One in five American nonprofits has less than three months of cash in the bank. This is what happens to them next.
We followed 258,398 nonprofits year by year, then checked them against the IRS list of revoked exemptions to see which ones actually shut down. Nonprofits with under three months of cash lose their tax-exempt status at nearly 3 times the rate of similar nonprofits their size. And once a nonprofit is in trouble, raising more money does not get it out. Cutting spending does.
| Runway | How many months a nonprofit could keep going if the money stopped. Its unrestricted cash divided by what it spends in a month. Every nonprofit already files both numbers. Almost nobody does the division. |
| Unrestricted cash | Money with no donor strings attached — formally, net assets without donor restrictions. Also called free cash. This is the number the whole report turns on. |
| Underwater | Owes more than it owns — runway below zero. |
| Critical | Under 3 months of runway. |
| Fragile | 3 to 6 months. |
| Stable | 6 to 12 months. |
| Healthy | More than 12 months. |
| In the danger zone | Underwater or Critical — under 3 months of cash. 22.5% of the sector is here, and 6.3% is already underwater. |
| Running a deficit | Spent more than it brought in that year. Straight off the filing. |
| One dominant funder | A single source provides most of the money. We count it when one source is 80%+ of revenue. |
| Zero paid staff | Nobody received a W-2 that year. The work may still be done by contractors or a fiscal sponsor — it does not mean the nonprofit is dormant. 24,391 nonprofits with no staff spend over $2M a year. |
| Lost its status | The IRS took away tax-exempt status. We only say this when the nonprofit appears on the IRS revocation list by name and date. We never infer it from a missing filing. |
Under three months of reserves means nearly 3 times the likelihood of losing tax-exempt status, verified against IRS revocation records. Rising to 7.8× among the largest nonprofits.
83.4% of underwater nonprofits are still underwater a year later. 60.8% are still in danger four years on. Staying power held at 82–85% straight through the pandemic.
Growing revenue made no measurable difference to whether a nonprofit got out of trouble. What did separate the ones who escaped: they had spread their funding across more sources, and they had built finance and admin capacity before the trouble started — not during it.
Nonprofits spending 93%+ on programs — 7% or less on overhead — hold 6.4 months of cash. Those putting 40%+ into the back office hold 10.3. Investing in overhead moves with better financial health, not worse.
Nonprofits with 40%+ of net assets donor-restricted hold 4.9 months of usable runway against 14.2 for lightly-restricted peers. Restriction transfers risk onto the grantee.
Aggregate runway improved, but the gap between the healthiest and thinnest tenth widened 58%. Sector averages describe fewer real nonprofits every year.
“Low-risk auditee” is a federal designation: nonprofits with clean recent Single Audits get lighter scrutiny on their government funding. Those nonprofits hold 4.5 months of cash against 6.0 for everyone else. The label governs how billions get watched, and it points the wrong way on solvency.
In the worst modelled branch, $2.3B of part-time finance hires protects $71.4B of annual programme delivery and 384,046 jobs.
Going-concern opinions — when a CPA formally doubts a nonprofit can keep going — line up with our cash measure: nonprofits their auditors formally doubt hold about a third of the cash of those they do not. Two completely separate sources, same answer.
Charity raters score overhead. Funders score program ratios. Boards read revenue. None of those predict which nonprofits are still here in six years. We know, because we checked all of them against 258,398 nonprofits tracked through time.
“The sector has spent thirty years chasing a number that goes the wrong way. The leaner the overhead, the shorter the life.”
This report is deliberately structured in three parts, in ascending order of usefulness. What is true takes up a quarter of it. What happens next and what to do about it take up the rest.
Among nonprofits where we can measure cash fairly, median runway is 9.4 months and 22.5% hold under three.
Among bands we can measure properly, runway falls at every step as nonprofits get bigger: 10.3 at $500K–1M, 9.8 at $1–5M, 8.8 at $5–25M, 6.7 above $25M. Committed cost bases, leverage and reimbursement-based revenue. Scale buys stability of income, not cash — and the largest institutions are the thinnest.
Housing & Shelter is the emergency nobody has named. 51.6% ran a deficit — worst of any sector — with 41.0% under three months. More than half of America's housing nonprofits spent more than they raised, during a housing crisis.
The finding funders will like least. The sector pushes nonprofits toward low overhead, and low overhead is exactly what stops them building a cash cushion. Nonprofits spending 93% or more on programs — so 7% or less on running the place — hold 6.4 months of cash. Nonprofits putting 40% or more into the back office hold 10.3 months. Every step between those two points moves the same way. So the leaner the overhead, the shorter the life. A funder who insists on a low overhead ratio is not making their grant go further. They are moving risk onto the grantee, and the grantee is the one who runs out of money.
