The ARPA cliff: what happens now that the emergency HCBS money is gone
States planned to spend an estimated $37.1 billion on home and community based services, most of it on the workforce. It was one-time money spent on a permanent problem, and the deadline has arrived.
Bonuses, retention payments, a wage add-on, a new program, extra training staff, a scheduling coordinator. Providers know exactly which line items came from Section 9817 because they had to report it. Ask, in writing, what was funded that way and what happens to it now. The answer tells you which parts of your adult child’s current support are structurally at risk this year, which is a very different question from whether the provider is doing a good job.
What it was
Section 9817 of the American Rescue Plan Act, a temporary enhanced federal medical assistance percentage for home and community based services.
Where it went
Workforce recruitment and retention was the largest planned spending category nationally at $26.3 billion, with $3.9 billion more for workforce training.
$37.1B
What it did not include
MACPAC found the legislation lacked evaluation requirements. States spent billions with no obligation to measure what worked, which is now the exact reason it is hard to argue for renewal.
24
The extensions
CMS granted 24 states additional time beyond March 31, 2025, with the longest approval running through September 30, 2026. After that there is no further runway.
The sustainability gap
About two thirds of states, 33 of them, described how they would maintain some initiatives after ARPA. Only about a third of direct care workforce activities were planned for continuation.
The double hit
The expiration lands in the same period as federal Medicaid reductions and new eligibility requirements. Providers are absorbing both at once.
What MACPAC found, January 2026
All 50 and the District of Columbia
An estimated $37.1 billion
$26.3 billion, the largest category
$3.9 billion
March 31, 2024
March 31, 2025
24 states, longest to September 30, 2026
33, about two thirds
Devon’s provider handed out retention bonuses two years running. Fifteen hundred dollars, twice, to the direct support professionals who stayed. Two of them stayed. Then the money ended and the bonuses ended and both of them left within five months of each other. Devon’s mother asked what changed. Nothing changed. That was the problem.
Find out what your family is standing on
Questions to ask now
- 1Was any part of this service funded with ARPA money?
- 2Which specific line items, and for how long?
- 3What ends when the funding ends?
- 4What is the plan to sustain it, in writing?
- 5Has staffing changed since the funding wound down?
- 6What is the current vacancy rate at this agency?
What to document this year
- Every service that has been reduced since 2025
- Every retention bonus or add-on that stopped
- Staff turnover at your provider, by month
- Authorized hours against delivered hours
- Any program that closed or shortened its days
- What your provider told you the reason was
- Dates of every notice you received
- What you had to cover yourself, in hours
One-time money spent on a permanent problem buys exactly as long as the money lasts. Knowing which parts of your support were built on it is not pessimism, it is planning.
Somewhere around 2022 a lot of families noticed things briefly got better. A retention bonus showed up in a direct support professional’s check. A program hired a second scheduler. A provider raised its posted wage and actually filled the shift. Nobody explained where that came from, and almost nobody explained that it had an expiration date printed on it.
What Section 9817 actually was
The American Rescue Plan Act, passed in 2021, included Section 9817, which gave states a temporary increase in the federal share of Medicaid spending on home and community based services. States had to reinvest the freed-up money in HCBS, submit spending plans, and report on what they did.
Every state took it. All fifty and the District of Columbia. Across the country, states planned to spend an estimated $37.1 billion in combined state and federal funds on HCBS reinvestment.
The largest single category was workforce recruitment and retention at $26.3 billion nationally, with another $3.9 billion for workforce training. Which is to say, states looked at the money and correctly identified that the binding constraint on home and community based services is that nobody will do the work at the wage.
They were right about the diagnosis. The problem was the prescription had a shelf life.
The structural mistake, said plainly
This was one-time money spent on a permanent problem.
A retention bonus is one-time money. A signing bonus is one-time money. A temporary wage add-on funded by a temporary federal match is one-time money wearing a wage’s clothes.
None of those things change what a job pays. They change what a job pays this year. And a direct support professional deciding between your provider and a warehouse is not comparing this year. They are comparing the next five.
I do not think states were foolish to do it. Given a two-year window and a workforce hemorrhaging people, paying retention bonuses was a defensible emergency response. The failure was not in the spending. It was in the absence of any plan to convert temporary money into permanent structure while the temporary money was still flowing.
The evaluation problem, which is worse than it sounds
MACPAC published an issue brief in January 2026 examining how Section 9817 was implemented. The finding that should worry every advocate is this: the legislation lacked evaluation requirements.
States spent tens of billions of dollars with no obligation to measure whether it worked.
Think about what that means for the next fight. When somebody proposes a permanent HCBS workforce investment, the first question in the hearing room will be what we got for the last one. And the honest answer, for most states, is that nobody was required to find out.
That is not a small procedural gap. It is the reason a historic investment produced no durable evidence base, and it is why the argument for renewal is harder now than it was in 2021 despite four years of spending.
If you ever wondered why advocates get so insistent about data collection requirements in legislation, this is why. Money without measurement funds one cycle. Money with measurement funds an argument.
What happens now
The original spending deadline was March 31, 2024, extended to March 31, 2025. CMS granted 24 states additional extensions, with the longest approval running through September 30, 2026.
So for a minority of states, some of this money is still moving right now, and will not be by the end of the year. For most, it has already stopped.
The expiration lands in the same window as federal Medicaid reductions and new eligibility requirements taking effect. Providers are absorbing a funding cliff and a compliance burden simultaneously, which is how you get the closure letters described elsewhere on this site.
What that means for your family is concrete. Anything in your adult child’s support that was funded by ARPA and not converted into a permanent rate is at risk this year. Not eventually. This year.
Your practical next steps
This week. Email your provider and ask directly which parts of your adult child’s service were funded through ARPA Section 9817, and what happens to those parts now. Providers had to report this, so the answer exists.
Ongoing. Document every reduction since 2025 with dates. Retention bonuses that stopped, staff who left, program days that shortened, hours authorized that went undelivered. This is the evaluation nobody required, and you are the only person in a position to conduct it for your own family.
When you advocate. Make the permanence argument, not the amount argument. More temporary money produces another cliff in three years. Ask specifically for wage floors in statute, rate methodologies with inflation adjustment built in, and reporting requirements attached to any new investment.
The emergency money is gone. The emergency is not. That gap is the entire policy conversation for the next several years, and families who can document what they lost when the money stopped are the only people who can prove it existed.