Question: What is the delta of an at-the-money European call option with respect to volatility?

Note that $$\frac{\partial\Delta}{\partial\sigma} = N'(d_1) \frac{\partial d_1}{\partial\sigma} = N'(d_1) \frac{- d_2}{\sigma} = \frac{-N'(d_1)d_2}{\sigma}$$ where $N(\cdot)$ is the CDF of the standard normal distribution. I am not able to deduce anything from this equation.

This QFSE post states that higher volatility for in-the-money option will have lower delta whereas higher volatility for out-of-the-money options will have higher delta.

Based on this website, it seems that higher volatility will lead to $\Delta = 0.5.$ But I am not able to show this.

  • $\begingroup$ The question is strangely asked. I would interpret it as “first-order sensitivity (delta) to volatility”, i.e. vega, not $\frac{\partial \Delta}{\partial \sigma}$. $\endgroup$ – siou0107 Mar 23 at 16:32
  • 1
    $\begingroup$ As an aside: the "holy trinity" of strikes are d2=0, ATMF, d1=0, and the holy trinity of Greeks are Vega,Vanna, and Volga. Understand the behaviour of these three Greeks at those three strikes, and you understand 95% of the volatility smile. $\endgroup$ – ilovevolatility Mar 24 at 10:21
  • $\begingroup$ How to obtain that ATMF implies $d_2=0$? $\endgroup$ – Idonknow Mar 24 at 14:14
  • $\begingroup$ d_2 = 0 is not the ATMF strike. The answer by @siou0107 is the answer to your question. $\endgroup$ – ilovevolatility Mar 24 at 14:21
  • $\begingroup$ But he doesn't answer when the call is ATMF. $\endgroup$ – Idonknow Mar 24 at 14:41

Based on your computation, you can observe that the $N’$ term is always positive, between 0 and 0.4. As $\sigma$ is always positive, you can focus on the $-d_2$ term. When $d_2 > 0$, i.e. call is ITM, delta has a negative sensitivity to volatility ; conversely for OTM call. That is in line with your remark.

| improve this answer | |
  • $\begingroup$ What about at-the-money? $\endgroup$ – Idonknow Mar 23 at 16:41
  • $\begingroup$ At the money it is fairly insensitive, as $d_2 \approx 0$ $\endgroup$ – siou0107 Mar 23 at 19:04
  • $\begingroup$ How do you obtain that $d_2\approx 0$ for at-the-money option? I have $d_2 = \left( \frac{r}{\sigma} - \frac{\sigma}{2}\right) \sqrt{\tau}$ only. $\endgroup$ – Idonknow Mar 24 at 8:40

In the following, I am assuming the BS73 model and I assume that "ATM" means

$$ S = Xe^{-r\tau} $$ The pricing formula for a European call then becomes $$ \tag{1} O\propto N\left(+\frac{1}{2}\sigma\sqrt{\tau}\right)-N\left(-\frac{1}{2}\sigma\sqrt{\tau}\right) $$ times some scaling factor which is irrelevant for our purpose. Clearly, $$ Vega\equiv\frac{\partial O}{\partial \sigma}=\frac{1}{2}\sqrt{\tau} \cdot{} n\left(\frac{1}{2}\sigma\sqrt{\tau}\right)+\frac{1}{2}\sqrt{\tau} \cdot{} n\left(-\frac{1}{2}\sigma\sqrt{\tau}\right) $$ Leading us to $$ \tag{2} \frac{\partial O}{\partial \sigma}=\sqrt{\tau}\frac{e^{-\frac{1}{2}\left(0.5\sigma\sqrt{\tau}\right)^2}}{\sqrt{2\pi}} $$ Thus:

  • For longer maturities, the Vega is larger than for smaller maturities
  • For all practical purposes (i.e. $IV<75\%$, $\tau<1yr$, you can approximate the ATM Vega to $$ \tag{3*} Vega \approx \sqrt{\frac{\tau}{2\pi}} $$
| improve this answer | |
  • 1
    $\begingroup$ I think you are answering different question. $\endgroup$ – Idonknow Mar 24 at 14:15
  • $\begingroup$ I interpreted your question " What is the delta of an at-the-money European call option with respect to volatility?" as the first derivative (hence Delta) w.r.t. Volatility. $\endgroup$ – Kermittfrog Mar 24 at 15:12
  • $\begingroup$ I think you are calculating Vega. But what I want is how delta changes if volatility changes? $\endgroup$ – Idonknow Mar 24 at 15:26
  • $\begingroup$ Funny enough: The result is the same, scaled by a factor of 2. $\endgroup$ – Kermittfrog Mar 24 at 18:52

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.