Runway swings more than five times over between sectors, and by seven months between states. Pick your state and your size to see the number that actually applies to you.
Sector and state are almost always reported one at a time. Put them together and you get the cut that actually applies to a real nonprofit — because a housing nonprofit in Illinois and a housing nonprofit in Georgia are not running the same business.
One row is red all the way across. Mental Health & Crisis is thin in every single state we can measure — there is no state where it holds a comfortable cushion. Housing swings wildly by geography. Mental health does not swing; it is just short everywhere.
Housing & Shelter runs from 5.6 months in Illinois to 15.8 in Georgia — a nearly threefold gap inside one mission area. Mental Health & Crisis is thin everywhere, but New York (3.9 months) is a different order of risk from Illinois (7.4).
Housing in Indiana is the single worst cell we can measure with confidence: 3.8 months of runway and a 49.5% deficit rate across 220 nonprofits.
This matters for how support is targeted. A national housing initiative and an Illinois housing initiative are not the same intervention, and the data says the second is where the need is.
| Sector | State | Orgs | Median runway | Deficit rate |
|---|---|---|---|---|
| Housing | IN | 220 | 3.8 | 49.5% |
| Mental Health | NY | 121 | 3.9 | 46.3% |
| Religion-Related | CO | 61 | 4.1 | 47.5% |
| Employment | TX | 68 | 4.2 | 30.9% |
| Education | UT | 121 | 4.3 | 28.1% |
| Mental Health | PA | 105 | 4.4 | 35.2% |
| International | MD | 132 | 4.5 | 49.2% |
| International | FL | 80 | 4.6 | 45.0% |
| Crime | NY | 92 | 5.2 | 34.8% |
| Mental Health | CA | 190 | 5.2 | 26.3% |
Five years of IRS filings, stitched together, let us follow the same nonprofits from 2017 to 2024 — over a million yearly records. Two things happened. Cash got better. Staffing fell apart.
| Tax year | Observations | Median runway | Under 3mo | Deficit rate | Single-source | Zero staff |
|---|---|---|---|---|---|---|
| TY2017 | 65,829 | 7.2 mo | 30.0% | 39.4% | 45.6% | 28.3% |
| TY2018 | 127,702 | 9.9 mo | 24.1% | 39.2% | 49.7% | 37.9% |
| TY2020 | 84,278 | 10.4 mo | 22.4% | 28.1% | 40.3% | 30.0% |
| TY2021 | 159,289 | 11.9 mo | 19.5% | 29.7% | 44.8% | 35.7% |
| TY2022 | 248,100 | 11.1 mo | 20.9% | 40.3% | 48.4% | 36.9% |
| TY2023 | 221,825 | 11.4 mo | 21.0% | 40.1% | 47.6% | 38.6% |
| TY2024 | 96,351 | 12.0 mo | 20.7% | 37.8% | 48.5% | 41.8% |
Median runway rose from 7.2 months in TY2017 to 12.0 in TY2024. The deficit rate fell to 28.1% during relief and rebounded to 40.3% the moment it ended — a twelve-point swing on one policy variable.
But the share of nonprofits with zero paid staff at all climbed from 28.3% to 41.8% across the same period. The sector got more liquid and less staffed simultaneously.
Given that zero staff is the biggest warning sign you can actually fix for shutting down in our survival model, that is not the sector getting healthier. Relief money landed on balance sheets, it is being spent down now, and the staffing it was supposed to protect never came back.
Twenty-four of twenty-five sectors improved on runway. Only Public Safety & Disaster went backwards. But deficits rose in fifteen sectors, led by Diseases & Disorders (+6.2pp) and Civil Rights & Advocacy (+5.5pp) — nonprofits that built reserves during relief and are now spending them.
Every closure in this report is a nonprofit that appears on the IRS list of revoked exemptions by name and date. We do not infer a closure from a missing filing, because a nonprofit can stop filing for several reasons and most of them are not death.
Nonprofits under three months of reserves face nearly 3 times the likelihood of revocation. Even the three-to-six-month band carries 1.754×.
Revenue concentration follows at 1.422×, deficits at 1.257×. Scale is powerfully protective: 0.424× per tenfold increase in revenue.
And zero paid staff is null — 0.906×. That surprised us. It predicts shrinkage, not death: staffless nonprofits downgrade to the 990-N rather than losing exemption. On the rough stand-in it looked like a top risk factor. Against confirmed loss of tax-exempt status it disappears.
Take the numbers with you: the report as a PDF, the 50-state and 22-sector tables as a spreadsheet, and the validity audit showing every correction we made and why.
