3
$\begingroup$

Suppose I have a time series process $\{X_t\}$ that is strictly stationary in the sense that the joint distribution of $[X_{t_1},...,X_{t_k}]$ and $[X_{t_1+a},...,X_{t_k+a}]$ are the same for any set of integers $t_1,...,t_k$ and any integer $a$. If in addition, I know that this process have mutually independent observations. Does this imply that $\{X_t\}$ is an i.i.d. process? My guess is that it is. Because strict stationarity implies that every term in this time series have the same distribution as $X_1$, and thus when independence also hold, it is an i.i.d. process. Does this look correct?

$\endgroup$
2
  • $\begingroup$ "Because strict stationarity implies that every term in this time series have the same distribution as $X_1$" This is not what strict stationarity means. This contradicts the definition that you gave, which is about the joint distribution, and not the marginals. $\endgroup$
    – mhdadk
    Commented Apr 23, 2023 at 23:05
  • 2
    $\begingroup$ @mhdadk Yeah, it's about joint distribution, but I meant it IMPLIES marginals are have identical distribution. Just take $t_1,...,t_k$ as 1, the definition says $X_1$ has the same distribution as $X_{1+a}$ for any $a$. What's wrong with this? $\endgroup$ Commented Apr 23, 2023 at 23:16

1 Answer 1

4
$\begingroup$

Yes, that is correct. Strict stationarity implies a common marginal distribution for the variables in the series, which is the ID part in IID. If you combine this with an assumption of independence you then get IID.

$\endgroup$
5
  • $\begingroup$ Thanks a lot! This is very helpful. $\endgroup$ Commented Apr 24, 2023 at 4:02
  • $\begingroup$ But isn't the assumption of "mutual independence" in the question weaker than independence? That distinction appears to be the crux of the matter. $\endgroup$
    – whuber
    Commented May 6, 2023 at 15:38
  • $\begingroup$ @whuber: Isn't mutual independence the stronger one (i.e., independent conditional on any of the observations in the series)? I was under the impression that pairwise, triplewise, etc., were the weaker and mutual independence is full independence. $\endgroup$
    – Ben
    Commented Oct 3, 2023 at 5:02
  • $\begingroup$ You could be right: I had understood "mutual" as meaning "pairwise," for otherwise why modify the word "independence" at all? $\endgroup$
    – whuber
    Commented Oct 3, 2023 at 13:10
  • 1
    $\begingroup$ @whuber: For now I'll leave the answer unchanged, since I think that mutual independence is the strong form. Like you, I always take the unqualified term "independence" as meaning the strong (mutual) form, so don't need the qualifier, but perhaps OP wanted to specify to be sure. $\endgroup$
    – Ben
    Commented Oct 3, 2023 at 15:02

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.