CORPORATE GOVERNANCE, DISCLOSURE QUALITY, OWNERSHIP STRUCTURE, AND FIRM VALUE. Ferdinand T. Siagian Sylvia Veronica Siregar Yan Rahadian

Transcription

1 CORPORATE GOVERNANCE, DISCLOSURE QUALITY, OWNERSHIP STRUCTURE, AND FIRM VALUE Ferdinand T. Siagian Sylvia Veronica Siregar Yan Rahadian Department of Accounting Faculty of Economics, University of Indonesia Indonesia

2 INTRODUCTION This study investigates simultaneous relations among disclosure quality, Good Corporate Governance (GCG) practice, and firm value. Specifically, this study examines whether firms compliance to Financial Statement Reporting and Disclosure Standard (P3LKE) for public companies in Indonesia issued by the Indonesian Capital Market Regulation Body (BAPEPAM) affects firm value measured by price to book value (PBV) ratio. We also examine whether firm s GCG practice and ownership structure affect value. Simultaneously, we examine whether firm value affects disclosure quality and GCG practice. Finally, we examine whether ownership structure affects GCG and disclosure quality and whether GCG practice affects disclosure quality. Negative association between disclosure and cost of equity [11] and cost of debt [43] implicitly shows that firm quality of disclosure positively affects its value. [26] and [22] find consistent evidence that disclosure quality affects stock price positively. However, we argue that firm with good performance shown by its high market value will tend to disclose more and therefore, will have a higher disclosure quality. If that is the case, there are two-way relations between firm value and disclosure quality and therefore, the two variables are jointly determined. Other variables that have been found to affect the disclosure level are GCG practice and ownership structure. [25] find positive association between GCG practice and firms disclosure level, both voluntary and mandatory. Using a sample of Singapore and Hong Kong firms, [13] find that firms controlled by family that are common in Hong Kong and Singapore have lower level of disclosure. Corporate governance studies find that GCG practice does affect firm value. [18], [30], [23], [3], and [33] find that GCG practice increases stock return. Their studies use Ordinary Least Square (OLS) regressions in their tests. Using simultaneous equation, [9] find that GCG practice positively affects firm value. They also find reverse causality; higher value firms adopt better corporate governance practice. Using OLS is not appropriate if there are twoway relations (simultaneity) between variables because the coefficient will be bias and inconsistent. 1 [8] examine simultaneous relations among Tobin s Q, Board composition, and managerial ownership. They find that the three variables are jointly determined. Their finding is consistent with findings in [17] that governance structure and performance are jointly determined. Our study extends the above studies by simultaneously examine the interrelationships among four variables; GCG practice, disclosure quality, and firm performance. We develop a GCG 1 See [24] for complete discussion about simultaneity and bias in OLS coefficients.

3 index to be a proxy for firm s GCG practice. Specifically, we develop a comprehensive checklist that is based on the Indonesian Institute of Corporate Directorship (IICD) index and calculate the score for all public firms. We also develop an index that measures firm s compliance to LP3KE issued by BAPEPAM that we use as a proxy for firm s disclosure quality. Studies in corporate governance literature not only examine the association between GCG practice and firm value but also factors that affect the implementation of corporate governance. One factor that affects corporate governance is ownership structure. [34] find positive association between institutional ownership and corporate governance score for public firms in Britain. The positive association is influenced by institutional investors pushing the firm to disclose more about the implementation of GCG. Institutional ownership is also found to positively affect the firm value [39]. Because majority ownership by family is very common in Indonesia we also examine whether family ownership is associated with disclosure quality, GCG practice, and firm value. There are conflicting results in this area of study. Several studies find that familycontrolled firms have higher values than other firms ([38]; [47]; [5]) but other study finds otherwise [27]. Previous studies have separately examine the association between corporate governance and firm value, corporate governance and reporting quality, reporting quality and firm value, ownership structure and firm value, and corporate governance and ownership structure. Extending previous studies, we examine whether there are simultaneous associations among the variables and conduct our tests using simultaneous equations. Our results indicate that there are interdependencies between GCG practice, disclosure quality, and firm value. We find that GCG practice and institutional ownership positively affect firm value. We do not find that firm value affects GCG practice and contrary to our prediction, institutional ownership negatively affects corporate governance practice. Finally, none of the variables in one of our models (GCG practice, ownership structure, or firm value) explains the variation of disclosure quality. Our study contributes to the literature in the followings: 1. Previous studies have not examined simultaneity among GCG practice, disclosure quality, and firm value. Our study examines the relations simultaneously. 2. We examine the two-way relations between (1) corporate governance and firm value and (2) quality of reporting and firm value. Previous studies only examine two way relations between corporate governance and firm value. The two-way relations between quality of reporting and firm value has not been studied as far as we know. 3. Compliance to P3LKE has not been studied before in Indonesia. We create P3LKE compliance checklist to measure the firms compliance to the disclosure rules. We use the result of the checklist as our proxy to quality of reporting. Our findings support the hypothesis that firms follow P3LKE because it is mandatory (none of our

