CHINAS FINANCIAL MARKET INTEGRATION

WITH THE WORLD

Anders C. Johansson
Stockholm School of Economics








CERC Working Paper 10
June 2009











Postal address: P.O. Box 6501, S-113 83 Stockholm, Sweden.
Office address: Hollndargatan 30 Telephone: +46 8 736 93 60 Telefax: +46 8 31 30 17
E-mail: [email protected] Internet: http://www.hhs.se/cerc
Chinas Financial Market Integration
with the World
Anders C. Johansson

Stockholm School of Economics

This version: May 2009
Abstract
It is commonly argued that Chinas nancial markets are eectively in-
sulated from the rest of the world. To see if this is true and to better un-
derstand Chinas nancial development, we analyze Chinas integration
with major nancial markets. Using conditional copulas, we show that
China has experienced an increasing level of integration with several ma-
jor nancial markets during the last decade, even though the countrys
nancial markets are commonly seen as being insulated. Furthermore,
the level of integration has increased with several major markets during
the current nancial crisis. The results and possible reasons for the in-
creasing integration are analyzed and the implications for policymakers
and market participants are discussed.
Keywords: China; Financial market integration; Codependence; Cop-
ula
JEL Classication: F30; G15

E-mail: [email protected]. Stockholm School of Economics, P.O. Box 6501, 113

83 Stockholm, Sweden. Financial support from the Bank of Sweden Tercentenary Foun-
dation, the Swedish Foundation for International Cooperation in Research and Higher Ed-
ucation, and the Swedish School of Advanced Asia-Pacic Studies (SSAAPS) is gratefully
acknowledged.
1
1 Introduction
Financial integration has been increasing around the world during the last
decades.
1
Reasons for this include the tendency for investors turning to other
markets for diversication and higher returns as well as nancial liberalization
in a large number of both developed and developing economies. Private capi-
tal ows across national borders have increased fast and developing countries
with their expected high levels of economic growth have been recipients of a
signicant amount of these ows. There are signicant potential benets to
be had as a result of increasing nancial integration for both developed and
developing countries. The possibility of diversication of savings should have
a direct positive eect on individuals wealth. Consumption smoothing is an-
other potential benet (see Obstfeld, 1994, among others). However, many
regional nancial crises and episodes of so-called nancial contagion show that
more open access to international nancial markets also brings with it potential
risks for individuals and the economy as a whole. Even though most nancial
crises are the direct result of misaligned fundamentals (Agenor, 2003), these
crises have shown the importance of having developed strong domestic nan-
cial systems with ecient regulatory overview and supervision. In the case
of China, the government is taking nancial liberalization step by step. For
instance, instead of opening up for cross-border investments immediately, the
country has created programs for qualied domestic investors that are allowed
to invest abroad, as well as qualied foreign investors allowed to invest in the
domestic nancial markets. Another example of controlled cross-border trans-
actions is the fact that the Chinese market has been divided into shares of
dierent classes, where A-shares are for domestic investors, while B-shares are
for foreign (and now also domestic) market participants. Overall, it seems clear
that the Chinese government sees the potential benets of a higher level of -
nancial integration with the rest of the world. The high domestic savings rates
and the reforms of the domestic pension system also create a strong demand
for alternative investments besides those available in the domestic nancial
markets.
The aim of this paper is to analyze Chinas nancial integration with neigh-
boring countries as well as other regions around the world. To better under-
stand how such integration changes over time, we focus on nancial market
integration, and more specically integration between the Chinese equity mar-
ket and other equity markets. One of the novelties in this paper is the focus
on Chinas nancial integration with nearby markets as well as regions. An-
other, and perhaps more important novelty is the new approach we use to study
dependence between dierent markets. To analyze Chinas nancial market in-
tegration with the rest of the world, we use both unconditional and conditional
1
For an interesting discussion on nancial globalization over the past two decades, see
Moshirian (2008). Also, for a discussion on regional economic and nancial integration in
Asia and Chinas role in that process, see Moshirian (2009).
2
copulas. Copulas have recently become popular in the nance literature due
to a number of reasons. It can be argued that copulas are more informative
measures of dependence between variables than standard linear correlation.
This is the case when the joint distribution of the variables is nonelliptical
and the typical linear correlation measure is not enough to describe the de-
pendence between the variables (Patton, 2006). Copulas have been used to
study the asymmetric nature of dependence between nancial variables (e.g.
Patton, 2006) as well as contagion (Rodriguez, 2007). The empirical results
in this paper indicate that China is moving from a very low level of nancial
market integration. The last two decades show signs of how Chinas nan-
cial markets are becoming increasingly dependent on markets in the rest of
the world. A number of reasons for why this development is taking place are
discussed, including the gradual opening up of the domestic markets to the
outside world, cross-listings, the internationalization of Chinese rms, and the
general economic integration around the world.
The rest of the paper is organized as follows: Section 2 discusses related
literature on nancial integration and the Chinese nancial markets. Section 3
introduces the methodology of copulas and how it is used to study the depen-
dence structure among nancial variables. Section 4 goes over the empirical
results. Section 5 discusses possible reasons for Chinas changing nancial
market integration with the rest of the world, and Section 6 concludes the
paper.
2 Literature Review
The literature on market integration is rich and includes both theoretical and
empirical studies. Important studies include Solnik (1974), Stehle (1977), Stulz
(1981), Jorion and Schwartz (1986), Eun and Janakiramanan (1986), Harvey
(1991), Sellin and Werner (1993), Bailey and Jagtiani (1994), Stulz and Wasser-
fallen (1995), and Carrieri, Errunza, and Hogan (2007). How can we dene
nancial market integration? Following Bekaert and Harvey (1995), markets
can be said to be fully integrated if assets with identical risk have identical
expected returns regardless of in which market they are traded. Early research
looked at how nancial integration is taking place over time and across borders
in both developed and developing parts of the world. Bekaert and Harveys
(1995) study on a large number of developed countries and twelve emerging
markets showed that some emerging markets seem to be more integrated with
other markets even though there are strong investment restrictions in place.
On the other hand, their study indicates that other emerging markets are seg-
mented, even though foreign investors have relatively free access to investment
opportunities. Their results thus indicate the possibility of increasing nancial
market integration over time even when a market appears to be quite insu-
lated from the rest of the worlds markets. Their ndings are interesting and
of signicant important to this study, since we focus on a market that is com-
3
monly seen as being very much insulated from the rest of the world. There are
also a number of studies that look at the gradual nature of market integration.
Bekaert (1995) analyzes regulatory changes and market integration. His results
indicate that set dates with regulatory changes do not signicantly relate to
market integration, indicating that integration is gradual in nature. Similarly,
Bekaert, Harvey, and Lumsdaine (2002) identify endogenous break dates in
the level of market integration and show that these dates do not necessarily
correspond well with dates of ocial market reforms.
In terms of nancial integration in general, a common argument in research
on nancial integration in East Asia is that most of the integration taking place
is not so much on a regional level, but rather on a global level. For instance,
Eichengreen and Park (2005) show that East Asian countries have stronger
nancial ties with Western Europe and the U.S. than neighboring countries in
the region. Joen, Oh and Yang (2005), looking at nancial market data, nd
support for this argument. As a latecomer on the stage of regional nancial
markets, China constitutes an interesting case. This paper therefore looks at
how the Chinese stock market relates not only to important markets close by,
such as Hong Kong and Japan, but also to the rest of Asia, Europe, and the
U.S.
There is a growing number of studies that focus on Chinas nancial rela-
tionship with neighboring economies. Shin and Sohn (2006) analyze trade and
nancial integration in Asia and include China in their sample. To identify
possible regional nancial integration, they focus on co-movements in interest
rates using standard correlation measures. They nd that the current level
of nancial integration in the region is low and that it is dicult to identify
the possible eects that nancial integration has on macroeconomic variables
such as output and consumption. In a study on spillover eects among nan-
cial markets, Johansson and Ljungwall (2009) focus on the equity markets in
the so-called Greater China region (China, Taiwan, and Hong Kong). They
show that there are spillover eects between the markets and that the Chinese
market is aected by its neighbors even though it is commonly seen as an in-
sulated market. Cheung, Chinn and Fujii (2003) study nancial integration in
the Greater China region. Using data on interbank rates, exchange rates, and
prices, they show that there is an increasing trend in nancial integration in the
region, especially between China and Hong Kong. In a related study, Cheung,
Chinn and Fujii (2005) analyze nancial integration in Greater China and how
its economies are integrated with the rest of the world. Their ndings indicate
that the money markets in the Greater China region are increasingly linked
to the worlds money markets, but that the banking sector on the mainland
remains insulated from banking sectors in other countries. Finally, Wang and
Di Iorio (2007) show that there is an increasing integration between Chinas
A-share market and Hong Kongs stock market, but that there is no evidence
that the Chinese A-share market is becoming more integrated with the world
market.
4
A number of reasons for why Asias nancial integration is lagging behind
that of regions such as Europe have been discussed in the literature on regional
nancial integration. When it comes to China, Masson, Dobson and Lafrance
(2008) argue that China is less nancially integrated with the rest of the world,
even though the country is becoming more and more important for the world
economy. The authors list a number of reasons for why this may change over
the coming decades, including Chinas increasing investments abroad, the con-
tinued development of the domestic nancial system, a development towards
a more exible exchange rate system, and the possibility of a continuation of
the relaxation of capital controls. These reasons for anticipating Chinas in-
creasing level of nancial integration with the rest of the world are important,
as is a more detailed analysis of why nancial integration is already increasing
as seen in this study. This is why a separate section in this paper focuses on
possible reasons for why Chinas nancial integration with the rest of the world
is increasing and why this trend is likely to continue.
3 Methodology
In this section, we introduce the concept of copulas and extend it to a con-
ditional specication of co-dependence. The discussion is kept brief and most
of the content is based on Nelson (1999) and Patton (2006). The theory of
copulas allows us to decompose a joint distribution of two or more variables
into their marginal distributions and the dependence function, i.e. the copula.
In this study, we use bivariate copulas and will therefore focus the following
discussion on a bivariate setting. It should be noted that the theory is easily
extended to a multivariate setting. Following the original work on copulas by
Sklar (1959) and the explanation by Patton (2006), we can decompose the
joint distribution into the marginal distributions:
F
XY
(x, y) = C(F
X
(x), F
Y
(y)), (1)
where F
XY
is the cumulative distribution function and F
X
(x) and F
Y
(y) are
the marginal distributions, respectively. If we instead use density functions,
the relationship can be written as:
f
xy
(x, y) = f
x
(x) f
y
(y) c(F
x
(x)F
y
(y)). (2)
If the distribution function F
XY
(x, y) is a continuous multivariate distribution
function, Sklars theorem shows that we can separate the marginal distribution
for the two variables from that of the dependence structure. There are several
reasons for why copulas in many situations are preferable to other measures of
dependence (see, for example, Rodriguez, 2007, for an excellent discussion on
the statistical properties of copulas). Here, it suces to say that copulas are
advantageous compared to another commonly used measure of dependence,
namely correlation. While correlation only captures the linear relationship be-
tween two variables, copulas contain all the information about the dependence
5
between them. As Rodriguez (2007) points out, copulas are especially useful
when it comes to analyzing the information about the tails of the joint distri-
bution of the variables in question. Carrieri, Errunza, and Hogan (2007) show
that correlation is not an appropriate measure of nancial market integration
because it underestimates the degree of integration. Since copulas measures
the full dependence between variables, we believe that our approach is better
suited to look at nancial market integration.
Patton (2006) developed the concept of copulas to allow for a time-varying
behavior of the dependence structure between variables. This concept is per-
haps especially useful when analyzing economic and nancial variables, as it
allows us to understand how the dependence structure evolves over time. Let
us assume that there is a conditioning variable, W, and that its dimension
is 1. Patton (2006) denes a conditional copula of (X, Y )|W = w, where
X|W = w F
X|W
(|w) and Y |W = w F
Y |W
(|w) is the conditional joint
distribution function of U F
X|W
(X|w) and V F
Y |W
(Y |w), given W = w.
Here, the variables U and V are the so-called probability integral transforms
of X and Y given W. The probability integral transforms U and V have a
Unif(0, 1) distribution (Fisher, 1932, and Rosenblatt, 1952). According to
Patton (2006), conditional copulas have the same properties as unconditional
copulas. He also shows how to extend Sklars original theorem to the condi-
tional case.
There are many ways to parameterize conditional copulas (for dierent
approaches, see Patton, 2006, and Rodriguez, 2006). Here, we follow the
approach in Patton (2006). Focusing on the conditional normal (Gaussian)
copula and considering the bivariate setting, we can write the dependence
function as:
C(u, v|) =
_

