Scribd Logo
Upload

Welcome to Scribd!

  • 43 views

    Mathematical Modelling and Financial Engineering in The Worlds Stock Markets

    Uploaded by

    Vijay Parmar
    This document discusses mathematical modelling and financial engineering of derivatives in stock markets. It begins by explaining why financial derivatives are necessary tools to hedge against risks like foreign exchange fluctuations. It then provides examples of option pricing models, like the Black-Scholes model, and derives the Black-Scholes partial differential equation. The document also discusses the Greeks (Delta, Gamma, etc.), which are the measures of how sensitive an option's price is to changes in the underlying variables. More advanced models are noted that incorporate factors like dividends, multiple interest rates, and different option exercise types.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Mathematical Modelling and Financial Engineering in The Worlds Stock Markets

    Uploaded by

    Vijay Parmar
    43 views7 pages
    This document discusses mathematical modelling and financial engineering of derivatives in stock markets. It begins by explaining why financial derivatives are necessary tools to hedge against risks like foreign exchange fluctuations. It then provides examples of option pricing models, like the Black-Scholes model, and derives the Black-Scholes partial differential equation. The document also discusses the Greeks (Delta, Gamma, etc.), which are the measures of how sensitive an option's price is to changes in the underlying variables. More advanced models are noted that incorporate factors like dividends, multiple interest rates, and different option exercise types.

    Original Title

    MITSTalk4

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document discusses mathematical modelling and financial engineering of derivatives in stock markets. It begins by explaining why financial derivatives are necessary tools to hedge against risks like foreign exchange fluctuations. It then provides examples of option pricing models, like the Black-Scholes model, and derives the Black-Scholes partial differential equation. The document also discusses the Greeks (Delta, Gamma, etc.), which are the measures of how sensitive an option's price is to changes in the underlying variables. More advanced models are noted that incorporate factors like dividends, multiple interest rates, and different option exercise types.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    43 views7 pages

    Mathematical Modelling and Financial Engineering in The Worlds Stock Markets

    Uploaded by

    Vijay Parmar
    This document discusses mathematical modelling and financial engineering of derivatives in stock markets. It begins by explaining why financial derivatives are necessary tools to hedge against risks like foreign exchange fluctuations. It then provides examples of option pricing models, like the Black-Scholes model, and derives the Black-Scholes partial differential equation. The document also discusses the Greeks (Delta, Gamma, etc.), which are the measures of how sensitive an option's price is to changes in the underlying variables. More advanced models are noted that incorporate factors like dividends, multiple interest rates, and different option exercise types.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 7

    MATHEMATICAL MODELLING AND FINANCIAL

    ENGINEERING IN THE WORLDS STOCK MARKETS

    VIJAY K. PARMAR

    Abstract. This is a sample document which shows the most important features of the AMS Journal Article class.

    1. Introduction
    What are Financial Derivatives and why are they necessary? In short they are
    insurance policies against adverse movements in the underlying stock price at a
    future date. For example, if , say Reliance Coorporation wants to buy rolling stock
    (ie trains) for the Mumbai Metro in 6 months time from a company in the Eurozone and the price agreed is 2 bn euros. Reliance assets are in Indian Rupees. So
    there is clearly some Foreign Exchange Risk which needs to be managed. It can
    either
    Buy 2bn euros today at todays FX rate and pay the company now and get
    delivery in 6 months time-not very good idea. The company will just take
    the 2bn euros put it in a bank account for 6 months. Take the 2bn plus the
    interest.
    Reliance itself can buy 2bn euros today and put it in the bank account.
    Then after 6 months pay 2bn euros keep the interest. Little better. But
    not very good business sense-tying up a companies assest in a bank account
    for 6 months!!!
    Enter into a contract either be commited to buy 2bn euros at a xed FX
    rate - if the rate is higher or lower Reliance is commited to buy 2 bn euros
    at that rate. Or an option which can be exercised if the actual rate higher
    than that (strike X) agreed 6 months prior or expires if the actual rate is
    lower than the strike rate X.
    Clearly, the latter option is NOT free. There is a price that needs to be paid. Too
    low then the writer loses money (arbitrage) too high nobody will buy the option.
    The correct price wherein the writer does not lose money and people will buy the
    option is termed the arbitrage-free price.
    1

