Gold and the Reserve Illusion

The European Central Bank's June report made the news cycle on Monday: gold has overtaken US Treasuries as the world's top reserve asset. Gold now accounts for 27% of all global central bank reserve assets at the end of 2025, up from 20% a year earlier. Treasuries sit at 22%. The euro lingers at 15%. It reads like the headline event of a generation.

Except - and this is the part I keep coming back to - the mechanics behind the move don't match the drama.

The 7-percentage-point swing in gold's share of reserves came from two things: a massive revaluation of existing gold holdings as the price soared, and a shrinking global reserve pool. Central banks bought 863 tons of gold in 2025, which is solid buying but actually 20% less than the 1,092 tons they accumulated in 2024. At the same time, gold's price climbed roughly 55% during the year, breaking above $4,000 an ounce in October and then hitting a record above $5,500 this January. When the metal you already own doubles in value, your reserve composition shifts on paper even if you haven't moved a single bar.

That distinction matters because it changes what the headline actually tells us. The market tends to interpret this as central banks executing a dramatic, deliberate pivot away from US debt. In reality, the shift is partly an accounting effect - existing gold became enormously more valuable - layered on top of a genuine but quieter strategic reorientation.

The shrinking pool

There's a second mechanical fact the headline obscures. The ECB report flagged a substantial drawdown in the total size of global reserve assets in 2025 - one of the largest in recent years. When the total pool shrinks, the remaining assets can shift their relative percentages sharply without anyone making big new allocation decisions. It's the difference between watching the contents of a deflating balloon rearrange themselves and watching someone deliberately swap the ingredients.

Why did the pool shrink? Part of it is exchange-rate moves - the dollar has been strong, which depresses the dollar-denominated value of reserves held in other currencies. Part of it is genuine selling. Foreign central banks reduced their Treasury holdings during periods of market stress in 2025, notably following US tariff announcements and the escalation of conflict in the Middle East. Turkey alone sold $22 billion of foreign government securities from its reserves in the wake of the Iran war. But the drawdown wasn't universal or catastrophic - foreign official demand for Treasuries remains present, just declined from a cyclical peak.

What this means for the reader is that the reserve system isn't collapsing. It's contracting and rebalancing. The difference feels smaller on Monday morning but is much larger when you're thinking about what happens next.

Why gold, and why now

So if revaluation did much of the heavy lifting, why are central banks still buying - and why gold specifically?

The answer has less to do with portfolio optimization and more to do with what central banks can't do with US Treasuries: they can't rely on keeping them. Since Russia's reserves were frozen in 2022, the concept of a "safe" reserve asset has developed a hidden cost. Treasuries pay yield, but they carry sovereign seizure risk. Gold pays nothing, but it's unfreezable. For a central bank in Poland, India, Turkey, or Saudi Arabia - each of which sits inside a different geopolitical risk zone - that trade-off is structural, not tactical.

The World Gold Council's survey found that 76% of central bank respondents expect gold to hold a higher share of reserves five years from now, up from 69% last year. More revealing: the share of respondents who actively manage their gold reserves rose from 37% to 44%. These aren't institutions parking gold in a vault and forgetting about it. They're thinking about rotation, valuation, and timing - the way a portfolio manager would.

The dollar that isn't going away

Here's where I think the analysis usually goes off track. Even with gold's surge and Treasury's retreat, dollar-denominated assets still make up roughly 42% of global reserve holdings. That's a long way from the euro's 15%, the yen's roughly 5%, or the renminbi's still-modest share. The dollar system hasn't been dethroned. It's been challenged by an asset class that doesn't pay, doesn't settle, and doesn't integrate with any payment system.

That's the odd part. Gold is winning the reserve allocation war - or part of it - precisely because it does none of the things a reserve asset is supposed to do. It earns no yield. It settles no trade. It intermediates no payments. It just sits there and is safe from confiscation. If that's the main reason central banks want it, then what we're looking at isn't monetary innovation. It's insurance.

This is worth holding onto, because it tells you something about how far the shift can go. Insurance works well until the insured event actually happens. If central banks need to move money during a crisis - to intervene in their currency, to finance deficits, to settle obligations - they'll still need liquid, tradeable, interest-bearing assets. Gold doesn't do that. Treasuries do.

What to watch

The question I'm tracking isn't whether gold keeps climbing. The question is whether the revaluation-driven share shift persists once gold prices stabilize, or whether active accumulation fills in behind the accounting move.

There are two paths. If central banks continue buying at or near current levels - roughly 850 to 900 tons a year - the gold share will compound upward, and the Treasuries share will slowly grind down. That's a real structural trend, and it would mean the "insurance" framing is underestimating a genuine portfolio migration.

Alternatively, if buying slows and the price rally normalizes, gold's reserve share could plateau or even retreat if the total reserve pool expands again. That would suggest the shift is more cyclical than permanent - a revaluation event layered on top of temporary geopolitical stress.

The ECB's own framing gives a hint. Their report said the international role of the euro "grew moderately" in 2025, but warned there is "no room for complacency". The institutions tracking this data know the picture is messier than any single percentage point suggests.

What I think the article should have been titled is: "Central banks are buying insurance they can't earn a yield on, and the math of a revalued metal made it look bigger than it is." That's less dramatic than gold dethroning Treasuries. It's also, to my mind, closer to what's actually happening to the architecture of global money.

The deeper question - the one I find myself circling back to - is what happens when enough central banks conclude that the yield on Treasuries isn't worth the sovereignty risk. At that point, it's no longer about reserve composition. It's about whether the dollar system can absorb a world where its own creditors are systematically hedging against it. We're not there yet. But the insurance policy is getting expensive, and it's getting larger.