1
(u)

_

1
(v)

1
2
_
(1
2
)
exp
_
(r
2
2rs + s
2
)
2(1
2
)
_
drds, (3)
where 1 < < 1. In equation (3),
1
is the inverse of the standard
normal cumulative distribution function. Patton (2006) suggests the following
evolution equation for
t
:

t
=

t1
+
1
10
10

j=1

1
(u
tj
)
1
(v
tj
)
_
. (4)
To keep
t
in (1, 1) at all times,

is the modied logistic transformation,
dened as:

1 e
x
1 + e
x
= tanh
_
x
2
_
. (5)
The second part in the parenthesis on the right-hand side of equation (4)
is included to allow for persistence in the dependence parameter. The third
part is meant to capture the variation in the parameter and is composed of the
6
mean of the product of the last 10 observations of the inverse of the cumulative
distribution functions of u and v.
In practice, there are three basic steps involved when one wants to obtain
the dependence structure between two variables. First, the two marginal distri-
butions are modeled using a suitable specication. Then, a probability integral
transform is used to convert the standardized residuals of the marginal distri-
butions into Unif(0, 1) distributions. Finally, the uniform distributions are
used to obtain the dependence structure between the two variables.
2
. Since
we are focusing on the dependence structure between pairs of equity mar-
kets, a natural choice for the marginal distributions are models that take het-
eroscedastic features of most nancial markets into account. We choose to
model the conditional volatility in the dierent markets using three possible
GARCH specications: the standard GARCH developed by Bollerslev (1986);
the threshold, or TGARCH model, developed by Glosten, Jaganathan, and
Runkle (1993); and the exponential, or EGARCH, model developed by Nelson
(1991). We apply all three specications to all markets and choose the most
suitable specication for each market based on Akaikes information criterion.
The three specications can be described as follows:
1. GARCH

2
t
= +
q

j=1

2
tj
+
p

i=1

2
ti
. (6)
Here,
2
t
is the conditional variance,
2
t1
are previous periods forecast vari-
ances, and
2
ti
are the news about volatility from previous periods, measured
as the squared residuals from the mean specication.
2. TGARCH

2
t
= +
q

j=1

2
tj
+
p

i=1

2
ti
+
r

k=1

2
tk
I

tk
, (7)
where I

t
= 1 if
t
< 0 and zero otherwise. This means that bad news, i.e.

ti
< 0 are allowed to have a dierent impact compared to good news, or

ti
> 0. Asymmetric features in volatility exists if
i
= 0.
2. EGARCH
log(
2
t
) = +
q

j=1

j
log(
2
tj
) +
p

i=1

ti

ti

+
r

k=1

tk

tk
. (8)
Again, asymmetric features in volatility exists if
i
= 0. This specication
allows for the leverage eect to be exponential, rather than quadratic as in the
TGARCH specication above.
2
I am grateful to Andrew Patton for generously providing a Matlab program for the
estimation of unconditional and conditional copulas.
7
To allow for thicker tails commonly observed in the distribution of nancial
time series, we use a Generalized Error Distribution (GED) instead of a normal
distribution when modeling and estimating the marginal distributions. The
GED was popularized in nancial time series analysis by Nelson (1991). A
random variable x
t
has a GED with zero mean, unit variance and tail parameter
(where > 0) if its probability density function can be written as:
g(x
t
) =
exp
_
(1/2)|x
t
/|

