Security Analysis and Portfolio Management

Lec 7

What is an efficient market?

An efficient capital market is one
in which security prices adjust rapidly to the arrival of new information and, therefore, the current prices of securities reflect all information about the security this is referred to as an informationally efficient market

4 Assumptions for fficiency

fficient market re!uires that
a large number of profit maximizing participants analyze and value securities, each independently of the others

New information regarding securities

comes to the market in a random fashion, and the timing of one announcement is generally independent of others

Profit-maximizing investors adjust security prices rapidly to reflect the effect of new information.
lthough the price adjustment may be imperfect, it is unbiased" #his means that sometimes the market $ill o%erad&ust and other times it $ill underad&ust, but you cannot predict $hich $ill occur at any gi%en time

'inally, because security prices ad&ust to all ne$ information, these security prices should reflect all information that is publicly a%ailable at any point in time" #herefore, the security prices that pre%ail at any time should be an unbiased reflection of all currently a%ailable information including the risk in%ol%ed in o$ning the security

#herefore, in an efficient market, the expected returns implicit in the current price of the security should reflect its risk, !hich means that in%estors $ho buy at these informationally efficient prices should recei%e a rate of return that is consistent $ith the percei%ed risk of the stock"

 #his scenario implies that informationally efficient markets re!uire some

minimum amount of trading and that more trading by numerous competing in%estors should cause a faster price ad&ustment, making the market more efficient

 random walk hypothesis,

$hich contended that changes in stock prices occurred randomly

'ama di%ided the o%erall efficient market hypothesis ( M)* and theempirical tests of the hypothesis into three subhypotheses
depending on the information set in%ol%ed+ (,* $eak-form M), (.* semistrong-form M), and (/* strong-form M)"

Weak-Form Efficient Market Hypothesis

#he weak-form EMH assumes that
current stock prices fully reflect all security market information, including the historical se!uence of prices, rates of return, trading %olume data, and other market-generated information, such as odd-lot transactions, block trades, and transactions by e0change specialists"

1ecause it assumes that current market prices already reflect all past returns and any other security market information,
this hypothesis implies that past rates of return and other historical market data should ha%e no relationship $ith future rates of return (that is, rates of return should be independent*"

#herefore, this hypothesis contends that you should gain little from using any trading rule that decides $hether to buy or sell a security based on past rates of return or any other past market data"

Semistrong-Form Efficient Market Hypothesis

#he semistrong-form EMH asserts that security prices adjust rapidly to the release of all public information"
that is, current security prices fully reflect all public information. #he semistrong hypothesis encompasses the $eak-form hypothesis, because all the market information considered by the $eak-form hypothesis, such as stock prices, rates of return, and trading %olume, is public" Public information also includes all nonmarket information,
such as earnings and di%idend announcements, price-to-earnings (P2 * ratios, di%idend-yield (32P* ratios, pricebook %alue (P214* ratios, stock splits, ne$s about the economy, and political ne$s"

#his hypothesis implies that in%estors $ho base their decisions on any important ne$ information
after it is public should not derive above-average risk-adjusted profits from their transactions, considering the cost of trading because the security price already reflects all such ne$ public information"

Strong-Form Efficient Market Hypothesis

#he strong-form EMH contends that stock prices fully reflect all information from public and private sources. #his means that no group of investors has monopolistic access to information rele%ant to the formation of prices" #herefore, this hypothesis contends that no group of in%estors
should be able to consistently deri%e abo%e-a%erage risk-ad&usted rates of return"

#he strongform M) encompasses both the $eak-form and the semistrongform M)" 'urther, the strongform M) e0tends the assumption of efficient markets, in $hich prices ad&ust rapidly
to the release of ne$ public information, to assume perfect markets, in $hich all information is costfree and a%ailable to e%eryone at the same time"

#$%#% N& '$%()#% *+ $++,-,$N# . '/$# 01P*#0$%$%

Weak-Form Hypothesis: ests and !esults
5esearchers ha%e formulated t$o groups of tests of the $eak-form M)" #he first category in%ol%es statistical tests of independence bet$een rates of return" #he second entails a comparison of risk-return results for trading rules that make in%estment decisions based on past market information relati%e to the results from a simple buy-and-hold policy, $hich assumes that you buy stock at the beginning of a test period and hold it to the end"

