Sei sulla pagina 1di 28

Efficient Capital Markets: A Review

of Theory and Empirical Work


Author: Eugene F. Fama
Course: Corporate Finance Theory
Instructor: Prof. S. Sowmya
Indian Institute of Management, Lucknow

1
PAPER OUTLINES
I. INTRODUCTION TO EFFICIENT CAPITAL MARKETS
II. THE THEORY OF EFFICIENT MARKETS
A. Expected Return or “Fair Game” Models
a. The Submartingale Model
b. The Random Walk Model
III. THE EVIDENCE
A. A Weak Form Tests of the Efficient Market Model
B. Tests of Martingales Models of the Semi-strong Form
C. Strong Form Tests of the Efficient Markets Models
IV. SUMMARY & CONCLUSION
2
I. INTRODUCTION (TO EFFICIENT MARKET)
• Role of Capital Market: Allocation of ownership of economy’s capital
stock.
• Ideal Market: Price provide accurate signal for resource allocation.
• Efficient Market: Prices always “fully reflect” available information.
• The Paper: A review of theoretical and empirical literature on
“Efficient Market Model”.
• How security prices get adjusted with relevant information
• Weak form tests: information is just historical prices.
• Semi-strong form tests: whether prices adjust to publicly available information.
• Strong-form tests: whether investors or group have monopolistic information.

3
II. THE THEORY OF EFFICIENT MARKETS
• The definitional statement of efficient market is not testable empirically.
• The process of price formation must be specified in more detail to make
the model testable.
• Most of the available work is based on assumption that “the condition of
market equilibrium can be stated in terms of expected returns”.
• The equilibrium expected return, which is conditional on some information
set, on a security is a function of its “risk”.
• Therefore, a general class of “expected return theories” can be described
as:

4
II. THE THEORY OF EFFICIENT MARKETS…
(A) Expected Return or “Fair Game” Model

• The expected return is a mathematical concept and depends on the


distribution of returns, therefore, might not implied by the notion of
market efficiency.
• Major empirical implication of the assumption (1): It rules out trading
system based only on information in i.e. system that have excess
profit or return:

5
II. THE THEORY OF EFFICIENT MARKETS…
(A) Expected Return or “Fair Game” Model

• Also,

• Here and both are “fair game” w.r.t. information .


• Two special cases of the model (1):
• The Submartingale Model, and
• The Random Walk Model

6
II. THE THEORY OF EFFICIENT MARKETS…
(B) The Submartingale Model

• Here price sequence follows a Submartingale w.r.t. .


• This has important empirical implication: No trading rules, based only
on information cannot have greater expected profit than a policy
of buying and holding the security.
• Why important?: Testing of such rules will provide empirical evidence
on the efficient market model.

7
II. THE THEORY OF EFFICIENT MARKETS…
(C) The Random Walk Model
• It was assumed in the efficient market model that successive price changes are
independent.
• Also, was the usual assumption that successive price changes are identically
distributed.
• When something is independent and identically distributed (i.i.d.), then it
constitutes a random walk model. The model says:

• General model (1) says mean of the distribution of return is independent of .


• Random walk model says the entire distribution of return is independent of .
• Empirical tests of the “random walk” model are in fact tests of “fair game” model.

8
III. THE EVIDENCE
(A) Weak Form Tests of the Efficient Markets Model
1. Random Walk and Fair Games: Background
• In 1900, Bachelier gave the “fundamental principle” for the behaviour of prices. It said that
speculation should be a “fair game”. (later on, the process implied is a martingale)
• There were empirical evidences in middle 1950s and early 1960s, that the behaviour of prices
could be well approximated by a random walk.
• For these evidences, economists provided a theory of efficient markets stated in terms of
random walk, but it usually implied a more general “fair game” model.
• Kendall (in 1953), Working (in 1934), and Roberts (in 1959) suggested empirically that series
of speculative prices may well be described by random walk. But they didn’t attempt to
provide any economic rational for this.
• Samuelson (in 1965) and Mandelbrot (in 1966) provided the model and rigorous theory that
price fluctuation is random.

