CORPORATE GOVERNANCE, ANALYST FOLLOWING AND FIRM VALUE Sulaman Mousell Damascus Unversty, Damascus, Syra and Khaled Hussaney* Strlng Unversty, Strlng, UK Ths paper s accepted for publcaton at: Corporate Governance: The Internatonal Journal of Busness n Socety * Correspondence should be addressed to Dr. Khaled Hussaney, Accountng & Fnance Dvson, Strlng Management School, Unversty of Strlng, Strlng FK9 4LA, UK. Emal: Khaled.Hussaney@str.ac.uk. 1
Abstract Purpose: We examne the mpact of a frm s corporate governance mechansms on the number of fnancal analysts followng UK frms. We also examne the potental effect of the number of analyst followng frms n the Unted Kngdom (UK) on the assocaton between corporate governance mechansms and frm value. Methodology: We use multple regresson models to examne the assocaton between corporate governance, analyst coverage and frm value for a large sample of UK frms lsted n London Stock Exchange wth fnancal year-ends between January 2003 and December 2008. Fndngs: We fnd that the aggregate level of corporate governance qualty s postvely assocated wth the number of analysts followng UK frms. We also fnd that compensaton score s the man component that affects the number of analysts followng UK frms. Our results suggest that fnancal analysts are partcularly concerned wth how much compensaton that executves and drectors receve. Ths s consstent wth Jensen and Mecklng (1976) who argue that CEO compensaton can be used as effectve mechansms for mtgatng agency costs. Hence hgher levels of CEO compensaton attract attracts more fnancal analysts to follow the frm. Surprsngly, when we examne the jont effect of both corporate governance qualty and the number of analysts followng on frm value, we fnd no sgnfcant effect for both varables on frm value. Orgnalty: We contrbute to pror research by provdng the frst emprcal evdence on the mpact of dsaggregated levels of corporate governance on analysts followng and frm value for a large sample of UK frms. Keywords: corporate governance; analyst followng, frm value, Unted Kngdom. Classfcaton: Research paper. 2
1. Introducton The present paper ams to examne the effect of a frm s corporate governance (CG) mechansms on the number of analysts followng UK frms. It also examnes the potental mpact of analysts followng UK frms on the assocaton between CG mechansms and frm value. The study uses the agency theory to examne these assocatons. Based on the agency theory, the exstence of hgh qualty CG mechansms should mtgate the agency costs n the relaton between the prncpal (shareholders) and ther agents (managers) (Jensen and Mecklng, 1976). Hgh qualty CG mechansms are hghly valued by the stock market. For example, pror research fnds that CG s a key determnant for frm valuaton (Lang et al, 2003), cost of equty (Bowen et al, 2006), market lqudty (Roulstone, 2003) and frms credt ratngs (Ashbaugh-Skafe et al (2006). Another stream of lterature studes on the usefulness of CG mechansms to the stock market tend to focus on fnancal analysts. Ths s because fnancal analysts are wdely consdered to be a very nfluental user group n stock markets. They represent and nfluence nvestors beleves and actvtes (e.g., Schpper, 1991; Hrst et al., 1995; and Lang and Lundholm, 1996). Ths lne of research fnds that CG affects the accuracy of analyst forecasts (Bhat et al, 2006 and Byard et al, 2006); the number of analysts followng frms and frm value (Lang et al., 2004; Haw et al., 2004; Yu, 2007 & 2010 and Boubaker and Labegorre, 2008). On the current study, we extend ths lne of lterature by examnng the assocaton between corporate governance, analyst followng and frm value n the UK. The frst verson of the UK Corporate Governance Code was ntroduced by the Cadbury Commttee In 1992; ts paragraph 2.5 s stll the classc defnton of the context of the Code (FRC, 2010): Corporate governance s the system by whch companes are drected and controlled. Boards of drectors are responsble for the governance of ther companes. The shareholders role n governance s to appont the drectors and the audtors and to satsfy themselves that an approprate governance structure s n place. The responsbltes of the board nclude settng the company s strategc ams, provdng the leadershp to put them nto effect, supervsng the management of the busness and reportng to shareholders on ther stewardshp. The board s actons are subject to laws, regulatons and the shareholders n general meetng. The present paper utlzes the unque UK regulatory framework n whch t combnes hgh standards of corporate governance wth relatvely low assocated costs (FRC, 2006). In 3
