Medicaid unwinding ex-parte renewals

Ex parte renewal is the automated step that state Medicaid agencies run to renew coverage without asking the enrollee for anything. It matches the enrollee's records against wage, tax, SNAP food assistance, unemployment, and Social Security data. In 30 states, it evaluated renewals by household instead of by person, the unit federal rules require.

How it is used

A pandemic-era protection kept people enrolled in Medicaid from March 2020. It ended on 31 March 2023. States then had to redetermine eligibility for their entire Medicaid and CHIP population over a roughly year-long unwinding. CHIP is the Children's Health Insurance Program.

Federal rules require states to try ex parte renewal first, before asking for a paper form. Done correctly, it protects enrollees, because it renews them without any paperwork.

KFF, a health policy research organization, reported that about 55 percent of retained coverage was renewed ex parte as of October 2023. That rose to 61 percent of renewals by September 2024. States ranged from about 99 percent in North Carolina to about 3 percent in Wyoming.

What went wrong

On 30 August 2023, the Centers for Medicare & Medicaid Services (CMS), the federal agency over Medicaid, notified states of a defect. Their systems ran ex parte renewals at the household or family level. Federal rules require the individual level.

This setting is called the unit of determination. When any one household member could not be renewed automatically, the whole household was sent a form. If the state did not get the form back, it ended coverage for everyone in the household on procedural grounds. That means over paperwork, not a finding that anyone was ineligible. The sources do not say why forms were not returned.

The case file says the system was accurate in the ordinary sense. It did exactly what it was configured to do. The error was in the specification, so no accuracy measure would have surfaced it.

Who was harmed

Children were hit hardest. Their income limits for eligibility are far higher than adults', so a child often stayed eligible when a parent did not.

The Office of the Assistant Secretary for Planning and Evaluation (ASPE), in the federal health department, projected that about 74 percent of disenrolled children would still be eligible. That is a projection, not an audit afterward.

How it was caught

CMS found the defect in 30 states. It ordered them to pause procedural disenrollments, reinstate coverage, adopt mitigation CMS approved, and reprogram to person-by-person renewal.

On 21 September 2023, CMS announced that nearly 500,000 children and other people would regain coverage. That figure adds up estimates the states reported to CMS. It is not an independently audited count.

The case file calls one move the most effective: a halt authorized in advance. CMS ordered states to pause procedural disenrollments for affected people.

Who decides

The states run the renewal systems and reprogrammed them. CMS oversees them. Section 5131 of the Consolidated Appropriations Act, 2023 gave CMS mandatory monthly state reporting and the power to order procedural disenrollments paused.

For states that did not comply, it added a cut to federal Medicaid funding, civil monetary penalties, and corrective action plans. An interim final rule of 6 December 2023 put these powers into regulation.

What the oversight did and did not do

The federal loop worked. Its monthly reporting and its power to pause terminations caught the defect and forced reinstatement across 30 states within months.

But it caught the defect after mass terminations, not before. Its check ran on totals the states reported each month, not on each decision when it was made.

The case file contrasts this with Michigan's MiDAS, which issued tens of thousands of false fraud findings. It also contrasts Australia's Robodebt, which raised hundreds of thousands of debts by an unlawful averaging method. There, corrections came years later from courts and a Royal Commission.

The wider unwinding

The defect was one part of a much larger loss of coverage. KFF recorded about 25.2 million people disenrolled as of 12 September 2024. Of those disenrollments, 69 percent were for procedural reasons.

The Government Accountability Office (GAO) found about 27 million disenrolled in the first 18 months. The two totals cover different periods and use different methods, so they do not conflict.

Some states failed the opposite way. They used ex parte renewal too little, which shifted the work onto paper forms and state processing backlogs, both prone to error.

What this case asks

The case file says the useful governance moves here are not accuracy fixes. The first is an independent person-by-person check of the household-level result.

The second is a check of each termination against the person's own eligibility before disenrollment. The third is a faster review rhythm, so a population-scale mistake is caught while terminations are still happening.

The case file also credits the monthly reports, with their counts and reasons for disenrollment, as the honest data that made the catch possible.

What the available tools can and cannot address

A failure pathway is a link between two parts of the network where a mistake made by one part can be passed on to the other. A pathway counts as closed once it passes on only a few mistakes. It need not stop them all.

This case's budget is 12 units. Each tool costs units from it.

Explore (No Targets) sets no targets. Under Service Targets Only, the targets can be met. The cheapest way is one tool, Vet connections, costing 3 units.

Vet connections lets one system retrieve or copy another's records only where that is explicitly granted. On this network, mistakes stop passing along three pathways. They are Cross-agency data matched for renewal, Prior enrollment records reused in matching, and Termination applied to coverage and managed care. It also limits how readily one mistake repeats along the pathway named One rule set for every renewal.

The first of those is the data match federal rules require. Closing it means mistakes stop passing along it, not that the match stops.

Under Service and Safety Targets, the targets include closing every failure pathway, and they can be met. The cheapest combination uses Vet connections, Understand the system, and Store less data.

At this level, Understand the system costs 4 and cuts Store less data from 3 to 2. The three cost 9 of the 12 units. Every combination that meets these targets includes all three.

Understand the system funds study of what the deployment is doing, and caseworkers correct more mistakes. Store less data limits what the renewal system and caseworkers write into the coverage record.

None of them uses Check with a second model or Check copied records. Those are the two tools that add the person-by-person checks this case is about.

Under All Governance Targets, this case is not fully addressable with the available tools. That level also asks for a larger, steadier gain in service, meaning the benefit the deployment delivers. Every combination within the budget was checked.

The best one that closes every failure pathway and meets the target on the deployment's environmental footprint falls just short of that service target.