This is the finding that should change how the sector thinks. Across 519,283 year-to-year transitions covering 247,947 nonprofits, we measured the probability of moving between health states. Distress turns out to be almost permanent.
83.4% of underwater nonprofits are still underwater next year. 70.6% of critical nonprofits stay critical. Meanwhile 90.0% of healthy nonprofits stay healthy.
Only 5.4% of nonprofits in the danger zone reach Stable or Healthy within a year. And only 1.1% of healthy nonprofits fall into danger.
We checked whether these are really two separate populations. They are not — the cash range has one peak with a long tail, not two humps. But position is remarkably sticky: 87.9% of nonprofits below six months of cash are still below it a year later, and 92.7% above are still above. One sector, one range, and most nonprofits stay roughly where they sit on it.
The escape rate from the danger zone falls with organizational size: 11.2% for nonprofits under $100K, 4.9% at $1–5M, and just 1.6% above $25M.
A large distressed nonprofit is the most trapped entity in the sector. Its cost base is committed, its revenue is contractual, and its balance sheet cannot turn quickly. Small nonprofits are fragile but nimble; large ones are stable until they aren't, and then they're stuck.
This is the number to take to a board. Financial trouble in this sector is not a bad year you ride out — for three in five nonprofits it is still there four years on. 83% of underwater nonprofits are still underwater after one year. Waiting is not a plan.
The sector's default response to financial trouble is to raise more money. We took 52,388 nonprofits that were in the danger zone at first observation and tracked them for three or more years to see what actually distinguished the 19.4% that escaped.
The ones that got out grew revenue by 29.5%. The ones that stayed stuck grew by 24.1% — and a slightly higher share of the stuck grew at all (66% against 62%). Account for size, sector and age and extra revenue growth changes the chance of getting out by nothing measurable. What separated them was spending: the ones that got out cut it by 16% while the stuck raised it by 5%.
What did predict escape: being diversified before the crisis (1.249×) and already having paid staff (1.136×).
Both are structural conditions established beforehand. Neither is something you can do once you're in trouble.
“Prevention is not the cheaper option. On this evidence it is the only option.”
Six questions. The weights come from the numbers in this chapter, and we checked them on nonprofits the model had never seen. You get a risk band and a ranked list of what to fix first, with how much each move is worth.
Here is the line that matters: six months of cash. 66.6% of nonprofits are above it and 33.4% are below. Between 2017 and 2024 the sector's average cash position improved — and the gap between the best-off tenth and the worst-off tenth grew by 58%. The middle did not hollow out so much as the two ends pulled away from each other.
Newer nonprofits are not weaker. Comparing founding cohorts at the same age (1–8 years), those founded 2019–2023 hold 5.6 months against 4.4 for the 2005–2010 group. The formation surge is not producing a more fragile vintage.
Revenue concentration has not worsened. Single-source share moved 45.6% to 46.3% across seven years. A large, stable risk — not a growing one.
We report both because a trend report that only finds trends is not measuring carefully.
Applying the measured transition matrix forward from the tax-year 2024 distribution gives a projection that assumes nothing bad happens — no recession, no funding cut, no donor retreat. Just the sector's own observed churn.
The current 20.7% danger-zone share is better than equilibrium. The transition dynamics pull toward 27.4% — more than one in four nonprofits financially distressed — and they get there without a single adverse event. The sector's present condition is being propped up by relief-era reserves that are draining.
IRS determinations rose from 47,900 in 2018 to 119,873 in 2025 — a 150% increase. On trend that is 168,618 new nonprofits a year by 2030. Among nonprofits large enough to file a full Form 990, 46.2% now report zero paid staff at all — nobody on a W-2. On that rate, roughly 339,655 of the new nonprofits will run with no employees. Smaller filers do not report staffing, so this is a floor, not a ceiling. We are creating nonprofits far faster than we are creating the capacity to run them.
Three choices, twenty-seven endings. We measured how fast spending really follows revenue down, using nonprofits that lived through it. Pick your four years before you read what we advise.
Every recommendation below carries a measured effect size from this dataset. Nothing here is inferred from best practice, sector convention, or what sounds responsible.
| Intervention | Measured effect | Evidence |
|---|---|---|
| Employ someone in a finance role | +2.1 months | Holds within every revenue band up to $25M |
| Move off single-source revenue | +5.5 months | 1.249× for escaping distress — the only lever that works after the fact |
| Negotiate unrestricted or general-operating funds | +10.4 months | 40%+ restricted vs 1-15% restricted |
| Grow the board past 8 members | +6.1 months | Tiny boards track with founder-controlled finances |
| Add any paid staff | 1.716× → 1.00 | The biggest risk on this list a nonprofit can act on itself |
Growing revenue changed nothing measurable. The ones that got out grew 29.5%; the stuck grew 24.1% — and more of the stuck grew at all. Cutting spending is what separated them.
