Вы находитесь на странице: 1из 17

Acknowledgement

I hereby take the opportunity to thank Metrics 4Analytics Pvt. Ltd. for providing
me with summer training at its facility in Hyderabad. This association, however
short lasting, has provided me with a comprehensive insight into the various
sectors of the US share market.
I would like to express my deep gratitude to Mr. Anjaneyulu Marempudi, Founder
& CEO, Metrics4 Analytics Pvt. Ltd. alongwith Mr. Rajeev Gupta, Director -
Research & Analytics, Metrics4 Analytics Pvt. Ltd. and Mr. Sanjay Banka,
Director - Research, Metrics4 Analytics Pvt. Ltd., who gave me the opportunity to
pursue my training in the Finance department at Metrics 4.

I would like to thank my faculty guide Dr SV Seshaiah of IBS Hyderabad for his
constant guidance and suggestions which again were very important as they
formed the foundation for the project undertaken.

I owe special thanks to my colleagues during tenure of this project, Ms. Srilatha,
Mr. Aman, Mr. Abhishek and Ms. Dipti for their constant support, guidance and
encouragement in exploring the nuances of the project without which, it would not
have been possible for me to accomplish this study.

Any omission in this brief acknowledgement does not mean lack of gratitude.

IBS, Hyderabad Gaurav Agarwal

07BS1398

May 2008

1|Page
Table of Contents
Acknowledgement................................................................................................................................... 1
Abstract: ................................................................................................................................................. 3
Methodology........................................................................................................................................... 4
Introduction ............................................................................................................................................ 5
Limitations of study ............................................................................................................................. 5
Objectives ............................................................................................................................................... 6
Event Window ..................................................................................................................................... 7
Estimation Window ............................................................................................................................. 7
Normal Return..................................................................................................................................... 7
Abnormal Return ................................................................................................................................. 7
Mean AR.............................................................................................................................................. 8
Labeling ............................................................................................................................................... 8
Pre & Post event Abnormal Returns..................................................................................................... 8
t-test: Paired Two Sample for Means ................................................................................................... 8
Findings: ............................................................................................................................................. 9
Volume Volatility ................................................................................................................................... 10
Average Volumes traded ................................................................................................................... 10
Abnormal Volume Volatility ............................................................................................................... 10
Cumulative Volume Volatility............................................................................................................. 10
Mean Volume Volatility ..................................................................................................................... 10
Pre & Post event Volume volatility..................................................................................................... 11
t-test: Paired Two Sample for Means ................................................................................................. 11
Findings: ........................................................................................................................................... 11
Market Adjusted AR .............................................................................................................................. 12
Findings: ............................................................................................................................................ 12
CAR Window ......................................................................................................................................... 12
Volume Volatility ................................................................................................................................... 14
Findings: ............................................................................................................................................ 14
CAR Window ......................................................................................................................................... 14
Share Price Return & Volume Volatility .................................................................................................. 15
References:............................................................................................................................................ 16

2|Page
Abstract:
The report titled “Impact of Buzz on Share Price-Volume Volatility” refers to
finding and analyzing the impact of the event type – buzz on the companies‟ share
prices and volumes being traded in the US stock market.

The whole project was divided into two phases:


Phase I: Price Return & Volume Volatility — In this phase I have tried to
analyze the impact of buzz on the share price return and the volumes traded taking
102 samples over a period of 2007-2008. It gives a detailed insight into the
abnormal returns that the market has provided to such an event over a period of
years.

Phase II: Create a Model for the same -- After the analysis of the event from the
specified sample size I have tried to put the same in a model that can be used for
making better decisions regarding any investment/trading activity on the happening
of the event analyzed.

The aim of this project is to study the impact of buzz and create a model that can
be used further on by the investors.

This analysis part would include both the qualitative and quantitative analysis. We
need to study the effect of both qualitative and quantitative data in order to carry
out an effective and efficient research.

3|Page
Methodology
Methodology adopted for the project is exploratory and analytical.

Data collection & analysis:


The buzz will be collected by referring to company websites. The buzz is related to
a particular company and the information related to its share prices and volumes
traded have been collected from relevant websites.

Looking for alternatives:


I have done a detailed analysis of the impact of buzz on the share prices by taking a
sample of 102 buzz that have occurred over the period of 2007 and 2008. I have
tried to create a model to factor in the effects of the buzz and give the investors
more detailed and better insight into the behavior of buzz. I have used a model to
calculate the volume volatility before and after the occurrence of the event. I will
now try to use a model for measuring the price volatility in my final project.

Documentation:
After all the analysis, I have done the final documentation of my project.

