In this paper, we examine the price performance in response to changes in earnings per share in the Shanghai Stock Exchange. We make a comparison between the A-share market and B-share market in terms of the price reaction to earnings announcement over a 3-year period from 2009 through 2011. The earnings announcements are informative. Specifically, positive abnormal returns are associated with satisfying earnings news while negative abnormal returns are accompanied by disappointing earnings news. Moreover, abnormal returns occurred the days prior to and the days after the earnings disclosure date. This may suggest that the capital market of China is not efficient. In addition, there is overreaction to the good news in the A-share market while there is overreaction to the bad news in the B-share market. We attribute the overreaction to good news to the fact that individual investors tend to be over optimistic to good news. The overreaction to bad news may be a consequence of the downturn trend in the B-share market in recent years.
Keywords: event study; earnings announcements; information content; Chinese stock market.
Introduction
From the angle of information economics, accounting and financial reporting is essential in an efficient capital market. Specifically, the reporting provides information to the market, which ultimately influences the security price. Among a variety of accounting and financial reports, the earnings announcements play a vital role. Basically all firms all over the world issuing shares have to disclose earnings reports, at least, annually.
The key explanation for releasing earnings reports focuses on market imperfections due to information asymmetries. Managers are supposed to have superior information concerning on the firm's operation and profitability over a specific period. Therefore, earnings announcements deliver valuable information to the market since it is a reflection of the financial conditions and management result of the company. Accordingly, investors prognosticate the future returns and risks on the firm and make their investment decisions.
Because of the explanation mentioned above, the information content of earnings announcements has been a major topic both in academic research and practical analysis. The focus is to evaluate the effect of the announcements on the equity price of the firms over a relatively short period, namely, event window. Event study provides a powerful setting to this evaluation because it builds a bridge through the event (e.g. earnings announcements) and the prices of the firm's equity.
Event studies are based on the notion of "efficient capital markets (Fama, 1970)". Fama (1991, p.383) also states that "security prices fully reflect all available information". As new announcements are available to the public, investors are expected to impound the information into the firm's stock price. Thus we can capture the impact of the new earnings announcement on the firm's value. In consequence, the incremental effect of the announcement on the value of the firm can be observed.
Event study methodology has been used widely in numerous studies on the valuation effects of earnings announcements. Ray Ball and Philip Brown (1968), Kross and Schroeder (1984), etc. find that stock prices respond positively to satisfying earnings announcement and negatively to disappointing earnings news in the U.S. stock market. Many other studies also explore the information content of earnings announcements in several non-U.S. markets (e.g., Brown (1970) for Australian firms, Firth (1981) for UK firms, Frost and Pownall (1994a) for firms listed in the US and UK). In addition, DeFond, Hung and Trezevant (2006) examine the cross-country differences in the information content of annual earnings announcements. They find that structural factors in the financial reporting environment can be used to explain the differences. They deduce that earnings announcements are more informative in countries with higher quality earnings and better enforced trading laws and stronger investor protection institutions. Whereas, the announcements will convey less information when a country witnesses more frequent interim financial reporting. These studies motivate us to extend this line of research to the Chinese Stock markets.
Chinese stock market is significantly different from other mature stock markets in accounting standards, financial reporting environment, corporate governance and so all. So it makes sense for us to investigate the information content of Chinese earnings announcements and examine whether there is any difference from other developed and mature markets.There are two exchanges, namely, the Shanghai Stock Exchange (SHSE) and the Shenzhen Stock Exchange (SZSE), where two types of shares-A-shares and B-shares---are listed. A-shares are strictly issued to domestic Chinese citizens. B shares were originally issued to attract foreign investors and could be traded only by foreign firms, causing rigid segmentation between A-share and B-share markets. Besides, companies issuing B-shares need to prepare two sets of financial statements based on two accounting standards, i.e. the IAS and the PRC GAAP, which are used to prepare accounting reports for the B-share holders and A-share holders, respectively. But the PRC GAAP and IAS numbers are released on the same day. Both domestic and international investors can access the two kinds of data freely, which makes it possible to use both the PRC GAAP and IAS data to make investment decisions.
Gao and Tse (2003) find that with the earnings number based on PRC GAAP, investors can get significantly positive abnormal returns in the good news event window and negative abnormal return in the bad news event window only in the B share market. But in the A-share markets, no significantly positive abnormal returns can be observed in the face of good news. They attribute this result to the market segmentation in the Chinese stock market.
