I agree. The problem with these kind of studies is that they aren't coming up with explanations and then trying to test them; they're simply looking for correlations, which advances knowledge not one jot.
In the case of caffeine I suspect the main benefit is that drinkers of tea and coffee eat less food, and rich food damages the blood vessels of the brain (as it does to the heart).