Where do fifty of America's biggest Jewish givers send their money?

Four signatures cover more than a third of the money

A "donor" here is a private foundation, which is the one giving vehicle American law forces to disclose grant by grant. Its annual return has to list every grant paid, with the recipient's name, city and amount. We built a roster of the family foundations known for substantial Jewish giving, ranked them by grants actually paid, and took the top fifty. Forty-four filings cover fiscal 2024; six already cover fiscal 2025. That definition does a lot of work, and we come back to what it excludes at the end.

The picture below is all fifty checkbooks drawn to scale. Each rectangle is one foundation, and its area is the dollars it paid out in its latest filed year. Color marks what the filing lets us see: blue means the grant book is itemized and readable line by line, orange means the foundation filed its entire grant list as a paper attachment the public file does not contain, and the one green block, Koum, itemized most of its book but not all of it.

Notice how fast the rectangles shrink. Helmsley ($440.9 million) and Schusterman ($333.5 million) together outweigh the bottom thirty foundations combined, and the top four, adding Marcus ($227.5 million) and Singer ($200.0 million), are 36 percent of everything.

Twenty-six of them run on money that stopped arriving

Geographically the fifty cluster hard. Ten are run from New York City, seven from Chicago, five from San Francisco, five more from greater Los Angeles, three from Boston, and the rest scatter out to Tulsa, Milwaukee, and Boca Raton. The fortunes underneath them read as a rough history of American Jewish money. Older endowments came out of apparel, retail, and real estate: Levi Strauss money sits behind three separate Haas and Goldman funds, Home Depot behind Marcus, L Brands behind Wexner, Schottenstein Stores behind Schottenstein, hotels behind Helmsley and every Pritzker entity. Newer money is finance and tech, which is how Elliott Management, Baupost, Blackstone, Highfields, Omega, Pershing Square, Broadcom, and WhatsApp all end up on a list of Jewish grantmakers.

Where the money came from turns out to matter less than whether any new money is still coming. The return also lists everyone who put money in during the year, and for private foundations that list is public. Twenty-six of the fifty reported no meaningful contributions at all. Those are pure endowments: the donor is finished giving, often because the donor is dead, and a board and staff spend the fund down at whatever pace they choose. The law sets a floor of roughly five percent of assets a year, and eighteen of the fifty sit within about a point of it.

Koum pays out 2.5 percent of $3.3 billion, which is the habit at its most cautious. The last column of the table answers a question filings rarely get asked out loud: who wrote the checks into the checkbook this year.

That column also catches the biggest money story of the year, and it is not a grant. While the Marcus Foundation was paying out $227.5 million, Bernie Marcus and his estate were putting $538.6 million in: $294.0 million from the estate and two personal gifts of $120.1 million and $107.3 million recorded in his final months. A foundation taking in more than twice what it pays out, in the year its donor dies, is swelling rather than winding down.

Below are the ten largest single checks written into any of the fifty this year. Dark bars are money arriving from an estate, a trust or a family fund; pale ones are a person writing a check under their own name. Count the dark bars: five of the ten, which is how large American fortunes usually move into philanthropy.

Five destinies, and the two dark ones are growing fastest

The chart below opens each year up, one bar per year on a single scale. The blue-green segments are money the public can follow: to the roughly 4,950 recipients only one, two or three of the fifty fund, which we call their own worlds; to the 137 organizations four or more of them fund, the shared center; and straight across the ocean to Israeli addresses. The orange segment is money whose recipients exist only on paper attachments the IRS does not scan. The gray segment went to a named recipient that is itself anonymous, a donor-advised fund or another family vehicle, where the destination is invisible by design.

Read the two dark segments across the three years. Together they carried $582.8 million, then $627.3 million, then $822.7 million: 21.4 cents of every dollar, then 20.2, then 24.4. The gray segment alone more than doubled, from $91.1 million to $235.2 million. Whatever else changed in these books over three years, the share that leaves public view at the very first hop is now at its highest point in the panel. Hop eleven goes after that money door by door, to establish how much of it a reader can actually recover.

The gray segment is mostly one man. Paul Singer's foundation paid out $200.0 million in the latest year, and $161.5 million of it, four dollars of every five, went to a single recipient: the JP Morgan Charitable Giving Fund, a donor-advised fund. Another $25 million went to a second Singer family vehicle. Across the panel his transfers into such funds ran $68.0 million, then $110.0 million, then $186.7 million. The most closely watched political donor in Jewish philanthropy is, on paper, fully transparent, and a rising share of his money disappears into the one legal structure built not to say where it goes.

