On October 15, the Obama administration significantly downgraded its estimate of how many people will enroll in exchange plans next year. The administration now expects only 10 million exchange enrollees at the end of 2016. Charles Gaba, a statistical expert who closely tracks Affordable Care Act (ACA) enrollment and who made fairly accurate projections for 2014 and 2015, is somewhat more optimistic. He projects enrollment at 12.2 million people by the end of next year.
These low estimates should rock the health policy community. The Congressional Budget Office only four months ago projected there would be 20 million exchange enrollees next year. This CBO estimate was 3 million people fewer than its estimate issued just after the 2012 Supreme Court case that preserved the law and 1 million fewer than its March 2010 estimate, produced days before the law was passed by the House of Representatives.
CBO’s early projection of 2016 exchange enrollment was actually lower than predictions by other experts who modeled the ACA. The table below shows 2010 projections for 2016 from the RAND Corporation, the Office of the Chief Actuary at the Centers for Medicare and Medicaid Services, and the Urban Institute.
A major question for the health policy community is why the early consensus was so much more optimistic than today’s reality. While there are several factors that explain how the expert community likely erred so significantly, the most plausible explanation is that exchange plans are much less attractive than experts had projected.
Two Key Changes Affect 2010 Estimates of the Law
In June 2012 the Supreme Court ruled in National Federation of Independent Business v Sebelius that although the individual mandate penalty represented an unconstitutional exercise of Congressional power under the commerce clause, the mandate, if interpreted as a tax, was permissible because Congress has the power to tax people without insurance. This case also made the ACA’s Medicaid expansion optional for states. Prior to the ruling, states were compelled to expand Medicaid to everyone up to 138 percent of the federal poverty level (FPL) or face the loss of all federal support for their Medicaid programs.
As a result of this ruling, exchange enrollment increased since the law authorizes premium tax credits and cost sharing subsidies for people who earn income between 100 and 138 percent of the FPL and who are ineligible for Medicaid. Therefore, many people in non-Medicaid expansion states who earn between 100 and 138 percent of the FPL qualify for subsidized exchange plans. The ruling caused CBO to increase projected exchange enrollment by about 3 million people while reducing projected Medicaid enrollment.
While the Supreme Court ruling boosted exchange enrollment, a transitional policy announced by the administration in late 2013 dampened it. In response to public outcry from people losing their insurance in the fall of 2013, President Obama permitted state insurance commissioners to allow insurers to extend plans for a year. The administration further extended this policy through September 30, 2017. Thirty-five states are allowing these transitional policies to remain in force.
By allowing people to remain on their existing plans, the administration reduced the number of people shopping for exchange coverage. However, because of the transient nature of the individual market and that the transition policy only applied to people who selected a plan after the ACA was signed into law, the policy likely did not depress exchange enrollment as much as the 2012 Supreme Court decision enhanced it. Therefore, unforeseen policy changes are not the main reason why the experts got their projections of exchange enrollment so wrong.
Exchange Plans Unattractive to Young, Healthy, and Non-Poor
In searching for an explanation of why the experts appear to be so wrong, it is important to remember that incentives matter. The ACA’s core features were designed to make insurance cheaper for people who were older, relatively sick, and poorer. In doing so, it made insurance a much worse deal for the young, relatively healthy, and non-poor. It should not be shocking that these people are voting with their feet and largely shunning ACA coverage.
A recent study by Mark Pauly, Adam Leive, and Scott Harrington – three economists at the University of Pennsylvania’s Wharton School – quantified how bad a deal ACA plans represent for most people without insurance. Their estimates show that ACA coverage becomes a significantly worse option as people earn higher income all else equal.
Their paper shows that the ACA makes a typical person who is uninsured and earns about $40,000 worse off by about $2,000 to $3,000. The Wharton economists conclude that: “[t]he minority of high risks among the middle class uninsured may gain, but most uninsured will lose and, according to our estimates, will prefer to remain uninsured at the current penalty levels for violating the individual mandate.”
In March, the health care research firm Avalere published a graphic showing how enrollment in ACA plans drops significantly as income increases. The Pauly, Leive, and Harrington findings explain why exchange plans mostly enroll people below 200 percent of the FPL ($23,540 for a single person in 2015).
The recently released risk corridor data, which I wrote about last week, demonstrates that insurers enrolled a disproportionate number of older and sicker people in ACA plans in 2014. As a brief reminder, the risk corridors were intended to transfer money from insurers that made excess profits on ACA plans to insurers with excess losses on these plans. It turns out that a lot more insurers made excess losses. In 2014, insurers with excess profits owed $360 million and insurers with excess losses claimed $2.9 billion.
Using the risk corridor data and formula, I estimate that insurers selling ACA plans lost at least $4 billion last year, an amount equal to 12 percent of the premiums they collected on these plans. These losses came despite a nearly $8 billion back-end subsidy these insurers received to pay for high cost enrollees through the ACA’s reinsurance program.
The ACA is a large, complicated piece of legislation whose effect was difficult for policy experts and modelers to predict. It now seems extremely likely that many prominent policy experts and modelers have badly overestimated exchange enrollment, at least in the near term. Other factors that I will explore in the future are also at play, but the linchpin is that exchange plans are not nearly as desirable as the expert community, with the exception of a relatively small number of experts including several Apothecary contributors, thought they would be.
The magnitude of how wrong the initial predictions appear to be should cause legislators and the public to approach future estimates of the impact of major legislation cautiously. The fact that people find exchange plans so much less attractive than experts assumed when the law was passed will hopefully convince some supporters of the ACA that the law needs to be revisited, and likely fundamentally changed.