79

FIRM PERFORMANCE AND STOCK RETURNS:

AN EMPIRICAL STUDY OF THE TOP PERFORMING
STOCKS LISTED ON SHANGHAI STOCK EXCHANGE

Jin Dehuan, Shanghai University of Finance and Economics

Zhenhu Jin, Valparaiso University

ABSTRACT

This study attempts to determine whether and to what extent stock prices in the Chinese stock
market are driven by firms’ operating performance. We use a sample that consists of the top 10%
performers listed in the Shanghai Stock Exchange from 1996 to 2000 in terms of annual stock
returns. We use several measures to proxy for the firms’ operating performance. Simple and
multiple regressions are performed to determine, at the firm level, whether and to what extent these
variables are related to stock price changes and what portion of the stock price movements can be
explained by these firm performance variables. Our results indicate that while firm performance
measures had some explanatory power of the stock price changes in the first two years during our
testing period, the operating performance measures’ explanatory power of the stock price
movements generally declined as the stock prices went up. Our results suggest that the significant
stock price increases from 1998 to 2000 were not driven by firms’ operating performance. Instead,
they were driven by several non-performance factors.

INTRODUCTION

The relationship between a firm’s stock price movements and its operating performance has
been a subject of considerable research interest in recent years but the results of the studies are not
conclusive. Lamont (1998), Shiller (1984), Fama and French (1988), and Sivakumar and Waymire
((1993) find that the firm’s operating performance can explain a significant portion of the stock price
movements but Lev (1989) conducts a review of several studies on the information content of
earnings and reports that earnings changes are only weakly related to contemporaneous stock returns
and that the operating performance is not significantly related to contemporaneous stock returns.
This paper extends this line of research into the Chinese stock market, which has experienced
significant growth since the Shanghai Stock Exchange opened on Dec. 19, 1990. The purpose of
the study is to determine whether and to what extent stock prices in the Chinese stock market are
driven by firms’ operating performance. This is of particular interest because the Chinese stock
market is widely viewed as a very speculative, policy driven market. On the macro level, there
appears to be a disconnect between the overall stock market returns and the GDP growth. From 1993
to 2003, the Shanghai Stock Exchange Composite Index experienced a cumulative increase of

Academy of Accounting and Financial Studies Journal, Volume 12, Number 1, 2008
80

79.53%, from 833.8 to 1497.04 (The annual returns of Shanghai Stock Exchange Index from 1991
to 2000 are presented in Table 1). But during the same time period, the real GDP increased from
3.46 trillion RMB to 11.66 trillion RMB, a cumulative increase of 236%. Clearly, the stock market
has not been a good indicator of the macro economic development even though the number of listed
companies in Shanghai Stock Exchange increased from 169 to 559 and the total trade volume
increased from 64 billion shares to 231 billion shares from 1994 to 2000. In our study, we use a
sample that consists of the top 10% performers listed in the Shanghai Stock Exchange from 1996
to 2000 in terms of annual stock returns. We use several measures to proxy for the firms’ operating
performance. These proxies are changes in earnings per share, changes in sales, profit margin, total
asset turnover, returns on equity, and returns on assets. Simple and multiple regressions are
performed to determine, at the firm level, whether and to what extent these variables are related to
stock price changes and what portion of the stock price movements can be explained by these firm
performance variables. Our results indicate that while firm performance measures had some
explanatory power of the stock price changes in the first two years during our testing period, the
operating performance measures’ explanatory power of the stock price movements generally
declined as the stock prices went up. Our results suggest that the significant stock price increases
from 1998 to 2000 were not driven by firms’ operating performance. Instead, they were driven by
several non-performance factors.

Table 1: annual Returns of Shanghai Stock Exchange Index

Year End of Year Annual Change in %
1996 917.02 65.14
1997 1194.10 30.22
1998 1146.70 -3.97
1999 1366.58 19.18
2000 2073.48 51.73

The paper proceeds as follows. Section one provides a literature review on stock returns and
firms’ operating performance. Section two presents data, the empirical methodology, and the
results. Section three summarizes the study and concludes.

LITERATURE REVIEW

Many studies have been conducted on the relationship between firms’ operating performance
and the stock returns but the results have not been conclusive. Lamont (1998) studies the
relationship between earnings and expected returns. He reports that both dividends and earnings
have the ability to forecast returns and that earnings contain information because they are correlated
with business conditions. He also finds that higher volatility of earnings is not noise but related to

