Corporate Governance and Institutional Ownership

JOURNAL OF FINANCIAL AND QUANTITATIVE ANALYSIS Vol.46,No.1,Feb.2011,pp.247–273 COPYRIGHT2011,MICHAEL G.FOSTER SCHOOL OF BUSINESS,UNIVERSITY OF WASHINGTON,SEATTLE,WA98195 doi:10.1017/S0022109010000682Corporate Governance and Institutional OwnershipKee H.Chung and Hao Zhang∗AbstractIn this study we examine the relation between corporate governance and institutional own-ership.Our empirical results show that the fraction of a company’s shares that are held by institutional investors increases with the quality of its governance structure.In a similar vein,we show that the proportion of institutions that hold afirm’s shares increases with its governance quality.Our results are robust to different estimation methods and alternative model specifications.These results are consistent with the conjecture that institutional in-vestors gravitate to stocks of companies with good governance structure to meetfiduciary responsibility as well as to minimize monitoring and exit costs.I.IntroductionCompanies use a broad set of tools to attract more investors to their stocks in the hope that a greater investor base raises share price and increasesfirm value.1 For instance,the board of directors of Ciena Corporation approved a reverse stock split in2006based on the belief that the resulting higher share price would attract more institutional investors.2Our study explores another important mechanism that can affect the investor base of a company.Specifically,we analyze the role of corporate governance as a means to attract institutional investors.Althoughfield∗Chung,keechung@,School of Management,State University of New York at Buffalo, 360Jacobs Management Center,Buffalo,NY14260;Zhang,hzhang@,Saunders College of Business,Rochester Institute of Technology,105Lomb Memorial Dr.,Rochester,NY 14623.The authors are solely responsible for the content.The paper benefitted greatly from many valuable comments and suggestions of Hendrik Bessembinder(the editor)and Paul Schultz(associate editor and referee).The authors also thank Jiekun Huang,Jing Jiang,Kenneth A.Kim,Choonsik Lee,Fang Liu,Carl Hsin-han Shen,and Lingyin Zhu for valuable comments and discussion,and Institutional Shareholder Services(ISS)for providing corporate governance data.1Merton(1987)suggests that an increase in thefirm’s investor base lowers investors’expected returns and increases the market value of its shares.2Many companies explicitly state that the main goal of their reverse splits is to attract more in-stitutional investors.See /Site/News Events/pr content.asp?news id=954 &cat id=1and /portal/site/google/?ndmViewId=news view&newsId= 20080922005721&newsLang=en.247248Journal of Financial and Quantitative Analysissurvey results indicate that corporate governance has become an important invest-ment criterion of institutional investors,3there is only limited empirical evidence on the issue.In this study,we shed further light on the relation between insti-tutional ownership and corporate governance using comprehensive measures of governance quality.Corporate governance has been a subject of numerous studies during the last2decades.Williamson(1985)and Grossman and Hart(1986)hold that man-agerial opportunism reduces the amount of capital that investors are willing to contribute to thefirm.Shleifer and Vishny(1997)note that corporate governance deals with various constraints that either managers put on themselves or investors put on managers to reduce the agency problem.Gompers,Ishii,and Metrick (2003)show that better corporate governance leads to greaterfirm values and higher stock returns.Chung,Elder,and Kim(2010)show that better governance results in higher stock market liquidity.Badrinath,Kale,and Ryan(1996),Falkenstein(1996),and Huang(2009) show that institutional investors prefer stocks that have higher market liquidity and lower return volatility.Other studies show that institutional investors prefer stocks of companies with better disclosure(Bushee and Noe(2000)),stocks of larger companies(Gompers and Metrick(2001)),stocks of companies that pay cash dividends or repurchase shares(Grinstein and Michaely(2005)),and stocks of companies with better managerial performance(Parrino,Sias,and Starks (2003)).None of these studies examines the effect of corporate governance on institutional ownership.Dahlquist,Pinkowitz,Stulz,and Williamson(2003),Giannetti and Simonov (2006),Ferreira and Matos(2008),and Leuz,Lins,and Warnock(2009)an-alyze the effect of ownership structure on institutional investors’stock selec-tion decisions.Dahlquist et al.find no relation between the ratio of control to cashflow rights and the holdings of foreign investors.Giannetti and Simonov show that both foreign and domesticfinancial institutions are reluctant to hold shares of companies that have high control to cashflow rights ratios of prin-cipal shareholders.Ferreira and Matos show that institutions hold fewer shares of companies that have more closely held ownership structure.Leuz et al.find that U.S.institutions invest less in foreignfirms