4|Page
Introduction
The share market is a function of price and volume volatility. This volatility is
caused by various fundamental and technical factors. All the factors that affect the
share price movement can be attributed to either the fundamental or the technical
factor as per its attributes. For example – If a company announces acquiring of a
big contract, it will have an impact on the share price-volume volatility. This
variable can be attributed to the fundamental factor as it shows strong growth in
company and it can be inferred that the company is adding strong value to its
bottom-line, in other words, has got strong fundamentals. On the other hand, if we
take the example of a company‟s announcement of potential acquisition of a
company. The share price-volume volatility caused in such a case is attributable to
technical factors since the company‟s share price is affected by the positive or
negative sentiments (as the case may be) of the people in relation to the potential
M&A announcement.
Companies try to declare events when they believe it is best suited for them so as
to have the required effect on its share prices. But it is not possible. In the process,
the trustworthy sources of market such as Wall Street Journal, Reuters etc. declare
the news which the company is most likely to announce at a later date. Such earlier
reporting of any event from any source other than that of the company itself is
captured in our company as „Buzz‟. These may be opinionated statements or may
even be a report that the company planned to declare at a later date.

Limitations of study

The project is subject to following limitations:

The project has been prepared from secondary data.

The project includes only one event type – “buzz”.

5|Page
Objectives
In the project I have tried to complete the objectives proposed:

Understand the impact of different buzz on the share price-volume


return and volatility.
Analyze the share price return and volume volatility to understand the
impact of the buzz.
Identify the events whose buzz have more impact on the share prices
of the company.
Attempt to create a model for better prediction of the impact of a buzz
on the share prices.

6|Page
Event Window
Event Window refers to the time period which is in immediate precedence and
succession to the event date. It is classified into pre-event and post-event
observation days. In other words, event window is used to understand since when
did the actual impact on the share price started occurring and till what time it
remained.
Buzz are short-term events and does not have an impact for a very longer period. I,
thus, decided to take the event window as t+13 and t-13 days where „t‟ represents
the event date.

Estimation Window

The next step was to decide on the estimation window. Estimation Window refers
to the time frame which we take to determine the normal return that a stock
provides when there is no event occurring. Generally we take the estimation period
starting one month prior to the actual event date and can extend it backwards to the
time frame suited for the analysis.

For my purpose I have taken my estimation window as -30 days to -60 days to
obtain the normal return provided by the stock when there is no other event taking
place.

Normal Return
Normal Return is the return that a company is expected to provide to its share-
holders if no event occurs for the company. It is very important to know the normal
returns to calculate the abnormal returns that the company provides during the
event window. Normal return is calculated for the stock in the estimation window.
The formula used for the calculation of normal return is:
Average Return = Average of Mean Returns
Normal Returns = {(today‟s closing share price – yesterday‟s closing share
price)/yesterday‟s closing share price}*100

Abnormal Return
Abnormal Return is the return that the company provides to its shareholders during
the event window when compared to the normal returns that the company would
have provided otherwise. The formula used to calculate abnormal return is:
7|Page
Daily Return = {(today‟s closing share price – yesterday‟s closing share
price)/yesterday‟s closing share price}*100

AR = Daily Return - Average Return

Mean AR

Mean AR is the average of abnormal returns on a cumulative basis provided on a


given day during the event window by the sample taken for consideration. It is
because we have to identify the impact of buzz on a cumulative basis since it is not
possible to get a clear picture if we apply it only on a single company. All the ARs
for each company on the given date were brought together in one excel sheet and
the following formula was used:

Mean AR = Average of ARs across all 102 events.


Labeling

We label the pre-event window in such a way that the last day is labeled as -14 and
the event date is always labeled as 0 to make identification easier. The post-event
window is labeled starting from 1 to the last date. In my case, the labeling starts
from -14 days and continues to +14 days.

Pre & Post event Abnormal Returns

Pre-event ARs refer to the average of Mean AR from -1 to -14 days.


t-test: Paired Two Sample for Means
On seeing the charts and the means one may easily comment on the significance of
the buzz and its impact on the share price. But it is very important to statistically
prove the significance of the buzz pre-event and post-event. I have used paired
sample t-test in my case. This test is useful when we compare effect of a particular
event on the sample. In my case, event is the buzz, and so I compared the pre-event
and post-event abnormal returns during the event window to understand if it is
statistically significant.
Null hypothesis: There is significant difference in the pre-event and post-event
abnormal returns.
Alternative Hypothesis: There is no significant difference in the pre-event and
post-event abnormal returns.

