0
$\begingroup$

Lets say we have a volatility surface for the SPX at time t with spot S. We consequently know the price of some call option at maturity T with strike K. What is the risk neutral expectation of the option price at t+1day, if spot moves 1% ( E[option price | S*1.01] ) ?

$\endgroup$
5
  • $\begingroup$ Be more specific. What is '1' here, infinitesimal / very small, or 1 month, 1 year? $\endgroup$
    – Frido
    Apr 13, 2023 at 13:54
  • $\begingroup$ Lets say 1 day. $\endgroup$
    – Rodrigo
    Apr 13, 2023 at 14:07
  • $\begingroup$ Why not just use the delta and gamma then? $\endgroup$
    – Frido
    Apr 13, 2023 at 14:30
  • $\begingroup$ Because volatility also moves. $\endgroup$
    – Rodrigo
    Apr 13, 2023 at 15:29
  • $\begingroup$ See answer below. $\endgroup$
    – Frido
    Apr 13, 2023 at 15:32

1 Answer 1

1
$\begingroup$

Your question is unclear / lacks relevant detail, but I suspect what you're really asking is what happens to the vol change when the spot change is given.

Assume, as an example, the following dynamics: \begin{align} dS_t &= \sigma_t S_t dW_t \\ d\sigma_t &= \alpha \sigma_t ( \rho dW_t + \sqrt{1-\rho^2} dZ_t) \end{align} with $dW dZ = 0$.

You'd like to calculate $E_t[ C(S_{t+dt},\sigma_{t+dt},K) | dS_t = c]$. Now $$ C(S_{t+dt},\sigma_{t+dt},K) = C(S_t,\sigma_t, K) + dC(S_t,\sigma_t,K) $$ with $$ dC(S_t,\sigma_t,K) = \Delta dS_t + \nu d\sigma_t $$ with $\Delta$ the delta of the option and $\nu$ the vega.

Given $dS_t = c$ then \begin{align} dS_t &= c\\ d\sigma_t &= \frac{\rho \alpha}{S_t} dS_t + (\cdot) dZ_t \\ & = \frac{\rho \alpha}{S_t} c + (\cdot) dZ_t \end{align} Thus \begin{align} E_t[ C(S_{t+dt},\sigma_{t+dt},K) |dS_t = c] &= C(S_t,\sigma_t,K) + c \Delta + \frac{\rho \alpha}{S_t} c \nu + E[ (\cdot) dZ_t] \\ &= C(S_t,\sigma_t,K) + c \Delta + \frac{\rho \alpha}{S_t} c \nu \end{align}

$\endgroup$
7
  • $\begingroup$ The problem now becomes: What is the risk neutral estimate of the spot-vol correlation $\endgroup$
    – Rodrigo
    Apr 13, 2023 at 15:33
  • $\begingroup$ That's a different question but it's well known that correlation is proportional to the ATM slope. $\endgroup$
    – Frido
    Apr 13, 2023 at 15:35
  • $\begingroup$ Well, you framed your answer in a way that left one variable to which we do not know the risk neutral estimate: the spot vol correlation, which should indeed be somewhat proportional to the skew. $\endgroup$
    – Rodrigo
    Apr 13, 2023 at 15:45
  • $\begingroup$ Yeah ok, and you're welcome by the way. $\endgroup$
    – Frido
    Apr 13, 2023 at 15:49
  • $\begingroup$ Didn't mean to sound smug, sorry if was the case, thanks for taking your time to reply. $\endgroup$
    – Rodrigo
    Apr 13, 2023 at 16:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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