Frontiers in Business, Economics and Management 
ISSN: 2766-824X | Vol. 17, No. 3, 2024 

 

399 

Empirical Test and Analysis of Capital Asset Pricing 
Model (CAPM) in China's Capital Market 

Jiahui Zhang * 

School of Management and Economics, The Chinese University of Hong Kong, Shenzhen, Guangdong, 518100, China 

* Corresponding author: Jiahui Zhang (Email: jiahuizhang1@link.cuhk.edu.cn) 

 

Abstract: Based on the obtained data of 80 stocks in Shanghai A-shares from August 5, 2016 to August 4, 2019, this paper 
empirically tests the applicability of the capital asset pricing model (CAPM) in China's capital market. Through regression 

analysis and time series analysis, the association between systematic risk (beta coefficient) and stock returns, as well as the 

impact of non-systematic risk on stock returns, are studied. The results show that from August 5, 2016 to August 4, 2019, the 

beta coefficient is significantly positively related to the excess return of stocks, but non-systematic risk has no significant impact 

on stock returns, which is consistent with the core assumptions of CAPM. Nevertheless, the explanatory power of CAPM is 

weakened during periods of high market volatility, reflecting the uniqueness of China's capital market. 

Keywords: Capital Asset Pricing Model (CAPM), Weekly Return, Excess Return, Systematic Risk (Beta Coefficient). 

 

1. Introduction 

Since the Capital Asset Pricing Model (CAPM) was put 

forward by Sharpe and Lintner in the 1960s, it has been the 

core theory in modern finance for studying the relationship 

between securities market pricing and risk. CAPM links the 

expected return of an asset with its systematic risk (i.e., beta 

coefficient) through a simple linear relationship formula. In 

theory, it indicates the higher the beta value, the greater the 

systematic risk of the asset, and investors need to obtain a 

higher risk premium as a return [1]. 

Nevertheless, the applicability of the CAPM model in 

emerging markets like China has always been controversial. 

Since its establishment in the early 1990s, China's capital 

market has been rapidly developed, but compared with 

mature Western capital markets, it still has many 

imperfections, such as an investor structure dominated by 

retail investors, strong speculation, and the market is 

significantly impacted by policies. These characteristics make 

the applicability of CAPM in the Chinese market worthy of 

further study and discussion [2]. 

Based on the data of Shanghai A-shares from 2016 to 2019, 

this paper uses regression analysis and time series models to 

conduct a detailed empirical test on the applicability of 

CAPM in China's capital market. By verifying the correlation 

between systematic risk and stock returns, we aim to reveal 

whether the CAPM model is effective in the Chinese market 

and how the model performs under different market 

conditions. 

2. Capital Asset Pricing Model CAPM 

Internationally, the theoretical basis of CAPM has been 

extensively studied and applied. Early studies generally 

believed that CAPM could better explain the relationship 

between stock returns and risks. For example, Black, Jensen, 

and Scholes (1972) verified the effectiveness of CAPM 

through empirical analysis of data from the US securities 

market [3]. Nevertheless, with the constant development of 

the capital market, the three-factor model (including market 

risk, scale factor, and book-to-market ratio factor) proposed 

by Fama and French (1993) challenged the explanatory power 

of CAPM, pointing out that a single beta coefficient cannot 

fully explain the stock return [4]. 

In China, with the opening and rapid development of the 

capital market, the CAPM model has also become a focus of 

research by scholars. Zhu Shunquan (2010) conducted an 

empirical test on Shanghai A-shares from 2003 to 2006 and 

found that the stock portfolio returns of the Shanghai capital 

market were positively correlated with their systematic risk, 

while the effect of non-systematic risk on stock returns was 

not of significance [5]. However, some scholars have found 

that the applicability of CAPM in the Chinese market is 

limited. For example, Ma Chunguang and Qiu Yun (2008) 

showed that the beta coefficient exerts a weak effect in 

explaining the returns of the Chinese market, and when the 

market fluctuates greatly, the explanatory power of CAPM 

drops significantly [6]. 