2
(+1)/
(1/)
, (9)
where
=

2
2/
(1/)
(3/)
. (10)
When the GED parameter is equal to 2, the probability density function
reduces to that of the standard normal distribution. If < 2, the density
has fatter tails than the normal density. Finally, if > 2, the density has
thinner tails than the normal density. We thus expect to see values for the
GED parameter to be lower than 2 for several of the markets.
To complete the specication of the marginal distributions, we also need
to model the mean. Initial tests indicate that a few of the markets exhibit
signicant serial correlation in returns. To fully model the marginal distribu-
tions, we therefore adjust the mean specication to allow for autoregressive
behavior when needed. When no serial correlation is found, only a constant is
used when we estimate the mean equation. Having gone over the basic princi-
ples behind standard and conditional copulas as well as the estimation of the
marginal distributions, we now turn to the empirical results.
4 Empirical Results and Analysis
Before analyzing the empirical results, we introduce the data used in this study.
As a proxy for the Chinese stock market, we use the Shanghai Composite Index
(SHCI). The SHCI is arguably the most commonly used index when studying
Chinas equity markets. The S&P 500 index and the Nikkei are used for the
U.S. and Japan markets, respectively. Similarly, the Hang Seng Index is used as
a proxy for the Hong Kong stock market. In order for us to be able to analyze
Chinas nancial market integration with Asia and Europe over time, the MSCI
Asia (excluding Japan) and the MSCI Europe indices are used as proxies for
the two regions. Finally, the MSCI World index is used as a proxy for the
world market. All indices are gathered from Datastream and are of monthly
frequency. The log returns in percentages are computed as 100 log(P
t
/P
t1
),
where P
t
is the price of the index at time t.
As discussed in the previous section, before the full dependence structure
can be analyzed, we need to model the marginal distributions for each of the
8
variables in the data sample. We test each of the three GARCH specications
discussed earlier in order to identify the most suitable model for each of the
seven stock indices.
The results of the estimations of each of the dierent marginal distribu-
tions are presented in Table 1. To save space, we only include the estimation
results for the optimal specication for each market. The AIC indicates that
a standard GARCH(1,1) is preferable for all time series, except for Japan,
for which an EGARCH(1,1) specication is used. Several of the series show
signs of serial correlation. Suitable lagged values of the returns are therefore
included in the mean equation for China, Asia, and Hong Kong. The Ljung-
Box statistic indicates that the serial correlation in the standardized residuals
from the estimations is insignicant in all of the seven cases. We can there-
fore conclude that we have dealt adequately with autocorrelation in the mean.
Similarly, the Ljung-Box statistic for the variance equations shows that there
are no signs of residual heteroscedasticity in any of the squared standardized
residuals. The conditional variance thus seems to be correctly specied in all
cases. The parameter for lagged volatility in the variance equation, , indi-
cates that volatility is very persistent in all seven cases, a common feature in
international equity markets. In the case of Japan, the leverage parameter is
very signicant. However, the parameter is positive, indicating that positive
news have a stronger eect than negative news. This result diers from the
leverage eect, for which negative news have a stronger eect on subsequent
volatility than positive news. The GED parameter is signicant in all seven
estimations. is lower than 2 in all cases except for Japan, indicating that
most of the series have distributions with thicker tails compared to the stan-
dard normal distribution. To conclude, the mean and variance specications
seem to be correctly specied in all seven cases.
[TABLE 1 HERE]
Having estimated each of the marginal distributions, we then apply the proba-
bility integral transform to convert the standardized residuals of the marginal
distributions into Unif(0, 1) distributions. To check that the specications
for the marginal distributions are correct, we also apply a test suggested by
Diebold, Gunther, and Tay (1998). If the marginal distributions are speci-
ed correctly, the probability integral transform of the residuals will be i.i.d.
Unif(0, 1). Diebold, Gunther, and Tay (1998) focus on the rst four moments
of the transformed series, or (u
t
u)
k
, where u
t
is the transformed series and
k = 1, 2, 3, 4. We therefore create time series for the four moments for each
of the seven transformed series and look at their autocorrelation functions. If
the marginal distribution is incorrectly specied for the monthly returns se-
ries, then the autocorrelation function will reveal dependence in the dierent
moments of the transformed time series. An overview of the autocorrelation
function for each of the four transformed series and their four moments is pre-
sented in Table 2. The table shows the autocorrelation for 5, 10, 15, and 20
9
lags, respectively. None of the probability integral transforms have any remain-
ing dependence in the rst four moments. The specied marginal distributions
thus seem to capture the volatility dynamics of the processes as well as serial
correlation in the mean. This means that we can use the transformed series
when estimating the dependence structure between China and the other six
markets.
[TABLE 2 HERE]
Knowing that we have modeled the marginal distributions correctly, we now
turn to the estimation of the copulas. We start with unconditional copulas
and look at the dependence structure between Chinas nancial market and
each of the six other markets. Table 3 shows the dependence parameter when
applying the normal copula. The rst row shows the level of dependence for
the whole period. The next three rows present the dependence structure when
we divide the sample into three subperiods of equal length. The dependence
between China and each of the six markets is relatively modest for the whole
sample period, an expected result if one considers how insulated the Chinese
equity markets have been from the outside world. However, what is more
interesting is the change over time in each of the six relationships. Looking at
the dependence parameter and comparing between the dierent subperiods, it
is clear that there is an increase in nancial market integration between China
and all six markets. The dependence is very low and even negative during the
rst subperiod, and then increases signicantly during the two next subperiods.
These results indicate that Chinas nancial markets have gone from being very
much insulated and unaected by the outside world to becoming increasingly
integrated with world markets over time.
[TABLE 3 HERE]
To better understand how dependence between the markets behave over time,
the conditional copula approach is used on the six pairs of variables. Figure
1 shows the results when we apply a conditional normal copula to China and
each of the six other markets. The time paths of the conditional dependence
parameters support the initial results in Table 3. Chinas nancial integra-
tion with the six markets is increasing over time. However, the movements
vary signicantly depending on market. There was a general decrease in the
integration between China and Asia and between China and Hong Kong fol-
lowing the initial increase in market volatility during the Asian nancial crisis
in 1997-1998. The eect of the Asian nancial crisis is much less pronounced
in the dependence between China and Europe, China and Japan, China and
the U.S., and China and the world. These results indicate that China was not
as severely hit by the nancial crisis as other countries in the region. Further-
more, the gure indicates that the current nancial crisis that originated in the
U.S. is causing a general increase in the dependence between China and other
nancial markets. The only market that China is not experiencing a signicant
increase in integration with during the last two years is that of Japan.
10
[FIGURE 1 HERE]
Besides the apparent increases in dependence between China and dierent in-
ternational nancial markets, it is worth noting that the increase in nancial
integration seems to be especially strong between China and a few particular
markets. The relationship between China and Hong Kong is getting stronger,
an expected result that will be discussed in more detail in the following sec-
tion. The dependence between China and Asia is also increasing signicantly,
indicating a strong trend in nancial integration. Finally, and somewhat sur-
prisingly, the conditional dependence between China and Europe seems to be
increasing steadily over time, especially during the period 2002-2008. The -
nancial market integration taking place between China and Japan seems more
stable by comparison.
5 Causes behind the Increasing Financial In-
tegration
The general view of Chinas nancial markets is that they are almost com-
pletely insulated with a very limited relationship to international nancial
markets. To some extent, this view is correct. For instance, Johansson (2009b)
shows that systematic risk in the two Chinese stock markets (Shanghai and
Shenzhen) when measured against the world market is limited. The results
indicate that there are a number of sharp and short-lived increases in system-
atic risk in these markets during the last decades and connect these changes to
volatility increases due to shifts in government policy rather than a changing
correlation pattern with the world market. However, as we have argued earlier
in this paper, correlation is only one way of measuring dependence, and in some
cases, there are alternatives that may be better suited to measure dependence
between two nancial markets.
The empirical results in the previous section highlight how the nancial
market integration between China and the rest of the world tends to increase