One check runs the other way. The visible three quarters really is complete: in 42 of the 44 books we can read, the grants add up to the foundation's own reported total to the exact dollar. Mandel is off by $1,800 on $77 million, and only Koum has a real hole, $11.0 million it never accounts for. What this record contains is accurate. The problem is what it is permitted to leave out.

A sixth of the money converges on the same 137 organizations, year after year

Mostly, yes. Over half the total goes to causes only one family cares about. The web is in what is left over. The fifty books name about 5,100 recipients in the latest year, and 137 of them appear in four or more of the books. Those 137 organizations collected $534 million, about one dollar in six of everything the fifty gave.

That number barely moves. Across three filings the count of organizations funded by four or more of the fifty ran 138, then 136, then 137, while the dollars behind it ran $479 million, $562 million, $534 million. Individual foundations churn heavily over the same period, which makes the stability of the membership list the more surprising half of the finding.

The grid below shows the twelve most shared in the latest filing. Each row is one recipient. Each column is one of the fifty foundations, in the same size order as the blocks in hop one. A filled dot means that foundation made at least one grant to that recipient that year. The count at the right is how many of the fifty funded it, and the dollar figure is what they gave it together.

Look at the top row. Hillel, the campus organization, collected checks from twenty-four of the fifty once its national body and its individual campus chapters are counted together, from Helmsley down to the Walter and Elise Haas Fund, totaling $26.0 million. Birthright Israel got seventeen. So did the Jewish Funders Network, which is not a cause at all but the club where these funders meet, and its $4.9 million in dues and program support may be the clearest single signal of who considers themselves part of organized Jewish philanthropy.

Two out of every three visible dollars leave the Jewish world entirely, in all three years

So far the recipients have been a crowd. To find out what is in it, we sorted all 25,199 grants into nineteen categories by what the recipient organization actually does, and checked the biggest names in each category by hand. This only works on money whose destination the filing names, so the hidden money is set aside here. In the latest year that leaves $2.54 billion to sort.

Below, the categories are stacked into two bars on one scale: blue for organizations that are Jewish by mission, green for secular ones.

Look at how little of the chart is blue. Of the $2.54 billion we can trace, Jewish organizations received $841 million, or 33.1 percent. The other 66.9 percent went to universities, hospitals, museums, charter schools, climate groups, and food banks with no Jewish mission at all. The single largest category is American higher education at $514.3 million, ahead of Israel and Israel-facing giving at $429.9 million.

The ratio holds across the panel, and it moves the right way slowly. The Jewish share of traceable dollars ran 30.2 percent, then 32.7, then 33.1. Three of the nineteen categories grew by more than half over the three years: higher education from $337.2m to $514.3m, Israel and Israel-facing from $325.0m to $429.9m, and the campus, camp and young adult bucket from $38.4m to $57.9m. Two shrank by more than a fifth: global development and relief from $47.4m to $26.5m, and the unclassifiable long tail from $410.4m to $350.6m.

Two cautions on this chart. It classifies by the recipient's mission, not by what the money bought, so a gift to a university that funds a Jewish studies chair counts as higher education. And the second segment of the green bar, the one labeled everything else, is thousands of organizations with a median grant well under $50,000 that no honest reading can sort, most of them small and local. Neither caution moves the headline. Even if every dollar in that segment turned out to be Jewish, which it plainly is not, the secular share would still be the majority in every year of the panel.

Nine dollars to a secular university for every one to a Jewish school

The two longest segments in the last chart were higher education and Jewish education, so it is worth opening both of them up and putting the contents side by side. To make the comparison fair we counted only American institutions on both sides, moving Israeli universities out to the Israel column where they belong, and we counted university medical schools as universities, since a gift to a medical school is a gift to the university that runs it.

Sorted that way, the fifty foundations sent $497.1 million to 265 American secular colleges and universities in the latest year. They sent $80.8 million to every form of American Jewish education combined: day schools, yeshivas, seminaries, and the organizations that train the teachers and write the curriculum. Narrow it to schools that actually educate Jewish children, the K-12 day schools and yeshivas, and the figure is $50.9 million.

The chart below counts the money in blocks. Each small square is $5 million. The top row is secular higher education, the middle row is all Jewish education, and the bottom row is Jewish K-12 schools alone. All three rows use the same square, so the comparison is a matter of counting. Hop ten runs the same comparison across all three filings, where the gap turns out to move a great deal more than a single year suggests.

The gap runs through participation as well as dollars. Forty-two of the fifty foundations wrote at least one check to an American secular university, so only eight skipped that world entirely, and the count was forty-two in each of the three years. Twenty-two wrote a check to a Jewish K-12 school, which means twenty-eight of these fifty Jewish family foundations funded no Jewish school at all last year. That count moved from twenty-two to twenty-one to twenty-two across the panel, so it is a settled habit rather than a bad year.

