The Structural Shift in Global Capital Flows Every Investor Needs to Understand

The Dollar's Slow Retreat, By the Numbers

Bretton Woods is the 1944 agreement that put the dollar at the center of the postwar financial system and the consensus built on it stood for eight decades. It's not going to collapse this year. It's eroding slowly in a way that's easy to miss if you only check the numbers once a year. The dollar's share of global foreign exchange reserves that is the specific mix of currencies that central banks around the world hold as an emergency fund has fallen from about71 percent at its peak in 2000 to 56.9 percent in the third quarter of 2025. Fourteen percentage points in a quarter of a century sounds gradual and it has not been distributed evenly. The pace has accelerated noticeably since 2022 and that timing is no coincidence

This is the part that I think is most often overlooked. The dollar hasn't lost reserve share because it is weak. For much of this period the dollar has remained strong. It lost share because the alternatives quietly improved a little and because something specific happened in 2022 that changed the way every finance ministry in the world thinks about the security of holding dollars in the first place

The Reserve Freeze That Changed the Calculus

When Russia invaded Ukraine in 2022 the United States and its allies did something that had never happened before on that scale. They froze approximately $300 billion of the Russian central bank's foreign exchange reserves held in Western institutions. Overnight a reserve of assets that Russia's central bank had treated as the safest possible in the world became inaccessible

I think this is the most important fact in all of history more important than any deficit figure that follows. Central bank reserves exist for a purpose. They need to be there when a country needs them during a currency crisis a war a natural disaster whatever the emergency. The full value of a reserve asset is that you can actually use it at that time. The freeze showed that dollar assets held in Western institutions come with a condition that no one had fully discounted before: security is only maintained as long as one stays aligned withthe foreign policy of the United States. That is not a footnote. That is the total change of the product

Every central bank you can imagine in a serious dispute with Washington and that's a longer list than most people assume looked at that freeze and updated their models. You don't have to wait for a war with the United States to want a little less exposure to an asset that can be deactivated with a phone call. A modest shift toward something that can't be frozen that way starts to look like cheap insurance even for a country that never plans to use it

How a Country Ends Up Holding Someone Else's Money

Where do the reserves really come from? The mechanism explains much of what follows. A country that sells to the rest of the world more than it buys executes what economists call a current account surplusChina has historically been the biggest example of this by a wide margin but oil exporters and heavy manufacturing exporting economies show the same pattern

That country's central bank then has a lot of dollars there because most global trade even trade that has nothing to do with the United States is still invoiced and settled in dollars. The central bank has to put that money somewhere and the deepest and most liquid market to deposit it in has historically been for US Treasuries. Economists call this recycling.Surplus dollars return to dollar assets which is the literal mechanical process by which a trade surplus is converted into a reserve

This is also why the reserve role of the dollar and the United States' low borrowing costs are the same story told from two angles. When the rest of the world wants to hold dollar reserves it has to buy something and the largest and safest thing available is US government debt. This is additional demand for Treasury bonds that has nothing to do with what the Federal Reserve is doing domestically and the additional demand for a bond drives down its yield. Economists sometimes call this the privilegedollar exorbitant: cheap borrowing because the entire world needs a place to park its trade surpluses. If reserve demand recedes at the margin even slowly the natural expectation is that this discount will also shrink. No one can say precisely how many basis points that is worth. Here I would treat any safe number like someone guessing with extra decimals

The Fiscal Math and the Iran War Acceleration

Place three more forces on top of the frozen reserve and the picture becomes clearer. The first is America's own fiscal trajectory. The federal government has run a structural deficit of 6 to 7 percent of GDP during the peacetime economic expansion which is historically unusual. Deficits of that size typically appear during a war or a recession not when unemployment is low and markets are near highs. The Congressional Budget Office projects that debt to GDP will exceed 130 percentby 2035 under current policy. Reserve managers also read the CBO projections and the logic of keeping their emergency fund in the currency of the world's largest and fastest-growing debtor becomes a little less compelling each year that trend continues

