**3. Empirical study**

#### **3.1. Data and variable measurement**

In this study we analyze the impact of corporate governance on the financial performance of a sample of EM firms and ADRs based in 22 emerging countries over a period extending from 2008–2014. Our analysis is restricted to this seven-year period because of the scarcity of reliable governance data before 2008. The countries are drawn from the 22 best performing emerging markets for the year 2014, as ranked by the *Bloomberg Visual Data*. The ranking is based on each country's GDP growth, inflation, the level of government debt, the annual change in government debt, currency purchasing power, and total investment as percentage of GDP. The 22 emerging countries in ranked order are: China, South Korea, Malaysia, Chile, Thailand, Panama, Peru, Latvia, Poland, Czech Republic, Columbia, Turkey, Hungary, Russia, Brazil, Philippines, Mexico, Indonesia, South Africa, Morocco, India and Egypt. Unfortunately, Morocco, Latvia, Egypt and the Czech Republic have been dropped from the sample because of missing firm-year observations. Moreover, our sample of ADRs excludes Malaysia, Chile, Thailand, Panama, Peru, Poland, Colombia, Philippines and India, either because these ADRs do not exist or because the data are missing for these countries. In total, the sample comprises 10,045 non-cross listed EM firm-year observations of companies that are publicly traded on the stock market of 18 EM countries and 610 cross-listed ADR firm-year observations of companies trading on the NASDAQ and NYSE in the US.

Additionally, firms will forge formal ethics policies to enhance the decision-making process as well as their market image. In line with the findings in the extant literature, and following

**H4:** EM Firms with a greater number of committees will exhibit better financial performance

**H5:** EM firms with formal ethics policies will exhibit better financial performance over time.

Prior literature documents a negative relationship between contemporaneous risk and governance; perhaps because establishing an effective corporate governance regime protects shareholders and creditors and serves to reduce the expropriation of power by managers and controlling shareholders [43–45]. Earlier studies also document risk as an important determinant of a firm's financial performance [46, 47]. For firms in the emerging markets, however, identifying an appropriate measure of risk can present problems when estimating the cost of capital because of the time-varying nature of the integration of these markets into the global economy. Nonetheless, the exposure of a firm to the risks inherent in its local market remains an important element of the overall risk assessment [48–50]. Although, research has demonstrated that both market risk and governance have a significant impact on the financial performance of firms, it remains necessary to develop a more complex and dynamic theoretical model than has been previously considered; a model that reflects how market risk moderates the governance-performance relationship in the emerging markets. We therefore posit

**2.3. The moderating effect of risk on governance-performance relationship**

the following hypotheses for each of the aforementioned governance indicators:

nance score and its financial performance over time.

financial performance among EM firms over time.

**3. Empirical study**

**3.1. Data and variable measurement**

pendent directors and its financial performance over time.

tees and financial performance among EM firms over time.

ethics policy and financial performance among EM firms over time.

**H6:** Greater market risk negatively moderates the relationship between an EM firm's gover-

**H7a:** Greater market risk negatively moderates the relationship between an EM firm's inde-

**H7b:** Greater market risk positively moderates the relationship between CEO Duality and

**H7c:** Greater market risk negatively moderates the relationship between number of commit-

**H7d:** Greater market risk negatively moderates the relationship between the presence of an

In this study we analyze the impact of corporate governance on the financial performance of a sample of EM firms and ADRs based in 22 emerging countries over a period extending

Bonding Theory, we hypothesize the following:

154 Financial Management from an Emerging Market Perspective

over time.

