Where every number on this site comes from, how it is calculated, what we clean, and what we refuse to claim.
Everything here is derived from public IRS filings. There is no proprietary data, no survey, and no self-reported information from sponsors or charities.
The Form 990 data is used in the parsed form published by the National Center for Charitable Statistics at the Urban Institute, which converts the IRS e-file XML into tabular form. Tax years 2019 to 2024 are included.
Returns take time to reach the public. The IRS releases e-filed returns on a rolling basis, and it can take two to three years for the bulk of a tax year's filings to be processed. NCCS advises that “the most recent year in the panel should always be treated as a partial year.”
That shows plainly in the coverage:
| Tax year | Sponsors with a filing |
|---|---|
| 2019 | 1,983 |
| 2020 | 2,175 |
| 2021 | 2,185 |
| 2022 | 2,155 |
| 2023 | 1,968 |
| 2024 | 343 |
Each sponsor is shown at its most recent available filing, and those years differ between organizations. Sector totals therefore blend years and should not be read as a point-in-time snapshot, nor compared year over year across the whole index. Per-sponsor trends are sound, because each organization is compared to itself.
Where a sponsor's latest filing is two or more years old, its page says so prominently. A missing year means no return has been released yet — not that the organization stopped filing or stopped granting.
DAF sponsor. Any organization filing Schedule D, Part I reporting non-zero donor-advised fund activity. That is a wider net than the sector's own count of roughly 1,500, because it includes every filer rather than a curated list — universities, hospitals, and federations that operate a DAF alongside their main work are included. It is a transparent, reproducible definition: anyone with the same files gets the same list.
Payout rate. DAF grants paid, divided by DAF assets held at the end of the fiscal year. Calculated only where assets are above zero and grants are not negative. This is the sector's conventional measure and it has real limits — a sponsor that receives one large gift and grants it out quickly shows a very high rate, and a pass-through platform holding almost nothing can exceed 100%. It measures flow against balance, not quality.
DAF-to-DAF transfer. A grant from one DAF sponsor to another. For the aggregate figure we count only recipients whose donor-advised fund programme is their primary activity — DAF assets of at least $50M and at least twice what they received, or their own DAF grants above $50M. Without that restriction, ordinary grants to universities that happen to run a small DAF inflate the number roughly fourfold. The runs a DAF program label on individual rows is separate and purely factual: it marks any Schedule D filer.
Cause. The NTEE major group assigned by the IRS in the Business Master File. The IRS never assigned a code to many smaller organizations, so 81% of grant dollars carry one. The remainder is filterable as “No IRS cause code” rather than hidden.
Filed data contains errors. We correct nothing silently — every rule below is visible on the page it affects.
Filers routinely enter dollar amounts in the “number of funds” box. One organization reports over four million accounts against $4M in assets. Summed as filed, the sector appears to hold 36.2 million donor-advised fund accounts; the real figure is around 2.6 million. We flag a count as unusable when it exceeds 250,000 or implies an average account below $500, and exclude it from totals rather than publishing it. 116 sponsors are affected.
Schedule D asks for totals, which cannot be below zero. 26 filings nonetheless report negative assets, grants, or contributions — a sign-convention error in the return rather than a real value. We display these exactly as filed, flag them, suppress the payout rate, and exclude them from sector totals. We never flip the sign, because we do not know what the filer intended.
$391.1M of grants are reported against placeholder identifiers — 14 of them, the largest being EIN 000000000, used by 35 separate sponsors. These are real grants with unusable identifiers, so they are excluded from the ranked recipient tables rather than attributed to a fictitious organization.
Sponsors type recipient names by hand, and they vary. One filer's typo once made EIN 53-0196605 appear as “The Bahamas RED Cross Society”. Where an organization is in the Business Master File we use the IRS's own name, city, and state; 80% of recipients are. For the rest we take the most commonly filed value across all sponsors.
One sponsor, SDG Impact Fund, reports $10.10 billion in donor-advised fund assets — 3.3% of every dollar counted on this site, and enough to rank it fifth. Nearly all of it arrived in a single year, and in June 2026 the Georgia Secretary of State issued an emergency order barring the organization from operating as a charity in that state.
We have left the figure in, because this site reports what organizations filed with the IRS and flags what is questionable rather than quietly deleting it. But you should know what it does to the totals:
| Including it | Excluding it | |
|---|---|---|
| Total DAF assets | $304.59B | $294.48B |
| Aggregate payout | 20.44% | 21.14% |
| Median payout | 12.67% | 12.68% |
This is a good illustration of why we lead with the median rather than the aggregate. A single disputed filing moves the sector total by ten billion dollars and the aggregate payout by seven-tenths of a point. It moves the median by one hundredth of a point.
Roughly one recipient in five has no current BMF record, and this is not a warning sign in most cases. State universities and other governmental units are exempt under §170(c)(1), and churches under §508(c)(1)(A), without ever applying — so they may never appear. Among the largest recipients carrying this note, most are public universities. It can also mean an organization merged, dissolved, or that the EIN was mistyped. We state the fact and leave the interpretation to you.
DAF+ is operated by Adam Riding. He is a partner in Crewe Foundation Services, a firm that manages and administers charitable organizations, and he serves on the board of one of them — Crewe Endowment Foundation (EIN 84-3378692), a donor-advised fund sponsor that appears in this index like any other sponsor.
The board seat is the part worth stating plainly rather than leaving to be found. A sponsor’s payout rate is substantially a product of decisions its board makes about grants and distributions — and this site publishes, benchmarks and ranks that rate. Directors are named in Part VII of Form 990, including the returns linked from every sponsor page here, so this was never going to stay quiet. It should not have to.
We are telling you before you find it, because on the measure this site leads with, Crewe scores well:
| Fiscal year | DAF assets | Grants | Payout |
|---|---|---|---|
| FY2022 | $13.7M | $1.4M | 9.9% |
| FY2023 | $10.9M | $6.9M | 63.8% |
Its most recent payout rate of 63.8% is roughly 5.1× the 12.5% median, placing it in the top 5% of the 488 sponsors holding between $5M and $50M.
What we do about it:
A high payout rate in one year is not necessarily a virtue, and this page explains why elsewhere: a sponsor that receives a single large contribution and grants it out quickly will show a high rate without that saying much about how it is run. Read Crewe's number with the same scepticism as anyone else's.
If a figure here is wrong, we want to know. Every sponsor page links to the source filing it was built from, so a discrepancy is usually traceable to either the return itself or our handling of it — and we will say which. Filing errors have to be corrected with the IRS; handling errors we fix and note. Get in touch.
Built from IRS Form 990 Schedule D and Schedule I for tax years 2019–2024 via the National Center for Charitable Statistics parsed e-file tables, and the IRS Exempt Organizations Business Master File and Publication 78. Covering 3,078 donor-advised fund sponsors and 332,900 grant recipients. This page is generated from the same data as the rest of the site, so the figures on it stay in step.