CharitySense Data

What the foreign-grants numbers actually count

The same foreign-activity form holds grant amounts and investment values. We read every available latest full return, then isolated one filing year to see how the different numbers relate. The answer changed the question.

By the CharitySense Data team

Published Updated

$46.5B
itemized foreign grants in current latest full returns for tax year 2024
277,653
latest full returns in the 2024 cohort
$221.0B
investment-only rows, reported as book value
9,202
filers with itemized foreign-grant rows
41.2%
of grant-row dollars from the ten largest filers

01/One column holds two

One column holds two kinds of money

We began with a deceptively simple question: how much money do American nonprofits send overseas? Schedule F seemed to offer an answer. Add its regional amounts in the current tax-year-2024 full returns and we get $330.9 billion. But the IRS instructions give that column two jobs. It records expenditures for foreign activity and the year-end book value of investments held through foreign entities.

Rows labeled only as investments add to $221.0 billion, or 66.8% of the column total. A value held at year-end is different from a dollar spent during the year. The old version of this article treated that combined column as a spending headline and subtracted investments to reach a grant estimate. That arithmetic was wrong. We have removed it.

What the investment label means
The IRS tells filers to report investments separately and enter their total book value at the end of the tax year. Some filed activity descriptions combine investments with another activity. We count only rows labeled solely as investments for the $221.0 billion figure, leaving mixed descriptions in the broader Part I total.
Four filing measures, one 2024 latest-return cohort

Dollar billions. These bars overlap and must not be added. Select one to see what it counts.

Source: CharitySense analysis of current latest full Form 990 filings for tax year 2024, read September 28, 2026. See Behind the Data for counts and limits.

02/grant rows answer

The grant rows answer a narrower question

The grant detail sits elsewhere on Schedule F. Parts II and III list assistance for foreign organizations and individuals, including certain domestic intermediaries designated for foreign beneficiaries. Across 9,202 filers in our 2024 latest-return cohort, their 84,996 rows add to $46.5 billion. About $35.5 billion is recorded as cash grants; the remaining $10.9 billion is noncash assistance.

$46.5B
itemized foreign assistance
$35.5B
cash grants listed
$10.9B
noncash assistance listed
$38.8B
separate main-return foreign-grants line

Even this figure needs a careful verb. A filer using accrual accounting may report the present value of a grant payable in future years. The filing records an amount granted, rather than confirming when a recipient received it or what happened afterward. We cannot turn $46.5 billion into “money that reached people” from these forms alone.

The main Form 990 has its own foreign-grants expense line, totaling $38.8 billion across this same cohort. Among 8,848 filers with both an itemized grant table and that expense line, 6,010 match to the nearest dollar and 2,838 differ. We keep the two measures separate. Reporting thresholds and accounting treatment can matter. The aggregate gap by itself does not tell us which explanation applies to a particular organization.

03/Concentration survives a full

Concentration survives a full count, but looks different

Earlier drafts used repeated random samples to describe a sector supposedly dominated by three foreign-activity filers. The whole available 2024 latest-return cohort tells a calmer story. Among filers with a positive Part I total, the ten largest account for 9.0% of the sum, while the top three account for 3.1%. Sampling a rare giant can make its share look enormous when most other filers are left out.

Grants are more concentrated. The ten largest filers represent 41.2% of the positive itemized grant amounts; the top three represent 23.6%. Those shares invite us to look closely at a few large programs, but they do not justify the old claim that ten filers move two-thirds of all foreign grants.

Counting the full available cohort changed the story more than choosing another sample did.

04/Investment domicile and grant

Investment domicile and grant region point in different directions

The geography on these forms is easy to misread. Of the investment-only amount, $146.5 billion, or 66.3%, is filed under Central America and the Caribbean. The IRS assigns an investment's region by the foreign entity's legal domicile. That region does not describe where a charity's services were delivered.

The largest regional group in the itemized grant rows is instead Sub-Saharan Africa: $13.9 billion, or 29.9% of the grant-row total. For grants to an organization, the region can describe its principal foreign office or, if it has none, where the funds will be used. We grouped recognizable variations of the IRS region names, but $2.7 billion in grant rows still has a missing or ambiguous region. A region label is not a country-level destination, much less an impact measure.

This distinction also limits a new kind of analysis. Our map of named organizational money relationships can show many domestic grant connections. The IRS instructs Schedule F filers to leave the recipient-name columns blank for foreign organization grants. We cannot responsibly fill those names from a regional amount, so the foreign-grant rows remain a geography-and-purpose view rather than a recipient network. Our grantee diligence essay explores what a named relationship can and cannot establish.

05/Behind the Data

Behind the Data

We read the current latest full Form 990 for each of 485,109 organizations on September 28, 2026. To avoid adding different tax years together as if they were one year's giving, the main figures here use only the 277,653 organizations whose latest full return has tax year 2024. Another 72,261 already have a later full return, and some organizations have an older latest return. Thus our 2024 group is a defined current-latest cohort, not every 2024 filing ever received or an estimate of all American foreign giving.

Part I is the sum of numeric regional amounts on 39,248 rows from 12,268 filers in that cohort. Its investment figure uses only investment-only activity labels. The grant figure adds cash and noncash amounts in 84,996 Part II and III rows from 9,202 filers, retaining filed negative adjustments. The separate main-return foreign-grants number sums that filed expense line, including the 10,052 filers with a positive value. We compared the two grant amounts filer by filer and treated rounded-dollar agreement as a match. Concentration ranks positive filer totals within the same cohort, with their positive amounts as the denominator. For geography we grouped obvious variations of published IRS region names; missing and ambiguous labels stay unassigned. We also checked our named-organization relationship map, which cannot turn Schedule F's unnamed foreign recipients into traceable endpoints. We did not infer countries or foreign recipient identities.

The filing's activity description is supplied by the organization. Recent years are still arriving, and older latest returns remain in the available population. Schedule F's grant-detail rules include thresholds and accrual reporting, so row totals cannot prove cash delivery or impact. Dollar billions are rounded independently to one decimal place; displayed cash and noncash amounts may differ from the displayed total by $0.1 billion. These numbers describe filings, not a judgment about any organization.

Sources

Ask the research team

What would you want us to pull from the filings? Tell us, and a future piece may answer it.

Final Thoughts

The old $40 billion answer came from scaled samples and a subtraction the form does not support. The current filings offer a firmer, narrower observation: $46.5 billion in itemized foreign grants among organizations whose latest full return is for tax year 2024. The broader foreign-activity column also contains year-end asset values; the separate expense line measures something else again. Each answers a different question.

We would like to understand how these patterns change across complete filing vintages and which kinds of assistance are hardest to see in the public record. For a wider view of nonprofit money, read in or out: where the money goes and the shape of nonprofit money.

About the author

The CharitySense Data team

Research team

We read public nonprofit filings and other sources to understand how the sector works. This blog shares what we find, along with the limits of the evidence. Each article explains how its figures were measured and cites the records or primary sources behind them.