addton, t s argued that the UK outperforms other countres n terms of governance standards, whle complance costs are estmated to be lower n the UK compared to other countres wth comparable standard (FRC, 2006). Ths context enables the present study to contrbute to exstng corporate governance research by examnng whether corporate governance mechansms affect analyst followng and frm value. Ths research s mportant for ts potental polcy mplcatons. It helps to nform regulators (and managers) about the benefts of mprovng the qualty of corporate governance mechansms to frms. In ths way t provdes part of the nformaton needed for a more cost-beneft analyss of the use of dfferent corporate governance mechansms by UK frms. We test to see f levels of corporate asymmetrc nformaton (measured by the number of analysts followng UK frms) can be reduced through corporate governance mechansms. We also test the jont effect of both corporate governance mechansms and the number of analysts followng on UK frm value. Analyst followng s consdered as one of the most mportant determnants of corporate valuaton (.e., Claessens et al, 2002; Lang et al, 2003), the cost of equty captal (.e., Bowen et al., 2008), market lqudty (.e., Attg, et al. 2006), and dvdend payment decson (.e., Basddq and Hussaney, 2010). In addton, pror research fnds that frms wth hgh qualty corporate governance mechansms are more lkely to be followed by a large number of fnancal analysts and that hgher analyst followng s assocated wth hgher valuaton for frms facng governance problems (Lang et al, 2004). Lang et al (2004) s the frst to study the assocaton between corporate governance, analyst followng and frm value n the UK. Ther analyss was focus on a sample of 638 frms wth fnancal year ends on 1996. However, they dd not examne a full set of corporate governance mechansms. Another stream of research n the UK examnes the assocaton between corporate governance and corporate voluntary dsclosure (.e. Hussaney and Al-Najjar, 2010; Hussaney and Wang, 2010; L, et al, 2008). However, these studes use the level of corporate dsclosure as a proxy for the number of fnancal analysts followng UK frms. In addton, they lmt ther analyss to a subset of corporate governance mechansms. In the present paper, we respond to a recent call for research that explans the mpact of corporate governance mechansms on analyst followng for a longer tme perod and larger sample sze (Yu, 2010). In partcular, we nvestgate whether frms that possess strong 4
nternal and external corporate governance mechansms have better captal market performance as a result of more analysts followng and better frm value relatve to frms wth weak corporate governance mechansms. We contrbute to corporate governance research by offerng new evdence on the assocaton between corporate governance, analyst followng and frm value for a large sample of UK frms. An mportant feature of ths paper s that t uses data of the largest corporate governance data provder to nsttutonal nvestors, Corporate Governance Quotent (CGQ) offered by RskMetrcs Group. Ths database contans 55 CGQ governance factors whch span eght categores of corporate governance ncludng board, compensaton, takeover and audt. We use an overall measure of corporate governance qualty score as well as ndvdual sub-scores to examne whch CGQ governance factors are sgnfcantly assocated wth analyst followng and frm value for UK frms. Usng a sample of 1514 UK lsted frms; we fnd that there s a postve assocaton between the qualty of overall corporate governance mechansms and the number of fnancal analysts coverng UK frms. Ths suggests that fnancal analysts have greater ncentves to follow better-governed UK frms. However, we fnd that not all corporate governance mechansms affect analyst followng. In partcular, we fnd that compensaton mechansm s the only sgnfcant varable n all regresson models. In addton, followng Lang et al. (2004) we use a regresson model that takes nto account the endogenety between analyst followng and frm value. However, we fnd no valuaton effect when fnancal analysts cover UK frms wth good corporate governance mechansms. Ths suggests that fnancal analysts serve as a montorng devce and act as substtute mechansm for corporate governance when valung frms. The rest of the paper s organsed as follows. Secton 2 revews relevant lterature and develops our research hypotheses. Secton 3 descrbes data and the sample selecton. Secton 4 descrbes the research desgn. Secton 5 presents the emprcal results. Secton 6 concludes and suggests lnes for further research. 2. Relevant Lterature and Hypotheses 2.1 Analyst followng and corporate governance There are three reasons why qualty of corporate governance may affect the number of analysts followng the company. Frst of all, Fan and Wong (2002) and Wllekens et al. 5