4 variables explains the variation in P3LKE compliance) despite the fact that the compliance score is still low. 4. The use corporate governance index in researches in Indonesia is still rare because such index for all public firms is not available. We develop a corporate governance checklist that we use as a proxy for the level of corporate governance practice. The remainder of the paper is organized as follows. We draw the theories about the subjects and develop the hypotheses in section 2. Section 3 describes the research method, sample selection procedure, and descriptive statistics. We present the results of the tests in section 4 and conclude the study in section 5. LITERATURE REVIEW AND HYPOTHESES DEVELOPMENT DISCLOSURE QUALITY AND FIRM VALUE In 2002 Indonesian Capital Market Regulation Body (BAPEPAM) issued a Financial Statement Reporting and Disclosure Standard (P3LKE) [46] that sets the minimum disclosure requirements for firms publicly traded in Jakarta Stock Exchange (JSX). The purpose of this regulation is to improve the disclosure quality of firms traded in JSX. There are 13 standards for 13 different industries in P3LKE as presented in Table 1. Table 1. Industries of Indonesian Financial Statement Reporting and Disclosure Standard (P3LKE) No Industry 1 Manufacturing 2 Investment 3 Hospital 4 Toll Road 5 Hotels 6 Restaurants 7 Telecommunication 8 Construction 9 Trade 10 Transportation 11 Real Estate 12 Farming 13 Forestry [40] find that during the Asian Crisis, firms with high quality of disclosures show better performance than firms with lower quality. [7] also find that during Korean crisis in 1997 firms with high disclosure quality experienced the lowest stock price decline. In sum, these studies find that there is a positive association between firm s disclosure quality and performance. [11] and [43] find that disclosure reduces cost of equity and cost of debt respectively which implicitly shows that firm quality of disclosure positively affects its value.

5 We hypothesize that Indonesian firms with higher quality of disclosures shown by their compliance to P3LKE will have higher value shown by higher price-to-book value. If firms with higher value are firms with higher performance, we argue that the management of these firms will have the tendency to disclose more in order to provide more information to the market about their performance. Therefore, they will have higher disclosure quality. We hypothesize that firm value positively affects disclosure quality. In other words, we hypothesize that there are two-way relations between firm value and disclosure quality. H1: Firms with higher disclosure quality will have higher value. H2: Firms with higher value will have higher disclosure quality. CORPORATE GOVERNANCE AND FIRM VALUE Using firms in five East Asian countries [40] find that GCG practice positively affects firm performance during Asian crisis. [30] use corporate governance rank data of firms in 14 developing countries. They find evidence that support the contention that corporate governance practice is positively related to operating performance and market value. Consistent with the above studies [3] so find positive impact of corporate governance practice on stock returns. The above studies use OLS regression. [9] state that OLS will result in coefficient bias if there are simultaneous relations among variables. They find that GCG practice positively affects firm value but they also find reverse causality; higher value firms adopt better corporate governance practice. [1] examine seven control mechanisms using simultaneous equations. Using OLS, they find that insider shareholding, outside director, debt, and corporate control activity each affects firm performance if separately regressed. However, if regressed together using OLS, the significant coefficient of insider shareholding disappears and when regressed simultaneously, the significant coefficients of debt and corporate control activity also disappear. Their finding shows that there are interdependencies among the variables. [15] use simultaneous equations to model CEO ownership, CEO compensation, and Tobin s Q. Using OLS and 3SLS they find that the three variables are endogenous variables where Tobin s Q positively affects CEO ownership and CEO compensation, and CEO ownership affects Tobin s Q. Black et al [10] develop a corporate governance index for all public firms in Korea for year Using OLS, 2SLS, and 3SLS they find evidence that higher corporate governance index is associated with higher Tobin s Q. Using simultaneous equations [9] find consistent results that corporate governance positively affect firm value. They also find reverse causality where firms with high value adopt better good corporate governance practice. From the above findings we argue that there is a two-way relation between corporate governance and value for firms in Indonesia. H3: Firms with better GCG practice have higher value. H4: Firms with higher value have better GCG practice.

6 OWNERSHIP STRUCTURE AND FIRM VALUE Family ownership [20] show that firms with family ownership are more efficient because monitoring costs in family-controlled firms are smaller. [38] find that firms controlled by family have higher value, are more efficient, and have less debt. [47] also find a positive association between family ownership proportion and stock return and profitability. Using sample firms in Thailand, [45] finds that family-controlled firms perform better than firms without controlling ownership or those controlled by the government. [45] also finds that family-controlled firms were not the reason for Asian crisis in [4] find that the costs of debt of family-controlled firms are lower than other firms because firms with family ownership have structures that reduce the agency problem between shareholders and creditors. Creditors seem to believe that family-controlled firms give more protection on the creditor s interests. [5] test whether family-controlled firms routinely take actions that impair minority shareholders. They find that in contrary, minority shareholders benefit from the family ownership. Using a sample of Indonesian firms, [6] shows that family-controlled firms, government-controlled firms, and financial institution-controlled firms have lower agency problems than firms without major ownership. He argues that in family-controlled firms, there are less agency problems because of fewer conflicts between the principal and the agent. Although the above studies find that there is positive relation between family ownership and firm value, other studies find otherwise. [5] find that public firms controlled by family, amount to one-third of all public firms in the US, have higher Tobin s Q and Return on Assets. However the positive relation weakens once the family ownership reaches 30%. This finding indicates that if the family domination is getting higher, there is a point where the impact on value will become negative. [27] find negative relation between family ownership and performance and that very dominant ownership can be used by the controlling family to transfer wealth from the minority shareholders. [37] states that the benefit from family controlling the firm is mainly for firms without majority ownership. Family control reduces agency problem between owners and managers but is potential to create conflict between the controlling family and minority shareholders if protection to the minority shareholders is weak. Conglomerate firm One factor that may affect the shareholders protection is whether the firm is part of a conglomerate or not. In conglomerate firms the shareholders wealth are not concentrated in one firm but are spread out in many firms. This will increase the probability of expropriation of the minority shareholders wealth. Kim and Yi (2005) study the impact of the existence of business affiliation on earnings management. They find that the magnitude of earnings management is higher for firms with business affiliation that firms without business affiliation. This indicates that firms with business affiliation give the controlling shareholders more incentives and opportunity to manage earnings for their own benefits. Although family-controlled may have a value benefit, it may not apply to conglomerate firms. Therefore, there may be an offsetting effect between the positive impact on value from