Statistical #ests of 6ndependence

M) contends that
Security returns o%er time should be independent of one another
because ne$ information comes to the market in a random, independent fashion and security prices ad&ust rapidly to this ne$ information"

#$o ma&or statistical tests ha%e been employed to %erify this independence" 'irst, autocorrelation tests of independence measure the significance of positive or negative correlation in returns o%er time" 3oes the rate of return on day t correlate with the rate of return on day t 2 3, t 2 4, or t 2 567 #hose who believe that capital markets are efficient would expect insignificant correlations for all such combinations" Se%eral researchers ha%e e0amined the serial correlations among stock returns for se%eral relati%ely short time hori7ons including , day, 4 days, 8 days, and ,9 days" #he results typically indicated insignificant correlation in stock returns o%er time ALWA:S ;<=S63 5 #5A=SA;#6<= ;<S#S

 #he second statistical test of independence is the runs test" #iven a series of price changes$ each price change is either designated a plus (>* if it is an increase in price or a minus (?* if it is a decrease in price" #he result is a set of pluses and minuses as follo$s+ >>>?>??>>??>>" A run occurs $hen t$o consecuti%e changes are the same@
t$o or more consecuti%e positi%e or negati%e price changes constitute one run" When the price changes in a different direction, such as $hen a negati%e price change is follo$ed by a positi%e price change, the run ends and a ne$ run may begin"

#o test for independence, you $ould compare the number of runs for a gi%en series to the number in a table of e0pected %alues for the number of runs that should occur in a random series" Studies that ha%e e0amined stock price runs ha%e confirmed the independence of stock price changes o%er time

#ests of #rading 5ules

#echnical analysts do not e0pect a set number of positi%e or negati%e price changes as a signal of a mo%e to a ne$ e!uilibrium in the market" #hey typically look for a general consistency in the price trends o%er time" Such a trend might include both positi%e and negati%e changes" 'or this reason, technical analysts belie%ed that their trading rules $ere too sophisticated and complicated to be properly tested by rigid statistical tests

 Ad%ocates of an efficient market hypothesi7ed that in%estors could not deri%e abnormal profits abo%e a buy-andhold policy using any trading rule that
depended solely on past market information"

5esults of Simulations of Specific #rading 5ules

6n the most popular trading techni!ue, filter rule$ an investor trades a stock when the price change e%ceeds a filter value set for it" As an e0ample, an in%estor using a A percent filter $ould en%ision a positi%e breakout if the stock $ere to rise A percent from some base, suggesting that the stock price $ould continue to rise" A technician $ould ac!uire the stock to take ad%antage of the e0pected continued rise" 6n contrast, a A percent decline from some peak price $ould be considered a breakout on the do$nside, and the technician $ould e0pect a further price decline and $ould sell any holdings of the stock and possibly e%en sell the stock short

 Studies of this trading rule ha%e used a range of filters from B"A percent to AB percent" #he results indicated that small filters $ould yield abo%ea%erage profits before taking account of trading commissions" )o$e%er, small filters generate numerous trades and, therefore, substantial trading costs" When these trading commissions $ere considered, all the trading profits turned to losses" Alternati%ely, trading using larger filters did not yield returns abo%e those of a simple buy-and-hold policy #hese results generally support the $eak-form M)

Semistrong-Form Hypothesis: ests and !esults

5ecall that the semistrong-form M) asserts that
security prices ad&ust rapidly to the release of all public information@ that is, Security prices fully reflect all public information"

Studies that ha%e tested the semistrong-form M) can be di%ided into the follo$ing sets of studies+
%tudies to predict future rates of return using available public information beyond pure market information such as prices and trading volume considered in the weak-form tests #hese studies can in%ol%e either time-series analysis of returns or the crosssection distribution of returns for indi%idual stocks"

Ad%ocates of the M) $ould contend that

it $ould not be possible to predict future returns using past returns or to predict the distribution of future returns using public information

 %econd way to test %% $.08 $vent studies that examine how fast stock prices adjust to specific significant economic events. corollary approach would be to test whether it is possible to invest in a security after the public announcement of a significant e%ent and e0perience significant abnormal rates of return"

