The price of a call with a stock with Bachellier process as its underlying and zero interest rate is giving by: $$C(t)=(S(t)-K)\Phi(\frac{S(t)-K}{\sigma \sqrt{T-t}})+\sigma \sqrt{T-t} \phi(\frac{S(t)-K}{\sigma \sqrt{T-t}})$$ How do I compute/derive/proof the time $t$ Delta-hedge ratio: hence compute $dC(t)/dS(t)$?

I have tried to reach to the delta by looking at the answers to this question: Bachelier option delta = probability of exercise? but there's some differentiation in there which is not clear for me.


Zero interest rate and drift so $S(T) = S(t) + \sigma (W(T)-W(t))$ and $\frac{d S(T)}{dS(t)} = 1. $ $$ C(t) = E_t[(S(T) - K)^+] $$ $$ \frac{dC(t)}{dS(t)} = \frac{d}{dS(t)} E_t[(S(T) - K)^+] = E_t[\frac{d}{dS(t)} (S(T) - K)^+] = E_t[\frac{d S(T)}{dS(t)} \text{Indicator}(S(T) > K)]= E_t[\text{Indicator}(S(T) > K)]= \text{Prob}_t[S(T) > K]=\Phi(\frac{S(t)-K}{\sigma \sqrt{T-t}}) $$

  • $\begingroup$ $\frac{d}{dS(t)} E_t[(S(T)-K)\textbf{1}_{S(T) > K}]= E_t[\frac{d}{dS(t)}(S(T)-K)\textbf{1}_{S(T) > K}]$... This is indirect application of Leibniz, right? In that case; we need to ensure contuinuty. As far as I know: $(S(T)-K)\textbf{1}_{S(T) > K}$ is not contnous? $\endgroup$ – Sanjay Mar 21 '18 at 20:21
  • $\begingroup$ I forgot to tag you before just wanted to make sure if you see the comment $\endgroup$ – Sanjay Mar 21 '18 at 21:13
  • $\begingroup$ Yes. $x \mapsto (x-K)^+$ is continuous and differentiable. The derivative is discontinuous but it does not matter. $\endgroup$ – Antoine Conze Mar 22 '18 at 7:42
  • $\begingroup$ It does not appear correct that $\frac{d}{dS(t)}(S(T)-K)\pmb{1}_{S(T)>K} = \pmb{1}_{S(T)>K}$. $\endgroup$ – Gordon Mar 22 '18 at 15:15
  • $\begingroup$ $\frac{d}{dS(t)} (S(T)-K)1_{S(T)>K} = \frac{dS(T)}{dS(t)} 1_{S(T)>K}$ $\endgroup$ – Antoine Conze Mar 22 '18 at 15:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.