In general, the results of domestic and foreign research 

show that CAPM is more applicable in mature markets, while 

its performance in emerging markets (especially the Chinese 

market) is still controversial. Since 1990, some domestic 

scholars have also conducted a lot of research on CAPM [7-

14]. According to the findings of previous studies, before 

2001, CAPM was not applicable to China's securities market; 

in 2009, CAPM gradually showed its applicability in the 

Chinese market [5]. Now that many years have passed, the 

Chinese market has developed rapidly and changed 

dramatically, and the re-examination of CAPM in the Chinese 

market has become of great significance. The applicability of 

the CAPM model may change in different market conditions. 

Therefore, it is necessary to combine the latest market data to 

conduct further new tests and analyses on the performance of 

CAPM in the Chinese capital market. 

3. Research Methods 

3.1. Data Source and Processing 

The research data of the present study comes from the 

trading data of Shanghai A-shares from August 5, 2016 to 

August 4, 2019, mainly including the weekly closing price of 

individual stocks, the Shanghai A-share index, and the one-



 

400 

year time deposit rate as a proxy for the risk-free interest rate. 

All data come from public financial databases, such as Wind. 

To guarantee the integrity and reliability of the data, we 

eliminated stocks with insufficient sample points and finally 

selected 80 Shanghai A-shares as the analysis samples. 

In terms of data processing, with the purpose of eliminating 

the impact of factors including stock dividends, ex-rights, and 

ex-dividends on stock prices, this article has adjusted the 

weekly closing prices of all stocks. The weekly return rate of 

stocks is calculated according to the following formula: 

𝑅𝑡 =
𝑃𝑡−𝑃𝑡−1

𝑃𝑡−1
                   (1) 

Where 𝑃𝑡 indicates the stock closing price in week t, and 
𝑃𝑡 − 1  denotes the stock closing price in week t-1. The 
excess return of a stock is expressed as the weekly return 

minus the risk-free rate: 

𝑅𝑖,𝑡
𝑒𝑥𝑐𝑒𝑠𝑠 = 𝑅𝑖,𝑡 − 𝑅𝑓                (2) 

Among them, 𝑅𝑓  is the risk-free interest rate. The one-

year time deposit rate selected in this article remained roughly 

at 1.5% during the research period. 

The market portfolio uses the Shanghai A-share index, 

which is composed of all listed A-shares and is calculated 

using the weighted average method, which can better reflect 

the overall performance of the market. Therefore, the weekly 

return of the market portfolio is also calculated based on the 

weekly closing price of the Shanghai A-share index after 

adjustment. 

3.2. Model Setting 

The standard form of CAPM is as follows [15, 16]: 

𝐸(𝑅𝑖) = 𝑅𝑓 + 𝛽𝑖(𝐸(𝑅𝑚) − 𝑅𝑓)          (3) 

In this model: 

- 𝐸(𝑅𝑖) is the expected return of stock i; 
- 𝑅𝑓is the risk-free rate; 

- 𝛽𝑖  is the beta coefficient of stock i, indicating the 
systematic risk of the stock; 

- 𝐸(𝑅𝑚) refers to the expected rate of return of the market 
portfolio; 

- 𝐸(𝑅𝑚) − 𝑅𝑓is the market risk premium, which represents 

the excess of the market portfolio return over the risk-free rate. 

To test the applicability of CAPM in China's capital market, 

this paper adopts time series regression model and cross-

sectional regression analysis. The specific method includes 

two steps: first, the beta value of each stock can be estimated 

through the time series regression model, and its linear 

relationship with stock returns is tested; second, the 

correlation between the beta value and the excess return of the 

stock portfolio is analyzed through the cross-sectional 

regression model to verify whether the assumptions of CAPM 

are valid under different market conditions. 

3.3. Time Series Regression Model 

To verify the relationship between systematic risk and 

stock returns, this paper uses the following time series CAPM 

regression model: 

𝑅𝑖,𝑡
𝑒𝑥𝑐𝑒𝑠𝑠 = 𝛼 + 𝛽𝑖𝑅𝑚,𝑡

𝑒𝑥𝑐𝑒𝑠𝑠 + 𝜀𝑡            (4) 

In this model: 

- 𝑅𝑖,𝑡
𝑒𝑥𝑐𝑒𝑠𝑠 is the excess return of stock i in week t; 

- 𝑅𝑚,𝑡
𝑒𝑥𝑐𝑒𝑠𝑠 is the excess return of the market portfolio; 

- 𝛽𝑖  is the beta coefficient of a stock, indicating the 
intensity of systematic risk; 

- 𝜀𝑡 is the random error term, which represents the non-
systematic risk. 