The financial and corporate governance data on the EM firms and ADRs were accessed through the *Bloomberg Database.* Financial performance is measured by the return on assets ratio (ROA), estimated by dividing a company's annual net income by its total assets. In the literature, ROA is frequently used as a measure for profitability and financial performance, especially in studies investigating corporate governance structure [51–53]. A firm's Governance Score was attained from the Bloomberg Professional database and is based on the extent of a company's governance disclosure, where the score ranges from 0.1 for companies that disclose a minimum amount of governance data to a maximum of 100. The *Independent Directors* variable is measured as the number of independent directors on the board divided by the board size; *Committees* is measured as the total number of committees (including audit, nomination and compensation committees) established by the company and ranges between a value of 0 and 3; *CEO Duality* is employed as a dummy variable that takes a value of 1 if the CEO of the company also serves as the chairman of the board and 0 otherwise; *Ethics policy* also is a dummy variable that takes a value of 1 if the company has a formal ethics policy and 0 otherwise. Finally, we include a market-based measure of risk, the CAPM-beta, as a moderator between governance and performance, because of its frequent use in the literature [30, 43, 44, 54]. For each firm, beta (*β<sup>t</sup>* ) is estimated for each year *t* based on daily stock prices: a beta greater than 1 indicates that a stock is more volatile than the market.

To control for potentially confounding, exogenous factors, we include a number of company, industry and country-level variables. At the firm level, we control for leverage because of its significant impact on performance through the cost of capital and capital budgeting decisions [55, 56]. Leverage is measured as the ratio of total debt to total assets. We also include Tobin's Q to control for investment and growth opportunities, because it diverges considerably across countries and captures important elements of firm performance [57, 58]. Tobin's Q is measured as the sum of a firm's market capitalization, total liabilities, preferred equity and minority interest divided by total assets. We also control for firm size, which is computed as the log of a firm's total assets. Industry effects are captured using nine indicator variables; one for each Industry Classification Benchmark (ICB) code. The nine relevant industries include oil and gas, basic materials, industrials, consumer goods, health care, consumer services, telecommunications, utilities, financials and technology. Finally, we employ an indicator variable for each country to control for possible fixed effects related to the national origin of a firm.

#### **3.2. Descriptive statistics and correlations**

Descriptive statistics for the cross-listed ADRs and non-cross listed EMs are presented in **Table 1**. The results show that ADR firms generate a higher governance disclosure score than the EM firms. Also, cross-listed ADRs tend to have a larger proportion of independent board members, are almost twice as likely to have a formal ethics policy as EM firms and tend to have more committees. Conversely, EM firms are much more likely to have to have a CEO that also serves as chairman of the board (i.e., CEO duality). In comparing the estimated betas of the two groups of firms, it becomes apparent that ADRs are riskier than EM firms, but this may stem from the residual elements of local market segmentation [48–50]. Finally, the profitability of EM firms appears greater over this period than that of the ADR firms, suggesting that there is a notable difference, not only in the governance structure of these firms, but also in the performance between cross-listed ADRs and non-cross-listed EM firms.



This table provides descriptive statistics for the cross-listed ADRs and non-cross listed EM firms in the sample. A firm is defined cross-listed if it was continuously cross-listed in the US market for the period 2008–2014. All cross-listed and non-cross listed firms are publicly traded EM firms that are from *Bloomberg*'s Best Emerging Markets 2014 list.

**Table 1.** Descriptive statistics cross-listed ADRs and non-cross-listed EM firms.

The Pearson correlation coefficients and corresponding significance levels are reported in **Table 2**. Prior to estimating the determinants of financial performance presented in the theory and hypotheses above, we perform a preliminary screening for collinearity. **Table 2** reveals no evidence of collinearity among the independent variables; no Pearson correlation exceeds │0.418│ and the variance inflation factors in later multivariate estimations remain small. Moreover, in line with the theory, the correlation estimates indicate that the Governance Score, Independent Directors, the Number of Committees and CEO Duality are significantly and positively correlated with financial performance (*p* < 0.01), while the coefficient estimate on CEO Duality is negative but insignificant. An examination of the correlation estimates between risk, performance and governance, reveals that the estimated betas are positively and significantly correlated with CEO Duality (*p* < 0.10) and the Number of Committees (*p* < 0.05), while they are negatively correlated with Ethics Policy. Although we expect a negative correlation between risk and the Number of Committees, the positive correlation might be related to the relatively large size of the firms establishing a greater number of committees. That is, given the positive and significant correlation between firm size and risk (*p* < 0.01), a positive correlation between risk and the Number of Committees is not surprising.

