Pmlect 2 A
Pmlect 2 A
An ecient capital market is one in which security prices adjust rapidly to the
arrival of new information and, therefore, the current prices of securities reect
all information about the security. This is referred to as an informationally
ecient market. (In other words, an ecient market is a market in which
all transactions have net present value equal to zero). Alternatively, it can be
said that the price of any asset is always equal to its present value, so that the
return for an investment is equal to the equilibrium return for a given level of
risk. All that is required for a market to be ecient is that current market prices
reect available information. If a market is ecient with respect to some piece of
information, then that piece of information can not be used to identify a positive
NPV investment.
The ecient market hypothesis (EMH) asserts that prices for assets are
ecient with respect to available information. The hypothesis implies that no
investment strategy based on current or historical information produces extraor-
dinary large prots. With thousands of investment advisory services, mountains
of information, and millions of investors, the adjustment of prices to new infor-
mation is almost instantaneous.
Assumptions made for the requirements of an ecient market include:
The combined eect of the rst two assumptions means that one would expect
price changes to be independent and random and the expected returns implicit
in the current prices of the security should reect its risk, because security prices
adjust to all new information that is publicly available at any point in time (i.e.,
security prices that prevail at any time should be an unbiased reection
of all currently available information, including the risk involved in owning the
security).
The early work related to ecient capital markets was based on the random
walk hypothesis, which contended that changes in stock prices occurred randomly.
Fama (1970) presented the ecient market theory in terms of a fair game model.
The model requires that the price-formation process be specied in enough detail
so that it is possible to indicate what is meant by fully reect. Available models
equilibrium prices formulate prices in terms of rates of return that are dependent
on alternative denitions of risk. All such expected returns theories of price
formation can be described as follows:
where:
E = expected value operator
Pj;t = price of security j at time t
P j;t+1 = price of security j at time t+1
rj;t+1= the one period percent rate of return for security j during period t +1
t = the set of information that is assumed to be fully reected in the
security price at time t
The equation states that the expected price of security j, given the full set of
information available, is equal to the current price times 1 plus the expected return
on security j, given the set of available information. The set of information should
include all current and past values of any relevant variables such as ination,
interest rates, earnings, GDP, and so fourth. In addition, it is assumed that
this information set includes knowledge of all the relevant relationships among
variables.
Financial economists generally identify three forms of market eciency, based
on the kinds of information which might be expected to inuence stock prices.
2
and transactions by exchange specialists or other unique groups. Since this hy-
pothesis assumes that current market prices already reect all past returns and
any other security-market information, this means that it is pointless to analyze
past prices in an attempt to predict future prices. In other words, past rates of
return and other market data should have no relationship with future rates of
return. Such an evaluation procedure is called technical analysis or (charting).
Weak-from eciency implies that technical analysis can not be used successfully
to forecast future prices and therefore that technical analysts do not earn extraor-
dinary prots. There is a great deal of evidence indicating that nancial markets
are weak-form ecient.
3
2.4. Security Prices and Random Walks
The ecient market hypothesis states that the current stock price fully reects
relevant news information. While some of the news is expected, much of it is
unexpected. The unexpected portion of the news, by denition, arrives randomly
- the essence of the notion that security prices follow a random walk because of
the random nature of the news. Some days the news is good, some days it is bad.
You cannot predict specics of the new with much accuracy. When the news
relevant to a particular stock is good, people adjust their estimates of future
returns upward or they reduce the discount rate they attach to these returns.
Either way the stock price goes up. Conversely, when the news is bad, the stock
price goes down.
Substantial uncertainty even surrounds news that seems reasonably predictable.
An article in Forbes reported the results of a study showing that over the period
1973-1990, the average error made by security analysts in forecasting the next
quarters corporate earnings for the rms they covered was 40%. On an annual
basis, the average error was never less than 25%. From 1985 to 1990, the average
error was 52%, indicating that the analysts forecasting ability had not improved
over the period.
Many people misunderstand what the random walk idea really means. It does
not say that stock prices move randomly. Rather, it says that the unexpected
portion of the news arrives randomly, and that stock prices adjust to news, what-
ever it is. In a famous analogy, a drunk staggers from lamppost to lamppost with
a point of departure and a target destination. The path of the drunk shows a
trend from one post to the next. Along the way, however, the rather is erratic.
