent drector on the board, and the extremely abbrevated annual reports are not audted. These factors could lead to ts long-term strategc goals and prortes beng compromsed, and thus the NSSF may not n effect behave as a long-term nvestor... The effect of foregn nvestors on CSR dsclosure We fnd that holdngs by QFIIs (foregn nvestors) exert a postve mpact on the probablty of frm CSR dsclosure, consstent wth other emergng market emprcal studes (e.g. Oh et al., 0; Khan et al., 0). The QFIIs are prncpally large nsttutonal nvestors wth longterm nvestment phlosophes, meanng they enjoy greater salence n ther relatonshp wth executves; thus, ther CSR demands are more lkely to be met. Furthermore, the QFIIs are more vocal than domestc nsttutonal nvestors, therefore ntensfyng ther salence n the mnds of managers (Huang & Zhu, 0). In addton, we fnd that the postve relaton between the QFIIs and the lkelhood of CSR dsclosure s weakened consderably across

18 Page of frms controlled by the state. Ths may be because SOEs have more avalable resources, and as a result pay less attenton to mnorty shareholders such as QFIIs, thus reducng ther salence to managers.. CONCLUSION Drawng upon stakeholder salence theory, ths paper examnes the mpact of dfferent types of ownershp on the lkelhood of CSR dsclosure n Chnese lsted frms. The results hghlght that where frms are requred to dsclose by the stock exchanges, the lkelhood of ssung CSR reports s hgher n SOEs than n non-soes. The results also suggest that foregn ownershp has a postve mpact on the presence of CSR reports, though ths effect s weaker n SOEs. Our study also provdes several unexpected yet nsghtful results. Frst, no sgnfcant relatonshp s found between state ownershp and the lkelhood of CSR dsclosure, confrmng exstng studes (e.g. Dam & Scholtens, 0). We argue that SOEs have the least need to use CSR dsclosure to seek preferred status due to ther nherent poltcal legtmacy. Second, nsurance company ownershp exerts a negatve mpact on the lkelhood of CSR dsclosure, suggestng that they perceve producng a CSR report as merely a short-term cost. Our study has mportant mplcatons for both polcy and socety. The results reveal that the ssue of varyng CSR gudelnes by dfferent authortes may have led to nconsstent dsclosure practces across frms. To enhance the quantty and qualty of CSR dsclosure, the Chnese government should provde more detaled gudance n terms of requred CSR report content and form (e.g. a standalone report). Ths should ensure CSR dsclosure comparablty, but more mportantly, frm accountablty to all stakeholders, thereby enhancng trust and facltatng the shared values on whch to buld a more cohesve socety. Second, our results shed addtonal lght on the postve effects of foregn ownershp on the lkelhood of CSR

19 Page of dsclosure. Chnese polcy-makers may consder allowng more QFIIs nto the domestc securtes market to facltate socally responsble nvestments. Thrd, Chnese frms should provde greater CSR tranng to ther managers, accountants and audtors. Our study also has some practcal mplcatons. Frst, our results confrm that dfferent owners may have dvergent and even competng perspectves on CSR. Understandng ths should help corporate managers to recognse how partcular types of shareholders wll value ther efforts regardng CSR actvtes and dsclosure; therefore mprovng corporate responses to those shareholders through CSR goal congruence. Second, our study should provde some gudance to help Chnese lsted frms to drect ther efforts more effectvely by, for example, dsclosng more CSR nformaton when attemptng to attract QFIIs or long-term nvestors. Our results should nevertheless be nterpreted wth a degree of cauton. Frst, the sngle year dataset employed may somewhat restrct the generalsaton of our fndngs. Second, ths study makes the smplfyng assumpton that CSR engagement and CSR dsclosure are synonymous. Thrd, our study examnes the causal effects of ownershp on the lkelhood of CSR dsclosure, though CSR performance tself s not measured. Fnally, whle the study examnes how ownershp mpacts upon CSR dsclosure, t may also be nsghtful to understand how CSR dsclosure s affected by other stakeholders, such as customers and the accountng professon.

20 Page of NOTES: The two stock exchanges requre that SZSE 0 Index component frms, SSE Corporate Governance Index frms, and SSE frms lsted overseas must provde a CSR report. Two addtonal models examne the mpact of manageral ownershp on the lkelhood of CSR dsclosure where the percentage of shares owned by drectors and senor managers s employed to measure manageral ownershp. However, the unreported results are nsgnfcant and do not alter the results for the other varables, ndcatng that manageral ownershp has lttle mpact on the CSR dsclosure n Chna.