over time during the 1990s and 2000s. Why do we come to this conclusion even
though the markets have been so closed to the outside world? In this section
we discuss a number of possible explanations, ranging from factors that can
have had an impact early on to factors that tend to grow in importance as
the Chinese economy as well as its nancial markets mature. We begin with
possible causes that are directly related to the nancial markets, including
dierent share classes, cross listings across national borders and a continued
gradual nancial liberalization. We then discuss more general economic causes,
often a result of Chinas growing importance in the world economy. These
factors include the worlds increasing reliance on China as a manufacturing
powerhouse, the internationalization of Chinese rms and how globalization
results in similar eects on rm value regardless of where companies are listed
11
or have their respective headquarters.
3
5.1 Share Classes
As mentioned earlier, due to capital account restrictions in China, it is com-
monly argued that the domestic stock market is insulated from the rest of the
world. There are, however, several reasons for why Chinas stock market can
be expected to be aected by shocks in international markets. The rst reason
we discuss here concerns that of share classes. Chinas two stock exchanges in
Shanghai and Shenzhen were established in 1990 and 1991, respectively. Be-
ginning in 1992, the Chinese Securities and Regulatory Commission (CSRC)
allowed domestic companies to issue what was called Special Shares. The
standard A-shares have been available to domestic investors, while foreign in-
vestors have invested in the Special Shares, later to be called B-shares. To facil-
itate trading in B-shares for foreign investors, these shares are denominated in
U.S. dollars in Shanghai and in Hong Kong dollars in Shenzhen. Even though
the dierent share classes have resulted in a seemingly segmented domestic
stock market, it is natural to assume that there exists at least an indirect
causal relationship between Chinese and foreign markets due to the possibility
for international investors to trade in B-shares.
As an example, let us assume that there is a market-wide shock in China.
This will aect both A- and B-shares simultaneously. The shock will then prop-
agate into other markets through the eect it has on international investment
portfolios when the value of B-shares changes. We can therefore assume that
a domestic market-wide shock has indirect eects on international stock port-
folios and thus indirect eects on other markets. As another example, assume
that there is a sudden shock in international markets that aects B-shares
in China. For instance, a shock in international markets may result in the
repositioning in international investment portfolios, thus changing the overall
demand for Chinese B-shares. Even though A- and B-shares are commonly
seen as segmented markets, the Chinese government has imposed changes in
the regulations that have weakened this segmentation. In 2001, it was decided
that Chinese investors are allowed to invest in B-shares as well as A-shares.
This change, put into eect on February 19, 2001, had a dramatic eect on the
relationship between A- and B-shares. There was a sharp decline in the dis-
count on B-shares as a result of the regulatory change, from an average of 75%
down to 8% in just a few months (Karolyi and Li, 2003). Chan, Menkveld, and
Yang (2007) analyze how information discovery is taking place in the Chinese
stock market. Their ndings indicate that before the regulatory change, the
3
As Masson, Dobson, and Lafrance (2008) point out, the potential move away from a
xed exchange rate regime is most likely to increase future nancial integration between
China and the rest of the world signicantly. The issue of Chinas exchange rate regime and
its eect on the domestic economy has been discussed in numerous studies. We therefore do
not analyze the potential eects on the countrys future nancial integration if a change in
the exchange rate regime takes place further here.
12
A-share market lead the B-share market in information discovery. After 2001,
the causality was reversed. Wang and Di Iorio (2007) analyze the integration
between A- and B-shares, Hong Kong, and the world market. Their results
suggest that there is an increasing integration between A- and B-shares. It
can therefore be assumed that changes originating in international markets
can have an indirect eect on both A- and B-shares in China.
5.2 Cross Listings
Besides B-shares, foreign investors are able to invest in Chinese companies that
are cross-listed, that is, they are listed in China as well as abroad. The per-
haps strongest reason for why companies decide to list their shares in another
country is that they seek to exploit the existence of a lower cost of capital.
In a recent study on Chinese cross-listings, Zhang and King (2008) nd that
limitations of the domestic markets is a strong motivating factor for Chinese
companies to list abroad. Their results also indicate that Chinese companies
are motivated to raise capital in foreign equity markets since they provide
a better access to capital markets and thus provide a lower cost of capital.
Also, Zhang and King (2008) argue that the dierent legal and economic en-
vironments play a role in the decision to list abroad.
4
There are a number of
dierent venues for Chinese companies to list in, but the most common one
is still that of Hong Kong. Shares of companies that are based in China and
choose to list in Hong Kong are often called H-shares. There are also so-called
red chips listed in Hong Kong. Red chips are shares of companies that have
the major part of their business located in China but decide to list abroad for
dierent reasons. Historically, cross-listed shares tend to be the largest and
most successful mainland companies, with a higher level of transparency and
better corporate governance than the average mainland company.
The fact that many of the companies that are listed abroad are also listed
domestically combined with the fact that these companies tend to be the largest
and most important ones in the Chinese economy makes it natural to assume
that they create a direct relationship between the Chinese market and interna-
tional markets. A recent study by Johansson and Ljungwall (2009) show that
there are linkages between the markets in the so-called Greater China region.
Cross-listings is most probably one of the reasons for the spillover eects found
in that study. Wang and Di Iorios (2007) analysis on market integration in-
dicates that the A-share and Hong Kong markets are getting more integrated
over time. The authors credit this development to the increased economic in-
tegration taking place between Hong Kong and the mainland. The signicant
number of cross-listings and the increasingly complex relationship between the
markets in Hong Kong and the mainland constitute plausible reasons for the
increasing relationship between Chinas equity markets and world markets.
4
For a more detailed and thought-provoking analysis on why rms decide to list in other
countries, see Karolyi (2006).
13
5.3 Financial Liberalization
Since the beginning of the economic reforms in the end of the 1970s, nan-
cial reforms have taken place gradually. In response to the growing needs of
well-functioning nancial markets as the real economy grows, nancial reforms
have gained momentum during the last decade. Lee and Wong (2009) di-
vides the nancial reforms in China into two separate phases. The rst phase,
spanning from 1978 to 1997, involved the remaking of the domestic banking
system, the establishment of the two stock exchanges, the development of the
foreign exchange market, and the gradual development of the domestic bond
market. During the second phase, from 1997 and forward, the focus has been
on the deregulation of the domestic banking sector, general improvements of
the nancial markets, and increasing the integration with the global nancial
system.
During the second phase of nancial reforms, the Chinese government has
initiated policy changes that should have signicant eects on the relation-
ship between Chinese and global nancial markets. Besides letting domestic
investors trade in B-shares, the government now allows Chinese rms to open
foreign exchange accounts. The implementation of rst the Qualied Foreign
Institutional Investor (QFII) program and later also the Qualied Domestic
Institutional Investor (QDII) program is also more than likely to have a signif-
icant eect on the level of nancial integration between China and the rest of
the World. The QFII scheme was introduced in 2002 as a way to let a number
of experienced foreign institutional investors get access to the Chinese domes-
tic markets. The program sets a quota for how much the foreign institutions
are allowed to invest, thus so far limiting the impact on the domestic market.
Nevertheless, the movement towards more open domestic markets is clear. The
QDII program was introduced in 2006 and originally covered only xed-income
products. However, the program was later expanded to cover equity products
as well. There are macroeconomic structural reasons as well as eects for indi-
vidual domestic investors that indicate the importance of the QDII program.
The hot topic of Chinas external imbalances with the resulting inux of foreign
liquidity and the pressure on the domestic currency to appreciate are obvious
signs of the need for a more open capital account. Similarly, the possibility
for domestic investors to invest in a much broader range of assets can have a
very signicant long-term wealth eect. As of 2006, regulations from the State
Administration of Foreign Exchange (SAFE) state that Chinese individuals
are allowed to exchange up to $20,0000 or the equivalent of foreign currency
every year.
It is clear that the second phase of nancial liberalization in China will
have eects on several levels. In terms of nancial market integration, it can
be assumed that the gradual opening of the Chinese capital account will result
