Newsletter · 2026-09-15 · Ashutosh Agarwal

Crowded Dollar Longs Explain Why the Dollar Won't Rally Before Fed and BOJ Hikes - The Dollar Brief - Week of September 15, 2026

The Dollar Brief for the week of September 15, 2026 (podcasts recorded September 8 to 14, 2026): a synthesis of the week's macro and FX podcasts on why the most crowded dollar-long positioning in five years has left few buyers ahead of an all-but-certain Fed hike and a Bank of Japan hike two days later, alongside State Street's call that dollar-yen has likely topped, Nellie Liang's measured case for stablecoins as a Treasury bill buyer, Mike Green's contrarian defense of the Treasury buyback, and the $1.3 trillion '$5,000 dividend.'

The Dollar Brief

Week of September 15, 2026: Crowded Dollar Longs Explain Why the Dollar Won't Rally Before Fed and BOJ Hikes

We are two days from a rare double-header. On Wednesday the Federal Reserve is almost certain to raise interest rates. On Friday the Bank of Japan is expected to raise its own. And yet the story of the week isn't either meeting; it's the dog that won't bark. Everything that is supposed to lift the US dollar has now happened at once: a hawkish Fed chair, a strong jobs report, a hot inflation print, oil back above $100, and Treasury yields brushing 5%. The dollar just sits there. This week a growing group of podcasts stopped calling that a mystery and started calling it what it is: a crowd. Almost everyone is already long the dollar, so there's no one left to buy it. Let's get into it.

(Quick plain-English glossary. The "dollar" is the US dollar; the "DXY" measures it against a basket of other rich-world currencies. "Long the dollar" means you own it and are betting it rises; when a bet gets "crowded," almost everyone is already on the same side, which usually means the move is nearly over. A "Treasury" is a loan to the US government; its interest rate is its "yield," and yields rise when bond prices fall. The "FOMC" is the Fed's rate-setting committee, meeting September 16-17; a "hike" raises its short-term rate. The "BOJ" is the Bank of Japan. The "carry trade" is borrowing cheaply in a low-rate currency (for years the Japanese yen) to buy higher-yielding things elsewhere; when it "unwinds," people rush to buy the yen back. A "buyback" is the Treasury repurchasing its own older bonds to try to push long-term rates down. A "stablecoin" is a digital token meant to always be worth one dollar, backed by real dollars and Treasury bills. "PPP," purchasing-power parity, is a rough estimate of a currency's fair value. "Commitment of Traders," or CoT, is a weekly US report showing how speculators are positioned.)

