“Unexpected changes in corporate earnings reports

and its impact on stock prices”

by
David Rascon

MSc Finance Dissertation.

2008

Bradford, United Kingdom, Class of 2008.

 STATEMENT OF AUTHENTICITY
I have read the University Regulations relating to plagiarism and certify that this
dissertation is all my own work and do not contain any unacknowledged work from any
other sources.

WORD COUNT: 11,707

1
 David Rascon. 07024650
“Unexpected changes in corporate earnings reports and its impact on stock prices”

Abstract.
The goal of this study was to find out if abnormal returns may be gained by using
the information released in the quarterly earning reports. In order to determine this, we
studied the Efficient Market Hypothesis created by Eugene Fama, that affirms in its
strong efficiency form that news are fully and instantaneously priced into stock prices,
and its relevant relationship with this study. In addition, we also studied relevant
academic information provided by Ball and Brown suggesting prices adjust gradually
rather than instantaneously. We were observing stock prices movements around the day
of the event, and analyzed these movements using a window event methodology. After
carefully observing the direction, duration and magnitude of this movements we were
able to establish if the information contained in these reports was priced instantaneously
as suggested by Fama’s Efficient Market Hypothesis, or if the information was priced
gradually, giving enough time to investors to participate in the short time uptrend. The
outcome of the event window methodology gave us solid and strong elements to believe
surprises in earning reports were priced gradually over a short period of time, not
instantaneously. This gradual movement can be seen as an opportunity for investors to
take advantage of this short trend, and gain abnormal returns in this process. Our
conclusions suggest a profitable and consistent strategy for short term trading can be
achieved using this kind of information, and proving at the same time, the market is
efficient but it does not act as fast as the EMH suggests.

Keywords: abnormal returns, earnings reports, window event, trading strategies, EPS.

2
 Preface.
This study analyze the impact of earnings reports on stock prices, by observing
how market participants react to positive, negative and neutral earnings related news in
the short term. The methodology used to measure these changes in stock prices was the
window event methodology. This window is 50 days long, and compiles information of
80 companies traded in the U.S.A. stock exchanges. After watching the behaviour of
stock prices around this type of events it was possible to determine that news are
absorbed gradually rather than instantaneously. This fact is a contradiction to the strong
form of EMH. This gradual change in stock prices takes place in a period of 10 to 20
days, giving investors the chance to gain an abnormal return by buying a portfolio of
equities whose reports were positive. Since creating wealth is the main goal of any fund
manager, and the performance of their portfolios is often compared against the yield of
market indexes, we believe this study is relevant, mainly because it provides a good
strategy to outperform the markets using relevant information, a trading strategy that rely
on a company’s change of intrinsic value due to the change of earnings, a trade fully
based in fundamental analysis.

3
 TABLE OF CONTENTS

STATEMENT OF AUTHENTICITY…………………………… 2
ABSTRACT………………………………………………………... 3
PREFACE………………………………………………………….. 4
LIST OF TABLES…………………………………………………. 5
LIST OF GRAPHICS……………………………………………… 6
ABBREVIATIONS………………………………………………… 7
INTRODUCTION……………………..……………….….……..……… 10

Chapter 1. LITERATURE REVIEW.

1.1 Background……………………………………..…………… 12
1.2 Efficient Market Hypothesis……………..…………………. 14
1.3 Market Equilibrium Models…………..…………………….. 18
1.4 The Ball and Brown study…………….…………………..… 20
1.5 Random Walk Theory……………………………………….. 23

Chapter 2. RESEARCH STRATEGY

2.1 Hypothesis…………………………………………………… 25
2.2 Research Design……….…………..………………………… 25
2.3 Research Data…………………………………..…………… 26
2.4 Sampling……………………………………………………... 27
2.5 Window event methodology………………………..………. 29

4
Chapter 3. ANALYSIS OF DATA
3.1 Window event Analysis and price behaviour……………….. 32
3.2 Cumulative Average Abnormal Return……………………… 33

Chapter 4. CONCLUSIONS
4.1 Introduction…………………………………………………. 37
4.2 Review of Methodology…………………………………….. 37
4.3 Summary of Results………………………………………….. 38
4.4 Final Conclusions…………………………………………….. 39
4.5 Answers to key Questions……………………………………. 39

APPENDIX A. Dissertation Proposal……………………………………. 41

APPENDIX B. Positive surprises, Portfolios 1…….…………………….. 49
APPENDIX C. Positive Surprises, Portfolio 2…………………………... 50
APPENDIX D. Stock Prices, Window Event……………………………. 51
APPENDIX E. Returns per Stock, Window Event……………...………. 58
BIBLIOGRAPHY………………………………………………………. 65

5
LIST OF TABLES.

Table 1.4.1 Beaver et Al Study. Residual changes in EPS and Stock prices.
Table 2.4.1 Sampling, list of companies taken for the window event study.
Table 3.1.1 CAAR. Information taken from day zero to day 25.

LIST OF GRAPHICS.

Graphic 3.1.1 CAAR, per group.

Graphic 3.1.2 Portfolio of companies taken from group #1, positive surprises.

6
ABBREVIATIONS
CAAR Cumulative average abnormal returns.
EMH Efficient market hypothesis.
EPS Earnings per share.
NASDAQ National Association of Securities Dealers Automated Quotation
NYSE New York Stock Exchange
S&P 500 Standard and Poor’s, 500 companies, Market Index
USA United States of America.

7
INTRODUCTION.
Traditional academic literature embraces the study of financial statements as an
important part of the investment decision making process. Investors use considerable
amount of resources analyzing information and historical data of public listed companies.
Many finance professionals agree that sometimes the release of new information affect
the intrinsic value of a given company and causes an intermediate change in stock price
trend. Allegedly, this price trend changes because market participants are discounting this
new information gradually over time.
When discounting this information, some investors use fundamental analysis, this
type of analysis attempt to find value on stock prices trough the study of internal and
external factors, such as earnings, growth, and company performance over time, ratios,
business cycles, monetary policy and global economic trends. There are a large variety of
events that might change investor’s expectations; these events may be related to interest
rates changes determined by central bankers, inflation, GDP, government regulations,
mergers, acquisitions and earnings reports among others.
We are particularly interested on analyzing the changes on stock prices due to
positive, neutral or negative surprises (changes) in quarterly earnings corporate reports.
We believe this type of report is one of the most important pieces of information because
they convey valuable data concerning firm’s activities, cash flows, returns, expenses,
financial performance.

8
 We will analyze return and abnormal return data in the days around corporate
quarterly earnings announcements. We will be observing the direction, speed and
magnitude of changes due to positive, neutral and negative surprises. The intention of this
study is to determine if a trading strategy may be built by trading stocks that are affected
by earning surprises on the day of the announcement, analyze and quantify the possible
returns of this strategy. Event window methodology will be employed in this study in
order to determine returns and abnormal returns around the announcement day. The
sample will be taken out of public companies traded in the NYSE, Nasdaq and Amex.
Taking into consideration comments made by well known financial market
professionals and practitioners on the financial press, we have reasons to believe the
market is not efficient, and many opportunities to gain abnormal returns are available
when these announcements are made. Hopefully we will be able to determine if the
release of new information has some value, or it has already been discounted by the
markets by the time this information is known.

9
1. LITERATURE REVIEW.
1.1 Background.
At the end of every quarter, companies listed in the organized markets have the
obligation to report their performance to the public (investors). Included in earning
reports are items such as, gross and net income, sales, earnings per share among some
other key metrics. The majority of the companies file this report on the months of
January, April, July and October. These reports are widely observed by all market
participants.
These reports contain relevant information of the company’s performance within
that period, Financial statements are provided to investors so they can observe the ability
of the company to gain sustainable profits and growth. The balance sheet, the income
statement and the cash flow statement can provide valuable information that help
investors to take educated guesses about the company future. Some of the most followed
parameters used to measure performance are the gross profit margin ratio, net profit
margin ratio and EPS. EPS is calculated as follows.

