Dear Judith: - University of Michigan
Current Draft: September, 2002
First Draft: April, 2002
Comments welcome
To Steal or Not to Steal:
Firm Attributes, Legal Environment, and Valuation
Art Durnev* and E. Han Kim**
GEL Classification: G32 (Financial Policy; Capital and Ownership Structure), K23 (Corporation and Securities Law)
Keywords: Corporate Governance, Investment Opportunities, External Financing, Ownership, Legal Environment
ABSTRACT
A simple model is presented in which the controlling shareholder chooses the quality of corporate governance under different legal regimes. The model identifies three firm-specific attributes that affect governance: investment opportunities, reliance on external financing, and ownership structure. Using firm-level corporate governance and transparency data on 859 firms in 27 countries, which reveal a wide variation in corporate governance and disclosure practices across firms within countries, we find results that are consistent with the hypotheses derived from the model. Firms with profitable investment opportunities, more reliance on external financing, and more concentrated ownership have higher-quality governance and disclose more. And firms with higher governance and transparency ratings are valued higher and invest more. Moreover, these relations are stronger in countries that are less investor-friendly, demonstrating that individual firms do adapt to poor legal environments to achieve efficient governance practices.
This paper examines the relation between firm-specific factors and corporate governance practices and how governance practices are in turn related to firm valuation. To this end we develop a simple model that shows firms with good investment opportunities, greater reliance on external financing, and higher ownership concentration practice better governance. The model also predicts that the relations are stronger in legal environments that are less investor friendly and that firms with better governance are valued higher. We test these predictions with newly released data on governance and disclosure practices for 859 firms in 27 countries. The results are consistent with the predictions of the model.
Previous studies have examined the effects of legal environment on corporate governance and the relation between corporate governance and firm valuation. Specifically, it has been shown that better legal protection for investors is associated with higher value of stock market (La Porta et al. (1997)), higher valuation of listed firms relative to their assets or changes in investments (Claessens et al. (2002), La Porta et al. (2002), Wurgler (2000)), and larger listed firms in terms of their sales and assets (Kumar et al. (1999)). Furthermore, industries and firms in better legal regimes rely more on external financing to fund their growth (Rajan and Zingales (1998) and Demirgüç-Kunt and Maksimovic (1998)).
Although these studies provide valuable insights into the effects of regulatory environment, they do not address firm-level issues such as what drives different governance practices across firms within a given legal regime and how corporate governance affects individual firm valuation. Current studies attempting to address these issues include Black (2001) and Black et al. (2002), who demonstrate a strong relation between corporate governance and firm valuation in Russia and Korea, and Doidge et al. (2001), who show that foreign firms listed on U.S. stock markets are valued higher. For a more complete survey on international corporate governance see Denis and McConnell (2002).
The relation between corporate governance and firm valuation has also been studied for U.S. firms (e.g., Bhagat and Brickley (1984), Bhagat and Jefferis (1991), Denis et al. (1997), Karpoff (1998), Demsetz and Lehn (1998), Denis (2001), Gompers et al. (2002)). Most of these studies show mixed results with the exception of Gompers et al., who find that corporate governance provisions related to takeover defenses are significantly related to firm valuation.
The contribution of this paper is three-fold. First, we identify three firm attributes that may influence individual firms’ choice of the quality of corporate governance within a given legal environment and empirically examine the hypothesized relations. Second, we extend previous studies on the relation between corporate governance and firm valuation using a large sample of firms for 27 countries. Finally, we document that the relations between firm attributes, governance practices, and firm valuation are stronger in less investor friendly countries, suggesting individual firms adapt to poor legal environments to effect efficient governance practices.
To identify the relevant firm attributes, we provide a simple model in which the controlling shareholder faces a trade-off between diverting resources for private benefits and sharing gains with minority shareholders by investing in profitable projects. Diversion of resources results in rejection of profitable projects, which lowers the value of the controlling stockholder’s share of the firm value. The diversion also imposes on the controlling shareholders private costs that depend on the quality of legal environment.
The model predicts that the quality of corporate governance in a given legal environment is positively related to the availability of profitable investment opportunities, concentration of ownership, and the need for external financing. It also shows firm valuation is positively related to the quality of governance practices. Equally important, it predicts that the above relations are stronger in weaker legal regimes.
The basic intuitions underlying the model are simple. One is less likely to commit crime if one has something valuable to lose–profitable investment opportunities; one does not spit into the well that one drinks from–external financing; one does not steal from oneself–ownership concentration; good rules and effective enforcement reduce thievery–legal environment. As for the interplay between firm attributes and legal environment, good corporate governance driven by private incentives becomes a more important complement to regulation when regulation is less effective. And of course, good corporate governance is valued higher where it is scarce–the stronger relation between firm valuation and corporate governance in weaker legal regimes.
To examine these issues we use data on corporate governance and transparency practices on 859 firms in 27 countries. The data show a wide variation in governance and disclosure practices. The variation appears to be greater in weaker legal regimes, dispelling the stereotype that all firms in weak legal regimes suffer from poor corporate governance and all firms in strong legal regimes practice uniformly high-quality corporate governance.
The data are consistent with the predictions of the model. We find that firms with better investment opportunities, higher concentration of ownership, and more reliance on external financing practice better governance. Furthermore, these relations are stronger in legal regimes that are less investor-friendly. Thus, in Coase’s sense (Coase (1960)) firms located in countries with poor legal protection appear to show the adaptability to achieve efficient governance practices.
The data also reveal that firms with better governance invest more and enjoy higher valuation. This is consistent with the contemporaneous findings of Klapper and Love (2002) who examine the relation between firm valuation and corporate governance using a set of data that partially overlaps with our data.
Section I presents the simple model, followed by hypotheses and empirical design in Section II. Section III describes data, Section IV reports empirical results, with Section V providing robustness checks. The concluding section contains summary and implications.
I. A Simple Model
This model assumes an environment that is similar to Johnson et al. (2000) and Shleifer and Wolfenzon (2002). The primary purpose of the model is to provide motivation for empirical tests and as such it is devoid of many interesting real world complexities like determinants of initial ownership and capital structure in different legal regimes[1]. Nor does it allow for a simultaneous determination of investment and financing decisions. Following the Modigliani and Miller (1958) tradition of examining investment and financing decisions separately, we first assume financing is given and then allow for external financing by assuming investment is given. The model does not consider the reputation-type issues discussed in Diamond (1991), Maksimovic and Titman (1991), and Gomes (2000).
In defining corporate governance practice, we take Shleifer and Vishny’s view, “Corporate governance deals with the ways in which suppliers of finance to corporations assure themselves of getting a return on their investment.” (Journal of Finance, 1997, p. 737). Thus, the quality of governance practice is defined as the degree to which non-controlling, outside shareholders get their fair share of return on investment, or more specifically, as (1-d) where d is the proportion of cash flows that the controlling shareholder diverts for private benefits at the expense of other shareholders.
Defined as such, a good governance system would make it difficult for the controlling shareholders to steal d, which depends on attributes such as independence and accountability of the board of directors, financial incentives and managerial disciplines for value creation, enforcement of managerial responsibility and accountability, timely and accurate disclosure of relevant information, assurance to maintain auditors’ independence, and protection of minority shareholders[2]. Our definition of the quality of corporate governance captures these attributes and, hence, focuses on the overall practice of corporate governance instead of the standard notion of legally binding contracts that are designed to shape and constrain management actions.
We consider a single period model in which the profit per unit of physical capital invested in project j is equal to [pic], where [pic] and the cost of a unit of capital is standardized at 1. We assume that [pic] is linear and decreasing in j for all firms with each firm having a maximum of [pic], which varies across firms[3]. Thus, [pic] is the measure that distinguishes the level of profitable investment opportunities across firms. With this definition, the gross return for the jth unit of capital invested can be written as [pic]. For brevity and without loss of generality, we assume that the interest rate is zero and investors are risk-neutral. Thus, if a firm takes all non-negative NPV projects, it will invest until [pic], and the last unit of capital invested will be [pic]; that is, the firm’s total investment will be [pic], and its market value will be [pic]. Figure 1 summarizes the investment opportunity, as well as the firm’s optimal investment and market value when d is suppressed to be zero.
Finally, we assume that the firm will liquidate at the end of the period when the controlling shareholder, who holds ( fraction of the firm, collects her share of liquidating dividends. The ownership fraction α must be large enough to give her the controlling interest.
A. Internal Financing
To illustrate the effect of profitable investment opportunities on governance practice, we start with a firm that hasnd internal funds, [pic], where e is a constant that indicates whether the firm has sufficient funds to invest in all non-negative NPV projects (e ( 0) or not (e < 0). The controlling shareholder makes the investment decision subject to the budget constraint F [4]. She is also free to divert any portion of F for private benefits but there are associated costs.
The costs of diversion are of three types. The first is the (expected) penalties imposed on the controlling shareholder when the diversion is illegal and she gets caught. Such penalties take the form of fines, jail terms, and loss of reputation that may hurt her future business or employment opportunities. The second is the cost of diverting corporate resources, requiring such transaction costs as bribes to internal and external agents. Additionally, private consumption or conversion of the diverted resources into cash equivalents incurs deadweight loss because the consumption value is often less than the replacement costs.
We assume the sum of all these costs is a fixed fraction, c, of the amount diverted[5]. These costs vary across countries due to differences in de jure and de facto regulation, and across industries within a legal regime due to differences in the nature of assets (e.g., tangible vs. intangible) and business models involved. For instance, the costs of diversion may be higher for tangible assets that are more visible and for firms in regulated industries[6].
In this setup, the controlling shareholder may not invest in all positive NPV projects; instead, she will invest in project j only if her share of the liquidating dividends from the project is greater than the after-cost diversion,
[pic] [1]
The right-hand side of [1] incorporates the possibility that if the rate of return from the project, [pic], is negative, the controlling shareholder may keep the money rather than invest into a negative return project. For example, if [pic] but [pic] she will neither divert nor invest; instead, she will retain the funds within the firm, of which she has claim to α fraction. Thus, the controlling shareholder will invest up to the point where[7]:
[pic] [2]
Hence, the optimal level of investment for the controlling shareholder is[8]
[pic] [3]
The funds remaining after the investment, [pic], will be diverted if the after-cost diversion is greater than the controlling stockholder’s share of liquidating dividends from it: (F-j*)(1-c)>a(F- j*) or [pic]. Thus, the optimal amount of diversion D* is equal to [pic] if [pic], and 0 otherwise. Since [pic], it follows from Eq.[3] that:
[pic] [4]
Equation [4] shows that if the cost of diversion is high relative to non-controlling shareholders’ ownership such that [pic], there will be no diversion and all the leftover funds will be kept within the firm. This is because for each dollar of diversion, the cost c is greater than the wealth transfer from the minority shareholders who own [pic] fraction of cashflow rights. Diversion takes place only if the cost, c, is less than the wealth transfer per dollar of diversion, [pic][9].
Thus, our measure of corporate governance, d, the proportion of cash flows that are diverted as a percentage of the firm’s endowment, [pic], is:
[pic] [5]
It follows from [5] that d* is negatively related to the cost of diversion c. Therefore, in more investor-friendly countries (high-c countries), firms will divert less and have better corporate governance[10].
A.1 Investment Opportunities
Proposition 1: Controlling shareholders of firms with more profitable investment opportunities divert less for private gains.
Proof: Since [pic] represents the profitability of investment opportunities, we take the partial derivative of d* with respect to [pic]:
[pic] [6]
This derivative is non-positive because diversion takes place only when [pic]for reasons explained earlier.
Q.E.D.
The intuition behind Proposition 1 is simple. When investment opportunities become more profitable, the controlling shareholder’s share of return from investment increases relative to the benefits from diversion. Thus, firms with profitable investment opportunities will practice low d, i.e. high quality governance.
Conversely, when a firm suffers a substantial drop in profitable investment opportunities, the controlling shareholders will divert more corporate resources. Johnson et al. (2000) document such behavior before the Asian financial crisis and the media alleges similar actions by the top management of Enron and other firms with subsequent scandals prior to their bankruptcy filings[11].
A.2 Legal Regimes
Proposition 2: A given increase in profitable investment opportunities has a lesser impact on diversion in a country with high costs of diversion than in a country with low costs of diversion.
Proof: This result follows because the derivative of Eq. [6] with respect to c is non-negative,
[pic] [7]
Q.E.D.
Proposition 2 implies that the sensitivity of diversion to investment opportunities falls as the cost of diversion rises. That is, the positive relation between investment opportunities and the quality of corporate governance is stronger in weaker legal regimes[12]. To illustrate, consider two countries, say the U.S., which has relatively low tolerance for diversion and imposes high cost, and Russia, which is more tolerant and imposes lower costs. Proposition 2 implies that the same increase in profitable investment opportunities will have a smaller negative impact on d in the U.S. than in Russia.
To demonstrate numerically, assume the cost of diversion, c, is equal to 0.6 in the U.S. and 0.3 in Russia and the excess cash, e, is equal to 0 in all cases. Also consider a firm that has low investment opportunity [pic], and the controlling shareholder owns 30%. The payoff to the controlling shareholder can be defined as [pic] where the market value, MV, is the present value of gross returns from all projects that are undertaken. Since[pic], the optimal diversion is [pic] if the firm is located in Russia, but [pic] if in the U.S. Now, if the firm’s investment opportunity improves to [pic], the optimal diversion will decrease from 0.67 to 0.61 for the firm in Russia, whereas for the U.S. firm the decrease will be smaller, from 0.17 to 0.15 (see Figure 2).
This example can also be used to illustrate the relation between variation in the quality of corporate governance and legal regime[13].
Corollary 1: If the mean and variation in profitable investment opportunities across firms are held constant, then variation in diversion is greater in weaker legal regimes [14].
A.3 Ownership
Proposition 3: Controlling shareholders of firms with higher ownership divert less corporate resources for private gains.
Proof: Differentiating Eq. [5] with respect to controlling shareholder’s ownership we obtain
[pic] [8]
Q.E.D.
Proposition 3 is a restatement of the well-known Jensen and Meckling (1976) agency argument that entrepreneurs with higher ownership divert less.
Proposition 4: A given increase in controlling shareholder’s ownership has a lesser impact on diversion in a country with high costs of diversion than in a country with low costs of diversion.
Proof: Differentiating [8] with respect to the cost of diversion c,
[pic] [9]
Q.E.D.
Proposition 4 implies that the sensitivity of diversion to ownership concentration falls as the cost of diversion rises. That is, the positive relation between ownership and the quality of corporate governance is stronger in weaker legal regimes. This is because with weaker legal protection for investors, concentrated ownership becomes a more important tool to resolve the agency conflict between controlling and minority shareholders[15].
A.4 Investment and Valuation
Proposition 5: Firms with low diversion invest more and are valued higher.
Proof: The relation between the level of corporate investments and governance is already established in the preceding analysis. By combining equations [3] and [5] the optimal investment can be written as:
[pic] [10]
Since the market value of the firm, MV, is the present value of gross returns from projects,
[pic] [11]
Equations [10] and [11] show that both the optimal investment and the market value of the firm either increase or remains unchanged as d* decreases. Q.E.D.
B. External Financing
There are several reasons for external financing to be related to corporate governance. The first follows from what we have already shown: Firms with profitable investment opportunities will have better corporate governance. If profitable firms invest more and more investment leads to greater external financing, firms with greater external financing are likely to have better corporate governance.
According to Demirgüç-Kunt and Maksimovic (1998), however, the prediction would be the opposite. They argue that profitable firms have more internally generated funds and hence rely less on external financing. Thus, to isolate the impact of external financing from that of profitability of investment opportunities, we assume investment is given and analyze how d is related to external financing. In this setting, firms relying more on external financing have greater incentives to convince investors that they will be protected through good corporate governance, because new investors rationally anticipate diversion by the controlling shareholders.
Proposition 6: For a given level of profitable investment opportunities, controlling shareholders of firms with greater dependence on external financing will divert less corporate resources for private benefits.
Proof: See Appendix A.
II. Empirical Hypotheses and Design
A. Hypotheses
The preceding theoretical discussions are summarized in the following hypotheses:
Everything else being equal,
1. The average quality of corporate governance is lower and the variation is greater in less investor-friendly legal regimes.
2. Firms with more profitable investment opportunities practice higher-quality corporate governance.
3. Firms with greater reliance on external financing practice higher-quality corporate governance.
4. Firms with more concentrated ownership practice higher-quality corporate governance.
5. Firms with higher-quality corporate governance will invest more and will be valued higher.
6. The above relations are stronger in less investor-friendly legal regimes[16].
B. Empirical Design
To test hypotheses 2, 3, 4, and 6 we regress individual firms’ corporate governance and transparency scores on measures of profitable investment opportunities, reliance on external financing, ownership concentration, and countries’ legal regime, while controlling for industry and other firm characteristics. Specifically, we estimate the following cross-sectional regression using OLS,
[pic][S1]
where c stands for country; i, industry; j, firm; k, control variables; I, the number of industries; and K, the number of control variables. CG is corporate governance or transparency scores; INV_OPP, investment opportunities; EXT_FIN, reliance on external financing; OWNERSHIP, concentration of ownership; Z, control variables; and d, industry dummy. The strength of a country’s legal regime is denoted by LEGAL, and INV_OPP*LEGAL, EXT_FIN*LEGAL, and OWNERSHIP*LEGAL are interaction terms of legal regime with investment opportunities, external financing, and ownership concentration, respectively.
