1. Trang chủ
  2. » Giáo Dục - Đào Tạo

Recognizing a BrownianMotion

6 123 0
Tài liệu đã được kiểm tra trùng lặp

Đang tải... (xem toàn văn)

THÔNG TIN TÀI LIỆU

Thông tin cơ bản

Định dạng
Số trang 6
Dung lượng 117,4 KB

Nội dung

Chapter 23 Recognizing a Brownian Motion Theorem 0.62 (Levy) Let B t; 0  t  T; be a process on ; F ; P , adapted to a filtration F t; 0  t  T , such that: 1. the paths of B t are continuous, 2. B is a martingale, 3. hB it=t; 0  t  T , (i.e., informally dB t dB t=dt ). Then B is a Brownian motion. Proof: (Idea) Let 0  stT be given. We need to show that B t , B s is normal, with mean zero and variance t , s ,and B t , B s is independent of F s . We shall show that the conditional moment generating function of B t , B s is IE  e uB t,B s     F s  = e 1 2 u 2 t,s : Since the moment generating function characterizes the distribution, this shows that B t , B s is normal with mean 0 and variance t , s , and conditioning on F s does not affect this, i.e., B t , B s is independent of F s . We compute (this uses the continuity condition (1) of the theorem) de uB t = ue uB t dB t+ 1 2 u 2 e uB t dB t dB t; so e uB t = e uB s + Z t s ue uB v dB v + 1 2 u 2 Z t s e uB v dv : |z uses cond. 3 233 234 Now R t 0 ue uB v dB v  is a martingale (by condition 2), and so IE  Z t s ue uB v dB v      F s  = , Z s 0 ue uB v dB v + IE  Z t 0 ue uB v dB v      F s  =0: It follows that IE  e uB t     F s  = e uB s + 1 2 u 2 Z t s IE  e uB v     F s  dv : We define 'v =IE  e uB v     F s  ; so that 's=e uB s and 't=e uB s + 1 2 u 2 Z t s 'v  dv ; ' 0 t= 1 2 u 2 't; 't=ke 1 2 u 2 t : Plugging in s ,weget e uB s = ke 1 2 u 2 s =k = e uB s, 1 2 u 2 s : Therefore, IE  e uB t     F s  = 't=e uB s+ 1 2 u 2 t,s ; IE  e uB t,B s     F s  = e 1 2 u 2 t,s : CHAPTER 23. Recognizing a Brownian Motion 235 23.1 Identifying volatility and correlation Let B 1 and B 2 be independent Brownian motions and dS 1 S 1 = rdt+ 11 dB 1 +  12 dB 2 ; dS 2 S 2 = rdt+ 21 dB 1 +  22 dB 2 ; Define  1 = q  2 11 +  2 12 ;  2 = q  2 21 +  2 22 ;  =  11  21 +  12  22  1  2 : Define processes W 1 and W 2 by dW 1 =  11 dB 1 +  12 dB 2  1 dW 2 =  21 dB 1 +  22 dB 2  2 : Then W 1 and W 2 have continuous paths, are martingales, and dW 1 dW 1 = 1  2 1  11 dB 1 +  12 dB 2  2 = 1  2 1  2 11 dB 1 dB 1 +  2 12 dB 2 dB 2  = dt; and similarly dW 2 dW 2 = dt: Therefore, W 1 and W 2 are Brownian motions. The stock prices have the representation dS 1 S 1 = rdt+ 1 dW 1 ; dS 2 S 2 = rdt+ 2 dW 2 : The Brownian motions W 1 and W 2 are correlated. Indeed, dW 1 dW 2 = 1  1  2  11 dB 1 +  12 dB 2  21 dB 1 +  22 dB 2  = 1  1  2  11  21 +  12  22  dt =  dt: 236 23.2 Reversing the process Suppose we are given that dS 1 S 1 = rdt+ 1 dW 1 ; dS 2 S 2 = rdt+ 2 dW 2 ; where W 1 and W 2 are Brownian motions with correlation coefficient  .Wewanttofind = "  11  12  21  22  so that  0 = "  11  12  21  22 "  11  21  12  22  = "  2 11 +  2 12  11  21 +  12  22  11  21 +  12  22  2 21 +  2 22  = "  2 1  1  2  1  2  2 2  A simple (but not unique) solution is (see Chapter 19)  11 =  1 ;  12 =0;  21 =  2 ;  22 = q 1 ,  2  2 : This corresponds to  1 dW 1 =  1 dB 1 =dB 1 = dW 1 ;  2 dW 2 =  2 dB 1 + q 1 ,  2  2 dB 2 = dB 2 = dW 2 , dW 1 p 1, 2 ; 6=1 If  = 1 , then there is no B 2 and dW 2 = dB 1 =dW 1 : Continuing in the case  6= 1 ,wehave dB 1 dB 1 = dW 1 dW 1 = dt; dB 2 dB 2 = 1 1 ,  2  dW 2 dW 2 , 2dW 1 dW 2 +  2 dW 2 dW 2  = 1 1 ,  2  dt , 2 2 dt +  2 dt  = dt; CHAPTER 23. Recognizing a Brownian Motion 237 so both B 1 and B 2 are Brownian motions. Furthermore, dB 1 dB 2 = 1 p 1 ,  2 dW 1 dW 2 , dW 1 dW 1  = 1 p 1 ,  2 dt,dt=0: We can now apply an Extension of Levy’s Theorem that says that Brownian motions with zero cross-variation are independent, to conclude that B 1 ;B 2 are independent Brownians. 238 . We can now apply an Extension of Levy’s Theorem that says that Brownian motions with zero cross-variation are independent, to conclude that B 1 ;B 2 are. the moment generating function characterizes the distribution, this shows that B t , B s is normal with mean 0 and variance t , s , and conditioning

Ngày đăng: 18/10/2013, 02:20

TÀI LIỆU CÙNG NGƯỜI DÙNG

  • Đang cập nhật ...

TÀI LIỆU LIÊN QUAN

w