Since February 2001, Chinese citizens are allowed to purchase B shares, which eliminate the rigid segmentation between A-share and B-share markets gradually. This change may affect the existing information content of earnings announcement, which forms the idea of this paper.
In this paper, event study methodology is adopted to investigate the impact of earnings announcements on the A- and B- share markets. The main focus is the different price reactions of the A- and B- shares in response to different types of earnings announcements (satisfying or disappointing) based on the PRC GAAP standards.
This paper contributes to the existing literature in several ways. First, we test whether an earnings announcement has an impact of on the market value of a firm after the financial crisis of 2008. As for as we know, this paper is the first to investigate the information content of earnings report in Chinese stock market since 2008 although there were a large body of researches on this topic before 2008. Second, we examine the incremental value relevance of earnings announcements. This method allows us to assess the efficiency of the market to absorb information that is partly disclosed before the event date. Third, we investigate whether there are any different reactions in response to different types of earnings news in the A-share and B-share market. Therefore, we make a comparison between the two market concerning on the information content. Finally, we identify the factors that have influenced the information content of earnings announcements in China.
Our results indicate that positive abnormal returns on the announce dates (-1, 0, 1) are associated with satisfying earnings announcements, whereas abnormal returns are negative in the case of earnings per share decreasing in the two markets. Furthermore, we find that the extent of reactions to different news varies in the two markets, i.e. A-share market and B-share market. In addition, there are sluggish and lagging market reactions in the two markets. Finally, our studies imply the insider trading and incomplete information system are the determinants of the abnormal price performance around the earnings announcement dates.
The remainder of the paper proceeds as follows. Section 2 presents a brief literature regarding event study and earnings announcements. Section 3 details the Chinese stock market. Section 4 describes the data and outlines the structure of the methodology employed. Section 5 presents the empirical findings for the sample. Section 6 provides some further explanations. The paper closes with the concluding discussion in section 7.
Literature review
This section begins with a review of the literature on the information content of announcements, followed by a brief examination of related research on the Chinese stock market.
2.1 General literature on the information content of announcements
The work of Ball and Brown (1968) and Fama et al. (1969) provides a framework for the event study methodology that is in use today. Ball and Brown investigate the valuation of earnings announcements. They conclude that the content of the annual report is considerable. But the annual report is not a timely medium because of interim reports and prompt media. Fama et al. studies the effect of the compounding events on the result. He examines the effects of stock splits after separating the effects of the simultaneous dividend inceases. Beaver (1968) examines the information value of earnings announcement from the perspective of common stock investors. He finds that an earnings report is said to have information content conditional on it leads to a change in investors' expectations on the firm's future returns such that the current equity price changes. Brown and Warner (1980, 1985) compare different methodologies that are employed in event studies to measure equity price performance and introduced abnormal performance into data. They suggest that the market model can perform well under various conditions. Monthly data and daily data are employed when considering implementation issues in the 1980 paper and 1985 paper.
MacKinlay (1997) reviews and summarizes the methods of event study. He set up a sample using the quarterly earnings announcements for the 30 firms in the Dow Jones Industrial Index through a five-year period. He concludes that the market gradually learns about the forthcoming announcement and investors can get abnormal returns over the event window. Besides the basic methodology of event study, he also develops some further issues, such as modifying the null hypothesis, power of the test and nonparametric tests, etc.
Alford and Jones, Leftwish, etc. (1993) compare the relative informativeness of accounting earnings in a number of countries using the U.S. market as a bench mark. Their investigation proves that the differences in the accounting standards, disclosure practice, etc. can result in apparently differences in the information content of the earnings announcement. Amir, Harris and Venuti (1993) make a comparison of the value-relevance between the U.S. and non-U.S. GAAP and state that the U.S. GAAP is more useful to the American investors. Chung and Lee (1998) analysed the Japanese stock market and find that different types of investors respond to the earnings announcement in different ways.
An event study performed by Bhattacharya, Daouk, Jorgenson, etc. shows that extending the investigation to emerging stock markets is of great significance. They examine the Mexican stock market and find that there are no abnormal returns and trading volumes over the event window. Evidence are provided to suggest that 'unrestricted insider trading' should be blamed because they make equity price fully incorporate the information before it is released. Since the insider trading is quite common in emerging market like Chinese stock market, it is an important academic topic to examine the information content of these markets.