Academy of Accounting and Financial Studies Journal, Volume 12, Number 1, 2008
 81

expected returns. Shiller (1984), and Fama and French (1988) use regressions of returns on the
lagged dividends and earnings yield and find that they have explanatory power. Sivakumar and
Waymire (1993) study stock price behavior in relations to earning reports by 51 NYSE firms. They
find that the announced earnings are positively related to stock returns. Roll (1988) regresses
individual stock price changes on contemporaneous news events and finds that less than 40% of the
variance of price changes can be explained by the regressions. Fama (1990) finds that two-thirds
of the variance of aggregate stock price changes can be explained by the variables proxying for
corporate cash flows and investors’ discount rates. Campbell and Ammer (1993) contend that the
use of contemporaneous regressions to explain asset price variability is appealing though this
approach does not provide much information on the channels through which macroeconomic news
variables affect asset prices. Lev (1989) conducts a review of several studies on the information
content of earnings and reports that earnings changes are only weakly related to contemporaneous
stock returns. Su (2003) analyzes the dynamic behavior of risks and returns in the Chinese stock
market and finds that the volatility of stock returns to be high in China relative to developed markets
and that returns are positively auto-correlated to greater extent in Chinese stock markets than in
developed markets. He also finds that some of the government policies in regards to the stock
market contribute to the market volatilities. Jin and Li (2003) also study the causes of high volatility
and irrational price movements in the Chinese stock market and identify poor disclosure and
inaccurate financial reports by the listed firms as some of the causes for high volatility.

DATA, METHODOLOGY, AND RESULTS

Our sample consists of 10% of the firms that were listed in the Shanghai Stock Market from
1996 to 2000. The firms selected were the top 10% performers in terms of annual stock returns. The
total number of listed firms in Shanghai Stock Exchange and our sample size each year during our
study period are provided in Table 2. The average year-end Price/Earning ratios of the stocks listed
on the Shanghai Stock Exchange from 1995 to 2000 are presented in Table 3.

Table 2: Sample size each year in the testing period

Year # of listed firms Sample size % of total number of listed
at beginning of year firms
1996 184 20 10.87
1997 287 30 10.45
1998 372 40 10.75
1999 425 40 9.41
2000 471 50 10.62

Academy of Accounting and Financial Studies Journal, Volume 12, Number 1, 2008
82

Table 3: Average P/E ratios from 1995-2000

End of Year 1995 1996 1997 1998 1999 2000
Average P/E Ratio 16.32 32.65 43.43 34.38 38.14 59.14

To be included in the sample, a firm must have been listed for more than a year and some
key financial measures are available.
We use the following measures as proxies for firms’ operating performance:

Total Asset Turnover Ratio (TAT) = Sales / Total Assets

Change in Earnings per Share (CES) =
(Current Year Earnings per Share / Previous Year Earnings per Share ) - 1
Profit Margin (PM) = Net Income / Total Sales
Return on Assets (RA) = Net Income / Total Assets
Return on Equity (RE) = Net Income / Shareholders’ Equity
Change in Sales (CS) = (Current Year Total Sales / Previous Year Total Sales) - 1

We use simple and multiple regressions to test whether the above variables can, individually,
or collectively, explain the annual stock returns of the sample firms in each of the five years from
1996 to 2000. To ensure the multinomial normality of the independent variables and to avoid
possible multicollinearity, all the variables are tested for univariate and multivariate normality. The
results indicate that no variables need to be transformed to ensure normality. The sample firms were
the best performing stocks in our testing periods in terms of annual stock returns. In our regression
analyses, the annual returns of the individual firms are used as dependent variables and the firm
performance measures listed above are used as independent variables. We are not only interested
in which variables are statistically significant but also in the R-squared, which provides some
information on the explanatory power of the regression models.
Results from simple regressions indicate that return on equity (RE) had significant
explanatory power in four of the five year testing period but the explanatory power dropped from
0.4009 in 1996 to 0.0985 in 1999. Our results also indicate that changes in earnings per share (CES)
was significant in three of the five year testing period, 1996, 1997, and 1999. Interestingly, the
explanatory power of the variable also declined over the years, from 0.4830 in 1996 to 0.1336 in
1999. Other variables, such as profit margin (PM), return on assets (RA), changes in sales (CS), and
total asset turnover (TAT), were significant either in one or two of the five year testing period. These
results are presented in Table 4.
Results from multiple regressions are presented in Table 5 and they indicate that different
variables had significant explanatory powers in each of the four years in our testing periods. In 1996,
the variable, changes in earnings, (CES), was statistically significant. In 1997, two variables,
changes in sales (CS) and return on equity (RE) were significant. In 1998, Return on equity (RE)

Academy of Accounting and Financial Studies Journal, Volume 12, Number 1, 2008
 83

was again the variable that was statistically significant. In both 1999 and 2000, profit margin (PM)
was the only variable that was statistically significant.

Table 4: Results of Simple Regressions

(Variables which are statistically significant)
Variables Year R2 F t a
Total Asset Turnover TAT 1996 0.2652 6.4969 2.5489 0.05
1996 0.4830 16.8190 4.1011 0.01
Change in Earnings Per Share (CES)
1997 0.1318 4.2497 2.0615 0.05
1999 0.1336 5.8612 2.4210 0.05
1999 0.2119 10.2185 3.1966 0.01
Profit Margin (PM)
2000 0.0770 4.0035 2.0009 0.10
1996 0.2968 7.5978 2.7564 0.05
Return on Assets (RA)
1999 0.0890 3.7127 1.9268 0.10
1996 0.4009 12.0479 3.4710 0.01
1997 0.1516 5.0015 2.2364 0.05
Return on Equity (RE)
1998 0.0941 3.9490 1.9872 0.10
1999 0.0985 4.1539 2.0381 0.05
Change in Sales (CS) 1997 0.2142 7.6331 2.7628 0.05

