Stochastic Calculus

Lesson 9 of 11

Girsanov's Theorem

Two probability worlds, the same paths: change the measure and the drift moves, the volatility doesn’t.

The idea

Reweight the paths of a Brownian motion with drift and you can make the drift disappear. Under a new measure Q\mathbb Q, with density dQdP=e−θWT−12θ2T\frac{d\mathbb Q}{d\mathbb P} = e^{-\theta W_T - \frac12 \theta^2 T}, the process Wt+θtW_t + \theta t is a standard Brownian motion. The two measures are equivalent: they agree on which events have probability zero, so the paths are the same and only their weights change.