efficiency

1

MARKETEFFICIENCY

ANDTHE

BEHAVIOUROFSECURITIESPRICES

Objectives

Thismoduleisdevotedtothestudyofbehaviourofcuritiesprices

ndofthismodulestudents

shouldbeableto

(1)Clearlyunderstandwhatismeantbytheconceptofefficiencyin

curitiesmarkets

(2)UnderstandFama'sconceptofmarketefficiencyasatheoryfor

describingthebehaviourofcuritypricesandthemeaningofmarket

efficiencystandardsasdefinedinFama'sEfficientMarketHypothesis

(3)DescribehowvariousformsoftheEfficientMarketHypothesiscan

betestedandtheoutcomesofsuchtestscarriedoutintheliterature.

(4)Beawareofevidencesofmarketinefficienciesormarketanomalies

reportedintheliterature

Readings

pter12

e:""

PleadownloadthisfromtheBlackboard‘ArticleFolder’

Furtherreferences

2

Fama-"EfficientCapitalMarkets:..."JournalofFinanceMay1970

Fama-"EfficientCapitalMarkets:II"er1991

3

Thetopicxaminedinthismoduleareorganidasfollows:

ceptofefficientcapitalmarkets

icientMarketHypothesis

ationsforinvestorswhenmarketsareefficient:

fmarketefficiency

anomaliesorevidenceofmarketinefficiencies

4

Efficiencyoffinancialmarkets

Therearetwodimensionstothemeaningofefficiencyinmarkets,

whichcanbedescribedas

(1)Operationalefficiency,and

(2)FunctionalEfficiencyorinformationalandvaluationefficiency

(1)Operationalefficiency

Thekeyelementsthatmakeamarketoperationallyefficientaremarket

liquidity,orderlinessandlowcostsoftrading.

Liquiditymeansinvestorscandispooftheirholdingsquicklyand

withoutsacrificinglargepricediscountsfromprevailingmarketprices.

Factorsthatcontributetomarketliquidityaredepth,breadthand

resilience.

Liquidityofthemarketisindicatedby

Breadthofmarket(tradingvolumesatprevailingprices)

Depthofmarket(volumesofbuyordersbelowandvolumesofll

ordersabovetheprevailingprice)

Marketbreadthanddepthhelpstoensureresilience.

Marketdepthmeanstheabilityofthemarkettoabsorbtemporary

imbalancesbetweencuritiessupplyanddemandwithoutleadingto

largepricechangesthroughthetradingactivitiesofmarketmakers.

Marketmakersmuststandreadytobuyupcuritieswhenthesupply

ofcuritiexceeddemand,orrundowntheirinventoriesofcurities

breadthmeanstradingvolume

andtheexistenceofadequatecompetitionamongmarketmakersto

ence

meanstheabilityofthemarketpricetorecoverfromunusuallylarge

llorbuyorders.

5

Marketorderlinessisanotherimportantaspectofanoperationally

derly

aincompetitive

eofprice

anipulation

occurswhensomeparticipantshavesignificantmarketpowerorwhen

malpracticessuchasfrontrunningoccurs.

Lowtransactioncostsprovidesathirdcontributiontooperational

anslowtaxes,brokeragecommissionsandbid-ask

spreads.

(2)FunctionalEfficiency

RelatestoInformationalandvaluationefficiency

Amarketisinformationallyefficientifinformationthathaveabearing

onthevalueofcuritiesarereadilyavailabletomarketparticipants.

uationally

efficientmarketthepricesofastswillbeclototheirintrinsicor

fundamentalvalues.

Fama'sformaldefinitionofmarketefficiency-(JF1970)

Amarketifficientrelativetoaninformationtiftheprice

expectationsformedonthebasisoftheinformationtisanunbiad

predictoroftheactualpricesubquentlyrealid.

Let



P

t

=priceattimet



t

=theinformationtavailabletoinvestorsattimet.

EP

ttt

(



)

1

=expectationoffuturepricebadontoday's

informationt.



t

=thedeviationoftheactualpricefromthe

expectedprice(thepredictionerror)

thismeans

6



t+1

=



P

t+1

-EP

ttt

(



)

1



Ifthepredictionerrorisunbiadthenthemarketifficient.

E

t

(

t+1

)=0

STANDARDSOFMARKETEFFICIENCY

(TheEfficientMarketHypothesis-Fama(1970))

Inordertoestimateandalsotestthedegreeofefficiencyofaparticular

market,weneedtodefinestandardsofefficiencyasayardstickof

measurement.