2.2 Empirical literature on the Chinese stock market
There are a wide variety of literatures investigating the correlation structure and the equity price premium puzzle in the A-share versus B-share market. A seminal study by Bailey (1994) indicates that limited investment channels for Chinese investors can partially explain the price premium in the A-share market. Ma (1996) attributes the price premium to the investors' attitudes towards risk, their expectations towards the future returns and the correlation between the B-share market and the foreign stock market. In a paper of Sun & Tong (1998), the model of differential demand elasticity is employed to explain the premium puzzle. Specifically, the demand for B-share market is quite elastic because of the existence of the H-share market and the red-chip market.
There are also a number of empirical studies focusing on the segmentations of the Chinese stock market. Chui and Kwok (1998) use the cross-correlation analysis to show that the returns in the A-share result from the B-share market. Yang (2003) adopts the VAR model and indicates that the A-share and B-share market is not linked in the long run. Furthermore, foreign investors in the Shanghai B-share market are better informed than the domestic investors in the A-share market.
Recently, the information content of accounting reports based on different accounting standards (IAS and PRC GAAP) for the stock prices in the Chinese stock market has also been investigated. Bao and Chow (1999) find that earnings based on the PRC GAAP and IAS can both deliver information for the investors and lead to abnormal returns. But the information content of the earnings based on IAS is greater than that based on PRC GAAP. In one of his papers, Chen (2002) suggests that IAS is more informative to the B-share investors while the investors in the A-share pay more attention to the PRC GAAP.
Details about Chinese stock market
The Stock Exchanges and B-share market
There are two official national exchanges in China, namely, the SHSE established on December 19, 1990 and SZSE founded on July 3, 1991. Over the last 20 years, the two exchanges expanded rapidly. At the end of 2010, there were 894 and 1165 companies listed on the SHSE and SZSE, respectively. The stock market values were RMB 19701 billion and RMB 8428 billion, respectively.
From 1992, the Chinese government allowed a selected list of firms to issue tradable shares to foreign investors, which were called B-shares so as to distinguish them from those issued only to the Chinese citizens, i.e. A-shares. The B-shares listed on the two exchanges are dominated in the U.S. dollars on the SHSE and Hong Kong dollars on the SZSE. At the end of 2010, 918 A-shares and 54 B-shares were listed on the SHSE, and 472 A-shares and 54 B-shares were listed on the SZSE.
So as to maintain the control over local firms, the B-shares were issued only to foreign investors at the first few years. The foreign investors were not allowed to hold A-shares while the Chinese citizens could not invest in the B-shares market and could not buy foreign currencies freely. As a consequence, Chinese citizens had little investment alternatives which may be an explanation for the price premium of A-shares relative to B-shares. And also, this strict segmentation lead to the phenomenon that individual investors dominated in the A-share market while the foreign institutional investors dominated in the B-share market.
Since the institutional investors are more specialized in analysing the information, there may be different reactions between A-share market and B-share market in response to the earnings announcement. A large body of studies investigate on this topic and their conclusions are essentially similar. They find that in the B-share market, positive abnormal returns are associated with good earnings news while negative abnormal returns go with bad earnings news. However, this pattern is not so clear in the A-share market. (Chen, Firth and Kim (2002), Gao and Tse (2004)).
Chinese stock market witnessed a great change in 2001, i.e. Chinese citizens were allowed to buy B-shares. After this implementation, a number of domestic investors poured into the stock market and stimulated the growth of the B-shares market. Nonetheless, the performances of many listed firms on the B-share market since 2001 have been not satisfying and the international investors have many investment alternatives, such as the H-shares and the A-shares. The interest of international investors on the B-shares waned and the B-share market has been increasingly marginalized. There have been few B-share initial public offerings since 2001 and many small listed firms on the B-share market could not trigger new funds. The performance of B-share market is not satisfying in these recent years. The destiny of the whole B-share market has been a tough question these years under the shares-delisting mechanism of China.