(2005) argue that frms tend to dsclose a better qualty nformaton f they mplement effectve corporate governance mechansms. Ths wll consequently reduce the cost for an analyst to follow such frms. Second, frms wth effectve corporate governance mechansms are more able to drectly montor ther managers, whch n turn, algn the managers nterests wth the shareholders nterests and reduce the agency problem arsng from the separaton of ownershp. Thrd, McNchols and O Bren (1997) argue that analysts allocate ther efforts on the bass of ther expectatons of the frm s future performance and they are nclned to add frms they vew favourably. Klapper and Love (2004) and Durnev and Km (2005) suggest that analysts prefer to follow better-governed frms because they are assocated wth better management, and hence better future economc vablty. For the above three reasons, we expect a postve assocaton between the number of analyst followng and the qualty of corporate governance. Our frst hypothess s stated as follows: H 1 : Analyst followng s postvely assocated wth the qualty of corporate governance. 2.2 Analyst followng, corporate governance and frm value There s an mportant role for fnancal analysts n conveyng frms performance to the market. Analysts substtute for corporate governance and play a governance role n the market. Ths substtuton effect has been documented by two recent studes. Lang et al. (2004) provded evdence that the negatve effect of lower nvestor protecton on valuaton s mtgated when analyst followng s greater. Knyazeva (2007) fnds that analysts substtute corporate governance n ther mpact on frm performance. If the number of analyst followng plays a dscplnary role on msbehavng managers and rewards value creaton, the presence of nternal governance mechansms should lower the margnal beneft of more analyst followng, and vce versa. Hence, we antcpate a negatve jont mpact of analyst followng and corporate governance on valuaton. Our second hypothess s stated as follows: H 2 : The relaton between the qualty of corporate governance and valuaton s less pronounced for frms wth greater analyst followng than for frms wth lower analyst followng. 6
3. Data and Sample Our sample of companes comprses the FTSE All Share consttuent frms. Corporate governance data s collected from the largest corporate governance data provder to nsttutonal nvestors, Corporate Governance Quotent (CGQ) suppled by RskMetrcs Group. CGQ provdes corporate governance rankng for more than 7,500 frms worldwde snce 2003. Therefore, we our sample perod starts wth 2003. Year 2008 s the most recent year at the tme of undertakng the analyss of the present paper. Ths database contans 55 CGQ governance factors whch span eght categores of corporate governance ncludng board, compensaton, takeover and audt. Table 1 provdes a summary of corporate governance ratngs varables provded by RskMetrcs.. Insert Table 1 here Analyst followng data s collected from FactSet, a unque dataset provded by FactSet Research Systems Inc., whch provdes analytcs to nvestment professonals to download, combne, and manpulate fnancal data for nsght nto global market ncludng UK stock market. We merge the analyst followng data wth corporate governance data usng companes Sedol numbers. We merge 2852 frm-year observatons for the number of analyst followng wth 2514 frm-year observatons from QCG usng Sedol numbers. Ths leads to an ntal sample of 1577 frm-year observatons. To test hypothess H 1, we use the natural logarthm of total assets as a proxy for frm sze. We collect total assets from Worldscope (tem no. WC02999). We remove frms wth mssng total assets data, ths ends up wth a sample of 1514 frm-year observatons (see Panel A n Table 2). To test hypothess H 2, we collect accountng data from Worldscope. In partcular, we calculate Tobn s Q as total assets (Worldscope tem WC02999) less the book value of equty (Worldscope tem WC05476) multpled by the number of shares outstandng plus the market value of equty (MV) dvded by total assets. We use Total debt (Worldscope tem WC03255) to compute total Debt/total Assets rato (DA) and Captal expendture (Worldscope tem WC04601) to compute Captal expendture/total Assets rato (CA). We remove frms wth mssng accountng data, ths leads to a sample of 1505 frm-year observatons (see Panel B n Table 2). Insert Table 2 here 7
