The Efficient Mar-

ket Hypothesis

McGraw-Hill/Irwin 8-1
Copyright © 2010 by The McGraw-Hill Companies, Inc. All rights reserved.
Learning outcomes
By the end of the lecture students should able:
To understand EMH
Understand the three forms of EMH
Implications of EMH to arbitragers and Speculaters

8-2
8.1 Random Walks and
the Efficient Market Hy-
pothesis

8-3
Efficient Market Hypothesis (EMH)
Do security prices accurately reflect information?
Informational Efficiency
 _____________________________________
Are price changes consistently predictable?

Allocational Efficiency

Are accu-
prices correct in that they _________
________________________ rately with the
reflect the (relative) valueassociated
security?

 Huge implications concerning the answers to these

questions ~ abnormal profits or waste of resources

8-4
Implications of Efficiency
• Allocational efficiency
– If markets are not allocationally efficient then per-
haps there is a ________________________
role for greater government
___________
intervention in capital markets.
Possible rules changes to attempt to improve
allocational efficiency
– Tax on trading activity
– More taxes on short holding period returns
– Changes in corporate compensation

8-5
Implications of Efficiency
Informational efficiency
If markets are not informationally efficient

Investors may not be able to trust that market

prices are up to date and investors should
then conduct their own research (or hire a
researcher) to validate the price.
Privileged groups of investors will be able to
consistently take advantage of the general
 public.
Active strategies could outperform passive
strategies.

8-6
EMH and Competition
* Competition among investors should imply that
stock prices fully and accurately reflect publicly avail-
able information very quickly. Why?
Else there are unexploited profit opportunities.
* Once information becomes available, market partici-
pants quickly analyze it & trade on it & frequent, low
cost trading assures prices reflect information.
Questions arise about efficiency due to:
• Unequal access to information
• Structural market problems
• Psychology of investors (Behavioralism) ~ irra-
tional explanations like weather may have an
impact on investment performance or people
may be easily fooled, etc
8-7
Random Price Changes
Why are price changes random?

In very competitive markets prices should react to
only NEW information

 Flow of NEW information is random
Therefore, price changes are random

Idea that stock prices follow a “Random Walk”

8-8
Random Walk and the EMH
Random Walk: stock price ______________________
changes are unpredictable
A “pure” random walk implies informational efficiency

Stock prices actually follow a ____________________

submartingale process

Expected price change should be positive over time
But random changes are superimposed on the posi-
tive trend

E(pricej,t) > E(pricej,t-1) t = time period

8-9
 Random Walk with Positive Trend
Security
Prices

Evidence on Random
Walk idea later.

Time
8-10
Forms of the EMH
Prices reflect all relevant information
Vary the ________________
information set
Weak
The relevant information is historical data such
as past prices and trading volume.

If the markets are weak form efficient, use of

such information provides no benefit at the
margin.

8-11
Forms of the EMH
Semi-strong

The relevant information is "all publicly available

information, including the past data and infor-
mation just released to the public."

If the markets are semi-strong form efficient,

then studying studying earnings and growth
forecasts and reacting to news provides no net
benefit in predicting price changes at the mar-
gin.

8-12
Forms of the EMH
Strong

The relevant information is “all information”

both public and private or “insider” information.

If the markets are strong form efficient, use of

any information (public or private) provides no
benefit at the margin.
SEC Rule 10b-5 limits trading by corporate in-
siders (officers, directors and major sharehold-
ers) using the insider information. Insider trad-
ing must be reported.

8-13
Relationships between forms of
the EMH

• Notice that semi-strong efficiency implies weak

_______________________________
form efficiency holds (but _____________)
__________________ NOT vice versa
• Strong form efficiency would imply that
both semi-strong and weak form efficiency hold
__________________________________________.

8-14
8.2 Implications of the
EMH (for Security Analy-
sis)

8-15
Types of Stock Analysis & Rela-
tionship to the EMH
Technical Analysis:
Technical Analysis using prices and volume
information to predict future price changes
TA assumes prices follow predictable trends
If the markets are efficient, will technical analysis
be able to consistently predict price changes?