#### **3.3. Empirical model**

include oil and gas, basic materials, industrials, consumer goods, health care, consumer services, telecommunications, utilities, financials and technology. Finally, we employ an indicator variable for each country to control for possible fixed effects related to the national origin

Descriptive statistics for the cross-listed ADRs and non-cross listed EMs are presented in **Table 1**. The results show that ADR firms generate a higher governance disclosure score than the EM firms. Also, cross-listed ADRs tend to have a larger proportion of independent board members, are almost twice as likely to have a formal ethics policy as EM firms and tend to have more committees. Conversely, EM firms are much more likely to have to have a CEO that also serves as chairman of the board (i.e., CEO duality). In comparing the estimated betas of the two groups of firms, it becomes apparent that ADRs are riskier than EM firms, but this may stem from the residual elements of local market segmentation [48–50]. Finally, the profitability of EM firms appears greater over this period than that of the ADR firms, suggesting that there is a notable difference, not only in the governance structure of these firms, but also

in the performance between cross-listed ADRs and non-cross-listed EM firms.

**Variables Mean Median Std. Dev. %25 %75**

ROA 6.161 5.006 9.139 1.922 9.818 Governance score 46.844 44.643 7.742 42.857 51.786 Committees 2.074 2 .966 1 3 Independent directors 44.178 42.860 13.739 33.333 54.550 CEO duality 0.237 0 0.425 0 0 Ethics policy 0.433 0 0.496 0 1 Leverage 0.237 0.200 0.179 0.093 0.346 Firm size 7.203 7.156 2.082 5.850 8.487 Tobin's Q 0.424 0.300 0.607 0.020 0.751 Beta 0.954 0.972 0.738 0.781 1.141

ROA 5.633 4.895 9.370 1.285 9.151 Governance score 50.248 50 9.560 42.857 57.143 Committees 2.219 3 0.953 1 3 Independent directors 48.063 44.444 19.784 33.333 62.500 CEO duality 0.160 0 0.368 0 0 Ethics policy 0.827 1 0.379 1 1

of a firm.

**EM firms**

**ADRs**

**3.2. Descriptive statistics and correlations**

156 Financial Management from an Emerging Market Perspective

In this study a two-step Generalized Least Squares (GLS) random effects model is employed to capture both cross-sectional and time-series variation in the data. We implement a double transformation of the data to correct for autocorrelation as well as potential heteroskedasticity problems, in which the data are corrected for unobservable, firm-specific and time-invariant effects [59, 60]. The Two-step GLS random effects regression analyses are specified from a general model that represents return on assets (ROA), our proxy for financial performance, as a function of various corporate governance indicators and control variables. In the first model, financial performance is defined as a function of an overall governance disclosure score as shown in Eq. (1):

$$\text{ROA}\_{\text{it}} = \beta\_0 + \beta\_1 \text{Covernment\\_Score}\_{\text{it}} + \text{Controls}\_{\text{it}} + \varepsilon\_{\text{it}} \tag{1}$$


cSignificance at the 0.10 level.