9
III. THE EVIDENCE
(A) Weak Form Tests of the Efficient Markets Model
2. Test of Market Efficiency in the Random Walk Literature
• Fair game model imply “impossibility” of trading system based on information only.
• Like a random walk, serial covariance of a “fair game” are also zero.
• If is a fair game, then its serial covariance:

• This shows that for all lags, the serial covariance between lagged values of a “fair
game” variable are zero. Thus, observations of a “fair game” variables are linearly
independent.
• But the “fair game” model does not necessarily imply that serial covariance of one-
period return are zero.
• But this is not recognized in the random walk literature as it is assumed in random
walk that the entire distribution of return is stationary through time. 10
III. THE EVIDENCE
(A) Weak Form Tests of the Efficient Markets Model
2. Test of Market Efficiency in the Random Walk Literature
• From Table 1, there is no evidence of substantial linear dependence between
lagged price changes or returns.
• But, there are evidence of statistically significant linear dependence in Table.
• Statistically significant deviations from zero covariances are not necessarily a
basis for rejecting the efficient market model.
• It is also unlikely that small absolute levels of serial correlation that are always
observed can be used as the basis of substantially profitable trading systems.
• So what degree of serial correlation would imply trading rules with profits?
• Alexander using his filter test concluded that filter cannot beat buy-and-
hold and this is a support for the efficient market hypothesis.
11
III. THE EVIDENCE
(A) Weak Form Tests of the Efficient Markets Model
3. Other tests of independence in the Random Walk Literature
• The basic model of market equilibrium is the “fair game” expected return
model, with a random walk arising when additional environmental conditions
are such that distributions of one-period return repeat themselves through
time.
• Therefore, violations of pure independence assumption of random walk are
expected.
• Violation 1 (Osborne and Fama):
• Large daily price changes tend to be followed by large daily price changes. The sign of
successor changes are apparently random. This is a denial of random walk model but not
of the market efficiency hypothesis.

12
III. THE EVIDENCE
(A) Weak Form Tests of the Efficient Markets Model
3. Other tests of independence in the Random Walk Literature
• Violation 2 (Niederhoffer and Osborne):
• Their data indicated that reversals (pairs of consecutive price changes of opposite sign
i.e. and ) are from two to three times as likely as continuations (pairs of
consecutive price changes of same sign i.e. and ).
• Violation 3 (Niederhoffer and Osborne):
• A continuation is slightly more frequent after a preceding continuation than after a
reversal i.e. let indicate occurrence of a positive price change, given two
preceding changes. Then events and are slightly more frequent than
events and .
• But these findings do not imply market inefficiency. Rather entire analysis
based on proprietary trading book imply market inefficiency w.r.t. strong form
tests of the efficient market model.

13
III. THE EVIDENCE
(A) Weak Form Tests of the Efficient Markets Model
4. Distributional Evidence
• Nature of the distribution of the price changes is a burning issue in the
random walk literature.
• Bachelier first proposed a normal distribution of price changes, which from
transaction to transaction are i.i.d. random variables, with finite variance.
• Osborne, Moore, and Kendall all observed fat tails in return distribution i.e.
high proportions of large observations compared to normal distribution.
• Mandelbrot later suggested that these departure from normality can be
explained by a more general form of Bachelier model, in particular if one does
not assume finite variance of distribution of price changes.

14
III. THE EVIDENCE
(A) Weak Form Tests of the Efficient Markets Model
5. “Fair Game” Models in the Treasury Bill Market
• After efficient market models of Mandelbrot and Samuelson, Roll’s work is first weak
form empirical work that is consciously in the “fair game” rather than random walk
tradition.
• We know that the “fair game” properties of the general expected return model apply
to:

• The variation in common stock returns about their expected value is so large that
relative to any changes in the expected value that the latter (changes in the expected
value) can safely be ignored. But this is not the case with Treasury Bills.
• Therefore, to test the “fair game” model on Treasury Bills requires explicit economic
theory for the evolution of expected returns through time.
15
III. THE EVIDENCE
(A) Weak Form Tests of the Efficient Markets Model
5. “Fair Game” Models in the Treasury Bill Market
• Roll used three theories of term structure viz. the pure expectations hypothesis,
market segmentation hypothesis (liquidity preference hypothesis), and market
segmentation hypothesis of Kassel.
• The model is:
• This says that one week “futures” rate for observed from the term
structure at is expected “spot” rate for at plus a “liquidity premium”.
• The conditional expectation required is of the form:

• After extensive testing, Roll concludes that (i) two market segmentation hypotheses
fit the data better, and (ii) that as far as his tests are concerned, the market for
Treasury Bills is efficient.

16
III. THE EVIDENCE
(A) Weak Form Tests of the Efficient Markets Model
6. Tests of a Multiple Security Expected Return Model
• How a security is priced vis-à-vis another security?
• Sharpe and Lintner theory is used and the Sharpe-Lintner model is:

• The stated goal of Sharpe-Lintner work has been to determine the extent to
which the return on a given security are related to the returns on other
securities.
• Kendall found that common stock price changes do not seem to be serially
correlated, but there is a high degree of cross-correlation between the
simultaneous returns of different securities.
17
III. THE EVIDENCE
(A) Weak Form Tests of the Efficient Markets Model
6. Tests of a Multiple Security Expected Return Model
• Fama, Fisher, Jensen, and Rolls (FFJR) tested the following “market model”
suggested by Markowitz:

• The observed properties of the “market model” are consistent with the
expected return efficient market model.

