I wonder if there's also a social component with moderate drinking along with other factors; being able to enjoy drinks with colleagues after work or catch up with friends at a bar are culturally acceptable activities that seem to align with the relationship between a close social circle and longevity/health, especially amongst middle age and older adults.
Here again, though, you have the same correlation/causation problem. People without a close social circle may not have one because they're too sick to get out much, or they're on a drug regimen that makes them drowsy or grouchy.
Why is that a safer bet when randomized control trials with animals finds the same thing?