TL;DR

The dollar won't rally because the trade is a crowd, not a mystery. NewEdge Wealth's Cameron Dawson: dollar-long positioning "peaked about a month ago, a really, really powerful, super crowded dollar long positioning… the most crowded that they've been in the last five years," and "peaks in dollar longs typically coincide with peaks in the dollar" (On The Tape with Danny Moses, Sep 10).
The pros admit it and keep leaning short anyway. Real Vision's Macro Mondays host said the dollar "has underperformed a lot versus everything that's been happening… we know that the market is heavily leaning into the dollar story," and stuck with "a short dollar lean" (Real Vision, Sep 14).
The Fed is "painted into a corner." J.P. Morgan's David Kelly: markets have priced a 92% chance of a hike, the 10-year has gone "from 4.67% to 4.96%" in six weeks, and "if the Fed doesn't deliver one, both Chairman Warsh and the FOMC will lose serious credibility" (Notes on the Week Ahead, Sep 14).
A hike may not even lift the dollar, and could cap it. History says the dollar is "very, very firm" into a first hike and then "could go anywhere"; with a hike already priced, the surprise risk is skewed the other way (Real Vision, Sep 14).
The yen has quietly become the dollar's other half. State Street's Michael Metcalfe puts a "70% chance we've seen" the top in dollar-yen, with the yen "more than 40% cheap" and fair value between 110 and 120 (Street Signals, Sep 10). A 25bp BOJ hike Friday is "fully priced" (Nomura – The Week Ahead, Sep 11).
Positioning has already flipped in the yen. The latest CoT report showed "a 100,000 contract shift away from yen shorts… the dollar-yen overall position is almost flat now, finally… slightly positive for the yen," while "euro longs have been very heavily reduced" (Saxo Market Call, Sep 14).
The dollar is "leaking." Exante Data's Jens Nordvig: strip out the usual drivers and "there's some little residual that is hard to explain… leaking in a weaker direction," and on debt metrics "the worst… it's actually the United States now. Japan used to be the worst" (The Compound and Friends, Sep 11).
A genuinely new, credible plumbing story: stablecoins as a Treasury buyer. Former Treasury and Fed official Nellie Liang: the two biggest stablecoins hold "79% of… reserve assets" in Treasury bills, so "every dollar of stable coins creates… 80 cents of demand for T-bills," and after netting, roughly "0.6" of net-new T-bill demand per dollar, pulled mostly "from reduced bank deposits" and "from abroad" (Macro Musings, Sep 14).
A contrarian defense of Bessent. Simplify's Mike Green argues the Treasury's buyback is plain "debt management," not yield-curve control: "we can issue $1 of current coupon paper and retire $2 of low coupon paper… in aggregate, we are significantly shrinking the debt," and points out US default-insurance costs and inflation break-evens have not risen (Wealthion, Sep 10).
The quieter Fed is its own dollar risk. Goldman's Jan Hatzius on Kevin Warsh's less-transparent Fed: "you'll have worse guesses on what the Fed will do, which would probably mean more volatility in rates markets… volatility that serves no good economic purpose" (Goldman Sachs Exchanges, Sep 9).
The fiscal match into a nervous bond market. The President's $5,000 "dividend" (roughly $1.3 trillion) would, if enacted, "explode" the deficit "to three" trillion and risk "a huge blow up" in bonds (The Rest Is Politics: US, Sep 10).

What's new

The real reason the dollar won't rally: it's a crowd, not a riddle

For weeks, dollar bulls have been scratching their heads at the same puzzle: good news for the dollar keeps arriving, and the dollar keeps not moving. This week several podcasts landed on the same unglamorous answer: there's simply no one left to buy it.

The clearest version came from Cameron Dawson, chief investment officer at NewEdge Wealth, on Danny Moses's On The Tape (Sep 10). "Higher yields haven't meant chasing into the dollar," she said, "because people are already there." Her measure of "there" is striking: dollar-long bets "peaked about a month ago, a really, really powerful, super crowded dollar long positioning… the most crowded that they've been in the last five years." Why that matters, in plain terms: "peaks in dollar longs typically coincide with peaks in the dollar. Troughs in dollar shorts coincide with troughs in the dollar." Some of that crowd has since unwound, she noted, "but it's not at the point where it's short or underweight yet at all." Translation: the dollar may have more air to come out before it's washed clean.

You heard the same read, from the opposite chair, on Real Vision's Macro Mondays (Sep 14). The host walked through his own history study: the dollar is "very, very firm" going into a first rate hike, "and then it could go anywhere." His honest observation on the present: "the dollar has underperformed a lot versus everything that's been happening. Energy prices up… the Fed suddenly catching up to the rate hikes we see elsewhere… very high bond yields… But it's not responding the way it would typically. And that's very telling for how the market is already positioned." Because "the market is heavily leaning into the dollar story," he's kept "a short dollar lean." He framed the whole setup as a coiled spring pointing the wrong way for the consensus: "right now the market is long dollars. It's very short bonds, starting to pile into energy longs… betting against the equity markets. So you've got the setup for a complete wrong-footing of everyone if something magically happens," for instance a Middle East de-escalation that drops oil "by $30 a barrel."

