Option pricing using TR-BDF2 time stepping method
|
|
- Clarence Ward
- 5 years ago
- Views:
Transcription
1 Option pricing using TR-BDF time stepping method by Ming Ma Aresearchpaper presented to the University of Waterloo in partial fulfilment of the requirement for the degree of Master of Mathematics in Computational Mathematics Supervisor: Peter.A.Forsyth and George Labahn Waterloo, Ontario, Canada, 01 c Ming Ma Public 01
2 I hereby declare that I am the sole author of this report. This is a true copy of the report, including any required final revisions, as accepted by my examiners. I understand that my report may be made electronically available to the public. ii
3 Abstract The Trapezoidal Rule with second order Backward Difference Formula (TR-BDF) time stepping method was applied to the Black-Scholes PDE for option pricing. It is proved that TR-BDF time stepping method is unconditionally stable, and compared to the usual Crank-Nicolson time stepping method, the TR-BDF shows fewer oscillations when computing the derivatives of the solution, which are important hedging parameters. iii
4 Acknowledgements Iwouldliketoexpressmyutmostgratitudeandappreciationofthisresearchexperience to the following people, who have contributed much to the completion of my project and also the development of my research skills, by providing me an educational and enlightening training: I would like to express my sincere thanks to Professor Peter.A.Forsyth and Professor George Labahn, for their continuous supervision and guidance during my project. Above all, I am most grateful for their encouragement for me to acquire diverse insights. iv
5 Dedication This is dedicated to the one I love. v
6 Table of Contents List of Tables List of Figures viii x 1 Introduction 1 Formulation 3.1 Basic background European options American options Black-Scholes model European options pricing Boundary conditions American options pricing Discretization Semi-discretization in time Spatial discretization Penalty method for American option pricing The Crank-Nicolson time stepping method and Rannacher smoothing vi
7 4 Von Neumann stability analysis of the TR-BDF time stepping method: European case 16 5 Numerical Tests European option case Numerical results Analysis American option case Numerical results Analysis and conclusion Summary 43 APPENDICES 44 A Definition of A-stable and L-stable 45 References 46 vii
8 List of Tables 3.1 Positive coefficient algorithm Penalty method for American option pricing Data for European Put Option Value of a European Put. Exact solution: Change is the difference in the solution from the coarser grid. Ratio is the ratio of changes on successive grids Data for American Put Option Value of an American Put. Change is the difference in the solution from the coarser grid. Ratio is the ratio of changes on successive grids Data for American Put Option Value of an American Put. Change is the difference in the solution from the coarser grid. Ratio is the ratio of changes on successive grids Data for American Put Option Value of an American Put. Change is the difference in the solution from the coarser grid. Ratio is the ratio of changes on successive grids Data for American Put Option Value of an American Put. Change is the difference in the solution from the coarser grid. Ratio is the ratio of changes on successive grids Data for American Put Option Value of an American Put. Change is the difference in the solution from the coarser grid. Ratio is the ratio of changes on successive grids Data for American Put Option viii
9 5.14 Value of an American Put. Change is the difference in the solution from the coarser grid. Ratio is the ratio of changes on successive grids ix
10 List of Figures 5.1 Value, delta(v S ), and gamma(v SS )ofaeuropeanput,σ=0.8, T =0.5, r=0.1, K=100. Left: TR-BDF time stepping method, right: Crank- Nicolson time stepping method with Rannacher smoothing. Top: option value (V ), middle: delta (V S ), bottom: gamma(v SS ) Value, delta(v S ), and gamma(v SS ) of an American Put, σ=0., T =0.5, r=0.1, K=100. Left: TR-BDF time stepping method, right: Crank- Nicolson time stepping method with Rannacher smoothing. Top: option value (V ), middle: delta (V S ), bottom: gamma(v SS ) Value, delta(v S ), and gamma(v SS ) of an American Put, σ=0.3, T =0.5, r=0.15, K=100. Left: TR-BDF time stepping method, right: Crank- Nicolson time stepping method with Rannacher smoothing. Top: option value (V ), middle: delta (V S ), bottom: gamma(v SS ) Value, delta(v S ), and gamma(v SS ) of an American Put, σ=0.4, T =5.00, r=0.03, K=100. Left: TR-BDF time stepping method, right: Crank- Nicolson time stepping method with Rannacher smoothing. Top: option value (V ), middle: delta (V S ), bottom: gamma(v SS ) Value, delta(v S ), and gamma(v SS ) of an American Put, σ=0.3, T =0.50, r=0.04, K=100. Left: TR-BDF time stepping method, right: Crank- Nicolson time stepping method with Rannacher smoothing. Top: option value (V ), middle: delta (V S ), bottom: gamma(v SS ) Value, delta(v S ), and gamma(v SS ) of an American Put, σ=0., T =1.00, r=0.05, K=100. Left: TR-BDF time stepping method, right: Crank- Nicolson time stepping method with Rannacher smoothing. Top: option value (V ), middle: delta (V S ), bottom: gamma(v SS ) x
11 5.7 Value, delta(v S ), and gamma(v SS ) of an American Put, σ=0.1, T =1.00, r=0.0, K=100. Left: TR-BDF time stepping method, right: Crank- Nicolson time stepping method with Rannacher smoothing. Top: option value (V ), middle: delta (V S ), bottom: gamma(v SS ) xi
12 Chapter 1 Introduction By holding an option, the holder obtains the right but not the obligation to enter into a transaction involving an underlying asset at a predetermined price at a specific date. The predetermined price is known as the strike price and the specified date is known as the maturity or expiry date of that option. There are different types of options. A call option gives the holder the right to buy an underlying asset while a put option gives the holder the right to sell the asset. European options can only be exercised at maturity whereas American options may be exercised any time by the expiry date of the option. Regardless of the different types of options, valuation and hedging of this type of financial contracts are always of importance. Different numerical methods can be used to calculate the price of the option. For example, the valuation of different types of options can be modelled as calculating the numerical solutions to corresponding partial differential equations (PDE). By assuming the price of the underlying asset follows a Geometric Brownian Motion, it was shown by Black and Scholes that the valuation of options can be done by solving a second order PDE with time and price of the underlying asset as two independent variables [3]. While the Black-Scholes equation is able to provide a closed form solution for pricing Europeans options, numerical methods are required for the case of American options. The PDEs are discretized and solutions are determined using a discrete set of time steps. When using numerical methods, it is always possible to have inaccuracies in the solutions, particularly if there are discontinuities in the payoff of the option, or its derivative. As 1
13 an example, when using the Crank-Nicolson time stepping method to solve the discretized system, one often encounters spurious oscillations in the Greeks (i.e. the approximate values of the first and second order derivatives of the option prices). Though Rannacher smoothing [13] canbeadoptedtoreducetheoscillationsforeuropeanoptionscase,itdoes not work well for American options. The Trapezoidal Rule with second order Backward Difference Formula (TR-BDF) can be classified as a fully implicit Runge Kutta method with second order accuracy. It has awiderangeofapplicationsinmanydifferentareassuchaselectronics[8], biology [14], mechanical engineering [] andelectricalengineering[1]. Since the TR-BDF method is mathematically L-stable (see appendix A for the definition of L-stability and A-stability), it has stronger stability properties than the Crank- Nicolson time stepping method which is only A-stable [1], we will use this method as the time stepping method to derive the solution to the option pricing problems under the Black-Scholes model. In this way, we expect oscillations in Greeks to be damped for both European and American options. The principle aims of this paper are as follows: Use the TR-BDF time stepping method to derive the solution to option pricing problems under the Black-Scholes model and thus price European and American options. Use Von Neumann stability analysis to analyse the stability properties of the TR- BDF time stepping method. Compare the results obtained from the TR-BDF time stepping method and the Crank-Nicolson time stepping method with Rannacher smoothing in terms of stability of Greeks and rate of convergence.
