Paci Fic-Basin Finance Journal: Xiangyi Zhou, Weijin Zhang, Jie Zhang

Pacific-Basin Finance Journal 20 (2012) 247–270
Contents lists available at SciVerse ScienceDirect
Pacific-Basin Finance Journal

journal homepage: www.elsevier.com/locate/pacfin
Volatility spillovers between the Chinese and world

equity markets
Xiangyi Zhou a,⁎, Weijin Zhang a, Jie Zhang b
a
Jinhe Center for Economic Research, Xi'an Jiaotong University, PR China
b
Department of Economics, Texas A&M University, United States
a r t i c l e i n f o a b s t r a c t
Article history: We propose measures of the directional volatility spillovers between

Received 23 February 2011 the Chinese and world equity markets based on Diebold and Yilmaz's
Accepted 2 August 2011
(2011b) forecast-error variance decompositions in a generalized
Available online 16 September 2011
vector autoregressive framework. It was found that the US market
had dominant volatility impacts on other markets during the
Keywords:
China
subprime mortgage crisis. The other markets were also very volatile,
World equity markets and driven by bad news, their massive volatilities were transmitted
Vector autoregression back to the US market. The volatility of the Chinese market has had
Variance decomposition a significantly positive impact on other markets since 2005. The
Spillover index volatility interactions among the markets of China, Hong Kong, and
Financial crisis Taiwan were more prominent than those among the Chinese,
Western, and other Asian markets were. The major correction of the
JEL classification:
Chinese stock market between February and July 2007 significantly
G15
contributed to the volatility surges of other markets. Owing to the
F36
restrictions on foreign investment, the Chinese stock market was
not considerably affected in terms of market volatility during the
subprime mortgage crisis.
© 2011 Elsevier B.V. All rights reserved.
1. Introduction
As globalization and financial liberalization are enabling international financial markets to become
more correlated and connected than ever before, an understanding of the correlations and interactions
among various financial markets is crucial for investors, financial institutions, and governments.
Early literature focused on the correlations among the financial markets of developed countries (see for
example, Eun and Shim (1989), Hamao et al. (1990), King et al. (1994)). The cited papers show that
⁎ Corresponding author at: Xi'an Jiaotong University, Jinhe Center for Economic Research, No. 28 Xian Ning West Road, Xi'an,
Shaanxi Province, PR China.
E-mail addresses: martonzhou2003@yahoo.com.cn (X. Zhou), softmagic211@gmail.com (W. Zhang),
seraphj15@econmail.tamu.edu (J. Zhang).
0927-538X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.pacfin.2011.08.002
248 X. Zhou et al. / Pacific-Basin Finance Journal 20 (2012) 247–270
developed financial markets are interconnected and that the volatility of the US stock market is transmit-
ted to other developed markets. In the last 20 years, with the development of emerging financial markets,
financial economists have become increasingly interested in the relationship between emerging and de-
veloped markets and its meaning to financial liberalization and global integration. For example, Bekaert
and Harvey (1995) estimated the degree of integration between major emerging markets and world
equity markets from 1969 to 1992. Bekaert and Harvey (1997) found that capital market liberalization
often leads to a higher correlation between local and international markets. Janakiramanan and Lamba
(1998) and Ng (2000) found that before 1996, volatility in the US and Japanese equity markets spilled
over significantly to the stock markets of the Pacific Basin, including Hong Kong, Korea, Malaysia, Singapore,
Taiwan, and Thailand.
As a young startup, the Chinese stock market has developed at a fast pace. According to statistics from
the Chinese Securities Regulatory Commission (CSRC), as of the end of May 2010, the total market value of
the Shanghai and Shenzhen stock markets had reached 3.07 trillion USD, ranking third in the world behind
the NYSE and the NASDAQ. With this type of growth, many scholars are paying more attention to the cor-
relations between the Chinese and international markets. As one of the first papers on the Chinese stock
market, Bailey (1994) found that although some world market indicators could enable the explanation
of the characteristics of the Shanghai and Shenzhen markets, these two markets were not integrated
with world equity markets in the early 1990s. Furthermore, Wang and Firth (2004) proposed that from
1994 to 2001, the direction of the return spillover was from developed markets such as the US, the UK,
and Japan to markets in Greater China, such as mainland China, Hong Kong, and Taiwan. After the 1997
Asian financial crisis, the volatility spillover has been bidirectional between the Greater China markets
and the developed markets. Wang and Di Iorio (2007) found little evidence that the Chinese A stock
index was better correlated with the MSCI world index, indicating that the Chinese A stock market was
isolated from world markets between 1994 and 2004. Lin et al. (2009) argued that from 1992 to 2006,
the Chinese A stock market had no significant correlation with world equity markets, whereas the B
stock market had some connection with Western markets and more correlation with Asian markets.
In the volatility spillover literature, the common econometric methodologies are the multivariate
Generalized AutoRegressive Conditional Heteroskedasticity (GARCH), Regime Switching (RS) and Stochastic
Volatility (SV) models. The multivariate GARCH model is most commonly used by researchers. Departing
from the methods above, Diebold and Yilmaz (2009) provided new measures of return and volatility spill-
overs of international stock markets based on forecast-error variance decompositions in a vector autoregres-
sive framework (DY 2009). Diebold and Yilmaz (2011a) discussed the return and volatility spillover among
five American countries using this method. Yilmaz (2010) used the same method to evaluate the return
and volatility spillover among major Asian countries. More importantly, Diebold and Yilmaz (2011b)
further improved the DY 2009 method and used the upgraded model (DY 2011) to explore the spillover
among major American financial assets including stocks, bonds, foreign exchanges, and commodities from
1999 to 2009, with particular attention to the volatility interaction during the subprime mortgage crisis.
Our paper differs from the previous literature in that we are the first to use the DY 2011 framework
to shed light on the volatility spillover between the Chinese and world equity markets. The DY 2011
method that we used has several advantages over other models. First, this method does not depend
on the Cholesky factor identification of VAR; therefore, the results of variance decomposition do not
hinge on the sequence of the variables. Hence, the DY 2011 is superior to the DY 2009. Second, not
only can the DY 2011 be used to gauge the magnitude of the volatility spillover, it can indicate the di-
rection of the spillover as well. In other words, it may provide the value of the directional spillover be-
tween any two markets, between one market and any set of (regional) markets, or between one market
and global (all) markets. Third, as Diebold and Yilmaz pointed out, the DY 2011 avoids the controversial
issues associated with the definition and existence of episodes of contagion (see the debate in Forbes
and Rigobon (2002)).
In addition to exploring the volatility correlation between the Chinese and any other market, as in the
previous literature, our paper investigates both the magnitude and direction of the volatility spillover be-
tween China and the Greater China markets, between Chinese and the Asian markets, and between China
and the global (all) market. This could provide a more vivid picture of the position and power of the
Chinese stock market in the world arena. In addition, because we can easily define and measure the direc-
tion of the spillover by variance decomposition, we can measure the impact of the Chinese market on any
X. Zhou et al. / Pacific-Basin Finance Journal 20 (2012) 247–270 249
other market or markets, the impact of any market or markets on the Chinese market, and the net impact
of the Chinese market on any other market or markets. This distinctive feature provides a better estima-
tion of the directional spillover than a measure of the significance of coefficients under the special variance
structures in multivariate GARCH models.
Our study also provides a new perspective on the literature on financial integration and liberalization.
We discuss the volatility correlation among the Greater China markets, Asian markets and global markets,
focusing on the effect of the recent financial crisis on volatility transmission. Diebold and Yilmaz (2009)
did not address this issue owing to the limitations of their data and methodology. Although Yilmaz
(2010) discussed the correlation among Asian markets and his data range covered this crisis, he only
used the DY 2009 model and provided the total spillover index rather than any directional index. Another
limitation in his paper is that the DY 2009 method depends on the Cholesky factor identification of VAR;
therefore, the results are too sensitive to variable ordering. Adopting the new DY 2011 method allows us
to evaluate the volatility correlations among global markets from the perspective of the Chinese stock
market. In this respect, our paper could be regarded as a good supplement to Diebold and Yilmaz's
(2011b) investigation of volatility spillovers across different assets during the crisis.
In our study, we found that from 1996 to 2009, the Chinese stock market was scarcely affected by other
markets because it was not completely open to foreign investors. Before 2005, the Chinese market was
only slightly affected by other markets. After 2005, the Chinese stock market had a significant volatility
spillover to other markets, the net effect of which was positive, indicating that the influence of the Chinese
stock market had increased substantially. The volatility interactions among the Chinese, Hong Kong, and
Taiwanese markets are more prominent than those among the Chinese, Western, and other Asian markets
are; this confirms the effect of financial market integration in the Greater China region. In general, volatil-
ity spillovers among the Chinese, Japanese, and Indian markets are more distinctive than those among the
Chinese, US, and UK markets are; this indicates that the connections and correlations among Asian stock
markets have become increasingly evident in recent years.
Regardless of whether it affected or was affected by other markets, the US stock market was highly cor-
related with other markets in terms of volatility, especially during the subprime mortgage crisis. Other
markets were also very volatile, and, driven by bad news, their massive volatilities were transmitted
back to the US market. The major correction of the Chinese stock market between February and July
2007 significantly contributed to the volatility surges of other markets. Owing to the restrictions on for-
eign investment, the Chinese stock market has not been greatly affected in terms of market volatility dur-
ing the crisis.
The remainder of the paper is organized as follows. Section 2 outlines the methodology and model used
in this paper. Section 3 introduces the data and provides descriptive statistics. Section 4 presents our sub-
stantive results and discussion. Section 5 concludes the paper.
2. Methodology
We first present the concept and measure of the spillover index from Diebold and Yilmaz (2009). Con-
Pp
sider covariance stationary N variables VAR(p), xi ¼ Φi xt−i þ εi , where ε ∼ (0, Σ) is the vector of IID dis-
i¼1 P∞
turbances. We can turn the VAR into a moving average (MA) representation, that is, xt ¼ Ai εt−i where