    VIJAY K. PARMAR

    2. Options and other Financial Derivatives


    2.1. Options. Mathematically, how do we price such options? Well, if {St } is the
    price of the asset at time t (in this case the exchange rate) and the agreed price we
    want to buy for termed strike X then clearly the payo at maturity T or value
    of the call option is
    CT = max(ST X, 0).
    Intuitively, if the FX rate ST is higher then the agreed price X then we want to
    exercise otherwise we let it expire (i.e. do nothing). The big question is we want
    to know what the value is today!!! Working back using continuous compounding
    at rate {rt } is
    C0 = exp(rT ) max(ST X, 0).
    This is ne if {rt } is constant. In reality it is unlikely to be constant, it is likely to
    be stochastic. So working back over innitessimal periods of time the value today
    is
    (
    )
    T
    C0 = exp
    rs ds max(ST X, 0).
    0

    But this is uncertain - we dont know ST or the evolution of the spot rate {rt }
    between now and maturity T . So we need to take an average over all the possible
    values. So
    [
    (
    )
    ]
    T
    C0 = E exp
    rs ds max(ST X, 0)
    0

    where E[] is the expectation operator. Similarly, if we want to sell we want a payo
    of
    PT = max(X ST , 0)
    and the value of the put option is
    [

    rs ds) max(X ST , 0) .

    P0 = E exp(
    0

    In practice there will be a model for the spot rate {rt } and {St }. That is, they
    satisfy Stochastic Dierential Equations (SDEs)
    drt =(rt , t)dt + (rt , t)dWt
    (2.1)

    dSt =A(St , t)dt + (St , t)dWt

    Example 1. Common spot rate processes are


    Ho-Lee: drt = (t)dt + dWt . So simply,
    t
    rt = r0 +
    (t)dt + Wt .
    0

    (
    )
    t
    rt N r0 +
    (t)dt, .
    0

    Hull-White: drt = (t)((t) rt )dt + (t)dWt . Itos lemma can be used to


    show the solution is
    t
    )
    (
    t
    t
    t
    eu dWu
    rt = e r0 +
    1e
    + e

    FINANCIAL ENGINEERING

    which has distribution


    (
    )
    ) 2 (
    )
    (
    rt N et r0 +
    1 et ,
    1 e2t .

    Cox-Ingersol-Ross: drt = a(t)(b(t) rt )dt + rt dWt


    Example 2. There are other models of stock price behaviour such CEV
    dSt = St dt + St dWt
    The constant parameters , satisfy 0, 0. Usually the most common is
    = 1.
    In general, whatever the payo g(T ) we want at time T the value now is clearly
    [
    (
    )
    ]
    T
    C0 = E exp
    rs ds g(T ) .
    0

    In practise, how to we calculate this? In simple cases, when for example


    drt =(rt , t)dt + (rt , t)dWt
    (2.2)

    dSt =St (a(t)dt + (t)dWt )

    We can directly calculate this expectation.


    Example 3. If g(T ) = 1 in the Ho-Lee model fro spot rates we have the price of a
    zero coupon bond
    (
    )
    1
    1
    C0 = exp r0 T + T 2 + 2 T 3 .
    2
    6
    2.1.1. Put-Call Parity. We notice that at time T
    max(S X, 0) max(X S, 0) = S X
    That is
    CT PT = ST X
    By discounting backwards
    Ct Pt = St erT X
    This is called the put-call parity. In other words, the put and call prices are
    always related. So once the call prices have been calculated no extra work needs to
    be done to calculate the put prices.
    This FPK equation can be and was obtained initially by a completely nancial
    argument by 2 economists Black and Scholes.
    2.2. Fixed Income Derivatives. Bonds and Swap derivatives. Just in the US in
    2013 it was worth $40 trillion.
    2.3. Credit Derivatives. Typical instruments include Credit Default Swaps . The
    market in 2007 was worth $13 trillion. Now it is half of that.
    2.4. Energy Derivatives.
    2.5. Weather Derivatives.
    2.6. OTC and Exchange Traded Options. Some options like the Reliance example given above are specially tailor-made for the client by the writer (usually a
    nancial institution). Others have standardized features