Investors watch for these results and they usually compare them against the
expectations set by the consensus of professional analysts. If the reports of the company
surprise the investment community by providing better than expected results, the price of
the stock will usually go higher. In the other hand, if this information reflects poor
performance of the company and the results are disappointing and behind analyst’s
expectations the most likely scenario is that the price of the stock will go down. These
positive and negative surprises are powerful triggers for short term stock price
movements.

10
 Even if the results of a company are strong and robust, but lower that the so called
“consensus of analysts”, prices of stocks will go down. If the company provides very
disappointing results but they are above consensus, the stock will go up. For example
Ford Motor in April 2008, analysts were forecasting an EPS of -$.60, and the result was
-$0.09. The company loose money, but they loose less money than previously estimated,
stock prices went up during the following 20 days.
Earnings surprises may occur because of different reasons. First, it is very
difficult to make an accurate forecast since all companies are subject of hard to predict
forces and economic changes. Sales, costs, exchange rates and other factors and
conditions may change abruptly, making this process even harder. Second, most of the
analyst set estimations too high at the beginning of the year, and spend the rest of the year
making downward adjustments. This is attributed to the fact that most analysts work for
big brokerage firms, and brokerage firms need to encourage investors to buy stocks, and
in order to push the investor to buy, they try to set high expectations, high rewards.
Another factor is the herding effect, most of the analyst will try to make close
calls to their peers, because of the fear of being wrong, and the “herd” is not always right.
How fast this information will be reflected on stock prices once the “new” information is
released is one of the questions of this study. Before we analyze the data we will review
the Efficient Market Hypothesis in order to understand how the markets are supposed to
work from the academic perspective.

1.2 Efficient Market Hypothesis.

The efficient market hypothesis was developed by Professor Eugene Fama at the
University of Chicago in the early 60’s. The efficient market hypothesis assumes that
11
financial markets are "informationally" efficient, meaning that all known information is
already reflected in the prices of stocks, bonds, currencies. This hypothesis states that is
not possible to outperform the market by using information the market already knows. In
this case information is defined as anything unknowable that may affect prices,
information that may appear randomly in the future.
We are talking about unpredictable information such as geopolitical events,
change in regulations, corporate events, earnings announcements, changes in monetary
policy, natural catastrophes, etc.
The EMH requires market participants, investors, to have rational expectations.
And that the majority of them update their expectations according to news whenever new
information is released or publicly known. This hypothesis also states that any of the
investors might be wrong, but the market as a whole cannot be wrong. It is important to
make the distinction that EMH does not require investors to be rational, but to have
rational expectations, when investors are faced with new information some of them may
under react and some of them may overreact, these reaction should follow a random walk
and follow a normal distribution pattern so that the net effect on the market prices cannot
be used in a reliable form to make abnormal profits.
The efficient Market Hypothesis, as discussed by Coutts and Sheik (2002) can be
formally stated:

Where is the information set available at time , P*is the expected price based upon the
information of , sp P* is not correlated with variables in the information set so that:

Assuming a zero constant equilibrium return and risk neutrality. Test for the
market efficiency theory are focused on whether or not the forecasted error is
uncorrelated with the variables contained within the information set . In an efficient
market the time series of stock prices are uncorrelated with the variables in the

12
information set . The EMH is subdivided in three forms, strong form efficiency, semi-
strong form efficiency, weak form efficiency.
In the strong form, share prices reflect all public and private news and there is no
chance of making abnormal returns or outperform the market using news, it also implies
there is no group of investors with access to private information that will allow them to
gain abnormal returns. In this kind of efficiency there is no fund manager who can
outperform the market during long periods of time in a consistent manner. Out of
thousands of investors some of them will earn abnormal returns or outperform the
market, but it is because of statistical reasons, not financial decisions.
Researchers have tested this form of efficiency by analyzing the performance of
four major groups of investors. These four groups are, company insiders, stock exchange
specialists, equity analysts, and fund managers. It has been demonstrated in different
studies, that corporate insiders consistently enjoyed abnormal return profits, apparently
this type of investors , on average, were sell their stock before prior to low return
earnings announcements, and hold their investments prior strong announcements.
Stock exchange specialists have monopolistic access to important information
about unfilled limit orders, and they are able to gain abnormal returns using this
information, they have a tendency to outperform the market indexes return. Security
analyst’s opinions are useful, they have a tendency to outperform the market, because
some of them have the ability to select undervalued stocks, there are studies that prove
top 100 ranked companies made by a consensus of analysts have a tendency to
outperform the market.
Last, Fund managers, studies indicated only 33% of fund managers can
outperform the buy and hold strategy, this give us reasons to believe that money
managers, in average, often find difficulties to gain abnormal returns, this information
supports the EMH and suggests all information is priced instantaneously.
The semi-strong form efficiency affirms that all information is digested by the
market very rapidly, with no chances of gaining excess returns by using this information.
It also states that fundamental analysis and technical analysis are not reliable tools that
can help a fund manager to make excess returns. The adjustments of prices because new
information is almost instantaneous in this efficiency form and these adjustments have a
reasonable and sizeable movement.
13
 Studies that have tested the semi strong form of EMH can be divided by two,
studies that predict future rates of returns using prices and trading volume, these studies
can involve either time series analysis of returns and cross section distribution of returns.
This type of study advocates the idea that it is not possible to predict future returns using
past returns or by predicting the distribution of returns. The second study used, use event
study methodology, most of them imply the information is adjusted rapidly giving no
chance to profit abnormal returns from this news.
Last, the weak efficiency form states historical share prices won’t help in the
process of picking up “winners”, there are no patterns for prices in assets, so Technical
analysis has no use on investment decisions, but some forms of fundamental analysis can
be used to make an excess return strategy. This form implies all movements in prices
follow a random walk and future movements are determined by information that will be
revealed unexpectedly in time and form.
Two major statistical test have been employed to test the independence of
securities returns over time, The run test have confirmed the independence in changes of
stock prices over time, while the autocorrelation tests measure the significance of positive
or negative correlation over time, those who believe the market s are efficient would
expect a insignificant correlations for such combinations.

Some finance practitioners, market commentators and economists disagree with

these forms of efficiency hypothesis, they believe strong efficiency cannot be achieved
due to some factors as slow diffusion of information, the great power of some investors,
such as institutional investors, the existence of the so called sophisticated investors. The
way the market reacts to news is not absorbed instantly according to them. For example a
decrease of the interest rate made by a central bank can change prices in assets,
movements in prices that can last, hours, even months. These types of news are not
absorbed instantaneously, allowing market participants to gain excess returns thanks to
this unexpected news.
A participant of the market that have inside information, or public information
before is released to the rest of the market need to act in a very discreet manner in order
to not be caught by regulators, so it can be assumed the people with this information have

14
influence in prices before news are released to the rest of the public. The most important
discrepancy of this hypothesis is related to behavioral finance.
The EMH in its strong form states that there is no fund manager capable of
outperform the market during long periods of time in a consistently manner. We believe
this is totally wrong, unreal. Some fund managers have the ability of recognizing a trend
in its early stage and overweight their investments in the industries they believe will be
most benefited. Or vice versa, underweight those industries where growth does not look
good. Good fund managers should be able to recognize threats and opportunities, and
they do. We can also make the case that is not that difficult to recognize industry leaders
and invest in them.
The semi-strong form affirms all information is digested by the market almost
instantaneously, we believe this is not true, we will try to prove it by analyzing the price
movements of the stock prices, we have reasons to believe these new events are
discounted over a period of 10 days, not instantaneously, giving the chance to any
investor to gain abnormal returns over short period of time, allowing the investors to
create short term high profitable trading strategies.

