Empirical Results: Do Good Corporate Governance Stocks Provide

IV. Empirical Results and Analysis

4.2 Empirical Results: Do Good Corporate Governance Stocks Provide

and the ADR benchmark assets might imply that the three G-index stock portfolios can improve our ADR benchmark assets.

4.2 Empirical Results: Do Good Corporate Governance Stocks Provide Better Diversification Benefits?

In this section we use the mean-variance spanning test to examine whether adding G-index portfolios to a set of benchmark assets based on the ADRs enlarges the investment opportunity set. First, we use the 12ADRs as the benchmark asset, and Table 5 presents empirical results from the mean-variance spanning and the intersection tests.⁹

In Table 5, we find that no matter if we are examining the equally-weighted case or the value-weighted portfolios, the results of the Likelihood ratio test, the Lagrange multiplier test, and the Wald tests are all significant, and we can reject the null hypothesis. Our empirical results indicate that no matter if one adds the entire G-index portfolio or adds the Democracy and the Dictatorship portfolio to our benchmark assets, the test statistics are significant as well. In other words, the previous three G-index portfolios can all help to expand the mean-variance frontier. We also use the Kan and Zhou (2001) step-down Wald test as a robustness check. If W1 is more significant than W2, which means that the expansion of the mean-variance frontier mostly comes from the change in the tangency portfolio. If W1 is less significant than W2, this implies that the expansion comes mostly from the change in the global minimum-variance (GMV) portfolio. Panel A and Panel B of Table 5 presents that all of the W1 are less significant than W2, which implies that the expansion of these portfolios comes primarily from the change in the global

9 The risk-free rate we used is the average of one-month T-bill rate from 1990 to 2005 collected from the

website of French, which is 0.33% per month. The optimal portfolio weights are reported in Appendix.

minimum-variance (GMV) portfolio.¹⁰

Furthermore, we also follow DeRoon and Nijman (2001) to test the mean-variance intersection. The intersection occurs when the original mean-variance frontier and the new mean-variance frontier have only one point in common. As Table 5 reports, either the equally-weighted or value-weighted results for WI are all insignificant, and thus we cannot reject our null hypothesis. In other words, the mean-variance frontier of augmented assets has only one point in common with the mean-variance frontier of benchmark assets based on the reference point of risk-free rate.

As we mentioned before, the previous mean-variance spanning tests only examine whether the expansion of the mean-variance frontier is significant. Bekaert and Urias (1996) suggest that the economic significance of the shift in the mean-variance frontier can be evaluated by the change in the Sharpe ratio. The Sharpe ratio, also known as the “reward to variability” ratio, measures the slope of the line from the risk-free rate to any portfolio in the mean-standard deviation plane (Sharpe (1994)). The values of the percentage change in the Sharpe ratio are inversely related to the p-values of the first step–down test (W1), and so does the intersection test whose test statistics involve the difference in the squared maximal Sharpe ratios. In other words, if we fail to reject the hypothesis in the first step-down test or the intersection hypothesis, then a small percentage change in the Sharpe ratio will occur.

Table 6 presents the results of change in the Sharpe ratio for the tangency portfolio. In the portfolio which sums up all the G-index stocks, the change in the Sharpe ratio of the

10 Tobin (1958) introduces the well-known separation property and argues that portfolio choice can be separated into two steps: (1) the determination of the optimal tangency portfolio; (2) the construction of the mix between the risk-free asset and the optimal tangency portfolio, dependent on investors’ preferences.

Therefore, investors are more likely concerned with the change in the tangency portfolio than the global minimum-variance portfolio.

value-weighted case is 0.6542. This is more than the 0.1145 change in the Sharpe ratio of the equally-weighted case. In the Democracy (G<=5) portfolio, the change in the Sharpe ratio of the value-weighted case is 0.0327, and the change in Sharpe ratio of the equally-weighted case is 0. This represents that the Sharpe Ratio has no increase. In the Dictatorship portfolio (G>=14), the change in the Sharpe ratio of value-weighted case is 0.0327. This is greater than the 0.0286 change ratio of the equally-weighted case. The previous results indicate that the changes in the Sharpe Ratio of the value-weighted cases are approximately greater than those of the equally-weighted cases.