We expect firms with more profitable investment opportunities, greater demand for external financing, and higher ownership concentration to have better corporate governance (Hypotheses 2, 3, 4). We also expect that the relation of profitability, external financing, and ownership to corporate governance is stronger in countries that belong to weaker legal regimes (Hypothesis 6). That is, we expect coefficients (1, (2, (3, and (1 to be positive and (2, (3, and (4 to be negative.
To examine the relation between individual firm valuation or investment with corporate governance, we again control for countries’ legal regime, industry, and firm characteristics and estimate the following cross-sectional regression using OLS:
[pic] [S2]
Because we expect firms with better corporate governance to be valued higher and invest more and this relation to be stronger in weaker legal regimes (Hypotheses 5 and 6), we expect (1 and (1 to be positive and (2 to be negative.
We first run these regressions using OLS and report the results. However, the inferences that can be drawn from OLS regression results are limited because of potential econometric problems. We conduct various robustness checks on endogeneity, regression model specification, and alternative definition of main variables and describe the results in Section V.
III. Data
A. Corporate Governance and Disclosure Scores
To measure the quality of corporate governance and disclosure practices for individual companies, CG, we rely on scores provided by two sources: Credit Lyonnais Securities Asia (CLSA) Corporate Governance Scores and Standard & Poor’s Transparency Ranking.
A.1 CLSA Corporate Governance Scores
CLSA issued a report on the corporate governance of 494 companies in 24 countries in March 2001[17]. Firms are selected based on size (large) and investor interest (high). The corporate governance scores are based on answers from financial analysts to 57 questions, which are used to construct scores on a 1-100 scale, where a higher number indicates better corporate governance. All questions have binary answers (yes/no) to reduce analysts’ subjectivity. In tabulating the scores, CLSA attempts to differentiate corporate governance attributes from country-level legal environment. Their report states, “Our scores do not mark down a company simply for being in a country that might be perceived to have a weak regulatory or legal framework." (CLSA Emerging Markets, 2002, p. 9).
Scores on the 57 questions are grouped into seven categories: discipline (managerial incentives and discipline towards value maximizing actions), transparency (timely and accurate disclosure), independence (board independence), accountability (board accountability), responsibility (enforcement and management accountability), fairness (minority shareholder protection), and social awareness. These seven attributes are then averaged by CLSA to construct the composite governance index, where each of the first six attributes is weighted 15% and social awareness is weighted 10%.
To check the validity of CLSA scores, Khanna et al. (2002) manually construct a corporate governance “scandal index" for a group of Indian firms covered by CLSA. The "scandal index" is based on the number of media reports revealing shareholder expropriation, tax evasion, and price fixing. They report that companies with low CLSA corporate governance scores are more likely to have scandals reported in media.
A.2 Standard & Poor’s Transparency Ranking
Standard & Poor’s recently released its survey of corporate disclosure practices for 568 companies in 16 emerging markets and 3 developed countries. It reports whether or not a particular firm discloses relevant information on 98 possible items that would be of interest to investors: 28 items on ownership structure and investor relations (ownership), 35 items on accounting and financial policies (disclosure), and 35 items on board and management structure and process (board). We calculate the number of items disclosed in each category in 2000 and assign scores from 0 to 28 for ownership, from 0 to 35 for disclosure, and from 0 to 35 for board. We also compute an aggregate transparency score, aggregate, ranging from 0 to 98, by totaling the scores of the three categories[18].
If a firm has more disclosures on ownership related items and fewer disclosures on items concerning accounting and financial policies, for example, the implication is that the firm has less to hide and relatively sound practices on issues concerning ownership. Conversely, reluctance to reveal practices concerning accounting and financial policies implies relatively unsound practices in that category.
The disadvantage of the S&P ranking is that it depends only on the number of disclosures and does not reflect the content. It is best viewed as a measure of transparency and not a comprehensive measure of corporate governance. The advantage is that the S&P rankings in each category are objective, whereas the CLSA rankings are partially based on opinions of financial analysts. Appendix B contains a detailed description of both.
A.3 Consistency across CLSA and S&P Scores
To determine whether companies scored high on corporate governance by CLSA are also scored high on disclosure by S&P, we identify 203 companies that are ranked by both agencies. Table I reports correlation coefficients between attributes of CLSA and S&P scores. Most of the attributes of CLSA scores are highly correlated with each other except for social. Different categories of S&P scores are also significantly correlated with each other, indicating that firms that disclose more in one category tend to disclose more in other categories.
The correlation across CLSA and S&P rankings reveal that CLSA’s composite index is significantly correlated with S&P’s aggregate score. To check whether the correlation is driven by country and industry differences, we regress CLSA’s composite index on S&P’s aggregate score with country and industry dummies. The relation remains significant, confirming the consistency between the two rankings.[19]
Although many CLSA individual attributes are not correlated with S&P’s, the correlations are positive and significant when they are measured on overlapping characteristics. For instance, S&P’s measure of disclosure on accounting and financial policies (disclosure) is significantly correlated with CLSA’s measure of transparency (transparency); S&P’s disclosure on board and management structure and process (board) is significantly correlated with CLSA’s ranking on board accountability (accountability), enforcement and management accountability (responsibility); and S&P’s disclosure on ownership structure and investor relations (ownership) is significantly correlated with CLSA’s transparency (transparency) and investor protection (fairness). These correlations suggest that CLSA scores are largely consistent with the objective measure reported by S&P[20].
B. Other Firm-specific Variables
Because much of the firm-level data originate from financial statements and accounting practices that vary across countries, it is difficult to directly compare the data in our sample. Consequently, all the regression specifications contain legal regimes and industry dummies as independent variables. Because one of the key distinguishing characteristics in legal regimes is accounting standards, the legal regime variable controls for their differences to some extent. Additionally, industry dummies help control for different accounting practices across industries. Any remaining noise would weaken the power of our tests.
Most of the firm-level data are obtained from the Worldscope. All variables are measured in U.S. dollars.
B.1 Investment Profitability, Reliance on External Financing and Ownership
To measure the opportunity for profitable investments, INV_OPP, we use return on invested capital (ROIC)[21]. Reliance on external financing, EXT_FIN, is measured as the sum of new equity and changes in long-term debt over capital expenditures, where new equity issue is defined as changes in book value of equity less changes in retained earnings. Ownership concentration, OWNERSHIP, is defined as the percentage of shares held by major shareholders, either individuals or corporations, that hold more than 5% of outstanding shares.
For all these variables we use two-year averages for 1999 and 2000 to reduce the impact of outliers. The 1998 data are not used because of the 1997-1998 financial crises.
B.2 Valuation and Investment
To run regression specification S[2] we use the two-year average (2000-2001) Tobin’s Q as the measure of firm valuation. As in Doidge et al. (2001) and La Porta et al. (2002), we define Tobin’s Q as the sum of total assets plus market value of equity less book value of equity over total assets. The market value of equity is the number of common shares outstanding times the year-end price. We measure investment, INVEST, as the two-year average (2000-2001) of capital expenditures over total assets.
Since Tobin’s Qs and the level of investment may not be directly comparable across countries, it would be natural to control for country fixed effects. However, it is not feasible to estimate country fixed effects and LEGAL at the same time as there is no within country variation in LEGAL. We deal with this problem in the robustness section by repeating our estimation without LEGAL but with country fixed effects.
We separate time periods during which dependent and independent variables are measured in order to reduce endogeneity. Specifically, we use 2000-2001 two-year average for Q and INVEST; 2000 for CG; and 1999-2000 two-year average for INV_OPP, EXT_FIN, and OWNERSHIP.
C. Legal Regime Variables
To measure investor protection, INVESTOR, we use the antidirector rights (shareholder rights) index defined in La Porta et al. (1998), Claessens et al. (1999) and Pistor et al. (2000), which ranges from 0 to 6[22]. This investor-protection measure reflects only de jure regulation, and countries like India and Pakistan that may not have the best de facto investor protection score INVESTOR values of 5, the highest in our sample. To measure the strength of de facto regulation we use the rule of law index, ENFORCE, from La Porta et al. (1998), Claessens et al. (1999), and Pistor et al. (2000) as a proxy for law enforcement. The rule of law is the assessment of the law and order tradition of a country and ranges from 0 to 10[23].
There is little correlation between de jure and de facto measures of regulation. The correlation coefficient between INVESTOR and ENFORCE is only 0.05 with p-value = 0.82. Thus, we multiply INVESTOR by ENFORCE to construct a measure that reflects both aspects of regulation and define it LEGAL [24]. Finally, the legal origin of company law or commercial code of each country (English Common Law, French Civil Law, German Civil Law, or Scandinavian Civil Law), ORIGIN, is also taken from La Porta et al. (1998) and Claessens et al. (1999)[25].
D. Control Variables
Several control variables are used in S[1] and S[2]. One-digit SIC industry dummies (di) are included in both regression specifications to account for the differences in assets structure, accounting practices, government regulation, and competitiveness, all of which may affect corporate governance and firm valuation. We use two-digit dummies as a robustness check.
We also control for firm size, SIZE, defined as logarithm of total assets. Because larger firms tend to attract more attention and may be under greater scrutiny by the public, size may affect governance structure. Size also proxies for firm age, and older and larger firms tend to have higher book-to-market value ratio. As a robustness check, we use the log of sales as a proxy for firm size because sales data are less sensitive to differences in the accounting standards across countries.
Research and development expenditure scaled by total assets, R&D, is used to control for differences in intangibility of corporate resources, which may be related to cost of diversion. Moreover, companies with high R&D expenditures tend to be high-growth firms and may enjoy high valuation.
Finally, we use export intensity, EXPORT, defined as the revenue generated from shipping merchandise to another country for sale, scaled by total sales. This measure is used to control for differences in exposure to globalization pressures in the product market. Companies that conduct more business globally may feel more pressure to conform their corporate governance to global standards (see Khanna et al. (2002)). All control variables are two-year averages, 1999-2000.
Table II summarizes definitions of variables and their sources.
E. Sample Construction
The original CLSA and S&P samples contain 494 firms in 24 countries and 568 firms in 19 countries, respectively. Most of the financial information for these companies is obtained from Worldscope. When the identity of companies is ambiguous, we check it in the ISI (Internet Securities, Inc.) Emerging Markets database. If the ambiguity cannot be resolved, the companies are dropped from the sample, resulting in 475 firms in the CLSA sample and 557 firms in the S&P sample[26].
Sample sizes are reduced further when relevant variables for each regression are unavailable from Worldscope. When INV_OPP and EXT_FIN enter as independent variables in S[1], 15 and 16 companies, respectively, are dropped from CLSA and S&P samples due to missing data. When the regression is run with OWNERSHIP as the independent variable, 95 and 99 companies, respectively, are dropped from the CLSA and S&P samples because ownership data are missing.
When regression specification S[2] is run for Tobin’s Q as the dependent variable, 5 and 7 companies, respectively, are dropped from the CLSA and S&P samples because of missing data. When we use INVEST as the dependent variable 7 and 8 companies, respectively, are dropped from the CLSA and S&P samples.
If a firm has all major financial variables except R&D and export, we assume zero for those two variables. In other words, we assume that when a company does not report these variables it is because R&D spending or sales generated through export are negligible. Dropping companies with missing data for R&D and export would reduce our sample size considerably and may bias our sample towards technology-oriented firms. As a robustness check and as suggested by Himmelberg (1999), we also use two dummy variables which take values of 1 when a firm does not report R&D or export. These dummies control for the possibility that non-reporting firms are different from reporting firms.
IV. Empirical Results
In this section we first provide summary statistics, compare the mean and variance of governance and transparency rankings across legal regimes and show correlation among main variables. We then conduct regression analysis on how corporate governance is related to investment profitability, reliance on external financing, ownership concentration, and legal regimes. Finally, we examine the relation of corporate governance to firm valuation and investment.
A. Data Analysis
A.1 Relation between Variation in Corporate Governance and Legal Regime
Table III provides summary statistics by country for legal regime variables, CLSA corporate governance rankings, and S&P transparency rankings. To examine the relation between legal regime and governance and transparency rankings (Hypothesis 1), the average CLSA composite governance scores are regressed on LEGAL while controlling for average investment profitability (ROIC). The same regression is run for the variance of corporate governance rankings with an addition of the variance of investment profitability and the number of firms as the control variables. Countries with fewer than two firms are deleted in both regressions.
Figure 3A presents a partial regression plot, which shows a significant positive relation between the strength of legal regime and the average quality of corporate governance at the country level. This is not surprising given the previous findings on corporate governance at the country level.
Figure 4A shows a partial regression plot on the variance of corporate governance scores. It shows a negative relation between the variance of corporate governance and the strength of legal regime but the relation is not statistically significant. One of the reasons for the weak relation could be that the composite score includes many aspects of corporate governance, while LEGAL mostly measures one aspect--investor protection and its enforcement. For instance, when we use the measure of investor protection in CLSA (fairness) and relate its variance to LEGAL the relation becomes highly significant (Figure 4B).
An alternative measure for LEGAL is the minimum governance score observed in a country, which can be viewed as the minimum quality of governance practice that the country’s legal system tolerates. Figure 4D plots the partial regression results that show a significant relation between this alternative measure of legal environment and variation of governance rankings.
The same regressions were repeated for S&P transparency scores and the results are displayed in Figures 3C, 3E and 4C and 4E[27]. The results are qualitatively the same, except for 4E[28].
These results clearly dispel the stereotypical notion that all firms in weak legal regimes suffer from poor corporate governance while firms in strong legal regimes practice uniformly high-quality corporate governance. To the contrary, there seems to be as much–perhaps more–variation in the quality of corporate governance in weaker legal regimes.
AF.2 Correlation between Main Variables
Table IV, Panel A, shows correlation coefficients among main variables. The overall measure of CLSA corporate governance scores, the composite index (COMP), is highly correlated with investment opportunities, external financing, and Tobin’s Q. COMP is also positively correlated with all legal regime variables and the correlation is the largest for the combined measure, LEGAL. However, COMP is uncorrelated with investment, ownership, firm size, R&D, and export intensity.
The S&P aggregate score shows stronger relations. It is not only correlated with all the variables with which CLSA scores are correlated but also with ownership concentration, firm size, R&D, and export intensity.
Panel B of the same table shows correlation coefficients between legal and firm-specific variables. Consistent with the earlier findings (La Porta et al. (1997), Rajan and Zingales (1998), and Demirgüç-Kunt and Maksimovic (1998)), firms in countries with better investor protection and law enforcement are more reliant on external financing and enjoy higher valuation[29]. Firm size, R&D, and exports are also correlated with Tobin’s Q, investment, S&P rankings, and with each other, confirming our reasons to control for these variables in regression analyses.
how corporate governance is related to investment profitability, reliance on external financing, ownership concentration, and legal regimes. Then we examine the relation of corporate governance to firm valuation and investment.
B. Regression Results
For regression specification [S1] separate regressions are run for INV_OPP and EXT_FIN, and for OWNERSHIP. These variables are separated for two reasons. First, using all three as independent variables in addition to their interaction terms with LEGAL creates a severe multicollinearity problem. Second, using all three variables in a single regression substantially reduces the sample size because the ownership data is unavailable for 95 and 99 firms in the CLSA and S&P samples, respectively. As a robustness check, we estimate S[1] with INV_OPP, EXT_FIN, OWNERSHIP, and their interaction terms altogether. For this regression, we control for multicollinearity by centering these variables (subtracting the sample mean from each observation). All p-values reported in tables are based on robust standard errors.
B.1 Investment Opportunities, External Financing, and Ownership
Table V reports the results of regression [S1] with INV_OPP and EXT_FIN for CLSA scores. The results are supportive of our hypotheses. Both investment opportunities and external financing are significantly, positively related to the composite index. The strength of legal regimes is also positively related to the index.
The interaction terms of legal regime with investment opportunities and external financing are also supportive of the hypothesis that positive relations for investment opportunities and external financing are stronger (weaker) in weaker (stronger) legal environment. Although the interaction term INV_OPP*LEGAL is not significant for the composite index, perhaps due to multicollinearity, the test of joint significance indicates that INV_OPP and INV_OPP*LEGAL are jointly significant.
The table also displays the results for seven governance attributes as the dependent variable. Investment opportunities are related to discipline, responsibility, and fairness, while external financing is related to transparency, independence, responsibility, and fairness. It appears that profitable firms tend to stress financial incentives and discipline for managers, enforcement and managerial accountability, and minority shareholder protection in their governance structure. Firms that are more reliant on external financing appear to gravitate toward governance structures stressing transparency, board independence, enforcement and managerial accountability, and minority investor protection – important concerns to new investors.
When social awareness is used as the dependent variable, however, none of the firm-specific independent variables of interest are significant. Although social awareness attracts a great deal of public attention, corporate profitability and external financing are not related to it. Apparently, firms do not become more socially responsible when they become more profitable or more reliant on external financing.
The regression results of [S1] with ownership concentration are reported in Table VI. The regressions also contain the OWNERSHIP2 term to account for possible non-linearity between ownership concentration and corporate governance as in Himmelberg et al. (1999) and McConnell and Servaes (1999). The significant positive coefficient for OWNERSHIP and the significant negative coefficient for the OWNERSHIP2 term imply that corporate governance improves with the concentration of ownership but at a decreasing rate. This is consistent with earlier findings of Morck et al. (1988) and McConnell and Servaes (1990) who argue that greater ownership concentration by insiders may align their interests with those of minority shareholders but it also may result in a greater degree of managerial entrenchment for greater consumption of private benefits.
