Wait, why would it be more efficiently allocated today? The number of small states has only gone up over time and it's also halfway through a census cycle.
I don't think the words I used conveyed the thing I was trying to convey.
What I had in mind is this: If every state exactly tracked the national vote, then winning 50%+1 of the national two-party vote gives you 100% of the EVs. If, on the other hand, states are bifurcated into deep rep states and deep blue states, then winning the exact same number of votes nationally would not result in 100% of the EVs, and depending on which states are red and which are blue, you could even lose if your opponent more efficiently translated votes into EVs.
Okay, now imagine that states have all different proportions of underlying support. We can't measure underlying support, it's a theoretical quantity. Apply a shock to the system that swings every state uniformly (or proportionately) in some direction by some amount. If there are a lot of states in the 47-53% range, then that shock swings comparatively lots of EVs per percentage point of vote swing--if there are a lot of states in the <40% >60% range, then that shock swings comparatively fewer EVs per percentage point of vote swing.
Applying the same logic to this election, my point was that I
suspect--and I don't have data for this although it would be fairly easy to do--that a uniform swing (or proportional swing) in national popular vote does not buy you as many EVs as it did in 1984 or 1936. I guess what I mean is I suspect there are fewer swing states now than there were then. Partisan self-sorting has been very well documented over the last few decades, which is what people talk about when they say polarization. I think the evidence on geographic partisan self-sorting is a little weaker. I'd have to check the literature so this is a more vague intuition.
Of course a complication is that we don't expect the swing to be uniform/proportional, especially if we believe Trump is going to cause a major bump in Hispanic turnout and we know Hispanic voters are allocated in a very lopsided way. But yeah, my
sense is that if you gave Hillary 57-59% of the vote, you wouldn't see a 500+ EV blowout.