i¼0
N × N coefficients matrix, Ai, obey Ai = Φ1Ai− 1 + Φ2Ai− 2 + … + ΦpAi− p, with A0 an N × N identity matrix
and Ai = 0 for i b 0. We use MA to forecast the future with the H-step-ahead. The forecast error variances
of each variable are decomposed into parts attributable to the various system shocks. Specifically, we
can evaluate the fraction of the H-step-ahead error variance in forecasting xi, which is due to shocks to
xj, ∀ j ≠ i for each i. It is necessary to obtain orthogonal innovations for variance decomposition, which
could be achieved by the identification schemes of Cholesky factorization. Therefore, the results of the var-
iance decomposition depend on the ordering of the variables.
Furthermore, Diebold and Yilmaz (2011b) used the generalized VAR framework proposed by Koop
et al. (1996) and Pesaran and Shin (1998) (hereafter, KPPS) and produced a variance decomposition in-
variant to ordering. Diebold and Yilmaz defined “own variance shares” as the fraction of the H-step-
ahead error variances in forecasting xi due to shocks to xi for i = 1,2,..,N and “cross variance shares” as
the fractions of the H-step-ahead error variances in forecast xi due to shocks to xj, for i, j = 1, 2,…, N, such
that i ≠ j. The KPPS H-step-ahead forecast error variance decomposition is
2
g
σii −1 ∑h¼0
H−1
ei′Ah Σej
θij ðHÞ ¼
H−1 ′
; ð1Þ
∑h¼0 ei Ah ΣAh′ ei
where Σ is the variance matrix for the error vector ε, σii is the standard deviation of the error term for the
ith equation, and ei is the selection vector with 1 as the ith element, and 0 otherwise. According to the
P
N
characteristics of generalized VAR, 1 we have θgij ðH Þ≠1. We normalize each entry of the variance decom-
j¼1
position matrix by the row sum as
g
g θij ðH Þ
θ̃ ij ðH Þ ¼ N
; ð2Þ
g
∑ θij ðH Þ
j¼1
P
N g P
N g
where θ̃ ij ðHÞ ¼ 1 and θ̃ ij ðH Þ ¼ N.
j¼1 i; j¼1
We can construct a total volatility spillover index as follows:
X
N
g N
θ̃ ij ðHÞ ∑ θ̃ ij ðH Þ
g
i; j ¼ 1 i; j ¼ 1
g i≠j i≠j
S ðH Þ ¼ ·100 ¼ ·100: ð3Þ
N
g N
∑ θ̃ ij ðH Þ
i;j¼1
The index is used to measure the contributions from the spillovers of volatility shocks across various
financial markets to the total forecast error variance.
As the KPPS framework solves the ordering problem, we can calculate the directional volatility spill-
over index using the normalized elements of the variance decomposition matrix. Here, we present the di-
rectional volatility spillovers received by market i from all other markets j as:
N
g
∑ θ̃ ij ðH Þ
j¼1
g j≠i
Si· ðHÞ ¼ N
·100: ð4Þ
g
∑ θ̃ ij ðH Þ
j¼1
Likewise, we can calculate the directional volatility spillovers transmitted by market i to all other mar-
kets j as:
N
g
∑ θ̃ ji ðH Þ
j¼1
g j≠i
S·i ðHÞ ¼ N
·100: ð5Þ
g
∑ θ̃ ji ðHÞ
j¼1
1
Diebold and Yilmaz (2011b) explained, “as the shocks to each variable are not orthogonalized, the sum of contributions to the
variance of forecast error (that is, the row sum of the elements of the variance decomposition table) is not necessarily equal to one.”
We can also obtain the net volatility spillover from market i to all other markets j by calculating the dif-
ference between Eqs. (5) and (4) as
g g g
Si ðH Þ ¼ S·i ðH Þ−Si· ðH Þ: ð6Þ
Similarly, we can measure the net pairwise volatility spillover between markets i and j as:
0 1
B θ̃ g ðH Þ θ̃ ji ðH Þ C
g
g B ij C
Sij ðH Þ ¼ BN − N C·100: ð7Þ
@ g g A
∑ θ̃ ik ðH Þ ∑ θ̃ jk ðH Þ
k¼1 k¼1
The method above, from Diebold and Yilmaz (2011b), primarily uses the “row normalization” to obtain
the spillover indices. Alternatively, we can normalize the elements of the variance decomposition matrix
with the column sum of these elements and compare the results with those of the “row normalization” for
verifying that the indices are robust to these changes.
In contrast to Eq. (2), we present the basic spillover index as:
g
g θij ðHÞ
θ̃ ij ðH Þ ¼ N
ð8Þ
g
∑ θij ðHÞ
i¼1
P
N g P
N g
where θ̃ ij ðHÞ ¼ 1 and θ̃ ij ðHÞ ¼ N.
i¼1 i;j¼1
The total volatility spillover index is:
N N
g g
∑ θ̃ ij ðHÞ ∑ θ̃ ij ðHÞ
i; j ¼ 1 i; j ¼ 1
g i≠j i≠j
S ðH Þ ¼ ·100 ¼ ·100: ð9Þ
N
g N
∑ θ̃ ij ðH Þ
i;j¼1
We can use these “column normalized” indices to calculate the volatility spillover across world equity
markets, Asian stock markets, and Greater China stock markets, as well as discuss issues such as globaliza-
tion and financial integration.
3. Data and descriptive statistics
The dataset we use covers the period from February 1, 1996 through December 30, 2009 and includes
the daily stock market indices in local currencies. Our data providers are Yahoo Finance and Bloomberg.
We select the S&P 500 and FTSE 100, as New York and London are the major world finance centers. The
major stock indices in France and Germany are chosen as the main stock indices in Europe. Because
China is an Asian country, we select stock indices in Japan and India as the major Asian market indicators,
stock indices in Korea and Singapore as the two small Asian market indicators, and the Greater China stock
indices of Hong Kong, Taiwan, and Shanghai. 2 Overall, our dataset contains 11 stock indices of various
countries or regions with 2669 pieces of daily data in each index.
2
The SSE Composite Index is an indicator for the overall movement of stocks in the Shanghai Stock Exchange. It is regarded as the
most common index for the Chinese stock market, because most large Chinese firms are listed on the Shanghai Stock Exchange. At
the end of 2008, the total market value of the Shanghai market was approximately 9.7252 trillion RMB, whereas the total market
value of the Shenzhen market was approximately 2.4115 trillion RMB. The latter accounts for approximately 25% of the former.
Therefore, we only use the SSE Composite Index as the indicator of stock markets in mainland China.
Consistent with the range-based method in Diebold and Yilmaz (2011b), 3 we estimate the daily vari-
ance using daily high and low prices. For market i on day t we get
h i2
2 max min
σ̃ it ¼ 0:361 ln Pit − ln Pit
where Pitmax is the highest price in market i on day t and Pitmin is the lowest daily price.
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi We further obtain the
annualized daily percent standard deviation (volatility) as σ̂ it ¼ 100 365·σ̃ it 2 and plot the volatilities of
six stock markets against that of the US in Fig. 1. The descriptive statistics are provided in Table 1.
Several findings were obtained, which are presented as follows. (1) As emerging markets, the Chinese
and Indian stock markets are more volatile than the US market. The Shanghai market is the most volatile
among the 11 markets, which is consistent with many studies on the Chinese stock market (e.g., Lin et al.,
2009). Moreover, the means of the volatilities of the Chinese, German, Indian, and Korean stock markets
are above 20%, whereas the Chinese, Indian, and Korean markets have median volatilities above 19%. (2)
The volatilities of developed markets, such as the UK, Japan, and Hong Kong, have relatively synchronized
fluctuations with the US market; this is especially true for the UK market. The Chinese, Indian, and Korean
markets have higher volatility than the US market, and their fluctuations are distinct from the movements
of the US market. (3) Volatility spikes occurred in almost all markets in October 2008 following the col-
lapse of Lehman Brothers. The volatility of the US market also reached its climax at that time, significantly
dwarfing the volatility of the Chinese market.
We compare the volatility of the US market with the VIX index 4 for the sample period (the figure is
available upon request) and find that our volatility measure is highly consistent with the VIX index.
Since the VIX is widely considered to be the best gauge of the volatility of the US market, both in academia
and in industry, this indicates that our range-based volatility measure is relevant and appropriate.
As Fig. 1 shows, the volatility hikes of the US market occurred in September 1998, July 2000, and Octo-
ber 2008, which illustrates the effect of the LTCM bailout, the burst of the IT bubble, and the bankruptcy of
Lehman Brothers. The volatility of the Chinese market was above 20% during most of the sample period,
which was much higher than that of the US market. There were volatility spikes in the Chinese market
in 1996, 1997, 2007, and 2009, signifying the impact of the Asian financial crisis in the 1990s and the re-
cent financial crisis.
4. Empirical results
4.1. Non-synchronous trading problem
It is obvious that stock market transactions are not synchronous (trading hours of various stock mar-
kets are given in Table 2). The volatility spillover literature must address this non-synchronous trading
problem before any meaningful results can be obtained.
Soriano and Climent (2006) summarized the solutions to the non-synchronous trading problem from
previous literature. They categorized international stock markets into three types based on their trading
hours: totally overlapping, non-overlapping, and partially overlapping markets. With respect to the first
type, there is no need to adjust for time, as these stock markets are synchronous and the close-to-close re-
turn could be useful. As for the second type, we must use open-to-close and close-to-open returns in order
to overcome the problem of non-overlapping trading (see Hamao et al., 1990 and Bae and Karolyi, 1994 for
details). The third type is the most complicated. Some common solutions involve the analysis of the data
from American Depositary Receipt (ADR) markets (Wongswan, 2003), dividing international markets into
several regions with non-overlapping trading hours (Melvin and Peiers, 2003) or using the data (e.g.,
weekly or two-day returns) from non-overlapping markets (Forbes and Rigobon, 2002).
3
Diebold and Yilmaz (2011b) referred to the research of Parkinson (1980) and others to determine the range-based estimation of
the daily volatility of stock markets.
4
VIX is the ticker symbol for the Chicago Board Options Exchange Market Volatility Index, a popular measure of the implied vol-
atility of S&P 500 index options. It represents one measure of the market's expectations of stock market volatility over the next 30-
day period.
US and CN US and HK
140 160
US US
CN HK
120 140
120
100
100
Index (%)
Index (%)
80
80
60
60
40
40
20
20
0 0
Feb96 Apr97 Jul98 Aug99 Nov00 Dec01 Mar03 Apr04 May05 Jun06 Jul07 Sep08Nov09 Feb96 Apr97 Jul98 Aug99 Nov00 Dec01 Mar03 Apr04 May05 Jun06 Jul07 Sep08 Nov09
Ending Date of window Ending Date of window
US and JP US and IN
140 200
US US
JP 180 IN
120
160
100 140
120
Index (%)
Index (%)
80
100
60
80
40 60
40
20
20
0 0
Feb96 Apr97 Jul98 Aug99 Nov00 Dec01 Mar03 Apr04 May05Jun06 Jul07 Sep08Nov09 Feb96 Apr97 Jul98 Aug99 Nov00 Dec01 Mar03 Apr04 May05 Jun06 Jul07 Sep08 Nov09
US and TW US and UK
140 140
US US
TW UK
120 120
100 100
Index (%)
Index (%)
80 80
60 60
40 40
20 20
0 0
Feb96 Apr97 Jul98 Aug99 Nov00 Dec01 Mar03 Apr04 May05 Jun06 Jul07 Sep08 Nov09 Feb96 Apr97 Jul98 Aug99 Nov00 Dec01 Mar03 Apr04 May05 Jun06 Jul07 Sep08Nov09
Fig. 1. Daily stock market volatilities.
As shown in Table 2, the 11 stock exchanges are spread widely across Asia, Europe, and America, with
heterogeneous trading hours. The stock markets in the UK, France, and Germany can be considered to be
completely overlapping markets in Europe, whereas the stock markets in Japan, China (Shanghai), India,
Hong Kong, Taiwan, Korea, and Singapore can be considered totally overlapping markets in Asia. The US
market does not overlap with the Asian markets; however, it somewhat overlaps with the European mar-
kets. The Asian markets are somewhat overlapping with the European markets. Considering this, we adopt
Table 1
Volatilities-descriptive statistics.
This table provides the summary statistics of volatilities in various markets. “CN” is China, “FA” is France, “GM” is Germany, “HK” is
Hong Kong, “IN” is India, “JP” is Japan, “KR” is Korea, “SP” is Singapore, “TW” is Taiwan, “UK” is the United Kingdom, and “US” is the
United States.
CN FA GM HK IN
Mean 23.21 18.97 21.12 19.53 23.48