The interaction term with LEGAL indicates that the positive relation between ownership and corporate governance is stronger in weaker legal regimes. This is consistent with Proposition 4 and with the intuition that in weaker legal regimes concentrated ownership becomes a more important tool to resolve agency conflict between controlling and minority shareholders.
Table VII reports results from the same exercise with S&P scores as the dependent variable. The results are generally consistent with our findings from the CLSA scores. The interaction terms exhibit the expected signs but the significance is weaker; however, they are significant in joint tests. Thus, the results seem to be robust whether the quality of corporate governance is measured by a partially subjective, intuitively appealing method employed by CSLA or an objective but limited method employed by S&P[30].
Taken together, these results suggest that it is not only the difference in legal environment that matters, but also firm-level differences in profitability, external financing, and ownership concentration make a difference in a firm’s choice of corporate governance. Furthermore, these firm-specific attributes matter more as the legal environment becomes less investor friendly.
B.2. Valuation and Investment
To test the hypothesis that corporate governance and valuation are positively related, we run regression specification [S2] with Tobin’s Q as the dependent variable. Independent variables are CLSA scores or S&P rankings, legal regime, an interaction term of legal regime with corporate governance or disclosure scores, return on invested capital (ROIC), firm size, R&D expenditures, and export. ROIC is added to control for a possible spurious relation between corporate governance and valuation because profitability is likely to influence both valuation and corporate governance.
Table VIII reports results based on CLSA scores. Consistent with the hypothesis, firms with higher-quality corporate governance are valued higher. The composite governance score is positively related with firm valuation, and of the seven attributes, four show significant relations with firm valuation. The social awareness score again shows no relation to valuation, indicating that socially responsible firms do not necessarily enjoy higher valuation.
The LEGAL variable also has the expected positive sign, which is consistent with the findings of La Porta et al. (2002). However, it is not significant for the composite index and five of the seven attributes. The last column in Table VIII shows that the coefficient on LEGAL becomes significant when the same regression is run without the governance scores. This is perhaps due to the high correlation between CLSA scores and legal regimes (see Table IV). Apparently, the corporate governance rankings provided by CLSA contain more information on firm valuation than the legal environment in which a firm is located.
The interaction term has the expected negative sign and is significant for three of the attributes. In joint significance tests the interaction term is significant for the composite index as well as for four attributes, suggesting the positive relation between corporate governance and valuation is weaker in stronger legal regimes. This explains why previous studies based on U.S. data found mixed results.
All control variables are of expected sign and highly significant. Firms with high ROICs enjoy higher valuation as do firms of smaller size, greater R&D expenditures, and more export orientation.
The same regressions are run with S&P scores and the results are reported in Panel A of Table IX. They are qualitatively the same. Firms that disclose more are valued higher and their disclosure practices seem to have a stronger link to firm valuation than does the legal environment in which a firm is located.
Finally, to investigate the relation between corporate governance and the level of investment activities, the same regression is run using investment as a dependent variable. Table X and Panel B of Table IX report regression results for CLSA and S&P scores. The results are similar and consistent with the predictions of the model. Firms that have higher-quality corporate governance, or those that disclose more, tend to invest more, and this relation is stronger in weaker legal regimes.
V. Robustness
Our results remain robust to a battery of checks addressing endogeneity, the possibility of firm-level observations not being independent, heteroscedasticity, multicollinearity, alternative definitions of main variables, and the problem of outliers.
A. Endogeneity
There is an endogeneity problem in our regression analyses. For example, firms that had good governance in the past may have better investment opportunities and raise more external capital, which results in correlation between the regression error term and explanatory variables making the estimates inconsistent. As stated earlier, we lag all independent variables to lessen this problem. To address this issue further, we estimate regression specification [S2] using the Instrumental Variables (IV) approach with the legal origin (English, French, German, or Scandinavian) as instruments for corporate governance scores. La Porta et al. (1998) argue that investor protection is better in common law countries, implying firms located in countries with English legal origin have better corporate governance[31]. The legal origin is used as the instrumental variable because it is reasonable to assume that the legal origin is to some extent exogenous to individual firm valuation and investment[32]. For IV regression we use the following specification,
[pic] [S3]
where CG is either CLSA or S&P scores, dc are country dummies, di industry dummies, and Z are control variables that include size, R&D expenditures, and export intensity. Legal origin dummies, ORIGIN, are used as instruments for firm-level corporate governance, CG[33].
Table XI reports and compares IV regression coefficients to those of the OLS regression for CLSA scores when the dependent variable is valuation; Table XII reports results for investment as the dependent variable[34]. At the bottom of these tables we report the (2 statistics of the Hausman specification test (Hausman (1978)) of null hypothesis that the OLS estimator is a consistent and efficient estimator of the true parameter[35]. The IV regression results show that the majority of CLSA scores are positively and significantly related to valuation and investment. Moreover, the Hausman test indicates that coefficients for corporate governance in OLS regressions are not significantly different from those in the IV regressions for most cases.
The results using S&P scores are reported in Tables XIII and XIV and they are consistent with those using CLSA scores. In the IV regression, all scores are positively related to valuation and investment, and coefficients estimated in the IV regressions are not significantly different from those estimated using simple OLS regressions[36].
B. Regression Specifications
Each company-specific observation is treated as independent in all regressions. However, it is natural to expect observations within the same country to be correlated due to similar institutional, legal, and socio-economic factors. Ignoring the inter-firm correlations could make our estimates inefficient. We deal with this problem by adjusting coefficients standard errors assuming that firm-specific observations are correlated within each country but are independent across countries and that there is heteroscedasticity of the error term within and across countries[37]. Correlation between the independent variables and their interaction terms with LEGAL can cause multicollinearity problem making the estimates inefficient. As suggested by Jaccard et al. (1990), this correlation can be reduced by centering the original variables[38]. Thus we re-estimate S[1] and S[2] after centering main independent variables and adjusting coefficients standard errors to account for within-countries correlations.
Tables XV and XVI report the estimation results for the CLSA composite index and the S&P aggregate index, respectively. Although both tables show slight changes in standard errors, the magnitude and overall significance for most of the results remain unchanged. As an additional robustness check we assume firm-specific observations are correlated across industries and the results are qualitatively the same.
As mentioned earlier, Tobin’s Q and the level of investment are not comparable across countries and we cannot use country fixed effects along with LEGAL. This problem is partially addressed when we compare IV and OLS regressions in Tables XI-XIV, where all OLS regressions are run without LEGAL but with country fixed effects. Tables XI-XIV show that almost all attributes of CLSA and S&P scores are significantly related to valuation and investment[39].
Finally, several CLSA scores contain a few companies that score either 0 (minimum possible) or 100 (maximum possible). For those scores, a more appropriate estimation of specification S[1] would be to use a limited dependent variable approach (Tobit regression). The results in Tables V and VI do not change if we use Tobit regression instead of simple OLS.
C. Alternative Variables
Our results are also robust to alternative definitions of independent variables or to addition of more control variables. It is possible, for instance, that corporate governance and transparency scores are more related to other proxies of legal environment than to investor protection and rule of law. Several alternative variables are used to measure de jure protection of capital suppliers and de facto enforcement. Specifically, we substitute investor protection with creditor protection and rule of law with measures of efficiency of the judicial system, risk of expropriation, risk of contract repudiation, and absence of corruption[40]. In addition, following the principal component analysis outlined in Berkowitz et al. (2002) we combine investor and creditor protection to construct a single capital providers’ protection index; for enforcement, we combine efficiency of the judicial system, rule of law, absence of corruption, risk of expropriation, and risk of contract repudiation to derive a single index. Finally, we redefine LEGAL as a country minimum CLSA or S&P scores. The results remain virtually unchanged with alternative definitions for LEGAL.
Our external financing measure is defined as a proportion of investment funded by new equity issues and changes in long-term debt. Using only new equity issues over investment as a measure of the reliance on external financing does not change our results; neither does including reliance on short-term debt in addition to long term debt.
As mentioned earlier, we instrument current values of INV_OPP and EXT_FIN by their lagged values to reduce the endogeneity problem. Using contemporaneous measures does not change our findings nor does using two-digit industry dummies instead of one-digit ones. The findings also remain valid when we define firm size using sales instead of assets and when we include dummies when R&D or export data is missing.
We also use OWNERSHIP and OWNERSHIP2 as additional control variables in specification S[2] because ownership concentration is related to valuation (Morck et al. (1988) and La Porta et al. (2002)), investment, and corporate governance scores. Finally, we include INV_OPP, EXT_FIN, and OWNERSHIP, as well as their interaction terms with LEGAL all together in S[1]. Our results are robust to all of the above.
D. Outliers
Because outliers could affect our results, all but corporate governance and legal regime variables are calculated as two-year averages. As a further check, we apply different methodologies to trim outliers: Hadi’s (1992, 1994) multivariate method (with a percent cut-off) and Cook’s D statistics. We also drop the top and bottom 1% observations of the main variables and Winsorize the main variables at 1%. None of these procedures change the results.
Several countries contain data for only a few companies. As an additional robustness check we rerun all regressions dropping countries with fewer than 3 firms[41]. The qualitative results remain the same.
In sum, our results do not appear to be driven by the endogeneity problem (in S[2]) or outliers, and they are robust to alternative estimation techniques, variables definitions, and more control parameters.
V. Summary and Implications
This paper analyzes how firm-specific attributes interact with legal environment in firms’ choices of governance practice. With a simple model we illustrate that profitable investment opportunities, reliance on external financing, and more concentrated ownership lead to better governance practice and that the effects are stronger in weaker legal environments. We also show that firms with better governance are valued higher and invest more.
To determine the empirical validity of these implications we use two newly-constructed sets of data on the quality of corporate governance and transparency. One data set relies on an intuitively appealing, yet partially subjective method, while the other relies on a restrictive but objective method. Consequently, all tests are conducted for both sets of data. All regressions are run with control variables that include industry dummies, size, R&D expenditures, and export. The results show that aforementioned firm attributes are positively related to the quality of governance practice.
These positive relations are stronger in countries with weaker legal regimes. Apparently, firms in weak legal regimes structure their own governance to take better advantage of profitable investment opportunities, as do firms with greater reliance on external financing to overcome the deleterious effects of poor legal protection on their ability to raise external capital. Firms in weaker legal regimes also seem more reliant on ownership concentration to resolve agency conflict between controlling and outside shareholders as a response to a lack of investor protection.
These results have implications for the debate concerning the Coase argument (Coase (1960)). Our results on legal regimes confirm the La Porta et al. (1998) basic thesis that law matters in determining corporate governance. But our results on firm-specific attributes also illustrate that firms located in countries with poor legal protection do adapt to achieve efficient governance practice.[42]
We find that companies with better corporate governance and greater disclosure enjoy higher valuation (Tobin’s Q) and invest more. For instance, a 10-point increase (out of 100) in CLSA corporate governance scores increases a firm’s market value by 13.3%, while a 10-point increase (out of 98) in S&P transparency scores increases a firm’s market value by 16.3%[43]. These results are consistent with practitioners’ views of the worth of good corporate governance in emerging economies.[44]
Our results also indicate that the positive relation between governance and valuation is weaker in stronger legal regimes. This explains why previous studies based on U.S. data show mixed results on the relation between corporate governance and firm valuation.
One governance attribute that consistently shows no relation to firm-specific variables or to firm valuation is social awareness. Firms do not seem to become more socially responsible when they become more profitable or more reliant on external financing, perhaps because they believe it is not important to investors. Indeed, our evidence indicates that while investors value other governance attributes, they attach little value to the criteria CLSA uses to define social awareness, which are child labor, political legitimacy, environmental responsibility, equal employment policy, and ethical behavior. Child labor, for example, is a controversial issue to someeconomists. They debate whether child labor in developing economies is damaging to those societies, as the alternatives could be starvation, prostitution, or drug peddling.
Our results imply that economic policies affect corporate governance. To the extent that pro-growth policies generate more profitable investment opportunities to corporations, the controlling shareholders will have greater incentives to improve governance practices. Strong legal protection of investors can also be beneficial. If strengthening disclosure rules, corporate transparency, and managerial accountability improve the functioning of new issues markets and enhance investor confidence, firms are more likely to rely on external financing, which our results imply will provide an impetus for firms to improve governance.
Our results also have implications for the debate on whether globalization leads to convergence in corporate governance (see Coffee (1999), Bebchuk and Roe (1999), Berglöf and von Thadden (1999), Khanna et al. (2002)). With increasing globalization of trade and capital flows, national boundaries and legal jurisdictions are becoming less effective in defining corporate behavior in governance and finance. Therefore, any debate over convergence must consider firm-specific factors and their interplay with legal systems in determining corporate governance.
Finally, caveats are in order. Although we have attempted to address endogeneity, a full treatment requires time-series analyses of changes in corporate governance practices, a task we plan to pursue upon sufficient accumulation of data over time. On the theoretical level, we are able to identify only three firm-specific attributes related to corporate governance; however, other variables of greater importance may exist, which only further research can reveal.
Appendix A
Proof of Proposition 6
Consider a firm that has decided to invest I but does not have internal funds to finance it. The lack of internal funds necessitates the firm to finance the investment by selling 1-β fraction of the firm. The amount the firm must raise is I/(1-d) such that when the controlling shareholder diverts dI/(1-d) the firm will be left with I for investment. Under these assumptions the controlling shareholder’s payoff is:
[pic]. [A1]
Since new investors earn only the equilibrium rate of return on their investment, which is zero by assumption,
[pic] [A2]
Substituting Eq.[A2] to Eq. [A1],
[pic] [A3]
Differentiating [A3] with respect to d[45],
[pic] [A4]
If c > 1 - α, Eq.[A4] is negative and the optimal d = 0. When the cost of diversion, c, is greater than the wealth transfer that d can generate from the existing shareholders, 1 – α, the controlling shareholder will find it optimal not to divert any corporate resources.
One special case of such a situation is when a firm is 100% owned prior to the external financing, i.e., α = 1. Here there are no minority shareholders, only the cost of diversion. Hence the obvious solution is d = 0. New shareholders will be protected ex-ante because they will price the expected diversion at the time of purchasing the new shares. The controlling shareholder will find it in her self-interest to convince new investors that there will be no diversion.
If α < 1 and c < 1 – α, Eq.[A4] is positive and there is an incentive to maximize d. As can be seen from Eq.[A2], however, increasing d means the controlling shareholder must sell a greater fraction of the firm but when β falls below a certain point, she will lose control of the firm. Therefore the maximum fraction of the firm she sells to new investors is bounded by a minimum βmin, below which the controlling shareholder loses the control of the firm[46]. Substituting βmin into Eq.[A2], we obtain:
[pic] [A5]
Since I determines the amount of external financing, taking partial derivative of d* with respect to I,
[pic] [A6]
Q.E.D.
Appendix B
CLSA Corporate Governance Scores and S&P Transparency Rankings
1. CLSA Corporate Governance Scores
The CLSA corporate governance scores are based on how analysts rate a company on 57 elements under seven major aspects of corporate governance in 2000. Below is the summary of those attributes.
