Why Belgium Holds More U.S. Debt Than Saudi Arabia, and What That Actually Means
Belgium holds $477.3 billion in U.S. Treasury securities as of December 2025, ranking it fourth among all foreign holders of U.S. federal debt, ahead of Canada, France, and Taiwan. Belgium’s GDP is roughly $600 billion. The idea that a country with an economy that size has organically accumulated nearly half a trillion dollars in U.S. sovereign paper, driven by domestic institutional demand, does not survive casual scrutiny. Belgium’s appearance near the top of the foreign holdings table is not evidence of Belgian appetite for Treasuries. It is evidence of how Euroclear, the Brussels-based international securities settlement system, processes custody of assets held by clients from dozens of other countries.
Treasury’s country-level data track the location of the asset — specifically, the geographic location of the custodian — not the nationality of the beneficial owner. A Chinese wealth fund, a Middle Eastern sovereign vehicle, a Swiss private bank, or a Korean pension fund that holds U.S. Treasuries through Euroclear’s infrastructure appears in the Belgian figure. This is not a flaw in Treasury’s methodology so much as a structural feature of how global securities markets work. Custody and ownership are legally distinct, and the data reflect custody location because that is what can be systematically observed across international settlement systems.
The same logic explains Luxembourg at $435.1 billion (fifth place), the Cayman Islands at $421.2 billion (seventh place), and Ireland at $340.7 billion (ninth place). None of these jurisdictions has a domestic savings pool remotely large enough to account for their ranked positions. All four are major financial centers or fund domiciliation hubs whose beneficial owners are distributed globally. The Cayman Islands position, in particular, reflects the hedge fund and alternative investment vehicle industry concentrated there. Ireland’s position reflects the domiciliation of large numbers of UCITS funds and other investment vehicles that are managed elsewhere but registered in Dublin.
This attribution problem is consequential for the analysis most often applied to foreign holdings data. The standard geopolitical read of the table — Japan and China at the top, therefore these are the strategic relationships that matter — is broadly correct at the top of the distribution where the distortion is smaller. Japan’s $1.185 trillion and China’s $683.5 billion do reflect substantial real holdings by entities with a genuine nexus to those countries, even if some portion flows through intermediary custodians. The attributional noise is larger in relative terms for mid-table entrants like Belgium and Luxembourg, where the domestic entity interpretation would be almost entirely wrong.
It also means that trends in country-level figures are harder to interpret than the headline movement suggests. When Belgium’s holdings rise, it does not mean Belgian institutions have become more enthusiastic about U.S. debt. It may mean that third-country investors who custody through Euroclear have increased their positions. When a financial center’s holdings fall, it may reflect outflows by clients from any number of countries, none of which will appear in the bilateral headline data. Tracking those movements requires access to beneficial ownership data that is not publicly released.
CRS notes this explicitly in its report: some of the largest apparent holders are international financial centers whose clients are presumably mostly from third countries. The footnote example — a Chinese investor custodying through a Belgian bank whose assets appear under Belgium, not China — is illustrative but understated. The scale of the attribution distortion across the entire top-ten list is substantial.
The practical implication is that country-level foreign holdings data are more useful for tracking broad directional trends among the top three or four genuine holders than for constructing precise bilateral exposure maps. The U.S. Treasury market is globalized in ways that its own reporting infrastructure cannot fully capture. The $9.2 trillion foreign holdings total is a reliable number. The detailed country table beneath it is a useful approximation with known and unquantifiable distortions built in at every level below the top line.