Appunti1

Embed Size (px)

Citation preview

  • 8/11/2019 Appunti1

    1/7

    Finance

    Optionpricin

    g:theBlack&Scholesmodel

    Ro

    ber

    toRen

    `o

    [email protected]

    Dipar

    timento

    diEconom

    iaPo

    litica

    ,Un

    iversi

    t`adiSiena

    Novem

    ber

    ,20

    02

    p

    .1/42

    Introductiontofinance

    FinanceisasubsetofEco

    nomics,whichisquitepeculiarfortwo

    reasons:

    Alotofdataareavailable

    ......

    .....

    thanmodelsareverysophisticated

    .

    Novem

    ber

    ,2002

    p

    .2/42

    Introductio

    ntofinance

    Tradedassetsinfinancialmark

    etsare:

    stocks(equity)

    bond

    currencies(foreignexchange)

    commodities:gold

    ,cotton,

    etc.

    derivatives

    Novem

    ber

    ,2002

    p

    .3/42

    Derivatives

    Derivativesarefinancialcontractswhosepayoffatfuturedates

    dependsuponthevalueofanotherasset,calledunderlying

    .

    Themostheavilytrad

    edassetsarefuturesandoptions.

    Novem

    ber

    ,20

    02

    p

    .4/42

    Futures

    Afuturescontractgivesyoutheobligationtobuyanassetina

    futuredate.Ininstitutiona

    lmarkets,thecounterpartshavetopay

    aguaranteewhensigning

    thecontract.Theresnoinitialprice.

    Futuresaremoreliquidthantheirunderlying.

    Futuresaredifferentfrom

    forwardbecauseofcontractual

    differencesregardingguaranteesandthewaythepayoffis

    delivered.

    Novem

    ber

    ,2002

    p

    .5/42

    Options

    Anoptionisafinancialcontractwhichprovidestherighttobuy

    (sell)anassetinafuturedatea

    tagivenprice.

    Calloption:

    C

    T

    S

    T

    K

    Putoption:

    P

    T

    K

    S

    T

    Novem

    ber

    ,2002

    p

    .6/42

  • 8/11/2019 Appunti1

    2/7Absenceofar

    bitrageandcompletenes

    s

    Weassumethatinthemarketisisnotpossibletorealizea

    positivepayoffwithnorisk.

    Suchacontractwouldbean

    arbitragecontract.Th

    isassumptioniscalledabsenceofarbitrage

    .

    Amarketissaidtobecompleteifitispossibletoreplicateany

    derivativewithabuy-and-sellstrategyaccomplishedwiththe

    underlying.

    Novem

    ber

    ,20

    02

    p

    .7/42

    Perf

    ectmarket

    Wewillworkinmarketin

    which:

    theresnoarbitrage

    therearenotransactioncosts

    assetscanbeshort-selledandboughtinthedesired

    quantities

    theresnodefaultrisk

    Novem

    ber

    ,2002

    p

    .8/42

    MoneyMarketAccount

    TheMoneyMarketaccountisamodelofarisklessasset,whose

    payoffiscertain

    dB

    t

    r

    t

    r

    t

    B

    t

    dt

    shortrate,spotrate

    Theshortrateisthestatevariableadoptedininterestratemodels.

    Novem

    ber

    ,2002

    p

    .9/42

    MoneyMarketAccount

    Ifr

    t

    iddeterministi

    c,theevolutionofB

    t

    isdeterministic,the

    solutionoftheordina

    rydifferentialequationis

    B

    t

    B

    0

    exp

    t0

    r

    s

    ds

    Ifr

    t

    isconstant,itcoincideswiththeyieldtomaturity.

    Novem

    ber

    ,200

    2

    p.1

    0/42

    FIB30timeseries

    Letslookthetimeseriesofanasset(FIB302000-2

    001

    ,daily

    data):

    Novem

    ber

    ,2002

    p

    .11/42

    Asset

    pricing

    Thepriceofafinancialassetis

    modeledviaastochastic

    differentialequation:

    dS

    t

    S

    t

    tS

    t

    S

    dt

    t

    S

    t

    S

    dW

    t

    drift

    volatility

    Novem

    ber

    ,2002

    p

    .12/42

  • 8/11/2019 Appunti1

    3/7

    TheBlac

    kandScholesmodel

    IntheBlackandScholesmodeltherearetwoassetsSandB

    ,

    whichfollowthefollowingdynamics:

    dB

    t

    rB

    t

    dt

    dS

    t

    S

    t

    dt

    S

    t

    dW

    t

    r

    areconsta

    nts!

    Sistheriskyasset,tipycallyastock.

    Bistheriskless

    asset.

    TheBlackandScholesmodelcanbeusedtovaluederivatives.

