Section V · Geopolitics

The week reserves stopped being safe

February 2022 didn't change gold's chemistry. It changed what a dollar reserve is — conditional — and every reserve manager noticed at once.

In late February 2022, the G7 froze roughly $300 billion of the Bank of Russia's foreign reserves — the assets of a G20 central bank, immobilized in a weekend, with the bulk (~$185–200bn) sitting at Euroclear earning interest now being channeled to Ukraine^[1]^[6]. Whatever one thinks of the policy, its monetary consequence was immediate and is still compounding: a dollar reserve is now visibly a liability of a government that can cancel it. Gold in a domestic vault is the one reserve asset with no counterparty and no off switch. The official sector began buying ~1,000t a year within months, and has not stopped.

Russian reserves frozen

$300bn

Feb 2022 — the precedent that reprices custody risk

CBs planning to add gold

43%

WGC 2025 survey — a record; 24% in 2023

USD share of reserves

56.8%

IMF COFER Q4 2025 — lowest since ~1994

Expect USD share to fall

73%

of surveyed central banks, over the next 5 years

Surveyed intent

The most telling chart in the study

The WGC asks reserve managers each year whether they intend to add gold in the next twelve months. The series — 24%, 29%, 43% — is a straight line up and to the right^[2], and 95% of respondents expect global official gold holdings to keep rising. Intent surveys can be cheap talk; this one has been validated by four consecutive years of realized buying at roughly double the pre-2022 norm. The 2026 edition lands mid-June — a key watch item.

Central banks planning to increase gold reserves

WGC annual reserve-manager survey, % of respondents

De-dollarization, measured honestly

Erosion, not collapse

The dollar is not being replaced — it is being hedged. Its share of allocated reserves drifted to 56.8% in Q4 2025, the lowest since the mid-1990s, but the euro (20.3%) and renminbi (2.0%) are not absorbing the difference^[3] — gold is. BRICS+ economies hold 17.4% of official gold, up from 11.2% in 2019^[5]; India has repatriated 274+ tonnes home since 2023; and the PBOC, with gold still only 9% of its reserves against a 60–70% norm for Western central banks, added at an accelerating pace into the spring (5t in March, 8t in April)^[4]. The arithmetic that matters for the forecast: if China alone were to lift its gold allocation from 9% toward even 20%, at current prices that is years of additional demand at the present global pace. No BRICS currency, payment rail, or treaty needs to succeed for this flow to continue — it is unilateral, already funded, and politically costless for the buyer.

The risk premium leg is smaller and faster: conflict headlines add and subtract a few percent (the January spike rode Middle-East and trade-policy stress). We deliberately weight it least — risk premia mean-revert; custody re-pricing does not. The structural question for 2026–28 is singular: does any plausible political settlement restore unconditional trust in dollar reserves? Unfreezing $300bn and forswearing the tool would be the bear case for gold. No such offer is on any table.

Sources & citations