with large insider block own-ership.Li,Ortiz-Molina,and Zhao(2008)show that institutions avoid investing in companies with dual-class shares.Because these studies focus on only one dimension of corporate governance(i.e.,ownership structure),however,they of-fer limited insight on the relation between corporate governance and institutional ownership.Bushee,Carter,and Gerakos(2010)analyze whether institutional investors tilt their portfolios towardfirms with preferred governance mechanisms.The au-thors conclude that although institutional investors have incentives to tilt their 3Survey results indicate that institutional investors prefer companies with good governance struc-ture.McKinsey&Company(2002)surveyed more than200institutional investors in31countries and showed that institutional investors put corporate governance quality on a par withfinancial indicators when evaluating investment decisions.Chung and Zhang249 portfolios towardfirms with better governance mechanisms,there is no signifi-cant relation between institutional ownership and corporate governance.4 McCahery,Sautner,and Starks(2010)conduct a survey to elicit institutional investors’views on country-level investor protection andfirm-level corporate gov-ernance mechanisms.Theyfind that among the institutions that responded to the survey,corporate governance is important to their investment decisions,and a number of them are willing to engage in shareholder activism(e.g.,80%of the institutions are willing to vote with their feet by selling their shares).They also show that the preferences for governance mechanisms vary across the institutional investor types.We shed further light on the behavior of institutional investors by analyz-ing the relation between corporate governance and institutional ownership.In contrast to prior research that relied on a single governance attribute,our study employs comprehensive measures of corporate governance quality that are con-structed from50governance factors for a large number of panies.Our study provides answers to the following questions:Do institutional investors pre-fer stocks of companies that have better governance structure?If so,does the institutional investors’preference toward better-governed companies vary across different types of institutions?What kinds of corporate governance provisions are most attractive to institutional investors?Does the relation between cor-porate governance and institutional ownership vary with thefirm’s information environment?The ordinary least squares(OLS)regression results show that the percent-age of afirm’s shares that are held by institutional investors is positively and significantly related to its governance quality across all types of institutional in-vestors.The relation between institutional ownership and corporate governance is economically significant:An increase in governance quality from the25th to 75th percentile results in an increase in institutional ownership by14%.We show that the governance provisions that are most effective in attracting institutional investors concern either the composition/operation of the board of directors or provisions that are designed to strengthen shareholder rights.The importance of corporate governance quality in institutional investors’stock selection decisions is smaller for those stocks that are followed by more analysts and have lower information asymmetry.We perform a battery of robustness tests andfind that our results are not sensitive to different estimation methods,different proxies of governance quality,or additional control variables.To determine whether the positive relation between governance quality and institutional ownership is driven by reverse causality(i.e.,institutional investor activism causesfirms to adopt better governance mechanisms),we also employ the2-stage least squares(2SLS)method using instrumental variables that are re-lated to governance quality measures,but unlikely to be correlated with residuals in the2nd-stage regression.The results show that the coefficients on the instru-mented governance quality measures are statistically significant and positive in 4Bushee et al.(2010)use corporate governance data provided by the Investor Responsibility Research Center(IRRC)that mainly cover corporate takeover defense strategies.As explained later, our data cover much broader arrays of corporate governance.250Journal of Financial and Quantitative Analysisthe2nd-stage regressions,indicating that the positive relation between institu-tional ownership and corporate governance is driven,at least in part,by institu-tional preference toward companies with better governance structure.The paper is organized as follows.Section II explains why institutional own-ership is likely to increase with corporate governance quality.Section III describes data sources and presents descriptive statistics.Sections IV and V present empir-ical results.Section VI provides a brief summary and concluding remarks.II.Institutional Ownership and Corporate Governance As in prior research,we define institutional ownership