14 Chapter Formulation.1 Basic background.1.1 European options AEuropeancalloptionisthemostbasicexampleoffinancialderivatives. Byholding aeuropeancalloption,onehastheright,butnottheobligationtobuyanunderlying asset at a specific maturity or expiry time T in the future at a specific strike price K. On the other hand, by holding a European put option, instead of buying, one has the right, but not the obligation to sell an underlying asset at the specific expiry day and strike price. The payoff of a European option can be written mathematically in the following form: max(k S, 0), for put options Option Payoff = (.1) max(s K, 0), for call options where S denotes the price of the underlying asset..1. American options The key difference between an American option and a European option is that an American option can be exercised at any time before its maturity or expiry date. When exercising the 3
15 American option, the payoff is the same as the European option. However, due to American option s early exercise feature, it is always priced no less than a European option with the same expiry date and strike price, otherwise an arbitrage opportunity is created.. Black-Scholes model In the year of 1973, for the purpose of pricing financial derivatives accurately, a partial differential equation was derived by Black and Scholes [3]. This equation is now referred to as the Black-Scholes equation, which is the most fundamental equation of the current mathematical finance studies. The following assumptions should be kept in mind when using the Black-Scholes equation: The price of the underlying asset follows geometric Brownian motion with constant drift and volatility. The risk-free rate of return is a constant and cash can be borrowed or lent at this rate. There are no arbitrage opportunities existing in the market. There are no transaction costs when purchasing or selling the underlying assets. Short selling is permitted in the market...1 European options pricing APDEforpricingEuropeanoptionswasderivedbyBlackandScholes: Consider an underlying asset with price S and assume the price follows the log-normal stochastic process ds = µsdt + σsdz, (.) 4
16 where µ is the drift rate, σ is volatility, and dz is the increment of a Wiener Process which is defined as: dz = φ dt (.3) where φ N(0, 1) follows the standard normal distribution and dt is defined as the increment of time. Suppose we construct a hedging portfolio which consists of a long position in one option whose value is given by V and a short position in a number of (α shares) underlying asset. Then the value of the portfolio is given by: In a small time dt, P P + dp,wehave: P = V αs (.4) dp = dv (α)ds (.5) Considering that α actually depends on S, ifwetakethetruedifferentialofp,weobtain: dp = dv (α)ds Sd(α). (.6) Since we are not allowed to peek into the future, so α can not contain any information about the future asset price movements. As a result, Ito s lemma is used here, and we have: dv =(µs V S + σ S V S + V )dt + σs V dz. (.7) t S Substituting equation (.) and (.7) into (.6), we obtain: dp = σs( V S V α)dz +(µs S + σ S V S + V t αµs)dt (.8) If we let α = V,theriskwhicharisefromtherandomnessofthepriceoftheunderlying S asset can be fully hedged. As a result, we can make this portfolio risk-less over the time interval dt and the change of the value of the portfolio is deterministic. Therefore, by holding the portfolio, according to the no-arbitrage principle, a risk-free rate of return should be obtained: dp = rpdt (.9) where r is the risk-free interest rate. As a result, the following equation is obtained: V t + 1 σ S V V + rs rv =0, (.10) S S 5
17 which is the Black-Scholes equation. and Define L as: τ as: LV = 1 σ S V V + rs S S rv, (.11) τ = T t (.1) where T is the expiry date, so τ is the time variable running backwards. Then we can rewrite the Black-Scholes equation as: V τ = LV. (.13).. Boundary conditions The usual boundary conditions are typically: and, as S 0, the Black-Scholes PDE reduces to the ODE: put S V 0 (.14) call S V S (.15) V τ We can simply solve this ODE at S =0. rv =0. (.16)..3 American options pricing Since American options have the feature of early exercise, which means the holder of an American option can choose to exercise at any time by the expiry date of the option and receive a payoff: Payoff = P (S, τ) (.17) 6
18 the American option pricing problem can be viewed as a linear complementarity problems (LCP) [7]. The payoff of an American option can be denoted as: V (S, τ =0)=max(K S, 0), Option Payoff = V (S, τ =0)=max(S K, 0), for put options for call options (.18) where K is the strike price at which the transaction is carried out. The price of an American option cannot be less than its payoff, otherwise there is an arbitrage opportunity existing in the market. In addition, because the American option may not be exercised at the optimal time by the holder, the value of the portfolio created may not be able to increase at the risk-free rate of return. With the two constraints above, the linear complementarity problem can be stated as: min[v τ ( 1 σ S V V + rs S S rv ),V P ] 0 (.19) 7
19 Chapter 3 Discretization There is no analytical solution to the linear complementarity problem in equation (.17), so in order to price American option, numerical techniques are required. 3.1 Semi-discretization in time First, we define the discretization of time as: (τ n ) n {0,...,N}, and set τ = τ n 1 τ n where we have τ 0 as the time of option expiry and τ N is the valuation time. We will first semi-discretize the V τ term in equation (.17). For example, the Crank- Nicolson time stepping method would result in: V n+1 = V n + τ L(V n )+L(V n+1 ), (3.1) However, this method is not a monotone scheme, hence it may be prone to oscillations when computing the Greeks. Recently, the TR-BDF method has been proposed to alleviate this problem. TR-BDF algorithm uses the following time semi-discretization: For the TR-BDF time stepping method, there are two stages at each time step: the first stage is the Trapezoidal method and the second stage is the Backward Difference Formula method which is applied to the first stage output and initial value at τ n.sowecan 8 The
20 obtain the value at τ n+1. By using the TR-BDF time stepping method to discretize the V τ term, we have where and 0 <α<1. V n+1 = 1 1 ( α) α V V = V n + α τ (1 α) V n +(1 α) τl(v n+1 ), (3.) α L(V n )+L(V ) (3.3) If we choose α to be 1, equation (3.3) will become the Crank-Nicolson time stepping method. Now we define the local truncation error (LTE) as: l n+1 = y n+1 y(t n+1 ) (3.4) assuming the value y n computed in previous step is equal to the exact solution for y at time t = t n. If we choose α to be, then the local truncation error is minimized [1], because the divided-difference estimate of the local truncation error is: where LT E n+1 = C τ 3 nv (3) (3.5) C = 3α +4α. (3.6) 1( α) Although the TR-BDF method has two stages, it is still a one step method. 3. Spatial discretization Define the discretization of the price of underlying asset by (S j ) j {0,...,m}, h j = S j S j 1. Using central, forward and backward difference method, we have: 9
21 for central difference, for forward difference, and n V S j n V S j = V n j+1 V n j 1 h j+1 + h j (3.7) = V j+1 n Vj n (3.8) h j+1 for backward difference. In addition, we have V S n j n V = V j n Vj 1 n (3.9) S j h j = h jvj+1 n (h j+1 + h j )Vj n + h j+1 V n h j h j+1 (h j+1 + h j ) j 1. (3.10) Substituting these discrete approximations into the Trapezoidal stage as well as the BDF stage, we can get tridiagonal linear systems for the unknown values Vj and V n+1 Define the vectors V n+1 =[V0 n+1 V1 n+1...vm n+1 ], V n =[V0 n V1 n...vm] n, V =[V0 V1...Vm] and let tridiagonal matrices ˆM and ˆN be defined so that for row j, wehave: j. and [ ˆMV n ] j = α τa j V n j 1 + α τ(a j + b j + r) V n j α τb j V n j+1, (3.11) [ ˆNV n ] j = (1 α) τa j V n α j 1 + (1 α) τ(a j + b j + r) α V n j (1 α) τb j V n α j+1. (3.1) The a j and b j are defined as: 10
22 a central j = b central j = σ Sj (S j S j 1 )(S j+1 S j 1 ) rs j S j+1 S j 1 σ Sj (S j+1 S j )(S j+1 S j 1 ) + rs j S j+1 S j 1 (3.13) (3.14) in the case of central differencing, a forward j = b forward j = σ Sj (S j S j 1 )(S j+1 S j 1 ) σ Sj (S j+1 S j )(S j+1 S j 1 ) + rs j S j+1 S j (3.15) (3.16) in the case of forward differencing, or a backward j = b backward j = σ Sj (S j S j 1 )(S j+1 S j 1 ) rs j S j S j 1 σ Sj (S j+1 S j )(S j+1 S j 1 ) (3.17) (3.18) in the case of backwards differencing. It is important to ensure that all a j and b j are positive for stability reasons. The following algorithm in Table (3.1) is adopted to decide between central or upstream (forward or backward) discretization at each node. Thus, for the Trapezoidal stage we have [I + ˆM]V =[I ˆM]V n, (3.19) and for the BDF stage we have [I + ˆN]V n+1 1 =[ α( α) ]V (1 α) [ α( α) ]V n. (3.0) The boundary conditions will be considered in next section. 11