NO

8-16
Basic Types
Support of
andTechnical Analysis
resistance levels
 Support level:
 A price level below which it is supposedly un-
likely for a stock or stock index to fall.
 Resistance level:
 A price level above which it is supposedly un-
likely for a stock or stock index to rise.

A resistance level may arise at say $31.25 if a stock repeatedly rises

to $31.25 and then declines, indicating that investors are reluctant to
pay more than this price for the stock.
A stock price above $31.25 would then indicate a 'breakout' which
would be a bullish signal.

8-17
Types of Stock Analysis & Rela-
tionship to the EMH
Fundamental Analysis:
using economic and accounting information to
predict stock price changes
Will fundamental analysis be able to consistently
predict price changes?

If the markets are only weak form efficient?

Fundamental Analysis CAN predict price changes
If the markets are semi-strong or strong form efficient?
Fundamental Analysis CANNOT predict price changes

8-18
Fundamental Analysis
Fundamental analysis assumes that stock prices
should be equal to
the discounted value (PV) of the expected future cash
flows the stock is expected to provide to investors.

Fundamental analysis is thus the

“art” of identifying over- and undervalued securities

based on an analysis of the firm's future prospects.

8-19
Fundamental Analysis
Fundamental analysis varies in technique but gener-
ally focuses on
forecasting the firm's future dividends or earnings,

discounting those future cash flows by the required

rate of return (usually obtained from the CAPM),

and comparing the resulting estimated price with the

current stock price.

8-20
Fundamental

Analysis
greater
If the estimated value is ______ than the current price
an investor should ___ the
buystock since it is ___________
and since its price should ________ (eventually)
undervalued increase to the
"true" or "fundamental" value uncovered by the ana-
lyst.

If the estimated price is ____ than the current price

less sold short) because the
the stock should be ____(or
sold
stock is currently __________ by the market.
overvalued
In either case if the analyst is correct the investor
should receive an ________________.
“abnormal return”

8-21
Fundamental Analysis
For FA to add value, your forecast must be better than
the consensus forecast.
Not enough to find a good company, you must find a
company that is better than others believe, i.e., mis-
priced.

8-22
Implications of Efficiency for Active or Pas-
sive Management

Assumes inefficiency,
Active Management use technical and/or
 fundamental analysis
Security analysis to pick securities

Timing strategies

Investment Newsletters
Passive Management
Consistent with semi-
 strong efficiency
 Buy and Hold portfolios
Index Funds or Index ETFs
(Exchange Traded Funds)

8-23
 Market Efficiency and
Portfolio Management
Even if the market is efficient (=> “the price is
right ” & you cannot beat the market ex-ante),
a role still exists for portfolio management
 Identify risk & choose appropriate risk level
 Tax considerations
 Other considerations such as liquidity needs or di-
versify away from the client’s industry.

Aside) EMH=> The price is right and you get what

you paid for. No ripoffs & No great deals! However,
you still need to know what you want (e.g., read
consumer reports, etc). You don’t want to buy a car
with A/C in Alaska. 8-24
8.3 Are Markets Efficient?

8-25
 Empirical Tests of Informational
Efficiency
Recall that over time stock prices tend to follow a
submartingale process
____________________ => a randomly chosen portfolio
of securities is expected to have a positive return.

8-26
 Empirical Tests of Informational
Efficiency
This means that when trying to figure out if some
portfolio manager is earning abnormal returns we
compare their performance to a
try to ____________________________
______________________ in an ex-ante sense.
randomly chosen portfolio
I.E. they must ____________________________, or in
outperform the random portfo-
practice they mustliobeat ____________________
some benchmark rate
of return
_______ on a consistent basis.

8-27
 Empirical Tests of Informational
Efficiency
a. Can anyone consistently earn an abnormal return?

b.
Do investors systematically misinterpret informa-
tion?
This says that EMH=> investors do not repeat the
same mistakes over and over in an irrational fash-
ion.
For example, sometimes they may overestimate the impact
on earnings of some event and sometimes they underesti-
mate the impact on earnings but on average the estimates
are unbiased.
8-28
Empirical Tests of Informational Efficiency
Event studies
Examine how quickly information is integrated into prices
around an informational event.
EMH suggests a rapid assimilation of information into prices.