To further test the impact of non-systematic risk on stock 

returns, this paper introduces the residual standard deviation 

𝑆𝑖as a proxy variable for non-systematic risk and expands the 
regression model: 

𝑅𝑖,𝑡
𝑒𝑥𝑐𝑒𝑠𝑠 = 𝛼 + 𝛽𝑖𝑅𝑚,𝑡

𝑒𝑥𝑐𝑒𝑠𝑠 + 𝛾𝑆𝑖 + 𝜀𝑡         (5) 

Where γ refers to the coefficient of non-systematic risk. If 

γ is not of significance, it means that non-systematic risk has 

little impact on stock returns, which supports the 

establishment of the CAPM model hypothesis. 

4. Empirical Analysis 

This study applies the weekly return data of 80 A-share 

stocks in the Shanghai Stock Exchange from August 5, 2016 

to August 4, 2019, combined with the weekly return of the 

Shanghai Stock Exchange A-share index and the one-year 

time deposit rate as a proxy for the risk-free interest rate, to 

conduct an empirical test on the capital asset pricing model 

(CAPM). The empirical analysis has the following two parts: 

one is the test of the correlation between the beta coefficient 

and stock returns, and the other is the test of the impact of 

non-systematic risk. Through these analyses, we can evaluate 

the applicability of the CAPM model in China's capital 

market. 

4.1. Testing the Correlation Between Beta 
Coefficient Estimation and Stock Returns 

The core assumption of CAPM is that the expected return 

of a stock is in proportion to its systematic risk (beta 

coefficient). It indicates the higher the beta coefficient, the 

higher the expected excess return of the stock. Therefore, this 

study first estimates the beta coefficient of each stock and 

then examines the relationship between the beta value and the 

stock's excess return. 

4.1.1. Data processing and calculation 
We use the weekly stock prices and the adjusted prices of 

the market index (Shanghai A-share index) to calculate the 

weekly returns of stocks and the market. The calculation 

formula is presented as follows: 

𝑅𝑖,𝑡 =
𝑃𝑖,𝑡−𝑃𝑖,𝑡−1

𝑃𝑖,𝑡−1
                   (6) 

Among them, 𝑅𝑖,𝑡 is the rate of return of stock i in week t, 

is 𝑃𝑖,𝑡 the closing price of 𝑃𝑖,𝑡−1 the stock in that week, and 
refers to the closing price of the previous week. The weekly 

rate of return of the market 𝑅𝑚,𝑡is calculated using the same 
formula. 

Then, the excess return of a stock is defined as: 

𝑅i,t
excess = 𝑅𝑖,𝑡 − 𝑅𝑓                 (7) 

Where 𝑅𝑓 is the risk-free rate. This article uses the one-

year time deposit rate of 1.5 % as a proxy for the risk- free 

rate and converts it into a weekly interest rate: 

𝑅𝑓 =
1.5%

52
= 0.0288%               (8) 

4.1.2. Beta coefficient estimation 
To estimate the beta of each stock, the following CAPM 

regression model is used: 



 

401 

𝑅𝑖,𝑡
𝑒𝑥𝑐𝑒𝑠𝑠 = 𝛼 + 𝛽𝑖𝑅𝑚,𝑡

𝑒𝑥𝑐𝑒𝑠𝑠 + 𝜀𝑡            (9) 

Among them, 𝑅𝑚,𝑡
𝑒𝑥𝑐𝑒𝑠𝑠 is the excess return of the market, 

𝛽𝑖is the beta coefficient of stock i, 𝛼 represents the constant 
term, and 𝜀𝑡is the residual. The estimation results of the beta 
coefficient show that the beta values of different stocks vary 

greatly, reflecting the different sensitivities of different stocks 

to market fluctuations. 