Among already-audited nonprofits, those with an audit committee hold 10.1 months against 9.7 without. Governance theatre with no measurable cash effect.
Nonprofits at 93%+ program spend hold 6.4 months of runway. Those under 60% hold 10.3. The metric charity raters reward is linked to a shorter life, not a longer one. We tested whether these nonprofits are simply more restricted or more government-funded. They are not, so we report the pattern and not a cause.
The standard reserve guidance is a single number applied to every nonprofit regardless of age or size. It is the most widely repeated piece of nonprofit financial advice, and it is not supported by the distribution of what nonprofits actually hold.
| Nonprofit age | 25th percentile | Median | 75th percentile | Orgs |
|---|---|---|---|---|
| <5 years old | 1.7 | 6.7 | 23.6 | 11,209 |
| 5-10 years old | 2.1 | 7.1 | 20.8 | 14,584 |
| 10-20 years old | 2.9 | 8.7 | 23.8 | 26,437 |
| 20-35 years old | 3.9 | 11.4 | 32.3 | 35,828 |
| 35-60 years old | 5.4 | 13.1 | 31.6 | 36,408 |
| 60+ years old | 7.6 | 17.1 | 38.4 | 18,288 |
A five-year-old nonprofit at the median holds 6.7 months. A sixty-year-old holds 17.1. Telling both to target three to six months tells one it is failing and the other it is finished.
Use the table. If you are under ten years old, the 25th percentile of your group is roughly two months — meaning a quarter of nonprofits like you operate below that. That is not permission to be thin. It is context for whether you are unusual.
What the age pattern really shows: deficit rates are flat across age (30.8% for the youngest, 37.7% for the oldest). Age does not make nonprofits better managed. It gives them longer to accumulate a buffer against the same mistakes.
Weights are derived from a logistic model fitted on 167,755 nonprofits tracked from 2019 to 2025, then calibrated against actual six-year shutting down rates. This is not a quiz — the point values are regression coefficients.
| Risk band | Nonprofits | Disappeared within 6 years | Share of sector |
|---|---|---|---|
| 0-7 Low | 57,004 | 2.2% | 34.0% |
| 8-14 Moderate | 43,197 | 4.4% | 25.8% |
| 15-21 Elevated | 37,137 | 6.4% | 22.1% |
| 22-28 High | 22,997 | 9.8% | 13.7% |
| 29+ Severe | 7,420 | 14.6% | 4.4% |
The prescriptions in this report are cheap relative to what they prevent. Here is the arithmetic, using the worst modelled branch so the comparison is conservative rather than flattering.
32,878 nonprofits filing a full Form 990 report no finance staff at all. One part-time finance hire each, at a loaded cost of $70,000, comes to $2.3B a year across the sector.
In the worst branch of our model, the difference between building that capacity and doing nothing is 8,101 nonprofits, 384,046 jobs, and $71.4B of annual programme delivery.
That is 31 dollars back for every dollar spent, and it is the low estimate. The hires cost money every year. What they prevent is losing skills and know-how that take years to build back.
This report has been corrected three times. We publish the corrections rather than quietly absorbing them, because each one changed a conclusion — and because the pattern of the errors is itself the report's argument.
The original parser assigned zero to net assets without donor restrictions whenever the field was absent. It is absent for a legitimate reason: Form 990-EZ and 990-PF filers never report it. About a third of the sector was counted as having no reserves at all. Running both parsers over identical files, they agree on 97.7% of records where each returns a value — but the original produced 5,700 phantom zeros in a single archive.
Having fixed the zeros, we reported that the smallest nonprofits held the deepest reserves — a median of 41.9 months under $100K. That was an artifact too. The IRS permits the short Form 990-EZ only when revenue is under $200K and assets are under $500K, so a small-revenue nonprofit filing a full 990 is by rule unusually asset-rich. Only 12.5% of sub-$100K filers appear in the balance-sheet sample, and their assets run 6.9× revenue against 1.1–1.3× elsewhere. Split that band at the $500K asset line and the halves read 16.7 months versus 120.
Runway is now reported only for nonprofits at $500K+, where coverage runs 76–87%. Median 9.4 months, not 11.1. And the pattern runs downward with size, not upward.
