You have a history of coming out of nowhere in the middle of discussions, misinterpreting my motivations, and attacking me as a person before the issue. Just like you did in this thread, first posting 'William, shut up' before editing your post or even mention what you were referring to. I actually think we would get along in real life, better than it seems online.
The best argument so far imho. But I honestly I think this may be a case of Ti versus Te. See, while it might make sense to have a more consistent system like that, I would rather view it objectively for profit. For example, I don't consider the risk that someone with blue eyes becoming an alcoholic is nearly as likely as a woman taking maternity leave. As a brief example:
I'm interviewing two people. Ceterus paribus, same experience, background, etc., but a woman and a guy in their 20s. Let's guess that the woman has a 40% chance to take maternity leave at some point in her 20s, which will cost me 3 months of labor. The blue-eyed man has a 2% of becoming an alcoholic because of his predisposition, much less likely, and it won't cost me nearly as much as 3 months of lost labor completely, AND that would be more easily identifiable as his work begins to drop, and I would have time to find a replacement. If there's a 40% chance I will lose money versus a 2% chance, I would consider the woman to be 20 times riskier, if the costs were the same.
Basically, I think many employers are more concerned with maternity leave as opposed to hiring drunks, at least for professionals in their 20s. Drunkenness might be a bigger problem among part-time or lower-wage jobs, but those positions rarely offer maternity leave anyway, imho.
So good argument, but I believe the cost in the decision is more important than any arbitrarily consistent value system.