That's the point though. Even if this fact is true it doesn't mean that drinking is healthy for these moderate drinkers. There could be other reasons that they're living longer and that drinking is actually working against them, not for them.
That's not the kind of inference we can derive from the data though.
The conclusion is that, even if we know that alcohol is not good for our health, moderate drinking has not been correlated to higher mortality or shorter life span, so we can conclude that, at the time of this writing, there are no evidence that show that moderate drinking is worse than not drinking for our health.
Regardless of the causes, the data simply doesn't show any correlation whatsoever with worsening health conditions, it shows, however, a correlation with a mildly longer average life span, which has not been explained, yet.