Black-Scholes Characteristic Function(CF)¶
CF is defined as:
It’s the expected value of a complex exponential, and it uniquely determines the distribution of the random variable X — like a frequency signature of the distribution.
Deriving the Black-Scholes Characteristic Function¶
We assume that:
That’s just the log of the stock price under GBM, which is normally distributed.
Then using the Definition of the CF, substitute in the normal PDF and solve:
Then combine the exponentials, we get:
Where:
- \(\hat{\mu} = s_0 + \left(\mu - \frac{1}{2}\sigma^2\right)T\)
- \(\hat{\sigma} = \sigma \sqrt{T}\)
This is just a normal PDF wrapped in a complex exponential.
To solve the integral, we rewrite the exponent using a completing-the-square trick:
We end up with:
Substituiting back:
This is the closed-form CF of \(\log S_T\) in the Black-Scholes model.
This is actually the same as the CF of a normal distribution:
So this confirms that:
- \(\log S_T\) is normal
- The CF matches the theory
We derived the characteristic function of \(\log S_T\) under Black-Scholes by plugging the normal distribution into the definition of CF and simplifying the integral using a completing-the-square trick. This gives us a clean formula that’s perfect for plugging into Fourier pricing methods.