Inside the secular column the money is concentrated in institutions that were already rich, and it got more so. About 58 percent of it, $287.8 million, went to two dozen elite private universities in the latest year, against $157.2 million and $156.5 million in the two years before, so the elite share of a growing pot nearly doubled in one step. Columbia alone drew grants from fifteen of the fifty, counting every spelling of its name that appears in the filings.

The table below puts the two lists next to each other, ranked. The left column is where secular higher-education money went, the right is where Jewish education money went, and the rows are not paired: they are simply the biggest recipients on each side, so you can read across and see what a comparable rank buys in each world.

Two things this does not prove. It is one year, and a capital campaign can put a single university at the top of a list for reasons that have nothing to do with a donor's priorities over a decade. And a foundation that skips Jewish schools may be funding Jewish life in other ways entirely, through camp, campus, or Israel, all of which are separate categories in the chart before this one.

What it does show is a pattern in where the largest Jewish fortunes in America put educational money when they have the choice. The institutions that educate Jewish children are funded by half of these donors at a tenth of the scale. The institutions that educated the donors, or that carry their name on a building, are funded by most of them at ten times the scale.

Thirty-three grants carry more money than the other 8,517

Both, and the gap between them is the point. The median grant is $50,000. The average is $393,579, almost eight times larger, and whenever an average sits that far above a median it means a handful of enormous entries are dragging it upward.

Below, each row is a size band. The left bar counts grants, the right bar counts dollars, and the story is in the mismatch between them.

Look at the top row and the bottom row together. The 1,449 grants under $10,000 are 16.9 percent of all the grants written and carry one tenth of one percent of the money. The 33 grants over $10 million are four tenths of one percent of the grants and carry 38.9 percent of the money. Put plainly, thirty-three checks moved more money than the other 8,517 combined.

This shape is structural rather than seasonal. Run the same bands over the two earlier filings and the share of grants under $10,000 comes out at 17.5 percent, 17.5 percent, and 16.9 percent. What moves is the top band: the twenty-five largest grants carried 30.6 percent of the money in the first year, then twenty-nine carried 35.1 percent, then thirty-three carried 38.9 percent. The small end of these books stays exactly where it is while the big end keeps getting bigger.

The same lopsidedness shows up when you count organizations instead of grants, and it has been getting worse. Ten recipients out of 5,240 collected 29.9 percent of all the money in the latest year, against 26.7 percent the year before and 25.4 percent the year before that. The top hundred collected 56.9 percent, up from 52.7. To reach 90 percent you need a thousand organizations, and the remaining four thousand split what is left. Three years is a short window, and in that window every concentration measure moved the same direction.

And the crowd is a crowd of strangers. Of the 5,240 recipients, 4,518 of them, or 86 percent, appear in exactly one of the fifty grant books. Those single-funder organizations took $2.19 billion, 65 percent of everything, up from $1.74 billion three years earlier. The shared center from two sections ago, the organizations that four or more foundations fund together, held exactly 140 members in each of the three years on this wider count, and $543 million in the latest one. So the picture is one enormous private world of fifty separate philanthropies, with a small dense knot of communal institutions at the middle that everyone pays into, and the private world is growing faster than the knot.

Geography sharpens it further, and it also shows how unstable a one-year map is. Sorting the traceable dollars by the recipient's own address, New York organizations take $619 million in the latest year, a quarter of everything, and they took a quarter in each of the other two years as well: $537m, $616m, $619m. California is steady in second. After that the map jumps around. Pennsylvania sat at $62 million and $65 million and then leapt to $210 million, almost entirely because Carnegie Mellon sits in Pittsburgh and received one $122.7 million grant. Massachusetts ran $122m, then $198m, then back to $132m. Michigan went $51m, $127m, $47m. Only the top two places on this map are places; the rest are the addresses of whichever large gift happened to land that year.

In the chart below, the length of each row's line is how far that place moved in three years.

One dollar in eight, and the biggest direct Israel funder has no living donor

Two kinds of Israel giving show up in these filings. The first is a wire to an Israeli address: the filing lists the recipient's city as Jerusalem or Tel Aviv and the country as Israel. That is the green segment in hop three, and it is oddly flat: $169.9 million, $163.7 million, $172.1 million across the three years. The second kind never leaves the United States on paper: a check to an American Friends organization, to Birthright, to PEF, to a friends-of-a-hospital fund. Counting every recipient whose mission is Israeli or Israel-facing brings the latest-year total to $429.9 million, or 12.8 percent of everything the fifty gave, and it is still a floor, because none of the hidden $587 million is counted in it.