Our survival model used organizational age and NTEE sector as predictors. Both come from the IRS Business Master File — and the IRS removes a nonprofit from that file when it revokes the exemption. Those variables are therefore missing by construction for exactly the nonprofits that died.
| Outcome group | Nonprofits | Matched to Business Master File |
|---|---|---|
| Still filing 990 or 990-EZ | 174,151 | 96.8% |
| Stopped filing (the old rough stand-in) | 12,099 | 29.4% |
| Verified revoked, 2019 onward | 3,781 | 12.3% |
Dropping unmatched rows silently discarded 6,624 of 8,848 outcome events — 75%. The model was fitted almost entirely on survivors. What exposed it was an impossible coefficient: a 99% reduction in risk per decade of age.
What we cannot test. Organizational age cannot be related to organizational death using any IRS source, because of that deletion. Age remains valid for describing living nonprofits, where it drives a real 2.6× gap in runway. No survival claim in this report rests on it.
We publish this because it is the report's thesis in miniature. Nonprofit financial data is easy to misread, the errors run toward alarm, and almost nobody checks. Three separate analytical traps, each of which produced a plausible and publishable wrong answer. If a dedicated analysis with a purpose-built parser hit all three, a board reading a quarterly packet has no chance.
Averages hide you. Pick your size and where your cash sits today, and this walks the five questions that matter — using only nonprofits in the same band and the same position.
Recolour the map by any metric, then pick up to six states. The comparison below is a single scorecard rather than one chart per metric — rows stay in the same order so you can read a state across, and the shading is each state's position among all 51, so a colour means the same thing in every column.
Everyone is told to diversify and nobody is told what the target is. So we measured it: the mix that nonprofits your size run when they are financially healthy, against the mix the thin ones run.
National medians are useless for planning. Pick your state and revenue band.
Three things decide the next four years: what Washington does, what donors do, and what nonprofits do about their own back office. Pick all three and we model what happens to the 135,324 nonprofits holding cash today. Twenty-seven endings. Every number comes from the filings in this report. When revenue falls, we do not guess how fast spending follows — we measured it on 20,906 nonprofits that lived through a real drop.
Each of these is measured in this report and each contradicts something the sector currently believes. They are written to be argued with.
83.4% of underwater nonprofits are still underwater a year later. 60.8% are still in danger four years later. The sector talks about cash-flow problems as episodes. The data says otherwise.
Growing revenue made no measurable difference to whether a nonprofit escaped trouble. The ones that stayed stuck actually grew revenue slightly more often than the ones that got out (66% against 62%). The sector's whole theory of change — raise more — does not move the thing it is meant to fix.
Nonprofits spending 93%+ on programs — 7% or less on overhead — hold 6.4 months of cash. Those putting 40%+ into the back office hold 10.3. Every rating agency that rewards a low overhead ratio is rewarding fragility.
40%+ restricted assets means 4.9 months of usable runway against 14.2 for lightly-restricted peers. Restriction is not stewardship. It is risk transfer onto the grantee.
Median runway runs from 6.7 months at under-fives to 17.1 at sixty-plus. One number for every nonprofit tells young ones they are failing and old ones they are done.
Low-risk auditees hold 4.5 months of runway against 6.0 for everyone else. The designation governing $2 trillion in federal spending is linked to less cash, not more.
One-year escape rate falls from 11.2% under $100K to 1.6% above $25M. Emergency funding for big distressed nonprofits buys time, not recovery.
Under three months of reserves means nearly 3 times the chance of losing tax-exempt status. A finance hire is worth up to 3.4 months of runway and the effect holds in every revenue band.
Formations up 150% since 2018, on trend for 168,618 a year by 2030, while the zero-staff share climbed from 28.3% to 41.8%. This is a formation problem, not a funding problem.
Matched against the IRS revocation list, the biggest warning sign a nonprofit can actually do something about is holding less than three months of cash — nearly 3 times the chance of losing tax-exempt status, rising to nearly 8 times at $25M and above. Not overhead. Not program ratio. Not revenue growth, which has no measurable effect on escaping trouble at all. And distress is 83.4% persistent year to year, so by the time it is visible on an annual report it is already entrenched. Runway is computable from a filing every nonprofit already submits, it predicts the outcome that matters, and no funder dashboard or rating agency tracks it. That is a measurement failure, and measurement failures are what infrastructure fixes — which is what we build.
Mazlo co-founder Kian walks through the nine claims, the transition data behind them, and what to do this quarter.
The full report as a PDF, the 50-state and 22-sector tables as a spreadsheet, and the validity audit documenting every correction we made. Built for people writing a board paper or a funding case.
The tools nonprofits use to manage money were built for businesses. Mazlo builds what the sector actually needs — banking, fund tracking, grant management and approvals in one place.
See what we're building →Get in touch