VIEWS: 0 PAGES: 42 POSTED ON: 7/19/2013
Agency Conflicts, Expropriation and Firm Value: Evidence from Securities-Market Regulation in China Henk Berkman Massey University, Auckland Rebel A. Cole DePaul University, Chicago Jiang Fu The University of Auckland Summary and Key Findings • Event study of regulatory changes intended to improve corporate governance in China • Announcement of regulations to protect investors resulted in higher firm valuations. (LLSV JF 2002) • Securities-market regulation created significant value for investors in a civil-law country with weak judicial enforcement (Glaeser, Johnson and Shleifer QJE 2001). • In a rule-based civil law country, regulation in form of “bright lines” are more effective than “broad standards.” (Black and Kraakman HLR 1996) Placement in the Finance Literature: Corporate Governance • Shleifer and Vishny (JF 1997): How do minority shareholders and creditors ensure that they receive a pro-rata return on their investment? • Johnson, La Porta, Lopes de Silanes and Shleifer (AER 2000): How can minority investors be protected from “tunneling,” which they define as ”the expropriation of profits and assets by controlling shareholders.” The Changing Focus of Corporate Governance: • Old Focus: Solutions to principal-agent problems arising from the separation of ownership from control. – How do we align interests of a firm’s managers with those of the firm’s shareholders? – Research focused on firms located in the U.S., where diffuse ownership is the norm. – Jensen and Murphy (HBR 1990): Median CEO ownership of U.S. firms was only 0.25%. The Changing Focus of Corporate Governance: • 1990s: (U.S.) researchers became aware that, outside of the U.S. and Japan, dispersed ownership is the exception rather than the rule. • Instead, ownership is concentrated in the hands of a few families or the State. • “Corporate Ownership Around the World,” – La Porta, Lopez de Silanes, Shleifer and Vishny (JF 1999) The Changing Focus of Corporate Governance: • Where ownership is characterized by controlling shareholders, the focus of corporate governance shifts: • New Focus: Protection of the rights of minority shareholders. • How to prevent controlling shareholder, rather than the firm manager, from expropriating the wealth of minority shareholders. The State as the Controlling Shareholder • Of particular interest is the case where the controlling shareholder is the State. • La Porta et al. (JF 1999) report that the State is the largest block holder for at least one quarter of the 20 largest traded firms in the 27 wealthy countries that they study. • Even in the U.S., there are large SOEs: – U.S. Postal Service: largest U.S. employer – Amtrack, the commuter railroad system – Air Traffic Control System – Public school systems across the U.S. The State as the Controlling Shareholder • Jones, Megginson, Nash, and Netter (JFE 1999): Share-issuance privatization (SIP) is the primary method for privatizing SOEs, and the State maintains control in most SIPs. – 1977-97: 556 SIPs, raising $232 Billion. – Mean/median size: $556 million / $104 million – Mean/median float: 43.9% / 35.0%. • This means that the State remained the controlling block holder in the majority of all SIPs. • Bottom Line: The State remains an important player in the control of many large and important firms around the world. The State as the Controlling Shareholder • State control creates a principal-agent problem because cash-flow rights are separated from control rights. – Government officials hold of the control rights. – Cash-flow rights belong to the taxpayers. – How do we protect the taxpayers from expropriation by government bureaucrats? Law and Finance: La Porta, Lopez-de-Silanes, Shleifer and Vishny • JFE 2000: “. . . legal protection of investors as a potentially useful way to think about corporate governance.” – Stronger protection => Better governance • JF 2002: Better governance => Higher valuation • JPE 1998: investor protection is largely a function of legal protection and the quality of enforcement. – Common law countries provide superior protection relative to civil law countries – Countries with poor investor protection develop substitutes, such as regulations mandating or prohibiting certain actions. Law and Finance in China • Stock markets developed during the 1990s, as more than 1,000 SOEs were privatized through share issuances. • Civil-law tradition with weak investor protection. • Developing judicial system, where courts