Our insignificant results of the step-down and intersection tests demonstrate that the expansions of frontiers mostly come from the change in the global minimum-variance portfolio, and that the change in the tangency portfolio which investors are more interested in is not as apparent. These results can also be confirmed in Table 6, the percentage change in the standard deviation of the GMV portfolio is -18.6% for adding all G-index stocks to the equally-weighted case and -12.5% for adding all G-index stocks to the value-weighted case. The standard deviations of the GMV portfolios for the Democracy portfolio (G<=5) and the Dictatorship portfolio (G>=14) have negative percentage changes as well. The extension is mostly due to the GMV portfolio but not the tangency portfolio which denotes that the newly added portfolios can only reduce the risk but are unable to improve the investment opportunity set.

Panel A of Figure 2 shows the expansion of the mean-variance frontier when adding each of G-index stocks portfolio to the equally-weighted benchmark assets for 12ADRs.

Panel B of Figure 2 presents the improvement of the mean-variance frontier that takes value-weighted 12ADRs as benchmark assets. Significantly, the extensions of the mean-variance frontiers at the equally-weighted benchmark portfolios are more apparent

than in the value-weighted benchmark portfolios. The significant results for the cases of equally-weighted portfolios, compared to value-weighted cases, can be verified from the previous test statistics results in Table 5. Obviously, they are more statistical significant for equally-weighted portfolios than for value-weighted ones. No matter in the equally-weighted or the value-weighted cases, we can intuitively find that taking all of the G-index stocks as test assets has the largest impact on diversification benefits. Nevertheless, our results indicate that not only stocks of well-governed firms (G<=5) but also stocks of badly-governed firms (G>=14) can expand the mean-variance frontier. Eventually, stocks of badly-governed firms (G>=14) will provide more diversification benefits than stocks of well-governed firms (G<=5). Additionally, the outcome that well-governed firms and badly-governed firms have similar impact on the original benchmark assets can be documented by the previous correlation reported in Table 4. The correlations between well-governed firms (G<=5) and badly-governed firms (G>=14) is 0.6719, which is much higher than other correlations in the table, implying that the results of well-governed firms will intuitively resemble that of badly-governed firms.

Overall, we find that the previous three G-index stock portfolios can provide significant diversification benefits by reducing risk, especially for the equally-weighted case in this section. Investors who add all G-index stock portfolios or well-governed/

badly-governed firm stocks will only reduce risk but cannot make their investment opportunity sets more progressive.

4.3 Results for 13ADRs as Benchmark Portfolios plus the G-Index Stock Portfolios Table 7 presents the mean-variance spanning and intersection test results for adding

various G-index stock portfolios to benchmark assets that are composed of 13ADRs. No matter in the equally-weighted or the value-weighted portfolios, we reject the null hypothesis by using the Likelihood ratio test, the Lagrange multiplier test, and the Wald test.

All of the statistics are significant, implying that adding G-index stock portfolios can expand the mean-variance frontier. Using the mean-variance intersection test, we can find that the original mean-variance frontier of benchmark assets and the new mean-variance frontier have only one point in common. Besides, the step-down test results in Panel A and Panel B of Table 7 demonstrating that all the statistics of W1 in each portfolio are less significant than W2. That is to say, the expansions of mean-variance frontiers are due to the change in the GMV portfolio rather than in the change of the tangency portfolio. These results can be confirmed in Table 8, and the percentage change in the standard deviation of the GMV portfolio is -13.9% for adding all G-index stocks in the equally-weighted case and -11.7% for adding all G-index stocks in the value-weighted case. In addition, the standard deviations of GMV portfolios for the Democracy portfolio (G<=5) and the Dictatorship portfolio (G>=14) also have negative percentage changes. As above, these results emphasize that adding G-index stock portfolios can only reduce the risk but fail to increase returns. Also worthy of mention is that adding the Democracy portfolio (G<=5) to the 13ADRs benchmark assets will obtain absolutely no change in the Sharpe ratio for the tangency portfolio (0%). This once again documents the low contribution of the tangency portfolio.

Panel A of Figure 3 shows the expansion of the mean-variance frontier from adding the various G-index stock portfolio to the equally-weighted benchmark assets for 13ADRs.

Panel B of Figure 3 presents the improvement of the mean-variance frontier that takes value-weighted 13ADRs as benchmark assets. Similar to above, adding various G-index

stock portfolios can extend the mean-variance frontier, even for the Dictatorship portfolio (G>=14) as well. The frontiers which add stocks of well-governed firms (G<=5) extend less than the frontiers adding stocks of badly-governed firms (G>=14), no matter in the equally-weighted or in the value-weighted cases. Test statistics in Table 7 confirm these results, and the outcomes of badly-governed firms (G>=14) are more significant. In summary, in this section we document that adding the various G-index stock portfolios to the benchmark assets that are composed of 13ADRs can provide significant diversification benefits by reducing risk but with no creation of extra returns.