8|Page
Findings:

On running the t-test I found that my null hypothesis got rejected, i.e. alternative
hypothesis comes out to be true. In other words, there is no significant difference
in the pre-event and post-event abnormal returns.
It implies that for the sample of 102 companies, the event window that I have taken
into consideration is not a representative of the actual impact of buzz on the share
prices. So I modified my event window to t+5 and t-5 days and re-run the test with
the same hypothesis. Again, it was statistically proved that there is no significant
difference in the pre-event and post-event abnormal returns. It again implies that
for the sample of 102 companies, the event window is not a representative of the
actual impact of buzz on share prices. Again, I modified my event window to t+2
and t-2 days and re-run the test with the same hypothesis. Again, it was statistically
proved that no significant difference exists in the pre-event and post-event
abnormal returns.

9|Page
Volume Volatility
The trading volumes are a very important parameter in understanding the impact of
any news on the share prices. The volumes are the first indicators of any impact of
any news on the share prices. The trading volumes differ as the investors‟
confidence keeps varying with the perceptions. It is very important for us to
identify how the investors‟ confidence varies pre-event and post-event.
I have taken the same 102 events as my sample database to work upon.

Average Volumes traded

Average volumes traded refer to the average volumes that trade in the stock under
normal conditions. I calculated the same for my estimation window.
Abnormal Volume Volatility

Since we are calculating volatility we should consider the movement of the same in
percentage increase or decrease terms. It is because volatility is a ratio. It is just a
number. It may be positive or negative. While calculating the abnormal returns we
used the daily price return averages and used the same for our purpose but in
volume volatility this cannot be applied. I have thus taken the ratio of the volumes
traded during the event window and the average volume traded during the
estimation window and subtracted it by 1 to get the percentage form of the same.
Abnormal Volume Volatility = (Volume traded on the day/Average Volumes
traded during the estimation window)-1

Cumulative Volume Volatility


After calculating the abnormal volume volatility, I put them all in one worksheet
and calculated Mean Volume Volatility.
Mean Volume Volatility

The average of the volume volatility for each day in the event window was taken
separately to get a broader picture of the volume volatility due to buzz as an event
on stock prices.

10 | P a g e
Pre & Post event Volume volatility

The pre-event mean was used to compare with the post-event mean volume
volatility so as to identify the visible difference between the two.
t-test: Paired Two Sample for Means

On seeing the charts and the means one may easily comment on the significance of
the buzz and its impact on the volumes traded. But it is very important to
statistically prove the significance of the buzz pre-event and post-event. I have
used paired sample t-test in my case. This test is useful while analyzing the impact
of a particular event on a particular sample. In my case, event is the buzz, and so I
compared the pre-event and post-event volume volatility during the event window
to understand if it is statistically significant.
Null hypothesis: There is significant difference in the pre-event and post-event
volume volatility.
Alternative Hypothesis: There is no significant difference in the pre-event and
post-event volume volatility.

Findings:

On running the t-test I found that my null hypothesis got rejected, i.e. alternative
hypothesis comes out to be true. In other words, there is no significant difference
in the pre-event and post-event abnormal returns. So I modified my event window
to t+5 and t-5 days and re-run the test with the same hypothesis. Again, it was
statistically proved that there is no significant difference in the pre-event and post-
event abnormal returns. Again, I modified my event window to t+2 and t-2 days
and re-run the test with the same hypothesis. Again, it was statistically proved that
no significant difference exists in the pre-event and post-event abnormal returns.

11 | P a g e
Market Adjusted AR
Since the normal return method has not provided any significant results I have
calculated the abnormal return after adjusting the returns with the market return to
get a clear picture on the significance of abnormal returns in the pre-event and
post-event window.
We have used the macro file provided by our company guide to calculate the
market adjusted abnormal return for our purposes. Table 1 shows the market
adjusted returns for the sample I have used to work upon.
It clearly shows the different event days and their corresponding Mean AR. The
same has been standardized to find the standard error on each day during the event
window. The respective Z scores have also been assigned to check the
significance. The symbol “~” represents the event day is not significant whereas
the symbol “*****” represents the event day is significant.

Findings:

The event buzz has significant impact on the share price returns on Day 0, -9 and
+10. It proves that the share prices of a company are impacted by the event
“Buzz”.

It can be concluded that an investor should consider the impact of buzz while
investing in the market over any buzz. It will help him to be more specific and
clear in making his/her investment decisions.

CAR Window

The market adjusted return method provides us information about the significant
days in the event window but it does not provide an insight as to when should an
investor invest. We therefore use the CAR window to identify the right time for the
investor to invest in the market based on the event type being analyzed.

CAR window represents Compounded Average Return window for the samples on
a compounded basis for different combinations of the event window. We try to
take all possible combinations in the event window to identify the best possible

12 | P a g e
frame for the investor to be in the market in relation to the event type being
analyzed.

In my event case, the investor who invests for the period of (-2, +15) gets the best
abnormal return from the market.

