In my first look at the Congressional Budget Office's estimate for the House GOP's health reform , a.k.a. the American Health Care Act (AHCA), I pointed out that the agency predicts most of the people set to "lose" coverage would in fact be choosing to stop buying insurance once the individual mandate was removed. It turns out the truth about the GOP bill's effect on coverage is even more nuanced than partisans are making it out to be.
The question of how many people wouldn't have coverage in the future hinges in large part on how many people are expected to have coverage in the future under current law, i.e. Obamacare. After all, the CBO isn't projecting how many fewer people will have coverage compared with today, but rather compared with its projections for the future. And the one thing we know about the CBO is it's had a really hard time predicting the future when it comes to insurance coverage purchased on the Obamacare exchanges. Just look at this graph, which comes from Avik Roy writing at Forbes :
As you can see, the CBO's projections for the exchanges have consistently been waaaaaay too high. For 2015, it was off by about 30 percent. For 2016, it originally forecast -- on three different occasions -- that over twice as many people would buy insurance than the number who actually did so (more than 20 million vs. 10 million). Even when it revised its estimate down to 12 million, it was off by 20 percent.
For 2017, the CBO predicted a 50 percent surge in exchange enrollment with some 5 million new sign-ups, when in fact the increase was only 10 percent with 1 million. For next year, the CBO forecast another big jump in enrollment, even though nothing about our experience with the exchanges up till now supports that prediction.
The upshot, as Roy notes is this: CBO's baseline of future enrollment, based on actual data so far, likely "is off by 7 to 8 million in future exchange enrollment; hence, the impact of the AHCA is also off by the same amount."
That's roughly one-third of the CBO's projected decrease in exchange-based coverage under the GOP plan. That is about the same proportion as the CBO projected for Medicaid enrollment drops in non-expansion states that, for some reason, the CBO assumed would eventually change course and expand the program after all.
So, one-third of the enrollment "drop" can be chalked up to unrealistic expectations about the future effect of Obamacare, and about half of it is attributable to people choosing not to buy insurance once the individual mandate goes away. That leaves about one-sixth of the decrease possibly being real -- about 6 million out of the headline number of 24 million.
Roy's numbers are slightly different from mine -- and, unlike the CBO, he doubts that many people would actually drop coverage due to the elimination of the individual mandate -- but he ultimately arrives at nearly the same place :
"You add all that up—7 million off on future exchange enrollment, around 9 million off on the individual mandate’s power, and 3 million off on future Medicaid expansions—the CBO’s estimate of the impact of the AHCA on coverage is off by 19 million, and that the real impact of the AHCA on coverage is negative 5 million."
There's no doubt the effect on those 5 million or 6 million people stand to be worse off, and Roy goes on to write about ways the AHCA could be amended to address them while keeping in place most of the big deficit savings from repealing Obamacare.
But there's also no doubt the discussion about the bill would be vastly different if more people understood and honestly reported on factors such as the CBO's poor track record of estimating coverage gains and the reasons why some people would drop -- not "lose" -- their coverage.