Let's recall four standard reasons for a correlation of two variables:
(1) Coincidence. One rose over some time period, and the other also rose over the same period, despite being completely unrelated. (The world murder rate has been going down, and atmospheric carbon dioxide levels have been going up. This is a negative correlation.)
(2) A causes B. (Cigarette smokers display lung cancer at elevated rates)
(3) B causes A. (Cigarette smokers display lung cancer at elevated rates)
(4) Root cause. Some other phenomenon causes both A and B. (Children in homes containing books display intelligence at elevated rates)
This should be obvious, but arguing against option (2) doesn't mean arguing for option (3).
Well, I can't actually produce a number for this, but how would you feel about a correlation between first digit of US zip code and incidence of Tay-Sachs disease?
Note: Tay-Sachs is something that happens to Jews, and Jews are concentrated in New York.
Does the zip code connection satisfy your demand of "in this context"?
My point is that "where you were born" is a broader term than "the zip code where you were born." Obviously your zip code doesn't cause anything. But your zip code determines things like what the schools are like in the area, what the income demographics are like in the area, etc. When people say "where you were born determines where you end up" they're referring to that larger definition of "where you were born" not literally your zip code. That's what makes many versions of (4) collapse into (3).
Though I think your Tay-Sachs example is one that arguably does not.
