We regard S&P 500 as the independent variable, TAIEX and OTC as the dependent
variables. We use E-VIEWS to consider the quantile regression between the returns of S&P
500 and the returns of TAIEX and OTC respectively. Every 0.05 quantile from 0.05 to 0.95
(left to right) for each sample was chosen. Moreover, 0.01 and 0.99 quantiles are chosen,
because we want to know the highest and the lowest stock prices. The interception and slope
estimates are under 95% confidence interval and there are 21 results of quantile regression in
each sample period.
Table 2 to Table 11 document the empirical results of the influences from the returns of
US stock market to the returns of close-to-open stock market of Taiwan over the sample
period from 1995 to 2009.
Table 2 presents the results from 1995/11/7 to 1997/6/30, and most of the coefficients of
slope are significant positive at 5% or 10% significant level, except for θ= 0.05, 0.55, 0.6,
0.75, 0.8, 0.9. Moreover, it becomes non significant negative correlation where θ= 0.01.
After 1997/7/2 (Table 3 to Table 5), there are significant positive correlations at 5%
significance level. And Table 6 shows that there is significant positive correlation between the
returns of US stock market and the close-to-open returns of TAIEX at 5% significance level
from 1995 to 2009.
< Table 2 is inserted about here >
< Table 3 is inserted about here >
< Table 4 is inserted about here >
< Table 5 is inserted about here >
< Table 6 is inserted about here >
Table 7 to Table 11 present the correlations between US stock market and the
close-to-open returns of OTC from1995/11/7 to 2009/3/26. In Table 7, there is non significant
positive correlation except for θ= 0.05, 0.9, 0.95 at 5% significance level and θ= 0.35 at
10% significance level. After 1997/7/2 (Table 8 to Table 10), there are significant positive
correlations at 5% significance level. Table 11 shows that there is significant positive
correlation between the returns of US stock market and the close-to-open returns of OTC at
5% significance level from 1995 to 2009.
< Table 7 is inserted about here >
< Table 8 is inserted about here >
< Table 9 is inserted about here >
< Table 10 is inserted about here >
< Table 11 is inserted about here >
Table 12 to Table 21 document the empirical results of the influences from the returns of
US stock market to the returns of open-to-close stock market of Taiwan over the sample
period from 1995 to 2009. Table 12 shows all of the coefficients of slope are significant
negative. Moreover, we find that there are significant negative at θ= 0.05 at 5% significance
level, and at θ= 0.15, 0.85, 0.9, 0.95, 0.99 at 10% significance level. We know that there are
negative correlations between the returns of US stock market and the returns of open-to-close
stock market of TAIEX from 1995/11/7 to 1997/6/30, though significant negative correlations
only occurred at the highest or lowest price.
< Table 12 is inserted about here >
Table 13 shows all of the coefficients of slope are significantly negative. Moreover, we
find that half of the quantiles are significant negative correlations. When θ= 0.01, 0.05, 0.1,
0.25, 0.8, 0.85, there are significant negative correlations at 5% significance level, and there
are significant negative correlations at θ= 0.15, 0.35, 0.75 at 10% significance level.
< Table 13 is inserted about here >
After 2000/7/3, almost all the coefficients of slope are significant negative, except for θ= 0.99 from 2000/07/03-2007/02/27 and θ= 0.05, 0.1, 0.15 from 2007/3/1 to 2009/3/26.
There seems to be a significant negative trend when time goes by. Table 16 shows that there is
significant negative correlation between the returns of US stock market and the open-to-close
returns of TAIEX at 5% significance level from 1995 to 2009.
< Table 14 is inserted about here >
< Table 15 is inserted about here >
< Table 16 is inserted about here >
Table 17 shows that all the coefficients of slope are negative, and we find that there are
significant negative at θ= 0.01, 0.05 at 5% significance level, and at Q= 0.45, 0.9, 0.95 at
10% significance level. In Table 18, Table 19 and Table 20, almost all the coefficients of slope
are significant negative at 5% significance level, except for θ= 0.2, there is non significant at
10% level of significance from 1997/7/2 to 2000/6/30 (Table 18), and there is non significant
at θ= 0.1 at 10% level of significance from 2007/03/01-2009/03/26 (Table 20). Table 21
shows that there is significant negative correlation between the returns of US stock market
and the open-to-close returns of OTC at 5% significance level from 1995 to 2009.
< Table 17 is inserted about here >
< Table 18 is inserted about here >
< Table 19 is inserted about here >
< Table 20 is inserted about here >
< Table 21 is inserted about here >
Table 22 to Table 31 are documented the empirical results of the influences from the
returns of US stock market to the returns of close-to-close stock market of Taiwan over the
sample period from 1995 to 2009. Table 22 shows that most of the coefficients of slope are
non significant positive correlations, however, the coefficients of slope are non significant
negative correlations when θ= 0.1, 0.01, 0.8, 0.85, 0.9, 0.95. There is significant negative
correlation at θ= 0.01 at 5% significance level, but there is significant positive correlation at
θ= 0.99 at 10% significance level. After 1997/7/2 (Table 23, Table 24 and Table 25), there are
almost significant positive correlations between the returns of US stock market and the
returns of TAIEX at 5% or 10 % significance level, except for θ= 0.95, 0.99 from 1997/7/2
to 2000/6/30 (Table 23) is non significant at 10% level of significance. Table 26 shows that
there is significant positive correlation between the returns of US stock market and the
close-to-close returns of TAIEX at 5% significance level from 1995 to 2009.