23 For j=0,...,m If(a central j a j = a central j b j = b central j ElseIf(a forward j a j = a forward j b j = b forward j Else a j = a backward j b j = b backward j EndIf EndFor 0andb central j 0)then 0andb forward j f 0)then Table 3.1: Positive coefficient algorithm 3.3 Penalty method for American option pricing In order to solve the American option pricing problem, which is also a linear complementarity problem, we can rewrite equation (.17) as a single equation with a non-linear penalty term Q(V,P), where P (S, τ) is the payoff of an American option: V τ = 1 σ S V V + rs S S where the penalty term Q(V,P) isdefinedas: rv + Q(V,P), (3.1) Here ρ is chosen to be large enough so that: Q(V,P) =ρ max(p V,0). (3.) or V τ = 1 σ S V V + rs S S rv if V>P, (3.3) V τ > 1 σ S V V + rs rv if V = P. (3.4) S S Therefore, Q(V,P) =ρ, where = Q(V,P)/ρ, assumingq is bounded. 1
24 For the TR-BDF time stepping method, the discrete penalized equations are: ( α)v n+1 j 1 V α j (1 α) τ V j V n j α τ + (1 α) α V n j = 1 [(LV ) j +(LV ) n j ]+q j (3.5) = (LV ) n+1 j + p n+1 j (3.6) where the penalty term q j and p n+1 j are defined as: and q j = 1 (P α τ j Vj )Large, if Vj <P j 0, otherwise (3.7) p n+1 j = ( α) (P (1 α) τ j V n+1 j )Large, if V n+1 j <P j 0, otherwise (3.8) where Large is a large number and P j is the payoff at j th node. and To solve those non-linear equations, we define diagonal matrices Q and P as: Q(V ) ij = P (V n+1 ) ij = Large, if i = j and V j 0, otherwise <P j ( α)large, if i = j and V n+1 j <P j 0, otherwise (3.9) (3.30) Then we are able to rewrite the non-linear equations as: and [I + ˆM + Q(V )]V =[I ˆM]V n +[ Q(V )]P, (3.31) [I + ˆN + P (V n+1 )]V n+1 1 =[ α( α) ]V (1 α) [ α( α) ]V n +[ P (V n+1 )]P. (3.3) 13
25 Next, we will use the following algorithm to price American options with variable timesteps. Let (V ) k and (V n+1 ) k be the k th estimate for V and V n+1 respectively. Let (V ) 0 = V n and (V n+1 ) 0 = V n. The algorithm for pricing an American option with variable timesteps can be stated as: While τ<t For k = 0,...until convergence [I + ˆM + Q (V ) k ](V ) k+1 =[I ˆM]V n + Q (V ) k P If EndFor max (V j )k+1 (Vj )k j max(1, (V j )k+1 ) <tol quit For l =0,...until convergence [I + ˆN + P (V n+1 ) l ](V n+1 ) l+1 1 =[ ]V [ (1 α) ]V n +[ P (V n+1 ) l ]P α( α) α( α) n+1 (Vj ) If max l+1 (V n+1 j ) l <tol quit j max(1, (V n+1 j ) l+1 ) EndFor τ = τ + τ MaxRelChange = max j τ = dnorm MaxRelChange V n = V n+1 EndWhile V n+1 V n max(1, V n+1, V n ) τ Table 3.: Penalty method for American option pricing 14
26 In the algorithm, T is time to maturity, dnorm is timestep size control parameter, tol is tolerance value and Large is defined to be 1 tol. 3.4 The Crank-Nicolson time stepping method and Rannacher smoothing As mentioned above, if we choose the value of α to be 1, then the Crank-Nicolson time stepping method can be written as the first stage of the TR-BDF time stepping method.[1]. The Crank-Nicolson time stepping is known to be only A-stable [1]. As a result, spurious oscillations in the Greeks can be introduced [9]. Rannacher smoothing, which adds two backward Euler steps before the Crank-Nicolson, is used to smooth off the payoff at maturity and reduce the oscillation problem [13]. We will illustrate that, though Rannacher smoothing is effective for European options, it does not work as well for American options. 15
27 Chapter 4 Von Neumann stability analysis of the TR-BDF time stepping method: European case It has been shown that the Crank-Nicolson time stepping method is unconditionally stable [10]. Here, we would also like to study the stability properties of the TR-BDF time stepping method. In this chapter, Von Neumann stability analysis is carried out for the TR-BDF time stepping method. The coefficients are assumed to be constant and the grid to be equally spaced in log S coordinates. By using the change of variable: we can change the Black-Scholes equation x =logs, S =exp(x) (4.1) into the form of: V τ = 1 σ S V SS + rsv S rv (4.) where V (x, τ) =V exp (x),τ. V τ = 1 σ V xx +(r 1 σ )V x rv (4.3) 16
28 The trapezoidal stage and second order backward difference stage of the TR-BDF time stepping method can be written as: V j V n j α τ = 1 [1 σ ( V j+1 V j + V j 1 x )+(r 1 σ ) V + 1 [1 σ ( V n j+1 V n j + V n j 1 )+(r 1 x σ ) V j+1 V j 1 x rv j] n j+1 V n j 1 rv n j ], x (4.4) and V n+1 j = 1 1 α α V j (1 α) V n j +(1 α) τ[ 1 α σ ( V n+1 j+1 V n+1 j + V n+1 j 1 ) x +(r 1 σ ) V n+1 j+1 V n+1 j 1 rv n+1 j ]. x (4.5) If we let: then the equations above can be rewritten as: V j[1 + (a + b + r)α τ ] α τ bv j+1 α τ av j 1 a = 1 1 σ x (r 1 1 σ ) x, b = 1 1 σ x +(r 1 (4.6) 1 σ ) x, and = V n j [1 (a + b + r)α τ ]+α τ bv n j+1 + α τ av n j 1, (4.7) V n+1 j [1 + (a + b + r) 1 α α 1 α n+1 τ] τbv α j+1 1 α α τav n+1 j 1 17
29 = 1 α( α) V (1 α) j α( α) V n j. (4.8) Let V n =[V n 0, V n 1,...,V n M] be the discrete solution vector to equation (4.4) and (4.5). Assume the initial solution vector is perturbed by: V 0 = V 0 + E 0 (4.9) where E n =[E n 0,...,E n M ] is the perturbation vector. Since V satisfies the equations: V j [1 + (a + b + r)α τ ] α τ b V j+1 α τ a V j 1 and = V j n [1 (a + b + r)α τ ]+α τ b Vj+1 n + α τ a Vj 1, n (4.10) V n+1 j [1 + (a + b + r) 1 α α τ] 1 α α τb V n+1 j+1 1 α α τa V n+1 j 1 = 1 V α( α) j (1 α) V α( α) j n, (4.11) by subtracting equations (4.10) and (4.7), (4.11) and (4.8) accordingly, we have: E j [1 + (a + b + r)α τ ] α τ be j+1 α τ ae j 1 and = E n j [1 (a + b + r)α τ ]+α τ ben j+1 + α τ aen j 1, (4.1) 18
30 E n+1 j [1 + (a + b + r) 1 α 1 α τ] α α τben+1 j+1 1 α α τaen+1 j 1 = 1 (1 α) α( α) E j α( α) En j. (4.13) Here, by using Von Neumann stability analysis approach, we want to prove that the initial perturbation is bounded when the number of steps becomes large. In order to use the Fourier transform method, we assume that the boundary conditions can be replaced by periodic conditions. As a result, the inverse discrete Fourier transform of Ej n is defined as: E n j = 1 X N N (C k ) n π exp (i jk), (4.14) N k= N +1 where C k is the discrete Fourier coefficient of E, i is 1andthewidthofthedomainis defined as X N =x N x N +1. LetW =exp(π N i)sothatwecanwriteen j as: E n j = 1 X N N k= N +1 (C k ) n W jk. (4.15) The inverse discrete Fourier transform of E n+1 j and E j are then given by: and E n+1 j = 1 X N N k= N +1 (C k ) n+1 W jk, (4.16) E j = 1 X N N k= N +1 C kw jk. (4.17) Now we substitute ( ) into (4.1) and (4.13). For equation (4.1) we get: 19
31 = 1 X N 1 X N N N k= N +1 C kw jk [1 + (a + b + r)α k= N +1 (C k ) n W jk [1 (a + b + r)α while for equation (4.13) we have: τ ] α τ b 1 X N α τ a 1 X N τ ]+α τ b 1 X N +α τ a 1 X N N k= N +1 C kw (j+1)k N k= N +1 C kw (j 1)k N k= N +1 (C k ) n W (j+1)k N k= N +1 (C k ) n W (j 1)k, (4.18) 1 X N N (C k ) n+1 W jk 1 α [1 + (a + b + r) α τ] 1 α α τb 1 k= N +1 = 1 α( α) 1 X N X N 1 α α τa 1 X N N CkW jk (1 α) α( α) k= N +1 N k= N +1 (C k ) n+1 W (j+1)k N k= N +1 (C k ) n+1 W (j 1)k 1 X N N k= N +1 (C k ) n W jk. (4.19) For all Fourier components C k,wewilllookatthemseparately. Forequation(4.18),for each k we have: C k W jk [1 + (a + b + r)α τ ] α τ bc k W (j+1)k α τ ac k W (j 1)k =(C k ) n W jk [1 (a + b + r)α τ ]+α τ b(c k) n W (j+1)k + α τ a(c k) n W (j 1)k, (4.0) 0
32 and for equation (4.19), we have: (C k ) n+1 W jk [1 + (a + b + r) 1 α α τ] 1 α α τb(c k) n+1 W (j+1)k 1 α α τa(c k) n+1 W (j 1)k Dividing both equations by (C k ) n W jk,equation(4.0)becomes: = 1 α( α) C kw jk (1 α) α( α) (C k) n W jk. (4.1) Ck [1 + (a + b + r)α τ (C k ) n ] α τ b C k (C k ) W k α τ n a C k (C k ) W k n =[1 (a + b + r)α τ ]+α τ bw k + α τ aw k, and equation (4.1) becomes: (4.) By factoring out becomes: C k [1 + (a + b + r) 1 α α τ] 1 α α τbc kw k 1 α α τac kw k and equation (4.3) becomes: = 1 C (4.3) k (1 α) α( α) (C k ) n α( α). C k (C k ) n and C k from the two equations above respectively, equation (4.) Ck α τ α τ [1 (a + b + r) ]+ = bw k + α τ (C k ) n [1 + (a + b + r) α τ ] α τ bw k α τ aw k aw k, (4.4) Since C k = 1 Ck α( α) (C k (1 α) ) n α( α) [1 + (a + b + r) 1 α 1 α τ] α α τbwk 1 α α τaw k. (4.5) a = 1 1 σ x (r 1 1 σ ) x,b= 1 1 σ x +(r 1 1 σ ) x, (4.6) 1
33 we can simplify the expressions by noting: a + b + r = σ x + r, and bw k + aw k = σ x cos( π N k)+i 1 x (r 1 σ )sin( π N k). Thus we can write equation (4.4) as: C k (C k ) σ [1 ( + r)α τ x = ]+ 1[ α τσ cos( π α τ )k + i x N n [1 + ( σ + r)α τ ] 1[ α τσ cos( π x x N and equation (4.5) as: (r 1 x σ )sin( πk)] N (4.7) k)], k)+i α τ x (r 1 σ )sin( π N C k = [1 + ( σ + r) 1 α x α 1 Ck α( α) (C k (1 α) ) n α( α) 1 α τ] [ τσ cos( π τ k)+i (r (4.8) 1 α x N x σ )sin( π k)]. N Substituting (4.7) into (4.8), we can obtain the expression of C k : C k = [1 1+(1 α) α( α) ( σ x τ +r)α ]+ 1+(1 α) α( α) 1 [ α τσ x cos( π α τ k)+i N x (r 1 σ )sin( π N k)] [1+( σ τ x +r)α ] 1 [ α τσ x cos( π α τ k)+i N x (r 1 σ )sin( π N k)] [1 + ( σ + r) 1 α 1 α τ] [ τσ cos( π x α α x N k)+i τ x (r 1 σ )sin( π N k)]. (4.9) The norm of E n is: E n = N j= N +1 E n j (E n j ) (4.30) where (E n j ) is the complex conjugate of E n j.since