And the positioning isn't just anecdote. Saxo Bank's John Hardy, reading the latest Commitment of Traders report on the Saxo Market Call (Sep 14), described "an absolutely titanic turnaround" in yen positioning ("like a 100,000 contract shift away from yen shorts"), leaving "the dollar-yen overall position… almost flat now, finally… slightly positive for the yen," with "euro longs… very heavily reduced." The crowd is quietly starting to file out.

The eve of the double-header, part one: the Fed painted into a corner

The Fed decides Wednesday, and the most useful walk-through of why it will almost certainly hike even if it doesn't want to came from J.P. Morgan Asset Management's David Kelly on Notes on the Week Ahead (Sep 14), who borrowed Fiddler on the Roof's "on the other hand" device to lay out both sides.

The case against hiking is real: wage growth "came in at just 3.1%… the weakest gain in any month since May of 2021, and the fifth consecutive month in which real wages fell"; the three-month average of job gains is "just 71,000"; and the oil-driven inflation is a "classic temporary supply shock" that hasn't moved long-term inflation expectations (the market's 10-year expectation is "2.36%, up just 0.11% since the start of the year"). But the case for hiking is now about credibility, not the economy. Warsh "asserted that the FOMC has no tolerance for persistently elevated inflation," then "foresworn forward guidance," and after Friday's hot core CPI the market has priced "a 92% chance of a hike," with the 10-year up "from 4.67% to 4.96%" in six weeks. As Kelly put it, quoting a colleague: "there is clarity in forward guidance and awkwardness in forward silence." Having said nothing to talk the market down, the Fed now has to deliver or "lose serious credibility." His kicker on the vote: a decision that may look close beforehand could still print "10-2, 11-1, or even a 12-0" as members close ranks in front of the President.

The forward-looking piece, and the one that matters most for this newsletter, is Kelly's tail: if growth and inflation cool into 2027, the Fed could "raise rates just once more, or not at all," which "could limit any further increase in long-term yields and allow a resumption of a longer-term dollar decline." In other words, the base case at one of the world's largest asset managers is that this hike is near the end of the road, and the dollar drifts down from here.

The eve of the double-header, part two: the yen is now the dollar's other half

Two days after the Fed, the Bank of Japan takes its turn, and this week the yen and the dollar are joined at the hip. The most quotable call came from State Street's global head of macro strategy, Michael Metcalfe, on Street Signals (Sep 10), which matters because State Street sees a huge slice of the world's actual custody flows. His verdict: "I'm going to say there's a 70% chance we've seen" the top in dollar-yen. His reasoning is history plus value: of "13" major turning points in dollar-yen "in the past 50 years… 9 of them have occurred around intervention," and this intervention "is going to be followed up by the BOJ really tightening policy." On value, the yen has been "more than 40% cheap" on State Street's purchasing-power model, and he'd target a move toward 120, maybe 110. One tactical wrinkle worth stealing: if you want to bet on yen strength, the dollar may not be the best thing to sell it against: "if value really is coming back as a factor, I think you'd prefer to buy the yen against the Swiss than the US dollar."

J.P. Morgan's FX team, on their At Any Rate BOJ preview (Sep 14), explained why the yen suddenly behaves normally again. Before July, markets feared the BOJ was "behind the curve," so rate-hike expectations went with yen weakness. After the July intervention, and the read that Washington was "urging the BOJ to accelerate rate hikes," that fear eased, and "rising rate-hike expectations have begun to translate into yen strength." They're "tactically bullish on yen" (long yen versus the Swedish krona since August 21), but flagged the catch: if the BOJ under-delivers on Friday, the "behind the curve" fear "could rise again," and the yen could fall instead of rise. Medium-term they still see the yen drifting back toward "155 to 165." A 25bp hike itself is "fully priced," per Nomura (Sep 11), so the drama is all in the signals: whether hawk Takata pushes for 50bp, whether reflationists dissent, and how hawkish Governor Ueda sounds, which "will depend partly on the Fed's decision beforehand."

