I feel the need to repost something I wrote in this thread:
As I said before, correlation does not imply causation. (Please note - if you haven’t already - that I’m using “imply” to mean logical implication.) Statistical regression can determine the probability distributions of dependent variables as functions of independent variables, but that in no way establishes a causal link. All it can do is help one make more accurate forecasts. Otherwise, one must conclude that reducing the number of pirates over time has been a, if not the, cause of (alleged) global warming.
Note that the above doesn’t take into account any flaws with the variables under consideration (such as GDP, in the context of econometrics).