4. Research Desgn 4.1 Models employed to test H 1 To test H 1, we employ the followng model: Followng a0 a1cg a2sze (1) Where: Followng s defned as the number of analysts followng UK frms collected from FactSet dataset. CG s the aggregated CG rankng provded by CGQ. Sze s the natural logarthm of total assets CGQ also provdes sub-scores for each frm n four partcular governance areas; board, takeover defences, executve and drector compensaton and ownershp, and audt. These subs-cores are expressed n numbers from 1-5 (5 ndcates frm n the top quntle n a governance area whle 1 ndcates frm s n the bottom quntle n a governance area). In order to understand whch of these components better explans the number of analyst followng, we run model (1) by replacng the overall rankng of corporate governance wth sub-scores pertanng to dfferent aspects of corporate governance. We expect a postve sgn wth each ndvdual corporate governance varable. That s we expect b 1 through b 4 to be postve n the followng model: Followng b0 b1 Board b2compensaton b3takeover b4 Audt b5sze (2) Where: Followng s defned as the number of analysts followng UK frms collected from FactSet dataset. Board s the Board Composton sub-score Composton s the Executve and Drector Composton sub-score Takeover s the Ant-Takeover Provsons sub-score Audt s Audt commttee/audt fees/audt Rotaton/Audtor Ratfcaton sub-score Sze s the natural logarthm of total assets Moreover, we construct another aggregate measure of corporate governance (NCG) from the addton of the four sub-scores (Board+Compensaton+Takeover+Audt) and re-run regresson (1) as follows: Followng c0 c1ncg c2sze (3) 8
We control for frm sze n the above three models usng the natural logarthm of total assets (n mllons). Based on pror lterature (e.g., Lang et al. 2003, 2004), we expect a postve sgn for Sze. 4.2 Models employed to test H 2 Wth a slght modfcaton to Lang et al. (2004) to accommodate corporate governance varables, we run the followng regresson model: Tobn Q d0 d1cg d2followng d3cgfollowng d4sze d5da d6ca (4) Tobn s Q s used to proxy for the market valuaton of assets and s computed as total assets less the book value of equty plus the market value of equty dvded by total assets. CG s measured by the corporate governance rankng publshed by QCG. Analyst followng (Followng) s defned as n equaton (1). In model 4, we use three control varables. Sze s the natural logarthm of total assets. Sze s expected to be postvely related to Tobn s Q. To control for the possblty that credtors are able to lessen manageral agency problems or that debt provdes valuable tax shelds, we use the rato of total debts to total assets (DA). We control for frm s potental nvestment opportuntes usng the rato of captal expendture to total assets (CA). We expect both DA and CA to have a postve sgn (Lang et al., 2004). We use another varable, (NCG), as a proxy for the qualty of corporate governance whch s the sum of the four dmensons (Board, Followng, Compensaton, and Audt), and we re-run the followng new model: Tobn Q h0 h1 NCG h2followng h3 NCGFollowng h4sze h5 DA h6ta (5) We further partton NCG nto ts components and use each of ths component one at a tme nstead of NCG n equaton (5), to examne whch of these components has more mpact on frm value. The results are reported n columns (7-13) of Table 8. We expect these components to have a postve mpact on frm value. Based on hypothess H 2, we expect the coeffcents of both d 3 and h 3 to be negatve and sgnfcant. 9