In the second model, each of four governance variables are incorporated into the governance – performance regression for each of the samples as in Eq. (2):

$$\begin{array}{l} \text{ROA}\_{\text{it}} = \beta\_0 + \beta\_1 \text{CEO\\_Duality}\_{\text{it}} + \beta\_2 \text{Independent\\_Direcitors}\_{\text{it}}\\ + \beta\_3 \text{Committees}\_{\text{it}} + \beta\_4 \text{Ethics\\_Poicy}\_{\text{it}} + \text{Controls}\_{\text{it}} + \varepsilon\_{\text{it}} \end{array} \tag{2}$$

The third model introduces beta as a moderator variable for risk into the governance-performance regression and includes an interaction variable for the governance disclosure score with Beta as exhibited in Eq. (3):

$$\begin{array}{l} \text{ROA}\_{\text{it}} = \beta\_0 + \beta\_1 \text{Govername\\_Score}\_{\text{it}} + \beta\_2 \text{Beta}\_{\text{it}}\\ \text{+ } \beta\_3 \text{Beta}\_{\text{it}} \text{ \*} \text{Covername\\_Score}\_{\text{it}} \text{ \*} \text{Coverolds}\_{\text{it}} + \varepsilon\_{\text{it}} \end{array} \tag{3}$$

Finally in the fourth model, we employ a moderator variable, the CAPM Beta. This governance-performance regression includes an interaction term for each of the four governance indicators with Beta as displayed in Eq. (4):

$$\begin{array}{l} \text{ROA}\_{\text{it}} = \beta\_0 + \beta\_1 \text{CEO\\_Duality}\_{\text{it}} + \beta\_2 \text{Independent\\_Direators}\_{\text{it}}\\ \quad + \beta\_3 \text{Committeess}\_{\text{it}} + \beta\_4 \text{Ethics\\_Policy}\_{\text{it}} + \beta\_5 \text{Beta}\_{\text{it}}\\ \quad + \beta\_s \text{Beta}\_{\text{it}} \text{\\_CEO\\_Duality}\_{\text{it}} + \beta\_\gamma \text{Beta}\_{\text{it}} \text{\\_Independence\\_Direators}\_{\text{it}}\\ \quad + \beta\_s \text{Beta}\_{\text{it}} \text{\\_Compmitteess}\_{\text{it}} + \beta\_s \text{Beta}\_{\text{it}} \text{\\_Ehics\\_Policy}\_{\text{it}} + \text{Controls}\_{\text{it}} + \varepsilon\_{\text{it}} \end{array} \tag{4}$$

As discussed in the Data and Variable Measurement section, the sample employed to estimate Eqs. (1) and (3) comprise 10,045 non-cross listed EM firm-year observations and 610 crosslisted ADR firm-year observations. However, in Eqs. (2) and (4), the sample is reduced to 5709 non cross-listed EM firms and 324 ADRs, as a result of missing observations for the four corporate governance indicators.

## **4. Results**

**ROA** 1.000

ROA Governance Score

0.097

1.000

0.000a

Committees

0.034 0.009a

Independent Directors

0.039 0.002a

> CEO Duality

−0.001

0.943

> Ethics Policy

0.047 0.000a −0.161 0.000a

0.000a

0.342

0.018b

−0.119

−0.012

−0.031

0.065 0.000a

0.000a

1.000

−0.094

Leverage Firm Size

−0.019

0.144

> Tobin's Q

Beta

0.303 0.000a −0.053

0.000a

aSignificance at the 0.01 level.

bSignificance at the 0.05 level.

cSignificance at the 0.10 level.

**Table 2.**

Pearson correlation for the full sample.

0.951

0.029b

0.198 This table provides Pearson Correlation coefficients and corresponding *p*-values for the cross-listed ADRs and non-cross listed EM firms in the sample.