7 family-controlled and the negative impact on value from conglomerate status. The offsetting effect would not exist for firms with family ownership that are not part of conglomerate. H5: Value of firms with high proportion of family ownership and non-conglomerate are higher than that of other firms. Institutional Ownership According to [21], institutional investors have the interests to monitor management actions because they have economic interest in that firm. Large ownership gives the institutional investors large incentives to actively monitor and influence management actions and policies. [39] find evidence that support positive relations between institutional ownership and Tobin s Q. From the above discussions, we hypothesize: H6: Firms with higher institutional ownership have higher value. OWNERSHIP STRUCTURE AND CORPORATE GOVERNANCE Key mechanisms of effective governance according to [29] are ownership structure, board structure, remuneration of the CEO and management, auditing and information, and market for corporate control. [34] find that institutional ownership positively impact governance score for firms in Britain. One possible explanation is because institutional investor forces the firm to disclose more on the firm s corporate governance practice. We hypothesize that firm with high proportion of institutional ownership will implement better GCG practice. H7: Firms with higher proportion of institutional ownership have higher corporate governance practice. [16] find that family dominates the ownership of public firms in several Asian countries such as that in Hong Kong, Indonesia, Malaysia, Korea, and Thailand. They also find that the ownership concentration causes low firm value. This indicates the possibility of expropriation of the minority shareholders by the majority shareholders. Their study also indicates that firms with concentrated ownership do not have adequate GCG practice that should protect the minority shareholders from expropriation by the majority shareholders. However, family owners may have more power to force the implementation of GCG to protect their stake in the firm. We hypothesize that there is association between GCG practice and family ownership although we do not predict the direction. H8: Corporate governance practice is associated with high proportion of family ownership and non-conglomerate status. OWNERSHIP STRUCTURE AND DISCLOSURE QUALITY Controlling institutional investors have significant stake in firm s performance. It is very important for them to know whether the management is doing on their behalf. Therefore, these investors will force the management to produce high quality disclosure. Several studies connect ownership structure to quality of reporting. [13] examine the association between ownership structure and voluntary disclosure using public firms in Hong Kong and Singapore. They find that outside ownership is positively associated with disclosure level but firms controlled by family that are very common for Hong Kong and Singapore have lower disclosure level. H9: Firms with higher proportion of institutional ownership have higher disclosure quality.

8 [19] study the impact of GCG practice on voluntary disclosure. They use ownership structure and board composition as the proxy for corporate governance. They find that the lower managerial and government ownership, the higher is the level of disclosure. We hypothesize that non-conglomerate firms controlled by family will disclose more in order to provide information to minority shareholders. H10: Disclosure quality of firms with high proportion of family ownership and non conglomerate are higher than that of other firms. CORPORATE GOVERNANCE AND DISCLOSURE QUALITY [25] find that corporate governance positively affect both voluntary and mandatory disclosure. In Britain, Peasnell et al (2000) find that outside directors is effective in reducing earnings management only after the publication of Cadbury Committee corporate governance guidelines; outside directors are not effective in reducing earning management before the publication. [2] study whether corporate governance mechanism is associated with the probability of financial restatement. They find that the existence of independent directors with financial knowledge reduces the probability of restatement. They also find that firms whose CEO is from the founding family have higher probability to restate their financial statements than firms whose CEO is from outside. Because one of the GCG principles is transparency, it is likely that firms that implement GCG practice will disclose more. H11: Firms with higher corporate governance practice have higher disclosure quality. RESEARCH METHOD SAMPLE SELECTION Our sample consists of firms traded in Jakarta Stock Exchange (JSX) in 2003 and 2004 that have submitted P3LKE to BAPEPAM with complete financial data. Table 2 shows the sample selection process. Table 2. Sample Selection Firm-Year with P3LKE checklist Firm-Year with CG checklist Firm-Year with both checklists Firm-Year with negative book value Firm-Year with incomplete financial data Total firm-year in the final sample We develop our disclosure quality checklist from BAPEPAM Financial Reporting and Disclosure Standards (P3LKE). We also develop a new corporate governance checklist that combines items in IICD, [41], and checklist from S&P s and National University of Singapore research. Financial data are obtained from financial reports from JSX database.

9 MODEL SPECIFICATION To test our hypotheses, we use the following simultaneous equations: CGit = a0 + a1pbvit + a2dfam it + a3instit + a4govtit + a5forit + a6sizeit + a7 GROWTH it + a8lnageit + a9dyeart + ε it (1) QREPit = b0 + b1 PBVit + b2dfam it + b3instit + b4cgit + b5govtit + b6 FORit + b7 SIZEit + b8growth it + b9opini it + b10 DYEARt + ε it (2) PBVit = c0 + c1cg it + c2dfam it + c3instit + c4qrepit + c5govtit + c6forit + c7 SIZEit + c8growth it + c9levit + c10 DYEARt + c11roa it + εit (3) where: CG = Corporate Governance Index PBV = Price to Book Value Ratio QREP = P3LKE Compliance Index DFAM = Family Ownership Dummy variable (1 if > 50% shares owned by a family and the firm is not a conglomerate, 0 otherwise) INST = Percentage share owned by institutional investor GOVT = Percentage share owned by Indonesian government FOR = Percentage share owned by foreign investor SIZE = Log of Total Assets at year-end GROWTH = Average growth of sales in the last three years LNAGE = Log of Listing Age OPINI Accounting Firm Opinion Dummy variable (1 for unqualified opinion, 0 otherwise) LEV = Leverage measured by Debt-to-Equity Ratio DYEAR = Year Dummy (1 for year 2003, 0 for year 2004) Our hypotheses predict the following: H1: c4 > 0 H2: b1 > 0 H3: c1 > 0 H4: a1 > 0 H5: c3 > 0 H6: c2 > 0 H7: a3 > 0 H8: a2 0 H9: b3 > 0 H10: b2 > 0 H11: b4 >0 VARIABLES DEFINITION Ownership Structure Variables First, we classify our sample into two groups; conglomerate firms and non-conglomerate firms. Then we classify our sample into groups of firms with high or low family ownership. The definition of family ownership follows the definition used in [6]; all ownership greater than 50% that is not owned by public firm, government, or financial institution. Firms with family ownership more (less) than 50% are considered high (low) family ownership. Combining the two classifications, we get 4 groups (conglomerate firms with high family ownership, conglomerate firms with low family ownership, non-conglomerate firms with high family ownership, and non-conglomerate firms with low family ownership). We then assign 1 for non-conglomerate firms with high family ownership and 0 for the other three