5. Emprcal Results 5.1 Analyst followng and corporate governance Table 3 provdes summary statstcs of the varables used n the regresson models. The number of analyst followng ranges from the mnmum of 0 to the maxmum of 43 wth a medan of 9 whch ndcates that 50% of frms have less than 9 analysts followng. The Corporate governance rankng (CG) has a hgh average of 85.664 and a medan of 87.8 whch shows that majorty of frms n the sample are of hgh corporate governance qualty whch s expected gven that all frms n CGQ are ncluded n MSCI EAFE ndex. The second measure of corporate governance (NCG) ranges from 6 to 20 wth a medan of 18 whch agan confrms the lower cross sectonal varaton n corporate governance qualty due to sample bas to hgh qualty corporate governance frms. The four sub-scores of corporate governance ranges from 1 to 5 wth a medan of 5 for Board, Takeover, and Audt and a medan of 4 for compensaton. Ths ndcates a hgher level of cross sectonal varaton n Compensaton compared to the three remanng components of corporate governance. Insert Table 3 here Table 4 shows the correlaton coeffcents between the varables used n the regresson models (1, 2 and 3). It can be notced that there are sgnfcant relatonshp between all the varables n the regresson models apart from the relatonshp between Compensaton and both Takeover and Sze. Insert Table 4 here Tables 5 & 6 show the emprcal results related to hypothess 1. Table 5 reports the results wthout usng year dummes; whle Table 6 shows the results after the ncluson of year dummes n the regresson models. The results reported n columns (1-4) show that there s postve mpact for the qualty of corporate governance on analyst followng wth a coeffcent of 0.701 on NCG whch s sgnfcant at 1% level of sgnfcance. Addng sze to the model n column (4) ncreases the explanatory power of the model as measured by adjusted R-squared from 3.34% to 33.75% but reduces the coeffcent on NCG to 0.419 although t remans sgnfcant at 1% level of sgnfcance. Columns (5-8) show that corporate governance components have a sgnfcant mpact on analyst followng but only negatve n the case of Takeover. When regress all corporate governance components on analysts followng (model 10), we fnd that Board, compensaton, and takeover are 10
sgnfcant whle Audt s the only nsgnfcant. When Addng Sze to the model n column (11) only Board and compensaton reman sgnfcant and the explanatory power of the model notceably ncreased to 33.98%. Gven the sgnfcant correlaton between Board and Sze, model 13 suggests Compensaton and Sze are the man factors that determne the number of analysts followng. Ths fndng mples that analysts are partcularly concerned about how much compensaton that executves and drectors receve. The results n Table 5 are consstent wth the correlaton analyss n Table 4..The postve assocaton between analyst followng and corporate governance qualty n the UK s consstent wth Yu (2007, 2010) fndngs for common law countres whch UK belongs to. Ths ndcates that analysts play a more actve role n countres wth stronger nvestor protecton and more effectve legal systems (Chang et al. (2000); Barnve et al. (2005)). Insert Tables 5 & 6 here 5.2 Analyst followng, corporate governance and frm value Table 7 shows descrptve statstcs of the varables used to test hypothess 2. The average number of analysts followng a frm s 10.223. The frm wth the largest analyst followng s covered by 43 analysts and the frm wth the smallest number of analyst followng s covered by zero analysts. The mean of corporate governance rankng (CG) s 85.682 wth a standard devaton of 11.754. These rankngs range from 0 to 100, wth a medan of 87.9, the 25 th percentle of 80, and the 75 th percentle of 94.2 whch ndcates that the sample s domnated wth hgh qualty corporate governance. Ths s expected gven that the sample s covered by MSCI EAFE ndex. The mean of Tobn s Q s 0.501, wth a standard devaton of 2.123. It ranges from -45.808 to 3.491, wth a medan of 0.642, the 25 th percentle of 0.450 and the 75 th percentle of 0.775. Insert Table 7 here Table 8 provdes the correlaton coeffcents among the varables. Frst, CG s postvely assocated wth Tobn s Q whle s NCG s negatvely assocated wth Tobn s Q and the same apples for ts four components apart from Audt. Sze and DA are the only varables wth sgnfcant relaton wth Tobn s Q. Insert Table 8 here 11
