Consider the stochastic process $X$ that satisfies the SDE \begin{equation} X_t=1+\mu\int_0^t X_s \, ds+\sigma\int_0^t X_s \, dB_s, \end{equation} with $\mu$ and $\sigma$ constants. Consider another process $Y$ given by \begin{equation} Y_t=X_t^\beta, \end{equation} with $\beta\ge 2$. Find the stochastic differential equation for process $Y$.

I am looking at the following solution: \begin{equation} dY_t=\beta X_t^{\beta-1}dX_t+\beta(\beta-1)X_t^{\beta-2}(dX_t)^2\\[10pt] dY_t=\beta X_t^{\beta-1}(\mu X_t \, dt+\sigma X_t \, dB_t)+\beta(\beta-1) X_t^{\beta-2} \sigma^2 X_t^2 \, dt\\[10pt] dY_t=\beta\mu X_t^\beta \, dt+\sigma\beta X_t^\beta \, dB_t + \beta(\beta-1) \sigma^2 X_t^\beta \, dt\\[10pt] dY_t=(\beta\mu+\beta(\beta-1)\sigma^2)Y_t \, dt+\sigma\beta Y_t \, dB_t \end{equation}

Assuming this solution is right, my $\textbf{question}$, or the problem I'm having difficulty fully understanding is, why is the second spatial derivative not multiplied by $\frac{1}{2}$? Unless I'm totaly spacing, should that term not be \begin{equation} \left[\frac{\beta(\beta-1)}{2}\right]X_t^{\beta-2}(dX_t)^2 \end{equation} from an application of the Ito formula? Thank you.


You have a typo in your first equation, it should be $$dY_t=\beta X_t^{\beta-1}dX_t+\beta(\beta-1)X_t^{\beta-2}(dX_t)^2$$

But it looks to me that you are correct, and the above equation is wrong.

If $Y_t =f(X_t)$, with $f(x)=x^{\beta}$, I do expect the $\frac{1}{2}$:

$$dY_t=df(X_t)=\frac{\partial f(X_t)}{\partial x}dX_t + \color{red} {\frac{1}{2}} \cdot\frac{\partial^2 f(X_t)}{\partial x^2}(dX_t)^2$$


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.