Megan McArdle at The Atlantic see’s it this way
1) Thanks to reconciliation instructions, they needed to improve the budget impact by at least $1 billion in the sidecar. They improved it by exactly $1 billion. Which goes back to what I’ve now said several times: the CBO process has now been so thoroughly gamed that it’s useless.
2) The proposed changes increase spending dramatically, most heavily concentrated in the out-years. The gross cost of the bill has risen from $875 billion to $940 billion over ten years–but almost $40 billion of that comes in 2019. The net cost has increased even more dramatically, from $624 billion to $794 billion. That’s because the excise tax has been so badly weakened. This is of dual concern: it’s a financing risk, but it also means that the one provision which had a genuine shot at “bending the cost curve” in the broader health care market has at this point, basically been gutted. Moreover, it’s hard not to believe that the reason it has been moved to 2018 is that no one really thinks it’s ever going to take effect. It’s one thing to have a period of adjustment. But a tax that takes effect in eight years is a tax so unpopular that it has little realistic chance of being allowed to stand.3) As I expected, the size of the magic asterisk–the modern equivalent of David Stockman’s infamous “savings to be named later” in the Reagan budgets–has had to be beefed up to offset the new spending.4) Medicare Advantage is being effectively outlawed–in some areas, the reimbursements will actually be below those of the fee-for-service programs.