13 | P a g e
Volume Volatility

The method first applied did not provide a comprehensive insight into the
significance of the event “buzz” on the volume volatility. However, the method
which was used second did prove the significance of the event type “buzz” on the
share volume volatility in the market.

Findings:

During the event window, only day -2 and day +12 have been found to be not
impacted by the event type “buzz”. All the other days have shown to have
significant impact of the event over the share volume volatility.

It can be concluded from the same that the buzz start making its impact on the
share volume 15 days before it is actually out in the market and continues to have
the impact till +15 days in the market. It can also be inferred that the market
discounts the buzz earlier and people start trading in the stock as soon as the
rumors are out and look for opportunities to make money.

CAR Window

Again, from the table we just know the days which have significant impact of the
event type under consideration here. But this is not enough for the investor to make
proper decisions in the market. He/she should also know when is the correct time
for him/her to enter or exit the market. This is possible through CAR window.

Compounded Annual Return Window, simply CAR window is a significant tool in


analyzing when an investor should enter/exit the market to make the best possible
investment decision related to the event type under consideration.

It implies that the people start trading significantly in the stock over the complete
event window and always look for opportunities to get the best return to their
investment.

14 | P a g e
Share Price Return & Volume Volatility

It is important to mention here that the share price return model and the volume
volatility calculation model should be used by an investor together. Looking at just
one of these models may not help the investor to make proper decision. The share
price return is always identified as the single most important factor based on which
the investor makes his decisions for the market. But volumes are the indicator of
the price changes that may occur in the near future. The price return is generally in
sync with the volume volatility. On careful study of the share market, we will
always find the price movement has always been complemented with a significant
volume change. The price and volume movement scenario can be understood for
the events if an investor looks into both price return and volume volatility together.

In the event type “buzz” the investor can make a decision on when he should make
his investment to get the best results keeping in mind the volumes being traded in
the market.

15 | P a g e
References:
Financial Econometrics, Damodar N. Gujarati, Introduction, Pgs – 8 – 15.
Financial Econometrics, Damodar N. Gujarati, The Nature of Regression
Analysis, Pgs – 22 – 36.
Financial Econometrics, Damodar N. Gujarati, Two Variable Regression
Analysis: Some Basic Ideas, Pgs – 47 – 55.
Financial Econometrics, Damodar N. Gujarati, Two Variable Regression
Analysis: The Problem of Estimation, Pgs – 63 – 80.
Financial Econometrics, Damodar N. Gujarati, Classic Normal Linear
Regression Model, Pgs – 112 – 117.
Financial Econometrics, Damodar N. Gujarati, Two-Variable Regression:
Interval Estimation and Hypothesis Testing, Pgs – 124 – 151.
Financial Econometrics, Damodar N. Gujarati, Extensions of the Two-
Variable Linear Regression Model, Pgs – 169 – 172.
Financial Econometrics, Damodar N. Gujarati, Multiple Regression
Analysis: The Problem of Estimation, Pgs – 208 – 211, 218 – 228, 237 –
239.
Financial Econometrics, Damodar N. Gujarati, Multiple Regression
Analysis: The Problem of Inference, Pgs – 263 – 266, 277 – 279.
Financial Econometrics, Damodar N. Gujarati, Dummy Variable Regression
Models, Pgs – 308 – 317, 321 – 323.
Financial Econometrics, Damodar N. Gujarati, Multicollinearity: What
Happens if the Regressors are Correlated?, Pgs – 345 – 358, 363 – 367.
http://www.eventvestor.com/index.php
www.finance.yahoo.com
www.finance.google.com
http://www.cboe.com/micro/vix/vixwhite.pdf
Research Papers in Applied Finance by ICFAI Journal ; Investment Horizon
and Volatility by T P Madhusoodanan
Research Paper in “Vindication of The Event Driven Investment Strategy”
by Adam T. Samson
Research Paper on “The Impact of CEO Turnover on Equity Volatility” by
Matthew J. Clayton, Jay C. Hartzell, Joshua V. Rosenberg
Research Paper on “Econometrics of Event Studies” by S.P. Kothari
Research Paper on “News Events and Price Movements. Price Effects of
Economic and Non-economic publications in the mass media” by Thomas
Schuster

16 | P a g e
Research Paper on “What is news? What firm-specific information releases
drives market prices?” by Paul Ryan and Richard J. Taffler
Research Paper on “Is More News Good News? Media Coverage of CEOs,
Firm Value, and Rent Extraction” by Bang Nguyen-Dang
Research Paper on “Market Dynamics and Stock Price Volatility” by
Honggang Li and J. Barkley Jr.
Research Paper on “Event Studies: Practical Issues and Solutions” by Cliff’s
Notes dated February 20, 2008.

17 | P a g e

Вам также может понравиться