What happens when the money stops

How we counted

Every executive director has done the math on the grant that pays for a third of the staff. The honest answer to what happens when it ends has always been "it depends", because nobody could see enough organizations, for enough years, to say anything else. Since 2017 the IRS has published nearly every nonprofit's grants list and financials as data, year after year. Link the grant lines from funders to the filings of the organizations that received them and the same relationship can be followed through time, with the recipient's own accounts on the other side.

We looked for pairs of funder and grantee where the grant was at least 10% of the grantee's spending that year, for three or more years running, from a private foundation, a community foundation or a donor-advised fund sponsor. A hospital system paying its own physician group is not a grant ending, and the first pass was full of those. Then two groups: relationships that kept going at that level, and relationships where the funder stopped and did not come back for three years. And then the grantee, two years on: still filing a return? Spending down by a quarter, or by half?

The number

When a grant like that ends for good, 32% of organizations are not filing two years later or have cut spending by more than a quarter. When the same kind of grant continues, 11%. Gone or halved: 22% against 4%. Three years on it is 32% against 11%; the damage does not recover.

Not filing is a blunt measure, so we traced every organization that stopped. Of the 177, 89 left no trace at all. The other 88 show at least one sign of life, and often more than one: 74 were still receiving grants under their own tax ID, 34 turned up under a near-identical name with a new one, 12 filed a return a year or two later. Count every organization with a sign of life as fine and the 32% becomes 27%. The 11% does not move.

What the IRS's own lists say, years later

Our outcome is read from the organizations' filings. A harder test is to take the same organizations to the lists the IRS and the states keep today: Publication 78, the list of organizations eligible to receive deductible gifts (revised 2026-08-11); the Exempt Organizations Business Master File (revised 2026-08-10); the automatic-revocation list; and, for California organizations, the Attorney General's registry and the Franchise Tax Board's revoked list. Five to seven years after the grant ended, 16% (95% interval 14% to 18%) of the organizations that lost it are no longer on Publication 78, against 6% of those whose grant continued. Among organizations that were healthy when the grant ended (spending not down more than a tenth on the year, positive net assets, no large deficit), 11% against 4%; among those still filing two years on, the ones our own measure called fine, 9% against 5%. OFAC and the Internal Revenue Bulletin were screened too and, as they should, return nothing for either group.

Were they dying anyway?

This is the first thing anyone should ask, so we did. Organizations about to lose a big grant were a little weaker in that last year than organizations whose grant continued: 17% were already shrinking, against 10%; 36% ran a deficit, against 28%.

Restrict it to organizations that were healthy when the grant ended, meaning spending not down more than a tenth on the year, positive net assets and no large deficit, and it is 23% (95% interval 20% to 25%) against 9%. Match each cliff to continued relationships of the same year, size, share and form type, and it is 32% against 11%.

There is one more test. Sometimes the funder itself disappears: it spends down, the donor dies, it stops making grants to anyone. Nothing about the grantee caused that. Those organizations were in trouble 23% of the time, against 33% when a still-active funder chose to drop them and 11% when the money continued. Roughly half the gap is funders leaving organizations that were already struggling. The other half is the cliff. Call it two to two and a half times the risk, for an organization that was doing fine.

Nobody steps in

The comforting story is that other funders notice and fill the gap. They do not, on average. Organizations that lose a big funder are twice as likely to lose all of their other grantmakers the following year (11% against 6%). Fewer pick up a new funder (64% against 72%). 20% have the lost money fully replaced by other grantmakers within two years, which sounds like something until you notice that 17% of organizations whose grant continued grew by that much anyway.

Where does the money come from, for the ones that make it? Mostly other contributions and, more than anything else, government grants, which rose by 4.6% of the budget over two years for these organizations against 1.8% for the rest. Total revenue two years on was 3.5% of budget lower than the year the grant ended; for organizations whose grant continued it was 13.8% higher.

Three things we did not expect

The names

The largest cliffs the method finds are recognizable stories. Each of these was checked against coverage; the grant figures are from the filings, the events from the sources linked.

When it is the government

On the Form 990, government grants have their own line. We ran the same design: organizations with government grants at 10% or more of spending for three years, then under 2%. The raw result is 25% in trouble two years later against 7%. But the raw number hides three different things.

Foundations do not backfill government money. Grantmaker dollars rose by even a quarter of what the government took away in 10% of cases, and by half in 7%. Among organizations that were healthy when the money ended, 20% were in trouble against 6%. Exposure is not small: in FY2023, 70,027 of the 321,910 Form 990 filers drew 10% or more of their budget from government grants, $295 billion in all, and 22,534 drew three quarters or more.

USAspending.gov lets us do this by agency, for federal money specifically. Among 1,161 organizations whose federal obligations fell below a quarter of their recent average after two or more years at 10% or more of budget, 27% were gone or a quarter smaller two years on, against 5% when the money continued (20% against 4% among healthy organizations; 6% for the matched control). The agency matters.

When Health and Human Services was the top funder that went away, 40% of organizations were in trouble; when it was the Department of Agriculture, 12%. Cliffs in 2021 and 2022 are largely pandemic relief winding down as designed, so the base rate used below is from 2019 and 2020: 35% overall, 25% for healthy organizations, 5% when the money continued.

Then there is 2025

We took every nonprofit that got 10% or more of its budget from federal awards in both 2023 and 2024: 7,306 organizations, $85 billion a year in obligations against $229 billion in spending. In calendar 2025, 1,668 of them saw their federal obligations fall below a quarter of their recent average, and 1,193 got nothing at all. That is 22.8% of the exposed population (95% interval 22% to 24%). In every year from 2019 to 2024, the same measure ran between 7.5% and 11.5%.

Among organizations whose main federal funder was USAID, 72% fell off. The National Endowment for the Humanities, 69%. The EPA, 56%. The State Department, 54%. Health and Human Services, 7%. Is that just money that was booked early? The award-level dates say no: the last quarter of 2024 was unusually heavy, by about $4.8 billion or 3.4% of the year, and January 2025 was heavy too, which flatters 2025 rather than hurting it. If the 2019 and 2020 rates hold, between 423 and 588 of those 1,668 organizations will be gone or a quarter smaller by the time their 2027 returns are filed, against 66 to 80 if the money had continued. The historical rates probably understate it: the old cliffs included re-competes and expirations, and these are terminations.

What to do with this

For funders: an exit is a design decision, and the shape of it matters. Organizations whose funder tapered rather than stopped came through at 15%, close to the continued rate. Being the only grantmaker in the room is the single biggest risk factor for the organization you fund (49% in trouble when you leave), which is a reason to bring others in before you go, not after.

For organizations: the line is 10% from one source, not three years. Six funders is materially safer than two (15% against 38%). And the government line on your own 990 is now the one to watch.