34 E n j = 1 X N N k= N +1 (C k ) n W jk, (4.31) we have E n = N X N N k= N +1 ( C k ) n. (4.3) Thus, in order for E n to be bounded, each Fourier coefficient C k n needs to be bounded, when n approaches infinity. Consequently, we need: C k 1. (4.33) In our case, we can find the value of C k from equation (4.9): C k = {1 1+(1 α) α( α) [ α τ α τ r+ {1+[ α τ α τ r+ σ x (1 cos( π N k))]} +{ 1+(1 α) α( α) σ x (1 cos( π N k))]} +{ 1 {1+ (1 α) [ α τ r + α τ σ (1 cos( π α( α) x N k))]} + { (1 α) α( α) If we write: M = 1+(1 α),n = α( α) then, since 0 <α<1, we have M>0,N >0. Also 1 α τ x (r 1 σ )sin( π N k)} α τ x (r 1 σ )sin( π N k)} 1 α τ x (r 1 σ )sin( π N k)}. (4.34) (1 α) α( α), (4.35) P = α τ r + α τ σ (1 cos(π k)), x N (4.36) Q = 1 α τ x (r 1 σ )sin( π k), N (4.37) 3
35 where P>0. Then C k can be written as: C k = (1 MP) +(MQ) (1+P ) +Q (1 + NP) +(NQ). (4.38) We can further simplify the equation above: C k = 1+M P + M Q MP 1+(N +1)P +(N +1)Q + R (4.39) where R =(N +1)P +(N + N)P 3 +4NP +N Q P + N (P 4 + Q 4 )+N Q P +NPQ. (4.40) Notice that R>0. A little manipulation shows that and so we have: N +1=M (4.41) C k = (1 + M P + M Q ) MP (1 + M P + M Q )+R. (4.4) Since MP > 0, we have C k < 1andso C k < 1. As the initial perturbation is bounded as the number of steps increases, we prove that the TR-BDF time stepping method is unconditionally stable. 4
36 Chapter 5 Numerical Tests In this chapter, we solve the Black-Scholes equation to price European options and American options. Both the TR-BDF time stepping method and the Crank-Nicolson time stepping method are used, so we can compare the two in terms of the stability of Greeks (V S, V SS )andrateofconvergence. For the American option pricing problem, we formally state the problem as a linear complementarity problem: min[v τ ( 1 σ S V V + rs S S where P (S, τ) isthepayoffconditionwhichas: rv ),V P ] 0. (5.1) P (S, τ =0)=max(K S, 0) for a put (5.) P (S, τ =0)=max(S K, 0) for a call (5.3) with K the strike price. The boundary conditions at S 0is: And boundary conditions for large S is: min[v τ rv, V P ] 0. (5.4) 5
37 V 0,S ; for a put (5.5) V S, S ; for a call (5.6) We will use the penalty method [7] to solve this linear complementarity problem. 5.1 European option case In this section, we will use both the TR-BDF and the Crank-Nicolson time stepping methods to solve the Black-Scholes equation and thus price European options. In addition, the properties of Greeks stability and rate of convergence of both methods are compared Numerical results With the data assumed in Table 5.1, a convergence study for the option price was carried out. A convergence table, which has a series of non-uniform grids, is used. In addition, variable timestep sizes are used and the initial timestep size on the coarsest grid is τ = T/5. In the convergence study, on each grid refinement, new grid nodes are added halfway between the original ones. The new grid has twice as many nodes as the previous grid, the timestep control parameter dnorm is halved, and the initial timestep size is divided by four. 6
38 Variables Values σ 0.8 r 0.10 Time to expiry (T ) 0.5 years Strike Price $100 Initial asset price S 0 $100 Table 5.1: Data for European Put Option Nodes Timesteps Value Change Ratio TR-BDF Crank-Nicolson (Rannacher smoothing) Table 5.: Value of a European Put. Exact solution: Change is the difference in the solution from the coarser grid. Ratio is the ratio of changes on successive grids. 7
39 Figure 5.1: Value, delta(v S ), and gamma(v SS )ofaeuropeanput,σ=0.8, T =0.5, r=0.1, K=100. Left: TR-BDF time stepping method, right: Crank-Nicolson time stepping method with Rannacher smoothing. Top: option value (V ), middle: delta (V S ), bottom: gamma(v SS ). 8
40 5.1. Analysis As can be seen from Table 5., both time stepping methods converge. In addition, second order convergence was obtained by both methods. It is known that, the Crank-Nicolson time stepping method is only A-stable [1] and can introduce spurious oscillations in the Greeks [9]. As can be observed in the Figure 5.1, there are no oscillations in delta and gamma for the Crank-Nicolson time stepping method using Rannacher smoothing. On the other hand, no oscillations are observed in delta and gamma for the TR-BDF method as well. 5. American option case In this section, we will use both the TR-BDF and the Crank-Nicolson time stepping methods to solve American option pricing problem with a penalty method. In addition, the stability of Greeks and rate of convergence of both methods are compared. A convergence study of pricing different American option examples was carried out in a similar way as the European case. The convergence table, which has a series of nonuniform grids, is used. The tolerance value tol is chosen to be 10 6.Inaddition,variable timestep sizes are used and the initial timestep size on the coarsest grid is τ = T/5. In the convergence study, on each grid refinement, the new grid has twice as many nodes as the previous grid, the timestep control parameter dnorm is halved, and the initial timestep size is divided by four. The convergence table, and figures of the Greeks are given in next section. 9
41 5..1 Numerical results Variables Values σ 0. r 0.10 Time to expiry (T ) 0.5 years Strike Price $100 Initial asset price S 0 $100 Table 5.3: Data for American Put Option Nodes Timesteps Value Change Ratio TR-BDF Crank-Nicolson (Rannacher smoothing) Table 5.4: Value of an American Put. Change is the difference in the solution from the coarser grid. Ratio is the ratio of changes on successive grids. 30
42 Figure 5.: Value, delta(v S ), and gamma(v SS ) of an American Put, σ=0., T =0.5, r=0.1, K=100. Left: TR-BDF time stepping method, right: Crank-Nicolson time stepping method with Rannacher smoothing. Top: option value (V ), middle: delta (V S ), bottom: gamma(v SS ). 31
43 Variables Values σ 0.3 r 0.15 Time to expiry (T ) 0.5 years Strike Price $100 Initial asset price S 0 $100 Table 5.5: Data for American Put Option Nodes Timesteps Value Change Ratio TR-BDF Crank-Nicolson (Rannacher smoothing) Table 5.6: Value of an American Put. Change is the difference in the solution from the coarser grid. Ratio is the ratio of changes on successive grids. 3
44 Figure 5.3: Value, delta(v S ), and gamma(v SS ) of an American Put, σ=0.3, T =0.5, r=0.15, K=100. Left: TR-BDF time stepping method, right: Crank-Nicolson time stepping method with Rannacher smoothing. Top: option value (V ), middle: delta (V S ), bottom: gamma(v SS ). 33
45 Variables Values σ 0.4 r 0.03 Time to expiry (T ) 5.00 years Strike Price $100 Initial asset price S 0 $100 Table 5.7: Data for American Put Option Nodes Timesteps Value Change Ratio TR-BDF Crank-Nicolson (Rannacher smoothing) Table 5.8: Value of an American Put. Change is the difference in the solution from the coarser grid. Ratio is the ratio of changes on successive grids. 34
46 Figure 5.4: Value, delta(v S ), and gamma(v SS ) of an American Put, σ=0.4, T =5.00, r=0.03, K=100. Left: TR-BDF time stepping method, right: Crank-Nicolson time stepping method with Rannacher smoothing. Top: option value (V ), middle: delta (V S ), bottom: gamma(v SS ). 35
47 Variables Values σ 0.3 r 0.04 Time to expiry (T ) 0.50 years Strike Price $100 Initial asset price S 0 $100 Table 5.9: Data for American Put Option Nodes Timesteps Value Change Ratio TR-BDF Crank-Nicolson (Rannacher smoothing) Table 5.10: Value of an American Put. Change is the difference in the solution from the coarser grid. Ratio is the ratio of changes on successive grids. 36
48 Figure 5.5: Value, delta(v S ), and gamma(v SS ) of an American Put, σ=0.3, T =0.50, r=0.04, K=100. Left: TR-BDF time stepping method, right: Crank-Nicolson time stepping method with Rannacher smoothing. Top: option value (V ), middle: delta (V S ), bottom: gamma(v SS ). 37
49 Variables Values σ 0. r 0.05 Time to expiry (T ) 1.00 years Strike Price $100 Initial asset price S 0 $100 Table 5.11: Data for American Put Option Nodes Timesteps Value Change Ratio TR-BDF Crank-Nicolson (Rannacher smoothing) Table 5.1: Value of an American Put. Change is the difference in the solution from the coarser grid. Ratio is the ratio of changes on successive grids. 38
50 Figure 5.6: Value, delta(v S ), and gamma(v SS ) of an American Put, σ=0., T =1.00, r=0.05, K=100. Left: TR-BDF time stepping method, right: Crank-Nicolson time stepping method with Rannacher smoothing. Top: option value (V ), middle: delta (V S ), bottom: gamma(v SS ). 39
