Whenever yen breaks
The U.S. intervention in the yen market might have been smaller than signaled
On July 31, the United States and Japan intervened jointly to strengthen the yen — the first bilateral U.S.-Japan joint yen purchase since the Asian Financial Crisis in 1998. Japan spent an estimated $87 billion across two separate interventions (one solo, one joint with the U.S.) on July 30 and 31, one of the largest currency defenses on record. How much did America spend? Well, without direct confirmation from Treasury or the Fed, the now-famous Reuters shot of Treasury Secretary Bessent’s notepad reading “To do: Buy Japanese Yen (JPY) $5–10 bil.” became the number that everyone believed America spent. The press then reported that the U.S. sold euros rather than dollars to support the yen.
Now, while Secretary Bessent confirmed that the U.S. was intervening and will intervene again to support the yen, he never publicly confirmed the size of the intervention. The market and press pegged the intervention at $5–10 billion from the leaked notepad. However, while poking around in Treasury’s weekly International Reserve Position report, I could find no evidence of a large-scale yen purchase or significant euro sales. The report includes foreign currency holdings from Treasury’s Exchange Stabilization Fund (ESF) and the Fed’s System Open Market Account (SOMA). On July 24, the last print before the intervention, euro holdings stood at $25.94 billion. On July 31, they were at $26.26 billion. On August 7 — a full week after execution and presumably three days after settlement — they stood at $25.89 billion.
As a rule I don’t do math in public, but this is not what a $10 billion sale looks like.
Millions, not billions
The data imply the U.S. intervened on a scale closer to a few hundred million dollars than a few billion. In our analysis of the data between the July 31 and August 7 reporting dates, euro holdings fell $367 million while yen holdings rose $624 million (before FX valuations). Adjusting for the euro’s appreciation from July 31 to August 7, our back-of-the-envelope calculation implies the euro sale came to roughly €380 million, about $440 million. The Federal Reserve’s weekly Factors Affecting Reserve Balances report also showed a marginal increase in foreign currency holdings (where the aggregate SOMA FX holdings data lives), gaining only $422 million in the week following the intervention.
I thought maybe the discrepancies would appear in derivatives or FX valuations. No dice. The weekly Treasury report showed that short positions in FX derivatives (which is where a short euro forward position would appear) barely moved. All else equal, for the exchange rate to hide a $10 billion sale inside an otherwise relatively flat book, the euro would have needed to rally from 1.15 to roughly 1.84 in a week! Meanwhile, here we are still at around 1.15.
I can’t get every number to reconcile — again, math in public — but any way you cut the data, there is nowhere for $5–10 billion to hide. After FX valuations, roughly $440 million of euros were sold against roughly $640 million of yen bought. The gap could suggest part of the yen was purchased with dollars, which would complicate the media’s claim that only euros were used specifically to spare the dollar.
There has been no official confirmation on which account sold the euros. Shahin Vallée and Mark Sobel both read the operation as Treasury-only — “seemingly without any co-investment” from the Fed, in Vallée’s careful phrasing — which would break with precedent from 2000 (joint defense of the euro) and 2011 (G7 intervention to prevent destabilizing appreciation in the yen after the Fukushima earthquake), when Treasury and the Fed split interventions evenly.
Perception is reality
Without some econometrics, it is hard to quantify the exact impact of the joint intervention relative to other factors such as Japan’s massive solo intervention the day prior or softer U.S. economic data. USD/JPY strengthened (lower value means stronger yen) from near ¥164 (yen per dollar) to around ¥155 (yen per dollar) since July 30, but has given back a portion of those gains, with USD/JPY retracing to the ¥159/158 area.
Markets often respond to the idea that capacity exists rather than any actual deployment. Mario Draghi’s 2012 “whatever it takes” remains the canonical case. In July 2012, amidst the highest point of stress in the European Debt Crisis, Draghi, the then-President of the European Central Bank, promised with that one phrase that the ECB would not let the euro fail. This commitment and the belief that the ECB would follow through succeeded in stabilizing markets. A program to stabilize the euro was launched after Draghi’s statement but never activated — markets just needed to believe the capacity was there.
Something much smaller but somewhat similar just happened with Bessent and Japan. The ambiguity about the details of the U.S. intervention had two results. First, markets filled in some of the gaps with a rather negative version of events: the press reported that unnamed European officials were irritated (though no official said so publicly); and commentators assumed Japan had already tapped FIMA and was expanding the Fed’s balance sheet (though the MOF said it would only use it for future intervention and the facility has remained unused for eight straight weeks and counting). And second, the ambiguity left the market with nothing but a photo op and a press conference that confirmed U.S. commitment to stabilizing the yen, which markets treated as evidence that capacity and intent to intervene were present — this may have amplified the impact of a much smaller euro sale.
Glass half full?
Much of the commentary around the intervention interprets it as a story about American weakness. The Treasury market can be held hostage. Washington’s options for currency intervention are constrained. Reserves that can’t be sold are not reserves. The FT’s Katie Martin called the Treasury market America’s “squishy underbelly” and Barry Eichengreen issued similar warnings. These concerns are real. U.S. debt is approaching $40 trillion.
But I don’t think this intervention is a death knell. For one, the Treasury market is not the easy hostage that some suggest. The largest holders of Treasuries at this moment are those who probably do not want to trigger a selloff in the market. Selling Treasuries to get dollars to buy your own currency will trigger a sharp appreciation in your currency, which is exactly what Japan wanted. It is hard to imagine China wanting to engineer a sharp appreciation of the RMB by doing the same. It’s also worth noting that our other adversaries aren’t holding enough Treasuries to be meaningful. We froze Russia’s reserves.
Building more capacity for currency interventions is possible but complicated. Treasury’s ESF foreign-currency holdings are modest, requiring the Fed to serve as the firepower in any future intervention — warehousing or SDR certificate monetization to fund the ESF, the FIMA repo facility to fund allied central banks, or maybe a solo SOMA intervention. Using these facilities will require Fed approval and ultimately expand the balance sheet. The Fed can sterilize the reserve impact so the operations do not ease monetary conditions, but balance sheet expansion in a period of elevated inflation is complicated. However just because the U.S. is constrained does not mean it is powerless.
We won’t really know what happened until the ESF’s July and August financial statements or the Fed’s Q3 foreign exchange operations report come out later this year. Maybe they will show the U.S. really did sell $10 billion of euros. I am skeptical.
Whatever the final number, I think history will view this particular intervention as demonstrating Washington’s willingness to defend an allied country’s currency with just a little bit of money and a lot of talk. And to be clear, that is not a bad thing. Mario Draghi’s “whatever it takes” was not a bluff. This intervention is certainly not the scale or breadth of “whatever it takes.” But maybe it is our “whenever yen breaks” moment. Dollar statecraft ultimately depends not just on resources, but also on markets believing that the resources will be deployed if needed.