A woman I helped last spring lost her coverage in April and had it back in July. She was eligible in April. She was eligible in May and June, when she did not have it. Nothing about her situation changed at any point in those four months.
What happened is that a renewal notice was mailed to an address she had moved out of in February, and the form inside it was due in eighteen days.
There is an enormous, careful, genuinely thoughtful literature about who should be eligible for public benefits in America — income thresholds, asset tests, work requirements, cliff effects, the whole architecture. Legislatures fight about it for years. I have read a lot of it and some of it is excellent.
I have come to think that almost none of it describes what actually determines who has coverage, and that the thing which does is the renewal interval.
Set the interval at twelve months and a certain caseload results. Set it at six and the caseload falls, by a lot, and not because anybody became ineligible. It falls because every renewal is a filter that catches people on the basis of whether a piece of paper found them — whether they moved, whether the mail forwarded, whether the eighteen days landed during a hospital stay or a double shift or a bad week, whether the online portal worked on a phone, whether they had the pay stubs from the right two months.
None of that is in the eligibility rules. All of it determines eligibility in practice, and it does so with a precision that no legislature could have written down.
I want to name what this filter selects for, because it is not random and calling it administrative churn makes it sound like weather. It selects against people who move often, which means renters. It selects against irregular income, because irregular income is harder to document than a salary. It selects against people caring for someone else. It selects against anyone whose life is already absorbing more disruption than it has capacity for — which is to say, it works exactly hardest on the people whose circumstances the programme was written to describe.
The defence of shorter intervals is real and I have written it myself, in a memo, for people who were paying me to write it. Circumstances change, people do become ineligible, and a programme that only checks once a year pays benefits to some people who no longer qualify. That is a genuine cost and it is measurable, and the people making that argument are not inventing it.
But the comparison, as it is usually run, prices one side and not the other. The savings from a shorter interval appear on a budget line. The cost — people who remain eligible, lose coverage, go without care for three months, and come back through an intake process that costs the state more than the continuous enrolment would have — appears on nobody's line at all, and the hours the applicant spends are priced at zero because there is no ledger they could appear on.
There is a version of this argument that says the churn is the intent, and I do not make it, because I do not have to. It does not matter whether anyone designed it. The mechanism works identically either way, and it is currently the single largest determinant of who is covered in this country, operating entirely outside the debate that everyone thinks is about who is covered.
She has coverage now. The gap cost her one appointment she rescheduled and one she did not, and the file at the office is closed, and in October there will be another form.