10 groups. Institutional investor (INST) is non-affiliated financial institutions such as insurance companies, banks, pension funds, and investment banking. Corporate Governance Variable Corporate Governance variable (CG) is based on corporate governance index that we develop using the corporate governance checklist. This checklist is completed using firm s financial statements (secondary data). We adapt the checklist from IICD, [41] and [44]. The questions in the checklist are divided into 5 groups according to OECD principles [42]: 1. Rights of shareholders 2. Equal treatment of shareholders 3. Role of stakeholders 4. Disclosure and transparency 5. Board responsibilities Firm Value Variable We use Price to Book Value (PBV) as the proxy for firm value that we calculate as follows: Pr icepershare PBV = BV pershare Disclosure Quality Variable The quality of disclosure variable (QREP) is measured using P3LKE checklist that we develop using standards issued by BAPEPAM. BAPEPAM issues P3LKE for 13 industries, and therefore we develop 13 different checklists. Each checklist has approximately 650 items. For each item, we read a firm s notes to Financial Statements to measure its disclosure compliance. There are three possible answers to the questions in the checklist; Yes, No, or Not Applicable (N/A). The score is calculated as follows: # Yes QREP = (# items # N / A) In calculating the score we realize the potential bias because sometimes it is difficult to differentiate between Not Applicable (the firm does not have the item) and No (the firm has the item but does not disclose). Control Variables In Equation 1, government ownership (GOVT) is included as control variable. This variable shows the percentage of shares owned by the Indonesian government. [36] state that stateowned firms may have weaker governance than other firms because they have objectives related to the country s welfare and not just the firm objectives. State-owned firms receive fund from the government but the government is usually a long-term investor and does not monitor its investment in an active manner. [31] study firms with US government ownership of more than 35% and find that the performance of these firms are not significantly different from other firms in the same industry. However, publicly traded state-owned firms are among the best and therefore, most

11 likely to implement a good corporate governance. We expect that government ownership may affect the implementation of good corporate governance principles for firms in Indonesia. Besides controlling for government ownership, we also control for foreign ownership (FOR). This variable represents the percentage of shares owned by foreign investors. [35] find that foreign investors are less likely to invest in firms with bad corporate governance. [7] find that during the 1997 crisis in Korea, firms with the higher foreign ownership experience the lowest decrease in share price. We argue that foreign ownership is associated with GCG level. Size is included as a control variable because larger firms tend to have more complex agency problem and hence voluntarily choose more strict governance. We contend that firm size positively affect GCG implementation. We use log of total assets as the proxy for size. Firms with high growth rate (GROWTH) needs large external financing and therefore will feel the need to improve their governance if they expect that better corporate governance reduces cost of capital. We use the average growth of sales in the last three years as our proxy for firm growth. Listing time (LNAGE) may affect the quality of governance because firms that have been listed for a while will become mature and may implement better governance. Following [9] we use the log of the age of being listed. Dummy year (DYEAR) is included as control variable to control for differences in adjusted mean of the dependent variable between different years. In Equation 2 (dependent variable = quality of disclosure) we include Public Accountant Opinion (OPINI) that represents the quality of audit as control variable. We content that higher quality of disclosure will be associated with higher audit opinion. This is a dummy variable with 1 for firms with unqualified opinion and 0 for other opinions. High profitable firms tend to have high value. Therefore, we include ROA as a control variable in Equation 3 (dependent variable = PBV). Following Yermack (1996), Schmid (2003), and [9], SIZE and GROWTH are also included as control variables in Equation 3. Leverage can affect firm value negatively or positively. The tax benefit of debt and management discipline by debt can cause positive impact of debt on firm value, but higher default risk due to higher debt can cause negative impact on firm value. The net impact is an empirical test. We use Debt-to-Equity ratio as the proxy for firm leverage. Dummy year (DYEAR) is included to control the possibility that the value is driven by year. ANALYSIS METHOD To test whether there is simultaneity among the variables we use Hausman Specification Test. We test the existence of simultaneity because if there is simultaneity, OLS will create