The drunk wanders fright and left, perhaps occasionally out into the street or
into a building wall. We cannot determine the precise route in advance. The
same is true of a security price and its consequent return. Over the long run,
security returns are consistent with what we expect, given their level of risk. In
the short-run, however, many ups and downs seem to cloud the long-run outlook.
The evidence on the EMH is mixed. Results from some studies have supported
the hypothesis and indicate that capital markets are ecient, while other results
have not been consistent with the hypothesis and have revealed some anomalies
related to these hypotheses.
4
3.1. Weak-form hypothesis tests and results
Statistical Tests of independence:
The EMH contends that security returns over time should be independent of
one another because new information comes to the market in a random, indepen-
dent fashion, and security prices adjust rapidly to this new information.
Autocorrelation tests of independence measure the signicance of positive
or negative correlation in returns over time. The serial correlation measures the
correlation between price changes in consecutive time periods, whether hourly,
daily, or weekly, and is a measure of how much the price change in any period
depends on the price change over the previous period. Does the rate of return
on day t correlate with rate of return on day t 1,t
2, or t 3? A serial
correlation of zero would therefore imply that price changes in consecutive time
periods are uncorrelated with each other, and can thus be viewed as a rejection of
the hypothesis that investors can learn about future price changes from past ones.
A serial correlation that is positive and statistically signicant could be viewed
as evidence of price momentum in markets, and would suggest that returns in
a period are more likely to be positive (negative) if the prior periods returns
were positive (negative). A serial correlation that is negative and statistically
signicant could be evidence of price reversals, and would be consistent with a
market where positive are more likely to follow negative returns and vice versa.
From the view of investment strategy, serial correlations can be exploited
to earn excess returns. A positive serial correlation would be exploited by a
strategy of buying after periods with positive returns and selling after periods
of negative returns. A negative serial correlation would suggest a strategy of
buying after periods with negative returns and selling after periods with positive
returns. Since these strategies generate transaction costs, the correlations have to
be large enough to allow investors to generate prots to cover these costs. It is
therefore entirely possible that there is serial correlation in returns, without any
opportunity to earn excess returns for most investors.
Those who believe that capital markets are ecient would expect insignicant
correlations for all such combinations. Results indicates insignicant correlations
in stock returns over time, but recently, some studies considered portfolios of
stocks with dierent market values (size) have indicated that the autocorrelation
is stronger for portfolios of small stocks.
Example 3.1. In this example, we calculate the daily returns on TSE300 from
11/16/1987 to 11/14/2002 as log(TSE300t /TS300t1): The returns are graphed
accordingly. Then we generate the following correlation matrix between todays
returns, and those at (t 1), (t2), (t 3), (t 4) and (t5). In other workds, we
5
6
-2
-4
-6
-8
-10
1988 1990 1992 1994 1996 1998 2000 2002
RTSE300
present the correlation matrix of the returns on Monday, with those of Tuesday,
Wednesday, Thursday, Friday of the same week, and Monday of the previous
week. The table shows that yesterdays returns explain about 10% of todays
returns and the coecient is positive (i.e., about 0.108) (note: signicance tests
should be carried further in this case). The matrix is as follow:
6
for independence, you compare the number of runs for a given series to the number
in a table of expected values for the number of runs that should occur in a random
series. If the actual number of runs is greater than the expected number, there is
evidence of negative correlation in price changes. If it is lower, there is evidence
of positive correlation.
Studies have conrmed the independence of stock prices changes over time
using runs tests. These procedures have been repeated for stocks traded on the
OTC market, and the results supported the EMH, however, studies that examined
price changes for individual transactions on the NYSE found signicant serial
correlations. A 1966 study by Niederhoer and Osborne of price changes in the
Dow 30 stocks assuming daily, four-day, nine-day, and 16-day return intervals
provided the following results:
Dierencing interval
Daily Four-day Nine-day Sixteen-day
Actual runs 735.1 175.7 74.6 41.6
Expected runs 759.8 175.8 75.3 41.7
7
2. Event studies that examine how fast stock prices adjust to specic signi-
cant economic events. A corollary approach would be to test whether it is
possible to invest in a security after the public announcement of a signi-
cant event and experience signicant abnormal rates of return. Advocates
of EMH contend that it would not be possible for investors to experience
superior risk-adjusted returns by investing after the public announcement of
any signicant information and paying normal transaction costs.