In the last stages of a bull market, investors start to ignore fundamentals and
become more irrational, markets are driven by investors that don’t care about
fundamentals, that are just following the trend, and want to be part of the easy money,
ignoring P/E ratios, price to book ratios and some other relevant information that can
alert them against “bubbles”. Behavioral finance states that investors are not motivated
by valuations, they are most of the time affected by expectations, they take decisions
based on their beliefs of where the market is going to, their expectations. EMH assumes
investors are rational at all time, and this might not be possible.
Another contradiction to the EMH is the fact that there are some famous well
known investors as Warren Buffet, Jim Rogers and Peter Lynch among others, which
have been able to consistently outperform the markets over long periods of time
exploiting inefficiencies in the market, and making money when the market is not giving
the correct price to certain assets.
There is no doubt irrational investment behavior occurs from time to time, one of
the most relevant examples can be the market crash of October 1987, when no important
15
news where released in that month and the indices loose more than 20% of their value.
Hence, EMH assumes market participants are rational, and this might not be true, at least
not at all times.
One more contradiction to EMH is that investors tend to over react or under react
to news and events, mispricing information in this process, giving us the opportunity to
doubt again the EMH. DeBondt and Thaler (1985) found that past winners tend to be
future losers and vice versa. And they attribute these reversals to investors over/under
reaction. They think investors tend to give a lot of weight on past performance and forget
that company’s performance tend to be mean-revert. Again, investors sometimes are
carried away by expectations mispricing the real value of firms, creating an opportunity
and efficiency in the market.

1.3 Market Equilibrium Models Used to test Market Efficiency.

There are some models of market equilibrium that can be used to test market
efficiency; the following three, are the most important models.
A. Expected returns are positive. This model simply states that expected return of “j”
from time “t” to time “t-1” is always higher than Zero, it is positive. According to
this model, if the market is efficient and this model is correct then any analyst
who forecast a negative return or a recommendation to sell is incorrect. Some
analysts feel that the market is not efficient and this motivate them to expect
negative expected returns on certain equities. Such type of analyst represents a
good source of tests of market efficiency.
B. According to Keown et al (2005)
“Analyst would agree with the proposition that the market always sets
prices so that its assessed expected returns are positive. But they would disagree
with the proposition that the market is efficient.
They feel that in setting prices, the market sometimes neglects relevant
information or draws incorrect inferences from it, so that sometimes the true
expected returns are negative. They feel that they see more information or are
better able to analyze available information than the market. Such analysts
represent a potentially fertile source of test of market efficiency.

16
 If they record the times when they asses’ negative expected returns on
securities, then one can simply compute the returns that are later realized. One or
few such observations are not much evidence for or against market efficiency; but
as a history of the predictions of analyst is built up, a reliable average return for
periods when he assesses negative expected return can be obtained.
If the average is negative and if the sample of predictions is sufficiently
large to make the negative average return a low probability event if true expected
returns are positive, then we can conclude that the analysts is able to identify
period when true expected return are negative.
If we are willing to stick by the model of market equilibrium which says
that the market always sets prices so that its expected returns are positive, then
the predictions of the analysts establish that the market sometimes either neglects
available information in setting prices or analyzes information incorrectly. In
either case, the analysts are living evidence for the existence of market
Inefficiency”.

C. Expected returns are constant. At any given time “t-1” the market sets the price of
a security “j” in such a form that its assessment of the expected return on the
security,
So that expected returns on a security “j” remain constant trough time, then any
analyst that predicted the opposite is wrong. But if the analyst is capable of
predicting that returns will not be constant and the market prove him right, and
then we could have evidence that the market is not efficient. In this case the
equation stated before does not hold, and the market is presumable inefficient.
1.4 The Ball and Brown Study.
Accounting and finance researchers have been analyzing and investigating for
decades, the relationship between earning announcements and prices of equities. A study
by Ball and Brown (1968) provided the first evidence of the adjustments experimented by
stock prices during the release of earnings related news.
Ball and Brown raised the question of whether or not positive unexpected changes
in earning report results were associated with positive changes in stock prices, and
whether or not negative surprises affect stock prices. Even though there are too many
factors that can affect prices of stock, Ball and Brown isolated the effect of earnings by
observing at cross-sectional changes in earnings and the correspondent change in prices.
As stated by Ball and Brown (1968) “it has also been demonstrated that stock
prices, and therefore rates of return from holding stocks, tend to move together. In one
17
study it was estimated that about 30 to 40 percent of the variability in a stock’s monthly
rate of return over the period of March 1944 through December 1960, could be
associated with market-wide effects. Market-wide variations on stocks returns are
triggered by the release of information which concerns all firms. Since we are evaluating
the income report as it relates to the individual firm, its contents and timing should be
assessed relative to changes in the rate of return on the firms stocks net of market wide
effects. The impact of market wide information on the monthly rate of return from
investing one dollar in the stock of firm “j” may be estimated by its predicted value from
the linear regression of the monthly price relative of firm “j” common stock on a market
index of returns”.

This design provided the confidence to assume that all these changes were
attributed only by earning reports and not other factors. The study used a sample of 261
companies over the period of 1946 trough 1968. They used annual earnings
announcements and they classified them in two, favorable and non favorable, using a
simple expectation model. Their findings provide relevant support for the influence of
earnings. Their study shows that the increment or decrement of stock prices was gradual
rather than instantaneous as the efficient market hypothesis would affirm. And also found
that the changes experimented by stock prices follow the direction of the surprise.
A positive surprise cause the stock price to raise and a negative surprise cause the
stock price to drop. Although, they were capable of determining the relationship between
the direction of the movements relative to the direction of the announcements, they stated
they study couldn’t determine the magnitude of this changes, it was not possible to
measure and create a reliable way of forecasting the size of the surprise against the
magnitude of the change.
They raised several questions as well, one of them was how is the market capable
of determining future changes in profits of particular companies, what media was used
for his purpose, and if the market knows something is going before the day of the
announcement, why they wait until the earnings announcement was released to make a
decision.
According to Hirst and Hopkins (2000) “Specifically, having advance knowledge
that a company will have a positive earnings change yield, on average, a 7% market
adjusted return on that stock. Likewise having advance knowledge that a company will
have a negative change yields, on average, a 9% negative market adjusted return on that
stock”.
18
 Beaver, Clark and Wright (1979) went even further than Ball and Brown, they
decided to investigate not only the direction of the changes but also the size and
magnitude of them, they analyzed a group of companies, and this group was formed by
276 publicly traded companies. They divided this group into 25 portfolios; each portfolio
had companies with different magnitude of change. This study determined there was no
relationship between the magnitude of the changes in announcements and the prices of
stocks, and determined the reaction was not uniform.
We concur with the findings of Beaver, Clark and Wright, that there is no
relationship between the magnitude of the changes in announcements and the magnitude
in changes of stock prices, we believe the direction of the surprise will be the same but
magnitude cannot be established. Mainly because there are several factors that intervene
in the size of the stock price change, factor like future earnings guidance, comments
made by the company CEO, these factors cannot be measured, the reaction can be strong
or weak, There are no educated or scientific ways to measure these factors, reaction may
vary from case to case, giving no opportunity to create a numerical relationship.
The following table shows the average residual change in earnings and the
average residual change price for the top 5 portfolios and the bottom 5 portfolios, this
table is part of the Beaver et al Study. Table 1.4.1