12 bias and inconsistent and therefore we need to use Two-Stage Least Squares (2SLS) [24]. Table 3 shows the result of the simultaneity test. Table 3. Hausman Specification Test To From CG (1) QREP (2) PBV (3) CG (1) (0.0713) * QREP (2) (0.0157) (0.3334) (0.0323) ** PBV (3) (1.6136) (0.0546) * (0.0323) ** ** Test variable is significant at 5% level * Test variable is significant at 10% level Table 3 shows that there are interdependencies between variables especially between QREP and PBV. All the tests show significant results at 5% or 10% levels and therefore, we use the Two Stage Least Squares (2SLS) instead of the Ordinary Least Squares (OLS). RESULTS DESCRIPTIVE STATISTICS Table 4 presents the descriptive statistics for variables used in our hypothesis tests. The average corporate governance score is while the average for the weighted score (WCG) is The Standard deviations of the two variables are similar. The average of PBV is with a high standard deviation (1.4355). This indicates that the PBV of the firms in the sample vary and are not concentrated in low or high PBV. P3LKE compliance score (QREP) is relatively small with average of only

13 Table 4. Descriptive Statistics Mean Std. Dev. Minimum Maximum CG WCG PBV QREP INST GOVT FOR SIZE GROWTH (0.7000) LNAGE LEV Dummy = 1 Dummy = 0 OPINI 91.15% 8.85% DFAM 19.27% 80.73% CG = unweighted corporate governance index, WCG = weighted corporate governance index, PBV = price to book value ratio, QREP = P3LKE index, DFAM = family owners dummy variable (1 if family ownership > 50% and not conglomerate firm), INST = perc institutional ownership, GOVT = percentage government ownership, FOR = percentage foreign ownership, SIZE = log total assets, AGE = listing age, OPINI = external auditor opinion dummy variable (1 if unqualified opinion, 0 if other opinion), LEV = leverage, GROWTH = growth opportunity Institutional ownership and government ownership of the firms in the sample are relatively small, 7.71% and 1.55% respectively. Foreign ownership is relatively higher than government ownership. On average, only 19.35% of the sample firms shares are owned by foreign investors. The firm s leverage also varies with average of and standard deviation of Most of the firms in the sample (91.15%) receive unqualified opinions and only small percentages (19.27%) of these unqualified opinion firms are nonconglomerate with family ownership more than 50%. Table 5 presents the Pearson Correlation for variables in our model. It shows that firm value is positively correlated with corporate governance practice. We also find that quality of reporting is positively correlated with size; large firms have better quality of disclosure. Corporate governance practice is negatively associated with family ownership, but positively correlated with government ownership, foreign ownership, size, and age of listing. We do not find significant correlation between institutional ownership and corporate governance practice.

14 HYPOTHESES TESTING RESULTS We present the result of the tests in Table 6. Inconsistent with the hypothesis that firm value positively affects good corporate governance practice, we find that PBV does not affect corporate governance index. This may indicate that firms with high value are not likely to improve their corporate governance to improve their value further. We also do not find support that family ownership and non-conglomerate status (DFAM) positively affects GCG practice. Specifically, there is no significant difference in corporate governance index between non-conglomerate family-owned firms and other firms in our sample. This result may be due to the sample characteristics where less than 20% of the sample firms are categorize in the family-owned non-conglomerate group. Other ownership structure variable is institutional ownership (INST). Contrary to our expectation the sign of the coefficient for institutional ownership is negative and significant. We predict that the existence of institutional ownership will improve firm governance and hence higher governance index. However, the regression result shows a negative association between the two variables. One possible explanation is that institutional ownerships in the sample firms are still low and therefore, these investors are not monitoring the firm s corporate governance well enough. [12] finds that short-term focus and active institutional investors cause management choose short-term oriented actions. [14] state that corporate governance implementation is costly and therefore in the short-term it may be more beneficial if the firm does not implement good corporate governance. It is possible that institutional investors in Indonesian firms are shortterm oriented investors that may be more beneficial with low good governance implementation. One other explanation is that firms with low governance have more room for improvements. It is possible that institutional investors invest in low GCG firms in order to get the highest benefit if the firm can improve its governance.

15 Table 5. Pearson Correlation Table Pearson Correlation Sig. (2-tailed) CG WCG QREP PBV DFAM INST FOR GOVT SIZE GROWTH LNAGE OPINI LEV CG (**) (**) -.176(*) (0.11).198(**).274(**).520(**) (**) (0.01) (0.02) WCG (**) -.169(*) (0.09).199(**).269(**).522(**) (0.01).193(**) - (0.04) QREP 1.00 (0.09) (0.02) (*) (0.01) PBV (*) (*).256(**) (**) DFAM 1.00 (0.11) -.270(**) (0.08) -.180(*) -.157(*) -.298(**) 0.11 (0.07) INST (**) (0.02) (0.03) (0.04) 0.03 (0.07) FOR 1.00 (0.05) (**) 0.03 (0.13) GOVT (**) (**) 0.05 (0.01) SIZE (**) (0.13) GROWTH 1.00 (0.06) (0.13) (0.06) LNAGE 1.00 (0.11) OPINI (**). - N = 192 ** Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed). -