3.2 The corporate disclosure and information system in the Chinese stock market
Chinese financial reporting requirements treat domestic and international investors differently. Firms issuing A-shares are required to report under the Chinese accounting standards, i.e. the PRC GAAP, whereas firms issuing B-shares are required to apply both IAS and PRC GAAP standards. However, each firm need to release annual report each year. According to D.Su (2003), the major characteristics of the annual reports include: (1) an emphasis is placed on reporting the company's short term objectives and strategies; (2) the focus is on the company's technological innovation and material events and legal proceedings; (3) little emphasis is placed on the company's business opportunities and risks; (4) the requirements for financial statements are relatively simple.
Furthermore, domestic and international reports tend to be audited by different audit firms. Specifically, domestic reports are audited by Chinese auditors who are less likely to be qualified and independent. By contrast, financial reports prepared by international firms tend to be audited by either Big Five or large Hong Kong firms, which avoid a number of problems facing local Chinese firms.
Disclosure can only be effective when accompanied by a complete information system. A complete information system not only includes the financial reporting and auditing standards but also financial media in general and efficient financial intermediaries. Nonetheless, the annual report is the main information resource for domestic investors in the Chinese stock market. Besides this, investors have to collect information-including insider information---all by themselves, which is both costly and inefficient. Unlike the mature capital markets, there are quite few professional analysts to provide information to the public. As a consequence, no information alternativeness may result in overreaction to the annual earnings report.
Data description and methodology
Data description
In this paper, we want to investigate the information content of the annual earnings content based on the PRC GAAP standard. What's more, we need to compare the price performances in the A- share and B-share markets in response to changes in earnings per share and examine the market reactions to good earnings news and bad news. So this study will focus on the annual earnings reports of the listed firms in the Shanghai Stock Exchange over a three-year period from January 2009 to December 2011. This corresponds to the annual earnings for year 2008 through year 2010.
After identifying the event, next comes to determine the selection criteria for including a given firm in the study. There are several criteria employed to screen the data.
A firm included in the sample data should issue both A-share and B-share in the Shanghai Stock Exchange. Only by this can we compare the different reactions concerning to one earnings announcement between the A-share market and B-share market.
Firms must be continuously listed on the A-share market and B-share market for 3 years prior to the announcement dates and one year after the announcement dates. Price data should be available for the period commencing 45 days prior to the announcement date and 35 days subsequent to the event date.
Firms belonging to financial, utility and service industries should be deliberately exclude from the sample. This is due to the fact that the capital and revenue structures of these firms are significantly different from those of firms from other industries. Their price performances are known to be sensitive to the government policy.
Shares have to be actively traded. According to the pioneering paper of Brown and Warner (1985), thin trading can lead to biased estimation of normal return. So it is essential to exclude those firms without transactions for over 30 days in the estimation window.
Confounding event is defined as another event that occurs at roughly the same time as the event of interest. Including this kind of event may lead to wrong interference about the true cause of a change in equity price. Hence, firms included in the sample should experience no merger & acquisition, no stock splits or dividends during the period under study. To capture the confounding events, news published about the firms should be checked. When selecting firms, we checked the website of the Shanghai Stock Exchange and China Securities Daily. In addition, shorter windows can limit the impact of the confounding event. This procedure trimmed the final sample significantly.
According to these criteria, 97 earnings announcements from 38 firms on the SHSE based on the PRC GAAP reporting standards survive on the screen. The 97 earnings earning announcements are consist of 53 increased EPS (earnings per share) and 44 dropped EPS.
For each announcement, we need to collect several pieces of data: the date of announcement, the earnings per share (EPS), daily stock prices, and daily market indices. Daily stock prices (A-share and B-share), price indices (Shanghai A-Share Index and B-Share Index) were collected from the DataStream. Dates of announcements and the news of dividend and split were collected from the website of SHSE.
4.2 The research methodology
The methodology employed here is the standard event study methodology. Before introducing the methodology, it is essential to define the null hypothesis. In this paper, we assume that the earnings announcement has no impact on the market performance. If the statistical diagnostics reject the null hypothesis, we can conclude that the earnings announcement convey information to the investors. Otherwise, the earnings announcement is less informative. Clear about this point, we specify the research method in the following part.
4.2.1 Classifying the event
First, we have to classify the event. If earnings announcements deliver information to the investors, we would expect there should be abnormal changes in the equity price in respond to the changes in the EPS over the event window. Specifically, stocks with EPS increasing should, on average, earn positive abnormal returns whereas stocks with declined EPS should, on average, earn negative abnormal returns. To capture this association, we divide our announcements into two categories: positive news and negative news. If the EPS in the report is higher than that of the previous year, the announcement is designated as positive news. By contrast, if the EPS is equal to or less than that of the previous year, the report is designated as negative news. Therefore, we get 53 positive events and 44 negative events.