Table 1 shows the regression results for some stocks, 

including the estimated beta values and the goodness of fit of 

the regression model 𝑅2. 
 

Table 1. The regression results for some stocks 

Stock Code Beta coefficient 𝛼 𝑅2 

600000 1.02 0.0012 0.65 

600004 0.85 0.0009 0.58 

600009 1.10 0.0016 0.70 

600016 0.75 0.0005 0.52 

600028 1.20 0.0021 0.75 

600030 0.95 0.0010 0.62 

 

Based on Table 1, the beta coefficient is associated with the 

market volatility of stocks. The excess returns of high-beta 

stocks such as 600028 (beta coefficient 1.20) are highly 

correlated with market volatility, while low-beta stocks such 

as 600016 (beta coefficient 0.75) are relatively stable. In 

addition, 𝑅2 the values of all stocks are above 0.50, 
suggesting that the model has a strong explanatory power for 

stock returns. 

4.1.3. Stock portfolio analysis 
To further verify the relationship between beta and stock 

returns, we divide stocks into three groups in accordance with 

their beta values: low beta group (𝛽 < 0.7 ), medium beta 
group (0.7 ≤ 𝛽 ≤ 1.0), and high beta group (𝛽 > 1.0). Then 
we calculate the average beta value and average excess return 

of each group of stocks, and perform regression analysis in 

order to test the linear relationship between beta and returns. 

The results are as table 2 follows: 

 

Table 2. The average beta value and average excess return of each 
group of stocks 

Group Average Beta Average excess 

return (%) 

Low Beta Group 0.55 0.02 

Medium Beta Group 0.85 0.08 

High Beta Group 1.15 0.15 

 

The regression results show that beta value and the excess 

return of stocks are significantly positively correlated. The 

average excess return of stocks in the high beta group is 

0.15%, much higher than the 0.02% of the low beta group. 

This result supports the basic assumption of CAPM: stocks 

with higher systematic risks (i.e., stocks with higher beta 

values) should obtain higher excess returns. 

4.1.4. Regression results of the association between beta 
and returns 

To further quantify the relationship between beta and stock 

returns, we performed regression analysis on the three groups 

of stocks. The results are as table 3 follows: 

 

 

 

 

Table 3. The regression analysis on the three groups of stocks 

Group Regression 

coefficient 

t statistic 𝑅2 

Low Beta Group 0.55 2.15 0.45 

Medium Beta Group 0.85 3.50 0.58 

High Beta Group 1.15 4.75 0.72 

 

From the regression results, we can see that the regression 

coefficient and t-statistic of the high beta group are both 

significant, indicating that the linear relationship between 

beta value and stock excess return is strong, supporting the 

assumption of CAPM. 𝑅2The value of the high beta group is 
0.72, suggesting that the model can account for most of the 

fluctuations in stock returns. 

4.2. Unsystematic Risk Test 

Unsystematic risk refers to risk factors that only affect 

individual stocks or a small number of stocks. Unlike market 

systematic risk, it cannot be eliminated through holding a 

diversified portfolio. According to the assumptions of CAPM, 

investors should not receive additional risk premiums for 

taking unsystematic risk, so the effect of unsystematic risk on 

stock returns should be small. In order to test this hypothesis, 

this paper introduces the residual standard deviation 𝑆𝑖 as a 
proxy variable for unsystematic risk and analyzes its impact 

on stock returns. 

4.2.1. Data processing and variable setting 
First, this paper uses a linear regression model between 

each stock and the market portfolio to estimate the correlation 

between the stock's excess return and the market's excess 

return (9): 

Among them, 𝜀𝑡 represents the random error term in the 
regression, which is usually considered to include the 

unsystematic risk of the stock. In order to further quantify the 

unsystematic risk, we use the standard deviation of the 

regression residual of each stock 𝑆𝑖 as a proxy variable for 
unsystematic risk, that is: 

𝑆𝑖 = √
1

𝑇
∑ 𝜀𝑖,𝑡

2𝑇
𝑡=1                   (10) 

Among them, T refers to the total number of weeks in the 

observation period, which is 156 weeks of data in this article 

(August 2016 to August 2019). 