For anyone who missed the summer backstory, Deep Values (Sep 12) reconstructed how Washington rescued the yen from a 40-year low near 164, dangling the Fed's FIMA repo facility (which lets Japan pawn its Treasuries for cash rather than sell them) and, when the New York Fed did step in, selling euros to buy yen so as not to touch the dollar-yen pair directly. A useful caveat on that facility came from J.P. Morgan's own preview: it isn't really built for big, sustained intervention: its tenor is only "overnight and one week," and the cost is "OIS plus 25 basis points," which is "quite high" (At Any Rate, Sep 14).

Hanging over all of it is Treasury Secretary Bessent's now-infamous boast: "I am the house now… I have asymmetric information" about what the Bank of Japan will do. Jens Nordvig, who's "known Scott Bessent for a long time," decoded it charitably on The Compound and Friends (Sep 11): Bessent "was pretty confident he could… orchestrate that the hike would come at the next meeting," so he "was willing to do the intervention on the expectation that the hike is coming," and, indeed, a September BOJ hike went from "three or four basis points" priced to more than 90%. Others heard something uglier. On The David Lin Report (Sep 14), long-time bear Peter Grandich called it plainly: "this administration is not afraid to talk about insider trading and do it" (Grandich is an avowed gold bull and dollar bear, so weigh the framing).

The dollar is "leaking": a slow puncture, not a blowout

The most memorable image of the week came from Jens Nordvig on The Compound and Friends (Sep 11). Build a normal model of the dollar (short-term rates, risk appetite) and lately "there's some little residual that is hard to explain… that is leaking in a weaker direction… in the last couple of months." His deeper point is about debt. Rank the top 30 economies by debt metrics and "guess who has the worst… Japan and US. It's actually the United States now. Japan used to be the worst." The dollar is "certainly more resilient than everything else because" it's still the reserve currency ("so is it special? Yes"), "but it's not trading fantastically well now." You can see the same logic in currencies with clean balance sheets doing quietly well: "Switzerland… Singapore… Australia is a country that actually doesn't have a lot of debt."

The quieter Fed is its own dollar risk

There's a slow-burn story under the meeting: Kevin Warsh is running a deliberately less-transparent Fed: shorter statements, less forward guidance, a smaller role for the "dot plot" of rate projections. Goldman Sachs's chief economist Jan Hatzius pushed back hard on Goldman Sachs Exchanges (Sep 9). Losing clarity on how the Fed reacts to data, he warned, means "you'll have worse guesses on what the Fed will do, which would probably mean more volatility in rates markets and arguably broader financial conditions… volatility that serves no good economic purpose." He also dismantled Warsh's stated rationale (that less guidance lets markets send the Fed cleaner signals): "markets price what they think the Fed will do, not what they think the Fed should do." Former Fed Vice Chair Don Kohn had more sympathy for trimming forward guidance but doubts the silence lasts: Warsh's July press conference "was not a good look… short-term rates go down and long-term rates go up," a sign the market was building in inflation and risk premiums it couldn't otherwise pin down.

Why does this belong in a dollar letter? Because a currency is, at bottom, a bet on an institution. The more the market has to "price not knowing," the wider the risk premium it demands, and a dollar that grinds sideways on great news is exactly what that looks like. Institutional Risk Analyst's Chris Whalen took the fear to its logical end on The Julia La Roche Show (Sep 12): "what happens if they raise short-term rates and the long end goes higher? Everybody is used to the idea that the Fed can get what it wants… But what if policymakers have lost the ability to control long-term interest rates?" If they have, he said, "the ability of governments to finance their operations by issuing more debt is going to be very much circumscribed," the kind of thing that ultimately shows up in the currency.

Is Bessent's buyback shrewd or doomed? The contrarian gets a hearing

Most of the week trashed the Treasury's bond buybacks as too small to matter: the 10-year "closing in on 5%" after Bessent tripled his repurchase pace to $6 billion and actually bought about $5.2 billion (Squawk Box Europe Express, Sep 11). Chris Whalen called it "a failure… he focused attention on something we didn't need to focus on" (The Julia La Roche Show, Sep 12).

