USAspending records $6.28bn obligated for work performed in the Philippines between FY2018 and FY2025. $3.87bn of it is Social Security and veterans' benefits paid to people who live there; $982.2m is foreign assistance. And asking the same question a different way moves USAID's figure by 9.0×, which is the part worth reading twice.
"US money in the Philippines" is usually read as aid. The register does not agree. 61.4% of it is direct payments — money sent to individuals rather than spent on programmes or contracts.
of the total comes from the Social Security Administration — $3.85bn. Its largest single programme, Social Security Retirement Insurance, is $3.00bn on its own.
is contract spending — $1,610.6m, most of it Defense. Grants, the instrument most people picture when they think of aid, are $734.9m.
across every agency whose Philippine programmes are foreign assistance — USAID, State, and the food, health, labour and justice programmes run by departments that mostly do something else.
USAspending can be filtered on where the work is performed or on where the recipient is registered. Both are reasonable readings of "money going to the Philippines", and they are not close.
$811m of USAID work is performed in the Philippines; $90.6m is paid to recipients registered there. The rest goes to implementing partners registered in the United States.
Social Security reads identically under both filters, because a pension is paid to a person at the address where they live. That is why the benefits figure on this page is the one number no choice of filter can argue with.
$6.28bn by place of performance against $4.45bn by recipient location. Picking one silently would have been the easiest mistake on this page to make.
This matters beyond the arithmetic. A page that had filtered only on recipient location — the more obvious choice, and the one the API documentation reaches for first — would have found $90.6m of USAID money and published a comparison between pensions and aid that was out by most of an order of magnitude. Neither filter is wrong. They answer different questions, and the difference is largest for exactly the agency the headline is about.
Everything else on this page uses place of performance, because "money spent
on the Philippines" is the question being asked. Both readings ship in
data/us-ph-money/us_ph_basis.csv.
The programme names settle any remaining ambiguity about what this money is. Three of the four largest are Social Security lines.
Social Security retirement, survivors and disability come to $3.84bn between them. USAID's foreign-assistance line is $635.3m, and the two food programmes — Food for Progress and Food for Education, both run by Agriculture rather than by an aid agency — add $66.2m.
The obvious next chart is the total by year, and it rises steeply through the last decade. Almost none of that rise is spending.
and not before, as reporting requirements widened. It did not start paying Filipino pensioners that year; it started disclosing it here.
It reported up to $304.3m in a single year across 7 years — $1.44bn in total — and then stops. It did not stop paying Filipino veterans.
sits before the window this page uses, and is not comparable with what follows in either direction. Every headline here is computed on FY2018 onward, and is still missing whatever Veterans Affairs no longer files.
The contract side is a different country from the pension side: logistics, fuel, ship repair and base support, most of it Defense.
to RESEARCH TRIANGLE INSTITUTE — a US research institute, not a Philippine organisation. It is the clearest single illustration of why the two readings differ.
of contract spending is classified as Facilities Support Services. Warehousing, fuel and ship repair follow it.
of the money sits under a placeholder recipient, because agencies report aggregated individual payments rather than naming pensioners. Correct practice, and a hard limit on reading the recipient table as a list of organisations.
Four limits, and the first one bounds every figure on the page.
The API returns null for outlays under this filter, so every number here is an obligation: money the government committed. For a monthly pension the two are nearly the same. For a multi-year defence contract they are not.
Place of performance and recipient location differ by 1.4× overall and 9.0× for USAID. This page uses place of performance throughout and publishes both, rather than presenting one as the answer.
Sorting agencies into benefits, assistance, defence and other is this page's reading, taken from each agency's own Philippine programme lines. It ships as a column, and the build fails if an agency above half a per cent of the total is left unclassified.
62.2% of the money goes to a placeholder recipient standing in for individual people. There is no way to see who, and there should not be.
Every figure on this page traces to one of these, through a CSV in
data/us-ph-money/. Each is checked against its source query on every
build.
Federal obligations to recipients located in the Philippines for every fiscal year from 2008 to 2025, split by award type, and the same series filtered to each of the five agencies followed individually — Keyless: no account, no token, no form. POST-only with a JSON filter body, which is the whole shape of this API — the category is a path segment and the filters are the payload. The award-type split is read from the fields this response already returns rather than from six separately filtered queries, because award_type_codes has no bucket for indefinite-delivery vehicles and those queries did not sum to the total The link is the endpoint's published contract rather than the endpoint itself, which answers 405 to a browser because it is POST-only.
Which agencies, which programmes, which recipients and which industries the money reaches, over the headline window — The same POST body against five different path segments. Each returns a ranked slice rather than the full list, so every table here is a top-N and the checks assert that a partial table never claims more than the whole The link is the endpoint's published contract rather than the endpoint itself, which answers 405 to a browser because it is POST-only.
The licence this page relies on: a “government work is something created by a U.S. government officer or employee as part of their official duties” and such works are “not subject to copyright protection” — The caveat that shaped the other federal sources in this series — that a government site may carry third-party content — barely applies here. Obligation amounts, agency names, programme names and recipient names are all facts agencies report about their own spending, and no descriptive prose is fetched or published
Would have provided the award-level detail behind every figure here as a single archive — The documented path at files.usaspending.gov/database_download/ answers 404, and the custom download builder produces a file asynchronously rather than at a stable URL. So the API is not a convenience on this source, it is the only reproducible way in — which is why every figure here is an aggregate the API computes rather than a sum over rows
Would have shown money actually paid out rather than committed — The API returns null for total_outlays under this filter, so every figure on this page is an obligation: money the government committed. For Social Security benefit payments the two are close; for a multi-year defence contract they are not, and the page says so rather than letting “obligated” be read as “spent” The link is the endpoint's published contract rather than the endpoint itself, which answers 405 to a browser because it is POST-only.
Would have named who receives the 86.7% of this money that is reported under a placeholder recipient — Not available and correctly so. Agencies report aggregated individual payments under stand-in names rather than naming pensioners, so the largest “recipient” in the data is a placeholder. That is good privacy practice and a hard limit on reading the recipient table as a list of organisations, which the page states rather than works around
I wrote a companion post — “Most American Money In The Philippines Is Pensions, Not Aid” — in simple, everyday words.