For any of these tests, you need to adjust the securitys rates of return for the
rates of return of the overall marked during the period considered. A 5 percent in
a stock during the period surrounding an announcement is not meaningful until
you know what the aggregate stock market did during the period and how this
stock normally acts under such conditions. If the market had experienced a 10%
during this period, the 5% return for the stock may be lower than expected.
Studies prior to 1970 recognized the need to make such adjustments for market
movements. That entails substracting the market return from the return for the
individual security to derive its abnormal rate of return, as follows:
8
for a stocks specic risk characteristics and overall market rates of return during
the period.
The time series tests assume that in an ecient market the best estimate of
future rates of return will be the long-run historical rates of return. The tests try
to determine whether there is any public information that will provide superior
estimates of returns for a short-run or a long-run horizon. The results of these
studies have indicated that there is a limited success in predicting short-horizon
returns, but the analysis of long-horizon returns has been quite successful. It is
found that the aggregate dividend yield (D/P) as a proxy for the risk premium on
stocks, indicated a positive relationship between the D/P and future stock-market
returns.
Other studies also found that the (1) default spread (the dierence between
the yields on lower-grade and Aaa-rate long-term bonds); and (2) term structure
or horizon spread (the dierence between the long-term Aaa yield and the yield
on 1-month Treasury bills, can be used to predict stock returns and bond returns
and have even been useful for predicting returns for foreign common stocks.
The two variables -dividend yield (D/P) and default spread - have been sig-
nicant, because when they are high, it implies that investor are expecting a high
return on stocks and bonds, and this occurs when the economic environment has
been poor, as reected in the growth rate of output.
Quarterly earnings studies: the question is, is it possible to predict future
returns for a stock based on a publicly available quarterly earnings reports? Some
studies consistently have failed to support the semistrong EMH.
Standardized Unexpected Earnings (SUE): normalize the dierence between
actual and expected earnings by the standard deviation of the regression estimate
used to derive expected earnings:
1. Some authors found that Big SUEs are associated with abnormal stock
price moves. They found that 51% of the price adjustment occurred after the
earnings announcement.
2. Reinganum conrmed the previous authors results for a longer time period.
Calender Studies: Do seasonal patterns in returns exist?
1. The January Anomaly: downward pressure placed on stocks in late De-
cember and a following rise in prices in early January.
a. Branch: Due to tax selling, investors could buy in December and sell in
January to earn excess prots.
b. Dyl: Stocks that had declined the previous year earned abnormal returns
in January.
9
c. Roll: Pattern is for the last day of December and the rst four trading
days of January. Small stocks seemed to do especially well.
d. Keim: Inverse relationship between size and abnormal returns, with the
strongest results for January, and most of the eect in the rst week.
e. Tinic and West: Strong seasonal relationship between risk and return, most
signicant in January.
Other Calendar Eects:
1. A weekend eect research showed that Mondays returns were signi-
cantly negative.
2. Harris: for large rms the negative Monday eect occurred before the
market opened, while for the smaller rms it was a trading day eect. The only
dierences were in the rst 45 minutes of trading. Prices also tended to rise on
the last trade of the day.
Predicting Cross Sectional Returns
Studies attempt to nd out what types of publicly available information are
useful in predicting future returns.
Price-Earnings Ratios and returns: some have suggested that the mar-
ket tends to overvalue high growth high P/E stocks, while undervaluing low
P/E stocks. Basu found that low P/E stocks had higher returns and lower
risk.
The size eect: Banz, and Reinganu: size and not the P/E ratio accounts
for the eect found by Basu. Some of the problem may stem from the use of
the CAPM to generate the expected returns. Roll found that small rms
measured betas are lower than they should be.
Book Value - Market Value ratio: Rosenberg, Reid and Lanstein found
a positive relation between returns and the ratio of the book value of a
rms equity to the market value of the equity. Fama and French also found
the positive relationship even when other factors were included. Kothari,
Shanken, and Sloan found the relationship between BV/MV and returns may
be periodic and probably not signicant in the long run.
Stock split studies: Fama, Fisher, Jensen, and Roll: found that stock
splits alone should not aect returns because they do not add to rm value.
Any eect would be due to other factors, such as higher adjusted dividend,
which have information content.
10
Initial public oerings: studies found that on average, initial public
oerings are underpriced by 16%. The market seems to respond within
one day to the mispricing. Hanley and Wilhelm found that institutional
investors reap a large part of short-term prots from IPOs.