Portfolio #1 was formed with companies that experimented the largest decline in
earnings (unexpected). Portfolio #10 was formed with stocks whose reports had the
largest positive earnings surprise. As we can observe in the table, there is an asymmetric
19
reaction, and these reactions are not uniform for all levels of negative or positive
surprises. Beaver at al did not address what are the reasons behind the magnitude of the
reaction on stock prices to a certain level of change in EPS. This reaction is known as the
“earnings response coefficient” (ERC). Apparently there is no way to determine the size
in the change of stock prices in relationship to the magnitude of unexpected changes in
earnings returns, but there is a directional relationship between them.
It is worth mentioning, that there is a significant difference between the
founding’s established by Eugene Fama and the statements made by Ball and Brown.
Fama assumes all new information is absorbed by market participants almost
instantaneously, a scenario where abnormal returns can be made only if you get lucky,
and an environment where fund managers are not capable of consistently outperform the
indexes. Ball and Brown study shows that price increment or decrement in stock prices
are gradual rather than instantaneous.
Even though the size or magnitude of the change in prices will remain unknown,
the direction of the trend is known, allowing investors to be part of the trend and obtain
abnormal gains. We could make the case that Ball and believed in a very weak form of
market efficiency, a form where the changes take place in days rather than seconds. This
significant difference might be the key to understand why it is possible to outperform the
market in a consistent manner, in a world where information travels instantaneously
trough the internet, but market participants take their time to price this new information.
1.5 Random walk theory.
Random walk theory has been a relevant point concerning the value of historical
data of stock prices, and how these values can predict, or not, the future value of the
securities. The theory has two important axioms, the first one state that price changes are
dominated by some type of probability distribution, and that successive price changes are
independent.
It is important to mention that according to Young E (1971), “considerations of
the type of probability distribution generating the process have been primarily concerned
with portfolio risks where the Markowitz expected return variance concept of an efficient
portfolio does not hold for some distributions”.

The independence assumption is related to the statement that knowledge of

previous prices has no value in predicting future prices. The independence assumption
20
can be represented as follows:
)

In the most elementary sense, the change in price for a unit of time is completely a
random variable, bringing up a scenario where past prices are not helpful to predict future
prices.
In order to put the Random walk theory into perspective, we will describe, in a
brief and general manner, the two major approaches used to predict future prices. These
approaches are known as the Technical analysis approach, and the fundamental or
intrinsic value analysis approach. Technical analysts affirm that prices have memory and
history tends to repeat itself. Patterns can be used to recognize trends and future prices of
stock, this kind of analysts basically assume that successive price changes in stock prices
are dependent, following a sequence of price changes. We could say this kind of analysis
is surrounded by mysticism, a sort of an astrologist of the financial markets, not reliable.
Fundamental analysis approach is assumes that in any point in time an individual
security has an intrinsic value which depends in the earning potential of this security. The
earnings potential are estimated and valued upon the outlook of the industry, the
economy, growth, etc. Using this information an analyst is able to predict future prices.
We could say that fundamental analysis is more accurate because prices of stock tend to
reach their intrinsic value.
Eugene Fama (1965) affirms that “Successive price changes might not be strictly
independent, but the actual amount of dependence might be so small as to be
unimportant”.

Taking into consideration that the future is uncertain, and that news in the future
can be positive or negative, news might change the intrinsic value of a company in a
matter of seconds, we can assume the direction of prices in both directions, is uncertain.
If it is uncertain then a random walk makes sense.
In the other hand, once a new piece of information is released to the market, this
information may affect the intrinsic value of the security forming a short term period of
certainty, causing a trend where random walk not longer applies, at least for the

21
upcoming short term future. In the case of this study, we try to exploit this fact to gain
abnormal return while taking a short term high reward low risk momentum approach.
2.0 Research Strategy.
2.1 Hypothesis.
This project aims to study the influence of surprises in quarterly corporate
earnings reports on share prices. Observe how many days it takes for the market to digest
this information, study the behaviour of prices after the information is known by market
participants. All these observations will be helpful to determine if profits can be made by
buying or selling shares the day of the announcement, and analyze the size of the gains or
losses relative to the neutral, positive or negative surprises. By determining the number of
days it takes the market to absorb the new information, we will be also testing how
efficient and how fast market efficiency really is.
2.2 Research Design.
The research will be carried out by Positivism research philosophy, assuming the
role of an objective analyst with an emphasis on quantifiable observations and statistical
analysis. Inductive research approach will be applied in this study. Inductive approach is
the adequate way to conduct this study since we do not have an initial hypothesis. We
decided to first observe the outcome of the study, analyze it and establish a valuable
conclusion at the end. We will be able to determine if according to this study, gaining
abnormal returns by buying or selling stock the day of the announcement is possible or
not, and if it is possible quantify and measure the gains.
We will follow the cross-sectional studies research strategy, employing
quantitative methods. The software that will be used to process the data is Microsoft
Excel 2007. The study will be a picture, a “snapshot” of the phenomena. We will be using
a group of companies and reports from the first quarter of 2008 the detailed list of
companies and its characteristics are further explained in the following paragraphs.

2.3 Research Data.

22
 It is mainly based in secondary data. Using raw data in the case of historical data
such as stock prices, and compiled data in the case of earnings expectations.
According to Saunders, Philip and Thornhill, (2003) “Documentary secondary
data are often used in research projects that also use primary data collection
methods. However, you can also use them on their own or with other sources of
secondary data, in particular for historical research. Research based almost
exclusively on documentary secondary data is termed archival research and,
although this term has historical connotations, it can refer to recent as well as
historical documents”.
In this sense we will attempt to conduct a research focused in a particular period
of time, the reports of the first quarter of 2008. Most historical data such as stock prices,
estimated earnings per share, and earnings per share reports will be collected from
Thomson markets, Bloomberg, and electronic databases from the New York Stock
Exchange and NASDAQ Exchange. Athens online, Google scholar and the library of the
Bradford School of Management will be the main sources of documentary data applied in
this project.
2.4 Sampling.
Since we can not look at individual cases because prices of shares might change
due to a large number of different factors, a representative sample will be required. First
we determined the size of the sample using a level of confidence of 95% (Z=1.96) and an
error of 1, estimating a total population of 5000. The result was a sample size of 68. We
increased the sample size to 80 to decrease the margin of error. After that we choose a
combination of quota sampling and judgement sampling as a sampling design.
Using a stock screener we first got a group of companies that are listed in the
U.S.A exchanges and trade more than 1 million shares per day. We can say the first
criterion was to take only companies with good marketability, acceptable liquidity. Some
companies where trading is thin, can experiment high fluctuations with low volume. We
prefer not to consider this type of companies for this study. After that, we separated this
big group into four groups. The criterion for each group was market capitalization.
The first group was made with companies that have market capitalization from
$.2 billion to $5 billion; the second one was formed with companies whose market
capitalization lies between $5 billion to $25 billion. Third group companies from $25

23
billion to $75 billion, and the fourth group was formed with companies that have more
than $75 billion in market capitalization.
After making these 4 groups, we took out 20 companies from each group. The
criteria used to take these twenty companies was industry category, each sub-group has
companies from different industries such as consumer discretionary, consumer staples,
energy, financials, health, industrials, information technology, materials,
telecommunications, retail and utilities. In this last phase of the sampling we used
judgement sampling design, getting 20 companies from different industries. We end up
with a sample of 80 companies. We are certain that this is a representative group of the
universe of companies we want to test, which are companies with high liquidity from
different industries and different sizes (market capitalization).
The following table shows a complete list of the companies that are part of the
sample. Table 2.4.1.