−0.001

0.028

−0.017

−0.021 0.103c

0.046b

0.526

0.000a

0.053b

1.000

−0.026

−0.008

0.122

−0.025

0.054b

0.000a

0.000a

0.000a

0.599

0.001a

0.000a

1.000

−0.025

−0.117

−0.161

−0.056

−0.007

0.042

−0.138

0.000a

0.000a

0.000a

0.000a

0.000a

0.000a

1.000

0.402

0.210

−0.054

−0.143

0.215

−0.133

0.000a

0.000a

0.000a

0.110

1.000

0.418

0.146

0.172

0.000a

0.103c

0.000a

1.000

−0.021

−0.062

0.021

0.107

0.000a

0.000a

1.000

0.226

0.263

0.000a

1.000

158 Financial Management from an Emerging Market Perspective

0.327

**Governance score**

**Committees**

**Board independence**

**CEO duality**

**Ethics policy**

**Leverage**

**Firm size**

**Tobin's Q**

**Beta**

#### **4.1. The direct effect of governance factors on financial performance**

To test hypotheses 1 through 5, we examine the influence of governance on firm performance using Eqs. (1) and (2). The results of estimating of Models (1) and (3) are presented in **Table 3** and reveal the impact of the governance score on the financial performance (i.e., ROA) of EM firms and ADRs, respectively. The estimate of Model (1) provides evidence supporting hypothesis 1 in that an increase in the governance score *(p* < 0.05) is likely to increase the financial performance of an EM firm, while the estimate of Model (3) shows a positive but insignificant coefficient on the governance score variable for the ADRs. The estimates of Models (2) and (4) provide the results of the four corporate governance indicators on financial performance for EM and ADR firms, respectively. Model (2) reveals no evidence of a relationship between the individual corporate government variables and performance for EM firms. Model (4), however, shows that CEO Duality negatively affects the financial performance of cross-listed ADR firms (*p* < 0.10).


This table shows the outcomes of cross-sectional time-series two-step GLS regressions of Return on Assets (ROA) on the governance score, four individual governance standards and control variables. It reports the standardized coefficients on the independent variables as well as standard errors in parentheses.

\* Significance at the 0.1 level.

\*\*Significance at the 0.05 level.

\*\*\*Significance at the 0.01 level.

**Table 3.** Two-step GLS regressions of governance factors on return on assets.

Overall, we find that stronger governance (as measured by the overall Governance Score) is associated with better financial performance among non-cross-listed EM firms, but we find no evidence of an impact on the performance of cross-listed ADRs. Also, there is no evidence of a positive effect by any of the individual governance rules on the performance of EM firms. CEO Duality, however, appears to be associated with weaker performance among the ADR firms over time.

#### **4.2. The moderating effect of risk on governance-performance relationship**

We examine the results of hypotheses 6 and 7 using the interaction terms as specified in Eqs. (3) and (4). **Table 4** provides the estimates of the regressions that include interactions between Governance Score and the four governance indicators with a moderating variable, Beta. Model (1) indicates that, when the moderating variable is introduced, the direct effect of the Governance Score on firm performance remains significant and positive (*p* < 0.05), while the interaction term has a significant but negative association with performance in EM firms (*p* < 0.05), providing evidence for hypothesis 6. Similarly, in Model (2) we observe that, although the direct effects of all four governance indicators are not significant, the moderating effect of beta on Ethics Policy, Number of Committees and performance are significant and negative, providing evidence in support of hypotheses 7c and 7d. The results further reveal that the independent variables Ethics Policy and Number of Committees are significant, conditional on the inclusion of the interaction variable, Beta into the model. Nevertheless, we find no evidence in support of the moderating effect of risk on the governance-performance relationship in Models (3) and (4) among the ADR firms.


Overall, we find that stronger governance (as measured by the overall Governance Score) is associated with better financial performance among non-cross-listed EM firms, but we find no evidence of an impact on the performance of cross-listed ADRs. Also, there is no evidence of a positive effect by any of the individual governance rules on the performance of EM firms. CEO Duality, however, appears to be associated with weaker performance among the ADR

This table shows the outcomes of cross-sectional time-series two-step GLS regressions of Return on Assets (ROA) on the governance score, four individual governance standards and control variables. It reports the standardized coefficients

**EM firms ADRs**

(0.005) (0.014)