51 Variables Values σ 0.1 r 0.0 Time to expiry (T ) 1.00 years Strike Price $100 Initial asset price S 0 $100 Table 5.13: Data for American Put Option Nodes Timesteps Value Change Ratio TR-BDF Crank-Nicolson (Rannacher smoothing) Table 5.14: Value of an American Put. Change is the difference in the solution from the coarser grid. Ratio is the ratio of changes on successive grids. 40
52 Figure 5.7: Value, delta(v S ), and gamma(v SS ) of an American Put, σ=0.1, T =1.00, r=0.0, K=100. Left: TR-BDF time stepping method, right: Crank-Nicolson time stepping method with Rannacher smoothing. Top: option value (V ), middle: delta (V S ), bottom: gamma(v SS ). 41
53 5.. Analysis and conclusion As can be observed from the examples in last section, results obtained by both time stepping methods converged to the same value and both methods achieved second order convergence. Figure compare option values, delta (V S ), gamma (V SS ) for both time stepping methods. It can be observed that though the value and delta are similar for both methods, Rannacher smoothing cannot reduce all the oscillations in gamma. So there are oscillations observed in the gamma corresponding to the early exercise boundary for the Crank-Nicoson time stepping method. This is due to the stability properties of the Crank-Nicolson time stepping method [4]. However, for the TR-BDF time stepping method which has better stability properties, there are still no oscillations observed in Greeks. So, with comparable speed of convergence, the TR-BDF time stepping method offers better results for Greeks, which are of practical importance as commonly used hedging parameters. 4
54 Chapter 6 Summary In this paper, the Trapezoidal Rule with the second order Backward Difference Formula (TR-BDF) time stepping method was studied for option pricing. The Crank-Nicolson time stepping method, as an alternative method, was compared against this method in the study. We first derived the solution to the option pricing problem under the Black-Scholes model using the TR-BDF time stepping method which is a second order fully implicit Runge Kutta method. Then Von Neumann stability analysis of the TR-BDF time stepping method was carried out. It was known that the Crank-Nicolson time stepping method is algebraically unconditionally stable, and in our analysis, it was proved that the TR-BDF time stepping method is also unconditionally stable. When pricing options, the Crank-Nicolson time stepping method can introduce spurious oscillations, particularly in Greeks. We did numerical tests and showed that, though Rannacher smoothing can fix this problems for European options priced using the Crank- Nicolson time stepping method, it does not work well for American options. Obvious oscillations corresponding to early exercise boundaries can still be observed in gamma in the American option case. On the other hand, the TR-BDF time stepping method, with better stability properties, does not introduce any oscillations in Greeks and has comparable speed of convergence as the Crank-Nicolson time stepping method with Rannacher smoothing. 43
55 APPENDICES 44
56 Appendix A Definition of A-stable and L-stable Consider a linear model problem: where λ is a complex number. u (t) =λu(t) (A.1) We say a numerical method is A-stable if its stability region contains the whole left half plane. If we define z = λ t, A-stable can be written as: {z C : Re(z) 0}. For example, forward Euler scheme leads to a discretization of u j+1 =(1+λ t)u j for the problem stated above and its stability region is 1+z < 1. By definition, it is not A-stable because the stability region is a disc of radius centred at the point -1. L-stable is stronger than A-stable, we define a numerical method is L-stable if it is A-stable and u j+1 u j 0as z. 45
57 References [1] R.E. Bank, W.M. Coughran, W. Fichtner, E.H. Grosse, D.J. Rose, and R.K. Smith. Transient simulation of silicon devices and circuits. Computer-Aided Design of Integrated Circuits and Systems, IEEE Transactions on, 4(4): ,1985. [] K.J. Bathe. Conserving energy and momentum in nonlinear dynamics: A simple implicit time integration scheme. Computers & structures, 85(7): ,007. [3] F. Black and M. Scholes. The pricing of options and corporate liabilities. The Journal of Political Economy, pages ,1973. [4] T.F. Coleman, Y. Li, and A. Verma. A Newton method for American option pricing. Journal of Computational Finance, 5(3):51 78,00. [5] S. Dharmaraja. An analysis of the TR-BDF integration scheme. PhD thesis, Massachusetts Institute of Technology, 007. [6] S. Dharmaraja, Y. Wang, and G. Strang. Optimal stability for trapezoidal backward difference split-steps. IMA Journal of Numerical Analysis, 30(1): ,010. [7] P.A. Forsyth and K.R. Vetzal. Quadratic convergence of a penalty method for valuing american options. SIAM Journal on Scientific Computation, 3:096 13,00. [8] C.L. Gardner, W. Nonner, and R.S. Eisenberg. Electrodiffusion model simulation of ionic channels: 1d simulations. Journal of Computational Electronics, 3(1):5 31, 004. [9] M. Giles and R. Carter. Convergence analysis of crank-nicolson and rannacher timemarching. Journal of Computational Finance, 9(4):89,006. [10] M. Grasselli and D. Pelinovsky. Numerical mathematics. Jones & Bartlett Learning,
58 [11] F. Le Floc h. TR-BDF for stable American Option Pricing. Journal of Computational Finance, 013. Available at SSRN: or [1] R.J. LeVeque. Finite Difference Methods for Ordinary and Partial Differential Equations. Society for Industrial and Applied Mathematics (SIAM), 007. [13] R. Rannacher. Finite element solution of diffusion problems with irregular data. Numerische Mathematik, 43():309 37,1984. [14] R. Tyson, LG Stern, and R.J. LeVeque. Fractional step methods applied to a chemotaxis model. Journal of mathematical biology, 41(5): ,000. [15] P. Wilmott. Frequently asked questions in quantitative finance. Wiley, 009. [16] H. Windcliff, P.A. Forsyth, and K.R. Vetzal. Analysis of the stability of the linear boundary condition for the black-scholes equation. Journal of Computational Finance, 8:65 9,
Infinite Reload Options: Pricing and Analysis
Infinite Reload Options: Pricing and Analysis A. C. Bélanger P. A. Forsyth April 27, 2006 Abstract Infinite reload options allow the user to exercise his reload right as often as he chooses during the
More informationResearch Article Exponential Time Integration and Second-Order Difference Scheme for a Generalized Black-Scholes Equation
Applied Mathematics Volume 1, Article ID 796814, 1 pages doi:11155/1/796814 Research Article Exponential Time Integration and Second-Order Difference Scheme for a Generalized Black-Scholes Equation Zhongdi
More informationPDE Methods for Option Pricing under Jump Diffusion Processes
PDE Methods for Option Pricing under Jump Diffusion Processes Prof Kevin Parrott University of Greenwich November 2009 Typeset by FoilTEX Summary Merton jump diffusion American options Levy Processes -
More informationShort-time-to-expiry expansion for a digital European put option under the CEV model. November 1, 2017
Short-time-to-expiry expansion for a digital European put option under the CEV model November 1, 2017 Abstract In this paper I present a short-time-to-expiry asymptotic series expansion for a digital European
More informationGreek parameters of nonlinear Black-Scholes equation
International Journal of Mathematics and Soft Computing Vol.5, No.2 (2015), 69-74. ISSN Print : 2249-3328 ISSN Online: 2319-5215 Greek parameters of nonlinear Black-Scholes equation Purity J. Kiptum 1,
More informationA distributed Laplace transform algorithm for European options
A distributed Laplace transform algorithm for European options 1 1 A. J. Davies, M. E. Honnor, C.-H. Lai, A. K. Parrott & S. Rout 1 Department of Physics, Astronomy and Mathematics, University of Hertfordshire,
More informationMAFS Computational Methods for Pricing Structured Products
MAFS550 - Computational Methods for Pricing Structured Products Solution to Homework Two Course instructor: Prof YK Kwok 1 Expand f(x 0 ) and f(x 0 x) at x 0 into Taylor series, where f(x 0 ) = f(x 0 )
More informationProject 1: Double Pendulum
Final Projects Introduction to Numerical Analysis II http://www.math.ucsb.edu/ atzberg/winter2009numericalanalysis/index.html Professor: Paul J. Atzberger Due: Friday, March 20th Turn in to TA s Mailbox:
More informationOption Pricing Models for European Options
Chapter 2 Option Pricing Models for European Options 2.1 Continuous-time Model: Black-Scholes Model 2.1.1 Black-Scholes Assumptions We list the assumptions that we make for most of this notes. 1. The underlying
