With the January 1, 2027 enforcement date approaching and the member-notice window expected to run roughly June 30 to August 31, 2026, plan leaders need a concrete number for one question: how much of our revenue is actually exposed? Modeling at-risk revenue is not guesswork if you segment the membership the right way.
Segment before you forecast
Start by dividing the membership into three buckets. The first is members not subject to the requirements at all, by age, eligibility category, or other statutory exclusion. The second is members subject to the requirements but likely exempt, such as caregivers, pregnant members, and those with qualifying disabilities. The third is members who are subject and must actively report compliance. Only the second and third buckets carry procedural risk, and they carry very different kinds of it.
For the likely-exempt bucket, the risk is documentation failure: the exemption exists but is not recorded in time. For the must-report bucket, the risk is reporting failure: the member qualifies through work, study, or volunteering but cannot complete the attestation. Quantifying each separately produces a far more defensible forecast than a single blended churn rate.
Attaching dollars to each segment
Once segmented, multiply each at-risk segment by its expected procedural-loss rate and the PMPM value of those members. The Arkansas precedent, where roughly one in four affected enrollees lost coverage, offers an upper-bound scenario for the must-report segment if no engagement infrastructure is in place. A well-run plan should model a base case well below that, but planning against the precedent rather than wishful thinking is what makes the number credible to a board.
The output should be a range, not a point estimate: a worst case assuming minimal intervention, a base case assuming standard outreach, and a best case assuming proactive exemption capture and multilingual, deadline-aware contact. The spread between worst and best case is, in effect, the value at stake in the engagement decision, and it is usually large enough to dwarf the cost of acting.
Two data hygiene points determine whether the model is trustworthy. First, language and contact-reachability data must be current; a member you cannot reach in a language they understand is functionally at higher risk regardless of eligibility. Second, the model should be refreshed as the 2026 notice window unfolds, because real reporting and exemption-capture rates during that window are the best leading indicator of what 2027 enforcement will actually cost. Modeled early and updated often, at-risk revenue becomes a managed exposure rather than a year-end surprise.