    Novem

    ber

    ,200

    2

    p.1

    3/42

    Contingen

    tclaimvaluation

    Letusconsidertheproble

    mofpricingaderivativeissuedint

    whichpromisesapayoffa

    tT

    tgivenby

    S

    T

    .

    Forexample

    ,fortheCalloption

    S

    T

    S

    T

    K

    .

    Let

    s

    F

    s

    S

    s

    bethederivativepriceattimet

    s

    T.

    Novem

    ber

    ,2002

    p

    .14/42

    Contingentclaimvaluation

    Itoslemmaallowsustocomputetheevolutionof

    t

    S

    d

    t

    t

    t

    dt

    t

    dW

    t

    t

    Ft

    SF

    S

    1 2

    2S2

    2F

    S2

    F

    t

    S

    F

    S

    F

    Novem

    ber

    ,2002

    p

    .15/42

    Portfolio

    ConsideraportfoliocomposedbytheassetSandthederivative.

    Let:

    uSbethefractio

    nofwealthinvestedinS.

    ubethefractio

    nofwealthinvestedin

    Thevalueoftheportfolioisthengivenby:

    V t

    uS

    t

    VS

    t

    u

    t

    V

    t

    then:

    dV

    V uS

    u

    dt

    V

    uS

    u

    dW

    t

    IfuS

    u

    0the

    portfolioisriskless.

    Novem

    ber

    ,200

    2

    p.1

    6/42

    Arbitr

    ageandPDE

    Nowweusetheassumptionofabsenceofarbitrage;sincethe

    portfolioisriskless,

    itsdrifthastobeequaltor.

    uS

    u

    r

    SolvingforuSandu

    ,kee

    pinginmindthatuS

    u

    1wehave

    Ft

    rSFs

    1 2

    2S2

    2F

    S2

    rF

    0

    F

    T

    S

    S

    T

    WethenfoundaPDEforthepriceofanyderivative.

    Novem

    ber

    ,2002

    p

    .17/42

    Arbitrage

    Someconsiderations

    Ourresultsdependsonly

    ontheabsenceofarbitrage.Thisis

    notastonishing,

    sincethe

    valuationequationforFneedsthe

    dynamicsofStobecomp

    uted

    .

    Thisresultisvalidonlyif

    thederivativeisactivelytraded.

    Thisresultisstillvalidif

    anddependont

    S.

    Novem

    ber

    ,2002

    p

    .18/42

  • 8/11/2019 Appunti1

    4/7

    RiskNeutralvaluation

    Letslookforaprobabilisticinterpretationofthisresult

    .

    TheFeynman-Kacfo

    rmulasaysthatthesolutionofthePDEis

    givenby:

    F

    tS

    t

    er

    T

    t

    EI

    X

    T

    t

    whereXisthesolutionoftheSDE:

    d

    X

    rrXdt

    XdW

    t

    X

    t

    S

    t

    TheonlydifferenceinthediffusionofXwithrespecttoSisthat

    thedriftisgivenbyr

    insteadof.

    Novem

    ber

    ,200

    2

    p.1

    9/42

    RiskNeutralValuation

    IfXwouldbethepriceof

    anasset,thaninvestorsarerisk-neutral

    Indeed

    ,investorwouldequallytradeX

    ,whichhasdriftrbutis

    uncertain,

    andB

    ,whichhasdriftrandnouncertainity.

    Actually

    ,investorsarerisk-averse.

    Theytakeriskypositionsonly

    ifthempromiseanexcess

    return.

    Novem

    ber

    ,2002

    p

    .20/42

    Girsanovtheorem

    DenotebyPtheprobabilityun

    derwhichWt

    isaBrownian

    motion.

    Girsanovtheoremsaysthatfor

    anyprocess

    t

    ,then

    B

    t

    W

    t

    t 0

    s

    dsisaBrownianmotionunderthe

    probabilityPwithdensity

    L

    t

    exp

    t0

    s

    dW

    s

    1 2

    t0

    s

    2ds

    withrespecttoP

    .

    Novem

    ber

    ,2002

    p

    .21/42

    R

    iskPremium

    Wethencall

    theprobabilityunderwhichW

    isaBrownian

    motion.

    Weinterpret

    XastheassetSundertheprobability

    .

    iscalledriskneutralprobability.FollowingGirsanovs

    theorem,

    therelation

    betweenW

    andWis:

    W

    t

    W

    t

    r

    r

    t

    Riskpremium:exces

    sreturnperriskunit

    .

    Riskismeasuredvia

    volatility

    .

    Novem

    ber

    ,200

    2

    p.2

    2/42

    RiskNeutralValuation

    Then,

    followingFeynman-Kacformula

    ,wecanwrite:

    F

    t

    S

    t

    er

    T

    t

    EI

    S

    T

    t

    Letusconsiderthediscou

    ntedprice:

    P

    s

    S s

    F

    T

    S

    T

    er

    T

    s

    Theaboveformulasaysth

    at,

    t

    s

    T:

    P

    t

    EI

    P

    s

    t

    Suchaprocessiscalleda

    martingale,and

    isalsocalled

    equivalentmartingalemea

    sure

    .