as the fraction of a firm’s shares that are held by institutional investors.Hence,by definition,institu-tional ownership of a company is1–the fraction of its shares held by noninsti-tutions(i.e.,individual investors).If we were to examine institutional investors’preference toward a certain group of companies,we would need to consider why institutional investors’preference for those companies is likely to be greater than that of individual investors.Below we present our conjectures on why institu-tional investors’gravitation toward stocks of companies that have better gover-nance structure is likely to be stronger than that of individual investors.Institutional investors(e.g.,banks,insurance companies,and pension funds) have strongfiduciary responsibilities.Del Guercio(1996)shows that many insti-tutional investors tilt their portfolios to stocks that are viewed as prudent invest-ments.Grinstein and Michaely(2005)suggest that institutions avoidfirms that do not pay dividends,because a“prudent”stock should have a history of stable dividend payments.Because investors in poorly governed companies are likely to face large risks of expropriation and other self-dealing problems,they may not earn fair rates of return or may even fail to preserve the invested capital(i.e.,fail to meet the prudent-person rule).Hence,institutional investors’fiduciary respon-sibilities give them a strong incentive to choose stocks of companies with good governance structure.5Despite the free-rider problem,institutional investors have a much stronger incentive to monitor companies that they own than do individual investors be-cause of their larger stakes in those companies,especially if exit is costly(i.e., large trading costs).Bushee and Noe(2000)suggest that institutional investors preferfirms with better disclosure rankings to reduce monitoring costs.To the extent that companies that have better governance structures require less outside monitoring,institutional investors are likely to prefer companies with better gov-ernance mechanisms to those with poor governance mechanisms.Institutional investors are more likely to prefer stocks with high secondary market liquidity than are individual investors because the price impact of trades has greater consequences to institutional investors.The price impact has greater consequences to institutional investors than to individual investors because in-stitutional investors typically trade in larger quantities than individual investors.5Hawley and Williams(2000)provide evidence that thefiduciary duty of institutional investors facilitates their predisposition to companies with good governance mechanisms.Chung and Zhang251 Chung et al.(2010)argue that good governance improvesfinancial and opera-tional transparency and thus reduces information asymmetry between insiders and outside investors.Theyfind thatfirms with better corporate governance exhibit higher stock market liquidity and lower trading costs(e.g.,narrower quoted and effective spreads as well as the smaller price impact of trades).This is another rea-son why institutional investors are more likely to prefer stocks of better-governed companies than individual investors.In short,we conjecture that institutional investors are more likely to prefer stocks of better-governed companies than individual investors are because better-governed companies are likely to require less monitoring,have higher stock mar-ket liquidity,and more easily meetfiduciary responsibilities.In what follows, we test our conjecture by examining whether afirm’s institutional ownership increases with its governance quality.III.Data Sources,Variable Definition,and Descriptive StatisticsOur initial study sample includes all stocks listed on the New York Stock Exchange(NYSE),American Stock Exchange(AMEX),or NASDAQ between January2001and December2006.We obtain stock-market-related data from both the Center for Research in Security Prices(CRSP)database and the NYSE’s Trade and Quote(TAQ)database,accounting data from Standard&Poor’s(S&P’s) Compustat,analyst coverage data from the Institutional Brokers’Estimate Sys-tem(IBES),institutional holding(13f)data from the Thomson Reuters Owner-ship Database,and corporate governance data from ISS.We exclude observations with missing variables and Winsorize Compustat-based and13f-based variables at the1st and99th percentiles to reduce the influence of extreme observations and possible data errors.We excludefinancial companies(Standard Industrial Clas-sification(SIC)codes6000–6999)and utilities(SIC codes4900–4949)from the study sample.Ourfinal sample consists of12,093firm-year observations.We limit our study period to2001–2006because the ISS corporate governance data are available only for this period.We measurefirm i’s institutional ownership in year t(INST OWNERSHIP i,t) by the ratio of the number of shares held by institutional investors to the total number of shares outstanding.Similar to the coding methods used in Brown and Caylor(2006),(2009)and Chung et al.