Auto-enrollment trade-offs

Sarah Kliff at Vox highlights one area of plausible discussion among liberals and some conservative Senators on health care. She looks at the idea of automatic enrollment with opt-outs from a catastrophic plan. This idea is part of Cassidy-Collins state option.

Republican legislators and policy experts are kicking around a novel way to increase health coverage: automatically enrolling millions of uninsured Americans into low-cost insurance plans….

And unlike Republicans’ other ideas, automatic enrollment is the rare health proposal that doesn’t reflexively alienate liberals. They are generally enthusiastic about policies that would lead to greater coverage.

“It’s a viable idea,” says Andy Slavitt, who ran Medicare under President Obama and is an ardent Affordable Care Act advocate. “What’s appealing about it to Republicans and to Democrats is you want people to have free choice but not be free riders.”

There is a major operational challenge of assignment. I don’t think it is as big of a deal as others make it out to be as I conceptualize it as effectively similar to Medicaid presumptive eligibility with retroactive payments. A region could be set up and insurers could bid on providing coverage to effectively the uninsured for an estimated pot of money. We don’t need an ugly database tracking enrollment.

But that is a detail.

The key thing is to look at the trade-offs between an opt-in and out-out auto-enrollment program.

The ACA is an opt-in program. People have to sign up for an Exchange plan, they have to sign up for Medicaid. Not everyone signs up. Auto-enrollment basically has everyone sign up.

We need to hold money constant for a minute to see the implications. The first pass will not be an ACA vs AHCA analysis. It will be an illustration. We’ll get a little more complicated in the second iteration of analysis.

The first scenario is for 1,000 people to be covered. The total healthcare cost for this group averages $5,000 per person per year. The total subsidy budget is $2,000 per person. These numbers are not too important but the intuition illustrated by these numbers will be.

Scenario 1-A: Opt-in
500 people sign up. The $2,000,000 subsidy budget means each person who signed up receives a $4,000 subsidy. This buys them on average an 80% actuarial value(AV) plan. The people who signed up are happy. The people who did not sign up are not happy as they have 0% actuarial value coverage.

Scenario 1-B: Auto-enroll with opt-out
990 people do not opt-out. The average subsidy is $2,020. This buys a baseline policy of 40.4% (AV). The people who previously had 80% AV plans are mad. The people who did not sign up in 1-A are happy if they did not opt-out.

This is the fundamental trade-off. An opt-in system for a given level of funding and holding everything else constant, will deliver higher benefits to the people who opt-in than a near-universal system.

Scenario 2: Things get complicated

We can’t quite hold everything constant. Expected and actual costs are not uniformly distributed. Some people are very expensive to cover in a year. These very expensive people are further split into people who have a one-off non-recurring event and people who have recurring events. Some people barely touch the medical system in a year. In an opt-in system, we can make a strong assumption that the people who sign up are motivated to sign up. This is part of the logic of the new rule to change open enrollment so that it ends on December 15th so healthy enrollees are on tap to pay for twelve months of coverage instead of ten.

With this assumption that an opt-in system encourages some selection, the $2000 subsidy won’t buy an 80% pooled AV plan. It might buy a 65% or 70% AV plan under favorable assumptions on cost distribution. Less favorable cost distributions bring down the pooled AV to someplace in the high 40s or low 50s if we assume significant foresight on the buyers. Adding everyone else to the same pool with the same level of subsidy increases the out of pocket maximums but at a slower rate than in a universe of random distribution of buyers and healthcare costs.

This is a key intuition to keep with automatic enrollment. Any capped subsidy level will lead to a lower actuarial value for the same amount of public spending than an opt-in system.

And as a matter of explicit policy preferences that can be a defensible choice, but let’s just be aware of that facet of this choice.