The second is the internationalization of the renminbi China's direct response to the vulnerability exposed by Russia's freeze. China settled about 53 percent of its own international trade in renminbi in 2025 up from less than 20 percent just five years earlier

The third is the most recent and visible: the Iran War of 2026. When the United States and Israel began military operations against Iran in late February 2026 the Strait of Hormuz was effectively closed. That forced countries that were not aligned with either side to move more quickly on bilateral trade and payments agreements that revolve around the dollar. Saudi Arabia quietly expanded yuan payment options for its Chinese oil buyers a real if partial step.to move away from an agreement that has fixed the price of Saudi oil in dollars since the 1970s. None of this means that the Iran war alone caused de-dollarization. It means that a trend that was already in motion got a boost

Gold's New Job in the System

Gold's role in this story is its own little mechanism which is worth separating from the monetary issues mentioned above. Central banks bought 1,037 tons of gold in 2023 and another 1,014 tons in 2024 the highest two-year total on record. Gold then rose above $5,400 an ounce after the Iran war broke out in early 2026 as investors sought a security asset.safe haven that is completely outside the dollar Treasury complex

The argument for gold in a reserve manager's world is genuinely simple simpler than most of the arguments in this article. Gold has no issuer. No one can freeze it with a sanctions order because there is no counterparty on the other side of the trade like there is with a Treasury bond or a bank deposit. It has also appreciated substantially over the last decade which doesn't hurt anyone in favor of holding more. The absence of counterparty risk and a rising price are worth more in a world where security of assetsReserves turned out to be conditional on the politics of what they were worth in the world before 2022

Two Kinds of Money Moving Across Borders, and Why That Distinction Matters

There is a distinction that is softened in most comments about de-dollarization and I think it is one of the most important on the entire topic. Money crosses borders in two very different ways

Portfolio flows They are purchases of financial assets - stocks bonds currencies - that a buyer can sell back in one day if they change their mind. People sometimes call this hot money for a reason. Foreign direct investment or FDI is the opposite: a company that builds a factory buys a majority stake in a foreign company and opens a regional headquarters. That kind of commitment can't be undone before lunch. It takes years to build and years to exit

Most of the main statistics thrown around in de-dollarization arguments are portfolio flow figures because that data is reported quickly and moves a lot making the graph more dramatic. Data on FDI is slower more uneven and much harder to turn into an interesting weekly story but I would argue that it says more about whether structural change is really happening because FDI reflects a decision that no one takes lightly. I want to be honest about a limit here. I don't have figures before me.reliable FDI specific to this trend and I'm not going to make up some just to fill the paragraph. What I can tell you is which type of flow deserves more weight when you see a headline and it's the slow one not the fast one

This connects to an older and related idea called home bias: the well-documented tendency of investors individual and institutional to own far more assets from their own country than that country's actual participation in global markets would justify. Home bias has been declining for decades helped by index funds and global ETFs that make international diversification a checkbox rather than a project. It is declining slowly because many of the reasons behind it are structural themselves. A pension fund with dollar-denominated liabilities wants dollar-denominated assets thatequal. Currency risk is real and expensive to fully hedge. Regulation and tax treatment often favor domestic holdings. None of that is reversed in a single news cycle no matter how dramatic that cycle may seem

A Worked Example: What Reserve Reallocation Actually Looks Like

Let's make the idea of reallocation concrete because saying that central banks are diversifying means very little without accompanying figures not even illustrative

Suppose a central bank call it Countrymillion in gold and 25 billion in yuan and other currencies. Add them up and you get the full 500 billion which is just a check that the initial allocation is internally consistent