Again, ad%ocates of the M) $ould e0pect security prices to ad&ust rapidly, such that it $ould not be possible for in%estors to e0perience superior riskad&usted returns by in%esting after the public announcement and paying normal transactions costs

 Ad&ustment for Market ffects

'or any of these tests, you need to ad&ust the securityCs rates of return for the rates of return of the o%erall market during the period considered"

6f one assumed that the indi%idual stocks should e0perience returns e!ual to the aggregate stock market" #his assumption meant that the market ad&ustment process simply entailed
subtracting the market return from the return for the indi%idual security to deri%e its a&normal rate of return$ as follows: '&normal !eturn ( !eturn from Security ) !eturn from Mkt

1ut each security has a risk factor, $hich has to be considered "
'&normal !eturn ( !eturn from security ) E%pected returns
0pected return+ function of securityCs relationship to market"

5esults of 5eturn Prediction Studies

#he time-series analysis assumes that in an efficient market the best estimate of future rates of return will be the long-run historical rates of return" #he point of the tests is to determine $hether any public information $ill pro%ide superior estimates of returns for a short-run hori7on (one to si0 months* or a longrun hori7on (one to fi%e years*" #he results of these studies ha%e indicated limited success in predicting shorthori7on returns, but the analysis of long-hori7on returns has been !uite successful" A prime e0ample is di%idend yield studies" After postulating that the aggregate di%idend yield (32P* $as a pro0y for the risk premium on stocks, they found a positi%e relationship bet$een the 32P and future stock market returns" Subse!uent authors found that the predicti%e po$er of this relationship increases $ith the hori7on, that is, di%idend yields $ere better at predicting long-run returns"

 Se%eral studies ha%e considered t$o %ariables related to the term structure of interest rates+
a default spread,
which is the difference between the yields on lower-grade and Aaarated long-term corporate bonds (this spread has been used as a pro0y for a market risk premium*, and

the term structure spread,

which is the difference bet$een the long-term Aaa yield and the yield on one-month #reasury bills"

#hese %ariables ha%e been used to predict stock returns and bond returns

 #he reasoning for these empirical results is as follo$s+ When the t$o most significant %ariablesDthe di%idend yield (32P* and the default spreadDare high,
it implies that in%estors are e0pecting or re!uiring a high return on stocks and bonds" =otably, this occurs during poor economic en%ironments, as reflected in the gro$th rate of output"

A poor economic en%ironment also implies a lo$-$ealth en%ironment $herein in%estors percei%e higher risk for in%estments" As a result, for in%estors to in%est and shift consumption from the present to the future, they $ill re!uire a high rate of return" 6t is suggested that, if you in%est during this risk-a%erse period, your subse!uent returns $ill be abo%e normal" 6n contrast, $hen these %alues are small, it implies that in%estors ha%e reduced their risk premium and re!uired rates of return and future returns $ill be belo$ normal

 Quarterly Earnings Reports Studies that address quarterly reports are considered part of the times-series analysis" Specifically, these studies !uestion $hether it is possible to predict future returns for a stock based on publicly a%ailable !uarterly earnings reports" #he typical test e0amined firms that e0perienced changes in !uarterly earnings that differed from e0pectations" #he results generally indicated abnormal returns during the ,/ or .9 $eeks following the announcement of a large unanticipated earnings changeD referred to as an earnings surprise" #hese results suggest that an earnings surprise is not instantaneously reflected in security prices. #his implies that earnings surprises and earnings re%isions can be used to predict returns for indi%idual stocks" #hese results are e%idence against the M)

The January Anomaly

Several years ago, ranch proposed a unique trading rule for those interested in taking ad%antage of ta0 selling" 6n%estors (including institutions* tend to engage in ta0 selling to$ard the end of the year to establish losses on stocks that ha%e declined" After the ne$ year, the tendency is to reac!uire these stocks or to buy other stocks that look attracti%e" #his scenario $ould produce do$n$ard pressure on stock prices in late =o%ember and 3ecember and positi%e pressure in early Eanuary" Such a seasonal pattern is inconsistent $ith the M) since it should be eliminated by arbitrageurs $ho $ould buy in 3ecember and sell in early Eanuary"