SAMPLE. (80 Companies)

Report Date $ $
Company Ticker Market Cap Window EPS EPS
Symbol (Billions, USD) Event = 0 Estimate Actual
1 3m Co. MMM 52.96 24-Apr-08 1.35 1.38
2 Accenture Ltd. ACN 23.71 27-Mar-08 0.56 0.64
3 Advanced Microdevices Inc AMD 4.68 27-Apr-08 -0.51 -0.51
4 Air Prods & Chems Inc. APD 21.66 23-Apr-08 1.2 1.23
5 Alcoa Inc. AA 32.36 7-Apr-08 0.48 0.44
6 Alpha Natural Res. ANR 6.65 5-May-08 0.17 0.39
7 Altria Group Inc. MO 43.14 24-Apr-08 0.37 0.37
8 America Movil S.A.B. AMX 97.81 24-Apr-08 0.88 0.74
9 American Express Co. AXP 51.76 25-Apr-08 0.81 0.84
1
0 American International Group. AIG 84.75 8-May-08 -0.76 -1.41
11 American Superconductor. AMSC 1.79 8-May-08 -0.11 -0.04
1
2 Arena Resources Inc. ARD 1.80 8-May-08 0.45 0.51
1
3 At&T Inc. T 214.89 22-Apr-08 0.74 0.74
1
4 Boeing Co. BA 53.37 23-Apr-08 1.35 1.61
1
5 Bucyrus Int. BUCY 5.81 24-Apr-08 0.49 0.55
1
6 Buffalo Wild Wings BWLD 0.56 29-Apr-08 0.36 0.36
1
7 Caterpillar Inc. CAT 49.74 18-Apr-08 1.33 1.45
1 Cepheid Inc. CPHD 1.62 1-May-08 -0.08 -0.03
24
8
1
9 CF Industries. CF 9.04 24-Apr-08 2.13 2.77
2
0 Chesapeake Energy Corp. CHK 33.10 1-May-08 0.93 1.09
2
1 Chipotle Mexican Grill. CMG 2.88 23-Apr-08 0.48 0.52
2
2 Cisco Systems Inc. CSCO 155.18 6-May-08 0.36 0.38
2
3 Citigroup Inc. C 109.35 18-Apr-08 -0.95 -1.02
2
4 Coca-Cola Co. KO 125.85 16-Apr-08 0.63 0.67
2
5 Concur Technologies. CNQR 1.56 30-Apr-08 0.11 0.13
2
6 CSX Corp. CSX 26.37 15-Apr-08 0.75 0.8
2
7 Cybersource Corp. CYBS 1.20 29-Apr-08 0.14 0.16
2
8 Darling International. DAR 1.31 8-May-08 0.21 0.26
2
9 Du Pont. DD 42.45 22-Apr-08 1.28 1.31
3
0 Exterran Holdings Inc. EXH 4.73 7-May-08 0.68 0.78
3
1 Exxon Mobil Corp. XOM 469.91 1-May-08 2.14 2.03
3
2 Flir Systems Inc. FLIR 5.31 25-Apr-08 0.22 0.24
3
3 Ford Motor Inc. F 14.57 26-Apr-08 -0.6 -0.09
3
4 General Dynamics Corp. GD 34.31 23-Apr-08 1.29 1.42
3
5 General Electric Co. GE 288.76 11-Apr-08 0.51 0.44
3
6 General Motors Corp. GM 9.19 30-Apr-08 -1.6 -0.62
3
7 Hess Corp. HES 41.17 30-Apr-08 2.02 2.34
3
8 Hewlett-Packard Co. HPQ 117.67 20-May-08 0.85 0.87
3
9 Hologic Inc. HOLX 6.00 1-May-08 0.24 0.26
4
0 Home Depot Inc. HD 46.17 20-May-08 0.37 0.41
4
1 Honeywell International Inc. HON 41.66 18-Apr-08 0.82 0.85
4
2 Illumina Inc. ILMN 4.56 17-Apr-08 0.24 0.26
4
3 Intel Corp. INTC 120.66 15-Apr-08 0.25 0.25
4
4 Intl. Business Machines Corp. IBM 174.03 16-Apr-08 1.45 1.65
4
5 Johnson & Johnson. JNJ 185.55 15-Apr-08 1.2 1.26
4
6 JP Morgan Chase & Co. JPM 136.86 16-Apr-08 0.64 0.68
4
7 Juniper Networks inc. JNPR 12.66 24-Apr-08 0.25 0.27
4
8 Lehman Brothers Holding Inc. LEH 15.06 18-Mar-08 0.71 0.82
4
9 Lincoln Electric Co. LECO 3.58 23-Apr-08 1.21 1.24
25
5
0 Lindsay Corp. LNN 1.54 17-Mar-08 0.47 0.79
5
1 Macy's M 7.81 14-May-08 0.02 0.02
5
2 Mastercard Inc. MA 38.28 29-Apr-08 2 2.59
5
3 Mcdonald's Corp. MCD 67.94 22-Apr-08 0.7 0.81
5
4 McMoran Exploration Co. MMR 1.68 17-Apr-08 0.32 0.47
5
5 Merck & Co. Inc. MRK 75.15 21-Apr-08 0.86 0.89
5
6 Microsoft Corp. MSFT 369.43 24-Apr-08 0.44 0.47
5
7 National Oilwell Varco. NOV 34.69 30-Apr-08 1.09 1.11
5
8 Nvidia Corp. NVDA 6.37 8-May-08 0.38 0.36
5
9 Pfizer Inc. PFE 120.22 17-Apr-08 0.66 0.61
6
0 Potash Corp. POT 71.91 24-Apr-08 1.52 1.74
6
1 PPL Corp. PPL 19.02 2-May-08 0.62 0.61
6
2 Priceline Com Inc. PCLIN 4.92 8-May-08 0.6 0.76
6
3 Procter & Gamble PG 201.19 30-Apr-08 0.9 0.92
6
4 Pulte Homes Inc. PHM 2.87 23-Apr-08 -0.77 -2.75
6
5 Red Hat Inc. RHT 4.27 17-Mar-08 0.19 0.2
6
6 Research in Motion Ltd. RIMM 79.82 2-Apr-08 0.7 0.72
6
7 SanDisk Corp. SBDK 5.45 17-Apr-08 0.26 0.21
6
8 Savient Pharmaceuticals. SVNT 1.34 7-May-08 -0.32 -0.33
6
9 Sirius Satellite Radio Inc. Siri 3.93 12-May-08 -0.07 -0.07
7
0 Sprint Nextel Corp. S 23.38 2-May-08 0.22 0.18
7
1 Supervalu Inc. SVU 7.11 17-Apr-08 0.71 0.73
7
2 Sysco Corp. SYY 18.18 28-Apr-08 0.41 0.43
7
3 The Walt Disney Co. DIS 64.11 6-May-08 0.51 0.58
7
4 United States Steel Corp. X 20.34 29-Apr-08 1.81 1.77
7
5 United Technologies Corp. UTX 67.03 17-Apr-08 1.01 1.03
7
6 Verizon Communications Inc. VZ 103.31 28-Apr-08 0.61 0.61
7
8 Wachovia Corp WB 36.48 14-Apr-08 0.4 -0.14
7
7 Wal-Mart Stores Inc. WMT 234.69 13-May-08 0.75 0.76
7
9 Weatherford Int. WFT 30.51 20-Apr-08 0.5 0.51
8
0 Williams Companies Inc. WMB 22.60 1-May-08 0.52 0.57

26
 2.5 Window Event Methodology.
In order to analyze the behaviour of prices and the gains and losses during the
observation period, event study methodology will be employed. This methodology is
generally used to determine the economic impact of different general corporate events.
These events might be related to changes in interest rates, mergers, economic cycles,
earnings. In this case, the goal is to determine the influence of quarterly earning reports
on stock prices. Under Fama’s efficient market hypothesis, the economic effect should be
fully reflected in the share price the same day of the announcement. The purpose of this
particular paper is to study if this really takes places the day of the announcement, or
during several days after the event.
According to Elton, Gruber, Brown and Goetmann (2007) the event study methodology
needs to follow 8 steps, these steps are further detailed below.
Step 1. Collect a sample of firms that had a surprise announcement.
A sample of 80 companies has been selected using the criteria and methodology
described before; this group of companies will be used for the study. A complete
list of the sample is available for review on appendix, exhibit “A”.
Step 2. Determine the precise day of the announcement and designate this day as day “
“Zero”.
The announcements of the corporate earnings reports take place in different days,
these events occur at different points in the calendar time. “Day Zero” is the day
where the announcement was made. In the case of Pfizer day zero is April 17,
while General Electric’s day zero is April 11. Eexact dates of the announcements
for all companies sampled can be found with the sampling list in the appendix.
This study will use intervals of one day. All reports are from the first quarter
earnings of 2008.