Median 19.05 15.79 17.83 16.13 19.76
Maximum 133.21 100.29 127.89 157.53 197.43
Minimum 2.97 3.32 0.00 3.88 4.58
Std. dev. 15.49 11.77 14.70 12.77 14.61
Skewness 2.20 1.97 1.97 2.71 2.66
Kurtosis 10.28 8.77 9.46 17.61 17.37
Observations 2669 2669 2669 2669 2669
JP KR SP TW UK US
Mean 18.72 24.80 17.01 18.76 17.10 17.37

Median 16.42 20.70 13.84 16.18 14.27 14.38
Maximum 134.80 181.84 148.66 109.06 111.21 125.17
Minimum 3.44 3.87 2.94 1.68 2.67 2.75
Std. dev. 11.08 15.28 12.14 10.92 11.25 11.89
Skewness 2.85 2.16 3.03 1.86 2.19 3.00
Kurtosis 19.00 12.33 19.81 9.12 11.27 18.55
Observations 2669 2669 2669 2669 2669 2669
a simple method that is similar to Cai et al.'s (2009) method in order to solve this non-synchronous trading
problem: the daily volatilities of the US and European markets are lagged one day in estimation of the VAR
model together with the current volatilities of the Asian markets. However, our results show that there is
little difference before and after this adjustment. Nonetheless, in our paper, we mainly present the empir-
ical results with this adjustment and discuss the economic implications based on them.
4.2. The full-sample volatility spillover table
We calculate θijg(H) in Eq. (2) of Section 3 with the full sample and obtain the first part of Table 3. The
ij th entry is the estimated contribution to the forecast error variance of market i resulting from innovations
to market j. The sum of the variances in a row (column), excluding the contribution to its own variance
(diagonal variances), denotes the magnitude of the influence from the volatilities of other markets (impact
on the volatilities of other markets), and the “from minus to” differences are the net volatility spillovers.
Moreover, owing to row normalization, the sum of the variances in a row is 100%. The last row in the
table represents the contribution to the volatilities of all markets from this particular market. The total
Table 2
Trading hours of various stock exchanges.
Stock exchange Greenwich time Local time
New York 14:30–21:00 9:30–16:00

London 8:00–16:30 8:00–16:30
Paris 8:00–16:30 9:00–17:30
Frankfurt 8:00–19:00 9:00–20:00
Tokyo 0:00–6:00 9:00–15:00
Shanghai 1:30–3:30, 5:00–7:00 9:30–11:30, 13:00–15:00
Bombay 3:30–10:00 9:00–15:30
Hong Kong 2:00–4:30, 6:30–8:00 10:00–12:30, 14:30–16:00
Taiwan 1:00–5:30 9:00–13:30
Seoul 23:00–6:00 8:00–15:00
Singapore 1:00–9:00 9:00–17:00
Table 3
Volatility spillover index (row and column normalization).
FA HK SP US TW IN GM KR JP UK CN From others
FA 22.56 3.95 3.98 11.09 3.94 3.16 24.83 4.03 8.27 11.55 2.62 77.44
HK 2.36 39.55 12.27 4.18 4.62 6.71 2.53 12.47 7.96 2.53 4.81 60.45
X. Zhou et al. / Pacific-Basin Finance Journal 20 (2012) 247–270

SP 3.28 14.24 36.21 5.86 3.60 6.33 4.05 10.27 9.13 3.74 3.30 63.79
US 9.81 5.15 5.94 28.87 3.30 4.94 14.10 4.68 10.69 9.18 3.35 71.13
TW 3.75 6.15 3.69 4.10 42.42 6.82 6.10 13.69 6.27 2.57 4.44 57.58
IN 1.76 5.60 3.85 3.15 3.55 60.26 1.37 6.27 6.00 1.74 6.45 39.74
GM 15.94 2.94 3.63 10.89 3.48 1.86 37.04 4.28 7.91 10.85 1.18 62.96
KR 1.64 8.68 7.28 3.28 5.67 5.37 2.76 53.02 8.13 1.52 2.66 46.98
JP 3.95 7.56 6.34 6.39 4.51 7.27 5.38 11.88 38.55 3.23 4.93 61.45
UK 13.71 4.86 5.66 12.67 3.37 3.95 19.76 3.93 8.65 20.35 3.09 79.65
CN 1.09 3.14 1.87 1.27 2.35 5.05 0.56 1.16 2.52 0.97 80.03 19.97
To others 57.30 62.26 54.50 62.88 38.39 51.46 81.43 72.68 75.54 47.87 36.84
Including own 79.86 101.82 90.71 91.75 80.81 111.72 118.47 125.69 114.09 68.22 116.87 Total spillover index: 58.29
FA HK SP US TW IN GM KR JP UK CN From others Including own
FA 28.25 3.88 4.39 12.09 4.88 2.83 20.96 3.21 7.25 16.92 2.25 78.66 106.91
HK 2.96 38.85 13.53 4.56 5.72 6.01 2.13 9.92 6.98 3.71 4.11 59.63 98.47
SP 4.11 13.98 39.92 6.38 4.45 5.67 3.42 8.17 8.00 5.49 2.82 62.49 102.41
US 12.29 5.06 6.54 31.46 4.08 4.42 11.90 3.73 9.37 13.46 2.86 73.71 105.17
TW 4.69 6.04 4.07 4.47 52.50 6.10 5.15 10.89 5.50 3.76 3.80 54.47 106.97
IN 2.21 5.50 4.24 3.43 4.39 53.94 1.16 4.99 5.26 2.55 5.52 39.25 93.19
GM 19.95 2.89 4.00 11.87 4.31 1.66 31.27 3.41 6.94 15.91 1.01 71.94 103.21
KR 2.05 8.53 8.02 3.58 7.01 4.81 2.33 42.18 7.12 2.23 2.28 47.95 90.13
JP 4.95 7.42 6.99 6.97 5.59 6.51 4.54 9.45 33.79 4.73 4.22 61.37 95.16
UK 17.17 4.77 6.24 13.81 4.17 3.53 16.68 3.12 7.58 29.83 2.65 79.73 109.56
CN 1.37 3.08 2.06 1.38 2.91 4.52 0.47 0.92 2.21 1.42 68.48 20.34 88.82
To others 71.75 61.15 60.08 68.54 47.50 46.06 68.73 57.82 66.21 70.17 31.52 Total spillover index: 59.05
255
volatility spillover index, which appears in the lower right-hand corner of the table, is computed as the
sum of all variances in the 11 × 11 matrix minus the sum of the diagonal variances.
g
Alternatively, as shown in the second part of Table 3, we calculate θ̃ ij ðH Þ in Eq. (8) by the method of
column normalization, making the sum of the variances in every column equal to one. It is obvious that
the sum of variances in each row is no longer equal to one. Similarly, we can obtain the “to” and “from”
directional spillovers, net volatility spillover, and total volatility spillover index.
In the first part of the row normalization, the ij th entry in the 11 × 11 matrix is the estimated contribu-
tion to the forecast error variance of market i resulting from innovations to market j. For example, the el-
ement in row 4, column 11 is the volatility contribution from the Chinese market to the variance of the US
market (3.55%). The element in row 11, column 4 is the volatility contribution from the US market to the
variance of the Chinese market (1.27%). The sum of the variances in column 4 is the total contribution from
the US market to all other markets in terms of volatility spillover (62.88%). The sum of the variances in row
4 is the total contribution to the US market spillover from all other markets (71.13%). Total spillovers from
other markets to a particular market are indicated in the right-most column. The spillovers from this mar-
ket to the other markets are shown in the second-to-last row in the table. The total volatility spillover
index is in the lower right-hand corner of the table, which is calculated as the average of the spillovers
from all other markets, that is, 641.14/11 = 58.29. This indicates that in the full sample, approximately
58.29% of the forecast error variances are due to volatility spillovers among different markets.
As the first part of Table 3 shows, we find that the markets affecting others the most are the German,
Korean, Japanese, the US, and Hong Kong markets, whereas the French, the US, and the UK markets are the
most affected by others. The contribution of the Chinese market to other markets is 36.84%, and the effect
of the other markets on the Chinese market is 19.97%, which is lower than the contribution of the other
markets.
As the second part of Table 3 shows, we find that the markets affecting other markets the most are the
French, UK, US, German and Japanese markets, while the French, US, German, and UK markets are the most
affected by others. The contribution of the Chinese market to the others is 31.52%, and the effect of the
other markets on the Chinese market is 20.34%, which is low compared with those of other markets.
Hence, the results are robust to methods of row and column normalization. Most open markets, i.e.,
those with high spillover with other markets, include the US, the UK, German, French, and Japanese mar-
kets, whereas the Chinese and Indian markets are the most closed in our sample.
4.3. Dynamic rolling — sample analysis
The full-sample static analysis alone is insufficient, because no single number can reflect the dynamic
volatility relationship between two countries, much less be used to evaluate the impact of the financial cri-
sis on volatility spillover across various markets. Considering this inefficiency, we forecast future volatility
spillovers with a 10-day horizon based on the VAR (2) model 5 using 100-, 500-, and 1000-day rolling sam-
ples 6 (hereafter win100, win500, and win1000, respectively). We find that the fluctuation of the volatility
spillover in the 100-day rolling sample is too vehement for us to detect any meaningful pattern. Hence, we
focus primarily on the results of volatility spillover using 500-day and 1000-day rolling samples.
4.3.1. Spillovers in regional markets
4.3.1.1. All 11 countries. The total volatility spillover among the 11 markets during the sample period is pre-
sented in Fig. 2. The red and blue lines denote the results from row and column normalization, respective-
ly. We find that these lines are very close, indicating that the two methods are similar in the calculation of
total spillover indices.
5
The Bayes Information Criterion (BIC) is used to select the order of the VAR.
6
In Section 4.4, we examine the results for forecast horizons varying from 5 days to 10 days as well as calculate the dynamic in-
dices for orders 2 through 5 of VAR. All results are fairly robust to the different choices of the order of VAR and the forecast horizon.
GlobalSpillindex , Window=500 GlobalSpillindex , Window=1000

