R
RedCoat
Member
Hi, please could somebody help with applying Ito's Lemma to Geometric Brownian Motion? Apologies for it being horrendous to read, I couldn't seem to get the maths notation to display in my post. I will use pd to show partial derivatives.
Suppose we have dSt = St (mu dt + sigma dZt) where Zt is a Wiener process.
Then to apply Ito's Lemma we let dQt = dSt / St = mu dt + sigma dZt, so this is now in a form to apply Ito's Lemma.
Ito's Lemma says that if dXt = mu dt + sigma dWt then df(Xt) = (mu pdf/pdXt + 0.5 * sigma^2 pd^2f/pdXt^2) dt + sigma pdf/pdXt dWt
Substituting,
df(Qt) = (mu * pdf/pdQt + 0.5 * sigma^2 * pd^2f/pdQt^2) dt + sigma pdf/pdQt dZt
Let f(Qt) = log(Qt)
Then pdf/pdQt = 1/Qt and pd^2f/pdQt^2 = -1/Qt^2
Substituting again
= df(Qt) = (mu/Qt - 0.5sigma^2 / Qt^2) dt + sigma/Qt dZt
Now, looking at (for example) the solution to Q3 on the April 2017 exam, I think my Qts should = 1 here, but I'm not sure why! Any help would be hugely appreciated. Think I'm making a huge mess of this.
Suppose we have dSt = St (mu dt + sigma dZt) where Zt is a Wiener process.
Then to apply Ito's Lemma we let dQt = dSt / St = mu dt + sigma dZt, so this is now in a form to apply Ito's Lemma.
Ito's Lemma says that if dXt = mu dt + sigma dWt then df(Xt) = (mu pdf/pdXt + 0.5 * sigma^2 pd^2f/pdXt^2) dt + sigma pdf/pdXt dWt
Substituting,
df(Qt) = (mu * pdf/pdQt + 0.5 * sigma^2 * pd^2f/pdQt^2) dt + sigma pdf/pdQt dZt
Let f(Qt) = log(Qt)
Then pdf/pdQt = 1/Qt and pd^2f/pdQt^2 = -1/Qt^2
Substituting again
= df(Qt) = (mu/Qt - 0.5sigma^2 / Qt^2) dt + sigma/Qt dZt
Now, looking at (for example) the solution to Q3 on the April 2017 exam, I think my Qts should = 1 here, but I'm not sure why! Any help would be hugely appreciated. Think I'm making a huge mess of this.