< Table 22 is inserted about here >
< Table 23 is inserted about here >
< Table 24 is inserted about here >
< Table 25 is inserted about here >
< Table 26 is inserted about here >
In Table 27, most of the coefficients of slope are non significant positive correlations
between the returns of S&P 500 and the close-to-close returns of OTC from 1995/11/06 to
1997/06/30, except when θ= 0.01, 0.05, 0.35, 0.6, 0.8, 0.9, 0.95, there are non significant
negative effects. Furthermore, there is significant negative correlation at θ= 0.01 at 5%
significance level. After 1997/7/2, there is significant positive correlation at 5% significance
level, except for θ = 0.95 and 0.99, there are non significant positive effects at 10%
significance level. Table 31 shows that there is significant positive correlation between the
returns of US stock market and the close-to-close returns of OTC at 5% significance level
from 1995 to 2009.
< Table 27 is inserted about here >
< Table 28 is inserted about here >
< Table 29 is inserted about here >
< Table 30 is inserted about here >
< Table 31 is inserted about here >
We analyze how US stock mark influences Taiwan stock market at different quantiles by
using the quantile regression. And it can give us the suggestion what we should react as the
price goes up or down. This paper not only confirms the prior studies that Taiwan stock
market is influenced by the stock price of the US, but also discovers the overreaction effect.
We summarize the results in Table 32, 33, 34, 35, 36, 37, 38, 39, 40 and our findings are as
follows:
1. For the close-to-open returns, there is significant positive effect from US to Taiwan stock
markets from 1995/11/7 to 2009/3/26. However, the price change spillover effect exists
when the stock price goes up or down greatly from 1995/11/7 to 1997/6/30 (Table 32 and
33). There seems to be a significant positive trend not only in TAIEX but also in OTC as
time goes by. In other words, we can say that there is the spillover effect from US stock
market to the close-to-open return of Taiwan stock market. Table 35 reports the slope
estimate of both the quantile regression and ordinary least squares method from 1995 to
2009. We can know that both of the slope estimates are positive. In other words, there are
positive correlations not only between S&P 500 and close-to-open returns of TAIEX, but
also between S&P 500 and close-to-open returns of OTC. Moreover, we can find that the
slope estimates are different in most of the quantiles. OLS might underestimate the slope
estimate when price goes down but overestimate it when price goes up.
2. For the open-to-close returns, there is significant negative correlation between US and
Taiwan stock markets from 1995/11/7 to 2009/3/26. In other words, there are
overreaction effects both in TAIEX and OTC. For TAIEX, the negative effect is stronger
when the stock price goes up greatly from 1995/11/7 to 1997/6/30, however, the negative
effect is stronger when the stock price goes down greatly from 1997/7/2 to 2000/6/30.
From 2000/7/3 to 2009/3/26, the negative effect is significant. For OTC, the negative
impact is stronger when the stock price goes up or down greatly from 1995/11/7 to
2000/6/30. From 2000/7/3 to 2009/3/26, the negative effect is significant. Table 37
reports the slope estimate of both the quantile regression and ordinary least squares
method from 1995 to 2009. We can know that both of the slope estimates are negative. In
other words, there are negative correlations not only between S&P 500 and open-to-close
returns of TAIEX, but also between S&P 500 and open-to-close returns of OTC.
Moreover, we can find that the slope estimates are different in most of the quantiles. OLS
might overestimate the slope estimates when price goes up or down greatly, but otherwise
OLS underestimates the slope estimates.
3. For the close-to-close returns, there is significant positive correlation between US and
Taiwan stock markets from 1995/11/7 to 2009/3/26, except for the returns from
1995/11/7 to 1997/6/30. There is a negative effect from U.S. to Taiwan when the stock
price goes down greatly from 1995/11/7 to 1997/6/30. Furthermore, there seems to be a
significant positive trend not only in TAIEX but also in OTC as time goes by. And we can
conclude there is the spillover effect from US stock market to the close-to-close return of
Taiwan stock market. Table 40 reports the slope estimate of both the quantile regression
and ordinary least squares method from 1995 to 2009. We can know that both of the slope
estimates are positive. In other words, there are positive correlations not only between
S&P 500 and close-to-close returns of TAIEX, but also between S&P 500 and
close-to-close returns of OTC. Moreover, we can find that the slope estimates are
different in most of the quantiles. OLS might underestimate the slope estimate when price
goes down greatly, but otherwise overestimates the slope estimates.
< Table 32 is inserted about here >
< Table 33 is inserted about here >
< Table 34 is inserted about here >
< Table 35 is inserted about here >
< Table 36 is inserted about here >
< Table 37 is inserted about here >
< Table 38 is inserted about here >
< Table 39 is inserted about here >
< Table 40 is inserted about here >