I was working through my textbook and found this problem that I got stuck at: Consider the AR(2) Model $$X_t = \phi_1X_{t-1}+\phi_2X_{t-2}+\epsilon_t$$ We can assume $\phi_2 > 0$, so the roots of the polynomial are real numbers. Show that the inequalities for the values of $\phi_1$ and $\phi_2$ which ensure it is causal stationary are given by: $$\phi_1 + \phi_2 < 1$$ $$\phi_2 - \phi_1 < 1$$ $$|\phi_1|<1, \phi_2 < 1$$

So here's what I've done so far:

Firstly I rearranged the equation to include the backshift operator $$(1-\phi_1B-\phi_2B^2)X_t = \epsilon_t$$ Then I used the quadratic equation for the roots $$z=\frac{\phi_1±\sqrt{\phi_1^2+4\phi_2}}{-2\phi_2}$$

Except I'm not sure where to go from here. How do I calculate the inequalities that ensure it's causal stationary from here? How do I figure out the conditions?

Thanks heaps in advance!


1 Answer 1


In this framework, for causal stationarity, you need $|z|>1$. This means, $$|\phi_1\pm\sqrt{\phi_1^2+4\phi_2}|>2|\phi_2|=2\phi_2$$ Consider $\phi_1>0$ case first. Then, for the first root, we have $\phi_1+\sqrt{\phi_1^2+4\phi_2}>2\phi_2$, which means $\phi_1^2+4\phi_2>4\phi_2^2+\phi_1^2-4\phi_1\phi_2 \rightarrow \phi_2-\phi_1<1$. For the other root, we have $-\phi_1+\sqrt{\phi_1^2+4\phi_2}>2\phi_2$, which yields $\phi_2+\phi_1<1$.

For $\phi_1<0$, you obtain the same inequalities; and these two can be summarized as $\phi_2+|\phi_1|<1$, which also means $|\phi_1|<1$ and $\phi_2<1$. Don't forget that $\phi_2$ is positive.


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.