Overview

We examined the impact of technical and policy expertise provided to states by the U.S. Facing Financial Shocks Team on Medicaid renewal rates during the Pandemic Health Unwinding. In models of outcomes by state over unwinding months, we estimate that the interventions raised ex parte rates by 21 points, overall renewals by 8 points, and reduced procedural denials by 9 points. We test modeling assumptions and corroborate the shift in outcomes using an event history framework.

Findings

Procedural terminations, which account for account for over half of Medicaid terminations, are driven by implementation

From April 2023 to August 2024, roughly 24 million Medicaid beneficiaries lost coverage during the 'unwinding', or the ending of the pandemic rule that disallowed states from disenrolling beneficiaries (KFF, 2024). About 70 percent of those who lost coverage, however, did so for 'procedural' reasons, such as missing paperwork. While some are likely not eligible, these coverage losses are mostly among those still eligible and needing coverage. 

While those numbers are grim, it was striking that as the unwinding progressed, a smaller and smaller share of enrollees lost coverage each month. Why did outcomes seem to get better over time? It turns out that states were all learning about how to improve their renewal processes during unwinding. We studied a concerted effort to make these kinds of improvements, led by a Facing Financial Shocks Team made up of USDS and CMCS staff. Over the course of a single month of going live with implementation improvements in the four states for which data are available (CA, NY, SC, WI), the team replicated the impact of about one year’s improvements elsewhere in the country.

Consistent outcome metrics were critical for driving priorities and identifying what works

The federal government pushed for close attention to unwinding outcomes through a common format and definitions for reporting outcomes, a centralized website where all state data would be available, and the expansion of policy waivers that states could use to improve ex parte and paper (manual) renewal rates. 

The FFS team viewed high ex parte rates in states like Oregon and Washington as inspiration for what was possible and borrowed best practices from those states.

States often lack the capacity to review and quickly modify systems that they outsource, floating technical + policy expertise can make a big difference

Many states wanted to expand their ex-parte renewals, but struggled with a lack of expertise to do so. By offering up teams of consultants with policy and technical expertise to assist states, the federal government also played an active role in driving some of the implementation of improvements to renewal processes.

The teams found that they could not only assist in comprehensive use of policy options, such as auto-renewals for people who had been found to be at 100% of the Federal Poverty Line or lower, they could spot errors in code that excluded people from auto-renewals, and they also could restructure code to systematically examine opportunities for auto-renewal before having the system send an enrollee a renewal form.

Approach

For point estimates of changes that accounted for trends and state differences, we modeled monthly, state-level renewal percentage rates using a beta regression with state fixed effects, controls for reported variations in state roll-outs, and an indicator for months after which interventions went live in states, which occurred at different points for different states. We used an event history framework to test our modeling assumptions and for statistical plots. 

Data

• The key data for our estimates came from the Medicaid PHE unwinding website, where we relied on the updated data files because some states had a large share of pending cases in the data that had not yet been updated. 

• State plans for the timing of when they would handle less likely eligible cases were submitted to CMS and released in mid 2023.

• Month-by-month requests to pause procedural denials are reflected in this CMS slide from June 2024. 

• We captured the end of unwinding for states (when they would begin to be omitted from analysis) based on this May 2024 slide from CMS.

Funders: Gates Foundation, Walmart Foundation