4.2.2 Measuring the changes in the equity prices
Central to an event study is to measure the changes in the equity prices which can be realised by calculating the abnormal return. To do this, it is essential to specify an event window and an estimation window. A 16-day event window is employed, made up of 5 post-event days, the event day, i.e. Day 0, and 10 post-event days.
There is one point need to be specified here. In China, a number of firms release their annual report on non-trading day. In this case, we set the data (daily price and price index) of the announce date, i.e. Day 0, the same as that of the previous trading day and the next trading day after the disclosure date is set as Day 1. Similarly, the previous trading day is set as Day -1. For instance, if firm A releases its annual report on 01/05/2010, which is not a trading day. In this case, the announce date is set as Day 0 and all the data on this day is copied from the previous trading day, Day -1. The next trading day, which is 04/05/2010, is set as Day 1. Under this circumstance, the market reaction to the news is indeed on Day 1.
Day 0 is the day the news of announcement is published on the China Securities Daily and the website on SHSE. In many cases, the news is announced on the previous day and reported the next day. If the announcement is made before the market is closed, the market's response to the news actually happens on Day -1. If the announcement is released after the market is closed, then the response actually occurs on Day 0. Therefore, it is meaningful to investigate the 3-day announcement period market reaction, t=-1, t=0 and t=1.
For each announcement, 20 trading days prior to the event window and 25 trading days subsequent to the event window are used as the estimation period. In this paper, we include the post event window data in the estimation period data. According to Craig Mackinlay (1997), we can achieve the goal of increasing 'the robustness of the normal market return measure to gradual changes in its parameters. ' by doing this.
Abnormal (excess) return is the difference between the actual ex-post return and the normal return. We can calculate it using the following equation:
(1)
Where is the abnormal return on share i on day t and is the normal return on share i on day t. is the actual return on day t. can be calculated using the continuously compounding measure, i.e.,
(2)
Where and are the prices of share I on day t and t-1, respectively.
To measure the abnormal return, we have to follow the following steps: (a) predicting the normal return over the event window in the absence of the event; (b) aggregating the abnormal return through time and across the observations; (c) testing the null hypothesis that the abnormal return is zero.
Measurement of the normal return using market model
The normal return is defined as the expected return over the event window if the event did not happen. There are basically two models to calculate the expected return----constant mean return model and market model. These two models follow from the statistical assumption that assets returns are jointly multivariate normal and independently and identically through time. Although this assumption is strong, it is empirically reasonable (Mackinlay, 1997). Besides, Brown and Warner (1980, 1985) suggest that the results yielded from constant mean model is similar to those of market model. Therefore, we employ the market model to estimate the normal return in this paper.
The market model implies that the returns on a specific firm are related to the returns on a market portfolio. The returns on a market portfolio can be proxy by returns on market indices. In this paper, we adopt the returns on the Shanghai A-share index and Shanghai B-share index to calculate the market portfolio returns on the Shanghai A-share market and B-share market. Using the continuously compounding method, the index return can be calculated as:
(3)
Where is the market index on day t.
Hence, the expected return can be estimated using the equation:
(4)
can be estimated by the OLS method using the daily return data on the estimation period, that is [-25,-5) and (10,35]. Under the assumption of jointly independently, identically, normally distributed asset returns, the OLS is efficient.
Given the market model parameter estimates, we can measure and analyse the abnormal return. The abnormal return is
(5)
Where
The abnormal return is the disturbance term of the market model estimated on an out of sample basis (Mackinglay, 1997). Under the null hypothesis that the event has no impact on the equity price, the distribution of abnormal return of a given event in the event window is
(6)
Aggregation of abnormal return
The abnormal returns should be aggregated so as to draw the overall conclusion for the effect of the event. The aggregation should be along two dimensions---through the event window and across the observations of the event. For this aggregation, we assume that there is not any clustering, so that the abnormal returns and cumulative returns will be independent across securities.
Hence, given the 53 events of positive news, the average aggregated abnormal return for period is
(7)
And its variance is
(8)
Similarly, for the 44 events of negative news,
(9)
Using these equations, the abnormal returns for any event period can be analysed.