The second kind is where the movement is. Across the panel the full Israel figure ran $325.0 million, then $361.3 million, then $429.8 million, a rise of 32.3 percent against 23.4 percent for the books as a whole, so Israel's share of the total went from 11.9 percent to 11.6 and then up to 12.8. The number of distinct Israel-facing recipients rose from 254 to 318. Almost none of that growth arrived by wire. It came through American intermediaries: American Friends of Tel Aviv University added $24.5 million between the first year and the last, the National Library of Israel's American arm added $14.8 million, PEF added $12.0 million, and American Friends of Ariel University added $10.8 million. October 7, 2023 falls inside the final year for the forty-four foundations on the ordinary calendar, and the final step is the larger of the two.

Below are the ten foundations that wired the most directly to Israeli addresses last year.

Look at the top bar before reading the label. The largest direct funder of Israeli institutions among the fifty is the Helmsley Charitable Trust, a fortune with no living donor: Leona and Harry Helmsley are long dead, and it is their trustees who run an Israel program that wired $78.2 million, 17.7 cents of every Helmsley grant dollar, to Israeli hospitals and universities: Ben-Gurion, Hadassah, Sourasky.

Somewhere between a hundred and nobody at all

The filings never state a headcount, but they do show the paychecks, and the paychecks describe two kinds of institution operating under identical law. One kind is a staffed professional shop. Helmsley reports at least 109 employees earning over $50,000 and paid its CEO, Sarah Paul, $1.03 million; running the charitable program cost $36.6 million on top of the grants, about eight cents for every grant dollar. Weinberg reports 63 such employees and paid president Rachel Monroe $1.07 million. Mandel paid CEO Jehuda Reinharz $1.35 million, the highest salary anywhere in the set. Jim Joseph carries the heaviest overhead relative to its size, $18.9 million of charitable spending on top of $63.0 million in grants, or thirty cents of operations per grant dollar, which is roughly what it costs to run an education funder that does its own evaluations and convenings rather than just cutting checks.

The other kind is a kitchen table. The Wexner Family Charitable Fund moved $42.0 million with no reported operating cost at all: every disbursed dollar left as a grant, and Leslie Wexner, listed as president at one hour a week, took nothing. Leon Cooperman signs his own checks in Boca Raton, unpaid. Helen Zell moved $63.0 million as unpaid president. Miriam Adelson runs $66.9 million as unpaid trustee on operating costs of $66,000, a tenth of a cent per grant dollar. Geffen's entire apparatus is two people, one of whom is paid.

The operations column below is charitable spending other than grants, taken from the filing's own disbursements column. It excludes investment management costs, and it cannot see whatever a family office absorbs off the books, so treat it as a floor on the cost of the machine rather than the whole bill.

The two fastest-growing large lines are secular universities and money with no named destination

Everything up to here rests on one filing per foundation, and a single year can mislead in a dozen ways. So we went back twice. Every one of the fifty has three consecutive filings in the public record, which gives 25,199 grants worth $9.20 billion, all counted the same way. For forty-four foundations the three years are fiscal 2022, 2023 and 2024. For the six whose books close later they are 2023, 2024 and 2025.

Start with the size of the pile. Total giving rose from $2.73 billion to $3.37 billion, an increase of 23.4 percent across two steps, and it did not arrive evenly: up 14.1 percent in the first step, then 8.2 percent in the second. The number of itemized grants moved much less, from 8,162 to 8,550, a rise of 4.8 percent. Money grew about five times faster than paperwork, so the average grant went from roughly $334,000 to roughly $394,000. These fifty are writing larger checks rather than reaching more organizations.

The eight panels below carry the measures this piece has been building on, counted the same way in every year. Each is a small chart with three points, earliest on the left, and the figure under the title is the change across all three.

Read the Jewish and secular panels against each other. Money to Jewish organizations rose 29.7 percent over the three years while money to secular organizations rose 13.7 percent, so the Jewish share of traceable dollars drifted from 30.2 percent to 33.1 percent. Almost all of that movement happened in the first step and then nearly stopped, which is a caution against reading it as a trend. Both underlying numbers are still growing, so nothing in the panel describes a retreat from secular giving.

The fourth panel is the one to stop on. Money whose destination is never named went from $582.8 million to $627.3 million to $822.7 million, a rise of 41.2 percent and the steepest of any large line. Three books did nearly all of it: Singer's transfers into donor-advised funds ran $68.0m, $110.0m, then $186.7m, Pritzker Traubert's single attached line ran $17.5m, $13.2m, then $120.5m, and Bader's ran $32.7m, $45.3m, then $68.9m. The untraceable share of the panel fell from 21.4 percent to 20.2 and then rose to 24.4, which means the completeness of the public record depends less on the disclosure rules than on which of these particular donors happened to be busy.