More informationCS 774 Project: Fall 2009 Version: November 27, 2009
CS 774 Project: Fall 2009 Version: November 27, 2009 Instructors: Peter Forsyth, paforsyt@uwaterloo.ca Office Hours: Tues: 4:00-5:00; Thurs: 11:00-12:00 Lectures:MWF 3:30-4:20 MC2036 Office: DC3631 CS
More informationIntensity-based framework for optimal stopping
Intensity-based framework for optimal stopping problems Min Dai National University of Singapore Yue Kuen Kwok Hong Kong University of Science and Technology Hong You National University of Singapore Abstract
More informationHedging with a Correlated Asset: An Insurance Approach
Hedging with a Correlated Asset: An Insurance Approach by Jian Wang A thesis presented to the University of Waterloo in fulfillment of the thesis requirement for the degree of Master of Mathematics in
More informationChapter 5 Finite Difference Methods. Math6911 W07, HM Zhu
Chapter 5 Finite Difference Methods Math69 W07, HM Zhu References. Chapters 5 and 9, Brandimarte. Section 7.8, Hull 3. Chapter 7, Numerical analysis, Burden and Faires Outline Finite difference (FD) approximation
More informationAdvanced Numerical Methods
Advanced Numerical Methods Solution to Homework One Course instructor: Prof. Y.K. Kwok. When the asset pays continuous dividend yield at the rate q the expected rate of return of the asset is r q under
More informationPDE Methods for the Maximum Drawdown
PDE Methods for the Maximum Drawdown Libor Pospisil, Jan Vecer Columbia University, Department of Statistics, New York, NY 127, USA April 1, 28 Abstract Maximum drawdown is a risk measure that plays an
More informationCONVERGENCE OF NUMERICAL METHODS FOR VALUING PATH-DEPENDENT OPTIONS USING INTERPOLATION
CONVERGENCE OF NUMERICAL METHODS FOR VALUING PATH-DEPENDENT OPTIONS USING INTERPOLATION P.A. Forsyth Department of Computer Science University of Waterloo Waterloo, ON Canada N2L 3G1 E-mail: paforsyt@elora.math.uwaterloo.ca
More informationarxiv: v1 [q-fin.cp] 1 Nov 2016
Essentially high-order compact schemes with application to stochastic volatility models on non-uniform grids arxiv:1611.00316v1 [q-fin.cp] 1 Nov 016 Bertram Düring Christof Heuer November, 016 Abstract
More informationHedging with a Correlated Asset: Solution of a Nonlinear Pricing PDE
Hedging with a Correlated Asset: Solution of a Nonlinear Pricing PDE H. Windcliff, J. Wang, P.A. Forsyth, K.R. Vetzal June 14, 2005 Abstract Hedging a contingent claim with an asset which is not perfectly
More informationFinite Difference Approximation of Hedging Quantities in the Heston model
Finite Difference Approximation of Hedging Quantities in the Heston model Karel in t Hout Department of Mathematics and Computer cience, University of Antwerp, Middelheimlaan, 22 Antwerp, Belgium Abstract.
More informationDefinition Pricing Risk management Second generation barrier options. Barrier Options. Arfima Financial Solutions
Arfima Financial Solutions Contents Definition 1 Definition 2 3 4 Contenido Definition 1 Definition 2 3 4 Definition Definition: A barrier option is an option on the underlying asset that is activated
More information1.1 Basic Financial Derivatives: Forward Contracts and Options
Chapter 1 Preliminaries 1.1 Basic Financial Derivatives: Forward Contracts and Options A derivative is a financial instrument whose value depends on the values of other, more basic underlying variables
More informationNUMERICAL METHODS OF PARTIAL INTEGRO-DIFFERENTIAL EQUATIONS FOR OPTION PRICE
Trends in Mathematics - New Series Information Center for Mathematical Sciences Volume 13, Number 1, 011, pages 1 5 NUMERICAL METHODS OF PARTIAL INTEGRO-DIFFERENTIAL EQUATIONS FOR OPTION PRICE YONGHOON
More informationHedging Under Jump Diffusions with Transaction Costs. Peter Forsyth, Shannon Kennedy, Ken Vetzal University of Waterloo
Hedging Under Jump Diffusions with Transaction Costs Peter Forsyth, Shannon Kennedy, Ken Vetzal University of Waterloo Computational Finance Workshop, Shanghai, July 4, 2008 Overview Overview Single factor
More information2.3 Mathematical Finance: Option pricing
CHAPTR 2. CONTINUUM MODL 8 2.3 Mathematical Finance: Option pricing Options are some of the commonest examples of derivative securities (also termed financial derivatives or simply derivatives). A uropean
More informationIEOR E4703: Monte-Carlo Simulation
IEOR E4703: Monte-Carlo Simulation Simulating Stochastic Differential Equations Martin Haugh Department of Industrial Engineering and Operations Research Columbia University Email: martin.b.haugh@gmail.com
More informationCalibration Lecture 4: LSV and Model Uncertainty
Calibration Lecture 4: LSV and Model Uncertainty March 2017 Recap: Heston model Recall the Heston stochastic volatility model ds t = rs t dt + Y t S t dw 1 t, dy t = κ(θ Y t ) dt + ξ Y t dw 2 t, where
More informationSTOCHASTIC CALCULUS AND BLACK-SCHOLES MODEL
STOCHASTIC CALCULUS AND BLACK-SCHOLES MODEL YOUNGGEUN YOO Abstract. Ito s lemma is often used in Ito calculus to find the differentials of a stochastic process that depends on time. This paper will introduce
More informationPricing American Options Using a Space-time Adaptive Finite Difference Method
Pricing American Options Using a Space-time Adaptive Finite Difference Method Jonas Persson Abstract American options are priced numerically using a space- and timeadaptive finite difference method. The
More informationPractical Hedging: From Theory to Practice. OSU Financial Mathematics Seminar May 5, 2008
Practical Hedging: From Theory to Practice OSU Financial Mathematics Seminar May 5, 008 Background Dynamic replication is a risk management technique used to mitigate market risk We hope to spend a certain
More informationAdvanced Topics in Derivative Pricing Models. Topic 4 - Variance products and volatility derivatives
Advanced Topics in Derivative Pricing Models Topic 4 - Variance products and volatility derivatives 4.1 Volatility trading and replication of variance swaps 4.2 Volatility swaps 4.3 Pricing of discrete
More informationMath 416/516: Stochastic Simulation
Math 416/516: Stochastic Simulation Haijun Li lih@math.wsu.edu Department of Mathematics Washington State University Week 13 Haijun Li Math 416/516: Stochastic Simulation Week 13 1 / 28 Outline 1 Simulation
More informationFINITE DIFFERENCE METHODS
FINITE DIFFERENCE METHODS School of Mathematics 2013 OUTLINE Review 1 REVIEW Last time Today s Lecture OUTLINE Review 1 REVIEW Last time Today s Lecture 2 DISCRETISING THE PROBLEM Finite-difference approximations
More informationValuation of performance-dependent options in a Black- Scholes framework
Valuation of performance-dependent options in a Black- Scholes framework Thomas Gerstner, Markus Holtz Institut für Numerische Simulation, Universität Bonn, Germany Ralf Korn Fachbereich Mathematik, TU
More informationPackage multiassetoptions
Package multiassetoptions February 20, 2015 Type Package Title Finite Difference Method for Multi-Asset Option Valuation Version 0.1-1 Date 2015-01-31 Author Maintainer Michael Eichenberger
More informationWrite legibly. Unreadable answers are worthless.
MMF 2021 Final Exam 1 December 2016. This is a closed-book exam: no books, no notes, no calculators, no phones, no tablets, no computers (of any kind) allowed. Do NOT turn this page over until you are
More information2.1 Mathematical Basis: Risk-Neutral Pricing
Chapter Monte-Carlo Simulation.1 Mathematical Basis: Risk-Neutral Pricing Suppose that F T is the payoff at T for a European-type derivative f. Then the price at times t before T is given by f t = e r(t
More informationBluff Your Way Through Black-Scholes
Bluff our Way Through Black-Scholes Saurav Sen December 000 Contents What is Black-Scholes?.............................. 1 The Classical Black-Scholes Model....................... 1 Some Useful Background
More informationTEACHING NOTE 98-04: EXCHANGE OPTION PRICING
TEACHING NOTE 98-04: EXCHANGE OPTION PRICING Version date: June 3, 017 C:\CLASSES\TEACHING NOTES\TN98-04.WPD The exchange option, first developed by Margrabe (1978), has proven to be an extremely powerful
More informationEFFICIENT MONTE CARLO ALGORITHM FOR PRICING BARRIER OPTIONS
Commun. Korean Math. Soc. 23 (2008), No. 2, pp. 285 294 EFFICIENT MONTE CARLO ALGORITHM FOR PRICING BARRIER OPTIONS Kyoung-Sook Moon Reprinted from the Communications of the Korean Mathematical Society
More informationAmerican Options; an American delayed- Exercise model and the free boundary. Business Analytics Paper. Nadra Abdalla
American Options; an American delayed- Exercise model and the free boundary Business Analytics Paper Nadra Abdalla [Geef tekst op] Pagina 1 Business Analytics Paper VU University Amsterdam Faculty of Sciences
More informationMASM006 UNIVERSITY OF EXETER SCHOOL OF ENGINEERING, COMPUTER SCIENCE AND MATHEMATICS MATHEMATICAL SCIENCES FINANCIAL MATHEMATICS.
