The Forecasting Ability of the Chinese Stock Market and the U.S. Stock Market on Each Other

Full text

The Forecasting Ability of the Chinese Stock Market
and the U.S. Stock Market on Each Other
Chujun Zhou1,
1 Department of Statistical Science, University College London, Gower Street London, WC1E 6BT
* zcakhoa@ucl.ac.uk
ABSTRACT
As the largest stock markets in the world, the relationship between the Chinese stock market and U.S. stock market
have always been the hottest topic. If the U.S. stock market can predict and provide additional forecasting information
for the Chinese stock market beyond that contained in Chinese economic variables, investors should incorporate the
U.S. stock market variable into their information set to enhance the accuracy of their return forecasts. This paper
mainly incorporates the data after China joined WTO because Chinese economy developed much faster than before
after it joined WTO. By employing OLS and cointegration tests, this analysis presents evidence of the increasing
interaction between the Chinese and the U.S. stock markets. This finding can be valuable to investors in a way that
investors shall incorporate the information in the other market to make the right decision about investment in the
domestic market.
Keywords: Chinese stock market, U.S. stock market, WTO, mutual interaction, OLS model
1. INTRODUCTION
With the rapid development of international capitals,
the macroeconomic relations among different countries
in the world have become closer, and the development
of macroeconomics has also shown a significant
correlation. The interconnectedness between the
capitalist markets of various countries has increased as
the global economy has become more integrated. With
the integration of international finance, stock markets in
different countries are showing a similar trend of rising
or falling. The stock market in different countries have a
similar trend in some specific circumstances. The
fluctuations of market yields and price in different
countries have a correlation which result in the stock's
prices in different market following a relatively stable
trend.
Investors all over the world pay more attention to the
linkage of the international stock markets, especially the
stock markets between China and the United States. The
United States, the world's largest market, has a big
influence in the world capital market. The United States
and China have been the largest trading partner for
many years. In addition, China's capital market has
opened to the rest of the globe, and local and
international economic linkages are close. China’s
capital market has become one of the most important
part of the global capital market.
Chinese central government have attached great
importance to the financial sector after 2015–2016
Chinese stock market turbulence. Public pension plans
continue to shift into the stock market. Internet has
penetrated many aspects of people's lives. More and
more Chinese people have investment in the stock
market [1]. Understanding the linkage between Chinese
and U.S. stock markets is important. It can help
investors to analyze investment portfolios and the
structure of the securities market thereby identifying
market trends and increasing investment portfolio
returns. Meanwhile, it can also help listed companies to
develop strategies and achieve internationalization of
capital markets.
To narrow the topic down and make it more specific,
this paper will focus on the years after China joined
World Trade Organizations (WTO). After China joined
WTO in the end of 2001, China became fully integrated
into the organization over the next several years.
Because US is the world’s largest economy and is one
of China’s largest trading partners, it is reasonable to
use the U.S. stock market as a proxy for global
economic activity. Hence, it would be of interest to
study the relationship between the two markets in the
years around financial crisis.
Advances in Economics, Business and Management Research, volume 648
Proceedings of the 2022 7th International Conference on Financial Innovation and Economic Development (ICFIED 2022)
Copyright © 2022 The Authors. Published by Atlantis Press International B.V.
This is an open access article distributed under the CC BY-NC 4.0 license -http://creativecommons.org/licenses/by-nc/4.0/. 433

2. LITERATURE REVIEW
2.1. U.S. stock market history
The development history of the U.S. stock market
has four historical periods in the past 200 years. The
first stage is from the end of the 18th century to 1886,
the U.S. stock market initially developed. The first stock
exchange, the Philadelphia Stock Exchange, was
founded in 1790. The US stock market was rapidly
developed from 1886 to 1929. The third stage is from
the Great Depression in 1929 to 1954.
2.2. Chinese stock market history
The first off-exchange trading counter was
established in Industrial and Commercial Bank of China
by Shanghai International Trust & Investment
Corporation on September 26, 1986. The first securities
company was established in Shenzhen Special
Economic Zone on September 27, 1987. On November
26, 1990, the Shanghai Stock Exchange (SSE) was
formally established. Shenzhen Stock Exchange trial
opened on December 1, 1990. Shenzhen Stock
Exchange trial opened on December 1, 1990. After that,
the first trading day of the Shanghai Stock Exchange
was opened on December 19, 1990. In July 1991, the
Shenzhen Stock Exchange officially opened.