Step 3. Define the period to be studied.

27
 This study will consider 50 days. Twenty five days before the event and twenty
five days after the event. Since the main objective of this paper is to determine if
successful short term trading can be made to gain abnormal returns using
surprises in earnings announcements, we believe 50 days is enough time to
determine if a short term trading can be made. Since we don’t know the results
yet, we also decided to analyze 25 days prior the announcement in the case that
the market get ahead of itself and it might be pricing these surprises before they
actually happen, in the remote case that this info is leaked to some investors
before the earnings are released to the whole public.

Step 4. For each of the firms in the sample, compute the return on each of the days being
studied.
For each firm in the sample, a daily return will be calculated for each of the 50
days of the window using the following formula.

Step 5. Compute the abnormal return for each of the days being studied for each firm in
the sample.
The return of the S&P 500 index will be used as the expected return. The index

model will be used to calculate the abnormal return. Using the following formula.

Step 6.. Compute for each day in the event period the average abnormal return for all of
the firms in the sample.
Once we have determined the return of the each company over the 50 day period, and we
have calculated the abnormal return we will proceed to calculate the average abnormal
return for all of the firms using the following formula.

28
Step 7. Calculate the cumulative abnormal return from the middle of the window (day
zero) to the end of the window period (day 25).
Step 8. Examine and discuss the results.
Once the analysis is performed, the results will be examined and conclusions will
be drawn, and we will be able to establish the influence of these reports on stock price in
current times, where the information flow is really fast and electronic trading is in use.
3.0 Analysis of data.
3.1 Window event Analysis and price behaviour.
After performing the window event study using the methodology previously
mentioned, we found that group #1, formed by those companies whose earnings reports
were positive, above expectations, got a cumulative abnormal return of 7.51% over a
period of 26 days. While companies in group #3 whose reports were negative, behind
expectations, suffered a negative cumulative abnormal return of -3.17%. Companies
where earnings were exactly the same as expected, group #2 gained a 4.75%.
After observing the behaviour of prices around the day of the event, we could see
that news was gradually absorbed by the market. It took days for the market to digest
these reports and fully price the news into stock prices. For example in group #1, where
reports where better than expected, the average gain in the first day of the event was
2.45%, but those companies gained an extra 5.1 % during the following days. Giving us
reasons to believe that news were not fully and instantaneously priced as assumed by the
EMH, and suggesting that buying the stock after the spike of the first day was still a
profitable trade. In group #3 stock prices felt an average of 2.91%, after few days they
totally recovered but finished loosing value at the en of the period, this movement
suggest that after some short period of volatility priced was adjusted finishing with a loss
at the end of the event period. The behaviour of prices in group #2 was quite interesting,
they appreciated during a period of 10 days approximately, after that they started to
decrease, suggesting some market participant over react after the news were released, and
prices got adjusted by the market after this so called overreaction. Some companies that

29
reported positive surprises yielded a negative return and vice versa, apparently the market
in these cases put more attention to the future earnings guidance given by company
officials who made some warnings about future negative or positive expectations for the
next trimester reports.
In the following graphic, we can observe the behaviour of stocks. Graphic 3.1.1

30
 Average Cumulative Abnormal Return
From day Zero to day 25.

Positive Neutral Negative

Day Surprises Surprises Surprises
31
 0 2.43% 1.50% -2.92%
1 3.86% 4.48% -2.93%
2 4.43% 4.38% -2.70%
3 4.26% 4.59% -3.04%
4 4.88% 3.79% -1.86%
5 4.96% 3.44% -1.58%
6 5.07% 4.94% -0.86%
7 5.00% 4.39% -1.20%
8 5.67% 4.93% -1.44%
9 5.67% 5.21% -1.08%
10 5.93% 6.74% -0.26%
11 6.41% 5.55% -0.14%
12 7.09% 6.56% 0.07%
13 7.48% 6.92% -0.06%
14 7.52% 6.53% -0.08%
15 7.31% 7.11% 0.68%
16 7.48% 5.97% -0.06%
17 6.99% 5.79% -0.61%
18 6.88% 4.99% -0.36%
19 7.07% 5.01% 0.19%
20 7.05% 5.19% -0.57%
21 7.67% 4.99% -1.44%
22 7.83% 5.34% -2.64%
23 7.82% 5.69% -3.29%
24 7.27% 5.14% -2.85%
25 7.52% 4.76% -3.17%

Table 3.1.1 CAAR, from day zero to day 25.

This information leads us to believe abnormal returns can be made by using these
earning reports with the proper strategy. Challenging the EMH that assumes fund
managers cannot earn abnormal returns in a consistent basis. This information also
confirms the findings made by Ball and Brown (1968), where they affirm information is
gradually absorbed by the markets, and priced during a period of days rather than
instantaneously as affirmed by Eugene Fama (EMH).

32
 Although stock prices follow a random walk, these prices are influenced by
earnings reports in a significant way, giving us a chance to forecast future movements
and profit from this price “spikes”.

We could make the case the magnitude or size of stock prices movements after
surprises in earning reports remain unknown, but the direction, in average, will follow the
direction of the surprise. Therefore when surprises are positive the direction of the prices
will be to the upside, causing a sudden appreciation of the company stock price, while
negative surprises will cause a downtrend in prices. In the case of neutral news, the
direction was in the upside; our interpretation of this fact is that investors buy these
equities hoping the growth rate will remain intact commanding a premium in the price of
the stock.

In order to verify if abnormal returns can be made using the earning reports , we
built a couple of portfolios, in a randomly manner we took 22 companies out of group
#1 and made two subgroups of 11 companies. Day zero, the day of the event, was not
considered, usually the news are released one day before after the market is closed or
early in the morning, when the market opens prices usually gap higher giving no
opportunity to but cheap, taking this into consideration we decided to built the portfolio
with the closing prices of day zero. This gives more certainty that the entry point will be
easy to reach. The total yield is lower, but is the realistic way to do it. A table containing
the information of the portfolios is available in the appendix #2.

The strategy of creating an equal weighted portfolio by buying the stocks at the
closing price of day zero, and holding it for 25 days gave us an average abnormal return
of 3.34% for portfolio #1 and 3.23% for portfolio #2. Confirming us, a short term
strategy can be built, using the information given by the companies in their earning
reports. Once again, proving information is not digested by the market fully and
instantaneously. Stock prices are adjusted gradually giving us a chance to profit from
earnings news.

33
 The following graphic shows the performance of portfolios in the period between
days 1, one day after the earning report was issued, and day 25, the end of the window.
Graphic 3.1.2.