16 Table 6. Regression Results Model 1 CG it = a 0 + a 1 PBV it + a 2 DFAM it + a 3 INST it + a 4 GOVT it + a 5 FOR it + a 6 SIZE it + a 7 GROWTH it + a 8 LNAGE it + a 9 DYEAR t + ε it Model 2 QREP it = b 0 + b 1 PBV it + b 2 DFAM it + b 3 INST it + b 4 CG it + b 5 GOVT it + b 6 FOR it + b 7 SIZE it + b 8 GROWTH it + b 9 OPINI it + b 10 DYEAR t + ε it Model 3 PBV it = c 0 + c 1 CG it + c 2 DFAM it + c 3 INST it + c 4 QREP it + c 5 GOVT it + c 6 FOR it + c 7 SIZE it + c 8 GROWTH it + c 9 LEV it + c 10 DYEAR t + ε it Est. Coeff. Est. Coeff. Est. Coeff. (Probability) (Probability) (Probability) Predicted CG Predicted QREP Predicted PBV Variable Sign Sign Sign N Intercept ** (0.3436) (0.2128) (0.0196) CG ** (0.3086) (0.0454) PBV (0.8975) (0.5597) QREP (0.8902) DFAM +/ / / (0.5262) (0.4464) (0.1764) INST *** * (0.0072) (0.3261) (0.0507) GOVT *** * (0.008) (0.3193) (0.0742) FOR *** * (0.0013) (0.3152) (0.0697) SIZE *** * (0.0000) (0.2831) (0.0631) GROWTH ** (0.5740) (0.5425) (0.0370) LNAGE (0.8374) OPINI (0.3147) LEV *** (0.0009) DYEAR * (0.4844) (0.2869) (0.0794) R Adjusted R F-statistic Prob (F-statistic) *** *** 1 We exclude 11 observations that we consider outliers CG = unweighted corporate governance index, PBV = price to book value ratio, QREP = P3LKE index, DFAM = family ownership dummy variable (1 if family ownership > 50% and not conglomerate firm, 0 otherwise), INST = percentage institutional ownership, GOVT = percentage government ownership, FOR = percentage foreign ownership, SIZE = log total assets, AGE = listing age, OPINI = external auditor opinion dummy variable (1 if unqualified opinion, 0 if other opinion), LEV = leverage, GROWTH = growth opportunity, DYEAR = dummy year (1 for 2003 and 0 for 2004) *** Test variable is significant at 1% level ** Test variable is significant at 5% level * Test variable is significant at 10% level Three control variables in Equation 1, GOVT, FOR, and SIZE, significantly affect CG. All the three variables have positive signs. The positive signs of GOVT and FOR variables show that firms with higher government and foreign ownership tend to implement better GCG. We also find that larger firms tend to have better corporate governance. Other control variables, GROWTH, LNAGE, and DYEAR are not significant.

17 Table 6 also shows that none of the explanatory variables in Equation 2 (dependent variable = QREP) is significant. One possible explanation is as follows. We use P3LKE score as our proxy for disclosure quality. Because P3LKE is mandatory, it is possible that the firms try to strictly follow the standard. In that case, it may be logical to find that none of the independent variables explain the variation in QREP variable. Other explanation is that we may include bias in our calculation of the score. We also find that QREP variable is not significant in Model 3 (dependent variable = PBV). This finding shows that compliance to P3LKE may not be considered an important factor in improving the firm value. One possible explanation is that the market does not consider P3LKE compliance as value relevant because it is mandatory. Consistent with our expectation, CG positively affects PBV. This result means that higher corporate governance practice cause higher firm value. The magnitude of the slope is also large showing that an increase in corporate governance index is associated with a large increase in firm value. We also find that institutional ownership variable, INST, is positive and significant. This finding shows that firm with higher institutional ownership have higher value. It is possible that institutional investors put some control on the firm. Combined with the result of the first equation, institutional owners seem to put some control on the firm but do not give enough attention on the firm corporate governance. We find that DFAM does not affect PBV showing that family ownership does not affect firm value for our sample firms. Control variables FOR, GOVT, and SIZE are significant but with different signs from our expectations. For example, higher foreign ownership is associated with lower firm value. One way to interpret this result is that assuming foreign investors are well-informed and conduct more extensive portfolio analysis; they seem to invest in undervalued firms. Negative coefficient for GOVT and SIZE are not that surprising. Negative image of government ownership affects the firm value. Large firm tends to have high book value and therefore, small PBV. The other three control variables, GROWTH, DYEAR, and LEV, are significant and positive. Consistent with our expectation the higher a firm GROWTH rate the higher is its PBV. DYEAR variable shows that year 2003 PBV is higher than that of year the leverage variable (LEV) is positive showing that the tax and additional control benefit of debt seem to dominate the cost of higher default risk in our sample. Consistent with previous study, the higher the leverage level the higher the value of the firm due to tax-shields and leverage as an additional control to discipline the management. SENSITIVITY ANALYSIS Weighted CG Index In calculating the CG index in the main model we do not use weight. In this sensitivity analysis, we follow the IICD and use the following weights: 1. Rights of shareholders 20% 2. Equal treatment of shareholders 15% 3. Role of stakeholders 15% 4. Disclosure and transparency 25% 5. Board responsibilities 25%