(2010),we calculate governance scores for eachfirm by assigning1point for each governance provision that meets the minimum standard provided in ISS Corporate Governance:Best Practices User Guide and Glossary(2003).Brown and Caylor(2006)construct their governance index using51governance standards(out of61)included in the ISS database. Chung et al.(2010)select24governance standards that are most closely re-lated tofinancial and operational transparency.Following Chung et al.,we ex-clude the“Director Education”category.Although Brown and Caylor(2006)omit firms with dual-class shares,we include the dual-class standard in our governance index construction because Li et al.(2008)show that it is an important determi-nant of institutional investment decisions.In this way,we construct our gover-nance score,GOV SCORE1i,t,with the maximum value of50and the minimum252Journal of Financial and Quantitative Analysisvalue of0.The Appendix presents the list of governance categories and stan-dards that are used in our study.To assess the sensitivity of our results to differ-ent measures of governance quality,we also construct the2nd governance score, GOV SCORE2i,t,using only the36governance standards that are included in the ISS database for our entire study period,from2001to2006.6We also calculate the following variables(forfirm i in year t)that are in-cluded in our regression analyses:the market value of equity(MVE i,t),the stan-dard deviation of quote-midpoint daily returns(VOLATILITY i,t),the average ra-tio of monthly trading volume to the number of shares outstanding(TURNOVER i,t), the average stock price(PRICE i,t),the continuously compounded annual stock return(RETURN i,t),the average effective bid-ask spread(BASPREAD i,t),the number of years since thefirmfirst appeared in the CRSP database(FIRM AGE i,t), the ratio of total debt to the book value of total assets(LEVERAGE i,t),the annual dividend yield(D YIELD i,t),the ratio of net income to the book value of total assets(ROA i,t),Tobin’s q ratio(TOBIN’S Q i,t)calculated using the method in Chung and Pruitt(1994),and the asset tangibility ratio(TANGIBILITY i,t)cal-culated using the method in Berger,Ofek,and Swary(1996)and Almeida and Campello(2007).Table1presents descriptive statistics on governance scores,institutional ownership,and other stock characteristics for our study sample offirms.The mean(median)value of GOV SCORE1is23.65(24),indicating that our sam-plefirms meet about half of the governance standards.7The mean(median)value of GOV SCORE2is20.11(20).The mean(median)value of our samplefirms’in-stitutional ownership is56.31%(61.23%).The mean and median values of market capitalization are$3,600million and$446million,respectively,indicating high skewness in the distribution offirm size.The mean(median)value of the effective bid-ask spread is0.0083(0.0035),and the mean(median)value of TOBIN’S Q ratio is1.7182(1.2707).The mean(median)value of asset tangibility is0.5181 (0.5146),indicating that our samplefirms hold large amounts of tangible assets.IV.Empirical ResultsThis section presents the results of univariate tests and regression analyses on the relation between corporate governance and institutional ownership.A.Univariate TestsIn each year,we sortfirms according to governance scores and group them into governance-score quintiles.8We then aggregate allfirm-year observations within each quintile across the6-year study period.Table2presents the mean and6The number of governance standards included in the ISS database was much smaller prior to enactment of the Sarbanes-Oxley Act in2002.7Brown and Caylor(2006)and Chung et al.(2010)report similar results.The99th percentile value of our GOV SCORE1is35,which is also in line with the maximum governance score of38(out of 51)reported in Brown and Caylor(2006).8We report the results based on GOV SCORE1.We also conduct the same test using GOV SCORE2andfind qualitatively identical results.Chung and Zhang253TABLE1Descriptive StatisticsTable1presents descriptive statistics on governance scores,institutional ownership,and other stock characteristics. GOV SCORE1and GOV SCORE2denote the governance score for eachfirm in a specific year.We determine whether a particular governance standard is met using the minimum standard provided in ISS Corporate Governance:Best Practices User Guide and Glossary(2003).We then obtain the governance score for eachfirm by awarding1point for each gover-nance standard that is met.INST OWNERSHIP is the ratio of the number of shares held by institutional investors to the total number of shares outstanding,MVE is the market value of equity,VOLATILITY is the standard deviation of quote-midpoint daily returns,TURNOVER is the average ratio of monthly trading volume to the number of shares outstanding,PRICE is annual average stock price,RETURN is the continuously compounded annual stock return,BASPREAD is the mean ef-fective bid-ask spread,FIRM AGE is the number of years since thefirmfirst appeared in the CRSP database,TOBIN’S Q is the Tobin’s q ratio calculated using the method in Chung and Pruitt(1994),TANGIBILITY is the asset tangibility ratio calculated using the method in Berger et al.