But the most interesting voice was the dissenter. Simplify's Mike Green, a genuine bond-market mechanic, argued on Wealthion (Sep 10) that the market has this backwards. What Bessent is doing "is just debt management, and it makes perfect sense": the Treasury can "issue $1 of current coupon paper and retire $2 of low coupon paper… a modest increase in our cash interest expense, but in aggregate, we are significantly shrinking the debt." Calling it a "financial crime of the century," he said, "gives you some idea how milquetoast we have become." Is it yield-curve control? Only in the trivial sense that consolidating your credit-card debt is: "Yield curve control in the classic sense… the Fed and Treasury conspire to keep rates artificially at a certain level. There's no sign that's what we're engaged in." His most important point cuts against the whole "the world is dumping America" narrative: if this were a credit or inflation scare, it would show up in market prices, and it hasn't. "US CDS" (the cost to insure against a US default) "has contracted over the past several months," and inflation break-evens "do not" reflect the fear either. In his telling, higher yields are a buyer-composition story, not a loss-of-faith story.

The other side, from allocators who think the market always wins these fights: Strive's Matt Cole (a former CalPERS bond manager, now running a Bitcoin-heavy firm, so note the bias) on The Wolf of All Streets (Sep 10). A buyback used "as a marketing exercise does not work," he said: "the market says, okay, they see the problem, but they're not stepping in with the size necessary to fix" it. His path: the 10-year breaks into "the 5% to 6% range where he ultimately blinks," at which point Bessent has to deploy real firepower (the Treasury's ~$1 trillion cash account) rather than a "laughable" $6 billion against "$40 trillion of debt." His forecast, wearing the crypto bull's hat: the DXY falls "about 30%" over three to five years as the government chooses to "preserve the dollar as the reserve currency for as long as you can" while accepting its slow decline.

The genuinely new thread: stablecoins as a captive Treasury buyer, done properly

Last week this newsletter flagged a gold dealer's excited take that stablecoins are "secret QE" (ITM Trading, Sep 11; an interested source, since he discloses owning a big stablecoin operator). This week the same idea got a far more credible airing from Nellie Liang, a former Under Secretary of the Treasury and long-time Fed economist, on David Beckworth's Macro Musings (Sep 14), and the numbers are worth knowing precisely because they're soberer than the hype.

Her starting point: stablecoins are, on net, "a positive for the role of the dollar," because if the US doesn't offer a digital dollar, "other countries are developing payment systems" and America would "lose… our share of cross-border transactions." On the debt, the mechanics are concrete. The two largest stablecoins already hold "79% of… reserve assets" in "Treasury or Treasury repo," so mechanically "every dollar of stable coins creates… 80 cents of demand for T-bills." But the net-new demand is smaller, because some of the money would have bought T-bills anyway. Working through where the cash comes from (banks, money-market funds, physical currency, or from abroad), she lands on "about a 0.6" of net-new T-bill demand per dollar of stablecoin growth, across scenarios where the market grows to "0.9 trillion, 1.9 trillion, [or] 4 trillion." The catch, and it's an important one for Main Street: that new demand comes mostly "from reduced bank deposits" and "from abroad," and if it's "smaller community banks that would lose deposits… one could imagine some disruptions" to small-business lending. So yes, the GENIUS Act does wire stablecoin growth into T-bill demand, but it's roughly 60 cents on the dollar, not a magic money machine, and it comes with a bill attached to community banking.