1. Discipline
1.1 Explicit public statement placing priority on corporate governance
1.2 Management incentives toward a higher share price
1.3 Sticking to clearly defined core business
1.4 Having an appropriate estimate of cost of equity
1.5 Having an appropriate estimate of cost of capital
1.6 Conservatism in issuance of equity or dilutive instruments
1.7 Ensuring debt is manageable, used only for projects with adequate returns
1.8 Returning excess cash to shareholders
1.9 Discussion in annual report on corporate governance
2. Transparency
2.1 Disclosure of financial targets, e.g., three- and five-year ROA/ROE
2.2 Timely release of Annual Report
2.3 Timely release of semi-annual financial announcements
2.4 Timely release of quarterly results
2.5 Prompt disclosure of results with no leakage ahead of announcement
2.6 Clear and informative results disclosure
2.7 Accounts presented according to IGAAP
2.8 Prompt disclosure of market-sensitive information
2.9 Accessibility of investors to senior management
2.10 Website where announcements are updated promptly
3. Independence
3.1 Board and senior management treatment of shareholders
3.2 Chairman who is independent from management
3.3 Executive management committee comprised differently from the board
3.4 Audit committee chaired by independent director
3.5 Remuneration committee chaired by independent director
3.6 Nominating committee chaired by independent director
3.7 External auditors unrelated to the company
3.8 No representatives of banks or other large creditors on the board
4. Accountability
4.1 Board plays a supervisory rather than executive role
4.2 Non-executive directors demonstrably independent
4.3 Independent, non-executive directors at least half of the board
4.4 Foreign nationals presence on the board
4.5 Full board meeting at least every quarter
4.6 Board members able to exercise effective scrutiny
4.7 Audit committee that nominates and reviews work of external auditors
4.8 Audit committee that supervises internal audit and accounting procedures
5. Responsibility
5.1 Acting effectively against individuals who have transgressed
5.2 Record on taking measures in cases of mismanagement
5.3 Measures to protect minority interests
5.4 Mechanisms to allow punishment of executive/management committee
5.5 Share trading by board members fair and fully transparent
5.6 Board small enough to be efficient and effective
6. Fairness
6.1 Majority shareholders treatment of minority shareholders
6.2 All equity holders right to call general meetings
6.3 Voting methods easily accessible (e.g., through proxy voting)
6.4 Quality of information provided for general meetings
6.5 Guiding market expectation on fundamentals
6.6 Issuance of ADRs or placement of shares fair to all shareholders
6.7 Controlling shareholder group owning less than 40% of company
6.8 Portfolio investors owning at least 20% of voting share
6.9 Priority given to investor relations
6.10 Total board remuneration rising no faster than net profit
7. Social awareness
4.1 Explicit policy emphasizing strict ethical behavior
4.2 Not employing the under-aged
4.3 Explicit equal employment policy
4.4 Adherence to specified industry guidelines on sourcing of materials
4.5 Explicit policy on environmental responsibility
4.6 Abstaining from countries where leaders lack legitimacy
2. S&P Transparency Ranking
S&P scores are based on transparency and disclosure, which is evaluated by searching for the inclusion of 98 possible information items (‘attributes’). These 98 attributes were selected after examination of the annual report and accounts of leading companies around the world and the identification of the most common disclosure items in 2000. These attributes are then grouped into three subcategories: (i) Ownership structure and investor relations (28 attributes), (ii) Financial transparency and information disclosure (35 attributes), (iii) Board and management structure and process (35 attributes).
|I |Ownership Structure and Investor Relations |II |Financial Transparency & Information Disclosure |
| |Does the company disclose: | |Does the company disclose: |
|1 |number of issued and outstanding ordinary shares disclosed? |1 |its accounting policy? |
|2 |number of issued and outstanding other shares disclosed (preferred, non-voting)? |2 |the accounting standards it uses for its accounts? |
|3 |par value of each ordinary share disclosed? |3 |accounts according to the local accounting standards? |
|4 |par value of each other shares disclosed (preferred, non-voting)? |4 |accounts according to an internationally recognized accounting standard (IAS/US GAAP)? |
|5 |number of authorized but unissued & outstanding ordinary shares disclosed? |5 |its balance sheet according to international accounting standard (IAS/US GAAP)? |
|6 |number of authorized but unissued & outstanding other shares disclosed? |6 |its income statement according to international accounting standard (IAS/US GAAP)? |
|7 |par value of authorized but unissued & outstanding ordinary shares disclosed? |7 |its cash flow statement according to international accounting standard (IAS/US GAAP)? |
|8 |par value of authorized but unissued & outstanding other shares disclosed? |8 |a basic earnings forecast of any kind? |
|9 |top 1 shareholder? |9 |a detailed earnings forecast? |
|10 |top 3 shareholders? |10 |financial information on a quarterly basis? |
|11 |top 5 shareholders? |11 |a segment analysis (broken down by business line)? |
|12 |top 10 shareholders? |12 |the name of its auditing firm? |
|13 |description of share classes provided? |13 |a reproduction of the auditors' report? |
|14 |review of shareholders by type? |14 |how much it pays in audit fees to the auditor? |
|15 |number and identity of shareholders holding more than 3%? |15 |any non-audit fees paid to auditor? |
|16 |number and identity of shareholders holding more than 5%? |16 |consolidated financial statements (or only the parent/holding co)? |
|17 |number and identity of shareholders holding more than 10%? |17 |methods of asset valuation? |
|18 |percentage of cross-ownership? |18 |information on method of fixed assets depreciation? |
|19 |existence of a Corporate Governance Charter or Code of Best Practice? |19 |a list of affiliates in which it holds a minority stake? |
|20 |Corporate Governance Charter / Code of Best Practice itself? |20 |a reconciliation of its domestic accounting standards to IAS/US GAAP? |
|21 |details about its Articles of Association. (e.g., changes)? |21 |the ownership structure of affiliates? |
|22 |voting rights for each voting or non-voting share? |22 |details of the kind of business it is in? |
|23 |way that shareholders nominate directors to board? |23 |details of the products or services produced/provided? |
|24 |way shareholders convene an EGM? |24 |output in physical terms? (number of users etc.) |
|25 |procedure for putting inquiry rights to the board? |25 |characteristics of assets employed? |
|26 |procedure for putting proposals at shareholders meetings? |26 |efficiency indicators (ROA ROE etc.) |
|27 |review of last shareholders meeting? (e.g., minutes) |27 |any industry-specific ratios? |
|28 |calendar of important shareholders dates? |28 |a discussion of corporate strategy? |
| | |29 |any plans for investment in the coming year(s)? |
| | |30 |detailed information about investment plans in the coming year(s)? |
| | |31 |an output forecast of any kind? |
| | |32 |an overview of trends in its industry? |
| | |33 |its market share for any or all of its businesses? |
| | |34 |a list/register of related party transactions? |
| | |35 |a list/register of group transactions? |
2. S&P Transparency Ranking (continued)
|III |Board and Management Structure and Process |
| |Does the company disclose: |
|1 |a list of board members (names)? |
|2 |details about directors (other than name/title)? |
|3 |details about current employment/position of directors provided? |
|4 |details about previous employment/positions provided? |
|5 |when each of the directors joined the board? |
|6 |classification of directors as an executive or an outside director? |
|7 |a chairman's name? |
|8 |detail about the chairman (other than name/title)? |
|9 |details about role of the board of directors at the company? |
|10 |a list of matters reserved for the board? |
|11 |a list of board committees? |
|12 |the existence of an audit committee? |
|13 |the names on the audit committee? |
|14 |the existence of a remuneration/compensation committee? |
|15 |the names on the remuneration/compensation committee)? |
|16 |existence of a nomination committee? |
|17 |the names on the nomination committee? |
|18 |the existence of other internal audit functions besides the Audit Committee? |
|19 |the existence of a strategy/investment/finance committee? |
|20 |the number of shares in the company held by directors? |
|21 |a review of the last board meeting? (e.g., minutes) |
|22 |whether they provide director training? |
|23 |the decision-making process of directors' pay? |
|24 |the specifics of directors' pay (e.g., the salary levels etc.)? |
|25 |the form of directors' salaries (e.g., cash, shares, etc.)? |
|26 |the specifics on performance-related pay for directors? |
|27 |the decision-making of managers' (not board) pay? |
|28 |the specifics of managers' (not on board) pay (e.g., salary levels etc.)? |
|29 |the form of managers' (not on board) pay? |
|30 |the specifics on performance-related pay for managers? |
|31 |the list of the senior managers (not on the board of directors)? |
|32 |the backgrounds of senior managers? |
|33 |the details of the CEO's contract? |
|34 |the number of shares held by the senior managers? |
|35 |the number of shares held in other affiliated companies by managers? |
REFERENCES
Bebchuk, L., and M. J. Roe, 1999, A theory of path dependence in corporate ownership and governance, Stanford Law Review 52, 127-170.
Berglöf, E., and E. L. von Thadden, 2000, The changing corporate governance paradigm: Implications for transition and developing countries, in Boris Pleskovic and Joseph E. Stiglitz, ed.: Annual World Bank Conference on Development Economics (World Bank, Washington, D.C.).
Berkowitz, D., K. Pistor, and J. F. Richard, 2002, Economic development, legality, and the transplant effect, European Economic Review, forthcoming.
Bhagat, S., and J. A. Brickley, 1984, Cumulative voting: The value of minority shareholder voting rights, Journal of Law and Economics 27, 231-284.
Bhagat, S., and R. H. Jefferis, 1991, Voting power in the proxy process: The case of antitakeover charter amendments, Journal of Financial Economics 77, 301-333.
Black, B., 2001, Does corporate governance matter? A crude test using Russian data, University of Pennsylvania Law Review 149, 1231-2150.
Black, B., H. Jang, and W. Kim, 2002, Does corporate governance matter? Evidence from Korean market, paper presented at the 2nd Asian Corporate Governance Conference in Seoul, Korea.
Claessens S., S. Djankov, J. Fan, and L. Lang, 2002, Expropriation of minority shareholders in East Asia, Journal of Finance, forthcoming.
Claessens, S., S. Djankov, and T. Nenova, 1999, Corporate growth and risk around the world, unpublished manuscript.
CLSA Emerging Markets, 2001. Saints and Sinners: Who’s got religion? (Hong Kong).
CLSA Emerging Markets, 2002. Make Me Holy…But Not Yet!, (Hong Kong).
Coase, R., 1960, The problem of social cost, Journal of Law and Economics 3, 1-44.
Coffee, J. C. Jr., 1999, The future as history: The prospects for global convergence in corporate governance and its implications, Northwestern University Law Review 93, 641-657.
Demirgüç-Kunt, A., and V. Maksimovic, 1998, Law, finance, and firm growth, Journal of Finance 53, 2107-2137.
Demsetz, H., and K. Lehn, 1998, The structure of corporate ownership: Causes and consequences, Journal of Political Economy 93, 210-227.
Denis D. J., D. K. Denis, and A. Sarin, 1997, Agency problems, equity ownership, and corporate diversification, Journal of Finance 52, 135-167.
Denis D. K., 2001, Twenty-five years of corporate governance research…and counting, Review of Financial Economics 10, 191-212.
Denis, D. K, and J. J. McConnell, 2002, International corporate governance, unpublished manuscript.
Diamond, D., 1991, Monitoring and reputation: The choice between bank loans and directly placed debt, Journal of Political Economy 99, 689-721.
Doidge, C., G. A. Karolyi, and R. Stulz, 2001, Why are foreign firms listed in the U.S. worth more?, NBER Working paper 8538.
Gomes, A., 2000, Going public without governance: managerial reputation effects,
Journal of Finance 55, 615-646.
Gompers, P. A., J. L. Ishi, and A. Metrick, 2001, Corporate governance and equity prices, NBER Working paper 8449.
Hadi, A. S., 1992, Identifying multiple outliers in multivariate data, Journal of the Royal Statistical Society, Series B 54, 761-771.
Hadi, A. S., 1994, A modification of a method for the detection of outliers in multivariate samples, Journal of the Royal Statistical Society, Series B 56, 393-396.
Hausman, J. A., 1978, Specification tests in econometrics, Econometrica 76, 1251-1271.
Himmelberg, C. P., R. G. Hubbard, and D. Palia, 1999, Understanding the determinants of managerial ownership and the link between ownership and performance, Journal of Financial Economics 53, 353-384.
Jaccard, J. J., R. Turrisi, and C. K. Won, 1990. Interaction Effects in Multiple Regression. (Sage, Newbury Park, CA).
Jensen M., and W. Meckling, 1976, Theory of the firm: Managerial behavior, agency costs and ownership structure, Journal of Financial Economics 3, 305-360
.
Johnson, S., and A. Shleifer, 2000, Coase and corporate governance in development, unpublished manuscript.
Johnson, S., 2002, Institutions, corporate governance, and debt: What have we learnt?, manuscript prepared for a conference in Hanyang University, Seoul, Korea.
Johnson, S., P. Boone, A. Breach, and E. Friemdan, 2000, Corporate governance and the Asian financial crisis, Journal of Financial Economics 58, 81-112.
Karpoff, J., 1998, The impact of shareholder activism on target companies: A survey of empirical findings, unpublished manuscript.
Khanna, T., J. Kogan, and K. Palepu, 2002, Globalization and corporate governance convergence? A cross-country analysis, unpublished manuscript.
Klapper, L. F., and I. Love, 2002, Corporate governance, investor protection, and performance in emerging markets, World Bank working paper 2818.
Kumar, K., R. Rajan, and L. Zingales, 1999, What determines firm size?, NBER Working paper 7208.
La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R. W. Vishny, 1997, Legal determinants of external finance, Journal of Finance 52, 1131-1150.
La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R. W. Vishny, 1998, Corporate ownership around the world, Journal of Finance, 471-517.
La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R. W. Vishny, 1998, Law and finance, Journal of Political Economy 106, 1112-1155.
La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R. W. Vishny, 2002, Investor protection and corporate valuation, Journal of Finance 57, 1147-1170.
Lins, K., 2002, Equity ownership and firm value in emerging markets, The Journal of Financial and Quantitative Analysis, forthcoming.
Maksimovic, V., and S. Titman, 1991, Financial policy and reputation for product quality, Review of Financial Studies 4, 175-200.
McKinsey Quarterly, 1996, Putting a value on corporate governance, Number 4. Available at
McConnell, J. and J. H. Servaes, 1990, Additional evidence on equity ownership and corporate value, Journal of Financial Economics 27, 595-612.
Mitton, Todd, 2000, A cross-firm analysis of the impact of corporate governance on the East Asian financial crisis, Journal of Financial Economics, 215-241.
Modigliani, F. and M. Miller, 1958, The cost of capital, corporation finance and, and the theory of investment, American Economic Review 48, 261-297.
Morck, R., A. Shleifer, and R. Vishny, 1988, Management ownership and corporate performance: An empirical analysis, Journal of Financial Economics 20, 293-315.
Pistor, K., M. Raiser, and S. Gelfer, 2000, Law and finance in transition economies, The Economics of Transition 8, 325-368.
Rajan, R., and L. Zingales, 1998, Financial development and growth, American Economic Review 88, 559-586.
Rajan, R., and L. Zingales, 2001, The great reversals: The politics of financial development in the 20th Century, NBER working paper 8178.
Shleifer, A., and R. W. Vishny, 1997, A survey of corporate governance, Journal of Finance 52, 737-783.
Shleifer, A., and D. Wolfenzon, 2002, Investor protection and equity markets, Journal of Finance, forthcoming.
Vardanian, R., 2001, Russian corporate governance: Reality, perception, risk, and reward, Troika Dialog Report, available at
Wurgler, J., 2000, Financial markets and the allocation of capital, Journal of Financial Economics 58, 187-214.
Figure 1. Investment opportunity for a firm with a maximum profit of [pic]and its optimal investment and market value when diversion is suppressed to zero
This figure presents a firm’s investment opportunity. The area 0ABC is equal to the firm’s market value, [pic], if the firm invests in all non-negative NPV projects.
Figure 2. Controlling shareholder’s payoff as a function of diversion d with different investment profitability and legal regimes
This graph plots the payoff to controlling shareholders as a function of diversion d for the same ownership α = 0.3, excess cash e = 0, and different cost of diversion c and profitability [pic].
US, LOW PROFITABILITY: cUS=0.6; [pic](black dotted line)
US, HIGH PROFITABILITY: cUS=0.6; [pic] (black solid line)
RUSSIA, LOW PROFITABILITY: cRUS=0.3; [pic](gray dotted line)
RUSSIA, HIGH PROFITABILITY: cRUS=0.3; [pic](gray solid line)
The corresponding optimal levels of diversion are dHUS = 0.15; dLUS = 0.17; dHRUS = 0.61; and dLRUS = 0.67.
As profitability changes from [pic] to [pic], the optimal level of diversion drops by [pic] in the U.S. and by [pic] in Russia.
[pic]
Table I
Correlation Coefficients between Attributes of CLSA Corporate Governance Scores and S&P Transparency Ranking
This table reports simple correlation coefficients between main attributes of CLSA and S&P samples. CLSA sample size is 494 companies. S&P sample size is 568 companies. The correlation coefficients between S&P and CLSA attributes are based on Spearman rank order correlation coefficients and they are based on 203 companies. Numbers in parentheses are probability levels at which the hypothesis of zero correlation can be rejected. Coefficients significant at 10% (based on two-tail test) are in bold face. Refer to Table II for variables definitions.