    VIJAY K. PARMAR

    3. Black-Scholes Equation Derivation


    Consider a portfolio of an asset S and number of call options C on that asset
    (ie a option to buy that asset S at a price X at a time T ). That is consider
    = S + C
    In a standard market the asset price evolves according to the log-normal process
    dSt
    = rdt + dWt
    St
    and let us assume that r = const and = const then we need to calculate the
    process that the call option C = C(S, t) obeys. In other words we need to know
    what dC =?. Now since 3.1 is an Ito process then we can invoke Itos Formula and
    nd
    [
    ]
    C
    1
    C
    2C
    C
    + rS
    + 2 S 2 2 dt + S
    dWt
    dC =
    t
    S
    2
    S
    S
    So now we want to create a portfolio which over a short period of time dt is
    RISK-FREE. What do we mean by risk-free? NO UNCERTAINTY! That is, no
    dependance on {Wt }. If it is risk-free over a short period of time dt then then it
    evolves DETERMINISTICALLY at the risk-free rate r. That is
    (3.1)

    d = rdt
    That is,
    [[

    C
    t
    [[
    C
    SdWt +
    t

    S(rdt + dWt ) +

    dS + dC =r(S + C)dt
    ]
    C
    1
    C
    C
    + rS
    + 2 S 2 2 dt + S
    dWt =r(S + C)dt
    S
    2
    S
    S
    ]
    ]
    2
    C
    1
    C
    C
    + rS
    + 2 S 2 2 dt + S
    dWt =rCdt
    S
    2
    S
    S
    2

    So the next thing is that we need to get rid of the uncertainty is dWt . This is easy.
    C
    We just take = 1/
    . Then we get the Black-Scholes equation
    S
    C
    C
    1 2 2C

    + rS
    +
    rC(S, t) = 0.
    2 S 2
    S
    t
    Next question, now that we have got the B-S equation. We need to solve it.
    Remark 1. Under simplfying assumptions (r = const and = const) we can
    transform this into the heat equation
    C
    2C
    =
    ,
    C = C exp(rt),
    t
    S 2
    So we see it a parabolic equation.

    (3.2)

    S = log S (r 1/2 2 )t

    By observing that it can be reduced to the standard heat equation (3.2) we can
    employ standard methodologies to derive the solutions as
    C0 = SN (d+ ) XerT N (d )
    where
    d =

    log S/X + (r 2 /2)T

    .
    T

    FINANCIAL ENGINEERING

    This is the call price of a European option with strike X on an asset St


    maturing at time T .
    We have made some VERY STRONG ASSUMPTIONS r = const and = const.
    If we relax them then the above risk-free argument still holds but the solution
    methodology becomes very complicated.
    Remark 2. This work along with that of Merton won the 1997 Nobel Prize in Economics. The level of the mathematics is amazingly low!! But this pioneering work
    revolutionized share trading around the world. It is interesting that both Scholes and
    Merton were the founders of LTCM an American Hedge Fund.Initially successful
    with annualized profit of over 21pc in its first year, 43pc in the second year and
    41pc in the third year, in 1998 it lost $4.6 billion in less than four months following
    the 1997 Asian financial crisis and 1998 Russian financial crisis requiring financial
    intervention by the US Federal Reserve to bail it out.
    4. General Black-Scholes Equation
    In general, for more complicated spot rate processes {rt } and more realistic share
    asset price {St } models directly expanding this expectation is near on impossible.
    [
    (
    )
    ]
    T
    Theorem 1 (Feymann-Kac). C0 = E exp 0 rs ds max(ST X, 0) is a solution to the following Fokker-Plank-Kolmogorov equation
    1 2
    2C
    C
    C
    (S, t) 2 + r(S, t)S
    +
    r(S, t)C(S, t) = 0.
    2
    S
    S
    t
    5. Risk Sensitivity and Greeks
    Clearly, C = C(S, t, r, ) so as each of these parameters changes the price of the
    option will change as will the value of the portfolio. These Greeks are dened as
    the partial derivatives w.r.t. each of the arguments. For our simple toy model we
    have
    Delta
    =
    Gamma
    =
    Vega
    =

    C
    = N (d1 )
    S

    2C
    N (d1 )

