Market efficiency refers to the ability of the market to reflect all the information that are of interest to the stock in the stock price. A highly efficient market not only prices in the information continuously but also accurately (Barnes, 2009). However market efficiency has been a controversial topic in financial economies with several contradicting theories and empirical evidences that have been produced by the researchers. The efficient market hypothesis, which is considered the most important theory in this area of study asserts that an efficient market includes all the possibly available information including the privately-held information in the stock prices (Dothan, 2008). The semi-strong form of the efficient market hypothesis asserts that the price of a security adjusts instantaneously to any new information that is released to the market. Due to this reason no investor may be able to earn significant positive returns acting on the basis of new information that is released (Jarrow and Larsson, 2011). This brief research attempts to study the impact of stock-specific information on the movement of share prices of the company and consequently upon the returns generated by the stock. This research uses market model as the theoretical basis and event study as the empirical method for accomplishing this objective.
2. Research Aim and Hypothesis
The primary aim of this research is to analyse the impact of firm-specific information on the stock returns of the firm. From an analytical perspective, this research attempts to ascertain if stock-specific information could have significant impact on the returns provided by the company’s shares over short-term.
This research uses Virgin Media Inc as the company to be studied and the event of acquisition of Virgin Media by British Sky Broadcasting as the case study (Reuters, 2010). Four events related to this acquisition are identified and the stock returns surrounding these events are analysed. Following is the hypothesis that are identified and tested in this empirical research for each one of the four events separately.
H1: The information related to acquisition of Virgin Media by British Sky Broadcasting has significant impact on the stock returns of Virgin Media’s shares.
3.1. Market Model
The previous chapter has identified the important aim and the hypothesis to be tested for this research. In order to analyse the impact of specific information on the performance of the stocks of Virgin Media, this research uses empirical research. It is necessary for the empirical research to have a theory as its foundation and an appropriate method for testing the hypothesis (Eckbo, 2008). The theory that is used in this case is the market model. What explains the returns generated by the stocks of a company has remained an important question for several years in academic research as well as in investment practice. Capital Asset Pricing Model (CAPM) is widely considered the origin of the market model. CAPM (Capital Asset Pricing Model) asserts that the expected return of a stock contains a risk free rate of return and a part that represents the undiversified risk of the security (Elton et al., 2009). This is calculated as the product of the beta of the security and the return from the market portfolio. CAPM has been a popular model for explaining the returns of the security. It asserts that he investors are compensated in the form of returns not for bearing overall risks but only the non-diversifiable risks of the security. Due to this reason beta, which measures the sensitivity of the returns of the stock to the market returns, is taken as the slope of the linear equation. However there are some important limitations of CAPM model. The biggest limitation of CAPM has been that the model has failed to hold its group during empirical tests (Kürschner, 2008). It is notorious for its empirical invalidity. Researchers have attempted to identify alternative theories that can also withstand the empirical tests. Market model, APT and Fama-French three factor models are some of the most popular alternatives that have been identified (MacKinlay, 1995).
Market model decomposes the stock returns into two parts (Li and Pincus, 2008). The first part consists of a term that represents the part of return that can be attributed to the association of the security with the market index. Beta is used as the measure of the sensitivity of stock returns to returns of market index. The second part of the model contains the return that is attributable to the stock-specific or company-specific risks. Thus according to market model there are two kinds of information that may be considered pertinent to a stock’s price movement. They are the information related to the market as a whole and the information that are directly related only to that particular company (Keller, 2008). The market model can be shown through the following expression.
Rj = aj + ßj Rm + ej
Rj is the return from stock j
ßj is the slope which is the sensitivity of change in stock returns to change in Rm
Rm is the returns from market index
aj is the constant term
ej is the error term
In this research the above given model is used as the basis for modelling the stock returns of Virgin Media. In this research market model is used to classify the stock-specific information from the other information. This helps this research is to ascertain if the stock-specific information has a significant influence on the overall stock returns.
3.2. Research Method
Empirical research requires a suitable research method to carry out the analysis of data and to draw observations on the basis of the data. As stated in the previous section this research uses market model as the basis for studying the stock returns. This research uses event study analysis as the research method for calculating abnormal returns during and after the event. Event study is a popular method of research often used when specific events such as dividend announcements, stock splits, mergers and acquisitions etc., are studied closely by the researchers (Campbell et al., 2010). The event study method involves the use of an initial estimation period to model the relationship between the returns of the market and the returns of the stock. The parameters estimated are used as the basis for calculating the expected returns of the stock in the event period and the post-event period. The difference between the actual observed returns and the expected returns is taken as the movement that is due to that specific information that is released into the market. Thus event study involves the definition of there periods – estimation period, event period and post-event period.