Israel is the measure most readers will want a trend for. It ran $325.0 million, $361.3 million, $429.9 million, up 11.2 percent and then 19.0 percent, so its share of all giving went 11.9 percent, 11.6, 12.8. October 7, 2023 falls inside the last of the three years for the forty-four foundations on the ordinary calendar, and that is the larger of the two steps. The rise is real without being a surge: a third more money over three years, against a book that grew by a quarter. It is also not a wire transfer story. Money sent directly to Israeli addresses barely moved at all, $169.9m, $163.7m, $172.1m, while the American intermediaries grew: American Friends of Tel Aviv University up $24.5 million across the panel, the National Library of Israel's American arm up $14.8 million, PEF up $12.0 million. Private family foundations run on multi-year pledges and quarterly board meetings, which makes them slower instruments than the federation emergency campaigns covered earlier in this series, and the shape here is consistent with that.

A correction, and what the third year does to the education gap

When this piece first ran with two years it reported that secular higher education rose while Jewish K-12 schools slipped, widening the gap from about $7.90 to $9.60 for every dollar. Two things were wrong with that sentence. The smaller problem was ours: we had been counting the Dana-Farber Cancer Institute as a Jewish day school, on the strength of its name, and Yeshiva University as a school rather than a university. Fixing both lowers the Jewish K-12 totals across the three years and drops the count of foundations funding a Jewish school in the latest year from twenty-six to twenty-two.

The larger problem was that two points do not make a line. With three years in view, secular higher education ran $317.2 million, $404.3 million, $497.1 million, and Jewish K-12 schools ran $14.5 million, $49.8 million, $50.9 million. The ratio between them was 21.9 to 1, then 8.1 to 1, then 9.8 to 1. Across the full panel the gap closed dramatically and then reopened a little. Our earlier sentence caught the reopening and missed the closing, which is the failure mode a two-point line invites.

Watch the short column in the chart below. It triples between the first pair and the second, then stops.

One more correction runs the other way, and it is not about education. Our earlier count of Israel giving was missing several large American Friends organizations, most of them the friends-of-an-Israeli-university kind. Corrected, the Israel series reads $325.0m, $361.3m, $429.9m, and the sentence we published saying Israel giving grew more slowly than the book as a whole was wrong. It grew faster.

That pale block carries most of the story. The Zell Family Foundation gave $1.5 million to Jewish schools in the first year, $34.0 million in the second and $10.0 million in the third. The second-year figure is $24.1 million to the Bernard Zell Anshe Emet Day School in Chicago and $10.1 million to the Rochelle Zell Jewish High School, with nearly all of the third year going to Rochelle Zell again. Both schools carry the family name. One family recapitalizing two schools it had already built accounts for the entire apparent breakthrough in the middle year.

Strip Zell out and a different pattern shows up in the latest year, one that is more encouraging than the middle year was. Jewish K-12 giving by the other forty-nine ran $13.0 million, $15.8 million, then $40.9 million, a rise of about 159 percent in the final step. The names behind it are mostly new to this column. The Gottesman Fund reached $10.8 million, including $5.4 million to the Shefa School in New York. Schottenstein reached $9.3 million, including $6.4 million to Ner L'Horeinu. Marcus reached $7.5 million, almost all of it the $7.1 million that went to the Weber School in Atlanta. Meanwhile the count of foundations funding any Jewish school moved from twenty-two to twenty-one to twenty-two. What changed was the size of a handful of checks rather than the number of donors willing to write one.

The panel also lets us watch each foundation move rather than each measure. Thirty-seven of the fifty gave more in the second year than in the first, and twenty-nine gave more in the third than in the second. Only twenty-one rose in both steps and only five fell in both, which leaves twenty-four that changed direction at least once. The median foundation's giving moved 20.2 percent in the first step, 13.3 percent in the second and 21.5 percent across the full three years. Seven of the fifty more than doubled. None of the fifty ended the panel below half of where it started, so the churn runs upward more often than down.

The grid below gives each foundation its own small chart, blue if it ended higher than it started and orange if it ended lower. Each panel is scaled to its own foundation, so a steep line in a small foundation is still a small amount of money.

Three years settle the shape of the money. Two dollars in three leave the Jewish world in every year of the panel, the undisclosed share sits in a band between a fifth and a quarter, Israel holds between eleven and thirteen percent, and secular universities took at least eight times what Jewish schools took in every year we looked at, and twenty-two times in the first. What three years cannot settle is any individual ranking, since a foundation's place on a top-fifty list turns on when an estate closes or a pledge lands.