Table 5: Results of Multiple Regressions

(Variables that are statistically significant
1996 1997 1998 1999 2000
Sample Size 20 30 40 40 50
Explanatory Variables CES RE PM PM
R2 0.4830 0.3312 0.0941 0.3233 0.0770
F 16.8190 6.6850 3.9490 8.8367 4.0035
t 4.1011 2.6929 1.9872 2.4672 2.0009
a 0.01 0.05 0.10 0.05 0.10

It is interesting to note that the R-squared of the multiple regressions followed the similar
pattern of the simple regressions. In 1996, the R-squared of our multiple regression was 0.4830 but
in 2000, it was only 0.0770. In other words, the explanatory power of the performance measures in
our regression models decreased during our testing period from 1996 to 2000 while the market index

Academy of Accounting and Financial Studies Journal, Volume 12, Number 1, 2008
84

moved up, from 917 to 2073. The correlation coefficient of the R-squared and the year end indices
were -0.6728(figure 1). When the market index was at 917 in 1996, the performance measures were
able to explain approximately 50% of the price movements. When the stock market took off and the
index soared to 2073 in 2000, the explanatory power of these variables dropped to less than 10%.
This suggests that the significant increase in market valuation from 1996 to 2000 was not based on
the operating performance of the firms listed on the exchange. Rather, the increase in market
valuation was based on factors that were not related to the operating performance of the firms listed
on the exchange. This interpretation is, in part, supported by the dramatic increase in the average
price to earning ratios from 1995 to 2000. During the five years, the price to earning ratio increased
from 16.32 to 59.14, an increase of 262.37%. This dramatic increase in price earning ratio clearly
shows that earnings did not increase while prices soared. At the same time, the declining R-squareds
from 1995 to 2000 in our regression models indicate that the important performance measures in our
models had less and less explanatory power of the stock price movements. Obviously, stock prices
were not driven by earning increases or other performance factors. Rather, they were driven by other
factors. We believe the disconnect between the stock price and firm performance may be due to the
following reasons. Starting from 1996, the benchmark one-year CD rate was cut repeatedly and a
lot of savers moved their money into the stock market from their bank accounts. In 1998, scores of
open-ended funds were launched and regulatory changes allowed state-owned enterprises to invest
in the secondary market. All these result in a significant flow of funds into the stock market. At the
same time, because the market was only established in 1991, there were only limited number listed
companies. The severe imbalance between the supply and demand for stocks caused the stock price
to go up significantly. In addition, the unique ownership structure of the listed companies and the
trade restrictions may also have contributed to the disconnect. During our testing period, about 60%
of the outstanding shares of the listed companies were owned by the state and these shares were not
allowed to be traded. In addition, investors were not allowed to sell stocks short even if they believe
that the stock price will go down. These institutional restrictions are not conducive to an efficient
market in which market prices reflect the intrinsic value of the stock. One limitation of the study
is that it only covers from 1996 to 2000. The Chinese stock market has undergone significant
changes but the result of the study provides important insight on the relationship between firms’
operating performance and their stock performance.

CONCLUSION

Using a sample that consists of the best performing 10% of the stocks listed on Shanghai
Stock Exchange from 1996 to 2000, we study the relationship between firms’ operating performance
and their stock returns. We find that while firm performance measures had some explanatory power
of the stock price changes in the first two years during our testing period, these firm performance
measures’ explanatory power of the stock price movements generally declined as the stock prices
went up. Our results indicate that the significant stock price increases from 1998 to 2000 were not
driven by firms’ operating performance. Rather, the stock price increases were due to other macro
economic factors and market imperfections.

Academy of Accounting and Financial Studies Journal, Volume 12, Number 1, 2008
 85

REFERENCES

Fama, Eugene & Kenneth French (1988). Dividend yields and expected stock returns. Journal of Financial Economics,
109, 241-266.

Fama & Eugene (1990). Stock returns, expected returns, and real activity. Journal of Finance. 45, 1089-1108.

Campbell, John & John Ammer (1993). What moves the stock and bond markets? A variance decomposition for long-
term asset returns. Journal of Finance. No.1, 3-37.

Lamont & Owen (1992). Earnings and expected returns. Journal of Finance. No.5, 1563-1587.

Lev., B. (1989). On the usefulness of earnings and earnings research: lessons and directions from two decades of
empirical research. Journal of Accounting Research (Supplement). 153-192.

Roll & Richard (1988). R2. Journal of Finance. No. 43, 541-566.

Shiller & Robert (1984). Stock prices and social dynamics. Brookings Papers on Economics Activitiy. 457-498.

Sivakumar, Kumar & Gregory Waymire (1993). The information content of earnings in a discretionary reporting
environment: evidence from NYSE industrials, 1905-10. Journal of Accounting Research. Vol 31, No. 1, 62-91.

Su, Dongwei (2003). Chinese stock markets, a research handbook. World Scientific New Jersey.

Academy of Accounting and Financial Studies Journal, Volume 12, Number 1, 2008
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.