Falling dollar reserve share doesn’t prove sovereign Bitcoin buys
NY Fed analysis (Sept. 2, 2026) found a lower dollar share in official reserves does not establish that central banks bought Bitcoin. Evidence requires disclosed allocations, funding sources and executed purchases.
The New York Federal Reserve’s Sept. 2, 2026 analysis of IMF COFER data found that a decline in the dollar’s share of global official foreign-exchange reserves does not prove central banks purchased Bitcoin. The dollar’s share fell from 64% at year-end 2015 to 56% at year-end 2025.
Researchers Linda S. Goldberg, Oliver Hannaoui and Sneha Parthasarathy break the change into two channels: shifts in countries’ currency preferences within their portfolios and changes in the total size of individual countries’ reserves that change each country’s weight in the global average.
The report shows a small number of large reserve portfolios can move the aggregate share. For 2015–2019, 76 countries with complete endpoint data contributed minus 1.2 percentage points from preferences and minus 1.5 percentage points from reserve-size changes. For 2019–2023, 62 countries with complete data contributed plus 0.3 percentage points from preferences and minus 0.5 percentage points from reserve-size changes. Data for China, Russia, Mexico and Morocco were missing for 2023.
Under assumptions that match the observed 2.3-percentage-point global drop, the researchers inferred a combined negative 2.0-percentage-point contribution from preferences for a broad group of countries. The report notes that aggregate reserve-share statistics do not identify which assets replaced dollars.
Staff Report 1087, first issued in March 2024 and revised in February 2026, separates reserves held for liquidity from an investment portion above those needs. The report models liquidity needs with measures such as short-term external debt or three months of imports and finds that when reserves exceed those measures, countries are more likely to diversify holdings beyond liquidity buffers.
The New York Fed’s analysis does not measure or estimate any purchases of Bitcoin by sovereign entities, nor does it model potential price effects on Bitcoin.
The authors cite a public example: on Nov. 13, 2025, the Czech National Bank announced a $1 million digital-asset test portfolio that included Bitcoin, a dollar stablecoin and a tokenized deposit and explicitly stated the test amount was outside its international reserves.
The report sets out the evidence required to demonstrate sovereign Bitcoin demand: a disclosed allocation to Bitcoin, identification of the funding source for that allocation, records of executed purchases and an official classification that holdings are part of international reserves rather than separate investments. “A shrinking dollar share supplies none of those details,” the paper states.
Articles by this author