Three ways to disappear, and they do not work the same way

Every section so far has treated the undisclosed money as a single block, and it is really three. The $822.7 million that named no destination in the latest year left the record by three different doors, which work differently enough that lumping them together hides the part worth being angry about. So we went after each one separately, to find out how much of it a determined reader can actually recover.

The short answer is almost none. Of the $822.7 million, we could follow $25.0 million one hop further and name what happened to it. That is three cents on the dollar. One route we have not yet closed could raise it, and we come back to that at the end.

The chart below splits the dark money by door, one bar per year. Watch the orange block stay roughly where it is while the gray one grows behind it.

The first door: a statement that was never filed

Six foundations wrote something like "See Attached Statement" where the recipient's name goes, and gave a total. We assumed, as most readers would, that the attachment was somewhere in the filing and that our parser had missed it. So we went and looked, filing by filing, at every document the IRS released alongside each return.

There is no attachment. Not a hidden one, not a badly named one. The Schusterman filing comes with ten separate schedules, itemizing the foundation's investments, its land, its loans from officers, its other expenses, its other income and its taxes. None of them is a grant list, and the same is true of all six. Whether the list survives anywhere else is a question we come back to at the end of this section, because the answer is not what we assumed when we started.

What those filings do itemize is worth reading twice. The Aviv Foundation's return breaks out its accounting fees, its legal fees, an explanation of its employee compensation and its corporate stock holdings, across sixteen schedules, while the $20.6 million it gave away appears as one line with no recipient. The Pritzker Traubert Foundation itemizes its amortization and its depreciation, and reports $120.5 million of grants as a single unnamed entry. These are careful filings, detailed about everything except where the money went.

Nobody here broke a rule. A foundation may attach its grant list as a separate document instead of typing the recipients onto the form, and it may attach that document as a PDF. The catch is at the other end: when the IRS publishes these returns as data, it leaves the PDFs out. Grantmakers.io, which builds foundation profiles from the same IRS files, prints a note where the Pritzker Traubert grant list should be: "The IRS does not currently provide attachments in the electronic dataset from which this site is based."

The people who work with this data have told the IRS as much. The Aspen Institute told the IRS as much last year, writing that a grant list handed over as a PDF is "not an acceptable alternative" and does not satisfy the 2019 law that made electronic filing mandatory so these returns would be readable. Six years after that law, a $333.5 million grant book can still be filed in a way that satisfies it and discloses nothing.

What fills the gap is whatever the foundation chooses to say voluntarily, and the choosing is the point. Schusterman's website names twenty-two grantees in Tulsa, with no amounts attached to any of them, against a book of $333.5 million. The same site sets out the terms: "All public disclosures of a Schusterman Family Philanthropies grant, contract or any other relationship" require the foundation's "express written permission." A grantee who names its own funder without asking is in breach. The public record says nothing, and the private record is licensed.

The second door: a named recipient that is a wall

The donor-advised fund route is different, and in some ways stranger. Here the filing does name the recipient. It just names an institution whose business is not saying where the money goes next.

Paul Singer's foundation is most of this block. Over the three filings it moved $68.0 million, then $110.0 million, then $161.5 million into the JP Morgan Charitable Giving Fund, a donor-advised fund administered by National Philanthropic Trust. In the latest year that single grant was four dollars of every five the foundation paid out.

National Philanthropic Trust files a public grant list, and it is useless for this purpose. NPT held $56.1 billion in assets at the end of its June 2025 year and took in $17.8 billion of revenue, so Singer's $161.5 million entered that pool and became indistinguishable from it. Every grant NPT makes is disclosed and not one of them can be traced back to the donor who advised it, which is the design working as intended rather than a defect in it.

Two features of the law make this route attractive beyond privacy. A grant from a private foundation into a donor-advised fund counts toward the foundation's mandatory five percent annual payout, and the fund that receives it has no payout requirement of its own. Money can therefore satisfy a legal obligation to give it away and then sit. This is not a niche practice: one analysis of IRS data found private foundations sent $2.6 billion to national donor-advised funds in 2021, up from about $1 billion in 2018, and that for 167 foundations such grants made up their entire qualifying distribution for the year.

The third door, and the one piece we could follow

The smallest block is the only one with a floor under it. When a foundation gives to another foundation, the receiving one has to file its own return and list what it does next. The second hop is public by law.

So we followed the largest one. Alongside the $161.5 million that went into the donor-advised fund, Singer's foundation sent $25.0 million to The Singer Family Charitable Foundation, a separate family vehicle in Delaware. That foundation's own filing for the year it arrived shows $28.4 million of revenue coming in and $3.7 million of grants going out. Its assets went from $3.1 million to $14.9 million.

