6th ICFC, 2014 THE RELATIONSHIP BETWEEN CORPORATE GOVERNANCE AND THE STOCK PRICE OF LISTED COMPANIES IN MAI USING FELTHEM – OHLSON VALUATION MODEL Supranee Sugaraserani September 4 , 2014 6th ICFC, 2014 Do the investors use publicized CG Rating in trading stocks in MAI? Company Summary SCG Symbol Resources/ Energy & Utilities CG SET Score: SAHACOGEN (CHONBURI) PUBLIC COMPANY LIMITED Data as of 31/07/2014 SCG Price 52 week P/E P/BV Paid-up Market Cap EBITDA (B.) 5.00 High/Low 5.25 / 3.58 (X) 16.63 (X) 1.91 (MB.) 955.00 (MB.) 4,775.00 (MB.) 202.96 6th ICFC, 2014 THE OBJECTIVE OF THE STUDY To examine the relationship between corporate governance (CG) and the stock price of listed companies in the Market for Alternative Investment (MAI) by using a valuation model 6th ICFC, 2014 Corporate Governance (CG) Rating • annual surveys on the state of corporate governance by the Thai Institute of Directors (IOD) • since 2001 (2544) • The sources for scoring are – company annual report, annual information filling (Form 56-1), notice and minutes of companies' shareholders meeting, company website, information on SET/SEC database, and other publicly available information. • After scoring, listed companies are classified into six groups. 6th ICFC, 2014 Figure 1: Classified CG scores, Logo, and description Score Range number of Logo (publicized) Description 90-100 80-89 70-79 Excellent 60-69 50-59 Satisfactory Lower than 50 Very Good Good Pass No logo given N/A Source: Corporate Governance Report of Thai Listed Companies (CGR) 2006. 6th ICFC, 2014 Previous studies - CG • Black (2001), Gompers, Ishii and Metrick (2003), found positive correlation between CG and firm performance. • Drobetz, Schilhofer and Zimmerman (2004), Durnev and Kim (2007), Klapper et al. (2002), Black, Jang and Kim (2006) found positive impact of CG to firm‘s value and investors gave their attention to performance of the good CG firms, especially in countries with weaker legal systems. • Morey et al. (2009) studied the correlation between governance quality and share prices in emerging markets, using samples from 21 countries, they found that improvements in CG were associated with higher share prices. 6th ICFC, 2014 Basic FO valuation model Pt = ₀+₁BVAt +₂BVLt + ₃AEt + ѵ + + Where: • Pt = • BVAt = • • • • BVLt AEt ¹, ² ℰ = = = = stock price at period t book value of assets at period t book value of liabilities at period t abnormal earnings of period t other non-accounting information tolerance value 6th ICFC, 2014 Main assumptions of FO model • The relevance of publicly accounting information and firm‘s market value. • The PV of expected future dividends determines the intrinsic firm value (considering a fair game). 6th ICFC, 2014 • Clean Surplus Relation - concept derives from conservative accounting. • Linear Information Model - the relationship of control variables must exist in a linear equation (Ota, 2002). 6th ICFC, 2014 Pt = ₀+₁BVAt +₂BVLt + ₃AEt + ѵ + + Qualitative data Decoded using 4-level rating scale YES Test Clean Surplus Relation 6th ICFC, 2014 Transformed model Pt = β0 + β1CGt-1BVAt + β2CGt-1BVLt + β3CGt-1AEt + β4 CGt-1 + ℰ 6th ICFC, 2014 Hypotheses H1: The relationship of CG rating meets the assumption of Clean Surplus Relation H2: The relationship of stock price, CG rating, firm‘s size, liability, and abnormal earning must exist in a linear equation. 6th ICFC, 2014 Limitations • Population The data were 45 listed companies in MAI which their securities were first traded before 2008 (2551) and had reported and disclosed of financial information and others non-accounting information, through the year 2012 (2555). continuously, 6th ICFC, 2014 • Variable Measurement Variable Pt BVAt Definition Firm’s value at period t Firm’s size at period t Calculation and Measurement Market price per share at the end of period Common logarithm of moving average of the book value of assets at the end of period BVLt Liability at period t Common logarithm of moving average of the book value of liabilities at the end of period AEt Operating efficiency or The percentile ranks of abnormal earnings. abnormal earning at Abnormal earning is the difference between period t moving average of operating profit after tax and a return, at risk free rate, to the moving average of equity at the beginning