EugeneFamahasdefinedthreelevelsofmarketefficiencyonthebasis

oftheamountofinformationthatisbuiltinto(orimpoundedin)market

prices.

historical

informationhistorical

information

allpublic

information

allpublic

information

historical

information

allpublicand

private

information

rmEfficiency

Onlyhistoricalinformationsuchasthehistoryofpastpricepatterns

arereflectedorbuiltintothecurrentmarketprice.

7

Theimplicationisthatinvestorscannotuanyknowledgeofpastprice

trendsorpatternstopredictfuturepricechangesandtherebydevelop

tradingstrategiestoearnabnormalreturns.

rongFormEfficiency

sthatallcurrently

publiclyavailableinformationisalreadyfullyreflectedinmarket

prices.

Theimplicationisthatinvestorscannotuanypubliclyavailable

informationalreadyknowntothemarkettodevelopstrategiestoearn

abnormalreturns.

FormEfficiency

sthatallinformation,

whetherpubliclyorprivatelyheld,includingthowithcorporate

insidersormarketspecialists,arefullyreflectedinmarketprices.

Theimplicationisthateveninvestorswithinsiderinformationcannot

utheirinformationtoearnabnormalreturns.

Somepropertiesassociatedwithanefficientmarket

(i)Pricechangeswillresultonlyfromnewinformationthathavean

effectonprentandfuturecurityreturnsratherthanonexisting

information.

(ii)Marketpriceswillreacttonewinformationquicklyandaccurately

(unbiadly)

8

(iii)Marketpriceswillfollow(orbecloto)arandomwalkprocess.

P

t

=P

t-1

+d+

t



t

isanindependentandidenticallydistributed(iid)riesofrandom

errors,disthedriftinprice

(iv)Marketpricesofcuritieswillgenerallyreflecttheirtrueintrinsic

values.

Somefactorsdrivingmarketstoefficiencyandwhywecanexpect

financialmarketstobeefficient

(i)Lawsthatcompelfirmstodisminateimportantinformationquickly

tothemarket

(ii)Anefficientandtechnologicallyadvancedinformationnetwork

(iii)Thestrongcompetitionamonganalystsandinvestorsdrivesprices

towardfficiency.

Largenumbersofinvestorsalllookingforabnormalprofit

opportunities,willbytheirownactions,competeawaysuch

opportunities.

(iv)Investorsandanalystsareeducated,knowledgeableand'smart'.

(v)Theindependenceoftheactionsofinvestors.

Thelawoflargenumberswillensurethattheneteffectofuncorrelated

tradingactionsofinvestorswillresultintheaveragepricesbeing

accurate.

9

(vi)Doinsidertradinglawshindermarketefficiency?

INSIDERTRADINGANDTHELAW

AmendmentstotheCorporationsLawintroducedinAugust1991

Whoisaninsider?

Aninsiderisonewhoposss'pricensitiveinformationwhichis

notgenerallyavailable'.

Aninsiderneednotbeconnectedtothefirmunderreference.

Whatisinsidertrading?

Tradingbadoninsiderinformationorcommunicatinginsider

informationtoanotherwhomighttradeonthatinformationisillegal.

Implicationsforinvestorsandthelikelyeffectivenessofinvestment

strategiesifmarketsaretrulyefficient:

(i)Predictingpricechangesbadonhistoricalinformationorpastprice

patternswillbeimpossible.

Therefore'Technical'analysisbadonanalysinghistoricalprice

'markettiming'strategiesmaybeof

littlebenefit

(ii)Sincemarketpriceswilladjusttonewinformationveryquicklyand

willaccuratelyreflectfundamentalvaluesingeneral

10

Anactivestocklectionstrategybadonfundamentalstockanalysis

foridentifyingunderandoverpricedstockswouldnotbeeasy.

(a)Securitieswillplotonthecuritymarketlinegiventhatast

valuationtheoriessuchastheCAPMiscorrect.

(b)Investorscanonlyhopetoearnanormalreturnfromtheir

lreturnisthereturncommensuratetothelevelof

riskintheinvestmentaccordingtotheCAPM.

(iii)Passiveinvestmentstrategiessuchasinvestinginanindexfundor

otherbuyandholdstrategieswouldbethemostappropriate.

SomealternativetotheEfficientMarketHypothesisfordescribing

thebehaviourofthestockmarket

ketOverreactionHypothesis(orthewinner-lor

hypothesis)DebondtandThaler(JF1985)

Atheorybadonirrationalinvestorbehaviour.

ionalSpeculativeBubbleshypothesisBlanchardandWatson

(1982)

Atheorybadonrationalinvestorbehaviourwhichatthesametime

canleadtothedeviationofmarketpricesfromtheirfundamental

values.

TESTINGFORMARKETEFFICIENCY

(1)Testsofweakformmarketefficiency

Canpastreturnsbeudtopredictfuturereturns?