|CLSA Corporate Governance Scores |S&P Transparency Ranking | |
|TRANS |INDEP |ACCOUNT |
|CLSA Corporate Governance Scores | | |
|Discipline |DISC |Measure of managerial incentives and discipline towards value maximizing actions. Source: 2000 CLSA Corporate Governance |
| | |Scores. Range: 0-100. |
|Transparency |TRANS |Measure of timeliness and accuracy of financial information disclosure. Source: 2000 CLSA Corporate Governance Scores. Range:|
| | |0-100. |
|Independence |INDEP |Measure of board independence. Source: 2000 CLSA Corporate Governance Scores. Range: 0-100. |
|Accountability |ACCOUNT |Measure of board accountability. Source: 2000 CLSA Corporate Governance Scores. Range: 0-100. |
|Responsibility |RESP |Measure of enforcement and management accountability. Source: 2000 CLSA Corporate Governance Scores. Range: 0-100. |
|Fairness |FAIR |Measure of minority shareholder protection. Source: 2000 CLSA Corporate Governance Scores. Range: 0-100. |
|Social Awareness |SOCIAL |Measure of social awareness. Source: 2000 CLSA Corporate Governance Scores. Range: 0-100. |
|Composite |COMP |Composite corporate governance score. It is equal to |
| | |0.15*DISC+0.15*TRANS+0.15*INDEP+0.15*ACCOUNT+0.15*RESP+0.15*FAIR+0.1*SOCIAL. Source: 2000 CLSA Corporate Governance Scores. |
| | |Range: 0-100. |
|S&P Transparency Ranking | | |
|Ownership |OWN |Measure of transparency of ownership structure and investor relations. Source: 2000 S&P Transparency Ranking. Range: 0-28. |
|Disclosure |DISCL |Measure of financial transparency and information disclosure. Source: 2000 S&P Transparency Ranking. Range: 0-35. |
|Board |BOARD |Measure of transparency of board and management structure and processes. Source: 2000 S&P Transparency Ranking. Range: 0-35. |
|Aggregate |AGGR |Aggregate S&P Transparency score. It is equal to OWN+DISC+BOARD. Range: 0-98. |
|Firm-level Variables | | |
|Firm valuation |Q |2000-2001 average of Tobin’s Q. Tobin's Q is the sum of total assets plus market value of common stock less book value of |
| | |equity over total assets. The market value of equity is the number of common shares outstanding times year-end price. Source:|
| | |Worldscope. |
|Firm level of investment |INVEST |Capital expenditures over total assets, 2000-2001 average. Source: Worldscope. |
|Firm reliance on external financing |EXT_FIN |Proportion of investment financed externally, 1999-2000 average. It is defined as the sum of new equity issues and changes in|
| | |long-term debt over capital expenditures. New equity issue is change in book equity minus change in retained earnings. |
| | |Source: Worldscope. |
|Firm investment profitability |INV_OPP |Return on invested capital, 1999-2000 average. Return on invested capital = (Net Income before Preferred Dividends + |
| | |((Interest Expense on Debt - Interest Capitalized) * (1-Tax Rate))) / (Last Year's Total Capital + Last Year's Short Term |
| | |Debt & Current Portion of Long Term Debt). Source: Worldscope. |
|Ownership concentration |OWNERSHIP |Ratio of shares held by major shareholders, either individuals or corporations, that hold more than 5% of outstanding shares,|
| | |1999-2000 average. Source: Worldscope. |
|Firm size |SIZE |Logarithm of total assets, 1999-2000 average. Source: Worldscope. |
|Firm research and development expenditures |R&D |Research and development expenditures over total assets*100, 1999-2000 average. Source: Worldscope. |
|Firm export intensity |EXPORT |Exports over total sales*100, 1999-2000 average. Exports is defined as the revenues generated from the shipment of |
| | |merchandise to another country for sale. Source: Worldscope. |
|Legal Regime Variables | | |
|Legal origin |ORIGIN |The legal origin of the company law or commercial code of each country (English common law, French civil law, German civil |
| | |law, and Scandinavian civil law). Source: La Porta et al. (1998) and Claessens et al. (1999). |
|Investor protection |INVESTOR |Anti-director index. Range 0-6. An index is formed by adding 1 when (1) the country allows shareholders to mail their proxy |
| | |vote to the firm, (2) shareholders are not required to deposit their shares prior to the general shareholders’ meeting, (3) |
| | |cumulative voting or proportional representation of minorities mechanism is in place, (4) an oppressed minorities mechanism |
| | |is in place, (5) the minimum percentage of share capital that entitles a shareholder to call for an extraordinary |
| | |shareholders’ meeting is less than or equal to 10 percent, or (6) shareholders have preemptive rights that can be waived only|
| | |by a shareholders’ vote. Range: 0-6. Source: La Porta et al. (1998), Claessens et al. (1999), and Pistor et al. (2000). |
|Rule of law |ENFORCE |Assessment of the law and order tradition of the country. Average between 1982 and 1995 for all countries except Russia for |
| | |which it is measured in 1998. Range: 0-10. Source: La Porta et al. (1998), Claessens et al. (1999), and Pistor et al. (2000).|
|Quality of legal regime |LEGAL |Defined as INVESTOR*ENFORCE. |
Table III
Summary Statistics of Legal Regime Variables and Corporate Governance and Transparency Scores by Country
ORIGIN is legal origin; INVESTOR is anti-director index; ENFORCE is rule of law; LEGAL is INVESTOR*ENFORCE. Spread is the difference between maximum and minimum. of the scores. N is the number of firms in the country.
| | | |S&P Aggregate Transparency Ranking, AGGR |
| |Legal Regime Variables |CLSA Composite Corporate Governance Score, COMP | |
|Country |ORIGIN |INVESTOR |
| | |Firm-level variables |
|0.18 |0.07 |Q |
|(0.00) |(0.09) | |
|0.02 |-0.05 |INVEST |
|(0.63) |(0.18) | |
|0.21 |0.09 |EXT_FIN |
|(0.00) |(0.00) | |
|0.13 |0.21 |INV_OPP |
|(0.00) |(0.00) | |
|0.07 |0.10 |OWNERSHIP |
|(0.14) |(0.04) | |
|0.02 |0.28 |SIZE |
|(0.60) |(0.00) | |
|0.00 |0.12 |R&D |
|(0.95) |(0.00) | |
|-0.06 |-0.14 |EXPORT |
|(0.17) |(0.00) | |
| | |Legal Regime variables |
|0.30 |0.17 |INVESTOR |
|(0.00) |(0.00) | |
|0.26 |0.39 |ENFORCE |
|(0.00) |(0.00) | |
|0.40 |0.38 |LEGAL |
|(0.00) |(0.00) | |
Panel B. Correlation Coefficients between Legal Regime Variables and Firm-level Variables
|Q |INVEST |EXT_FIN |INV_OPP |OWNER |SIZE |R&D |EXPORT |Legal Regime variables |
|0.185 |-0.113 |0.138 |0.168 |-0.031 |-0.129 |0.036 |-0.062 |INVESTOR |
|(0.00) |(0.02) |(0.00) |(0.00) |(0.54) |(0.00) |(0.43) |(0.18) | |
|-0.040 |-0.002 |0.184 |-0.144 |0.015 |0.167 |0.108 |0.043 |ENFORCE |
|(0.39) |(0.96) |(0.00) |(0.00) |(0.77) |(0.00) |(0.02) |(0.36) | |
|0.081 |-0.098 |0.221 |-0.005 |0.003 |0.045 |0.058 |-0.048 |LEGAL |
|(0.08) |(0.03) |(0.00) |(0.91) |(0.95) |(0.33) |(0.21) |(0.30) | |
| | | | | | | | |Firm-level variables |
| |0.21 |0.15 |0.57 |0.080 |-0.44 |0.26 |0.22 |Q |
| |(0.00) |(0.00) |(0.00) |(0.12) |(0.00) |(0.00) |(0.00) | |
| | |-0.27 |0.20 |0.114 |-0.19 |0.17 |0.13 |INVEST |
| | |(0.00) |(0.00) |(0.02) |(0.00) |(0.00) |(0.00) | |
| | | |0.10 |-0.105 |0.01 |0.03 |0.04 |EXT_FIN |
| | | |(0.00) |(0.04) |(0.87) |(0.49) |(0.36) | |
| | | | |0.114 |-0.36 |0.23 |0.18 |INV_OPP |
| | | | |(0.03) |(0.00) |(0.00) |(0.00) | |
| | | | | |-0.13 |-0.02 |-0.078 |OWNER |
| | | | | |(0.01) |(0.65) |(0.13) | |
| | | | | | |-0.21 |-0.17 |SIZE |
| | | | | | |(0.00) |(0.00) | |
| | | | | | | |0.41 |R&D |
| | | | | | | |(0.00) | |
Table V
OLS Regression of CLSA Corporate Governance Scores on Investment Opportunities, Reliance on External Financing, and Control Variables
This table reports the results of cross-section regression:
[pic]
where c indexes country; i indexes industry; j indexes firm; and t indexes time. CG is one of CLSA Corporate Governance Scores (COMP, DISC, TRANS, INDEP, ACCOUNT, RESP, FAIR, SOCIAL) in 2000. d are industry dummies (coefficients are not reported). INV_OPP (investment opportunities) is defined as return on invested capital, 1999-2000 average; EXT_FIN (external financing) is 1999-2000 average ratio of the sum of new equity issues and changes in long-term debt over capital expenditures; LEGAL is defined as INVESTOR*ENFORCE where INVESTOR measures investor protection and ENFORCE (enforcement) is the rule of law. Z are control variables: SIZE is the log of total assets,1999-2000 average; R&D is research and development expenditures scaled by total assets times 100, 1999-2000 average; and EXPORT is export scaled by sales times 100, 1999-2000 average. Numbers in parentheses are probability levels at which the null hypothesis of zero coefficient can be rejected. Coefficients significant at 10% level (based on 2-tailed test) are in boldface. All p-values are based on robust (heteroscedasticity consistent) standard errors. Sample size is 460 firms. We drop companies that do not have one of the following items: total assets, sales, ROIC, long-term debt, book equity, or retained earnings. If all variables are available except R&D expenditures and export, we set those two to zero.
|Dep Variable: CLSA Corporate Gov Scores |COMP |DISC |TRAN |INDEP |ACCOUNT |RESP |FAIR |SOCIAL |
|INV_OPP |10.625 |24.777 |6.454 |3.843 |1.947 |18.898 |21.320 |5.594 |
| |(0.07) |(0.01) |(0.38) |(0.71) |(0.86) |(0.02) |(0.06) |(0.62) |
|EXT_FIN |2.590 |1.044 |6.033 |4.100 |0.006 |3.862 |2.536 |0.248 |
| |(0.01) |(0.47) |(0.00) |(0.02) |(0.90) |(0.04) |(0.10) |(0.87) |
|LEGAL |0.609 |0.411 |0.488 |1.065 |0.312 |1.028 |0.803 |-0.080 |
| |(0.00) |(0.01) |(0.00) |(0.00) |(0.05) |(0.00) |(0.00) |(0.62) |
|INV_OPP*LEGAL |-0.152 |-0.498 |-0.284 |-0.419 |0.128 |-0.402 |-0.054 |0.186 |
| |(0.55) |(0.10) |(0.48) |(0.10) |(0.79) |(0.24) |(0.40) |(0.68) |
|EXT_FIN*LEGAL |-0.064 |-0.061 |-0.095 |-0.133 |-0.007 |-0.081 |-0.067 |-0.008 |
| |(0.07) |(0.07) |(0.10) |(0.07) |(0.90) |(0.03) |(0.31) |(0.87) |
|SIZE |0.418 |1.032 |1.662 |-0.660 |3.392 |-1.002 |-1.876 |0.425 |
| |(0.35) |(0.13) |(0.03) |(0.48) |(0.00) |(0.16) |(0.05) |(0.61) |
|R&D |-0.081 |-1.475 |-1.394 |2.956 |1.213 |-1.372 |-0.971 |-4.862 |
| |(0.94) |(0.39) |(0.50) |(0.22) |(0.56) |(0.43) |(0.70) |(0.02) |
|EXPORT |-0.044 |0.017 |-0.085 |-0.028 |-0.065 |-0.128 |-0.046 |0.074 |
| |(0.29) |(0.78) |(0.23) |(0.74) |(0.40) |(0.06) |(0.60) |(0.38) |
|F test statistics of joint significance |4.510 |13.290 |2.070 |4.230 |0.500 |4.270 |2.440 |0.080 |
|INV_OPP=0 and INV_OPP*LEGAL=0 |(0.01) |(0.00) |(0.18) |(0.01) |(0.60) |(0.02) |(0.09) |(0.92) |
| | | | | | | | | |
|EXT_FIN=0 and EXT_FIN*LEGAL=0 |5.170 |2.500 |2.090 |7.870 |0.020 |2.800 |4.790 |1.320 |
| |(0.01) |(0.03) |(0.10) |(0.00) |(0.98) |(0.06) |(0.01) |(0.27) |
|F test statistics of overall significance |5.800 |4.540 |18.820 |34.430 |2.680 |20.120 |6.430 |1.740 |
| |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.04) |
|Regression R2 |0.216 |0.097 |0.109 |0.195 |0.075 |0.239 |0.143 |0.058 |
|Number of Companies |460 |
Table VI
OLS Regression of CLSA Corporate Governance Scores on Ownership Concentration and Control Variables
This table reports the results of cross-section regression:
[pic]
where c indexes country; i indexes industry; j indexes firm; and t indexes time. CG is one of CLSA Corporate Governance Scores (COMP, DISC, TRANS, INDEP, ACCOUNT, RESP, FAIR, SOCIAL) in 2000. d are industry dummies (coefficients are not reported). OWNERSHIP is the ratio of shares held by major shareholders, either individuals or corporations, that hold more than 5% of outstanding shares, 1999-2000 average; OWNERSHIP^2 is the squared term of ownership; LEGAL is INVESTOR*ENFORCE where INVESTOR measures investor protection and ENFORCE (enforcement) is the rule of law. Z are control variables: SIZE is the log of total assets, 1999-2000 average; R&D is research and development expenditures scaled by total assets times 100, 1999-2000 average; and EXPORT is export scaled by sales times 100, 1999-2000 average. Numbers in parentheses are probability levels at which the null hypothesis of zero coefficient can be rejected. Coefficients significant at 10% level (based on 2-tailed test) are in boldface. All p-values are based on robust (heteroscedasticity consistent) standard errors. Sample size is 380 firms. We drop companies that do not have one of the following items: total assets, sales, or ownership. If all variables are available except R&D expenditures and export, we set those two to zero.
|Dep Variable: CLSA Corporate Gov Scores |COMP |DISC |TRAN |INDEP |ACCOUNT |RESP |FAIR |SOCIAL |
|OWNERSHIP |37.456 |31.493 |72.108 |20.432 |47.449 |28.243 |56.422 |17.801 |
| |(0.01) |(0.07) |(0.00) |(0.37) |(0.03) |(0.10) |(0.02) |(0.36) |
|OWNERSHIP^2 |-24.818 |-9.622 |-50.543 |-5.742 |-54.166 |-10.917 |-40.152 |-11.500 |
| |(0.05) |(0.58) |(0.01) |(0.79) |(0.01) |(0.52) |(0.09) |(0.51) |
|LEGAL |0.822 |0.989 |0.669 |1.205 |0.099 |1.212 |0.981 |0.548 |
| |(0.00) |(0.00) |(0.03) |(0.00) |(0.73) |(0.00) |(0.00) |(0.01) |
|OWNERSHIP*LEGAL |-0.536 |-1.130 |-1.477 |-0.420 |0.168 |-0.432 |-0.522 |-0.254 |
| |(0.06) |(0.00) |(0.00) |(0.42) |(0.73) |(0.22) |(0.32) |(0.47) |
|SIZE |0.293 |0.470 |0.300 |-1.261 |3.023 |-1.325 |-0.606 |1.454 |
| |(0.49) |(0.49) |(0.71) |(0.17) |(0.00) |(0.05) |(0.52) |(0.06) |
|R&D |-0.177 |-0.319 |0.347 |0565 |-0.952 |-0.173 |-0.446 |-2.171 |
| |(0.75) |(0.71) |(0.01) |(0.69) |(0.26) |(0.85) |(0.75) |(0.05) |
|EXPORT |-0.052 |0.000 |0.091 |-0.144 |-0.020 |-0.011 |-0.075 |-0.063 |
| |(0.14) |(0.99) |(0.23) |(0.03) |(0.79) |(0.11) |(0.41) |(0.29) |
|F test statistics of joint significance |4.020 |4.520 |6.950 |0.440 |3.140 |1.360 |2.800 |0.480 |
|OWNERSHIP=0 and OWNERSHIP*LEGAL=0 |(0.02) |(0.01) |(0.00) |(0.64) |(0.04) |(0.26) |(0.06) |(0.61) |
|F test statistics of overall significance |12.780 |42.4220 |48.650 |47.560 |15.130 |98.500 |16.170 |73.040 |
| |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |
|Regression R2 |0.229 |0.069 |0.085 |0.190 |0.092 |0.289 |0.130 |0.115 |
|Number of Companies |380 |
Table VII
OLS Regression of S&P Transparency Ranking on Investment Opportunities, Reliance on External Financing, Ownership Concentration and Control Variables
This table reports the results of cross-section regressions:
[pic](Panel A)
[pic] (Panel B)
where c indexes country; i indexes industry; j indexes firm; and t indexes time. CG is one of S&P Transparency Rankings (AGGR, OWN, DISCL, BOARD) in 2000. d are industry dummies (coefficients are not reported). INV_OPP (investment opportunities) is defined as return on invested capital, 1999-2000 average; EXT_FIN (external financing) is 1999-2000 average ratio of the sum of new equity issues and changes in long-term debt over capital expenditures; OWNERSHIP is the ratio of shares held by major shareholders, either individual or corporations, that hold more than 5% of outstanding shares, 1999-2000 average; OWNERSHIP^2 is the squared term of ownership; LEGAL is INVESTOR*ENFORCE where INVESTOR measures investor protection and ENFORCE (enforcement) is the rule of law. Z are control variables: SIZE is the log of total assets, 1999-2000 average; R&D is research and development expenditures scaled by total assets times 100, 1999-2000 average; and EXPORT is export scaled by sales times 100, 1999-2000 average. Numbers in parentheses are probability levels at which the null hypothesis of zero coefficient can be rejected. Coefficients significant at 10% level (based on 2-tailed test) are in boldface. All p-values are based on robust (heteroscedasticity consistent) standard errors. Sample size is 541 (Panel A) and 458 (Panel B) firms. In Panel A we drop companies that do not have one of the following items: total assets, sales, ROIC, long-term debt, book equity, or retained earnings. In panel B we drop companies that do not have one of the following items: total assets, sales, or ownership. If all variables are available except R&D expenditures and export, we set those two to zero.