In this case four events related to acquisition are considered. They are as follows:
|Event 1||4-Jun-10||British Sky Broadcasting buys virgin media channel for £160 million (Announcement).|
|Event 2||13-Jul-10||British Sky Broadcasting completes virgin media TV deal.|
|Event 3||20-Jul-10||British Sky Broadcasting deal for Virgin media goes to OFT.|
|Event 4||5-Oct-10||The OFT has approved the acquisition of Virgin media’ TV channel business by British Sky Broadcasting.|
For all the four events, only one estimation period is used. In this case the estimation period spans for 2 years preceding the first event. Therefore the estimation period is from 4 June 2008 to 3 June 2010. The event period and the post-event period have to be identified separately for the four events. The event period is taken as the three days beginning with the actual day of event. Two additional days includes because it is possible that the impact of the news event may have lasted for a few more trading sessions. The post-event period begins on the fourth day of the event and last till the end of the 3 months following the event day. Following table shows the event period and the post-event period for each event in the sample.
|Event||Event Date||Event Period||Post-Event Period|
|Event 1||4-Jun-10||4-Jun-10 to 6-Jun-10||7-Jun-10 to 4-Sep-10|
|Event 2||13-Jul-10||13-Jul-10 to 15-Jul-10||16-Jul-10 to 13-Oct-10|
|Event 3||20-Jul-10||20-Jul-10 to 22-Jul-10||23-Jul-10 to 20-Oct-10|
|Event 4||5-Oct-10||5-Oct-10 to 7-Oct-10||8-Oct-10 to 5-Jan-10|
Using the estimation period from 4 June 2008 to 3 June 2010, the intercept and slope are estimated employing the following market model.
Rj = aj + ßj Rm + ej
For the event period and the post-event period, a and ß from the above model are applied and the expected returns are calculated as follows.
Abnormal return is calculated as the difference between expected return and the observed returns during the event period and the post-event period.
ARj = Rj – E(Rj)
In order to calculated the cumulative abnormal returns, the total return generated by an investment of $1 made the day before the event day till the end of post-event period is calculated. This is known as the cumulative abnormal return (CAR). Average Abnormal Return (AAR) is calculated as the simple average of the AR for the entire period from event till the end of post-event period.
The hypothesis that is tested in this case is that the mean of abnormal returns generated by the stock during the event period and post-event period is significantly different from zero. If it is found to be valid then it is concluded that the stock-specific information has a significant impact on the stock returns. If this hypothesis fails then the conclusion would be that the events considered in this research have not had a significant impact on the stock returns.
This research requires stock price information related to Virgin Media as well as a suitable market index. In this case NASDAQ 100 index is taken as the market index. Virgin Media is traded on the NASDAQ stock exchange. All the data are taken on a daily periodicity to ensure sufficient observations to perform the calculations. One of the important concerns while obtaining data of Virgin Media was that the returns had to be adjusted for dividends. Therefore the adjusted closing price, which takes into account the dividend adjustments, is used as the basis for calculation of the returns.
The daily returns are calculated using the following formula.
Rt = ln(St) – ln(St-1)
This presents the continuously compounded rate of return for the stock price, which is considered appropriate for the empirical studies.
The next section of the report presents the observations from the analyses performed on the data sample.
4. Findings and Analysis
4.1. Estimation of Market Model
The first step in event study method used in this research is to estimate the market model and to estimate the values of a and ß of the stock. The model used is already shown in the previous section. Using the daily returns data pertaining to 4 June 2008 till 3 June 2010, the market model is estimated using Ordinary Least Squares (OLS) method of linear regression. The important results of the estimation are shown below.
|Adjusted R Square||0.61683|
|Total Number Of Cases||504|
|Ri = 0.0004 + 1.8338 * Rm|
|Coefficients||Standard Error||LCL||UCL||t Stat||p-level|
From the above table it can be seen that the coefficient of Rm, which is beta is 1.8338. The beta value of more than 1 indicates that the stock can be considered significantly high risk in nature. The t-statistic of the coefficient is 28.4731, which falls outside the area covered between the lower limit and upper limit at 95% confidence level as shown in the table. On the basis of this observation, the null hypothesis that beta is equal to zero can be rejected at 5% significance. The intercept, which is the alpha, has a value of 0.0003. Since the p-value of alpha is 0.7940 and it is more than 0.05, the null hypothesis that alpha is equal to zero is not rejected at 5% significance level. In other words, the alpha is insignificant while beta is significantly positive.
The overall validity of the model can be tested using F-test. The F-statistic is 810.7182 and the associated p-value is 0.000. Since the p-value is less than 0.05, the null hypothesis that all the coefficients are equal to zero can be rejected at 5% significance level. The R-squared value shown in the table indicates the extent to which the market return is capable of explaining the changes in the stock returns. In this case the R-squared value of 0.6175 indicates that Rm is able to explain about 61.75% of the changes in Rj. This can be considered fairly high. Thus the model is found to be able to significantly model the stock returns. The estimate model is shown below.
Ri = 0.0004 + 1.8338 * Rm
Since the data used in this estimation is time series, there is a risk of heteroskedasticity and autocorrelation (Hamilton, 1994). These two have to be checked before finalising the estimation model. The autocorrelation can be tested using the Durbin-Watson statistic.
The Durbin Watson statistic is estimated using the following formula.
The value of d is calculated as 2.0372. Since this value is not significantly different from 2, it is concluded that there is no significant autocorrelation in the residuals. The heteroskedasticity can be observed from the following chart showing the residuals and the Rm in scatter plot.
The above chart clearly shows that there is no pattern that can be observed from the distribution of residuals about the Rm. Therefore it can be concluded that the residuals do not have different variances. In other words, they are homoskedastic in nature.
On the basis of the above tests, it is concluded that the model estimated using the linear regression analysis is suitable for estimation of the expected returns during the event period and the post event period. The next section is concerned with the calculation and analysis of abnormal returns.