Then the average abnormal return can be aggregated through the event window to form the cumulative abnormal return. For any interval in the event window
Testing the null hypothesis
The null hypothesis, as we mentioned before, is that the event has no impact on the behaviour of stock returns (mean and variance). We consider the student t-statistic to test the null hypothesis. The statistic for CAR is calculated as:
The null hypothesis will be rejected if is in the critical region, that is. To make it simple, we set the critical values with for and as -1.96 and 1.96, respectively.
This test statistic depends on the statistical assumption (jointly, independently, normally distributed).Brown and Warner (1985) suggests that the test has high power when this assumption is achieved.
Empirical results
Table 1 and 2 display abnormal returns and cumulative abnormal returns around the event date. In table 3, the t-statistics for CARs are also presented. The trend of CAR is portrayed in Figs.
results about abnormal return
Table1. Abnormal returns for A-share market [1]
Table2. Abnormal returns for B-share market
From table 1, we can see that for the A-shares, abnormal return is significant on Day -3 for disappointing earnings announcement and on Days +1，+5 and +10 for positive earnings news. As far as abnormal return values are concerned, we find that the abnormal returns are negative in response to the bad news in the 3-day event period although none of them is significant. Similarly, the abnormal returns for the good news are positive for the 3-day event period, which is consistent with the bad news effects found in other stock markets and previous studies on the Chinese stock market.
From table 2, we can see that for the B-share market, abnormal return is significant on Days -3 and +1 for negative earnings news and on Days -5, 0, +5 and +7 for positive earnings news. For the negative news, the abnormal returns are negative for the 3-day event period while the abnormal returns are positive for the positive earnings news over the 3-day event period. Together with the results in table 1, this tends to give us the idea that positive abnormal returns are earned in the cases of increased earnings per share whereas negative abnormal returns are associated with decreased earnings. This idea will be further tested in the following part.
Trends of cumulative abnormal returns
As shown in the figures, the A-share CARs for the disappointing earnings before the event date are positive. They start to decline 3 days before the report is released and this decrease persists for 6 days to Day 3 before the CARs become randomly but still negative. This indicates that in the A-share market, investors cannot correctly anticipate the bad news and cannot adjust timely to the new news. However, the t-tests for average 3-day (-1, 1) and 16-day (-5, 10) A-share CARs cannot reject the null hypothesis. These two results imply that the investors in the A-share market are not sensitive to the bad news. Another explanation is that the bad news we define here is not what domestic investors consider bad news to be.
On the other hand, the CARs for positive earnings news in the A-share market begin to increase 5 days before the event date. And this increase is sharp and persists till Day 10. Furthermore, the t-test for the 3-day CAR and 16-day CAR significantly reject the null hypothesis that the announcement has no impact on the performance of price. This result can lead us to draw the conclusion that Shanghai A-share market is not semi-strong efficient as far as the earnings announcements for the sample stocks over the time period is considered. This result can also reveal the fact that domestic investors are over optimistic on the good news.
Table 3. T-test for CARs on the A-share market and B-share market
When we consider the B-share market, we can find that the CARs for the disappointing earnings announcements tend to decline 3 days before the announcement date and this decline will continue to Day 10. This indicates that in the B share market, investors are sensible to the negative earnings news. They can anticipate the bad news. After the announcements are released, the speed of adjustment is slow. In addition, the 3-day and 16-day CARs test reject the null hypothesis at 1% level. Contrast to the bad news, the CAR for the satisfying news starts to increase 5 days before the disclosure date of earnings information. Then it remains stable around a positive value which is approximately 2.5% after Day +1. The 3-day CAR is 1.6% with a t-statistic of 3.571 which also reject the null hypothesis of efficient market. The test for 16-day CAR also rejects the null hypothesis. These results indicate that the investors in the B-share market are sensitive to both positive and negative news. They can anticipate the news but the adjustment speed to the announcement is slow. Compared to good news, there tends to be overreaction to the bad news in the B-share market.
