@gordonliao
Post
Post 1 of 14
The dollar's share of FX reserves is at a two-decade low. Treasury convenience yields have turned negative. End of dollar dominance? No — but financial innovation could reshape the system. Our new Jackson Hole Economic Policy Symposium paper, w/ @EswarSPrasad & Tony Zhang 🧵 
Image from X post
Post 2 of 14
The case for decline is real: the dollar's share of global FX reserves fell from 72% in 2000 to 57% today. And the convenience yield on Treasuries — the premium investors pay for safety and liquidity — has compressed to zero, even negative at long maturities. 
Image from X post
Post 3 of 14
But dollar remains strong in settlements and debt Int'l payments: the dollar is 59% and steady FX turnover: ~90% of all trades involves the dollar. Int'l debt: ~60% of issuance & holding Dollar rivals EUR, JPY, GBP all lost ground 
Image from X post
Post 4 of 14
The real action is outside of the traditional G4 currencies. The RMB's share of SWIFT reached 4.5% in 2024, and it's also expanding CIPS, its SWIFT alternative. Smaller reserve currencies are gaining while the traditional majors fade
Post 5 of 14
A quiet shift every EM policymaker should track: central bank reserve accumulation has leveled off while the private sector now does the accumulating. Self-insurance is being privatized — and private flows respond to frictions, and to the technology that removes them. 
Image from X post
Post 6 of 14
Where does privatized self-insurance go? Into U.S. assets. Net foreign inflows into U.S. securities are running near record highs, and the U.S. now absorbs 36.5% of the rest of the world's external assets — a share still rising. 
Image from X post
Post 7 of 14
Financial innovation (instant payment systems, tokenized assets, stablecoins) will change the landscape. In principle: a level playing field. Less need for vehicle currencies, direct access to any market. In practice? We built a model of currency competition that show nuances
Post 8 of 14
The core friction: debt must be settled in the currency it's denominated in. Obtaining a safe dollar claim exactly when you need it is a search problem — easy in deep markets, costly in thin ones. A stablecoin is settlement technology: redeemable at par, minimum search. 
Image from X post
Post 9 of 14
Result 1: settlement tech cuts both ways. When an EM improves its own market infrastructure — better matching, deeper float — local-currency issuance becomes viable for more of its firms. Innovation at home strengthens the home currency. Infrastructure is currency policy. 
Image from X post
Post 10 of 14
Result 2: dollar stablecoins deepen dollar use. In our calibration, the share of EM firms issuing in dollars also increases. Demand doesn't leave the system — it's rerouted through issuer reserves into Treasuries, deepening the very market that anchors dollar dominance.
Post 11 of 14
Result 3: settlement technology is now a margin of currency competition. A rival currency running a stablecoin the dollar doesn't match can capture EM issuers outright; but with a large enough issuance base, no tech edge is needed. Market depth and technology are substitutes. 
Image from X post
Post 12 of 14
Multipolarity sounds appealing. But splitting global issuance into shallower pools thins every market: average EM firm welfare roughly halves in our bipolar scenario. And in a liquidity crunch, central bank backstops are weaker when there's no longer a lender of last resort. 
Image from X post
Post 13 of 14
Three futures for the international monetary system:
1 - Greater concentration — most likely on today's path 2
True multipolarity — best, but needs deep markets, strong institutions, innovations 3. Fragmentation — the riskiest The outcome isn't preordained. Policy decides.
Post 14 of 14
Paper + slides: https://www.kansascityfed.org/documents/18555/prasad.pdf https://www.kansascityfed.org/documents/18553/prasad\_handout.pdf Presented at the Federal Reserve Bank of Kansas City's Jackson Hole Economic Policy Symposium, Aug 27–29, 2026. Grateful to the organizers and discussants — and to my co-authors @EswarSPrasad and Tony Zhang.