4.0 Conclusions.
4.1 Introduction.
Financial statements are an important part of the investment decision making
process. Earning reports are one of the most influential pieces of information on equity
values. Positive, neutral or negative surprises in quarterly earnings corporate reports can
cause an abruptly change in prices. Around corporate quarterly earnings announcements
meaningful short time trends take place; these trends follow the direction of the surprise.
Although the magnitudes of these changes are uncertain, the direction of them is not.
34
After analyzing this fact, a new question appears how fast does prices affected by these
surprises are. How long does it take for the market to digest them? And how reliable
these new short term trends are? According to the efficient market hypothesis developed
by Professor Eugene Fama at the University of Chicago in the early 60’s. In its strong
form, The efficient market hypothesis assumes that financial markets are
"informationally" efficient, reflecting instantaneously any news, providing difficulties to
gain abnormal returns from this surprises. In this scenario prices change so rapidly that
give no opportunity to react. But in reality, after analyzing the window event we can
clearly see that although market efficiency exists, the speed of the change is far from
being instantaneous, as stated by the EMH. Thus, giving the opportunity to profit from
these short lived trends, giving fund managers a chance to outperform the market while
gaining abnormal returns.
4.2 Methodology overview.
The results of our study might be contradictory of the EMH when it states that
there is no fund manager capable of outperform the market during long periods of time in
a consistently manner. As our study shows, earning reports might be used to gain
abnormal returns. Fund managers should be able to use this information and outperform
the market indexes, taking advantage of these short lived trends. Or underweight those
companies where earnings were a deception to the market.

Our window event methodology results also give us elements to believe that Ball
and Brown studies are closer to reality. Ball and Brown affirm effect of earnings reports
might be isolated from market wide effects. This design provided the confidence to
assume that all these changes were attributed only by earning reports and not other
factors. Ball and Brown study also shows that the increment or decrements of stock
prices are gradual rather than instantaneous as the efficient market hypothesis would
affirm. And also found that the changes experimented by stock prices follow the direction
of the surprise.
Our findings also show that Random walk theory might be right in the sense that
there is no certain way to know if earning reports will bring positive, neutral or negative
surprises, this uncertainty may cause a random walk, because price can go either way, it
35
will depend in the unknown outcome. But once the outcome is known, a short term trend
might be established giving certainty and destroying the random walk element. In plain
words, once a new piece of information is released to the market, this information may
affect the intrinsic value of the security, forming a short term period of certainty, causing
a trend where random walk not longer applies, at least for the upcoming short term
future.
Although we believe the results of this study are relevant, due to time constraints,
and difficult to access historical data, it was not possible to analyze more trimesters, but
we feel comfortable that this study is representative due to the adequate size of the
sample and the methodology taken.
4.3 Summary of results.
The results of window event study using shows that group #1, companies whose
earnings reports were positive, above expectations, got a cumulative abnormal return of
7.51% over a period of 26 days. While companies in group #3 whose reports were
negative, behind expectations, suffered a negative cumulative abnormal return of -3.17%.
Companies where earnings were exactly the same as expected, group #2 gained a 4.75%.

4.4 Final Conclusions.

Many conclusions can be drawn from this data, this information lead us to believe
abnormal returns can be made by using these earning reports with the proper strategy.
Challenging the EMH that assumes fund managers cannot earn abnormal returns in a
consistent basis. This information also confirms the findings made by Ball and Brown
(1968), where they affirm information is gradually absorbed by the markets, and priced
during a period of days rather than instantaneously as affirmed by Eugene Fama (EMH).

Although stock prices follow a random walk, these prices are influenced by
earnings reports in a significant way, giving us a chance to forecast future movements
and profit from this price “spikes”. The magnitude or size of stock prices movements
after surprises in earning reports remain unknown, but the direction, in average, will
follow the direction of the surprise.

36
 The performance of the two portfolios made out of companies whose reports
where positive, show an abnormal return, in a relative high reward low risk environment,
if we take into consideration the statistical behaviour of prices after such surprises. The
strategy of buying the stocks at the closing price of day zero, and holding it for 25 days
gave us an average abnormal return of 3.34% for portfolio #1 and 3.23% for portfolio #2.
Confirming us, a short term strategy can be built, using the information given by the
companies in their earning reports. Assuming we can do this exercise 4 times per year, we
could be earning a yield higher than 12%, above the average yearly return of the S&P
500, which is 7, 29 %.

4.5 Answers to key questions.

After a carefully analysis of the data, and conclusions were established, we can
successfully respond to the key questions of this research, which are summarized in the
following paragraphs.

A) Why is it important to observe and measure the changes in price experienced by

company’s stock prices when its earnings report is released to the market?
It is important to observe and measure the changes in prices because they give us
information about the behaviour of them around this kind of event, we can say
prices change rapidly but not instantaneously, we can determine a short term
trend, a short period of time where prices don’t follow a random walk, giving
market participants a great opportunity to gain abnormal returns.
B) Can we measure the intensity of these changes?
Although the direction of the price change is directly correlated to the direction of
the surprise, and the average response to these changes was established, we don’t
have sufficient information to establish the relationship of a unit of surprise and a
unit of change. Once the outcome of the earning report is know, we can be certain
of the direction that prices will follow, but the magnitude of changes may remain
uncertain in an individual basis.
C) What is the average change in the stock price of a company whose results were
above expectation against those companies who reported downside surprises? The
average abnormal return of stock whose reports were positive was 7.51% while
the average abnormal return of companies whose reports were negative had a
negative abnormal return of 3.17%.

37
 D) Can we do short term trading and obtain abnormal returns using this information?
What is the risk? Can we measure this risk? Can we measure the reward?
As explained before, the strategy of buying the stocks at the closing price of day
zero, and holding it for 25 days gave us an average abnormal return of 3.34% for
portfolio #1 and 3.23% for portfolio #2. Confirming us, a short term strategy can
be built, using the information given by the companies in their earning reports. In
other to measure the risk, we think more events should be studied, a study that
includes stocks traded in other markets and more events should be tested, in other
to successfully establish the risk involved in this strategy.
E) How fast news is priced into stocks in the year 2008?
Our study showed evidence that prices move gradually rather than instantaneously.

Appendix 1. Dissertation Proposal.

Dissertation Proposal
Academic Year: 2007 – 2008.
Programme: MSc Finance

“The impact of unexpected changes in corporate quarterly earnings

reports on stock prices”

Canditate: David Rascon

Supervisor: Wenxuan Hou

38
 Statement of Authenticity
Total Words: 1,643

“I certify that this assignment is the result of my own work and does not exceed the word count noted
above.”

Signature:
David Rascon

Bradford, United Kingdom, May 26, 2008.

39
40
1.0 Introduction.
All public companies listed in organized stock exchanges are subject to certain
regulations, among these regulations we find that these companies are required to provide
financial statements on a quarterly basis, being the income statement report the most
important of them. The purpose of releasing financial statements is to provide
information about the company’s financial performance, its strengths and financial
positions. Market participants evaluate the company’s stock price, trends of growth and
future expectations depending on the content released on this reports. When this
information is released the market immediately tries to price this news into the stock
price. If news were considered to be neutral, meaning the results are very close to the
results expected by investors the change in the price of the stock will not vary much, but
in the cases when these news or results are way below or above market expectations,
share prices will suffer drastic and unusual changes in price. There are plenty of studies
that evaluate this variations, but its worthwhile to re-examine and re-evaluate these
changes because of the new environment that we are experiencing in recent times, times
where information flows more rapidly and easy trough the internet, times where
electronic trading and quantum trading practices are growing. Now, transaction costs are
very low also, we will try to find out if it is possible to profit from these announcements
in a repetitive basis, and how profitable can it be.

1.1 Research objectives.

41
 A) The objective of this study is to observe and measure the changes experienced by
stock prices during its quarterly earning report announcement.
B) Evaluate the intensity of these changes. Under reaction and over reaction of stock
market participants.
C) Compare the changes of prices of companies whose results were close to
expectations against to those companies whose results were different than
expectations.
D) Determine if an abnormal return can be obtained by using this information.

1.2 Research key questions.

F) Why is it important to observe and measure the changes in price experienced by

company’s stock prices when its earnings report is released to the market?
G) Can we measure the intensity of these changes?
H) What is the average change in the stock price of a company whose results were
close to expectation against those companies who reported significant
upside/downside surprises?
I) Can we do short term trading and obtain abnormal returns using this information?
What is the risk? Can we measure this risk? Can we measure the reward?
J) How fast news is priced into stocks in the year 2008?

