Black-Scholes Model - Solutions Throughout this exercise you may use assume (where appropriate) the following results without proof d1 = d2 = N (x) = log (S=E) + r D + p T t log (S=E) + r D p T t Z x 2 1 =2 p e d 2 1 1 2 2 1 2 2 (T t) (T t) , and where S 0 is the spot price, t T is the time, E > 0 is the strike, T > 0 the expiry date, r 0 the interest rate, D is the dividend yield and is the volatility of S. 1. The Black–Scholes formula for a European call option C (S; t) is given by C (S; t) = S exp( D(T t))N (d1 ) E exp( r(T t))N (d2 ): a) By di¤erentiating with respect to S and show that the delta and vega are given by r d12 T t ( D(T t)) 2 = e( D(T t)) N (d1 ) ; and = Se e : 2 Note: @d1 @d2 = @S @S and @d2 @d1 = @ @ p T t So = @C @S ( D(T t)) ( D(T t)) = e = e = e( D(T t)) ( D(T N (d1 ) + Se t)) 1 p e 2 1 @d1 N (d1 ) + p Se( 2 @S | D(T d12 2 t)) e @d1 @S E exp( r(T d12 2 Ee( {z r(T t)) =0 N (d1 ) because the term in the bracket above is zero. 1 e 1 t)) p e 2 ! d 2 2 2 } d22 2 @d2 @S = @C @ = Se( D(T t)) = Se( D(T t)) = = r r T t 2 T t 2 Se( d12 2 1 p e 2 d12 2 1 p e 2 D(T ( D(T Se @d1 @ t)) t)) e e Ee( r(T @d2 p + T @ d12 2 + d12 2 t)) d22 2 1 p e 2 t 2 1 p Ee( 2 @d2 1 4 ( D(T p Se | @ 2 r T t ( r(T = Ee 2 t)) @d2 @ r(T e t)) d22 2 e d12 2 {z Ee( =0 t)) e d22 2 ! 2. Given that S is de…ned by the SDE dS = a (S; t) dt + b (S; t) dW (2.1) where a and b are given functions of S and t; show using Itô’s lemma that any function V (S; t) satis…es the SDE dV = @V 1 @2V @V +a + b2 2 @t @S 2 @S dt + b @V dW @S where we have assumed that all partial derivatives exist. Hence derive the partial di¤erential equation @V 1 @2V + b2 2 = r V @t 2 @S S @V @S (2.2) for the fair price of an option based on a security S which satis…es (2:1) with r the risk-free interest rate. Show (by substitution) that V (S; t) = e t S 2 is a solution of (2:2) provided b2 = ( r) S 2 and is a constant. The …rst part of this problem is trivial. Follow the derivation of the BSE as done in the notes. The only di¤erence here is that a (S; t) and b (S; t) replace S and S: For the second part simply substitute V (S; t) = e t S 2 in (2:2) ; the following terms are needed @V = @t e t e 1 S 2 + b2 2 t S2; 2e @V = 2e @S t S 2 + b2 2 t = r e = @2V = 2e @S 2 S; t S2 2e rS 2 ! b2 = ( t t S2 r) S 2 : @d2 @ r(T t)) e d22 2 } 3 5 3. The Black–Scholes formula for a European call option C (S; t) is C (S; t) = S exp( D(T t))N (d1 ) E exp( r(T t))N (d2 ) From this expression, …nd the Black–Scholes value of the call option in the following limits: a. (time tends to expiry) t ! T , > 0 (this depends on S=E); exp ( r(T t)), exp( D(T t)) ! 1 We know d12 d12 log (S=E) + r D 21 2 (T t) p T t r D 12 2 (T t) log (S=E) p p = + T t T t 1 2 (T t) log (S=E) p p + r D = 2 T t T t 1 2 p p r D log (S=E) log (S=E) 2 p + +O = T t= p T t T t T t 8 8 < S E S>E < 1 S>E p log (S=E) 0 S=E 0 S=E ! p +O so C ! T t ! : : T t 0 S<E 1 S<E = b. (volatility tends to zero) ! 0+ , t < T ; (this depends on S exp( D(T t))=E exp( r(T t))) Again start with what we know, i.e. d12 = = = log (S=E) + r D 21 2 (T p T t log (S=E) (r D) (T t) p p + T t T t log (S=E) (r D) (T t) p p + T t T t t) 1 2 (T t) p 2 T t p 1 T t 2 log (S=E) + (r D) (T t) p +O( ) T t Now we employ a small trick in the …rst quotient d12 ! t))=E exp( r(T t))) p +O( ) T t 8 Se( D(T t)) > Ee( r(T t)) < 1 ! 0 Se( D(T t)) = Ee( r(T t)) : 1 Se( D(T t)) < Ee( r(T t)) h i ! max Se( D(T t)) Ee( r(T t)) ; 0 = so C log (S exp( D(T 3 4. Suppose S evolves according to the SDE dS = Sdt + S dW where and are positive constants. Given that the interest rate is zero, derive the corresponding Black–Scholes partial di¤erential equation (PDE) for the option based upon this asset S (you are not required to solve any equation). Write this PDE in terms of the greeks. We know that if S evolves according to the stochastic di¤erential equation (SDE) dS = a (S; t) dt + b (S; t) dW and V = V (S; t) then Itô gives @V 1 @2V @V + S + S2 @t @S 2 @S 2 dV = Then set up a portfolio d = dV dS: So d = =V dt + S @V dW @S S ) in one time-step (we hold @V @V 1 @2V + S + S2 @t @S 2 @S 2 @V dW @S dt + S …xed) ( Sdt + S dW ) therefore we take = @V @S to eliminate the risk associated with the portfolio (to cancel out terms with dX), which gives d = 1 @V @2V + S2 @t 2 @S 2 dt portfolio now riskless. No arbitrage tells us that we are guaranteed return at risk free rate, so 1 @V @2V + S2 @t 2 @S 2 and r = 0 so Using greeks dt = r dt @V 1 @2V + S2 = 0. @t 2 @S 2 @V @t = and @2V @S 2 = allows us to write this pde as 1 + S2 2 = 0. 5. The value of an option V (S; t) satis…es the Black–Scholes equation. Write the option value in the form V (S; t) = exp( r(T 4 t))q(S; t): (5.1) Show that the function q(S; t) satis…es the equation @q 1 + @t 2 2 S2 @2q + (r @S 2 D)S @q = 0: @S This is the backward Kolmogorov equation, used for calculating the expected value of stochastic quantities. Substitute @V @t @V @S @2V @S 2 = exp( r(T = exp( r(T = exp( r(T @ q(S; t) + rV (S; t); @t @q t)) & @S @2q t)) 2 @S t)) from (5:1) into the BSE, all the exponentials cancel out and the above equation is left. Thus the value of an option can be expressed in the form V (S; t) = exp ( r(T t)) E [Payo¤(S)] where E[x] means the expected value of x . This is not a real expectation, but taken under the risk-neutral random walk (so r replaces ) and forms the basis of Monte Carlo methods applied to …nance. More on this later. 5