It's one thing buying too much bacon and another actually using it for personal consumption. You might as well be cooking it for your friends or your dog. This does not make sense.
You're clearly not an actuary or -- more importantly -- an insurance exec. Just because there isn't a perfect causal link doesn't mean that the correlation isn't useful for minimizing risk.
I'd wager that at least a double digit percentage of young males are as careful as young females in their driving behavior. From the insurance perspective, who cares? The risk pool still pays a premium.