(1996)and Almeida and Campello(2007),LEVERAGE is the ratio of total debt to the book value of total assets,D YIELD is the annual dividend yield,ROA is the ratio of net income to the book value of total assets,and S&P500equals1if thefirm is included in the S&P500index,and0otherwise.PercentileVariables Mean Std.Dev.1st5th25th50th75th95th99th GOV SCORE123.6462 5.801612.000014.000019.000024.000028.000033.000035.0000 GOV SCORE220.1069 4.508411.000013.000016.000020.000024.000028.000029.0000 INST0.56310.27800.00480.05580.35010.61230.79480.94070.9847 OWNERSHIPMVE3,600.8516,546.4610.6927.54146.74446.111,549.7713,765.9764,220.51 (in$millions)VOLATILITY0.03090.01490.01010.01340.02050.02780.03780.05810.0799 TURNOVER 1.6462 2.08140.08790.19870.6246 1.1640 2.0397 4.43798.7576 PRICE21.010626.4974 1.2659 2.35547.284915.819728.550054.258378.1139 RETURN0.21580.5872–0.7120–0.5095–0.14490.11340.4243 1.3387 2.4259 BASPREAD0.00830.01480.00040.00060.00140.00350.00970.03050.0567 FIRM AGE17.314716.2267 1.0000 2.0000 6.000011.000023.000045.000078.0000 TOBIN’S Q 1.7182 1.52300.18870.47180.8684 1.2707 2.0376 4.46347.9432 TANGIBILITY0.51810.17420.12230.23580.40550.51460.61630.83830.9430 LEVERAGE0.18480.20330.00000.00000.00370.14040.29560.54730.8380 D YIELD0.00830.03590.00000.00000.00000.00000.00710.03680.0886 ROA0.00850.1399–0.5167–0.3055–0.01290.03890.07890.15970.2561 S&P5000.18340.27870.00000.00000.00000.00000.0000 1.0000 1.0000TABLE2Institutional Ownership and Governance ScoreIn each year,we sortfirms according to governance scores and group them into governance-score quintiles.We then ag-gregate allfirm-year observations within each quintile across the6-year study period.The1st column in Table2presents the mean and median institutional ownership offirms within each quintile.To determine whether the relation between institu-tional ownership and corporate governance varies across different types of institutions,we repeat the previous procedure using data for each of the following5types of institutions based on Thomson’s classification:i)bank trust departments; ii)insurance companies;iii)investment companies;iv)independent investment advisors;and v)others(e.g.,pension funds,foundations,and university endowments).We perform the t-test(Wilcoxon test)using data for each year to deter-mine whether the mean(median)value of institutional ownership forfirms that belong to the highest governance-score quintile is statistically different from the corresponding value forfirms that belong to the lowest governance-score quintile. The last2rows present the mean value of each statistic across years.Type5(otherinstitutions:Type4pension funds, Total Type2Type3(independent foundations, Institutional Type1(insurance(investment investment universityOwnership(banks)companies)companies)advisors)endowments) Governance-ScoreQuintile Mean Med Mean Med Mean Med Mean Med Mean Med Mean Med 1(lowest)0.42200.40310.04870.03240.02030.01070.06890.04820.23090.21460.07700.0576 20.48050.48850.05160.04060.02300.01240.07830.05810.26040.25450.08730.0699 30.56270.61090.06650.05960.02440.01550.09000.07450.29750.30030.10140.0849 40.64410.70570.08240.07980.03010.02010.10360.09400.33140.33790.10900.0970 5(highest)0.70770.74860.09820.09400.03280.02310.12200.11380.34360.34350.11670.1044 Mean t-values16.8212.64 5.8110.7010.738.42 (Student’s t-test)Mean z-values13.7913.318.7710.689.859.89 (Wilcoxon test)254Journal of Financial and Quantitative Analysismedian institutional ownership offirms within each quintile.The results show that institutional ownership increases monotonically with governance scores.The mean institutional ownership offirms that belong to the lowest governance-score quintile is42%,whereas the correspondingfigure is71%forfirms that belong to the highest governance-score quintile.We perform the t-test(Wilcoxon test)using data for each year to determine whether the mean(median)value of institutional ownership forfirms that belong to the highest governance-score quintile is statistically different from the corre-sponding value forfirms that belong to the lowest governance-score quintile.The results show that the difference in the mean(median)institutional ownership be-tween the2quintiles is statistically significant at the1%level in all sample years. The last2rows in Table2present the mean value of each statistic across years.To determine whether the relation between institutional ownership and cor-porate governance varies across different types of institutions,we repeat the pre-vious analysis using data for each of the following5types of institutions based on Thomson’s classification:i)bank trust departments;ii)insurance companies; iii)investment companies;iv)independent investment advisors;and v)others (e.g.,pension funds,foundations,and university endowments).9The results show thatfirms with higher governance scores exhibit greater institutional ownerships across all types of institutions.For example,for Type1institutions(banks),the mean and median institutional holdings offirms that belong to the lowest governance-score quintile are4.87%and3.24%,respectively,whereas the cor-respondingfigures forfirms that belong to the highest governance-score quintile are9.82%and9.40%,respectively.The difference in the mean(median)institu-tional ownership between the2quintiles is statistically significant at the1%level in each year of our study period across all types of institutions.B.Regression Results Using the Level VariablesAlthough the results in Table2are consistent with the conjecture thatfirms with better governance structure attract more institutional investors,the relation between the2variables may be driven by their respective correlations with other variables.Falkenstein(1996)and Huang(2009)show that mutual funds prefer stocks with higher market liquidity.Chung et al.