2.3. Comparison of the correlation between
Chinese and U.S. stock markets
One of the most important characteristics of the
modern financial system is economic integration and
financial globalization. The stock price is not only
affected by domestic information, but also by spillover
effects due to the transmission of international
information. Many researchers studied the information
the linkage and relationship between Chinese and U.S.
Stock Markets. Cheng and Glascock did not find the
evidence of long-term relationship between the Chinese
and US stock markets, but they found that the
correlation between the two markets has increased after
the financial crisis of 2007-2008 [2]. Qiu and Ye
analyzed the data of SSE index and the S & P 500 and
use multifractal detrended cross-correlation analysis
(MF-DCCA) method to analyze the interaction between
Chinese and U.S. stock markets [3]. It was found that
the interaction between Chinese and U.S. stock markets
during the financial crisis. The U.S. stock market has
shown an increasing spillover effect to the Chinese
stock market. Bissoondoyal-Bheenick et al.,
investigated the stock market volatility spillover among
countries: U.S., China, and Australia. They found that
there is a bilateral causality relationship between these
countries at the market index level from 2007 to 2016
[4]. Zhang et al. built a time varying parameter vector
auto-regression (TVP-VAR) model to investigate the
dynamic spillover effects between U.S. stock volatility
and China's stock market crash risk [5]. Their results
indicated that China's stock market crash risk increases
as the U.S. stock volatility rises. Furthermore, China's
stock market crash risk will be exacerbated by the
country's improvement in financial market openness,
short-term capital outflows, and RMB depreciation.
3. MODELING
The research focuses on the forecasting ability of the
opening of stock price in Chinese stock market and the
U.S stock market on each other. My hypothesis is that
the opening of stock price in Chinese stock market can
predict the daily return in the U.S stock market and vice
versa. To test the hypothesis, some other control
variables need to be included in the model. I will
include interest rate and currency exchange rate in my
model. Therefore, my model looks like this:
𝑃
1𝑡 = 𝛽0+ 𝛽1∗ 𝑃2𝑡 + 𝛽2∗ 𝐼𝑖+𝐶𝐸𝑖+ 𝜇
(1)
where P1 represents the opening stock price in one
market and P2 represents the opening stock price in the
other market. I represent the interest rate and CE stands
for currency exchange rate. I will run two regressions:
one tests the forecasting ability of the U.S. stock market
on Chinese stock market and the other tests the
forecasting ability of the Chinese stock market on the
U.S. stock market.
The model this paper was going to use is Ordinary
Least Square (OLS) model. Since the data displays
time-series characteristics, it will use Dicky Fuller test
to test whether the variables are unit root. To test
whether there is a difference between the forecasting
ability on the first Friday of each month, on which the
job report get announced, it will also include a Friday
dummy variable. Additionally, it will test the out-of-
sample forecasting ability of the model.With the rapid
development of international capitals, the
macroeconomic relations among different countries in
the world have become closer, and the development of
macroeconomics has also shown a significant
correlation. The interconnectedness between the
capitalist markets of various countries has increased as
the global economy has become more integrated. With
the integration of international finance, stock markets in
different countries are showing a similar trend of rising
or falling. The stock market in different countries have a
similar trend in some specific circumstances. The
fluctuations of market yields and price in different
countries have a correlation which result in the stock's
prices in different market following a relatively stable
trend.
Advances in Economics, Business and Management Research, volume 648
434

4. DATA
This paper analyzed stock returns predictability of
the Chinese, stock returns predictability of the Chinese
stock market and the U.S. stock market from 2002 to
2012, especially after China joined WTO and after the
recent financial crisis occurred. This paper used
Shanghai composite index to represent the Chinese
stock market. The economic reforms started in 1978 led
to the rebirth of the stock markets in China. The
Shanghai Stock Exchange and the Shenzhen Stock
Exchange are the two major emerging capital markets in
China. The Shanghai Stock Market was officially
opened in 1990 and the Shenzhen Stock Market in 1991.
