Central banks have spent four years buying gold. This week one of them spent the summer moving it.
On September 2, De Nederlandsche Bank said it had transferred about 86 tonnes from vaults in New York and Ottawa to London. Governor Olaf Sleijpen put the motive in one sentence: improve tradability, strengthen resilience and preparedness, expect never to need the metal, refuse to leave it where it cannot be used quickly in a crisis. Finance minister Eelco Heinen said the plan was withheld until completion because it was vital public interest. That is how official gold is handled when the topic is no longer accounting.
The attached market note framed the episode as “crisis preparedness” plus a Goldman warning on geographic concentration. Both halves are real. Neither half is a signal that Canadian mines will sell more ounces next quarter. The signal is about where a bar has to sit if a reserve manager wants to pledge it, swap it, or sell it before a weekend becomes a political event.
This is not investment advice. Official gold does not trade like a junior. It also does not stay put when the legal and logistical map changes.
What DNB Actually Did
The Netherlands holds 612.4 tonnes. Before the move, the split was roughly 30.8% at the cash centre in Zeist, 18.1% in London, 31.3% in New York, 19.7% in Ottawa. After the move Zeist was unchanged at 30.8%. London rose to 32.1%. New York and Ottawa were each cut to 18.5%. Bloomberg’s table in circulation this week matches the bank’s own percentages.
Eighty-six tonnes did not all cross the Atlantic as bars. DNB’s statement is more precise than the headlines. About 59 tonnes in New York were sold and replaced with London-standard metal in London. More than 27 tonnes were shipped physically from the United States and Canada to Zeist, and a similar quantity of good-delivery metal went from Zeist to London so DNB would not have to remelt bars. Combining sales, purchases and transport, the bank said, spread the operational risk of a relocation that ran from March to August.
Some reports parsed the North American legs as about 78 tonnes out of New York and about seven out of Ottawa. The official wrap is 86 tonnes from the combined U.S. and Canadian stock to London. Either way, Ottawa was not a rounding error. It was a custody node that just lost share so London could gain it.
Why London? DNB’s public answer is liquidity. Gold at the Bank of England meets international trade standards and is “regarded as the world’s most easily tradable gold.” Bars in the U.S. and Canada, the bank said, could not be used as quickly or directly in a crisis. Sleijpen’s other line is the one reserve managers will quote: keeping a larger share in London “strengthens the function of gold as an anchor of trust.”
France Already Ran the Sale-and-Replace Play
DNB is not the first large European name to shrink New York. Banque de France sold 129 tonnes stored at the New York Fed — about 5% of its stock, bars that did not meet current LBMA-style standards — and bought high-specification metal in Europe. Then-governor François Villeroy de Galhau said the residual New York parcel was not the right standard, that a long risky haul was worse than a sale-and-buyback, and that the operation was not a political gesture. The sale booked about €11 billion of capital gain in 2025. France’s stated gold stock stayed 2,437 tonnes. The metal changed address and quality. The headline tonnes did not.
That distinction matters. “Repatriation” in the French case was mostly a quality and location upgrade executed through the market. “Relocation” in the Dutch case is a liquidity upgrade executed through the market plus a smaller physical lift. Neither announcement said the Federal Reserve had lost a bar. Both said North American storage was the wrong place for metal you might need to trade tomorrow.
Germany Has Not Joined the Queue
The Bundesbank still holds on the order of 3,350 tonnes, with about 1,236 tonnes — roughly 37% — in New York. President Joachim Nagel has rejected the idea that those bars are in legal jeopardy. He has said he has no doubt the gold is safely stored at the New York Fed, and that the United States would hurt itself most if it put that legal status in question and thereby risked confidence in financial markets.
Advocacy groups in other EU countries want metal home. Michael Jäger of the European Taxpayers Association has argued Germany should bring gold back and has used language about U.S. unpredictability. That is politics. Nagel’s sentence is the official German position until it isn’t. A Canadian reader should not flatten “the Dutch moved Ottawa metal” into “every European bar in New York is leaving.” The map is fragmenting, not emptying in one night.
Goldman: Location Is Now a Reserve-Management Variable
Lina Thomas’s commodities work, which Goldman has been circulating around the same $4,900 year-end target, puts vault geography on the same page as official buying. The location of central-bank gold, she notes, is increasingly top of mind. A World Gold Council survey used in that research has the Bank of England as the most preferred custodian — on the order of 57% of reserve managers in the 2026 reading — with the New York Fed still important because it sits on the main settlement networks for swaps, leasing and immediate access. Domestic storage, the BIS, Banque de France and, increasingly, China show up as the rest of the menu.