(a)Testingforrialcorrelation

11

Ex:firstorderautocorrelation



1

1

1

1

2

1

1

1













T

rrrr

t

t

T

t

)

(b)Testingfilterrulesforstocktrading

(2)Testsofmi-strongformmarketefficiency(EventStudies):

Testinghowquicklyandaccuratelycuritypricesrespondtonewly

releadpublicinformation?

t-1tt+1

Time(days)

Stockprice

delayedreaction

overreaction

efficientrespon

Eventstudymethodology

Atestofmistrongefficiencyiswhetherthestockpricereactiontoan

event,takingplaceondayt(suchasabetterthananticipatedearnings

announcement)bringsforthanimmediatepricereactionorwhether

theresponlagsontodayt+1andt+2etc.

12

Testprocedure

(1)Selectasampleoffirmsmakingforexample,betterthananticipated

earningsannouncements.

(2)Testtheirpriceresponsondayt,t+1,......(wheret=announcement

date)

(3)Butpriceswillchangeanywayduetooverallmarketchanges(with

orwithouttheannouncement).

(4)Needtoisolateandexaminepricechangesolelyduetothe

announcementeffect.

Obrvedreturn(OR)=Returnduetoannouncement(AR)+Normal

return(NR)

Normalreturn(orexpectedreturn)isthereturnbadontherelationof

thefirm'sreturntothatofthemarketandcouldbemeasuredby

applyingthemarketmodel(characteristicline).

Thenormalreturnondaytforfirmibadonthecharacteristiclineis

givenby

R

it

=a

i

+

i

R

mt

+e

it

E(R

it

)=a

i

+

i

E(R

mt

)

AR=OR-NR

=R

it-

E(R

it

)

=R

it

-[a

i

+

i

E(R

mt

)]=e

it

Theabnormalreturnsaretheregressionresiduals

13

(5)Calculatetheaverageabnormalreturns(AAR)

Averageoftheresidualsfortheparticulardayacrossallthesample

firms

(6)Calculatethecumulativeaverageabnormalreturns(CAAR)overa

timeinterval.

Thesumoftheaverageabnormalreturnsoververaldayst+1,t+2...

etc.

(7)AretheCAARssignificantlydifferentfromzero?

Example:TheCAARpatterninanefficientmarket

t-1tt+1

Time(days)

alreturn

0x

x

x

Example:InthestudybyFoster,OlnandShevlin:

Stockswithlargepositiveearningssurpriarnedabnormalreturns

fromupto60dayspriortotheearningsannouncementandupto60

daysaftertheearningsannouncement.

Whatdoesthisimplyaboutmarketefficiency?

14

(3)Testsofstrongformmarketefficiency

(i)

testwhetherabnormalreturnsaremadebyoutsidersfollowinginsiders'

festudy,Seyhunstudy

(ii)TestingwhetherabnormalreturnsaremadebyNYSEspecialists

EMPIRICALEVIDENCEOFMARKETINEFFICIENCIESORMARKET

ANOMALIES

(1)Testsofmarketpredictability

(i)Predictabilityofshorttermreturns

Overall,testsofrialcorrelation,runstestsandfilterrulesfindthat

weakformefficiencyislargelyvalidated.(Fama1965,FamaandBlume,

LoandMacKinlayetc.)

(ii)Predictabilityoflonghorizonreturns

FamaandFrench1988andPoterbaandSummers1988findnegative

correlationinlonghorizonreturns.

sit

necessarilyinvalidatemarketefficiency?

(iii)Predictorsofaggregatestockmarketreturns

15

FamaandFrench1989andCampbellandShiller1988findthatvariables

suchasdividendyield,defaultyieldspreadcanpredictvariationin

stockmarketreturns.

(2)Crossctionalanomalies

Arethecrossctionalanomaliestheresultofmarketinefficiencyorthe

resultofastpricinganomalies?

(i)TheSmallfirmeffect(highreturnsofsmallfirmspeciallyin

January)

Banz1981,Reinganum1983Keim1983

(ii)ThelowP/Estrategy(highreturnsoflowP/Estocks)

Basu1977

(iii)Themarkettobookvalueratio

FamaandFrench1992

(iv)Theneglectedfirmeffect

ArbelandStrebel1983,AmihudandMendelson1986

(v)TheValueLinestockrankingsystem

ValueLineclaimsthattheperformanceofstocksoverthenext12month

periodcanbepredictedifstocksarerankedinaccordancewiththe

followingcriteria.

Therankisbadonacompositeof(1)relativeearningsmomentum(2)

Earningssurpri(3)Nonparametricvalueposition

16

(3)SeasonalAnomalies

(i)TheJanuaryeffect(highreturnsinJanuary)

(ii)Theweekendeffect(neverllonMondays)

French(1980)