In conclusion，the earnings announcements have impact on the equity price both on the A-share and B-share market. This effect includes lagging response and sluggish market reaction. This implies that there is insider trading before the announce date such that a small component of investors can acquire the insider information and trade on it, making abnormal return. But a large part cannot access to this information. So the abnormal returns change gradually. After the report is released, all the investors are informed of the information, they adjust their expectations and revise their investment decisions. Therefore, the CAR increases dramatically. But the adjustment speed is such slow that the values of CAR are still significantly positive (negative) on Day +10 for satisfying (disappointing) news. The signal of abnormal returns has positive relationship with the property of news. That is, when the earnings news is positive news, investors can get a positive cumulative abnormal return. By contrast, when the news is a negative one, investors will get a negative cumulative abnormal return. This phenomenon can be understood easily. Most investors perceive that when the EPS of the stock increases (decreases) in the report, the future profitability of this firm will increase (decrease). Thus, they buy (sell) the share and the stock price will increase (drop), thereby earning a positive (negative) cumulative abnormal return.
In addition, A-share market investors are more sensitive to good news. Hence, in face of positive earnings announcement, the CARs increase dramatically and there is an overreaction in the market. Compared with A-share market, investors in the B-share market are more sensitive to bad news. This may be related with the overall situation of B-share market. This will be discussed further in the following section.
Further explanations
We believe that there are some reasons that have contributed to domestic investor's overreaction to earnings announcements: First, insider trading among government officials and managers. Second, most investors in Chinese stock market are short term individual investors who are more likely to speculate based on sentimental factors. Furthermore, the overreaction to negative news in the B-share market has something to do with the out-of-favour B-share market.
Insider trading
The information content of accounting numbers is contingent on satisfying market efficiency. Nonetheless, emerging Chinese capital market is widely known to be inefficient because of the numerous institutional frictions imposed by regulatory process and unequal access to information. Some of key institutional features are listed below.
Firstly, the government remains the majority shareholder. A sizable percentage of shares are owned by government officials and regulators. Although several laws restricted insider trading, they are not enforced. Besides, the 'insider trading' is difficult to define. As a result, it is very likely that insiders trade on the information before it is released to the public. In this case, the earnings numbers would have been impounded in share prices preceding to the disclosure date of annual reports.
Secondly, there is asymmetric information concerning various components of earnings between the managers of a firm and outsider investors. The managers tend to have inside knowledge or superior information. Other than macroeconomic variables, the outside investors are expected to depend on the managers to disclose the information. If the disclosed reports are informative and revised frequently over interim, there will be little earnings surprise or significant stock market reactions. If there is an adequate incentive for disclosure, the managers will disclose the information voluntarily and timely. Otherwise, regulatory agencies have to compel the managers to disclose information such that the quality of information is doubted. Therefore, the managers can trade on the information themselves or by their relatives before the announcement is released.
Given the information environment described previously, we can conclude that insider trading and manipulation of trade by officials and managers might be responsible for impounding accounting numbers into stock prices prior to public disclosure.
investors' sentiments and incomplete information system
Unlike other developed stock markets, the information system in the Chinese stock market is incomplete. There is no professional agent to make earnings forecast about the valuation of equity. In other words, investors lack alternative information resources other than the published accounting reports. The number of investment specialists and institutional investors (e.g., mutual funds and pension funds) is quite small. Consequently, price may be less informative and accounting information may contain more surprises in the market.
It is widely known that most individual investors in the Chinese stock market have a short term horizon. The number of long term investors is relatively small. With short-term horizon, investors tend to be more concerned with current period accounting numbers and may overreact to accounting information. Hence, they are more likely to revise their beliefs about future profitability based on the announcements. Therefore, if a firm releases positive earnings news, investors will believe that the company will earn more profit in the future and they will buy that share. Consequently, the price of that share will go up. By contrast, if there is disappointing news, the share price will go down since investors are doubted about the firms' future profitability.
Reason for the overreaction to disappointing news in the B-share market.
Established in 1992, B-share market was once the only way for international investors to enter into Chinese stock market. So the investors in the B-share market were mostly international institution. Therefore, in many early studies on the efficiency of B-share market, the conclusions are basically the same. That is, the B-share market is efficient and there is no significant abnormal return before and after the announcement is released. (Su 2003, Gao & Tse 2003, Chen, Firth and Kim2002, etc)
Since 2001, domestic investors were allowed to trade in this market. Trading volume soared to an unprecedented level at US$ 79.60 billion---more than nine times that of the previous year. At the same time, some qualified foreign investors are allowed into much larger A-share market. Apart from the B-share market, international investors can also invest in the H-share market and A-share market, etc. Consequently, many international investors fled from the B-share market as the domestic investors entered into the B-share market. So the investors have been mainly domestic investors in B-share stock market since then. As we mentioned before, domestic investors are short-term horizon and their trading strategy is mainly buy-and-sell. In this case, investors tend to overreact to negative news due to reasons we mentioned before.