18
III. THE EVIDENCE
(B) Test of Martingale Models of the Semi-strong Form
• Semi-strong form tests: whether current price “fully reflect” all
publicly available information.
• Each test is concerned with the adjustment of security prices to one
kind of information generating event.
• Each test brings the supporting evidence for the model.
• Study of stock splits was the initial major work in this area done by
FFJR.
• There are only few major types of information generating events on
which studies have been done.
19
III. THE EVIDENCE
(B) Test of Martingale Models of the Semi-strong Form
1. Splits and the Adjustment of Stock Prices to New Information
• Split in themselves are not necessarily sources of new information. FFJR
presumed that split may often be associated with the appearance of more
fundamentally important information.
• The approach of FFJR relies on “market model” (12). In this model, if a stock
split is associated with abnormal behaviour, this would be reflected in the
estimated regression residuals for the month surrounding the split.
• The average residual of over all split securities for month m is:
• The cumulative average residuals is:
• They also defined and as cumulative average residuals for the split
followed by “increased” and “decreased” dividends.
20
III. THE EVIDENCE
(B) Test of Martingale Models of the Semi-strong Form
1. Splits and the Adjustment of Stock Prices to New Information
• Plots of , , and :

21
III. THE EVIDENCE
(B) Test of Martingale Models of the Semi-strong Form
2. Other Studies of Public Announcements
• Variants of residual analysis developed by FFJR is used by others to study different
kinds of public announcements, and all of these also support the efficient market
hypothesis:
• Ball and Brown studied the effect of annual earning announcements:
• Used residuals from regression of firm’s earning on the average earnings of all firms to classify
the firm’s earning as “increased” or “decreased” relative to the market.
• Concluded: Only 10-15% information in announcement is not anticipated within a month.
• Waud studied the effects of announcements of discount rate changes by Fed:
• In this case residuals are deviations of the daily returns on the S&P500 index from average
daily return.
• There is statistically significant “announcement effect” on stock return for the first trading day
following an announcement.
• He concluded that market anticipates the announcement.
22
III. THE EVIDENCE
(B) Test of Martingale Models of the Semi-strong Form
2. Other Studies of Public Announcements
• Scholes studied the effect of large secondary offering of common stock and
new issues of stock:
• Average secondary issues are associated with a decline of 1%-2% in the cumulative
average residual returns for the corresponding common stocks.
• This is not due to selling pressure rather results from negative information implicit in the
fact that someone is trying to sell a large block of a firm’s stocks.
• Sometimes corporate insiders have important information about their firm that is not yet
publicly known.
• Thus Scholes’ evidence provide support for the efficient market model in the semi-strong
form sense, but also provide some strong-form evidence against the model.

23
III. THE EVIDENCE
(C) Strong Form Tests of the Efficient Markets Models
• There are contradictory evidences against this model in preceding
discussions. This model of market efficiency is strictly not valid.
• Now the relevant question is:
• Does it pay for the average investor to expend resources searching out little
known information?
• Are these activities generally profitable?
• Who are the people in the investment community that have access to “special
information”?
• Jensen extensively studied the management of open end mutual
funds.
24
III. THE EVIDENCE
(C) Strong Form Tests of the Efficient Markets Models
1. Theoretical Framework (used by Jensen)
• Major goals:
• Whether “in general” fund managers have access to special information.
• Whether some funds are better at uncovering such special information than others.
• Major problem in framework:
• Developing a norm against which performance can be judged.
• Jensen used Sharp-Lintner model of equilibrium expected returns:

• (19) can be plotted as:

25
III. THE EVIDENCE
(C) Strong Form Tests of the Efficient Markets Models
1. Theoretical Framework
• Market line is the naïve
strategy, which investors might
follow.
• The performance of a mutual
fund is then measured relative
to this naïve strategy.

26
III. THE EVIDENCE
(C) Strong Form Tests of the Efficient Markets Models
2. Empirical Results
• Jensen’s findings:
• Funds, “in general”, do not perform well considering loading charges (pure salesman’s
commission), management fees, other costs, and risk level of the portfolio compared to
investing in a combination of riskless asset and market portfolio.
• If we ignore loading charges, even then it seems funds do not perform well.
• There is hardly any evidence of special information among mutual funds managers.
• There are perhaps some individual funds, which performed consistently well
and this is an evidence against the efficient market model. These kind of
funds escaped Jensen’s search.
• Jensen argued that results are striking evidence in favour of efficient markets
model.

27
THANK YOU

28

Potrebbero piacerti anche