18 Table 7 shows the result of the weighted corporate governance. The three tests show consistent results with the main analysis. In other words, the use of weighted CG does not affect the main findings. Table 7. Regression Results Using Weighted Corporate Governance Index Model 1 WCG it = a 0 + a 1 PBV it + a 2 DFAM it + a 3 INST it + a 4 GOVT it + a 5 FOR it + a 6 SIZE it + a 7 GROWTH it + a 8 LNAGE it + a 9 DYEAR t + ε it Model 2 QREP it = b 0 + b 1 PBV it + b 2 DFAM it + b 3 INST it + b 4 WCG it + b 5 GOVT it + b 6 FOR it + b 7 SIZE it + b 8 GROWTH it + b 9 OPINI it + b 10 DYEAR t + ε it Model 3 PBV it = c 0 + c 1 WCG it + c 2 DFAM it + c 3 INST it + c 4 QREP it + c 5 GOVT it + c 6 FOR it + c 7 SIZE it + c 8 GROWTH it + c 9 LEV it + c 10 DYEAR t + ε it Est. Coeff. Est. Coeff. Est. Coeff. (Probability) (Probability) (Probability) Predicted WCG Predicted QREP Predicted PBV Variable Sign Sign Sign N Intercept ** (0.7082) (0.1642) (0.0181) WCG ** (0.3086) (0.0454) PBV (0.6846) (0.5094) QREP (0.5037) DFAM +/ / / (0.5208) (0.6498) (0.3712) INST ** * (0.0145) (0.3540) (0.0540) GOVT *** * (0.0081) (0.3332) (0.0840) FOR *** * (0.0008) (0.3237) (0.0773) SIZE *** * (0.0000) (0.2602) (0.0665) GROWTH ** (0.5215) (0.7976) (0.0405) LNAGE (0.6758) OPINI (0.3377) LEV *** (0.0017) DYEAR * (0.4356) (0.2982) (0.0819) R Adjusted R F-statistic Prob(F-statistic) *** *** 1 We exclude 11 observations that we consider outliers WCG = weighted corporate governance index, PBV = price to book value ratio, QREP = P3LKE index, DFAM = family ownership dummy variable (1 if family ownership > 50% and not conglomerate firm, 0 otherwise), INST = percentage institutional ownership, GOVT = percentage government ownership, FOR = percentage foreign ownership, SIZE = log total assets, AGE = listing age, OPINI = external auditor opinion dummy variable (1 if unqualified opinion, 0 if other opinion), LEV = leverage, GROWTH = growth opportunity, DYEAR = dummy year (1 for 2003 and 0 for 2004) *** Test variable is significant at 1% level ** Test variable is significant at 5% level * Test variable is significant at 10% level

19 Tobin s Q Previous studies use Tobin s Q as proxy for firm value. In this sensitivity analysis we use Tobin s Q instead of PBV that we calculate as follows: ( MVE + DEBT) Tobin' sq = ( BVE + DEBT) where: MVE = market value of equity DEBT = book value of debt BE = book value of equity We use book value of debt for both the numerator and denominator because market value of debt is not available in Indonesia. Table 8 shows the regression result. The results are different from the main tests, especially for model 1 and 3. Like PBV, in Equation 1 Tobin s Q does not affect CG. INST and FOR variables that were previously significant are now insignificant. In Equation 3, none of the variables are significant. It is possible that there is a omitted variable problem in the model. ANALYSIS OF INDEXES P3LKE Compliance Index We find an improvement in compliance to the P3LKE from in 2003 to in However, our mean-difference test shows that the improvement is not statistically significant. Based on industry, we find improvements in investment, toll road, hotels, restaurants, transportation, and farm industries. In manufacturing, construction, trade, real estate, and forestry we find lower index score for The P3LKE compliance is still low and does not show significant improvement. This finding should become the concern of the BAPEPAM that issues the P3LKE. BAPEPAM needs to monitor more closely the disclosure quality of Indonesian public firms and analyze the problems firms are facing in complying with the P3LKE. Corporate Governance Index Although there is improvement in the score average from 2003 (0.6413) to 2004 (0.6483), the difference is not statistically significant. We think that the average score of 66% is still very low and need to improve their CG score. 2 2 Details results are available upon request

20 Table 8. Regression Results Using Tobin's Q Model 1 CG it = a 0 + a 1 TQ it + a 2 DFAM it + a 3 INST it + a 4 GOVT it + a 5 FOR it + a 6 SIZE it + a 7 GROWTH it + a 8 LNAGE it + a 9 DYEAR t + ε it Model 2 QREP it = b 0 + b 1 TQ it + b 2 DFAM it + b 3 INST it + b 4 CG it + b 5 GOVT it + b 6 FOR it + b 7 SIZE it + b 8 GROWTH it + b 9 OPINI it + b 10 DYEAR t + ε it Model 3 TQ it = c 0 + c 1 CG it + c 2 DFAM it + c 3 INST it + c 4 QREP it + c 5 GOVT it + c 6 FOR it + c 7 SIZE it + c 8 GROWTH it + c 9 LEV it + c 10 DYEAR t + ε it Est. Coeff. Est. Coeff. Est. Coeff. (Probability) (Probability) (Probability) Predicted CG Predicted QREP Predicted TQ Variable Sign Sign Sign N Intercept (0.3644) (0.9632) (0.1463) CG (0.5956) (0.2870) TQ (0.7030) (0.5597) QREP (0.8986) DFAM +/ / / (0.7741) (0.4872) (0.5804) INST (0.3038) (0.4062) (0.4018) GOVT ** (0.0281) (0.5131) (0.3329) FOR (0.2710) (0.3903) (0.3935) SIZE * (0.0635) (0.3816) (0.3578) GROWTH (0.4717) (0.7594) (0.2511) LNAGE (0.9835) OPINI (0.3382) LEV (0.5981) DYEAR (0.4865) (0.5698) (0.3472) R Adjusted R F-statistic Prob(F-statistic) *** We exclude 11 observations that we consider outliers CG = unweighted corporate governance index, TQ = Tobin's Q, QREP = P3LKE index, DFAM = family ownership dummy variable (1 if family ownership > 50% and not conglomerate firm, 0 otherwise), INST = percentage institutional ownership, GOVT = percentage government ownership, FOR = percentage foreign ownership, SIZE = log total assets, AGE = listing age, OPINI = external auditor opinion dummy variable (1 if unqualified opinion, 0 if other opinion), LEV = leverage, GROWTH = growth opportunity, DYEAR = dummy year (1 for 2003 and 0 for 2004) *** Test variable is significant at 1% level ** Test variable is significant at 5% level * Test variable is significant at 10% level CONCLUSION AND LIMITATION CONCLUSION This study examines simultaneous relations among corporate governance, disclosure and reporting quality, and firm value. Our results show that there are interdependencies between CG, QREP, and PBV. This suggests that OLS should not be used and therefore, we use 2 SLS in our tests.