| |Panel A |Panel B |
|Dep Variable: S&P Transparency Ranking |AGGR |OWN |DISCL |BOARD |AGGR |OWN |DISCL |BOARD |
|INV_OPP |15.397 |8.944 |3.428 |3.025 | | | | |
| |(0.09) |(0.04) |(0.17) |(0.44) | | | | |
|EXT_FIN |9.899 |2.966 |1.421 |4.986 | | | | |
| |(0.03) |(0.03) |(0.37) |(0.03) | | | | |
|OWNERSHIP | | | | |33.616 |16.547 |9.036 |8.033 |
| | | | | |(0.01) |(0.00) |(0.03) |(0.17) |
|OWNERSHIP^2 | | | | |-28.962 |-16.985 |-6.080 |-5.897 |
| | | | | |(0.01) |(0.00) |(0.07) |(0.26) |
|LEGAL |0.313 |0.106 |0.076 |0.130 |0.231 |0.059 |0.070 |0.102 |
| |(0.00) |(0.00) |(0.00) |(0.00) |(0.04) |(0.16) |(0.07) |(0.10) |
|INV_OPP*LEGAL |-0.033 |-0.008 |-0.048 |-0.073 | | | | |
| |(0.86) |(0.94) |(0.26) |(0.16) | | | | |
|EXT_FIN*LEGAL |-0.110 |-0.043 |-0.026 |-0.039 | | | | |
| |(0.49) |(0.44) |(0.67) |(0.72) | | | | |
|OWNERSHIP*LEGAL | | | | |-0.081 |-0.002 |-0.044 |-0.003 |
| | | | | |(0.70) |(0.97) |(0.54) |(0.76) |
|SIZE |1.282 |0.307 |0.658 |0.317 |1.261 |0.340 |0.692 |0.229 |
| |(0.00) |(0.01) |(0.00) |(0.07) |(0.00) |(0.01) |(0.00) |(0.17) |
|R&D |0.029 |0.141 |0.112 |0.041 |0.179 |0.088 |0.084 |0.007 |
| |(0.03) |(0.00) |(0.02) |(0.63) |(0.11) |(0.02) |(0.06) |(0.93) |
|EXPORT |-0.050 |-0.022 |-0.012 |-0.016 |-0.002 |-0.008 |-0.002 |0.008 |
| |(0.10) |(0.00) |(0.04) |(0.07) |(0.89) |(0.23) |(0.60) |(0.24) |
|F test statistics of joint significance |3.050 |7.770 |5.750 |1.010 | | | | |
|INV_OPP=0 and INV_OPP*LEGAL=0 |(0.05) |(0.00) |(0.00) |(0.37) | | | | |
| | | | | | | | | |
| | | | | | | | | |
|EXT_FIN=0 and EXT_FIN*LEGAL=0 |4.930 |4.960 |0.600 |7.110 | | | | |
| |(0.01) |(0.01) |(0.55) |(0.00) | | | | |
|OWNERSHIP=0 and OWNERSHIP*LEGAL=0 | | | | |5.220 |10.680 |2.840 |1.060 |
| | | | | |(0.01) |(0.00) |(0.06) |(0.35) |
|F test statistics of overall significance |13.280 |14.780 |14.490 |6.080 |7.800 |8.41 |9.860 |3.030 |
| |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |
|Regression R2 |0.236 |0.232 |0.250 |0.144 |0.177 |0.179 |0.240 |0.083 |
|Number of Companies |541 |458 |
Table VIII
OLS Regression of Valuation on CLSA Corporate Governance Scores and Control Variables
This table reports the results of regression:
[pic]
where c indexes country; i indexes industry; j indexes firm; and t indexes time. Valuation is 2000-2001 average Tobin’s Q defined as total assets plus market value of equity less book value of equity over total assets and market value of equity is the number of common shares outstanding times year-end share price. CG is one of CLSA Corporate Governance Scores (COMP, DISC, TRANS, INDEP, ACCOUNT, RESP, FAIR, SOCIAL) in 2000. LEGAL is INVESTOR*ENFORCE where INVESTOR measures investor protection and ENFORCE (enforcement) is the rule of law. Z are control variables: ROIC (return on invested capital), 1999-2000 average; SIZE is the log of total assets, 1999-2000 average; R&D is research and development expenditures scaled by total assets times 100, 1999-2000 average; and EXPORT is export scaled by sales times 100, 1999-2000 average. Numbers in parentheses are probability levels at which the null hypothesis of zero coefficient can be rejected. Coefficients significant at 10% level (based on 2-tailed test) are in boldface. All p-values are based on robust (heteroscedasticity consistent) standard errors. Sample size is 470 firms. We drop companies that do not have one of the following items: total assets, sales, ROIC, book equity, number of shares outstanding, or year-end price. If all variables are available except R&D expenditures and export, we set those two to zero. Coefficients on CG are multiplied by 100.
|Dependent variable |Tobin’s Q |
|COMPOSITE |1.383 |
Table IX
OLS Regression of Valuation and Investment on S&P Transparency Rankings and Control Variables
This table reports the results of regression:
[pic] (Panel A)
[pic] (Panel B)
where c indexes country; i indexes industry; j indexes firm; and t indexes time. Valuation is 2000-2001 average Tobin’s Q defined as total assets plus market value of equity less book value of equity over total assets and market value of equity is the number of common shares outstanding times year-end share price. Investment is 2000-2001 average ratio of capital expenditures to total assets. CG is one of S&P Transparency Rankings (AGGR, OWN, DISCL, BOARD) in 2000. d are industry dummies (coefficients are not reported). LEGAL is INVESTOR*ENFORCE where INVESTOR measures investor protection and ENFORCE (enforcement) is the rule of law. Z are control variables: ROIC (return on invested capital), 1999-2000 average; SIZE is the log of total assets, 1999-2000 average; R&D is research and development expenditures scaled by total assets times 100, 1999-2000 average; and EXPORT is export scaled by sales times 100, 1999-2000 average. Numbers in parentheses are probability levels at which the null hypothesis of zero coefficient can be rejected. Coefficients significant at 10% level (based on 2-tailed test) are in boldface. All p-values are based on robust (heteroscedasticity consistent) standard errors. Sample size is 550 firms (Panel A) and 549 (Panel B). In Panel A we drop companies that do not have one of the following items: total assets, sales, ROIC, book equity, number of shares outstanding, or year-end price. In Panel B we drop companies that do not have one of the following items: total assets, sales, ROIC, or capital expenditures. If all variables are available except R&D expenditures and export, we set those two to zero. Coefficients on CG are multiplied by 100.
| |Panel A |Panel B |
|Dependent variable |Tobin’s Q |INVESTMENT |
|AGGREGATE |1.353 | |
Table X
OLS Regression of Investment on CLSA Corporate Governance Scores and Control Variables
This table reports the results of regression:
[pic]
where c indexes country; i indexes industry; j indexes firm and t indexes time. d are industry dummies (coefficients are not reported). Investment is 2000-2001 average ratio of capital expenditures to total assets. CG is one of CLSA Corporate Governance Scores (COMP, DISC, TRANS, INDEP, ACCOUNT, RESP, FAIR, SOCIAL) in 2000. LEGAL is INVESTOR*ENFORCE where INVESTOR measures investor protection and ENFORCE (enforcement) is the rule of law. Z are control variables: ROIC (return on invested capital), 1999-2000 average; SIZE is the log of total assets, 1999-2000 average; R&D is research and development expenditures scaled by total assets times 100, 1999-2000 average; and EXPORT is export scaled by sales times 100, 1999-2000 average. Numbers in parentheses are probability levels at which the null hypothesis of zero coefficient can be rejected. Coefficients significant at 10% level (based on 2-tailed test) are in boldface. All p-values are based on robust (heteroscedasticity consistent) standard errors. Sample size is 468 firms. We drop companies that do not have one of the following items: total assets, sales, ROIC, or capital expenditures If all variables are available except R&D expenditures and export, we set those two to zero. Coefficients on CG are multiplied by 100.
|Dependent variable |INVESTMENT |
|COMPOSITE |0.127 | | | | | | | |
| |(0.00) | | | | | | | |
|DISCIPLINE | |0.052 | | | | | | |
| | |(0.13) | | | | | | |
|TRANSPARENCY | | |0.060 | | | | | |
| | | |(0.04) | | | | | |
|INDEPENDENCE | | | |0.038 | | | | |
| | | | |(0.05) | | | | |
|ACCOUNTABILITY | | | | |0.021 | | | |
| | | | | |(0.42) | | | |
|RESPONSIBILITY | | | | | |0.034 | | |
| | | | | | |(0.10) | | |
|FAIRNESS | | | | | | |-0.009 | |
| | | | | | | |(0.66) | |
|SOCIAL AWARENESS | | | | | | | |0.040 |
| | | | | | | | |(0.09) |
|LEGAL |0.002 |0.001 |0.000 |0.000 |0.000 |0.001 |0.000 |0.001 |
| |(0.00) |(0.44) |(0.65) |(0.83) |(0.72) |(0.37) |(0.52) |(0.24) |
|CG*LEGAL |-0.003 |-0.001 |-0.001 |-0.001 |0.000 |-0.001 |0.000 |-0.001 |
| |(0.00) |(0.23) |(0.34) |(0.18) |(0.93) |(0.10) |(0.75) |(0.10) |
|ROIC |0.007 |0.015 |0.017 |0.022 |0.016 |0.016 |0.018 |0.018 |
| |(0.48) |(0.16) |(0.10) |(0.03) |(0.13) |(0.13) |(0.10) |(0.10) |
|SIZE |0.000 |0.000 |0.000 |0.000 |0.000 |0.000 |0.000 |0.000 |
| |(0.91) |(0.90) |(0.84) |(0.89) |(0.85) |(0.83) |(0.84) |(0.78) |
|R&D |0.009 |0.010 |0.010 |0.010 |0.010 |0.010 |0.010 |0.010 |
| |(0.08) |(0.07) |(0.04) |(0.05) |(0.04) |(0.06) |(0.04) |(0.05) |
|EXPORT |0.000 |0.000 |0.000 |0.000 |0.000 |0.000 |0.000 |0.000 |
| |(0.46) |(0.55) |(0.49) |(0.67) |(0.56) |(0.56) |(0.62) |(0.57) |
|F test statistics of joint significance |15.480 |1.180 |4.390 |7.810 |1.490 |2.190 |0.120 |5.450 |
|CG=0 and CG*LEGAL=0 |(0.00) |(0.31) |(0.01) |(0.00) |(0.23) |(0.10) |(0.89) |(0.00) |
|F test statistics of overall significance |32.260 |37.090 |38.000 |38.230 |36.350 |36.800 |36.360 |35.580 |
| |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |
| Regression R2 |0.364 |0.330 |0.343 |0.351 |0.333 |0.330 |0.328 |0.332 |
|Number of Companies |468 |
Table XI
Comparison of Coefficients Estimated Using Instrumental Variables Approach vs. OLS: Valuation on CLSA Corporate Governance Scores and Control Variables
This table reports and compares the results of regression:
[pic]
using OLS and IV estimation approaches, where c indexes country, i indexes industry; j indexes firm; and t indexes time. dc and di are country and industry dummies, respectively (coefficients are not reported). We use legal origin (English, French, or German), ORIGIN , as instruments for the corporate governance scores in IV estimation. Valuation is defined as total assets plus market value of equity less book value of equity over total assets where market value of equity is the number of common shares outstanding times year-end share price. CG is one of CLSA Corporate Governance Scores (COMP, DISC, TRANS, INDEP, ACCOUNT, RESP, FAIR, SOCIAL) in 2000. Z are control variables: SIZE is the log of total assets, 1999-2000 average; R&D is research and development expenditures scaled by total assets, 1999-2000 average times 100; and EXPORT is export scaled by sales times 100, 1999-2000 average. Numbers in parentheses are probability levels at which the null hypothesis of zero coefficient can be rejected. Coefficients significant at 10% level (based on 2-tailed test) are in boldface. All p-values are based on robust (heteroscedasticity consistent) standard errors. Sample size is 470 and 469 firms in OLS and IV regression, respectively. ORIGIN is not defined for Russia, thus, we drop it from IV estimation. Hausman (1978) specification test compares the IV estimates with OLS estimates. The null hypothesis is that the OLS estimator is a consistent and efficient estimator of the true parameter. P-val is the probability level at which the null hypothesis can be rejected. The Hausman statistic is calculated as [pic]where [pic]is the coefficient vector from the IV regression, [pic]is the coefficient vector from the OLS regression, [pic]is the covariance matrix of the coefficients from the IV regression, and [pic]is the covariance matrix of the coefficients from the OLS regression. Numbers in parentheses are probability levels at which the null hypothesis of zero coefficient can be rejected. Coefficients significant at 10% level (based on 2-tailed test) are in boldface. Coefficients on CG are multiplied by 100.
|Dependent variable |Tobin’s Q |
| |OLS |IV |OLS |IV |OLS |IV |OLS |IV |
|COMPOSITE |1.622 |1.019 | | | | | | |
| |(0.00) |(0.00) | | | | | | |
|DISCIPLINE | | |1.008 |0.496 | | | | |
| | | |(0.00) |(0.42) | | | | |
|TRANSPARENCY | | | | |0.569 |1.664 | | |
| | | | | |(0.02) |(0.00) | | |
|INDEPENDENCE | | | | | | |0.481 |0.902 |
| | | | | | | |(0.03) |(0.00) |
|SIZE |-0.230 |-0.228 |-0.241 |-0.197 |-0.240 |-0.244 |-0.232 |-0.200 |
| |(0.00) |(0.00) |(0.00) |(0.07) |(0.00) |(0.00) |(0.00) |(0.00) |
|R&D |0.195 |0.193 |0.218 |0.188 |0.197 |0.191 |0.199 |0.197 |
| |(0.07) |(0.07) |(0.04) |(0.00) |(0.07) |(0.09) |(0.07) |(0.10) |
|EXPORT |0.008 |0.008 |0.007 |0.014 |0.009 |0.010 |0.008 |0.010 |
| |(0.07) |(0.06) |(0.09) |(0.25) |(0.06) |(0.03) |(0.07) |(0.05) |
|F test statistics of overall significance |20.120 |24.240 |43.350 |21.760 |24.950 |42.530 |8.310 |14.210 |
| |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |
| Regression R2 |0.384 |0.382 |0.389 |0.272 |0.368 |0.344 |0.372 |0.339 |
|Number of Companies |470 |469 |470 |469 |470 |469 |470 |469 |
|Hausman test statistics, (2 |0.570 | |2.270 | | |1.050 | |2.220 |
| |(0.74) | |(0.00) | | |(0.31) | |(0.19) |
(Continued)
|Dependent variable |Tobin’s Q |
| |OLS |IV |OLS |IV |OLS |IV |OLS |IV |
|ACCOUNTABILITY |0.115 |0.189 | | | | | | |
| |(0.66) |(0.53) | | | | | | |
|RESPONSIBILITY | | |0.407 |0.642 | | | | |
| | | |(0.10) |(0.04) | | | | |
|FAIRNESS | | | | |0.271 |0.235 | | |
| | | | | |(0.10) |(0.00) | | |
|SOCIAL AWARENESS | | | | | | |0.151 |0.164 |
| | | | | | | |(0.34) |(0.21) |
|SIZE |-0.240 |-0.279 |-0.226 |0.096 |-0.231 |-0.192 |-0.239 |-0.260 |
| |(0.00) |(0.00) |(0.00) |(0.79) |(0.00) |(0.00) |(0.00) |(0.00) |
|R&D |0.200 |0.206 |0.193 |0.191 |0.198 |0.190 |0.183 |0.119 |
| |(0.07) |(0.09) |(0.08) |(0.00) |(0.07) |(0.07) |(0.09) |(0.00) |
|EXPORT |0.008 |0.009 |0.008 |0.010 |0.008 |0.006 |0.008 |0.006 |
| |(0.08) |(0.05) |(0.08) |(0.22) |(0.09) |(0.19) |(0.09) |(0.48) |
|F test statistics of overall |7.670 |24.310 |21.290 |3.950 |14.440 |26.360 |16.440 |43.110 |
|significance | | | | | | | | |
| |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |
| Regression R2 |0.362 |0.276 |0.366 |0.333 |0.365 |0.236 |0.371 |0.315 |
|Number of Companies |470 |469 |470 |469 |470 |469 |470 |469 |
|Hausman test statistics, (2 |0.330 | |1.140 | |1.050 | |3.170 | |
| |(0.88) | |(0.53) | |(0.31) | |(0.05) | |
Table XII
Comparison of Coefficients Estimated Using Instrumental Variables Approach vs. OLS: Investment on CLSA Corporate Governance Scores and Control Variables
This table reports and compares the results of regression:
[pic]
using OLS and IV estimation approaches, where c indexes country, i indexes industry; j indexes firm; and t indexes time. dc and di are country and industry dummies, respectively. We use legal origin (English, French, or German), ORIGIN , as instruments for corporate governance scores in IV estimation. Investment is defined as capital expenditures over total assets, 2000-20001 average. CG is one of CLSA Corporate Governance Scores (COMP, DISC, TRANS, INDEP, ACCOUNT, RESP, FAIR, SOCIAL) in 2000. Z are control variables: SIZE is the log of total assets, 1999-2000 average; R&D is research and development expenditures scaled by total assets times 100, 1999-2000 average; and EXPORT is export scaled by sales times 100, 1999-2000 average. Numbers in parentheses are probability levels at which the null hypothesis of zero coefficient can be rejected. Coefficients significant at 10% level (based on 2-tailed test) are in boldface. All p-values are based on robust (heteroscedasticity consistent) standard errors. Sample size is 468 firms. ORIGIN is not defined for Russia, thus, we drop it from IV estimation. Hausman (1978) specification test compares the IV estimates with OLS estimates. The null hypothesis is that the OLS estimator is a consistent and efficient estimator of the true parameter. P-val is the probability level at which the null hypothesis can be rejected. The Hausman statistic is calculated as [pic]where [pic]is the coefficient vector from the IV regression, [pic]is the coefficient vector from the OLS regression, [pic]is the covariance matrix of the coefficients from the IV regression, and [pic]is the covariance matrix of the coefficients from the OLS regression. Numbers in parentheses are probability levels at which the null hypothesis of zero coefficient can be rejected. Coefficients significant at 10% level (based on 2-tailed test) are in boldface. Coefficients on CG are multiplied by 100.
