ضرایب ریسک شایع در بازده بازارهای سهام آسیا در حال ظهور

This paper examines the application of the Fama and French's (1993) three-factor model in three Asian emerging markets (Hong Kong, Singapore and Taiwan). The empirical evidence is consistent with the US findings that the model can explain most of the variations in average returns. However, we find that the main contributing factor is the contemporaneous market excess returns. The impact of the size effect and book-to-market (BE/ME) factor is limited and in some cases insignificant. When the three-factor model is modified by using lagged market excess returns instead in order to check for the predictability of the market factor, the explanatory power of the model drops substantially but both the risk factors for size and BE/ME are now able to contribute significantly in explaining the cross-sectional variations of stock returns. Their explanatory powers are strongest for small-size with high BE/ME portfolios. The robustness of our results is also checked for the separation of up and down markets periods and January effect.

Several studies documented that average return is related to firm size, book-to-market equity ratio (BE/ME), earnings to price ratio (E/P), cash flow to price ratio (C/P) and past sales growth. Banz, 1981, Basu, 1983, Cook and Rozeff, 1984, Davis, 1994, De Bondt and Thaler, 1987, Keim, 1983, Lakonishok and Shapiro, 1984, Reinganum, 1982 and Rosenberg et al., 1985, and Lakonishok et al. (1994)) provided evidence on these firm characteristics in explaining the average stock returns. Since these patterns in the behavior of stock prices cannot be explained by the Capital Asset Pricing Model (CAPM) of Sharpe, 1964 and Lintner, 1965, they are typically called anomalies. Fama and French (1992) found that size and BE/ME play dominant roles in explaining the cross-sectional variations in US stock returns. Fama and French (1993) showed that size and BE/ME proxy for the security's loadings in priced factors within a three-factor model. The three factors are the returns on the market portfolio and those on two zero net-investment portfolios: long in portfolio of small-size stocks and short in portfolio of big-size stocks (SMB) and long in portfolio of high BE/ME stocks and short in portfolio of low BE/ME stocks (HML). They found that the three factors provide a good job in explaining the cross-section of average stock returns. Fama and French (1996) further showed that the three-factor model captures returns regardless of the construction methods of portfolios, i.e. based on E/P, C/P, and past sales growth. Daniel and Titman (1997) examined the irrational pricing against the three-factor model of Fama and French, 1993 and Fama and French, 1996. They argued that expected returns are not related to an asset's covariance with any economic risk factor but rather with firm specific characteristics. They rejected the three-factor model, but not the characteristic model. However, Davis et al. (2000) documented that the three-factor model explains the value premium, as measured by HML, better than the characteristic model of Daniel and Titman (1997). They argued that the results of Daniel and Titman (1997) are due to their short sample period. Daniel et al. (2001) replicated the tests of Daniel and Titman (1997) in the Japanese stock market and provided evidence rejecting the three-factor model but not rejecting the characteristic model. Previous empirical work has discovered that US stock returns are largely explained by size and BE/ME effects. In Asian emerging markets, Chui and Wei, 1998 and Ho et al., 2000, and Lam (2002) showed that significant size and BE/ME effects are observed in Hong Kong. In fact, Ho et al. (2000a) also suggested that the CAPM may indeed be misspecified as beta plays no role after examining the equilibrium risk-return relationships in the Hong Kong stock market. Wong and Lye, 1990 and Lau et al., 2002 found that Singaporean stock returns are related to firm size. Chui and Wei (1998) also found no significant firm size and BE/ME effects in Taiwan. However, it should be noted that the above articles did not employ the exact Fama and French (1993) three-factor model in their analysis in that no zero net-investment portfolios are formed for size and BE/ME factors. They simply employed the market capitalization and book-to-market ratio directly in their regression models. To the best of our knowledge, except those studies by Drew and Veeraraghavan, 2001 and Drew and Veeraraghavan, 2003, there is probably no study to test the robustness of the same model in the Asian emerging stock markets. Drew and Veeraraghavan (2003) investigated the robustness of the Fama and French (1993) three-factor model in Hong Kong, Korea, Malaysia and Philippines. They documented that size and value effects exist for all four markets under investigation and concluded that the multi-factor model approach provides a parsimonious description of the cross-section of returns for these Asian markets over the 1990s. This paper helps provide more empirical evidence of the model in three Asian markets. This paper makes no attempt to provide any argument whether the three-factor model of Fama and French, 1993 and Fama and French, 1996 or characteristic model of Daniel and Titman (1997) is superior but has the following two purposes. The first one is to examine the fitness of the three-factor model in three Asian emerging equity markets (Hong Kong, Singapore and Taiwan). Besides adding evidence on two new markets (Singapore and Taiwan) as compared to the work of Drew and Veeraraghavan (2003, hereafter DV), our paper is also different from theirs in several ways even for the common market under study (i.e. Hong Kong). First, we use a longer sampling period (7/1986–12/1998) than DV (12/1993–12/1999). Second, we employ a different source of data from theirs (Pacific-Basin Capital Markets Databases vs. Datastream). These two points are important as noted by Campbell et al. (1997) because using different sampling periods, different data sources and different markets can help in checking the true out-of-sample performance of the mutli-factor model. Third, besides employing time-series regressions as DV did, we also perform time-pooled cross-sectional regression analysis. Fourth, we form 9 (3 sizes time 3 BE/ME values) portfolios instead of 6 (2 sizes time 3 BE/ME values) as in DV. Fifth, in running OLS regression, we adjust for the effects of heteroskedasticity and first-order autocorrelation which was not done in DV (DV only checked for the existence of autocorrelation and concluded that autocorrelation does not exist in their sample) and hence, our results are considered to be more reliable in terms of these potential errors. The second objective of this paper is to investigate the three-factor model when the overall contemporaneous market factor is replaced by the lagged market factor. This extension of the model has not been examined in previous studies, but may provide new insight for us to better understand the role plays by the market factor. We all know that portfolios returns are highly correlated with the contemporaneous market returns and the relevance of other factors may be missed under statistical tests. Hence, the proposed extension helps check whether the market factor does have a dominant role in these Asian markets. Furthermore, by using the lagged market factor, we can see how large its predictive power is, if any, on portfolios returns. This investigation is more relevant to emerging markets when compared with those developed markets like US and UK, as previous studies have indicated that significant serial autocorrelation may exist in emerging markets returns. We further enhance our analysis by performing robustness check with respect to two effects: up and down markets separation and the January effect. The rest of the paper is organized as follows: Section 2 describes the data and methodologies. A sub-section on institutional features of the three markets is also included. Section 3 reports the empirical results while Section 4 concludes the paper.

