The Associated Press asked 47 states and the District of Columbia how many children were waiting for childcare assistance. Twenty-three states and DC said they were keeping a list. Three more said they were turning eligible applicants away without keeping one.
In 2023 the number of states with waiting lists was eight.
Indiana had nearly 37,000 children on its list this spring. Minnesota had about 9,000. Marin County, California, had 380. These are children whose families meet the eligibility criteria — low income, homelessness, foster care — for a benefit funded by the federal government and administered by the states.
Eligible and unserved are not a contradiction
This is the part that confuses people encountering the programme for the first time, and it is worth stating precisely, because the confusion is doing political work.
Childcare assistance is not an entitlement. An entitlement pays everyone who qualifies, and its cost is whatever the number of qualifying people multiplied by the benefit turns out to be — Medicaid works this way, and so does Social Security. A capped programme has a fixed appropriation, and when the appropriation runs out it stops enrolling regardless of how many people qualify.
Under a cap, eligibility does not entitle you to the benefit. It entitles you to be considered for it, which in practice means a position in a queue whose movement depends on somebody ahead of you leaving.
Nobody is denied. That is the point of the design and also its cruelty: a denial can be appealed, and a waiting list cannot.
What changed was the money, not the need
The lists mostly began after autumn 2024, when $28bn of pandemic childcare aid expired and Congress declined to continue it.
That sequence matters for reading the numbers. The count of waiting children did not rise because more families became poor or because more parents went to work. It rose because a temporary appropriation ended on the date it had always been scheduled to end, and the states that had used it to enrol families now had more enrolled families than money.
This is the same mechanism this desk described in local government yesterday, where three quite different problems are all being called a budget gap. A one-time transfer spent on recurring costs produces a cliff, and the cliff is fully predictable at the moment of the spending decision. The alternative — decline the money, or spend it only on one-time things — would have meant not enrolling the families in the first place. Neither option was good. Only one of them produces a waiting list.
The list is a measurement artefact
Here is the trap in this data, and the AP's own framing is careful about it.
A state with a long waiting list looks worse than a state with none. But three of the states in this survey have no list because they turn applicants away at the door, and the AP could not get data from three states at all. A waiting list is an administrative choice to record unmet demand. Not keeping one does not reduce the demand; it removes it from the record.
So the honest reading of "23 states and DC" is a floor, not a total, and the honest reading of any individual state's number is that it measures both its shortfall and its willingness to write the shortfall down. Indiana's 37,000 is a large number partly because Indiana counts.
That distinction recurs across public administration and it is almost always read backwards. The jurisdictions that publish queue data — permitting backlogs, interconnection queues, court delays — get criticised for the queues, while the ones that quietly fail to serve people look clean.
The cost that does not appear anywhere
A parent who cannot get childcare and cannot afford it does one of a small number of things: leaves work, leaves education, reduces hours, or arranges informal care of uncertain quality. The AP found parents forgoing necessities, falling behind on bills, relying on acquaintances, and bringing children to work or class.
None of that is captured as a cost of the programme's cap. The appropriation is what it is; the foregone earnings, foregone training and reduced labour force participation land on households and on state tax bases, spread thinly enough that no line item records them.
Ruth Friedman, who ran the federal Office of Child Care under President Biden, located the origin of the funding structure in an older assumption: that childcare was the family's responsibility. A capped programme is what that assumption looks like once it has been partly abandoned but not replaced — public money, offered to a fraction of the people the criteria describe.
The number that would settle it
Not the waiting list. Publish the share of eligible children actually served, state by state, and publish it annually.
Every state can compute it, because both terms are known: the eligibility criteria define the denominator and enrolment is the numerator. Almost none of them present it, and the reason is that the resulting fraction is the clearest possible statement of how much of a stated commitment is being met.
A waiting list tells you how many people a state was willing to count. A served-share tells you what the promise is actually worth.
The count of 23 states and the District of Columbia with waiting lists as of spring 2026, the comparison with eight states in 2023, the three additional states turning eligible applicants away, the figures of nearly 37,000 children in Indiana, about 9,000 in Minnesota and 380 in Marin County, California, the expiry of $28bn in pandemic childcare aid in autumn 2024 and Congress's decision not to extend it, the collection of data from 47 states and DC, and the remarks of Ruth Friedman, who headed the Office of Child Care under President Biden, are from an Associated Press investigation published on 3 September 2026. The analysis is our own.
Topics nationalstate and localpublic financefamilies