Two types of stocks are traded in the two markets: class
A and class B. Class A shares are restricted to Chinese
citizens and denominated in Chinese currency yuan or
Renminbi (RBM), while class B shares can be bought
and sold only by foreigners and are settled in foreign
currencies (US dollars for Shanghai, Hong Kong dollars
for Shenzhen). By the end of 1995, there were 135
companies and 161 stocks listed on the Shenzhen Stock
Exchange. Regardless of some inevitable difficulties in
its infancy stage, the rapid development of China's
capital markets has generated interest among academics,
investors and regulators. In this research, I focus the
analysis on the relationship of share A markets with four
other markets. There are at least two reasons for
examining class A share markets. First, the class B
shares market has been losing its appeal to foreign
investors while the class A share market dominates that
of class B shares in terms of the number of listed
companies, trade volume and market capitalization.
Second, it allows us to address an interesting issue: how
the volatility of a market, which is largely closed to
foreign investors, is related to the volatility of foreign
markets —— the U.S. market. It got the data of
Shanghai composite index on Shanghai stock exchanges
This paper use Standard & Poor's 500 (S&P 500) to
stand for the U.S. stock market. It is an American stock
market index based on the market capitalizations of 500
large companies having common stock listed on the
New York Stock Exchange or NASDAQ Stock
Exchange. The S&P 500 is widely used as a measure of
the general level of stock prices, as it includes both
growth stocks and value stocks. It got the data of S&P
500 on Yahoo finance.
The ideal data contains all the opening price of the
Shanghai Composite Index and S&P 500, interest rate
and exchange rate from 2002 to 2012. However, due to
different dates for holidays in China and in the U.S.,
some data does not match. Any data that is missing in
one country will lead the whole data set containing
stock return, interest rate, and currency exchange rate to
be dropped. A table that describes the data is attached.
The missing data will cause some breaks in the time
interval. For example, for May 1st to May 7th, Chinese
people are celebrating National day as holidays.
Therefore, the market is closed during these days while
the U.S. stock market remains open. During these days,
investors in the U.S stock market react to the market
without the information in the Chinese stock market.
However, the breaks are relatively short compared to
this time scale containing a large data set.Besides the
two major variables mentioned above, it also include the
exchange rate of the currencies and interest rate in each
country. This paper got these data on Federal Reserve
Economic Research (FRED).The basic statistics are
shown in table 1:
Table 1. Descriptive Statistics
Variable
Obs
Mean
Std. Dev.
Min
Max
S&P 500
2545
118.52
18.96
2622.03
19531.2
Shanghai Composite Index
2545
2187.68
1004.48
1011.5
6092.06
Interest Rate in the U.S.
2545
4.34
0.85
2.08
6.79
Interest Rate in China
2545
7.02
0.92
4.55
8.96
Currency Exchange rate
2545
7.78
0.62
6.6
8.28
5. EMPIRICAL EVIDENCE
In our empirical application, I first analyze whether
the stock price in China can predict the stock price in
the U.S. before and after China joined WTO and
financial crisis occurred. In the model, the independent
variable is the stock price of S&P 500 and the
independent variables are the stock price of Shanghai
Composite Index, interest rate in the U.S. and the
currency exchange rate between yuan and dollar, job
report dummy.
Because the data is time-series data, whether the
stock price is unit root is tested by using augmented
Dicky-Fuller test. The MacKinnon approximate p-value
of S&P 500 is 0.5729 and the p-value of Shanghai
Composite Index is 0.3256, indicating that unit root
problem exists in the stock prices. Although the two
markets’ stock prices are unit root processes, there
might still be a relationship between the two. The
cointegration test is conducted to see whether the two
variables are cointegrated. The p-value is 0.0058,
indicating that there is evidence of cointegration. I also
Advances in Economics, Business and Management Research, volume 648
435

test whether interest rates in both countries are unit root
processes. According to the Dicky-Fuller test, the p-
value of interest rate in the U.S. is 0.1455 and that of the
interest rate in China is 0.2341. However, the
cointergration tests indicate that there is a conintegration
relationship between the interest rate and the stock
prices in each market. Hence, the OLS regression is not
a spurious regression. According to the regression
result, the positive impact of Shanghai Composite Index
on S&P 500 is significant.
Then, I analyzed the impact of the price of S&P 500
on Shanghai composite index. In the model, the
independent variable is the stock price of Shanghai
Composite Index and the independent variables are the
stock price of S&P 500, interest rate in China and the
currency exchange rate between yuan and dollar. The
cointegration test is conducted to see whether the two
variables are cointegrated. The p-value is 0.0063, so we
can reject the null and conclude that there is no evidence
of cointegration. According to the regression result, the
positive impact of S&P 500 on Shanghai Composite
Index is significant. Hence, there is a mutual forecasting
relationship between the two markets.