have been reluctant to step in to protect investors. • Relatively strong securities market regulator: the China Securities Regulatory Commission (CSRC) Regulation and Finance: Regulatory vs. Judicial Enforcement • Glaeser, Johnson and Shleifer (QJE 2001): “ . . . Government regulation may be the cheapest way to bring about efficient allocation of resources.” – Securities market regulation can protect minority investors from expropriation by controlling shareholders. – “In economies with a developing judicial system, . . . the cost of regulatory enforcement may be cheaper than that of judicial enforcement.” – Empirical evidence from comparison of financial regulation in Czech Republic and Poland during 1990s. Stricter enforcement in Poland relative to Czech Republic led to superior development of financial markets. Regulation and Finance: “Bright Lines” vs. “Broad Standards” • Black and Kraakman (Harvard LR 1996): “Bright-Line” regulations are preferable to “Broad- Standard” regulations in transitional economies with civil law judiciaries. • Bright Lines: provide clear boundaries as to what is required and what is legal. • Broad Standards: set forth murky boundaries as to what is required and what is legal, relying upon the judicial interpretations by the courts to adapt the standards to an evolving regulatory environment. Regulation and Finance: “Bright Lines” vs. “Broad Standards” • Civil-Law judiciaries are less likely than Common- Law judiciaries to provide needed interpretations of “Broad-Standard” regulations, leaving regulations unenforced. • Regulations should rely upon enforcement by direct participants (e.g., shareholders and directors) to reduce the need for enforcement by indirect participants such as the judges and lawyers. Why Study China? • At least 1.3 billion reasons . . . . • Second largest economy on PPP-adjusted basis. • Chinese stock market expected to surpass Nikkei as second largest world equities market. • China makes a most interesting contrast to disastrous privatization efforts undertaken in the former Soviet Bloc: – Mass privatization in Eastern Europe using vouchers failed miserably. – Why has China’s approach worked so much better? What went wrong in Russia? • Black, Kraakman, Tarassova (Stanford LR 2000) • Need for time to develop institutions/infrastructure. – Legal system that protects minority shareholders and creditors – Accountants to audit financial statements – Prosecutors to go after law breakers – Judges and courts to arbitrate disputes – Informed investors to allocate capital – Securities regulators to set and enforce rules Why did China’s approach to privatization work better? • Staged privatization: – Slowly develop the necessary infrastructure and institutions – Give these institutions time to “practice” and develop expertise while the State maintains ultimate control and prevents wholesale looting. Staged Privatization in China: Numbers and Market Caps of Listed Firms 1993 -2000 Staged Privatization in China: State Ownership of Listed Firms 1993-2000 Key CSRC Regulatory Changes: 2000-2001 • New regulations for annual shareholder meetings that improved the voting rights and other powers of minority shareholders. • Prohibition against issuance of debt guarantees to shareholders or subsidiaries owned by shareholders. • Limitations on asset transfers between related parties. • Requirement that each listed firm include independent directors on its Board and provide them with resources necessary to carry out their duties. Data • Stock price data from Datastream for 500 trading days ending June 5, 2001. • Accounting data from 1999 annual reports. • In total, we have data for 850 firms. • Event windows: 1 day before CSRC issuance of each regulation through 1 day after first published report in financial press. Event-Study Methodology • We estimate cumulative mean-adjusted returns: Market Return t = β0 + β1 * Event 1 + β2 * Event 2 + β3 * Event 3 + β4 * Event 4 + ε t – Market Return t is the return on the equally weighted market portfolio of firms listed only on the Chinese stock exchanges on day t (excludes firms also listed in Hong Kong). – β 0 captures the mean return. – β J , J = 1 to 4, estimate cumulative mean-adjusted returns associated with each event window. Event-Study Methodology – Event J , J = 1 to 4, are dummy variables that take on a value of 1/N for each event window, where N is the number of days included in the window. – This definition means that the betas are the total CARs