    Novem

    ber

    ,2002

    p

    .23/42

    RiskNeutralProbabilityandarbitrage

    TherelationwefoundfortheB

    lackandScholesmodelismuch

    moregeneral:

    Fundamentaltheoremofmathematicalfinance:

    Inthemarket,

    therearenoarbitrageopportunitiesifandonlyif

    thereexistsaprobability

    wh

    ichisequivalenttoPunderwhich

    discountedpricesaremartingalesunder

    .

    Themarketiscompleteifandonlyif

    isunique.

    Theabovereasoningsshowtha

    ttheBlackandScholesmodelis

    completeandfreeofarbitrage.

    Novem

    ber

    ,2002

    p

    .24/42

  • 8/11/2019 Appunti1

    5/7

  • 8/11/2019 Appunti1

    6/7

    The

    hedgingstrategy

    Sinceforthestrategy

    u

    u0theportfoliovalueisgivenby:

    dV

    t

    V

    t

    u

    0

    t

    r

    u

    t

    dt

    V

    t

    u

    t

    t

    dW

    t

    wehave:

    u

    t

    SF

    SF

    u0

    t

    Ft

    1 2

    2S2

    2F

    S2

    rF

    Ifweasku0

    u

    1wehave:

    Ft

    rSF

    S

    1 2

    2S2

    2F

    S2

    rF

    0

    whichisthePDEfor

    theoptionprice!

    Novem

    ber

    ,200

    2

    p.3

    1/42

    Th

    eGreeks

    Thepricesensitivitywith

    respecttotheparametersismeasured

    bytheso-calledGreeks:

    F

    S

    2F

    S2

    Fr

    Ft

    F

    Novem

    ber

    ,2002

    p

    .32/42

    he

    dging

    Whenwehedgeanoptioninth

    eBlackandScholesmodel,

    we

    have:

    u

    SF

    thanisexactlythenumberofstocksSthatIhavetobuyfora

    givennumberofoptionsF

    ,inordertohavearisklessportfolio.

    Followingthat,

    intheBlackan

    dScholesmodelhedgingiscalled

    hedging.

    Novem

    ber

    ,2002

    p

    .33/42

    Blackand

    Scholes:theempirics

    TheBlackandScholesmodelisparamountlyused

    ,startingfrom

    theearlySeventies.I

    tisthenaturalbenchmarkforanyoption

    pricingmodel.

    Itowe

    sitsfortunetoitssimplicity:itsthe

    simplestdiffusionmo

    delwithNormalreturns.

    Nowweshouldasko

    urselvesifitsassumptionarerespectedby

    financialmarkets,andconsequentlyifitcanbeusedsafelyornot.

    Novem

    ber

    ,200

    2

    p.3

    4/42

    Areinterestratesconstant?

    Interestratesvarythrough

    time:

    Novem

    ber

    ,2002

    p

    .35/42

    Isvolatilit

    yconstant?

    Volatilityisnotconstant.

    Novem

    ber

    ,2002

    p

    .36/42

  • 8/11/2019 Appunti1

    7/7

    ArereturnsNormal?

    Lookingatthetimes

    eriesitseemsnot.

    ...

    Novem

    ber

    ,200

    2

    p.3

    7/42

    ArereturnsNormal?

    Indeedtheyarenot,thedistributionisleptokurtic(fattails):

    Novem

    ber

    ,2002

    p

    .38/42

    Arereturns

    independent?

    TheACFofreturnsisnullafteroneday!Thisiscoherentwith

    theassumptionsofindependen

    treturns,but.

    ..

    Novem

    ber

    ,2002

    p

    .39/42

    Arereturnsindependent?

    TheACFoftheabsolutevalueofreturnsshowsdependencefor

    verylongperiods(1-2months).

    Novem

    ber

    ,200

    2

    p.4

    0/42

    Impliedvolatilityandmoneyness

    TheBlackandScholesformulaimpliesthatimpliedvolatility

    doesnotdependonthemo

    neynessS

    K.

    Impliedvolatilityshouldb

    ethasameforCalloptionatthemoney

    (S

    K

    1),inthemoney(S

    K

    1),outofthemoney(S

    K

    1).

    Sucharesultwouldbean

    importantpointmadebytheBlackand

    Scholesmodel,

    irrespectiv

    elyoftheempiricalfactswhichit

    cannotaccountfor.

    Novem

    ber

    ,2002

    p

    .41/42

    Smile

    effect

    Impliedvolatilitydependsonm

    oneyness:

    Novem

    ber

    ,2002

    p

    .42/42