4.4 Robustness Check

So far, our previous results all indicate that adding stocks of well-governed firms (G<=5) to our ADR benchmark assets cannot significantly provide more diversification benefits than stocks of badly-governed firms (G>=14). Even stocks of badly-governed firms will shift the mean-variance frontier more than stocks of well-governed firms. This is in contrast to our prior expectations. In addition, they identically gain less extra returns for a unit increase in standard deviations attributed to a low percentage change in the Sharpe ratio for a tangency portfolio during the period from 1990 to 2005. In this section, we consider the issue of “time-period-specificity” proposed by Core, Guay, and Rusticus(2006), and divide our full-sample period into two sub-sample periods (1990-1999, 2000-2005). The cumulative raw returns reported in Figure 1 indicate that nearly all of the positive returns to the hedge portfolios documented by GIM for the 1990-1999 time period occur from 1997 to 1999 (Core, Guay, and Rusticus(2006)). After 2000, the hedge portfolio earns negative returns, which is regarded as an influence of the Internet bubble around April 2000. We add the Democracy portfolio and the Dictatorship portfolio to the 12ADRs benchmark

assets in the two sample periods respectively. The results are displayed in Table 9 and Table 10.

We can find evidence from Panel A in Table 9 that the test statistics of the intersection test (WI) and the step-down test (W1) for the Democracy portfolio are much more significant than that for the Dictatorship portfolio during the sample period 1990-1999. This can be confirmed in Panel A of Table 10, the percentage changes in the Sharpe ratio of the Democracy portfolio are 11.488% for the equally-weighted case and 4.4673% for the value-weighted case. This is much higher than the changes for the Dictatorship portfolio (which are 0.1538% for the equally-weighted case and 0.0419% for the value-weighted case, respectively). Therefore we can document that in the period from 1990 to 1999, investors adding stocks of well-governed firms to their benchmark portfolios can not only reduce more risk but can also earn more extra returns than adding stocks of badly-governed firms.

In other words, stocks of well-governed firms will improve the investment opportunity set more apparently than stocks of badly-governed firms, which entirely corresponds to our prior expectations.

Panel B of Table 9 reports the test statistic results of the sub-sample period 2000-2005.

Though the test statistics of the intersection test (WI) and the step-down test (W1) for the Democracy portfolio are still more statistically significant than that for the Dictatorship portfolio, they are not as significant as before (compared with the results in Panel A for the period 1990-1999). The insignificant test statistics of WI and W1 in Panel B reveal that the expansion of frontiers are mostly due to the change in the global minimum-variance portfolio, therefore we acquire low percentage changes in the tangency portfolio. The percentage decreases in the standard deviation of the GMV portfolio in Table 10 are quite obvious but the percentage changes in the Sharpe ratio are quite small both in Democratic and Dictatorship cases. Briefly, during the period from 2000 to 2005, neither the

well-governed firms nor the badly-governed firms could improve the investment opportunity set as efficacious as in the period from 1990-1999. The expansion of mean-variance frontiers when adding G<=5 or G>=14 portfolios to the equally-weighted/value-weighted benchmark assets during the sample period 1990-1999 are reported in Panel A and B of Figure 4. Panel C and D in Figure 4 then display the shift of the mean-variance frontiers taking equally-weighted/value-weighted 12ADRs as benchmark assets from 2000 to 2005.

In summary, following the original GIM sample period ranging from 1990-1999, stocks of well-governed firms can improve the investment opportunity set more than stocks of badly-governed firms can as we expected. However, during 2000-2005, both can reduce the risk of the investment opportunity set only and can acquire low extra returns. Our expected differences between well-governed and badly-governed firms seems to be unclear.

An obvious candidate to consider is the influence of the “Internet bubble” around April 2000. As Figure 1 presents, nearly all of the positive returns to the hedge portfolios occurred from 1997 to 1999 and was negative after 2000 helping to provide evidence explaining our results as above. Overall, the “time-period-specificity” suggested by Core, Guay, and Rusticus (2006) likely play a role in explaining the difference between well-governed and badly-governed firms cause conspicuously different impacts on diversification benefits contributing during 1990-1999.