 !rice"Earnings Ratios Several studies have e#amined the relationship bet$een the historical price-earnings *+,E- ratios for stocks and the returns on the stocks" Some have suggested that lo$ P2 stocks $ill outperform high P2 stocks because gro$th companies en&oy high P2 ratios, but the market tends to o%erestimate the gro$th potential and thus o%er%alues these gro$th companies, $hile under%aluing lo$-gro$th firms $ith lo$ P2 ratios" A relationship bet$een the historical P2 ratios and subse!uent risk-ad&usted market performance $ould constitute e%idence against the semistrong M), because it $ould imply that in%estors could use publicly a%ailable information regarding P2 ratios to predict future abnormal returns

 %eglected &irms
Arbel and Strebel considered an additional influence beyond si7eD
attention or neglect"

#hey measured attention in terms of the number of analysts $ho regularly follo$ a stock and di%ided the stocks into three groups+
(,* highly follo$ed, (.* moderately follo$ed, and (/* neglected"

#hey confirmed the small-firm effect but also found

a neglected-firm effect caused by the lack of information and limited institutional interest" #he neglected-firm concept applied across si7e classes

%ent Studies
Stock Splits 6P<s Listing on Stock 0changes World economic e%ents Fey political e%ents

#he hypothesis recei%es almost unanimous support from the numerous e%ent studies on a range of e%ents including stock splits, initial public offerings, $orld e%ents and economic ne$s, accounting changes, and a %ariety of corporate finance e%ents"
About the only mi0ed results come from e0change listing studies

Strong-Form Hypothesis: ests and !esults

#he strong-form M) contends that stock prices fully reflect all information, public and private. #his implies that no group of in%estors has access to private information that will allow them to consistently e0perience abo%e-a%erage profits" #his e0tremely rigid hypothesis re!uires not only that stock prices must ad&ust rapidly to ne$ public information
but also that no group has access to pri%ate information"

 6n%estigators ha%e tested this form of the M) by analy7ing the performance of the follo$ing four ma&or groups of in%estors+ (,* corporate insiders, 94: stock exchange specialists, (/* security analysts at ;alue )ine and elsewhere, and 97: professional money managers.

;onclusions 5egarding the Strong-'orm M)

#he tests of the strong-form M) generated mi0ed results, but the bulk of rele%ant e%idence supported the hypothesis" #he results for t$o uni!ue groups of in%estors (corporate insiders and stock e0change specialists* did not support the hypothesis
because both groups apparently ha%e monopolistic access to important information and use it to deri%e abo%e-a%erage returns

,.P),- #,*N% *+ $++,-,$N# - P,# ) . '/$#%

Efficient Markets and echnical 'nalysis #he assumptions of technical analysis directly oppose the notion of efficient markets" A basic premise of technical analysis is that stock prices mo%e in trends that persist" #echnicians belie%e that $hen ne$ information comes to the market, it is not immediately a%ailable to e%eryone
but is typically disseminated from the informed professional to the aggressi%e in%esting public and then to the great bulk of in%estors"

Also, technicians contend that in%estors do not analy7e information and act immediately" #his process takes time" #herefore, they hypothesi7e that stock prices mo%e to a ne$ e!uilibrium after the release of ne$ information in a gradual manner, $hich causes trends in stock prece mo%ements that persist

 6f the capital market is $eak-form efficient as indicated by most of the results, then prices fully reflect all rele%ant market information so technical trading systems that depend only on past trading data cannot have any value. <y the time the information is public, the price adjustment has taken place" #herefore, a purchase or sale using a technical trading rule should not generate abnormal returns
after taking account of risk and transaction costs"

Efficient Markets and Fundamental 'nalysis

'undamental analysts belie%e that, at any time, there is%a basic intrinsic %alue for the aggregrate stock market, %arious industries, or indi%idual securities
and that these %alues depend on underlying economic factors"

#herefore, in%estors should determine the intrinsic %alue of an in%estment asset at a point in time by e0amining the %ariables that determine %alue such as current and future earnings or cash flo$s, interest rates, and risk %ariables" 6n%estors $ho engaged in fundamental analysis belie%e that, occasionally, market price and intrinsic %alue differ but, e%entually, in%estors recogni7e the discrepancy and correct it