in an increase in the integration between the domestic nancial markets and
the world market. The empirical results in this study support this assumption,
with evidence of continuous increases in nancial market dependence between
14
China and many of the major international markets since the mid-1990s.
5.4 Chinas Increasing Trade and FDI Inows
During the last decades, China has gone from playing a marginal role in the
world economy into becoming one of the most important global economic
forces. A long period of unprecedented growth in foreign direct investment
(FDI) has resulted in strong capital inows as well as dierent spill-over ef-
fects to the Chinese economy. Another phenomenon in the Chinese economy,
partly a result of the increasing FDI inows and the presence of foreign compa-
nies in the country, is the very strong and persistent increase in exports. The
combination of the increases in FDI inows and exports means that the domes-
tic economy is becoming more and more intertwined with the world economy,
or at least with the economies with which China is engaged in signicant trade
and the economies from which China is receiving the major parts of its FDI
inows.
Chinas trade with the rest of the world has not only increased dramati-
cally, with an export growth of over 500% since 1992 (Amiti and Freund, 2008).
The trading partners and the composition of the trade is also of considerable
interest. Chinas economic development is often compared to that of India. In
terms of trade, the two large economies are very dierent, however, as pointed
out in several studies (e.g. Bussi`ere and Mehl, 2008). For instance, while
Chinas share in total global trade has increased to 7% in 2006, Indias share
has remained at a modest 1%. Another and perhaps more important dierence
is the trade linkages that have developed between these two countries and the
rest of the world. Bussi`ere and Mehl (2008) show that China is responsible
for a signicant share of total foreign trade in goods in the U.S., the U.K,
Japan, and the euro area. Similarly, Chinas general trade intensity is higher
as are its bilateral trade linkages. Bussi`ere and Mehl (2008) argue that this is
largely because China today constitutes an integral part of the Asian produc-
tion chain, a commonly used term for the well-developed regional production
network. A clear sign of this is that a large part of Chinas total imports
come from neighboring countries in the region, while no such pattern can be
found for India. Overall, Chinas trade links are one of the main causes for the
countrys phenomenal growth during the last decades.
The increase in economic ties between China and the rest of the world is
likely to result in increased nancial integration with global markets. As a very
large number of domestic rms are dependent on contracts with multinational
rms, adverse shocks to the demand for foreign rms goods abroad can have
an indirect but equally signicant impact on a number of domestic rms. For
instance, if a company that is listed on a foreign stock exchange suddenly is
aected by negative sentiments in global consumer markets, a company listed
on one of the Chinese stock exchanges that produce and sell products to the
foreign company is negatively aected as well. In this very simple example, a
15
negative impact on the market value of the foreign rm would be followed by
a similar loss in market value of the Chinese rm. It is natural to assume that
the development of the Chinese economy into the worlds manufacturing center
results in an indirect increase in nancial market integration, even though the
causal eect is dicult to measure in absolute terms.
5.5 The Internationalization of Chinese Firms
The topic of the current and future internationalization of Chinese rms is
indeed both sensitive and intriguing. Chinas rst sovereign wealth fund han-
dled by China Investment Corporation and several other state-owned enti-
ties responsible for investing abroad have made many international observers
uneasy. However, Chinas outward foreign direct investment (OFDI) is still
modest in size. Also, a majority of the investments overseas have focused on
mining and manufacturing, with a 65% and 25% share of total OFDI, respec-
tively (Sch uler-Zhou and Sch uller, 2009). Sch uler-Zhou and Sch uller argue
that the current patterns of Chinas OFDI indicate that investment choices
are heavily inuenced by the government. Leaving the political dimensions
out of the analysis, the focus of the following discussion is on the eects that
Chinese rms interest abroad may have on nancial integration.
Even though the size of OFDI from China up to now is relatively modest,
the trend is clear. The Chinese government has formulated a policy that spells
out the desire to create a number of national corporate champions that will be
able to compete globally. One of the rst examples of the expansion of Chi-
nese rms abroad is Lenovo, the computer manufacturer that acquired IBMs
laptop division in 2004. Besides straightforward acquisitions of brand names
such as in the case of Lenovo and IBM, several of Chinas largest companies
are increasingly turning outwards in their pursuit to expand their operations.
For instance, Boisot and Meyer (2009) point out that one reason for why
an increasing number of Chinese companies that are relatively small by in-
ternational standards choose to internationalize early in their development is
that these companies face increasing costs due to protectionism and inecient
domestic logistics. A natural next step is therefore to internationalize their
operations in the pursuit of more ecient institutions. Examples of large Chi-
nese companies that have already internationalized their operations include
the white-goods manufacturer Haier, the telecom solutions provider Huawei,
and the steel company Baosteel.
As Chinese companies expand across national borders and regions, they
start to compete with multinational companies from other countries. Even
though this holds true for the domestic arena in China, limited access to the
Chinese market in many industries as well as the limited purchasing power
in the country up to recently have to some extent inhibited such competition
inside China. The current internationalization of Chinese rms thus marks
a new turning point for the countrys integration with the world economy.
16
As Chinese companies begin to compete in international markets, they are
more likely to be aected by similar shocks as their international competitors.
Take the telecom solutions industry as an example. As Huawei is becoming
a serious contender to multinational companies like Ericsson and Nokia, these
companies are bound to be aected by changes in the global demand for their
products and services. This means that a severe adverse demand shock in a
large market can be expected to result in a decrease in rm value not only for
the competing international companies but also for the Chinese counterpart.
This development should hold true for a number of sectors in which Chinese
companies has put extra energy into trying to build up an international pres-
ence. For instance, China was a net importer of steel for a long period of
time. Recently, this has changed, as large Chinese steel companies are com-
peting globally with other multinational steel companies. This development
indicates that Chinese companies and other multinational companies will be
aected similarly by changes in market demand. This is perhaps particularly
true when it comes to the raw material sector, an industry where Chinese
companies are currently expanding aggressively around the world. Overall,
the internationalization of Chinese rms is most likely to result in stronger co-
movements between them and international rms, thereby strengthening the
relationship between the domestic and international nancial markets.
6 Conclusion
Studies on nancial integration show that a countrys nancial markets tend
to become more integrated with international markets as nancial liberaliza-
tion is initiated. Financial integration has important implications not only for
domestic and foreign investors, but also for policy makers who need to under-
stand the factors that inuence movements in nancial markets. In this study,
we focus on Chinas nancial market integration with a number of important
international markets. We limit the analysis to equity markets, partly because
the access to reliable data, and partly because the Chinese equity market has
received signicant attention both domestically and abroad. We propose the
use of copulas when measuring nancial integration, since it has been shown
that copulas are able to capture dependence that is not identied by regular
linear correlation measures. We therefore rst model and estimate the marginal
distributions for each of the seven equity markets using conditional volatility
models. We then use the residuals from the marginal estimation to obtain
the copulas. Using both unconditional and conditional copulas, we are able to
show that Chinas nancial market is becoming increasingly integrated with a
number of the major international markets included in the study. Furthermore,
the trend of increasing nancial integration has continued during the current
nancial crisis that originated in the U.S. The support for increased nancial
integration begs the question why such a development is taking place.
We provide a number of plausible explanations for why China is becoming
17
more nancially integrated with the world. Beginning with structural features
in the nancial markets, we argue that the changing nature of the before so
segmented share classes is resulting in closer ties between A-, B-, H-shares
and thereby with shares in other markets. Another possible reason for the
increasing nancial integration is that more and more Chinese companies are
listed both in domestic markets and abroad. One result of this phenomenon
may well be an increasing integration with international markets, since inter-
national investors get access to Chinese companies. Sceptics may argue that
these markets are perfectly segmented. However, we believe that investor sen-