This paper has examined the relations between a market factor together with two proxies for the risk factors related to size and book-to-market equity ratio and stock portfolios' returns using the three-factor model of Fama and French, 1993 and Fama and French, 1996 over the period July 1986 to December 1988 in three Asian emerging markets: Hong Kong, Singapore and Taiwan. The empirical evidence is consistent with the US findings reported by Fama and French, 1993 and Fama and French, 1996 and with those of four Asian markets studied by DV (2003) that the model largely explains the variations in average returns when using the contemporaneous market factor. However, the impact of the size effect and BE/ME factor is very limited and insignificant in most cases. When the three-factor model is modified by using lagged market factor instead, the explanatory ability of the model drops substantially but both SMB and HML are now able to contribute significantly in explaining the time-pooled cross-sectional variations in stock returns. Our results provide some insights on the three-factor model. When the contemporaneous market factor is employed, it is the dominate factor over the other two factors: SMB and HML. Only when lagged market factor is used that SMB and HML can reflect their importance in explaining the stock returns variations. Furthermore, their explanatory powers are strongest for small-size with high BE/ME portfolios. This paper also checks the robustness of the model on two effects: separation of up and down markets periods and the January effect. Our results find that the conditional model based on up and down markets does not help in improving the explanatory power of the three-factor model when the contemporaneous market factor is employed. However, when the lagged market factor is used in the model, the conditional extension does help in increasing the explanatory power, particularly in the Singaporean stock market. We find no evidence of January effect in the Hong Kong stock market but our results indicate a January effect exists in the BE/ME premium in Singapore and a reverse January effect exists in the market risk premium when the lagged market factor is used in the model. The results found in the paper are interesting and should provide new insights to our understanding of the Fama and French three-factor model. The results also have implications to portfolio managers as investing in small firm size and/or high book-to-market ratio stocks should be rewarded by generating higher average returns. However, they should also be aware that though the traditional CAPM could be mis-specified as size and BE/ME are important additional risk factors, the market factor is still the dominating factor in these Asian emerging markets.