1、From book page 385, ffBy Q(t)=M(t)Q(t)D(t)We can calculate dQ(t)=Q(t){[R(t)−Rf(t)]dt+σ(t)dW(t)}23(9.3.16)We want to know if Usingthis representation: Q(t)=M(t)Q(t)D(t), can we get the same result?dQ(t)=d(M(t)Q(t)D(t))=Q(t)D(t)dM(t)+M(t)d(Q(t)D(t))dM(t)=R(t)
2、M(t)dtsolved(Q(t)D(t))d(Q(t)D(t))=dQ(t)D(t)+dD(t)Q(t)=dQ(t)D(t)−R(t)D(t)Q(t)dtSod(M(t)Q(t)D(t))=dQ(t)No matter which way we use, it’ll always go back todQ(t).And we’ve assumed dQ(t)has the form:2dQ(t)=γ(t)Q(t)dt+σ(t)Q(t)[ρ(t)dW(t)+1−ρ(t)dW(t)]212And byfd(D(t
3、)M(t)Q(t))has no dt part 2fwe can matchσ(t)ρ(t)Θ(t)+σ(t)1−ρ(t)Θ=R(t)−R(t)+γ(t)(9.3.10)2122Bring it back to dQ(t)We get fdQ(t)=Q(t)[R(t)−R(t)+σ(t)dW(t)]232wheredW(t)=ρ(t)dW(t)+1−ρ(t)dW(t)312