timents inuence movements in these markets similarly. For instance, if there
are prolonged negative sentiments in the mainland markets, this is most likely
to cause a similar drop in cross-listed shares, thus inuencing both domestic
and international investment portfolios. Furthermore, the implementation of
the QFII and QDII programs brings with it new channels for integration in the
sense that both domestic and international investors are getting access to the
same markets. It can be expected that these gradual steps of nancial liberal-
ization have caused and will continue to cause signicant increases in Chinas
nancial integration with the rest of the world. Besides the issues related to
the structure of the nancial markets and the nancial reforms, we also discuss
additional plausible reasons for why Chinas markets is becoming more inte-
grated with the international markets. Trade and FDI are two such reasons,
constituting channels for increasing economic ties across national borders and
thus an increasing sensitivity for shocks in dierent markets. Similarly, we ar-
gue that the ongoing internationalization of Chinese rms, albeit at a modest
level so far, may constitute a signicant factor behind the current, and not the
least future level of nancial integration.
To conclude, the empirical results in this study indicate that nancial inte-
gration is increasing between China and several major international markets.
Furthermore, a number of possible reasons for the increase in nancial integra-
tion discussed in this paper indicate that nancial integration between China
and the rest of the world is likely to continue to increase as long as the Chinese
government continues its gradual reforms. Increasing nancial integration has
strong implications for both investors and policy makers. As domestic investors
are getting access to international nancial markets, they need to understand
how these markets relate to the domestic markets in order to construct op-
timal portfolios. Similarly, foreign investors need to understand how Chinese
markets are aected by shocks in other markets, and whether the importance
of these eects will increase. For policy makers, it is important to understand
how the data-generating process in the domestic nancial markets is changing.
When the two stock exchanges in Shanghai and Shenzhen were established in
the early 1990s, they were in many ways truly insulated from international
markets. Policy makers in China were thus able to focus on domestic matters
and domestic causes of market instability. As nancial integration between
China and the rest of the world increases, however, these policy makers are
18
facing an increasingly complex situation in which not only domestic shocks af-
fect the market. Now, they also need to understand how shocks and negative
sentiment in markets such as the U.S., Europe, and the rest of Asia propagate
to the Chinese stock market.
19
References
[1] Agenor, P.-R., 2003. Benets and Costs of International Financial Integra-
tion: Theory and Facts, World Economy, 26, 1089-1118.
[2] Amiti, M. & Freund, C., 2008. The Anatomy of Chinas Export Growth,
World Bank Policy Research Working Paper, No. 4628.
[3] Bailey, W. & Jagtiani, J., 1994. Foreign Ownership Restrictions and Pre-
miums for Intemational Investment: Some Evidence from the Thai Capital
Market, Journal of Financial Economics, 36, 57-88.
[4] Bekaert, G., 1995. Market Integration and Investment Barriers in Emerging
Equity Markets, World Bank Economic Review, 9, 75-107.
[5] Bekaert, G. & Harvey, C.R., 1995. Time-Varying World Market Integration,
Journal of Finance, 50, 403-44.
[6] Bekaert, G., Harvey, C.R. & Lumsdaine, R.L., 2002. The Dynamics of
Emerging Market Equity Flows, Journal of International Money and Fi-
nance, 21, 295-350.
[7] Boisot, M. & Meyer, M.W., 2008. Which Way through the Open Door?
Reections on the Internationalization of Chinese Firms, Management and
Organization Review, 4, 349-65.
[8] Bollerslev, T., 1986. Generalized Autoregressive Conditional Heteroskedas-
ticity, Journal of Econometrics, 31, 307-27.
[9] Bussi`ere, M. & Mehl, A., 2008. Chinas and Indias Roles in Global Trade
and Finance: Twin Titans for the New Millenium?, European Central Bank
Occasional Paper Series, No. 80.
[10] Carrieri, F., Errunza, V. & Hogan, K. 2007. Characterizing World Market
Integration through Time, Journal of Financial and Quantitative Analysis,
42, 915-940.
[11] Chan, K., Menkveld, A.J. & Yang, Z. 2007. The Informativeness of Do-
mestic and Foreign Investors Stock Trades: Evidence from the Perfectly
Segmented Chinese Market, Journal of Financial Markets, 10, 391-415.
[12] Cheung, Y.-W., Chinn, M.D. & Fujii, E., 2003. China, Hong Kong, and
Taiwan: A Quantitative Assessment of Real and Financial Integration,
China Economic Review, 14, 281-303.
[13] Cheung, Y.-W., Chinn, M.D. & Fujii, E., 2005. Dimensions of Financial
Integration in Greater China: Money Markets, Banks, and Policy Eects,
International Journal of Finance and Economics, 10, 117-32.
20
[14] Diebold, F.X., Gunther, T.A. & Tay, A.S., 1998. Evaluating Density Fore-
casts with Applications to Financial Risk Management, International Eco-
nomic Review, 39, 863-83.
[15] Sellin, P. & Werner, I., 1993. International Barriers in General Equilib-
rium, Journal of International Economics, 34, 2107-138.
[16] Eichengreen, B. & Park, Y.C., 2005. Financial Liberalization and Capital
Market Integration in East Asia, In Park, Y.C., Ito, T., & Wang, Y. (Eds.),
A New Financial Market Structure for East Asia, Cheltenham: Edward
Elgar.
[17] Eun, C.S. & Janakiramanan, S., 1986. A Model of International Asset
Pricing with a Constraint on the Foreign Equity Ownership, Journal of
Finance, 41, 897-914.
[18] Fisher, R.A., 1932. Statistical Methods for Research Workers, Edinburgh:
Oliver and Boyd.
[19] Glosten, L.R., Jaganathan, R. & Runkle, D., 1993. On the Relation be-
tween the Expected Value and the Volatility of the Normal Excess Return
on Stocks, Journal of Finance, 48, 1779-1801.
[20] Harvey, C., 1991. The World Price of Covariance Risk, Journal of Finance,
46, 111-57.
[21] Joen, J., Oh, Y. & Yang, Y., 2005. Financial Market Integration in East
Asia: Regional or Global?, Korea Institute for International Economic Pol-
icy.
[22] Johansson, A.C., 2009. An Analysis of Dynamic Risk in Greater Chinas
Equity Markets, Journal of Chinese Economic and Business Studies, forth-
coming.
[23] Johansson, A.C. & Ljungwall, C., 2009. Spillover Eects among the
Greater China Stock Markets, World Development, 37, 839-51
[24] Jorion, P. & Schwartz, E., 1986. Integration Versus Segmentation in the
Canadian Stock Market, Journal of Finance, 41, 603-13.
[25] Karolyi, G.A., 2006. The World of Cross-Listings and Cross-Listings of the
World: Challenging Conventional Wisdom, Review of Finance, 10, 99-152.
[26] Karolyi, G.A. & Li, L., 2003. A Resolution of the Chinese Discount Puzzle,
Unpublished Manuscript.
[27] Lee, J. & Wong, A., 2009. Impact of Financial Liberalization on Stock
Market Liquidity: Experience of China, Hong Kong Monetary Authority
Working Paper 03/2009.
21
[28] Masson, P., Dobson, W. & Lafrance, R., 2008. Chinas Integration into
the Global Financial System, Bank of Canada Review, Summer 2008, 17-29.
[29] Moshirian, F., 2008. Globalisation, Growth, and Institutions, Journal of
Banking and Finance, 32, 472-79.
[30] Moshirian, F., 2009. Can an Asia Pacic Community, Similar to the Eu-
ropean Community, Emerge?, Journal of Banking and Finance, 33, 2-8.
[31] Nelson, D., 1991. Conditional Heteroskedasticity in Asset Returns: A New
Approach, Econometrica, 59, 347-70.
[32] Obstfeld, M., 1994. Risk Taking, Global Diversication, and Growth,
American Economic Review, 84, 1310-29.
[33] Patton, A., 2006. Modeling Asymmetric Exchange Rate Dependence, In-
ternational Economic Review, 47, 527-56.
[34] Rodriguez, J.C., 2007. Measuring Financial Contagion: A Copula Ap-
proach, Journal of Empirical Finance, 14, 401-23.
[35] Rosenblatt, M., 1952. Remarks on Multivariate Transformation, The An-
nals of Mathematical Statistics, 23, 470-72.
[36] Shin, K. & Sohn, C.-H., 2006. Trade and Financial Integration in East
Asia: Eects on Co-Movements, World Economy, 29, 1649-69.
[37] Sch uler-Zhou, Y. & Sch uller, M., 2009. The Internationalization of Chi-
nese Companies: What Do Ocial Statistics Tell Us about the Chinese
Outward Foreign Direct Investment?, Chinese Management Studies, forth-
coming.
[38] Sklar, A., 1959. Fonctions de repartition `a n dimensions et leurs marges,
Publications de lInstitut Statistique de lUniversite de Paris, 8, 229-31.
[39] Solnik, B., 1974. The International Pricing of Risk: An Empirical Inves-
tigation of the World Capital Market Structure, Journal of Finance, 29,
48-54.
[40] Stehle, R., 1977. An Empirical Test of the Alternative Hypothesis of Na-
tional and International Pricing of Risky Assets, Journal of Finance, 32,
493-502.
[41] Stulz, R.M., 1981. On the Eects of Barriers to International Investment,
Journal of Finance, 36, 923-34.
[42] Stulz, R.M. & Wasserfallen, W., 1995. Foreign Equity Investment Re-
strictions, Capital Flight, and Shareholder Wealth Maximization, Review of
Financial Studies, 8, 1019-58.
22
[43] Wang, Y. & Di Iorio, A., 2007. Are the China-Related Stock Markets
Segmented with both World and Regional Stock Markets?, Journal of In-
ternational Financial Markets, Institutions and Money, 17, 277-90.
[44] Zhang, C. & King, T.-H.D., 2008. The Decision to Cross-List: The Case
of Chinese IPOs and ADRs, Unpublished Manuscript.
23
T
a
b
l
e
1
:
M
a
r
g
i
n
a
l
D
i
s
t
r
i
b
u
t
i
o
n
E
s
t
i
m
a
t
i
o
n
s
C
h
i
n
a
A
s
i
a
E
u
r
o
p
e
H
o
n
g
K
o
n
g
J
a
p
a
n
U
S
A
W
o
r
l
d
M
o
d
e
l
A
R
(
2
)
-
G
A
R
C
H
(
1
,
1
)
A
R
(
2
,
7
)
-
G
A
R
C
H
(
1
,
1
)
G
A
R
C
H
(
1
,
1
)
A
R
(
7
)
-
G
A
R
C
H
(
1
,
1
)
E
G
A
R
C
H
(
1
,
1
)
G
A
R
C
H
(
1
,
1
)
G
A
R
C
H
(
1
,
1
)
M
e
a
n
E
q
u
a
t
i
o
n
c
0
.
4
2
0
0
.
7
1
1
*
0
.
9
9
0
*
*
*
1
.
0
6
9
*
*
-
0
.
6
4
2
0
.
8
9
9
*
*
*
0
.
8
5
4
*
*
*
(
0
.
4
8
9
)
(
0
.
3
8
5
)
(
0
.
2
5
1
)
(
0
.
4
2
2
)
(
0
.
4
1
6
)
(
0
.
2
4
2
)
(
0
.
2
4
4
)
r
i
,
t