Committees 0.002 0.000

Independent directors −0.004 −0.006

CEO duality 0.013 −0.122\*

Ethics policy −0.011 −0.078

Leverage −0.088\*\*\* −0.086\*\*\* −0.094\*\*\* −0.146\*\*\*

Firm size 0.060\*\*\* 0.081\*\*\* −0.042 −0.023

Tobin's Q 0.112\*\*\* 0.131\*\*\* 0.291\*\*\* 0.357\*\*\*

Industry fixed effects Yes Yes Yes Yes Country fixed effects Yes Yes Yes Yes Constant 0.210\*\*\* 0.118\*\* 0.565\*\*\* 0.589\*\*\*

Observations 14,045 5709 610 324 R2 0.111 0.088 0.559 0.721 Chi2 1018.5 559.5 225.6 259.5

on the independent variables as well as standard errors in parentheses.

**Table 3.** Two-step GLS regressions of governance factors on return on assets.

Governance score 0.010\*\* 0.009

160 Financial Management from an Emerging Market Perspective

**(1) (2) (3) (4)**

(0.006) (0.028)

(0.012) (0.035)

(0.017) (0.065)

(0.016) (0.077)

(0.005) (0.008) (0.029) (0.034)

(0.009) (0.012) (0.041) (0.042)

(0.005) (0.007) (0.028) (0.035)

(0.037) (0.056) (0.145) (0.177)

firms over time.

Significance at the 0.1 level. \*\*Significance at the 0.05 level. \*\*\*Significance at the 0.01 level.

\*


This table shows the outcomes of cross-sectional time-series two-step GLS regression model for the moderating effect of risk (measured by beta) on the relationship between return on assets (ROA) and the governance score, as well as four individual governance standards. It reports the standardized coefficients on the independent variables as well as standard errors in parentheses.

\* Significance at the 0.1 level.

\*\*Significance at the 0.05 level.

\*\*\*Significance at the 0.01 level.

**Table 4.** The moderating effect of risk on governance-performance relationship.

Overall, the results indicate that the existence of a formal ethics policy in the emerging markets may signal a sincere commitment to a strengthened governance structure, which improves the profitability of EM firms. However, in the presence of relatively high market risk (i.e., Beta), the performance enhancing attributes of an ethics policy are mitigated. Given that Governance Score is a continuous variable while the Number of Committees is a categorical variable, we provide graphical analyses to depict the interaction between risk (Beta) and the Governance Score as well as the interaction between risk and the Number of Committees in **Figures 1** and **2**, respectively. **Figure 1** shows that lower risk EM firms enjoy greater benefits in the form of performance from more rigorous governance standards than the higher risk firms. Similarly, **Figure 2** reveals that lower risk EM firms see greater performance gains from adding committees.

#### **4.3. Supplementary analyses**

Given the frequent use of market capitalization of equity as an alternative performance measure in the accounting and finance literature, we assess the robustness of the governance-performance

Corporate Governance and Financial Performance in the Emerging Markets: Do ADRs Perform… http://dx.doi.org/10.5772/intechopen.72297 163

**Figure 1.** The moderating effect of risk on Governance Score-Performance relationship in the EMs.

**Figure 2.** The moderating effect of risk on Committees-Performance relationship in the EMs.

Overall, the results indicate that the existence of a formal ethics policy in the emerging markets may signal a sincere commitment to a strengthened governance structure, which improves the profitability of EM firms. However, in the presence of relatively high market risk (i.e., Beta), the performance enhancing attributes of an ethics policy are mitigated. Given that Governance Score is a continuous variable while the Number of Committees is a categorical variable, we provide graphical analyses to depict the interaction between risk (Beta) and the Governance Score as well as the interaction between risk and the Number of Committees in **Figures 1** and **2**, respectively. **Figure 1** shows that lower risk EM firms enjoy greater benefits in the form of performance from more rigorous governance standards than the higher risk firms. Similarly, **Figure 2** reveals that lower risk EM firms see greater performance gains from adding committees.

