I find it difficult to work with $\sigma$-algebra generated by several variables... Here's my question :
Let $\lbrace X_{j}, j\in \mathbb{Z} \rbrace$ be a sequence of random variables, such that for all finite set $J \subset \mathbb{Z}$, and for all $m \in \mathbb{N}$, the random vectors $(X_{j}, j\in J)$ and $(X_{j+m}, j\in J)\, $ follow the same distribution.
For all $j \in \mathbb{Z}$, let $\, \mathcal{M}^{j}_{-\infty} \,$ be the $\sigma$-algebra generated by $\lbrace X_{i}, i \leq j\rbrace \,$, and $\mathcal{M}^{+\infty}_{j} \,$ the $\sigma$-algebra generated by $\lbrace X_{i}, i \geq j\rbrace \,$.
We define $\alpha_{n} = \alpha(\mathcal{M}^{0}_{-\infty}, \mathcal{M}^{+\infty}_{n})$, where for all sub-tribes $\mathcal{A}, \mathcal{B}$, $$\alpha(\mathcal{A}, \mathcal{B}) = \sup_{A \in \mathcal{A}, B \in \mathcal{B}} |P(A \cap B) - P(A)P(B) |$$
Now I want to show that for all $j \in \mathbb{Z}$, $$\alpha(\mathcal{M}^{0}_{-\infty}, \mathcal{M}^{+\infty}_{n}) = \alpha(\mathcal{M}^{j}_{-\infty}, \mathcal{M}^{+\infty}_{j+n})$$
I understand the idea : it follows from the fact that $(X_{i}, 1 \leq i \leq j)$ and $(X_{i}, n-1\leq i \leq j+n)$ follow the same distribution. But I don't know how to formalize it...
Edit : I don't know if it helps, but I forgot an assumption : $X_{0}$ has a zero mean.