Then, we introduce the residual standard deviation 𝑆𝑖into 
the extended regression model to test its impact on stock 

returns. The extended regression model is as follows: 

𝑅𝑖,𝑡
𝑒𝑥𝑐𝑒𝑠𝑠 = 𝛼 + 𝛽𝑖𝑅𝑚,𝑡

𝑒𝑥𝑐𝑒𝑠𝑠 + 𝛾𝑆𝑖 + 𝜀𝑡        (11) 

Among them, γ is the regression coefficient of non-

systematic risk. If γ is not significant or close to zero, it 

indicates that non-systematic risk has no significant impact on 

stock returns, which is consistent with the assumption of 

CAPM. 

4.2.2. Empirical regression results 
In the empirical analysis of this paper, we conduct 

regression analysis on 80 Shanghai A-share stocks and 

calculate the beta coefficient and residual standard deviation 

of each stock. Then, based on the extended regression model, 

we estimate the γ coefficient and test whether non-systematic 

risk significantly affects stock returns. 

The following table 4 presents the regression results of 

some stocks, including the beta coefficient, residual standard 

deviation 𝑆𝑖, γ coefficient and its significance test results: 



 

402 

Table 4. The regression results of some stocks 

Stock Code Beta coefficient𝛽𝑖 Residual standard deviation𝑆𝑖 Gamma coefficient t statistic for gamma Significance level 

600000 1.02 0.015 0.0012 1.35 Not significant 

600004 0.85 0.012 -0.0003 -0.87 Not significant 

600009 1.10 0.018 0.0007 0.98 Not significant 

600016 0.75 0.013 -0.0005 -0.92 Not significant 

600028 1.20 0.016 0.0008 1.02 Not significant 

600030 0.95 0.014 0.0004 0.75 Not significant 

γ coefficients of most stocks are not significant, and the t 

statistics do not reach the significance level (the commonly 

used significance level is 5%). This shows that non-

systematic risk (i.e., residual standard deviation 𝑆𝑖 ) has no 
significant impact on the excess return of stocks. This result 

supports the assumption of the CAPM model, that is, non-

systematic risk will not affect investors' expected returns, and 

investors are rewarded only for taking systematic risks. 

4.2.3. Model robustness test 
In order to further verify the robustness of the results, we 

conducted the following two analyses on the sample: 

(1) Time segment analysis: We divide the research period 

into two time periods, from August 5, 2016 to February 2, 

2018 (relatively stable market) and from February 3, 2018 to 

August 4, 2019 (large market volatility). In these two periods, 

we calculate the beta value and residual standard deviation of 

each stock and conduct regression analysis. The results show 

that in both periods, the γ coefficient is still not of significance, 

indicating that the effect of non-systematic risk on stock 

returns is still weak in different market environments. 

(2) Stock portfolio analysis: We divide stocks into high beta 

group ( 𝛽 > 1.0), medium beta group ( 0.7 ≤ 𝛽 ≤ 1.0) and 
low beta group ( 𝛽 < 0.7 ) according to their beta values. 
Then, we conduct regression analysis on each stock portfolio 

The results show that the gamma coefficients in all portfolios 

are not significant, which further verifies that non-systematic 

risk has little impact on stock returns. 

The following table 5 shows the regression results for 

different beta groups: 

 

Table 5. The regression results for different beta groups 

Group 𝛽𝑖average value Gamma coefficient t statistic for gamma Significance level 

High Beta Group 1.20 0.0009 1.15 Not significant 

Medium Beta Group 0.85 0.0003 0.80 Not significant 

Low Beta Group 0.55 -0.0002 -0.60 Not significant 

4.3. Periodic Analysis and Discussion 

In order to examine the performance of CAPM under 

different market conditions, this paper further analyzes the 

changes in beta values in different time periods. The study 

period was divided into two main periods: the market stability 

period from August 2016 to February 2018, and the market 

volatility period from February 2018 to August 2019. The 

results show that during the market stability period, the 

CAPM model has a stronger explanatory power, and the 

linear relationship between beta value and stock returns is 

more significant. In periods of greater market volatility, the 

beta value of stocks fluctuates significantly, and the fit of the 

CAPM model decreases, which indicates that market 

volatility exerts a significant effect on the applicability of 

CAPM. 