2
0
.
1
6
3
*
*
*
0
.
1
2
3
*
(
0
.
0
5
9
)
(
0
.
0
6
9
)
r
i
,
t

7
0
.
1
2
8
*
0
.
1
7
0
*
*
*
(
0
.
0
7
3
)
(
0
.
0
7
0
)
V
a
r
i
a
n
c
e
E
q
u
a
t
i
o
n

3
.
7
9
8
2
.
1
9
1
0
.
9
3
4
2
.
2
8
4
0
.
3
4
7
*
*
*
0
.
5
1
3
0
.
7
7
4
(
3
.
1
8
7
)
(
1
.
7
5
9
)
(
1
.
0
0
5
)
(
1
.
9
9
9
)
(
0
.
0
8
5
)
(
0
.
5
5
3
)
(
0
.
8
6
6
)

0
.
1
1
3
0
.
1
5
4
*
*
0
.
1
8
4
*
*
*
0
.
1
7
8
*
*
-
0
.
1
3
8
*
*
*
0
.
1
9
5
*
*
0
.
1
5
9
*
*
*
(
0
.
0
7
6
)
(
0
.
0
6
7
)
(
0
.
0
7
0
)
(
0
.
0
6
9
)
(
0
.
0
3
4
)
(
0
.
0
8
0
)
(
0
.
0
5
9
)

0
.
8
6
9
*
*
*
0
.
8
1
0
*
*
*
0
.
7
9
8
*
*
*
0
.
8
1
1
*
*
*
-
0
.
1
3
2
*
*
*
0
.
8
0
4
*
*
*
0
.
8
1
4
*
*
*
(
0
.
0
6
9
)
(
0
.
0
9
5
)
(
0
.
0
7
7
)
(
0
.
0
7
5
)
(
0
.
0
3
4
)
(
0
.
0
7
8
)
(
0
.
0
7
9
)

0
.
9
3
4
*
*
*
(
0
.
0
1
9
)
G
E
D
p
a
r
a
m
e
t
e
r
1
.
0
4
0
*
*
*
1
.
5
0
0
*
*
*
1
.
3
8
5
*
*
*
1
.
2
8
5
*
*
*
2
.
0
5
2
*
*
*
1
.
4
5
9
*
*
*
1
.
4
6
0
*
*
*
(
0
.
0
8
2
)
(
0
.
2
1
2
)
(
0
.
1
8
6
)
(
0
.
1
8
1
)
(
0
.
3
2
1
)
(
0
.
1
7
2
)
(
0
.
1
6
7
)
L
B
(
1
0
)
9
.
6
0
6
7
.
9
2
7
3
.
6
9
3
5
.
4
9
2
6
.
0
7
8
7
.
7
8
6
3
.
6
1
5
[
0
.
4
7
6
]
[
0
.
6
3
6
]
[
0
.
9
6
0
]
[
0
.
8
5
6
]
[
0
.
8
0
9
]
[
0
.
6
5
0
]
[
0
.
9
6
3
]
L
B
2
(
1
0
)
0
.
2
4
9
1
3
.
2
2
5
2
.
9
9
8
8
.
1
5
3
3
.
8
5
9
1
.
5
7
6
1
.
6
7
9
[
1
.
0
0
0
]
[
0
.
2
1
1
]
[
0
.
9
8
1
]
[
0
.
6
1
4
]
[
0
.
9
5
3
]
[
0
.
9
9
9
]
[
0
.
9
9
8
]
N
o
t
e
:
*
,
*
*
,
a
n
d
*
*
*
i
n
d
i
c
a
t
e
s
t
a
t
i
s
t
i
c
a
l
s
i
g
n
i

