Two numbers are doing all the work in the de-dollarization story, and both are being read backwards.
The first: the dollar's share of global official foreign exchange reserves fell from 64 percent in 2015 to 56 percent in 2025. The second: the yuan's growth as a payment currency has lost its momentum over the last two years.
Read together, they form a tidy narrative. The dollar is bleeding reserve share, a rival is rising, the world is slowly walking away.
Except the mechanisms underneath both numbers point the other direction. The reserve "decline" isn't a global exit — it's a handful of large holders doing idiosyncratic things. And the yuan "stall" isn't the currency losing steam — it's the metric losing sight of where the money actually moves.
The reserve decline isn't a global exit — it's a handful of large holders doing idiosyncratic things. The yuan stall isn't the currency losing steam — it's the metric losing sight of where money moves.
Start with reserves, because a New York Fed decomposition just took the aggregate apart.
The eight-point drop looks like a trend. It reads like every central bank quietly trimming its dollar pile in unison. It is not that.
Look at the cross-section of countries — not the weighted aggregate, but the raw count — and roughly equal numbers increased and decreased their dollar holdings over each four-year window. There is no dominant directional shift. Half the room leaned in, half leaned out.
That balance is the whole story, because it means the aggregate number is dragged by weight, not by consensus. Big holders swing the average; the median country barely moves.
The Fed splits the drop into two channels. One is the "preferences channel" — a central bank actively reallocating out of dollars into something else. The other is the "reserve change channel" — a bank growing or shrinking its total pile at a dollar share different from the global average, which moves the aggregate mechanically without anyone changing their mind about the dollar.
From 2015 to 2019, the dollar share fell three points. The decomposition splits that almost evenly: 1.2 points from preferences, 1.5 points from reserve size changes.
Now watch who's inside those numbers.
The 1.2-point preferences decline — the part that actually looks like a vote against the dollar — was, per the New York Fed, dominated by China and Russia. Two players, specific circumstances, accounting for most of the "active reallocation away from the dollar." Turkey, Peru, and Spain added maybe negative 0.2 points each. Meanwhile Sweden and Switzerland moved the other way, raising their dollar shares.
The reserve-change side is even more telling, because Switzerland was, per the New York Fed, the largest negative contributor there — and Switzerland was buying dollars.
Here's the trick. Switzerland ran large-scale reserve accumulation over 2015-19, expanding its total pile. But because Switzerland's dollar share started below the global average, adding all those reserves mechanically dragged the cross-country aggregate down — even as Switzerland raised its own allocation to dollars.
Sit with that. One of the biggest drivers of the "dollar decline" was a central bank increasing its dollar holdings. The aggregate fell because Switzerland got bigger, not because Switzerland got out.
The same country can push both ways at once. China, per the New York Fed, contributed to the preferences decline while contributing positively through reserve drawdowns. The headline number smears these opposing forces into a single downward line that no individual behavior actually describes.
Strip away the concentration and the arithmetic, and what's left is a cross-section that's flat. Bloomberg's read of the study is blunt: the shift from the dollar is overstated. The dollar's status in official portfolios is, in the Fed's phrasing, largely intact.
Now flip to the yuan, because the same logic — the metric misleading, not the reality — applies with a twist.
The payment-momentum stall is real in the data everyone watches. Peterson's Martin Chorzempa lays out a mixed picture, and the yuan still punches far below China's economic weight. On the surface: growth faded, story over.
But there's a divergence in that data that has never appeared before. China's Cross-Border Interbank Payment System — CIPS, the domestically-run rail for yuan settlement — is running at record volumes. At the same time, SWIFT-visible yuan payments are falling.
That gap is the tell. For the first time, the two metrics move in opposite directions.
The reading that fits, per Chorzempa's account: yuan settlement may be migrating onto rails that Western tracking can't see. SWIFT shows you the payments that route through SWIFT. It shows you nothing about the payments that don't.
So the "stall" is partly a measurement blind spot. The activity didn't disappear — it moved somewhere the yardstick doesn't reach. The yuan isn't losing steam so much as slipping out of the frame.
Which raises the real question: what is CIPS actually for?
Not displacement. Insurance. Chorzempa's framing is that yuan internationalization looks less like a bid to dethrone the dollar and more like an insurance policy — plumbing to be activated for large-scale payments if sanctions ever force China off the dollar system.
That reframes everything, including mBridge, reportedly a multi-central-bank settlement platform. Note who's reportedly inside it — the UAE, among others. This isn't a China-only project; it is, arguably, a coalition building parallel payment rails against a shared fear.
And here is the paradox that ties both numbers together.
The rails only need to exist because being cut off from the dollar is catastrophic. CIPS and mBridge can be read as hedges against the dollar's weaponization — which is to say, hedges against the dollar's continued centrality. You don't build an escape hatch from a room nobody's trapped in.
The value of the insurance, arguably, is a function of how much everyone still depends on the thing they're insuring against. Every dollar spent on sanction-proof plumbing is a confession that the dollar still runs the world.
The exposure numbers make the dependence concrete. Foreign holdings of US securities hit $35.33 trillion as of June 2025 — up 41.9 percent from $24.89 trillion three years earlier.
That is not a world retreating from dollar assets. That is, reportedly, a 41.9 percent increase in three years.
Equities are the largest slice at $19.84 trillion, with long-term debt at $13.84 trillion. Money is still flowing in, not draining out — the aggregate stock points up and to the right, hard.
So put the two mismeasured numbers side by side.
The reserve share fell because a few large holders — China, Russia, Switzerland — did specific things, one of which was buying more dollars. The cross-section is flat; the flight is a statistical artifact of concentration and portfolio-size mechanics.
The yuan payment share stalled because settlement is migrating to rails the trackers can't observe. The currency isn't fading; the visibility is.
Neither number measures what the headline says. And both, read correctly, describe a system where the dollar's gravity is intact — where even the countries most motivated to leave are still tightly tied up in dollar-based networks, quietly building the exits they hope never to use.
The de-dollarization narrative wants a rival and a retreat. What the mechanisms show is stranger and more durable: the escape hatch and the dependence are the same structure, and building one reinforces the other.
The dollar isn't being displaced. It's arguably being feared — which, for a reserve currency, may be the more secure position.