Column 2 n Table 9 shows that CG and Followng come wth the expected postve sgn but nsgnfcant coeffcents. Ths postve relatonshp s consstent wth Lang et al. (2003), Klapper and Love (2004), Lang et al. (2004). The results from corporate governance components n columns (11) and (13) are smlar. In the models wthout nteracton terms (n unreported results), Followng has a sgnfcant mpact on frm valuaton but of the wrong sgn (negatve). Moreover, Sze s the only varable that s persstently sgnfcant n all models. Table 9 provdes emprcal test on the conjecture that the postve market valuaton effect of corporate governance s condtonal on the number of analyst followng. The results n column (2) and (4) support H 2. As expected, the nteracton term CGFollowng and NCGFollowng has a negatve coeffcent but t s nsgnfcant. Ths s nconsstent wth Yu (2007) who fnd postve but stll nsgnfcant coeffcent. Insert Table 9 here 6. Concluson Our paper bulds on lterature that nvestgates the lnk between corporate governance, analysts followng and frm value. We extend ths work n four mportant ways. Frst, we nvestgate the assocaton between corporate governance mechansms and the number of analysts followng UK frms. In partcular, we examne the extent to whch better corporate governance qualty attracts more fnancal analysts for coverng UK frms. Second, we try to dentfy whch component of corporate governance s more lkely to affect analysts followng. Thrd, we test the jont effect of corporate governance mechansms and analysts followng on UK frm value. Fnally, we study the effect of dfferent corporate governance sub-categores on frm value n the presence of analysts followng. The paper uses a sample of 1514 UK frms lsted n London Stock Exchange. It fnds that the overall level of corporate governance qualty s postvely assocated wth the number of analysts followng UK frms. Among the four ndvdual components of corporate governance, we fnd that compensaton score seems to be the man component that affects analyst followng. Our paper also shows that both better corporate governance qualty and greater analyst followng have no sgnfcant effect n enhancng UK frm value. Our results 12
add to the understandng of the expected role of fnancal analysts as addtonal montorng mechansm n the stock market. An nterestng ssue for future work would be to examne the extent to whch the qualty of corporate governance mechansms affect fnancal analysts stock recommendatons (.e. buy, sell or hold) and ther earnngs forecasts accuracy. In addton, t s worth to nvestgate the extent to whch fnancal analysts actually use corporate governance nformaton n reports. Usng a content analyss approach, researchers can dentfy whch CG mechansms are reported n analysts reports and the extent to whch these mechansms are mportant for the stock market partcpants (.e. current and future nvestors). 13
References Attg, N., Fong, W., Gadhoum, Y. and Lang, L. (2006). Effects of large shareholdng on nformaton asymmetry and stock lqudty. Journal of Bankng and Fnance 30, 2875-2892. Barnv, R., Myrng, M., and Thomas. W. (2005). The assocaton between the legal and fnancal reportng envronments and forecast performance of ndvdual analysts. Contemporary Accountng Research 22 (December): 727-758. Basddq, H. and Hussaney, K. (2010). Is asymmetrc nformaton assocated wth UK dvdend polcy? Journal of Appled Accountng Research, forthcomng. Boubaker, S. and Labegorre, F. (2008). Ownershp structure, corporate governance and analyst followng: A study of French lsted frms. Journal of Bankng & Fnance, 32(6): 961-976. Bowen, R., Chen, X. and Qang, C. (2008). Analyst followng and the cost of rasng equty captal: evdence from underprcng of seasoned equty offerngs. Contemporary Accountng Research, 25 (3): 641 946. Brown, L. and Caylor, M. L. (2004). Corporate governance and frm performance, Workng paper, Georga. State Unversty, Unted State of Amerca. Avalable at SSRN: http://ssrn.com/abstract=586423 or do:10.2139/ssrn.586423. Chang, J., Khunna, T. and Palepu. K. (2000). Analyst actvty around the world. Workng paper. Harvard Busness School. Claessens, S., Djankov, S., Fan, J. P. H. and Lang, L. H. P. (2002). Dsentanglng the ncentve and entrenchment effects of large shareholdngs The Journal of Fnance 57 (6): 2741-2771. Durnev, A., and Km, H. (2005). To steal or not to steal: frm attrbutes, legal envronment, and valuaton. The Journal of Fnance 60: 1461-1493. Fan, J., and Wong, T. (2002). Corporate ownershp structure and the nformatveness of accountng earnngs n East Asa. Journal of Accountng and Economcs 33: 401-425. FRC (Fnancal Reportng Councl), (2006). The UK Corporate Governance Code, June. FRC (Fnancal Reportng Councl), (2010). The UK Corporate Governance Code, June. Haw, I., Hu, B and Hwang, L. and Wu, W. (2004). Ultmate ownershp, ncome management, and legal and extra-legal nsttutons, Journal of Accountng Research, 42(2): 423 462. Hussaney, K. and Al-Najjar, B. (2010). Future-orented nformaton: determnants and use. Journal of Appled Accountng Research, forthcomng. 14