So of the $25 million that left the first foundation's books as a grant paid, about $3.7 million reached working charities that year and roughly $11.8 million became the second foundation's balance sheet. The law is stricter here than for the donor-advised fund route: a grant to another non-operating private foundation does not automatically count toward the payer's five percent unless the recipient elects to treat it as distributed out of corpus and passes it on. The money is neither lost nor hidden, just slower than the word "granted" makes it sound, and this is the only one of the three doors where a reader can find that out.

The diagram below follows both paths. The black bar at the left is everything the foundation paid out, and each ribbon peeling off it is a grant, drawn to its size. The hairline down the right is where the public record stops.

Put the three doors back together and the trend inside the block matters as much as its size. The paperwork share fell from 84 percent of the dark money to 71 percent across the three filings, and not because the attachments shrank; they grew from $491.7m to $587.5m. It fell because the other two doors grew faster. The donor-advised fund share went from 13 percent to 23 percent, and the foundation-to-foundation share from 3 percent to 5 percent. The one door a reader can see through is the smallest of the three, and the two that close completely are where the growth is.

None of this is a scandal in the legal sense, and it is worth being precise about that. Every foundation named here filed what it was required to file. The problem is that the requirement, applied to some of the largest charitable fortunes in American Jewish life, produces a public record in which a third of a billion dollars can move without naming a single recipient, and in which the largest single grant anywhere in these books is a transfer into an account whose purpose is to end the trail.

One route we have not closed

First the correction. When this section went up it said flatly that the attachment is not there, and that is true of the data the IRS publishes: we checked every schedule filed with each of the six returns and none is a grant list. It may not be true of the scanned image of the return, which we have not been able to get hold of. Candid, which assembles foundation grant data from these filings, looked at 521 affected returns and found the grant list missing from both places in 212 of them and missing from the data but sitting in the scan in the other 309. So roughly three in five of these cases are recoverable by someone willing to read a PDF, and we have not closed that route on our six. The real recovery figure could be higher than three cents, and finding out is the next piece of work rather than a caveat to wave at.

At least eight cents of every American foundation dollar goes somewhere unnamed

Everything above describes fifty foundations and says nothing about whether a quarter of the money going dark is remarkable or ordinary, which a reader is entitled to know. A national baseline exists, in the same comment letters that flagged the problem to the IRS.

In 2021, at least 1,957 American foundations gave away more than $5 billion without naming who got it, according to Candid, which tracks this. American foundations gave $90.88 billion in total that year, so that is at least 5.5 cents in every dollar disappearing the same way Schusterman's does. Another $2.6 billion went into donor-advised funds, which is 2.9 cents more. Together, at least 8.4 cents of every American foundation dollar goes somewhere unnamed.

Both national numbers are floors. They count the cases somebody has already found, and the donor-advised fund figure leaves out community foundations and Jewish federation funds, which our fifty use. The real national rate is higher than 8.4 percent, so the gap below is narrower than it looks.

In the chart below, further right means less disclosure. Notice that in both panels every dot sits to the right of the national line.

Both things are true, then. The practice is national and legal everywhere, and these fifty did not invent it. They also run well ahead of the country on both routes. This particular set of donors, who hold a large share of the money that funds American Jewish institutional life, discloses less of it than American foundations generally do.

The comparison settles one more thing. Those national figures exist because people noticed the same problem years ago and told the IRS about it, in formal comments with counts and dollar figures attached. Nobody has to discover that this happens. The law that was supposed to make these returns readable is now six years old.

Political money is the one thing a foundation legally cannot give

The question arrives as soon as you read the names. Some of the fifty sit among the largest political donors in the country, and none of that money appears anywhere in this piece. We did not miss it. It cannot be there.

A private foundation is barred from political campaign spending, and the bar has teeth. The tax code puts an excise tax on a foundation's expenditures for political campaign intervention, and it taxes lobbying too, both the direct kind and the grassroots kind. Public charities get some room to lobby. Private foundations get a tax bill. So the campaign money moves through vehicles that have nothing to do with a foundation return: personal checks to candidates and to super PACs, which the Federal Election Commission publishes by name, and 501(c)(4) social welfare organizations, which may endorse candidates and pay for independent expenditures and which never publish who funded them.

The families describe the split themselves. Schusterman's own account of its structure says the Philanthropies covers "the charitable grantmaking and programmatic work of the Charles and Lynn Schusterman Family Foundation, as well as the advocacy activities supported by the Schusterman family." The advocacy sits beside the foundation rather than inside it, which is exactly why it sits outside this piece.

So there is a fourth kind of invisibility here, and it is not the kind Hop eleven described. Those three were gaps inside a system built to disclose. This one is a different filing system altogether, and it cuts both ways. Federal campaign contributions are itemized by donor name and published, which makes them more public than anything in these grant books. A 501(c)(4) donor is more private than anything in them.