The de-dollarization whisper

What surfaced on reserve managers this week came from openly gold-bullish voices, so treat it as an interested corner rather than settled fact. On Palisades Gold Radio (Sep 10), trader Francis Hunt claimed "South Korea, for the first time in 13 years, [is] buying gold, dumping treasuries," and that Norway's giant wealth fund is "reducing American holdings, buying gold." His framing of de-dollarization is worth quoting because it's the right mechanism even if the cheerleading is loud: "when you sell your treasury, they are not leaving it in dollars in the stock market anymore. They're taking it home." That rhymes with a harder data point from a credible source last week: Norway's sovereign fund saying it is "moving some of our safety portfolio away from US sovereign bonds into other US credit assets" (Squawk Box Europe Express, Sep 11). The signal to hold onto: even the trimming that is happening is a rotation within dollars (out of Treasuries, into US credit) more than a flight out of dollars, for now.

The fiscal match: a $5,000 "dividend" tossed at a jumpy bond market

Into a bond market that has suddenly rediscovered deficits, the President has tossed a very large one. On The Rest Is Politics: US (Sep 10), Prof G Markets' Ed Elson explained why the bond market flinched where the stock market shrugged: bond investors "are thinking about all the things that could go wrong," and they've watched "an unbelievable sell-off… the 30-year Treasury yield climbing back above 5.3%, the 10-year… at 4.83%, the highest since 2023." His math on the "$5,000 to every voter" promise: a cost near "$1.3 trillion" that "would explode" this year's roughly $2 trillion deficit "to three," which would trigger "just a huge blow up" in bonds if actually enacted. He's skeptical it will be (Trump "promised a $5,000 Doge dividend… a $2,000 tariff dividend… a $1,000" health-care dividend, none delivered), but "the most important thing is the fact that he's willing to say it." A telling footnote: the earlier tariff "refunds" flowed to companies, not citizens: "Walmart got nearly $3 billion. Apple got more than $2 billion." On Balance of Power (Sep 10), even Republicans were wary: House Judiciary Chair Jim Jordan said he "wants to see details before committing," and the Vice President's claim that tariffs are "already paying down debt" was flagged as factually wrong, since the debt keeps growing.

The debate

Where the podcasts genuinely disagreed this week.

Has the dollar already topped, or is it just resting before a hike-driven pop? A live split. The "it's topping" camp is now crowded with flow-watchers: Cameron Dawson (On The Tape, Sep 10) on five-year-extreme long positioning; Real Vision's "short dollar lean" (Real Vision, Sep 14); Nordvig's "leaking" dollar (The Compound and Friends, Sep 11). The counter, still respectable, is J.P. Morgan's FX desk, which stayed "tactically constructive" on the dollar into the meetings on the grounds that it's cheap (At Any Rate, Sep 11). The tell that settles it: whether the dollar can rally on an actual hike, or whether, as the crowd argument predicts, the good news is already spent.

Does a Fed hike even help the dollar? Genuinely unresolved, and counter-intuitive. Real Vision's history study says the dollar is strong into a first hike and then "could go anywhere" (Real Vision, Sep 14). David Kelly goes further: he sees the hike as near the end of the cycle, which would "allow a resumption of a longer-term dollar decline" (Notes on the Week Ahead, Sep 14). Robert Savage's operator view is that intervention plus a BOJ hike work against the dollar via the carry trade (below). So the classic "higher rates, stronger currency" reflex may simply not fire this time, because everyone already positioned for it.

Is Bessent's buyback shrewd debt management or a doomed defense? Sharply contested, and the smartest disagreement of the week. Mike Green's contrarian defense, that it's rational liability management, and the fear-narrative isn't showing up in CDS or break-evens (Wealthion, Sep 10), runs straight into Matt Cole's "the market always wins these" (The Wolf of All Streets, Sep 10) and Whalen's "a failure" (The Julia La Roche Show, Sep 12). Both can't be right. The referee: whether Bessent is ever forced to deploy real "size" from the Treasury's cash account, the moment that would prove the market called his bluff.

Has dollar-yen made its high? Flagged as high-conviction but not certain. State Street's Metcalfe is at "70%" yes, on value and the history of interventions (Street Signals, Sep 10), and J.P. Morgan is tactically long yen (At Any Rate, Sep 14). The honest caveat both make: it hinges on the BOJ delivering Friday. Under-deliver, and the "behind the curve" fear, and yen weakness, comes straight back.