Assessing performance of professional managers

Can professional managers, using their resources and tools,
“beat” the market after considering risk?
EMH suggests professionals will not outperform the market.

Testing a trading rule

Testing whether a rule that uses available information can
earn abnormal returns after considering the risk and cost of
using the rule.
EMH suggests that such rules will not work.

8-29
How Tests Are Structured
1. Examine prices and returns around some material
announcement

8-30
 Individual Abnormal Returns Surrounding
the Event in an Efficient
_______________
Market
= abnormal
return

-t 0 +t

Announcement Date
Earnings above forecast for example

8-31
 Individual Abnormal Returns Surrounding
the Event in an Inefficient
_______________
Market
= abnormal
return

-t 0 +t

Announcement Date
Earnings above forecast for example

8-32
How Tests Are Structured (cont.)
2. How do we determine if the returns are abnormal?
Market Model approach
a. rt = a + b(rindex,t) + et
Estimate a and b coefficients
b.
Abnormal Return or AR = (Actual - Expected)
ARt = et = Actual return – [a + b(rindex,t)]

8-33
How Tests Are Structured (cont.)
2. continued:
c. Cumulate the abnormal returns over time:
Add up the ARs over time

-t 0 +t
In this case there appears to be information leakage
before the announcement date (day 0), is this a viola-
tion of EMH? Maybe, but wait!
8-34
Issues in Examining the Results
Magnitude Issue
small changes in performance
Even __________________________ ___ may be worth-
while for managers of large investments.
Eg. $5 billion dollar portfolio. Use research to improve results
by 1/10 of a percent (or 10 basis points) per year
= $5 million in value.

The problem is that these __________________________

small changes in performance
would be virtually
_________________
impossible to measure since the stan-
dard deviation of many portfolios ______________.
is 20% or more

8-35
Issues in Examining the Results
(Sample) Selection Bias Issue
“I have this foolproof new trading scheme that will
make me millions. I want to tell everyone about it.”
We rarely hear about the schemes that don’t really
work!

Lucky Event Issue

If 10,000 people flip fair coins 50 times we can ex-

pect some people to flip 75% or more heads. => Do
they have special skills to produce such results?
In a large group of stock analysts, some will be correct
most of the time in their picks, and they will look very
smart even though their results are due to a pure chance!
8-36
Issues in Examining the Results
Possible Model Misspecification
 Results have to be adjusted for the risk of the
given stock or strategy.
 This means that tests of efficiency are necessarily
joint tests of the model (e.g., CAPM) used to mea-
sure risk and market efficiency.
 Results counter to efficiency may be due to the
fact the right model was not used to measure
the risk and hence the expected return.

8-37
Counter Evidence: Some Apparent Predictors of
Broad Market Returns
Fama and French
Aggregate returns tend to be higher for firms with
higher dividend yields
Campbell and Shiller

Aggregate returns tend to be higher for firms with
higher earnings yields
Keim and Stambaugh
Changes in bond credit spreads can predict
market returns
Each of these may also be consistent with
changing risk premiums (=> failing to capture
the “correct” risks) and may have nothing to say
about market efficiency. 8-38
Bubbles and Market Efficiency
Periodically stock prices appear to undergo a ‘speculative
bubble.’
A speculative bubble is said to occur if prices do not equal the
intrinsic value of the security.

Does this imply that markets are not efficient?

 Very difficult to predict if you are in a bubble or not
and when the bubble will burst.
 Stock prices are estimates of future economic perfor-
mance of the firm and these estimates can change
rapidly.
 Risk premiums can change rapidly and dramatically.

8-39
Bubbles and Market Efficiency
With hindsight it appears that there are times when
stock prices decouple from intrinsic or fundamental
value, sometimes for years.
Prices eventually conform to intrinsic value. => EMH

8-40
Summary: What Does the Evidence Show?