(2010)find thatfirms with better governance structure tend to have higher stock market liquidity(e.g.,narrower quoted and effective spreads).Consequently,the relation between institutional ownership and governance scores may be due to their respective correlations with liquidity.In addition,prior research identifies a number of variables that af-fect institutional investors’portfolio decisions.Gompers and Metrick(2001)find that institutional investors prefer stocks of larger companies,while Grinstein and Michaely(2005)show that institutional investors preferfirms that pay smaller dividends among dividend-payingfirms.Badrinath et al.(1996)show that insti-tutional investors preferfirms with safety net characteristics,such as established (old)companies,low return volatility,and lowfinancial leverage.Gompers et al.9CDA/Spectrum’s classification of institutional investors is not reliable from1998forward(e.g., see Chen,Harford,and Li(2007)).Therefore,we use the pre-1998classification.Chung and Zhang255 (2003)show thatfirms with stronger shareholder rights exhibit higher stock re-turns and better operating performance.To examine the relation between institutional ownership and corporate gov-ernance after controlling for the factors that are associated with both or either of the2variables,we estimate the following regression model:(1)INST OWNERSHIP i,t=β0+β1GOV SCORE i,t+β2Log(MVE i,t) +β3VOLATILITY i,t+β4TURNOVER i,t+β5Log(PRICE i,t)+β6RETURN i,t+β7BASPREAD i,t+β8FIRM AGE i,t+β9TOBIN’S Q i,t+β10TANGIBILITY i,t+β11LEVERAGE i,t+β12D YIELD i,t+β13ROA i,t+β14S&P500i,t+εi,t,where all variables are the same as defined in Section III.Panel A of Table3presents the OLS regression results with clustered stan-dard errors at thefirm level that are estimated from12,093firm-year observations (we include year dummy variables in the regression model).The clustered stan-dard errors correctly account for the dependence in the data common in a panel data set and produce unbiased estimates.See Petersen(2009)for a detailed de-scription of this method in the context offinancial research.To assess the sensi-tivity of our results with respect to different estimation methods,we also estimate regression model(1)using the Fama-MacBeth(1973)method and report the re-sults in Panel A.10Columns1and2present the results using GOV SCORE1and columns3and4present the results using GOV SCORE2.The results indicate that the coefficients on both GOV SCORE1and GOV SCORE2are positive and statistically significant,indicating that the positive relation between institutional ownership and corporate governance quality remains intact even after controlling for other possible determinants of institutional ownership.The relation between institutional ownership and corporate governance is economically significant:An increase in governance score from the25th to75th percentile results in an increase in institutional ownership by14%.11It is important to note that annual ISS governance data are gathered from proxy statements that are usually collected during thefirst half of each year and thus they may capture the governance structure during the second half of the pre-vious year.12Note also that our institutional ownership variable is constructed based on the year-end holding data.Hence,the positive relation between institu-tional ownership and governance score shown in Table3is more likely to indicate10Following Fama and MacBeth(1973),wefirst run cross-sectional regression using data for each year and then perform the t-test using the estimated regression coefficients across years.For each variable,we show the mean regression coefficient across years and its t-value.The high t-values are largely due to small variation in the estimated regression coefficients across years.11Note that0.0088×(28–19)/0.5631=14.1%,where0.0088is the estimated coefficient on GOV SCORE1in Table3(Panel A),28is the75th percentile value of GOV SCORE1,19is the25th percentile value of GOV SCORE1,and0.5631is the mean aggregate institutional ownership of our samplefirms.Similarly,we obtain0.0104×(24–16)/0.5631=14.8%using GOV SCORE2.12For example,Amylin Pharmaceuticals,Inc.began to meet the governance standard that“the positions of chairman and CEO should be separated,”from September2003.However,this change shows up in the ISS database only from2004.。