The trades in play

These are speakers' own stated positions and views, not advice.

Lean short the dollar into the meetings. Real Vision's Macro Mondays kept "a short dollar lean," arguing the good news is already in the price and the crowd is dangerously one-sided (Real Vision, Sep 14).
Buy the yen, and maybe not against the dollar. J.P. Morgan is long yen versus the Swedish krona (At Any Rate, Sep 14); State Street's Metcalfe prefers buying the yen "against the Swiss" if value is the driver, targeting a move toward 120 (Street Signals, Sep 10).
Position for a 10-20% dollar decline; own foreign shares. Veteran FX trader Robert Savage (an operator, not a pundit): a 10-20% dollar correction over "the next year or two" is "a very rational view given how overvalued it's been," and "generally, when the dollar goes down, foreign shares go up" (Wealth Building With Options, Sep 8).
Own core fixed income, international equities, and alternatives. David Kelly's house view is that this hike is near the top, yields cool, and a weaker dollar rewards "core fixed income for yield, international equities for total return, and alternatives" (Notes on the Week Ahead, Sep 14).
Own gold and Bitcoin as debasement hedges, from openly interested sources. Strive's Matt Cole sees the DXY down ~30% over three to five years and Bitcoin as the beneficiary (The Wolf of All Streets, Sep 10); gold traders on Palisades Gold Radio (Sep 10) say to keep stacking. Both are talking their book.

Read-throughs

The dollar's silence is the signal. When a currency won't rise on a hawkish Fed, strong jobs, hot inflation and $100 oil, the message isn't in the news; it's in the positioning. The most crowded dollar-long trade in five years has almost no marginal buyer left, which is exactly why good news lands with a thud. Watch the dollar's reaction to Wednesday's hike: a shrug or a fade would confirm the crowd thesis.
Watch the sequence, not just the outcomes. Fed Wednesday, BOJ Friday. A Fed hike first gives Governor Ueda cover to sound more hawkish and defend the yen, so the FOMC may move dollar-yen as much as the BOJ's own decision does. The yen is downstream of Washington this week.
"Strong dollar" and "leaking dollar" are both true at once. The dollar is still the most resilient currency in the room because it's the reserve currency, and it's still slowly losing the automatic trust that made it one. Nordvig's "leak," Norway's rotation out of Treasuries (but not out of dollars), and Mike Green's point that CDS and break-evens aren't flashing red all describe the same in-between state: erosion, not exodus.
The plumbing is quietly rewiring who buys US debt. Stablecoins add roughly 60 cents of net-new T-bill demand per dollar of growth (Liang): real, but no panacea, and paid for partly by community-bank deposits. It's a structural tailwind for short-term Treasuries that does nothing for the long end, which is precisely where the pain is.
The fiscal and the political are now one trade. A $1.3 trillion pre-election giveaway, oil above $100, a Fed hiking into a slowdown, and a bond market that's rediscovered the deficit all point at the same date on the calendar: the November midterms. Every week the Strait of Hormuz stays contested, Wednesday's decision, and the dollar's footing, gets harder.

What changed this week

The dollar's non-rally got a name: crowding. The explanation shifted from "mysterious" to "positioning is a five-year extreme," with hard CoT evidence that the yen side has already begun to flip.
The double-header came fully into view. Fed Wednesday (92% priced, "painted into a corner") and BOJ Friday (25bp "fully priced"), with the yen's fate hinging on the Fed going first.
A credible, quantified stablecoin case replaced the hype. Nellie Liang's ~0.6 net-new-T-bill-demand estimate, and its cost to community banks, reframed "secret QE" as a real but modest, short-end-only tailwind.
A serious contrarian defense of the buyback appeared. Mike Green's "it's just debt management, and the fear isn't in CDS or break-evens" gave the bull case for the dollar a spine it lacked last week.