and not the average daily CAR, which would be the case if we used simply zero-one dummies. – We also estimate CARs using a market model, where the return on a portfolio of 24 Chinese firms that trade only on the Hong-Kong exchange serve as the market proxy. Table 2: Cumulative Returns by Event Intercept 0.000 -0.000 (0.09) (0.97) Event 1: Shareholder meeting 0.095 a 0.109 a (0.01) (0.01) Event 2: Related Guarantees 0.008 0.001 (0.83) (0.98) Event 3: Asset Transfer 0.032 0.024 (0.57) (0.68) Event 4: Independent Directors 0.026 0.021 (0.37) (0.48) Hong-Kong Return 0.075 a (0.01) Cross-Sectional Differences in CARs • Market-wide changes in value may obscure differential impacts upon firms with strong and weak governance. • Our Hypothesis: Firms with weak governance are more reliant upon legal and regulatory protection than are firms with strong governance. • Shareholders of firms with weak governance should benefit disproportionately from the new regulations improving investor protection. Cross-Sectional Differences in CARs • To test this hypothesis, we first calculate a firm-specific measure of governance based upon firm value as measured by Tobin’s Q. • Based upon Gompers, Ishii and Metrick (QJE 2003), who model Tobin’s Q as a function a vector of firm characteristics W i and a vector of corporate governance variables X i: Qi=a+bXi+cWi +ei Cross-Sectional Differences in CARs • First, we estimate b X i by orthogonalizing Q against W i , which we proxy by standard explanatory variables that appear in the literature: size, leverage, growth prospects and industrial classification. • What is left, we call Corporate Governance Q, or CG-Q. (see eq. 4 and Table 3.) • Bertrand, Mehta and Mullainathan (QJE 2002) use a similar approach in one of their tests of tunneling at Indian firms. Cross-Sectional Differences in CARs • Next, we estimate CARs for each firm using the same approach as in Table 2. • We then estimate a WLS regression of the form: • CAR i, event j = γ 0 + γ 1 CG-Q i + e i • The weights are the standard errors from the regressions used to estimate CAR for each firm. • Results appear in Panel A of Table 4. Table 4 Panel A Event 1: Event 2: Event 3: Event 4: Shareholder Related Asset Independent Meeting Guarantees Transfer Directors Intercept 0.001 0.002 0.004 0.002 (0.30) (0.34) (0.26) (0.21) -0.002 c -0.004 a -0.005 a 0.000 CG-Q (0.09) (0.01) (0.01) (0.99) Cross-sectional Differences in CARs: Econometric Considerations • We also use a portfolio time-series regression to test the cross-sectional differences in CARs. • This approach deals with the cross-correlations in firm returns that are likely to occur when the event date and event window are the same for all firms. • It provides unbiased estimates of coefficients along with standard errors that fully account for the cross- sectional heterosckesticity and cross-security dependence. (Sefcik and Thompson JAR 1986). CARs for “High CG-Q - Low CG-Q” Portfolio • We sort our sample firms into quantiles based on CG-Q. • We then calculate daily returns for a portfolio that is long on the highest quantile and short on the lowest quantile. • Hence, this portfolio return measures the difference in returns on the high CG-Q quantile portfolio and the low CG-Q quantile portfolio. • For quantiles, we use 5, 3 and 2 (fifths, thirds and halves). CARs for “High CG-Q - Low CG-Q” Portfolio Return (CG-Q Hight – CG-Q Lowt ) = β0 + β1 * Event 1 + β2 * Event 2 + β3 * Event 3 + β4 * Event 4 + β5 * Market Returnt + ε t where: QHight is the return for day t on an equally weighted portfolio of the highest quantile firms based upon industry-adjusted Q; QLowt is the return for day t on an equally weighted portfolio of the lowest quantile firms based upon industry-adjusted Q Market Return t is one of two proxies for the market return: the return on the HK portfolio. For each event, we expect β J to be negative. Table 4 Panel B Event 1: Event 2 Event 3: Event 4: Shareholder Related Asset Independent Meeting Guarantees Transfer Directors 5 Groups -0.024 c -0.032 a -0.054 a -0.001 (0.07) (0.01) (0.01) (0.92) 3 Groups -0.018 c -0.025 a -0.049 a 0.001 (0.07) (0.01) (0.01) (0.91) 2 Groups -0.015 b -0.022 a -0.049 a 0.001 (0.04) (0.01) (0.01) (0.85) Multivariate Cross-Sectional Analysis • We also examine how CARs co-vary with several directly observable firm characteristics, including several proxies for the quality of corporate governance. • Following La Porta et al. 2002, we assume that the “ultimate owner” of the firm has effective control over the firm. • We then consider three variables that might mitigate the ultimate owner’s incentive or ability to expropriate minority shareholders. 