V. Conclusions

The purpose of this study is to investigate the impact of adding G-index stocks to the benchmark assets composed of ADRs from the perspective of asset allocations for the period from 1990-2005. We apply the mean-variance spanning test and the intersection test to analyze whether the well-governed firm stocks can provide significant diversification benefits.

Our results do not entirely reveal the positive answers. As our previous results revealed, almost all of the test statistics indicate that investors who invest in the various G-index stocks are able to expand their mean-variance frontiers relative to investments in benchmark portfolios composed of 12ADRs or 13ADRs. Besides, according to the insignificant results of WI andW1, investors who add all G-index stock portfolios or the well-governed/

badly-governed firm stocks will provide diversification benefits only by reducing risk but cannot improve their investment opportunity set. However, our evidence indicates that the stocks of badly-governed firms will improve the investment opportunity set more than the stocks of well-governed firms which is absolutely in contrast to our prior expectations. This finding might imply that the corporate governance index may not be a useful measurement to distinguish the diversification benefits when investing in the full sample period from 1990-2005. In other words, corporate governance has no direct relation with the degree of mean- variance frontier’s improvement in this sample period.

For explanation, we consider the problem of “time-period-specificity” proposed by Core, Guay, and Rusticus (2006), dividing our full-sample period into two sub-sample periods (1990-1999, 2000-2005). Our result of sub-sample periods reveal that following the original GIM sample period 1990-1999, the stocks of strong governance firms can significantly improve the investment opportunity set more than the stocks of weak

governance firms. However, the significant difference in diversification benefits between well-governed and badly-governed firms diminishes immediately during 2000-2005. Thus, our consideration of the influence of the “Internet bubble” around April 2000, which is the

“time-period-specificity” suggested by Core, Guay, and Rusticus (2006), helps to provide evidence to confirm our conclusion. Overall, the reason that the well-governed firm stocks are unable to contribute more diversification benefits than the poorly-governed firm stocks in the entire sample period from 1990-2005 show significant differences in diversification benefits during 1990-1999, and could be due to the time-period-specificity nature for structure change.

References

Alaganar, V.T., and Ramaprasad Bhar, 2001, Diversification gains from American depositary receipts and foreign equities: Evidence from Australian stocks, Journal of International Financial Markets, Institutions and Money 11, 97-113.

Ambrose, Brent, and William Megginson, 1992, The role of assets structure, ownership structure and takeover defense in determining acquisition likelihood, Journal of Financial and Quantitative Analysis 27, 575-589.

Barber, Brad M, and John D. Lyon, 1996, Detecting long-run abnormal operating performance: The empirical power and specification of test statistics, Journal of Financial Economics 41, 359-399.

Bebchuk, Lucian A., and Alma Cohen, 2005, The costs of entrenched boards, Journal of Finance Economics 78, 409-433.

Bebchuk, Lucian A., Alma Cohen, and Allen Ferrell, 2005, What matters in corporate governance? Working paper, Harvard Law School.

Bekaert, G. and M. Urias, 1996, Diversification, integration and emerging market closed-end funds, Journal of Finance 51, 835-869.

Brown, Lawrence D. and Marcus L. Caylor, 2004, Corporate governance and firm performance, Working Paper, Georgia State University.

Carhart, Mark M., 1997, On persistence in mutual fund performance, Journal of Finance, 52, 57-82.

Chen, H.-C., Ho, K.-Y., Lu, C., and Wu, C.-H., 2005, Real estate investment trusts: An asset allocation perspective, Journal of Portfolio Management: Real Estate Special Issue, 46-54.

Chen, H.-C., Ho, K.-Y., and Wu, C.-H, 2006, The portfolio diversification effects of initial public offerings (previously titled as: Initial public offerings: An asset allocation perspective), Working paper, National Central University, National Central University, Yuan Ze University, Taiwan.

Core, John E., Robert W. Holthausen, and David F. Larcker, 1999, Corporate governance, chief executive compensation, and firm performance, Journal of Finance Economics 51, 371-406.

Core, John E., Wayne R. Guay, and Tjomme O. Rusticus, 2006, Does weak governance cause weak stock returns? An examination of firm operating performance and analysts’

expectations, Journal of Finance 61, 655-687.

在文檔中公司治理指標運用於投資組合分散策略—以美國股市和ADR市場為例 (頁 28-0)