21 We find that PBV does not affect corporate governance index which indicates that firms with high value may not try to improve their corporate governance in order to increase their value. We do not find empirical evidence that support our hypothesis that family ownership improves corporate governance. This indicates that there is no significant difference in GCG index between firms with dominant family ownership and non-conglomerate and GCG index of other firms. Contrary to our prediction, institutional ownership negative affect corporate governance. It is possible that institutional investors in Indonesian firms are short-term oriented investors that may be more beneficial with low good corporate governance practice. Three control variables (GOV, FOR, and SIZE) are significant and consistent with our prediction. None of the variables in Equation 2 is significant. One possible explanation is that because P3LKE is mandatory, it is possible that firms follow the standard without considering other factors. In that case, it may be logical to find that none of the independent variables explain the variation in the P3LKE compliance. Equation 3 shows that QREP is not significant. This finding may indicate that P3LKE compliance is not considered important factor to increase market value by public firms in JSX or there is uniformity in complying with the P3LKE. Consistent with our hypothesis, CG and INST are positive and significant. Firms with higher corporate governance index have higher value. The coefficient is large showing that the impact of improving good corporate governance practice is large on firm value. Control variables FOR, GOVT, and SIZE are significant but with different sign. Higher foreign ownership is associated with lower firm value. One way to interpret this result is that assuming foreign investors are well-informed and conduct more extensive portfolio analysis, the value of those firms are the intrinsic value while other firms are actually overvalued. Negative coefficient for GOVT and SIZE are not that surprising. Negative image of governance ownership affects the firm value. Large firm tends to have high book value and therefore, small PBV. Three control variables (GOV, FOR, and SIZE) are significant and consistent with our prediction. In our sensitivity tests, we find generally consistent results except when we replace PBV with Tobin s Q. We find slight and insignificant improvement in P3LKE score from in 2003 to in This score is still relatively low. Similarly, there is insignificant change in CG score and overall, the score is still relatively low. LIMITATIONS Our corporate governance index is calculated using secondary data that may not reflect the true corporate governance implementation and effectiveness in a firm. Ideally we should also use primary data such as interviews, questionnaires, direct observation of governance activities, etc. There is also a possibility of bias in valuing disclosure quality because we have difficulties in deciding whether an item is not disclosed or it is not applicable for that firm.

22 REFERENCES [1] Agrawal, A. and C. R. Knoeber, Firm Performance and Mechanisms to Control Agency Problems between Managers and Shareholders Journal of Financial and Quantitative Analysis (1996), pp [2] Agrawal, A. and S. Chadha, Corporate Governance and Accounting Scandals Working Paper, University of Alabama (2004). [3] Alves, C. and V. Mendes, Corporate Governance Policy and Company Performance: the Portuguese Case Corporate Governance: An International Review (2004). [4] Anderson, R.C., S.A. Mansi, and D.M. Reeb, Founding Family Ownership and the Agency Cost of Debt Journal of Financial Economics (2003), p.263. [5] Anderson, R. C. and D. M. Reeb, Founding-family Ownership and Firm Performance: Evidence from the S&P 500 Journal of Finance (2003), pp [6] Arifin, Z., Masalah Agensi dan Mekanisme Kontrol pada Perusahaan dengan Struktur Kepemilikan Terkonsentrasi yang Dikontrol Keluarga: Bukti dari Perusahaan Publik di Indonesia, Unpublished Dissertation University of Indonesia (2003). [7] Baek, J-S, J-K Kang, and K.S. Park, Corporate Governance and Firm value: Evidence from the Korean Financial Crisis Journal of Financial Economics (2004), pp [8] Barnhart, S. W. and S. Rosestein, Board Composition, Managerial Ownership, and Firm Performance: An Empirical Analysis The Financial Review (1998), pp [9] Beiner, S., W. Drobetz, M.M. Schmid, and H. Zimmermann, An Integrated Framework of Corporate Governance and Firm Valuation Working Paper (2005). [10] Black, B.S., H. Jang, and W. Kim, Does Corporate Governance Affect Firm Value? Evidence from Korea Working Paper (2003). [11] Botosan, C. A., Disclosure Level and the Cost of Equity Capital The Accounting Review (1997), pp [12] Bushee, B. J., The Influence of Institutional Investors on Myopic R&D Investment Behavior The Accounting Review (1998), pp [13] Chau, G. K. and S. J. Gray, Ownership Structure and Corporate Voluntary Disclosure in Hong Kong and Singapore The International Journal of Accounting (2002), pp [14] Cheng, C. S. A. and A. Reitenga, Characteristics of Institutional Investors and Discretionary Accruals Working Paper (2001). [15] Chung, K. H. and S. W. Pruitt, Executive Ownership, Corporate Value, and Executive Compensation: A Unifying Framework Journal of Banking and Finance (1996), pp [16] Claessens, S., S. Djankov, J. Fan, L. Lang, Disentangling the Incentive and Entrenchment Effects of Large Shareholdings Journal of Finance (2002), pp [17] Demsetz, H. and K. Lehn, The Structure of Corporate Ownership: Causes and Consequences Journal of Political Economy (1985), pp [18] Durnev, A. and E. H. Kim, To Steal or not to Steal: Firm Attributes, Legal Environment, and Valuation Journal of Finance (2005), pp [19] Eng, L. L. and Y. T. Mak, Corporate Governance and Voluntary Disclosure Journal of Accounting and Public Policy (2003), pp [20] Financial Statement Roundtable, Financial Economists Roundtable Statement on Institutional Investors and Corporate Governance (1999), pp.5-7.