21 Page 0 of REFERENCES: Aggarwal, R., Hu, M., & Yang, J. (0). Fraud, market reacton, and the role of nsttutonal nvestors n Chnese lsted frms. The Journal of Portfolo Management, (), -. Blar, M. (). Ownershp and Control: Rethnkng Corporate Governance for the Twenty- Frst Century. Washngton, DC: Brookngs Insttuton. CASS-CSR (0). Whte book of Chnese CSR reports. The Chnese Academy of Socal Scences Corporate Socal Responsblty Research Centre. Avalable from: [Accessed September 0]. Chen, G., Frth, M., & Xu, L. (00). Does the type of ownershp control matter? Evdence from Chna s lsted companes. Journal of Bankng & Fnance, (), -. Chna Securtes Regulatory Commsson (0). QFII Quota Rased to USD Bllon and RQFII Plot Expanded n Sngapore and London. [Onlne] Avalable at ml [Accessed October 0] Cox, P., Brammer, S., & Mllngton, A. (00). An emprcal examnaton of nsttutonal nvestor preferences for corporate socal performance. Journal of Busness Ethcs,, -. Dam, L., & Scholtens, B. (0). Does ownershp type matter for corporate socal responsblty? Corporate Governance: An Internatonal Revew, 0(), -. Fan, J., Wong, T., & Zhang, T. (0). Insttutons and organzatonal structure: the case of state-owned corporate pyramds, Journal of Law, Economcs, and Organzaton, (),. Freeman, R. E. (). Strategc Management: A Stakeholder Approach. Boston: Ptman. Graves, S. B., & Waddock, S, A. (). Insttutonal owners and corporate socal performance. Academy of Management Journal, (), -. 0

22 Page of Hu, Y. Y, & Karbhar, Y. (0). Incentves and dsncentves of corporate envronmental dsclosure: evdence from lsted companes n Chna and Malaysa. Thunderbrd Internatonal Busness Revew, (), -. Huang, E. J., Lu, E. P., & Kao, G. W. (0). Investment regulaton, portfolo allocaton, and nvestment yeld n the U.S. and Chna nsurance ndustry. The Chnese Economy,, -. Huang, W., & Zhu, T. (0). Foregn nsttutonal nvestors and corporate governance n emergng markets: evdence of a splt-share structure reform n Chna. Journal of Corporate Fnance,, -. Johnson, R. A., & Greenng, D. W. (). The effects of corporate governance and nsttutonal ownershp types on corporate socal performance. Academy of Management Journal, (), -. Khan, A., Muttakn, M. B., & Sddqu, J. (0). Corporate governance and corporate socal responsblty dsclosures: evdence from an emergng economy. Journal of Busness Ethcs,, 0-. La Porta, R., Lopez-de-Slanes, F. & Shlefer, A. (). Corporate ownershp around the world, Journal of Fnance,, -. Lau, C., Lu, Y., & Lang, Q. (0). Corporate socal responsblty n Chna: a corporate governance approach. Journal of Busness Ethcs,, -. Lecke, S., & Pan, N. (00). A revew of the natonal socal securty fund n Chna. Pensons: An Internatonal Journal, (), -. L, D., Ln, H., & Yang, Y. (0). Does the stakeholders corporate socal responsblty (CSR) relatonshp exst n emergng countres? Evdence from Chna. Socal Responsbltes Journal, (), -.

23 Page of L, Q., Luo, W., Wang, Y., & Wu, L. (0). Frm performance, corporate ownershp, and corporate socal responsblty dsclosure n Chna. Busness Ethcs: A European Revew, (), -. Lu, N., Bredn, D., Wang, L., & Y, Z. (0). Domestc and foregn nsttutonal nvestors behavour n Chna. The European Journal of Fnance, 0(-), -. Marqus, C., & Qan, C. (0). Corporate socal responsblty reportng n Chna: symbol or substance? Organzaton Scence, (), -. Mtchell, R. K., Agle, B., R., & Wood, D., J. (). Toward a theory of stakeholder dentfcaton and salence: defnng the prncple of who and what really counts. The Academy of Management Revew, (), -. Natonal People s Congress (NPC), (00). Mutual Funds Law. The NPC, Bejng, Chna. Neubaum, D., O., & Zahra, S., A. (00). Insttutonal ownershp and corporate socal performance: the moderatng effects of nvestment horzon, actvsm, and coordnaton. Journal of Management, (), -. Oh, W. Y., Chang, Y. K. & Martynov, A. (0). The effect on ownershp structure on corporate socal responsblty: emprcal evdence form Korea. Journal of Busness Ethcs,, -. Orltzky, M. (00). Does frm sze confound the relatonshp between corporate socal performance and frm fnancal performance. Journal of Busness Ethcs, (), - 0. Roberts, R. W. (). Determnants of corporate socal responsblty dsclosure: an applcaton of stakeholder theory. Accountng, Organzatons and Socety, (), -.