Now suppose that the reserve committee of countrymillion which is 300 billion. This is a reduction of 75 billion dollars since 375 minus 300 equals 75. Gold's target is 15 percent of 500 billion or 75 billion an increase of 50 billion from where it started. The yuan and other currencies go from 25 billion to 10 percent of 500 billion or 50 billion an increase of25 billion. The euro cube does not move at all. Check the flows against each other: minus 75 billion dollars plus 50 billion gold plus 25 billion yuan and other currencies net to zero exactly what you would expect since the total fund size did not change in this simplified version

Spread evenly over five years that equates to about $15 billion a year in dollar assets $10 billion a year in gold and $5 billion a year in yuan and other currencies. Fifteen billion dollars a year seems like a big number until you compare it to the fund from which it comes. Three percent of the total stock of reserves per year for five years. This is what real structural change looks like fromin.No rush.Arithmetically modest in any given year.Obviously it's only big once you put it together.That's also exactly why it's so easy for someone who signs up once a quarter and doesn't realize it's happening

Case Study: What Central Banks Did After 1997

If we want to see how this reserve accumulation mechanism plays out before in a completely different context the Asian financial crisis of 1997 and 1998 is the cleanest precedent I know of

Thailand's currency collapsed first in the summer of 1997 when the country could no longer defend its peg to the dollar. Panic spread quickly: to Indonesia to Malaysia and above all to South Korea. South Korea's central bank burned through its foreign exchange reserves trying to defend the won and by the end of that year the reserves had fallen to the point that the country had to turn to the International Monetary Fund for an emergency bailout with painfulconditions.Austerity.Forced restructuring of politically connected conglomerates.Interest rate shocks that pushed the economy into a deep recession

The lesson that all affected countries learned from that experience was almost identical to the lesson taught by the reserve freeze in 2022. Never again be caught without enough reserves of your own to have to ask for help. In the decade that followed South Korea along with China Taiwan and much of the rest of the export-heavy Asian countries built up reserves that dwarfed anything they had before the crisis. Part of that was simply the natural byproduct of running large current account surpluses and recycling the reserves.But their scale and urgency were a direct and deliberate response to what had happened in 1997. Economists later gave this pattern a name the global savings glut and argued that it was one of the forces that helped keep U.S. interest rates unusually low throughout the 2000s as much of that newly created reserve flowed directly into Treasury bonds

The parallel with today is not exact. No one had reserves frozen in 1997. Countries simply ran out of them through ordinary crisis mechanisms. But the behavioral response rhymes closely. A bad experience with insufficient reserves changes reserve policy over a generation not over a news cycle. If that pattern holds here the current shift away from dollar concentration is not something that will be reversed the moment the war with Iran ends. It is a policy adjustment with a much longer half-life than that

The Counterargument: Most Structural Shifts Get Declared and Then Fade

I've made a real case for structural change so let me argue against myself for a minute because structural change is one of the most overused phrases in finance and I don't want to be the next person to cry wolf

Every few years someone declares the end of the dollar's dominance. It happened when the euro was launched in 1999 when genuinely serious people were projecting that a unified currency backed by an economy roughly the size of the United States would become a true peer-to-peer reserve currency within a generation. It happened again after the 2008 financial crisis when the crisis originated within the United States and yet curiously capital flooded into dollar assets anyway. It happened again.in 2020 when the Federal Reserve printed trillions of dollars in emergency stimulus and a wave of commentary insisted that the currency's reserve status must eventually be devalued. None of those predictions were crazy. All of them came early at the very least.separate rather than a unified market and that structural gap has mattered far more than most people predicted in 1999

The honest pattern in this data going back decades is that the most apparent de-dollarization is cyclical not structural. It tracks risk appetite. When the world feels calm capital shifts to higher-yielding more adventurous places and the dollar's share by various measures declines. When something breaks capital rushes back to the deepest most liquid most boring market in the world which remains for now the US Treasury market and theDollar participation rebounds again. Iran's war itself is a small piece of evidence of the skeptical case. In fact the dollar strengthened in the days immediately following the first attacks despite all the talk in this article about the war accelerating de-dollarization in the long run. That's the safe-haven cyclical reflex working exactly as it always does right in the middle of a story I'm telling you that is structural