80 75
75
70
70
65
Index (%)
Index (%)
65
60
60
55
55
50
Row Normalization Row Normalization
Column Normalization Column Normalization
45 50
Oct98 Nov99 Dec00 Dec01 Jan03 Feb04 Jan05 Feb06 Jan07 Feb08 Mar09 Jun01 Jul02 Jul03 Jul04 Aug05 Aug06 Jul07 Aug08 Aug09
Fig. 2. Total volatility spillovers (global markets).
We highlight some major events in these plots:
1. From 1999 to 2001, the spillover index rose from 50% to 65% in the win500 chart, revealing that the
burst of the IT bubble in the US greatly contributed to the spillover hike among global equity markets.
2. The spike of the spillover index in October 2006 indicates the first signs of the subprime issues. At that
time, housing sales slumped and the housing default rate soared, especially for those houses that were
supported by subprime lending.
3. The period of the highest volatility correlations of global stock markets was from July 2007 through
December 2008. The spillover reached its climax in September and October 2008, when Lehman
Brothers' bankruptcy was announced. From November 2008 through August 2009, the spillover de-
creased gradually, and stabilized thereafter.
4.3.1.2. Asia and Greater China. Volatility spillovers among the Asian stock markets and among the Greater
China markets, including Shanghai, Hong Kong, and Taiwan, are presented in Figs. 3 and 4, respectively.
The Asian and Greater China spillover indices generally surged during the sample period because of in-
creasing financial integration among the Asian countries. This finding is confirmed by the previous litera-
ture. For example, Chi et al. (2006) analyzed the integration of the East Asian stock markets based on the
ICAPM framework and proposed that this degree of integration continued to rise from 1991 to 2005.
AisaSpillindex , Window=500 AisaSpillindex , Window=1000

60 55
55
50
50
45
45
Index (%)
Index (%)
40 40
35
35
Row Normalization Row Normalization
30
Column Normalization Column Normalization
30
25
20 25
Fig. 3. Total volatility spillovers (Asian markets).

Fig. 4. Total volatility spillovers (Greater China markets).
Moreover, Johansson and Ljumgwall (2008) noted that the three stock markets in the Greater China region
were significantly interdependent in terms of returns and volatilities from 1994 through 2005.
With respect to the ranges of the spillover index, the Asian index was almost half of the global index,
and the Greater Chinese index was almost half of the Asian index. This shows that the financial integration
of Asia with global markets was relatively low and that the financial integration of Greater China was even
lower than that of Asia as a whole. The explanation for this may hinge on the lack of openness of Asian
markets in comparison with Western markets. In particular, as an important player in Greater China
and Asia, the Chinese market has not yet released its restrictions on foreign investments.
The major events behind the high spillover indices in Asia and the Greater China region include the
burst of the IT bubble in June 2000, interest rate cut by the federal government between February and
July 2004, and Lehman Brothers' bankruptcy in September and October 2008. Interestingly, we could
not find spillover hikes from February to July 2007 in the Asian spillover plot, although we detected spill-
over surges in the Greater China plot during this period. This phenomenon is consistent with the volatility
fluctuations discussed in Chapter 3. In other words, the Chinese market appeared very volatile in as early
as February 2007, whereas the Hong Kong and Taiwanese markets experienced little fluctuation during
that time. Thus, the major spillover hikes in the Greater China region may be owing to volatility transmis-
sion from the Chinese market to the Hong Kong and Taiwanese markets.
4.3.2. Volatility spillovers between the Chinese market and other markets
4.3.2.1. China vs. the World. Fig. 5 illustrates the volatility spillover between the Chinese market and the
other ten markets. The blue line denotes the spillover from the Chinese market to the other markets.
The red line denotes the spillover from the other markets to the Chinese market. The black line is the
net spillover from the Chinese market to the other markets.
The volatility contribution of the Chinese market to the world equity markets is concentrated between
20% and 70% (win500) and 20% and 60% (win1000), whereas the index fluctuated between 20% and 40%
before 2005 and continued to increase afterward. The highest value appeared in July 2007, and the second
highest in August 2008. These findings indicate that the Chinese market has become increasingly impor-
tant in world equity markets since 2005 because of many major financial liberalization and open market
policies.
The volatility spillover from the world markets to the Chinese market was between 20% and 40%. It
fluctuated randomly and without any clear pattern, demonstrating that the Chinese market was a closed
market and not significantly affected by the other markets in terms of volatility. The win500 index was rel-
atively high in 2000 owing to the mild impact of the burst of the IT bubble on the Chinese market.
We found the net spillover from the Chinese market to the world markets in the range of − 20% to 60%
(win500) and −20% to 35% (win1000). Initially, from 1998 to 2005, it was below zero and advanced to
Fig. 5. Spillovers between the Chinese market and the other ten markets.
above zero subsequently, with spikes in 2007 and 2008, illustrating that the significance of the Chinese
market has emerged gradually since 2005.
The previous literature is consistent with the findings above. For example, Lin et al. (2009) found that
the Chinese A stock market had no significant relationship with world equity markets from 1992 to 2006,
as indicated in the DCC-GARCH models. Wang and Di Iorio(2007) showed that the Chinese A stock market
was isolated from the world markets between 1994 and 2004, and did not find any empirical evidence to
suggest a connection between the Chinese A stock index and the MSCI world market index.
Hence, the great enhancement of the role played by the Chinese stock market in the world arena be-
came evident after 2005. Moon and Yu (2010) argued that the year 2005 was a key period for the Shanghai
market. Prior to December 2005, the unexpected volatility of the S&P 500 index had nothing to do with the
volatility of the Shanghai index. Since December 2005, the Chinese Shanghai market has had a significant
and symmetric influence on the volatility of the US market. Moon and Yu postulated that the reform of
non-tradable shares and the first floating of the RMB exchange rate in 2005 after 10 years of a fixed ex-
change rate regime introduced this structural change.
Two vital reforms to the Chinese financial markets were initiated in 2005. First, on May 29, 2005, the
Chinese Securities Regulatory Commission (CSRC) published The Circular on Relevant Issues Concerning
the Pilot Reform of Non-Tradable Shares of Listed Companies and officially began non-tradable share reform
on December 12, 2005, which led to huge volatilities in the Chinese stock market in January 2006. Second,
the RMB exchange rate was adjusted to 8.11 RMB/USD on July 21, 2005, the first adjustment after ten years
of a fixed exchange rate at 8.28 RMB/USD. Thereafter, the RMB exchange rate has been on a gradual path of
appreciation; this would certainly induce structural changes to the international fund flow and have a no-
ticeable effect on the Chinese and world equity markets.
With respect to the policy effects in 2005, we construct a structural change model in the traditional
VAR framework and confirm the significance of the two reforms on volatility spillovers (Table 4). In
each model, in addition to the two-period volatilities of each market, we add different dummy variables
and interaction terms as the independent variables in VAR. We determine that two institutional reforms
of the Chinese stock market have a significant impact on the volatilities of various markets, especially on
the stock markets of the UK, Korea, India, Taiwan, Singapore, and Hong Kong.
Moreover, the likelihood ratio test of the hypothesis that the structural variables are insignificant is in-
vestigated. The likelihood ratio (LR) statistics indicate that all structural changes in the six models have
significant effects on the volatilities at the 1% significance level.
In contrast to the Chinese market, the volatility contribution of the US stock market to the world mar-
kets was concentrated between 60% and 80%, 7 which was three times that of the Chinese market at the
7
The spillover plot of the US and the world markets is available upon request.
260
Table 4
VAR with structural changes.
This table displays the estimation results of the VAR (2) of volatilities in all markets. We specify three dummy variables: DEXRt is a dummy that equals 1 if the time period is after the exchange rate
reform and 0 otherwise; DSOEt is a dummy that equals 1 if the time period is after the reform of non-tradable shares and 0 otherwise; and DSOE2t is a dummy that equals 1 if the time period is
between two reforms and 0 otherwise. DCNEXRt = DEXRt × CNt-1, DCNSOEt = DSOEt × CNt-1, and DCNSOE2t = DSOE2t × CNt-1. In the table, “CN” is China, “FA” is France, “GM” is Germany, “HK”
is Hong Kong, “IN” is India, “JP” is Japan, “KR” is Korea, “SP” is Singapore, “TW” is Taiwan, “UK” is the United Kingdom, and “US” is the United States. Standard errors are in brackets. *, **, and ***
indicate significance at 10%, 5%, and 1% levels, respectively.
Dependent FA HK SP US TW IN GM KR JP UK CN
Variable
Model 1
X. Zhou et al. / Pacific-Basin Finance Journal 20 (2012) 247–270