The RMSE of the first model, in which Shanghai
Composite Index price forecasts the stock price of S&P
500, is 20.58. This is a pretty small RMSE, indicating
that the out-of-sample forecasting ability is good.
Similarly, the RMSE of the second model, in which the
stock price of S&P 500 forecasts Shanghai Composite
Index price, is 15.72.
Besides the interaction between the two stock
markets, interest rate and currency exchange rate also
play a significant role. According to the regression
result, in the first model, the estimated coefficient of
interest rate is positive and statistically significant (p =
0.000). Whether central banks should set the short-term
nominal interest rate in response to stock price
fluctuations remains in debate.
The estimated coefficient of currency exchange rate
is negative and statistically significant (p = 0.032). The
regression result of the first model is included in table 2:
Table 2 The regression result
Independent Variable
Coef.
Robust
Std. Err.
p |
S&P 500
0.0056***
0.0035
0
Interest rate in China
0.266***
0.0384
0
Currency Exchange rate
0.18***
0.00723
0
*Significant at .1 level; **significant at .05 level; ***significant at .01
level
The second model do the same steps. Both of them
pass the Ramsey test and Pagan test, meaning that the
regressions do not have the problem of autocorrelation
and the functional forms are appropriate.
6. CONCLUSION
In this paper, by employing OLS and cointegration
tests, it presents evidence of the increasing interaction
between the Chinese and the U.S. stock markets, two of
the largest and most important stock markets in the
world currently.
Because there is no overlap of trading hours between
China’s stock market and the U.S. stock market, the
stock price during trading sessions in these two markets
are driven by different information sets. As a result, the
stock prices, which are mainly dependent on trading
activities, on those two stock markets may not show
significant correlations. However, this research offers
new evidence of the interaction between the two
markets. We show that since 2002 the stock prices in the
U.S. market indexes contain important overnight
information, which can forecast the opening stock price
of the Chinese stock market at the next opening.
Similarly, the stock price in the Chinese market indexes
can also forecast the opening stock price of the U.S.
market at the next opening. This finding can be valuable
to investors in a way that investors shall incorporate the
information in the other market to make the right
decision about investment in the domestic market.
There are limitations in this research. First of all,
significant forecasting ability of the two stock markets
after China joined WTO. However, due to data
limitations, we do not know this effect significantly
differs from the scenario before China joined WTO. If
we could have included ten more years before China
joined WTO, we would be able to test whether there is a
significant difference. Additionally, there might be other
variables that could explain the dependent variables but
are not included in the model. Moreover, further
research can be done to test whether the stock price in
one market granger causes the stock price in the other
market. To test this, we can conduct a Vector
Autoregressive model, in which the lagged values of the
dependent variable are included. It is hoped that these
estimates provide a departure point for more refined
models and data.Figures and tables should be placed
Advances in Economics, Business and Management Research, volume 648
436

either at the top or bottom of the page and close to the
text referring to them if possible.
REFERENCES
[1] Liu, H. (2018). Analysis of the Differences and
Linkage between Chinese and American Stock
Markets. American Journal of Industrial and
Business Management, 8(03), 700.
[2] Cheng, H., & Glascock, J. L. (2006). Stock market
linkages before and after the Asian financial crisis:
evidence from three greater China economic area
stock markets and the US. Review of Pacific Basin
Financial Markets and Policies, 9(02), 297-315.
[3] Qiu, Y., & Ye, C. (2019). Multifractal Analysis of
the Interaction between Chinese and American
Stock Markets. Open Journal of Statistics, 9(01),
143.
[4] Bissoondoyal-Bheenick, E., Brooks, R., Chi, W., &
Do, H. X. (2018). Volatility spillover between the
US, Chinese and Australian stock markets.
Australian Journal of Management, 43(2), 263-285.
[5] Zhang, P., Gao, J., Zhang, Y., & Wang, T. W.
(2021). Dynamic Spillover Effects between the US
Stock Volatility and China’s Stock Market Crash
Risk: A TVP-VAR Approach. Mathematical
Problems in Engineering, 2021.
Advances in Economics, Business and Management Research, volume 648
437