1. Cash-flow rights of the ultimate owner 2. Control rights of the ultimate owner relative to other block holders (specifically, the 2nd and 3rd largest block holders). - ln [ (Shares Largest) / (Shares 2nd and 3rd Largest) ] 3. The issuance of B-shares to primarily institutional foreign investors. Multivariate Cross-Sectional Analysis • We also test whether ultimate State control is a plus or minus by including a dummy variable indicating State control. • Xu and Wang (China Ec Rev 1999) suggest that the State is more likely to expropriate minority shareholders than are private block holders. • The theoretical model of Perotti (AER 1995) suggests the opposite, as does empirical research on privatization in the former Soviet bloc countries (Black et al (SLR 2000); Coffee (J Corp Law 1999)). • Our hypothesis: Cash-flow rights belong to taxpayer rather than government officials who exercise control rights, so the State is less likely to expropriate than a private controlling block holder. • Firm size and leverage are included as control variables. Multivariate Cross-Sectional Analysis First, we estimate the following WLS regression model: CARi, event J = β0 + β1,J * CF-Rights of Ultimate Owner i + β2,J * Large Shareholder Dominance Ratio i + β3,J * State is Largest Dummy i + β4,J * B-Shares Dummy i + β5,J * Leverage i + β6,J * ln (Assets) i + εiJ Results appear in Table 5 in the WLS columns. Multivariate Cross-Sectional Analysis: Econometric Considerations • As before, there are econometric problems with this WLS regression approach. • Given the contemporaneous event periods for the sample firms, there is cross-correlation in the firm return processes from which the CARs are estimated. • More complicated here than with CG-Q because we have multiple explanatory variables. • Sefcik and Thompson (1986) propose a method to obtain unbiased estimates of the cross-sectional regression coefficients, along with standard errors that fully account for cross-sectional heteroscedasticity and cross-security dependence. Multivariate Cross-Sectional Analysis: Portfolio Time-Series Regression This portfolio time-series approach involves three steps: 1. Orthogonalize each of our six explanatory variables against the others and a set of 25 industry dummies. 2. For each of the six orthogonalized variables OV, we form a portfolio that is long the highest quantile of that variable and short the lowest quantile of that variable. 3. Regress the returns of this high-long/low-short portfolio on the market return and a set of four dummy variables representing each event window: Multivariate Cross-Sectional Analysis: Portfolio Time-Series Regression So our regression model is: Return (High Quantile) t – Return (Low Quantile) t =β0 + β 1 * Event 1 + β 2 * Event 2 + β 3 * Event 3 + β 4 * Event 4 + β 5 * Market Return t + εt The betas from these time-series regressions are equivalent to the cross-sectional betas from the WLS regression of CARs on the explanatory variables. Multivariate Cross-Sectional Analysis: Portfolio Time-Series Regression Summary • In aggregate, the four regulatory events we study resulted in a statistically significant 16 percent increase in firm value. • Moreover, firms with weak governance benefited disproportionately relative to firms with strong governance. • For three of the four events, we find statistically significant cross-sectional differences in CARs that are supportive of theories of corporate governance. • Only for our “broad-standard” event, the requirement of independent directors, do we fail to find significant results. Conclusions • Our results provide new evidence supporting the theoretical model of La Porta et al. 2002, which predicts that better investor protection is rewarded with higher firm valuations. • Our results demonstrate that regulations to protect minority shareholders in a civil-law country with a weak judiciary can create significant value for minority shareholders, consistent with Glaeser, Johnson and Shleifer (2001) • Our results are consistent with Black and Kraakman (1996), who argue that “bright-lines” regulations are preferable to “broad standards” in a transitional economy with a weak judiciary.
Pages to are hidden for
"From State to State Improving corporate governance when the "Please download to view full document