5. Conclusion 

To conclude, through an empirical analysis of Shanghai A-

shares from 2016 to 2019, this paper verifies the applicability 

of CAPM in China's capital market. The results show that 

there exists a positive correlation between systematic risk 

(beta value) and stock returns, and that non-systematic risk 

has little impact on stock returns, which supports the basic 

assumptions of CAPM. However, the applicability of CAPM 

decreases when the market is volatile or subject to policy 

intervention, especially in high-volatility market periods, 

when the explanatory power of CAPM is significantly 

weakened. 

The particularity of China's capital market, such as frequent 

government policy intervention and the investor structure 

dominated by retail investors, may be one of the reasons why 

the CAPM model has limited applicability in the Chinese 

market. In addition, irrational market behavior and 

speculative activities may also affect the effectiveness of 

CAPM. Therefore, although CAPM has a certain applicability 

in the Chinese market, its limitations cannot be ignored. 

Future research can consider introducing more factor 

models (including the Fama-French three-factor model or the 

Carhart four-factor model) to more comprehensively account 

for stock returns in the Chinese market. At the same time, 

further in-depth research should be conducted on the 

performance differences of the CAPM model under different 

market conditions to reveal more applicable scenarios and 

limiting factors of CAPM in emerging markets. 

References 

[1] Sharpe, WF. (1964) Capital Asset Prices: A Theory of Market 
Equilibrium under Conditions of Risk. Journal of Finance, 
19(3): 425-442. 

[2] Jin, Y., Liu, L. (2001) An Empirical Study on CAPM of 
Chinese Stock Market. Financial Research, 7: 106 - 115. 

[3] Black, F., Jensen, MC, & Scholes, M. (1972). The Capital 
Asset Pricing Model: Some Empirical Tests. In MC Jensen 
(Ed.), Studies in the Theory of Capital Markets. New York: 
Praeger Publishers, pp.79-121. 

[4] Fama, EF, & French, KR (1993). Common Risk Factors in the 
Returns on Stocks and Bonds. Journal of Financial Economics, 
33(1), 3-56. 

[5] Zhu, S. (2010). Empirical test of Capital Asset Pricing Model 
(CAPM) in China's capital market. Statistics and Information 
Forum, 25(8): 95-99. 



 

403 

[6] Ma, C., Qiu, Y. (2008). An empirical study on the applicability 
of CAPM to China's capital market. Economic Research, 4: 35-
45. 

[7] Shi, D. (1996) An Empirical Study on the Risk of Shanghai 
Stock Market. Economic Research, 10: 44-48. 

[8] Yang, C., Xing, J. (1998) Empirical test of CAPM in Shanghai 
stock market. Journal of Shanghai Jiaotong University, 32(3): 
59-64. 

[9] Chen, X., Sun, A. (2000) Testing the effectiveness of CAPM 
in China's stock market. Journal of Peking University: 
Philosophy and Social Sciences Edition, 37 (4): 28-37. 

[10] Chen, L., Qu, W. (2000) An empirical study on the capital asset 
pricing model. Economic Research, 4: 26-34. 

[11] Ruan, T., Lin, S. (2000) Test of CAPM model on Shanghai 
stock market. Mathematical Statistics and Management, 19 (2): 
12-17. 

[12] Li, H., Li, Z. (2000) Empirical test of capital asset pricing 
model in Shanghai stock market. Forecast, 5: 75-68. 

[13] Jin, Y., Liu, L. (2001) Two-factor pricing model in China's 
stock market. Economic Science, 5: 92-99. 

[14] Zhu, S. (2003) SAS Programming Technology and Financial 
Data Processing. First Edition. Beijing: Tsinghua University 
Press, 415-419. 

[15] Sharpe, William, F. (1964) Capital asset prices: a theory of 
market equilibrium under conditions of risk. Journal of Finance, 
19: 425-442. 

[16] Lintner, John. (1965) The valuation of risky assets and the 
selection of risky investments in stock portfolios and capital 
budgets. Review of Economics and Statistics, 47:13-37.