Technical Analysis (TA)

Stocks do not follow a pure random walk, so there is hope
for technical trading strategies. => However, it should be
more than a pure random walk, given some positive return
expected.
Most TA rules utilize short term trading strategies that gen-
erate excessive transaction costs and are not profitable.

There appears to be some long term trend reversals.

8-41
Summary: What Does the Evidence Show?

Fundamental Analysis
Appears to be difficult to consistently generate
abnormal returns using fundamental analysis.
This is because the analysis/investment industry is so
competitive.
May help you avoid seriously overvalued invest-
ments.

8-42
Summary: What Does the Evidence Show?

Fundamental Analysis

Without fundamental analysis the markets would

surely be inefficient, &
Abnormal profit opportunities would exist,
Leading to profitable fundamental analysis

Grossman & Stiglitz

AER, 1980
8-43
Summary: What Does the Evidence show?

Anomalies Exist
 Small Firm in January Effect
 Book to Market Ratios
 Long Term Reversals
 Post-Earnings Announcement Drift (Momentum)

8-44
Behavioralism bias
Motivation
Stock prices in the 1990s did not appear to
match “fundamentals,”
e.g., high price earnings ratios
Evidence of refusal to sell losers (Why?)
Economics discipline is exploring behavioral
aspects of decision making

8-45
What does it all mean?
Technical Analysis:
It may be an item in your toolkit
but be careful relying on it too
Your choices much.


Pick stocks yourself, based on fundamental anal-

ysis, but diversify



Pick one or more mutual funds
Unlikely to consistently earn + abnormal returns
Pros paying attention to market and firm conditions

Index or otherwise passively diversify.

8-46
Selected Problems

8-47
Problem 1
Zero, otherwise returns from the prior period could be
used to predict returns in the subsequent period.

8-48
Problem 2
No. Why?
Maybe due to a higher risk?
Maybe due to constant surprises beyond expected sur-
prise earnings.

8-49
Problem 3
Not necessarily, Why?
It could indicate information leakage (=> a violation of
EMH) or it could indicate that splits occur during price
runups (=> still consistent of EMH).

8-50
Problem 4
No, Why?
You won’t get + abnormal returns if the economic cy-
cle is predictable, the news will already be incorpo-
rated in the stock’s price.

8-51
Problem 5
Buy it
This is a “Value Investment” where you believe the
stock will perform better than the market and if you
are correct you will earn a positive abnormal return.

8-52
Problem 6

a. The small firm in January effect

b. Might work, but it might not
i. Doesn’t hold every year
ii. Would lead to “underdiversification”
iii. Higher trading costs of these stocks might wipe out
any gains
iv. Since there is no solid theoretical reasoning for this
the ‘extra return’ might just be a risk premium.

8-53
Problem 7

a. Consistent, expect about half to outperform the

market by chance
b. Violation, earn + AR by investing with last year’s
winners
c. Probably consistent, but it depends. I might be able
to use an option strategy to take advantage of this.

8-54
Problem 7

Violation, you have exploitable price momen-

tum persisting into February

d.
Violation, the reversal offers an exploitable
opportunity, namely buy last week’s losers

8-55
Problem 8

i. Implicit in the dollar-cost averaging strategy is the

notion that stock prices fluctuate around a “normal”
level. Otherwise, there is no meaning to statements
such as: “when the price is high.”

ii. How do we know, for example, whether a price of $25

today will be viewed as high or low compared to the
stock price six months from now?

8-56
Problem 9
The market expected earnings to increase by more than
they actually did.

8-57
 Problem 10

a.  The prices of growth stocks may be consistently bid too high due to in-
vestor overconfidence.
 Investors/analysts may extrapolate recent earnings (and dividend) growth
too far into the future and thereby inflate stock prices, forcing poor re-
turns eventually on growth portfolios.
 At any given time, historically high growth firms may revert to lower
growth and value stocks may revert to higher growth, changing return
patterns, this may happen over an extended time horizon.

8-58
 Problem 10

b. Enough investors should prefer value stocks to growth

stocks and bid up the prices of value stocks and drive
down the prices of growth stocks until the “extra” return
on the value stocks was eliminated.

8-59