What the charitable record does show is the edge of the same argument

Money for public argument does appear in these books, as long as the organization receiving it is a charity. Two of the nineteen categories hold most of it: policy and civic organizations on the secular side, and Jewish advocacy and defense on the other. Together they came to $99.0 million in the latest year, about three cents of every dollar the fifty gave.

That figure is falling, which surprised us. Across the three filings the two categories ran $102.6 million, $118.7 million, then $99.0 million, so the latest year is the lowest of the three in dollars, in a book that grew 8.2 percent. As a share it went 3.8 percent, 3.8 percent, 2.9 percent.

The names inside those categories are the recognizable ones. The Anti-Defamation League drew checks from sixteen of the fifty, more than any other advocacy body in the set, for $6.1 million. The American Jewish Committee drew eleven, for $0.7 million, which is a lot of funders and very little money. The Milken Institute took $41.8 million across the three years from four of them, and the Maccabee Task Force took $26.5 million across the three years from exactly one. Smaller sums went to the Foundation for Defense of Democracies, $6.9 million in the latest year, and to PragerU, $3.0 million.

One entry is worth naming on its own, because it sits closest to the line. The American Israel Education Foundation took $9.0 million from these fifty in the latest year. It is the charitable organization affiliated with AIPAC, and among other things it pays for the educational trips to Israel that members of Congress take. It is a 501(c)(3), the money is a legal charitable grant, and it funds the educational arm of a lobby. That is the sharpest form the visible half of this takes.

What we cannot tell you is the other half. We have not matched these families against campaign finance records, so this piece makes no claim at all about how much any of them gives politically, or to whom. That is a different dataset and a different piece, and anyone who tells you the foundation filings answer it is reading the wrong document.

The Wexner Foundation gave $177,334. The Wexner money gave $42 million.

The three-year panel settles how much a single snapshot can be trusted in aggregate. It does not settle three specific traps that survive any number of years, and the three panels below show one example of each. In each panel, the two bars are two numbers that sound like the same thing without being it.

The left panel is the brand problem. Ask anyone in Jewish organizational life to name a great philanthropic name and Wexner comes back fast, yet the Wexner Foundation's own filing shows $177,334 in grants paid, because it is an operating foundation that runs fellowships rather than writing checks. The paying entity is the Wexner Family Charitable Fund, two lines away in the IRS file at $42.0 million, including $10 million to Ohio State and $10 million to the Wexner Foundation itself.

The middle panel is the accounting problem. Mandel booked $214.9 million of grants in fiscal 2024 but moved only $77.0 million in cash; the rest is commitments recorded for future years. The right panel is the estate problem, and the dossier already gave it away: while the Marcus Foundation was paying out $227.5 million, Bernie Marcus and his estate were pouring in $538.6 million. A single-year snapshot catches this foundation mid-transformation, so its rank today says little about its rank in five years. Rales runs the same dynamic in reverse, emptying rather than filling: $131.8 million out of a $53 million foundation, $122.7 million of it in one grant to Carnegie Mellon.

Mostly to places you can name. Of $3.37 billion in the latest year, $2.54 billion is traceable to about 5,100 specific organizations, and the tracing holds up: 42 of the 44 readable books reconcile to the dollar. Of that traceable money, two dollars in three leave the Jewish world altogether, for universities, hospitals, and museums. American higher education is the largest single category at $514.3 million, ahead of Israel and Israel-facing giving at $429.9 million. Eighty-six percent of the recipients are funded by exactly one of the fifty, and thirty-three checks moved more money than the other eight thousand five hundred and seventeen. Twenty-six of the fifty took in no new money this year at all, which means the checkbook now runs on whatever the donor left behind.

Three consecutive filings say the shape is not a fluke. Total giving rose 23.4 percent across the panel, the Jewish share of traceable dollars moved under three points, the shared center held at 137 organizations give or take one in all three years, and the gap between secular universities and Jewish schools closed sharply before reopening a little. Underneath that steady surface seven of the fifty more than doubled their giving, so the aggregate holds while any individual ranking does not.

Then there is the money that steps out of public view at the first hop: 21 cents of every dollar in the first year of the panel, 20 in the second, 24 in the third. We went after it and recovered $25.0 million of $822.7 million, three cents on the dollar. Most of the rest sits either in an attachment the IRS does not publish as data or in a fund whose disclosures are complete and whose attribution is impossible. The most complete public map of Jewish mega-giving in America has a hole in it the size of its second-largest donor, and over three years the hole grew faster than the map. Nothing about that hole is an accident or a violation. It is simply what the disclosure rules permit, and any donor who prefers not to be read has an easy and legal way out.