MASM006 UNIVERSITY OF EXETER SCHOOL OF ENGINEERING, COMPUTER SCIENCE AND MATHEMATICS MATHEMATICAL SCIENCES FINANCIAL MATHEMATICS May/June 2006 Time allowed: 2 HOURS. Examiner: Dr N.P. Byott This is a CLOSED
More information( ) since this is the benefit of buying the asset at the strike price rather
Review of some financial models for MAT 483 Parity and Other Option Relationships The basic parity relationship for European options with the same strike price and the same time to expiration is: C( KT
More informationTHE BLACK-SCHOLES FORMULA AND THE GREEK PARAMETERS FOR A NONLINEAR BLACK-SCHOLES EQUATION
International Journal of Pure and Applied Mathematics Volume 76 No. 2 2012, 167-171 ISSN: 1311-8080 printed version) url: http://www.ijpam.eu PA ijpam.eu THE BLACK-SCHOLES FORMULA AND THE GREEK PARAMETERS
More informationThe Pennsylvania State University. The Graduate School. Department of Industrial Engineering AMERICAN-ASIAN OPTION PRICING BASED ON MONTE CARLO
The Pennsylvania State University The Graduate School Department of Industrial Engineering AMERICAN-ASIAN OPTION PRICING BASED ON MONTE CARLO SIMULATION METHOD A Thesis in Industrial Engineering and Operations
More informationHedging Credit Derivatives in Intensity Based Models
Hedging Credit Derivatives in Intensity Based Models PETER CARR Head of Quantitative Financial Research, Bloomberg LP, New York Director of the Masters Program in Math Finance, Courant Institute, NYU Stanford
More informationMonte Carlo Based Numerical Pricing of Multiple Strike-Reset Options
Monte Carlo Based Numerical Pricing of Multiple Strike-Reset Options Stavros Christodoulou Linacre College University of Oxford MSc Thesis Trinity 2011 Contents List of figures ii Introduction 2 1 Strike
More informationAn Adjusted Trinomial Lattice for Pricing Arithmetic Average Based Asian Option
American Journal of Applied Mathematics 2018; 6(2): 28-33 http://www.sciencepublishinggroup.com/j/ajam doi: 10.11648/j.ajam.20180602.11 ISSN: 2330-0043 (Print); ISSN: 2330-006X (Online) An Adjusted Trinomial
More informationAdvanced Numerical Techniques for Financial Engineering
Advanced Numerical Techniques for Financial Engineering Andreas Binder, Heinz W. Engl, Andrea Schatz Abstract We present some aspects of advanced numerical analysis for the pricing and risk managment of
More informationKing s College London
King s College London University Of London This paper is part of an examination of the College counting towards the award of a degree. Examinations are governed by the College Regulations under the authority
More informationComputational Finance Finite Difference Methods
Explicit finite difference method Computational Finance Finite Difference Methods School of Mathematics 2018 Today s Lecture We now introduce the final numerical scheme which is related to the PDE solution.
More informationLecture 8: The Black-Scholes theory
Lecture 8: The Black-Scholes theory Dr. Roman V Belavkin MSO4112 Contents 1 Geometric Brownian motion 1 2 The Black-Scholes pricing 2 3 The Black-Scholes equation 3 References 5 1 Geometric Brownian motion
More informationMulti-period mean variance asset allocation: Is it bad to win the lottery?
Multi-period mean variance asset allocation: Is it bad to win the lottery? Peter Forsyth 1 D.M. Dang 1 1 Cheriton School of Computer Science University of Waterloo Guangzhou, July 28, 2014 1 / 29 The Basic
More informationAmerican Equity Option Valuation Practical Guide
Valuation Practical Guide John Smith FinPricing Summary American Equity Option Introduction The Use of American Equity Options Valuation Practical Guide A Real World Example American Option Introduction
More informationPAijpam.eu ANALYTIC SOLUTION OF A NONLINEAR BLACK-SCHOLES EQUATION
International Journal of Pure and Applied Mathematics Volume 8 No. 4 013, 547-555 ISSN: 1311-8080 (printed version); ISSN: 1314-3395 (on-line version) url: http://www.ijpam.eu doi: http://dx.doi.org/10.173/ijpam.v8i4.4
More informationPricing Implied Volatility
Pricing Implied Volatility Expected future volatility plays a central role in finance theory. Consequently, accurate estimation of this parameter is crucial to meaningful financial decision-making. Researchers
More informationAn Introduction to Computational Finance
An Introduction to Computational Finance P.A. Forsyth June 17, 2003 Contents 1 The First Option Trade 2 2 The Black-Scholes Equation 2 2.1 Background.................................... 2 2.2 Definitions.....................................
More informationNumerical Methods in Option Pricing (Part III)
Numerical Methods in Option Pricing (Part III) E. Explicit Finite Differences. Use of the Forward, Central, and Symmetric Central a. In order to obtain an explicit solution for the price of the derivative,
More informationLecture 4. Finite difference and finite element methods
Finite difference and finite element methods Lecture 4 Outline Black-Scholes equation From expectation to PDE Goal: compute the value of European option with payoff g which is the conditional expectation
More informationUtility Indifference Pricing and Dynamic Programming Algorithm
Chapter 8 Utility Indifference ricing and Dynamic rogramming Algorithm In the Black-Scholes framework, we can perfectly replicate an option s payoff. However, it may not be true beyond the Black-Scholes
More information4. Black-Scholes Models and PDEs. Math6911 S08, HM Zhu
4. Black-Scholes Models and PDEs Math6911 S08, HM Zhu References 1. Chapter 13, J. Hull. Section.6, P. Brandimarte Outline Derivation of Black-Scholes equation Black-Scholes models for options Implied
More informationStochastic Volatility (Working Draft I)
Stochastic Volatility (Working Draft I) Paul J. Atzberger General comments or corrections should be sent to: paulatz@cims.nyu.edu 1 Introduction When using the Black-Scholes-Merton model to price derivative
More informationAmerican options and early exercise
Chapter 3 American options and early exercise American options are contracts that may be exercised early, prior to expiry. These options are contrasted with European options for which exercise is only
More informationBlack-Scholes-Merton Model
Black-Scholes-Merton Model Weerachart Kilenthong University of the Thai Chamber of Commerce c Kilenthong 2017 Weerachart Kilenthong University of the Thai Chamber Black-Scholes-Merton of Commerce Model
More informationA new PDE approach for pricing arithmetic average Asian options
A new PDE approach for pricing arithmetic average Asian options Jan Večeř Department of Mathematical Sciences, Carnegie Mellon University, Pittsburgh, PA 15213. Email: vecer@andrew.cmu.edu. May 15, 21
More information1. Introduction. Options having a payo which depends on the average of the underlying asset are termed Asian options. Typically the average is arithme
DISCRETE ASIAN BARRIER OPTIONS R. ZVAN, P.A. FORSYTH y AND K.R. VETZAL z Abstract. A partial dierential equation method based on using auxiliary variables is described for pricing discretely monitored
More informationPricing Barrier Options under Local Volatility
Abstract Pricing Barrier Options under Local Volatility Artur Sepp Mail: artursepp@hotmail.com, Web: www.hot.ee/seppar 16 November 2002 We study pricing under the local volatility. Our research is mainly
More informationLattice (Binomial Trees) Version 1.2
Lattice (Binomial Trees) Version 1. 1 Introduction This plug-in implements different binomial trees approximations for pricing contingent claims and allows Fairmat to use some of the most popular binomial
More informationValuing the Guaranteed Minimum Death Benefit Clause with Partial Withdrawals
1 2 3 4 Valuing the Guaranteed Minimum Death Benefit Clause with Partial Withdrawals A. C. Bélanger, P. A. Forsyth and G. Labahn January 30, 2009 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Abstract In
More informationLecture Note 8 of Bus 41202, Spring 2017: Stochastic Diffusion Equation & Option Pricing
Lecture Note 8 of Bus 41202, Spring 2017: Stochastic Diffusion Equation & Option Pricing We shall go over this note quickly due to time constraints. Key concept: Ito s lemma Stock Options: A contract giving
More informationHedging with Life and General Insurance Products
Hedging with Life and General Insurance Products June 2016 2 Hedging with Life and General Insurance Products Jungmin Choi Department of Mathematics East Carolina University Abstract In this study, a hybrid
More informationAn IMEX-method for pricing options under Bates model using adaptive finite differences Rapport i Teknisk-vetenskapliga datorberäkningar