This table shows the outcomes of cross-sectional time-series two-step GLS regression model for the moderating effect of risk (measured by beta) on the relationship between return on assets (ROA) and the governance score, as well as four individual governance standards. It reports the standardized coefficients on the independent variables as well as

**EM firms ADRs**

Beta\*committees −0.874\*\*\* −0.006

162 Financial Management from an Emerging Market Perspective

Leverage −0.090\*\*\* −0.091\*\*\* −0.120\*\*\* −0.184\*\*\*

Firm Size 0.061\*\*\* 0.081\*\*\* 0.013 0.030

Tobin's Q 0.112\*\*\* 0.131\*\*\* 0.290\*\*\* 0.339\*\*\*

Industry fixed effects Yes Yes Yes Yes Country fixed effects Yes Yes Yes Yes Constant 0.204\*\*\* 0.121\*\* 0.627\*\*\* 0.701\*\*\*

Observations 14,045 5709 610 324 R2 0.116 0.096 0.554 0.745 Chi2 1044.3 595.5 245.0 326.6

**Table 4.** The moderating effect of risk on governance-performance relationship.

**(1) (2) (3) (4)**

(0.006) (0.008) (0.029) (0.032)

(0.009) (0.012) (0.039) (0.038)

(0.005) (0.007) (0.027) (0.032)

(0.037) (0.055) (0.134) (0.161)

(0.241) (2.570)

Given the frequent use of market capitalization of equity as an alternative performance measure in the accounting and finance literature, we assess the robustness of the governance-performance

**4.3. Supplementary analyses**

standard errors in parentheses.

Significance at the 0.1 level. \*\*Significance at the 0.05 level. \*\*\*Significance at the 0.01 level.

\*

models with regard to our measure of financial performance by substituting the market-tobook value of equity ratio (M/B) for ROA [61–63]. Specifically, M/B is calculated as the ratio of the total market value of a firm's equity to its book value of equity; it is interpreted as the multiple by which the market values a firm's net assets. **Table 5** provides the results of estimating the models using M/B as the relevant performance metric. Results reveal that the positive effect of the Governance Score on performance remains significant (*p* < 0.10) for EM firms [Model (1)] and insignificant for the cross-listed ADRs [Model (3)]. An examination of the impact of the individual governance variables on M/B indicates that the parameter estimate on CEO Duality becomes negative and significant (*p* < 0.05) for EM firms [Model (2)], suggesting that CEO Duality weakens corporate governance mechanisms and reduces the long-term potential wealth creation of the firm. In line with the extant literature, the results also show that an increase in Independent Directors (*p* < 0.05) increases firm value [36]. In contrast, but consistent with Aebi et al. [8], we find a negative relationship between firm performance and the specific corporate governance variables, Number of Committees (*p* < 0.10) and Ethics Policy (*p* < 0.01) for cross-listed ADRs [Model (4)].

The equity of cross-listed firms are offered on the market as Level 1, 2 or 3 ADRs. Level 1 ADRs are traded over-the-counter and require the least amount of compliance and regulatory oversight by the SEC. Level 2 and 3 ADRs are required to adhere to all of the financial reporting standards mandated by SOX (2002). If compliant, level 2 and 3 ADRs trade on the major US stock exchanges (i.e., NASDAQ and NYSE). Given the regulatory differences between level 1


This table shows the outcomes of cross-sectional time-series two-step GLS regressions of market value (measured by market-to-book ratio) on the governance score, four individual governance standards and control variables. It reports the standardized coefficients on the independent variables as well as standard errors in parentheses.

\* Significance at the 0.1 level.

\*\*Significance at the 0.05 level.

\*\*\*Significance at the 0.01 level.

