CBO: Funding CHIP will cost* 10% as much...because killing the mandate made the alternative to CHIP cost that much more.

A new letter from the Congressional Budget Office to Sen. Orrin Hatch says:

Re: Updated Cost Estimate for S. 1827, the Keep Kids’ Insurance Dependable and Secure Act of 2017

Dear Mr. Chairman:

The Congressional Budget Office and the staff of the Joint Committee on Taxation (JCT) have updated their estimate of S. 1827, the Keep Kids’ Insurance Dependable and Secure Act of 2017, to account for the enactment of Public Law 115-97, which repealed the penalties related to the individual health insurance mandate starting in 2019, and to account for administrative action. The agencies now estimate that enacting this legislation would increase the deficit by $0.8 billion over the 2018-2027 period.

This one is quite a head-scratcher, but let me see if I can peel the onion.

First, it's vitally important to understand that how much the program would increase the deficit is NOT the same thing as how much it would cost in hard dollars. For instance, the actual cost of the CHIP program in 2016 was around $13.6 billion to cover an average of 5.9 million children per month (yes, there are around 9 million kids who are on the program for some portion of the year, but many are only enrolled for a few months). Between population growth and inflation, it's safe to assume that the actual cost of the CHIP program from 2018 - 2027 would likely total something like $150 billion, or around $75 billion for the 5 years that's on the table.

OK, so what's with the original $8.3 billion and the revised $800 million? Well, remember that these updates are relative to budget projections under previous circumstances.

S. 1827 would extend federal funding for the Children’s Health Insurance Program (CHIP) for five years, through fiscal year 2022. The bill also would make several other changes to CHIP, including a change in the federal matching rate for the program and extension of the requirement that states maintain eligibility levels as they were in 2010.

Compared with the October 20, 2017, cost estimate for the legislation, this cost estimate shows a cost over the next 10 years that is smaller by $7.5 billion. 1 The net change in the deficit is significantly smaller than the agencies’ prior estimate primarily because the offset to the cost of funding CHIP for five years is larger. That offset, which reduces direct spending related to the marketplaces, is higher for three main reasons:

• CBO and JCT expect that premiums for coverage through the marketplaces will be higher in the absence of the mandate penalties than they would otherwise have been. As a result, the federal cost of enrolling a child in coverage through the marketplaces will be higher. Thus, funding CHIP for five additional years—causing some children to be covered in that program rather than through the marketplaces—would result in a larger reduction in spending related to the marketplaces than in the prior estimate.

$13.6 billion divided by an average of 5.9 million children means the CHIP program costs the government around $2,300 per child per year.

Last fall, the CBO projected that if the CHIP program were allowed to be discontinued, several million parents would instead enroll their children in private policies via the ACA exchanges. Since all of these families are by definition relatively low-income (otherwise their kids wouldn’t qualify for CHIP in the first place), that means the children would receive heavy ACA exchange subsidies.

I don't know exactly how much the average CHIP-eligible income level enrollee is eligible for under a subsidized ACA plan, but as an example, the Kaiser Family Foundation subsidy estimator says that a family of (adults both age 30) with two young children earning 142% FPL likely qualifies for around $1,323/month in subsidies, which would cover 93% of a Silver plan's cost.

The point is that every kid who’s shifted from CHIP to a subsidized ACA exchange plan would cost the federal government nearly the same as if they just funded CHIP in the first place...and I believe CHIP is a far more comprehensive program anyway, so it would cost the government around the same amount to provide worse coverage to the child.

The actual CBO projection from last fall was apparently a net deficit of $8.5 billion over 5 years, or around $1.7 billion per year, I'm guessing this means that the average subsidy provided to the "CHIP-to-ACA" is actually slightly lower than $2,300/year; the point is that funding the CHIP program wouldn't cost much more than not funding it.

HOWEVER, that projection was made before the GOP repealed the individual mandate. Since the CBO projects that the mandate repeal will jack up ACA market premiums by an extra 10%, that means that APTC subsidies are also expected to go up by a similar amount...which means that not funding CHIP would cost the feds more yet...and by definition, funding CHIP would cost the government even less relative to not doing so...$7.5 billion over 5 years, to be precise, according to the CBO. That's around $1.5 billion per year.

• The agencies estimate that, without the penalties related to the individual mandate, a larger share of parents would be uninsured. When funding for CHIP is reduced under current law, some parents would seek to preserve coverage for their children and enroll them in a family policy through the marketplaces. Some of those parents would otherwise be insured and could add their children to their coverage, while other parents would be uninsured and enroll both themselves and their children, resulting in higher costs to the federal government relative to insured parents. In the revised estimate, more parents would fall into the latter category. Thus, funding CHIP for five additional years would result in a larger reduction in spending related to the marketplaces than in the prior estimate.

This is similar to the above, but refers to the parents who might remain uninsured as long as their kids are covered by CHIP...but who the CBO thinks would jump onto the exchanges so that their kids could be covered along with them on a family plan if CHIP isn't funded. I think the idea here is that ACA policies have to have at least one adult enrolled, so if you want your kid covered, you have to be on it as well--which, again, means higher APTC subsidies. Ironically, I think the CBO is actually suggesting that funding CHIP could actually result in some adults not being covered since they might skip coverage for themselves as long as their kids are in CHIP.

• The agencies now incorporate into their estimate final regulations related to how premiums differ by age that increase premiums for children’s coverage through the marketplaces. Those regulations cause federal spending per child in the marketplaces to be higher. Thus, funding CHIP for five additional years—reducing the number of children covered through the marketplaces—would result in a larger reduction in spending related to the marketplaces than in the prior estimate.

I think this has to do with that in ACA policies (starting this year), children's premiums are the same from birth through 14 years old, but then go up a bit each year from 15 and up, whereas under CHIP (I think) the funding is the same for each kid regardless of their age. I could be wrong about this one, however.

Anyway, the point is that the CBO isn't saying that CHIP only costs $27 per child ($800 million / 5 years / 5.9 million); it's saying that doing so would only cost $160 million per year more than not doing so, while also providing far more comprehensive coverage to those enrolled and ensuring that none of the 9 million (including partial-year enrollees) are lost in the cracks.

UPDATE: I changed the title based on this excellent summary explainer from Andy Slavitt:

CHIP got cheaper because the alternative got more expensive when the individual mandate was repealed. 2/

— Andy Slavitt (@ASlavitt) January 6, 2018

It took me 1,300 words to say what it took him just 14 words to. Amazing.