Ordeals and administrative burdens

In a recent article by Betsy Cliff and colleagues, they analyze the impact of premiums on Michigan Medicaid Expansion enrollment. Their NEJM commentary gives a good summary of the programmatic details:

Michigan was among the first states to institute premiums for people covered by Medicaid expansion….
It includes copayments (generally between $2 and $8) for certain medical services for all enrollees and premiums for those with incomes between 100% and 133% of the FPL. Premiums depend on household size and are capped at 2% of household income; for individual enrollees, premiums range from $21 to $29 per month…
Michigan enrollees can cut their premium amounts by half by agreeing to adopt or making progress toward adopting certain healthy behaviors, such as undergoing preventive screenings or quitting smoking…
Among people eligible for a premium invoice, we estimated a 2.3 percentage-point increase in disenrollment during the 6 months after an enrollee could first be invoiced. This change represents a 12% relative increase as compared with the program’s baseline attrition rate.5 When we examined trends according to enrollee health status, the increase in premium-induced disenrollment was evident among people with either below-median baseline medical spending or no diagnosed chronic diseases, but not among those with above-median baseline spending or at least one chronic disease

Michigan Medicaid Expansion has a feature where people can be billed for premium but non-payment of premium does not result in termination of coverage. This is different than the ACA market where non-payment of premium is a routine source of termination of coverage. So there is a significant decrease in enrollment due to the arrival of a bill that is elective to pay.

There are at least two ways to think about this.

One is to think that the premium represents an ordeal. Individuals who are highly motivated to maintain coverage will jump through the necessary hoops and climb over the barriers to keep their coverage. Individuals who are not motivated because they do not highly value their coverage won’t do so. In this framing, an ordeal is a rational way to allocate resources to individuals with a high willingness to pay. The people who most need/value a service will get it and the people who are basically shrugging their shoulders walk away.

The other way of looking at this is that this is an administrative burden. Coleman Drake, Sih-Ting Cai, Dan Sacks and I have a working paper that looks at the impact on enrollment in the Colorado ACA market when when people move from zero premium to very low (cup of coffee) monthly premiums. We see a massive drop in effectuated enrollment for January as people who have low dollar premiums make mistakes setting up payments. Many people correct most of these mistakes in February. Our findings imply that people get tripped up even if and when they want insurance. Adrianna McIntyre and her team at Harvard look at the impact of people re-enrolling when they go from zero to not zero in premiums. They found that moving people out of zero premium plans led to large enrollment losses precisely at the moment when people could be terminated for non-payment of premium.

On Twitter, one of the authors makes the following point that I’m struggling with:

This work—along with evidence from Amy Finkelstein, @nhendren82 Mark Shepard, Pietro Tebaldi, @Prof_HeidiAllen @LauraDague and others—shows pretty clearly that low-income, healthy individuals will decline health insurance without heavily subsidized premiums.

— Betsy Querna Cliff (@betsyqcliff) June 20, 2022

The commentary notes that the annual premiums (~<$360) for people who disenroll is massively less than the look-back period annual medical expenses ($2,000). The people who are disenrolling have significant medical need that is five or six times higher than their premiums. This to me suggests the ordeal interpretation is doing a very bad job of screening out people with very low needs for the program. The net cost of medical care for these individuals even with premiums is still massively subsidized. I think that the growing evidence base of the effect of moving people from $0 to cup of coffee premium levels suggests more of a burden mechanism rather than a targeting mechanism. Both are at play, as people who have metastatic cancer are likely to read every line on a renewal letter to jump through hoops compared to someone who saw their PCP once, and picked up a cheap antibiotic at CVS to take care of a case of pink-eye and otherwise does not routinely interact with the medical system. But the people who are dropping still are dropping coverage that has massive subsidies. And these drops are happening in multiple lines of business/coverage and in multiple states. I think getting a better understanding of the tensions between thinking about nominal premiums as ordeals and nominal premiums as administrative burden will be crucial in the policy process.