**Table 5.** Two-step GLS regressions of governance factors on market value.

and level 2 and 3 ADRs, we test the sensitivity of our results on the ADR sample. Eqs. (1) and (2) are re-estimated using a restricted sample that includes only level 2 and 3 ADRs. **Table 6** presents the results of Models (1) and (2) using ROA as the performance measure (dependent variable) and Models (3) and (4) using M/B. A comparison of the results of Models (3) and (4) using the full sample (in **Table 4**) with Models (1) and (2) using the restricted sample (in **Table 6**), reveals that both the sign and the statistical significance of the parameter estimates remain similar with regard to the governance-performance relationship when level 1 ADRs are excluded from the sample. Therefore, using ROA as the performance measure, level 2 and 3 ADRs exhibit no evidence of improved performance given an enhanced governance structure.

corporate governance variables, Number of Committees (*p* < 0.10) and Ethics Policy (*p* < 0.01)

The equity of cross-listed firms are offered on the market as Level 1, 2 or 3 ADRs. Level 1 ADRs are traded over-the-counter and require the least amount of compliance and regulatory oversight by the SEC. Level 2 and 3 ADRs are required to adhere to all of the financial reporting standards mandated by SOX (2002). If compliant, level 2 and 3 ADRs trade on the major US stock exchanges (i.e., NASDAQ and NYSE). Given the regulatory differences between level 1

**EM firms ADRs**

(0.006) (0.017)

Governance score 0.011\* 0.003

Committees −0.005 −0.060\*

Independent directors −0.011 0.082\*\*

CEO duality −0.044\*\* −0.018

Ethics policy −0.010 −0.259\*\*\*

Leverage 0.021\*\*\* −0.005 0.003 0.047

Firm size 0.008 0.045\*\*\* 0.031 0.061

Tobin's Q 0.745\*\*\* 0.826\*\*\* 0.913\*\*\* 0.999\*\*\*

Industry fixed effects Yes Yes Yes Yes Country fixed effects Yes Yes Yes Yes Constant 0.004 −0.062 −0.449\*\* −0.686\*\*\*

Observations 14,045 5709 610 324 R2 0.793 0.802 0.869 0.932 Chi2 15934.0 10435.4 1004.5 1186.0

the standardized coefficients on the independent variables as well as standard errors in parentheses.

**Table 5.** Two-step GLS regressions of governance factors on market value.

\*

Significance at the 0.1 level. \*\*Significance at the 0.05 level. \*\*\*Significance at the 0.01 level.

This table shows the outcomes of cross-sectional time-series two-step GLS regressions of market value (measured by market-to-book ratio) on the governance score, four individual governance standards and control variables. It reports

**(1) (2) (3) (4)**

(0.007) (0.033)

(0.015) (0.040)

(0.021) (0.079)

(0.020) (0.089)

(0.007) (0.010) (0.040) (0.043)

(0.009) (0.012) (0.053) (0.047)

(0.007) (0.009) (0.036) (0.040)

(0.035) (0.052) (0.184) (0.194)

for cross-listed ADRs [Model (4)].

164 Financial Management from an Emerging Market Perspective

When M/B is employed as a performance measure, a comparison of Model (4) in **Table 6**, using the restricted sample, with Model (4) in **Table 5**, using the full sample, reveals that the statistical significance of some of the parameter estimates decrease or disappear. When level 1 ADRs are excluded from the sample, the variables Independent Directors and Ethics Policy become statistically insignificant, while Number of Committees continues to have a negative and statistically significant impact on market value for level 2 and 3 ADRs. In addition, similar to the findings of Model (3) in **Table 5** that uses the full model, governance score still remains insignificant in Model (3) (**Table 6**) when level 1 ADRs are excluded from the sample.



This table shows the outcomes of cross-sectional time-series two-step GLS regressions of return on assets (ROA) and market value (measured by market-to-book ratio) on the governance score, four individual governance standards and control variables for the level 2 and 3 ADRs. It reports the standardized coefficients on the independent variables as well as standard errors in parentheses.

\* Significance at the 0.1 level.

\*\*Significance at the 0.05 level.

\*\*\*Significance at the 0.01 level.

**Table 6.** Two-step GLS regressions of governance factors on ROA and market value for level 2&3 ADRs.