Exchange listing: despite the potential increase in liquidity and exposure,
listing on a national exchange does not appear to lead to a permanent
increase in rm value. There appear to be prot opportunities around the
time of the announcement, as well as price declines following the listing.
There is no change in systematic risk. Dahran and Ikenberry conrmed
price declines after listing.
Unexpected world events and economic news: Reilly and Drzycimski
found that market adjusts to world events very rapidly. Pierce and Roley:
markets reaction to economic news did not persist past the day of the
announcement. Jain found that the impact of surprises in economic reports
reected in prices in about an hour.
Corporate events: if there is a real economic impact, then prices will
change in the direction indicated by the event. For instance, rms being
acquired in a takeover experience positive excess returns. The adjustment
is fairly rapid.
Summary on the Semistrong-form EMH
Event studies provide heavy support for the semistrong-form of the EMH.
Rate of return prediction studies do not support the EMH. Several anom-
alies were found that still persist.
11
Stock Exchange Specialists: because specialists have a monopoly on
certain types of market information (i.e., the limits book) one would expect
that they might be able to earn excess returns. Several studies support this
contention, though many were done in the early 1970s. More recent studies
show lower rates of return than before.
1. The evidence for corporate insiders and stock exchange specialists sug-
gests that these groups are able to beat the market, in part because of
their monopolistic access to certain information.
2. Some analystsrecommendations, especially to sell, may contain signif-
icant information.
3. Performance of professional money managers, especially mutual fund
managers, is consistent with the strong-form EMH.
Ecient market hypothesis could have some implications regarding the following:
12
Aggregate market analysis: EMH states that past data are not useful in
themselves in selecting superior stocks, so in order to take advantage of long-
run price moves one must do a superior job of identifying and estimating
the relevant variables that cause prices to move.
Industry and Company Analysis: the crucial factor here is to be able to
forecast the relevant variables that aect industry and rm performance
better than the market. The strong-form EMH tests suggest that there may
be superior analysts, and cross-sectional tests suggest that some variables
may have useful information for analysts. For fundamental analysts to do
better than the market is to : (1) identify the relevant variables that aect
a securitys return; and (2) do a better job than average in forecasting those
variables.
Portfolio Management: If one has a superior analysts, one should follow
their recommendations. Analysts should concentrate their energies on tak-
ing advantage of a possible neglected rm eect. In case a portfolio without
superior analysts, then one should measure clients risk preferences, then
build a portfolio that reects those preferences. The portfolio should be
completely diversied and one should be minimizing transaction costs.
The Rationale and Use of Index Funds: the EMH and the lack of
superior performance dictate that many portfolios should just match the
performance of the overall market. Market funds or index funds are designed
to mimic the market composition and performance.
An immediate and direct implication of an ecient market is that no group
of investors should be able to beat the market consistently using a common in-
vestment strategy. An ecient market would also carry negative implications for
many investment strategies:
13
3. In an ecient market, a strategy of minimizing trading would be superior
to a strategy that required frequent trading.
The EMT has some widely known and well-documented violations. Recall that
a market anomaly is any event that can be exploited to produce abnormal prof-
its. Market anomalies imply market ineciency. The following table identies
four categories of anomalies: seasonal, event, rm, and accounting anomalies.
Firm anomalies are anomalies that result from rm-specic characteristics (i.e.,
small rm eect). A seasonal anomaly depends solely on time. For example, the
January anomaly (or January eect). Event anomalies are price changes that occur
after some easily identied event, such as a listing announcement. Finally,
accounting anomalies are changes in stock prices that occur after the release of
accounting information.
14
Anomaly Event Anomalies
The more analysts recommending a stock
Analysts recommendations
the more likely it will go up
Insider trading The more insiders buying a stock, the more likely it is to go up
Prices rise after is announced that
Listings
a rm will be listed on an exchange
Prices continue to rise after Value Line
Value Line rating changes
places a security in its #1 category
Accounting Anomalies
P/E ratio Stocks with low P/E ratios tend to have higher returns
Stocks with larger-than anticipated earnings announcement
Earnings surprises
tend to continuet o rise even after the announcement
Price/sales ratio If the price-to-sales is low, then the stock tends to outperform
If the price-to-book-value ratio is low
Price/book ratio
then the stock tends to outperform
Dividend yield If the dividend yield is high, then the stock tends to outperform
Stocks of rms whose growth rate of earnings
Earnings momentum
is rising tend to outperform
15