2.0 Literature review.

In the development of this study, the following literature will be relevant.
1.1 Eugene Fama and his efficient market hypothesis; financial markets are
“informationally efficient”. Meaning that prices of traded assets are already
reflecting all available information. The efficient market hypothesis also declares

42
 that is not possible to consistently outperform the market by using information that
the market already knows. There are some controversial issues related to the
efficient market hypothesis. For example how fast this information can be delivered
to investors and the way that markets participants react to this information is also a
flaw. Investor might under/over react to this information. In the other hand,
behavioural finance states that when entering positions market participants,
investors, are not motivated by weather the fundamental valuation and price looks
cheap or expensive, but weather they are expecting the market to go higher or
lower.

2.2 The Ball and Brown (1968) study; this research study established the first
evidence of the adjustments of stock prices due to earning announcements, the size of the
sampling was two hundred and sixty one companies over the period of twenty two years
(from 1946 to 1968), in this study each event was classified as favourable or not
favourable, using an simple expectation model. Ball and Brown concluded in their studies
that only 10% to 15% of the information was not anticipated by the investors. In a
research made in 1972 by Kiger, evidence was presented showing substantial change in
volume and significant reaction to quarterly earnings announcements, this study was
made using a sample of securities traded in the New York Stock Exchange. Another study
made by May (1971) compares changes in prices on earning reports season against a
season with no reports made. His results show that the intensity of the change in stock
prices was higher on earning report season than no report season, suggesting more
volatility when earnings reports are released.
2.3 Joy, Litzenberg and McEnally (1977). Almost all studies related to the release
of quarterly information data state that this information is fully priced in to stock prices
within minutes once the information is known by investors. Or is already priced in,
before this information is released. On this research study they present evidence (over the
period studied) that the changes in earnings reports are not fully priced in, at the time of
release. They prove this changes cause a reaction among investors. Although they use an
event window of twenty six weeks, and this study intends to use a smaller event window,
it will help to compare some of the information.

43
 2.4 Elton, Gruber, Brown and Goetmann (2007). This book provides useful
information and formulas that will be needed to develop the window event methodology,
and information about the correct calculation of abnormal returns and cumulative
abnormal returns.
2.5 Francis, Schipper and Vincent. (2002). this research study tries to prove the
usefulness of earnings announcements, attributing the magnitude of changes in prices to
the magnitude of the unexpected change in earnings. The change in stock price is bigger
when the information released is significantly different to the information expected,
suggesting a correlation between changes in stock prices to the magnitude of changes in
earnings reports against expectations. In their study they calculated beta adjusted
abnormal returns three days before and three days after the announcement. Their test
confirms a positive association between market reaction and earnings announcements.

3.0 Research Methodology.

This study will use and combine deductive an inductive approach. Using
secondary data analysis of relevant archival data, collected from electronic sources such
as Thomson markets, Bloomberg, and other research studies. The collection of data may
include both qualitative and quantitative data.
3.1 Event study will be used in this study. According to Elton, Gruber, Brown and
Goetmann (2007) the event study methodology is as follows:
a) Collect a sample of firms that had a surprise announcement (the event). In the
case of this study we will understand as surprise event those earning reports that
were different (above/below) the consensus of analyst’s expectations. However
those companies whose earnings reports showed no surprises will be also
analyzed in order to compare both results.

44
 b) Determine the precise day of the announcement and designate this day as Zero.
The interval will be one day.
c) Define the period to be studied. This study will consider 20 days around the event.
Meaning ten days before the event and ten days after the event.
d) For each of the firms in the sample, compute the return on each of the days being
studied. for each firm in the sample
e) Compute the abnormal return for each of the days being studied for each firm in
the sample. The return of the S&P 500 index will be used as the expected return.
f) Compute for each day in the event period the average abnormal return for all of
the firms in the sample.
g) Calculate the cumulative abnormal return from the beginning of the period.
h) Examine and discuss the results. Once the analysis is performed, the results will
be examined and conclusions will be drawn, and we will be able to establish the
influence of these reports on stock price in current times, where the information
flow is really fast and electronic trading is in use.

3.2 Data sources.

3.2.1 Primary information. The compute of each day return, average abnormal
return, and cumulative abnormal return fro each firm (stock) will be taken from my own
calculations based on the information described below.
3.2.2 Secondary Information. Prices of stocks before and after the earnings
announcements will be taken from Thomson markets and Bloomberg. Earnings estimates
of the companies that are part of the sampling, and their results will be taken from
Bloomberg consensus of estimate earnings reports and tables.
4.0 Schedule.

Time Progress (2008)

Activity May June July August
Dissertation Proposal deadline 30th
Feedback on dissertation proposal 1st to 15th
Preparation trial/pilot 1st to 30th
45
 Analysis and interpretation of Data 1st to 10th
First Draft 1st to 30th
Revision and improvement of dissertation 1st to 15th
Dissertation deadline 20th

4.1 Resources.

This research will be elaborated using Windows programs such as Word 2007 and Excel
2007. The most important sources of information will be, related studies, academic
research, historical data from Thomson markets and Bloomberg.

Bibliography.

- Ball, R. and Kothari, S. Security Returns around Earnings Announcements. (1991). The
Accounting Review, vol 66, No 4. PP. 718-738. American Accounting Association.
- Ball , R., and Brown, P. (1968). An Empirical Evaluation of Accounting Income
Numbers. Journal of Accounting Research, Vol. 6, No2, PP.159-178.
- Brownlee, E.R., Ferris, K.R., and Haskins, M.E.,(1990). Corporate Financial Reporting.
2nd edition. Boston: Donelly and Sons Co.
- Chadwick, L, (2001). Accounting and Finance. Essex: Pearson publication.
- Dongcheol, K.,Myungsun, K.(2003). The impact of earnings Announcements on the
Share Price Behaviour of Similar Type of Firms. The Journal of Financial and
Quantitative Analysis, Vol.38, No 2, PP. 383-398
- Elton, Gruber, Brown and Goetmann, (2007). Modern Portfolio Theory and Investment
Analysis. New York: John Wiley and Sons, Inc.
- Elliot, B., and Elliot, J.,(2008). Financial Accounting and Reporting, 12th edition. Essex.
Pearson Education Limited.

46
- Emery, Finnerty and Stowe, (2007). Corporate Financial Management. New Jersey:
Pearson Prentice Hall.
- Francis, J., Schipper, L. and Vincent, L. (2002). Expanded Disclosures and the
Increased Usefulness of Earnings Announcements. The Accounting Review, Vol 77, No3.
American Accounting Association.
- Graduate Program Office. MA/MSc Dissertation Guidelines. (2007-2008). Bradford
University School of Management.
- Keown, J.A., Martin, J.D., Scott, F.D., and Petty, J. (2005) Foundations of Finance: The
Logic and Practice of Finance Management. New York. Prentice Hall.
- Jhon, J.A, Jhonson, D.G and Whitaker, D., (2006). Statistical Thinking in Business, 2 nd
edition. Boca Raton F.L. Capman and Hall.
- Joy, M., Litzenberg, R. and McEnally, R. (1977) The Adjustments of Stock Prices to
Announcements of Unanticipated Changes in Quarterly Earnings. Journal of Accounting
Research, vol 15, No 2. Blackwell Publishing.

- Saunders, M. Lewis, P., and Thronhill, A., (2003). Research Methods for Business
Students, 3rd Edition, Harlow: Pearson.
- Silver, M.S. (1992). Business Statistics, 2nd Edition, Berkshire, UK. McGraw-Hill.
- Watsham and Parramore, (2007). Quantitative methods in Finance. London: Thomson
Learning.
- Young, E. (1972 Random walk in stock prices: a test of the variance-time function.
Econometrica, Vol 37, No 5. PP 797-812. The Econometric society.

47