24 Page of Segel, D. S., & Vtalano, D. F. (00). An emprcal analyss of the strategc use of corporate socal responsblty. Journal of Economcs & Management Strategy, (), -. Van der Zee, E. (0). Soveregn wealth funds and socally responsble nvestment: dos and don ts. European Company Law, (), -. Wu, W., Johan, S. A., & Ru, O., M. (0). Insttutonal nvestors, poltcal connectons and the ncdence of regulatory enforcement aganst corporate fraud. Journal of Busness Ethcs,, 0-. Wu, W., Wu, C., Zhou, C., & Wu, J. (0). Poltcal connectons, tax benefts and frm performance: evdence from Chna. Journal of Accountng and Publc Polcy,, -00.

25 Page of Table : The proporton of frms publshng a CSR report by the CSRC ndustry classfcaton Industry No. of CSR report frms No. of Non- CSR frms Total number of frms Per cent of CSR report frms (%) Agrculture, forestry, lvestock farmng and fshery 0% Mnng % Manufacturng 0 % Electrc power, gas and water producton and supply % Constructon % Wholesale and retal trade 0 % Transport, storage and postal servce 0 % Accommodaton and caterng % Informaton transmsson, software and nformaton technology servces 0 0% Real estate 0 % Leasng and commercal servce % Scentfc research and techncal servce % Water conservancy, envronment and publc faclty management 0% Resdent servce, repar and other servces % Health and socal servce 0% Culture, sports and entertanment % Dversfed % Total 0 % Note: The labels CSR report frms and Non-CSR report frms correspond to the number of frms publshng a CSR report and those whch do not, classfed by the 0 CSRC ndustry classfcaton code. Total number of frms corresponds to the total number of frms n each ndustry class. Percentage of CSR frms s the Number of CSR report frms dvded by the Total number of frms for each ndustry.

26 Page of Table : Descrptve statstcs and correlatons for the model varables Varable Mean S.D. CSRRP SIZE.. 0. ROA LEVERAGE OWNCON SSE SE SOE MF (%) INSUR (%) NSSF (%) FOREIGN (%) Note: N =,. When the absolute value of a correlaton coeffcent s greater than or equal to 0.0, t s sgnfcant at the % level.

27 Page of Table : Probt models of the lkelhood of frm CSR report publcaton Varable Model Model Model Model Model Model Model Control varables SIZE 0.*** 0.*** 0.*** 0.*** 0.*** 0.00*** 0.*** (0.00) (0.00) (0.00) (0.0) (0.0) (0.0) (0.0) ROA.0***.***.***.***.***.00***.*** (.0) (.0) (.00) (.0) (.) (.) (.) LEVERAGE (0.) (0.0) (0.) (0.) (0.) (0.) (0.) OWNCON -0.00** -0.00** -0.00** -0.00** -0.00** -0.00** -0.00** (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) SSE -0.*** -0.** -0.0** -0.0** -0.** -0.** -0.** (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.) SE.***.***.0***.***.***.***.0*** (0.) (0.) (0.0) (0.) (0.) (0.) (0.) Shareholder varables H: SOE (0.0) (0.) (0.0) (0.0) (0.) (0.) H: SOE SE 0.** 0.** (0.) (0.) H: MF (0.0) (0.0) (0.0) (0.0) H: INSUR -0.0* -0.** -0.0* -0.0* (0.0) (0.0) (0.0) (0.0) H: NSSF (0.0) (0.0) (0.0) (0.0) H: FOREIGN 0.* 0.*** 0.*** (0.0) (0.) (0.) H: FOREIGN SOE -0.** -0.** (0.) (0.) Intercept -.*** -.*** -.*** -.*** -.*** -.*** -.*** (.0) (.0) (.0) (.0) (.0) (.00) (.0) Pseudo R Note: Standard errors are shown n parentheses; ndustry dummes are ncluded, but not reported here. * p < 0., ** p < 0.0, *** p < 0.0 level, two-taled coeffcent test (N =,).