Hussaney, K. and Wang, M. (2010). The mpact of corporate governance on future-orented dsclosures: A large-scale UK study. Workng paper, Strlng Unversty, UK. Klapper, L., and Love., I. (2004). Corporate governance, nvestor protecton, and performance n emergng markets. Journal of Corporate Fnance 10: 703-728. Knyazeva, D. (2007). Corporate governance, analyst followng and frm behavor. Workng paper. New York Unversty. Lang, M. H., Lns, K. V.and Mller, D. P. (2003). ADRs, analysts, and accuracy: does cross lstng n the Unted States mprove a frm's nformaton envronment and ncrease market value? Journal of Accountng Research 41(2): 317-345. Lang, M. H., Lns, K. V.and Mller, D. P. (2004). Concentrated control, analyst followng and valuaton: do analysts matter most when nvestors are protected least? Journal of Accountng Research 42: 589-622. L, J., Pke, R. and Hanffa, R. (2008), 'Intellectual captal dsclosure and corporate governance structure n UK frms', Accountng and Busness Research, 38 (2): 137-159. McNchols, M., and O Bren, P. (1997). Self-selecton and analyst followng. Journal of Accountng Research 35 (Supplement): 167-199. Wllekens, M, Bauwhede, H., Gaeremynck, A. and Gucht, L. (2005). The mpact of nternal and external governance mechansms on the voluntary dsclosure of fnancal and nonfnancal performance. Workng paper. Aston Busness School (UK). Yu, M. (2007). Analyst actvty and corporate governance: a global perspectve, PhD Dssertaton, Kent State Unversty, USA. Yu, M. (2010). Analyst followng and corporate governance: emergng-market evdence, Accountng Research Journal, 23 (1): 69-93. 15
Table 1: Ratngs Varables Summary Board Ant-Takeover Provsons 1 Board Composton 34 Ant-Takeover Provsons Applcable Under Country (local) Laws 2 Nomnatng Commttee Executve and Drector Compensaton 3 Compensaton Commttee 35 Cost of Opton Plans 4 Governance Commttee 36-37 Opton Re-prcng 5 Board Structure 38 Shareholder Approval of Opton Plans 6 Board Sze 39 Compensaton Commttee Interlocks 7 Changes In Board Sze 40 Drector Compensaton 8 Cumulatve Votng 41 Penson Plans For Non-Employee Drectors 9 Boards Served On CEO 42 Opton Expensng 10 Boards Served On Other Than CEO 43 Opton Burn Rate 11 Former CEO s 44 Corporate Loans 12 Charman/CEOs Separaton Progressve Practces 13 Board Gudelnes 45 Retrement Age for Drectors 14 Response To Shareholder Proposals 46 Board Performance Revews 15 Board Attendance 47 Meetngs of Outsde Drectors 16 Board Vacances 48 CEO Successon Plan 17 Related Party Transactons 49 Outsde Advsors Avalable to Board Audt 50 Drectors resgn upon job change 18 Audt Commttee Ownershp 19 Audt Fees 51 Drector Ownershp 20 Audtor Rotaton 52 Executve Stock Ownershp Gudelnes 21 Audtor Ratfcaton 53 Drector Stock Ownershp Gudelnes Charter/Bylaws 54 Offcer and Drector Stock Ownershp 22-27 Features of Poson Plls Drector Educaton 28-29 Vote Requrements 55 Drector Educaton 30 Wrtten Consent 31 Specal Meetngs 32 Board Amendments 33 Captal Structure Source: RskMetrcs Group - Corporate Governance Quotent (20 th December 2007) 16
Table 2: Sample selecton Panel A: Sample selecton for H1 No. of observatons Intal sample from FactSet 2852 Less: mssng data on QCG 1275 Equals: 1577 Less: Mssng data for accountng varables 63 Fnal Sample 1514 Panel B: Sample selecton for H2 No. of observatons Intal sample from FactSet 2852 Less: mssng data on QCG 1275 Equals: 1577 Less: Mssng data for accountng varables 72 Fnal Sample 1505 17
Table 3: Descrptve Statstcs for H1 Varable Mean Std dev Mn 1 st quartle Medan 3 rd quartle Max Followng 10.167 6.846 0 5 9 14 43 CG 85.664 11.747 0 80.8 87.8 94.2 100 NCG 17.767 1.801 6 17 18 19 20 Board 4.399 0.708 1 4 5 5 5 Compensaton 4.099 0.999 1 4 4 5 5 Takeover 4.782 0.653 1 5 5 5 5 Audt 4.487 0.749 1 4 5 5 5 Sze 14.082 1.938 10.146 12.693 13.832 15.073 21.596 Based on 1514 observatons 18