VAR (2)
DEXRt − 0.184 0.058 0.755⁎ 0.174 − 0.241 2.027⁎⁎⁎ − 0.260 − 2.575⁎⁎⁎ 0.267 2.120⁎⁎⁎ 3.828⁎⁎⁎
(0.396) (0.455) (0.417) (0.400) (0.441) (0.552) (0.456) (0.523) (0.431) (0.359) (0.616)
Model 2
VAR (2)
DSOEt − 0.159 0.004 0.739⁎ 0.167 − 0.255 1.993⁎⁎⁎ − 0.237 − 2.494⁎⁎⁎ 0.144 2.094⁎⁎⁎ 4.032⁎⁎⁎
(0.391) (0.450) (0.412) (0.395) (0.436) (0.545) (0.451) (0.517) (0.426) (0.355) (0.608)
Model 3
VAR (2)
DCNEXRt − 0.015 − 0.026⁎ 0.008 0.017 − 0.040⁎⁎⁎ 0.022 − 0.006 − 0.109⁎⁎⁎ − 0.019 0.068⁎⁎⁎ 0.027
(0.013) (0.015) (0.014) (0.013) (0.015) (0.018) (0.015) (0.017) (0.014) (0.012) (0.020)
Model 4
VAR (2)
DCNSOEt − 0.017 − 0.027⁎ 0.008 0.016 − 0.042⁎⁎⁎ 0.021 − 0.007 − 0.108⁎⁎⁎ − 0.021 0.067⁎⁎⁎ 0.032
(0.013) (0.015) (0.014) (0.013) (0.015) (0.018) (0.015) (0.017) (0.014) (0.012) (0.020)
Model 5
VAR (2)
DSOE2t 0.232 − 0.737 0.279 0.027 − 0.366 0.877 0.161 − 0.576 − 1.582 1.050 5.485⁎⁎
(− 1.476) (− 1.698) (− 1.557) (− 1.491) (− 1.647) (− 2.059) (− 1.702) (− 1.953) (− 1.609) (− 1.341) (− 2.298)
DEXRt − 0.178 0.040 0.761⁎ 0.174 − 0.250 2.048⁎⁎⁎ − 0.256 − 2.588⁎⁎⁎ 0.229 2.145⁎⁎⁎ 3.960⁎⁎⁎
(− 0.397) (− 0.457) (− 0.419) (− 0.401) (− 0.443) (− 0.554) (− 0.458) (− 0.526) (− 0.433) (− 0.361) (− 0.618)
Model 6
VAR (2)
DCNSOE2t − 0.036 − 0.033 0.000 − 0.003 − 0.045 0.013 − 0.021 − 0.038 − 0.049 0.012 0.093
(− 0.051) (− 0.059) (− 0.054) (− 0.052) (− 0.057) (− 0.072) (− 0.059) (− 0.068) (− 0.056) (− 0.047) (− 0.080)
DCNEXRt − 0.016 − 0.027 0.008 0.017 − 0.042⁎⁎⁎ 0.022 − 0.007 − 0.110⁎⁎⁎ − 0.020 0.069⁎⁎⁎ 0.030
(− 0.013) (− 0.015) (− 0.013) (− 0.013) (− 0.015) (− 0.018) (− 0.015) (− 0.017) (− 0.014) (− 0.012) (− 0.021)
beginning of the sample period, indicating the dominant position of the US market in the world. Moreover,
the volatility spillover from the US market to the world markets was very stable during the sample period.
Two mild spikes appeared, one in August 2007 and the other in October 2008, owing to the credit crunch
and bankruptcy of Lehman Brothers.
The spillover from the world markets to the US market was approximately 70% to 85% (win500) and
70% to 80% (win1000). The indices from July 2007 to October 2008 were slightly higher than those in
other periods were. It is evident that the spillover from the world markets to the US market was twice
that from the world markets to the Chinese market, again indicating the openness of the US market and
the lack of openness of the Chinese market.
The net volatility spillover from the US market to the world markets was in the range of −20% to 8%
(win500) and −10% to 2% (win1000). More importantly, the US net spillover was nearly negative for
the majority of the time, except from 2001 to 2004. The volatility of the US market was particularly affect-
ed by the world markets in 1999, 2006, and 2007. In September and October 2008, the net volatility spill-
over from the US market to other markets had sudden jumps, causing the index to rise slightly above zero
from August 2008 to August 2009. This illustrates that as a leading indicator of world markets, the US mar-
ket would have a huge impact on other markets during the financial crisis. However, the US market would
usually be more affected by the world markets due to its highly open nature.
4.3.2.2. China vs. Asia. Fig. 6 illustrates the volatility spillover between the Chinese market and major Asian
markets. The blue line denotes the spillover from the Chinese market to the Asian markets. The red line
denotes the spillover from the Asian markets to the Chinese market. The black line is the net spillover
from the Chinese market to the Asian markets.
The volatility spillover from the Chinese market to the Asian markets is concentrated between 10% and
140% 8 (win500) and 10% and 80% (win1000). The index was approximately 20% prior to 2005, but in-
creased quickly after 2005. The spillover hikes appeared between February 2007 and July 2007 and
again in October 2008. Once again, these results indicate that the Chinese market had little effect on the
Asian markets from 1999 to 2005. The spillover from the Chinese market to the Asian markets became
increasingly evident after 2005. In particular, the huge volatilities of the Chinese market in February
2007 were transmitted to the Asian markets, and were reinforced by the bankruptcy of Lehman Brothers.
The volatility spillover from the Asian markets to the Chinese market fluctuated between 10% and 40%
(win500) and 10% and 30% (win 1000) with no clear patterns, which confirms that the Chinese market
was hardly affected by the Asian markets owing to the Chinese market's lack of openness.
The net volatility spillover from the Chinese market to the Asian markets was between −20% and 120%
(win500) and − 20% and 60% (win1000). It remained negative from 2001 to 2005 and became positive
after 2005. Subsequently, it continued to rise between 2005 and 2007 and reached its zenith from January
to July 2007 and again from September to October 2008 before gradually declining each time.
4.3.2.3. China vs. Greater China. Fig. 7 illustrates the volatility spillover between the Chinese market and the
Greater China markets, including those of Hong Kong and Taiwan. The blue line denotes the spillover from
the Chinese market to the Greater China markets. The red line denotes the spillover from the Greater China
markets to the Chinese market. The black line is the net spillover from the Chinese market to the Greater
China markets.
The volatility spillover from the Chinese market to the Greater China markets is concentrated between
5% and 60% (win500) and 5% and 30% (win1000), which was lower, on average, than that from the Chinese
market to the Asian markets. The index was less than 10% from 1998 to 2004 and remained stable during
this period. It increased quickly after 2005 and reached its climax from February to July 2007, after which it
began to decline. The second highest spike was around August and September 2008. These findings are
similar to the results of spillovers between the Chinese and world markets and the results of spillovers be-
tween the Chinese and Asian markets.
8
The single item in Table 3 is less than 100%; however, we can get indices of more than 100% by summation. Thus, it is not sur-
prising to get some spillover indices of the Chinese market to the Asian markets above 100%.
CN and Asia, Window=500, Row Normalization CN and Asia, Window=1000, Row Normalization
160 100
Affected by Others
140 Affect Others
Net Affect Others 80
120
100 60
80
Index (%)
Index (%)
40
60
20
40
20 0
0
-20 Affected by Others
-20 Affect Others
Net Affect Others
-40 -40
Fig. 6. Volatility spillovers between the Chinese and Asian markets.
The volatility spillover from the Greater China markets to the Chinese market fluctuated between 5%
and 20% (win500) and 5% and 15% (win1000) with no clear pattern, indicating that the closed Chinese
market was rarely affected by the Greater China markets.
The net volatility spillover from the Chinese market to the Greater China markets was between −20%
and 60% (win500) and −10% and 25% (win1000). The index remained negative from November 1999
until July 2004, after which it rose continuously. It reached its first and second climaxes between February
and July 2007 and from August to September 2008, respectively, which is consistent with previous find-
ings on spillover between the Chinese market and other markets.
A number of previous studies have obtained findings that are similar to ours. For example, Hu et al.
(1997) found that the Taiwan and Hong Kong stock markets had no influence on the Chinese Shanghai
and Shenzhen markets in terms of volatility spillover from 1992 to 1996. Groenewold et al. (2004)
found that the Chinese market was isolated from the Hong Kong and Taiwanese markets between 1992
and 2001, although some evidence suggests that the Hong Kong market has weak predictive power
over the return of the Chinese market.
Fig. 7. Volatility spillovers between the Chinese and Greater China markets.
Fig. 8. Volatility spillovers between the Chinese and US markets.
4.3.3. Volatility spillovers between the Chinese and individual markets