The trade-off is the one Venezuela illustrated at the BoE in 2018: freezing or restricted access. Survey charts Goldman published show more banks planning to diversify overseas storage, not simply haul everything home. Full repatriation is expensive for smaller institutions and replaces one concentration with another. The emerging pattern is barbell custody — London for tradability, New York for market access, a domestic pile for politics — rather than a single fortress.
Thomas’s demand nowcast is the other half of the same note. Official buying in June ran about 57 tonnes, or about 100 tonnes on a three-month seasonally adjusted basis, against a pre-2022 average near 17 tonnes, with China the largest identifiable buyer. A 32-tonne inflow of monetary gold into London can look like a custody transfer rather than fresh buying if Bank of England official holdings rise by a similar amount. Goldman’s year-end $4,900 case still assumes roughly 50 tonnes a month of official demand in 2026 and 40 tonnes in 2027, anchored on emerging-market diversification after the 2022 freeze of Russian reserves. Vault-shifting does not cancel that bid. It changes where the bid is stored.
Why Ottawa Is the Canadian Line
Canadian Mining Report readers do not sit on the New York Fed’s gold committee. They do sit in a country that has long stored other people’s bars in Ottawa, next to a trade fight with the United States that DNB’s timing will be read against even if Sleijpen never mentioned tariffs.
Ottawa’s share of Dutch gold went from 19.7% to 18.5%. That is not a run on the Bank of Canada. It is a European reserve manager deciding that crisis liquidity lives in London, not in a North American vault whose legal and logistical path to the loco-London market is longer. If other mid-sized European holders copy the sale-in-New-York, buy-in-London template, Ottawa’s role as a diversification site shrinks at the margin. If they copy only the physical lift to Zeist, Canada still loses a customer for storage, not a customer for ounces.
Neither outcome feeds a Canadian mine. Custody fees and sovereign confidence are not mill throughput. The second-order effect is narrative: official gold that will not sit still is official gold that is still being treated as a reserve asset, not as a relic. That narrative already supports the structural bid Goldman, RBC and Schroders have been writing up. It does not tell you whether Friday’s payrolls print takes spot through $4,500.
What This Is Not
It is not proof that New York gold will be seized. Nagel said the opposite about German bars. It is not proof that France “fled” the Fed; France said the bars were the wrong standard and booked a gain. It is not a 86-tonne addition to world investment demand; a large piece was a location swap. It is not a reason to buy a TSX gold stock before the next FOMC.
It is evidence that reserve managers now price three risks at once: price risk, credit risk of the reserve currency, and jurisdiction risk of the vault. The third item used to be a footnote for everyone except the banks that already brought metal home in the 2010s. DNB just made it a press release.
Price Context, Without the Slogan
Spot gold this week has been rebuilding above $4,400 after a mid-July low and a Warsh-week fade, with some desks noting a bounce toward the 200-day average as ETF and options demand flickered back when hike odds softened. Goldman’s $4,900 year-end target still sits above the market. DNB’s 86 tonnes, at $4,400, is on the order of $12 billion of metal changing address. That is a lot of logistics. It is not a year of official buying by itself.
If more banks choose London good-delivery over North American storage, loco-London tightness can appear even when the global stock is unchanged. That is a basis and lease-rate story. Miners get paid loco whatever their offtake says, not loco a Dutch press release.
Conclusion
The Netherlands moved gold out of New York and Ottawa because London is where a bar can be used. France already sold the wrong bars in New York and bought the right ones in Europe. Germany says New York is still safe. Goldman says the map of vaults is now a reserve-management problem and still wants $4,900 on 50 tonnes a month of official demand.
For a Canadian audience the new fact is smaller and sharper: Ottawa just became a slightly less important drawer in someone else’s crisis plan. That is how de-dollarization looks when it is not a slogan. It looks like a custody change announced after the metal has already left.
Important information
This article is for informational and educational purposes only. It is not investment advice or a recommendation to buy or sell gold, mining equities, or any other instrument. Figures for DNB, Banque de France, the Bundesbank, and Goldman Sachs Research are taken from those institutions’ public statements and from contemporaneous reporting; they can be revised. Relocation of official gold is not the same as new official buying. Forward-looking price targets, including Goldman’s $4,900 year-end 2026 case, are scenarios and may prove incorrect. Consult a licensed adviser. The author and publisher accept no liability for actions taken on the basis of this article.