|Dependent variable |INVESTMENT |
| |OLS |IV |OLS |IV |OLS |IV |OLS |IV |
|COMPOSITE |0.043 |0.059 | | | | | | |
| |(0.00) |(0.02) | | | | | | |
|DISCIPLINE | | |0.020 |-0.002 | | | | |
| | | |(0.31) |(0.87) | | | | |
|TRANSPARENCY | | | | |0.045 |0.050 | | |
| | | | | |(0.00) |(0.09) | | |
|INDEPENDENCE | | | | | | |0.028 |0.025 |
| | | | | | | |(0.04) |(0.01) |
|SIZE |-0.003 |-0.002 |-0.003 |-0.003 |-0.003 |-0.003 |-0.004 |-0.004 |
| |(0.05) |(0.12) |(0.08) |(0.08) |(0.05) |(0.07) |(0.06) |(0.04) |
|R&D |0.008 |0.008 |0.008 |0.008 |0.009 |0.008 |0.008 |0.008 |
| |(0.14) |(0.15) |(0.15) |(0.15) |(0.12) |(0.15) |(0.13) |(0.14) |
|EXPORT |0.000 |0.000 |0.000 |0.000 |0.000 |0.000 |0.000 |0.000 |
| |(0.79) |(0.85) |(0.97) |(0.99) |(0.93) |(0.99) |(0.92) |(0.91) |
|F test statistics of overall |41.780 |24.050 |20.440 |19.520 |29.760 |22.000 |45.070 |35.740 |
|significance | | | | | | | | |
| |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |
| Regression R2 |0.407 |0.401 |0.388 |0.386 |0.324 |0.387 |0.352 |0.397 |
|Number of Companies |468 |468 |468 |468 |468 |468 |468 |468 |
|Hausman test statistics, (2 |1.660 | |0.980 | |0.270 | |1.100 | |
| |(0.99) | |(0.99) | |(0.10) | |(0.85) | |
(Continued)
|Dependent variable |INVESTMENT |
|Regression specification |OLS |IV |OLS |IV |OLS |IV |OLS |IV |
|ACCOUNTABILITY |0.017 |-0.455 | | | | | | |
| |(0.17) |(0.38) | | | | | | |
|RESPONSIBILITY | | |0.033 |0.031 | | | | |
| | | |(0.07) |(0.04) | | | | |
|FAIRNESS | | | | |0.003 |0.018 | | |
| | | | | |(0.81) |(0.46) | | |
|SOCIAL AWARENESS | | | | | | |-0.012 |0.289 |
| | | | | | | |(0.40) |(0.28) |
|SIZE |-0.003 |-0.001 |0.008 |0.008 |-0.003 |-0.003 |-0.003 |-0.004 |
| |(0.05) |(0.81) |(0.15) |(0.10) |(0.09) |(0.07) |(0.08) |(0.12) |
|R&D |0.008 |0.008 |0.000 |0.000 |0.008 |0.008 |0.008 |0.008 |
| |(0.10) |(0.00) |(0.98) |(0.99) |(0.15) |(0.00) |(0.13) |(0.00) |
|EXPORT |0.000 |0.000 |0.000 |0.000 |0.000 |0.000 |0.000 |0.000 |
| |(0.94) |(0.58) |(0.81) |(0.88) |(0.98) |(0.80) |(0.97) |(0.77) |
|F test statistics of overall |21.720 |12.450 |18.700 |44.710 |14.200 |6.930 |23.150 |8.390 |
|significance | | | | | | | | |
| |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |
| Regression R2 |0.390 |0.299 |0.387 |0.377 |0.387 |0.271 |0.388 |0.223 |
|Number of Companies |468 |468 |468 |468 |468 |468 |468 |468 |
|Hausman test statistics, (2 |3.990 | |2.198 | |0.518 | |1.180 | |
| |(0.99) | |(0.85) | |(0.46) | |(0.71) | |
Table XIII
Comparison of Coefficients Estimated Using Instrumental Variables Approach vs. OLS: Valuation on S&P Transparency Rankings
This table reports and compares the results of regression:
[pic]
using OLS and IV estimation approaches, where c indexes country, i indexes industry; j indexes firm; and t indexes time. dc and di are country and industry dummies, respectively. We use legal origin (English, French, or German), ORIGIN , as instruments for transparency scores in IV estimation. Valuation is defined as total assets plus market value of equity less book value of equity over total assets and market value of equity is the number of common shares outstanding times year-end share price. CG is one of S&P Transparency Rankings (AGGR, OWN, DISCL, BOARD) in 2000. Z are control variables: SIZE is the log of total assets, 1999-2000 average; R&D is research and development expenditures scaled by total assets times 100, 1999-2000 average; and EXPORT is export scaled by sales times 100, 1999-2000 average. Numbers in parentheses are probability levels at which the null hypothesis of zero coefficient can be rejected. Coefficients significant at 10% level (based on 2-tailed test) are in boldface. All p-values are based on robust (heteroscedasticity consistent) standard errors. Sample size is 550 firms. ORIGIN is not defined for Russia, thus, we drop it from IV estimation. Hausman (1978) specification test compares the IV estimates with OLS estimates. The null hypothesis is that the OLS estimator is a consistent and efficient estimator of the true parameter. P-val is the probability level at which the null hypothesis can be rejected. The Hausman statistic is calculated as [pic]where [pic]is the coefficient vector from the IV regression, [pic]is the coefficient vector from the OLS regression, [pic]is the covariance matrix of the coefficients from the IV regression, and [pic]is the covariance matrix of the coefficients from the OLS regression. Numbers in parentheses are probability levels at which the null hypothesis of zero coefficient can be rejected. Coefficients significant at 10% level (based on 2-tailed test) are in boldface. Coefficients on CG are multiplied by 100.
|Dependent variable |Tobin’s Q |
| |OLS |IV |OLS |IV |OLS |IV |OLS |IV |
|AGGREGATE |1.886 |4.406 | | | | | | |
| |(0.09) |(0.00) | | | | | | |
|OWNER | | |3.469 |16.158 | | | | |
| | | |(0.09) |(0.00) | | | | |
|DISCLOSURE | | | | |5.360 |13.500 | | |
| | | | | |(0.10) |(0.00) | | |
|BOARD | | | | | | |2.635 |10.045 |
| | | | | | | |(0.21) |(0.00) |
|SIZE |-0.169 |-0.191 |-0.158 |-0.179 |-0.177 |-0.214 |-0.159 |-0.177 |
| |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |
|R&D |0.268 |0.263 |0.271 |0.268 |0.269 |0.264 |0.268 |0.260 |
| |(0.07) |(0.07) |(0.07) |(0.07) |(0.06) |(0.06) |(0.07) |(0.07) |
|EXPORT |0.006 |0.006 |0.006 |0.006 |0.006 |0.006 |0.006 |0.005 |
| |(0.16) |(0.18) |(0.15) |(0.16) |(0.15) |(0.16) |(0.17) |(0.22) |
|F test statistics of overall |10.720 |15.440 |12.330 |11.800 |11.330 |12.150 |9.830 |17.670 |
|significance | | | | | | | | |
| |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |
|Regression R2 |0.444 |0.435 |0.441 |0.414 |0.445 |0.431 |0.442 |0.424 |
|Number of Companies |550 |550 |550 |550 |550 |550 |550 |550 |
|Hausman test statistics, (2 |0.310 | |5.260 | |0.170 | |3.130 | |
| |(0.66) | |(0.05) | |(0.84) | |(0.00) | |
Table XIV
Comparison of Coefficients Estimated Using Instrumental Variables Approach vs. OLS: Investment on S&P Transparency Rankings
This table reports and compares the results of regression:
[pic]
using OLS and IV estimation approaches, where c indexes country, i indexes industry; j indexes firm; and t indexes time. dc and di are country and industry dummies, respectively. We use legal origin (English, French, or German), ORIGIN , as instruments for transparency scores in IV estimation. Investment is defined as capital expenditures over total assets, 2000-20001 average. CG is one of S&P Transparency Rankings (AGGR, OWN, DISCL, BOARD) in 2000. Z are control variables: SIZE is the log of total assets, 1999-2000 average; R&D is research and development expenditures scaled by total assets times 100, 1999-2000 average; and EXPORT is export scaled by sales times 100, 1999-2000 average. Numbers in parentheses are probability levels at which the null hypothesis of zero coefficient can be rejected. Coefficients significant at 10% level (based on 2-tailed test) are in boldface. All p-values are based on robust (heteroscedasticity consistent) standard errors. Sample size is 549 firms. ORIGIN is not defined for Russia, thus, we drop it from IV estimation. Hausman (1978) specification test compares the IV estimates with OLS estimates. The null hypothesis is that the OLS estimator is a consistent and efficient estimator of the true parameter. P-val is the probability level at which the null hypothesis can be rejected. The Hausman statistic is calculated as [pic]where [pic]is the coefficient vector from the IV regression, [pic]is the coefficient vector from the OLS regression, [pic]is the covariance matrix of the coefficients from the IV regression, and [pic]is the covariance matrix of the coefficients from the OLS regression. Numbers in parentheses are probability levels at which the null hypothesis of zero coefficient can be rejected. Coefficients significant at 10% level (based on 2-tailed test) are in boldface. Coefficients on CG are multiplied by 100.
|Dependent variable |INVESTMENT |
| |OLS |IV |OLS |IV |OLS |IV |OLS |IV |
|AGGR |0.096 |0.183 | | | | | | |
| |(0.02) |(0.01) | | | | | | |
|OWN | | |0.145 |0.685 | | | | |
| | | |(0.18) |(0.00) | | | | |
|DISCL | | | | |0.263 |0.415 | | |
| | | | | |(0.02) |(0.03) | | |
|BOARD | | | | | | |0.155 |0.418 |
| | | | | | | |(0.05) |(0.01) |
|SIZE |-0.003 |-0.004 |-0.002 |-0.003 |-0.003 |-0.004 |-0.002 |-0.003 |
| |(0.28) |(0.22) |(0.39) |(0.28) |(0.23) |(0.22) |(0.36) |(0.28) |
|R&D |0.002 |0.002 |0.003 |0.003 |0.003 |0.002 |0.002 |0.002 |
| |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.01) |
|EXPORT |0.000 |0.000 |0.000 |0.000 |0.000 |0.000 |0.000 |0.000 |
| |(0.11) |(0.08) |(0.14) |(0.14) |(0.11) |(0.10) |(0.11) |(0.07) |
|F test statistics of overall |18.360 |21.070 |17.760 |18.310 |18.030 |18.610 |18.000 |23.570 |
|significance | | | | | | | | |
| |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |(0.00) |
| Regression R2 |0.379 |0.373 |0.374 |0.344 |0.380 |0.378 |0.376 |0.362 |
|Number of Companies |549 |549 |549 |549 |549 |549 |549 |549 |
|Hausman test statistics, (2 |0.920 | |3.280 | |0.414 | |0.960 | |
| |(0.43) | |(0.00) | |(0.38) | |(0.20) | |
Table XV
CLSA Sample
Regressions with Centered Independent Variables and the Presence of Within-Country Correlations
This table reports the results of regressions:
[pic](Panel A)
[pic] (Panel B)
[pic] (Panel C)
[pic] (Panel D)
where c indexes country; i indexes industry; j indexes firm; and t indexes time. CG is CLSA Corporate Governance Composite Score in 2000 minus the sample mean (in Panels C and D). Valuation is 2000-2001 average Tobin’s Q defined as total assets plus market value of equity less book value of equity over total assets and market value of equity is the number of common shares outstanding times year-end share price. Investment is 2000-2001 average ratio of capital expenditures to total assets. INV_OPP (investment opportunities) is 1999-2000 average of return on invested capital minus the sample mean; EXT_FIN (external financing) is 1999-2000 average ratio of the sum of new equity issues and changes in long-term debt over capital expenditures minus the sample mean. OWNERSHIP is 1999-2000 average percentage of shares held by major shareholders, either individuals or corporations, that hold more than 5% of outstanding shares minus the sample mean; OWNERSHIP^2 is the squared term of ownership; LEGAL is INVESTOR*ENFORCE minus the sample mean where INVESTOR measures investor protection and ENFORCE (enforcement) is the rule of law. Z are control variables: ROIC (return on invested capital), 1999-2000 average; SIZE is the log of total assets, 1999-2000 average; R&D is research and development expenditures scaled by total assets times 100, 1999-2000 average; and EXPORT is export scaled by sales times 100, 1999-2000 average. d are industry dummies (coefficients are not reported). P-values are based on standard errors that allow for within countries correlations. Numbers in parentheses are probability levels at which the null hypothesis of zero coefficient can be rejected. The estimator of coefficient’s variance when observations are correlated within groups (countries or industries in our case) but uncorrelated between groups is [pic], were M is number of groups, N number of observations, k number of estimated parameters, [pic] conventional estimator of variance, L likelihood function, β is the vector of parameters estimated and [pic]the contribution of the kth group to the scores [pic]. Coefficients significant at 10% level (based on 2-tailed test) are in boldface. Coefficients on CG are multiplied by 100.
|Panel |A |B | | |C |D |
|Dependent variable |COMPOSITE |COMPOSITE | | |Tobin’s Q |INVESTMENT |
|INV_OPP |7.292 |OWNERSHIP |25.509 |COMPOSITE |0.912 |0.042 |
| |(0.03) | |(0.05) | |(0.00) |(0.00) |
|EXT_FIN |1.179 |OWNERSHIP^2 |-24.818 |LEGAL |0.006 |0.000 |
| |(0.03) | |(0.05) | |(0.25) |(0.13) |
|LEGAL |0.484 |LEGAL |0.538 |CG*LEGAL |-0.002 |-0.003 |
| |(0.00) | |(0.00) | |(0.40) |(0.00) |
|INV_OPP*LEGAL |-0.152 |OWNERSHIP*LEGAL |-0.536 | | | |
| |(0.59) | |(0.07) | | | |
|EXT_FIN*LEGAL |-0.064 | | |ROIC |1.759 |0.007 |
| |(0.11) | | | |(0.00) |(0.41) |
|SIZE |0.418 |SIZE |0.293 |SIZE |-0.157 |0.000 |
| |(0.54) | |(0.55) | |(0.01) |(0.94) |
|R&D |-0.081 |R&D |-0.177 |R&D |0.184 |0.009 |
| |(0.93) | |(0.71) | |(0.06) |(0.11) |
|EXPORT |-0.044 |EXPORT |-0.052 |EXPORT |0.008 |0.000 |
| |(0.22) | |(0.28) | |(0.04) |(0.23) |
|F test statistics of joint significance | |F test statistics of joint significance | |F test statistics of joint| | |
| | | | |significance | | |
|INV_OPP=0 and INV_OPP*LEGAL=0 |2.86 |OWNERSHIP=0 and OWNERSHIP*LEGAL=0 |4.02 |CG=0 and CG*LEGAL=0 |14.96 |8.99 |
| |(0.08) | |(0.02) | |(0.00) |(0.00) |
|EXT_FIN=0 and EXT_FIN*LEGAL=0 |3.38 | | | | | |
| |(0.05) | | | | | |
|F test statistics of overall significance|75.50 |F test statistics of overall significance |94.88 |F test statistics of |662.75 |868.36 |
| | | | |overall significance | | |
| |(0.00) | |(0.00) | |(0.00) |(0.00) |
| Regression R2 |0.216 |Regression R2 |0.229 | Regression R2 |0.454 |0.364 |
|Number of Companies |460 |Number of Companies |380 |Number of Companies |470 |468 |
Table XVI
S&P Sample:
Regressions with Centered Independent Variables and the Presence of Within-Country Correlations
This table reports the results of regressions:
[pic](Panel A)
[pic] (Panel B)
[pic] (Panel C)
[pic] (Panel D)
where c indexes country; i indexes industry; j indexes firm; and t indexes time. CG is S&P Corporate Governance Aggregate Score in 2000 minus the sample mean (in Panels C and D). Valuation is 2000-2001 average Tobin’s Q defined as total assets plus market value of equity less book value of equity over total assets and market value of equity is the number of common shares outstanding times year-end share price. Investment is 2000-2001 average ratio of capital expenditures to total assets. INV_OPP (investment opportunities) is 1999-2000 average of return on invested capital minus the sample mean; EXT_FIN (external financing) is 1999-2000 average ratio of the sum of new equity issues and changes in long-term debt over capital expenditures minus the sample mean. OWNERSHIP is 1999-2000 average percentage of shares held by major shareholders, either individual or corporations, that hold more than 5% of outstanding shares minus the sample mean; OWNERSHIP^2 is the squared term of ownership; LEGAL is INVESTOR*ENFORCE minus the sample mean where INVESTOR measures investor protection and ENFORCE (enforcement) is the rule of law. Z are control variables: ROIC (return on invested capital), 1999-2000 average; SIZE is the log of total assets, 1999-2000 average; R&D is research and development expenditures scaled by total assets times 100, 1999-2000 average; and EXPORT is export scaled by sales times 100, 1999-2000 average. d are industry dummies (coefficients are not reported). P-values are based on standard errors that allow for within countries correlations. Numbers in parentheses are probability levels at which the null hypothesis of zero coefficient can be rejected. The estimator of coefficient’s variance when observations are correlated within groups (countries or industries in our case) but uncorrelated between groups is [pic], were M is number of groups, N number of observations, k number of estimated parameters, [pic] conventional estimator of variance, L likelihood function, β is the vector of parameters estimated and [pic]the contribution of the kth group to the scores [pic]. Coefficients significant at 10% level (based on 2-tailed test) are in boldface. Coefficients on CG are multiplied by 100.