Owing to limited space, we only discuss the volatility spillovers between the Chinese market and the
US, the UK, Japanese, Indian, Hong Kong, and Taiwanese markets. Other spillovers could certainly be
obtained by similar calculations.
4.3.3.1. China vs. the US. Fig. 8 illustrates the volatility spillovers between the Chinese and the US markets.
The blue line denotes the spillover from the Chinese to the US market. The red line denotes the spillover
from the US to the Chinese market. The black line is the net spillover from the Chinese to the US market.
Volatility spillovers from the Chinese to the US market varied across different rolling window widths.
They fell between 0% and 25% (win500) and 1% and 15% (win1000), which was fairly large compared with
that from the UK to the US market (4% to 15%), from the Japanese to the US market (5% to 15%), from the
Indian to the US market (0% to 10%), and from the Hong Kong to the US market (2% to 10%). In the win500
model, the index remained stable under 5% from 1998 to 2005, after which it began to rise. With a sharp
increase beginning in February 2007, it reached its first climax in July 2007 and a second in August 2008. In
the win1000 model, the index was stable from 2001 to 2007; however, after July 2007, it rose, then began
to fall in August 2008.
The volatility contribution from the US to the Chinese market was concentrated between 0% and 4%
(win500 and win1000) with no clear pattern. This was rather small compared to that of the US to the Japanese
market (4% to 10%), the US to the Hong Kong market (2% to 8% and 2% to 10%), the US to the UK market (8% to
15%) and the US to the Indian market (0% to 8% and 0% to 10%). These findings confirm that the Chinese mar-
ket was a closed market, as previously stated.
The net volatility spillover from the Chinese to the US market was between − 1% and 25% (win500) and
−1% and 12% (win1000); these are significant effects. Moreover, the index remained negative from 1999
to 2005 and continued rising after 2005 before reaching climaxes in July 2007 and October 2008.
The capital account of the Chinese market is still frozen, and the RMB is not a freely exchangeable cur-
rency. Although QFII 9 was proposed in December 2002, the total QFII quotas approved by the foreign ex-
change bureau only amounted to 15.73 billion USD as of September 30, 2009. Therefore, we argue that the
flow of foreign funds would not have a noticeable impact on the Chinese stock market, even though news
from the US and other major markets could influence the Chinese market to some degree. On the other
hand, according to statistics from the Chinese Securities Journal, 51 Chinese companies were listed on the
9
Qualified Foreign Institutional Investors (QFII) is a transitional policy measure taken by developing countries before full capital
account liberalization to attract foreign investment to local financial markets. The regulators and policymakers mandate some re-
quirements for foreign investors to enter the local market. Foreign investors must remit a fixed amount of foreign currency into
the country and transfer it into local currency. Moreover, these so-called QFIIs must trade using special accounts under the scrutiny
of financial regulators.
NYSE as of December 24, 2007; among them, 39 had a total market value of 1.5 trillion USD. 10 Further-
more, there were 204 Chinese companies listed on the US stock markets as of June 2008, including
some extremely large companies on the NYSE, such as the Chinese National Petroleum Corporation
(CNPC), Sino Petroleum Corporation (Sinopec), and China Mobile Communication Corporation; some
high-tech companies on the NASDAQ, such as Sina and Baidu; and other companies in several US OTC mar-
kets. We postulate that although there are not many companies cross-listed in the Chinese and US mar-
kets, the US-listed companies with “Chinese concepts” mainly conduct their business in China, and their
business partners are Chinese firms and institutions, including many locally listed companies. Hence, if
the Chinese stock market fluctuates violently, it would have a major impact on these “Chinese concept”
stocks in the US through changes in business interactions and the market situation in China, which
would lead, in turn, to noticeable volatility hikes in US stock markets. However, volatilities in the US mar-
kets would not be transmitted to Chinese stock markets in return because of the foreign investment
restrictions.
Spillover from the Chinese to the US market increased suddenly in February 2007, before the subprime
crisis had come into full play. We could also focus on the volatility spikes of the Chinese stock market in
February 2007, as shown in Fig. 1 of Chapter 4, when volatilities in the US market were dwarfed by
those in the Chinese market. We postulate that it was the major correction of the Chinese stock market
in February 2007 significantly affected the world equity markets, especially the US market, and the devas-
tating news of the Lehman Brothers bankruptcy may have exacerbated this adjustment in the end.
Our findings are consistent with the extant literature. For example, Wang and Firth (2004) showed that
the volatilities of the Shenzhen and Shanghai markets were not affected by current or delayed bad news
from developed markets after the 1997 Asian financial crisis. Moon and Yu (2010) found that unexpected
volatilities of the S&P 500 Index had little effect on the Shanghai market before December 2005, and that
the Shanghai market had a significantly symmetric impact on the volatilities of the S&P 500 Index after-
ward. Frank and Hesse (2009) demonstrated that the correction to the Shanghai market in February
2007, beginning of the subprime crisis, Bear Stearns rescue, and bankruptcy of Lehman Brothers caused
the correlations among financial assets across the globe to skyrocket at these particular points. More inter-
estingly, Frank and Hesse (2009, p. 14) posited that in February 2007, “The implied correlations between
the U.S. Libor-Overnight Indexed Swap (OIS) spread, a proxy for funding illiquidity, and the JP Morgan
Emerging Markets Bond Index Plus (EMBI+) sovereign bond spreads of Asia, Europe, and Latin American
countries sharply increase from 0.20 to almost 0.50 … All region spreads co-movements jump up around
the China stock market bust in similar magnitude.” In addition, using a DCC GARCH framework, Frank and
Hesse demonstrated that the correction to the Shanghai market led to a temporary spike of the correlation
measures in Brazil, Russia, and Turkey.
4.3.3.2. China vs. the UK. Fig. 9 illustrates the volatility spillovers between the Chinese and UK markets. The
blue line denotes the spillover from the Chinese market to the UK market. The red line denotes the spill-
over from the UK market to the Chinese market. The black line is the net spillover from the Chinese market
to the UK market.
Volatility spillovers from the Chinese market to the UK market varied across different models. They
were approximately 0% to 25% (win500) and 0% to 14% (win1000), similar to those from the Chinese mar-
ket to the US market. The volatility spillover from the US market to the UK market was in the range of 8% to
15% (win500) and 3% to 15% (win1000), which was, on an average, larger than that from the Chinese mar-
ket to the UK market, although it was more stable. The spillover from the Chinese market to the UK market
had the same pattern as that from the Chinese market to the US market. It fluctuated below 5% (win500)
from 1998 to 2005, then began to rise. It rose sharply from February 2007 and reached climaxes in July
2007 and August 2008 before falling. The win1000 index remained stable from 2001 to 2007 and increased
from July 2007 until August 2008.
The volatility from the UK market to the Chinese market fluctuated in the band of 0% to 4% without any
clear pattern, similar to that from the US market to the Chinese market, which was also from 0% to 4%. The
10
According to data from Wikipedia, NYSE's market capitalization of its listed companies was USD 12.25 trillion as of May 2010;
therefore, the market cap of Chinese companies listed on the NYSE makes up approximately 1/12 of the total market value.
Fig. 9. Volatility spillovers between the Chinese and UK markets.
volatility from the UK market to the US market was in the range of 4% to 15% (win500) and 5% to 15%
(win1000) during the same sample period.
The net volatility spillover from the Chinese to the UK market was between −1% and 22% (win500)
and − 1% and 12% (win1000). The index remained negative from 2002 to 2004 and became positive
after 2005. The climaxes were around February 2007 and August 2008. These findings are comparable
to those of the spillover between the Chinese and US markets. In fact, the volatility movement of the UK
market was well synchronized with that of the US market, as shown in the volatility figure in Chapter 3.
We can therefore infer that the huge volatilities of the Shanghai market in early 2007 might also have
been transmitted to the UK market.
4.3.3.3. China vs. Japan. Fig. 10 illustrates the volatility spillovers between the Chinese and Japanese mar-
kets. The blue line denotes the spillover from the Chinese to the Japanese market. The red line denotes
the spillover from the Japanese market to the Chinese market. The black line is the net spillover from
the Chinese market to the Japanese market.
Volatility spillovers from the Chinese market to the Japanese market varied across different models.
The spillover was between 2% and 18% (win500) and 2% and 11% (win1000), which was less than that
from the Chinese market to the US market or from the Chinese market to the UK market. Volatility
CN and JP, Window=500, Row Normalization CN and JP, Window=1000, Row Normalization
20 12
Affected by Others Affected by Others
Affect Others Affect Others
10
15 Net Affect Others Net Affect Others
10
6
Index (%)
Index (%)
5 4
2
0
-5
-2
-10 -4
Fig. 10. Volatility spillovers between the Chinese and Japanese markets.
spillovers from the US to the Japanese market were in the range of 2% to 10% (win500) and 5% and 10%
(win1000), which was, on an average, larger than that from the Chinese to the Japanese market, although
they remained stable during the sample period. Volatility spillovers from the Chinese to the Japanese mar-
ket remained stable from 2001 to 2006 and began to rise at the end of 2006. The volatility spillover
reached its climax from February to July 2007 and again from August to October 2008 before falling and
remaining stable afterward; this is similar to the volatility spillover of the Chinese market to the UK and
US markets.
The volatility spillover from the Japanese to the Chinese market was concentrated between 2% and 10%
(win500) and 2% and 5% (win1000). The volatility spillover from the UK or the US market to the Japanese
market was in the range of 0% to 4%, which is smaller than that from the Japanese market to the Chinese
market. The volatility spillover from the Japanese market to the US market was in the band of 5% to 15%
(win500) and 5% to 18% (win1000), which is much larger than that from the Japanese to the Chinese
market.
Net volatility spillovers from the Chinese market to the Japanese market varied across different models,
between −5% and 15% (win500) and −2% and 8% (win1000). The index was almost the same as that from
the Chinese market to the US or the UK markets. It remained negative from 2000 to 2006 and climaxed
from February to July 2007 and again from September to October 2008.
As a result, we propose that the Japanese market had more volatility spillover to the Chinese market
than did the US or the UK markets. Moreover, the Chinese market had less volatility spillover to the
Japanese market than to the US or the UK markets. Therefore, the Japanese market had a larger net vola-
tility spillover to the Chinese market than the US or the UK markets to the Chinese market. Similarly, Hu et
al. (1997) revealed that the Japanese and US stock markets had a noticeable impact on the Hong Kong, Tai-
wan, Shanghai, and Shenzhen markets from 1992 to 1996; however, the influence of the US market on
these markets exceeded that of the Japanese market. Nevertheless, using the new sample period of 1996
to 2009 for our study, we conclude that the impact of the Japanese market on the Greater China markets
has been outweighing that of the US market since 1996 owing to financial integration between China
and Japan.
4.3.3.4. China vs. India. Fig. 11 illustrates the volatility spillovers between the Chinese and Indian markets.
The blue line denotes the spillover from the Chinese market to the Indian market. The red line denotes the
spillover from the Indian market to the Chinese market. The black line is the net spillover from the Chinese
market to the Indian market.
Volatility spillovers from the Chinese market to the Indian market were concentrated between 0% and
14% (win500) and 2% and 12% (win1000), which is similar to the spillover from the US market to the Indian
market (0% to 10% and 0% to 8%). The spikes appeared between February and July 2007 and in October 2008.
CN and IN, Window=500, Row Normalization CN and IN, Window=1000, Row Normalization
25 14
Affected by Others Affected by Others
20 Affect Others 12 Affect Others
Net Affect Others Net Affect Others
15
10
10
8
5
Index (%)
Index (%)
6
0
4
-5
2
-10
0
-15
-20 -2
-25 -4
Fig. 11. Volatility spillovers between the Chinese and Indian markets.
The volatility spillover from the Indian market to the Chinese market was between 0% and 10%
(win500) and 2% and 10% (win1000), whereas the spillover from the US market to the Chinese market
was 0% to 4%. Thus, as compared to the US market, the Indian market had more impact on the Chinese
market, indicating the increasing trend of Asian financial integration. On the other hand, the spillover
from the US market to the Indian market was in the range of 0% to 10%, revealing that the US market
had more influence on the Indian market than it did on the Chinese market. We argue that because
India adopted the open-up policy much earlier than China did, a more open Indian market would be
more easily affected by other world markets. In fact, the Indian government implemented open-up poli-
cies for foreign investors in as early as September 1992. Foreign investors were able to enter the first
and secondary markets freely and invest in stocks, bonds, and other financial assets. In 1996, over 600 for-
eign companies had participated in the Indian stock market, and the total investment amounted to 2.6 bil-
lion USD. In contrast, the Chinese stock market has not yet totally opened, and the capital account is still
frozen. Thus, it is not surprising to find a low spillover from the US market to the Chinese market. On the
other hand, the volatility spillover from the Indian market to the US market was in the range of 0% to 10%,
which was no larger than that from the Chinese market to the US market, indicating that the Chinese mar-
ket had a greater volatility impact on the US market in spite of its lack of openness.
Net volatility spillovers from the Chinese market to the Indian markets varied across different models,
around −5% to 10% (win500) and − 4% to 8% (win1000). The index remained negative from 2002 to 2006.
It reached its climax from February to July 2007 before becoming negative again. This suggests that the im-
portance of the Chinese stock market significantly increased after 2005, and that the volatility spikes in
February 2007 may have been transmitted to the Indian market.
4.3.3.5. Shanghai vs. Hong Kong. Fig. 12 illustrates the volatility spillovers between the Shanghai and Hong
Kong markets. The blue line denotes the spillover from the Shanghai market to the Hong Kong market. The
red line denotes the spillover from the Hong Kong market to the Shanghai market. The black line is the net
spillover from the Shanghai market to the Hong Kong market.
Volatility spillovers from the Shanghai market to the Hong Kong market varied across different models,
which were between 0% and 40% (win500) and 2% and 20% (win1000), whereas the spillover from the US
market to the HK market was concentrated in the range of 2% to 8% and 2% to 10%. This indicates that as
compared to the US market, the Chinese market had a greater influence on the Hong Kong market in terms
of volatility spillover. It also demonstrates that the Shanghai and Hong Kong markets were well connected
and that Greater China's financial integration had some tangible effects during the sample period. Similar-
ly, Zheng et al. (2009) showed that volatilities from the Chinese stock market were transmitted to the
Hong Kong market from 1998 to 2009. We furthermore suggest that the spikes of spillovers from the
Shanghai market to the Hong Kong market occurred from February to July 2007 and August to October
2008, which are identical to those from the Chinese market to the other markets.
CN and HK, Window=500, Row Normalization CN and HK, Window=1000, Row Normalization
50 25
40 20
30 15
Index (%)
Index (%)
20 10
10 5
0 0
-10 Affected by Others -5 Affected by Others

