Making Sense of CBO's Health-Care Score

Making Sense of CBO's Health-Care Score
AP Photo/Carolyn Kaster

In its latest assessment of congressional proposals to reform health care, the Congressional Budget Office (CBO) addressed the version of the American Health Care Act (AHCA) of 2017 passed by the House of Representatives on May 4th. As with CBO’s March report, which looked at a previous version of the legislation written by several House committees, a single number became the headline.  

The headline-grabbing number stems from CBO’s estimate for how many fewer people would have health insurance if the proposal became the law of the land. In March, the number was 24 million according to The Washington Post, The New York Times, and CNN. This time, the same outlets report CBO’s 23 million.

Whether 23 or 24 million, the CBO number is a tricky one to understand.

The CBO estimate does not mean that many fewer people will have health insurance than those who have it in 2017. Instead, it is a measure of how many fewer people would have health insurance in 2026 than would have it if current law remains unchanged. This is not the same thing as saying that 23 or 24 million people would lose coverage — though explaining such conditional counterfactuals doesn’t make for catchy headlines. As I’ve noted in these pages, some of those included in the 23 million figure do not even have Medicaid yet. These “not yets” live in states the CBO projects would have expanded Medicaid eligibility in the future (though now won’t.)

Medicaid enrollees — whether those currently covered under the program or those likely to obtain it over the next decade — account for more than half of those in the CBO’s 23 million number. CBO’s analytic effort in this latest estimate focuses on what will take place in the private health-insurance market. On the Medicaid side, the CBO changed its assessment of what will happen this year and next, but made no change to its view on Medicaid enrollment in 2020 and beyond. In both their March and May estimates, the CBO says 14 million fewer people will have Medicaid in 2026.

This estimate gives no weight to incentives, present in the House-passed bill but not in the March version, created by per capita spending targets. The bill passed by the House sets limits on how much the federal government will share in costs per enrollee. These limits apply to each of several types of people who enroll in Medicaid, including the elderly, blind and disabled, children, and “expansion enrollees,” i.e., those who have become eligible to enroll thanks to the ACA.

For the first time, states will be at risk if the average cost per person exceeds a cap. The average gets calculated by dividing the total costs by the number of enrollees. Both numbers influence the average. If everyone had the same health-care costs, it wouldn’t matter. But the rationale for health insurance is the large variation in costs.         

Health-care costs have long been incredibly concentrated. The 5 percent of the population with the highest costs accounts for one half of total health-care expenditures. And the 50 percent of the population with the lowest expenses accounts for less than three percent of expenses.

Not all of the variation is random, as with the kind of risk that might be embodied in taking a fall and breaking one’s leg. Some people or groups have higher costs year after year. The idea of “adverse selection” is all about how much more likely it is for those who expect high need for medical services to enroll than those who expect low health-care use.

The Affordable Care Act gave permission for states to expand Medicaid eligibility with the federal government taking on a larger share of the expenses for those newly eligible. The AHCA proposes taking away the higher federal share. Previously, the CBO predicted that some states would respond by trimming eligibility. But the new per capita caps add a wrinkle.

If the lowest income people have higher costs, dropping those in the next higher income group would have an impact on average costs. When the denominator — the number of people enrolled — shrinks faster than the numerator — total costs— it causes total costs to go up. And if costs go up faster than the federal caps allow, states will be on the hook for 100 percent of those costs.

There is good reason to believe those with the lowest income have higher costs. A trip to the hospital means not being able to work during that period. Hospitals have a strong interest in seeing that everyone who gets hospital care and who is eligible for Medicaid applies for it. The same states that the CBO predicted would save money by moving to a less generous eligibility level will now have to balance two factors: the impact on the number enrolled and the impact on average costs.

It’s rational to think that states will behave differently than they would under the March version of the AHCA. But that behavior shift is not recognized in CBO’s most recent analysis.

Hanns Kuttner is a senior fellow at Hudson Institute.

Show commentsHide Comments

Related Articles