c
a
n
c
e
a
t
t
h
e
1
0
%
,
5
%
,
a
n
d
1
%
l
e
v
e
l
,
r
e
s
p
e
c
t
i
v
e
l
y
.
L
B
(
1
0
)
a
n
d
L
B
2
(
1
0
)
r
e
p
r
e
s
e
n
t
L
j
u
n
g
-
B
o
x
s
t
a
t
i
s
t
i
c
s
f
o
r
s
e
r
i
a
l
c
o
r
r
e
l
a
t
i
o
n
u
p
t
o
t
h
e
1
0
t
h
o
r
d
e
r
f
o
r
t
h
e
s
t
a
n
d
a
r
d
i
z
e
d
r
e
s
i
d
u
a
l
s
a
n
d
s
t
a
n
d
a
r
d
i
z
e
d
s
q
u
a
r
e
d
r
e
s
i
d
u
a
l
s
,
r
e
s
p
e
c
t
i
v
e
l
y
.
S
t
a
n
d
a
r
d
e
r
r
o
r
s
a
r
e
g
i
v
e
n
i
n
p
a
r
e
n
t
h
e
s
e
s
.
N
u
m
b
e
r
s
i
n
b
r
a
c
k
e
t
s
f
o
r
t
h
e
L
j
u
n
g
-
B
o
x
s
t
a
t
i
s
t
i
c
s
a
r
e
p
-
v
a
l
u
e
s
.
24
T
a
b
l
e
2
:
A
u
t
o
c
o
r
r
e
l
a
t
i
o
n
-
P
r
o
b
a
b
i
l
i
t
y
I
n
t
e
g
r
a
l
T
r
a
n
s
f
o
r
m
s
C
h
i
n
a
A
s
i
a
E
u
r
o
p
e
H
o
n
g
K
o
n
g
A
C
(
l
)
1
2
3
4
1
2
3
4
1
2
3
4
1
2
5
0
.
0
3
2
-
0
.
0
2
8
0
.
0
5
0
-
0
.
0
4
8
-
0
.
0
0
9
-
0
.
0
3
8
0
.
0
4
2
-
0
.
0
4
2
-
0
.
0
3
6
-
0
.
0
3
8
0
.
0
1
8
-
0
.
0
2
9
-
0
.
0
2
7
0
.
0
1
2
(
0
.
9
2
5
)
(
0
.
8
4
4
)
(
0
.
8
5
6
)
(
0
.
7
0
3
)
(
0
.
5
0
1
)
(
0
.
4
6
3
)
(
0
.
5
0
0
)
(
0
.
1
5
6
)
(
0
.
8
1
6
)
(
0
.
9
1
2
)
(
0
.
9
7
2
)
(
0
.
9
8
9
)
(
0
.
9
2
1
)
(
0
.
1
7
9
)
1
0
0
.
0
4
7
-
0
.
0
4
1
0
.
0
4
2
0
.
0
4
7
0
.
0
6
0
0
.
0
3
4
0
.
0
3
1
-
0
.
0
5
0
0
.
0
7
3
-
0
.
0
5
0
0
.
0
7
8
0
.
0
3
9
0
.
0
6
8
-
0
.
0
0
9
(
0
.
7
5
3
)
(
0
.
6
0
0
)
(
0
.
7
9
8
)
(
0
.
0
8
8
)
(
0
.
8
0
3
)
(
0
.
0
5
0
)
(
0
.
7
7
2
)
(
0
.
9
1
3
)
(
0
.
7
2
3
)
(
0
.
9
2
3
)
(
0
.
9
6
4
)
(
0
.
1
4
3
)
(
0
.
8
4
9
)
(
0
.
3
1
6
)
1
5
-
0
.
0
1
5
0
.
0
8
0
-
0
.
0
5
4
-
0
.
0
2
1
-
0
.
0
6
6
-
0
.
0
2
7
-
0
.
0
5
3
0
.
1
0
1
-
0
.
0
9
4
0
.
1
0
8
-
0
.
0
0
7
-
0
.
0
1
9
-
0
.
0
2
7
-
0
.
0
3
1
(
0
.
9
1
2
)
(
0
.
4
8
5
)
(
0
.
9
2
1
)
(
0
.
3
0
0
)
(
0
.
8
8
5
)
(
0
.
1
8
5
)
(
0
.
8
9
3
)
(
0
.
8
9
4
)
(
0
.
7
9
3
)
(
0
.
9
3
1
)
(
0
.
9
2
6
)
(
0
.
1
9
0
)
(
0
.
8
3
8
)
(
0
.
4
4
8
)
2
0
0
.
0
2
1
-
0
.
0
0
6
-
0
.
1
2
2
-
0
.
0
9
0
-
0
.
0
6
9
-
0
.
0
6
5
-
0
.
0
1
8
-
0
.
1
1
6
0
.
0
0
9
-
0
.
0
8
0
-
0
.
0
9
3
-
0
.
0
5
0
-
0
.
0
5
7
-
0
.
0
9
9
(
0
.
8
6
8
)
(
0
.
6
9
1
)
(
0
.
7
9
7
)
(
0
.
3
2
7
)
(
0
.
8
4
9
)
(
0
.
2
3
6
)
(
0
.
9
4
3
)
(
0
.
5
5
2
)
(
0
.
8
9
3
)
(
0
.
9
1
1
)
(
0
.
8
2
0
)
(
0
.
2
1
9
)
(
0
.
8
0
2
)
(
0
.
4
8
2
)
H
o
n
g
K
o
n
g
J
a
p
a
n
U
S
A
W
o
r
l
d
A
C
(
l
)
3
4
1
2
3
4
1
2
3
4
1
2
3
4
5
-
0
.
0
0
3
0
.
0
2
2
0
.
1
0
1
0
.
0
5
6
0
.
0
8
2
0
.
0
4
4
0
.
1
0
1
0
.
0
2
8
0
.
1
1
1
0
.
0
1
1
0
.
0
5
7
0
.
0
0
7
0
.
0
7
5
-
0
.
0
1
5
(
0
.
7
1
4
)
(
0
.
3
0
7
)
(
0
.
6
5
9
)
(
0
.
4
1
2
)
(
0
.
4
8
4
)
(
0
.
3
5
2
)
(
0
.
8
0
3
)
(
0
.
3
9
6
)
(
0
.
6
4
6
)
(
0
.
3
2
8
)
(
0
.
8
6
2
)
(
0
.
8
8
6
)
(
0
.
6
9
7
)
(
0
.
8
4
6
)
1
0
0
.
0
8
6
-
0
.
0
4
5
-
0
.
0
2
3
0
.
0
0
0
0
.
0
8
0
0
.
0
5
7
0
.
0
0
2
0
.
0
7
8
0
.
0
2
8
0
.
0
6
1
0
.
0
0
3
0
.
0
2
7
0
.
0
4
6
0
.
0
4
1
(
0
.
6
1
5
)
(
0
.
7
8
0
)
(
0
.
8
3
5
)
(
0
.
7
6
2
)
(
0
.
6
7
8
)
(
0
.
5
6
5
)
(
0
.
5
8
4
)
(
0
.
5
2
1
)
(
0
.
8
0
3
)
(
0
.
5
8
1
)
(
0
.
7
6
6
)
(
0
.
9
7
4
)
(
0
.
8
6
8
)
(
0
.
9
7
8
)
1
5
-
0
.
0
3
5
0
.
0
8
0
-
0
.
0
1
5
0
.
0
5
1
0
.
0
0
8
0
.
0
4
7
-
0
.
1
1
8
-
0
.
0
0
6
-
0
.
0
7
2
-
0
.
0
4
0
-
0
.
0
8
8
0
.
0
6
0
-
0
.
1
0
8
0
.
0
7
6
(
0
.
6
8
6
)
(
0
.
6
6
2
)
(
0
.
9
2
5
)
(
0
.
8
8
5
)
(
0
.
7
7
3
)
(
0
.
7
3
0
)
(
0
.
5
5
9
)
(
0
.
7
8
3
)
(
0
.
9
3
1
)
(
0
.
8
5
5
)
(
0
.
8
4
0
)
(
0
.
9
9
4
)
(
0
.
9
0
2
)
(
0
.
9
9
2
)
2
0
0
.
0
5
0
-
0
.
0
0
8
0
.
0
0
6
0
.
0
9
4
0
.
0
0
2
0
.
1
0
8
-
0
.
0
7
0
0
.
0
3
8
0
.
0
1
0
0
.
0
1
5
-
0
.
0
9
2
0
.
0
3
9
-
0
.
0
3
3
0
.
0
2
5
(
0
.
8
0
0
)
(
0
.
7
7
6
)
(
0
.
9
1
9
)
(
0
.
8
9
0
)
(
0
.
9
2
0
)
(
0
.
7
3
7
)
(
0
.
5
1
9
)
(
0
.
8
9
1
)
(
0
.
9
6
5
)
(
0
.
8
7
1
)
(
0
.
8
6
2
)
(
0
.
9
9
8
)
(
0
.
9
6
1
)
(
0
.
9
9
8
)
N
o
t
e
:
T
h
e
t
a
b
l
e
r
e
p
o
r
t
s
a
u
t
o
c
o
r
r
e
l
a
t
i
o
n
f
u
n
c
t
i
o
n
s
f
o
r
(
u
t

u
)
k
,
w
h
e
r
e
k
=
1
,
2
,
3
,
4
.
A
C
(
l
)
s
t
a
n
d
s
f
o
r
a
u
t
o
c
o
r
r
e
l
a
t
i
o
n
u
p
t
o
l
a
g
l
.
p
-
v
a
l
u
e
s
a
r
e
g
i
v
e
n
i
n
p
a
r
e
n
t
h
e
s
e
s
.
25
Table 3: Unconditional Dependence with China
Asia Europe Hong Kong Japan USA World
Whole Period 0.173 0.145 0.157 0.127 0.139 0.148
Period 1 -0.111 -0.210 -0.155 -0.112 -0.139 -0.181
Period 2 0.072 0.030 0.133 0.214 0.086 0.068
Period 3 0.513 0.513 0.479 0.286 0.428 0.485
26
F
i
g
u
r
e
1
:
C
o
n
d
i
t
i
o
n
a
l
D
e
p
e
n
d
e
n
c
e
-
.
4
-
.
2
.
0
.
2
.
4
.
6
.
8
9
4
9
6
9
8
0
0
0
2
0
4
0
6
0
8
A
s
i
a
-
.
4
-
.
2
.
0
.
2
.
4
.
6
.
8
9
4
9
6
9
8
0
0
0
2
0
4
0
6
0
8
E
u
r
o
p
e
-
.
4
-
.
2
.
0
.
2
.
4
.
6
.
8
9
4
9
6
9
8
0
0
0
2
0
4
0
6
0
8
H
o
n
g

K
o
n
g
-
.
4
-
.
2
.
0
.
2
.
4
.
6
.
8
9
4
9
6
9
8
0
0
0
2
0
4
0
6
0
8
J
a
p
a
n
-
.
4
-
.
2
.
0
.
2
.
4
.
6
.
8
9
4
9
6
9
8
0
0
0
2
0
4
0
6
0
8
U
S
A
-
.
4
-
.
2
.
0
.
2
.
4
.
6
.
8
9
4
9
6
9
8
0
0
0
2
0
4
0
6
0
8
W
o
r
l
d
27