Table 4: Correlaton Matrx Followng CG NCG Board Compensaton Takeover Audt Sze Followng 1 0.210* 0.184* 0.259* 0.100* -0.101* 0.152* 0.571* CG 1 0.712* 0.608* 0.391* 0.156* 0.480* 0.189* NCG 1 0.628* 0.655* 0.317* 0.660* 0.133* Board 1 0.147* -0.088* 0.446* 0.293* Compensaton 1-0.022 0.123* 0.016 Takeover 1 0.001-0.166* Audt 1 0.166* Sze 1 *** = sgnfcant at the 1% level; ** = Sgnfcant at the 5% level; *=Sgnfcant at the 10% level Table 5: The assocaton between corporate governance and analyst followng (H1) Varable Expected Sgn (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) Intercept? -0.303-22.519* -2.288-24.971* -0.865 7.349* 15.206* 3.939* 12.340* 1.116-23.074* -18.249* -20.730* CG + 0.122* 0.061* NCG + 0.701* 0.419* Board + 2.508* 2.161* 0.792* Compensaton + 0.688* 0.673* 0.413** 0.530* 0.624* Takeover + -1.054* -1.031* -0.835* -0.017 Audt + 1.388* 0.412 0.145 Sze + 1.948* 1.966* 1.919* 2.018* 2.013* Adj-R 2 4.34% 33.63% 3.34% 33.75% 6.67% 0.94% 0.95% 2.25% 1.85% 7.65% 33.98% 32.60% 33.38% N 1514 1514 1514 1514 1514 1514 1514 1514 1514 1514 1514 1514 1514 *Wthout year dummy control varables (NCG s the sum of board, compensaton, takeover, and audt), dependent varable s Analyst followng. *** = sgnfcant at the 1% level; ** = Sgnfcant at the 5% level; *=Sgnfcant at the 10% level 19
Table 6: The assocaton between corporate governance and analyst followng (H1) Varable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) Intercept 1.245-23.132* -1.247-25.286* -0.034 8.582* 12.594* 5.145* 10.301* -0.020-24.343* -19.027* -21.339* CG 0.112* 0.058* NCG 0.693* 0.397* Board 2.302* 2.105* 0.763* Compensaton 0.564* 0.576* 0.361** 0.463* 0.544* Takeover -0.410-0.455-0.534*** 0.079 Audt 1.238* 0.322 0.123 Sze 1.942* 1.951* 1.923* 2.002* 2.001* Year04 2.647* 2.852* 3.054* 3.058* 3.071* 3.083* 3.100* 3.077* 3.088* 3.068* 3.061* 3.081* 3.070* Year05-0.539 1.667** -0.734 1.582** 0.217-0.617-0.355-0.297-0.130 0.937 1.956* 1.611** 1.776* Year06-1.237 0.815-1.495*** 0.691-0.517-1.486** -1.076-0.824-1.074 0.026 0.953 0.774 0.735 Year07-1.808** -0.033-1.673** 0.065-0.894-1.569** -1.559-1.626** -1.073-0.047 0.521-0.072 0.339 Year08-0.451 1.322** -1.190 0.920 0.201-0.684-0.216-0.6689-0.206 0.654 1.367*** 1.280** 1.243*** Adj-R 2 6.47% 34.80% 5.96% 34.87% 8.26% 3.42% 2.92% 4.50% 3.49% 8.58% 34.95% 33.90% 34.42% N 1514 1514 1514 1514 1514 1514 1514 1514 1514 1514 1514 1514 1514 Wth year dummy control varables. *** = sgnfcant at the 1% level; ** = Sgnfcant at the 5% level; *=Sgnfcant at the 10% level 20
Table 7: Descrptve statstcs for H2 Varable Mean Std dev Mn 1 st quartle Medan 3 rd quartle Max followng 10.223 6.829 0 5 9 14 43 CG 85.682 11.754 0 80 87.9 94.2 100 NCG 17.768 1.802 6 17 18 19 20 Board Compensaton Takeover Audt 4.398 0.708 1 4 5 5 5 4.103 0.996 1 4 4 5 5 4.781 0.655 1 5 5 5 5 4.486 0.750 1 5 5 5 5 sze 14.082 1.940 10.146 12.693 13.830 15.067 21.597 DA 0.237 0.190 0 0.083 0.219 0.349 1.672 CA 0.050 0.056 0 0.015 0.037 0.065 0.860 Q 0.501 2.123-45.808 0.450 0.642 0.775 3.491 Based on 1505 observatons 21
Table 8: Correlaton Matrx Followng CG NCG Board Compensaton Takeover Audt Sze DA CA Q Followng 1 0.209* 0.185* 0.263* 0.097* -0.098* 0.154* 0.575* 0.065** -0.063** -0.015 CG 1 0.711* 0.608* 0.390* 0.157* 0.480* 0.191* 0.094* -0.012 0.001 NCG 1 0.628* 0.656* 0.318* 0.660* 0.135* 0.074* -0.049*** -0.013 Board 1 0.149* -0.0883* 0.444* 0.293* 0.111* -0.075* -0.004 Compensaton 1-0.020 0.126* 0.021 0.040 0.007-0.017 Takeover 1 0.001-0.166* -0.049*** 0.002-0.013 Audt 1 0.165* 0.062** -0.058** 0.007 Sze 1 0.195* -0.105* 0.055** DA 1 0.154* 0.056** CA 1 0.009 Q 1 *** = sgnfcant at the 1% level; ** = Sgnfcant at the 5% level; *=Sgnfcant at the 10% level 22
Table 9: The assocaton between corporate governance, analyst followng and frm value Varable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) Intercept -0.664-0.939-0.440-1.003-0.462-0.572-0.779-0.619-0.552-0.687-1.084-0.750-1.123 CG -0.001 0.002-0.018 Followng -0.021** 0.011-0.020** 0.040-0.020** -0.020** 0.005-0.020** -0.027-0.021** 0.011-0.021** 0.022 CGFollowng -0.001 NCG 0.014 NCGFollowng -0.003 Board -0.062-0.053-0.006 Boardfollowng -0.006 Compensaton -0.026-0.029-0.045 compfollowng 0.002 Takeover -0.015-0.010 0.070 Takefollowng -0.009 Audt 0.029 0.002 0.086 Audtfollowng -0.009 Sze 0.095* 0.095* 0.095* 0.095* 0.096* 0.098* 0.098* 0.094* 0.093* 0.094* 0.094* 0.094* 0.094* DA 0.480 0.481 0.487 0.482 0.492 0.490 0.485 0.482 0.482 0.475 0.484 0.475 0.459 CA 0.290 0.282 0.264 0.289 0.265 0.252 0.270 0.292 0.290 0.289 0.292 0.292 0.308 Adj-R 2 0.5% 0.45% 0.52% 0.48% 0.35% 0.52% 0.47% 0.51% 0.45% 0.50% 0.46% 0.49% 0.48% N 1505 1505 1505 1505 1505 1505 1505 1505 1505 1505 1505 1505 1505 *** = sgnfcant at the 1% level; ** = Sgnfcant at the 5% level; *=Sgnfcant at the 10% level 23