    =
    S 2
    S T t

    C
    = SN (d1 ) T t

    Theta
    =

    C
    SN (d1 )
    =
    rXer(T t) N (d2 )
    t
    2 T t

    Rho

    C
    = X(T t)er(T t) N (d2 )
    r
    A key role of a risk manager is to rebalance the portfolio so it the sensitivitiues
    of the overall portfolio is as close to zero as possible.
    =

    VIJAY K. PARMAR

    6. More Realistic Models


    Most stocks pay dividends -we need to incorporate them into the pricing
    equations.
    With FX options domestics and foreign interest rates need to be taken into
    account. Again each of these will have their own evolutionary models.
    6.1. More General Options. So far we have mentioned only exercise at the end
    of the life of the option or maturity. These are European options. If we allow
    more exible exercise during the life of the option we have American options or
    with some restriction Bermudean Options. The strike price X does not have to
    be given explicity up-front-it may depend of the performance of the stock price St
    during its life Eg
    X = max(St )
    or
    X = average(St )
    Then we have path-dependent options. These all add more complications to
    the pricing.
    7. Solving the Fokker-Planck-Kolmogorov
    So now concentrating on solving the this PDE with the boundary conditions
    is equal to nding the price of the call option. As the payo and the models for
    {rt } and {St } become more realistic (and hence more complicated) solving this
    analytically becomes almost IMPOSSIBLE. We then have to appeal to numerical
    methods to solve this. This can either be a discretization of the PDE, building
    a Binomial or Trinomial Tree or a Monte Carlo method to directly estimate the
    expectation.
    8. Trading Strategies
    With derivatives you gain much more leverage for the same amount of nancial
    commitment.
    8.1. How to make money. With the primary market there is only one way to
    make money (and loose money)-buy (take a long position) in the underlying stockif it goes up-you make money-if it goes down you loose money!! However, in the
    derivative market (secondary market) you can make money either
    if the stock goes up
    if the stock goes down
    if the stock hardly moves.
    How?
    Single stock+single option
    Spreads
    Bull Spread
    Bear Spread
    Buttery Spread
    Calendar Spread
    Diagonal Spread
    Combinations
    Straddles

    FINANCIAL ENGINEERING

    Strips
    Straps
    Strangles
    9. Were is this analysis used?
    In investment banks, insurance companies, hedge funds, FX companies. On the
    BSE Equity options, FX Options and IR options are traded.
    10. Careers
    Risk Management, Front oce trading, Product Control, Model Validation
    Quantitative Analysts. Starting salaries in London (50 lakh ruppees p.a.).
    10.1. Front Oce Trading. Here the role of the Quant is to support the trader.
    The trader formulates a suitable instrument to fulll the requirements of the client
    and agrees the details of the contract(knowing what is possible and what is not
    possible). The quant then has to build a model to price the instrument and risk
    manage the instrument. The model then has to under go a Due Diligence process
    where every eected department is consulted - Risk Management, Regulatory Department, Model Validation etc. After the approval is given then the mandate for
    trading is given specifying the volumes, restrictions and reserves which need to be
    held for unobservable parameters. Whilst the trade is live ie on the books daily
    P/L is calculated. This a very high pressure job environment.
    10.2. Model Validation. This is a back oce support role. A quant here has to
    INDEPENDENTLY valid the front oce trading models by building his own model
    and subjecting it to the same stress tests as the front oce model and comparing
    the two.
    10.3. Product Control. It is the role of the quant in product control to ag
    any unobservable parameters in the trading model. For example, volatility and
    calculate the appropriate level of reserves that need to be held for this. This is a
    VERY UNPOPULAR role. The more reserves you make the front oce hold the
    less prot there are seen to make!!!!
    10.4. Key Skills.
    MSc in Mathematical Finance, PhD in Mathematics, Physics, Engineering.
    Good grasp of Stochastic Calculus and Numerical Methods.
    Good programming skills in VBA, C++, C
    References
    [1] Black, F., Scholes Pricing of Options and Cormporate LiabilitiesJ. Political Economy, (1973)
    [2] Hull, J.C.,: Options, Futures and Other Derivative SecuritiesPrentice Hall
    (Former Head of Derivative Analytics, Product Control, HSBC London)

    You might also like