Consider a univariate stochastic process (time series) $X_t$. I am interested in conditions under which $\lim_{j\to \infty} \mathbb{E}_t[X_{t+j}]$ exists. For example if $X_t$ is a stationary process it seems that $\lim_{j\to \infty} \mathbb{E}_t[X_{t+j}]$ is constant. Is this correct? Are there other interesting cases? I have been told that the existense of this limit might be related to "conditional stationarity".

Intuitively $\mathbb{E}_t[X_{t+j}]$ should approach its long run unconditional mean (which is constant due to stationarity). However, I am not sure how to prove this result. An example of a strictly stationary process for which this property is true would be the standard AR(1) process $X_t=\rho X_{t-1}+\epsilon_t$, where $-1<\rho<1$ and $\epsilon_t$ is Gaussian white noise.


Another interesting case? If $X_t$ is a martingale, then for any $j$ $$E[X_{t+j}|X_t]=X_t$$ and so $$E[X_{t+j}]=E[X_t]$$ but you typically don't have stationarity (even approximately)

| cite | improve this answer | |
  • 1
    $\begingroup$ Thanks! Sure e.g. with a random walk, the limit is well-defined to be the most recent value. $\endgroup$ – fesman Jun 28 at 7:10

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.