Explanation
The post’s central argument is that “the dollar is losing reserve share” is not the same thing as “the dollar is losing its role as the operating system of global finance.” Dollar reserves have fallen from roughly 72% to 57% since 2000, and one traditional measure of Treasury specialness has deteriorated. Yet the dollar still sits on one side of ~90% of FX trades, accounts for about 59% of international payments, and dominates cross-border debt. Meanwhile the euro, yen and sterling have not filled the gap; the second tier is becoming more fragmented. The paper was presented at the Kansas City Fed’s 2026 Jackson Hole symposium; the authors are Gordon Liao, Eswar Prasad and Tony Zhang. Liao currently works at Circle, while Prasad is a Cornell economist and former senior IMF official, so this is essentially international-macro research applied to stablecoins rather than a generic crypto thesis. ([Federal Reserve Bank of Kansas City][1])
“Convenience yield” is the first unintuitive bit. A Treasury can be worth more than an otherwise equivalent asset because it is exceptionally liquid, safe and useful as collateral. Investors therefore accept a lower financial yield; that yield sacrifice is its convenience yield. The authors’ point is that this premium has shrunk or even reversed at long maturities. That is genuine evidence that Treasury “specialness” has weakened—but evidently not enough to undo the dollar’s enormous network effects.
The really interesting part is their model of why market depth perpetuates currency dominance. Suppose an Indonesian company has dollar debt coming due. It must obtain a dollar asset at the right moment. In a huge dollar market, finding someone willing to sell the needed safe dollar claim is almost guaranteed; in a thin rupiah market, finding the equivalent local claim can be difficult or expensive. Firms therefore have another reason to issue dollar debt in the first place. More dollar issuance makes the dollar market deeper, which makes dollars still easier to obtain: a self-reinforcing liquidity loop.
Stablecoins initially look as though they should destroy this advantage by making every currency equally easy to move. The authors’ counterintuitive result is: only if every currency gets comparably good infrastructure. Better rupiah settlement can indeed make rupiah borrowing more attractive. But if the innovation is primarily USDC-like dollar stablecoins, they make obtaining a dollar claim even easier. And the stablecoin issuer generally invests the backing assets in dollar assets such as Treasury bills. So payment demand that appears to bypass banks gets rerouted back into the underlying dollar capital market, potentially making the market supporting the dollar deeper still. This complements recent Fed discussion describing stablecoins as broadening access to dollar-denominated finance rather than forming a separate monetary universe. ([Federal Reserve][2])
That explains the final graph: technology and market size are substitutes. A smaller currency such as the euro would need substantially better settlement technology to overcome the dollar’s liquidity advantage. As its underlying issuance market approaches or exceeds the dollar’s size, less technological advantage is required. Network effects are therefore powerful but not immutable.
The “private self-insurance” point connects all of this. Emerging economies historically accumulated official FX reserves to protect themselves against crises. The authors argue that reserve accumulation has flattened while households and firms increasingly hold foreign safe assets themselves—disproportionately U.S. assets. Financial technology matters because it can make that private dollarization dramatically easier.
So the paper is basically saying: stablecoins do not inherently decentralize the international monetary system. If dollar markets start deeper and dollar stablecoins scale first, removing transaction frictions can strengthen rather than weaken dollar dominance. True multipolarity requires rival countries not merely to tokenize their currencies, but to build comparably deep, liquid capital markets and credible institutions.
[1]: https://www.kansascityfed.org/research/jackson-hole-economic-symposium/2026/?utm_source=chatgpt.com "Jackson Hole Economic Policy Symposium 2026 - Federal Reserve Bank of Kansas City" [2]: https://www.federalreserve.gov/econres/notes/feds-notes/fifth-conference-on-the-international-roles-of-the-u-s-dollar-stablecoins-digital-payments-and-the-ir-of-the-usd-20260716.html?utm_source=chatgpt.com "The Fed - Fifth Conference on the International Roles of the U.S. Dollar: Stablecoins, Digital Payments, and the International Role of the U.S. Dollar"