PROJEKTRAPPORT An IMEX-method for pricing options under Bates model using adaptive finite differences Arvid Westlund Rapport i Teknisk-vetenskapliga datorberäkningar Jan 2014 INSTITUTIONEN FÖR INFORMATIONSTEKNOLOGI
More informationLecture Quantitative Finance Spring Term 2015
and Lecture Quantitative Finance Spring Term 2015 Prof. Dr. Erich Walter Farkas Lecture 06: March 26, 2015 1 / 47 Remember and Previous chapters: introduction to the theory of options put-call parity fundamentals
More informationNumerical valuation for option pricing under jump-diffusion models by finite differences
Numerical valuation for option pricing under jump-diffusion models by finite differences YongHoon Kwon Younhee Lee Department of Mathematics Pohang University of Science and Technology June 23, 2010 Table
More informationEconomathematics. Problem Sheet 1. Zbigniew Palmowski. Ws 2 dw s = 1 t
Economathematics Problem Sheet 1 Zbigniew Palmowski 1. Calculate Ee X where X is a gaussian random variable with mean µ and volatility σ >.. Verify that where W is a Wiener process. Ws dw s = 1 3 W t 3
More informationAMH4 - ADVANCED OPTION PRICING. Contents
AMH4 - ADVANCED OPTION PRICING ANDREW TULLOCH Contents 1. Theory of Option Pricing 2 2. Black-Scholes PDE Method 4 3. Martingale method 4 4. Monte Carlo methods 5 4.1. Method of antithetic variances 5
More informationAn improvement of the douglas scheme for the Black-Scholes equation
Kuwait J. Sci. 42 (3) pp. 105-119, 2015 An improvement of the douglas scheme for the Black-Scholes equation FARES AL-AZEMI Department of Mathematics, Kuwait University, Safat, 13060, Kuwait. fares@sci.kuniv.edu.kw
More informationFourier Space Time-stepping Method for Option Pricing with Lévy Processes
FST method Extensions Indifference pricing Fourier Space Time-stepping Method for Option Pricing with Lévy Processes Vladimir Surkov University of Toronto Computational Methods in Finance Conference University
More informationSimulating Stochastic Differential Equations
IEOR E4603: Monte-Carlo Simulation c 2017 by Martin Haugh Columbia University Simulating Stochastic Differential Equations In these lecture notes we discuss the simulation of stochastic differential equations
More informationThe Black-Scholes Equation
The Black-Scholes Equation MATH 472 Financial Mathematics J. Robert Buchanan 2018 Objectives In this lesson we will: derive the Black-Scholes partial differential equation using Itô s Lemma and no-arbitrage
More informationNumerical Methods for Optimal Stochastic Control in Finance
Numerical Methods for Optimal Stochastic Control in Finance by Zhuliang Chen A thesis presented to the University of Waterloo in fulfillment of the thesis requirement for the degree of Doctor of Philosophy
More informationPricing theory of financial derivatives
Pricing theory of financial derivatives One-period securities model S denotes the price process {S(t) : t = 0, 1}, where S(t) = (S 1 (t) S 2 (t) S M (t)). Here, M is the number of securities. At t = 1,
More informationFrom Discrete Time to Continuous Time Modeling
From Discrete Time to Continuous Time Modeling Prof. S. Jaimungal, Department of Statistics, University of Toronto 2004 Arrow-Debreu Securities 2004 Prof. S. Jaimungal 2 Consider a simple one-period economy
More informationlast problem outlines how the Black Scholes PDE (and its derivation) may be modified to account for the payment of stock dividends.
224 10 Arbitrage and SDEs last problem outlines how the Black Scholes PDE (and its derivation) may be modified to account for the payment of stock dividends. 10.1 (Calculation of Delta First and Finest
More informationRichardson Extrapolation Techniques for the Pricing of American-style Options
Richardson Extrapolation Techniques for the Pricing of American-style Options June 1, 2005 Abstract Richardson Extrapolation Techniques for the Pricing of American-style Options In this paper we re-examine
More informationChapter 3: Black-Scholes Equation and Its Numerical Evaluation
Chapter 3: Black-Scholes Equation and Its Numerical Evaluation 3.1 Itô Integral 3.1.1 Convergence in the Mean and Stieltjes Integral Definition 3.1 (Convergence in the Mean) A sequence {X n } n ln of random
More informationA No-Arbitrage Theorem for Uncertain Stock Model
Fuzzy Optim Decis Making manuscript No (will be inserted by the editor) A No-Arbitrage Theorem for Uncertain Stock Model Kai Yao Received: date / Accepted: date Abstract Stock model is used to describe
More informationNumerical Solution of Two Asset Jump Diffusion Models for Option Valuation
Numerical Solution of Two Asset Jump Diffusion Models for Option Valuation Simon S. Clift and Peter A. Forsyth Original: December 5, 2005 Revised: January 31, 2007 Abstract Under the assumption that two
More informationThe Yield Envelope: Price Ranges for Fixed Income Products
The Yield Envelope: Price Ranges for Fixed Income Products by David Epstein (LINK:www.maths.ox.ac.uk/users/epstein) Mathematical Institute (LINK:www.maths.ox.ac.uk) Oxford Paul Wilmott (LINK:www.oxfordfinancial.co.uk/pw)
More informationCRANK-NICOLSON SCHEME FOR ASIAN OPTION
CRANK-NICOLSON SCHEME FOR ASIAN OPTION By LEE TSE YUENG A thesis submitted to the Department of Mathematical and Actuarial Sciences, Faculty of Engineering and Science, Universiti Tunku Abdul Rahman, in
More informationBinomial model: numerical algorithm
Binomial model: numerical algorithm S / 0 C \ 0 S0 u / C \ 1,1 S0 d / S u 0 /, S u 3 0 / 3,3 C \ S0 u d /,1 S u 5 0 4 0 / C 5 5,5 max X S0 u,0 S u C \ 4 4,4 C \ 3 S u d / 0 3, C \ S u d 0 S u d 0 / C 4
More informationExercises for Mathematical Models of Financial Derivatives
Exercises for Mathematical Models of Financial Derivatives January 24, 2 1. It is customary for shares in the UK to have prices between 1p and 1,p (in the US, between $1 and $1), perhaps because then typical
More informationADAPTIVE PARTIAL DIFFERENTIAL EQUATION METHODS FOR OPTION PRICING
ADAPTIVE PARTIAL DIFFERENTIAL EQUATION METHODS FOR OPTION PRICING by Guanghuan Hou B.Sc., Zhejiang University, 2004 a project submitted in partial fulfillment of the requirements for the degree of Master
More informationContinuous Time Mean Variance Asset Allocation: A Time-consistent Strategy
Continuous Time Mean Variance Asset Allocation: A Time-consistent Strategy J. Wang, P.A. Forsyth October 24, 2009 Abstract We develop a numerical scheme for determining the optimal asset allocation strategy
More information13.3 A Stochastic Production Planning Model
13.3. A Stochastic Production Planning Model 347 From (13.9), we can formally write (dx t ) = f (dt) + G (dz t ) + fgdz t dt, (13.3) dx t dt = f(dt) + Gdz t dt. (13.33) The exact meaning of these expressions
More informationExtensions to the Black Scholes Model
Lecture 16 Extensions to the Black Scholes Model 16.1 Dividends Dividend is a sum of money paid regularly (typically annually) by a company to its shareholders out of its profits (or reserves). In this
More informationCS476/676 Mar 6, Today s Topics. American Option: early exercise curve. PDE overview. Discretizations. Finite difference approximations
CS476/676 Mar 6, 2019 1 Today s Topics American Option: early exercise curve PDE overview Discretizations Finite difference approximations CS476/676 Mar 6, 2019 2 American Option American Option: PDE Complementarity
More informationAN OPERATOR SPLITTING METHOD FOR PRICING THE ELS OPTION
J. KSIAM Vol.14, No.3, 175 187, 21 AN OPERATOR SPLITTING METHOD FOR PRICING THE ELS OPTION DARAE JEONG, IN-SUK WEE, AND JUNSEOK KIM DEPARTMENT OF MATHEMATICS, KOREA UNIVERSITY, SEOUL 136-71, KOREA E-mail
More informationAs an example, we consider the following PDE with one variable; Finite difference method is one of numerical method for the PDE.
7. Introduction to the numerical integration of PDE. As an example, we consider the following PDE with one variable; Finite difference method is one of numerical method for the PDE. Accuracy requirements
More informationLecture 11: Ito Calculus. Tuesday, October 23, 12
Lecture 11: Ito Calculus Continuous time models We start with the model from Chapter 3 log S j log S j 1 = µ t + p tz j Sum it over j: log S N log S 0 = NX µ t + NX p tzj j=1 j=1 Can we take the limit
More informationNumerical Methods and Volatility Models for Valuing Cliquet Options
Numerical Methods and Volatility Models for Valuing Cliquet Options H.A. Windcliff, P.A. Forsyth, and K.R. Vetzal Revised: February 14, 2006 First Version: September 13, 2004 Abstract Several numerical
More information