In recent years, the trading volume in B-share market is quite small and the function of financing nearly disappears. There have been few initial purchasing offering in the B-share market since 2001. Some small firms cannot even trigger new capital. Over time, the B-share market has become illiquid. The expectations on the future of B-share market are essentially pessimistic. Under the circumstance, any negative news may be exaggerated. Hence, when there is disappointing news, investors are panic about the future profitability and price will decrease, thereby significantly negative abnormal returns will occur.
By contrast, in the A-share market, although the cumulative abnormal returns are negative, the t-statistics are insignificant and we cannot reject the null hypothesis that annual news report has no effect on the share price. One explanation for this insignificance can be that A-share holders tend to be more tolerant to negative news. This may attributed to the expectations of the investors on the government's behaviour. Traditionally, the Chinese government is likely to intervene and "save" the market in the circumstance of downturn. It has become conventional wisdom among investors that government will save the market. This notion makes investors in the A-share market more aggressive and more optimistic. And the expectation of government saving market makes investors less sensitive to disappointing news and over react to positive news. So the t-statistics for bad news in the -share market are insignificant.
reasons for overreaction to positive news
As shown in Figs, we can observe that there is a dramatic increase in the CARs over the event period in the two markets, especially in the A-share markets. The sharp increase trend indicates the overreaction to good news among the investors of Chinese stock market. We maintain that the special makeup of Chinese stock investors can be used to explain this phenomenon partly.
In the Chinese stock market, the majority shareholders are individual investors, who are more risk averse. The individual investors appear to be over optimistic to good news. Behaviour finance study shows that when filtering information, individual investors are likely to focus on the information that can boost their confidence and selectively screen out those undermining confidence. This kind of behaviour may lead to the overreaction to good earnings news in the stock market.
Conclusion
We have conducted an event study based on the annual earnings announcements on the Shanghai stock market over a three-year period from 2008 and 2010. The annual earnings announcements in both A-share market and B-share market are value relevant. Negative cumulative abnormal returns are observed on the event window when the disappointing earnings news is released while positive cumulative abnormal returns are obtained when satisfying earnings news is published. We can observe both lagging and sluggish response to the earnings announcement. These findings indicate that the Chinese stock markets are not efficiency and price cannot reflect all the information released.
We also examine reasons for the existence of abnormal returns. The insider trading operated by government officials and firms' managers leads to the pre-event abnormal return. The incomplete information system and short-sighted behaviour of individual investors can partially explain the sluggish response to the earnings announcement.
In the B-share market, we can observe overreaction to negative news. This phenomenon may be due to the downturn in the B-share market. Investors are lack of confidence on the future profitability of B-share market. When a disappointing earnings report is released, the pessimistic expectation may be exaggerated, which may lead to significantly negative abnormal returns. By contrast, in the A-share market, the t-test statistics for CARs are insignificant. This may be attributed to investors' belief that government will save the market in the downturn, which makes them over optimistic to the market.
Overreaction to positive news exists on both markets, especially on the A-share markets. This can be explained by some findings in the behavioural finance. Individual investors tend to pay close attention to the information that can boost their confidence and screen out the ones damaging their confidence. As a consequence, investors are over optimistic to good news and earn considerable positive abnormal return.
Our empirical findings have practical implications for both investors and policy maker. Specifically, potential investors can exploit significant abnormal returns by trading around the announcement dates. The daily abnormal return on the announcement day can be 0.45% and 0.62% in the A-share market and B-share market, respectively in the case of EPS increases. For the policy maker, the financial information system in the Chinese stock market needs to be improved in the timeliness, transparency and completeness. Furthermore, the regulatory authorities must take steps to prevent insider trading as much as possible.
Future research should be directed to examination of daily turnover volume variation around the announcement dates since trading volume is an essential aspect of market performance. He and Wang (1995) indicate that trading volumes around the earnings announcement dates are value relevant in investigating the information content of earnings announcements. In addition, future studies can extend the research on the Chinese stock market to other emerging markets, so as to compare the information contents of earnings announcements between different emerging countries.