Affect Others Affect Others
Net Affect Others Net Affect Others
-20 -10
Fig. 12. Volatility spillovers between the Shanghai and Hong Kong markets.
The volatility from the Hong Kong market to the Shanghai market was in the range of 0% to 15%
(win500) and 0% to 10% (win1000), whereas the spillover from the US and UK markets to the Chinese
market was between 0% and 4% and the spillover from the Japanese market to the Chinese markets was
approximately 2% to 10% and 2% to 5%. This indicates that as compared to the US, the UK, or Japanese mar-
kets, the Hong Kong market had much greater volatility impact on the Chinese markets. Nevertheless, Hu
et al. (1997) proposed that the volatilities of Japanese markets have been transmitted to US markets, and
that neither the Taiwan nor Hong Kong markets had any concrete effect on the Chinese market from 1992
to 1996. Although our data span is 1996 to 2006, we postulate that the Hong Kong market began to influ-
ence the Chinese market significantly after 1996. On the other hand, the spillover from the Hong Kong
market to the US market was in the band of 2% to 10%, which is comparable to that from the Chinese mar-
ket to the US market.
We note that the climax of the spillovers from the Hong Kong market to the Shanghai market occurred
around 2000, indicating that the burst of the IT bubble had an effect on the Chinese market, mainly
through the Hong Kong stock market.
Net volatility spillovers from the Shanghai market to the Hong Kong market were in the range of −10%
to 40% (win500) and −5% to 15% (win1000). The spillovers remained negative from 1998 to 2005, imply-
ing that the Chinese market was greatly affected by the Hong Kong market during the Asian financial crisis.
The index reached its climaxes around February to July 2007 and October 2008.
4.3.3.6. Mainland China vs. Taiwan. Fig. 13 illustrates the volatility spillovers between the Chinese mainland
and Taiwanese markets. The blue line denotes the spillover from the Chinese mainland market to the Taiwanese
market. The red line denotes the spillover from the Taiwanese market to the Chinese mainland market. The
black line is the net spillover from the Chinese mainland market to the Taiwanese market.
Volatility spillovers from the Chinese mainland market to the Taiwanese market varied across different
models. They were between 0% and 20% (win500) and 2% and 14% (win1000), whereas the spillover from
the US market to the Taiwanese market was in the range of 2% to 7% and 2% to 8%. Thus, it seems that the
impact of the Chinese mainland market on the Taiwanese market was no less than that of the US market
on the Taiwanese market, suggesting that the Chinese mainland market played a leading role in the Greater
China stock markets. In addition, we find that the climaxes of the spillover from the Chinese mainland mar-
ket to the Taiwanese market occurred around February to July 2007 and October 2008, which is similar to
the findings in other markets.
The volatility spillover from the Taiwanese market to the Chinese mainland market was in the band of
2% to 15% (win500) and 2% and 8% (win1000), whereas the spillover from the US and UK markets to the
Chinese market was approximately 0% to 4%, indicating that as compared to the US or UK markets, the
Taiwanese market had a much greater influence on the Chinese mainland market. On the other hand,
the spillover from the Taiwanese market to the US market was between 2% and 10% and 2% and 12%,
CN and TW, Window=500, Row Normalization CN and TW, Window=1000, Row Normalization
25 14
Affected by Others
20 12 Affect Others
Net Affect Others
10
15
8
10
6
Index (%)
Index (%)
5
4
0
2
-5
0
-10
-2
Affected by Others
-15 Affect Others -4
Net Affect Others
-20 -6
Fig. 13. Volatility spillovers between the Chinese Mainland and Taiwanese markets.
which is identical to that of the Chinese to the US and UK markets. Considering these results, we could
propose that the integration of financial markets across the Taiwan Straits was so well developed during
the sample period that the connection between the Chinese mainland and Taiwanese markets was more
significant than that between the Taiwanese and the US markets or the Taiwanese and the UK markets.
The spillover from the Taiwanese market to the Chinese mainland market reached its climax around
July 2002 to January 2003, as volatilities caused by the SARS epidemic in the Taiwanese market were
transmitted back to the Chinese mainland market at that time.
The net volatility spillover from the Chinese mainland to the Taiwanese market was between − 10%
and 20% (win500) and − 5% and 10% (win1000). The index remained negative from June 2000 to July
2004, indicating that the Chinese mainland market was greatly affected by the Taiwanese market before
2004. After 2004, the significance of the Chinese mainland market began to emerge gradually. The cli-
maxes of the spillover index occurred around March 2007 and October 2008, implying that the Chinese
mainland market had a major impact on the Taiwanese market during these periods.
4.4. Robustness
In order to test for robustness, we must slightly modify our baseline model to determine if the variation
affects our main results. Thus, we make the following adjustments: (1) the US and European markets indi-
ces are not lagged one day; (2) we calculate the dynamic indices for orders 2 through 5 of VAR; and (3) we
examine the results for forecast horizons varying from 5 to 10 days. In each case, we calculate the spillover
indices from win100, win500, and win1000. Owing to limited space, we cannot display the various plots
depicting the volatility spillover between the Chinese and other markets. However, we can conclude that
our findings are robust to the above variations.
5. Conclusion
Using the generalized vector autoregressive framework of Diebold and Yilmaz (2011b), wherein
forecast-error variance decompositions are invariant to variable ordering, we proposed measures of
total volatility spillover, regional volatility spillover, and directional volatility spillover between the Chi-
nese and world equity markets between February 1996 and December 2009. Specifically, we investigated
the volatility spillover among 11 world markets, the volatility spillover among the Chinese, Greater Chi-
nese, and Asian markets, as well as the volatility spillover between the Chinese and major individual mar-
kets such as stock markets in the US, the UK, Japan, India, Hong Kong, and Taiwan.
In our study, we found that from 1996 to 2009, the Chinese stock market was hardly affected by world
markets in terms of volatility spillover. Prior to 2005, the Chinese market was slightly affected by other
markets. After 2005, the Chinese stock market had a significant volatility spillover effect on other markets,
a net effect that was positive, indicating that the influence of the Chinese stock market was greatly en-
hanced during those years. The volatility interactions among the Chinese, Hong Kong, and Taiwanese mar-
kets were more prominent than were those among the Chinese, Western, and other Asian markets, which
confirmed the effect of the financial market integration in the Greater China region. In general, volatility
spillovers among the Chinese, Japanese, and Indian markets were more distinct than those among the
Chinese, the US, and the UK markets, which implies that the connections and correlations among Asian
stock markets have become increasingly more evident in recent years.
Regardless of whether it affected or was affected by other markets, the US stock market was highly cor-
related with other markets in terms of volatility, especially during the subprime mortgage crisis. Although
other markets were also very volatile, driven by bad news, their massive volatility was transmitted back to
the US market. The major correction of the Shanghai stock market between February and July 2007 signif-
icantly contributed to the volatility surges of other markets. Owing to restrictions on foreign investment,
the Chinese stock market was not greatly affected with respect to market volatility during the recent glob-
al crisis.
For this study, we used stock indices from 11 markets in Asia, Europe, and North America. As our focus
is on the Chinese stock market, we did not consider the stock markets of South America or Australia; this is
left for future investigation. Moreover, the DY 2011 framework could be used to shed light on other inter-
esting issues, such as the spillover between financial markets and the macro economy, spillovers among
different stock markets within one country, spillovers among cross-listed stocks in various countries, and
spillovers with high frequency data.
Acknowledgments
We are very grateful to Xiaohui Hou, Yusen Kwoh, Ray Yeutien Chou, Vic Edwards, the editor and an
anonymous referee for their valuable comments. We appreciate the financial support from the third
phase of Projects “211” and “985” of Xi'an Jiaotong University. All remaining errors are our own.
References
Bae, K.H., Karolyi, G.A., 1994. Good news, bad news and international spillover of stock return volatility between Japan and the US.
Pacific-Basin Finance Journal 2, 405–438.
Bailey, W., 1994. Risk and return on China's new stock markets: some preliminary evidence. Pacific-Basin Finance Journal 2,
243–260.
Bekaert, G., Harvey, C., 1995. Time-varying world market integration. Journal of Finance 50 (2), 403–444.
Bekaert, G., Harvey, C., 1997. Emerging equity market volatility. Journal of Financial Economics 43 (1), 29–77.
Cai, Y., Chou, R.Y., Li, D., 2009. Explaining international stock correlations with CPI fluctuations and market volatility. Journal of Bank-
ing and Finance 33 (11), 2026–2035.
Chi, J., Li, K., Young, M., 2006. Financial integration in East Asian equity markets. Pacific Economic Review 11 (4), 513–526.
Diebold, F.X., Yilmaz, K., 2009. Measuring financial asset return and volatility spillovers, with application to global equity markets.
The Economic Journal 119, 158–171.
Diebold, F.X., Yilmaz, K., 2011a. Equity Market Spillovers in the Americas. In: Alfaro, R. (Ed.), Financial Stability, Monetary Policy, and
Central Banking, vol. 15. Bank of Chile Central Banking Series, Santiago, pp. 199–214.
Diebold, F.X., Yilmaz, K., 2011b. Better to give than to receive: predictive directional measurement of volatility spillovers. Interna-
tional Journal of Forecasting.
Eun, C.S., Shim, S., 1989. International transmission of stock market movements. Journal of Financial and Quantitative Analysis 24,
241–255.
Forbes, K.J., Rigobon, R., 2002. No contagion, only interdependence: measuring stock market comovements. Journal of Finance 57 (5),
2223–2261 (October).
Frank, N., Hesse, H., 2009. Financial spillover to emerging markets during the global financial crisis. IMF Working Paper, WP/09/104.
Groenewold, N., Tang, S.H.K., Wu, Y., 2004. The dynamic interrelationships between the greater China share markets. China Econom-
ic Review 15 (2004), 45–62.
Hamao, Y., Masulis, R.W., Ng, V., 1990. Correlations in price changes and volatility across international stock markets. Review of Fi-
nancial Studies 3, 281–307.
Hu, J.W.-S., Chen, M.-Y., Fok, R.C.W., Huang, B.-N., 1997. Causality in volatility and volatility spillover effects between US, Japan and
four equity markets in the South China Growth Triangular. Journal of International Financial Markets, Institution and Money 7,
351–367.
Janakiramanan, S., Lamba, A.S., 1998. An empirical examination of linkages between Pacific-basin stock markets. Journal of Interna-
tional Financial Markets, Institutions and Money 8 (2), 155–173.
Johansson, A., Ljumgwall, C., 2008. Spillover Effects among the Greater China stock markets. World Development 37 (4), 839–851.
King, M., Sentana, E., Wadhwani, S., 1994. Volatility and links between national stock markets. Econometrica 62, 901–933.
Koop, G., Pesaran, M.H., Potter, S.M., 1996. Impulse response analysis in non-linear multivariate models. Journal of Econometrics 74,
119–147.
Lin, K.-P., Menkveld, A.J., Yang, Z.S., 2009. Chinese and world equity markets: a review of the volatilities and correlations in the first
fifteen years. China Economic Review 20 (1), 29–45.
Melvin, M., Peiers, B., 2003. The global transmission of volatility in the foreign exchange market. The Review of Economics and Sta-
tistics 85 (3), 67–679.
Moon, G.H., Yu, W.C., 2010. Volatility Spillovers between the US and the China Stock Markets: Structural Break Test with Symmetric
and Asymmetric GARCH Approaches. Available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1325002 2010.
Ng, A., 2000. Volatility spillover effects from Japan and the US to the Pacific-Basin. Journal of International Money and Finance 19,
207–233.
Parkinson, M., 1980. The extreme value method for estimating the variance of the rate of return. Journal of Business 53, 61–65.
Pesaran, M.H., Shin, Y., 1998. Generalized impulse response analysis in linear multivariate models. Economics Letters 58, 17–29.
Soriano, P., Climent, F.J., 2006. Volatility Transmission Models: A Survey. Available at http://papers.ssrn.com/soL3/papers.cfm?
abstract_id=929953 2006.
Wang, S.S., Firth, M., 2004. Do bears and bulls swim across oceans? Market information transmission between greater China and the
rest of the world. Journal of International Financial Markets, Institutions and Money 14, 235–254.
Wang, Y., Di Iorio, A., 2007. Are the China-related stock markets segmented with both world and regional stock markets? Journal of
International Financial Markets, Institutions and Money 17 (3), 277–290.
Wongswan, J., 2003. Transmission of information across international equity markets. Federal Reserve-International Finance Discus-
sion Papers, No. 759.
Yilmaz, K., 2010. Return and volatility spillovers among the East Asian equity markets. Journal of Asian Economics 21 (3), 304–313.
Zheng, Y., Chen, H., Wong, W.-K., 2009. China's Stock Market Integration with a Leading Power and a Close Neighbor. Available at
SSRN: http://ssrn.com/abstract=1547665 2009.

