But it is evidence for causation.
Is it?
Yes. If A is correlated with B, is A more likely to cause B than if A is not correlated with B? Clearly yes.