of period. CGt-1 Corporate governance rating at previous period a) FO valuation model- using 4-level rating scale instead of the volume of logo. b) Testing clean surplus relation- a ceiling score of each group will be used instead of the volume of logo. 6th ICFC, 2014 RESULT OF THE STUDY H1: The relationship of CG rating meets the assumption of Clean Surplus Relation Using Simple Regression: 6th ICFC, 2014 Table 2: Simple regression analysis output Significant level at 0.05 (n=45) 6th ICFC, 2014 • The Corporate Governance rating has met the assumption of the Clean Surplus Relation . • The longer • As the constants’ coefficient is decreasing, Curve Fit is tested and the the predictions from IOD conveys its survey the equations without constants are higher the power of the better, so, the constants are impact of current period removed from equations. CG to following period CG (R2) 6th ICFC, 2014 The relationship of CG rating meets the assumption of Clean Surplus Relation and the null hypothesis is rejected. CG09 = 0.567CG08 ………..……… (2.1) CG10 = 0.751CG09 ……………..… (2.2) 6th ICFC, 2014 H2: The relationship of stock price, CG rating, firm‘s size, liability, and abnormal earning must exist in a linear equation. Using Multiple Regression: 6th ICFC, 2014 Table 3: Multiple regressions analysis output Significant level at 0.05 Year 2009 (n=45) Control variables Unstandardized Coeff. Collinearity Statistics B Stand. Err. Toler. VIF constant 4.010 1.206 Firm Size09 3.751 1.829 .003 287.784 Liability 09 .271 .972 .014 70.819 Ab. earning09 2.154 1.077 .441 2.269 CG08 -24.887 6.885 .009 113.034 R2 = .415 Std. Error of the Estimate = 3.012932 F = 7.091 Sig. = .000 Durbin-Watson = 1.584 Year 2010 (n=45) Control variables Unstandardized Coeff. Collinearity Statistics B Stand. Err. Toler. VIF constant 2.752 1.621 Firm Size10 3.228 1.708 .005 214.078 Liability 10 -.578 .957 .015 64.624 Ab. earning10 .867 1.096 .459 2.177 CG09 -15.801 6.274 .014 73.654 R2 = .255 Std. Error of the Estimate = 3.429150 F = 3.430 Sig.=.017 Durbin-Watson = 1.903 Year 2011 (n=43 : delete 2 companies’ data because of the high volatility in stock price) Control variables Unstandardized Coeff. Collinearity Statistics B Stand. Err. Tolerance VIF constant .709 1.682 Firm Size11 3.822 1.535 .005 206.24 Liability 11 -1.163 .951 .013 76.824 Ab. earning11 -.573 .907 .446 2.240 CG10 -14.694 5.458 .015 67.091 R2 =.283 Std. Error of the Estimate = 3.171487 F = 3.755 Sig.=.011 Durbin-Watson =1.696 6th ICFC, 2014 Test the autocorrelation - Compare DW values to the critical value • 2009: the result cannot conclude whether there is an autocorrelation or not. • 2010: the result can conclude that there is no autocorrelation. • 2011: the result cannot conclude whether there is an autocorrelation or not. 6th ICFC, 2014 The findings have proved that all control variables can explain the change in stock price and the relationship exists in a linear equation. The null hypothesis is rejected. P09 = 4.01(+/-3.0129) + 3.75BVA09 + 0.27BVL09 + 2.15AE09 – 24.89CG08 ..……. (3.1) P10 = 2.75(+/-3.429) + 3.23BVA10 – 0.58BVL10 + 0.87AE10 – 15.80CG09 ……… (3.2) P11 = 0.71(+/-3.171487) + 3.82BVA11 – 1.16BVL11 - 0.57AE11 – 14.69CG10 …….... (3.3) 6th ICFC, 2014 SKEPTICISM ISSUES FROM THE FINDING • FO model is predictive valuation model that emphasizes on growth. So, all control variables shall have positive correlation. 2 control variables that have negative correlation – CG rating and liability. a) CG rating – the finding implies that investors do not pay attention to the CG rating and do not use it for decision making (Bauer et al., 2004, Beiner et al., 2006). 6th ICFC, 2014 b) Liability - the liability and equity of listed companies in MAI at the time of the study may not be relevant to determine the firm‘s value and its operating results in the future (Modigliani and Miller, 1958). 6th ICFC, 2014 • Explanatory power of the models - show a weak power of explanation. During 2008-2011, Thailand had economic downturn and that would cause abnormalities to the capital market. The investors‘ decision to weigh the risks with the benefits received would be different from normal capital market (Fama, 1965, 1969). Typically, investors' behavior, securities could not be evaluated in isolation, but only as a group Markowitz (1952).