So what would really distinguish this round from previous false alarms? I think it's the combination of forces working at the same time rather than any one of them. The story of the euro in 1999 was that of a currency that tried to compete on convenience and depth but fell short on depth. The stories of 2008 and 2020 were monetary and fiscal shocks that did not change who could freeze whose assets. This time there is a demonstrated risk of confiscation a fiscal trajectory that obviously does not correct itselfalone and a geopolitical realignment all happening at the same time which is a different combination than any of the previous false alarms. I think that combination is what makes it more likely that this one is real. I want to be honest: I could be wrong and the best argument for predicting that this moment is different in finance is that it usually isn't

How I Actually Use This

That's how I really think about this when I look at a portfolio mine or one I'm studying and I want to be clear that this is my reading not a recommendation to buy or sell anything

The first thing I do is treat the dollar's decline as a multi-decade backdrop not a trade-off. Reserve stock numbers move perhaps a percentage point or two a year even in an accelerated period. If you try to time a single quarter around this story you're almost certainly using the wrong tool for the job. I think about the dollar's structural trends the same way I think about demographic trends: real directionally useful and almost useless in predicting what will happen nextTuesday

The second thing I do is separate the assets that benefit from a gradual diversification of reserves from the assets that benefit from the sharp panic into safe havens because this article contains both and they point in opposite directions in any given week. Gold benefits from both which is part of the reason I find the gold setup here really interesting and not just another commodity story. Emerging market assets and non-dollar fixed income benefit from the gradual story but are severely affected by the story.of acute panic exactly the dynamic that the Iran war demonstrated in a matter of days

The third thing and this is where I was wrong when I started reading about this a couple of years ago is that I used to treat headlines about dollarization as a single narrative that I should have an opinion on. I no longer do. Now I ask which of the four forces a given headline is really about: the reserve freeze precedent the fiscal trajectory the renminbi payment infrastructure or acute geopolitical tension because they operate on completely different timelines and respond to completely different triggers. Adding them together ishow you end up dismissing the entire story because part of it seems cyclical or overreacting to a short-term panic move as if it confirmed a decade-long thesis

What I'm really doing with this specifically is treating it as an input into my thinking about the home bias in my own head not just a textbook portfolio. It's easy for anyone who grew up in the United States to implicitly assume that dollar assets are the default safe option and that everything else is a speculative bet. The freezing of reserves is a genuine counter to that assumption at least for anyone other than the US government itself. I don't think that means abandoningI think it means not treating anything but dollars as automatically the riskiest side of a decision which is a bias I didn't notice until I started reading central bank reports for this article

The Bottom Line

The dollar's reserve ratio has fallen fourteen percentage points since 2000 with most of the recent movement concentrated from 2022 onwards and the mechanism behind that change is not mysterious once you separate its pieces. The freeze of Russia's reserves changed the way each central bank values the safety of dollar assets. The US fiscal trajectory makes holding reserves in the currency of the world's largest debtor a little less automatic each year.in renminbi and the 2026 Iran war have added momentum to that. Gold has surged in real demand because it is the primary reserve asset with no counterparty capable of freezing it. None of that means the dollar will cease to be the world's dominant reserve and transaction currency any time soon and the sharp safe haven reflex triggered by the Iran war is proof that the cyclical pull toward the dollar is still very much alive. What's really changing is the margin one or twopercentage points a year driven by forces that unlike most de-dollarization fears of the past are piling on top of each other rather than arriving one at a time. My honest read is that this one has a better chance of being real than the false alarms that preceded it precisely because it is not based on a single cause. I could still be wrong on that. The reasonable conclusion if you are building a portfolio view around any of this is to treat it as a slow present value that you are inclined toover the years not as a signal to trade in this quarter