|Panel |A |B | | |C |D |
|Dependent variable |AGGREGATE |AGGREGATE | | |Tobin’s Q |INVESTMENT |
|INV_OPP |9.881 |OWNERSHIP |7.000 |AGGREGATE |0.881 |0.031 |
| |(0.10) | |(0.05) | |(0.05) |(0.15) |
|EXT_FIN |5.332 |OWNERSHIP^2 |-28.962 |LEGAL |0.006 |0.000 |
| |(0.03) | |(0.01) | |(0.23) |(0.08) |
|LEGAL |0.414 |LEGAL |0.197 |CG*LEGAL |-0.029 |-0.007 |
| |(0.00) | |(0.00) | |(0.58) |(0.13) |
|INV_OPP*LEGAL |-0.033 |OWNERSHIP*LEGAL |-0.081 | | | |
| |(0.88) | |(0.76) | | | |
|EXT_FIN*LEGAL |-0.110 | | |ROIC |2.500 |0.046 |
| |(0.36) | | | |(0.07) |(0.03) |
|SIZE |1.282 |SIZE |1.261 |SIZE |-0.257 |-0.002 |
| |(0.00) | |(0.00) | |(0.12) |(0.42) |
|R&D |0.029 |R&D |0.179 |R&D |0.270 |0.003 |
| |(0.05) | |(0.17) | |(0.12) |(0.11) |
|EXPORT |-0.050 |EXPORT |-0.002 |EXPORT |0.006 |0.000 |
| |(0.17) | |(0.93) | |(0.10) |(0.91) |
|F test statistics of joint | |F test statistics of joint | |F test statistics of joint significance| | |
|significance | |significance | | | | |
|INV_OPP=0 and INV_OPP*LEGAL=0 |3.18 |OWNERSHIP=0 and |3.74 |CG=0 and CG*LEGAL=0 |7.08 |1.59 |
| | |OWNERSHIP*LEGAL=0 | | | | |
| |(0.08) | |(0.02) | |(0.00) |(0.23) |
|EXT_FIN=0 and EXT_FIN*LEGAL=0 |4.18 | | | | | |
| |(0.04) | | | | | |
|F test statistics of overall |103.08 |F test statistics of overall|32.42 |F test statistics of overall |157.43 |128.24 |
|significance | |significance | |significance | | |
| |(0.00) | |(0.00) | |(0.00) |(0.00) |
| Regression R2 |0.236 |Regression R2 |0.177 | Regression R2 |0.392 |0.257 |
|Number of Companies |541 |Number of Companies |458 |Number of Companies |550 |549 |
-----------------------
*Doctoral Student, Department of Finance, University of Michigan Business School, Ann Arbor, Michigan 48109-1234. Tel: (734) 358-2427. E-mail: adurnev@umich.edu.
** Fred M. Taylor Professor of Business Administration, University of Michigan Business School, Ann Arbor, MI 48109-1234. Tel: (734) 764-5222 & (734) 764-2282. Fax: (734) 763-3117. Email: ehkim@umich.edu.
We are grateful for helpful comments and suggestions by Sugato Bhattacharyya, Serdar Dinç, Mara Faccio, Michael Fuerst, Kathleen Fuller, Charles Hadlock, Woochan Kim, Florencio Lopez-de-Silanes, Vojislav Maksimovic, Todd Mitton, M. P. Narayanan, Andrei Shleifer, David Smith, Daniel Wolfenzon, Bernard Yeung, and participants of the Conference on International Corporate Governance at the Tuck School of Business, 2nd Asian Corporate Governance Conference in Seoul, Korea, Estes Park Conference; and seminars at the University of Michigan and KAIST where earlier versions of the paper were presented under different titles. We would like to thank Nick Bradley, Ian Byrne, George Dallas, and Laurie Kizik for providing us with S&P Transparency Rankings and CLSA Corporate Governance Scores, and Vlad Cherniavsky for excellent research assistance.
[1] See Johnson (2002) for the effects of legal regimes on corporate capital structure.
[2] These are the variables used by Credit Lyonnais Securities Asia (CLSA) in constructing their corporate governance rankings, which is one of our empirical proxies for the quality of corporate governance. The other proxy is Standard and Poor’s transparency rankings, which is based on the number of items a firm chooses to disclose concerning ownership structure and investor relation, accounting and financial policies, and board and management structure and process (see Appendix B).
[3] That is, firm i has its own [pic], but we do not put subscript i to [pic]for notational simplicity.
[4] This assumption of no external financing is relaxed later.
[5] We assume that diversion occurs before the investment decision is made (as in Johnson et al. (2000)) and the cost of diversion is linear in the amount of funds diverted. Johnson et al. (2000), La Porta et al. (2002), and Shleifer and Wolfenzon (2002) assume that the cost of diversion is convex. We do not need the convexity assumption, and the results do not change with a convex cost function.
[6] A potential weakness of our approach is that we assume c, the strength of legal regimes, is exogenous to the observed level of malfeasance, d. As witnessed during the recent corporate scandals following the Enron debacle, governments react to revelations of wide spread corporate misdeeds, with those more responsive to public opinions being more inclined to undertake legal reforms. Reforming legal systems, however, is a slow process because it inevitably becomes a political issue, with the controlling shareholders—and those who benefit by association—insisting on the status quo (see Bebchuk and Roe (1999) and Rajan and Zingales (2001)). Moreover, if the number of firms is sufficiently large, the controlling shareholder may behave as if her actions will have no effect on the cost of diversion. Alternatively, c can be interpreted as the controlling shareholder’s perceived cost of diversion that takes into account the effects of her action on future legal reforms.
[7] The budget constraint, [pic], may become binding if e is negative, i.e. the endowment is insufficient to fund all non-negative NPV projects. Although the derivations in the text assume non-negative e all the results continue to hold with negative e.
[8] If e < 0,
[pic] [3A]
The maximum investment is constrained by [pic] and one of the conditions in Eq.[3] changes from [pic] to [pic].
[9] When the cost of diversion is so small that [pic] the optimal diversion will be the entire endowment [pic] and all positive NPV projects will be rejected. This is because when [pic] the controlling stockholder’s share of the liquidating dividend from the highest return project, [pic], is less than the after-cost diversion, [pic].
[10] If e < 0,
[pic] [5A]
and all subsequent discussions and derivatives of d* in the text remain valid.
[11] Eq. [6] also shows that the partial derivative is zero when [pic], which means that when investor protection is so weak (i.e., c is so small) that the firm’s entire resources are diverted, profitability of projects makes no difference. However, even in such a legal regime, a sufficient increase in profitability ([pic]) will lead some firms to switch into the condition of [pic], where firms make positive investments. Thus, even in a virtual absence of investor protection, some firms will invest if profits are sufficiently large.
[12] The intuition for this proposition requires a two-step explanation: When the cost of diversion is high, the net benefit of diversion is low and the optimal tradeoff point at which the marginal benefit of diversion is equal to the marginal benefit of investment occurs at a relatively low level of d. It would take much greater increases in profitability to bring the trade-off point to an even lower level of d in a high c country than in a low c country where the trade-off point occurs at a relatively high level of d.
[13] Assume two-firm economies in which both firms L and H exist in the U.S. and in Russia such that the deviation in [pic] between the two firms in both countries is the same and equal to 0.2. But the deviation in d* is 0.06 in Russia and 0.02 in the U.S. Thus, if the variation in profitable investment opportunities across firms were the same, firms in Russia would show greater variation in corporate governance than those in the U.S. The variation in corporate governance also depends on the level of [pic] itself because d* is negatively related to [pic]. To illustrate, consider again a two-firm economy in which firm L’s [pic] is 3 and firm H’s [pic] is 3.3, the same 10% difference as above. In this case, however, d*L = 0.44 and d*H = 0.40 in Russia and, d*L = 0.11 and d*H = 0.10 in the U.S., reducing the deviation in d* to 0.04 in Russia and 0.01 in the U.S.
[14] An alternative scenario that provides the same prediction is that the law and country-level institutions set a minimum standard on the quality of governance that firms cannot go below. The minimum standard truncates the distribution of governance qualities in which the truncation point–the minimum standard–affects the observed variation. However, this explanation requires an assumption that the law is effective in enforcing and maintaining the minimum standard, which appears to have been broken rather blatantly in the recent corporate scandals in the U.S. In contrast, our approach requires no limit on d–the minimum standard, as we proxy the strength of regulation by the private cost c.
[15] Shleifer and Wolfenzon (2002) obtain a similar result under a more restrictive set of assumptions.
[16] It does not directly follow from our model that the relation of corporate governance with external financing, valuation, and investment is greater in less investor-friendly regimes.
[17] Although issued in March 2001, we treat CLSA scores as 2000 data because the companies were rated in 2000.
[18] This is equivalent to assigning an equal weight to each disclosed item.
[19] The regression is:
COMP = 0.17(AGGR + (i di + (c dc R2 = 0.45,
[0.06]
where di and dc are industry and country dummies (coefficients not reported), respectively, and the number in parentheses is the probability level at which zero coefficient can be rejected.
[20] It is possible that the firms in CLSA and S&P rankings may suffer from selection and reporting bias; namely, only firms with good governance practices may cooperate with CLSA’s survey and the ranking agencies may choose firms that are easier to assign scores. However, companies have incentives to cooperate because exclusion from the ranking may create ill reputation, and the ranking agencies have commercial interest to have a well-balanced portfolio of companies on their lists.
[21] Worldscope defines return on invested capital as (Net Income before Preferred Dividends + ((Interest Expense on Debt - Interest Capitalized) * (1-Tax Rate)) / (Last Year's Total Capital + Last Year's Short Term Debt & Current Portion of Long Term Debt).
[22] An index is formed by adding 1 when (1) the country allows shareholders to mail their proxy vote to the firm; (2) shareholders are not required to deposit their shares prior to the general shareholders’ meeting; (3) cumulative voting or proportional representation of minorities mechanism is in place; (4) an oppressed minorities mechanism is in place; (5) the minimum percentage of share capital that entitles a shareholder to call for an extraordinary shareholders’ meeting is less than or equal to 10 percent; or (6) shareholders have preemptive rights that can be waived only by a shareholders’ vote.
[23] ENFORCE is calculated as 1982-1995 average for all countries except Russia. For Russia, it is calculated in 1998. La Porta et al. (1997) also use the rule of law as a measure of law enforcement. An alternative is to use the efficiency of the judicial system index reported in La Porta et al. (1998), which is defined as assessment of the efficiency and integrity of the legal environment as it affects business. We decided to use the rule of law for two reasons. First, using the efficiency of the judicial system as a proxy for law enforcement would reduce our sample size because this variable is not defined for China, Hungary, Poland, and Russia in La Porta et al. (1998). Second, the two variables are highly correlated. The correlation between the rule of law and the efficiency of the judicial system is 0.64 (p-val = 0.00). As a robustness check we drop firms in countries for which this measure is not defined and use the efficiency of the judicial system as a measure of law enforcement.
[24] Our main results do not change if INVESTOR and ENFORCE enter [S1] and [S2] separately rather than as a product.
[25] INVESTOR, ENFORCE, and ORIGIN are defined in La Porta et al. (1998) for each country in our sample except China, Hungary, Russia and Poland. For these countries we rely on the variables reported in Claessens et al. (1999) and Pistor et al. (2000). ORIGIN is not defined for Russia.
[26] CLSA and S&P samples contain 102 and 109 financial institutions, respectively. It is customary to exclude them (SICs 6,000 through 6,999) because their financial and accounting statements are not comparable to those of non-financial firms. Although including industry dummies partially mitigates this problem, we repeat all regressions excluding banks and financial firms as a robustness check.
[27] For consistency we also report how the country average CLSA measure of investor protection (fairness) is related to LEGAL and how country average CLSA composite score is related to LEGAL defined as the country minimum governance score. Figure 3B and 3D indicates that the relations between the two are positive and significant.
[28] The results are also qualitatively the same if we use spread, defined as the difference between country maximum and minimum scores, instead of the variance of the scores.
[29] The correlation between LEGAL and the level of investment is negative, which implies firms in poor legal regimes invest more. Although this may be due to over-investment, it is difficult to determine the cause because the investment data do not distinguish between profitable and unprofitable investments.
[30] In contrast to the CLSA scores, S&P rankings are strongly related to the control variables; namely, larger firms and firms with more R&D expenditures tend to disclose more, while firms that are more export-oriented tend to disclose less.
[31]The average CLSA composite corporate governance scores in common law (English) countries and civil law (German and French) countries are 59.28 and 51.46, respectively, and they are significantly different from each other at 0.001% level. The average S&P aggregate transparency scores are 45.5 and 39.04 in common law and civil law countries, and the difference is again highly significant.
[32] Average values of Tobin’s Q are 1.96 and 1.48 in common law and civil law countries, respectively, and the averages are significantly different at 10% but not at 5%. The average level of investment is 0.088 in common law countries, which is not significantly different from an average level of investment of 0.115 in civil law countries at 10%.
[33] Unlike specification S[1] and S[2], S[3] does not control for the quality of legal regime, LEGAL, or past investment opportunities, ROIC, because both of them are endogenous. For identification purposes the number of instrumented variables, corporate governance scores in our case, cannot exceed the number of instruments. For this reason S[3] includes country dummies to control for country-specific characteristics.
[34] Since legal origin is not defined for Russia, we excluded Russia from Instrumental Variables estimation.
[35] Hausman specification test (Hausman (1978)) compares the IV estimates with OLS estimates. The null hypothesis is that the OLS estimator is a consistent and efficient estimator of the true parameter. It is calculated as [pic], where [pic] is the coefficient vector from the IV regression, [pic]is the coefficient vector from the OLS regression, [pic]is the covariance matrix of the coefficients from the IV regression, and [pic]is the covariance matrix of the coefficients from the OLS regression.
[36] Unfortunately, we could not use the IV approach for S[1] when investment profitability, reliance on external financing, and ownership concentration enter as independent variables because we are unable to identify suitable instruments that would be closely related to those variables but unrelated to firm corporate governance structure.
[37] The estimator of coefficient’s variance when observations are correlated within groups (countries or industries in our case) but uncorrelated between groups is [pic], were M is the number of groups, N the number of observations, k the number of estimated parameters, [pic] the conventional estimator of variance, L the likelihood umber of groups, N the number of observations, k the number of estimated parameters, [pic] the conventional estimator of variance, L the likelihood function, β the vector of parameters estimated and [pic]the contribution of the kth group to the scores [pic] (see Huber (1967)).
[38] For example, in CLSA sample, the correlation between the original EXT_FIN and EXT_FIN*LEGAL is 0.90 which drops to 0.23 after we center EXT_FIN and LEGAL.
[39] To conserve space we do not report the estimation results for S[1] with country fixed effects but without LEGAL and interaction terms. The results are similar to those reported in Tables V-VII. Specifically, INV_OPP, EXT_FIN, and OWNERSHIP are positively and significantly related to most of the CLSA and S&P attributes. Furthermore, regressions with country fixed effects in Tables XI-XIV do not include interaction terms. When we estimate S[1] and S[2] with country fixed effects along with interaction terms but without LEGAL itself, the results are similar to those reported in Tables V-X.
[40] Creditor protection, efficiency of judicial system, risk of expropriation, risk of contract repudiation, and corruption index are defined in La Porta et al. (1998).
[41] In the CLSA sample these countries are Argentina, Columbia, Czech Republic, Greece, Hungary, Peru, Poland, and Russia. In the S&P sample only New Zealand falls in this category.
[42] See Johnson and Shleifer (2000) for an excellent literature review on the debate concerning the Coasian argument in corporate governance.
[43] These numbers are based on the magnitude of OLS coefficients in Tables X and XIII. The percentage is calculated against the median of Tobin’s Q.
[44] According to a study by a Russian investment bank, Troika Dialog, “poor image of corporate governance is currently downgrading the value of quoted shares in Russia by approximately $50bln on a simple valuation model. That means a two-fold downgrading of the Russian market, as the total value of the country’s capital market is $47bln…” (see Vardanian (2001)). In a recent survey conducted by McKinsey, two-thirds of 119 large institutional investors said they would pay more for the stock of a company with good corporate governance (see the McKinsey Quarterly (1998)).
[45]We obtain Eq. [A4] because [pic], and hence, [pic].
[46] We thank Daniel Wolfenzon for pointing this out.
-----------------------
B
A
0
[pic]
[pic]
j
[pic]
C
1
................
................
In order to avoid copyright disputes, this page is only a partial summary.
To fulfill the demand for quickly locating and searching documents.
It is intelligent file search solution for home and business.
Related searches
- university of michigan admissions staff
- university of michigan admission requirement
- university of michigan sat scores
- university of michigan payroll office
- university of michigan application deadline
- university of michigan act requirements
- university of michigan entrance requirements
- university of michigan transfer deadline
- university of michigan philosophy dept
- university of michigan applicant portal
- university of michigan neurology
- university of michigan hospital neurology