Right, when you insure someone, the expected payout is just the average for that person's demographic with the same risk factors. But because of individual randomness, the standard deviation is much higher. So the expected payout for an individual could be $2000, but with a standard deviation of $0 to $20,000. With groups, the larger sample size creates a much smaller standard deviation, and you don't see nearly as much variance in costs. If the group is large enough, you should be able to model the costs on a Normal distribution. This allows you to even out the premiums across members.
Also, I believe the ACA is mandating community rating, which means a standard price for everyone regardless of risk factors.
Yup. Right on all counts.
I wouldn't mind if the ACA was set up not to create plans for states that didn't comply with the law - it should be all or nothing. It's hard to see the government creating more than a dozen effective exchanges for states that refuse to even work with them.
You get what you vote for.
Except I don't really think that people should be denied health insurance at an affordable price because 51% of the populace (or just the politicians) are goddamn morons. I guarantee that if you put the exchange up for a vote in Jersey they would want to do it, it's just Christie being a dummy and positioning himself for 2016.