Paci Fic-Basin Finance Journal: Xiangyi Zhou, Weijin Zhang, Jie Zhang

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Paci Fic-Basin Finance Journal: Xiangyi Zhou, Weijin Zhang, Jie Zhang

Hochgeladen von

Copyright:

Verfügbare Formate

Paciﬁc-Basin Finance Journal 20 (2012) 247–270

Contents lists available at SciVerse ScienceDirect

Paciﬁc-Basin Finance Journal

Volatility spillovers between the Chinese and world

Article history: We propose measures of the directional volatility spillovers between

3. Data and descriptive statistics

4.1. Non-synchronous trading problem

Fig. 1. Daily stock market volatilities.

Mean 23.21 18.97 21.12 19.53 23.48

Mean 18.72 24.80 17.01 18.76 17.10 17.37

4.2. The full-sample volatility spillover table

Stock exchange Greenwich time Local time

New York 14:30–21:00 9:30–16:00

X. Zhou et al. / Paciﬁc-Basin Finance Journal 20 (2012) 247–270

FA HK SP US TW IN GM KR JP UK CN From others Including own

4.3. Dynamic rolling — sample analysis

4.3.1. Spillovers in regional markets

GlobalSpillindex , Window=500 GlobalSpillindex , Window=1000

Fig. 2. Total volatility spillovers (global markets).

We highlight some major events in these plots:

AisaSpillindex , Window=500 AisaSpillindex , Window=1000

Fig. 3. Total volatility spillovers (Asian markets).

Fig. 4. Total volatility spillovers (Greater China markets).

X. Zhou et al. / Paciﬁc-Basin Finance Journal 20 (2012) 247–270

Fig. 6. Volatility spillovers between the Chinese and Asian markets.

Fig. 8. Volatility spillovers between the Chinese and US markets.

4.3.3. Volatility spillovers between the Chinese and individual markets

Fig. 9. Volatility spillovers between the Chinese and UK markets.

-10 Affected by Others -5 Affected by Others

Das könnte Ihnen auch gefallen